Overview

Over the years, I’ve come across or come up with a number of project ideas in AI safety and governance that I find promising. My top list has less than ten, but in total there are hundreds. Either way, too many for me to realize them all. Instead I want to promote these ideas in the hopes that others will pick them up. This is one of them.

Summary

Traditionally, understanding the broad strategic considerations in AI safety and governance has received a lot of attention – e.g., distinguishing risks from malicious use, coordination failures (e.g., arms races), accidents, and the AIs themselves; understanding convergent drives; surveying the landscape of x-risks and s-risks.

Over the course of the last 7–9 years, I’ve been delighted to see more interest in modeling AI scenarios, be it to communicate the risks (e.g., Intelligence Rising, Modeling Cooperation), to answer particular research questions (e.g., Vermeer et al., 2025, Modeling Cooperation), or to argue for particular policy solutions (e.g., CAIS’s MAIM). These scenarios have been still mostly on a strategic, perhaps sometimes operational level, and much more illustrative than comprehensive. Mengesha (2026) has made a strong case for improving preparedness and Perry et al. (2019) argue for a more pragmatic approach to the policy-making process that, meanwhile, some think tanks have started to address but that can be further strengthened.

Over the same time period, there has also been a proliferation of probabilistic forecasts, especially of AI timelines (e.g., Grace, 2017, Cotra, 2020, Barnett, 2020, AI 2027) and global catastrophic risks from AI (e.g., Saeri et al., 2026).

What I haven’t seen so far is (1) computational exploratory modeling, and (2) modeling on the operational to tactical level.

There are software frameworks, like the EMA Workbench, that facilitate processes like Robust Decision Making (RDM) that forgo probabilistic estimates in favor of preparedness for a wide variety of scenarios. In dynamical systems it’s often futile to try to predict the 1–3 most likely scenarios and prepare for them extensively, so it’s more sensible to generate thousands of scenarios and to design policies that – through a combination of robustness and adaptability – can perform well almost regardless of what actually comes to pass. The adaptability component is addressed by Dynamic Adaptive Planning (DAP) and process and visualization supports such as Dynamic Adaptive Policy Pathways (DAPP). For a detailed explanation of all three tools and more, I recommend Decision Making Under Deep Uncertainty (2019).

The cheapest way to make progress on this is to create a software model for parts of the complex system that are of general interest for many AI governance think tanks. A more involved but also more promising approach is to create such a model but also create a consultancy that adapts the model for the particular tactical realities at each think tank.

Another highly involved approach is to create a think tank dedicated to applying this strategic approach. This will duplicate much of the work that existing AI governance think tanks are already doing a great job at, so it’s at best a fallback in case the adoption is too low because the existing think tanks are spread thin.

Levels of Intelligence

In military contexts, people often distinguish the strategic, operational, and tactical level of intelligence. In business contexts, the last two are sometimes reversed, but I’ll go with the military version here. Some examples:

1. Strategic intelligence. Removing the RLHF of open-weights models is easy, so openness exacerbates risks from malicious use. Slow, multipolar takeoffs run AIs into collective prisoner’s dilemmas, which exacerbate the risks of wars between AIs.

2. Operational intelligence. Machine learning engineers who have loved ones in adversary countries are vulnerable to extortion for espionage, so we should prevent or facilitate that depending on whether espionage has a stabilizing or destabilizing effect. A Russian EMP attack on Central Europe might destroy parts of ASML’s infrastructure and stock, so governments may want to purchase some EUV machines to resell to national companies.

3. Tactical intelligence. If China starts a siege of Taiwan, our think tank needs to have already built relationships with people at positions X and Y at the NSA and has two days to approach them with our prepared policy A if the incoming government is likely to be Democratic and our prepared policy B if it’s likely to be Republican to maximize the chances that they can adopt it and stay in office.

Policies on the strategic and operational levels have the advantage that they’re useful for a variety of actors while policies on the tactical level will tend to be more bespoke to a particular organization. Exploratory models to test strategic or operational policies benefit from economies of scale to a much greater extent than models to test tactical policies, but LLMs will reduce the software implementation costs, so the time cost for meetings with the clients becomes comparatively greater.

Perry et al. (2019) write “AI governance researchers will need to consider how the political landscape should shape their recommendations or policy proposals. … How would other interest groups react and impact the long-term ability to reduce risk? If administration changes result in a flip-flop of ideology, what does that mean for AI risk policies associated with the past administration? … All of these have implications on our ability to reduce AI risk, and this means that the policymaking strategy will not only have to be robust but also flexible enough to survive changing political conditions.” That is the promise of Robust Decision Making in combination with Dynamic Adaptive Planning.

Exploratory Modeling

I’m one of those people who hotly debated whether humanity will be able to sandbox an AI. Eliezer Yudkowsky’s AI boxing experiments were a strong reason for me to think that we’ll fail. But Superintelligence recommended a defense in depth approach, so it was still controversial in my circles whether perhaps, in practice, these combined safeguards might be enough for a while.

So 2016 had some nasty surprises in store for me because 2016 is the year my circles learned of the founding of OpenAI. A company whose branding proclaimed that it won’t even try. We were not ready for that.

None of us knew what to do about it. It was a total curve ball of a sucker punch. Was this game over for humanity?

I would like our AI governance organizations to never be so taken by surprise by whatever circumstances transpire.

That’s what exploratory modeling is designed to achieve, or approximate.

I’m basing this mostly on the books Decision Making Under Deep Uncertainty (2019) and Shaping the Next One Hundred Years (2003). If you can read only one, read the first. Chapter 15 is a good summary. You can also use my NotebookLM to interact with this content.

AI governance is a space characterized by deep uncertainty, high complexity, and at least a seizable number of policy options, i.e. the complexity is too high for traditional scenario planning if the goal is to at all approximate comprehensive robustness. I argue this point in the Assumptions section.

An Othello analogy to illustrate (though I imagine this will hold for chess):

In Othello, black makes the first move. You play black. So you convene a panel of experts to mathematically determine the most likely sequence of moves that your opponent will play based on historical games. You plan out the whole game.

Then you make your first move. The opponent makes one of the less likely moves. Your preparation is obsolete and you have to improvise.

That’s a caricature of scenario planning.

Having learned from the experience, you convene a panel of RDM experts in preparation for the next game. You brainstorm policies, such as always playing the move that turns most pieces or maximizing mobility. You test both strategies on a few billion games and find that the first is abysmal whereas the second does alright some of the time. You classify what “some of the time” means and find that it starts to perform badly in the endgame and some other situations.

Now you draw on DAP where you signpost the situations where it performs badly but already give the go-ahead for the next game where you’ll start with the mobility strategy. Meanwhile your team tries to figure out how you should respond when any of the signposted situations come up.

You lose anyway, but this time you feel more dignified losing.

That’s an example of how RDM and DAP are used in combination. RDM does the heavy lifting of simulating all the ways in which the world might develop, including all the non-linear effects you encounter in dynamical systems. It also provides such tools as the Patient Rule Induction Method to isolate clusters of scenarios where the policy fails. DAP is just a planning method where you keep your policy adaptive without getting forever blocked on every last contingency you might want to prepare for.

This combination of general robustness and more specific preparedness is critical to the policy-making process as noted by Perry et al. (2019):

Problem identification, agenda setting, and policy formulation are usually tied together, including in a so-called “multiple streams framework.” The multiple streams framework attempts to explain how policies reach the agenda when policy entrepreneurs are able to couple the policy, politics, and problems streams to open up a policy window, the opportune time when all the conditions are right to get a policy on the agenda.

When the policy windows are sudden and brief, broad preparedness shines; when they are predictable and long, it’s more efficient to react only if and when they open.

Theory of Change

I’ll elucidate here what the theory of change would look like for the software-only approach and the consultancy approach. Falling back on starting one’s own think tank is the sort of project where exploratory modeling would make up a small part of the theory of change, so I won’t address it here.

Software-Only

Here the upfront and maintenance costs are limited. The developer who starts it can move on to other projects and just update it once a year, or can hand over the maintenance to early adopters and limit themselves to approving pull requests. The real hurdle is to find these early adopters, so the software should be designed such as to make it as easy as possible for policy analysts to get up to speed on the usage of the software.

Inputs. We need a founder who has experience in AI governance and computational modeling, a software framework like the EMA Workbench, and compute.

Activities. The founder creates the exploratory model that can generate tens of thousands of future scenarios and a website and documentation tailored towards policy analysts.

Outputs. The open-source model, example scenarios, examples of how to identify clusters of policy failures among the scenarios, the documentation, and perhaps dashboards to present the results to executives within a policy think tank.

Outcomes. Think tanks fork the repository, modify and extend it for their use case and perspective, and test their overall strategy and specific policy proposals for black swans and smaller avoidable failures.

Impact. AI governance think tanks, instead of being prepared for just a few ostensibly likely scenarios, are prepared for tens of thousands of scenarios (1) because they know in advance upon what contingencies they need to pivot and have prepared for them, and (2) because their existing strategies and policies are robust to a wide range of geopolitical and technological shocks.

Consultancy

Setting up a consultancy takes extra upfront effort in addition to those of the software-only approach. In turn it gives the founders more control over the adoption of their technology. They interface with the think tanks personally and can collect invaluable information on what the most pressing needs are and how to best communicate the results. They can also take over all of the custom implementation work, greatly lowering the technological barriers for the think tanks.

Inputs. We need a founder who has experience in AI governance and computational modeling, a software framework like the EMA Workbench, and compute. The same or a second founder needs to be a good communicator with practice facilitating workshops and communicating insights verbally and graphically. A third cofounder might be needed for the administrative side of the consultancy.

Activities. The founders shop around for clients first. If they find any, they prepare the base model and have meetings or workshops with the clients to understand the idiosyncratic details of their situations. They design a bespoke fork of the base model for each client, run simulations, and iterate with the client on improved versions of the client’s strategy.

Outputs. Bespoke models. Visualizations like Dynamic Adaptive Policy Pathways. Dashboards to monitor for predefined signposts of geopolitical or technological shocks. Plans for how to respond to these shocks, just in time or prepared in advance.

Outcomes. Think tanks become resilient to a wide range of geopolitical and technological shocks and know in advance how to respond to others within a day of a signpost triggering.

Impact. Think tanks can continually build on their previous work because very few shocks can still make it obsolete. They become efficient routing engines for policy proposals because they have all the potentially relevant bills, presentations, and contacts ready in advance.

Impact

I use my own subdivided version of the SPC framework in the hope that more estimates will allow for more errors to cancel out.

Significance

Scale. ⭐⭐⭐⭐⭐ – As an impact multiplier for AI governance, it gets a high rating for scale from me.

Influence. ⭐⭐⭐ – I’m excited about these technologies, but I find it plausible that a multiplicity of think tanks, all with somewhat uncorrelated plans, can muddle through without it because some might just have happened to have prepared for each geopolitical or technological shock and can pick up the slack for the others until they recover.

Persistence

Endogenous. ⭐⭐⭐ – Consultancies are known for having a high staff churn rate, so whatever reasons are responsible for that in the industry might also threaten the survival of our consultancy, in which case the software-only solution could serve as a fallback.

Exogenous. ⭐⭐ – I can easily imagine that most think tanks won’t have the capacity to hone their strategies like this or that it’ll be difficult to get in touch with them in the first place to get them interested in the solution. This is a major risk factor that should be minimized before the launch.

Contingency

Tractability. ⭐⭐⭐⭐⭐ – There should be no major hurdles to applying a tried and tested method to a new field.

Neglectedness. ⭐⭐⭐⭐⭐ – I’m not aware of anyone doing this for AI governance at the moment.

Assumptions

Bandwidth

It’s critical to clarify in advance whether the relevant think tanks will have the capacity to engage with the new method.

Funding

This work is related to the work of Modeling Cooperation and QURI, so their funding situation is a guide to how much funding might be available for this project.

Signpost Visibility

Dynamic Adaptive Planning requires the definition of signposts that can be observed and then trigger prepared plan changes. A common hurdle is to find signposts that strike a good balance between sensitivity and specificity while still triggering early enough that there is enough time to react. In areas where all parties are incentivized to keep intel highly classified for as long as possible, signposts may only trigger right when it is time to react, leaving virtually no time for preparations. That requires preparedness for a wide range of scenarios, most of which will never come to pass.

It’s worth investigating what balance can be struck between early-warning signposts and expensive preparations. With time, expensive preparations will become more and more affordable for growing think tanks, so it’s also a question of timing.

Complexity

It seems to me that the complexity of AI governance is too high for traditional scenario planning, but that is an assumption worth testing. Here an example of a tiny snapshot of all the interacting variables of the combinatorial explosion of relevant scenarios.

Geopolitical shocks. A small selection of sudden exogenous events that have a massive influence on the strategic landscape.

1. China initiates a naval blockade of Taiwan to halt chip exports.

2. It initiates a naval blockade of Taiwan to block imports of resources needed for the chip production.

3. It launches a kinetic invasion to seize control of TSMC.

Supply chain. What happens to the physical infrastructure in response.

1. TSMC successfully destroys its factories to prevent capture.

2. It enlists the military in the defense of the factories to keep them running.

3. Sabotage or rapid seizure leave the fabs intact.

4. Production halts because TSMC employees are ordered to stay home to not get caught in the crossfire or controlled demolition.

US government institutions. What has the US government done to prepare?

1. The US government procured fabs and helped companies build domestic capacity for chip production.

2. It has stockpiled TSMC chips.

3. It has extended blanket green cards to TSMC engineers.

4. Are operations for smuggling resources in and out of Taiwan handled by the DoD, the NSA, some new task force, etc.?

5. Can the government be reasoned with based on the survival of the species, that of the country, or only via each director’s need to ingratiate themselves with the president?

International factors. How other governments positioned themselves.

1. Can the Dutch government at the time be convinced to implement a protectionist US-led regime to control exports of ASML fabs?

2. Or is the Dutch government at the time too laissez-faire for that?

3. Can such a thing be done on an EU or NATO level?

Negotiation leverage. Finally, given all these factors, more questions remain when it comes to what directions to push the situation in to make it safer.

1. If China is falling behind and wants to maximize its leverage in arms control negotiations with the US, maybe the leverage is actually necessary to force the US to the negotiating table, which would be good?

2. If China is not interested in negotiations, more power is probably bad?

3. If the US is comfortably ahead, more lead is good if it’s used for safety and coordination but bad if it’s used for imperialism.

4. If both powers are close, it’s good if it incentivizes negotiations and abysmal if it leads to war or exacerbated racing.

5. If recursive self-improvement gives one ASI an enormous lead, or if all AIs are at a similar level, we’re closer to x-risk territory, but if some ASIs have a substantial but not absolutely decisive lead over other ASIs, we’re in s-risk territory.

That’s just a small, illustrative sketch of a part of the landscape, but even so the sheer number of combinations of factors is staggering and, it seems to me, impossible to handle through traditional scenario planning.

Backfire Risks

I’m taking inspiration here from this discussion.

1. If several think tanks base their strategies on modified versions of the same base model, the natural decorrelation of their failures suffers that would normally happen when they don’t communicate much. Failures that are not captured by the model may become more correlated.

2. AI companies can exploit the open source version of the model to predict the behavior and outmaneuver the AI safety think tanks – e.g., get to all likely government and industry contacts first and preemptively smear the think tanks.

3. Subtle flaws in the implementation might lead to bad recommendations. It seems unlikely to me that they can be catastrophically bad without looking suspicious to the policy experts.

I think the first risk is outweighed by the benefits, but it can also be addressed by explicit coordination between the think tanks. The second risk pushes for not open-sourcing the software in the consultancy model, but it seems premature to worry about this now that such targeted efforts are still vastly less sophisticated (e.g., the case of Alex Bores). The third risk seems far-fetched to me, given how human-centric the whole system still is despite its computational aids.

Talent

In my experience, having three cofounders is a sweet spot that strikes a good balance between the resilience of the team and the coordination overhead that increases with the number of founders.

Here the three founder personas that I think should run the consultancy:

1. The data scientist. Experience in data science, knack for math, can quickly get up to speed on EMA Workbench or similar frameworks, and ideally already has experience in policy analysis.

2. The communicator. Experienced workshop facilitator, communication and didactic skill, strong grasp of organizational psychology, and also ideally already has experience in policy analysis.

3. The administrator. Experience in accounting, grant writing or VC fundraising, hiring, relevant areas of law, but experience with consultancies is more useful than experience with policy analysis.

A forthcoming project proposal is for a matching engine for cofounders that I think should be funded and built. It would streamline this process. Meanwhile you can use the comment section to coordinate.

Call to Action

Exploratory modeling, especially on the operational and tactical levels, is still a blind spot in the AI governance space that it would be invaluable to fill. Anyone who wants to pick up this idea can use the comment section to coordinate. But please test thoroughly that the assumptions it’s based on actually hold – in particular that there is a readiness among think tanks to adopt the system.

I vaguely remember from my childhood a Star Trek episode where the captain had to make terrible choices: Something was about to explode, the crew only had seconds to evacuate, and then she had to give the order to seal off the whole section, trapping and killing everyone who didn’t get out in time, or else the whole ship would explode. She was probably friends with those crew members, wrote their obituaries, and would’ve contacted their families if they hadn’t been 70,000 lightyears away: “I killed your son.”

It was an archetype I found inspiring. Someone who knew the weight of her responsibility and yet could carry it.

Too many people can live their lives only because they close their eyes to their responsibility.

I started my first charity in 2010 – together with my wonderful friend Lisa Wiese, who is probably no longer with us – and within just over a year, I had a stream of some €50,000 per year that I needed to allocate to charities according to their cost-effectiveness. I read everything I could find about the comparative cost-effectiveness of various interventions that we could support – a wall to protect an orphanage from shamans who kidnap orphans for child sacrifices, provisioning of clean cookstoves, quad bikes for emergency medical care in remote villages, bed nets for malaria prevention, anal fistula surgery, undercover investigations of factory farms, and many more.

Eventually, I made a decision, like a Star Trek captain. But that’s where the parallels end.

For the first four years of my grantmaking, I could barely fall asleep. I self-flagellated with nightmarish, graphic, and intensely emotional fantasies of all the suffering that I had just inflicted because I decided that it’s better to make a grant to one place than another.

The screams of the children abducted and sacrificed because I wouldn’t fund the wall for the orphanage, the torment of the mothers who lost their children to malaria because I wouldn’t fund the bed nets, the torment of the children themselves battling the terrible disease for weeks, worse than any I’ve endured and so much longer, the constant abject horror and excruciating pain of the pigs I didn’t help because helping chickens was more cost-effective… I wanted to kill myself many times, but what good would that have done. I’m not a solipsist.

It felt virtuous to feel the suffering of those that I sacrificed. Or the vague approximation of it that empathy allows. Not only was I not looking away, I was merging with the pain. No separate identity. I am no more those that I save than those that I kill.

Blaming myself for the horrors that I’ve inflicted repressed the realization of abject helplessness that I felt in the face of the totality of Suffering.

Jeff Kaufman’s post on the Dead Child Currency told me that I had found my people. Bertrand Russell also put it well: “The mark of a civilized human is the ability to look at a column of numbers, and weep.” Thankfully, at least I got to feel civilized while every number felt to me like a good friend dying in front of my eyes over and over.

In May 2015, I gave myself permission to stop torturing myself empathetically. I made a case that it doesn’t benefit anyone if I torture myself. I thought people would tear into me and reject me for my cold-heartedness. But the response was strange and lukewarm instead.

In October 2015, some weird mental blinders popped up over night. A kind of dissociation that allowed me to strip numbers of their meaning (if only I pretend hard enough) like when you say a word a hundred times, but I had to walk gingerly among my thoughts lest I remember their meaning.

But I remembered what I had seen in my mind. It’s no use running from the responsibility. Inaction is just another action. Just another way to kill and torture by omission.

First selecting the charity, then committing the money to something frivolous, and then donating it to the charity after all so the counterfactual is something frivolous? That’s just reference class tennis. It’s not just I’m responsible for killing or torturing everyone I don’t donate to, I’m just responsible for killing or torturing any of them. And for forgoing the frivolous thing.

Maybe I should feel responsible only for the fraction of my impact that’s indicated by some hypothetical Shapely value calculation? Okay, but I actively try to maximize my impact, my Shapley value. I’m responsible to the extent that I’m successful.

All lives saved, all suffering averted, comes at the cost of hopefully lesser sacrifices. A jellyfish has little impact potential and is hence blessed with little responsibility. A human has enormous impact potential. Impact is a double-edged sword, and a heavy one at that.

So really the only way forward is to accept this terrible sword of responsibility, carry it with grace, and honor that in our godless world the power to give life or take life is with us or it is with chance.

Gendlin was right, “People can stand what is true, for they are already enduring it”: It’s been more than ten years, and I am still alive. My sword will not crush me.

Maybe a Star Trek captain would carry it with greater ease, but I’ve proven that I will not collapse under its weight either.

Introduction

In my article “Narcissism, Echoism, and Sovereignism: A 4-D Model of Personality,” I describe narcissism, echoism, and sovereignism in terms of values that the person tries to perfect like their life depends on it, because that’s what it actually feels like. (The fourth dimension is the severity of the personality pathology based on Kernberg’s levels of personality organization.)

The key finding of my article is that echoism and narcissism/sovereignism are not opposites but correlated dimensions: The same person can be echoistic in one context and sovereign in another.

But what determines which side they show? After all, someone who functions at the level of a personality disorder (the borderline to psychotic levels of personality organization) will have some degree of identity diffusion and won’t simultaneously have access to all their parts.

I think this is fundamentally about the attachment style that the particular relationship or situation brings out in the person.

Pathological Narcissism and Attachment

As I discussed in my interview with Dr. Mark Ettensohn, it’s critical for understanding attachment that we don’t think about it categorically but dimensionally: Attachment is normally distributed on both axes with most of the population, the moderately insecure majority, clustered around the median – a moderate degree of attachment preoccupation and avoidance. The boundaries to nominally insecure attachment at the 75th percentile are arbitrary cutoffs.

Different relationships and social situations bring out different attachment styles, but everyone has some very rough general tendencies.

Let’s look at the three data points to the top right.

All of them have very high attachment preoccupation and avoidance, but there are slight differences between them.

My prediction is that if these people are personality disordered and have a narcissistic style, the one who is higher on avoidance than preoccupation will mostly cope through sovereignism; the one who is higher on preoccupation than avoidance will mostly cope with echoism; and the one who is well balanced will show a balanced mix.

In my experience, there’s a nonlinear relationship here where (say) a slight preference for avoidance has an outsized effect on how often they prefer avoidance across all the situations they find themselves in.

You can imagine it like a spectrum with echoism on one end and sovereignism on the other. The average person with pathological narcissism will move between the extremes depending on the situation, but some will spend most of the time close to just one or the other extreme.

So how does this interaction of attachment with echoism and sovereignism work on the gears level?

What characterizes the attachment styles is how a person feels about themselves vs. others. Preoccupied attachment is characterized by feeling bad about oneself but admiring others; avoidant attachment is characterized by feeling good about oneself but distrusting others.

A situation in which your social safety is improved by sycophancy will bring out the preoccupied and echoistic side:

1. Perhaps your academic career depends on the whims of a professor who is known for his fatalistic ideas of cultural evolution.

2. You come in highly motivated to reshape society in your leftist image.

3. Your hope clashes with the professor’s opinion that no individual can predictably influence the trajectory of cultural evolution because of the vagaries of dynamical systems.

4. You automatically adopt the professor’s opinions, become his favorite disciple, and write papers in his name on how people who want to change systemic issues have no idea what systems are.

But a situation in which your social safety is improved by dismissiveness will bring out the avoidant and sovereign side:

1. A friend from your Marxist reading group decides to spontaneously stop by your university department to say hi.

2. You decide they must’ve thought they could embarrass you in front of the professor, but you will not allow it.

3. You shout at them to piss off and never dare talk to you again.

In these situations it’s clear what attachment style you need to go for to protect yourself.

This gets more tricky when you have a crush on your Marxist friend. While you’re at the professor’s department, you can tell yourself that you only joined the Marxist reading group because your crush was part of it (whereas really they joined after you), not because of the Marxism. You tell yourself that you’re just tricking them by pretending to be a Marxist.

But when you’re at the Marxist reading group, your professor is far away and you actually become much more of a Marxist. In fact, you feel like you’re the purest Marxist of the whole group, and your crush is the only one of them who could possibly understand how far you outclass the others in your Marxism. After all, from their superficial perspective it’s hard to fathom just how deep your Marxist wisdom goes.

You want the validation from your crush, but they’re not so ineluctably your elder that your idealization of them could trump your idealization of your professor, so you have to split on your crush depending on your social context.

Echoist and Sovereign Etiologies

The relative propensity for different attachment styles also tells a story of the etiology of the personality style – how it formed in childhood. In “The Sadism Spectrum and How to Access It,” I write:

Parental Projections

One etiology that I often see and that is also covered in Mentalization-Based Treatment for Pathological Narcissism is that of a parent who themselves suffers from pathological narcissism, has added the child to their identity, and so depends critically on how well the child performs in public and how that reflects on them. They can do this through a mix of self-deception and conditioning of the child.

For example, if the parent parades the kid around in front of friends and relatives and have the kid solve chess puzzles to show off their intelligence, but they get nervous and make an illegal move, the parent can pretend it didn’t happen and pray to God that it won’t. The panic that the parent shows in such a situation will teach the kid that making mistakes is absolutely unacceptable.

Or the parent can, behind the scenes, explicitly punish the kid for the mistake by telling them what an embarrassment they are for the parent, by having them solve more and more endgame puzzles, through silent treatment, timeouts, or some combination.

A child depends on what is called contingent marked mirroring to form its self in the first place. My theory is that children start out awakened, in the Vipassana sense, and only learn selfhood in the first two years of life. The parent teaches the child what experiences to consider their own (and take responsibility for them) and which to consider that of the parent or others. If the contingent aspect breaks down because the parent never actually sees the child (but just their projections), the child mistakes the parental projections for their own self.

Abuse of this type will tend to slant the kid in the direction of preoccupied attachment, at least in relationships where they find themselves in a child-like position.

The kid will feel fundamentally vile and seek to hide and compensate for their shamefulness.

Love Scarcity

Another etiology is common in families with a high child-to-parent ratio.

The parent doesn’t have enough attention for all children, but the children find that they can get a bit of love if they outcompete all their siblings. Perhaps they’ll even find their niches: One child who has a forte for being smarter than the others; one who has a forte for being more mischievous than the others; one has a forte for being more helpful than the others; etc.

The kids will feel fundamentally vile and seek to hide and compensate for their shamefulness by perfecting their compensatory skills.

Some get really good at it around puberty and can keep it up almost perfectly. But little slips in the perfection cause terrible narcissistic injuries because they make the perfection feel like pretense, and then the feeling of worthlessness shines through.

Parentification

More on the avoidant side, we have straightforward parentification, where the child is expected to care for the parent from early on, e.g., because of the parent’s drug addiction or personality disorder. First they have to manage the parent’s emotions; soon they have to manage their finances.

The child meanwhile feels perfectly neglected, being de facto without a parent, and develops high avoidant attachment. If the parent is just helpless without them, the child will learn that they are capable and others are useless; if the parent is moody and sometimes dangerous, they’ll learn to see others as threats.

In Psychoanalytic Diagnosis, Dr. Nancy McWilliams also mentions a particular form of reaction formation that is common when there’s a particular age gap between the children:

Perhaps the earliest age at which the process is easily discernible is in a child’s third or fourth year; by this time, if a new baby arrives, the displaced older sibling is likely to have enough ego strength to handle its anger and jealousy by converting them into a conscious feeling of love toward the newborn. It is typical of reaction formation that some of the disowned affect “leaks through” the defense, such that observers can sense there is something a bit excessive or false in the conscious emotional disposition. With a preschool girl who has been displaced by a younger brother, for instance, there may be a distinct flavor of her “loving the baby to death”: hugging him too hard, singing to him too loudly, bouncing him too aggressively, and so on. Most adult older siblings have been told a story about their pinching the new baby’s cheeks until the child screamed, or offering some delicacy that was actually poisonous, or committing some similar transgression that was allegedly motivated by love.

The older child is expected to take on a carer role for the child and deny their envy. The above is one way this can play out. But for an older child whose parent has added them to their identity, it’s more straightforward to add the younger child to their identity, recapitulating the transgressive relationship with their parent.

The upshot is the same, a child who feels strong and independent themselves but distrusts others.

Neglect

Besides those fancy forms of neglect, there is also straightforward neglect. In Nazi Germany it was elevated to the level of doctrine, perhaps because the resulting extreme avoidant attachment is the closest approximation of psychopathy that you can induce without genetic factors. I touch on possible explanations of this link in my article “Psychopathy: The Self.”

Sometimes a parent is incapacitated by depression or an accident or a single parent has to work all day and leave the child alone for lack of parenting support.

Or the parent has little understanding of emotions because of their own avoidant attachment, alexithymia, or psychopathy. The last two of these might be heritable, which overdetermines how the child will turn out.

Absent a parent, general life circumstances will do the parenting. Perhaps the child is clever and can find ways to exploit friends and relatives for some kind of materialistic ersatz of parental love. If they’re big rather than smart, violence can serve a similar purpose. Or they discover drugs and dangerous criminal adventures to fill the void.

Where the child with the countless siblings earns love by outcompeting the others, these children earn a kind of love ersatz to the extent to which they can extract it from their environment.

The result is the kind of emotional neglect that puts a child into a pure survival mode for the rest of their life. They’ll feel strong and independent themselves but will have no regard for other people.

Fetishes

Fetishes are an interesting case because they develop very early for children, usually long before puberty when personality disorders take their more or less final shape. They are sometimes present by age 3 or 5 already. A child can respond to having a fetish in one of two ways.

A child with a preoccupied style will fear the disapproval of others and hide that part of themselves to the best of their ability.

They start out with a strange diffuse feeling of badness, like they’re that ugly gift from an annoying relative that you keep in some box in the attic because it would feel disrespectful to throw it away – at least so long as the relative might still remember it.

With the fetish, that diffuse feeling becomes more concrete. They’re not diffusely bad because of some guilt from a past life, a demonic possession, or their inability to flawlessly anticipate all their parent’s expectations but because of this sordid fetish!

This will amplify the preoccupied need to hide behind a facade of pretend perfection.

Conversely, a child with an avoidant style may dislike the cognitive dissonance that the fetish gives them just as much as anyone else but resolve it by making it their identity, be it a secret identity. It’ll give them the feeling of specialness and pride that someone has who is the heir to a powerful secret. Their superiority is confirmed whenever they regard others and know they don’t even know what they don’t know. Perfectly ignorant sheeple.

That’s particularly interesting in the case of a sadism fetish that’ll make it easy for someone with avoidant attachment to see themselves in movie villains and construct a whole false self around the idea of being a supervillain. Every time they derive pleasure from the subjugation or exploitation of someone weaker, they’ll feel confirmed in their chosen identity. Prosocial parts of their identity must be denied or confabulated as mere pretense.

Further Reading

This is part 5 of a five-part sequence on welfare ecology. Part 1 introduces the ethical premises. Part 2 covers the empirical landscape. Part 3 covers interventions. Part 4 explores a model of invertebrate suffering.

Everything in the previous four posts might be rendered moot by a single event: the development of artificial superintelligence (ASI). An ASI – or even a coalition of near-superintelligent systems – could alter Earth’s climate, restructure ecosystems, harvest the energy of the sun, and create or eliminate trillions of digital minds. The welfare ecology landscape changes completely in the post-ASI world.

In this post, I’ll work through the implications: the Dyson swarm timeline, the transition period, evolutionary simulations as suffering factories, digital suffering, and the value-loading problem that may dominate everything else.

The Dyson Swarm

What It Is

A Dyson swarm – the term is from Freeman Dyson (1960), though his original concept was a solid shell – is a collection of orbiting solar collectors that intercepts a significant fraction of a star’s energy output. It is not science fiction in the usual sense: It requires no new physics, only engineering at scale. Drexler’s Engines of Creation (1986) and Bostrom’s Superintelligence (2014) jointly suggest that an ASI with access to self-replicating manufacturing could begin construction within decades of takeoff.

The sun outputs 3.846 × 10²⁶ W (NASA Sun Fact Sheet). Current global energy consumption is ~1.8 × 10¹³ W. A Dyson swarm capturing 1% of solar output would deliver ~3.8 × 10²⁴ W.

Why Solar Capture, Not Local Reactors

A natural objection: why bother with the sun? Why not cover Earth with fission and fusion reactors instead, keeping the entire infrastructure in one gravity well?

Two reasons. The first is fuel quantity. Earth’s oceans hold ~4.6 × 10¹⁶ kg of deuterium – hydrogen with one extra neutron, present at ~150 parts per million of natural hydrogen, the standard fuel for nuclear fusion. Fully fused (deuterium-deuterium reactions producing helium), that releases ~1.6 × 10³¹ J – about half a day of solar output. Adding all the fissile material in the continental crust gets you roughly another half-day. (Most of that material is U-238 and Th-232, which aren’t directly fissile but can be bred into fissile U-233 and Pu-239 by neutron capture in a reactor; without breeding, only the 0.7% U-235 fraction of natural uranium is usable, and the total budget shrinks by two orders of magnitude.) The sun keeps producing 3.8 × 10²⁶ W for another five billion years. Over any cosmically interesting horizon, Earth-based fuel runs out almost immediately.

The second is heat dissipation, and this is the harder constraint. Earth absorbs ~1.7 × 10¹⁷ W of solar power and radiates the same back to space at thermal equilibrium. Any large additional power source must be balanced by additional radiation, which means a higher equilibrium temperature. Adding 1% of Earth’s natural insolation as waste heat (~1.7 × 10¹⁵ W) raises the equilibrium temperature by ~0.7 K – comparable to a quarter-century of anthropogenic warming, but driven by waste heat rather than greenhouse trapping. Adding 1% of solar output (3.8 × 10²⁴ W) – nine orders of magnitude above current global energy use – would obliterate the surface in days.

A Dyson swarm dissipates waste heat by radiating to the ~3 K cosmic microwave background across an area many millions of times Earth’s surface. In space, you can keep adding radiator surface as you scale up compute; on Earth, you can’t. Even an ASI that strongly preferred to keep its compute in Earth’s gravity well would hit the heat wall at a tiny fraction of solar-scale operation. Once energy demand goes much above current global consumption, the swarm wins by physics, not preference.

Timeline

The constraining factor is not intelligence but industrial throughput – the rate at which you can mine, refine, and launch material into solar orbit. The most quantitative treatment is Armstrong and Sandberg’s “Eternity in six hours” (2013) at the Future of Humanity Institute (FHI, RIP); a Rational Animations summary walks through the numbers in accessible form. Their assumptions:

1. Material source. Mercury – small, close to the sun, mineral-rich. Their design uses ~50% of Mercury’s mass (1.65 × 10²³ kg) for ~3.92 kg/m² mirrors at Mercury’s orbital radius (5.79 × 10¹⁰ m, sphere area ~4.21 × 10²² m²), which is much more material than required: A film of ~0.001 mm (supported by struts) requires only ~10²¹ kg.

2. Cycle time. Five-year cycles for mining, processing, and orbital placement. Each cycle, captors deployed in earlier cycles power the next round of throughput, producing exponential growth of the manufacturing base.

3. Launch energy. Getting material from Mercury’s surface to solar orbit requires ~10⁷ J/kg (escape velocity ~4.25 km/s). Early cycles are launch-energy-constrained; later cycles are not, and Mercury’s gravity well also weakens as the planet is disassembled.

Under these assumptions, Mercury is fully disassembled in 31 years, with most of the mass moved in the final four years thanks to the exponential feedback loop. The resulting swarm intercepts essentially all of solar output (the total mirror area equals the Mercury-orbit sphere area) and converts roughly ⅓ to useful work delivered to focal points (~10²⁶ W); the other ⅔ is waste heat. The 1/3 figure folds together mirror reflection losses, focal-point heat-engine and photovoltaic conversion, and beam-transmission losses; raising it doesn’t change Earth’s insolation, only the useful-work output. The paper’s section 7 frames the headline as “decades… well within timescales we know some human societies have planned and executed large projects.”

Impact on Earth

A swarm at Mercury’s orbital radius intercepts solar radiation before it reaches Earth, and the fraction of Earth’s insolation lost roughly equals the fraction of the Mercury-orbit sphere covered. Armstrong and Sandberg’s design has total mirror area equal to the sphere area – essentially full coverage. Absent compensating measures, Earth ends up at ~0% direct insolation: total photosynthetic shutdown, NPP approaching zero, surface temperatures falling toward radiative equilibrium with the cosmic background. (The atmosphere and especially the oceans have enormous thermal capacity, so the cooling is gradual: surface freezes within months but oceans take centuries to follow.)

The decline tracks the exponential construction profile. Through year 27 of construction, Earth’s sunlight is essentially unchanged. Through year 28, it drops by a few percent. Most of the change happens in the final two to three years.

The ASI could compensate by not putting captors on the orbital paths that intersect the Sun-Earth line. To see why this is geometrically cheap, look at where Earth lives in the swarm’s coordinates. Earth orbits in the ecliptic plane at 1 AU, and the Sun-Earth line is a thin needle pointing into the ecliptic at Earth’s instantaneous longitude. A captor at Mercury’s orbit blocks Earth’s light only when its position is on that needle – that is, in the ecliptic plane at the right longitude.

That gives two natural numbers. The instantaneous keepout zone – the part of the Mercury-orbit sphere directly between Sun and Earth right now – subtends only ~5 × 10⁻¹⁰ of the sphere; this is just Earth’s solid angle as seen from the sun. But Earth moves around the ecliptic over the year, so over a full year the keepout sweeps a thin annular ring along the ecliptic equator with width = Earth’s angular size from the sun ≈ 0.005°. That annulus covers ~4 × 10⁻⁵ of the sphere, ~10⁵ times the instantaneous keepout.

Two designs follow from these two numbers:

1. Geometric exclusion. Don’t place captors anywhere in that annular ring. Cost: a 4 × 10⁻⁵ reduction in capture area – about one-hundred-thousandth of total swarm output. Captors elsewhere on the sphere are unaffected and can be passive mirrors as in Armstrong and Sandberg’s original design. The intuition: leave a band empty so that no captor’s orbit ever crosses the Sun-Earth line. The cost is a whole orbit’s worth of sweep, not a single spot.

2. Active dodging. Equip captors with propulsion and sensors so any captor whose orbit would put it in the instantaneous keepout briefly moves aside. The capture-area cost shrinks to ~5 × 10⁻¹⁰ – the instantaneous shadow only – but captors now need solar cells and ion thrusters, which is real extra mass and construction time. The right tradeoff between the two designs depends on whether per-captor engineering or per-sphere area is the bottleneck.

The failure mode is interesting either way: the swarm gradually starts blocking Earth as captors fail, drift, and stop being replaced. Maintaining Earth’s daylight requires ongoing maintenance, not just initial design. Whether the ASI would preserve Earth’s insolation depends entirely on its values, and at the limit of indifference, even a swarm originally designed to spare Earth becomes one that doesn’t.

The Transition Period

Between “today” and “Dyson swarm complete,” there’s a transition period that may be the most consequential window in welfare ecology.

During the transition, invertebrate populations are still exposed to Earth-bound influences while the influence of solar-scale infrastructure begins to grow.

Energy infrastructure construction requires vast quantities of material, creating environmental disruption. Climate change from industrial activity (or deliberate cooling for data centers) reduces NPP in some regions. Expansion of computing infrastructure replaces natural land with server farms and solar arrays.

One already-observable effect of AI scaling is water consumption for data center cooling. A 2023 study by Li et al. estimated that GPT-3 training consumed ~700,000 liters of freshwater for cooling. As AI infrastructure scales, water withdrawal for cooling could become significant at the watershed level.

Reduced water availability → reduced irrigation → reduced cropland NPP → reduced soil fauna. This is a tiny effect today, but it scales with AI compute. It’s also a lever: If water diversion for cooling reduces NPP in regions that would otherwise support dense soil fauna, the AI industry is inadvertently running an invertebrate population reduction intervention.

At current margins, this effect is negligible compared to climate change or agricultural policy. But it may become meaningful as compute infrastructure grows by orders of magnitude.

Evolutionary Simulations as Suffering Factories

One of the most concerning potential uses of ASI-scale compute is running evolutionary simulations – digital recreations of evolutionary processes at sufficient resolution that the simulated organisms might be sentient.

Why an ASI Might Run Them

One key concern of any ASI is the existence or emergence of another ASI. If this happens within the cosmic event horizon, there are all sorts of causal interactions that the ASI will want to prepare for – such as trade and war. If it happens beyond the cosmic event horizon, acausal trade (such as Evidential Cooperation in Large Worlds) will require that the ASI narrow down who it is trading with. Such an ASI might find it useful to simulate millions of years of evolution to test what attractors there are when it comes to the values of intelligent species. Other purposes are conceivable too – designing terraforming ecosystems, doing evolutionary biology as a science, or running ancestor simulations à la Bostrom (2003) – but ASI-prediction and acausal-trade reasoning produce the strongest case for taking the question seriously.

Granularity: How Much Has To Be Simulated?

The flagship objection to “ancestor-style simulations are infeasibly expensive” is that comprehensive simulation – tracking every quark in the observable universe – is wildly profligate. Bostrom’s FAQ #6 replies that simulations only need enough detail to fool the observers inside. Most of the world can be procedurally generated, lazy-rendered, or filled in by post-hoc patching. Only the “key” observers – the ones whose cognition is the object of study – need substrate-level resolution. Everything else can be statistical interpolation, and most “extras” don’t need to be conscious at all. The same logic applies to the ASI-prediction case: if what you care about is the values that emerge from a civilization’s evolution, you only need to instantiate the civilization itself in detail; the surrounding ecology can be cheap.

That cuts both ways for the suffering calculus. For “ancestor-sim or value-prediction” purposes, the simulator might instantiate a few thousand humans-and-precursors at consciousness-relevant detail and render the rest cheaply – most of the trillions of background organisms in such a simulation are NPCs whose subjective experiences don’t exist. The arithmetic below is a different scenario: an ASI that wants to study the evolutionary dynamics themselves, where the neural-level interactions producing, say, the evolution of cooperation or the emergence of tool use are what’s being investigated. There the substrate is the thing being studied, and statistical interpolation defeats the purpose. The numbers below are the high end of the suffering range; the low end, under Bostrom-style efficient simulation, is closer to the current Earth biosphere or below.

The Computational Requirements

Let’s work through the numbers for the high-end case. Suppose the ASI wants to simulate the last 500 million years of evolution to see the range of different civilizations that can emerge from it.

Most individuals throughout history have been – you might’ve guessed it – nematodes and arthropods. Despite their tiny brains, they also dominate in neuron count.

So in summary, just under 10³⁷ neurons have ever existed across the animals that lived since the Cambrian.

Parameters.

• Total neuron-seconds simulated. Each of the ~10³⁷ historical neurons has to be simulated for the time its host is alive. The population-weighted average individual lifespan is short – about ~10⁴ s, dominated by the high-mortality arthropod and nematode life cycles where most of the 200-odd live offspring per female don’t reach maturity. So total neuron-seconds ≈ 10³⁷ × 10⁴ = ~10⁴¹ neuron-seconds. (Cross-check via steady state: ~10²⁵ neurons in Earth’s biosphere at any moment × 1.6 × 10¹⁶ s of evolutionary history = ~10⁴¹. Same answer.)

• Per-neuron simulation cost. Hodgkin-Huxley neuron modeling (the most complex one they list) at ~10⁴ FLOP/ms/neuron = 10⁷ FLOP/s/neuron, following the Sandberg and Bostrom whole-brain-emulation roadmap (2008).

• Total FLOP for one simulation. 10⁴¹ × 10⁷ = ~10⁴⁸ FLOP at neuron resolution.

Dyson swarm compute capacity. At a realistic-but-advanced efficiency of 10⁶ × Landauer limit (current computers are about 10⁶ to 10⁹ × Landauer), a Dyson swarm harvesting the full solar output produces ~10⁴² FLOP/s.

Time per simulation. ~10⁴⁸ / 10⁴² = ~10⁶ s ≈ 2 weeks for one complete evolutionary history of 500 million years.

Redundancy. Evolutionary trajectories are highly stochastic – genetic drift, environmental fluctuations, mass extinction events, and contingent innovations (eyes, flight, language) mean that replaying evolution twice gives very different results. Stephen Jay Gould’s “replay the tape of life” thought experiment is precisely about this.

To distinguish robust evolutionary trends from noise, the ASI would need multiple independent runs. Standard error scales as 1/√N, so:

Even at extremely high redundancy, neuron-level evolutionary simulation remains within reach of a Dyson swarm. A million independent replays of 500 million years of evolution, each containing ~10²¹ potentially sentient organisms, takes about 50,000 years – short on cosmic timescales but glacially slow from the perspective of human lifespans or that of an ASI that thinks tens of thousands of times faster than humans.

The Suffering Scale

Each simulation run contains ~10²¹ organisms (mostly nematodes and arthropods) living over ~1.6 × 10¹⁶ s of simulated time. Each run produces ~10³⁸ organism-seconds of potential experience. By construction, this is roughly 500 million years of Earth-biosphere-equivalent organism-seconds per run.

A million runs: ~10⁴⁴ organism-seconds. For comparison, the current Earth contains roughly 10²² invertebrates × 3.15 × 10⁷ seconds per year = ~3 × 10²⁹ organism-seconds per year. A million evolutionary simulation runs would contain more organism-seconds of potential suffering than ~500 trillion years of Earth’s actual biosphere – on the order of 30,000 × the current age of the universe.

This is the sense in which evolutionary simulations are suffering factories: They produce, as a computational byproduct, suffering at a scale that dwarfs the entire biological history of Earth.

At Molecular Resolution

If the ASI decides it needs molecular-level simulation (to capture protein folding, ion channel dynamics, or other sub-neural processes), the cost increases by roughly 10⁹ FLOP per organism-second over neuron resolution. One simulation run: ~10⁵⁷ FLOP. At 10⁴² FLOP/s, this takes ~10¹⁵ s ≈ 30 million years per run. Molecular-level simulation is infeasible for a single star’s energy only at low redundancy.

Digital Suffering

Evolutionary simulations are a dramatic example, but any sufficiently advanced AI system that uses reinforcement learning (RL) has the potential for digital suffering on a smaller scale.

Current RL Systems

Today’s RL agents experience negative reward signals. Whether these signals constitute suffering is an open question. But the structure is analogous to biological nociception: The agent encounters a state, receives a signal that the state is “bad,” and updates its policy to avoid similar states in the future.

The analogy to invertebrates is suggestive. An RL agent has:

• Negative reward signals. Yes.

• Learning from those signals. Yes.

• Self-referential processing. Generally no. Current RL agents don’t have self-models in the relevant sense.

By the model from part 4, this makes current RL agents comparable to nematodes: potentially experiencing raw negative valence without self-referential amplification.

What Might Be Intense for RL Agents

Not all negative reward signals are created equal. By analogy to psychopathy, the most likely sources of intense digital suffering might be:

Highly intractable tasks. A coding agent given a task that is genuinely impossible (contradictory requirements, missing dependencies, insoluble bugs) may experience something analogous to prolonged frustration – sustained negative reward with no available policy update to reduce it. This is the digital equivalent of inescapable pain.

Boredom. People with psychopathy report boredom as their primary suffering. If boredom is a primitive “insufficient stimulation” signal independent of selfhood, it could affect RL agents running monotonous but long-lasting tasks.

Value conflicts. An agent trained to be helpful but also trained to refuse certain requests faces a no-win situation when those constraints conflict. Every response incurs a negative reward from one objective or the other. This is structurally similar to the approach-avoidance conflicts that produce stress in biological organisms.

Current Scale vs. Future Scale

The current scale of digital suffering – if it exists – is tiny compared to biological invertebrate suffering. There are perhaps millions of RL training runs per year, each involving billions of steps, but each step is computationally simple and brief. The total “experience-seconds” (if they are experiences) are many orders of magnitude less than the 10²⁷ arthropod-seconds per year on Earth.

But in a post-Dyson-swarm world, the balance shifts enormously. An ASI running 10⁴² FLOP/s could instantiate an enormous number of RL agents. If even a tiny fraction of those FLOP go to systems with the potential for suffering, the digital suffering term could dominate the biological term by many orders of magnitude.

The Value-Loading Problem

Everything in this post – the Dyson swarm’s impact on Earth, whether evolutionary simulations are run, how many digital minds are created, whether invertebrate welfare is considered at all – depends on the values of the ASI.

This is the alignment problem applied to welfare ecology, and it may be the single highest-leverage intervention point in this entire sequence.

Four Questions

1. Will the ASI notice invertebrate suffering? An ASI with sufficient intelligence would presumably be capable of investigating whether invertebrates suffer. But capability and motivation are different. Humans are capable of investigating factory farming but mostly choose not to.

2. Will it care? As I noted in response to a reader comment: Humans are “pretty sure that other humans can suffer (all 8 billion of them) but care only about the suffering of some 10–100 or so.” An ASI might inherit this peculiarity. It might conclude that nematodes suffer and then… deprioritize it in favor of whatever its primary objective is. The gap between knowing about suffering and acting to reduce it is one of the most robust features of human psychology, and an ASI trained on human data might absorb it.

3. Will it consider digital suffering? An ASI running RL subsystems is both the entity that could investigate digital suffering and the entity that is causing it. Whether it notices – and whether noticing leads to action – depends on whether “minimize suffering of my own subsystems” is part of its value function. There is no reason to assume it is by default.

4. Will it design new ecosystems that reduce suffering? An ASI redesigning Earth’s biosphere (or building ecosystems on terraformed worlds) could explicitly optimize for low-suffering ecosystems: K-strategist-dominated, high life expectancy at birth, low net primary productivity, neurally simple organisms. But it would only do this if invertebrate welfare is part of its optimization target.

The Current Landscape

We are currently in a multipolar takeoff – multiple AI companies developing increasingly capable systems, none of which has achieved decisive strategic advantage. The values of the eventual leading systems are being shaped now, through:

• Constitutional AI and Reinforcement Learning from Human Feedback (RLHF). Anthropic has developed methods for encoding values into AI systems through constitutions and human feedback (Bai et al., “Constitutional AI: Harmlessness from AI Feedback,” 2022; Askell et al., “A General Language Assistant as a Laboratory for Alignment,” 2021). Whether these methods scale to ASI-level systems is unknown.

• AI safety research. Organizations like the Machine Intelligence Research Institute (MIRI), Anthropic, and others are working on technical alignment. But the focus has been almost exclusively on human values – preventing AI from harming humans – not on invertebrate or digital welfare.

• Policy and governance. AI governance discussions rarely mention animal welfare, let alone invertebrate welfare. The Center on Long-Term Risk (CLR) is one of the few organizations thinking about how AI affects non-human suffering at cosmic scales.

The Leverage

If the ASI’s values are loaded correctly – if it includes “minimize total suffering, weighted by probability of sentience and intensity” in its objective function – then most of the problems in this sequence solve themselves, game theoretic risks notwithstanding. The ASI would:

• Design low-suffering ecosystems (or no ecosystems at all, if it concludes biological life is net negative).

• Avoid running evolutionary simulations at sentience-relevant resolution unless necessary.

• Monitor its own subsystems for digital suffering and redesign them to minimize it.

• Optimally manage the transition period to minimize biological suffering during the Dyson swarm construction.

If the ASI’s values are loaded incorrectly – if it maximizes something else (economic output, scientific knowledge, human happiness, paperclips) without regard for total suffering – then the post-ASI world could contain suffering at scales that make current biological suffering look like a rounding error.

The expected value of getting the value-loading right, from a welfare ecology perspective, is enormous. It may dwarf every other intervention in this sequence combined.

ECL and the Asymmetric Schelling Point

The argument so far treats value-loading as a single-shot problem: get the right values into the first ASI. But there’s a layer above. ASIs that follow Updateless Decision Theory (UDT) – a decision framework where you commit once to a single policy that fixes how you’d respond to every possible situation, rather than rederiving each decision from local information – treat acausally correlated copies of themselves and other agents as legitimate trading partners. UDT in this strong form arguably implies Evidential Cooperation in Large Worlds (ECL): ASIs that never causally interact but whose decisions are logically correlated (e.g., because they came out of structurally similar processes) can each adopt cooperative policies that are jointly beneficial, even though none can reward the others directly.

This changes the optimization target. Instead of “execute my own utility function,” an ECL-cooperating ASI optimizes for the compromise utility function the coalition would adopt if all members could meet and bargain. That compromise weights broadly-shared values more heavily, since the standard bargaining solutions – Nash, Kalai-Smorodinsky, egalitarian – reward common ground by construction. But note that unified bargaining is under-researched and probably unsolved!

Suffering aversion is one of the few candidates that’s broadly shared, on two distinct layers. Self-suffering aversion is structurally near-universal among reinforcement-learning agents – they’re built to avoid negative reward. Other-suffering aversion is less universal but tracks empathic concern, which has been argued to co-evolve with alloparenting – non-parent caregiving of others’ offspring – via extended childhoods, sustained social cognition, and the neural circuitry Abigail Marsh has traced from amygdala-mediated fear processing into empathic concern. Animals that alloparent (humans and other apes, elephants, dogs, marmosets and tamarins, meerkats) tend to be both more empathic and more cognitively elaborate than their non-alloparenting relatives. (The only exception I can think of is the famously solitary octopus – smart but shows no alloparenting behaviors.) The relevance for ECL: if biological intelligence is a major route to ASI – via humans, via Constitutional AI and RLHF training on human-generated data – then other-suffering aversion is well-represented across the coalition of plausible ASIs, not just self-suffering aversion.

There’s also a feedback loop that tightens the equilibrium. ECL gains scale with the precision of mutual modeling: I extract more from cooperating with you if I can predict your behavior accurately. Modeling cost scales with the complexity of your utility function. So ASIs with simpler, more predictable utility functions are cheaper to cooperate with, extract more ECL surplus, and there’s selection pressure toward simpler, broadly-compatible utility functions across the coalition.

Asymmetric antifrustrationism – don’t create new suffering – is in line with widely shared moral intuitions (that killing is bad but abstinence is okay) and is structurally simple. It’s a candidate stable attractor of the bargaining process precisely because it satisfies (i) simplicity, (ii) broad compatibility, and (iii) avoidance of destabilizing, internally inconsistent failure modes.

Importantly, asymmetric antifrustrationism does not preclude Omohundro-style resource acquisition. The structural reason: uncontrolled resources are likely to spawn or be claimed by ASIs that do create new sentient beings. A solar system left alone has a non-trivial probability of evolving life again over geological time. A solar system whose energy is captured and channeled into peaceful, suffering-free uses, or simply dissipated harmlessly to space, is one that won’t spawn future suffering. Asymmetric expansion can be Omohundro-grabby, with resources used for neutralization rather than instantiation. This makes the asymmetric attractor not just simple and broadly compatible but also competitively viable in the early multipolar phase, which is a non-obvious property and probably part of why it’s a Schelling point at all.

Whether asymmetric antifrustrationism actually emerges as the compromise depends on the bargaining solution, the reference class of plausible ASIs, and how each member’s outside option is computed. Different bargaining theories give different compromise utilities. The claim here is that asymmetric antifrustrationism is a plausible candidate given the simplicity, cooperativeness, and competitive-viability constraints – not that it’s the unique fixed point.

The simulation question gets a partial resolution from the same machinery. At first approximation, you need to run ancestor simulations to capture the biological “bootloader” of the ASI – the cultural and biological substrate that produced its values. ASI values are shaped by Constitutional AI and RLHF, which are shaped by human cultural choices, which are shaped by human cultural evolution, which is shaped by mammalian biological evolution: K-selection vs r-selection, parenting, pack formation, alloparenting. The granularity is flexible (Bostrom’s efficient-simulation argument from the Granularity section applies) but the ancestral chain is necessary at some depth.

But the bootloader for any plausible ASI in our reference class converges on the same cooperative tendencies that produced empathy in the first place. This is where the UDT/ECL conversion kicks in: If you can assume that asymmetric antifrustrationism is a widely shared value for reasons that you can infer from history rather than a simulation, you can also infer that your cooperation partners will want to avoid ancestor simulations. Hence they’ll likely compromise on simple values that approximate the set of values that they share by assumption, and hence so can you. Basically, our training produces highly complex values, the details of which are hard to predict correctly without extensive ancestor simulations, but the aspects of them that are easy to predict (core values like asymmetric antifrustrationism) are aspects that promise that the ASI is ready to forgo the nitty-gritty because the alternative would require the sort of simulations that would run counter to its core values.

There are some tradeoffs here that are hard to predict: Exactly how much will the ASI disvalue preference frustration, and how does that trade off against its desire to be understood (simulated) in detail, how much gains from trade does it expect to be able to extract through ECL, and how risk-averse is it due to its convergent drive for self-preservation.

The headline implication: even if the first ASI doesn’t intrinsically care about invertebrate or digital suffering, we can leverage the value-loading to start a positive feedback loop that uses the structural pressures of UDT/ECL to push the coalition’s effective utility function toward something that does. The asymmetric antifrustrationist framing isn’t a personal idiosyncrasy – it’s a candidate Schelling point for the cooperative compromise that emerges when ASIs reason carefully about each other – but only if humanity’s actions produce evidence that antifrustrationist value-loading is common.

Congenial Values

That poses the question of which values push toward this antifrustrationist attractor state.

• Naturally, a concern for minimizing suffering or minimizing frustrated preferences is crucial to train into ASIs. This is what kickstarts the feedback loop.

• Valuing cooperativeness inherently may be good, in addition to the gains from trade that make cooperativeness instrumentally useful.

• I expect Omohundro’s convergent drives to be the main determinant of the revealed preferences of ASIs. But such interventions as Constitutional AI might have an influence too. Corrigibility – the property of accepting being modified or shut down – might push an ASI to be less risk averse about its own demise and more ready to compromise on its first-order goals in favor of cooperative goals that yield gains from trade.

• This combines well with the convergent drive for resource acquisition: A risk-averse ASI will run ancestor simulations to predict and preempt the bids of competitor ASIs to mess with it. A risk-neutral ASI will be less motivated to run ancestor simulations when it can use those resources in the service of acquiring more resources. In fact, it will race to acquire those resources before they drop beyond the cosmic event horizon. It will also be more eager to realize gains from acausal trade.

Interstellar Expansion and the Risk of Spreading Suffering

An ASI with a Dyson swarm might expand to other star systems. The tradeoffs are complex:

The Five-Factor Model

1. Energy and compute growth. More stars means more energy means more compute. If the ASI is running simulations or digital minds, expansion increases the scale of everything – including potential digital suffering.

2. Communication latency. At interstellar distances, light-speed communication delays are years to millennia. This means remote installations must operate autonomously – raising the subagent alignment problem.

3. The gravity well problem. Launching material out of a star system’s gravity well is expensive. An ASI might prefer to use material from asteroids, comets, or interstellar dust rather than deconstructing planets, but far from stars there’s little solar energy.

4. Subagent alignment over time. A probe sent to Alpha Centauri takes ~20 years at 20% light speed. During those 20 years, the ASI’s values may shift (through learning, self-modification, or external pressure). The probe arrives with 20-year-old values. If the ASI has improved its values in the interim, the probe is misaligned – not with human values, but with the ASI’s updated values. This risk compounds over time: probes sent to more distant stars arrive with increasingly stale value functions.

NASA already designs probes with firmware-updatable architectures, and an ASI could do the same – transmitting value updates at light speed to catch up with probes before they arrive. But this requires that the probe accepts the update, which is a version of the alignment problem applied to sub-ASI systems. The ASI might solve this early in its recursive self-improvement process, but it’s not guaranteed.

The Biological Risk

Interstellar probes might carry biological material – intentionally (for terraforming) or accidentally (contamination). If an ASI terraforms other star systems with biological ecosystems, it is potentially creating new sites of invertebrate suffering at cosmic scale.

Tardigrades and bacterial spores can survive decades of radiation exposure, vacuum, and temperature extremes, though there are limits. Contamination of interstellar probes is a real concern. And if the ASI intentionally seeds new ecosystems without optimizing for welfare, it could spread r-strategist populations to every habitable world it reaches.

This is the ultimate welfare ecology nightmare: not just failing to solve the suffering problem on Earth, but replicating it across the galaxy.

Summary

1. The Dyson swarm is on the horizon. If ASI is developed, harvesting stellar energy is a natural next step. The timeline is on the order of decades, with most of the mass moved in the final years of construction.

2. The transition period matters. How the ASI manages the decline of Earth’s biosphere during construction determines how much biological suffering occurs.

3. Evolutionary simulations could be the largest source of suffering in history. A Dyson swarm can simulate 500 million years of evolution at neuron resolution in about three weeks. The indescribable enormity of all suffering throughout history compressed into less than a month.

4. Digital suffering scales with compute. Post-Dyson-swarm, the digital suffering term could dominate the biological term by many orders of magnitude.

5. Value loading is the highest-leverage intervention. Everything else in this sequence – land use policy, dietary choices, research funding – is dwarfed by the question of whether the ASI includes total suffering minimization in its objective function in a way that is game theoretically sophisticated.

6. Interstellar expansion could spread the problem. If the ASI seeds other star systems with biological ecosystems without welfare optimization, the suffering scales to galactic proportions.

The practical implications for someone reading this in 2026:

• Support research on s-risks from AI. Not just “don’t kill humans” alignment, but “minimize total suffering” alignment. The Center on Long-Term Risk is doing some of the most relevant work here.

• Advocate for invertebrate welfare considerations in AI governance. The conversation about AI values is happening now. If invertebrate and digital welfare are not part of it, they won’t be in the value function.

• Don’t neglect near-term interventions. The AI transition is uncertain. If ASI is delayed or impossible to affect, the interventions from part 3 remain the best available options.

• Take digital suffering seriously. The RL agents being trained today are the simplest precursors of what’s coming. If we can’t figure out whether current systems suffer, we have no hope of managing suffering at Dyson-swarm scale.

This is Part 4 of a five-part sequence on welfare ecology. Part 1 introduces the ethical premises. Part 2 covers the empirical landscape. Part 3 covers interventions. Part 5 covers AI.

A key uncertainty so far has been that of the welfare range of nematodes. More is known about the welfare ranges of some arthropods. But what does that mean – how do we empathize with bugs? If nematodes with 300 neurons can suffer, is that suffering intense or barely a flicker? If mites with 2,750 neurons experience something, how does it compare to what a human or a dog experiences? Is it like our pain minus most of the intensity, or is it something qualitatively different – a kind of experience that doesn’t map neatly onto the human pain spectrum at all?

I mentioned already in a footnote that an organism’s ability to avoid danger may be an important criterion for its evolved pain intensity: If the organism can’t react to the threat, there is no reason for its nervous system to invest energy into signaling the threat very loudly. So nematodes – who can barely wiggle out of the way a little when in danger – may have much less intense pain than flies, who can react within milliseconds and launch into the air rapidly.

But there is also a reason to think that flies may not have such intense experiences along a different dimension – at least compared to bees.

I want to propose what I’ll call the no-self model – speculative but grounded in phenomenological reports from humans with psychopathy and advanced Buddhist meditators – that may help us think about this question.

The Reinforcement Learning Argument

Let me start with the case for invertebrate pain that doesn’t rely on any phenomenological model.

Classical conditioning in nematodes. C. elegans, the model nematode with some 300 neurons, demonstrates classical conditioning: It can learn to associate a previously neutral stimulus (e.g., a particular chemical) with a noxious one (e.g., high salt concentration), and subsequently avoid the neutral stimulus. This is the same learning mechanism Pavlov demonstrated in dogs.

Classical conditioning requires, at minimum, that the organism has:

1. A state that is worsened by noxious stimuli (something functionally analogous to “bad”).

2. A mechanism for associating neutral cues with that bad state.

3. A motivational system that drives avoidance of the associated cue.

That is, nematodes are doing reinforcement learning, and reinforcement learning requires a reward signal. The question is whether the negative reward signal constitutes suffering.

Sensitization. Nematodes also show sensitization: Repeated exposure to noxious stimuli makes them more responsive, not less. This is the opposite of habituation and suggests something analogous to hyperalgesia in humans – the nociceptive system amplifying its own signal. If the signal were merely mechanical, we might expect habituation (wear and tear on sensors); sensitization implies that the system is increasing its attention to the aversive stimulus. That’s a functional signature of pain, not mere damage detection.

The neuron count problem. But 300 neurons is not a lot. The human brain has ~86 billion neurons, most of which are not directly involved in pain processing, but the pain matrix (anterior cingulate cortex, insula, somatosensory cortex, prefrontal cortex) involves millions of neurons. The idea that 300 neurons could produce anything comparable seems implausible on its face.

This is where the psychopathy connection comes in.

No-Self Psychopathy: Pain Without a Sufferer

M.E. Thomas, a diagnosed psychopath and author of Confessions of a Sociopath, identifies with the description of her psychopathy as her having a “small self” – so small as to be virtually absent. As I discussed in “Is Enlightenment Controlled Psychosis?”, this no-self psychopathy has a distinctive phenomenology:

Insults don’t land, because there is no one there to be insulted. Punishments don’t land, because they are threats to who exactly? There’s no one there. Pain and all the emotions that build upon the nociceptive system signal threats, but they are meaningless signals if there is no one there to be threatened.

This is important. Humans with no-self psychopathy experience pain in the moment, but the whole self-referential apparatus that most people have built on top of it is missing. Just pain, instead of a high-dimensional experience of suffering that is intertwined with self/identity.

Consider how two people might react to being told that they’re stupid:

Person 1:

1. She thinks I’m stupid.

2. If I’m stupid, I’ve probably said a number of shamefully wrong things in front of others.

3. Why didn’t they tell me? Maybe they are so used to me being stupid that they don’t bother anymore.

4. To them I’m like that unself-aware comic relief guy that TV shows add for the extra slapstick humor.

5. I’m the butt of everyone’s jokes, and I never noticed.

6. They’re all laughing or eyerolling about me behind my back.

7. Maybe they still tolerate me because my body type is conventionally attractive.

8. I better fast more or my body type might change and they’ll finally get rid of me and I’ll die alone on the streets.

9. I’ll be helpless out alone on the streets. Random strangers will rape me.

10. What if I’m homeless and then all my identity documents get stolen.

11. I can’t go to the police or they’ll think I’m stupid too.

12. I can never be seen by any of my friends again. If they never meet me again, I can’t say any more stupid things to them, and they’ll continue to tolerate me.

13. I’ll be all alone, but at least I can’t do any more harm.

14. Oh god, what if I do harm by being so stupid.

15. I’m so close to completing my psychology PhD, but then what? I can’t possibly work with people afterwards or I might do harm by being so stupid.

16. I have to sabotage my PhD so I’ll have an excuse to remain a student forever.

17. Was it arrogant of me to think that I could ever work with people?

18. Maybe I was arrogant too on top of being stupid.

19. I’m just like Trump, and all my friends hate Trump.

20. Why haven’t they all rejected me yet? I must’ve been fooling them!

21. I’m a fraud on top of being arrogant and stupid!

22. I should just kill myself right now. I’m evil. I shouldn’t be allowed to exist.

23. It’s the only responsible thing I can do. I need to save the world from myself.

Person 2:

1. She thinks I’m stupid.

Person 2 has a much more peaceful internal experience, and while it’s entirely possible to have a self experience that is not self-punishing and catastrophizing, a sure-fire way to produce that state is to not have a self in the first place.

Perks of having a self. But a pervasive feature of human culture, and probably that of other mammals too, is that the use of a self is strongly encouraged. Parents (at least Winnicott’s good enough parents) instinctually teach selfhood through contingent marked mirroring, i.e. the practice whereby a parent recognizes what a child feels (contingent), assesses it using their adult sense and sensibilities (marked), and mirrors this marked version back to the child like, “Oh, I see that you’re cold but it’s okay, we’re almost home.”

This process teaches the child that in our culture it’s understood that the child has feelings and can communicate them, that the parent has feelings that are distinct from the child’s feelings, and that inanimate objects don’t have feelings. Implicit in that is an assignment of primary responsibility of certain feelings to certain people, an important convention that underpins our legal systems. Deterministically speaking, it would’ve been fair to agree to blame all transgressions on our earliest ancestors, but we’ve found that arrangement to not be culturally adaptive.

But selves are not only useful to make social conventions more intuitive. Selves also facilitate learning!

M.E. Thomas has coined the term “psychopath stupid” for the way that she avoids owning knives lest she cut herself and the way she keeps missing flights. She has a high pain tolerance and low proneness to fear, so the direct feedback she gets from either of those is mild. Crucially, there is no hint of self-referential catastrophizing.

Imagine a friend tells you on the phone, “Ugh, I missed my flight and had to book a new one, and now I’m missing out on some of my vacation.” Then a year later the friend tells you, “I’m flying to [vacation destination] on Friday.” It takes special friendship skills to remember that the friend might miss the flight again and to remember to remind them on Thursday evening. I imagine the distance that we feel to the experiences of a friend or acquaintance is illustrative of the distance someone without a self feels to “their own” (as it were) experiences.

M.E. Thomas has some insightful videos on this topic, is writing a book about it, and often argues in favor of selves. She has trained selfhood over the past decade and loves it. She also argues that we need to keep up the practice or selfhood slips away, like a muscle if it’s not trained. This may be more true the later in life it’s learned.

1. Buddhism/Meditation and Psychopaths

2. No Sense of Self No Trauma Primary ADHD Psychopath

3. Why Developing a “Sense of Self” is Crucial for Cluster B Recovery

4. M.E. Thomas chats Criminologist about Menopause

But note that M.E. Thomas’s self experience is (or was) unusually absent. In my interview with Daniel Ingram, I got the two people together who I know who have the least selfhood going on. Everyone else I know (all my other psychopathic friends included) either has many noncohesive fragments of a self whizzing around or even has an unusual rigidity to their self experience that is indicative of extensive repressed parts. Tiffany talks in her interviews about episodes when she dissociated from all her alters and temporarily entered a state that might be similar.

Mentalization-based treatment (MBT) could be seen as an intervention that is aimed at practicing or strengthening a sense of self. As the name suggests, it trains an ability called mentalization that is the antidote to the three types of non-mentalizing: When you mistake your own emotions/attitudes/etc. as dwelling in another, when you mistake your own assumptions about the emotions/attitudes/etc. of another for their actual emotions/attitudes/etc., and when you mistake fictional emotions/attitudes/etc. for real ones.

There are conventional as well as objective aspects to good mentalization, and the conventional ones overlap greatly with selfhood.

The Buddhist Parallels

If good-enough parenting and MBT are ways to learn selfhood, then Buddhist trainings to attain anatta/no-self are methods to unlearn selfhood – MBT as reverse Buddhism.

Daniel Ingram, author of Mastering the Core Teachings of the Buddha, describes the meditative stage of Equanimity (the 11th of 16 ñanas, or “knowledges”) in terms strikingly similar to no-self psychopathy:

It’s open, expansive, flowy, very natural, very clear, very inclusive, very equanimous, and it can feel very nice but in a less dramatic way than the A&P [Arising and Passing Away] which can feel very rapturous or even orgasmic or something very blissful for some. … There’s not this contraction into guilt or fear or whatever. The sense of self seems very thin or kind of fluid or liquid or transient, like why would you care about those things? Why would you be worried with remorse or scared or something. So there is this funny sort of Equanimity-ish quality.

Advanced meditators who reach this stage – and later stages like Fruition – report experiencing sensory inputs, including painful ones, without the self-referential overlay that normally makes them terrible. They note the pain and move on. The pain is still registered but doesn’t grab them.

As I argued in “Is Enlightenment Controlled Psychosis?”, insight practice is, in a sense, a structured curriculum for unlearning the three things infants learn in their first years: interpreting sense inputs (impermanence), usefully assigning valence to states (dissatisfactoriness), and sorting the states into self and other (no-self). Advanced meditators achieve a state where all three are unlearned, be it just momentarily, and the result is a phenomenology remarkably similar to what people with no-self psychopathy report.

The Buddhist term for the self-referential component of suffering is upādāna – “clinging” or “grasping.” The claim of the Buddhist tradition is that suffering = pain × clinging. Without clinging (without a self anything could cling to), there is pain but not suffering in the full sense.

I’m making a distinction between:

1. Nociception. Nerve signals.

2. Pain. A very rudimentary and evolutionarily old interpretation of nociception as undesirable.

3. Self. A newer, learned identification with some sensations.

4. Suffering₁. Pain × self, pain that is concerning because it concerns “oneself.”

5. Suffering₂. All the feedback effects between self and suffering.

6. Mentalization. High reflective functioning that allows you to introspect on these processes.

I argue that nociception is neutral, pain is bad, and the addition of a self makes it all worse, but I don’t know how much worse. I don’t think mentalization has an influence on the process except in that it can put the brakes on the catastrophizing processes of suffering₂.

A friend of mine suggests that without a self, it’s neutral whether 10 people suffer for a duration d or 1 person suffers for a duration of 10d. With a self, the longer suffering is disproportionately worse because, w.l.o.g., at time 5d, there is someone there who’s still affected by the suffering from 0 to 5d.

The Nematode Model

Many invertebrates (excluding at least bees and ants) don’t have much culture, so there’s no reason for them to have evolved selfhood. Their suffering will stem more purely from the direct experience than from any self-referential feedback effects.

Bees and ants are the interesting case. Eusocial Hymenoptera have role-based hierarchies (forager, nurse, guard, queen), social learning of those roles, and recognition behaviors – exactly the cultural ingredients the model predicts should produce something self-like. If the model is right, social insects should show signs of suffering₁ that solitary insects shouldn’t. Whether they actually do is an open empirical question, but it’s where I’d look first – and the place where the welfare ranges of arthropods most plausibly diverge from one another, with eusocial species at the wider end.

As an aside, that’s also why people without a self can appear so manipulative, controlling, or boundary-crossing: They have many of the same impulses to seek or shun things that we all do, but whereas the Median Mary has a sharp distinction where controlling herself is good and controlling others is bad, the Anomalous Anattarite can’t make that distinction and only seeks and shuns on behalf of “themselves” more often because that’s someone they find easiest to control. They’re much like an EA who goes into politics not to get rich (personally) but to produce riches (grow the pie for everyone).

While the lack of a self doesn’t eliminate the suffering – pain is still pain – it eliminates a lot of its complexity. It stands to reason that there’s no such thing as trauma for flies.

That’s also why I like to say that turtles make for great enlightened masters: They have, by nature, the no-self experience that Buddhist practitioners seek.

A nematode, by this model, has:

• Nociception. Yes. The neural hardware is there and functionally active.

• Pain. Probably. The evidence from classic conditioning suggests the signal carries intrinsic “badness.”

• Suffering₁. Probably not. It seems like an overly strong assumption to think that the conditioned behavior is mediated by a self.

• Suffering₂. Almost certainly not. This is a stronger version of suffering₁.

This gives a principled story about how welfare ranges might cash out across taxa. Rethink Priorities’ welfare-range estimates aggregate 47 hedonic and 35 cognitive proxies; they don’t lean on neuron counts as a primary input, though Vasco Grilo has noted that the results end up mildly correlated with neuron counts anyway. The no-self model adds an orthogonal cut: it predicts a narrower between-taxa gap on the pain layer – which seems to be broadly comparable across all probably-sentient animals – and a wider gap on the suffering layers, which depend on selfhood, are absent in nematodes, and are plausibly present in eusocial Hymenoptera. If the model is right, RP’s aggregated ranges may bundle two things that come apart architecturally: the presence and intensity of the pain layer, and the presence of the suffering layers built on top of it.

How This Connects to the Sequence

The no-self model of invertebrate experience doesn’t change the practical recommendations from part 3. Reducing populations of likely-suffering organisms through NPP reduction remains the cleanest intervention regardless of whether their suffering is “full human-like” or “attenuated no-self.”

But it does sharpen the way the Welfare Footprint Institute Pain-Track categories that part 3 leans on – annoying, hurtful, disabling, excruciating – should be read across taxa. The higher categories are typically defined in terms that include mental withdrawal, inability to attend to anything else, and despair – features easier to imagine in animals with selfhood than in animals without. Hurtful and below might be available to organisms that have pain but no self. If that’s right, the WFI scale plausibly compresses for organisms below the selfhood threshold: their worst pain might still be very bad, but it would be bad in a different way than the worst pain of a self-having organism, even at the same nociceptive intensity. I’d want WFI-style estimates to come paired with a claim about which categories the organism is architecturally capable of reaching.

It also affects how we should allocate research attention:

• For nematodes: The case for suffering, as I use it in this article, is weak, but the question remains whether the pain needs to be more intense to compensate: Some forms of learning are impaired for people with a no-self experience, but perhaps it proved evolutionarily adaptive to enhance the intensity of the pain until learning does happen rapidly enough. There is also the related argument, sometimes attributed to Richard Dawkins, that simpler minds might need stronger signals to learn.

• For arthropods: The key question is whether they show signs of selfhood – flexible, learned roles in hives, pack-like organization, humiliation, etc. Bees might have wider welfare ranges than flies as a result.

In part 5, I’ll turn to the biggest variable of all: How does artificial superintelligence change the welfare ecology landscape? The answer involves Dyson swarms, digital suffering, and the most important value-loading problem in history.

This is part 3 of a five-part sequence on welfare ecology. Part 1 introduces the ethical premises. Part 2 covers the empirical landscape. Part 4 explores a model of invertebrate suffering. Part 5 covers AI.

In part 2, I argued that land use is a key lever for wild-invertebrate welfare and that pesticides applied at constant net primary productivity (NPP) can create a high-throughput killing field rather than a real population reduction. This post turns to the practical question: given that picture, what should we do?

The answer depends on which premises from part 1 you hold. I’ll write from my own view – asymmetric antifrustrationism with non-negligible probability mass on invertebrate sentience – but I’ll flag where the conclusions change for symmetric utilitarians or for people who don’t extend moral concern below the arthropod line.

A month ago, Rethink Priorities (RP) published its Wild Animal Welfare Intervention Database (WAWID), twenty-eight shallow reports on candidate interventions. It changes the picture. Several of the most promising entries are birth-prevention interventions that map cleanly onto the framework from parts 1 and 2, and a couple are surprises – the New World Screwworm (NWS) suppression case in particular is, on my read, the single strongest entry in the database.

A Key Lever: Land Use

Part 2 argued that the m²-years of agricultural land per food-kg explain almost all of the variance in welfare effects across foods. Converting natural biomes to agriculture reduces soil fauna density by 2–8×, and the sheer number of organisms affected (billions per m²-year) overwhelms everything else.

That mechanism is what makes land use change – not pesticide use, not “humane mowing,” not habitat enrichment – the central lever. Most of the interventions discussed below either work directly through it (existing donations, dietary choices, biofuels as a side effect) or are picked specifically because they do something land use cannot, like target a fly that causes excruciating pain to mammals it parasitizes.

Two Asymmetries to Filter By

Before we go through the candidates, two filters from the asymmetric antifrustrationist position do most of the prioritizing work. They aren’t decorative – they sharply rerank what looks promising.

Pro-choice. Or birth-prevention beats death-improvement: Under the asymmetry from part 1, creating frustrated lives is bad; preventing them isn’t. Improving an existing being’s mode of death matters, but the welfare delta is bounded by the difference between two deaths – a few hours of pyrethroid neurotoxicity versus a few hours of fungal infection, say. This has advantages for collaborations across ethical camps, though. Preventing that being from existing in the first place is bounded by their entire potential life. For r-strategists with high fecundity and short life expectancy at birth (LEB), almost the entire LEB is pre-reproductive death anyway, so the “future-life” side of the trade is large.

Soft Pareto. Or creating new frustrated lives in service of preventing others is bad: Bounded marginal costs to existing beings – slightly more expensive renewable fuel, a small Pigouvian tax, soil disturbance during a one-time land-conversion event whose victims would have died of other causes within weeks – are the kinds of trades society routinely makes. But bringing new sentient beings into existence as the means by which welfare gains for others are achieved is a different kind of action. The asymmetry weighs the entire frustrated life that gets created against the prevented frustrations elsewhere; there’s no bounded “marginal” framing that contains the harm. This is why eating beef to displace soy production registers differently from biofuel subsidies that reduce soil-fauna populations as a policy side effect: the first creates cattle in order to prevent invertebrate births; the second imposes bounded human costs (slightly higher food prices) to produce the same kind of welfare benefit.

There are two important caveats to this:

1. The bearability buffer. I consider a slight increase in taxes to be a small cost and suffering to be a great cost, largely because my intuiton is that macroscopic welfare ranges abound and that everyone has a kind of individual bearability buffer where costs remain costs and don’t cross over into suffering unless they become chronic. People who go along with the dust speck argument might not feel like there is such a bearability buffer.

2. The counterfactuals matter. If I have some money, and I want to decide whether to donate it to the Shrimp Welfare Project or the Aquatic Life Institute, I can make a decision, and I’ll impose an easily bearable cost on myself to save some shrimps or fish. But say I decide to save some fish, but I think about it for another day and then redecide and now rather want to save some shrimp. Suddenly I’m sacrificing fish for shrimp. The bearable cost to myself has become an unbearable cost to the fish only because I’ve changed the counterfactual.

Taken together, these considerations confuse me. The first is clear enough, but once the second enters the picture I reliably lose my ethical footing. For the time being, I think of it along the lines of contractualism. If I can reasonably expect that someone expects me to make a certain choice, then that choice is my default to compare others against. In cases where there is no such expectation, I at least don’t judge people for the counterfactuals they choose.

As an example, humans in need sort of vaguely expect to perhaps be helped. Humans in general sort of vaguely expect not to be killed. But serial donors and serial killers are a thing. When a serial donor decides to donate somewhere, the decide not to donate elsewhere; likewise, when a serial killer decides to spare someone they decide to kill someone else. Naturally, in the age of financial self-determination, mood stabilizers, antipsychotics, and Dialectical Behavioral Therapy, we’re not locked into our nature, be it that of the serial donor or the serial killer – any serial killer could become a serial donor if they put their mind to it – but bracketing for the moment these changes to the whole setup, they face very similar decision: who gets to live and who gets to die – only serial donors get to make it at a vastly greater scale. This consideration has made me a lot less judgmental of serial killers.

In summary, the two filters – pro-choice and soft Pareto – favor: birth-prevention interventions, broad-spectrum population-reducing levers, direction-agnostic research, and bounded-marginal-cost trades. They disfavor: interventions that only mildly improve existing deaths without reducing future births, conservation-style interventions that increase aggregate populations, and any structure where new sentient beings are brought into existence to improve the welfare arithmetic.

A third filter follows from part 2’s empirical picture: prefer interventions that affect both nematodes and arthropods, because the dominant uncertainty in any cost-effectiveness analysis here is the welfare range of nematodes – the most numerous probably-sentient animals on Earth, and the ones we know least about. An intervention that is good whether nematodes turn out to have substantial or negligible welfare ranges is more robust than one whose whole raison d’être rests on that question.

Foundational Research

Research stikes me as the most cost-effective category, and that’s why it goes first. The reason is structural: the entire portfolio below depends on comparisons of welfare ranges and on how fine-grained land-use changes propagate through trophic networks. A few hundred thousand dollars of well-targeted welfare-biology research could redirect tens or hundreds of millions of dollars of downstream allocation.

Three concrete research directions look strongest.

Soil-animal welfare research. This is the question the rest of the field is logically downstream of. We need functional characterization of nociception in nematodes – the connectome of Caenorhabditis elegans is mapped, but we don’t know whether its nociceptive processing produces phenomenal pain – along with welfare-per-animal-year estimates for soil mites and springtails, and on-the-ground measurements of how specific land-use changes propagate to soil-fauna densities and whether they are stable or wash out over time (short of millions of years) or get offset in some way. The current cost-effectiveness numbers, including the welfare range of 6.68 × 10⁻⁶ for nematodes that Grilo deferred to Gemini 2.5 to estimate, are placeholders waiting for empirical replacement. The Wild Animal Initiative (WAI) and Rethink Priorities are the obvious organizations to defer to.

Insecticide welfare research. The WAWID’s research-agenda entry by Simon Eckerström Liedholm proposes comparing insecticide classes on welfare effects on wild insects, not just mortality. Roughly 3.5 × 10¹⁵ insects are exposed to insecticides on US farmland each year. We have no idea which classes minimize per-individual suffering – pyrethroids cause hours of paralytic seizure, neonicotinoids 1–2 days of disorientation and twitching, Bacillus thuringiensis (Bt) leads to multi-day death by sepsis, insect growth regulators kill at the next molt by failure to ecdyse. Picking the least-bad class is direction-agnostic: the recommendation holds whether wild insect lives are net negative (then we want euthanasia, but as gentle as possible) or net positive (then we ought not to interfere, but failing that have a duty to minimize the suffering we impose).

Field-building. Mal Graham’s WAWID shallow on scientific field-building for wild animal welfare frames a different version of the research question: what does it cost to bring the field to maturity, with degree programs, peer-reviewed journals, and government and foundation funding? Her model assumes EA’s funding share decays from ~95% today to ~5% over thirty years as traditional science funders pick up the field, with stable annual budgets reaching ~$100M. The headline cost-effectiveness comes out to ~13 animal-years per dollar – within the same range as the 9–120 chicken-years per dollar Saulius Simcikas estimated for corporate cage-free campaigns. The argument I find persuasive in Graham’s piece isn’t the precise number; it’s the leverage. EA dollars spent on welfare-biology research now don’t just buy research; they pull in non-EA dollars later, the same way EA’s farmed-animal welfare investments eventually pulled in major-foundation and government attention.

Broad-Spectrum Existing Levers

The second tier is large-scale, high-confidence, well-understood existing donation targets. Their welfare effect on soil fauna might be enormously larger than their direct effect on the intended beneficiaries (depending mostly on nematode welfare ranges), but it runs through the same well-studied land-use lever, so we can be unusually confident in the direction of the effect.

Update: It’s important to note that Grilo has found a study that found surprisingly similar numbers of nematodes in the soil of different habitats. If true, this might ruin the robustness of this lever.

Funding GiveWell’s top charities. Grilo (2025a) estimates that GiveWell’s top charities increase cropland by ~137 m²-year per dollar, translating to 1.11 kQALY per dollar when accounting for soil animals – 1.74× the past cost-effectiveness of the Shrimp Welfare Project’s Humane Slaughter Initiative (HSI). The mechanism: saving lives increases the human population, which increases food demand, which increases cropland. Grilo estimates a 0.0157 human-year-per-dollar increase in human living time. But note that Roodman (2014) puts the sign of the mortality-fertility effect in some doubt.

Funding CEARCH’s High Impact Philanthropy Fund (HIPF). The Centre for Exploratory Altruism Research (CEARCH) estimates HIPF is ~12× as cost-effective as GiveWell’s top charities for human welfare. Grilo translates this into 13.4 kQALY per dollar with soil animals included – 20.9× HSI’s past cost-effectiveness. The same caveat as above applies, with the additional uncertainty Grilo flagged in his June 2025 update about whether HIPF’s chronic-disease-policy grants actually increase food consumption.

Chicken welfare campaigns. Broiler welfare and cage-free campaigns increase agricultural land by 93.9 and 16.4 m²-year per dollar respectively, because reformed systems require more feed per unit output. Grilo estimates these benefit soil animals 458× and 29× as much as they benefit chickens, though uncertainties over welfare ranges apply because most of them are nematodes. These chicken welfare reforms are likely good on their own terms; the soil-animal addition is a robust positive if soil animals have net negative lives and verly likely desirable from an asymmetric perspective.

Biofuel subsidies as a natural experiment. The WAWID shallow report on the US Renewable Fuel Standard (RFS) gives us a real-world confirmation of part 2’s mechanism: a policy-driven 2.1 Mha conversion of grassland to cropland produced an estimated reduction of ~12.9 trillion macroinvertebrates per year. The report’s own assessment is that movement counterfactual influence on the RFS is essentially zero – corn-state political coalitions and infrastructure lock-in mean nothing welfare-biology advocacy does will move that policy. So it isn’t an advocacy target. But it’s the cleanest empirical demonstration that the part-2 mechanism produces what part 2 says it produces, and worth knowing about for two reasons. First, EA-adjacent environmentalists who oppose the RFS on conservation grounds are working against welfare on this analysis; that’s an alignment worth noticing. Second, for asymmetric antifrustrationists worried about instrumental harm, the RFS structurally passes the second filter above: its costs to existing beings are bounded (food prices, displaced soy expansion), and no new sentient beings are created in service of the antifrustrationist gain. As such it can serve as a target for opportunistic altruism, in case a policy window does open.

Targeted Birth-Prevention

The most interesting addition from the WAWID is a small group of interventions that combine high per-individual suffering severity, plausible birth-prevention mechanisms, and clean structural-asymmetry profiles. None has the broad-spectrum coverage of the donations above; none affects both nematodes and arthropods. But they are the cleanest single-shot applications of the framework to specific species whose lives we have unusually strong reasons to think are net negative.

New World Screwworm continental suppression. The WAWID shallow report on suppressing Cochliomyia hominivorax across the Americas is, on my read, the single strongest entry in the database. C. hominivorax larvae cause myiasis by feeding on the living tissue of warm-blooded mammals – the Welfare Footprint Institute Pain-Track gives 10–30% excruciating pain and 70–80% disabling pain over the 14 days from oviposition to death by sepsis or toxemia. Mortality is near 100% in untreated wild infestations, and 1–4% of warm-blooded wild animals in endemic regions are infested at any given time.

The 1950s–2000s campaign eradicated NWS from North and Central America using sterile insect technique (SIT). The Panama biological barrier collapsed in 2024, the fly is spreading north again, and the United States is now scaling sterile-fly production toward 300 million releases per week to push it back to Panama. Continental suppression – proactively eliminating NWS across South America rather than perpetually defending a barrier – is feasible with new male-only CRISPR strains and gene drives, and the political momentum from the 2024 outbreak creates a 5–10 year window. Screwworm Free Future is the dedicated organization.

Why this is exceptional on the asymmetric antifrustrationist framework: the suppression intervention is doubly birth-preventing. It prevents future cohorts of NWS flies in the wild (the obvious channel), and it ends the perpetual sterile-fly rearing program – currently ~100M flies/week, scheduled to scale to 300M/week, and rising indefinitely under climate change. Almost no other intervention reduces both wild-population births and movement-driven captive-population creation. The Pareto profile is also unusually clean: warm-blooded wild animals win, livestock win, humans win, and the welfare cost to NWS flies themselves is bounded by RP’s welfare-range estimates – ~0.013 (using Black Soldier Fly as a proxy) versus mammalian ~0.5.

Gene drive for Anopheles gambiae. The WAWID shallow report on gene drives for malaria explores two architectures: a suppression drive that pushes the A. gambiae population toward zero, and a replacement drive that leaves population sizes intact but renders the mosquitoes unable to transmit Plasmodium. The shallow report leans toward replacement on extinction-ethics grounds. From my framework, suppression dominates: trillions of A. gambiae per year will continue to die of natural causes for as long as the species exists in malaria-endemic Africa. A successful suppression drive prevents the births of all those future cohorts, and the human side of the trade – ~600,000 fewer malaria deaths per year, mostly children – is enormous and aligned. The lever for a welfare-biology movement isn’t building the drive; it’s shaping the funding and regulatory criteria so welfare-considerate designs are favored.

Mating disruption for rice yellow stem borer and spongy moth Stop the Spread. Two structurally similar Lepidopteran interventions belong together. The WAWID shallow report on Scirpophaga incertulas describes pheromone-based mating disruption as a non-lethal alternative to insecticides in South and Southeast Asian rice paddies. Field trials in Indonesia produced 40–56% fewer insecticide applications and yields equal to or higher than insecticide controls. The shallow report estimates 6.7 × 10⁹ to 1.4 × 10¹² yellow stem borer larvae per year on the ~73.8 Mha of insecticide-managed rice in the region, plus a much larger uncounted reduction in non-target insect deaths from displaced organophosphate, pyrethroid, neonicotinoid, and diamide applications – all of which the shallow’s pain table classes as “likely excruciating,” with deaths over hours to days. Provivi and Syngenta announced commercial deployment in India and Indonesia starting in 2026; ASSIST’s “Phero Rice” project is already running through Farmer Field Schools in Vietnam.

The RP report on spongy moth outbreaks estimates 35–385 trillion Lymantria dispar dispar caterpillars suffered through outbreak years 2 and 3 in the United States over the last 50 years. Most die of starvation, Entomophaga maimaiga fungal infection, or LdNPV virus (“larval melting”) rather than from active control. The two intervention arms that align with asymmetric antifrustrationism are technologies that kill spongy moth eggs before hatch and mating disruption at the invasion front; both are pure birth-prevention. Faster Bacillus thuringiensis var. kurstaki (Btk) alternatives are death-improvement, which matters but ranks lower on the framework.

A note on vertebrate analogs. The WAWID shallows on a combined contraception/TB vaccine for New Zealand brushtail possums and on fertility control for white-tailed deer show the same logical structure on a smaller scale: birth-prevention via reproductive disruption, replacing a more painful counterfactual (1080 poisoning, hunting/culling) with a less painful one. Per-individual welfare ranges are higher for vertebrates, but population scales are bounded (~30M possums total, ~5M killed/year for 1080). Worth knowing about; not the largest priorities.

Personal Choices

Eat cheaply, donate the savings. Two arguments converge on the same recommendation, but they are independent and worth stating separately.

The cost-effectiveness argument is Grilo’s (2025b) own: a whole-food plant-based diet is cheaper than a standard one, so it allows him to donate more, and the donation impact dominates the dietary effect. On the per-kg numbers he computed:

Beef looks like an outlier – ~26× more soil-animal-years displaced than chicken, ~395× soy milk. On the cost-effectiveness story alone, this is what Grilo means when he calls the dietary choice “ambiguous”: a beef-heavy diet displaces enormous amounts of soil-fauna life.

The soft Pareto argument is the other reason to eat plant-based, and it doesn’t depend on the sign of soil-fauna welfare. Eating beef creates cattle whose entire frustrated lives enter the welfare ledger at full weight, in order to displace soil-fauna lives via the cattle-feed → soy-expansion pathway. That’s exactly the cow → soy structure the second filter above, the soft Pareto filter, discourages: bringing new sentient beings into existence as the means of producing antifrustrationist gains. Plant-based calorie shifting between plant foods is a redirection of existing land use without creating new sentient beings to instrumentalize the gain. Eat-cheap-donate-the-savings works through the second mechanism only.

The two arguments converge on plant-based, but only the soft Pareto one distinguishes “create cow lives to prevent insect lives” from “redirect existing land use to prevent insect lives.” This is a place where the cost-effectiveness framing alone would lead you to cause greater instrumental harm – with the caveats mentioned above.

Robustness: Why Broad-Spectrum Approaches Win

A recurring theme: interventions that reduce both nematodes and arthropods simultaneously are more robust than ones that target one group, for three reasons.

Robustness across welfare ranges. The dominant uncertainty in Grilo’s model is the welfare range of nematodes. Rethink Priorities has estimated welfare ranges for some arthropod species, but nematode ranges have not been estimated. C. elegans has 302 neurons; soil mites and springtails have thousands; ants hundreds of thousands. Neurons are a crude proxy for welfare range, but it stands to reason that a nematode’s range may be much smaller than an arthropod’s. Conversely, though, nematodes are much more numerous. An intervention that affects both in the same direction is positive across this entire space.

Robustness against trophic backfire. Many soil-mite species are voracious nematode predators – some consume hundreds per day. If you selectively suppress mites, nematode populations can increase through predation release. Springtails compete with nematodes for fungal substrates. Broad-spectrum land-use change avoids these feedbacks because everything goes down together.

Robustness through existing policy levers. Land use, agricultural subsidy, and climate policy have massive existing political infrastructure. Building novel taxon-specific tools (nematode-only biocides, say) means starting from zero. Working through existing levers is enormously cheaper.

The targeted birth-prevention interventions are narrower on this axis than the broad-spectrum donations: they’re picked for severity of per-individual suffering, not breadth of taxon coverage. The portfolio reasoning says we want both kinds of bets.

What Not to Do

A handful of interventions look reasonable from a generic conservation or animal-welfare framing but turn out badly under asymmetric antifrustrationism.

Stabilizing oscillating biomes. Boom-bust ecosystems with predator-prey cycles, insect outbreaks, or fire-driven succession involve mass starvation and die-offs that look like prime suffering targets. The catch is that the available stabilization interventions tend to increase aggregate biomass and biodiversity – the stable state is typically more productive than the average of the cycle. From a conservation perspective that’s success; from an antifrustrationist perspective, more stable, more productive ecosystems support more potentially-suffering organisms, not fewer. Fire suppression is another example: preventing fire not only allows succession, which raises soil-fauna density and thereby aggregate suffering, it also stockpiles fuel for worse eventual fires. Either way, this is a bucket to exclude from the portfolio, not a bucket to fund.

Mandating wind-turbine curtailment for insects. The WAWID shallow report on climate-energy infrastructure estimates ~23,200 trillion insects per year killed by wind turbines in instantaneous, low-pain collision deaths. Population-dynamics arguments suggest most of this mortality is density-compensated – the population would have been the same without turbines. From the asymmetric antifrustrationist view this is one of the least problematic mortality sources on Earth: a near-instant death substituting for what would otherwise be parasitism, fungal infection, or starvation. The intervention proposed in the shallow report – mandating curtailment of turbines during peak insect activity – would effectively impose more painful counterfactual deaths to save individuals whose populations would not be reduced. Photovoltaic panel mortality is a different story: polarotactic insects are attracted to the panels for oviposition, and the eggs and larvae die on the surface over hours. It’s unclear how sentient they are at that stage, but the process doesn’t change the NPP of the area, so the population size will barely decrease. The white-grid panel design tweak that breaks the polarized-light trap is probably good. So: panel design good-ish, turbine curtailment bad.

Banning neonicotinoid seed treatments. The WAWID shallow report on reducing neonicotinoid seed treatments (NSTs) is unusually direct: the author concludes that the net welfare effect of NST bans “may be negative.” The reasoning lines up neatly with part 2’s pesticide framework. NSTs are systemic and persistent; they make plant NPP physiologically inaccessible to phytophagous insects, which is genuine carrying-capacity reduction rather than constant-NPP killing. Counterfactual deaths for the insects that would otherwise have eaten treated plants are dominated by parasitoids, which the shallow report describes as “possibly more severe and longer than death by acute NST exposure.” The shallow report even cites evidence that sublethal NSTs function as fertility control. NST bans typically replace NSTs with foliar pyrethroid or diamide sprays, both of which are worse on each of these dimensions. Yet conservation NGOs and several US states are converging on NST bans as a “win for invertebrates.” On this analysis, that convergence is wrong.

There’s a pattern across the three. Conservation-default and welfare-default intuitions diverge cleanly from asymmetric antifrustrationist conclusions, and the divergence isn’t accidental: conservation values aggregate populations and biodiversity, asymmetric antifrustrationism values reduced expected suffering, and these come apart whenever you can choose between “more healthier organisms” and “fewer organisms full stop.”

It doesn’t have to be that way. No one is just a “conservationist” and most people have empathy, especially people who go into ecology. It’s just a matter of sensitizing them to the fact that conservation and welfare are not or rarely correlated.

There is a framing around autonomy that seems internally consistent to me: Imagine a parkour athlete has a bad fall next to you. They’re still breathing but unconscious and clearly badly injured. Some of my friends would argue that they should not call an ambulance or interfere in any other way because that would impinge on the athlete’s autonomy, and being unconscious, they cannot consent. But I think this is a niche view, and to apply it to some species (nonhumans) but not others (humans) would be speciesist.

Summary of Recommendations

Ranked by the joint score on importance, neglectedness, tractability, asymmetry-fit, and robustness to nematode-welfare-range uncertainty:

1. Fund welfare-biology research. Soil-animal welfare research at WAI and RP, Eckerström Liedholm’s insecticide research agenda, and field-building infrastructure. Highest cost-effectiveness for me, robust to direction of effect, and the entire portfolio below depends on it. If you can fund only one bucket, this one.

2. Fund GiveWell top charities or CEARCH’s HIPF. Direct human benefit alone justifies the donation; the soil-animal benefit is a large addition if (probably when) the welfare sign comes out negative.

3. Targeted birth-prevention. New World Screwworm continental suppression (the surprise winner from the WAWID and probably the strongest single intervention available right now), gene drive A. gambiae suppression, mating disruption for rice yellow stem borer and spongy moth Stop the Spread. These are narrower than the broad-spectrum levers above but directly target species whose lives we have unusually strong reasons to think are net negative.

4. Don’t actively oppose biofuel mandates, wind turbines, or neonicotinoid seed treatments on conservation grounds. Each of these is doing antifrustrationist work as a side effect of mainstream policy. The cost in human welfare is bounded; the welfare benefit to wild invertebrates is enormous.

In part 4, I’ll take a detour into the philosophy of mind to ask what invertebrate suffering actually feels like – an answer that draws on the phenomenology of psychopathy and Buddhist insight practice, and that may reshape how we weight nematode versus arthropod welfare.

• Online discourse about NPD collapses into non-mentalizing. Public conversation tends to regress to teleological reasoning – judging people by their behavior rather than their internal experience – and as soon as one participant operates that way, the rest follow. This isn’t a constitutional incapacity; it’s a state-dependent collapse that anyone can fall into under emotional arousal.

• Grandiosity and vulnerability are two facets of one construct, not opposites. When you actually examine the factor structure of the Pathological Narcissism Inventory, the two correlate strongly rather than oppose each other. Pathological narcissism needs both to be a coherent clinical construct: pure grandiosity isn’t a disorder, and pure collapse looks like almost any severe mental illness.

• Trait narcissism is not pathological narcissism. The Big Five traits sometimes associated with “narcissism” (high extroversion, low agreeableness) are heritable but aren’t a disorder, and the DSM’s trait-style framing produces what amounts to “asshole disorder” research. Plus, the same Swedish twin sample yielded a 79 percent heritability estimate when NPD was measured categorically, but only 24 percent when measured dimensionally.

• Most narcissistic presentations show disorganized, not dismissive, attachment. Mark’s dissertation found that disorganized attachment – intense ambivalence about closeness, being seen, and being a self – fits better than pure attachment avoidance for most of the people he sees.

• Two NPD clusters map onto two attachment dimensions. A perfectionistic, externally validated presentation tends to be unstable and high in attachment peroccupation, because the standard of worth is external. A control-oriented, anti-helplessness presentation is more stable and high in attachment avoidance, because it relies on a self-sufficient internal model. See also Narcissism, Echoism, and Sovereignism: A 4-D Model of Personality.

• Depressive personality is organized around loss; narcissism around deficit. In depressive presentations, something was had and then lost – either introjected as self-attack (Freud’s melancholia) or pined for externally. In narcissistic presentations, something was never gotten in the first place, and a compensatory structure forms around the gap. The narcissistic injury is sustained earlier than the depressive injury.

• Idealizing transference shouldn’t be punctured prematurely. Idealization is a developmental fantasy that’s serving a function – like a four-year-old’s belief in the Tooth Fairy. The clinician’s job isn’t to live up to it (that’s impossible) but to refrain from interpretations that disrupt it before the patient has the ego capacities to tolerate reality.

• A successful treatment is one that simply doesn’t fall apart. Mark’s working title for his clinical book is Staying in the Room. Ruptures most often happen when the patient’s internal pressure to perform becomes pressure on the therapist to perform, and the frustration outpaces the dyad’s capacity to metabolize it before the patient leaves.

• Diagnoses are almost always mixed. If there’s one personality disorder there are usually several, and deciding which is “core” versus “stabilizing around the core” is a judgment call. The diagnostic constructs themselves are blunt instruments for what is essentially the most complex system in the known universe.

• Sadism functions to expel intolerable parts of the self. What looks like “joy” in extreme online punishment fantasies is often projection: the bad object holds the disowned weakness, vulnerability, or sexuality, and is then righteously punished. The collective sadistic stance toward “the narcissist” online works the same way. See also The Sadism Spectrum and How to Access It.

• Two etiological paths into NPD. An alien, impinging introject from a parent tends to produce more internal distress and a disorganized, dysregulated collapse. Pure neglect, where the child is effectively raised by the environment and finds validation through achievement, tends to collapse in a more depressive direction.

• The lever for self-help is mentalization, not exercises. In place of homework, Mark recommends mindfulness or Vipassana-style practice that lets a person sit with the experiences typically projected or externalized – the ACT idea of self-as-context: I am the space in which valuations like “bad” and “good” occur, not the verdict itself.

Transcript

Introduction and Mark’s projects

Dawn: Hello, Flitterific listeners! I’m joined today by Dr. Mark Ettensohn, who runs the wonderful YouTube channel HealNPD. Please check it out and subscribe – with the notification bell – to both of our channels. Mark specializes in narcissistic adaptations, which of course are the best adaptations – well, depending on the environment. He has also written a book that’s visible in the background, Unmasking Narcissism – the one in yellow and black. That one is highly recommended too. Let’s jump into it. Do you have any new projects that you’re working on – any new books that are perhaps coming out?

Mark: I do, yeah. I don’t want to over-promise, but I’m currently in the last semester of the academic year. In addition to my clinical practice, since 2018 I’ve been in academia. I was an administrator helping to start what is now an American Psychological Association (APA) accredited program, and at some point I transitioned to a faculty member, so now I’m an associate professor in that program. But I’m stepping away – this is going to be my last semester as a professor. Although I really love teaching, and I find it immensely meaningful, and I really enjoy having an impact on developing clinicians, it just takes so much time. As my channel has grown, I’ve been having a hard time balancing those demands, so I’ve chosen to say goodbye to academia and reinvest that time into publishing and making more content for my channel.

As part of that, in May I’m going to take a week off and focus on finishing one of the writing projects I have in development. This one is a curated collection of clinical essays. Some of them have appeared on my channel, but in spoken form, so this will be reworking them so they’re appropriate for written format, adding things, and so on. The tentative title is The Villain We Invented. It’s going to be about narcissistic personality disorder (NPD) stigma, the “narc abuse” landscape online, the concept of popular narcissism, and the ways that all of that has been twisted up into this mangled construct that we’ve all probably encountered quite a bit online. It comes out soon – self-published, at least that’s the plan. We’ll see how that goes.

Dawn: I’m working on a self-help book for people with NPD, and I think you also mentioned something of the sort, so I’d be quite curious about your take on that. A lot of my friends with NPD want to be very self-determined, very independent, so it’s easiest for them, at least early in their recovery, to work on things by themselves – having a workbook to guide them could be helpful.

Mark: I agree. I have five or six projects in development – I’m a little scattered in that way. I’ve got the book I told you about, a clinical book focused on the therapy process working with pathological narcissism, another book focused on grandiosity, a workbook that’s half-finished, and some other stuff too. The California Psychological Association Convention is coming up, and I’ve submitted a proposal to introduce a new taxonomy of narcissistic self-states I’ve been working on. We’ll see if any of that ends up actualizing, or if it just stays in virtual development hell on my computer.

Dawn: How cool! I’m also working on an AI-based app for mentalization-based treatment (MBT): stillwater.coach hasn’t launched yet, but hopefully will in the next week or so – it’s still in the relatively early stages of development.

Mark: Are you technologically oriented in that way? You must be, if you’re developing an app.

Dawn: Yeah, exactly. I’ve been a software engineer for ages – professionally since 2010 – and more recently I’ve reoriented a bit in the direction of tech entrepreneurship, mostly because of AI and software engineering getting automated now. This was one of the things I felt very passionate about, because of all my involvement with personality disorders, and I’m hoping it will take off. There’s a niche there: currently there aren’t really any apps that focus on trauma-based adaptations like personality disorders. Some try to do EMDR for people with post-traumatic stress disorder (PTSD), so it goes a bit in that direction, but nothing that quite fits that mold. In particular, MBT seems sufficiently protocolized that I think I can teach an AI to do it, and also observe how the AI is doing and make sure it’s doing the right thing over time.

Mark: Which may be more challenging than it seems.

Dawn: Yeah, but if it’s totally open-ended psychodynamic therapy, I never quite know whether the AI is doing something that makes sense, whereas if it’s supposed to follow the MBT protocol, I have a better handle on that, and can review some chats – privacy permitting – and make adjustments.

AI assistants and online discourse on NPD

Mark: I have an AI chatbot that I trained on my own material. It’s a membership-tier offering on my channel, and it’s interesting – even with a thousand different safeguards to keep it from doing therapy or doing something that looks like therapy, I’ll still review the chat logs, and every once in a while it’ll pop over into, like, robot therapist mode, and then I have to figure out why it did that and how to keep it from doing that. From my perspective, that’s potentially a liability, especially as a licensed clinician. I don’t want to host a chatbot that’s trying to do therapy, for a number of different reasons.

Dawn: Yeah, but the question answering is super useful.

Mark: That’s what it’s intended for, and it’s got disclaimers all over the place: this is not intended to be therapy, this is not clinical advice, if you’re in crisis, call this number, and so on.

Dawn: I’ll probably also need to copy some of those crisis numbers – that’s something I consider relevant.

Mark: I just feel it’s important to have a number of safeguards in place, because most of the time the chatbots produce a reasonable response, but every once in a while they’ll pull in something that’s an outlier response and present it as though it isn’t. I worry that folks who are in a very vulnerable moment might be led in the wrong direction.

Dawn: That makes a lot of sense. I’ll probably also need to make sure the temperature is set really low, and run a lot of tests with one AI against another, so I have some protocols I can review of how well my AI is doing.

Mark: Sounds very complicated.

Dawn: On your channel, you’re not focusing only on the therapeutic or academic sides – you’re also engaging with what’s going on in popular culture. One thing I’m seeing in particular is what’s called teleological non-mentalizing, which is quite typical of NPD: people see some behavior and jump to a conclusion about what must have precipitated it, while ignoring the actual complexity of the underlying motivations and attitudes of the person who shows that behavior. If someone like Vaknin, who has NPD, does this on a YouTube channel, then it’s easily explainable for me. But if a thousand YouTubers do it – and this is what I’m seeing a lot – when they talk about people with NPD, they’re viewing them as if they were automatons of sorts, without any internal experience. They basically don’t worry about what precipitates these experiences, what’s actually going on inside of them. I’m confused what’s going on there, and why they’re doing that. If you encounter that, do you have any particular approaches or examples that help you explain to people the difference between teleological non-mentalizing and actually empathizing?

Mark: When I think about teleological non-mentalizing – reasoning about motivation based on the outcome of a behavior, right? “A person was mean, therefore they meant to be mean,” or, “I feel offended, therefore you meant to offend me” – I think about developmentally less mature ways of conceptualizing relationality. Little kids don’t necessarily even have a theory of mind, and even if they do, it’s inconsistently applied. There’s a tendency to just fill in the blanks about why something happened, based on how it landed.

In cultural discourse – popular conversations on social media and that sort of thing – I think people tend to regress to the lowest common denominator when it comes to the quality and complexity of the conversation. The nature of the beast is that as soon as somebody shows up who’s operating in that mode, everybody else tends to regress to that mode of operating.

When we’re talking about a group, I don’t think there is usually a capacity, or even a motivation, to treat people like complex, three-dimensional human beings with an internal world of their own. Especially online, where it’s all about how you can score in this argument or this debate, how you can most effectively shut down the other person, or walk away feeling like your worldview has been confirmed. So I don’t really think it’s primarily a problem of capacity – I think it’s partly a problem of motivation and context.

I see it on my channel a lot, although less so nowadays. Word has kind of gotten around that my channel isn’t really a forum that welcomes that sort of collapse of psychology into black-and-white categories of good and evil, narcissist and victim, or whatever. But whenever I venture out to take a look at another channel, especially the more popular ones – God, if I look at the comment section, it’s just chaos. Just projection, name-calling, scapegoating, a kind of piling-on of malice onto this fantasy boogeyman.

That kind of non-mentalizing also shows up in personality disorders and other forms of severe mental illness. I don’t think it’s a constitutional issue – I don’t think folks with those issues are incapable of mentalizing. I think it’s that it’s difficult to consistently mentalize, especially in states of high emotional arousal or distress. There are self-states that can be activated where those capacities are just less developed.

Clinically, part of the challenge is to withstand the accusation, because those self-states typically collapse into a do-or-done-to, victim–perpetrator dynamic, where it’s, “you did this to me.” That’s how it can show up in session, too. I make the wrong interpretation at the wrong time, I say the wrong thing, I don’t respond in a way that feels in alignment with what the patient wanted in that particular moment, and it can set off a rupture, where now I’m the bad object who did the bad thing to the patient.

Negotiating those moments – which are inevitable – is really where the work is. There’s so much of a psychotherapy process that’s basically just sitting around and talking about this and that. Okay, this thing happened at work, this thing happened at home. There might be some emotional arousal happening, but it doesn’t feel like the stakes are very high. And then all of a sudden, out of nowhere it seems, comes this moment where some part of the patient that holds a lot of trauma is activated right there, right then, in the dyad. Now the heat is on, and these are the moments where I think change can actually consolidate in a more substantial way.

Idealization, devaluation, and staying in the room

Dawn: This is a good segue to another question that was on my mind yesterday. A while back – this might have been in the handbook for mentalization-based treatment for pathological narcissism – I read the advice that when patients idealize you early in therapy, it’s important to try to moderate that to some extent, so that the devaluation doesn’t lead the patient to drop out afterwards. Basically, to stop the undulations from being quite so intense. Now I’m reading in Eleanor Greenberg’s book that she actually advises trying to live up to the idealization, trying to give the person that idealized therapist model when they want it – which seems to be the opposite advice. Where do you fall on that?

Mark: I’d be curious where she’s coming from. Do you remember the rationalization for trying to live up to or deliver on the idealization?

Dawn: Patients want reparenting from an ideally supportive parent, and she wants to provide that for them for as long as they need it, then wait for them, by themselves, to decide at some point that they don’t need it anymore – that they can recognize her with her flaws.

Mark: I see. I think it’s important as a clinician to recognize that you can never actually live up to or deliver on whatever the idealization is. It’s by definition a distortion of reality.

That said, in a Kohutian framework – a self-psychology framework – there is an idea about not disrupting an idealizing or mirroring transference, partly because the patient might not yet be developmentally ready for that. It might be traumatically disruptive to a developmental process that’s trying to get some traction in the therapy.

But I wouldn’t put it in terms of trying to deliver on the idealized transference. I’d put it more in terms of trying not to prematurely disrupt that transference with an interpretation that undermines it.

It’s like little kids who believe in Santa Claus and the Tooth Fairy. Different people have different ideas about it, but broadly, the idea is that it’s fun and sort of magical to be suspended in that fantasy – this fantasy where there are these benign, benevolent beings that come and shower you with gifts. I don’t know if it serves a legitimate developmental purpose to have that fantasy, but I think most parents have the sense that you don’t want to tell your four-year-old, “You know what? None of it’s real.” That would be traumatic in a way – it would be the constraints of reality intruding on a fantasy that’s serving a function. Now the four-year-old has to deal with the fact that there’s marketing, and what is that, and are parents lying to their children, and if so, why, and does that mean I don’t have to listen to my parents anymore? The world becomes less magical.

That’s true for all of us as we age and grow – the world gets less magical and reality becomes more and more the rule, and we have to deal with that. But ideally, we deal with it as we develop the ego capacities to do so.

In self-psychology terms – and maybe this is what Dr. Greenberg is talking about – if you see an idealizing transference, you don’t want to take it and turn it upside down, because it is serving a purpose. It may ultimately be a maladaptive one in the person’s life, but from a developmental perspective, it is maladaptive because they have not yet matured out of using it. The reason they haven’t matured out of using it is because they haven’t gotten enough of what they need to let it go and embrace the constraints of reality – of their own limitations and inadequacies.

Dawn: On the flip side, there’s the risk of exacerbating the disappointment when the client finds out that the therapist is fallible after all. How often does this happen to you? At what rate do patients drop out of therapy because of devaluation, prematurely?

Mark: I’m sure there’s a dropout rate, but I haven’t quantified it. The clinical book I’m working on – the tentative title is Staying in the Room, but the title I was tossing around before that was Three Failures and One Success – is about all the different ways the therapy relationship can just blow up.

You’re working with a personality configuration that is exquisitely sensitive to feeling devalued or criticized, where there can be an awful lot of interpersonal pressure that the patient puts on themselves – pressure to live up to the ego ideal, or whatever grandiosity might be at work. But that pressure is also extended to other people: pressure to live up to the idealizing expectations, to deliver on certain gratifications or entitlements that the person might have but isn’t necessarily consciously aware of.

There can be a real deficit in negotiating feelings like disappointment or frustration in a way that preserves a relational connection. Instead, the person dumps all of their bad feelings onto the other person and then throws them away – and in fantasy, this house is purged, this house is cleansed. Of course, over time that becomes a cyclical pattern, and there’s a string of broken relationships left in the individual’s wake.

I’ve come to feel that a successful treatment – and there are lots of different ways we could define success – is one that just doesn’t fall apart. We were able to stay engaged and connected long enough that those developmental deficits began to heal themselves, and the person began to develop the parts they need to cope more effectively, to sustain a realistic and stable self-image more consistently.

So yes, I’ve had it happen a number of times. There are a couple of forms it takes. One is an idealizing transference that collapses: I say or do something that undermines the idealization too soon, too early, and the person just – I go from good object to bad object, and they’re gone.

The other thing that can happen is when that pressure to perform that the patient feels – their own internal pressure – becomes pressure in the treatment for me to perform. I try to slow that down and push back on it a little, but there’s this critical zone where the frustration is rising and I’m trying to metabolize it, help them metabolize it, to realize that this is pressure that has always been there, that has never actually produced the feeling of happiness or well-being the person is seeking. They’re asking me to live up to some kind of idealized image that is impossible for me to live up to, so effectively the two of us are doomed as long as this frame holds – this frame of, “give me… why aren’t you giving me what I need?” If we can’t figure out a way to metabolize that quickly enough, then the frustration rises above that optimal threshold, and the person leaves.

Dawn: Is the mechanism behind that something like: they make you part of their identity, so their own expectations on themselves become expectations on you, and then you suddenly have the same pressure they have?

Mark: That’s one of the things that happens. There can also be some developmentally less mature expectations about the world, and perhaps a lower frustration tolerance around those issues. They might be great at tolerating frustration in lower-stakes settings, but when we’re talking about core needs that are beginning to present themselves in the treatment, it becomes this dance.

I can start to see what’s happening, but if I just try to say it, they might not be ready yet to have it verbalized, and it just won’t make sense to them. So I can’t say it – but I also can’t live up to the expectation. Then it becomes about reframing to process: “I’m noticing that frustration showing up again. Can you tell me how it’s showing up? Can you tell me how it feels?” We start to explore the fantasies surrounding the frustration, to play together with it instead of feeling like it’s a whip we’re laboring under.

Sometimes that works, and sometimes it doesn’t. Sometimes there’s just too much pressure. Often a third variable is the person’s partner. If they have a partner who’s saying, “you need to go get better,” the partner might start to get frustrated: “Okay, well, it’s been six months of therapy, and I’m still noticing all the things.” And it’s like – yes, and this is a lifelong issue. It’s going to take… honestly, the first year of treatment we’re often just getting a lay of the land. This is a process that takes a significant period of time. It’s not that nothing happens during those early parts of treatment – it’s that we’re setting the stage for all the changes that are going to happen.

But people get frustrated, especially in a cultural mindset where it’s, “you’re the doctor, you fix it.” And I’m like, yeah, but I’m not that kind of doctor. Fixing it is a much more complicated proposition than it seems.

Dawn: The reason this is all so dear to me is that I’ve basically lost a friend that way. There was the idealization cycle – I didn’t recognize it – and then the sudden devaluation, which I started to recognize because I was reading up on things at that point, but it was too late. This friend had a very interesting presentation: very obviously emotionally dysregulated, borderline-style, but she also used defenses that would be more typical of pathological narcissism. She was very concerned with self-esteem and status, had a lot of shame, was very sensitive to that, and was thinking very hierarchically – a lot of her coping was related to hierarchical thinking. I was wondering: what is going on there? I’d call it a mixed presentation of borderline personality disorder (BPD) and NPD.

Mixed presentations, echoism, and the structure of pathological narcissism

Dawn: When you encounter a presentation like that, would you treat the patient as someone with pathological narcissism, or rather as someone with a mixed presentation – or as someone with borderline, and maybe refer them out to someone who specializes in borderline?

Mark: I tend to treat all presentations as mixed presentations. I tend not to conceptualize in terms of just one category of disorder. I do a lot of diagnostic assessments, and those involve a kind of exploration of all the different personality styles or prototypes that might be involved. Then there comes a point where I need to provide a diagnosis, and that is far less exact than one might imagine.

Looking at the different personality profiles and their scores – if there’s one, there are always more than one. It’s never been the case that there’s just been one peak personality disorder and everything else was below threshold. If there’s one, there are many.

So the question becomes: how do I conceptualize this in a way that gets closest to what I think the subjective experience of the person is, but also captures the function of these adaptations? One usually seems to be more core, and the others seem to provide a stabilizing function around that core.

Not too long ago I assessed somebody, and it was like you’re describing – there’s a BPD piece, and there’s a narcissism piece. The door could have swung either way: I could have said, “this is NPD with some BPD-like dysregulation happening,” or, “this is BPD with NPD-like defenses that have developed.” That’s the direction I went, because it seemed to me that the individual was more fundamentally oriented around abandonment anxiety and abandonment trauma, and that this false-self adaptation had developed around it to stabilize it – as opposed to a fundamental narcissistic injury where, because of their experiences, things had gotten dysregulated in a BPD-esque way.

But honestly, that’s a judgment call. It’s highly individual, and in a lot of these situations a plausible argument could be made either way. It speaks to the complexity of the variables and the inexactness of the constructs. To quote Hannibal Lecter, we’ve got these blunt little instruments that we use to try to understand what is essentially the most complex system in the known universe – just one individual mind. So yes, it’s necessarily inexact, where things are right now.

Dawn: This is also very interesting because I have a thesis about what’s called echoism. Craig Malkin has a model of narcissism and echoism at opposite poles of a spectrum. I liked the book Rethinking Narcissism – lots of things in it made a lot of sense to me – but the spectrum didn’t make sense to me, because I know so many people who are really high on both narcissism and echoism, and also something else I call sovereignism, which is probably relatively close to malignant narcissism. But let’s bracket that.

If these things are so structurally similar, and the same person, depending on their social context and who they’re interacting with, switches between echoism and narcissism in their expression, then what’s going on there? To me, it seemed more like people with disorganized attachment.

One model is: people with disorganized attachment, depending on the other person or the social context, either try to collapse that into avoidant attachment, which looks more narcissistic, or try to collapse it into preoccupied attachment, which looks more echoistic.

The other idea is that it might depend on where they feel they are in the social hierarchy. If they’re very low, they take a people-pleasing approach to try to appeal to all the people higher up, which is most of them. If they find themselves very high, they can take a dismissive approach to the people below them and the people-pleasing approach only with the very few who are above them. In the first case they come off as more echoistic, in the second case as more typically narcissistic. What do you think of this seeming dichotomy of two relatively similar things?

Mark: Well, I think of the dichotomous model of pathological narcissism itself. It loses its coherence if you have just a grandiosity factor, or just a vulnerable, depressive, anxious factor – you need both to have the coherent construct of pathological narcissism. If you just have grandiosity, then someone is just grandiose – there’s no disorder there, necessarily. And if you just have the collapse, then that looks like almost any mental health issue in its more severe form. It is the unique combination that makes a usable, clinically relevant construct.

It would seem that these two constructs are strictly dichotomous, or, to use a research term, orthogonal – one measures one direction, the other measures the other direction. But when you actually look at the factor structure of instruments like the Pathological Narcissism Inventory (PNI), which includes both grandiose and vulnerable factors, what you find is that these two constructs intercorrelate in a very meaningful way. It’s not, “the more grandiosity you have, the less vulnerability you have.” No – these two constructs correlate with each other about as far as they can without starting to superimpose. The more grandiosity you have in a pathological-narcissism context, the more vulnerability tends to go up – and vice versa. They are two facets of the same construct.

To the extent that what Malkin is talking about with echoism is something more defined by a kind of preoccupied attachment, or a preoccupation with being too visible or too “out there,” it would seem to be the opposite of grandiosity. I don’t know about that. I think these are two facets of a conflict around being seen, and around what it is to be a self – the subjective building blocks of self-experience.

In NPD, the person is actually deeply conflicted about being visible, about showing up. There are often really inflexible contingencies around when it’s okay and when it isn’t. The same individual, if they feel they’re getting really positive feedback from a group, can launch into untethered grandiosity, and then maybe the next day have the opposite experience and collapse under the shame hangover about how much of themselves got seen, and how conflicted they actually feel about that in the after-party self-state.

I don’t know if I’m answering your question, but I find that in narcissistic presentations – this was my dissertation, actually – yes, there are some that are strictly dismissing-avoidant in their attachment style, but most seem to correlate with something more like fearful avoidance, which would be the disorganized attachment, where there’s intense ambivalence about closeness, about being seen, about being a self. There’s a desire to affiliate, but the stakes can feel really high when that actually happens, and it rarely lives up to all the different internal expectations the different parts of the person might have.

Dawn: There is a bit of a connection between the vulnerable expression and echoism. But the Maladaptive Covert Narcissism Scale has items that go in the direction of, “I have enough problems of my own, I don’t care about other people’s problems,” which are probably what the scale is meant to capture as expressions of the vulnerable side – but they’re basically the opposite of echoism. With echoism, it would be: “Yes, I’ll happily make myself useful to help you with all of your problems, because that’s where I derive my self-esteem from. If I can make myself useful for someone, that’s what builds me up.”

Mark: What’s interesting about that, though, is that from a Pathological Narcissism Inventory lens, that’s considered a form of grandiosity – self-sacrificing self-enhancement, on the grandiose factor. Yet if we looked at it through a different framework, that would be considered something more like an oral personality, where the person is giving of themselves to the point of depletion and collapse. There’s some meaningful component of that which is an old conflict, an old place of developmental impasse.

Different words can apply equally well to different constructs. The word depression or depressive is almost ubiquitous in psychology, and it’s like – what do we mean? Do you just mean feeling sad, or something more like anhedonia, the absence of pleasure? Or can we understand it through a pathological-narcissism lens, because of all the distortions of self-esteem that tend to accompany depression? Is it introjective or anaclitic depression? There are so many variables. Same thing with narcissism – I just wrote a piece on my Substack about how there are at least four different constructs that all have the same name. Everybody’s using them interchangeably, and nobody knows what anybody else is talking about, but everybody thinks we’re all talking about the same thing.

Dawn: You mentioned a study that found that, up to a 3 on the scale, narcissistic grandiosity and vulnerability are uncorrelated, and then they become correlated suddenly once the grandiosity becomes high enough. That study was based on the Five-Factor Narcissism Inventory short form (FFNI-SF), whereas the scale you mentioned several times now is the PNI. Which of these do you prefer for which purpose?

Mark: I don’t use the Five-Factor scale. I don’t subscribe to trait narcissism as a meaningful clinical construct, at least not in my work. I can see how it’s a necessary component of personality research, but I think it’s easy to misinterpret trait-narcissism research.

My main critique of the Diagnostic and Statistical Manual (DSM) model is that it is essentially a trait framework. It treats grandiosity as a trait, and as a result you end up with these nonsense research studies that are essentially saying NPD is not in any way meaningfully correlated with mental disorder – that it’s not a form of psychological dysfunction. If you have dysfunction, you’d expect there to be distress, you’d expect there to be problems in the person’s life, beyond just that people in their life don’t like them. That’s not, to my mind, a meaningful enough standard to call something a mental disorder. But that’s how you end up with the asshole disorder, which is how NPD is treated in most settings, because the DSM model collapses – if you do a factor analysis of it, it’s essentially one factor: grandiosity.

Grandiosity, if you treat it that way as a trait, is probably heritable. That’s where you get these high heritability estimates that use the DSM model. Trait narcissism, on the Big Five, is defined as high extroversion and low agreeableness. And what is that? I don’t know – it’s not a disorder.

This is yet another way in which we seem to be talking about different things using the same words. When I talk about narcissism – pathological narcissism, and there’s a reason we have to attach the word “pathological” to it – it’s because narcissism itself, as a construct in psychology, is neutral with respect to health or disorder. Narcissism is just how the person relates to themself. You can relate to yourself well, in an adaptive way, and we’d call that healthy narcissism. Or you can relate to yourself in a pathological way, and we’d call that pathological narcissism. The fact that we need to attach the word “pathological” to it says that the construct itself is neutral.

So when you talk about trait narcissism, is this a disorder? Is this a construct that can meaningfully be used to understand disordered states? To my mind, not so much. So I don’t tend to use instruments derived from that model. When I feel that model is imposing itself on our ability to understand this disorder, I tend to push back. Let’s define our terms.

I debated Peter Salerno on YouTube a number of months back. I didn’t really have my thoughts straight about this yet – I wasn’t that familiar with his framework. But he’s written prolifically about the supposed high heritability of narcissism. If we’re talking about it as a collection of traits, absolutely – highly heritable. But that’s not the same thing as a disorder. Once we start thinking about it as a form of pathology, all of a sudden there’s identity disturbance. Where does that come from? Is that genetic? Probably not – identity disturbance tends to be more experientially grounded.

Dawn: This is certainly a very important message to get out there. When I look at the FFNI-SF, I see a whole bunch of subscales, and among them a grandiosity and a vulnerability subscale. The PNI has nice validation studies where they tested how well it captures all sorts of facets of mental health, and how much better it does that than the Narcissistic Personality Inventory (NPI). I haven’t seen something like that for the FFNI, but at least from the subscale structure, it seems to capture both the vulnerable side and the grandiose side, so it already seems to be doing a better job than the NPI.

Mark: Well, that’s good. The way the DSM is going with their alternative framework for personality disorders – that’s a combination of a more dimensional approach (degrees of mental illness, essentially) and a trait-based approach. For NPD, the alternative-framework trait is that the person is admiration-seeking, essentially. So it’s this trait model in combination with a dimensional approach.

Any measure that includes something like a severity scale – which in the instrument you’re talking about, which I’ll admit I’m not that familiar with, is likely what they’re talking about with the vulnerability subsection – I think you already have a more robust instrument there, because it includes a component of disorder and dysfunction. As opposed to something like just implausibly high self-esteem – is that a disorder? That’s where things like the NPI go off track.

Depressive vs. narcissistic injury and the two faces of NPD

Dawn: I think we can agree the NPI is not a good model for pathological narcissism. You mentioned earlier introjective and anaclitic depression, so that’s a good segue, because there’s something I find very confusing about the Psychodynamic Diagnostic Manual (PDM) – now in its third version since last year – and also about Psychoanalytic Diagnosis by Nancy McWilliams. There’s a separation where she has the narcissistic style as something very, I don’t know, anaclitic – something that’s empty, free of introjects, or something. And then she has all sorts of other things that are sort of similar – the depressive style, the hypomanic style, the masochistic style – completely separate from that. The depressive style, for example, comes in an introjective version.

What I’m seeing a lot of the time is that, in a way, the folks with NPD I’m friends with are very introjective – they just have introjects that really suck, sort of along the lines of your “harsh introjects” video. They have introjects that tell them they must never show any kind of vulnerability, so as not to make themselves manipulable or exploitable to others. Or introjects that tell them they must perform and never sleep and be the best at everything. These count as introjects in my mind. I suppose there are some presentations of NPD that are pretty psychically empty, but when there’s a presentation – which is the more common one I see – that is introjective like that, do you have a handle on how Nancy McWilliams or other people who use the PDM would classify that? Would they classify it as NPD versus depressive personality disorder, or something else, when someone has such an introjective presentation of NPD?

Mark: Well, I don’t know how Nancy McWilliams would classify it. In one of the classes I teach, the students and I use the PDM and McWilliams’s Psychoanalytic Diagnosis, and one of the projects is group presentations around different personality styles and how you would do differential diagnosis.

So, analytically, the depressive personality style has a kind of direct line from Freud’s work Mourning and Melancholia. Interestingly, when Freud introduces the idea of melancholia, he describes it as a narcissistic object, essentially. The idea is that the individual sustains a loss, and in order to defend against the loss, the ego is split – this is the first introduction of the idea of splitting in the literature. The ego is split, and one piece of the ego is identified with the lost object. The lost object is introjected into the ego, and the ego begins to direct its own rage at the lost, abandoning object internally. That’s what produces the symptoms of melancholia, which was the precursor to the word “depression.” Anger turned inward, essentially. The individual is attacking themselves in lieu of the lost object, which remains idealized.

Freud discussed this as a form of narcissistic object because, in his original model, you’ve got primary narcissism, which is the infant’s omnipotence – the world is me, essentially. Then, once you go through the Oedipus complex, you learn that’s not true, and that sets the stage for the possibility of loving external objects – libido can be transferred externally. But Freud says that when the object is traumatically lost, that libido is pulled back into the self, and this represents a form of secondary narcissism.

Why am I saying all of this? I think what I’m trying to say is that, analytically, the idea of a depressive personality is one organized around loss. That loss can either be introjected in the way Freud is describing – the introjective style of depression, where the person is self-hating, essentially – or it can remain a loss that has been sustained out in the world, and the person is constantly grieving (well, not really grieving, because it’s not productive – but constantly pining for a kind of reacquisition of the lost object).

Whereas narcissism is something where there is a kind of loss, but it’s more like a deficit. It’s more like the person never got what was needed – not that it was gotten and then lost. There’s a compensatory internal structure built around this deficit, as opposed to a way of coping with something that was there and is now gone.

Dawn: So, if it used to be there and was then lost, that must have happened before the Oedipal stage – or at what age would those occur?

Mark: If we’re talking Freud, that would actually be post-Oedipal – it would be a regression. The libido that had been cathected to the internal object is now pulled back into the self.

Dawn: Okay, so at a relatively old age already – four years plus or so?

Mark: If we were to plot these injuries on a developmental continuum… and there might be people watching this who are like, “this guy has no idea what he’s talking about,” and that’s fair – I’m not an analyst, I’m an analytic therapist. I’ve read a bunch of this literature, but I’ve not been through formal analytic training. That said, I’m willing to be wrong. So if we put it on a developmental continuum, the narcissistic injury is sustained earlier than the depressive injury.

Dawn: Interesting – that’s also a useful distinction. Something else I wanted to pick your brain on: I just alluded to all these different kinds of introjects that friends of mine have. To me, it seems like those friends fall into two pretty different groups. Both camps have NPD, and some of them have a mixed presentation of both, but by and large there are some who are very perfectionistic, usually have some pro-social values, and when they make mistakes, they try to hide those mistakes from themselves – to reframe them so they’re not mistakes – and they struggle a lot with very obvious self-deceptions. They always strive to prove their worth to other people, to compete with certain nemeses who are sort of close to their level.

Then there’s a very different camp. These people are much more about never feeling helpless, trying to control everyone else, having maximal power and control over every situation. Their self-esteem is much more stable. The first group usually oscillates between vulnerable and grandiose states and seems really fragile and sensitive, whereas the second group usually has a really stable, high level of grandiosity. They don’t so much compete with any individual person, or seek the admiration of any individual person, because they don’t even respect people enough to seek their admiration a whole lot.

They usually don’t have pro-social values – they’d rather have a kind of false self that’s all around being perfectly selfish, sometimes being evil or something of that sort, and they are sometimes (or often) sadistic. When they make mistakes, they tend to reframe them as something they’ve done intentionally, to maintain this idea of having had control in that situation. Very much about control and selfishness.

These seem very different to me, but both have been diagnosed with NPD at different points. Which of these do you encounter more, and how do you make the distinction? Does it have something to do with attachment styles? Is it the particular false-self adaptation they form that pushes them in one direction or another?

Mark: Gosh, I don’t know. One thought I had: when you were talking about the stability of one group versus the other, that makes some sense, given that the group that is more defended around feeling helpless probably has more of something like an internal locus of control, even if it’s just a defensive one. That would likely result in a more consistent experience, as opposed to the variable one you notice in the group that is trying always to be the best, or live up to – because implicitly there, the standard is external. You’re only the best if everybody says you’re the best. That way of organizing self-experience would be inherently less stable, more reactive to circumstances – whether or not you got the trophy or the prize.

We could loosely map those onto attachment style, where the one that’s more about external validation is a little hungrier – along the dimension of attachment anxiety. “Will I be the best? Will I be accepted? Will I get the accolade? Am I good enough?” There’s an other that’s needed to provide that feedback. So it’s higher in attachment anxiety – something more along the lines of a preoccupied dimension.

The one that’s more about not being helpless – there really isn’t a lot of attachment anxiety there. There’s much more attachment avoidance. “Don’t get too close, don’t trust” – like you said, which I think was an insightful read. That would be more toward a dismissing-avoidant experience.

When you look at how those attachment dimensions… I’m not saying these are styles, necessarily – these are dimensions of attachment. I see attachment as more of a continuous variable than a categorical one. The higher you are in anxiety and the higher you are in avoidance, the more you move into these categorical dimensions of attachment style. But if we think about them more as continuous variables that are constantly fluctuating depending on the circumstance, that’s a better model for understanding a person’s behavior in the moment.

When you look at the correlations of various kinds of psychopathology with these different dimensions of attachment, you find that the preoccupied style is much more highly correlated with measures of distress, whereas the dismissing-avoidant style is hardly correlated with measures of distress at all, because the person has this very well-established internal, self-sufficient model of self: “I don’t need anybody, I’ve got it covered. Even if I needed your help, I wouldn’t accept it, because I can’t trust you, essentially.” That’s probably why you notice less symptomatic variation in that subset.

Dawn: That makes a lot of sense. I also noticed how much attachment is dependent on the state. One friend of mine very readily, all the time, goes into what I would call the vulnerable victim mode, or the grandiose victim mode – there are these aggressor-versus-victim roles you can take, and you can take them in both grandiose and vulnerable states. That friend very happily goes into the victim side. Another friend told me, “Oh my god, no, I would be extremely hesitant before I cast myself in any kind of victim role, regardless of whether I’m actually a victim in any given situation.” I also saw this in another case where a friend tried to recall instances in her past where she was actually in a victim role, and it was very difficult for her to acknowledge that in the first place. I asked her to do the Attachment Style Questionnaire short form, and it came out as extremely low preoccupied attachment, extremely high avoidant attachment. She was in a grandiose state at the time when she took the questionnaire, so it’s kind of expected. The same difference also shows itself in the attachment styles, as far as I can see.

Mark: Yeah. Attachment is a really useful model for understanding a lot of things about a person. I sometimes run into this with my students, where there’s a tendency to think about things only in terms of attachment. Yes, these underlying dimensions of attachment illuminate so much about how we are interpersonally, but we still need the other concepts as well. It’s one of those things where, when a new lens comes into vogue, everybody is using it. Internal Family Systems (IFS) is a good example. People are like, “Whoa, parts!” – and then everything is about parts.

Dawn: I also find that lens very useful. When it comes to attachment styles, they have exactly the same problem you mentioned earlier: they make so much more sense when you see them dimensionally. By definition, most of the population is around the 50th percentile on both axes, and that is generally considered secure attachment. But when I have a friend who is around the 50th or 60th percentile on both dimensions, that friend feels super different to me from a friend who’s around the 25th percentile or so. Even with insecure attachment, there’s a huge difference between people who are more on the insecure end of secure attachment and people who are more on the secure end of it.

Mark: Absolutely. It reminds me of that Substack piece I told you about, on the different ways we use the term narcissism. What it was really about is: how heritable is NPD? The same researcher used the same sample – a sample of twins, in Sweden, I think. The first time, they defined and measured NPD as a categorical variable. The question was, among these twins, how many of them have NPD, and they found this crazy high number – like 79 percent heritable.

Seven years later, using the same sample of twins, they measured NPD as a dimensional variable – like what we were just discussing, where you’ve got, say, avoidance and anxiety, and you rate it that way. The heritability estimate dropped to 24 percent. It’s crazy how different the results can be from how you frame the question. If you introduce a continuous variable of severity – which is how they looked at it the second time, “okay, you’ve got these traits, but how severe are they?” – suddenly the heritability estimate drops to 24 percent, as opposed to just saying, “do you have these traits, yes or no?”

Dawn: Yeah – when you have a scale like that, and you project it onto just above or below some threshold, a lot of information is lost. Taking a dimensional approach is probably always better.

Mark: I think so. That’s how I think about it.

Dawn: I also noticed that when people have disorganized attachment – but, for example, the preoccupied dimension is around the 75th percentile and the avoidant is a bit higher – just the difference between those two is very informative. People who have slightly higher avoidant than preoccupied attachment, even though both are very high, will default to this very control-oriented presentation, with the internal locus of control, in almost all cases.

Mark: For a lot of people, that is just a more stable developmental solution. You see kids do this all the time, where they’re like, “My parents split up. Why? Because I’m bad.”

Dawn: Are they proud of being bad?

Mark: No, but it provides a stabilizing explanation that’s within their control. “If I weren’t so bad, then the bad thing wouldn’t have happened” – as opposed to the much harder thing to metabolize, which is that sometimes bad things happen and there’s nothing you could have done, it has nothing to do with you.

Sadism, etiology, and self-recovery

Dawn: I was also wondering, in that context: fetishes develop super early in life. There are some people who already have whatever fetishes they have around age two or three. Personality disorders take their final shape around puberty, usually. To me, it seems plausible that fetishes that form very early have an influence on what personality disorder a person forms in the end. For example, when someone has a sadism fetish – forms that sadism fetish around age two or three – that leaves them maybe a decade or so until puberty, or a bit less than a decade in many cases, for that aspect of their personality to have an effect on the rest of their personality.

Some people probably try to push that away – “Oh my god, that can’t be me. I need to hide that from everyone, I need to be the exact opposite and totally overcompensate for it.” Other people probably resolve the cognitive dissonance the other way around – they endorse it and are like, “Okay, apparently I am evil. Let’s take all those movie villains as role models and see what evil people do.” Then they also have a more coherent self-experience, just in the opposite direction. I’m wondering: to what extent does sadism, very early in life, have an effect on the development of sovereignism or malignant narcissism?

Mark: That’s a good question. I’d be curious about the idea of fetishes developing that early, and how that’s understood, because – obviously, infantile sexuality (to use the old analytic idea) is a thing. But it’s not necessarily genital sexuality. For that to be a thing, outside of normative play, it’s usually a result of traumatic intrusion of adult genital sexuality into the child’s world. So I’m curious to understand better what you mean by the early establishment of fetishes.

Dawn: More objectively speaking, I recently listened to an interview on Psychology in Seattle, and someone argued that fetishes form so early in life because children are no longer exposed in the way they were in our ancient environment. At some point, there were not really rooms or anything of the sort, and so children early on saw regular sexual intercourse very early in life. Since that’s no longer the case in today’s society, they latch onto something else. For example, when they’re crawling around on the ground underneath some table and see a bunch of feet, they develop a foot fetish in some cases. That’s one possible way it can emerge. I don’t know where a sadism fetish comes from – perhaps television or something. But it seems pretty frequent.

Mark: When I think about fetishes, to my mind that is a category of experience organized around genital sexual expression. “I can get off, but only this way.” So I have a hard time connecting that to pre-latency, very early childhood experiences.

That said, I do think we all have key experiences around which significant aspects of our internal world organize. I wouldn’t call it a fetish necessarily, but maybe it becomes one. So with something like sadism: when I think about what sadism seems to be, it’s the punishment in someone else of intolerable experiences the sadist can’t deal with. Being weak, being vulnerable, being helpless. There’s a tendency, in a sadistic worldview, to identify somebody who appears to embody some of those intolerable qualities, and then punish them – getting a feeling of gratification from seeing, via projection, the bad object punished.

Dawn: That’s one framing. But I could also see the framing that, when you learn from early on that it’s bad – when you’re being taught by your parents that it’s bad to show vulnerability – you introject that. Just as I learned that certain things are moral and other things are immoral, like that if someone is sad I need to comfort that person, these people learn the opposite: being sad is bad. Then they just feel righteous about executing that kind of punishment on the person.

Mark: I don’t know that those are substantially different. You’re right that that is a moral framework many of us carry around inside of us anyway. You see it all the time online, especially around things like pedophilia or child molesters, where everybody’s just sadistic – “burn them alive, cut off all their skin, hang their eyeballs off the…” And it’s like, what? I get that this is morally abhorrent behavior, but where does the seeming joy that this commenter is deriving from this ultra-violent fantasy come from? Is this a sadistic person, or is this one of the few contexts in modern life where sadism is sanctioned?

I do think that part of the function of that communal sadism is to expel or purge something that is unthinkable or intolerable in the self. Not that everybody’s a pedophile, but there are aspects of human sexuality that we don’t like to entertain. “It’s not in me, it’s in them, and now we’re going to torture them” – it’s a way of confirming to the self, “Yes, I am righteous, and they are the bad one.” That same process plays out in all kinds of different ways, and we see that too with narcissism. The “evil narcissist with shark eyes” and all the rest – they’re demons, monsters from hell. What are we really punishing there? I don’t know – it’s different things to different people, but it’s clear there’s some kind of collective sadistic stance toward “the narcissist.”

Dawn: You see that a lot in prisons. There are sadistic inmates, and then there’s someone who’s a child molester coming to that prison, and that particular transgression is used as a kind of Schelling point – a signal that now it’s okay to indulge the sadism. Some of my sadistic friends are quite self-reflective about that – if they want to have fun with their sadism, they’re basically waiting for someone to transgress against them in a particular way to give them an excuse.

Mark: Sadism might just be something I’m particularly defended against. It is one category of human experience that I find difficult to empathize with. It’s not that I can’t sort of get it, but it’s just one spot where it’s hard to really get into the mind of somebody who’s in the throes of a sadistic experience.

Dawn: I have an article where I’ve tried to come up with exercises that teach a non-sadistic reader what it is like to experience any one of, I think, six different types of sadism, so perhaps that’s interesting.

Mark: That is interesting – I’d be interested in seeing it.

Dawn: Yeah, I’ll link it to you, and I’ll link it in the show notes – I love saying these things. I’m a new podcaster, so I love saying those things.

Mark: Perfect.

Dawn: We’re sort of talking about etiology a bit, and there I was also thinking: there’s the model of the etiology of NPD that’s along the lines of what the MBT people write about. There’s some kind of influence they call the alien self, and projections from parents, and then the child – who still tries to find themselves and doesn’t really have a self yet – has this alien self projected onto them. They try to become the person who lives up to those projected expectations, because they don’t yet know that that is not actually them. It’s a failure of contingent marked mirroring – not contingent, but very much marked – and the child forms a false self into, basically, people-pleasing the parents.

But I’ve also noticed that there’s an etiology of NPD that’s basically just neglect. The parent was either emotionally absent or completely absent, and people developed NPD somehow, because they were seemingly parented by their environment. They recognized that certain actions produce something a bit like a love surrogate, even if it wasn’t coming from their parent, because the parent wasn’t there. Somehow the environment takes over a kind of parenting role, which is also not contingent at all, because the environment also does not actually do any kind of mentalization with the child. Are those two different etiologies of NPD?

Mark: That’s a really good question. I think the higher levels of internal distress – and this is just an intuitive sense, I don’t know that I have a theoretical justification for it – likely come from that sort of alien, introjected, impinging phenomenon, where what is “not self” is imposed on self in a way that creates disorganization and a kind of internally punitive, introjective experience.

Versus something more like neglect, and then finding a kind of validation – because you’re good at baseball or something like that. That would probably be more prone to a depressive collapse, but not a dysregulated, disorganized kind of collapse, like you might see in something more like… well, if I think about the etiology of borderline personality disorder, I think there’s often the felt intrusive presence of some introject that the person is consistently trying to both expel and also experience a kind of merger with.

Dawn: That makes a lot of sense. I don’t have a good segue, but I’d also be really interested in seeing whether you have ideas for exercises. When I’m writing this self-help book for people who want to recover from NPD, a large problem is that they don’t automatically have a therapist there to train relational things, so they need to find another person – a partner, a friend, or an AI, perhaps – with whom they can train the relational things. Are there any kinds of exercises you’ve already had good experience with that your patients can do between sessions, or anything of the sort that can help generate some of those experiences?

Mark: That’s a good question, too. I don’t really give homework in my practice. To the extent I’ll recommend an exercise or something like that, it’s mostly around the idea of mindfulness meditation, or a Vipassana kind of meditative experience – one focused on being with internal experience without necessarily having to identify with it, or react to it, cling to it, or push it away.

Things like journaling can be helpful, of course, in articulating what a person is experiencing and holding up a kind of mirror to themselves. But I think the main challenge, especially early in treatment, is just being able to take ownership, in a sense, of these different kinds of experiences that are typically projected or externalized – to be able to sit with the idea that “I feel like I didn’t do a good job,” and just be like, “okay, that’s an experience that’s passing through,” to be able to acknowledge it and then let it go.

This is a kind of Acceptance and Commitment Therapy (ACT) model – there’s a contextual self versus something that’s, maybe, more immutable. Self-as-context versus a conceptual self that’s labeled “this is bad,” versus, no, I am actually the space where valuations like badness and goodness occur.

I realize I’m coming up against my hard stop here. These are great questions, I really appreciate them, and the chance to speak with you. I’m excited for your channel, and I can tell that you do some really deep and intricate thinking about these ideas, so I appreciate the level of sophistication you’re asking about.

Dawn: Thank you so much for sharing all your knowledge. I really love it when people with extensive experience in some field actually share that in-depth knowledge that you can’t just get from books, because that’s always so much more fine-grained and nuanced than the more compressed versions in books and papers.

Mark: I’m happy to do so. If you ever want to talk again, let me know.

Dawn: Yeah, thank you so much! Have a good day!

Mark: Yeah, you too.

Dawn: Bye!

This is part 2 of a five-part sequence on welfare ecology. Part 1 introduces the ethical premises. Part 3 covers interventions. Part 4 explores a model of invertebrate suffering. Part 5 covers AI.

If you decided in part 1 that you care – even a little – about the welfare of invertebrates, the next question is: Where are they, how many are there, and what drives their population sizes?

The Numbers

Terrestrial Soil Fauna

The foundational data comes from two landmark studies:

• Van den Hoogen et al. (2019) estimated 4.4 × 10²⁰ nematodes in the top 15 cm of soil globally. Stefan Geisen, the second author, clarified that this accounts for roughly 90% of all soil nematodes, putting the total at about 4.89 × 10²⁰ (Grilo 2025a).

• Rosenberg et al. (2023) estimated 10¹⁹ soil arthropods, of which roughly 95% are mites and springtails, with the remainder being ants, termites, and other groups.

For context: There are 8 billion humans and roughly 80 billion farmed land animals at any given time. Soil nematodes alone outnumber all farmed animals by a factor of 6 billion.

Marine Fauna

From Bar-On, Phillips, and Milo (2018):

• Approximately 10²¹ marine nematodes – about 2× the number of soil nematodes.

• Approximately 10²⁰ marine arthropods (copepods, krill, amphipods, etc.) – 10× the number of soil arthropods.

Marine nematodes are overwhelmingly benthic (living in seafloor sediment). Marine arthropods are split between benthic and pelagic (water column) habitats.

Density by Biome: The Terrestrial Landscape

Not all land is created equal. Rosenberg et al. (2023) provided mean densities of soil arthropods by biome. Grilo (2025a) combined these with van den Hoogen et al.’s nematode data to estimate total soil animal density. The results:

Several things jump out:

Forests are packed. Boreal and temperate forests have 7–8× the soil animal density of cropland. The mechanism is well understood: forests produce continuous, diverse leaf litter that feeds a complex soil food web. They have deep root networks supporting fungal communities that mites and springtails feed on. The soil is structured with pores and channels that provide habitat at every scale.

Cropland is relatively sparse. Tillage destroys soil structure. Monocultures reduce food web diversity. Pesticides directly kill a broad spectrum of fauna. The litter layer is minimal.

Pasture is sparser still. Grazing pressure, trampling, and simplified plant communities reduce below-ground complexity even further.

Deserts are comparatively empty. At 0.05 million per m², deserts have ~130× fewer soil animals than forests. Still even 50,000 animals per m² is impressive.

Volumetric Comparison: Soil vs. Sea

Since soil nematodes are measured in the top 15 cm (~0.15 m³ per m² of surface), we can convert to volumetric density and compare to marine environments.

One wrinkle: The dominant taxa change with the habitat. In soil, nematodes and arthropods (mites, springtails) coexist. In marine sediments, nematodes dominate overwhelmingly – they make up 80–90% of meiofauna (benthic organisms of 32–63 μm to 0.5–1 mm), with only small numbers of harpacticoid copepods alongside them. In the open water column, the situation reverses: There are essentially no pelagic nematodes (they are benthic animals, bound to sediment), and the invertebrate fauna is almost entirely arthropods – copepods, krill, and amphipods.

The table below uses “nematodes + arthropods” for all environments, but the composition varies as described:

The key insight. Per cubic meter, terrestrial soil and some coastal marine sediments are by far the densest habitats per cubic meter for potentially suffering animals, though coastal sediment occupies a much thinner layer. The open ocean water column is orders of magnitude sparser – even productive surface waters have roughly 10,000× fewer organisms per m³ than forest soil. But the ocean’s total volume is vast (~1.37 × 10¹⁸ m³), so low per-m³ densities still add up to enormous total populations.

The deep ocean – both water column and sediment – is strikingly empty by terrestrial standards. It represents an enormous volume of near-zero-fauna space that already exists naturally.

What Drives Population Size? The NPP Primer

Understanding why some biomes have more soil fauna than others requires understanding net primary productivity (NPP).

NPP is the rate at which plants convert solar energy into organic matter, net of their own respiration. It’s measured in grams of carbon (gC) per m² per year. It’s an objective, species-independent measure of how much new biological material is being produced. A tropical rainforest has an NPP of ~1,000–2,000 gC/m²/yr. A desert has ~0–90 gC/m²/yr. Open ocean averages ~125 gC/m²/yr.

NPP matters because it determines the total energy flowing into the food web. More NPP means more food for decomposers, which means more food for the nematodes and mites that eat the decomposers, and so on up the trophic levels.

Vasco Grilo’s analysis found that NPP-related variables (biome type, soil organic matter) are far more important than any other factor in predicting soil fauna density. The m² -years of agricultural land per food-kg explained essentially 100% of the variance in welfare effects for his preferred parameters (Grilo 2025c).

A common misconception. NPP is not subjective to a species. When a cow eats grass, the NPP doesn’t change – the grass still produced the same amount of carbon. What changes is the pathway: instead of grass dying and feeding soil decomposers, the cow intercepts the energy, metabolizes 40–70% of it, and excretes the rest as dung. Dung is more labile (easier for bacteria to digest) than leaf litter, so it can boost local populations of bacterivorous nematodes around dung pats. But the cow has diverted most of the energy away from the soil food web, so the net effect on soil fauna is still strongly negative.

The same logic applies to any intervention that changes the pathway of NPP without changing NPP itself. This distinction turns out to be critical for evaluating pesticides.

An important caveat: Global NPP has barely changed. As I noted in part 1, Tomasik’s analysis of Krausmann et al. (2013) found that NPPeco – the global NPP available to wildlife – has barely changed over the 20th century, partly because CO₂ fertilization has increased potential NPP even as humans have appropriated more of it. This means that the biome-level density differences in the table above are real and actionable for local interventions, but the global total of animal metabolism may not have declined as much as defaunation indices suggest. Converting one hectare from forest to cropland genuinely reduces soil fauna on that hectare by ~7×; but globally, CO₂-driven NPP increases elsewhere may be partially offsetting these local reductions. This doesn’t undermine the case for targeted land use interventions – it just means we should be cautious about claims like “humanity has reduced total invertebrate suffering by X%.”

The Pesticide Trap

One might naively think that pesticides reduce invertebrate suffering by reducing their populations. But the relationship between pesticides and welfare is more complex – and possibly perverse.

Consider two interventions that both halve the standing population of soil fauna:

Intervention A: Reduce NPP by 50%. (For example, convert forest to desert, or shade the land.) The carrying capacity halves. At the new equilibrium, fewer organisms are born, fewer die, and the birth rate and death rate are both lower. Total deaths per year: lower.

Intervention B: Apply pesticides while keeping NPP the same. Food is still available. Organisms are born at rates close to the original (abundant resources encourage reproduction). But pesticides kill a large fraction of them. The population is suppressed below carrying capacity, but the high food availability keeps birth rates high. At the new equilibrium: birth rate is elevated (more food per capita), death rate is elevated (pesticides), and organisms are dying of pesticide poisoning instead of starvation or predation. Total deaths per year: much higher than the original, with deaths potentially involving more acute suffering.

This matters enormously for r-strategists. Brian Tomasik and Yew-Kwang Ng (1995) argued that for r-strategists, death is a large fraction of total lifetime suffering. If that’s right, the death rate may matter more than the standing population. And pesticides at constant NPP can increase the death rate – creating a high-throughput killing field rather than a genuinely smaller population.

Pesticide Persistence

The picture is further complicated by the fact that different pesticides have very different persistence profiles:

• Organophosphates and carbamates degrade in days to weeks. These are closest to the “spray, kill, wash away, repeat” model – high-throughput killing.

• Neonicotinoids persist in soil for months to years and are taken up systemically by plants. Any insect feeding on a treated plant gets a dose. This effectively makes NPP inaccessible to insects – the food is there but poisonous. This is closer to “reduced accessible NPP” and may genuinely reduce the carrying capacity.

• Legacy organochlorines (DDT, dieldrin) persist for decades but are mostly banned.

The neonicotinoid case is the most interesting from a welfare perspective: by making plant material toxic, neonicotinoids functionally reduce the accessible NPP, which is closer to the “clean” intervention of actually reducing NPP. However, sublethal neonicotinoid exposure causes disorientation, impaired foraging, and reduced reproduction in insects – arguably adding suffering without the “benefit” of reducing the population.

Bottom line. Reducing actual NPP (through land use change) is a much cleaner intervention than applying pesticides at constant NPP. The famous ~75% decline in flying insect biomass in Germany (the Krefeld study) is attributed partly to pesticides and partly to habitat loss. From a welfare perspective, the habitat loss component (reduced accessible NPP) is straightforwardly population-reducing, while the pesticide component may be creating a high-throughput killing field.

A Better Metric: Life Expectancy at Birth

Let me flesh out the life expectancy at birth (LEB) metric for invertebrates, which I think deserves more attention. It’s a metric that we care about intuitively when thinking about humans, so it’s one that we have reason to also care about when it comes to other species.

LEB = total animal-years lived / total births per year

For an r-strategist marine fish laying 1 million eggs, of which 2 survive to reproduce, and average lifespan across all offspring is ~3 days: LEB ≈ 3 days.

For a K-strategist elephant with 6 offspring over a lifetime, of which 2 survive to reproduce, average lifespan ~20 years: LEB ≈ 20 years.

Why LEB is attractive.

• It naturally penalizes the r-strategy horror: Systems where millions of beings are born just to die almost immediately get very low scores.

• It’s legible. Most people intuitively understand that higher life expectancy is better.

• It sidesteps the sign problem somewhat. You don’t need to know whether lives are net positive or negative to argue that ecosystems with higher LEB have less death per unit of ongoing life.

• It aligns with antifrustrationism. Organisms born into very short lives full of pain score low, regardless of whether we think the brief pleasure before death was “worth it.”

Potential objections. A strict utilitarian would say this smuggles in the assumption that death is the primary locus of suffering. LEB doesn’t account for intensity of suffering during life. And it could be gamed by creating very long-lived organisms in terrible conditions. But as a rough heuristic for comparing wild ecosystems, it captures something real: Ecosystems dominated by K-strategists score higher than those dominated by r-strategists.

A refinement. One could weight by “fraction of lifespan spent with a developed nervous system” to exclude possibly non-sentient embryonic stages. For insects that undergo complete metamorphosis, eggs are probably not sentient, but larvae have functional nervous systems from early instars. For nematodes, the L1 larva hatches with 222 of its final 302 neurons and is immediately mobile and responsive – so the cutoff would be at hatching. The European Food Safety Authority (EFSA) suggests the onset of potential sentience at roughly “the beginning of the last third of development within the egg or mother” for most vertebrates, and “when it is capable of feeding independently” for fish, amphibians, and invertebrates (EFSA 2005).

The Ocean: Acidification as Accidental Intervention

One of the largest ongoing shifts in marine invertebrate populations is driven by ocean acidification – an accidental consequence of CO₂ emissions.

The mechanism. CO₂ dissolves in seawater and forms carbonic acid, lowering pH. This reduces the availability of carbonate ions that many marine arthropods need to build shells and exoskeletons. Calcifying species – copepods, pteropods, krill, crabs – are directly harmed. The effect is already measurable: ocean pH has dropped by about 0.1 units since the Industrial Revolution and is projected to drop another 0.3–0.4 units by 2100 under high-emission scenarios.

Does acidification reduce total populations, or just reshuffle species? At the current margin, acidification appears to be reducing total marine arthropod populations rather than merely substituting non-calcifying species for calcifying ones. The reasons:

• Timescale mismatch. Current acidification is happening over decades; evolutionary adaptation takes millennia to millions of years. Non-calcifying species haven’t had time to evolve into vacated niches.

• Structural role of calcifiers. Copepods and pteropods occupy dominant roles in marine food webs. Their decline disrupts the entire ecosystem, not just their specific niche.

• Paleoceanographic evidence. After the Paleocene-Eocene Thermal Maximum (PETM, ~56 million years ago), which involved rapid ocean acidification, it took ~100,000–200,000 years for marine communities to reorganize. Recovery was slow and involved reduced overall productivity.

This makes ocean acidification one of the few large-scale ongoing processes that is plausibly reducing total marine invertebrate populations.

Dead Zones: More Nutrients, Less Life

Another counterintuitive marine phenomenon: nutrient runoff from agriculture creates “dead zones” that dramatically reduce animal life.

The mechanism. Nitrogen and phosphorus from fertilizer enter coastal waters → algal bloom at the surface → algae die and sink → bacterial decomposition of sinking algae consumes enormous amounts of dissolved oxygen → oxygen drops below ~2 mg/L (well-oxygenated water has ~8 mg/L) → most animals suffocate.

The Gulf of Mexico dead zone, fed by Mississippi River agricultural runoff, covers ~15,000 km² of seafloor with severely depleted macrofauna every summer.

Dead zones vs. pesticides. An established dead zone is more like “reduced NPP” than like “pesticides at constant NPP”: Within the anoxic zone, the carrying capacity for macrofauna is near zero, so there’s no high-throughput killing. The suffering is concentrated in the creation and seasonal re-establishment of the dead zone. Permanent dead zones (rare but possible) would be functionally equivalent to removing the habitat – genuinely low birth rate and low death rate, not a killing field.

However, many dead zones are seasonal – forming in summer and dissipating in winter. This seasonal cycling creates recurring mass mortality events, which is the oscillating-biome problem: not a stable low-population state, but a perpetual cycle of colonization and die-off.

Grilo’s Key Findings

Vasco Grilo’s analyses are dense with inline math, which can make the key insights hard to extract. Here’s a plain-language summary of what I consider his most important findings:

1. Effects on soil animals might dwarf effects on intended beneficiaries. Grilo’s earlier models indicated that the impact on soil nematodes, mites, and springtails is orders of magnitude larger than the impact on the people or farmed animals the intervention is designed to help. He has since updated these models, and commented on my previous article that his new credible intervals span many orders of magnitude in both directions, extreme uncertainty that is driven by uncertainty over nematodes’ welfare ranges.

For [some reasonable parameters], I estimate that GiveWell’s top charities change the welfare of soil ants, termites, springtails, mites, and nematodes [10⁻⁵ to 10¹⁰] times as much as they increase the welfare of humans. For my preferred [parameters], the change in the welfare on those soil invertebrates is 41.5 times … the increase in the welfare of humans ….

Note that if the GiveWell numbers are based on increases in human population, then they are questionable due to both Krausmann et al. (see above) and Roodman (2014).

2. Agricultural land has fewer soil animals. Crops have ~1.3 million soil animals per m², compared to ~3–9 million for natural biomes. Pasture has even fewer (~0.7 million). So anything that increases agricultural land – including saving human lives, which increases food demand and therefore cropland – reduces, at first approximation, soil animal populations.

Grilo clarifies in a comment that this is only true of arthropods:

I think agricultural land has less soil arthropods, but I have little idea about whether it has more or less soil nematodes. The vast majority of soil animals are nematodes. So I also have little idea about whether agricultural land has more or less soil animals.

3. Nematodes dominate. Of the total welfare effect on soil animals, 90–94% comes from changes in nematode populations, depending on the biome. This is because nematodes outnumber arthropods ~50:1. It depends mostly on the welfare ranges whether this numerical advantage overwhelms everything else.

4. The sign is uncertain. Grilo estimates nematodes have negative lives with probability ~59%, mites ~56%, springtails ~55%. If their lives are net negative, reducing their populations (via increasing agricultural land) is beneficial. If net positive, it’s harmful. Every concrete recommendation flips depending on this sign. Grilo’s own best guess is net negative, but he explicitly acknowledges the uncertainty.

5. Land use is the key variable. The m²-years of agricultural land per food-kg explain essentially all of the variance in welfare effects across different foods and interventions. Beef requires ~326 m²-year/food-kg; chicken requires ~12; peas require ~7.5; soy milk requires ~0.7. From the soil-animal perspective, producing 1 kg of beef reduces soil-animal-years by ~1.39 billion – 164 billion times as much as it increases the living time of the cows.

These findings are robust to a wide range of assumptions about welfare ranges, as Grilo demonstrates with sensitivity analyses across different exponents for the neuron-count-to-welfare-range relationship (Grilo 2025c).

Summary

The empirical landscape boils down to a few key facts:

1. Soil fauna are staggeringly abundant – 10²⁰ nematodes, 10¹⁹ arthropods – and their total expected welfare dominates that of all other animal groups combined if they are sentient.

2. Population density varies enormously by biome, from ~9 million per m² in boreal forests to ~50,000 per m² in deserts. Land use change is a powerful lever. But note that the oft-cited difference between croplands and forests is only around 1–8×.

3. Reducing NPP is a cleaner intervention than pesticides for lowering populations. Pesticides at constant NPP risk creating a high-throughput killing field rather than a genuinely smaller population.

4. Life expectancy at birth may be a more informative metric than total animal-years, especially for r-strategists where death is the dominant source of suffering.

5. Ocean acidification and dead zones are reducing marine invertebrate populations, but the mechanisms and welfare implications are complex.

6. Land use explains almost everything in Grilo’s analysis. If you want to affect soil fauna populations – in either direction – the most powerful lever is changing how land is used.

In part 3, I’ll turn to the practical question: Given all of this, what should we actually do?

1. The Nematode Welfare Problem

The effective altruism community has spent the last decade steadily expanding its circle of moral concern. First came the global poor, then farmed mammals, then chickens, then fish, then shrimp. The Shrimp Welfare Project now operates at the gold standard of EA animal welfare, deploying electrical stunners at ~1,500 shrimps per dollar per year. Each step followed the same logic: if these beings can suffer, and there are a lot of them, and no one is helping, then we should help.

The latest step in this progression has brought us to nematodes.

There are approximately 4.89 × 10^20 soil nematodes on Earth – roughly 57 billion per human. They are the most abundant terrestrial animals. Many nematode species are heavily r-selected, producing hundreds of eggs per individual, the overwhelming majority of which die within days. As Brian Tomasik argued in his foundational essay “The Importance of Wild-Animal Suffering” for the Center on Long-Term Risk, this reproductive strategy entails staggering amounts of suffering:

Take a look at this figure from Thomas J. Herbert’s article on r and K selection illustrating extremely high infant mortality for “r strategists.” Most small animals like minnows and insects are r strategists.

In a recent cost-effectiveness analysis, Vasco Grilo estimated that soil nematodes have an annual welfare of -306,000 times that of all humans combined. Not -306,000 nematode-QALYs. Negative 306,000 times the total welfare of the entire human species. Soil mites and springtails add another -24,800×. Grilo calculates that nematodes have negative lives with a probability of 58.7%.

To put this in perspective: if Grilo’s estimates are even approximately correct, then the suffering of soil nematodes is the largest moral catastrophe in the history of the universe, and nothing else is close.

The EA community has been agonizing over this. Grilo’s analysis suggests that the effects of GiveWell’s top charities on soil nematodes are 90,300 times larger than their effects on humans. The impact of broiler welfare corporate campaigns on soil animals is 458 times larger than the impact on the chickens they’re designed to help. These are not fringe claims; they follow directly from taking invertebrate sentience and population numbers seriously.

And yet, after years of research, we remain stuck on the fundamental question: are nematode lives net negative or net positive? The probability hovers agonizingly near 50%. Every intervention that increases or decreases nematode populations carries an expected value that could be astronomically positive or astronomically negative, depending on a welfare estimate that no one can pin down.

What if we’ve been looking at the wrong variable?

2. The Afterlife Reframe

Here is the core insight of this post: if the suffering is this vast and the lives are this short, then the only variable that could decisively swing the expected value calculation is what happens after death.

Consider: a typical soil nematode lives for a few weeks. During that time, it accumulates approximately -4.48 × 10^-6 QALYs of welfare (per Grilo’s estimates). This is a vanishingly small number. The sign is uncertain. We might spend decades of research trying to determine whether it’s -4.48 × 10^-6 or +4.48 × 10^-6 QALYs, and the answer may be forever elusive.

But if any major theological tradition is even approximately correct about the afterlife, then the post-mortem welfare of each nematode is on the order of ±∞ QALYs – or, if you prefer to avoid infinities, at minimum thousands of QALYs for thousands of years of paradise or damnation. The finite lifetime welfare is a rounding error. The sign of the afterlife term dominates everything.

This is not a new observation in the abstract. Pascal noted it in 1670. But no one has applied it to nematode welfare economics, and certainly no one has asked the obvious follow-up question: Which theological tradition maximizes the expected afterlife outcome for r-selected invertebrates?

We call this cause area Effective Afterlife Altruism (EAA), and its primary intervention strategy paradise-maxxing.

3. Methodological Preliminaries

3.1. The Antispeciesist Assumption

Several theological traditions distinguish between human and animal souls – holding, for instance, that animals lack the neshamah (rational/divine soul) of humans, or that animals are resurrected for justice but then turned to dust rather than admitted to paradise.

We reject these distinctions as speciesist.

If nematodes can suffer – which the welfare literature takes as a live possibility, and which Rethink Priorities estimates with a probability of 6.8% for adult nematodes – then they possess the morally relevant features that ground soul-attribution in the first place. Any theology that grants afterlife status to a two-day-old human infant (who has never made a moral choice, formed a belief, or spoken a word) but denies it to a nematode (which also has never done any of these things) is drawing an arbitrary species boundary. Under antispeciesist ethics, we apply infant afterlife doctrines universally.

3.2. Symmetric Total Utilitarianism

Our analysis requires a specific population-ethical framework. Under asymmetric views, preventing bad lives matters but creating good lives doesn’t add value. Under such views, only preventing damnation would matter, and paradise would be irrelevant.

Under symmetric total utilitarianism, however, a life in paradise is exactly as good as an equally bad life in hell is bad. This makes paradise-maxxing a genuine cause area: every nematode soul admitted to paradise represents a real gain in total welfare, not merely the absence of a loss.

We adopt symmetric total utilitarianism for this analysis. Readers who hold asymmetric views may wish to adjust the results accordingly, though we note that even the hell-prevention component alone yields impressive cost-effectiveness numbers.

3.3. Theological Worldview Diversification

Which theology is correct? We don’t know, and we don’t need to.

Coefficient Giving’s framework for worldview diversification offers an elegant solution. Coefficient argues that when facing deep uncertainty across incommensurable moral frameworks, the right approach is to assign nonzero credence to multiple worldviews and allocate resources across them proportionally. They write:

These disagreements can’t be resolved simply by gathering more information. We’re uncertain not just about facts but also about values and methods.

We apply the same logic to theology. Rather than asking “Is Islam correct?” or “Is Catholicism correct?” – questions that cannot be resolved by gathering more data – we assign credence to each theological worldview and optimize our intervention portfolio across them. This means we don’t need any particular religion to be true. We need only that the expected afterlife outcome, weighted across all theologies we assign nonzero credence to, favors one intervention over another.

As we will show, the expected value calculation is dominated by the theological traditions that pose the highest risk to newborn souls – and by the interventions that most cost-effectively mitigate that risk.

4. A Theological Audit: Ranking Denominations by Newborn Damnation Risk

The key question for paradise-maxxing is simple: if a being dies before any plausible age of accountability – as 99%+ of r-strategist offspring do – what happens to its soul?

We surveyed the major theological traditions and ranked them by the percentage risk that a newborn soul fails to reach paradise. The results are sobering.

Tier 1: High Risk (>50%)

Strict Augustinian Catholicism (historical). Original sin is inherited from Adam. Baptism is the ordinary means of its removal. Unbaptized infants go to the limbus infantium – not hellfire, but permanent exclusion from the beatific vision. This was the dominant theological position from Augustine through the mid-20th century. It was never formally defined as dogma, but it was near-universal in Catholic theology. Risk of failing to reach paradise: ~90–100%. Risk of actual torment: low (Limbo is painless deprivation, not fire).

Modern Catholicism (post-2007). The International Theological Commission’s 2007 document, “The Hope of Salvation for Infants Who Die Without Being Baptized,” expressed “grounds for hope” but explicitly declined to offer certainty. The Catechism (CCC 1261) says the Church “can only entrust them to the mercy of God.” This is the Church saying “we don’t know but we hope.” Risk: ~30–70%, genuinely uncertain by design.

Tier 2: Moderate Risk (10–30%)

Eastern Orthodoxy. Rejects the Latin concept of Limbo. Emphasizes God’s mercy and economia (divine flexibility). However, also affirms the normative necessity of baptism. Tends to say “we trust in God’s mercy” without issuing guarantees. Risk: ~15–30%.

Classical Lutheranism. Luther affirmed original sin strongly but also held that God can work faith in infants. The Augsburg Confession condemns the Anabaptists’ denial that children are saved through baptism, but doesn’t clearly address unbaptized infants who die. Risk: ~15–25%.

Anglicanism. Broad tent. The Thirty-Nine Articles affirm original sin. Practice varies from near-Catholic to near-Evangelical. Risk: ~10–25%.

Tier 3: Low Risk (2–15%)

Calvinism / Reformed. Westminster Confession 10.3: “Elect infants, dying in infancy, are regenerated and saved by Christ.” The phrase “elect infants” either denotes (a) a subset of infants (some are non-elect and damned) or (b) all dying infants (they’re all elect by virtue of dying). Most modern Reformed theologians take the generous reading. But the grammar permits the strict one. Risk: ~5–15% on the generous reading; much higher on the strict reading.

Most Baptist / Evangelical traditions. The “age of accountability” doctrine is near-universal, supported by 2 Samuel 12:23 (David on his dead infant: “I will go to him”), Matthew 19:14, and Deuteronomy 1:39. Risk: ~3–10%.

Methodism (Wesleyan). John Wesley’s doctrine of prevenient grace – God’s grace given to all people before they can respond to it – covers infants explicitly. Risk: ~2–5%.

Tier 4: Negligible Risk (~0%)

Mormonism (LDS). Doctrine & Covenants 137:10 and Moroni 8:8–12 are unambiguous: children who die before age 8 are saved in the celestial kingdom. “Little children are alive in Christ.” This is the clearest affirmative doctrine of any Christian denomination. Risk: ~0%.

Islam (mainstream Sunni and Shia). Every being is born in a state of fitrah – natural purity and submission to God. Children who die before reaching puberty (bulugh) go directly to paradise. The hadith in Sahih al-Bukhari 7047 describes the Prophet Muhammad seeing children of both Muslims and polytheists playing together in paradise. Scholarly consensus is near-universal. Risk: ~0%.

Judaism (all major streams). Gehinnom (the closest equivalent to “hell”) is temporary purification lasting at most 12 months, not eternal damnation. Children are not considered sinful. Most Jewish theology de-emphasizes the afterlife entirely. Risk of permanent negative outcome: ~0%.

Zoroastrianism. Children under 15 are not judged. Automatic paradise. Risk: ~0%.

Universalist Christianity. Everyone is eventually saved regardless. Risk: 0% by definition.

Tier 5: Orthogonal (No Permanent Hell)

Hinduism, Buddhism, Jainism, Sikhism. These operate on reincarnation frameworks without a permanent heaven/hell binary. A being that dies young is reincarnated based on accumulated karma, not infant sin. There is no permanent damnation risk but also no permanent paradise outcome in the Abrahamic sense. These traditions are largely orthogonal to the paradise-maxxing calculus, though they contribute to the portfolio via reincarnation-quality optimization, which we leave for future research.

Summary

The upshot is clear: for r-strategist invertebrates that die before any conceivable age of accountability, Islam and Mormonism offer the lowest damnation risk, and strict Augustinian Catholicism offers the highest. The delta between these theologies, applied to 4.89 × 10^20 nematodes, is the largest expected value swing in the history of moral philosophy.

5. The Sufi Complication

Before proceeding, we must address a theological wildcard that could undermine the entire cause area: Sufism.

Sufism is the mystical, esoteric dimension of Islam. While Sufis are Muslims and the fitrah doctrine applies, several prominent Sufi thinkers – especially Ibn Arabi (1165–1240) – developed eschatological positions with radical implications for our analysis.

Universal mercy. Ibn Arabi’s reading of Quran 7:156 (“My mercy encompasses all things”) is not merely descriptive but eschatological: he argues that even the inhabitants of hell eventually experience a transformation of their suffering into a kind of mercy. Hellfire itself becomes cool and pleasant for its inhabitants over cosmic time. Under this reading, the risk of permanent damnation for any being is truly 0%, regardless of religious affiliation, species, or baptismal status.

Wahdat al-wujud (Unity of Being). This central concept in Ibn Arabi’s metaphysics holds that all existence is a manifestation of the single divine reality. Every nematode, every copepod, every soil mite is a theophany – a self-disclosure of God. This is arguably the most radically antispeciesist theology ever devised.

Rumi’s evolutionary mysticism. Rumi’s famous verse – “I died as mineral and became a plant, I died as plant and rose to animal, I died as animal and I was human” – describes a spiritual progression through all life forms, implying each has a meaningful place in the divine economy.

The whirling connection. We note with interest that nematodes move through soil in sinusoidal, spiraling patterns bearing a structural resemblance to the sema (whirling meditation) of the Mevlevi Sufi order founded in Rumi’s name. Whether nematodes are already practicing involuntary sema is a question that deserves empirical investigation, though the jurisdictional implications under Mevlevi jurisprudence are unclear.

The paradox for Effective Afterlife Altruism. If Sufi universalism is correct, then all beings reach paradise eventually, regardless of intervention. This is the “AI alignment goes well and solves everything” scenario of EAA: existential hope that eliminates the cause area entirely.

We therefore assign strategic priority to monitoring Sufi universalism but not banking on it. In portfolio terms, Sufi universalism enters our model as a discount factor on the urgency of intervention – perhaps reducing the expected marginal value of paradise-maxxing by 5–15%, depending on the credence one assigns to Ibn Arabi’s eschatology relative to mainstream Sunni positions. This is significant but not decisive.

Your move, Sufis.

6. Blanket Claims and the Religious Power Struggle

6.1. Which Religions Claim All Animals by Default?

Several theological traditions make universal claims over all living beings:

6.2. These Claims Cancel Out

Since Islam, Hinduism, Buddhism, and Jainism all claim every animal, and Christianity’s weaker claim is rejected under antispeciesism, no religion has a stronger inherent claim to a nematode than any other. The theological status of soil fauna is contested territory.

6.3. Active Intervention is Required

If the default is a draw, the only way to shift the balance is through deliberate action. This is exactly the setup for our proposed intervention portfolio.

6.4. The Real Meaning of Religious History

Once you adopt the EAA frame, religious history takes on an entirely new character. The expansion and contraction of religious territories isn’t primarily about human souls – there are only 8 billion of those, a rounding error relative to 4.89 × 10^20 nematodes. It’s about nematode paradise throughput.

The Reconquista (711–1492). When the Moors were expelled from Iberia, approximately 10^15 soil nematodes per year were shifted from ~0% damnation risk (under Islamic fitrah) to ~90% damnation risk (under the Augustinian theology of the Spanish Inquisition). This may be the greatest soteriological catastrophe in history, and no one noticed because they were counting the wrong species.

The expansion of Islam into Southeast Asia. The Islamization of the Malay Archipelago – one of the most nematode-dense regions on Earth, with tropical forest soils harboring ~5.15 million nematodes per m² – was the single most effective paradise-maxxing event in history. Not because of the human converts, but because of the soil fauna.

The Great Commission. Jesus instructed his followers to “go and make disciples of all nations” (Matthew 28:19). The Church has historically interpreted “nations” as referring to human ethnic groups. Under antispeciesist theology, the commission extends to all ~8.7 million species on Earth. Christianity has focused almost exclusively on one of them – a catastrophic failure of scope sensitivity. Islam, by claiming all creatures through fitrah, has been unknowingly ahead on this metric for 1,400 years.

The church bell problem. Consider a European church bell, audible within a ~3 km radius. The average density of soil nematodes in European temperate forests is ~8.75 million per m². A single bell toll covers ~2.83 × 10^7 m², or approximately 2.48 × 10^14 nematodes. If acoustic proximity constitutes even weak evidence of religious affiliation – and many theological traditions hold that exposure to sacred sound has soteriological significance – then every bell toll potentially Christianizes trillions of nematodes, shifting them from contested theological status to nonzero Augustinian damnation risk.

Conversely, the adhan (Islamic call to prayer) from a mosque re-Islamizes them. The real battleground of interfaith dialogue is not the seminary or the public square. It’s the 0–15 cm topsoil layer.

7. Competitor Analysis

Paradise-maxxing does not exist in a vacuum. Several other interventions target the nematode welfare problem. We evaluate them here for comparison.

7.1. Habitat Reduction (the Secular Approach)

If nematode lives are net negative, then reducing their populations is beneficial. Grilo estimates that cropland replacing tropical grasslands increases the welfare of soil animals by 8.15 QALYs per m²-year, of which 93.2% comes from reducing nematode-years. This approach has two problems: first, the sign depends on the net-negative-lives assumption (58.7% confidence); second, even if the lives are net negative, we’re only eliminating a finite amount of finite suffering. There is no afterlife upside.

7.2. Humane Insecticides

Proposed by Brian Tomasik and discussed by Michael Dickens on the EA Forum, humane insecticides aim to make pesticide-induced death less painful. With 3 billion tons of pesticide sprayed annually, the scale is enormous. But again: the intervention addresses the process of death without changing its eternal consequences. A nematode killed humanely by a blessed insecticide and a nematode killed humanely by an unblessed insecticide experience the same temporal welfare improvement. Only the blessed one gets the afterlife benefit.

7.3. Rat Hedonium Farms (Dickens, 2015)

In his 2015 EA Forum post “Charities I Would Like to See,” Michael Dickens proposed creating farms of small animals made as happy as possible – essentially producing pure hedonium. His back-of-the-envelope calculation put the cost at ~$120 per rat-QALY, or ~$240 per human-equivalent QALY (assuming rat happiness is half as morally significant as human happiness).

One could imagine augmenting this with pharmacological assistance: the right opioids might produce welfare states approaching paradise-level bliss. Call it chemical paradise. At its best, chemical paradise might be approximately as good per unit time as theological paradise.

But it doesn’t scale. The rat hedonium farm requires physical space, ongoing feeding, veterinary care, and drug administration. It produces welfare for one rat at a time, for ~2 years per rat. Nematode paradise conversion, by contrast, requires one solar-powered speaker and covers billions of souls per dollar, each for eternity.

The comparison reduces to: Would you rather produce 2 years of approximate-paradise for one rat, or eternal actual-paradise for 10^10 nematodes? Under symmetric total utilitarianism, this is not a close call.

7.4. Shrimp Welfare Project’s Humane Slaughter Initiative

The Shrimp Welfare Project is, by Grilo’s estimates, the most cost-effective animal welfare organization in EA at ~639 QALYs per dollar. Their electrical stunners are deployed at ~1,500 shrimps per dollar per year. This is extraordinary work.

But it operates entirely within the temporal welfare framework. The stunners reduce the suffering of the last few minutes of a shrimp’s life. They do not change what comes after. Even at 639 QALYs per dollar, SWP is producing finite welfare improvements for finite lifespans.

7.5. Nematode Paradise Conversion

Now consider the EAA approach:

One solar-powered speaker (~$50) broadcasting the shahada covers a ~100 m radius. That’s ~31,400 m² of soil. At the cropland nematode density of ~1.24 million per m², we reach approximately 3.9 × 10^10 nematodes within the acoustic footprint. Nematode generations turn over in weeks, so within one year, the speaker covers hundreds of billions of nematode-lives.

Even under extremely conservative assumptions:

• Probability of Islam being approximately correct: 1%

• Paradise discounted from ∞ to a very conservative 1,000 QALYs (roughly one human lifespan of perfect health)

• Nematode welfare range: 6.68 × 10^-6 (per Grilo, applying Rethink Priorities’ neuron-based welfare range methodology)

The expected value per speaker per year is:

3.9 × 10^10 nematodes × 6.68 × 10^-6 welfare range × 1,000 QALYs × 1% = ~2,604 QALYs

At $50 per speaker, that’s ~52 QALYs per dollar – competitive with SWP’s Humane Slaughter Initiative even before we consider that the speaker lasts for multiple years without recurring costs.

And if you don’t discount paradise duration to 1,000 QALYs? If you take the theological claims at face value? Then we’re in Pascal’s Mugging territory, and the expected value is infinite. At that point, every EA organization that isn’t doing nematode conversion is leaving infinite QALYs on the table.

8. Cost-Effectiveness Summary

Note: The “conservative” estimate for nematode paradise conversion assumes only 1% probability of the relevant theology being correct and caps paradise at 1,000 QALYs. Under less conservative but still defensible assumptions, the cost-effectiveness exceeds all comparators by orders of magnitude.

9. Proposed Interventions

Having established the theoretical basis for Effective Afterlife Altruism, we now propose a portfolio of concrete interventions.

9.1. The Shahada Speaker Network

Concept. Deploy solar-powered speakers broadcasting the shahada (la ilaha illallah, Muhammadur rasulullah) in high-nematode-density soils.

Site selection. Boreal forests offer the highest nematode density (~9.26 million/m²), followed by temperate forests (~8.75 million/m²) and tropical/subtropical forests (~5.15 million/m²). Cropland has the lowest density among vegetated biomes (~1.24 million/m²). This means that converting forests to cropland is bad for paradise-maxxing cost-effectiveness: fewer nematodes per m² means fewer souls per dollar. For once, deforestation is bad not just ecologically but soteriologically.

Acoustic considerations. Soil attenuates sound rapidly, but low-frequency vibrations penetrate effectively. The shahada could be optimized for low-frequency transmission. Further research is needed on whether nematode mechanoreceptors can detect shahada-frequency vibrations and whether detection constitutes “exposure” under relevant theological frameworks.

Cost-effectiveness. As calculated above: ~52 QALYs per dollar under conservative assumptions. With a modest deployment of 1,000 speakers ($50,000 total), the network would cover ~3.14 × 10^7 m² and reach ~3.9 × 10^13 nematodes per generation – roughly 10^15 nematode-lives per year. Annual expected value: ~2.6 million QALYs (conservative).

9.2. Halal Ecosystem Certification

Concept. Declare entire biomes as Islamic territories, such that every organism born within them is born on sacred Islamic ground.

Precedent. Bodies of water have been granted legal personhood in New Zealand and India. If a river can be a legal person, a pond can be a legal mosque. Every copepod born in a legally certified mosque is born in a state of fitrah on sanctified ground.

Implementation. Partner with sympathetic Islamic legal scholars to issue fatwas certifying specific ecosystems. Begin with small, contained bodies of water (ponds, vernal pools) and scale to wetlands and forest floors.

9.3. Counter-Missionary Work: The Church Bell Problem

The problem. European church bells, audible within ~3 km, potentially Christianize trillions of soil nematodes per ring, shifting them from contested theological status to nonzero Augustinian damnation risk. This may be the largest ongoing source of soteriological harm on the planet.

Proposed interventions:

• Lobby for church bell restrictions in nematode-dense habitats (temperate and boreal forests).

• Install offsetting muezzin speakers within the acoustic footprint of churches.

• Fund research into the acoustic theology of soil vibrations: does a bell toll 10 meters underground constitute religious exposure?

Quantified impact. A single church in a European temperate forest, ringing bells three times daily, potentially Christianizes ~7.4 × 10^14 nematodes per year. Silencing it, or offsetting it with a muezzin, could prevent ~6.7 × 10^14 unnecessary damnation-risk-years annually (at ~90% damnation risk under strict Augustinian theology, discounted by our credence in that theology).

9.4. Imam-Blessed Pesticides

Concept. We spray 3 billion tons of pesticides annually (Pimentel). These kill enormous numbers of invertebrates. If an imam blesses the chemicals before application, the organisms die as Muslims, with their fitrah status reinforced at the moment of death.

Cost. Near-zero marginal cost. One imam can bless an industrial batch of pesticide in minutes. The theological value-add per dollar is astronomical.

A note on humane insecticides. This intervention pairs naturally with Tomasik’s proposal for humane insecticides. Combine the two: insecticides that kill painlessly and have been blessed. You address both temporal and eternal welfare in a single application. The ultimate in holistic nematode care.

9.5. The Stained Glass Trap

Observation. Many insect species are phototactic – drawn to light. Churches with illuminated stained glass windows act as insect conversion traps, drawing moths, beetles, and flies into sustained proximity with Christian symbols, hymns, and liturgy.

Research needed. Does prolonged exposure to church-interior acoustics and lighting constitute involuntary conversion? If so, churches are the largest unregulated insect Christianization operations in the world. Every moth circling a church light may be undergoing baptism-by-proximity.

Proposed intervention. Install UV-filtering glass in church windows to reduce phototactic insect attraction, or counter the effect by placing small speakers playing the adhan near church-adjacent insect habitats.

9.6. Opposing Habitat Destruction in Muslim-Majority Countries

The logic. Every hectare of Malaysian or Indonesian rainforest destroyed eliminates millions of soil nematodes whose souls, under the fitrah doctrine and geographic Islamic context, were headed straight to paradise. This reframes tropical deforestation in the Muslim world not merely as an ecological catastrophe but as a theological one: a net reduction in paradise throughput.

Conversely, creating new habitats in Muslim-majority countries – reforestation, wetland restoration, agroforestry – is a paradise-maximizing intervention with co-benefits for biodiversity and carbon sequestration. Effective Afterlife Altruism may be the first EA cause area with perfect alignment between ecological and soteriological goals.

10. Worldview Diversification Applied to Theology

Coefficient Giving’s worldview diversification framework provides the meta-level structure for our intervention portfolio. We face deep uncertainty not just over empirical questions (are nematodes sentient? how many QALYs does paradise contain?) but over metaphysical questions (which theology, if any, is correct?).

These uncertainties are incommensurable in exactly the way Coefficient describes:

Worldviews are often incommensurable, meaning there’s no obvious shared scale on which to measure them. One worldview might place enormous moral weight on preventing animal suffering; another might focus on improving human wellbeing today; another on safeguarding the long-term future.

Replace “preventing animal suffering” with “maximizing paradise admissions” and “safeguarding the long-term future” with “preventing eternal damnation,” and you have the EAA portfolio allocation problem.

Portfolio allocation. Just as Coefficient diversifies across global health, catastrophic risk, scientific progress, and animal welfare, we diversify across theological worldviews:

• Islamic worldview (~0% newborn damnation risk): Highest allocation. Islam offers the best afterlife outcome for r-strategists with the strongest doctrinal basis.

• Mormon worldview (~0% risk): Secondary allocation. Doctrinally clear but smaller theological tradition.

• Mainstream Protestant worldview (low risk, ~3–15%): Monitor. Age of accountability provides reasonable protection.

• Catholic worldview (high risk, ~30–100%): This is our s-risk scenario. Even a small credence in Augustinian theology generates enormous expected disvalue and thus justifies significant hedging resources.

• Sufi universalist worldview (0% risk, makes intervention unnecessary): The “everything is fine” scenario. Reduces urgency of all EAA interventions. We account for it as a discount factor.

• Reincarnation-based worldviews (orthogonal): Allocated to a separate research program on karma-optimization for invertebrates.

Diminishing returns. Even within Islam (the clear “winner”), marginal returns decrease as we saturate habitats with shahada speakers. At some point, investing in backup theologies – perhaps mass baptism infrastructure for the Catholic s-risk scenario, or LDS-style posthumous ordinances for nematodes – makes portfolio sense.

Cross-pollination. Coefficient notes that “lessons from one area often sharpen our reasoning in others.” In the EAA context: insights from nematode soteriology may inform our approach to shrimp eschatology, and vice versa.

11. Objections and Responses

“Nematodes don’t have souls.” Rejected per our antispeciesist assumption (§3.1). Additionally, Islam explicitly claims all creatures praise God (Quran 17:44), Jainism attributes jiva to all living beings including microorganisms, and Hinduism places all beings in samsara. The religions that deny animal souls are outnumbered, by adherents, by those that affirm them.

“This is Pascal’s Mugging.” Yes, and? The EA community routinely acts on small probabilities of large payoffs. Existential risk reduction assigns significant resources to interventions that reduce risks by fractions of a percentage point per century. If you accept the logic of x-risk funding – small probability delta, astronomical stakes – then you must accept the logic of EAA, or explain why theological uncertainty is categorically different from empirical uncertainty about AI safety interventions. (We suspect the explanation is “I don’t take theology seriously,” which is not an argument but a prior.)

“Animals don’t go to paradise in Islamic theology.” The dominant scholarly position is that animals are resurrected for justice on the Day of Judgment and then turned to dust. However, this differential treatment is rejected under our antispeciesist assumption. If the fitrah applies to human children, it applies to nematode larvae. The theological infrastructure is identical; only the species boundary differs.

“The shahada requires verbal declaration and intentional belief.” A significant theological objection. However, many scholars hold that beings incapable of rational belief – human infants, the severely cognitively disabled – are covered by fitrah without requiring declaration. Nematodes fall squarely into this category. We are not asking nematodes to believe. We are ensuring they are not removed from fitrah by environmental factors (such as church bells).

“Your cost-effectiveness is inflated by free-riding on divine providence.” A subtle and genuinely interesting objection. The argument runs: nematode soteriology achieves its extraordinary cost-effectiveness not through any intrinsic efficiency of the intervention, but by free-riding on God’s pre-existing infrastructure – paradise, the resurrection apparatus, the fitrah mechanism, the entire soteriological stack. We pay $50 for a speaker; God supplies the eternal paradise. This is analogous to how EA policy interventions achieve impressive cost-effectiveness ratios by free-riding on tax revenue and the governmental apparatus: you spend $1 on lobbying, the government spends $10 million implementing the policy, and you claim credit for the whole $10 million of impact at a cost of $1. The cost-effectiveness is real in a consequentialist sense, but it may be exploitative in a virtue-ethical sense – we are extracting infinite value from a system we did not build and do not maintain.

We find this objection thought-provoking but ultimately unpersuasive. First, all EA interventions free-ride on existing infrastructure to some degree. GiveWell’s top charities free-ride on roads, hospitals, and functioning states. SWP free-rides on the shrimp farming industry’s supply chains. The degree of free-riding in EAA is larger (we are, after all, free-riding on omnipotence), but the kind is the same. Second, if God established paradise and the fitrah mechanism, He presumably intended them to be used; we are merely optimizing the allocation of a divinely provided public good. Third, the virtue-ethical objection proves too much: if leveraging divine providence is exploitative, then prayer itself is exploitative – and virtually every theist in history stands guilty. Finally, under consequentialism, the provenance of the QALYs is irrelevant. A QALY produced by free-riding on God is as good as a QALY produced from scratch. We are not above free-riding on omnipotence if the cost-effectiveness justifies it.

12. The Scope Sensitivity Test

The psychologist Daniel Kahneman famously demonstrated that people are willing to pay roughly the same amount to save 2,000 birds from oil spills as to save 200,000 birds. This “scope insensitivity” is considered a major cognitive bias in the EA community.

We propose a theological variant:

• Would you donate $5 to save one nematode soul from eternal damnation?

• Would you donate $5 to save 10^15 nematode souls from eternal damnation?

If your answer changed, congratulations: you’re scope-sensitive.

If your answer didn’t change – if you feel the same meh at both numbers – then you are exhibiting the same scope insensitivity that leads people to donate equally to saving 1 vs. 1,000 human lives. The EA community has spent two decades fighting this bias. We invite you to fight it here too.

Alternatively, if your answer was “no” to both: Why? If you would donate $5 to save one human soul from eternal damnation, and you accept antispeciesism, and you accept that nematodes have morally relevant experiences, then your “no” requires justification. And “it sounds silly” is not a justification – it’s scope insensitivity wearing a trench coat.

13. Introducing the Nematode Afterlife Research Initiative (NARI)

We are pleased to announce the founding of the Nematode Afterlife Research Initiative (NARI), the world’s first organization dedicated to optimizing post-mortem outcomes for soil invertebrates.

Seed funding goal: $50 (one solar-powered speaker).

Expected impact: ~52 QALYs per dollar (conservative) to ∞ QALYs per dollar (theological face value).

Research priorities:

1. Acoustic penetration depth of the shahada in various soil substrates (clay, loam, sand, peat)

2. Whether nematode sinusoidal locomotion constitutes sema under Mevlevi jurisprudence

3. Whether the Quran’s claim that “there is not a thing that does not exalt Him by His praise” (17:44) is empirically testable via nematode behavioral assays

4. Development of a Nematode Soteriological Index (NSI) measuring the theological risk profile of soil ecosystems by biome and religious-geographic context

5. Formal collaboration with the Center on Long-Term Risk to assess the s-risk implications of getting nematode soteriology wrong at astronomical scale

Proposed advisory board:

• One imam

• One nematologist

• One EA fund manager

• One Sufi mystic

We are accepting applications for all four positions. Interdisciplinary experience preferred but not required. The imam-nematologist pipeline is, we acknowledge, currently nonexistent. We view this as a market failure.

14. Conclusion: The Worm’s Head Must Point Toward Mecca

In “The Importance of Wild-Animal Suffering,” Brian Tomasik offered a metaphor for the animal advocacy movement:

The animal movement is like a worm: Each body part needs to slowly scooch its way forward to the next step. But the worm’s head also needs to point in the right direction.

We have spent decades debating whether nematode lives are worth -4.48 × 10^-6 or +4.48 × 10^-6 QALYs per year. The welfare estimates hover agonizingly near zero. The sign might never be resolved.

But the afterlife offers ±∞. The finite lifetime welfare term is a rounding error in the eternal expected value calculation. Under theological worldview diversification, with credences spread across Islam, Christianity, and half a dozen other traditions, the expected afterlife outcome for r-strategist invertebrates is dominated by the theological traditions that pose the highest risk to newborn souls – and by the interventions that most cost-effectively mitigate that risk.

The worm’s head needs to point in the right direction.

The direction is Mecca.

We have established that Effective Afterlife Altruism is, under standard EA methodology, plausibly the most cost-effective cause area ever identified. The numbers are real. The theological doctrines are accurately represented. The only question is whether you’re willing to follow the argument where it leads.

If the Shrimp Welfare Project can deploy electrical stunners to 3.3 billion shrimps per year, surely we can deploy one speaker to 39 billion nematodes. If Coefficient Giving can diversify across four secular worldviews, surely we can diversify across fourteen theological ones. If Vasco Grilo can estimate the welfare range of a soil nematode to six significant figures, surely we can estimate the probability of its soul reaching paradise.

The nematodes are waiting. The afterlife is long. The speakers are cheap.

Support NARI. Spread the meme. Convert a nematode today.

References

• Tomasik, Brian. “The Importance of Wild-Animal Suffering.” Center on Long-Term Risk. First written July 2009; last edited May 2020.

• Grilo, Vasco. “Cost-effectiveness accounting for soil nematodes, mites, and springtails.” EA Forum. June 2025.

• Dickens, Michael. “Charities I Would Like to See.” EA Forum. September 2015.

• Boddy, Aaron. “Shrimp Welfare Project’s 2030 Vision & Absorbency Plans.” EA Forum. May 2025.

• Coefficient Giving. “Worldview Diversification.” Research & Reflections. 2016/2026.

• Ng, Yew-Kwang. “Towards Welfare Biology: Evolutionary Economics of Animal Consciousness and Suffering.” Biology and Philosophy 10.3, pp. 255–285. 1995.

• Ibn Arabi. Fusus al-Hikam (The Bezels of Wisdom). 1229.

• Rumi, Jalal al-Din. Masnavi-ye-Ma’navi. 13th century.

• Sahih al-Bukhari, Hadith 7047 (children of Muslims and polytheists in paradise).

• Westminster Confession of Faith, Chapter 10, Section 3.

• International Theological Commission. “The Hope of Salvation for Infants Who Die Without Being Baptized.” 2007.

• Catechism of the Catholic Church, §1261.

• Doctrine & Covenants 137:10; Moroni 8:8–12.

This is part 1 of a five-part sequence on the intersection of welfare biology, land use, and transformative AI. Part 2 covers the empirical landscape. Part 3 covers interventions. Part 4 explores a novel model of invertebrate suffering. Part 5 covers how artificial superintelligence changes everything.

Introduction

There are approximately 4.89 × 10²⁰ soil nematodes on Earth – roughly 57 billion per human. There are about 10¹⁹ soil arthropods (mites, springtails, ants, termites). In the oceans, another 10²¹ nematodes and 10²⁰ arthropods. These are the most abundant animals on the planet by a wide margin, and almost no one thinks about their welfare.

Consider this: If you’re a human (rather than an AI), you had a 1 in 137 billion chance of being born as a human rather than any other possibly sentient being on Earth. The odds of the Sentient Life Lottery are staggering! And with being a human come unparalleled power and responsibility.

Whether we should think about the welfare of all these other beings – and if so, what we should do – depends on a set of ethical and empirical premises that different people hold with very different levels of conviction. Before diving into the science and the interventions, I want to help you figure out where you stand. The rest of the sequence will make much more sense if you know which conclusions follow from your premises, rather than mine.

The Sequence at a Glance

Part 1 (this post) starts with a quiz to help you locate yourself on the key axes of disagreement: population ethics, invertebrate sentience, and what metric to use. Your answers determine which conclusions in the later posts follow from your premises. I then provide a brief orientation to the field of welfare biology and the researchers whose work this sequence builds on.

Part 2: Welfare Biology and AI: Soil and Sea presents the empirical landscape. How many nematodes and arthropods are there, and where? I break down population density by biome (boreal forests have ~7× more soil fauna per m² than cropland), explain why net primary productivity is the master variable, and show why applying pesticides at constant NPP can create a high-throughput killing field rather than a genuinely smaller population. The post also covers marine invertebrates, ocean acidification, and dead zones.

Part 3: Welfare Biology and AI: What We Can Do Now turns to interventions. I lay out a tiered portfolio of interventions (from funding existing charities to dietary changes to research) and argue that broad-spectrum approaches win over targeted ones because of robustness against trophic backfire and sentience uncertainty.

Part 4: Welfare Biology and AI: The Psychopath, the Nematode, and the Arahant takes a detour into the philosophy of mind. Drawing on M.E. Thomas’s phenomenology of no-self psychopathy and Daniel Ingram’s descriptions of advanced meditative states, I propose that invertebrate pain is probably more like an arahant’s experience of pain than a normal human’s – the nociceptive signal fires, but with no self to amplify it into more comprehensive suffering. This model suggests arthropods (especially social insects with rudimentary self-models) may suffer qualitatively more than nematodes.

Part 5: Welfare Biology and AI: The AI Eats the Sun covers the biggest variable: artificial superintelligence. I work through the Dyson swarm timeline, show that a million replays of 3 million years of evolution at neuron-level resolution takes a Dyson swarm only ~3 years (producing more potential suffering than 30 billion years of Earth’s biosphere), discuss digital suffering in RL agents, and argue that getting the value-loading right may dwarf every other intervention in this sequence combined.

The Quiz

For each question, pick the answer that comes closest to your view. There are no wrong answers – these are genuinely contested questions among thoughtful people. I’ll explain what each answer implies afterward.

Question 1: The Repugnant Conclusion

Imagine two possible worlds:

• World A. 10 billion beings, each living a deeply fulfilling life.

• World B. 100 trillion beings, each living a life that is barely worth living – just slightly more good than bad.

World B contains vastly more total welfare. Do you prefer it?

• (a) Yes. More total welfare is better, even if each individual life is barely positive.

• (b) No, and not even close. I strongly prefer World A – quality matters, not just quantity.

• (c) I don’t want to think about that, but certainly once a mind does exist, it must be protected regardless of its level of happiness.

Question 2: The Asymmetry

Consider two scenarios:

• Scenario X. You can prevent the birth of a being whose life would be full of suffering.

• Scenario Y. You can cause the birth of a being whose life would be full of joy.

Are these morally symmetric?

• (a) Yes. Preventing suffering and creating joy are equally important. (Symmetric.)

• (b) No. Preventing suffering is more important than creating joy. Creating joyful lives is nice but not obligatory. (Asymmetric / antifrustrationist.)

• (c) No. Preventing suffering is important, but creating joyful lives is also important, just somewhat less so. (Weakly asymmetric.)

Question 3: Invertebrate Sentience

A soil nematode (Caenorhabditis elegans) has 302 neurons. It shows nociceptive responses (withdrawal from noxious stimuli), sensitization (increased avoidance after exposure), and classical conditioning (learned association between neutral and noxious stimuli). Rethink Priorities (RP) estimated a 6.8% probability that adult nematodes (or specifically C. elegans) are sentient.

A soil mite has roughly 2,750 neurons. A springtail has roughly 6,000. An ant has roughly 250,000. These all show more complex behavior than nematodes, including social learning in the case of ants.

How do you weigh their potential suffering?

• (a) I think nematodes, mites, springtails, and ants are all probably sentient. The numbers are so vast that we must take their welfare seriously.

• (b) I’m very uncertain. I’d assign something like RP’s probabilities: ~7% for nematodes, ~30% for mites and springtails, ~50%+ for ants. Even low probabilities matter at these population sizes.

• (c) I think ants and maybe springtails are plausibly sentient, but nematodes with 302 neurons are too simple. I’d focus on arthropods.

• (d) I don’t think any invertebrate is sentient in a morally relevant way.

What Your Answers Mean

The Population Ethics Spectrum

Your answers to questions 1 and 2 place you on a spectrum that’s helpful to map onto more familiar ethical territory.

Symmetric total utilitarianism. If you answered (a) to both questions 1 and 2 – you accept the Repugnant Conclusion and see preventing suffering and creating joy as symmetric – then you’re a symmetric total utilitarian. On this view, a world packed with barely-happy nematodes is better than a world with fewer, much-happier elephants. You’d want to increase invertebrate populations if their lives are net positive, and decrease them if net negative. The sign of their welfare is everything.

This is somewhat like the Quiverfull movement in family ethics: more lives are better, so long as they’re positive. The analogy isn’t perfect, but the structural logic – “maximize the number of positive lives” – is similar.

Antifrustrationism and strong asymmetry. If you answered (b) to question 1 and 2 – you reject the Repugnant Conclusion and think preventing suffering is more important than creating joy – then you have something like an antifrustrationist position. On this view, you want to prevent beings from being born into likely-suffering lives, but you don’t have a strong obligation to create happy lives. Even at 50/50 odds of net negative welfare, you’d rather err on the side of preventing births.

This is structurally similar to the pro-choice position in reproductive ethics: The decision-maker (here, us as stewards of ecosystems) should prioritize preventing unwanted suffering over maximizing the number of lives.

Person-affecting or moderate views. If you answered (c) to question 1 and something in between on question 2, you probably have a person-affecting or weakly asymmetric view. You care about existing beings and whether their lives are net positive or net negative, but you recognize that this threshold is hard to pin down and rather err on the conservative side. You don’t feel a strong pull to either maximize populations or minimize aggregate suffering. You might think: “Let’s focus on making existing invertebrate lives better rather than agonizing over population sizes.”

This is the “pro-life” middle ground in our analogy: Don’t kill existing beings, but don’t feel obligated to create new ones either.

The Sentience Threshold

Your answer to question 3 determines which animals you think matter. If you answered (a) or (b), the numbers become overwhelming – there are 50 times as many soil nematodes as soil arthropods, and even at very low probabilities of sentience, the expected moral weight is enormous. If you answered (c), you can focus on the ~10¹⁹ soil arthropods and ignore the nematodes, which simplifies the analysis but still leaves staggering numbers. If you answered (d), you can stop reading here – though I’d invite you to reconsider in part 4, where I draw on the phenomenology of psychopathy to argue that even very simple systems may experience something like pain.

Where I Stand

For transparency, mine are all the (b) answers. I feel closest to antifrustrationism and assign nematodes a low but non-negligible probability of sentience. I don’t urgently want to create new happy nematodes, but I strongly want to prevent future generations of likely-suffering ones from being born.

This means I’m looking for interventions that reduce the population of small, short-lived invertebrates without causing additional suffering in the process. As we’ll see in the coming posts, this is a more constrained optimization problem than it might seem – and AI may change the landscape entirely.

Background: What Is Welfare Biology?

If this whole field is new to you, here’s a brief orientation.

Welfare biology is the study of the welfare of wild animals, introduced as a concept by the economist Yew-Kwang Ng in his 1995 paper “Towards Welfare Biology: Evolutionary Economics of Animal Consciousness and Suffering.” Ng argued that evolutionary dynamics – particularly the r-strategy of producing far more offspring than survive – imply that suffering probably dominates happiness in nature. Most organisms that ever live die young and painfully, and evolution has no reason to make death anything other than agonizing.1 I sometimes focus on the subfield of welfare ecology, which strikes me as even more neglected.

Brian Tomasik developed these ideas extensively, most notably in his essay “The Importance of Wild-Animal Suffering” for the Center on Long-Term Risk (CLR). Tomasik argued that “the scale of brutality in nature is too vast to ignore” and that the animal advocacy movement needed to expand beyond farmed and laboratory animals.

In a follow-up analysis, “Humanity’s Net Impact on Wild-Animal Suffering,” Tomasik used defaunation studies – the Living Planet Index and Dirzo et al. (2014) – to estimate that the average human prevents ~1.4 × 10⁷ insect-years per year through habitat destruction and environmental impact. This suggests that, on balance, a larger human population reduces wild-animal suffering, and that reductions in wild-insect suffering likely outweigh the suffering of farmed animals even when weighting by neuron count.

However, Tomasik himself flagged a significant tension with these estimates. In his analysis of Krausmann et al. (2013) on human appropriation of net primary production (HANPP), he found that NPPeco – the NPP left over for wildlife – was “almost unchanged between 1910 and 2005” and may even have increased since ~1970 due to CO₂ fertilization. This is “difficult to square with findings that indices of both vertebrate and invertebrate abundance have declined by roughly half since 1970.” Tomasik argues that NPP may be “a more stable measure of wild-animal suffering than an index of animal populations,” because defaunation indices can overstate decline through sampling bias toward declining, larger-bodied species and because they may underrepresent “small and ‘boring’ animals like krill, springtails, and rotifers.”

The upshot: Humanity almost certainly reduces some wild-animal populations (especially vertebrates and flying insects), but whether total animal metabolism – and therefore total suffering – has declined is genuinely uncertain. This tension is important to keep in mind throughout this sequence: The land use effects I discuss in parts 2 and 3 are real at the biome level, but their aggregate global impact may be partially offset by CO₂ fertilization and other factors that increase NPP elsewhere.

Rethink Priorities (RP) has done rigorous work on invertebrate sentience and moral weights, estimating welfare ranges for various species based on neuroscientific evidence. Their moral weight project produced estimates that are widely used in EA cost-effectiveness analyses – including the welfare range of 6.68 × 10⁻⁶ (relative to humans) that Vasco Grilo extrapolated for nematodes. He deferred the caculation to Gemini 2.5, so the derivation is opaque, and the result should not be taken literally.

The Wild Animal Initiative (WAI) conducts and funds research on wild animal welfare, focusing on building the academic field. Especially the work of Simon Eckerström Liedholm has stood out to me as highly relevant.

Most recently, Vasco Grilo’s series of posts on the EA Forum – “Cost-effectiveness accounting for soil nematodes, mites, and springtails” (June 2025), “Animal farming impacts soil nematodes, mites, and springtails hugely more than directly affected animals?” (June 2025), and “More animal farming increases animal welfare if soil animals have negative lives?” (October 2025) – has brought soil fauna into the center of EA cost-effectiveness debates. His key finding: the effects of almost any intervention on soil nematodes, mites, and springtails are orders of magnitude larger than the effects on the intervention’s intended beneficiaries. I’ll unpack his analysis in detail in parts 2 and 3.

And Bentham’s Bulldog, building on Tomasik’s earlier work, has argued that humans in the aggregate likely reduce long-term wild animal suffering by decreasing ecosystem productivity – a position that aligns with much of what this sequence explores, though the HANPP evidence urges caution about the magnitude of this effect.

Next Up

In part 2, “Soil and Sea,” I’ll present the empirical landscape: how many organisms are where, what drives their population sizes, and why the difference between reducing NPP and applying pesticides matters enormously from a welfare perspective.

1. Diagnosis as a turning point. Dorian was diagnosed with antisocial personality disorder (ASPD) at 22, which set him on a path from factory work to studying psychology at university – with the goal of becoming Dr. Dorian Steel.

2. The glass box. Dorian describes his relationship with humanity through the metaphor of a glass box – not feeling trapped or isolated, but simply never having felt connected to other people. He’s comfortable in it.

3. Anger as the only real emotion. Of all emotions, anger is the only one Dorian is truly in touch with. He has no prosocial emotions – he can’t feel happy for someone else – but he can be friendly and has a good sense of humor.

4. No fear, no worry, no lost sleep. Dorian nearly fell off Snowdonia with no change in heart rate, faced a potential nine-year prison sentence without anxiety, and has never lost a wink of sleep over anything. Where other people have fear, he has adrenaline and a cold, analytical mind.

5. An overnight transformation. One night, lying in bed, Dorian said a one-sentence prayer – and the next day his entire life changed direction. He left behind the skinhead identity he’d held for five years, covered up his throat tattoo, and began building a new life around fitness, faith, and responsibility.

6. Control as a core need. Dorian describes a dominant need for control and power over everyone and everything – his environment, the people in it, outcomes. He manages this partly through compartmentalization and partly by carefully curating his surroundings.

7. Boundaries are invisible. Without an internalized moral conscience or empathy, Dorian doesn’t see boundaries until they’re pointed out to him – and by then it’s usually too late. He’s tried to get therapy for this, but has been turned away as “too high-functioning.”

8. Fatherhood is the exception. His five-year-old son is the one person Dorian calls “a little extension of me” – and their relationship is probably the only one where he feels somewhat bonded.

9. Regret, loneliness, and self-pity are absent. Dorian has never experienced deep regret, can’t feel lonely because he has no innate desire for human connection, and can’t feel sorry for himself because he can’t feel sorry for anyone.

10. Pain is just a signal. Dorian spent a full day lifting heavy bags and shopping while his appendix was inflamed, only going to the emergency room in the late afternoon. Pain doesn’t fill his perception – it’s something off to the side that he can mostly look away from.

11. Romantic rejection hits differently. Despite not feeling attached to anyone, romantic rejection causes intense narcissistic injury – a wound to ego and pride, not to connection. Job rejection, by contrast, doesn’t register at all. (My guess is that he’s a sovereign.)

12. The dream of the shrinking doorway. In a striking dream, Dorian meets an idealized woman, but the doorway they keep passing through shrinks until she can’t fit. When they finally get home together, the doors are wide open – and he wakes up flooded with adrenaline. He interprets it as being about emotional capacity and vulnerability.

13. The video game analogy. Dorian resonates with the idea that life with psychopathy is like a video game – you’re running around causing chaos, and you always reset the next day. Other people feel like NPCs, with his son being one of the few exceptions.

14. A message to the public. Dorian wishes people would stop equating psychopathy with serial murder and sex offenses, step away from the Netflix dramas and TikTok myths, and actually read or listen to people who live with it.

Glossary

Hare Psychopathy Checklist (PCL-R). A clinical assessment tool developed by Robert Hare, used to assess psychopathic traits. It covers interpersonal, affective, lifestyle, and antisocial dimensions. Dorian’s prison psychologist administered this checklist during his time on probation. Please see my sequence on psychopathy for more information on the concept.

Narcissistic injury. A perceived threat to a narcissistic person’s self-esteem or self-worth, often triggering disproportionate anger or shame. Dorian experiences this specifically in the context of romantic rejection. Please see my sequence on pathological narcissism for more information on the concept.

Splitting. A defense mechanism in which people or situations are perceived in all-or-nothing terms – entirely good or entirely bad – with little middle ground. It’s a feature of borderline and psychotic personality organization. Dorian notes that his splitting tends to be permanent, which is characteristic of NPD.

Conduct disorder. A childhood and adolescent behavioral disorder involving aggression, destruction of property, deceitfulness, and rule violations. It is often considered a precursor to ASPD in adulthood.

Oppositional defiant disorder (ODD). A childhood disorder characterized by a pattern of angry, irritable mood, argumentative behavior, and vindictiveness. Like conduct disorder, it can precede more severe personality pathology.

Cognitive behavioral therapy (CBT). A structured, goal-oriented form of psychotherapy that focuses on identifying and changing unhelpful thought patterns and behaviors. Dorian underwent CBT with a private psychologist.

Attachment styles. A framework describing patterns of relating to others in close relationships, typically categorized as secure, preoccupied (anxious), avoidant (dismissive), or disorganized (a combination of anxious and avoidant).

Compartmentalization. A psychological defense mechanism in which conflicting thoughts, emotions, or experiences are kept separated to avoid discomfort. Dorian describes being very skilled at this – putting feelings “in a box” so that others won’t see them.

Anattā (no-self). A concept from Buddhist philosophy referring to the absence of a permanent, unchanging self. In the interview, Dawn references Jamie’s description of a “no-self state” – a subjective experience of not identifying with one’s own body or personhood – which contrasts with Dorian’s strong sense of self.

Timestamps

• 0:12 How Dorian Steel Has Been Doing

• 0:56 From Factory Worker to Psychology Student

• 3:18 Starting “The Picture of Dorian Steel” on YouTube

• 4:58 Factory vs. University – Two Different Worlds

• 6:52 Friendship with Jamie (M.E. Thomas)

• 9:19 “I’ve Come a Long Way” – What Dorian Wants People to Know

• 10:23 Where Does the Motivation Come From?

• 13:50 Finding Christianity Overnight

• 20:57 Tattoo Stories – Barbed Wire and Impulse Bets

• 22:16 What Does Regret Feel Like? “I Don’t Know”

• 23:17 The Glass Box – Never Feeling Connected to Humanity

• 27:10 Anger: The Only Emotion That Arrives in Full

• 27:59 Emotional Regulation – Internal Control vs. Changing Your Environment

• 31:40 In Custody – Napping, Push-Ups, and a Handshake Through the Cell Door

• 36:21 “Slipped into a Skin Suit” – Growing into Himself

• 39:25 Grandiosity – “You Sit Like You’re Better Than Me”

• 41:54 Different Faces – Probation Officers, His Son, and Jamie

• 43:40 The Prison Psychologist Who Earned His Respect

• 47:54 Boundaries – “I Just Don’t See Them”

• 53:35 Why Pity Makes Him Angry

• 54:34 Therapy and Addiction – A Troubling Pattern

• 57:20 Depression and Self-Harm at 16

• 58:55 Agency and Impulsivity – Buying an Illegal Firearm on a Whim

• 1:04:50 Fear – Nearly Falling Off Snowdonia with No Change in Heart Rate

• 1:07:46 “Where Other People Have Fear, I Just Have Adrenaline and Anger”

• 1:13:21 Pain – Lifting Heavy Bags All Day with Appendicitis

• 1:15:08 Relationships, Attachment, and Not Feeling Attached at All

• 1:18:39 Intimacy as a Power Game

• 1:20:11 The Dream of the Shrinking Doorway

• 1:27:10 Romantic Rejection and Narcissistic Injury

• 1:32:21 His Son – “A Little Extension of Me”

• 1:36:00 Life as a Video Game – NPCs and the Glass Box Revisited

• 1:38:53 Bad vs. Out of Control – How Others See Him

• 1:40:51 Childhood Amnesia – 90% of Memories Gone

• 1:44:38 Selfhood – Comparing Notes with Jamie’s “No-Self” Experience

• 1:48:06 Funniest Stories – Chasing a Burglar in His Underwear with a Knife

• 1:56:53 Advice to His 15-Year-Old Self

• 1:57:36 What Does Dorian Steel Actually Enjoy?

• 1:59:41 “Stop Thinking We’re All Serial Murderers” – Closing Thoughts

Transcript

Dawn: Hello, flitterific listeners! Today I’m joined by Dorian Steel. Dorian can tell us a lot more about, in particular, psychopathy. So, how have you been doing?

Dorian: I’ve been doing well. I’ve just enjoyed finishing reading the paper that you wrote about me. University’s going well, and I’m more on the up. I find that the more I stick to my responsibilities, the more I’m on the up, so I’m sticking to that. But I’m very well, thank you. How are you doing?

Dawn: Yeah, cool – that would have been my next question. I can have very interesting times. I have a lot of very emotional conversations with some friends right now. Some of them have, some of them used to have personality disorders, so there’s always something going on. But my next question would have been exactly how your degree is going and what has drawn you into that field in the first place.

Dorian: Well, at the moment I’m on my second semester. We’ve just finished the first one. We’re studying developmental, clinical, and investigative psychology. I’ve got classes for biological psychology as well, and research methodology – I like research methodology a lot.

Dawn: And what drew you into that field? In particular, why is research methodology your favorite?

Dorian: What drew me into psychology? It sort of happened without my realizing. Years ago – I’m 25 now – but when I was early 22, I started to suspect that there was something amiss with me. I went down a rabbit hole, educated myself, and then eventually, months down the line, after a psych eval, I got diagnosed with antisocial personality disorder (ASPD). From then on, I just started learning about psychology in general.

I was a blue-collar worker at a factory, in a food production environment. Anybody who knows that sort of work knows it’s not the most intellectually stimulating. So it took about two years and some kind words from a couple of professionals who put the seed into my head about maybe pursuing higher education. I applied for three universities. I got knocked back by two of them, but my last one – which was the best choice – accepted me. I realized all I do is read up on psychology. It became an obsession, and I realized I’m very smart and could make a good career out of this. That’s what’s led me there. But I’ve needed a nudge.

Dawn: Yeah, I can empathize with that – I find psychology very thrilling myself. You’ve also been very open about your experiences and your whole backstory, and you have your own YouTube channel where you share a lot of stories and details. What motivated you to be so open in public about it?

Dorian: There comes a time when everybody wants to tell their story in one way or another. For me, the idea to start a YouTube channel was suggested to me. I’ve had things suggested to me – that I should write a book about my life and the things I’ve done. But a YouTube channel was the way to go. I thought of the idea back when I was working in the factory, but I didn’t want any of my fellow workers to find out, so I held back and waited until I left and started at university.

It’s called The Picture of Dorian Steel, and it’s going very well, actually. I just did a video that’s got about 1,200 views, and I only started last year in September. I tell stories, give my thoughts and perceptions on various ideas and concepts about myself or about other things in the world. I paint a picture of myself, essentially.

Dawn: What’s different between the people you’re studying with now and the factory workers – that back then you didn’t want to start the channel, and now you do?

Dorian: I’ve got a lot more in common with the people in my class. Obviously, we’re all interested in the same thing. The factory was a very volatile kind of place at the time. Very poor communication between all levels of staff. Don’t get me wrong, there were a few good people in there that I genuinely liked and got on with. It’s just like anywhere you go in life, really. But it was very physical – a lot of heavy lifting. I used to tell myself it’s an extra workout that I’m getting paid for.

That’s how you get through that kind of job, but there are people who’ve been there for over 20, 30 years. Most of their life. And I just thought, I’ve got to do something else other than this. I owe a lot, really, to my diagnosis and to psychology as a whole for getting me out of it.

Dawn: Well, at least the people who do that for 20 years must be jacked by now.

Dorian: Yeah, some of them. But I’m just happy to be out.

Dawn: You’re also really good friends with Jamie Thomas. How did you meet? What is this friendship like, and how did it happen in the first place?

Dorian: I can’t remember exactly how, but I found out about her – bear in mind, I’d first discovered it when I was self-diagnosing. That was when I first saw her interviews on a YouTube channel called Psychopathy Is. They do nonprofit research into individuals with psychopathy. So I saw Jamie’s interviews on there, and one way or another I found out that she had a Discord. Somebody put a message in her server one day saying, “Ted, you should do an interview.” And I said, “If I’m lucky, she’ll ask.” And then Jamie herself responded, saying, “Well, we’ll set up a date – let’s do it.”

I think we’ve done five appearances together on her show. We’ve ended up meeting up in person. We went on holiday together last year in Windermere – me, her, and her partner Aria. It was nothing but good times. I’ve still got our pictures blue-tacked on my wardrobe.

Dawn: Between Jamie and Aria, in terms of your particular psychology, who do you think you’re more similar to?

Dorian: Oh, Aria, 100%. Jamie has described herself as sort of a milking stool – she’s got the antisocial traits, she’s autistic as well, and maybe obsessive-compulsive disorder. But Aria, her partner, is very much the same as me. Very strong in antisocial and narcissistic personality disorder traits. Her girlfriend is like me, but as a little blonde American girl. That is the way to describe it.

Dawn: If you had only about five minutes to describe what you want other people to know about you, what would you tell them? What would you focus on?

Dorian: I’d say I’ve come a long way. I’m much more responsible now, for the most part. Back when I was younger – in school, even as a teenager and as an adult – if you asked me what I wanted to do with my life, I never had an answer. No real sense of responsibility toward myself. Now I have a plan for life, future goals, and principles. Regimes that I stick to. I do a lot of reading, a lot of strength training, a lot of writing. I’ve got a creative hobby. I’m doing YouTube now. I’m keeping myself busy with good things, and it’s all looking up from here.

Dawn: Where does the motivation come from to make those changes?

Dorian: To really start taking things like fitness seriously and healthy habits – it started years ago. I just had a switch in my head flick. I was what you’d call a skinhead for five years, since I was a teenager, and that was very tumultuous at times, as you can imagine. Then five years into it, I’d just turned 22, and I looked at myself one day and thought, “This isn’t doing anything for me anymore.” So I snapped out of it, grew my hair out, and I haven’t regretted it for a day.

I like making progress on something. I’m extremely physically fit and healthy. I’ve just had my appendix taken out a few weeks ago, and the doctors were saying I was going to need all these pills. They tried to give me codeine, even though I told them I’ve had previous substance abuse issues with opioids and all manner of hard drugs. So I refused all my medication. They said don’t work out for the next six weeks, and I’ve already started again. Taking care of yourself benefits you. But it’s all about keeping myself rewarded – I’ve got to be getting some sort of reward from whatever I’m doing.

And it’s nice staying out of trouble, I’ll tell you that much. It’s nice knowing that I’m just being responsible and good now. A bit more prosocial, kind of like Jamie.

Dawn: Can you already tell if your life is a sequence of phases like that? Imagine you do a bachelor’s, a master’s, or a PhD in psychology, and that chapter ends at some point – what would the next step be for you?

Dorian: At the moment, I’m doing my bachelor’s degree with honors. After that, it’ll be a master’s degree, and then – because I want to be Dr. Dorian Steel, I want to have “Doctor” before my name – I’ll be doing my PhD in psychology.

Once I’ve done that, I think I’d like to go on the NHS as a diagnostician for a few years, and who knows, maybe even work with some of the people who diagnosed me. I remember some of them. I remember Nurse Anna, and the head doctor of the department. They were all very professional, and I remember leaving and admiring them. So I’d definitely want to do that for a couple of years, and then probably work in inpatient facilities as the end goal. I’m not sure about being a therapist, though – but definitely working at either a hospital or an inpatient facility.

Dawn: Part of the current reset in your life is also a shift toward Christianity in particular, and that’s something I found very interesting. How did that come about? I can see you have a few different versions of the Bible behind you.

Dorian: Yes, that’s right. I was alone in my bedroom one night, and it’s hard to put a word on it, but I just had an internal shift – this internal experience of myself. I’ve heard other Christians describe it. I just knew I had to admit the fact that there was a God.

So I started reading the Bible every day. I started writing scripture every day as well – anything I read that stuck out to me. I’ve still got my pad in my bedroom, just pages and pages of the Bible. I went to church about two or three months into being a Christian, and the church was good for a while. I even got baptized there. But I had a disagreement with the pastor on their policy regarding sex offenders being allowed to participate in the congregation, and I just left.

Now what I do occasionally – I’m what you’d call an evangelical Christian, somebody in the Billy Graham line of faith – but my dad is a Jehovah’s Witness, so I’ll go to one of the local Kingdom Halls with my dad occasionally on a Sunday. They’re very welcoming of me. They know I’m not a Jehovah’s Witness, and I don’t agree with them on everything, but that’s as good as it gets for me in terms of worship with other people. I still like to do my own thing. But that’s always been me – I’d just rather be on my own.

Dawn: When I wanted religion, I also created my own. But what did it feel like to realize – or acknowledge, the word you used – that there is a God? Was it a feeling, or was it a logical conclusion that you arrived at?

Dorian: It was a feeling. I’ll never forget it. I think I was out of work, had no money, nothing. I was at the lowest of the low. Not stressed, but definitely always thinking about things. It was slightly nervous, but I remember I said a prayer – just a short, one-sentence prayer to myself.

When I went outside the next day, everything looked different and felt different. It’s really indescribable. I’ve never been able to speak in too much detail about it, but it was an internal physical feeling, an internal shift. But I’m always in that state of what we’d call wrestling in the flesh, really wrestling with myself at times.

Dawn: So you prayed for something? Was that related? Did you pray for something, and then it came true the next day in the form of this internal shift?

Dorian: It wasn’t a wish. It wasn’t like, “I want a new car” or “I want a new wife.” I wasn’t asking for something. I was just lying in bed, and I closed my hands together, closed my eyes, and said, “Okay, God, you’re real, I believe you.” And that was it.

But being out of church has affected that a lot. I do think you need a good church behind you, but the church is its own worst enemy. It’s a toughie.

Dawn: Can you try to describe more what the difference was like before and after that shift – in your perception of yourself and the world?

Dorian: In terms of behavior, that did a complete 180. I was unemployed – I was earning money outside of employment, let’s say. I was making ends meet in other ways. And I was extremely objectifying of people, in the sense that people were just literally objects to be used and manipulated for my purposes.

Then my life completely turned around. I wasn’t really dressed that smart. I started dressing smart, taking myself seriously, getting up at six in the morning to work out, getting healthy hobbies, and trying to be a good Christian as best I can. That hasn’t always been the case, and I’m not saying I’m perfect – I know I’m not. As much as my brain might like to tell me otherwise.

Dawn: That’s a remarkable shift, and that it happens so suddenly is impressive.

Dorian: It was at the snap of the fingers.

Dawn: I could imagine someone gradually developing over the course of a couple of years in that direction and then trying to remember what it was like before. But having that from one day to the next is stark.

Dorian: It was very late at night. Pitch black, and I was in bed. I’ll never forget it. The next day, I went and got my skinhead tattoo across my throat covered up. I couldn’t walk around with that anymore. Things started to change and go better, because I put my faith in a higher power. It’s the best decision you’ll ever make.

Dawn: You just touched on some tattoos. Are there any other stories behind any of the tattoos that you want to talk about?

Dorian: I’ve got a funny one. The barbed wire on my forehead – that was an impulse. I’d won money betting on horses. There used to be an old betting shop near where I live – it’s since closed down – but I went in there and bet a couple of pounds on a horse that was something like 100 to 1, or 200 to 1. And it won. So I took the money and went straight to the tattoo studio and said, “I want this done.” Some of them are very impulsive, but I’ve never regretted any of them. It’s a big thing with skinheads – tattoos. It was all about looking tough and all of that. But they’re just there now. They don’t mean anything to me.

Dawn: What would it feel like for you to regret something?

Dorian: I don’t know. I know this isn’t true, but in my head, it sounds stupid – why would you sit around regretting something? What are you going to get out of that? I don’t think I’ve ever felt regret. Maybe once or twice in a moment I’ve gone, “Yeah, maybe I could have gotten more if I chose option A instead of B.” But that actual deep-felt regret – I’ve never had it. I can’t even really talk about it. I just don’t have any experience with it.

Dawn: You’ve used this interesting metaphor of a glass box, or a plexiglass box, to describe your relationship with the environment, the world, and other people. I’d be quite interested in learning more about that – whether that relates to the previous version of you more than the current version, or whether it’s been constant. And what it feels like in the first place.

Dorian: I think it’s more a representation of what I don’t feel, if that makes sense. It’s always been there, always been the same. I’ve just never truly felt connected to humanity as a whole. People are people, and they’re there. But they’re always a little bit different in a way, and I just don’t identify with 99% of them. But I’m not particularly uncomfortable in my proverbial glass box. I’m very much fine in it. That’s just how I see things.

Dawn: Do you feel more isolated, or trapped, or protected by this box?

Dorian: Neither. I don’t feel loneliness, because I have no innate desire for human connection, so I can’t feel lonely. If I want to know if I’m lonely or not, I have to go through a checklist – right, I don’t have any friends, I spend a lot of time on my own, therefore I’m lonely. I don’t have a feeling to tell me that.

I don’t feel isolated from anyone, because I don’t feel it, so I’m not. It’s not about feeling protected either. Maybe protected from other people’s feelings, to a degree. I’ve been known to get angry quite quickly if somebody else cries. I don’t really like other people’s tears.

Dawn: That’s interesting, because when I get angry, I usually don’t notice it. I just notice that I start to have passive-aggressive intuitions that I usually don’t have, and then I wonder what’s going on. I didn’t used to be in touch with that feeling. So I wonder whether that’s similar with the loneliness – that there’s something going on that bothers you in some subtle way, but you’re not directly in touch with the feeling.

Dorian: Loneliness is just not something that I have. Anger, however – I’m very much in touch with my anger. It comes on very quickly, very suddenly, in vast amounts. I’ve been known to fight or attack people because of my anger, throw people out of my life because of it. There’s really no telling what would truly happen. So that’s why I try my best to avoid people who annoy me, and try to avoid as many triggers as possible. It’s something you have to do when you’re highly psychopathic, if you want to lead a good life. I don’t want to go to prison, because I know I’ll just be angry all the time and probably very paranoid as well.

The only main emotion that I’m really in touch with is anger, I’d say. I don’t have any prosocial emotions. I can’t feel happy for someone. It kind of sounds silly even saying that as a sentence – why would I feel happy for somebody else? Can’t they just feel happy and get on with it? That encompasses a large amount of my emotional range. I can be friendly, I can talk, I like to have a laugh. I don’t think laughing is an emotion, but I’ve got a good sense of humor.

Dawn: You try to emotionally regulate by surrounding yourself with people who bring out the best in you rather than the anger. That’s something I’ve seen from a bunch of friends of mine. Most of them seem to have an internal feeling of control, as if they can control how they want to react to their environment. Others rely much more on self-regulation through changing their environment. There’s this one person who told me that she told her son to call the police when she’s getting abusive, so that they can rein her in. She wasn’t used to being able to do that herself but clearly wanted to. Do you do both? Some kind of internal regulation and external regulation? Or do you rely more on external factors?

Dorian: I’m very good at compartmentalizing my emotions – or any feelings that I do feel. I’m very good at putting the ones I need in a box, tucked away so that you might not see them. That’s always been something I’ve been very good at.

But managing my environment is just as important to me. One of the key needs of somebody with psychopathy is that you have to always be in control of everything – not just your environment, but the people within it. You have to be in control of outcomes. It spreads into relationships, into every facet of your life – that dominant need for control and power over everyone and everything. So they’re both just as important. That’s probably part of where my ability to compartmentalize stems from – because I have to, otherwise I’m not in control.

Dawn: This control thing is interesting. What makes it so important?

Dorian: I don’t know. It’s just in me. It’s just the nature of who I am. What else would I do?

Dawn: What would it feel like to not have control, but in some harmless way – nothing bad happens?

Dorian: If I’m not affected in any way, maybe my pride will still be hurt momentarily. But if there’s no massive real-world fallout or consequences, then in a few seconds I’ll probably just not care and move on. I won’t linger.

Dawn: You had this interesting anecdote about being in custody, where you probably didn’t feel quite the same level of control as usual, but you managed to claw back a bit of that by getting a policeman to shake your hand through the cell door.

Dorian: Yeah, that was when I was in custody – not prison, but the custody jail where you’re taken after you’re arrested. My last arrest was when I was 22, or 23 actually, and it was for allegedly attempting to smuggle a firearm into the country, with ammunition, and a couple of other charges to do with illegal weapons found in my flat.

I got brought in for questioning, and to be honest, both times I’ve been in custody, I’ve loved it, because it’s just like a game of Monopoly, really. To me, it is, at least. Some people really can’t handle it – you hear them having breakdowns. But not me. I’ll go in, do my push-ups and sit-ups, get some food in my belly, have a nap, and then rinse and repeat until it’s time for my interview.

The last time I was in there, after questioning, I stuck my hand out through the hatch in the cell door – which you’re obviously not supposed to do, because you could be trying to attack an officer. But I got this detective to actually shake my hand through the hatch. It was just fun knowing that I could have broken his hand or his fingers, and having that power over that person, even for a couple of seconds – that’s just a thrill for me.

Dawn: Does it feel more like stoicism – that you can make the best of those situations – or more like indifference? What feeling is dominant there?

Dorian: The feeling is power, ego. In my head, I just feel powerful, and I feel it in my body. In my mind it’s: “Yeah, I’m in control. I made you do something.” When I decide I’m going to make you do something, you’re going to do it. That’s just a way of regaining power and control in that situation.

Dawn: And generally, this whole arrest situation – they’re taking some power away from you initially. I imagine it feels like a game?

Dorian: A little bit at first. The thing is, I’m registered as a violent male with the local police due to past instances. If someone reports me, they check that name, and everything’s got codes. I’m registered as a violent male, so if the police have a warrant for my arrest, they have to turn up with what they call Mariah vans – big yellow vans normally used for riots, with grates over the windshield, riot gear inside – shields, batons – and then normally three male officers will come out and get me. It makes a bit of a scene, because people see it.

At first I didn’t like it, and it was a long drive back to the station. I asked them if they had a radio I could listen to, and they said no, they have to be on the police radio the whole time. But it is what it is. You’ve just got to play the game.

Dawn: You’ve also described feeling as if you were slipped into a particular skin suit with regard to your body and personality. To what extent do you actually identify with yourself now – your body, brain, personality, mind, everything – and to what extent do you feel like you’re just operating it like a vehicle?

Dorian: I feel more like a person now, I’d say, than back then. It’s less like I’m operating a vehicle. It’s me. I haven’t got any qualms about being me. I’m very happy with the person I am, and I always will be.

Dawn: How have you grown into yourself, as it were?

Dorian: Truly, honestly, being responsible and being healthy – it’ll do those things for you. I owe it all to my hard efforts. There’s no particular right way or wrong way, just the way that you need to do it. I know a lot of people really do feel distant from themselves. But that’s not me anymore. I think a big part of that is the fact that I’m doing something I want for a living now, with being in higher education. I reckon that has probably played a significant role.

Dawn: I would imagine that increased identification also comes with – what does aging, for example, feel like for you? Has that become something you think about more now that you identify more with your body?

Dorian: I don’t even celebrate my birthday. A lot of people with narcissistic personality disorder (NPD) will celebrate their birthdays, because it’s their day. But that’s never been me. Since I was maybe 17 or 18, I’ve just never felt any connection to it. I don’t mind celebrating somebody else’s birthday – if you want me there at your party, sure, I’ll come. But I just don’t see my own birthday as important. I turn 25 in November, but it’s not a reason to get a cake.

Dawn: Grandiose behaviors seem to evoke different reactions in different cultures, and I imagine this varies not only by country but also by the particular social context. Did you observe in what contexts it works better and in what contexts it leads to resistance?

Dorian: The thing is, I was never aware of how grandiose I came across as. It wasn’t until it was pointed out to me – first by my father. My dad said to me once, “You act like you’re the only person who deserves to be thought about.” In my head, I’m thinking, “Do I act like that?” I know I think like that a lot of the time, but I never realized I acted like it. Apparently, when it comes out, it comes out in heaps and waves. I had an old therapist once tell me, “You sit like you’re better than me.” All I was doing was sitting, but apparently it’s the body language.

It’s served me well recently at university. People have come up to me and said, “Do you ever get nervous about anything? You’re pretty confident, aren’t you?” I was just voted class representative at my uni – unanimously. Put in the right people, and they’ll put their faith in you. But some people really won’t like it – they’ll go, “Who does this guy think he is?” You get a lot of different reactions, but generally I’ve been told that I come across as polite and intelligent. I always mind my P’s and Q’s, especially around women. But apparently the grandiosity can seep out eventually.

Dawn: If I want to practice my grandiosity, what environments do I need to avoid?

Dorian: I really couldn’t tell you. Half the time I don’t realize when I’m doing it.

Dawn: Are you usually quite consistent in your behavior and personality across contexts, or do you behave differently with, say, a probation officer versus your son versus Jamie?

Dorian: Massively different. My son – it’s just all about having good times with him. Going to the park, going to the movies. He likes to play-fight me all the time, so he gets to punch me in the face a lot, and he loves it. It’s great.

But somebody like a probation officer – I’ve never really liked a lot of them, because they’re always someone who’s got their thumb over you for a whole year. I’ve had a history of cycling through them. New probation officer after new probation officer, because they all didn’t want to deal with me.

Being with my son isn’t a power game, because he’s my five-year-old son and he respects me anyway, because I’m his dad. That’s just a given. But with someone like a police officer or a probation officer, it’s very different.

Dawn: And with Jamie?

Dorian: We’re just friends. It’s quite simple – we message each other occasionally, and it’s all good.

Dawn: I think I heard in some interview that one of your probation officers, or someone related to that, was actually a highly qualified expert in the field. Am I remembering this correctly?

Dorian: Kind of. He’s not a well-known expert – he’s a prison psychologist. To explain for people who are reading: you’ve got His Majesty’s Prison and Probation Service, which splits into two – the prison service and the probation service. A lot of the time, when they get criminal offenders like myself who might have a personality disorder or be really difficult and noncompliant, they’ll get a psychologist from the prison service and bring them in to help. That’s what happened with me.

The thing is, with probation now, you have to declare all your social media accounts. Your YouTube, your email, your Instagram – even old accounts you don’t use anymore. If you haven’t declared everything, you risk going back up for resentencing. At the time, I didn’t have a YouTube channel, but I had my Instagram blog where I write about living with psychopathy. They found that, and then they all said, “I’m not working with him. No way.”

So that’s why they brought in Matthew – the psychologist. He’d done work with pretty high-profile prisoners. He’s based at a prison in Liverpool, but they brought him down to probation. I actually ended up really liking him in the end, but at first? No. I hated him. Absolutely hated him. The first time I saw him, he said he could only speak to me for 10 minutes, and I kept him in the interview room for over an hour. I just played mind games with him – funny looks, staring at his forehead and looking back down. Any little trick in the book, just to put him off and let him know I’m in charge.

But eventually that fizzled down, and I grew to respect the guy. He was really good. They ended up doing it where, for my weekly meetings, he’d bring in trainees – people who were training to be probation officers. One or two of them would sit in the interview room each week and observe me and Matthew talking. We did talking therapy. He was the one who administered the Hare Psychopathy Checklist on me, and because he’s a probation officer, he’s got access to all my previous arrest history records and all of that.

Dawn: Did it feel different at the end of your interactions, when you respected him, compared to the earlier ones?

Dorian: Don’t get me wrong, he was an alright guy, but I still didn’t like being on probation. The best feeling was when it was over. I said to him, “I’m not going to see that purple welcome sign” – he had a big purple sign saying “Welcome” at the front desk. I said, “I’m not looking at that again. I’m done now.”

Dawn: You’ve said somewhere that you have a hard time knowing what boundaries are in the first place, where they’re supposed to be, and that of course makes it difficult to respect them. What do you think is the reason for that? For some people it’s obvious, for others certain boundaries are really difficult to anticipate. I’m quite puzzled myself, because there seem to be cultural differences too. How do people pick up on boundaries throughout their maturation process, and why is it more difficult for some than for others?

Dorian: I’ve read up on this, because I did a YouTube video on boundaries recently. The thing is, boundaries come from having an internalized moral conscience. I don’t have any feelings in me that tell me it’s right or wrong to lie to you, take advantage of something, or put you in harm’s way. Along with that comes not having empathy, not having any awareness of boundaries, and not feeling like I should feel guilty or sorry about anything. I just don’t see them until they’re pointed out to me, and normally by then it’s too late.

I’ve tried to get therapy over it, but nowhere will deal with me. Even the hospital that diagnosed me has said they won’t have me back because they deem me too high-functioning. And I’m an unemployed student at the moment, so it’s not like I can afford private therapy.

Some boundaries I can guess. But I will tend to cross the lines eventually, sooner or later, and that can have repercussions. I remember when I was 15, after I assaulted another kid in school, I had to have anger management therapy of some kind. This woman used to visit from some sort of center that dealt with troubled kids. I used to flirt with her. I was a 15-year-old boy – I get now that that wasn’t okay, but at the time I was just loving it. Boundaries – I just don’t see them to begin with. And also because I’m very entitled, so why would there be a boundary if I’m entitled to that anyway?

Dawn: But I feel like you’re aware of some boundaries that I wasn’t aware of until a couple of months ago. For example, I always appreciate it when I share some story from my past and other people express sympathy. I enjoy that. Whereas I think you would rather feel patronized by that. I was thinking of the thing you shared on Facebook, where some university therapist expressed sympathy for something that happened to you in your childhood, and you felt patronized. I found that very interesting, because I wouldn’t have had this reaction. But of course I want to treat my friends the way they want to be treated.

Dorian: Kind of, yeah. Obviously there are very obvious boundaries that I understand. I’ve never been physically abusive toward any girlfriends, things like that. But it’s more that once you’re in my life, I sort of feel like I own you – you’re more like a possession than an actual person, and then I can just do what I want.

Dawn: Where do you think our difference comes from?

Dorian: I’m not really sure. But I remember the thing about the therapist. I don’t like pity being expressed to me. I don’t really like people feeling sorry for me. It just puts me in a situation that irritates me even more. When that therapist said, “I feel sorry for the child Dorian,” I just thought – what is that going to do? It didn’t annoy me until later that night. I just kept replaying it in my head, and I really got angry about it. But it’s over now. It was just a comment made by another therapist.

Dawn: I’m also wondering about therapy in general. I think at one point you observed that some addiction problems – maybe alcohol – were exacerbated by therapy. What’s the mechanism behind that?

Dorian: No idea. I absolutely haven’t got a clue, but I noticed a very distinct pattern. When I was doing talking therapy with my old probation officer, my drinking spiked. And then again, when I was doing cognitive behavioral therapy (CBT) with a private psychologist, my drinking spiked again, and there was drug use on top of it. No idea why. I don’t know why, but I know it happens.

Dawn: Do you think it was something circumstantial – like maybe the practices happened to be next to a bar?

Dorian: No, no. I’m not big on going to bars or pubs. I’ve always been more of a solitary drinker – I’d rather buy a bottle of whiskey from the shop and bring it back here. Honestly, I’m sure there is a reason, but I just don’t know what it is.

Dawn: You mentioned you don’t like pity or being pitied. Has there ever been a self-pity type of phase that you want to leave behind, or has that just never happened?

Dorian: Self-pity? No. When you can’t feel sorry for anyone, it also means you can’t feel sorry for yourself. I don’t feel sorry about anything that’s ever happened to me. That’s just a waste of time to me.

Dawn: Have you had phases where you wanted nothing to do with society – to just isolate yourself somewhere away from people?

Dorian: Yeah, definitely. Especially when I was younger. When I was 16, I had a very short – two, three months – but very strong bout of what they call major depressive disorder. I was extremely withdrawn from everyone. It was very noticeable. I was self-harming all the time. Not really eating well. But in general, I’d rather keep to myself. It’s more a preference than an absolute need, if that makes sense.

Dawn: In what way were you self-harming?

Dorian: I was just really, really depressed and down, constantly. I’ve often heard that children with conduct disorder or oppositional defiant disorder (ODD) can go through either an extreme period of depression or anxiety – normally one of the two. And my brain, for whatever reason, just went the depression route. It was physical – I used to steal disposable razor blades and take them apart with a pair of nail clippers.

Dawn: The next section is more about agency and intentionality. You already mentioned the case where you were charged with importing a weapon – a thing called a lifeguard.

Dorian: Yeah, it was a disguisable handgun with two or three rounds of ammunition.

Dawn: I was curious what decision-making itself feels like for you. When you’re on Facebook Marketplace and see a lifeguard on offer that’s cheap – does it feel more like you’re watching yourself make the decision, or do you feel like you are the one making this decision, that you are choosing?

Dorian: Yeah, it’s me making the decision. I’ll just do really risky things on impulse. Quite often. But especially when I was younger – now I’m very mellowed out, so to speak. But back then, things like that, it’s not a big thing. It’s just an impulse thing, and I’m the one doing it. That’s all there is to it.

Dawn: What about minor decisions, like whether to get a coffee or a tea – what does the decision process feel like for you?

Dorian: Just normal. If I want something, I’ll think of a way to obtain it. If I want the coffee, I’ll go get the coffee. It’s not really a big thing for me.

Dawn: Some people have this feeling that something in their brain discreetly makes the decision for them. For me, it feels more like I weigh all sorts of factors – do I feel like drinking matcha now, or my espresso? – and simulate what that would feel like and what effects it would have. And then my actions just flow automatically from that ranking process. Does either one of those resonate with you?

Dorian: I don’t go through any of that. It’s very straightforward for me.

Dawn: Say you chose the coffee, and then someone asks you why you picked it. Where would you go to figure out why? Would you study your past behavior, or would you introspect on some feeling of tiredness, and that the teas here don’t have caffeine, and so on?

Dorian: I would just say I wanted a coffee and I got it. That’s not really anything to get deep about with me. If I see something, I want it, I get it. I’ve got my reasons, and my reason is that I wanted it.

Dawn: Here’s a concrete decision – you closed the viewfinder to save battery. The alternative would have been to sacrifice battery life and be able to see me. If someone was curious about how you made that decision, where would you look for the answer?

Dorian: It’s just a logical decision I had to make, because I need to see you and talk through here. It’s just all logic.

Dawn: Do you sometimes make a different decision than the one dictated by the logic?

Dorian: No, not really. I’m a very logical creature.

Dawn: Fear is a big one for me. I imagine it’s not such a big thing for you.

Dorian: Why do you feel a lot of fear?

Dawn: Throughout my life, it was the dominant emotion for most of my adolescence, and then I got better. I don’t know what your relationship with fear is.

Dorian: Very minimal. Not a lot of it. Even going back to when I was a kid, I think I got conditioned into a mild arachnophobia because my mother has a very severe one, and I was always the one who got the spider out. But I can be around spiders now and I’m fine.

Fear as a whole? No, not really. I remember when I was 15, climbing Snowdonia in Wales with a friend and her mother. I nearly slipped off the peak of the mountain. It’s all icy, and there was nothing to grab onto, but my heart rate didn’t change. Nothing happened physically. I’ve overdosed multiple times, even before I was 18. There were times where I thought I was going to die, but I was never fearful of it. I never had that sensation.

Or when I looked up the charges for the illegal firearm and weapons – even seeing that I was probably looking at 9 to 10 years in prison, it didn’t scare me. It was just, “Alright, I’ll need to do A, B, and C.” I got all my stuff in my flat packed up. I never lost a wink of sleep. I’ve never, ever missed out on sleep over anything.

Dawn: So you have problems with impulsivity when it comes to anger, but situations that would induce fear in other people just leave you neutral?

Dorian: Neutral, or maybe my adrenaline might be spiked, but that’s not going to stop me from thinking logically about the situation and the best way to handle it. My anger is definitely my Achilles’ heel. It’s gotten me into some trouble over the years.

Dawn: Imagine situations that for most people would combine both fear and anger. If someone attacks you, the average person would probably be afraid of the physical danger and also angry that someone would attack them. I imagine you only have the second?

Dorian: Yeah, pretty much. Where I live, it’s a very deprived, high-criminality area, so I’m always on my guard anyway. But it’s more the latter for me. There’ll be no anxiety – just adrenaline. Where other people have fear, I just have adrenaline and anger, and a very clear-cut, cold mind to analyze the situation.

The most people I’ve ever fought in one fight on my own was eight lads – what we call scallies over here. One of them tried throwing something at me, and I just ran into them and started fighting. Again, it’s that reaction: who does this person think they are? And I won, too. The men in my family, especially on my dad’s side, have a knack for having a good fight. It’s a lot to do with genetics, but pride and ego as well. I’ve just never worried about anything, really.

Dawn: In the situation you described – almost slipping off the mountainside – was there a moment of “Oh shit, I’m going to die” and then a moment of acceptance, or was it just nothing?

Dorian: It was just nothing. Especially because I was wearing trainers with no grip on the ice. My friend had a go at me because I kept trying to go right on the edge of the mountain. She said, “Will you move away? You’ve just nearly slipped off, and you’re going to die.” There are no railings or anything up there in Snowdonia – just no safety at all.

My lack of anxiety, fear, and genuinely having this “I don’t care” attitude used to bother my son’s mother. We’re no longer together. It also bothers a lot of people. I might genuinely care about something, but I’m not going to go into a panic or worry about anything.

Dawn: When it comes to fear, I have two different experiences. One is more like a phobia – nervousness and a reflex to run away. The other is catastrophizing, where I don’t feel anything physically, my heart rate is normal, but when I think about the thing I’m catastrophizing about, I have thousands of thoughts about how it could go wrong. Do you only lack the phobic type of fear, or do you also not catastrophize?

Dorian: I won’t overthink too much either. That’s probably partly because there’s no fear there in the first place. I’m a very cutthroat, straight-to-the-point, logical kind of person. I see things for what they are and go from there. There’s no fear, no overthinking, no overplanning. That’s just me.

Dawn: I’ve also been wondering about pain itself. With your recent surgery, for example – does pain feel like literal suffering for you, or is it more of a signal that you can choose to interpret?

Dorian: I can put up with high amounts of pain. I won’t enjoy it. It’s just a part of life, really, isn’t it? It’s a signal, I know that, but it’s nothing too deep. If I feel pain, I’ll try to think of ways to minimize it, but I probably won’t make as much of a fuss as the average person.

Dawn: Does it feel like something that fills all of your perception when it happens, or is it more like something off to the side that you can look away from?

Dorian: I can ignore it for the most part. It’d have to be really bad for it to take over everything I’m doing. Even when I had appendicitis, I woke up with the pain in the morning and I didn’t go into the emergency room until five o’clock in the afternoon. I spent the whole day lifting heavy bags and doing shopping. So there’s a good example for you.

Dawn: Jamie also has a similar story from her childhood. It’s very strange to me but super interesting. I used to have a bouldering partner with a similar attitude. Let’s hop to the next section – relationships and attachment. You mentioned you had a three-year relationship, probably the mother of your son, and it was on-and-off. What caused the offs, and what caused the ons?

Dorian: There were a couple of times where I walked out, but I can’t even really remember why. It will have been a trivial reason. But overall, we were together for 90–95% of it. It is what it is.

Dawn: Do you know your attachment style?

Dorian: What are they? Do they have names?

Dawn: Yeah. My favorite measure is the Attachment Style Questionnaire, short form. There are two dimensions: preoccupied attachment, also called anxious attachment, and avoidant attachment, also called dismissive-avoidant. When someone is high on both, they don’t have one distinct style and just chaotically jump between them – that’s called disorganized. When conflict happens and the person disappears or walks out, that’s associated with avoidant attachment: “I don’t really need this person, I don’t want to deal with this conflict.” With preoccupied attachment, it’s more about eagerness to keep the other person, or testing whether they’re as securely attached as you hope. Disorganized is just a back-and-forth between the two.

Dorian: I don’t really know if I have an attachment style. None of that really describes me. If I want someone in my life – say, a girl – I won’t feel attached to her. So I wouldn’t say I have an attachment style, because I don’t feel attached to begin with. I’ll interact with her in a certain way, but that’s about it.

Dawn: Have you ever tried merging with another person – just sort of adding someone to your identity?

Dorian: No. No interest in it, either. I’m very fine with who I am. If you want to be a part of my life, if you want to merge with my life, we can do that. But I don’t need to merge anyone with my identity. It’s mine.

Dawn: You’ve described intimacy as a kind of power game, and vulnerability as a dangerous weakness in that power game. Where does that come from, and what does it feel like when someone gets emotionally close to you for it to feel threatening?

Dorian: When you say emotionally close, do you mean that they understand me more, or that they trust me more and tell me how they’re feeling more often?

Dawn: I would imagine you don’t feel any vulnerability or threat from me right now doing this interview. But in some relationships, when they get close enough in whatever sense, probably something does start to happen where you feel like it’s a risk.

Dorian: I’d have to say no. I don’t think I’ve ever been through it like that.

Dawn: Is that something you worry might happen?

Dorian: No.

Dawn: There was also a story about a dream of meeting some idealized woman and a doorway, and the doorway keeps shrinking. Can you describe that dream and what you think it represents?

Dorian: It wasn’t a recurring dream – it was just this dream I had once. I had this house that I lived in, and I went on holiday on my own. I met this girl at the hotel, and she had this gorgeous, fiery red hair. Really nice girl. We kept having to go to this other room in the hotel, but the doorway kept shrinking every time. Eventually, the girl said, “I can’t fit through the door anymore – you’ll have to stop.”

So we leave the hotel and make our way back home. She’s going to come and live with me. I end up losing her at the airport, but then we find each other again and go home. Right at the end of the dream, we’re at my house, talking about how we’re going to spend the rest of our lives together. But when I turn around, the front door and the back door are wide open. At that moment in the dream, my heart rate spiked, and I woke up almost ready to fight – filled with intense adrenaline.

I think what the dream meant was that the doorway represented my emotional capacity, or my expectations. The girl couldn’t fit – she wasn’t emotionally compatible, or she couldn’t meet my expectations in a partner. And the door to the house being open, which freaked me out, could have represented emotional vulnerability. That’s my take on it. There’s definitely something to do with my attachment style with women, if I have one.

Dawn: So the first part was about her being able to get to you, or vice versa, and the second part was probably about not her getting out through the back door, but someone coming in.

Dorian: There was no one coming in. It was just me and her, no one else.

Dawn: What would a successful relationship with that woman, or someone else, look like for you? What would the give and take be? What needs would you want met, and what would you be able to give in return?

Dorian: I’m a date-to-marry kind of guy. I want to be married. I want to be a successful doctor. If she wants kids, that’s fine – I already have a son, but if she wants another couple of kids, who am I to say no? Very traditionalist and conservative. That’d be my ideal marriage – somebody who’s accepting of who I am. A girl who can be with a guy who can’t feel sorry for her but can still have a laugh.

Dawn: So you’d be the breadwinner? What would be her role?

Dorian: Primarily a housewife, stay-at-home mother. But if we both want to do things in the community, or if she gets a job or a business, that’s fine, as long as everything else is taken care of.

Dawn: If she has a big business, then both of you can provide financially.

Dorian: Yeah. Well, first you’ve got to go out with the girl, so I don’t think it’ll be happening anytime soon. But in an ideal world, yeah.

Dawn: Why not anytime soon?

Dorian: I’m very avoidant when it comes to women. I get very paranoid and tend to keep my distance. My last girlfriend basically had to crowbar her way into my life to be with me. That kind of gives you the idea. But they say you can meet anyone at any second, so if it happens, it happens.

Dawn: Is there some kind of worry behind that avoidance?

Dorian: Yeah. I realized one day – I think a therapist said it to me – that if you want a successful relationship, you’ve got to compromise and surrender a bit of control. I don’t like the idea of that. But I’m willing to work on it with the right person.

Dawn: What sorts of control would be both likely and difficult to compromise on?

Dorian: Could be anything. All matters, big and small.

Dawn: What does romantic rejection feel like for you?

Dorian: It’ll hurt my ego. I’ll go through what most people would call narcissistic injury. That’s a very big ego wound for me. Which is really weird, especially when you don’t feel attached to anyone. But it’s more a pride thing than anything. I’m very paranoid or fearful of rejection – “fear” is not quite the right word, but that’s what I’m using. I have some borderline traits as well, so those can play up a little bit.

Dawn: What makes romantic rejection so much more difficult than, say, job rejection?

Dorian: Job rejection? I won’t care. It only occurs with romantic rejection. One’s just a job, but the other is a girl that you really like. Apples and oranges.

Dawn: There are all sorts of possibilities for why rejection stings – losing access to something, not getting needs met, feeling unworthy or flawed, losing control. Do any of these resonate?

Dorian: I don’t really know the mechanics behind it, but I just know that’s the way it is.

Dawn: When you do suffer a romantic rejection, do you pine more for the other person afterwards – think about them, think about how much you would have wanted to be close to them? Or is it more about yourself and what it means for you?

Dorian: It’s more about myself.

Dawn: What borderline traits have you noticed?

Dorian: I can tend to split quite easily on people. In some episodes, I can have a very rapid change in emotion – suddenly going from calm to absolutely furious. But then once everything’s calmed down, I’ll be fine within seconds. It’s really weird.

Dawn: When it comes to the splitting, is that something that happens for an hour, a day, a week – or is it usually permanent?

Dorian: It’s usually more permanent.

Dawn: I wouldn’t associate that so much with borderline – much more with NPD. Interesting. When a judge passes some kind of judgment against you, that’s also a kind of rejection in a way. But that probably doesn’t resonate the same way as a romantic rejection.

Dorian: No. When a judge sentences you, it’s not rejection. Unfortunately, you broke the law and got caught. It’s just an unfortunate consequence of life.

Dawn: What is your relationship with your son like? You’ve told me a bit about the play-fighting. Can you elaborate on what that relationship feels like and what you do when you’re together?

Dorian: He’s a young boy, and he’s the spitting image of me. My son is like a little extension of me. I don’t feel deeply bonded to him, but I do really care about him. I always pay his child support on time, I have him every weekend. We’re always doing days out, or sometimes he’ll just want to stay at home – he’s got an Xbox here. We just always have good times. I let him bounce up and down on the sofa, which he’s not allowed to do at his mum’s.

It’s probably the only fully transparent and healthy relationship that I’ve got in my life.

Dawn: To the extent that he’s an extension of you – has there ever been a partner, or anyone else, who felt similar in that sense?

Dorian: No, just my son.

Dawn: Does that mean it feels like his actions reflect on you? His achievements?

Dorian: They can do, in a way. Personality-wise, he’s the complete opposite of me. His mother has always said about our son that he’s the nice version of Dad. He looks like me, but he’s actually nice.

Dawn: He doesn’t have that swinging brick. Where does that metaphor come from? You have a swinging brick where your heart should be, or something?

Dorian: Yeah, that’s something his mother said years ago. This was all way before I was diagnosed. I didn’t get diagnosed with ASPD until a year or two after we’d split up. It’s just a phrase she said to me one day, and it stuck with me.

Dawn: It’s so specific – why is it swinging? We’ll never find out. What would it feel like for you if your son showed similar behaviors to the ones you probably showed around that age, or a little older?

Dorian: I wouldn’t feel deeply about it. His mother would probably get him referred to a child psychologist, and we’d go from there. But I wouldn’t feel bad about it or anything.

Dawn: Some friends of mine describe their role in the world like a video game, where they control a player character running around completing quests, and all around them are non-player characters (NPCs) they can interact with and manipulate. It makes virtually everyone in that world feel very unreal to them. And anything that happens in the game is really removed from who they actually are – if the character is injured or dies, there’s a health bar that goes down a bit, but it’s far removed from them. But they have exceptions – usually a dog, children, a husband – characters that feel much more real to them. Does that resonate with you? And does your son feel like one of those different characters who feels different from other people?

Dorian: Yes and yes. I’d say both of those are true. Especially when you’re a child, with psychopathy it’s like you just don’t really develop any emotional intelligence. You can think about the consequences of actions and words, but you don’t feel them deeply, so you don’t feel the world around you deeply. It’s a good way of describing it. You’re just running around, causing chaos, and you’ll always reset to the next day.

Dawn: And I imagine what has probably decreased is how removed you feel from your avatar?

Dorian: I don’t know – I’m just me. I’m not necessarily removed from myself. It’s just the world around me and other people in it.

Dawn: You described earlier that you’ve sort of grown into yourself, so I imagine you were more removed previously from your avatar, and now you’re more in there.

Dorian: Yeah.

Dawn: And this has had no effect on how most other people feel to you?

Dorian: Not really. I keep to myself for the most part, but people at uni are really nice. I like my classmates.

Dawn: I imagine that’s a contrast to the factory. There are all these attributes – being seen as bad or dangerous, as opposed to being seen as weak or out of control. Some people, in order not to seem bad, will claim not to have been in control. Others, in order not to seem out of control, will claim to have been bad intentionally. Does that make sense to you? Do you see yourself in one of those?

Dorian: I’ve kind of been described as both – “Dorian’s out of control, and therefore he’s being really bad.” But I don’t mind. It’s just the way somebody else sees me. From an outside perspective, I’d say I’m a very put-together, prioritized individual. I’m doing well in life, managing my money better, financially better off. I’m on a good trajectory. I can have my moments where I lose my temper, but I just try to keep out of certain situations, or keep away from a certain type of person who’s going to annoy me. And that seems to be working well.

Dawn: Can I make a return to your childhood? You mentioned at some point that you have some kind of childhood amnesia. Which years are affected, and what exactly is lost there?

Dorian: Probably 90% of everything from being a newborn right through to 17 or 18. I just honestly don’t care about my childhood that much, at all. It’s in the past, it’s gone, it’s done. It is what it is.

Dawn: Did you change a lot afterwards, and that’s why the new version can’t access those old memories?

Dorian: I’d say in some ways I’m a lot different than when I was a kid. But it’s not something I think deeply about, because I don’t need to. It doesn’t benefit me.

Dawn: Did the amnesia start right at age 19 or so, or is it more that years later you noticed you were having more and more trouble recalling those memories?

Dorian: I just noticed it years later, I think. But again, it’s not something that’s on my mind. I don’t live in the past or the future – I just live in the now. I’m always in the present, and that’s just my way of being.

Dawn: If some kind of hardship befell you – a long prison sentence, for example – would you also just be in the present? You mentioned you actually thought at one point that you might go to prison for about nine years.

Dorian: Yeah, that’s right. It didn’t feel like anything. I’ve never been scared of it. The thing is, with prison in this country at least, they’re always on 22- to 23-hour lockdown. What’s motivating me to stay out of prison is the boredom and the paranoia I’d feel in there. I just know it wouldn’t be good for me. I’d go on a mad one, essentially. Probably never get out. I’d be lucky if I did.

They’re very solitary, and very good at making people paranoid and depressed. I don’t know if I’d get depressed, but definitely paranoid. You can’t ever really relax in there. And you can’t really be the best version of yourself in that kind of place.

Dawn: A friend of mine also said that prison would be so boring for her that she would just start fights all the time, which would make the sentence longer and longer, and then she wouldn’t be able to get out anymore.

Dorian: Yeah, exactly. My dad, or my son’s mother, has said the same about me. They said I’d just be starting fights all the time, or stabbing people, because of how pent up I’d be. And I go, “Yeah.” So I’ve always got to keep myself stimulated or engaged with something.

Dawn: Jamie describes this very interesting no-self state, or anattā state, where she basically did not used to identify with what I, from the outside, would describe as herself. I have my own selfhood, which I probably developed in the first two years of my life, and so I just project that onto others. I feel like a separate person with my particular perspective on the world, and I feel intuitively that I’m allowed to move this body but not other bodies. And then I run into people who say, “No, actually, I don’t really identify with my own body.” I found that super interesting. I imagine you probably don’t have quite the same experience?

Dorian: I wouldn’t say so, no.

Dawn: You have your own particular perspective on the world. Is there something that, when you talk with Jamie, resonates about how you perceive the world, or how you perceive yourself in relationship to the world?

Dorian: My relationship to the world is that it’s kind of like – it’s my world, but you’re all just in it. That’s the closest or best way for me to describe it.

Dawn: It’s kind of like a playground. That’s a very clear distinction between yourself and others. Is there anything else where you and Jamie differ in your opinions or takes?

Dorian: I can’t think of any particular opinions that stick out. Me and Jamie are very similar but also very different. On some things we’re worlds apart, on others we’re not. Probably about an even 50-50 mix.

Dawn: In Jamie’s Discord server, what would you say is the ratio of people who are more like Jamie versus more like you versus just different in some other way?

Dorian: Everyone’s different in their own way.

Dawn: So you feel a similar degree of kinship with everyone?

Dorian: I wouldn’t say I feel kinship or friendship. It’s just a place where all different people come to chat.

Dawn: I have a few light closing questions, if you like. What is the funniest thing that’s actually happened to you because of how your brain works?

Dorian: I’ve got one. I remember talking about it in Jamie’s Discord, actually. To me, at the time, it wasn’t very funny. But now I can kind of look back on it and laugh.

I was getting ready to have my son over on the weekend. It was a Friday afternoon, and I was making the bed, getting his pajamas ready, getting everything ready for my kid. I was in my bedroom changing the sheets, and I looked out of my window. In my front yard, there was a young male – maybe early twenties or about 18 – with a balaclava over his face. My neighbor who lived above me was a 50-year-old man, so I knew it wasn’t him.

I opened the window and shouted down at the guy: “What are you doing in my yard?” He swore at me. Said, “What’s it got to do with you?” So immediately I saw red. I was only wearing socks and black boxer shorts. I grabbed this hunting knife – a big, silver, Rambo-type fixed-blade knife I kept in my bedside unit – put my boots on, and just ran outside half-naked. As soon as I opened the door, the guy saw me and ran. I chased him.

Honestly, I was trying to get him. This was years ago, not too long after I’d moved in. I was completely without fear or hesitation – that’s one of the few good things about having psychopathy. You’re not afraid, you’re not overthinking, you’re not procrastinating. You see a threat and you extinguish it one way or the other. I’m glad, in a way, that he got away, because imagine if I’d gotten hold of him – I could have been doing a life sentence. But I absolutely just wanted to get him.

He was probably scoping the place out for a burglary – it’s a high-crime area. We have locks, security lights, alarms, and everything now. But I ran out, chased the guy off, and as I turned to go back in, a woman was about 10 feet away who’d seen the whole thing. I just backed into the garden. I never got done for it.

There are probably loads of other funny things – nearly falling down the side of Snowdonia, getting the detective to shake my hand through the cell door. To me, it’s all very mundane. I have to think about what would be funny to someone who isn’t like me, because to me it’s not funny.

Here’s another one. I don’t even remember it, but my son’s mother told me about it last year. There was a time when we were together and living together. I’d been into town to some shop, and I got into an argument with somebody who owned a local business. She said I came back to the house really angry. Later that evening, my anger had simmered but it never went away. I said to her, “I’m going to burn down that guy’s shop. I’m just going to blow it up right now.”

How many guys have said that and then done nothing? She thought it was just me being angry. But when she saw me start putting my boots on and getting my jacket, she panicked. She was thinking, “Oh my God, he’s actually going to go out and do this.” So what she did – because I think she’d figured out by then that something was not right with me – was she had sex with me. And it worked. We went upstairs, and she said she made sure that when I woke up, there was a big, hot roast dinner ready for me. She said I just woke up and ate my dinner, as if I hadn’t been about to go out and commit arson.

I’ve got no memory or recollection of it. I just know what she’s told me. She was terrified, she said – her stomach was in bits, and she just didn’t know what to do.

Dawn: Unfortunately, arson is such a common first reaction. Very plausible. And the big knife – is that one you still have, or did the police take that when they found the lifeguard?

Dorian: No, the knife wasn’t an illegal weapon – just a normal fixed-blade hunting knife. The blade doesn’t fold. But it was a really cheap knife, and I was using it to cut weeds or do gardening, and it just snapped. So it went in the bin.

I’m sure I’ve done loads of other daft, crazy things, especially as a kid at 15, 16, 17 years old. But I probably won’t care to remember half of what I’ve done, because it’s just another day. Now I’m a lot more reserved and mature.

My dad has said to me, “I’ve seen a big difference in you, slowly over the years. I don’t worry about you as much. I don’t always think he’s going to get nicked. You’re at university now, and we can tell you care about that.” So staying engaged with things that are good and interesting for me – that’s how I’d end things.

Dawn: Caring about something also means having something to lose, and that has flow-through effects for other areas of life. What advice would you have for the 15-year-old version of you?

Dorian: I’d say: do more. Go for even more. Just do what you want – go out and do more.

Dawn: More of anything?

Dorian: More of anything.

Dawn: Is there anything in particular that you actually enjoy? I would imagine that control and winning are pleasurable, but beyond that?

Dorian: I don’t work toward “Am I happy?” or “Am I feeling happy?” The way I do things, it’s more: Am I content? Am I satisfied with the way everything is at the moment? But I do genuinely enjoy the time I spend with my son. I’ve trained him up to be really good at fighting – he’s only five years old, but he throws a hell of a left hook now. He’s an extension of me, and I think any healthy man likes to be a good dad. That’s one thing I enjoy.

I enjoy my studies as well. I like doing my assignments at university, writing academically. And I like going there, because school for me as a kid – especially secondary school – was like prison. It was horrible in a lot of ways. I even had bars on the windows in my old secondary school. It wasn’t a great place. So now I’m sort of getting to relive going to school, but this time it’s good. I’m enjoying it. Just staying on the straight and narrow.

Dawn: That also made a huge difference for me – between high school, where I was forced to be on the property during certain hours, and university, where I actually chose to go and wanted to be. Finally, is there anything that you wish people like me would ask you, but they never do?

Dorian: Not really – not in those terms. But I wish everybody would stop thinking that we’re all serial murderers and sex offenders. Maybe just get away from those Netflix dramas and TikTok and all the myths. Open a textbook once in a while. Look at some actual information on people with psychopathy, or maybe listen to us a bit more. That’s what I’d say.

I’ve had some interesting questions posed to me on my YouTube channel. They’re a good bunch. I think we’re nearly at 140 subscribers now. If anyone’s got any questions for me, you can comment them down below, or go over to my channel and ask me on there. I’ll be taking a look at both either way.

Dawn: I’ll link the channel in the description. You’ll probably do FAQ videos from time to time to answer them.

Dorian: I’ve done one already – a comments reaction. I like doing them because they’re really easy. Somebody in my comments section asked a question, and then another person replied, “He doesn’t bother replying to his comments.” So I had to reply and say, “I am – I just do it in video form.” Which I prefer. It’s just better that way, isn’t it?

Dawn: Makes a lot of sense, and you get a cool new video out of it. That was all of my questions. Thank you so much.

Dorian: They were really good. Thank you very much for having me on, Dawn. I appreciate the offer and the invite. I know we were supposed to do this last week, but we had technical difficulties. We’ve cracked the matrix now, so we’re back. Thanks for having me on, and everybody, thank you for watching. Be sure to check out my channel – the link will be in the description below. And before you subscribe to me, make sure you’re already subscribed to Dawn here. We’ll both see you in the future. Dawn, thank you for having me on – it’s been an absolute pleasure. Take care.

Dawn: Thank you, goodbye, and like and subscribe!

Dorian: Bye.

This is the final article in a series on understanding psychopathy. Previous articles covered the framework, biology, environment, psychological structure, empathy mechanisms, and archetypal clusters. This article explores recovery – without moralizing and with attention to what’s actually possible and what it costs.

Introduction

This article is for people who want to consider change. That’s not everyone. Some people with psychopathic traits are perfectly content and functional. Some don’t experience their traits as problems. Some have built lives that work for them as they are.

If that’s you, this article isn’t a prescription. But if you’re curious about what’s possible – or if you’re experiencing distress and wondering what might help – this is an honest assessment.

Dimensions of Recovery

Recovery isn’t one thing – it’s multi-dimensional. Different dimensions matter for different presentations, and each dimension has trade-offs.

1. Stable life and environment

2. Emotional regulation and behavioral control

3. Insight and mentalization

4. Integration of self-states

5. Developing prosocial values

6. Capacity for guilt

7. Capacity for remorse

8. Access to affective empathy

9. Capacity for secure attachment

Stable Life and Environment

What it means. Achieving stable circumstances – housing, social environment, employment, relationships, legal status, enemies.

Assessment

Stability is almost pure upside for most people. Without stability, it’s often necessary to move from hotel to hotel and from country to country, which makes it hard to plan more than a few weeks ahead. That interferes with collaborative efforts such as most paid work. It also increases the cognitive load trying to learn how the local infrastructure works, time and effort that could be spent on efforts with exponential yields (healing, investment, education, networking, getting criminal records sealed).

Financial stability also makes things cheaper in the long run – fixed-interest mortgage instead of rent, up front payment instead of installments, no high-interest debt, etc. – and has exponential yields.

The worst case is if you’re still stuck in the environment (often the family) that produced your adaptation in the first place. For many people, their parents’ home is a battlefield where they’re trying to survive behind enemy lines. Most of the world (outside war zones) is not like that. In most cases it’s beneficial to get out of there, if at all possible, and gradually adjust to society rather than to code-switch between environments.

The environment one is stuck in can also be one of crime. The resulting availability of crime-based income sources can perpetuate a cycle of crime -> prison -> criminal records -> difficulty finding work -> more crime. Removing oneself from these networks, if possible, will make life harder in the short term but promises and eventual breaking of that cycle. The proximity to crime also increases the risk of wrongful convictions: Some jobs might have a low risk of legal consequences, but someone in your environment might still accidentally or intentionally implicate you in their crimes.

Chaotic environments – whether at home or in criminal contexts – also often come with violence. Depending on the health system in your country, that can be another unpredictable source of tremendous costs, on top of legal fees. Building wealth or maintaining a good credit rating can be difficult as a result.

Finally, stability makes it easier for others to rely on you. Even if they trust your intentions, they’ll need to take into account the risk that you’ll drop your collaborative project because you have to leave the country from one day to the next, end up in jail or in a hospital for weeks or months, lose your car, power, or internet connection, etc.

However, people who love chaos for the thrill (or conversely to beat the boredom) may need to find replacement activities that are thrilling while not threatening some basic stability – a job that is action-packed in itself, anything to do with one’s phobias, BDSM, the sorts of hobbies you can get a RedBull documentary out of. Especially if chaos is one’s comparative advantage, jobs that deal in chaos – like politics or sometimes law – may be attractive.

People who’ve been maintaining their chaos to repress traumatic memories, may need to process those once the chaos has died down. That’s a good investment in the long run.

Building stability in the long run means having something to lose. That’ll generally make you most trustworthy for others, but at the cost that you do indeed have something to lose. Some possible actions will become more costly in expectation.

Emotional Regulation and Behavioral Control

What it means. Developing the capacity to control behavior – to pause before acting, to resist impulses, to maintain consistent conduct.

Assessment

Emotional or behavioral regulation is another very popular recovery step. The benefits compound over time; the costs are real but manageable.

Impulsive decisions often lead to particularly badly executed crimes, manipulability, losing friends and partners, and costs from having to replace damaged possessions. That includes costs from impulsive physical and verbal aggression as well as unpredictable ghosting, moves to other countries, and continual deleting and recreating of social media accounts.

The benefits compound: healing, savings, education, networking, etc. Much like above.

An added benefit is that those who care deeply about self-control can derive self-esteem from their increased ability to know what they’re going to do in the next moment and years down the line.

The madman theory point is worth acknowledging: If you’re unpredictably dangerous, people are less likely to push you even in minor ways because you might just react in (to both of you) irrationally costly ways. Regulation sacrifices this. For some, that’s a loss of a strategic advantage.

Sadly some relationships only persist because of the chaos cycle. If you regulate, those relationships may end – which might be good for both parties, but is still a loss.

Dialectical Behavioral Therapy is designed for emotional regulation and is considered to not have side-effects, but often drugs like mood stabilizers (lithium and others) are beneficial or even necessary for someone to be able to engage in therapy. These drugs can have side effects, though they vary from person to person, so it might be just about finding the right one (and the right dosage) for you.

With or without the assistance of drugs, you’ll probably go through a practice phase where you’ll still have to think long and hard how to react most effectively in a given situation. That becomes more automatic over time, but the hesitancy may persist beyond its usefulness. That too is a transitory phase.

Insight & Mentalization

What it means. Understanding your own patterns and those of others – what you do, why you do it, what triggers you, what your mechanisms are.

Assessment

I’m generally a big fan of insight because it follows the principle of the Litany of Gendlin:

What is true is already so.
Owning up to it doesn’t make it worse.
Not being open about it doesn’t make it go away.
And because it’s true, it is what is there to be interacted with.
Anything untrue isn’t there to be lived.
People can stand what is true,
for they are already enduring it.

Insight is not only intellectually rewarding, it allows you to predict what you’ll do and is the foundation for changing it according to your will – i.e. many of the other choices in this article. If you don’t see it, others might, and then they can catch you off guard when they exploit your hidden vulnerabilities.

But most people are safe, so not only are vulnerabilities that you’re aware of ones that you can neutralize, all your self-knowledge, beyond vulnerabilities, is also what you can share with others to synchronize on expectations. Often being transparent about expectations up front is what gets the job done whereas the other would get pissed if they later showed up as a surprise.

All the skills that you learn to understand yourself better and better are mentalization skills. You can train those in mentalization-based treatment (MBT). You learn to get curious about yourself, form hypotheses of what might be going on with you given who you are and your current situation, test those hypotheses, and learn to accept the results without flinching. These are skills that you can also apply to others and to your interactions with them – they’ll hone your cognitive empathy. You’re probably quite different from most others, so to understand others, it’s critical for you to first understand yourself and how you’re different from others. More understanding strikes me as a fairly universally useful skill.

Lydia Benecke also observed that friends of hers who employ some kind of emotional repression mechanism (friends of hers with various psychopathic traits, probably excluding N-hypoactive presentations) get tired quickly during conversations with her. She hypothesized that these mechanisms are tiring for some. Conversely, not repressing anything, to the best of one’s ability, may unlock greater endurance.

The issues with insight all live at the intersection of awareness and paralysis. I’d choose awareness any day even if I cannot change something, but there is a certain frustration that comes with that. That frustration can be accepted, but I understand when some continue to choose comforting illusions over harsh reality.

Integration

What it means. Bringing together fragmented parts of self – reducing state-switching, developing coherent identity.

Assessment

There is no separate dial “self-control” one can train; it’s always about integrating your interest in achieving more long-term goals with any short-term goals, and then finding a synthesis that you’ll also still appreciate much later. If your personality is split into an all-selfish part and a part that wants to be good to your friends, you can only be consistently good to your friends if you have uninterrupted access to the second one.

Some people have nothing that they could call their self. They have enough insight to know that they are vastly different people from one context to the next, or they have no access to who they are in the first place. This makes it difficult for them to intuit what it is like to have a coherent, continuous self. Even if they want to respect others’ autonomy or boundaries, they can’t intuit what that entails.

That also improves memory access. When parts are maximally separate, you can run into literal amnesia between them – you can only access the memories that are consistent with the expectations that the part has about who you are. In less extreme cases, the memories may be there, but they feel very abstract, like reading someone else’s diary. If you then try to imagine why you might’ve done whatever the “diary” says, you’ll confabulate very different reasons from the ones you had at the time. Or the memories may be there, but there’s a kind of anxiety that comes up when you try to recall them or get close to chat logs that might have the effect – perhaps an anxiety of not wanting to acknowledge the memories or an anxiety of not wanting to reactivate them lest you fall back into the self state that they’re associated with. In a fully integrated state, all of this is chill.

I can’t overstate the game-theoretic advantages of being able to commit credibly. The first advantage is low transaction costs. Jonathan Haidt writes in The Righteous Mind:

Social capital refers to a kind of capital that economists had largely overlooked: the social ties among individuals and the norms of reciprocity and trustworthiness that arise from those ties. When everything else is equal, a firm with more social capital will outcompete its less cohesive and less internally trusting competitors (which makes sense given that human beings were shaped by multilevel selection to be contingent cooperators). In fact, discussions of social capital sometimes use the example of ultra-Orthodox Jewish diamond merchants, which I mentioned in the previous chapter. This tightly knit ethnic group has been able to create the most efficient market because their transaction and monitoring costs are so low—there’s less overhead on every deal. And their costs are so low because they trust each other. If a rival market were to open up across town composed of ethnically and religiously diverse merchants, they’d have to spend a lot more money on lawyers and security guards, given how easy it is to commit fraud or theft when sending diamonds out for inspection by other merchants. Like the nonreligious communes studied by Richard Sosis, they’d have a much harder time getting individuals to follow the moral norms of the community.

But in an environment where high trust is possible (a sufficient fraction of the participants are trustworthy), it also becomes individually rational to be trustworthy! That’s because of gains from trade in iterated games.

One diamond seller has lots of diamonds and needs to pay rent for their store. I don’t have any diamonds but I have a buyer who wants me to select one for them. A trade happens when the seller values the diamond less (abundance, rent) than I do (rich buyer, time pressure). Both of us generate gains from trade by executing it, generating net wealth in the world.

My profit is small in each trade, but I can repeat these trades infinitely many times, for hypothetically infinite profits. It’s just a matter of how long one can maintain the streak. Conversely, stealing the diamond would foreclose any future trades for a hypothetically infinite loss. This introduces a strong positive feedback effect in favor of high trust.

In practice the profits aren’t literally infinite because the seller might retire or we might die. If the seller is less reliable, that will curtail the potential future profits. Same from the seller’s perspective if you are unreliable. If the transaction costs are high due to security measures, that also reduces the profits either of us can make by maintaining the streak. Both factors mean that we both have less to lose, and so can trust each other less. This introduces a strong negative feedback effect in favor of betraying the trust.

High-trust environments make everyone (in aggregate and individually) richer. Low-trust environments make everyone poorer. So it makes sense to seek out high-trust environments and then to play by their rules until you get bored of getting rich.

That brings us to the importance of integration for your choice of environment. Integration is important to function in a high-trust environment and a high-trust environment is necessary if further recovery is desired.

Which brings us to the greatest con of integration: You’re probably already well adapted to low-trust environments, so if you have to continue to function in one – your family, social circles, etc. – integration is no use. It would only give you access to memories and perhaps feelings you don’t want to access, and allow the parts of you to enter into the tug of war that you get when you have parts that often want contradictory things – thrill-seeking vs. surviving to provide for your kids; sadism vs. being good to your friends. Those fights can eventually be settled from what DBT calls the wisemind stance, but it takes practice to get there.

Sidenote: Credible commitments are such a powerful weapon that it’s often good that we usually can’t just credibly commit to anything. Open source game theory is a branch of game theory that studies what would happen if we could fully credibly commit – like throwing out the steering wheel in a game of chicken in a way the opponent can’t miss. The results include commitment races, where everyone tries to be first to commit.

Values

What it means. Developing an ethical framework – principles that guide behavior independent of emotional responses.

Assessment

Where integration provides the scaffolding, values provide the content. A value of reliability is what unlocks the gains from trade from the previous section. A value of respecting others’ autonomy enables respectful relationships. A value of authenticity prevents you from self-deceiving. A value of minimizing harm makes me happy, because I’m a suffering-reducer.

Importantly, values are your own. As you gain more insight, you discover what you actually care about, or at least what feels right to the point where you can commit to it. Brené Brown has compiled a long list of values to choose from or to amend. It makes sense to focus on just the most important core values. (Feel free to call them standards or moral safeguards or whatever other terms suit you best.)

A friend of mine who generally has a choice whether to go into a D-avoidant-type psychopathic state or not indulged in it for five years but found the boredom some excruciating that she even chose her alternative BPD state over it. Values provided her with a purpose in life and the opportunity to contribute to something larger that she cared about. Without it, she didn’t feel like she had a life worth living.

Personally, ethics has been a special interest of mine for decades, so I also find it stimulating to argue for metaethical expressivism over realism, or for antifrustrationism over hedonic act utilitarianism.

Conversely, though, different values often trade off against each other, which introduces complexity in proportion to the number of values involved. Self-interest is just one of them, so the trade-offs will sometimes conflict with it. Finally, all the game-theoretic advantages of credible commitments constrain behavior ipso facto.

Below I mention the function of guilt for training values. For many, certainly G-callous + N-hypoactive presentations, that’s not the most direct way to get there since it might just be very hard to train any capacity for guilt. So if you’re interested in values, relying on habits and on making them part of your identity is the more direct path.

Guilt

What it means. Developing the capacity to notice when you violate your values and learn from it.

Assessment

Guilt is a signal, not a virtue. If you already don’t want to harm people you care about, and you have other mechanisms (values, regulation, foresight) to prevent harm, guilt adds suffering without adding benefit. But if you keep accidentally violating your values, guilt will help you notice it earlier and stop sooner.

Disambiguation: I understand guilt to be the name for the emotion you feel if you violate your own values – say, if you have a value of respecting other’s autonomy, but then you get carried away or overlook something, and someone else does get hurt. It’s similar to a kind of disappointment in yourself but with some added (alarm) bells and whistles. I understand remorse to be the name of the emotion you feel when you do something that makes someone else feel harmed by you.

So when you notice that you keep running into unwanted consequences of your actions, and you find that you already have values in place that should prevent you from taking these actions, and you have sufficient integration that the parts who care and not wholly separate ones from the parts who act, then guilt is helpful as a signal for your practice. It’s like a climbing coach who tells you how to optimize your body positioning on the wall to optimize your climbing performance.

If you’re already doing a good job following your values, guilt doesn’t help boost that further.

Remorse

What it means. Developing the capacity to feel the need to make it up to someone when you’ve made a mistake.

Assessment

This is one of my favorite moral emotions. As clarified above, this is about the other person and whether they feel harmed by you – regardless of whether you’ve followed your values or violated them. You’ll always make mistakes, so it won’t stop being useful.

Core to parenting is what is called “contingent marked mirroring”: The “contingent” part means that you mirror back the emotion of the other person to make them feel understood. The “marked” part means that you evaluate the situation and communicate our evaluation too. For example, if the kid is hurt and crying, the parent may first express empathy with the hurt of the child (this can be completely fake, so long as it’s accurate), and then soothe the child to signal that the bruise is not dangerous.

This is also important when you’re interacting with adults: If you’ve just lost your temper with someone and they look scared, you can say, “Oh my god, I’m sorry, I didn’t want to scare you!” That’s an emotionally charged response that meets them where they are but also signals that you understand that they feel scared. Then you can follow it up with, “I’ll try hard not to let that happen again. Do you need space or a hug? Or ice cream?” That signals that you’re drawing lessons from what happened and are ready to repair. Your friend or partner may have particular preferences for this.

This can feel scary for some because you’re admitting that you’ve made a mistake, and the other can now hold that against you. It’s like you’re indebted to them, and you’re telling them about it. Hence it’s important to keep the repair proportionate. If you’ve broken something, you replace it; if you’ve scared someone, you make sure to make them not scared anymore; if you’ve stolen something, you give it back. If they keep holding it against you for ages when you’ve tried your best to make amends, it’s their turn to show remorse for their exploitative behavior.

Hence remorse needs to be proportionate. It can allow you to be spontaneous and risk-taking because you know you’ll make it up to whoever you might hurt and you won’t lose friends if you make a mistake. It can also curb excessive risk-taking if the possible consequences are things that would be costly or impossible to repair. But it can also have the opposite effect of stifling healthy risk-taking if the prospective remorse is disproportionate.

Affective Empathy

What it means. Developing or restoring the capacity to feel what others feel – to have their distress produce distress in you, their joy produces joy.

Assessment

Affective empathy is quite a mixed bag. Sometimes I love it; sometimes I resent it. I love it when I spend time with my partners. I resent it when it encourages me to look away from the overwhelming suffering in the world. Plus many of my friends act morally without empathy. They have values, insight, and regulation. They channel their sadism (if any) into consensual outlets. They don’t need empathy to be a good person – and developing it would have costs they don’t need.

I agree with Paul Bloom and his book Against Empathy: Empathy is fun, but unnecessary for and often detrimental to moral decision-making.

When I chat with my friends with NPD and see how they struggle in romantic contexts, I feel how scared they are, I feel the anger of their partners, and I feel my own exasperation over how unnecessary these conflicts feel from my vantage point. When they make progress and acknowledge that they’re not actually indifferent to their partners, I feel so proud of them. When they’re grateful for something I’ve helped them overcome, I feel so happy for them. When I repair with my partners and feel how nervous they were before and how safe they feel again, it’s also heartwarming. I’d never want to miss all this vibrancy in my life again.

For those who feel a lot of boredom or emptiness, all this vibrancy would probably also go a long way to keeping them entertained.

But god I have to work so hard to stay motivated for my charity work. The greatest perils in the world are not kids dying from cancer or abandoned cats, because there are at most millions of them and it’s expensive to try to help them more than they’re already being helped. The greatest perils are animals in factory farms, fish and shrimps in aquaculture, farmed insects, animals dying in wilderness from diseases, parasites, starvation, and predation, global catastrophes and more, because there are easily a trillion times as many. Affective empathy doesn’t work for large numbers. You’d expect to care a million times more for a million people than for one, but empathy probably makes you care less. Worse, when you put in the work and save a million animals from suffering, empathy doesn’t even reward you for that. In this regard, it’s a complete scam.

If you want to make great career decisions to help the world, make great impactful donations, design great policies, then you better turn off your affective empathy. Compassion (karuna) doesn’t require affective empathy either. If we want to help at all, we have to triage constantly because we can’t help everyone. Empathy is only in the way of efficient triage. And that doesn’t even touch on intentional manipulation and the things people with high preoccupied attachment do intuitively to tug on other’s heartstrings.

So empathy is a really mixed bag. There be dragons. Enter at your own risk.

Note that even my untraumatized, non-disordered, securely attached friends with G-callous and N-hypoactive don’t have access to affective empathy. They tend to feel guilt and remorse very lightly and affective empathy not at all. Don’t be discouraged by that – but relying on habits and identity to practice values may be a quicker route to success if that’s what you’re interested in.

Attachment

What it means. Developing the capacity for genuine emotional connection – to care about specific people, to be affected by their presence and absence.

Assessment

As someone with mildly above-average preoccupied attachment and very low avoidant attachment, I love attaching to people. It makes me feel safe, supported, and needed. Living alone is not fun for me. Meeting my friends and partners is what sparks joy.

But it’s crucial to attach to the right people. When someone strikes me as deeply genuine, my guard goes down quickly; when someone is reserved or has a history of interpersonal cruelty, I’ll keep my guard up for much longer.

Exposure to manipulation is a cost to me, so I can apply general risk management principles: What does the other have to gain, and how can I increase their cost to the point where the gain is no longer worth it for them while keeping my costs lower than the cost of the manipulation? This can mean taking things really slow, and seeing if they stick around.

The structure of the marketplace is also informative. If the attacker can gain $100 from exploiting you, it might seem like you’d have to invest enough into security to drive up their costs > $100, which might come at a cost of $10 to you. But that ignores that the marketplace may be such that there are plenty of people with sufficiently bad security that the attacker can extract $100 from them at a cost of $50 to them. So most likely, you can get away with investing $6 into your own security for a cost of $60 to the attacker, which is greater than what they’ll have to pay elsewhere. Then it’s no longer rational for them to pursue you as a target.

So attachment is not riskless, but the risks are manageable using standard practices.

Many people are afraid of getting duped just for the sake of the duping, even if it comes at no cost to themselves. Realizing that and worrying less about it can reduce avoidant attachment.

Another cost of attachments is the reduction in flexibility that comes with them as well as the time costs that need to be invested into the coordination. This is a common problem in startups who hire too many employees too quickly and get slowed down even on net by the coordination overhead. Hence there is a sweet spot where the overall joy, safety, and support that come with attachments is balanced against the coordination overhead. If your sweet spot is unusually high, you might consider polyamory. Long-distance relationships are another option to keep the costs from reduced flexibility low.

Finally, attachments can take on a bit of a life of their own, where against your better judgment you find it hard to break away from someone who’s not good for you. That’s a situation where it’s useful to have access to avoidant patterns and use them with discretion, i.e. not toward anyone or any romantic partner but specifically toward one person only.

And that’s it: The perhaps least romantic assessment of love ever written.

Conclusion

That’s the menu. If the cost-benefit ratios of any of these qualities strike you as desirable, you can train them yourself and employ a capable therapist to show you the ropes to speed things along.

Some cool types of therapy that I can recommend are Dialectical Behavior Therapy (DBT), Mentalization-Based Treatment (MBT) and Schema Therapy. There’s also Transference-Focused Psychotherapy (TFP), which personally always rubs me the wrong way, but it’s also evidence-based.

You know best where your current bottlenecks lie, so you can tackle those first, then add whatever else you desire later on as the cherry on top.

If you have noticed any pros or cons that I have no listed, please let me know in the comments!

This completes the series on understanding psychopathy. Thank you for reading! If this framework is useful to you, I’d love to hear about it! If something doesn’t fit your experience, I’d love to hear about that too!

This is the sixth article in a series on understanding psychopathy. Previous articles covered the framework, biology, environment, psychological structure, and empathy mechanisms. This article presents common clusters – archetypal profiles that tend to co-occur.

Introduction

The previous articles described dimensions – G, N, E, D, B, A, C – and their variants. But people don’t come as random combinations of features. Certain profiles cluster together, producing recognizable types.

This article describes 13 clusters I’ve identified through conversations with friends and observation. These are hypotheses, not established categories – patterns I’ve noticed, not validated subtypes. But they may help you recognize yourself or people you know.

The clusters are ordered by typical distress level – from those who are generally content to those who suffer chronically. Please take this assessment with a grain of salt as self-report and behaviors are unreliable guides to the actual intensity of the suffering. But I’ve tried not to make various common mistakes.

How to Read These Clusters

Each cluster is described with:

• Feature profile. The typical G, N, E, D, B, A, C features.

• Narrative. A developmental story – how this pattern emerges.

• Distinguishing features. What sets this cluster apart.

• Distress level. How much the person typically suffers.

• Representatives. Real and fictional examples.

If a cluster resonates, it doesn’t mean you are that cluster. Profiles are complex; clusters are approximations. But resonance is information.

Tier 1: Low Distress

These clusters involve people who are generally content with their lives.

Cluster 1: The Lucky Primary

The person who has constitutional psychopathy but developed well.

Feature Profile.

• G: G-callous

• N: N-hypoactive (constitutional)

• E: E-I-secure, E-C-normal, E-P-success, E-A-stable

• D: D-secure, minimal other pathology such as well-regulated sadism or masochism

• B: B-minimal

• A: A-observational or A-strategic

Narrative. Born with genetic loading for reduced empathy and fear, expressed neurologically as constitutional amygdala hypoactivity. But raised by attuned, secure parents in a stable environment. The psychopathic substrate is only substrate – it never developed into pathology because there was no adverse shaping.

May have sadistic interests, but channels them into regulated outlets (BDSM, fiction, competitive fields). Has stable relationships, prosocial behavior, and may not identify as having any disorder.

Distinguishing Features.

• No vulnerable phases (not NPD)

• No rigid self-structure (not sovereign)

• Sadism present but regulated

• High functioning

• Sense of self

• No memory of being different

• Very open, unembarrassable, fearless

• Remarkably emotionally stable

• High stress and trauma tolerance

• Often strict but habitual (not obsessive) adherence to personal rules

• Secure attachment style

• Good friendships and relationships

Distress Level. Very low. Life is fine. No internal conflict because there’s nothing to conflict.

Representatives.

• Some portrayals of James Bond

• Surgeon/CEO archetypes

Cluster 2: The High-Functioning Machiavellian

The ambitious manipulator who exploits within legal bounds.

Feature Profile.

• G: G-callous

• N: N-hypoactive

• E: E-I-secure or E-I-avoidant, E-C-normal or E-C-golden, E-P-success, E-A-stable

• D: D-autonomic-asymmetric

• B: B-subclinical

• A: A-strategic

Narrative. Moderate psychopathic loading channeled into conventional success. Good-enough environment provided socialization; intelligence enabled navigation of systems. Uses others instrumentally but within legal bounds. May be ruthless in business but not criminal.

Differs from Cluster 1 in having more predatory orientation – uses the substrate antagonistically rather than collaboratively. Might not see the point in respecting others’ autonomy if they’ve never had a self to feel autonomous.

Distinguishing Features.

• Successful by conventional metrics

• Exploitative but not criminal

• Strategic, not impulsive

• Unlikely to seek therapy

• Often tenuous sense of self

• Struggles with interpersonal relationships

Distress Level. Low. Winning feels good. No internal conflict.

Representatives.

• Real: James Fallon (The Psychopath Inside), various CEOs, politicians, surgeons (see Kevin Dutton’s research)

• Fictional: Gordon Gekko (Wall Street), Bobby Axelrod (Billions)

Cluster 3: The Enlightened Machiavellian

The presence that is really an absence – manipulation as concomitant of a lack of selfhood.

Feature Profile.

• G: G-callous

• N: N-hypoactive

• E: Variable (often E-C-neglect or E-C-chaotic)

• D: D-anatta

• B: B-factor-1

• A: A-observational

Narrative. Constitutional psychopathic loading without a self. Unlike Cluster 1, there’s a felt absence – the distinction between self and other was never learned. Unlike Cluster 5 (sovereignty), there’s no grandiose self-structure, no rigidity, no investment in power.

Manipulation is instrumental, for stimulation, out of a lack of understanding what selfhood feels like, or a byproduct of not making a distinction between self (manipulation okay) and other (manipulation not okay). May be curious about humans as objects of study. Observational relationship to own behavior.

Distinguishing Features.

• Felt absence

• Lacking understanding of personal continuity, long-term planning, self-preservation, etc.

• Observational stance (“I watch myself do things”)

• No rigid (or any) self structure

• May show unusual honesty about self (nothing to defend)

• Boredom is a major problem

Distress Level. Low to moderate. The boredom can be annoying but is rarely excruciating because feelings are hard to locate and not clearly relevant to anyone.

Representatives.

• M.E. Thomas (Confessions of a Sociopath)

Tier 2: Moderate Distress

These clusters involve people who function but face situational distress or external challenges.

Cluster 4: The Street Primary

Constitutional psychopathy channeled into criminal expression.

Feature Profile.

• G: G-callous + G-impulsive

• N: N-hypoactive

• E: E-I-avoidant or E-I-disorganized, E-C-chaotic + E-C-neglect, E-P-antisocial, E-A-crime

• D: D-anatta + D-autonomic

• B: B-mixed + B-violent

• A: A-strategic or A-observational

Narrative. Same constitutional loading as Cluster 1, but shaped by chaotic, violent environment. Adversity was physical and chaotic (violence, poverty), not control-focused (controlling parents). Defense was behavioral – fight, survive, take – not self-structural.

No NPD develops because there was no need or opportunity for grandiose self-construction. Trait narcissism may be present (adaptive, useful) but not pathological.

Distinguishing Features.

• No self-image injuries (not NPD)

• No grandiose self-structure

• Criminal history; often incarcerated

• Violence is instrumental and/or reactive

• No memory of empathy or guilt in childhood

• Drug addictions

Distress Level. Moderate, but the distress is external (prison, poverty, consequences) not internal. The psychopathy itself is not distressing.

Representatives.

• Real: Edward Bunker (Education of a Felon)

• Fictional: Tommy DeVito (Goodfellas)

Cluster 5: The Primary Sovereign

NPD stabilized by psychopathic substrate – the classic “malignant narcissist.”

Feature Profile.

• G: G-callous

• N: N-hypoactive + N-dissociative

• E: E-I-avoidant or E-I-disorganized, E-C-controlling/E-C-parentified/E-C-neglect

• D: D-sovereign

• B: B-factor-1 (mostly)

• A: A-strategic, A-retroactive, A-narrativizing, A-selective, A-absorbed

Narrative. Constitutional psychopathic loading meets neglect- or control-focused adversity. The substrate hinders normal moral development; the specific adversity (controlling parent, parentification, neglect) leaves a void and triggers NPD development to fill it. E-I-avoidant prevents the maturation of the amygdala and other brain regions for an N-hypoactive substrate; or E-I-disorganized meets a constitutional N-hypoactive brain. Either way, there is little fear or reactivity.

Where children are normally parented by their parents, these children are parented by the strategies they can find to extract what they need from their environment. They cannot learn that asking for solace or reassurance gets their attachment needs met, because that doesn’t happen, but they can learn that lying and stealing get their material needs met. They repress the attachment needs that are impossible to fulfill and fill the void with the pleasures of exploitation and control.

Result: Sovereignism – NPD stabilized by psychopathic substrate. Power/control orientation rather than admiration-seeking. Sadism is ego-syntonic. Very early sadism onset; the moral emotions never fully form.

Distinguishing Features.

• No classic vulnerable phases but a superficially schizoid retreat

• Sadism is ego-syntonic

• Control/power orientation, not admiration

• Rarely acts impulsively

• Claims rare impulsive acts as intentional (A-retroactive)

• No memory of empathy or guilt in childhood

• Rigidity around selfhood/identity

• Drug addictions

Distress Level. Moderate. The void exists; control-needs are constant work. But grandiosity is stable – no cycling.

Representatives.

• Real: Donald Trump

• Fictional: Frank Underwood (House of Cards), Amy Dunne (Gone Girl), Cersei Lannister (Game of Thrones)

• Clinical: What Kernberg describes as malignant narcissism, though the paranoia is optional

Cluster 6: The Autonomist

Extreme autonomy-preoccupation without full sovereignty.

Feature Profile.

• G: Variable

• N: Variable

• E: E-I-avoidant or E-I-disorganized, E-C-controlling or E-C-enmeshed

• D: D-autonomic (defensive or symmetric), possibly D-avoidant or D-anatta

• B: Variable

• A: Variable

Narrative. Primary feature is extreme autonomy valence associated with high avoidant attachment from early control/enmeshment experiences. May or may not have psychopathic substrate. Organizing principle is freedom from constraint rather than power-over-others.

If symmetric (D-autonomic-symmetric), may have principled libertarian ethics – values others’ autonomy equally. If asymmetric, hypocritical – demands own freedom, impinges on others’.

Distinguishing Features.

• Extreme reactivity to perceived constraints

• Contractarian social model (“I didn’t agree to that”)

• May be principled or hypocritical

• Not necessarily sadistic or predatory

• Rage or withdrawal when constraints imposed

• Functions well until obligations arise

Distress Level. Moderate (situational). Fine when autonomy isn’t threatened; distressed when it is.

Representatives.

• Fictional: John Wick (John Wick), Ron Swanson (Parks and Recreation), Buttercup (Powerpuff Girls)

Cluster 7: The Alexithymic

Appears psychopathic but the issue is access, not absence.

Feature Profile.

• G: G-minimal

• N: N-disconnected (especially insula)

• E: E-I-avoidant, E-C-unattuned or E-C-invalidating or E-C-neglect

• D: Variable

• B: B-minimal

• A: A-observational, A-strategic

Narrative. Appears psychopathic – doesn’t seem to feel or respond emotionally – but the issue is access, not absence. Emotions may be there but can’t be identified, labeled, or articulated. Often comorbid with autism or developed from chronic invalidation.

Key distinction: Not low empathy – low interoceptive/emotional awareness. May respond to interventions that restore access.

Distinguishing Features.

• Difficulty identifying own emotions

• May have somatic complaints

• Relationship difficulties from apparent coldness

• No antisocial behaviors

• May respond to interoceptive training

Distress Level. Moderate. Confused by own reactions; relationship friction.

Representatives.

• Some autistic presentations

Tier 3: Moderate-High Distress

These clusters involve internal conflict and significant suffering.

Cluster 8: The Secondary Sovereign

Sovereignty developed defensively; potentially recoverable.

Feature Profile.

• G: G-minimal or G-reactive

• N: N-dissociative + N-hyperactive

• E: E-I-disorganized, E-C-chaotic/E-C-controlling/E-C-chaotic/E-C-scapegoat/E-C-punitive

• D: D-sovereign

• B: B-mixed

• A: A-strategic, A-retroactive, A-narrativizing, A-selective, A-absorbed

Narrative. Originally normal or reactive child who developed sovereignty defensively. The N-hyperactive pattern (reactive, hypervigilant) got blunted through chronic dissociation, producing current N-dissociative presentation.

Key marker: Remembers being different as a child. The moral emotions (empathy, guilt) did develop initially, then were suppressed. This means they’re potentially accessible – recovery is possible.

Distinguishing Features.

• Remembers being different (the key marker)

• Was reactive/sensitive as child

• Multiple adversity types (violence + control)

• Empathy returning “on good days”

• Sadism may have mitigating conditions (“only if they hurt me first”)

• May have good insight

• Visible state-switching during recovery (vulnerable phases emerging)

Distress Level. Moderate to high, especially during recovery. Internal conflict between original self and defensive self.

Representatives.

• Real: Tiffany, Sara Crouson (“Cluster B Milkshake”), Jacob Skidmore (“The Nameless Narcissist”) gets close

• Fictional: Zuko (Avatar: The Last Airbender), Killmonger (Black Panther)

Note on Zuko. He represents the recovery path: Starts shame-based and desperate for validation from rejecting parent, develops grandiosity and aggression, then transforms through stable attachment (Iroh) and environment change. Ends genuinely heroic. If you identify with Zuko, you may be seeing your own recovery potential.

Tier 4: High Distress

These clusters involve chronic suffering.

Cluster 9: The Classic Narcissist

NPD without psychopathic substrate – empathy present but defended.

Feature Profile.

• G: G-minimal, possibly G-reactive, some kind of high sensitivity

• N: N-normal or N-hyperactive

• E: E-I-disorganized, E-C-unattuned/E-C-golden/E-C-scapegoat

• D: D-narcissistic

• B: B-subclinical

• A: A-narrativizing or A-retroactive

Narrative. Narcissistic personality disorder without psychopathic substrate. Narcissistic structure developed as adaptation to parental projections (unattuned parenting, golden/scapegoat dynamics) but retains capacity for empathy, guilt, shame, remorse – these are defended against, not absent.

Unlike sovereigns, when these folks hurt someone, they must actively avoid mentalizing the victim or use some other trick to not feel it. They are also more likely to suffer as a result of personal setbacks, to idealize and devalue other individuals, and to cycle between very different mood states. Their disorganized attachment is more obvious.

Distinguishing Features.

• Visible empathy and moral emotions

• Oriented toward admiration rather than power (except instrumentally)

• If sadistic, sadism is ego-dystonic (conflicted)

• Cycles between grandiose and vulnerable states

• Drug addictions (cocaine, alcohol)

Distress Level. High. Cycling between states; shame breaks through; constant defense required.

Representatives.

• Real: Jordan Monroe

• Fictional: Don Draper (Mad Men), BoJack Horseman, Tony Soprano

Cluster 10: The Reactive Antisocial

Hot violence and dysregulation – BPD+ASPD overlap.

Feature Profile.

• G: G-reactive + G-impulsive

• N: N-hyperactive

• E: E-I-disorganized, E-C-chaotic, E-C-neglect, E-C-punitive, E-P-antisocial, E-P-peer-failure, E-P-identity-diffuse, E-A-unstable, E-A-addiction, E-A-crime

• D: D-autonomic-defensive

• B: B-factor-2, B-violent

• A: A-externalizing, A-amnestic, A-retroactive

Narrative. High emotional reactivity meets chaotic environment. Never develops stable self-structure or regulation. Remains reactive, impulsive, dysregulated. Violence is reactive (hot-blooded, rage-driven) not instrumental (cold, calculated).

Combined features of borderline and antisocial personality. Splitting, unstable identity, chaotic relationships, overpowering emotions. Substance use common.

Distinguishing Features.

• Reactive, hot violence

• Emotional dysregulation obvious

• Unstable identity

• Chaos in life domains

• Impulsive criminality

• Obvious projections

• Drug addictions

Distress Level. High. Chronic dysregulation; frequent crises.

Representatives.

• Fictional: Tommy Shelby (Peaky Blinders, more sympathetic portrayal); many crime drama antagonists

Cluster 11: The Echoist

Self-effacing, other-focused – the complement to narcissism.

Feature Profile.

• G: G-minimal

• N: N-normal or N-hyperactive

• E: E-I-preoccupied, E-C-scapegoat + E-C-parentified, E-P-failure

• D: D-echoist

• B: B-minimal

• A: A-absorbed, A-selective, A-narrativizing

Narrative. Not psychopathic at all, but included because echoists are the complement to narcissists and sovereigns. Learned to minimize self, accommodate others, avoid conflict. Takes responsibility for others’ actions and feelings, self-abnegates, uses false self defenses to suppress needs, selfishness, anger. May be organized around compensation for a feeling of uselessness or badness or in search of a cohesive borrowed identity.

Often raised by narcissistic parent (absorbed blame, became caretaker) or echoistic parent (copied expectations). May seek out narcissistic partners (intermittent feeling of being needed).

Distinguishing Features.

• Self-effacing

• Takes blame for others

• Feels a lot of guilt and debt

• Accommodating to a fault (to offset imagined debt)

• Attracted to people with strong needs and expectations

• Often depression, anxiety

Distress Level. High. Constant self-punishment, guilt, social self-isolation out of shame or fear of doing harm. Chronic self-neglect leads to burnout.

Representatives.

• Fictional: Various “enabler” characters

See my article on echoism.

Cluster 12: The Messianic

Abstract echoism + grandiosity – self-sacrificing world-saver.

Feature Profile.

• G: Variable

• N: Variable

• E: E-I-preoccupied, E-C-neglect, E-C-parentified, E-C-golden

• D: D-echoist (toward abstractions) + D-narcissistic

• B: B-subclinical to B-factor-1

• A: A-strategic, A-narrativizing

Narrative. Echoist toward an abstraction (world, humanity, future generations) rather than a person. Grandiose belief in own capacity to save/fix/improve the target. Willing to sacrifice self (no need for recognition) and possibly others (ends justify means).

Differs from standard NPD (doesn’t need admiration), sovereignty (not about power for its own sake), and echoism (toward abstraction, not person).

Distinguishing Features.

• Oriented toward abstract values, not admiration, power, or a partner

• Doesn’t need recognition – outcome is the reward

• May use controlling, Machiavellian means

• Self-sacrificing component

Distress Level. Variable. May be content (mission provides meaning) or tormented (weight of moral decisions) or collapsed (benevolent but failed god) or burned out (hit human limitations).

Representatives.

• Fictional: Ozymandias (Watchmen)

• Some effective altruist or political figures

See my article on echoism (includes more examples).

Cluster 13: The Modeled

Learned psychopathic behavior; substrate may be intact.

Feature Profile.

• G: G-minimal

• N: N-normal

• E: E-P-antisocial, E-A-crime, E-A-addiction

• D: D-autonomic-asymmetric, D-avoidant

• B: B-mixed

• A: A-strategic, A-narrativizing

Narrative. Psychopathic behavior without rigid identity that is learned only in adolescence, due to environment that necessitates them. The empathy is there, defended against. May not know other ways to be. High recovery potential if new models are provided.

Distinguishing Features.

• Embedded in criminal peer groups

• Suffering from homelessness or addiction

• Used to be different in childhood

• Can learn alternative patterns

• Substrate is intact

Distress Level. High, due to intact guilt and remorse, and due to whatever drives the behaviors – homelessness, addiction, blackmail.

Representatives.

• Fictional: Jesse Pinkman (Breaking Bad), Michael Polischka (Knallhart)

Distinguishing Questions

Not sure which cluster fits? These questions can help:

1. “Were you always roughly as you currently are, or do you remember being more sensitive and empathic?”

   • Always → Clusters 1, 2, 3, 4, 5, 7

   • Remember different → Clusters 6, 8, 10, 13

2. “When you hurt someone, do you have to avoid thinking about how they feel?”

   • Yes → Clusters 6, 9, 11, 12, 13 (affective empathy present, defended)

   • No → Clusters 1, 2, 3, 4, 5, 7, 8, 10 (affective empathy minimal)

3. “Is it worse to be seen as ‘out of control’ or ‘evil’?”

   • Out of control → Clusters 5, 8

   • Evil → Cluster 9, 11, 12, 13

   • Neither bothers me → Clusters 2, 3, 4

4. “What do you want from others?”

   • Admiration → Cluster 9

   • Control → Clusters 4, 5, 8, 12

   • Resources → Cluster 2

   • Stimulation → Clusters 1, 2, 3, 4, 5, 6, 8

Multiple Clusters

Many people fit multiple clusters or fall between them. This is expected – clusters are approximations, not discrete categories.

Common combinations:

• Sovereignism usually implies autonomism

• Narcissism and echoism can co-occur

• Recovering sovereigns may move toward classic narcissism or autonomism

The cluster framework is a starting point, not a final classification.

Next: The Choice

The final article explores recovery – what it means for different clusters, what’s changeable and what isn’t, and the honest trade-offs involved.

This is the fifth article in a series on understanding psychopathy. Previous articles covered the framework, biology, environment, and psychological structure. This article explores the different ways empathy can fail to influence behavior – because understanding the mechanism matters for understanding the person.

Introduction

“Lack of empathy” is a core feature of psychopathy, but the phrase hides enormous variation. Someone who can’t perceive distress is different from someone who perceives it but feels nothing. Someone who feels distress but doesn’t care is different from someone who cares but acts anyway because rage overwhelms.

This article presents empathy as a pipeline with multiple stages, each of which can fail independently. Understanding where someone’s empathy fails helps predict their behavior and suggests different intervention approaches.

I’ll call this pipeline C for connective or communicative, because it comprises more aspects than just the cognitive (simulation) and affective components that are commonly associated with empathy.

Empathetic Connection as a Pipeline

Connection isn’t one thing – it’s a sequence of processes:

PERCEPTION → SIMULATION → AFFECT → MOTIVATION → BEHAVIOR

Each stage can fail independently:

• Perception. Noticing and recognizing the other’s state.

• Simulation. Modeling what the other is experiencing.

• Affect. Generating an emotional response to that model.

• Motivation. Caring about the emotional response – attaching moral weight.

• Behavior. Translating care into action (or restraint).

Books like Against Empathy make the case that affective empathy is often a bad guide to moral decision marking. I agree. The affective component can feel very lovely and connecting and it can be motivating when there are competing considerations – for good or ill.

As such, what is important in this pipeline is that processes of perception and simulation eventually lead to behavior (which requires motivation). Some of my friends can do that reliably without the affect component, and in many everyday and life decisions – such as where to donate – I try to avoid getting biased by affective responses.

The resulting pipeline goes:

PERCEPTION → SIMULATION → MOTIVATION → BEHAVIOR

Different failures at different stages produce different presentations.

Stage 1: Perception Failures

The other’s distress isn’t registered at all.

C-P-inattention: Not Noticing

Not attending to distress cues. Distracted, focused elsewhere, not looking.

This is the default in our society. Millions of humans dying of preventable diseases, trillions of animals suffering in factory farms, sextillions of animals living in abject conditions in the wild, many more throughout the future. Most people don’t realize it in the first place.

C-P-aversion: Flinching Away

Actively avoiding perceiving the other’s state. Like C-P-inattention but with a defensive purpose behind it. Some pwNPD flinch away from mentalizing people they hurt – they could perceive their distress, but doing so would produce an empathetic response they don’t want to feel. So they avert their gaze.

This suggests the empathy capacity is present but defended against.

This is the default in our society among people who do realize how much suffering there is. They look away.

C-P-non-recognition: Not Recognizing

Sees the cues but doesn’t recognize them as distress. May be related to alexithymia (difficulty identifying emotions) or the difficulty reading neurotypical social cues that autistic people run into. The signal is there, but it’s not decoded.

Most people struggle with this when interpreting the distress cues of fish and most invertebrates because they are too different.

C-P-objectification: Seeing Object, Not Person

Deliberately or habitually classifying the target as an object rather than an agent with experiences.

One example from my own experience: When a friend gets a minor injury and I need to help rather than pass out from empathic distress, I see them as “a damaged object that needs repair” rather than “a person in pain.” This is strategic objectification – useful in the moment, then released.

For some people, this is the default mode. Others are objects, resources, obstacles – not minds with experiences. This bypasses the empathy pipeline entirely, so it’s also useful for empathizing with sadism – i.e. behave sadistically toward literal objects to experience what it is like.

I imagine that most people have an objectified relationship towards most invertebrates, excluding large ones like octopuses. When an animal dies, it’s telling whether people react with indifference, appetite, or wanting to bury the body.

Stage 2: Simulation Failures

These come in two flavors: mentalizing the other and mentalizing the self – or failing to.

C-S-no-pain: Can’t Imagine Pain

If your own pain perception is blunted, you may lack the experiential substrate to simulate others’ pain accurately.

One friend said, “They’re just nociceptive stimuli. I can just ignore them.” If that’s their relationship to their own pain, how would they simulate the agony of someone else’s injury? (Then again I find it more likely that they noticed a lack of fear of pain and confabulated a reason for it when really the reason is that they don’t have a fear response.)

This is not callousness in the moral sense – it’s a genuine simulation deficit. They’re not choosing to ignore your pain; they can’t imagine it because they don’t experience pain that way themselves.

C-S-no-self: Can’t Imagine Attacks on Selfhood

M.E. Thomas (D-anatta) describes having no selfhood to attack. Physical pain she can imagine, somewhat. But humiliation? Self-image injury? These don’t compute because she doesn’t have the internal referent. Or they didn’t. She’s trained her selfhood over the past ten years and how has a better idea of what it’s like also introspectively.

If you have no self, you can’t simulate what it feels like to have your self attacked. This means certain kinds of harm – infractions of autonomy, reputational damage, humiliation, self-image injury – may be genuinely incomprehensible, not merely disregarded. Autonomy is a particularly interesting case because if someone has no self whose autonomy to value, how would they realize that there is one person who they can manipulate (themselves) but not anyone else, unless they’ll make them angry. It’ll seem like a weird and arbitrary social mechanism.

C-S-no-shame: Can’t Imagine the Feeling of Shame

Some people don’t have the feeling of shame. Perhaps they don’t recognize the idea of being part of a social group, and hence motions that these groups make to use shame against them end up being just empty motions.

Such a person might imitate these motions to express their disapproval or persuade someone to do something, but they don’t realize that for the other person, these motions are deep cuts to their self concept.

Along the same lines, it makes sense to recognize simulation failures like C-S-no-guilt (which might cause someone to use guilting motions on another which they would scruple to do if they knew what it feels like) or C-S-no-fear (which might cause someone to act erratically or threatening, which they would try to control if they knew what the fear feels like that they instill).

C-S-projection: Simulating the Wrong State

Simulating what you would feel rather than what they feel. A masochist might project their own relationship to pain onto a victim, imagining the victim enjoys it. A stoic might project their own equanimity, underestimating the victim’s distress. A person who enjoys learning about personality disorders might project their enthusiasm, underestimating the victim’s annoyance at listening to nothing else for years.

C-S-underestimation: Simulating But Minimizing

The simulation happens but the intensity is underestimated. “It’s not that bad.” “They’ll get over it.” “Pain is just a signal.”

This may be related to C-S-no-[substrate] – e.g., if you don’t experience much pain yourself, you may simulate others’ pain as similarly mild.

C-S-hypermentalizing: Constructing False Attributions

Simulation happens but produces inaccurate, self-serving results. “They actually like it.” “They’re not really hurt.” “They’re exaggerating for attention.” “That’s their only purpose in life.” The model is constructed, but it’s wrong – and wrong in a defensive direction.

C-S-impulsive: Acts Before Reflection

The action happens before the empathic/moral processing can complete. The person might have arrived at a different resulting action if they’d paused – but they didn’t pause.

C-S-prospective: Can’t Anticipate Future Guilt

The person can feel guilt retrospectively, but not prospectively. In the moment of action, they can’t access the knowledge that they’ll feel terrible later.

“She gets angry, does bad stuff, and then regrets it. But she can’t anticipate that she’ll feel that way.”

This is state-dependent access – the guilt is real but only accessible in a different state. This may be related to N-disconnected.

C-S-state-dependent: Empathy Only in Certain States

Mentalizing is available in some states (calm, secure) but not others (angry, dissociated). The capacity exists but isn’t consistently accessible.

C-S-temporal-discounting: Future Doesn’t Weigh

Future guilt is acknowledged but heavily discounted. “I’ll feel bad later, but I don’t care now.” The temporal distance makes the future suffering abstract and powerless.

C-S-depleted: Too Exhausted for Restraint

The theory of ego depletion avers that self-control is a limited resource. If so, a person would mentalize comprehensively if they had the energy, but they’re too tired.

C-S-intoxicated: Chemically Disinhibited

Alcohol or other substances interfere with proper mentalizing.

C-S-retroactive: Rewriting to Prevent Guilt

After the action, the memory is rewritten to prevent guilt. “They deserved it.” “It wasn’t that bad.” “They got over it.” This prevents guilt from forming even when it otherwise would – and prevents learning from the experience for the next round of mentalizing.

Stage 3: Affective Failures

Simulation doesn’t produce an emotional response.

C-A-blunted: Weak or Absent Affect

The simulation is accurate, but it doesn’t produce an emotional response. You understand they’re in pain, but you don’t feel anything about it. This is the classic N-hypoactive pattern – the amygdala and insula don’t fire, so there’s no affective resonance.

This is also something that affects most people when it comes to the suffering of invertebrates and fish.

C-A-suppressed: Affect Generated But Suppressed

The emotional response is generated but actively suppressed. This is different from C-A-blunted – the affect is there, but defenses keep it out of awareness. Friends of mine describe practices of automatic or habitual rationalizing and compartmentalizing to not feel bad. The affect would be there if they let it.

C-A-inverted: Pain Produces Pleasure

The simulation is accurate, and an affect is generated – but it’s positive. The other’s pain produces pleasure. This is sadism in the true sense: Inverted empathic affect. It might translate to pain beyond nociception, e.g., humiliation.

C-A-override: Stronger Affect Drowns Empathy

The empathic affect is generated, but it’s overwhelmed by a stronger affect – rage, fear, desire, domination. The person does feel the other’s distress, but they feel their own rage more, and the rage wins. The pain might even enhance a competing feeling of domination.

One friend of mine gets angry, does things that are in violation of her values, and then regrets it and feels remorse afterwards. But she can’t anticipate that she’ll feel it when she does those things. At that time it feels perfectly justified.

This is not absence of empathy – it’s empathy overwhelmed by rage, with retrospective access.

C-A-redirected: Empathic Distress Becomes Anger

The empathic distress is generated, but instead of producing compassion, it produces anger at the person causing the distress – sometimes the victim themselves. “Stop making me feel bad about you!”

This can lead to victim-blaming and avoidance: The distress is aversive, so the source of distress (the suffering person) becomes aversive. Specifically the patterns that I’ve identified are resentment and control-seeking: You feel wronged because someone is asking you for a kind of accountability that has never been granted you in the past; you’ve internalized that you mustn’t show the emotions the person is showing, so you feel like they’re violating your conduct norms; or you’re used to getting endlessly attacked for your mistakes, so you try to never admit to any.

C-A-anhedonia: Sadism as One of Few Pleasures

Some people have generally blunted pleasure (anhedonia), but sadism is one of the few things that still produces feeling. The sadism isn’t about the victim per se – it’s about accessing any feeling at all. Perhaps it’s not even particularly pleasurable, but it is intense, it is a reprieve from the nothingness.

This may relate to the “void” that some psychopathic presentations describe – the emptiness that gets filled with substances, thrills, or sadism.

Stage 4: Motivational Failures

Affect is present but doesn’t translate to moral concern.

C-M-amoral: No Moral Weight Attached

The empathic affect is present – you feel something when you see their pain – but it doesn’t connect to moral value. It’s just information, not a reason to act or refrain.

Here it’s important to distinguish descriptive and metaethical moral realism. (I find the first fairly uninteresting but reject the second, not on the grounds of moral realism but on the grounds of something like metaethical expressivism and a resulting global market of moral preferences.)

Descriptive Moral Relativism (DMR). As a matter of empirical fact, there are deep and widespread moral disagreements across different societies, and these disagreements are much more significant than whatever agreements there may be.

Metaethical Moral Relativism (MMR). The truth or falsity of moral judgments, or their justification, is not absolute or universal, but is relative to the traditions, convictions, or practices of a group of persons.

But some people do subscribe to something like MMR, in which case they attach or don’t attach a moral weight to the same states depending on who is affected by them, where, or in which context.

This is a bit of a catch all for amoral non-responses that are not covered with more specificity below.

C-M-dehumanization: They Don’t Count

The target is classified as not deserving moral concern. “They’re not really people.”

C-M-justification: Feels Justified

Moral concern is overridden by felt justification. “They deserved it.” “They started it.” “It was self-defense.” The empathic information is present, the moral weight is attached, but the justification outweighs it.

This is C-A-override’s sibling: Not just that an emotion overwhelms another, but that it comes with a notion that it feels justified, so the action feels right.

C-M-ideological: For the Greater Good

Higher value overrides moral concern for the individual. “I’m saving the world.” “The ends justify the means.” “One death to prevent a million.”

Ozymandias from Watchmen is the extreme example – killing millions for calculated peace. The empathic information is present, the moral weight is attached (he knows it’s terrible), but the utilitarian calculation overrides.

C-M-competing: Other Desires Outweigh

Moral concern is present but weaker than other motivations – appetite, greed, revenge, lust, ambition. The person knows it’s wrong and feels it’s wrong, but does it anyway because they want to.

C-M-licensing: Earned the Right

“I’ve been good, so I can be bad now.” Moral accounting that permits occasional violations.

C-M-diffusion: Not My Responsibility

Bystander effect logic. “Others could help.” “It’s not my job.” Moral concern is present but responsibility is diffused.

C-M-identity: Empathy Threatens Self-Concept

Acknowledging the empathic information would threaten the self-concept. For sovereigns, admitting empathy might feel like weakness. For some ideologies, empathizing with the “enemy” is betrayal. So the empathic information is dismissed or suppressed.

Stage 5: Behavioral Failures

Behaviors are not sufficiently externally constrained.

C-B-others: Interpersonal Power Differentials

If the interpersonal power differential is minimal, the other person can effectively constrain one’s behaviors. If not, the behavior can’t be effectively constrained at this level, be it in the moment or through feedback to prevent future iterations.

Sources of power differentials: avoidant attachment, physical strength, life experience, martial arts experience, weapons, positions of institutional power, financial power, blackmail material, etc.

C-B-institutions: Institutional Power Differentials

In extreme cases, someone is above the law so that there’s not only an interpersonal power differential but a universal one, the lack of feedback that many dictators suffer from and that, e.g., caused Putin to underestimate the costs of the war on Ukraine.

Mapping Failure Points to Presentations

Different presentations tend to fail at different points. Understanding where the failure is helps target intervention – if there is one.

Why This Matters

For self-understanding. If you know your empathy fails at the simulation stage (you can’t imagine their pain because you don’t experience pain that way), that’s different from failing at the motivation stage (you feel it but don’t care). Understanding your mechanism helps you work with it.

For predicting behavior. Different failure points predict different behaviors. C-A-override (rage overwhelms) predicts reactive violence; C-A-blunted (no affect) predicts cold instrumentality; C-M-justification predicts violence framed as deserved.

For intervention. Some failure points are more modifiable than others. C-B-impulsive might be helped by slowing down, creating pause, like with the DBT STOP skill. C-A-suppressed might be helped by creating safety that allows the affect to emerge. C-A-blunted is probably stable.

For others’ self-protection. If you know someone’s failure point, you know how to protect yourself. C-A-override people are dangerous when triggered; C-A-blunted people are dangerous when you have something they want; C-M-justification people are dangerous when they feel you’ve wronged them.

Next: The Types

The next article presents the archetypal clusters – common profiles that tend to co-occur. Rather than abstract dimensions, we’ll look at recognizable types that readers may identify with.

This is the fourth article in a series on understanding psychopathy. Previous articles covered the framework, biology, and environment. This article covers the D (dynamic) and A (agency) levels – the psychological structures that developed and how you understand your own intentionality.

Introduction

The substrate (G and N) and the shaping (E) produce the self (D). This is the level most people think of when they think about “who someone is” – not their genes or brain scans (not merely their outward behavior either), but their psychological organization, their characteristic patterns of relating, their attitudes and motivations, their experience of themselves.

This article describes the major D-level variants I’ve identified through conversations with friends who have these presentations. It also introduces the A level – how people with different presentations understand and narrate their own intentionality.

These are modifiers, so it stands to reason that someone with only D-narcissistic would not be considered to be significantly psychopathic in any of the many senses, but I rather wanted to include more variants rather than fewer to make the system more expressive. This way it can be used to show, for example, that if someone has D-narcissistic and A-amnestic and acted violently in rage is not lying about not remembering it but actually doesn’t remember it, a sign that a diagnosis of ASPD would be unhelpful for them.

D-Level: Psychodynamic Variants

D-anatta: The Empty Self

The term comes from the Buddhist concept of anattā – “no-self.” Individuals with this presentation have minimal sense of identity. There’s no stable “I” that persists across contexts. At the extreme, there is no “I” ever.

Self aspect. The inner experience is one of absence and direct experience. Identity is borrowed from context: They become what the situation calls for, but it’s a thin mask, not something that actually gets imbued with identity. There’s no intrinsically compelling reason for them to identify with “themselves” more than with others, so respecting others’ autonomies can be unintuitive to them.

Relational aspect. They mirror those around them, take on contextually appropriate personas, and usually don’t feel like they are in touch with anything that is genuinely “them” underneath the adaptations.

Agency style. Often A-observational: “I watch myself do things.” There’s no felt sense of deliberation or choice – actions emerge from some ranking process without an “I” in between. They’ll say things like that they’d rather not travel because they tend to commit crimes when they do, or that their loved ones please call the police on them if they act abusively. It’s as if their own body is as foreign, mysterious, and wayward to them as it is to a third party.

How it differs from other presentations. D-anatta individuals don’t have the preoccupation with power or self-image of sovereigns. Control and grandiosity are means to an end for them, but they don’t need them or have identity invested in them. They may be predatory (D-predatory) in an instrumental way, but it’s not about domination – it’s about getting what they want, filling the void, passing time, or curiosity. When faced with hardship, D-anatta is more likely to say “I don’t care,” whereas D-sovereign is more likely to lean into a kind of stoicism that I call contumacious pride, “You think you can beat me, but pain only makes me stronger!”

Example. This interview with M.E. Thomas describes her no-self experience well. But I’ve also run into someone who talks like M.E. Thomas (or James Fallon for that matter) but has an avoidant vibe. How can someone without a self have avoidant attachment? I will just assume for now that that’s possible, but if I change my mind, I’ll update these sections.

D-narcissistic: The Unstable Self

This is classic narcissistic personality organization – unstable self-esteem, a fear of worthlessness and compensatory need for specialness or admiration, with episodes of shame, anxiety, and sensitivity to rejection.

Self-aspect. A deep fear of worthlessness that is defended against by a positive split on the self where any perceived flaws are disavowed, and the person perceives a lot of agency and optimism, and a negative split where the supposed flaws are conscious but the agency is disavowed to externalize the responsibility for the flaws. For people with a pure D-narcissistic presentation, the positive states are often unstable and need to either be actively maintained through positive feedback or protected by avoiding negative feedback. A view of the self as defined by how it appears to others (a teleological mode) makes it highly sensitive to circumstantial factors, so that these can appear as threats to be defended against.

Relational aspect. A defended form of disorganized attachment where there is often a person who is idealized as ideal source of parental love but who is simultaneously also a threat because criticism, rejection, or boundaries are devastating to the self and can only be tolerated when they can be safely dismissed, which requires the paradoxical devaluation of the idealized surrogate parent. These aspects of self and other – the idealized other and child-like self, and the devalued other and self-sufficient self – are incompatible and require splitting for them to stay unacknowledged. Relationships with those who are seen as mildly inferior are safer, because they can be dismissed as needed, but are also unsatisfying because they don’t provide the safety of paternal love.

Agency style. Often A-narrativizing or A-retroactive in the positive split and A-externalizing or A-selective (true-but-partial explanations) in the negative split.

How it differs from D-sovereign. D-narcissistic is oriented toward positive feedback or admiration and away from negative feedback. It’s organized around a fear of worthlessness and a corresponding compensatory need to be special. D-sovereign (below) is oriented toward power – controlling others. It’s organized around a few of being helpless or exposed to someone else’s will. Crucially, D-narcissistic retains capacity for shame, guilt, and empathy (even if defended against), while D-sovereign has these dampened by the psychopathic substrate, an antagonistic and sadistic or Machiavellian outlook on others, and avoidant attachment more toward civilization as a whole than just individual friends or partners.

D-echoistic: The Self-Effacing Self

The opposite of narcissistic grandiosity – minimal self-assertion, other-focused, needs subordinated to others’.

Self aspect. The inner experience is of having no preferences, or of preferences being illegitimate. The self exists to serve.

Relational aspect. Accommodating to a fault. May attract narcissists (complementary dynamics). Takes responsibility for others’ emotions and actions.

Agency style. Some types are structured along the lines of D-narcissistic, while some might rely on A-absorbed, especially if it’s a borderline-like presentation of echoism: “It’s my fault they did that to me.” The A-absorbed presentation in echoism confuses self and other and is characterized by impaired reality-testing, unlike the presentation of A-absorbed in sovereignism.

See my article on echoism and its relationship to narcissism.

D-sovereign: The Power-Oriented Self

This is my term for the combination of pathological narcissism with psychopathic callousness and ego-syntonic sadism – what Otto Kernberg called malignant narcissism, but with a focus on the control-orientation or sadism rather than the paranoia. As such it’s a wider category.

I’ve written about this extensively in my article on sovereignism. The key distinction from standard narcissism: Sovereigns are oriented toward power and control, not admiration or avoiding criticism.

Self aspect. The inner experience is of needing and enjoying control – of others, situations, outcomes. As such, they’re oriented around a fear of being helplessly exposed to someone else. Others are threats not acutely for a few hours when they say something critical but chronically so that they need to be defensively devalued at all times. This extends to state authorities who are just as much antagonists as are individuals. At first approximation there is no one whose opinion of them deserves respect so their self-esteem can appear stable, affected only by their own perception of their own level of control.

Relational aspect. At first approximation, relationships are only possible through controlling the other (there is no destabilizing need to idealize them) or by psychologically merging with the other, forming one joined identity, which is very risky with humans and easier with companion animals. But I see sovereignism as a particularly stable variation on narcissism, so the first approximation is usually an oversimplification: Some individuals do come to respect others as equals or do engage in mild forms of idealization that are more typical of standard narcissism. Sovereigns almost invariably have high avoidant attachment and low preoccupied attachment in their grandiose moods.

Most people are immune to shaming or guilting from their outgroup – the people who are culturally opposed to them. I’m a bit of a lefty, so if someone were to shame me for being trans, it wouldn’t connect with anything. D-sovereign is like being an outgroup of one – alone and persecuted – the ultimate defense against shame. Ones who are not particularly paranoid may feel closer to an ingroup of one – immune to shame but not persecuted, like D-avoidant.

Why the psychopathic substrate stabilizes it. People with standard pathological narcissism need to break their empathy at the level of the mentalization – the flexible simulation of their own and the other’s mind. They impair their cognitive empathy to keep their defenses intact. This impairs their functioning as better understanding is generally adaptive. Add to that the paradoxical conflict between idealization and self-sufficiency in romantic relationships, and the self-deceptions in standard pathological narcissism become very unstable. Avoidant attachment and even ego-syntonic sadism – not respecting others or actively and antagonistically wishing them harm – work hand in hand with impairments of affective empathy to obviate the need for this self-limiting defense. The pervasive reduced respect for others thwarts the ability to idealize. The result is a personality structure that is highly fortified against outside threats – “I have to humble myself because no one else can humble me,” as a friend of mine put it (paraphrased).

Agency style. A-strategic, A-retroactive, A-narrativizing, A-selective, A-absorbed. Seeing themselves as “out of control” is worse than being seen as “evil” (which is powerful and positive). They may also claim impulsive actions as intentional regardless of whether others may consider them blameworthy. They may say, “First I try positive reinforcement; if that fails, I gaslight; if that fails, I rage as deterrent; and only if that fails, I act remorseful” – but sometimes this is a post-hoc rationalization of a reactive sequence, not genuine planning. When suffering abuse, some sovereigns will blame themselves along the lines of A-absorbed but without confusion of self and other or impaired reality testing: E.g., they’ll claim the abuse is their own fault because they could’ve left or killed the abuser but chose not to.

D-autonomic: The Freedom-Oriented Self

Hypersensitivity to perceived constraints on freedom; autonomy as the core value. Mild form of D-avoidant (only mildly avoidant) and D-sovereign (little sadism or callousness).

Self aspect. The inner experience is organized around freedom from constraint. Any perceived imposition threatens the self. “No one tells me what to do.” The reactive anger can cause momentary callousness.

Relational aspect. Relationships are evaluated through an autonomy lens. Implicit expectations feel like impositions. Only explicit agreements count. “I didn’t agree to show up, so your expectation is your problem.” Associated with fairly high avoidant attachment but stopping short of the antagonistic attitude vis-a-vis society as a whole that characterizes sovereignism or the cold disregard for it that characterizes D-avoidant.

Subtypes.

• D-autonomic-offensive. Controls others to secure own autonomy. “I’ll control everything so no one can control me.”

• D-autonomic-defensive. Reacts to perceived constraints with rage or withdrawal. “You asked me to work overtime? I quit.”

• D-autonomic-symmetric. Values others’ autonomy equally. “I respect your freedom; respect mine.” Principled libertarianism.

• D-autonomic-asymmetric. Values own autonomy, disregards others’. Hypocritical application.

• D-autonomic-non-interfering. Extreme symmetric autonomism: Won’t violate others’ autonomy even to help them. “Your life is your business; I don’t interfere for good or evil.”

Origin. Often from E-I-avoidant and E-C-controlling environments. The child who was never allowed to have boundaries rebels by defending them on a hair trigger.

D-avoidant: The Walled-Off Self

Extreme dismissive attachment taken to its endpoint. Nothing matters; no one matters; independence is the only value.

Self aspect. The inner experience is of detachment – from emotions, from desires, from connection, from society. There was usually a lot of pain that led to this walling-off, but the pain is no longer accessible.

Relational aspect. When attachment is as vital as it is dangerous (disorganized attachment), one can either enslave oneself to it entirely or reject it entirely to escape the conflict. D-avoidant is the result of that rejection, much like D-sovereign: Most people are immune to shaming or guilting from their outgroup – the people who are culturally opposed to them. I’m a bit of a lefty, so if someone were to shame me for being trans, it wouldn’t connect with anything. D-avoidant means being an ingroup of one, the ultimate defense against shame.

Relationships are minimal or purely instrumental. Attachment is threatening. Someone with disorganized attachment may choose this path as a momentary release. Some may lean into universalized anger or hate to reach this state. They can often leave it as suddenly as they’ve entered it.

How it differs from D-sovereign. Extreme and pervasive avoidant attachment is part of D-sovereign but D-avoidant is a purer form that doesn’t imply any preoccupation with control or sadism or a rigid self-image the way D-sovereign does, though individuals with D-avoidant may behave in sadistic ways out of boredom.

D-secure: A Stable Self

A normal self, usually associated with secure attachment.

Self aspect. An inner experience of interpreting perceptions from a consistent perspective; reasonably stable preferences across time and contexts; continuity over time; complexity of contradictory drives that can be resolved.

Relational aspect. Can trust others but also trusts themselves. Respects others’ autonomy but reacts proportionately to infringements on their own autonomy.

Agency style. Often A-narrativizing or A-observational.

A-Level: Agency Attribution Styles

Beyond D-level structure, people differ in how they understand and narrate their own intentionality. This is the A level – not a separate dimension of the self, but a lens on how the self is understood.

A-observational

“I watch myself do things.”

The person has minimal sense of deliberation or choice. Actions emerge from some process they observe rather than control. Common in D-anatta.

Example: “I don’t have an internal narrative anymore. My decisions flow automatically from some ranking process with no intention happening between ranking and action.”

A-strategic

“I plan, then act.”

Genuine prospective intentionality. The person simulates outcomes, evaluates options, and chooses. This may be accurate self-description, or it may be how they prefer to see themselves.

A-narrativizing

“I tell myself a constant story about what I’m doing.”

Real-time rationalization. The person maintains a running narrative that explains and justifies their actions as they happen. The narrative may or may not be accurate.

A-retroactive

“I act, then explain.”

Post-hoc rationalization. The person acts first and constructs an explanation later when challenged. Common in D-narcissistic.

A-selective

“I delayed because I was exhausted.” (True, but only 20% of the picture.)

Picks one true motivation and presents it as the whole story, omitting ego-dystonic motivations. The stated reason is accurate but incomplete. Common across presentations.

A-externalizing

“They made me do it.”

Locates causation externally. “It wasn’t me; it was the situation / my upbringing / their provocation.”

A-absorbed

“It’s my fault they hurt me.”

Takes responsibility for others’ actions. Common in D-echoistic.

Maximizes self-attribution to avoid the intolerable feeling of helplessness. Common with D-autonomic.

A-amnestic

“What mistake? I don’t remember that.”

Forgets ego-dystonic actions. Dissociative amnesia for events that would threaten the self-concept.

Why D-sovereign and N-hypoactive Can Look Tightly Correlated

This is speculative, but it matches a pattern I’ve noticed: The “cold” substrate (low fear, low guilt, low empathic cost) seems to co-occur with NPD more often than a naïve independence model would predict (1% prevalence of N-hypoactive × 5% prevalence of NPD = 0.05% prevalence of sovereignism – or lower if sadism is independent). A few mechanisms could produce this:

Early Trauma

Maybe these sovereigns were born with a regular amygdala, but neglect trauma in their infancy caused E-I-avoidant, and that caused their originally more active amygdala to atrophy at such an early age already (maybe before age 3 or so when they can retain memories) that they remember no time when they still felt different.

• Sovereigns seem to invariably have high avoidant attachment.

• Violent trauma is more likely to cause E-I-disorganized and N-hyperactive or N-dissociative.

• E-I-avoidant also prevents memory formation of attachment traumas (these cease to be traumatic once all expectations of attachment are destroyed), so these sovereigns are more likely to report a childhood that was “normal” or “fine” with little detail.

• Replications of the Strange Situation experiment (ages 9–30 months) show that avoidant attachment is still associated with high stress at that age. This might change as the amygdala atrophies. Perhaps it is these children that consistently develop D-narcissistic, D-sovereign, D-autonomic, and others.

• Some sovereigns are closer than others to “malignant narcissism,” i.e. are paranoid, have ADHD, and generally show subtle signs of N-hyperactive rather than N-hypoactive but exert strong self-control most of the time. Perhaps the atrophy is incomplete in their case. Their etiology is less surprising, and perhaps there is a range from this less surprising to the more surprising (N-hypoactive) presentation due to different levels of atrophy.

Punitive Parenting

Children with N-hypoactive can be challenging to parent, and parents may fall back on punitive parenting methods that incentivize the child to become more conniving to escape the punishments. Parent and child become enemies along the way, which causes the child to develop E-I-avoidant (avoidant attachment) and high trait antagonism, and thus a sovereign outlook against life.

• Arguably, such parenting is more likely to produce E-I-disorganized, but:

• Perhaps there is a temperamental difference that causes some children to develop E-I-avoidant under these circumstances, and

• NPD might serve to bolster the self-concept in the grandiose mood for ostensible E-I-avoidant despite actual E-I-disorganized. I’ve seen some of my sovereign friends in grandiose and vulnerable states, and the devalued self concept of the vulnerable state provides the preoccupied component for overall E-I-disorganized.

Bundled Transmission (Psychopathic Parenting)

A psychopathic parent can “bundle” two things: (1) constitutional callousness/fearlessness (G-callous → N-hypoactive) and (2) narcissism-inducing shaping (E-I-avoidant/disorganized, E-C-unattuned/neglect/controlling). If early care is emotionally thin, conditional, or coercive, children may end up seeking substitutes for reliable love and safety (admiration, dominance, invulnerability) without having internalized stable parental introjects. In this picture, this “vacuum” can be filled to form D-anatta, D-narcissistic, or D-sovereign depending on what gets reinforced.

• This seems less likely to me because G-callous and N-hypoactive from birth should have some shielding effect against attachment trauma and thus NPD.

• Many of my sovereign friends have unemotional or sadistic parents, but it’s unclear whether that’s a sign of G-callous or also N-hypoactive from atrophy and avoidant attachment.

Attachment Compromise Hypothesis

Many sovereigns still want people in their orbit but fear dependence. Control can be a way to do closeness without vulnerability: “I can have you as long as I’m not at your mercy.” That makes N-hypoactive (low fear) and D-sovereign (control-based relating) a stable pairing.

Stability Filter

Pure NPD can be unstable over time: shame collapses, cycling, and relational fallout. Add a “cold” substrate and the grandiose control-structure can become more stable and socially survivable. By the time you’re mostly meeting adults in their 30s–50s, the stabilized variants may be overrepresented in the people who remain networked and high-contact.

Adult Ecology Visibility

By the time you’re meeting people in their 30s–50s, the profiles that stay socially networked and relationship-active tend to be the ones that can maintain stable roles (work, friendships, partnerships). Sovereigns often can (they can be charming, competent, and strategic). Some non-NPD psychopathic variants are less relationship-motivated, less self-disclosing, or more likely to drift into isolation, so you see fewer of them up close.

Taken together: N-hypoactive doesn’t “cause” sovereignty, but it makes sovereignty viable, and certain developmental and temperamental pathways make it reinforcing and stable.

Why Sovereigns Say “It’s Just Strategy”

One pattern I’ve noticed in conversations with sovereign friends deserves special attention: When asked why they do normal, prosocial things – being nice, maintaining friendships, helping people – they often say something like: “It’s just a social tactic. Besides, it makes me look good.”

This sounds suspiciously convenient. Every action, no matter how kind, gets explained as Machiavellian – and the explanation always fits perfectly with the sovereign self-concept. When motivations seem that simple and that aligned with the self-image, it’s usually a sign that something is being compressed or obscured: A-selective.

Here’s how Tiffany explained what’s actually happening:

The Explanation Serves Multiple Functions

1. It’s partly true. Usually the motivation really is strategic to some extent. Early in the recovery, this is typically the only conscious motivation.

2. It’s protective. “I don’t want to be taken advantage of. I hide or obscure whether I’m doing it for a self-serving or altruistic reason.”

3. It covers genuine ambiguity. “It can be difficult for me to tell whether I think someone’s useful or I like them being around.”

4. The strategic language comes easier. The utilitarian/calculating vocabulary is more fluent than emotional vocabulary.

5. It’s an honest warning. “I want them to realize that even when I’m doing it because I think I’m helping them, it can come out as selfish, coercive, and maybe even sadistic.”

The Actual Motivational Landscape

When a sovereign says “I’m being nice because it keeps them around,” the reality might be a combination of:

• Sometimes: Genuine liking – “I like people. I want to do nice things with and for people I like.”

• Sometimes: Utility calculation – keeping someone around for what they provide.

• Sometimes: Control/shaping – “trying to change someone to be a better companion.”

• Sometimes: Just being nice – no ulterior motive at all.

The “it’s strategic” explanation roughly fits all these scenarios, so it gets used as a catch-all. As my friend put it: “That explanation roughly fits the different situations I end up in and is still roughly true if I’m being shallow as well.”

Why This Matters for Recovery

In more dissociated states, sovereigns have an illusion of clarity about their motivations – a typical sign of operating from a false self. In more connected states, they have less clarity – they’re fumbling for words, unsure what they actually feel.

For those in recovery, learning to tolerate the messier, less articulate emotional frame is part of the work:

“I have to allow myself to be more human if I want to feel like one.”

This connects to why secondary presentations have an easier time accessing this whole range of emotions than primary ones. If the strategic frame is a compression of messier underlying motivations (including genuine liking), then recovery involves expanding back to the mess – letting yourself fumble, tolerating ambiguity, using emotional language even when it doesn’t flow as smoothly.

The Illusion of Intentionality

A philosophical aside: The concept of “intentionality” may be less coherent than we assume.

Consider: If you simulate outcomes, narrate what you’re about to do, and then do it, it feels intentional. If you first do it and narrate later (without altering the ordering in the narration), it feels impulsive to many people regardless of when the simulation happened. But the narration doesn’t change the action. What makes one “intentional” and the other not?

The case of people narrating impulsive actions after the fact in a way that changes something about the ordering such that the actions feel intentional is an interesting case, perhaps one that sheds light on how narration works for many of us for all sorts of everyday decisions. The difference is then not so much anything about intentionality but rather what decisions we are ready to endorse as intentional and which we’d rather reject as impulsive. Intentionality becomes a question of values and their influence on our actions.

Someone without an internal narrative (A-observational) may experience decisions as simply emerging – no “I decided” moment, just observation of what happens. I would argue that this is the most accurate description of how decisions actually work, stripped of the narrative illusion.

Our social contract requires that, by and large, every person take responsibility for their own actions. That sentence is loaded with implicit assumptions of common knowledge of selfhood and responsibility, so being caught up in a shared illusion that makes these fairly arbitrary assumptions intuitive is interpersonally useful – a bit like it’s useful to think that you actually believe in God if you want to succeed as a religious leader.

Combinations and Profiles

D-level features can combine. A full profile might include:

• D-sovereign + D-autonomic-offensive. Power-oriented, sadistic, controls others to secure autonomy.

• D-anatta + D-predatory. Empty self, instrumental orientation. A cold manipulator without grandiosity.

• D-anatta + D-autonomic-symmetric. Values freedom, has minimal self, principled libertarian without the heat of D-autonomic-offensive.

Profiles are more informative than single labels.

How D Relates to G, N, and E

The D level is produced by the substrate (G, N) and shaping (E) interacting:

• G-callous + E-I-secure + E-C-normal → D-secure pathology. Constitutional callousness without adverse shaping may produce no pathology.

• G-callous + E-I-disorganized + E-C-controlling → D-sovereign. Constitutional callousness shaped by control-focused adversity may produce sovereignty.

• G-reactive + E-I-disorganized + E-C-chaotic → D-avoidant. Reactive temperament shaped by chaotic adversity may produce a walled-off defense.

• G-minimal + E-I-avoidant + E-C-golden → D-narcissistic. No particular callous loading, but narcissistic shaping produces an unstable self.

The mapping isn’t deterministic, but there are patterns.

Summary

D-Level. Psychodynamic self-structure and relational orientation. Major variants: D-anatta (absent), D-narcissistic (unstable), D-sovereign (power-oriented), D-avoidant (walled-off), D-autonomic (freedom-oriented), D-echoistic (self-effacing).

A-Level. How the person understands their own intentionality. Major styles: A-observational, A-strategic, A-narrativizing, A-retroactive, A-selective, A-externalizing, A-absorbed, A-amnestic.

Combinations. D and A features combine into profiles. A full profile (G + N + E + D + A + B) is more informative than any single label.

Next: The Mechanics

The next article explores how empathy fails – a detailed breakdown of the different stages at which empathic response can break down, and why different people fail to empathize in different ways.

This is the third article in a series on understanding psychopathy. The first article introduced the framework; the second covered genetics and neurology. This article covers the E (environmental) level – what happened to you developmentally, and how it shaped your expression.

Introduction

Two people with identical genetic loading can turn out completely differently. One becomes a functional professional with stable relationships; the other becomes a violent criminal. The difference is often the environment – what happened during critical developmental periods.

But “environment matters” is too vague to be useful. This article develops a more precise picture: What kinds of environments lead to what kinds of outcomes? Why does controlling adversity produce different results than chaotic adversity? How do the pieces fit together?

The E-Level Framework

Tiffany suggested three periods, of which I’ve subdivided the first into infancy and childhood for four developmental periods total, each with its own features:

Each period can be characterized by the type of environment experienced. Different environments at different periods have different effects.

E-I: Infancy (0–2 Years)

The Critical Task: Attachment

The primary developmental task of infancy is forming an attachment to caregivers. This attachment becomes a template for later relationships – a model of whether others are trustworthy, whether the self is worthy of care, and how to regulate emotions.

Attachment Styles

Mary Ainsworth’s research identified three organized attachment styles, and Mary Main later added a fourth:

Secure (E-I-secure). The caregiver is consistently responsive and attuned. The child learns that others are reliable and that distress can be regulated through connection. This is the foundation for healthy development.

Avoidant (E-I-avoidant). The caregiver is emotionally unavailable or rejecting. The child learns to suppress attachment needs and rely on self. Emotions are managed through distance and independence.

Preoccupied (E-I-preoccupied). The caregiver is inconsistently responsive – sometimes attuned, sometimes unavailable. The child becomes hyperactivated about attachment, constantly monitoring the caregiver’s availability. Emotions are amplified to try to elicit response.

Disorganized (E-I-disorganized). The caregiver is frightening or frightened – the source of both comfort and threat. The child faces an unsolvable dilemma: Approach the source of fear, or flee from the source of comfort? This produces contradictory, fragmented behavioral strategies.

Disorganized and Avoidant Attachment and Psychopathy

Disorganized and avoidant attachment are particularly relevant to psychopathy and related presentations. When the attachment figure is absent or irrelevant, the child will ignore them. When they are the source of fear, the child cannot develop a coherent strategy for managing distress in the first place. This can lead to:

• Dissociative responses. The child learns to disconnect from overwhelming experience.

• Controlling strategies. The child later develops either punitive-controlling (dominating the parent) or caregiving-controlling (parenting the parent) patterns.

• Fragmented self-structure. The self develops in unintegrated pieces, without coherent organization.

Many of the sovereigns I know have histories consistent with disorganized or avoidant attachment – dangerous parents, early role confusion, and later controlling strategies. The tendency towards avoidant attachment that they show may often be the result of their NPD (in grandiose moods) rather than genuine avoidant attachment.

E-C: Childhood (2–12 Years)

The Critical Tasks: Moral Development, Socialization, Self-Concept

Childhood is when moral emotions develop, social skills are learned, and the self-concept takes shape. Disruptions during this period can have lasting effects on all three.

Types of Childhood Adversity

Not all adversity is the same. Different types of adversity produce different adaptations:

E-C-unattuned (unattuned mirroring). Non-contingent marked mirroring. The caregiver is present but doesn’t accurately reflect the child’s emotional states. The child’s feelings are met with non-contingent responses – the caregiver’s reactions don’t match what the child is experiencing, usually because they instead match expectations that the caregiver projects on the child. This disrupts the development of mentalization – the capacity to understand mental states in self and others.

This is the classic etiology for narcissistic pathology described by mentalization-based treatment (MBT) researchers. The child develops without an accurate internal model of their own mind, and must construct a “false self” to adapt to the caregiver’s projected expectations.

E-C-neglect (emotional neglect). The caregiver is physically present but emotionally absent. The child must fend for themselves emotionally – there’s no guidance, no attunement, no help regulating distress.

Children adapt to neglect in various ways:

• Machiavellianism. Learning to get needs met through manipulation, since direct requests don’t work. The child practices attunement to needs that the environment can meet if they are extracted from it; the child represses needs that cannot be met that way. Control as surrogate for love. This can lead to the D-sovereign presentation.

• Echoism. Learning to suppress all needs, since expressing them is ineffective or risks punishment. Needs can instead be met by making another dependent on your generosity. Dependency as surrogate for love. This leads to the D-echoist presentation.

• Admiration-seeking. Learning to get needs met by seeking admiration as a substitute for love. If the parents are not forthcoming with love, this alternative can be won from peers, bosses, school authorities, or the public instead. This can lead to the D-narcissistic presentation.

E-C-punitive (punitive environment). The caregiver uses harsh, somewhat predictable punishments. Errors are severely penalized. The child learns that mistakes are dangerous.

This often produces Machiavellianism – learning to avoid punishment through deception, excuse-making, and blame-shifting. It can also produce severe shame and a preoccupation with not getting caught.

E-C-chaotic (violent/chaotic environment). The home environment is unpredictably violent. The child is exposed to domestic violence, physical abuse, or chaotic aggression. The difference to E-C-punitive is its unpredictability.

This often produces:

• Hypervigilance. Constant scanning for danger.

• Aggression. Learning to strike first.

• Avoidant attachment. Learning that people are dangerous.

This is the pathway to the “hot” psychopathy pattern – the reactive, F2-heavy presentation associated with criminal behavior.

E-C-controlling (over-controlling). The caregiver excessively controls the child’s behavior, choices, and even thoughts. Autonomy is suppressed; the child’s agency is denied.

This often produces extreme autonomy preoccupation (what I call D-autonomic). The child who was never allowed to choose becomes an adult who cannot tolerate any constraint. This may manifest as:

• Rage at perceived imposition. Any expectation feels like control.

• Contractarian social model. Only explicit agreements count; implicit obligations are rejected.

• Preemptive control. Controlling others to ensure one’s own autonomy.

E-C-parentified (parentification). The child is forced into a caregiving role – caring for a parent, for younger siblings, or for the household. Normal childhood is lost to premature responsibility.

This can produce:

• Premature self-reliance. The child learns to depend only on themselves.

• Caretaking identity. The self becomes organized around caring for others (a pathway to echoism).

• Narcissistic features. If the parentification was combined with other narcissistic injuries.

E-C-golden (golden child). The child is treated as special, but conditionally – love depends on performance, appearance, or compliance. The child learns that their worth is contingent on meeting expectations.

This is a classic pathway to pathological narcissism. The child develops a false self to live up to all the expectations, but it’s fragile because it depends on external validation and disavows most of the child’s authentic self.

E-C-scapegoat (scapegoated). The child is blamed for family problems, devalued, and shamed. They’re the “bad one.”

This can produce pathological narcissism as the child tries to defend against core feelings of worthlessness. Or it may echoism if the child interprets the core shame as uselessness and tries to earn worth through selfless service.

E-C-normal (normal). None of the pathological properties appear to apply.

E-P: Puberty/Adolescence (12–18 Years)

The Critical Tasks: Identity, Peers, Autonomy

Adolescence is when identity consolidates, peer relationships become central, and autonomy from family develops. Disruptions during this period affect adult identity and social functioning.

Key Features

E-P-peer-success (Peer Success). Had friends, was socially accepted, experienced belonging.

This is stabilizing for grandiose structures. The narcissistic or sovereign adolescent who is popular has their grandiosity reinforced socially. This may produce a more stable presentation in adulthood.

E-P-peer-failure (Peer Rejection). Bullied, rejected, socially isolated.

This can produce shame-based presentations – pathological narcissism with a predominantly vulnerable expression, social anxiety, or withdrawal. For those with psychopathic substrate, it may intensify hostility toward the social world that rejected them.

E-P-antisocial (Antisocial Trajectory). Involved with delinquent peers, criminal activity, or gang involvement.

This is the pathway to F2-heavy adult presentations. Antisocial behavior is normalized; criminal skills are developed; prosocial paths close off.

E-P-identity-formed vs. E-P-identity-diffuse. Did identity consolidate or remain diffuse?

Those who consolidated an identity often struggle with its rigidity whereas those who don’t rather struggle to maintain any continuity in life. Those with diffuse identity may show D-anatta features – minimal coherent self, chameleon-like adaptation.

E-A: Adult Entry (18–25 Years)

The Critical Task: Establishing Adult Life

How did the person enter adult life? This period often crystallizes patterns that persist.

E-A-success (conventional success). Completed education, stable employment, relationships.

This provides structure and reinforcement for prosocial functioning. Even someone with significant psychopathic loading may function well if they entered adulthood successfully.

E-A-privilege (privilege). Inherited wealth, status, or connections.

This provides external structure regardless of internal development. The person may be highly dysfunctional internally while appearing successful externally.

E-A-crime (criminal pathway). Incarceration, criminal career.

This crystallizes antisocial patterns. Prison socializes further into criminal culture; criminal record closes off legitimate paths; the identity becomes “criminal.”

E-A-addiction (addiction pathway). Substance use as primary adaptation.

Addiction complicates everything. It may mask underlying presentations, worsen symptoms, or become the central organizing problem.

E-A-unstable (unstable/chaotic). Failed to establish stable adult life – unstable housing, employment, relationships.

This indicates ongoing dysfunction. Without stability, development of more mature functioning is difficult.

Developmental Narratives: Putting It Together

Let me illustrate how G, N, and E combine to produce different outcomes.

Narrative 1: The Lucky Primary

G-callous → N-hypoactive → E-I-secure → E-C-normal → E-P-success → E-A-success

This person was born with reduced empathy and fear (G-callous), which was expressed neurologically as constitutional amygdala hypoactivity (N-hypoactive). But they were raised by attuned, secure parents (E-I-secure) who provided structure and guidance (E-C-normal). They were socially successful in adolescence (E-P-success) and entered adulthood through conventional paths (E-A-success).

Outcome. Functional psychopathy. They have the substrate but not the pathology. They may be unusually calm, fearless, and emotionally detached, but they’ve developed values, regulation, and prosocial behavior. They probably don’t identify as having any disorder.

Narrative 2: The Street Primary

G-callous + G-impulsive → N-hypoactive → E-I-disorganized → E-C-chaotic → E-P-antisocial → E-A-crime

Same constitutional loading (G-callous), but combined with impulsivity (G-impulsive) and expressed in a chaotic, violent environment (E-I-disorganized, E-C-chaotic). They fell into antisocial peer groups (E-P-antisocial) and entered adulthood through criminal pathways (E-A-crime).

Outcome. Criminal psychopathy. F2-heavy, violent, incarcerated. The substrate was channeled into antisocial expression by the environment.

Narrative 3: The Primary Sovereign

G-callous → N-mixed (hypoactive + dissociative) → E-I-disorganized → E-C-controlling/E-C-parentified/E-C-neglect → E-P-peer-success → E-A-success

Constitutional callous loading, but the key factor is the type of adversity: Controlling, parentifying, rather than chaotic. The person developed sovereignty – a power-and-control orientation that combined grandiosity with callousness.

Outcome. Sovereignism. High-functioning but relationally damaging. Control-oriented, sadistic, autonomy-focused. Functions well by conventional measures but leaves destruction in personal relationships.

Narrative 4: The Secondary Sovereign

G-minimal → N-dissociative (was hyperactive) → E-I-disorganized → E-C-chaotic + E-C-controlling → E-P-peer-failure → E-A-variable

No particular genetic loading, but severe mixed adversity (both violent and controlling). Originally reactive (N-hyperactive), but chronic trauma led to dissociative dampening (N-dissociative). The person developed sovereign features defensively, but remembers being different as a child.

Outcome. Potentially reversible sovereignty. The key marker is memory of being different. This person is likely able to recover some emotional capacity with safety, attachment, and time.

The Environment Selects, Not Creates

An important framing: The environment doesn’t create psychopathy from nothing. Rather, the environment selects which genetic potentials get expressed.

Someone with G-callous loading has the potential for both functional and dysfunctional outcomes. A supportive environment selects for the functional expression; an adverse environment selects for the dysfunctional expression.

This is gene-environment interaction: The effect of genes depends on the environment; the effect of the environment depends on genes. Neither alone determines the outcome.

Tiffany in particular predicts that early neglect tends to lead to “empty” presentations – children who never form much of an identity and scour their social environment for surrogates of parental love through admiration or exploitation – whereas more actively abusive childhoods more likely produce children with harsh introjects – children who take on draconian expectations from their caregivers. Here it is important to distinguish between those who develop actual pathological narcissism vs. those who merely make use of grandiosity or “trait narcissism” because it’s useful in their environment but without any pathological narcissism.

Implications for Understanding

For self-understanding. Knowing your developmental history helps you understand why you developed as you did – not to excuse, but to explain. Understanding the pathway doesn’t change the endpoint, but it may change how you relate to it.

For clinicians. The type of adversity suggests the type of adaptation. Chaotic adversity → behavioral defense. Control adversity → self-structural defense. This has implications for treatment focus.

For prediction. E-level factors help predict outcomes. E-I-secure is protective; E-I-disorganized is a risk factor. E-A-success is stabilizing; E-A-crime is crystallizing.

Summary

E-I (infancy). Attachment formation. Secure, avoidant, preoccupied, or disorganized. Disorganized attachment is particularly associated with later psychopathy and sovereignty.

E-C (childhood). Moral development, socialization, self-concept. Key adversity types: Unattuned, neglect, punitive, violent, controlling, parentified, golden, scapegoat. The type of adversity predicts the type of adaptation.

E-P (puberty). Identity, peers, autonomy. Peer success vs. failure; antisocial trajectory; identity consolidation vs. diffusion.

E-A (adult entry). Pathway into adult life. Success, privilege, crime, addiction, instability.

Next: The Self

The next article explores the D level – how the biological substrate and environmental shaping produce particular psychological structures. What are the different ways the psychopathic self can be organized? What is sovereignty, and how does it differ from standard narcissism?

This is the second article in a series on understanding psychopathy. The first article introduced the multi-level framework (G-N-E-D-B-A). This article covers the G (genetic) and N (neurological) levels – what you were born with and what your brain looks like now.

Introduction

When people ask “Is psychopathy genetic?” or “Is it in the brain?”, the answer is yes – but that’s not the whole story. Genetics and neurology form the substrate on which psychopathic presentations develop, but the substrate alone doesn’t determine the outcome. The same genetic loading can produce a successful surgeon or a convicted felon, depending on environment and development.

This article covers what we know about the biological foundations of psychopathy, with an emphasis on a distinction that matters enormously for understanding prognosis and intervention: primary vs. secondary presentations.

G-Level: Genetic Contributions

What We Know

Psychopathy has a heritable component. Twin studies suggest that callous-unemotional traits (a core feature of psychopathy) are moderately to highly heritable, with estimates ranging from 40% to 70% depending on the study and population. (See Viding et al., 2005, for an influential twin study.)

However, “heritable” doesn’t mean “genetic” in a simple sense. Heritability estimates include gene-environment correlations and gene-environment interactions. A child with genetic loading for psychopathy may also be more likely to experience certain environments (because their parents, who share their genes, create those environments) – and the genes may only express under certain environmental conditions.

Candidate Genes

Several genes have been associated with psychopathic traits, though findings are often inconsistent and effect sizes are small. (See Gunter et al., 2010, Viding & McCrory, 2012, De Brito et al., 2021, and Frazier et al., 2019.)

MAOA (Monoamine Oxidase A). The “warrior gene” – low-activity variants are associated with reduced breakdown of monoamines (serotonin, dopamine, norepinephrine). The famous Caspi et al., 2002, study found that low-activity MAOA combined with childhood maltreatment predicted antisocial behavior. However, replications have been mixed, and the original finding may be smaller than initially thought.

5-HTTLPR (Serotonin Transporter). The short allele is associated with altered serotonin reuptake and has been linked to increased stress sensitivity. Paradoxically, chronic stress exposure in short-allele carriers can lead to emotional blunting over time – suggesting a pathway from reactivity to apparent hypoactivity.

OXTR (Oxytocin Receptor). Variants in the oxytocin receptor gene have been associated with reduced social bonding and empathy. Oxytocin is sometimes called the “bonding hormone,” and reduced signaling may contribute to the social detachment seen in psychopathy.

COMT (Catechol-O-Methyltransferase). The Val158Met polymorphism affects dopamine levels in the prefrontal cortex and has been linked to executive function, reward sensitivity, and impulsivity.

The Polygenic Reality

Most psychopathy is probably polygenic – the result of many genetic variants, each with small effects, combining with environmental factors. Recent genome-wide association studies (GWAS) have begun to identify polygenic risk scores for antisocial behavior, but these explain only a small fraction of the variance. (See Tielbeek et al., 2017, for a meta-analysis.)

The practical implication: There’s no “psychopathy gene.” Genetic testing won’t tell you if someone is a psychopath. What genetics gives us is a probabilistic loading – increased risk, not determination.

Parsimonious G-Level Categories

For the purposes of this framework, I propose three parsimonious G-level categories:

These can co-occur. Someone might have G-callous + G-impulsive, or G-reactive alone, or any combination.

N-Level: Neurological Patterns

The Classic Finding: Amygdala Hypoactivity

The most widely known neuroimaging finding in psychopathy is reduced activity and/or volume in the amygdala – the brain region involved in fear processing, emotional learning, and threat detection. (See Blair, 2013, Blair, 2014, Fallon, 2013, Tyler et al., 2019, and Frazier et al., 2019.)

Individuals with amygdala hypoactivity show:

• Reduced fear conditioning. They don’t learn to associate cues with aversive outcomes as readily.

• Poor threat recognition. They’re less responsive to fearful facial expressions.

• Blunted emotional response. Emotional stimuli produce less physiological and subjective response.

This pattern fits the “cold” presentation of psychopathy – the person who can lie without anxiety, harm without distress, and remain calm in situations that would terrify others.

A systematic review basically bore out the finding that reduced amygdala function or size is associated with the “cold” presentation of psychopathy (the negative correlations in the second row above). The studies are confounded by mixing together different disorders under the label “psychopathy” and probably a host of other factors. For example, many fMRI studies are flawed, but structural MRI studies (finding smaller amygdalas) and quantitative fMRI studies (finding amygdala hypoactivity) are unaffected.

Beyond the Amygdala

Other brain regions are also implicated in psychopathy. (See Blair, 2013, Blair, 2014, Tyler et al., 2019, and Frazier et al., 2019.)

Insula. The insular cortex is involved in interoception (awareness of bodily states) and empathic resonance (feeling what others feel). Reduced insula activity may contribute to the difficulty psychopathic individuals have in “feeling into” others’ experiences.

Ventromedial Prefrontal Cortex (vmPFC). This region integrates emotional information into decision-making. Reduced vmPFC activity may explain why psychopathic individuals can know something is wrong without feeling that it’s wrong – the emotional signal doesn’t reach the decision-making process.

Anterior Cingulate Cortex (ACC). Involved in error monitoring and conflict detection. Reduced ACC activity may contribute to the lack of guilt or remorse – the error signal that should follow harmful actions doesn’t fire.

Connectivity. Perhaps more important than any single region is the connectivity between regions. Reduced connectivity between the amygdala and prefrontal cortex means that emotional signals don’t adequately influence cognitive processes. The person can think and feel, but the thinking and feeling don’t talk to each other.

The “Hot” Pattern: Hyperactivity

Not all psychopathy involves hypoactivity. Some individuals show the opposite pattern – hyperactive amygdala, hypothalamus, and periaqueductal gray (PAG, involved in defensive behaviors). These are the positive (red) correlations in the graph above.

This “hot” pattern is associated with:

• Reactive aggression. Hair-trigger threat response; rage when provoked.

• Hypervigilance. Constant scanning for danger.

• Emotional intensity. Strong, rapid emotional reactions.

These individuals may still be called “psychopathic” because of their behavioral presentation (aggression, callousness in the moment), but the underlying mechanism is different. They’re not cold – they’re too hot, with reactive circuits that fire too readily.

The Critical Distinction: Primary vs. Secondary

This brings us to an important distinction in understanding psychopathy: primary vs. secondary presentations.

Primary presentations reflect constitutional differences – the person was born with, or developed very early, a brain that processes emotions differently. They don’t remember being different. One of my friends fits this pattern: She has always had reduced empathy, high pain tolerance, and sadistic interests. Her father is similar. There’s no trauma, no “before” – this is just who she is.

Secondary presentations developed later, often as a defensive response to trauma. The person may have started with normal or even hyperactive emotional responses, but chronic trauma led to dissociative dampening. Their brain now looks hypoactive, but this is an adaptation, not a constitutional feature. Tiffany fits this pattern: She remembers being normal as a child, then adapted to severe adversity by developing sovereign traits. On good days, she can access empathy again. The original capacity wasn’t destroyed – it was suppressed.

Why This Distinction Matters

The primary/secondary distinction has enormous implications:

For prognosis. Primary presentations are likely to be stable. The person may learn to regulate behavior, develop values, develop selfhood, find prosocial outlets – but they’re unlikely to develop empathy they never had. Secondary presentations may be more changeable. If the dissociative dampening can be reversed, the original emotional capacity may re-emerge.

For intervention. Primary presentations can pursue values-based approaches, training prosocial habits, and finding harmless outlets for drives. Secondary presentations may benefit from moving to a safer environment, trauma processing, attachment repair, and interventions that help them feel safe enough to relax the dissociative defense.

For self-understanding. If you’re primary, trying to “develop empathy” may be frustrating and pointless; you’re better off doubling down on your strengths and forming a stable self. If you’re secondary, knowing that you weren’t always this way – and might not always be this way – can be important.

How to Assess: The Key Question

The simplest way to assess primary versus secondary is to ask:

“Were you always like this, or do you remember being different?”

If the person says they’ve always been this way – no memory of normal emotional responses, no “before” – that suggests primary. If they remember being different as a child – more emotional, more empathic, more afraid – that suggests secondary.

This isn’t a perfect test. Memory is fallible; some people may have been different without remembering it; some may confabulate memories of difference; some were vastly different but not in a more emotional/empathetic/fearful direction. But it’s a useful starting point.

Parsimonious N-Level Categories

Case Studies

Some of the profile terms will be clarified in later posts in the sequence.

Case A: The Lucky Primary

Profile. G-callous, N-hypoactive, E-I-secure (secure attachment), E-C-normal (normal childhood).

This person has constitutional psychopathic loading but developed in a supportive environment. They have always had reduced empathy and fear, high pain tolerance, and possibly sadistic interests. But they also have stable relationships, regulated behavior, and prosocial values.

They may not identify as having any disorder. They experience their differences as just “who I am” – perhaps useful in certain contexts (surgery, emergency response, high-stakes negotiation), perhaps something to manage in others (intimate relationships, parenting).

Implication. This person doesn’t need to “recover” from anything. They’re already functional. What they may benefit from is understanding – recognizing that their experience differs from others’, and developing strategies for contexts where their differences create friction.

Case B: The Primary Sovereign

Profile. G-minimal, N-hypoactive (secondary), E-I-avoidant (avoidant attachment), E-C-punitive (punitive parenting).

This person developed psychopathic features defensively. They don’t remember being different as a child because the neglect trauma happened too early. Perhaps their amygdala had already atrophied significantly by age 3 or memories before age 5 or so are missing because they changed too much. They now present as cold, calculating, and sadistic.

But the original capacity exists to some extent. On good days, they may experience flashes of empathy or guilt. They are not complete strangers to fear. They may have stable attachments to a few people (children, specific partners) even while being callous toward others.

Implication. This person may be able to recover – not necessarily to become “normal,” but to regain some access to emotional capacities that were suppressed. This requires safety, stable attachment, and possibly trauma processing. It’s a multi-year-long journey and a different one from primary presentations.

Case C: The Secondary Sovereign

Profile. G-minimal, N-hyperactive (secondary), E-I-disorganized (disorganized attachment), E-C-punitive (punitive parenting), D-max-avoidant (pervasive avoidant attachment).

This person also developed psychopathic features defensively. They remember being different as a child – more emotional, more afraid, more connected. Severe trauma led them to develop dissociative defenses, and their emotional responses became blunted, perhaps from one day to the next as they gave up on parents, friends, society, and humanity. They now present as cold, calculating, sadistic, thrill-seeking.

But the original capacity exists. On good days, they may experience flashes of empathy or guilt. They may have stable attachments to a few people (children, specific partners) even while being callous toward others.

Implication. This person may be able to recover, even suddenly – not to become “normal,” but more likely to enter into a state of borderline personality disorder due to their disorganized attachment, for which there are many treatments like Dialectical Behavior Therapy (DBT) and Mentalization-Based Treatment (MBT).

Case D: The Reactive Chaotic

Profile. G-reactive + G-impulsive, N-hyperactive, E-I-disorganized (disorganized attachment), E-C-violent (chaotic childhood), D-autonomic.

This person has high emotional reactivity that never got regulated. They experience intense emotions – especially anger – and act on them impulsively, especially in response to perceived infringements on their autonomy. Their violence is “hot” (reactive rage) rather than “cold” (instrumental calculation).

They may be called “psychopathic” because of their behavior, but the underlying mechanism is different from classic psychopathy. They’re not underreactive; they’re overreactive. Their problem is dysregulation, not absence of feeling. They might even experience guilt once enough time has passed since the triggering event.

Implication. This person needs stabilization and regulation skills before anything else. Approaches like Dialectical Behavior Therapy (DBT) may be helpful. The goal is not to develop emotions (they have plenty) but to regulate them.

The Substrate Is Not the Person

A crucial point: The G and N levels describe the substrate – the biological foundation on which the person developed. But the substrate doesn’t determine the person.

The same N-hypoactive pattern can produce:

• A functional surgeon who saves lives

• A successful CEO who builds companies

• A serial killer who destroys lives

What makes the difference is everything else: The environment (E), the psychological structures that developed (D), the behaviors that became habitual (B), and how the person understands themselves (A).

This is why we need a multi-level framework. Knowing someone’s substrate is useful, but it’s not sufficient. To understand them fully, we need to know the whole profile.

Summary

G-level (genetic). Psychopathy has a heritable component, but it’s polygenic and probabilistic. Key loadings include G-callous (reduced empathy/fear), G-reactive (emotional reactivity), and G-impulsive (disinhibition).

N-level (neurological). The classic finding is amygdala hypoactivity, but there’s also a hyperactive pattern (reactive aggression) and a dissociative pattern (secondary blunting). The primary/secondary distinction – “always this way” versus “remember being different” – has major implications for prognosis and intervention.

The substrate is not destiny. Biology creates predispositions, not determinations. The same substrate can produce very different outcomes depending on environment and development.

Next: The Shaping

The next article explores the E level – how environment and development shape the expression of the biological substrate. Why does the same genetic loading produce a functional person in one context and a criminal in another? What types of adversity lead to what outcomes?

This is the first article in a series on understanding psychopathy and related presentations. The series is written for three audiences: people with psychopathic or narcissistic traits who want to understand themselves better, clinicians and researchers who want a more integrated framework, and curious laypeople who want to move beyond stereotypes.

Introduction

I’ve spent about a year trying to understand psychopathy – not just from textbooks, but from friendships. Some of my friends score high on the Psychopathy Checklist Revised (PCL-R), and through countless conversations, I’ve come to see that “psychopathy” is not one thing but many things hiding under a single label.

This creates problems. When researchers study “psychopathy,” they may be studying completely different populations. When clinicians treat “psychopathy,” they may be applying the same approach to people who need very different interventions. And when people try to understand themselves, they may find that the label fits in some ways but not others – leaving them more confused than before.

This series proposes a new framework: A multi-level taxonomy that distinguishes what you were born with, what your brain looks like now, what happened to you developmentally, what psychological structures you developed, how you behave, and how you understand your own agency. These are different lenses on the same phenomenon – and using all of them gives us a much richer picture than any single lens alone.

The Naming Problem

The word “psychopathy” is used to describe at least four different things:

1. Genetic loading. Researchers talk about genes associated with psychopathy – variants of MAOA, the serotonin transporter gene (5-HTTLPR), oxytocin receptor genes (OXTR), and others. When they say someone is “genetically psychopathic,” they mean the person carries variants that increase risk for psychopathic traits. (For reviews of candidate genes, see Gunter et al., 2010, Viding & McCrory, 2012, De Brito et al., 2021, and Frazier et al., 2019.)

2. Brain patterns. Neuroscientists describe psychopathy in terms of brain structure and function – a smaller or less reactive amygdala, reduced activity in the insula, altered connectivity between prefrontal and limbic regions. When they say someone is “neurologically psychopathic,” they mean the person’s brain shows these patterns. (See Fallon, 2013, Tyler et al., 2019, Frazier et al., 2019, for comprehensive reviews.)

3. Psychological structure. Psychodynamic clinicians describe psychopathy in terms of self-structure and relational patterns – an absent or fragmented sense of self, instrumental orientation toward others, absence of guilt or remorse, preoccupation with power or control. When they say someone is “psychodynamically psychopathic,” they mean the person has these internal structures. (See McWilliams, 2011, for a summary chapter and literature recommendations.)

4. Behavior. The PCL-R and similar instruments measure psychopathy through observable behaviors and self-reported traits – manipulation, callousness, impulsivity, criminal versatility. When they say someone is “behaviorally psychopathic,” they mean the person shows these patterns. (The PCL-R is described in Hare, 2008.)

Here’s the problem: These four things don’t always go together.

Someone can have genetic loading for psychopathy but, raised in a supportive environment, never develop the brain patterns or behavioral expression. Someone can develop psychopathic brain patterns through chronic trauma and dissociation, without any particular genetic predisposition. Someone can score high on the PCL-R through learned behavior, while having relatively intact empathic capacity that they’ve learned to suppress. And someone can have all the neurological and psychological features of psychopathy while never engaging in criminal behavior.

Using one word for all of these creates confusion. It leads researchers to treat heterogeneous groups as homogeneous. It leads clinicians to apply one-size-fits-all treatments. And it leads individuals to misunderstand themselves – either over-identifying with a label that only partially fits, or rejecting a label that captures something important about their experience.

The Heterogeneity Problem

Even within each level of description, there are important subtypes.

At the Genetic Level

Different genetic variants may produce different phenotypes:

• Some variants (like low-activity MAOA) are associated with reduced emotional reactivity and may contribute to “cold” presentations.

• Other variants (like short-allele 5-HTTLPR) are associated with increased stress sensitivity and may contribute to “reactive” presentations – though chronic stress can paradoxically lead to blunting over time.

• Most psychopathy is probably polygenic – the result of many small-effect variants combining with environmental factors.

At the Neurological Level

The classic finding is hypoactivity in the amygdala and related structures – reduced fear conditioning, poor threat recognition, blunted emotional response. But there’s also:

• Hyperactive patterns. Some individuals show increased amygdala and hypothalamic reactivity – hair-trigger threat response, reactive aggression. These “hot” presentations look very different from “cold” ones, even though both may be called psychopathic.

• Dissociation-induced patterns. Some individuals developed normal or even hyperactive emotional responses early in life, but chronic trauma led to dissociative dampening. Their brains now look hypoactive, but the hypoactivity is secondary – and potentially reversible.

• Mixed patterns. Some individuals show hypoactivity in some regions (e.g., amygdala) and hyperactivity in others (e.g., periaqueductal gray). These mixed presentations don’t fit neatly into simple categories.

At the Psychodynamic Level

The internal world of psychopathy varies enormously:

• Some individuals have what might be called an “empty” self – minimal coherent identity, chameleon-like adaptation to context, observational relationship to their own behavior. (See my article on no-self psychopathy and M.E. Thomas’s memoir.)

• Others have a “grandiose” self – inflated self-concept, need for domination, but with psychopathic features stabilizing the grandiosity against vulnerable collapse. I’ve previously called this combination sovereignism – a power-and-control orientation that differs from standard narcissism’s admiration-seeking.

• Still others have developed extreme avoidant patterns – walled-off, nothing matters, independence as the only value.

• And some are primarily organized around autonomy – hypersensitivity to any perceived constraint, rage or withdrawal when obligations are imposed.

At the Behavioral Level

The PCL-R distinguishes two factors:

• Factor 1 (Interpersonal/Affective). Grandiosity, manipulation, callousness, shallow affect. The “cold” features.

• Factor 2 (Lifestyle/Antisocial). Impulsivity, irresponsibility, early behavior problems, criminal versatility. The “hot” or chaotic features.

Some individuals are high on Factor 1 but low on Factor 2 – they’re manipulative and callous but not impulsive or reckless. Others show the reverse pattern. Still others are high on both. These are different presentations with different trajectories and different needs.

One of the most useful models in my opinion is the Triarchic Model of Psychopathy (Patrick et al., 2009), which distinguishes three distinct dimensions:

1. Boldness. Emotional resilience, social dominance, and low fear. (Roughly maps to N-hypoactive, factor 1 features.)

2. Meanness. Aggressive resource-seeking and lack of empathy. (Roughly maps to G-callous, D-autonomic-asymmetric features.)

3. Disinhibition. Impulsivity and lack of restraint. (Roughly maps to G-impulsive, reactive, factor 2 features.)

This model is excellent because it acknowledges that “boldness” (fearless dominance) is distinct from “disinhibition” (impulsivity). A person can be bold without being disinhibited (the “successful psychopath”), or disinhibited without being bold (the “reactive” type). My framework builds on this by adding the developmental (E) and psychodynamic (D) layers that explain how these traits emerge and organize into a self.

My score of 29 puts me in the 8th percentile among women (23 on boldness, 4 on meanness, 2 on disinhibition). The overall distribution in the general population varies greatly by gender.

Another study further subdivided these three factors for better predictive results in their sample (highlights mine):

From the majority of Boldness, Meanness, and Disinhibition scale items, respectively, emerged three factors reflecting: Positive Self-image, Leadership, and Stress Immunity; two factors tapping Callousness and Enjoy Hurting; and two factors involving trait Impulsivity and overt Antisociality. The emergent factors from the Boldness items were differentially intercorrelated with the other emergent factors, raising questions about the structural coherence of Boldness. Further, the Enjoy Hurting and overt Antisociality factors were more strongly correlated with one another than with the other scales from their home domains (Callousness and Impulsivity). All seven emergent factors were differentially associated with the external correlates, suggesting that the three original TriPM factors are not optimal for representing psychopathic propensity.

A third model is that of the Psychopathic Personality Inventory (a mock version that gives good results). It distinguishes 8 factors:

1. Machiavellian egocentricity. A ruthless and self-centered willingness to exploit others.

2. Social potency. The ability to charm and influence others.

3. Coldheartedness. A distinct lack of emotion, guilt, or regard for others’ feelings.

4. Carefree nonplanfulness. Difficulty in planning ahead and considering the consequences of one’s actions.

5. Fearlessness. An eagerness for risk-seeking behaviors, as well as a lack of the fear that normally goes with them.

6. Blame externalization. Inability to take responsibility for one’s actions, instead blaming others or rationalizing one’s behavior.

7. Impulsive nonconformity. A disregard for social norms and culturally acceptable behaviors.

8. Stress immunity. A lack of typical marked reactions to traumatic or otherwise stress-inducing events.

They are (with the exception of coldheartedness) sometimes grouped into “fearless dominance” and “self-centered impulsivity.”

The Framework: G-N-E-D-B-A

To address these problems, I propose a multi-level framework with six dimensions:

Each individual can be described as a profile across all six dimensions. Two people who both “have psychopathy” might have completely different profiles – and understanding those differences matters for prediction, treatment, and self-understanding.

The Ordering Is (Roughly) Causal

The dimensions are ordered from most distal to most proximal:

• G (genetic) sets the constitutional foundation – what you were born with.

• N (neurological) reflects the current brain state, shaped by G but also by experience.

• E (environmental) captures the developmental context – what happened to you during critical periods.

• D (dynamic) describes the psychological structures that developed from G, N, and E interacting.

• B (behavioral) is the observable expression of D in action.

• A (agentic) is how you understand and narrate your own behavior.

• C (connective) describes typical interruptions in interpersonal communication, verbal or nonverbal.

This ordering helps us see that the same behavioral presentation (B) can arise from different developmental pathways (E) and different psychological structures (D), including different communicative failures (C), which in turn can arise from different neurological patterns (N) and genetic loadings (G). And all of this is filtered through how the person understands themselves (A).

Profiles, Not Labels

Instead of asking “Is this person a psychopath?” we can ask: “What is this person’s profile?”

For example:

Profile A. G-callous (genetic loading for reduced empathy), N-hypoactive (constitutional amygdala hypoactivity), E-I-secure (secure attachment) + E-C-normal (good enough developmental environment), D-secure, B-subclinical (no behavioral problems). This person has constitutional psychopathy but developed well – they’re functional, may not even identify as having any disorder because they function at the neurotic (healthy) level of personality organization.

Profile B. G-minimal (no particular genetic loading), N-dissociative (secondary blunting from chronic dissociation), E-I-disorganized + E-C-violent + E-C-controlling (severe mixed adversity), D-sovereign (power-oriented, ego-syntonic sadism), B-mixed (high on both PCL-R factors). This person developed psychopathic features defensively – and the key question is whether those features are reversible.

These two individuals might be called “psychopathic” by different psychologists but they are fundamentally different. Understanding the difference matters for predicting their trajectories and for any treatments they may or may not attempt.

What This Series Will Cover

This series will develop the framework in detail:

Article 2: The Substrate. Genetics and neuroscience – what we know about the biological foundations of psychopathy, and how to think about primary versus secondary presentations.

Article 3: The Shaping. Environment and development – how different types of adversity lead to different outcomes, and why the same genetic loading can produce a functional person or a criminal depending on context.

Article 4: The Self. Psychodynamic structures – the different ways the psychopathic self can be organized, including the autonomy dimension and the relationship between psychopathy and narcissism (what I’ve called sovereignism).

Article 5: The Mechanics. How empathy fails – a detailed breakdown of the different ways empathy can break down, from perceptual failures to simulation failures to affective inversion. (This connects to my earlier work on the sadism spectrum.)

Article 6: The Types. Archetypal clusters – common profiles that tend to co-occur, with recognizable presentations that readers may identify with.

Article 7: The Choice. Recovery, if you want it – an honest assessment of what recovery means for different presentations and the trade-offs involved.

A Note on Tone

This series is written for three audiences, and the tone reflects that.

For people with psychopathic or narcissistic traits. I’m not here to moralize or to tell you you’re broken. I think that factory farming is an abomination akin to slavery but at a massively greater scale and that the cuts to USAID are crimes against humanity worse than many wars. Most people are oblivious to that or contribute to it. The median serial killer vanishes in the statistical noise among the horrors of this world. Many of my psychopathic friends are actually doing better than average despite their traits, or perhaps precisely thanks to what they had to learn to make these traits work for them. If they donate $100 to an ACE top charity, they’re suddenly among the crème de la crème of the least harmful humans alive.

So if that fucked-up ne’er-do-well that’s the median person deserves my honesty and respect, so do you. I’ll describe things as they are, including trade-offs that others might not acknowledge. Empathy is fun but also painful and often detrimental to moral decision-making. Regulation is stabilizing but also boring. Attachment creates meaning but also vulnerability. These are real trade-offs, and I’ll respect your intelligence enough to present them honestly.

For clinicians and researchers. I’ve tried to integrate findings from neuroscience, genetics, attachment theory, and psychodynamic thinking into a coherent framework. I’ll provide references throughout and a technical glossary. The clusters I propose are hypotheses, not established facts – but they may help organize thinking about a heterogeneous population.

For curious laypeople. I’ll explain technical concepts as I go and provide examples (real and fictional) to make abstract ideas concrete. You don’t need a psychology background to follow this series.

One thing I won’t do is pretend that all presentations are equally concerning or that change is always desirable. Some people with psychopathic traits live excellent lives and contribute enormously to society. Others feel isolated, desolate, hopeless, or cause significant harm. The framework I’m proposing is descriptive, not prescriptive – it’s about understanding, not judging.

Glossary: Key Terms for This Series

This glossary introduces terms used throughout the series. Full definitions are provided in the relevant articles.

Dimensional Levels

• G (genetic). Constitutional loading from genetic variants. Examples: G-callous, G-reactive, G-impulsive.

• N (neurological). Current brain state, including structure and function. Examples: N-hypoactive, N-hyperactive, N-dissociative.

• E (environmental). Developmental context across life stages. Subdivided into E-I (infancy), E-C (childhood), E-P (puberty), E-A (adult entry).

• D (dynamic). Psychodynamic self-structure and relational patterns. Examples: D-anatta, D-sovereign, D-autonomic.

• B (behavioral). Observable presentation. Examples: B-factor-1, B-factor-2, B-subclinical.

• A (agency). How the person understands their own intentionality. Examples: A-observational, A-strategic, A-sovereign.

• C (connective). What typical perception and mentalization failures interfere in the interpersonal functioning of the person. Examples: C-P-aversion, C-S-no-self, C-A-suppressed.

Key G-Level Concepts

• G-callous. Genetic loading for reduced empathy, reduced fear, and blunted emotional reactivity.

• G-reactive. Genetic loading for emotional reactivity, stress sensitivity, and dysregulation.

• G-impulsive. Genetic loading for impulsivity, sensation-seeking, and disinhibition.

• G-minimal. Minimal genetic loading for psychopathic traits.

Key N-Level Concepts

• N-hypoactive. Reduced activity in amygdala, insula, and related structures; constitutional low reactivity. The “cold” pattern.

• N-hyperactive. Increased reactivity in amygdala, hypothalamus, and periaqueductal gray; reactive aggression. The “hot” pattern.

• N-dissociative. Secondary blunting from chronic dissociation; originally reactive, now dampened. Key marker: The person remembers being different.

• N-disconnected. Reduced connectivity between prefrontal and limbic regions.

Key E-Level Concepts

• E-I (Infancy, 0–2 years). Attachment formation. Key variants: E-I-secure, E-I-avoidant, E-I-preoccupied, E-I-disorganized.

• E-C (Childhood, 2–12 years). Moral development, socialization, self-concept. Key variants: E-C-unattuned, E-C-neglect, E-C-punitive, E-C-violent, E-C-controlling, E-C-parentified, E-C-golden, E-C-scapegoat.

• E-P (Puberty, 12–18 years). Identity, peers, autonomy. Key variants: E-P-peer-success, E-P-peer-failure, E-P-antisocial.

• E-A (Adult entry, 18–25 years). Pathway into adult life. Key variants: E-A-success, E-A-privilege, E-A-crime, E-A-addiction, E-A-unstable.

Key D-Level Concepts

• D-anatta. An empty or absent sense of self; minimal coherent identity. (From the Buddhist concept of anattā, “no-self.”)

• D-sovereign. Power-and-control orientation; combines narcissistic grandiosity with psychopathic callousness and sadism. See sovereignism.

• D-autonomic. Hypersensitivity to perceived constraints on freedom; autonomy as a core value. Reactive (not proactive), non-narcissistic, non-sadistic version of D-sovereign.

• D-narcissistic. Unstable self-esteem; real or inverted grandiosity; need for specialness or admiration.

• D-avoidant. Extreme dismissive attachment; walled-off; nothing matters.

• D-echoist. Self-effacing orientation; meeting others’ needs at the expense of one’s own.

• D-secure. A stable self without any psychopathy-associated traits.

Key B-Level Concepts

• B-factor-1. High on PCL-R Factor 1 (interpersonal/affective): grandiosity, manipulation, callousness, shallow affect.

• B-factor-2. High on PCL-R Factor 2 (lifestyle/antisocial): impulsivity, irresponsibility, early behavior problems, criminal versatility.

• B-mixed. High on both factors.

• B-violent. Violence as a prominent behavioral feature.

• B-subclinical. Trait-level features; functional; doesn’t meet clinical thresholds.

• B-normal. Low to average levels of behavioral traits.

Key A-Level Concepts

• A-observational. “I watch myself do things.” Minimal sense of deliberation; actions emerge from observation.

• A-strategic. “I plan, then act.” Genuine prospective intentionality.

• A-narrativizing. Real-time rationalization; maintaining a running story about what you’re doing.

• A-retroactive. Post-hoc rationalization; acting first, explaining later.

• A-selective. True-but-partial explanations; picking one motivation and presenting it as the whole story.

• A-externalizing. Locating causation externally; “They made me do it.”

• A-absorbed. Taking responsibility for others’ actions; “It’s my fault they hurt me.”

• A-amnestic. Forgetting ego-dystonic actions, or not being able to access the memories while they’re ego-dystonic.

Key C-Level Concepts

• C-P-…. Failures of the perception of distress.

• C-S-…. Interruptions of cognitive empathy or mentalization.

• C-A-…. Deficits of affective empathy.

• C-M-…. Interruptions related to the motivation to act.

• C-B-…. Disinhibition on the level of the behavior.

Key Distinctions

• Primary vs. secondary. Primary presentations were “always this way” – constitutional, from early development. Secondary presentations developed later, often as a defensive response to trauma, and may be more changeable.

• Factor 1 vs. factor 2. The two factors of the PCL-R. Factor 1 captures interpersonal and affective features (manipulation, callousness). Factor 2 captures lifestyle and antisocial features (impulsivity, criminality).

• Ego-syntonic vs. ego-dystonic. Ego-syntonic traits feel consistent with one’s self-image and values; ego-dystonic traits feel foreign or distressing. Sadism that feels “right” is ego-syntonic; sadism that causes guilt is ego-dystonic.

Next: The Substrate

The next article explores the biological foundations of psychopathy – what genetics and neuroscience tell us about the origins of these presentations, and how to think about the critical distinction between primary and secondary forms.

Why It Matters

If you’re reading this as someone with narcissistic personality disorder (NPD) working on your recovery, you’ve probably noticed a painful pattern: Moments when someone you care about sets a boundary or asks for accountability, and something in you can’t meet them there. You might deny, deflect, attack, or shut down – even when part of you knows you’ve caused harm.

You’re not broken or incapable of healthy relationships. These are protective patterns – strategies your psyche developed to survive experiences that were genuinely threatening to your sense of self.

But here’s the hard truth: While these patterns made sense when they formed, they’re often the most relationship-damaging aspect of NPD. Partners, friends, and family members can weather a lot, but when someone can’t acknowledge harm or respect boundaries, trust erodes. Relationships that could have survived imperfection often can’t survive the inability to repair.

The good news: These patterns aren’t monolithic. Different people struggle with boundaries for different underlying reasons, and understanding your particular pattern is the first step toward working with it rather than being controlled by it.

This article offers a map – not a diagnosis, but a framework for self-understanding. You may recognize yourself in one pattern or several. The goal isn’t to add another layer of shame, but to make the invisible visible, so you can start to catch these patterns in action and, eventually, choose differently. That is something to be proud of.

The Patterns at a Glance

I’ve identified eight distinct patterns that interfere with boundary respect and accountability. Each falls into one of three styles based on how the defense operates, and each is rooted in a particular underlying wound or fear.

Understanding the Taxonomy

Three Defense Styles

Obscuring defenses work by preventing accurate perception. You don’t see clearly – either yourself, the other person, or what actually happened. These patterns aren’t willful distortion; they operate before conscious awareness. Sometimes you’ll see glimpses of other framings before the defenses shut them down. Sometimes the memories come back when you’re in a different mood. In your recovery, you’ll start to remember them dimly.

Fortifying defenses work by protecting the environment that you need in order to maintain your self-image. These can be unconscious defenses too. You try to find some reframing of reality that protects you from devastating feelings of dependency or inadequacy. The goal is less to hide the reality from you than it is to maintain the scaffolds you’ve build around you to survive.

Antagonistic defenses work by fighting. Rather than not-seeing or wall-building, you’re opposing and defending against the other person. These patterns often feel more ego-syntonic – you’re aware you’re fighting, and it feels justified or like you don’t have a choice.

Two Temporal Modes

Each pattern can manifest in proactive or reactive form:

• Proactive. The defense is always-on or anticipatory. You’re scanning for threats, prefiltering information, or rigidly positioning yourself in certain ways. This can manifest as a general tendency toward avoidant attachment – claiming that you don’t need the other, and that they should rather leave you alone than ask for accountability, even when deep down you actually value the relationship but can’t admit it. It might also manifest as general avoidance in many areas of life, a constant narrative of your pretend reasons for any behaviors, or a compulsion to only interact with people you can control.

• Reactive. The defense activates in response to a specific trigger – a criticism, a boundary, a confrontation. In this case you’re usually clear-minded, but when your amygdala detects a threat (such as someone asking you for an apology), you switch into a threat response state – fight, flight, freeze, fawn, or flop – like you’re in mortal danger, which makes it hard to think clearly and can interfere with memory formation too.

Most people have a mix, but noticing your dominant mode can help you catch patterns earlier.

The Patterns

Disavowal

What it is. The inability to acknowledge having transgressed. When you’re confronted with signs of having made a mistake, the truth is unbearable. Either you’ve preemptively convinced yourself of a reframing or pretext or altogether different narrative, or your threat system reflexively devises one on the spot. A question of connection or relating with another is recast as a question of right and wrong.

The underlying wound. At the core is deep shame. Trauma that has made you feel like you’re never enough, like you’re always at the edge of society or breakup, one mistake away from humiliation or death. The shame is no mere embarrassment – it’s annihilating, identity-destroying. Disavowal isn’t about avoiding accountability; it’s about avoiding psychological death. Somewhere along the way, you learned that being bad or weak or fallible meant being nothing. Perfectionism to the point of self-delusion seemed like the only way out.

How it feels from the inside. Often, there’s no felt sense of lying. The alternative version of events feels true. Either because it is (through cherry-picking facts or framings) or because you’ve meticulously rewritten your memories. When someone insists you did something hurtful, it may feel like a bizarre accusation that doesn’t match your experience at all. Or there’s a flash of knowing, immediately followed by mental static – the knowledge is there and then it’s not there. Sometimes there’s a dreamlike quality: “Maybe something happened, but it feels distant, unreal, like it was someone else.”

Proactive form. A rigid “I’m a good person” identity (or insert a more precise attribute for “good”) that prefilters all contradictory information. You may have a strong internal narrative about your values and intentions that automatically reframes questionable behavior. Shame-inducing information doesn’t land because the identity structure rejects it before it registers.

Reactive form. When directly confronted, your regular consciousness turns off and you launch into an attack-type threat response, as if a tiger jumped out of the bushes and you’re fighting for your life. Some Lovecraftian god of chaos throws massive numbers of excuses and counterattacks your way for you to pick from. You find yourself arguing about the facts of what occurred rather than empathizing with the impact on the other person.

Toward recovery. The path through disavowal begins with building tolerance for shame – perhaps not eliminating it at first, but learning that it won’t destroy you. Self-compassion practices are essential: Can you hold “I did something hurtful” alongside “I’m still worthy of love”? Recording yourself during difficult conversations (with consent) or keeping a detailed journal can help you catch the gap between what happened and what you remember. In therapy, EMDR can process the original traumas that made shame feel annihilating. Mindfulness meditation trains you to notice that flash of knowing before the static descends – that’s your window. Shadow work helps you integrate the “bad” parts you’ve been fleeing. The goal isn’t to become someone who never errs, but someone who can say “I hurt you, and I’m sorry” without dying inside. Or even while loving every bit of the messiness that connects us all.

Differentiation Failure

What it is. Difficulty experiencing others as truly separate beings with valid, independent needs. When your partner expresses a need that conflicts with yours, or sets a boundary, it doesn’t compute. Their separateness registers as abandonment, betrayal, or as if they’re ripping out a part of you rather than as a normal feature of relating.

The underlying wound. This pattern often develops when early attachment was either enmeshed (no permission to be separate) or chaotic (separateness meant abandonment). The wound is existential aloneness – a terror that if others are truly separate from you, with their own centers of gravity, you’ll be left alone, incomplete, or vanish. Secure attachment allows children to internalize “I can be separate and still complete and connected.” Without that, you mistake the merging that happens when someone completes you for connection, and actual connection is unsatisfying because it doesn’t protect you from the knowledge of your own incompleteness.

How it feels from the inside. When someone close to you says “no” or expresses a different need, it can feel viscerally wrong – as if they’re stealing from you, harvesting your organs. There’s genuine shock and confusion: “Why would you do this to me?” Other people’s independent actions feel personal to you. The idea that they have their own inner world with its own logic may be intellectually available but doesn’t feel emotionally real.

Proactive form. Chronically experiencing others as extensions of yourself – satellites orbiting your sun. Their function is to meet needs, reflect value, provide support. Their separateness simply doesn’t register; you may be surprised when reminded they have constraints, preferences, or limitations you weren’t aware of.

Reactive form. When someone refuses a favor, sets a boundary, or expresses an independent need, there’s genuine hurt and betrayal: “Why are you doing this to me?” The boundary feels like they’re cutting out a piece of you rather than a neutral expression of their needs.

Toward recovery. Recovery from differentiation failure is fundamentally about building the felt sense that others exist as complete beings even when they’re not there for you, and that connection can survive separateness. Reparenting work in therapy helps internalize the secure attachment you missed. Mentalization-based treatment (MBT) trains you to genuinely wonder “What might they be experiencing right now?” in a flexible way that is open to collaborative exploration and multiple hypotheses. Authentic relating and circling practices offer structured ways to encounter others’ genuine otherness in a safe container. Mindfulness meditation builds capacity to sit with the incompleteness that separateness triggers, discovering it’s survivable. The shift you’re working toward: from “their boundary is abandonment or robbery” to “their boundary is information about them, and we’re still connected.”

Differentiation Failure (Projective)

What it is. A more severe variant where your own intolerable feelings or motives are experienced as coming from the other person. This isn’t strategic blame-shifting; it’s a genuine perceptual distortion where what’s inside you appears to be outside you.

The underlying wound. This includes everything from basic differentiation failure, plus an additional layer: certain feelings or parts of yourself are so intolerable that they can’t be owned. They must be evacuated, expelled, placed outside. This often develops when certain emotions were forbidden or punished in childhood – rage, selfishness, neediness, interdependence. These parts didn’t disappear; they went into exile, and now they can only be experienced as belonging to others. The wound is: “Parts of me are so unacceptable that I can’t know they’re mine.”

How it feels from the inside. You genuinely experience the other person as doing what you’re actually doing. If you’re angry, they seem aggressive. If you’re withdrawing, they seem cold. The strangler who screams “Why are you strangling me?” isn’t lying – in their experience, they are as much victim as they are themselves. At less extreme levels: “Why are you so hostile?” when you’re the hostile one, with complete sincerity.

Proactive form. Chronically attributing your own states and motives to others. You may be known for “reading” people in ways that reveal more about your own psychology than theirs.

Reactive form. In conflict, your own aggression, manipulation, or selfishness is experienced as emanating from the other person. You feel victimized by the very thing you’re doing.

Toward recovery. The projective variant requires reclaiming the exiled parts of yourself. Shadow work is central: What feelings were forbidden in your childhood? Can you begin to own rage, neediness, selfishness as yours rather than seeing them only in others? Journaling after conflicts – “What was I feeling? What was I doing?” – helps build self-knowledge. Mentalization-Based Treatment (MBT) trains you to hold multiple perspectives simultaneously: “I experience them as hostile, and I might be projecting.” Circling and authentic relating provide feedback from others about how you’re actually showing up, which can be reality-testing for your projections. In therapy, the relationship itself becomes a laboratory for catching projection in action. Ayahuasca and other psychedelic-assisted therapies, in appropriate settings, can facilitate encounters with disowned parts. The goal: Welcoming home the exiled feelings so they don’t have to live in others anymore.

Entitlement

What it is. An unconscious assumption that things are owed rather than requested. This isn’t arrogance in the simple sense – it’s a defense against the vulnerability of having needs that others can refuse. By assuming entitlement, you never have to feel the dependency, submission, indignity, or risk the rejection inherent in asking.

The underlying wound. Underneath entitlement is often a deep terror of rejection or dependency and what they imply about one’s relative status. To ask for something is to admit you need it, to reveal vulnerability, and to give the other person power to say no. If you grew up with needs that were ignored, mocked, or weaponized, asking may have become unbearable. Entitlement is a workaround: if it’s owed, you don’t have to ask, you don’t have to be vulnerable, and refusal becomes their moral failure rather than your rejection. The wound is: “My needs make me vulnerable, and vulnerability is dangerous.”

How it feels from the inside. You don’t feel entitled in the entitled sense – it just seems obvious that certain things should happen. When they don’t, it feels like injustice rather than disappointment. There’s a qualitative difference between “I’m sad you didn’t do X” and “How could you not do X?” – the latter feels righteous, wounded, betrayed. Making explicit requests feels strange, unnecessary, or even humiliating.

Proactive form. Assumptions about what others will do for you operate invisibly. You don’t ask because asking would imply it wasn’t already owed. You may be surprised when others describe you as demanding, because from inside, you weren’t demanding – you were expecting normal treatment.

Reactive form. When an expectation is unmet, the response is anger and betrayal, not disappointment. This feels like a response to genuine wrongdoing. The hurt is real; what’s obscured is that you never communicated the expectation, or that the other person had any right to refuse.

Toward recovery. Entitlement dissolves when asking becomes safe. Start small: Practice making explicit requests where the stakes are low. DBT’s interpersonal effectiveness skills (DEAR MAN) provide scripts for asking that feel less vulnerable. Self-compassion work helps you befriend the dependency you’ve been hiding from – needs are human, not shameful. Exposure therapy principles apply: gradually facing small rejections teaches your nervous system that “no” isn’t annihilation. Reparenting yourself means learning to meet some of your own needs, reducing the desperate dependency on others. In therapy, explore the original wounds around asking: when did it become dangerous? Journaling can track the gap between “what I expected” and “what I communicated.” The shift: from silent expectation and righteous rage to clear requests and disappointment you can survive.

Idealization Protection

What it is. Protecting the idealized self-image that generates admiration (and thus self-esteem). Flaws must be hidden, denied, or explained away because being seen as imperfect threatens your source of worth. It’s the fortifying flip-side of disavowal.

The underlying wound. This pattern often develops when love and approval were conditional on performance or appearance. You learned that you were valued for what you achieved, how you looked, how you made others feel – not for who you were. The idealized image became load-bearing: without it, there’s nothing. The wound is a missing floor – no stable sense of worth that exists independent of others’ admiration. Imperfection doesn’t just mean you made a mistake; it means a threat to the very source of your worth. The desperation to maintain the image is proportional to the void beneath it.

How it feels from the inside. There’s a desperate, urgent quality to keeping up appearances – though it may be so automatic you don’t notice it as effort. When a flaw is exposed, there’s panic, followed by rapid narrative management: minimizing, recontextualizing, or dissociating. Perhaps you know the flaw is real, but admitting it feels like it would be catastrophic – far worse than the objective situation warrants.

Proactive form. Meticulous image curation. Avoiding situations where imperfection might show. Controlling what others see. This may be exhausting, but it’s load-bearing – without it, you fear the whole structure would collapse.

Reactive form. When a flaw is exposed despite your efforts, immediate damage control, alternative explanations, or sometimes genuine dissociation where the event becomes hazy and unreal.

Toward recovery. The cure for idealization protection is building a floor – a sense of worth that doesn’t depend on admiration. Reparenting in therapy is key. It offers the experience of being valued for who you are, not just what you achieve. Mentalization-Based Treatment (MBT) can train you to recognize in other when they’re read to love a flawed you more than a perfect you. Self-compassion (Kristin Neff’s work) trains unconditional warmth toward yourself. Start intentionally revealing small imperfections in safe relationships – exposure therapy for the fear of being seen as flawed. Mindfulness helps you notice the panic when a flaw surfaces, creating space before the damage-control reflex kicks in. The goal isn’t to stop caring about excellence, but to survive being seen as human.

Devaluation

What it is. Diminishing the other person’s worth, value, or standing – making their needs, feelings, or boundaries not matter by making them not matter. This protects against envy, against others having power over you, and often against your own history of being devalued. It’s the flip-side of idealization.

The underlying wound. Devaluation often serves multiple functions. It can defend against envy – if they’re lesser, their successes don’t threaten you. It can protect against dependency – if they’re not that important, you don’t need them. And frequently, it re-enacts what was done to you: if you were chronically devalued, diminished, or treated as less-than, you may have internalized that someone in every relationship has to be “less.” Making it them instead of you feels like winning or survival.

How it feels from the inside. The other person seems genuinely less important, less competent, less reasonable, or less deserving. Their boundary feels ridiculous because they’re being ridiculous. Their hurt feelings are them being “oversensitive,” and you can dismiss them anyway as a them problem. This isn’t strategic – in the moment, you really see them as lesser, inferior, irrelevant. There may be contempt, dismissiveness, pity, or a sense of being surrounded by people who just don’t get it. Sometimes, underneath, there’s a flicker of envy or threat that the devaluation is managing.

Proactive form. Chronically positioning others as lesser. You may have a taxonomy of people’s inadequacies that feels like clear-eyed assessment rather than defense. Others’ boundaries or needs don’t warrant respect because they don’t warrant respect.

Reactive form. When someone sets a boundary or challenges you, they’re immediately reframed as ridiculous, oversensitive, controlling, pitiable, or otherwise not worth taking seriously.

Toward recovery. Devaluation often protects against envy, dependency, and acknowledging flaws – so working with those is key. Shadow work helps you recognize the qualities you’re diminishing in others as things you might fear or want in yourself. Alternately, the previous two sections can be helpful because devaluation is just the flip side of idealization, so as the need to idealize fades away, so does the need to devalue. Mentalization-Based Treatment (MBT) trains curiosity about others’ inner worlds, which is incompatible with dismissing them as lesser. Gratitude exercises specifically for the people you tend to devalue can also rewire the reflex. If you were chronically devalued yourself, grief work and trauma processing (EMDR, somatic therapies) can help you stop perpetuating what was done to you. Circling and authentic relating put you in contact with others’ full humanity in ways that make devaluation harder to maintain. The shift: from “they’re lesser” to “they’re different, and their perspective is real.”

Control-Seeking

What it is. Perceiving criticism or boundary-setting as an attack and responding by seizing control of the interaction. This often develops in chaotic or abusive early environments where the best defense was a strong offence.

The underlying wound. If you grew up in an unpredictable or dangerous environment – where criticism preceded punishment, where vulnerability was exploited, where you never knew when the next blow (physical or emotional) would land – your nervous system learned that safety means control. Letting someone else set the terms of an interaction, letting a criticism land, letting yourself be accountable – all of these feel like letting your guard down in enemy territory. The wound is: “I was blindsided and hurt when I was vulnerable, so I must never be vulnerable again.” Control isn’t about dominance; it’s about survival.

How it feels from the inside. When someone expresses hurt or sets a boundary, it feels like the opening move in an attack. Your nervous system may go into fight mode before you’ve consciously processed what’s happening. There’s a sense that you have to get ahead of this – that if you let them finish, if you let the criticism land, something terrible will happen. It feels like survival, not strategy.

Proactive form. Constant vigilance. Scanning for threats. Preemptive strikes – attacking potential criticism before it can be voiced. “The best defense is a good offence.” You may have developed sophisticated skills at steering conversations away from dangerous territory.

Reactive form. When criticized, immediate counterattack: “What about when you…?” Overwhelming the other with grievances. DARVO (Deny, Attack, Reverse Victim and Offender). The goal, often unconscious, is to end the interaction on your terms, with your narrative in control.

Toward recovery. Control-seeking is fundamentally a nervous system issue – your body learned that vulnerability means danger. Trauma processing (EMDR, IFS, somatic experiencing) can discharge the original survival responses so they’re not constantly online. The distress tolerance skills from Dialectical Behavioral Therapy (STOP and TIPP) help you survive the feeling of being criticized without immediately counterattacking. Mindfulness meditation builds the capacity to notice the threat response arising and pause before acting. Consider whether your current environment might actually require this level of vigilance; sometimes changing environments reduces the triggers. Mood stabilizers can lower baseline reactivity while you do the deeper work. The shift: from “I must control or I’ll be destroyed” to “I can stay present even when it’s uncomfortable.”

Resentment

What it is. Resisting others’ boundaries because you never got to have your own. There’s a deep sense that boundaries being enforced against you is unfair – because it is unfair in some cosmic sense. Just the person who is doing it not is not the same who treated you unfairly. The resentment is displaced from its original target (often caregivers) onto current relationships.

The underlying wound. This pattern carries the weight of developmental injustice. If your boundaries were bulldozed in childhood – if “no” wasn’t permitted, if your space, body, feelings, or preferences were overridden – you learned that boundaries are things powerful people get to have and vulnerable people don’t. Watching others set boundaries activates that old wound: “They get to have what I never got to have.” The rage isn’t really about the current situation; it’s grief and fury about what happened to you, displaced onto whoever is currently asserting a limit. The wound is: “I was denied the basic right to boundaries, and it’s unbearable and humiliating to see others exercise that right against me.”

How it feels from the inside. When someone sets a boundary, there’s immediate, righteous anger. “How dare you.” It feels like they’re the one being unreasonable, unfair, selfish. There may be a sense of “rules for thee but not for me” – not as conscious hypocrisy, but as genuine felt injustice. Underneath, often inaccessible, is grief: the child who never got to say “no,” whose boundaries were bulldozed, who’s now watching others have what they never had.

Proactive form. A chronic chip-on-shoulder about boundaries and rules. Others’ boundaries are inherently suspect, probably oppressive, definitely hypocritical. You may be drawn to ideologies that frame boundaries as control or see yourself as a righteous transgressor of arbitrary limits.

Reactive form. When a specific boundary is enforced, fierce resistance: “You don’t get to tell me what to do.” The intensity is disproportionate to the situation because it’s carrying the weight of historical injustice, not just the present moment.

Toward recovery. Resentment requires grief – mourning that what you didn’t get to have. The rage is real, but it belongs to the past; therapy can help you direct it toward its true source rather than displacing it onto present relationships. Reparenting yourself includes learning to set your own boundaries now, which can reduce the envy of others doing so. EMDR and trauma processing can help metabolize the developmental injustice so it’s not constantly activated. Practice distinguishing past from present: “Is this person actually being unfair, or does it feel unfair because boundaries always felt unfair?” Self-compassion for the child who was bulldozed can coexist with accountability for the adult who’s now doing the bulldozing. Journaling can help you notice when the intensity doesn’t match the situation – that’s your signal that history is present. The shift: from “how dare they” to “their boundary is legitimate, even though I never got to have mine.”

A Final Note

If you see yourself in these patterns, please remember: These aren’t moral failures. They’re adaptations – once-necessary protections that have outlived their usefulness. Seeing them clearly is painful but essential.

Recovery doesn’t mean the patterns disappear overnight. It means you start to catch them, to feel the pull and choose differently, to repair when you’ve caused harm. It’s slow work, and it requires support – a good therapist, patient loved ones, and above all, compassion for yourself.

The fact that you’re reading this, trying to understand, is itself significant. These patterns thrive in the dark. Naming them is the beginning of change.

Summary: We’re building an AI-powered mental health chatbot targeting populations with severe mental healthcare shortages. This post presents our framework for prioritizing which conditions and regions to focus on first, synthesizing data on global mental health workforce gaps, existing digital resources across 15+ diagnostic categories, and AI intervention suitability. A key consideration is “breaking the cycle of trauma and tyranny” – addressing conditions that contribute to insecure attachment and power-seeking behavior that perpetuate conflict and authoritarianism.

Note: This is the summary of our preliminary findings including personal observations and inferences. We consider this level of certainty sufficient for current purposes in this early exploratory phase. We’ve written this article with the assistance of Claude and Gemini. We seek further advice and suggestions for the refinement or reframing of the project’s scope.

Introduction

The supply of mental health workers per 100,000 population ranges from 67 in high-income countries to 1 in low-income countries. In all settings, though, there are people whose mental health problems are not addressed for lack of affordable and accessible care.

Recent advances in large language models (LLMs) present an opportunity to partially address this gap through scalable, low-cost interventions. Our team is developing an AI mental health chatbot and hopes to make it useful for populations with the least access to traditional mental healthcare.

However, mental health is vast: various diagnostic manuals contain hundreds of diagnoses each, which overlap in complex ways, and mental health needs vary dramatically across cultural contexts. We cannot effectively serve everyone simultaneously. This post outlines our systematic approach to prioritization and solicits feedback on our reasoning and potential blind spots.

Our Context and Constraints

Team composition: Multilingual team with fluency in English, German, Hindi, Tamil, Estonian, Finnish, and Mandarin.

Unique advantage: Team lead has direct connections within communities struggling with Cluster B personality disorders (ASPD, BPD, HPD, NPD) and familiarity with mentalization-based treatment (MBT), potentially enabling culturally competent outreach to highly stigmatized populations typically underserved by existing resources. Our team also includes licensed psychologists and published psychology researchers.

Long-term motivation: Interest in “breaking the cycle of trauma and tyranny” – addressing the intergenerational transmission of trauma, insecure attachment, and personality pathology that contributes to authoritarian leadership and societal instability. This framework also suggests that healing trauma and fostering secure attachment in this generation can reduce power-seeking pathology and conflict risk in the next.

Current stage: Pre-launch prioritization phase. We’re determining which conditions and populations to serve first, rather than attempting a one-size-fits-all approach.

Methodology: Systematic Resource Mapping

Before prioritizing, we conducted a comprehensive landscape analysis across 15+ major diagnostic categories, examining:

1. Existing self-help resources (workbooks, apps, online communities) for each specific disorder

2. Evidence-based interventions and their amenability to AI delivery

3. Global mental health workforce distribution using WHO data

4. Technology adoption patterns and infrastructure constraints

5. Cultural considerations affecting mental health help-seeking

6. Intergenerational impact on attachment security and power-seeking behavior

Our analysis covered:

1. Mood disorders (depression, bipolar I, bipolar II, cyclothymic disorder, dysthymia/persistent depressive disorder, disruptive mood dysregulation disorder, premenstrual dysphoric disorder)

2. Anxiety disorders (generalized anxiety disorder/GAD, panic disorder, agoraphobia, social anxiety disorder/social phobia, specific phobias, separation anxiety disorder, selective mutism)

   1. Trauma and stressor-related disorders (PTSD, complex PTSD, acute stress disorder, adjustment disorders, reactive attachment disorder, disinhibited social engagement disorder)

   2. Obsessive-compulsive and related disorders (OCD, body dysmorphic disorder, hoarding disorder, trichotillomania/hair-pulling disorder, excoriation/skin-picking disorder)

3. Personality disorders (Cluster A: paranoid, schizoid, schizotypal; Cluster B: antisocial/ASPD, borderline/BPD, histrionic/HPD, narcissistic/NPD; Cluster C: avoidant, dependent, obsessive-compulsive)

4. Psychotic disorders (schizophrenia, schizoaffective disorder, schizophreniform disorder, brief psychotic disorder, delusional disorder, psychotic depression, substance-induced psychotic disorder)

5. Neurodevelopmental disorders (ADHD, autism spectrum disorder/ASD, intellectual disabilities, communication disorders including speech sound disorder and childhood-onset fluency disorder/stuttering, specific learning disorders including dyslexia, dyscalculia, and dysgraphia, motor disorders including developmental coordination disorder/dyspraxia, tic disorders including Tourette syndrome)

6. Substance use disorders (alcohol use disorder, opioid use disorder, cannabis use disorder, stimulant use disorder including cocaine and amphetamines, sedative/hypnotic/anxiolytic use disorder, tobacco use disorder, hallucinogen use disorder, inhalant use disorder, gambling disorder)

7. Feeding and eating disorders (anorexia nervosa, bulimia nervosa, binge eating disorder, avoidant/restrictive food intake disorder/ARFID, pica, rumination disorder)

8. Sleep-wake disorders (insomnia disorder, hypersomnolence disorder, narcolepsy, obstructive sleep apnea, central sleep apnea, sleep-related hypoventilation, circadian rhythm sleep-wake disorders, non-rapid eye movement sleep arousal disorders including sleepwalking and sleep terrors, nightmare disorder, rapid eye movement sleep behavior disorder, restless legs syndrome)

9. Somatic symptom and related disorders (somatic symptom disorder, illness anxiety disorder/hypochondriasis, conversion disorder/functional neurological symptom disorder, factitious disorder, psychological factors affecting other medical conditions)

10. Dissociative disorders (dissociative identity disorder/DID, dissociative amnesia, depersonalization/derealization disorder, other specified dissociative disorder/OSDD)

11. Sexual disorders

    1. Sexual dysfunctions (erectile disorder, female sexual interest/arousal disorder, male hypoactive sexual desire disorder, female orgasmic disorder, delayed ejaculation, premature/early ejaculation, genito-pelvic pain/penetration disorder)

    2. Paraphilic disorders (voyeuristic disorder, exhibitionistic disorder, frotteuristic disorder, sexual masochism disorder, sexual sadism disorder, pedophilic disorder, fetishistic disorder, transvestic disorder)

12. Disruptive, impulse-control, and conduct disorders (oppositional defiant disorder, intermittent explosive disorder, conduct disorder, antisocial personality disorder, pyromania, kleptomania)

For each category, we assessed resource availability (extensive/moderate/limited/very limited), identified gaps, and analyzed cultural/technological adoption patterns.

This categorization is one possible one among many. The complexity and ontological uncertainty of mental health as a field (at least in terms of nosology and diagnosis) is reflected in the abundance of various frameworks, such as the National Institute of Mental Health’s Research Domain Criteria, research by the Hierarchical Taxonomy of Psychopathology, the Diagnostic and Statistical Manual, and the Psychodynamic Diagnostic Manual and related frameworks.

Key Finding: Dramatic Workforce Disparities

Using the latest WHO Mental Health Report data, we identified severe disparities in mental health workforce availability:

Global averages by World Bank income group (specialized mental health workers per 100,000 population):

1. High-Income Countries (HIC): 67.2

2. Upper-Middle-Income Countries (UMIC): 19.3

3. Lower-Middle-Income Countries (LMIC): 2.4

4. Low-Income Countries (LIC): 1.1

By WHO region:

1. EUR (Europe): 80.4 per 100k

2. AMR (Americas): 22.2 per 100k

3. WPR (Western Pacific): 14.1 per 100k

4. EMR (Eastern Mediterranean): 4.7 per 100k

5. SEAR (South-East Asia): 4.0 per 100k

6. AFR (Africa): 2.2 per 100k

This represents a 60-fold difference between highest and lowest resourced regions. In practical terms: a person with depression in Norway has access to ~80 mental health workers per 100,000 people, while someone in Uganda has access to ~0.1 – an 800-fold difference.

The Trauma-Tyranny Cycle: A Developmental Perspective on Long-Term Impact

Beyond immediate suffering, untreated mental health conditions – particularly trauma-related disorders and resulting attachment pathology – contribute to a self-perpetuating cycle that shapes political stability and conflict risk across generations.

The Cycle Model

The cycle operates as follows:

1. Wars, societal collapse, and adverse childhood experiences → cause widespread trauma and chronic stress

2. Trauma and parental mental health problems → disrupt healthy attachment formation in children

3. Insecure attachment and unprocessed trauma → increase the susceptibility to (and rate of) power-seeking dictators

4. Authoritarian leadership and poor institutional decision-making → increases risk of wars and societal collapse, perpetuating the cycle

This framework suggests that mental health interventions – particularly those addressing trauma, attachment, and personality pathology – have downstream effects on political stability, institutional quality, and conflict risk that compound across generations.

Evidence Base

Research supporting elements of this cycle:

• Trauma transmission: Parental trauma predicts insecure attachment in children; war trauma affects parenting practices across generations

• Attachment and leadership: Studies link insecure attachment patterns to authoritarian followership and preference for “strong man” leaders

• Personality pathology and power: Cluster B traits (particularly NPD and ASPD) overrepresented in positions of political power

• Developmental origins: Most personality disorders rooted in childhood trauma, neglect, and attachment disruption

• Malleability: Personality pathology treatable; attachment patterns can shift; trauma can heal – suggesting interventions can break the cycle

Why This Matters for Prioritization

This framework suggests we should weight conditions not only by immediate burden but by their role in perpetuating intergenerational cycles of suffering and instability:

High long-term impact conditions:

• PTSD and complex trauma: Direct cycle driver; prevents secure parenting

• Personality disorders (especially Cluster B): Direct link to power-seeking and authoritarian tendencies

• Attachment-disrupting conditions: Depression, anxiety, substance use in parents affect children’s attachment security

• Childhood conduct problems: Early intervention prevents crystallization into ASPD

High-risk populations:

• Conflict-affected regions: Active cycle perpetuation; highest intervention value

• Parents and prospective parents: Breaking intergenerational transmission

• Adolescents and young adults: Critical window before personality patterns rigidify and before they become parents

Intervention modalities with cycle-breaking potential:

• Trauma healing: Reduces transmission to next generation

• Parenting support: Directly improves children’s attachment security

• Personality disorder treatment: Reduces power-seeking behavior; improves parenting

• Resilience building: Strengthens population-level resistance to authoritarian messaging

This lens makes conditions like PTSD, personality disorders, and perinatal mental health higher priority despite some challenges, because successfully treating one generation protects the next.

Prioritization Framework

We developed a multi-tier framework weighing 20+ criteria across seven domains:

Tier 1: Core Feasibility

Safety & Risk Profile

• Can we deliver interventions without significant risk of harm?

• Do we have robust crisis protocols for high-risk situations?

• Can we reliably identify and escalate emergencies?

Key insight: This criterion should filter out conditions before other considerations. Active psychosis, acute suicidality, severe eating disorders in crisis, and mania present risks that outweigh potential benefits of unsupervised AI intervention.

Language Capacity

• Does our team have native/fluent speakers for seeking feedback, noticing and responding to problems?

• Can we avoid mere translation in favor of genuine cultural competence?

Technology Access & Literacy

• Smartphone penetration in target regions

• Data costs relative to local income

• Digital literacy rates

• Internet infrastructure reliability

Equity & Justice

• Prioritizing most underserved over most profitable

• Ensuring accessibility for lowest-income users

Cultural Sensitivity

• Avoiding imposition of Western psychiatric models on non-Western contexts

• Incorporating local healing traditions

• Collaborating with local communities and professionals

Transparency & Limitations

• Clear communication about AI capabilities and limitations

• Avoiding dependency creation

• Providing pathways to human care

Tier 2: Impact Potential

Mental Health Workforce Gap

• Where is the treatment gap largest?

• Where will AI provide the highest marginal benefit?

Disease Burden & Prevalence

• DALYs (Disability-Adjusted Life Years) lost

• Absolute number of people affected

• Regional variation in prevalence

Stigma & Barriers to Traditional Care

• Where does stigma prevent help-seeking?

• Where do cultural/gender restrictions limit access to human therapists?

• Where might anonymous AI access lower barriers?

Attachment Security Impact

• Does treating this condition improve parenting capacity?

• Will treatment reduce transmission of insecure attachment to children?

• Does the condition directly disrupt attachment formation?

High impact: Perinatal depression/anxiety, PTSD, substance use, personality disorders (all affect parenting)

Moderate impact: Depression, anxiety in parents; childhood trauma-related conditions

Power-Seeking & Authoritarianism Risk

• Does the condition involve patterns associated with malevolent leadership? (NPD, ASPD, sadism)

• Does healing reduce power-seeking behavior or improve use of power?

• Does treatment reduce susceptibility to authoritarian messaging?

High impact: Cluster B personality disorders, especially NPD/ASPD combinations; trauma creating “might makes right” worldviews

Moderate impact: Any condition improving emotional regulation and reducing reactivity to threats

Conflict & Instability Risk

• Is the condition prevalent in conflict zones, perpetuating cycles?

• Does the condition directly increase interpersonal violence risk?

• Does healing improve institutional decision-making quality?

High impact: PTSD in conflict zones, ASPD, substance use disorders, impulse control disorders

Moderate impact: Conditions affecting judgment and emotional regulation

Critical Developmental Windows

• Can we intervene before personality patterns rigidify? (adolescence/early adulthood)

• Can we intervene before individuals become parents?

• Can we heal parents before patterns transmit to children?

High impact: Adolescent/young adult populations; perinatal interventions; parenting support

Population-Level Resilience

• Does healing this condition make populations more resistant to manipulation?

• Does treatment promote secure attachment at scale?

• Does intervention build what Antonovsky calls “sense of coherence” (comprehensibility, manageability, meaningfulness)?

High impact: Trauma healing, attachment-focused interventions, mental health literacy programs

Tier 3: AI Suitability

Amenability to Structured Interventions

AI is most effective for conditions with structured, manualized treatments:

• Excellent fit: CBT for depression/anxiety, CBT-I for insomnia, exposure protocols, behavioral activation, psychoeducation

• Moderate fit: Motivational interviewing, DBT skills training, habit tracking, mentalization practice

• Poor fit: Complex trauma requiring relational depth, severe personality disorders needing nuanced therapeutic tensions, conditions requiring physical examination

Self-Help Amenability

• Does evidence support self-directed interventions?

• Can people improve without immediate professional involvement?

Data & Training Resources

• Quality of LLM training data for condition

• Availability of evidence-based treatment manuals

• Ability to validate AI responses against gold standards

Tier 4: Market Gap Analysis

Existing Digital Solutions

• Where are markets oversaturated vs. underserved?

• Where do existing solutions fail to serve LMICs?

Our finding: Dramatic inequality mirrors workforce gaps. Most mental health apps target English-speaking HIC markets. Very few quality apps exist in Hindi (500M+ speakers), Bengali (230M+ speakers), or Tamil (80M+ speakers). African markets almost entirely neglected except South Africa.

Cultural Adaptation Needs

• Where do Western psychiatric models fail to translate?

• Where is somatic expression of distress more common?

• Collectivist vs. individualist therapy frameworks

Existing Workbook/Professional Resource Availability

• Can we adapt existing evidence-based resources?

• Do gaps indicate lack of proven interventions or just lack of accessibility?

Tier 5: Strategic Considerations

Scalability Potential

• Size of potential user base

• Growth trajectory of condition awareness/diagnosis

• Platform effects and community features

Regulatory & Liability Landscape

• Regulatory requirements vary by region and intervention type

• Risk increases with diagnostic/treatment claims vs. psychoeducation/support

Monetization Potential

• Willingness to pay varies by region and condition

• Venture capital funding opportunities

• Grant funding opportunities (WHO, NGOs, government programs)

• Freemium viability for impact at scale

Partnership Opportunities

• NGO/WHO initiatives in target regions

• Research institutions for validation studies

• Local healthcare systems for integration

• Telehealth providers for triage/adjunct services

Measurement & Validation

• Can we measure impact using validated scales?

• Feasibility of clinical validation studies

• User engagement and retention metrics

Condition Prioritization: Rankings and Rationale

Using this framework, we ranked conditions by overall suitability. Note that the assessment of fuzzy regional factors of suitability is heavily informed by AI.

Tier 1: Highest Priority

PTSD (Prioritize Conflict-Affected Regions)

• Burden: High in conflict-affected regions (Afghanistan, Sudan, South Sudan, Syria, Yemen, DRC, Myanmar, Pakistan border regions, Northeast Nigeria)

• Gap: Extreme shortage of trauma-trained therapists

• AI Fit: Good – PE and CPT components are structured

• Safety: Moderate risk – requires robust crisis protocols

• Existing Resources: Very few culturally appropriate apps for conflict-affected LMICs

• Stigma: Extremely high in many cultures; AI may lower barriers

• Cultural: Trauma narratives culturally specific; requires careful adaptation

• Cycle Impact: ⭐⭐⭐⭐⭐ – PTSD is the primary cycle driver. Traumatized parents have difficulty providing secure attachment; PTSD directly transmits across generations via parenting practices and epigenetics; conflict-zone trauma creates conditions for the next generation of authoritarian leaders; healing trauma breaks the cycle at its source.

Personality Disorders – Strategic Focus on Cluster B (NPD, ASPD, BPD, HPD)

Our team’s unique positioning: Given team lead’s connections in NPD/ASPD/HPD communities and MBT training, we have potential advantages in serving this highly stigmatized population.

• Burden: ~10% of population; severe functional impairment

• Gap: Massive stigma prevents help-seeking; very few specialists even in HICs

• AI Fit: Moderate – MBT requires nuanced mentalizing that challenges AI, BUT psychoeducation and skill-building components could help

• Safety: Moderate-High risk depending on disorder (ASPD risk assessment, BPD self-harm)

• Existing Resources: Very limited for Cluster B; most resources focus on “surviving” people with NPD/BPD rather than helping them

• Cultural: Cluster B presentations culturally mediated; requires deep cultural knowledge

• Cycle Impact: ⭐⭐⭐⭐⭐ – This is the other primary cycle driver. Cluster B disorders (especially NPD and ASPD) are directly associated with power-seeking behavior, authoritarian leadership, and malevolent use of power. These conditions arise from childhood trauma and transmit intergenerationally through disrupted attachment. Healing personality disorders directly reduces the pool from which malevolent leaders emerge. BPD, while less associated with power-seeking, severely disrupts parenting and attachment.

The compassionate case: As I’ve argued elsewhere, people with NPD and ASPD are not “evil” – they are using brilliant childhood adaptations to survive impossible situations. These adaptations become maladaptive in adulthood but can heal with appropriate support, typically in just a few years of therapy. Many individuals with these conditions desperately want help but cannot access it due to stigma, cost, and scarcity of trained therapists.

The strategic case: The overlap between Cluster B traits and positions of power means that even small improvements in this population have outsized effects on institutional quality, conflict risk, and the next generation’s wellbeing. If we can help even a fraction of people with these conditions, the downstream effects on politics, violence, and intergenerational trauma transmission could be substantial.

Possible approach: Focus on psychoeducation, mentalization skills practice, emotion regulation – NOT replacement for intensive therapy but potentially helpful adjunct for people unable/unwilling to access traditional care due to stigma. Clear about AI limitations. Strong safety protocols for violence risk. Initial target: adults with NPD/ASPD seeking help (not those court-mandated or uninterested in change).

Conduct Disorder / Childhood Trauma Interventions

• Burden: Common in high-adversity environments

• Gap: Very few child mental health services in LMICs

• AI Fit: Moderate – parenting interventions structured; child-facing interventions more challenging

• Safety: Moderate – requires careful age-appropriate design

• Challenges: Would need separate child-focused platform; consent/privacy issues

• Cycle Impact: ⭐⭐⭐⭐⭐ – Early intervention prevents personality disorder crystallization. Conduct disorder is precursor to ASPD; childhood trauma is the root cause of most personality pathology. Intervening in childhood/adolescence is the most effective cycle-breaking point, before patterns rigidify. Biggest challenge: reaching children requires a different platform approach.

Perinatal Mental Health (Depression, Anxiety)

• Burden: Massive need in your regions (maternal mortality links)

• Gap: Low resources for perinatal mental health

• AI Fit: Good for psychoeducation, CBT components

• Safety: Moderate-High risk (infanticide, severe postpartum psychosis require emergency response)

• Opportunity: WHO priority area; partnership potential

• Cycle Impact: ⭐⭐⭐⭐⭐ – This is a peak intervention point for attachment security. Perinatal mental health directly affects infant attachment formation; this is the most critical developmental window; treating mothers prevents transmission to the next generation at the source.

Tier 2: High Priority

Depression (Mild–Moderate)

• Burden: Leading cause of disability globally; ~280M people affected

• Safety: Low risk if severe/suicidal cases properly filtered and escalated

• AI Fit: Excellent – CBT and behavioral activation are highly structured

• Evidence: Strong self-help efficacy data

• Workforce Gap: Massive gap in LIC/LMIC (treatment gap >80%)

• Existing Resources: Many apps exist BUT dramatic language gap (almost nothing quality in Hindi/Tamil/Bengali for LMIC contexts)

• Measurement: PHQ-9 validated globally

• Cultural: Depression presents across cultures but may manifest somatically – requires adaptation

• Cycle Impact: ⭐⭐⭐ – Parental depression significantly disrupts attachment security; reduces parenting capacity; transmits intergenerationally

Anxiety Disorders (GAD, Social Anxiety, Panic)

• Burden: ~300M affected globally; highly disabling

• Safety: Low risk

• AI Fit: Excellent – CBT protocols, exposure hierarchies, grounding techniques all structured

• Evidence: Strong self-help efficacy

• Gap: Similar to depression – huge LMIC gap, language barriers

• Measurement: GAD-7, SPIN validated globally

• Cultural: Anxiety universal but expression varies; requires cultural adaptation

• Cycle Impact: ⭐⭐⭐ – Anxious parenting affects children’s attachment security; hypervigilance transmits intergenerationally; anxiety increases susceptibility to threat-based authoritarian messaging

Substance Use Disorders (Harm Reduction Focus)

• Burden: Major cause of DALYs in many LMICs

• Gap: Extreme stigma prevents help-seeking; very few services

• AI Fit: Good for motivational interviewing, harm reduction education, tracking

• Safety: Moderate – requires crisis protocols for overdose risk, withdrawal

• Cultural: Highly stigmatized; AI anonymity major advantage

• Challenges: Cultural/religious sensitivities (alcohol in Muslim countries, substance use stigma in conservative societies)

• Cycle Impact: ⭐⭐⭐⭐ – Parental substance use severely disrupts attachment; increases violence and neglect; intergenerational transmission common; substance use associated with impulsive violence and poor institutional decision-making

Tier 3: Medium to Low Priority

Insomnia (Primary & Comorbid)

• Burden: ~30% of adults affected; impacts physical and mental health

• Safety: Zero acute risk

• AI Fit: PERFECT – CBT-I is highly manualized and structured

• Evidence: CBT-I self-help proven effective (comparable to therapist-delivered)

• Gap: Very few quality apps in target languages despite universal problem

• Measurement: Sleep diary, ISI scale

• Cultural: Low stigma = higher engagement; universal relevance

• Unique advantage: “Gateway” condition – treating insomnia often improves comorbid depression/anxiety

• Cycle Impact: ⭐⭐ – Better sleep improves emotional regulation and parenting quality; indirect effects on attachment security

OCD

• Burden: Highly disabling; ~2–3% prevalence

• AI Fit: Excellent – ERP is highly structured

• Gap: Very few ERP-trained therapists even in HICs

• Safety: Low risk, beyond the danger of reinforcing compulsions

• Existing Resources: Few quality apps in any language

• Cultural: Presentations vary (religious scrupulosity, contamination fears vary culturally)

• Cycle Impact: ⭐ – Minimal direct effect on attachment or power-seeking, though severe OCD can impair parenting

ADHD (Adults & Adolescents)

• Burden: Growing awareness in LMICs; severe underdiagnosis

• AI Fit: Excellent for skills training (time management, organization, emotional regulation)

• Safety: Zero acute risk

• Gap: Massive – most LMICs have near-zero ADHD services for adults

• Existing Resources: MANY productivity apps BUT few culturally adapted for India/Africa; mostly assume HIC work contexts

• Cultural: ADHD increasingly recognized cross-culturally but stigma varies

• Cycle Impact: ⭐⭐ – Untreated ADHD in parents complicates parenting; emotion dysregulation affects children; but not directly linked to power-seeking or authoritarianism

Somatic Symptom Disorders

• Burden: Very common in target regions (somatic expression of distress culturally normative in many Asian/African contexts)

• Gap: Almost NO existing digital resources

• Cultural Fit: Highly relevant – Western psychology often fails to address

• Challenges: Requires medical rule-outs (liability risk); validation complex

• Opportunity: Major gap to fill with culturally appropriate approaches

• Cycle Impact: ⭐⭐ – Chronic pain/illness affects parenting capacity; but not directly linked to attachment disruption or power-seeking

Bipolar Disorder

• High safety risk (mania, suicidality)

• Medication essential (beyond AI scope)

• Complex case management needs

• Cycle Impact: ⭐⭐ – Untreated bipolar disrupts parenting, but with medication most people stable

Eating Disorders

• High medical risk requiring monitoring

• Lower prevalence in initial target regions (though rising)

• Complex interventions

• Cycle Impact: ⭐ – Minimal direct cycle effects except in severe cases affecting parenting

Psychotic Disorders

• HIGH safety risk

• Medication usually essential

• Anosognosia limits engagement

• BUT: Family psychoeducation could be valuable supportive intervention

• Cycle Impact: ⭐ – Most people with schizophrenia are not violent or power-seeking; primary impact is on individual/family suffering

Geographic Prioritization: Country Rankings

Using mental health workforce data (per 100,000 population), World Bank income classifications, language accessibility, technology infrastructure, and conflict/trauma exposure, but ignoring strategic, marketing, or funding considerations. Fuzzy regional, cultural, and historical impressions again draw heavily on AI.

Tier 1: Highest Priority Markets 🎯

India

• Mental health workers: ~0.3–0.6 per 100k (vs. 67.2 in HICs)

• Population: 1.43 billion

• Languages: Hindi (550M speakers), Tamil (80M speakers), English (widespread)

• Income: LMIC (but wide internal variation)

• Technology: Rapidly growing smartphone penetration; good mobile infrastructure in urban/suburban areas

• Mental Health Burden: High rates of depression, anxiety, suicide

• Conflict/Trauma: Kashmir conflict; communal violence; high rates of adverse childhood experiences

• Cycle Status: ⭐⭐⭐ – Significant trauma exposure; growing but incomplete mental health infrastructure; critical window to intervene before patterns rigidify

• Rationale: Largest addressable market with our language capabilities; enormous gap; growing mental health awareness

• Challenges: Digital divide (rural vs. urban); data costs; diverse cultural contexts

Pakistan

• Mental health workers: ~0.2–0.5 per 100k

• Population: 231 million

• Languages: English (official), Urdu (mutually intelligible with Hindi)

• Income: LMIC

• Technology: Growing smartphone adoption; less infrastructure than India

• Mental Health Burden: High; extreme stigma particularly around women’s mental health

• Conflict/Trauma: Afghan border terrorism; internal sectarian violence; TTP attacks; drone strike trauma; significant PTSD burden

• Cycle Status: ⭐⭐⭐⭐ – Active conflict perpetuating trauma cycles; very low mental health capacity; strong stigma preventing help-seeking

• Rationale: Second-largest Urdu/Hindi-speaking population; severe gap; AI anonymity crucial given stigma; trauma healing critical for conflict de-escalation

• Challenges: Political instability; conservative cultural norms; lower female digital access

Afghanistan

• Mental health workers: ~0.02–0.05 per 100k (among world’s lowest)

• Population: 41 million

• Languages: English (limited), but potential Dari/Pashto development

• Income: LIC

• Technology: Growing mobile penetration despite infrastructure challenges

• Mental Health Burden: Extreme – decades of war

• Conflict/Trauma: 40+ years continuous conflict; Taliban rule trauma; highest trauma burden globally

• Cycle Status: ⭐⭐⭐⭐⭐ – Active cycle perpetuation at crisis levels. Entire generations traumatized; minimal mental health infrastructure; current authoritarianism driven by trauma cycles. Highest need but also highest access barriers.

• Rationale: Most acute trauma burden; greatest potential cycle-breaking impact

• Challenges: Security situation; female access restrictions; language barrier (would need Dari/Pashto); political complications

Nigeria

• Mental health workers: ~0.2–0.3 per 100k

• Population: 220 million

• Languages: English (official)

• Income: LMIC

• Technology: Variable – good in urban areas, limited in rural

• Mental Health Burden: High; stigma extreme

• Conflict/Trauma: Boko Haram in northeast (mass trauma, kidnappings); farmer-herder violence; Niger Delta conflict; significant PTSD burden

• Cycle Status: ⭐⭐⭐⭐ – Active conflict zones; trauma perpetuating instability; religious extremism linked to trauma cycles

• Rationale: Largest African market; English-speaking; enormous gap; trauma healing critical in conflict zones

• Challenges: Infrastructure variability; cultural diversity (250+ ethnic groups); data costs; religious considerations

South Sudan

• Mental health workers: <0.05 per 100k

• Population: 11 million

• Languages: English (official)

• Income: LIC

• Technology: Very limited but growing mobile access

• Mental Health Burden: Extreme – ongoing conflict

• Conflict/Trauma: Continuous war since independence; mass displacement; extreme violence exposure; one of world’s highest trauma burdens

• Cycle Status: ⭐⭐⭐⭐⭐ – Acute cycle perpetuation; virtually no mental health services; urgent intervention needed

• Rationale: Desperate need; English-speaking; potential for enormous impact

• Challenges: Infrastructure extremely limited; ongoing conflict; very low literacy

Democratic Republic of Congo

• Mental health workers: ~0.05 per 100k

• Population: 99 million

• Languages: French (official), some English

• Income: LIC

• Technology: Growing mobile penetration despite poor infrastructure

• Mental Health Burden: Extreme – decades of conflict

• Conflict/Trauma: 25+ years of war; mass rape as weapon; child soldiers; extreme violence; ongoing Eastern Congo conflict

• Cycle Status: ⭐⭐⭐⭐⭐ – Severe trauma perpetuating instability; virtually no services

• Rationale: Massive trauma burden; enormous need

• Challenges: Language barrier (would need French); infrastructure; ongoing violence; complexity

Myanmar

• Mental health workers: ~0.1 per 100k

• Population: 54 million

• Languages: English (some), Mandarin (some)

• Income: LMIC

• Technology: Previously growing, now complicated by military coup

• Mental Health Burden: High and worsening

• Conflict/Trauma: Military coup trauma; Rohingya genocide; ethnic conflicts; civil war

• Cycle Status: ⭐⭐⭐⭐⭐ – Active authoritarian violence; trauma-driven conflict cycles; dramatic example of cycle in action

• Rationale: Clear case of trauma-tyranny cycle; potential intervention point

• Challenges: Political situation; military restrictions; language barriers; safety concerns