26 Comments

Yours is a radical Idea that psychiatric establishments aren't likely to welcome. "It's too radical, too reductionist, too mechanistic, too insensitive to the many subtleties of varieties of human personality". I could go on and on about all the reasons that psychiatry orthodoxy will come up with and I think there'll be some merit in all of them.

But when all arguments for both sides are laid out, I think yours would still emerge as the most forward-thinking.

One important danger I could think of that's inherent in your proposal is that of reshaping, to a nontrivial degree, the type of people, in terms of cognitive, emotional, behavioral inclinations, who are drawn to the profession. Currently, as well as previously, psychiatry (especially the clinical part) isn't a profession known to hold much appeal for people who are mathematically or statistically inclined. In fact, the predominant intellectual inclination among its practitioners is closer to humanities than science per se. And it's easy to see why this is the case: there's no other discipline in which humans, strictly speaking, are both the subject matter as well as the end object.

In reality and in practice, it's more accurate and appropriate to compare the broad field of mental health subjects to disciplines in humanities and arts than to those in science. Why? It has an idiosyncratic element that often transcends and defies the straight and narrow paradigm eschewed by hard sciences. While there's no doubt that the medical, nosological, theoretical, and diagnostic dimensions of the discipline owe everything to the scientific method, there are other dimensions (praxes, psychotherapy, clinical interview) whose successes depend on factors that do not necessarily emanate from or related to the scientific sentiment. Such factors like empathy, curiosity, interpersonal detachment, relational warmth, attentional patience, preference for people over things, preference for stories over statistics, etc. And it could be argued that this second dimension is more deterministic of treatment outcomes, at least in individual cases, than the former which admittedly could be more significant in ensuring better result when evaluated as statistical averages.

Harking back to the danger I mentioned earlier, what are the chances that adopting the numeric paradigm as against the narrative one won't change the professional landscape of the discipline so much that it ends up losing its 'human soul'? I don't know if you consider this a legitimate critique. And if you do, what would be your response?

Expand full comment

That is a very interesting question that is much bigger than psychiatry. Basically, it is about finding a suitable division of labor between humans and machines. I think that labor should be divided so humans do what humans do best and machines do what machines do best. Machines are better at remembering the entire history of psychiatry. Humans are better at actually talking to people.

So I think that ideally, machines should make the verdict of a person and tell what is actually wrong. A machine, or an objective-minded human. Then people can be sent to therapy to empathic humans. As far as I can see, telling what is wrong with people and helping people doesn't even need to be the same job.

A more extreme example of how to divide labor between humans and machines could be a daycare. On the one hand, we want daycares to be efficiently built and run. We don't want to repeat old mistakes in their designs. We want at least someone in the staff to know a bit about group psychology and what statistically prevents toddlers from mutiny. We want technically-minded and unemotional people to think about the security of the buildings and surroundings. On the other hand, we want emphatic, people-oriented people to work in daycares. The solution is to employ different people to organize the places and to care for children on a day-to-day basis. And in schools, teachers will still be needed to teach manners for human interaction, although a computer is probably better than a human at teaching math (at least to smarter kids).

So in the case of psychology, I think that human-oriented professionals will be reduced to psychotherapists. And that doesn't have to be a bad deal for either party.

Expand full comment

I find nothing to disagree with or argue against in your solution to the potential problem I raised. I think that'll be a brilliant division of expertise if it can be pulled off.

Expand full comment

I'm not Tove, but I'll say that the phrase *human soul* implies the existence of something that doesn't literally exist, and as soon as we try to say that a human soul may still have some poetic or evocative value, we're confronted with the implication that animals don't have it. Really it's a terrible idea and needs to go away, yet sadly can't go away, because of how deeply Abrahamic monotheism is interwoven throughout Western history, culture, and thought.

Expand full comment

Thank you for your reply. I put the expression 'human soul' in a quote precisely because I wasn't insensible of the objection you rightly raised. If you're a mental health clinician, especially one who does psychotherapy (I am one), you'd understand that there are a lot more about the actual process of psychotherapy that is "evocative", imaginative, and even superstitious. Mind you, I don't mean this as a criticism of therapy as it is practiced today (even though I don't think much will or should change in the future). It's these intangible human assumptive elements, rather than the cold light of scientific fact and stance, that make psychotherapy effective and such a transformative experience.

Ironically, as much as you're right about your commentary on "human soul", it's assertions like this that I worry about if the approach proposed by Tove comes to gain purchase in the field. First, it'll transform the human professional landscape, then it'll eventually transform the way psychotherapy (among other praxes) is done. Like I said, you're right about the soul thing but the treatment aspect of psychiatry, precisely psychotherapy, isn't necessarily about asserting cold fact or or reeling out statistical averages (as important as these things are). Psychotherapy is the somewhat atheory arena where radical individualism is and must be realized. Sometimes, in pursuit of the reality and fact that define a single life, one will have to say, condone, and encourage things that aren't necessarily scientifically proven.

Expand full comment

Yeah I enjoyed your comment, and because my usual way of relating to people is to state things I don't agree with, I picked the only part that I could really respond to to respond to.

In that vein, while I'm not sure that I understand exactly what you're saying here, I do suspect that you'd probably find my opinion very directly in opposition to yours. Radical individualism is an idea adjacent to the human soul; despite our possessing specific perspectives and individual desires, we're also indelibly part of a collectivistic, eusocial species whose members have poor survival prospects on their own and whose reproduction is so dependent on cooperation what we fall into an addictive social state called love as part of the mating process. Moreover, most of us are so non-unique that by the time we are old, we have met most of the upteen-hundred different people that exist three times already - this is an inevitable claim that follows from the trait-theoretic perspective on human beings that Tove is talking about.

Expand full comment

I don't think our perspectives are that diametric. I find that I don't disagree at all with the extended view you just expressed. What I contend is that the context matters. In the hyper self-focused context of psychotherapy, we can't deal as much with generalities or commonalities even though these are present all the time.

Psychotherapy is a space where we deal and travel in the fine marginal contours of individual differences using as our differentiating medium the shared tendencies or the human baseline. And by "radical individualism", strictly in the context of psychotherapy, I mean a process and method which has as its guiding principle the beliefs that (1) individual differences matter more than similarities (2) that an individual cannot be understood unless their points of uniqueness (as opposed to their normative qualities) are discovered (3) that the pathway to healing is individually charted (even though when correctly charted sometimes leads the client back to those collective practices and knowledge you rightly referred to). But it is the process to this destination rather than the destination itself that psychotherapy is concerned with and that process is radically individualistic.

Expand full comment

Machine learning could certainly find clusters, and associations with symptoms (if there are any symptoms separate from innate characteristics - it seems like the line between the two is very blurred), and classify people based on them (with relative probability) much more consistently than a psychiatrist. I think the fact all that inputs are so soft and squishy justifies the soft and squishy human element in making an assessment, but it does all end up seeming very arbitrary and hand-wavy.

Expand full comment

> This is my personality according to the Hexaco personality test.

Oh wow! I always knew you were much lower in Aesthetic Appreciation than I am, but I don't experience you as being as low in Agreeableness as that. Thinking back, I could tell I was definitely trying your patience early on. You're very good at gritting your teeth and not arguing with annoying people on your blog!

> When I presented this idea to Anders, he wasn't very impressed at all. What if people give different answers to standardized questions over time? he asked. Or in different cultures?

Except that self-other correlations are pretty good overall, and answers taken on different tests are similar.

Basically these tests are very much like thermometers with a mark at -100C and 100C, and no other marks on them. What if you put one thermometer in the sun, and another in the shade? What if you squint at the thermometer and think its 20C, and someone else thinks its 25C? "Oh no! Thermometers like that don't work, they are useless!" Sure, the thermometer will have trouble, but who cares? If it's below freezing, or above 40C, it works well enough that everyone can tell. Definitely its better than not having it at all.

I'll add that people at work all insist that I'm extremely Extraverted and high in Conscientiousness, much more than I rate myself (yes I've asked them to score me and yes they have obliged). And their descriptions of me are very consistent - I just got transferred to a new job location for a month, where people were very insistent on telling me "You are so positive, enthusiasic, and upbeat, you're so well organized" just like at my old job where I worked six years ago. Very likely this because work is just a place where the thermometer is in the sun. If you want enthusiastic, well-organized Apple Pie, you'll have to pay me for it.

Expand full comment

I have concluded that my aesthetic appreciation should be smaller than that of many other people, because some people say that they feel things when they look at abstract paintings and that is the point of it. I don't feel things from seeing colors and shapes (except for some aversion when it looks really, really bad). I just see things.

On some points I don't belong to the perfect target group for the Hexaco questions. In reality I'm both very cynical over people and still like them a lot (and act as if I do). That wasn't really an option for Hexaco, which only registered my cynism. Also, I can't be that super non-greedy, because then I wouldn't own anything and I do. I just don't want to live among rich people.

Expand full comment

Oddly enough, this is roughly what NIMH's RDoC framework (https://www.nimh.nih.gov/research/research-funded-by-nimh/rdoc/about-rdoc) is supposed to do: go back and start looking at psychological functioning using more basic/fundamental dimensions of human functioning (not necessarily clinical-specific categories/taxonomies).

I don't know that NIMH is pushing it as hard as they used to, but it appears they are still funding work in the area, which I think is good.

Expand full comment

An interesting measure you could capture this way is the difference between someone's self-report and how others perceive them. For example, I am sure my mother has no idea how often she comes across as hostile or angry to others.

Expand full comment

That would definitely be interesting. Although you would need a number of respondent to get a somewhat reliable picture of who is really the crazy one.

Expand full comment

Tricky for adults. With kids you could give separate questionaires to each parent and their teacher.

Expand full comment

Very neat! The basic idea is similar to the The Hierarchical Taxonomy of Psychopathology

(https://www.hitop-system.org/) — but your model is less convoluted and arguably more intellectually honest, given how little* we actually know about what lies behind these conditions. I can see why those trying to come up with dimensional alternatives to the DSM/ICD systems feel compelled to shoehorn the existing diagnostic labels into their models. But the high reliability of the existing diagnoses doesn't, I think, imply that they are going to fit neatly into a more valid alternative model.

*: To which many in the field would object that we actually know a whole lot — just look at the immense volume of research literature! But the ancients who thought Earth was the centre of the universe also had a lot of empirical knowledge of the heavens and constructed intellectually impressive models to explain it. I can imagine a venerable professor of the pre-Copernican age brushing off a young student's wondering if there might be simpler explanations: "We know this is all really complex and multi-factorial!"

Expand full comment

I'm very impressed with the idea - specifically mental disorders the way we do IQ and personality. In IQ research they find that the ability to many different types of puzzles are correlated. The more correlated with everything else a problem is the more "g loaded" it is. We could do the same with mental illness. It wouldn't even need to be academic experts doing it. All you would need is a statistically literate person who can convince a large number of people to do some surveys.

Expand full comment

Indeed there is some intriguing research on a "p factor" of psychopathology analogous to the "g factor" of IQ: https://pmc.ncbi.nlm.nih.gov/articles/PMC4209412/

Expand full comment

That is an interesting idea that I have never heard of before. It reminds me of a test that was put before me when I once applied for a job as a security guard. It was constructed to spot a person's weaknesses, from bad family relationships to hallucinations. I felt so offended by the test and the visibly bored man who was supervising it that I refused to give honest answers and thereby got no job. Possibly it was something like a P factor they were after.

Expand full comment

Yes. That kind of transparency really would be a revolution for psychiatry.

Expand full comment

Yours is a good suggestion, but seemingly premature (given it seems obvious but hasn't been adopted).

The generous explanation for "psychiatry’s taxonomy problem" is that it is far from a science. Are there many psychiatrists/psychologists who have the hubris to argue otherwise?

It is still struggling with exploring the 'landscape' of human nature and the instability in taxonomy is a product of varying attempts to 'slice and dice' the data from which to construct simplifying descriptive categorical models. Not being able to consistently identify and describe things (say the human equivalent of mountains, rivers and plains) makes it essentially impossible to craft scientific explanatory models of such.

I say "generous" because many could argue that presenting psychiatry as 'scientific' is really an exercise in relative professional status raising independent of identified efficacy of the profession's outputs (changed mental state of client).

Expand full comment

Data don't need to be perfect to be worth collecting. Systems for data collection and models for data interpretation will be developed and improved over time. Then systems for translating between old and new systems will be needed. Perfect? No. But at least better than attempting to translate between the 1994 concept of autism and the 2024 concept of autism.

Expand full comment

In more general terms I've been thinking about the effort we do to split/lump our world, and the preferences we are inclined to do (to split rather than lump & vice-versa). I.E. when we (and we each) prefer to split or lump. My suggestion is we consider our labours to do either/both and unite them in a circles as you have done ,i.e. split more in a lumpy cognisance.

Expand full comment

Something that bothers me about psychiatry - among very, very many things that bother me about psychiatry - is that psychiatrists group brain malfunction (schizophrenia, postpartum psychosis) and moral malfunction (borderline personality disorder) together under the heading of "mental illness." These don't belong together. Furthermore, treating them as though they were the same type of thing removes the possibility of moral reform for the latter.

Expand full comment

and the recognition that reform of narcissists is not possible within a human lifetime (that's part of the definition for ) and reform of psychopaths is not possible within the lifespan of the universe

Expand full comment

I agree with you. As a Christian, I also believe that when the lifespan of the universe is played out, everything will be remade (reformed). It's really the only hope for the psychopath, et. al.

Expand full comment

one doesn't have to be a anything to agree with you there :)

Expand full comment