[Up] [Map] [No prior] [Robot Wisdom home page]
The term "theory" is bandied about in ways that I personally find shocking. In the social sciences (not to speak of "literary theory," etc.), there is very little that merits the term, in my opinion. That's not a criticism of the fields; rather, of those in them who like to inflate their self-images by posturing about their "theories," which usually turn out on inspection to be either truisms or false or worse...
-Noam Chomsky
Here's a series of thought-experiments about the future of psychology, originally posted to comp.ai and several other newsgroups in June 1995. It meanders quite a bit, bringing up lots of points that I hope deserve debate.
I want to begin by trying to view the puzzle of human psychology from the broad perspective of evolution:
It's generally agreed that (in theory) we could precisely trace our family tree all the way back to the protozoa of the Precambrian. And we might, in theory, build an unabridged database that records the birth of each of our direct ancestors, starting way back when, in exact sequence.
(This would be at least a trillion individuals, or a million closely printed volumes. For any given time, it should include only members of a single species-- our 'parent species' in that age. Punctuated equilibrium theory suggests there will be infrequent bottlenecks when the parent population is reduced to a handful of individuals, and that virtually all evolution happens during these periods.)
Some of these database entries we might mark with a star, to indicate that that individual was the first to show a new behavior, in a way that was genetically linked, and favored by natural selection (and so passed on the its descendents). So we may imagine a companion database that traces these innovative adaptive behavior patterns, along the path to humanity. (We'll ignore quantitative changes, and focus only on qualitative differences.) So you might find:
1,000,000,000 BP: Increased movement in presence of pH above some level. 100,000,000 BP: Making noise with lungs to signal agitation to others. 1,000,000 BP: Using sharp-edged rocks to cut wood. etc.
And imagine going thru this timeline and tagging with a star all those behaviors that humans still exhibit. How many might this be? Shouldn't it be comparable (at least within a few orders of magnitude) to the number of active genes in the human genome? (Did I hear a mere 100,000 for this figure???)
Getting a sense of the contents and scale of this list is essential for setting the research agenda of the social sciences. The temptation will be great to throw up one's hands, imagining that it must consist of infinite numbers of trivialities, but imagine a single volume psych text with the 10,000 most important behavioral innovations-- how much of human behavior could just these 10,000 cover? What categories might they be divided into?
One way to categorize behaviors is by the psychological motivations they satisfy. Safety and hunger behaviors will account for many of these, but if you think about the sorts of ways animals cope with safety and hunger, the basics of these motives might be boiled down to a few pages of basic patterns, like: sit-real-still-and-wait-for-prey-to-wander-by. Sexual and social behaviors will account for many more, but again some overview of the basics ought to be expressible in a few dozen pages. So, beyond these, what other categories will be needed? How much of a 'dent' will these few dozen pages make?
Now imagine the intelligent computer of the future, and the (Cyc-like) knowledgebank such an AI would require. (Cyc is a 'super-expert system' being built in Austin by Doug Lenat and others. A FAQ is available at this site.) Cyc is expected to require tens of millions of rules, although a stripped-down kernel of the most important among these is expected to be much smaller. So how much overlap must there be between the essentials of our behavioral timeline, and this stripped-down Cyc knowledgebank? Would a 10,000-rule kernel need a very different emphasis, or are the same issues relevant for each?
Now imagine a project to review and digest the entire literature of the academic social sciences in the 20th century-- behaviorism, cognitive psych, social psych, etc. What sort of list would this produce? How much overlap would there be between such a digest, and our evolutionary database of behavior, or our knowledgebase for Cyc?
My contention is that 20th C social science is so out of touch that a full summary of its contributions might be on the order of a hundred rules! And this is due to a completely false and destructive model of how the social sciences must proceed...
So what has 20th C social science achieved? How much clearer is our understanding of human actions today, than it was in 1901? How much more sensitive have we become, to what makes people tick?
Were there any new insights in this century, that are helping make your daily relationships run more smoothly? Can you list them? If so, how many can you name? If none... why not???
I'm arguing that social science in this century has failed because it's been chasing a phantasm, and that this phantasm resulted from mis- mapping the methodology of the physical sciences onto the human domain.
But I'm not saying that no proper mapping will be possible, or that we can understand the brain without the scientific method! These false conventions can and should be unseated (no mean feat), and replaced with an approach that restores the process of knowledge accumulation, leading to a computer simulation of the human personality...
The basic model of scientific method is just this: hypothesis, then experiment. And this is mirrored in the basic structure of journal articles: this was our hypothesis, this is how we tested it, this was our result, and these are our conclusions.
In criticising such reports, one can argue that the results don't justify the conclusions, or that the experiment doesn't always produce those results, or that the experiment doesn't even test that hypothesis.
But one can also object that the hypothesis itself is ill-formulated.
And I fear this is so rarely a problem in the physical sciences, that
it's been overlooked, for the social sciences, as the most salient,
universal critique...
Because, in the social sciences, the object under study (the human brain) is so much more complex than other sorts of physical system, that no well-defined hypothesis has yet been proposed! How can you formulate a hypothesis about a system before you've begun to isolate its subsystems, or identify its subelements, or enumerate its inputs and outputs? How can you begin to take quantitative measurements before you know the qualitative dimensions of the phenomenon you're measuring?
With a physical system, you can physically isolate it and introduce changes discretely, but physically isolating a human doesn't simplify their brainstate to some 'zero' level! Stripping their experience of a its natural motivational context does not reveal any simpler substructures-- the structures of the brain evolved for coping with motivating contexts, and to understand its evolution we have to watch its behaviors in those contexts..
So what we have to do, simply enough, is watch our brains in action, and even more, play with them, testing their limits in every possible way! And this is our solemn duty as scientists-- to abandon all preconceptions, and try to discover the full range of things this complex system can do. And this is, in fact, exactly what's done in the natural sciences, at the comparable stage of understanding-- but this is not normally written up in the publications!
So I'm suggesting that to get psychology moving, one simple thing that's needed is a forum for sharing our experiences, observing (and experimenting with!) our brains. And there's no need to enforce any sort of rigor on this project-- occasional flawed reports are much less problematic than the risk of overlooking vast ranges of interesting psychological nuances.
And the language we ought to use for this is plain old ordinary language... not jargon. In fact, we need to acknowledge that literature possesses longstanding, finely-honed skills at using language to depict our brains at work ...and in fact already offers millions of pages of such observations, imaginatively rewoven into novels.
And I'm not saying we should be looking to novels for laws of behavior. All we'll be looking for is that first hypothesis-- a model of human actions, detailed enough to begin to test. So we're not reading novels and saying, things are always this way. We're saying, sometimes they might be this way, and our model must be flexible enough to account for a broad range of such plausible nuances...
And when you begin to accumulate an inventory of all the ways humans might act, you can begin to look for the larger scale patterns, and for the subelements and subsystems, the inputs and outputs.
And one way to pursue this should be to collect these descriptions, in some physical form like index cards or database records, and sort them into groups according to the elements they may contain.
And I suggest that for human behavior, during this early stage, one of the most useful ways of sorting these will be according to the motivations depicted-- food, safety, sex, esteem, etc. These should be chapter headings in our inventory/ textbook/ database.
And then, when we think we've gotten the broad shape right, we can start to look for simplifications that can be built into a first, self- contained computer simulation. Which ought, at the first, to present a simulated creature in a simulated field, satisfying hunger needs, and reproducing sexually.
And the hunger part of this is easy, and has been done many times (even in text adventure games). But anything beyond the simplest model of sexual reproduction will have to include sexual selection, which means judging the fitness of potential mates-- and this, I think, is where the matter has stalled...
In the first message in this series, trying to gain some perspective on the evolution of human behavior, I wrote:
>imagine a single volume psych text with the 10,000 most important >behavioral innovations. [...] One way to categorize behaviors is by >the psychological motivations they satisfy.
And some weeks earlier I'd posted a version of this outline for the psychology text of the future, trying to analyse one motive in each chapter, inventorying all the sorts of behavior humans show, involving that motive:
Hunger: growing food (what 'psychological adjustments' does this require), acquiring food (incl. hunting and gathering), competing for food, feeling hunger (the phenomenology), specificity of hungers, preparing food (what makes it appealing), consuming (rituals, taboos, etc), food-giving/sharing, excreting, fasting (what motivates, how it effects), starving (how we react, desperate action), overeating (why, results, solutions), food-poisoning (psychological effects, avoidance)
(A lot of this has been explored, I think-- mostly by anthropologists, though, so almost none of it shows up in intro psych texts! Novels should add tremendous amounts of detail, too.)
Within the inventory of behaviors, for each motive, we may distinguish other abstract classes of relationship, like:
developmental origins of motive; unsatisfied needs; objects and tools; abilities, actions; inability, inaction; satisfaction; taking more, consequences; taking less, consequences; helping others, obstructing others; competing needs, competing people, substitute needs
Returning to the other basic motives, keep these subclasses in mind, and also imagine what literature 'knows' on these topics, vs what the social sciences 'know':
Safety: parental care, infantile traume; feeling endangered/safe (phenomenology of), establishing and maintaining relationships with protectors; establishing and maintaining barriers to attackers (walls, distance); mounting counterthreats, fighting off attacks; repairing damage; neurotic anxiety, neurotic risk-taking, risk-taking as entertainment
Sex: infantile sexuality, adolescent sexuality; sexual desire; sexual skills; impotence; establishing and maintaining relationships with sex partners; jealousy and faithfulness; engaging partner; sexual responsibility; solitary sex; coerced sex, paid sex; taboo desires, guilt
Affection: family roots; loneliness, unlovability; establishing and maintaining affectionate relationships; expressions of disaffection, competition, conciliation; manipulated affection; leaving the nest; abandonment, bereavement
Esteem: family roots; low self-esteem; skills for maintaining esteem, execution and critique, attack and praise; vanity, coerced esteem, craving attention, showing off, bragging, humility; manipulating others via their vanity
Entertainment: boredom, risk-taking, play, story-telling, spectacles, contests, travel, music, preferences among entertainments; the role of entertainer, creating entertainments, performing
Expression: learning expression; the craving to communicate; media, shaping an expression; self-criticism and revision, publication, performance; interviewing; expression and esteem; pathologies
These classes and subclasses are acknowledged to be incomplete and partly redundant. But I believe these are close to being regularizable into a computer representation, using a very simple scheme based on the five categories: person, place, thing, motive, and modality.
All 'motive stories' must involve (at least) one person and one motive. Stories about objects and tools will involve "person, motive, thing". Competing-motives stories involve "person motive motive". And subcategories like helping-others, obstructing-others, and competing- people are distinguished by the pattern "person person motive".
Furthermore, I've proposed that these five categories have a handful of typical pairwise relationships. In particular, the relations between a person and a motive include:
person suffers motive person gratifies motive person indulges motive person denies motive
So it looks to me like this inventory of human behavior can already be maintained very precisely... which should imply that it can be animated on a computer with minimal difficulty...???
[The previous three sections of this argument I wrote offline, but this fourth one will never get written if I aim for polish, so I'm just going to improvise some thoughts...]
If you take the view that psychology requires self-knowledge (and what sane person could argue the opposite?!?), then a course of preparation for the human sciences ought to suggest certain strategies, beyond 'book learning', to maximize self-discovery...
But none of these are going to be compatible with grad school!!!
Because the first thing, for self-knowledge, is to be yourself-- you can't learn what you are if you're busy trying to conform to some image...
And this implies, too, being completely honest with yourself. Which, in my experience, usually means you have to be honest with others as well... which (in my experience) is exactly contrary to the sort of ego-games that grad school demands...
And, in the US of the 1990s, it also seems to put you at odds with society in general-- the pressure to conform seems in many ways to have grown enormously in the last few decades...
One's life has to become one's experiment, and a high level of personal detachment has to be cultivated, even as your natural emotional reactions are encouraged towards spontaneous expression. (This paradox of detached emotionalism will be the source of endless misunderstandings, I predict! ;^/
And it requires a curious sort of faith, to allow one's own being to dictate one's actions, when society is demanding conformity. One has to trust that our evolutionary heritage is wiser than society, and allow it to reveal its own wisdom no matter what convention claims...
So I'd suggest that the people, right now, who are most qualified to do human sciences must be people outside the 'system'-- and there will need to be some effort to appeal to them, and to create a new system that allows them the freedom to be at odds with convention, while still working together as a scientific community...
Another sort of skill one must cultivate is psychologically expressive language. Novels and even poems display an uncanny subtlety in describing mental states, and this 'wisdom of the words' is a priceless advantage for psychological self-exploration...
The skill of meditation is also very useful, encouraging one to watch the movements of the mind in a detached fashion.
And it's important to practice some creative art, to maintain a connection with the 'naive witness' within-- the key to art is learning to see one's creations as others see them, and if you lose touch with this, you've lost touch with yourself...
Newsgroups: rec.arts.books,comp.ai,sci.cognitive,sci.psychology.theory Subject: Quantifying literary progress (Was: Top 100 books list?) Date: 19 Aug 1995 13:12:43 -0500
Over on rec.arts.books, I've gotten into a minor flamewar about whether or not the history of literature demonstrates progress of expressiveness...
The debate on r.a.b is under the subjectline "The 100 best books?" because, in response to that query, I posted my list of 50 favorite authors.
And my list is unapologetically weak on the classics, which drew criticism on r.a.b, and I replied that I think modern writers have a great advantage over the classics, because of an expanded toolkit of themes, styles, and techniques...
But I'm quite surprised at how vigorously the opponents of this view are resisting it there, because I thought in the literary world this truth was generally understood.
I'm noticing now that their resistance has all the hallmarks of psychological denial... so I'm inclined to connect it back to the whole "Fiasco of Social Science" argument I posted a few months back. (Hence the broader crosspost.)
Here's the gist:
In article <d-turnb-180895125514@obiwan.ccsm.uiuc.edu>; on r.a.b, Doug Turnbull <d-turnb@students.uiuc.edu>; sigged: >"Every notion of progress is refuted by the existence of the Iliad. The >perfection of the first step makes any idea of progressive ascension >ridiculous." >----- R. Calasso
I assert that this view is a sentimental as claiming that the Gospels supply a complete blueprint for human ethics!
The authors I list as my favorites are my favorites because they offered new detailed observations of human psychology, that were instantly recognizable and convincing, but that I'd never seen mentioned before, so that now for the first time when similar events occurred I could say, "Yes, Iris Murdoch wrote about this in Nuns and Soldiers"...
And I view the history of literature as a quantifiable, growing inventory of behavioral descriptions, so that it will eventually become demonstrably false to claim that Homer or Shakespeare said all that needs to be said. The human behavioral repertoire is enormous and subtle, and our human-social denial system makes certain areas (especially the sexual) impossible to view with detachment... so literature has a long, long way to go, to capture a full portrait of humanity...
I wrote to Iris Murdoch once, asking her permission to compile a 'textbook of literary psychology' based entirely on her works, that would simply inventory every nuance in her works, chopping out all except these psychological descriptions, and re-sorting them into an overall conceptual scheme...
(Happily, she thought the idea made perfect sense!)
I haven't undertaken that task (though it would be large, it would also be very enjoyable), but I have done a comparable analysis of love poems and the literature of love. It follows the chronological course of EveryMan's EveryLoveAffair, from loneliness to courtship to rejection or acceptance, difficulties of relationship, and separation or bereavement. There are some 500 categories of description, with an average of 4 or 5 quotes for each nuance... [sample]
In the 'Fiasco' thread I argued that these nuances might be viewed from an evolutionary perspective, a giant index of the mutations that made the new behaviors possible, reconstructed (theoretically) as a timeline of our species' evolution...
Someone else in the r.a.b thread asked for particular examples of literary progess-- one that comes to mind is Robert Stone. I read about a panel at Northwestern where he was challenged about his depiction of human depravity in "Dog Soldiers" and "A Flag for Sunrise". The journalist described his demeanor as old-testament-prophet-like as he proclaimed that his duty, and every writer's duty, is to portray the world as it is without hypocrisy or sentimentalism...
[Up] [Map] [No next] [Robot Wisdom home page] (Feedback)