If reductionism is the hammer, what nails are out there?
EDIT: I’m moving this to the Discussion section because people seem to not like it (lack of upvotes) and to find the writing unclear. I’d love writing advice, if anyone wants to offer some.
Related to: Dissolving the question, Explaining vs explaining away
I review parts of the reductionism sequence, in hopes of setting up for future reduction work.
So, you’ve been building up your reductionism muscles, on LW or elsewhere. You’re no longer confused about a magical essence of free will; you understand how particular arrangements of atoms can make choices. You’re no longer confused about a magical essence of personal identity; you understand where the feeling of “you” comes from, and how one could in principle create many copies of you that continued “your” experiences, and how the absence of an irreducible essence doesn’t reduce life’s meaning.
The natural next question is: what other phenomena can you reduce? What topics are we currently confused about which may yield to the same tools? And what tools, exactly, do we have for such reductions?
With the goal of paving the way for new reductions, then, let’s make a list of questions that persistently felt like questions about magical essences, including both questions that have been solved, and questions about which we are currently confused. And let’s also list tools or strategies that assisted in their dissolution. I made an attempt at such lists below; perhaps you can help me refine them?
Some places where many expected a fundamental or non-reducible “essence”[1]:
1. Ducks.
Why it’s tempting to postulate an essence: Organisms seem to come in types. New organisms of the given type (e.g., new ducks) come into existence, almost as though the the type “Duck” had causal power. Humans are able to form mental concepts of “duck” that approximately mirror the outside predictive regularities.
2. Life.
Why it’s tempting to postulate an essence: Living creatures act very differently from dead creatures. A recently killed animal doesn’t move, loses its body heat, etc., even though its matter is in almost the same configuration. [2]
3. Free will.
Why it’s tempting to postulate an essence: Humans (among other things) are in fact organized to “choose” their actions in some meaningful sense. We (mostly) choose a single course of action in a relatively unified manner that responds to outside information and incentives. “Choice” also seems like a useful internal concept, but I’m not sure how to describe the details here.
4. Personal identity
Why it’s tempting to postulate an essence: People have personalities, plans, beliefs, bodies, etc. that approximately persist over time. Internally, we experience consistent memories that happened “to us”, we choose our own actions, and we anticipate future experiences.
5. Pain
Why it’s tempting to postulate an essence: We feel pain. We find ourselves motivated to avoid pain. We sometimes almost feel others’ pain, as when we wince and rub our thumbs after watching someone else smash their thumb with a hammer, and we often find ourselves motivated to avoid their pain as well. We can report verbally on the pain, modify our behavior to reduce the pain, etc.
6. Mathematics
Why it’s tempting to postulate an essence: Mathematics often pops up in science. It’s also “simple”, is at least somewhat culturally universal, is relatively easy to implement portions of in machines, has a nice notion of “proof” whereby we can often formally determine what is true, and can often determine true results without much contact with outside empirical data, and is something aliens might plausibly share with us.
7. Reality / existence / the physical world
Why it’s tempting to postulate an essence: Our perceptions are well predicted by imagining a set of fairly stable objects that we can see, touch, turn over in our hands, etc. and that retain their color, shape, heft, and other properties fairly stably over time. At higher levels of abstraction, too, the world is fairly lawful and coherent. [3]
What lessons for future reductions?
These examples suggest the following heuristics:
A. Even when it really, really feels like there should be an essence, there probably isn’t one.
B. Philosophical questions are just ordinary questions that one is particularly ignorant about; they are not questions about separate magisteria that must permanently be reasoned about in some special way.
C. People expect magical essences in places where there really are interesting empirical regularities. In order to understand those regularities, and to create a new set of concepts that better do the work that our old magical essences intuitions used to do, it is necessary to do real research. Ritual assertions that “It’s all physics” and “there aren’t essences” do not create the needed concepts and anticipations.
D. A reasonable first step, in tackling a new confusion, is to ask the why it feels like there is a question or concept there, and to list the empirical regularities, or cognitive artifacts, that contribute to that feeling.
These heuristics aren’t original; Eliezer noted them already in his reductionism sequence (which is very much worth reading). But I suspect that many apply these heuristics more to problems they already understand (“of course free will has no magical essence”) more than to problems we don’t yet understand (“of course there is no magical essence that distinguishes our actual, real world from imaginable physicses/worlds that aren’t real”).
I’m hoping that reviewing heuristics for reduction, and staring at solved and unsolved problems side by side, may help us with the unsolved problems (which I’ll attempt some steps toward in subsequent posts).
[1] I agree with SarahC’s point that humans seem predisposed to impute essences everywhere. Still, discussions about whether there’s a magical essence “free will” seem to pop up more often than discussions about whether there’s a magical essence “carpet”, “ocean”, “California”, or “female”. I mean, folks are interested in these other questions, and they have discussions about what meaning to use and how much that meaning cleaves nature at its joints, but they don’t generally expect a separate sort of essence that has causal powers and isn’t made out of atoms.
[2] People unacquainted with modern biology seem often to make predictive errors due to expecting an essence of life. For example, I had lunch the other day with a physics professor from a good university who thought that, even if we could assemble an atom-for-atom duplicate of a person’s exact physical state, it might well not act like a person for want of a soul. Another acquaintance was surprised to hear that scientists do in fact believe that a cell assembled in a test tube would act just like a cell assembled anywhere else.
[3] I’m less satisfied with this unpacking than with the others on the list. Can anyone do better?
Gender/sexuality. People really want essences here. Some people are still stuck on “male essence / female essence”, some manage to get as far as “attracted to male essence / female essence”. Reductionism could (dis)solve this issue.
I would say that it is not that we want essences in our sexuality, but that gender and sexuality are essentialist by nature: the sexual drive is built on top of the parts of our brains that essentialize/abstract/encapsulate, and so reducing the concept would involve modifying the human utility function to desire the parts, rather than the pretended whole.
Or, to put it another way: a heterosexual blegg is not 50% attracted to something with 50% blegg features and 50% rube features; it is attracted only to pure rubes, and the closer something is to being a rube, without exactly being a rube, the less attractive it is. This is basically the Uncanny Valley at work: some of our drives want discrete gestalts, and the harder they have to work to construct them, the less favorably they’ll evaluate the things they’re constructing on.
I have been vaguely thinking of writing a top level article about psychological essentialism for a while, but this seems like a good place to just point to it. There is a large body of research on the subject and the best summary of it I know is Susan Gelman’s “Essentialism in Everyday Thought”.
The process of imputing causation to visible characteristics from an unseen inner variable seems to be something humans subconsciously and automatically do in some situations. You can do studies of essentialist reasoning in children by asking “Suppose an X was raised by Ys, in case C would the X tend to do this X-ish things or this Y-ish things?” Very young children will give different answers for difference values of C, and in some cases (like with human language) they are clearly wrong which makes it seem like maybe they are executing a domain specific “complex essentialist cognitive module” rather than reasoning from observed evidence.
If you want to see some back and forth on the subject instead of just taking one author’s word for it, here is a 1993 criticism of the essentialist research program by Susan Jones and Linda Smith and a rebuttal by Susan Gelman.
My impression is that essentialist cognition is enormously common because it is enormously useful. There is an essence for ducks and its called the duck genome. Hypothesized essences have been mapped to one or more physical mechanisms so often that mostly I just assume that this can be done for all essences that are coherent. The mechanisms generally have approximately the pragmatic properties normally ascribed to essences (like immutability) but it doesn’t match human intuitions perfectly because, whether its a democracy or an ecology or a painting “it’s just an arrangement of atoms, right?”
After I internalized this view of the world, I gained a measure of sympathy for people who believed in supernatural theories because those situations became intelligible to me as people having working models for reality using the built in tools of human reasoning. “Evil ghosts in the swamp” model “malaria” well enough to explain why to avoid swamps while looking for food. “God” models “inscrutable optimization processes stronger than me” enough to help people make a kind of peace with the fact of limited power and explain the appearance of order in the world while looking for food.
When I need to “argue past the essence” without giving the whole cogsci backstory I try to point out the “magic sparks” buried in the theory and then denaturalize the sparks by hypothesizing vaguely realistic complex mechanisms in their place. Like sometimes people deploy the concept of IQ as working like a magic spark that is more or less bright, and you can denaturalize it by breaking it into pieces that can obviously vary independently like working memory capacity, vocabulary acquisition, “good nutrition, good sleep, and the time to cultivate your mind”, and so on.
But most of the time, a theory with a spark is perfectly good for making the predictions you need to avoid scary situations and find food. I would expect an AGI worth its salt to have some internal mechanism that functions roughly to “notice the sparks” and use them as appropriate and push deeper as dictated by opportunity and need :-)
This sounds very useful. Please post on this.
I would still encourage you to adapt this into a top level post :). (Easy for me to say . . .)
Essentialism also seems very prominent in human pleasure, per “how pleasure works” book.
As far as supernatural theories, I am in an interesting position as far as taking some meditation training from a martial arts teacher; on one hand I (surely hope!) am not “aligning meridians of the body” while doing the breathing exercises, on the other hand I don’t want to dismiss this incorrect “model” too early as it might be of further usefulness, given that these systems tend to be very ancient.
...a PHYSICS PROFESSOR??
That makes my head hurt.
And that makes me realize that, after all this time reading LW and OB, I must still be overestimating academia. Consider me updated.
The physics professors who taught me quantum mechanics seemed to believe there is something special making up the human mind that, unlike all other matter in the universe, cannot be put int a superposition, and thus causes the wave function to collapse to one of its component eigen states. I won’t say that it doesn’t bother me that some of them believe in souls, but I’ve gotten used to it.
The physics professor that taught me quantum mechanics not only thought ‘collapse’ was some wonderfully mysterious phenomenon, the fact that the most probable location to find an electron can be at r = 0, while the most probable radius is non zero was wonderfully mysterious to him.
It’s important to understand that there’s likely strong compartmentalization going on here. Having such beliefs is not mutually exclusive with being a strong scientific researcher. Such indications should not be read as overly strong indications of low levels of general competence.
If you’re doing the reductionism right there is probably only one nail out there! ;)
Er, I mean: what subjects are we currently confused about, in such a manner that we could become unconfused using the tools of reductionism?
I think that was intended as a joke, Anna!
Indeed. I can’t fault you for using a catchy headline. :)
Hunger. People go on diets, notice that the diets leave them hungry, and fail to take the next step and determine which of the body’s half-dozen hunger mechanisms is being triggered first.
Karl Smith of the blog Modeled Behavior frequently talks about body weight in a reductionist way (examples 1, 2, 3), without claiming to know how to fix obesity problems.
Anticipated experience
What about ethics? It seems that many people think there is some ‘moral bedrock’ somewhere—but is there really such a thing?
To me it seems that ethical questions are really about the tension between our knee-jerk moral intuitions and ethical frameworks (utilitarianism, deontology etc.). Increasingly elaborate theories are built out of the urge to somehow make our ‘moral compass’ seem logical, until someone comes up with some clever example where the theory somehow conflicts with our intuitions...
I know moral relativism is not universally popular, but can reductionism/rationalism lead to anything else?
That depends a lot on what you mean by “moral relativism”. Certainly rationality and reductionism need not imply taking morality less seriously. I liked what Eliezer’s Harry Potter had to say on the subject:
If you haven’t looked at it already, you might like Eliezer’s sequence on metaethics, which talks about how one can notice that our concerns are generated by our brains, and that one could design brains with different concerns, while still taking morality seriously.
I’m glad someone noticed that. It was NOT EASY to compress that entire metaethical debate down into two paragraphs of text that wouldn’t distract from the main story.
It sounds a lot as if you are suggesting that there is some essence to morality which transcends “concerns are generated by our brains”.
“still taking morality seriously” modifies “one [who can...]”, not “brains with different concerns”.
I’m not sure why you came to think there was some confusion on this point, so I will not presume to suggest where you went wrong in your reading.
One attractor in the space of moral systems that doesn’t have much to do with what could be engineered into brains is the class of moral systems that are favoured by natural selection.
There’s no such essence in a classical philosophical sense in naturalistic metaethics. Rather, the transcendence comes from morality-as-computation’s ability to be abstracted away from brains, like arithmetic can be abstracted away from calculators. “Concerns generated by our brains”, for example, breaks down when we try to instill morality into a FAI that might be profoundly non-human and non-person.
I read some of it, and after you mentioning it, I read some more. E.g. The Bedrock of Fairness touches on the issue of whether or there is this moral ‘essence’. Also, I liked Paul Graham’s What you can’s say, which discusses the way morals change.
Overall, I think the closest thing that comes to a ‘moral essence’ is that the set of moral intuitions (no matter how vaguely defined) is the best thing that evolutionary processes have been able to come up with. Hume’s is-ought problem does not really apply because there is no real ought.
The set of morals we ended up with is probably best summarized with the Golden Rule, which is a useful illusion in the same way that free will is, and similarly, for all practical purpose we can treat it as if it were real.
[ It’s an interesting though experiment to consider whether there could be other, radically different sets of morals that would lead to the same or better evolutionary fitness, while still being ‘evolutionary feasible’. ]
No. Next? :P
(Ok, to be fair ethics is actually a really good example. Quite possibly the best example, given that most of the other critical things are approximately reduced already. Just not that particular ethical question.)
Apply reductionism to itself.
Or to “rationalism”.
Or, to pick one currently popular example, to “blackmail”.
It seems to me that people often struggle to come up with a technical definition of some word which captures the “essence” of a concept. One particular example which I have some experience with is the definition of “life”. This activity can generate considerable emotion, and I don’t think that the reductionist explanation of “natural kinds” quite applies to this kind of dispute.
Maybe not reductionism vs essentialism in quite the way that Anna intends. But close enough to create confusion. In fact, I might advise Anna to attempt a taxonomy of different kinds of “essence” and different kinds of “reduction” so as to dispel some of the confusion.
You’re right—ethics should be on the list.
I’d had it there originally, and had then removed it on the theory peoples’ persistent tendency to postulate an essential and irreducible ethics had more to do with folks having strong and non-truth-seeking motives on the subject than with empirical regularities of a sort that essences could help predict.
But on reflection, one’s goals are confusingly different from other sorts of phenomena, so maybe even without strong emotions folks would expect magical essences here.
Which list? The list of things successfully reduced, or the list of candidates for reduction that you are asking us to help you build?
You wondered why people seem to be confused by this posting. I think it is because there are two lists being discussed here, and you have been extremely unclear in your transitions in distinguishing them.
Or maybe it is just me.
Essentialism seems to me to be more general than anti-reductionism. One can be a reductionist essentialist.
I take essentialism to be the view that there is a privileged classification scheme that applies to all things in the world, and that the position that a thing T occupies in this scheme is entirely determined by whether T possesses a certain package of essential qualities. All other qualities of T are accidental.
Thus, to determine whether T is an X, you just need to run down a checklist of qualities x_1, . . ., x_n and see whether T has all of them. The classification scheme is unambiguous; if T doesn’t have one of the x_i’s, then T is an altogether different kind of thing.
Furthermore, these essential qualities aren’t exceedingly exotic. Having every arrangement of atoms be a different essential quality wouldn’t count. The classification scheme that the essential qualities induce is well-approximated by the classification scheme that we use in everyday life. For example, the true classification scheme should include categories that closely align with our concepts of living vs. nonliving and male vs. female.
In practice, this means that things are best thought of as a bundle of essential qualities, plus a diff of accidental qualities. Good practical and ethical reasoning focuses far more on the essential qualities than on the accidental ones.
But taking this view doesn’t rule out being a reductionist. You could be an essentialist while still believing that to have essential qualities x_1, . . ., x_n means to have the arrangement of your atoms be in a certain set of possible arrangements. You would just hold that, for whatever reason, the possible arrangements of atoms can be classified with a scheme of the above sort.
Eliezer’s sequence on words is a good antidote to this kind of thinking.
Two different schools of “outsider science” leap to mind.
The first is Ayn Rand’s theory of concepts from her attempt to construct an epistemology to theoretical ground Objectivist arguments as being “about physical reality”. I’ve never actually read her work here, but Adam Reed has a summary of Rand’s theory of concepts which sounds very much like your proposed “reductionist essentialism”.
The second is something I discovered a few days ago via LW, which is the “Baraminology Study Group” which appears to be an attempt to figure out a classification scheme for animals that is both (1) physically grounded and (2) consistent with the young earth creationist assertion that breeding pairs of each “kind” (or “baramin” in their jargon) could all have fit on Noah’s Arc and subsequently diverged via “microevolution” into the animals we see today.
This almost makes me wonder if there might be as-yet-undetected human variation on this dimension the way people’s visualization abilities have substantial variation?
Perhaps hardcore realists lack any cognitive machinery for managing concepts other than cognitive essentialist machinery, or their essentialist module is turned up to 11? Or perhaps nominalists have somewhat broken cognitive essentialist machinery and have to fall back on more flexible but more expensive “general reasoning” faculties?
And now that I spell this out, I think it may offer an actually testable mechanistic hypothesis that could be used as the basis of a coherent experiment in psychoceramics! :-D
If I try to construct a “sociologically nonjudgmental” theory I think it would go something like this: “There is variation in human essentialist reasoning tendencies [from some source] though different study communities tend to be relatively homogeneous in this respect. When one or more people attempt to study a domain bringing radically different essentialist reasoning tendencies to bear on the subject their theories are almost necessarily incommensurable and communicative isolation is the result. Any sufficiently-low-status ‘communicatively isolated’ theoretical community is liable to be labeled ‘crackpot’, will be out-competed for grants and tenure, and tend to lack the time and money to socially enforce epistemic rigor.”
The theory’s testable hypothesis is that a psychological instrument to assess a person’s use and orientation towards cognitive essentialism would show large differences between such groups as:
Objectivists as a “cult” relative to mainstream 1960′s philosophers.
Baraminologists as a “cult” relative to mainstream academic biologists.
The anti science faction in the American Anthropological Association as a traditionalist schismatic faction in Anthropology.
And perhaps in the reverse direction (with a novel tendency successfully invading an existing discipline), the takeover of political science by behaviorists in the 1950′s and rational choice theorists in the 1980′s.
This needs to be a post.
Another large category of essences: Perception of all sorts, not just pain. Computer scientists have found out the hard way that vision doesn’t just happen.
(7) is a hard one. I think it was G.K. Chesterton [1] who said that seeing through everything is indistinguishable from being blind.
You’re using the standard of the universe as scientifically postulated (interacting matter/energy which can be understood through observation, experiment, and logic) as the tool for dissolving essences.
I suppose the risk is thinking that the universe assuredly is as currently conceived, and may not be a great deal weirder than that. On the other hand, unless it’s got a whole other layer of currently unimaginable weirdness, we’re still limited [2] to observation, experiment, and logic in investigating it.
[1] I realize that citing Chesterton makes the statement less credible, if anything, but I think it’s a legitimate point as well as clever, and I″m giving credit where it’s due.
[2] I’m inclined to think that if there is a great increase in processing capacity, there will be emergent qualities as unimaginable to us-now as formal logic is to an earthworm, but there’s no point in putting much thought into what the details might be.
When you make an effort to minimize or eliminate your belief in essences, how does it affect your life?
Talent/Intelligence
Why it’s tempting to postulate an essence: people seem to be good at different things, and for any particular thing, some people are better than others; and since we can’t read others’ minds, and people aren’t always skilled in introspection and explanation, it’s not always easy to find out what they know that we don’t.
Aesthetics/”Taste”
Why it’s tempting to postulate an essence: people have different “tastes”, and it’s difficult to argue them into changing their mind about what they like. Plus, people aren’t good at introspecting to discover via analysis why it is they like what they like, and are motivated not to do so because they’re afraid they won’t like it anymore (similar to the way people fear physicists’ reduction of rainbows). Plus, shared tastes are a common criterion used to divide people into tribes.
Music Theory
Why it’s tempting to postulate essences: certain periods/styles exhibit empirical regularities not shared by other periods/styles, and cataloging these regularities is felt to be most of what the subject matter consists of. The discipline doesn’t have a culture of reductionism, perhaps because irreducible “talent” (see above) is already assumed among those studying it, so there’s no need to systematically address the question of how to make music from non-music, or how to invent new styles. Also, people can sense that treating such questions systematically would involve difficult new abstractions, in contrast to the concrete concepts used to do cataloging, and they run away in fear.
People also seem to believe that genres have essences. And countries. And religions.
There are no magical essences, but there are many unexplained things. I’ll go so far as to claim that if you have really internalized LW’s reductionism lessons, you should be seeing mysteries everywhere you look!
It is one thing to say “duh, this thing is built out of these tiny blocks”. Feynman gives an example in his lectures (I have no idea if it’s still true today, or if it just reflects Feynman’s state of knowledge about physics): all attempts to derive ferromagnetism from the behavior of individual atoms, at the time Feynman was writing his lectures, gave the result with the wrong sign. Can you believe that? All physicists at the time assumed that the complexity of the situation—lots of tiny spinning charges interacting—should somehow make the correct sign emerge, if you “do all the calculations right”. Now does this remind us of anything, hmm.
It’s easy to note that something can in principle be built out of smaller building blocks, but much harder to explain how it was built. Do actual work, figure out the mechanism, like I tried to do in reducing “could”.
For example, do we really know anything about “personal identity” in the presence of copies? Can we build “personal identity”? Given a computer program that runs multiple copies of a person, can we make correct statements about the “subjective anticipation” of these copies? Which forms of anthropic reasoning work, and which don’t? Does the doomsday argument work? Many people hold strong opinions, but there doesn’t seem to be any definitively correct solution.
All the other examples on your list have the same problem. We cannot reduce mathematics—we don’t even know what a natural number is! (Which means we can’t teach it to a computer.) Reducing pain qualia—or “intelligence”, say—is likely correspondingly harder.
Edit: made tone less confrontational. Apologies if you read my comment before that.
The reference is volume II, chapter 37 “Magnetic Materials,” section 1.
Does anyone know about this? Has this been resolved?
I have no idea what an essence is supposed to be. I guess that’s why I’m having a hard time to understand your post.
That’s a symptom of having already successfully reduced. Pre-reduced language starts to become far harder to understand at a deep level!
If there’s one set of physical laws that determine everything, isn’t that a kind of essence?
On 7.: even if you were to successfully show that the real world reduces to some lower, possibly more/less complicated, level (as particle physics did, and nuclear physics did, and, well, physics really likes doing...) that next level is still “the real world” just understood in slightly more detail.
Which you can then try and reduce further. (I really hope it’s infinitely complicated :p)
In addition; 1-5 are all aspects of seven. So I’m really not sure what the point you’re trying to make.
We do now regard 1-5 as all being made out of the same stuff that 7 is made out of. But many folks persistently felt that there had to be separate kinds of essences involved in at least some of 1-5. I felt that way for some of them. The point is that I/they were mistaken; and the next point is to ask if there are any other questions about which we may be similarly mistaken or confused.
Does that make any more sense?
Even before 1-5 were succesfully reduced they were still seen as part of 7.
They were just seen as parts of 7 with different properties to the rest.
It doesn’t really clarify my main point of confusion, which is what you’re attempting to gain by listing 7 (ie. pretty much everything) as a single thing to be reduced
I agree that making particles out of smaller particles wouldn’t make me un-confused about the sense in which there is something rather than nothing.
That doesn’t mean there isn’t some set of concepts and reductions that would make me unconfused. Eliezer has a good description of how such processes can go at Righting a wrong question.
What do you mean by “the sense in which there is something rather than nothing.”?
That seems to imply that there is a sense in which there isn’t anything. But there not being anything is inconsistent with your existence, and by “cogito ergo sum”esque arguments, you can be certain of your own existence.
What about the nails scattered around here http://lesswrong.com/lw/oh/righting_a_wrong_question/ ?