Concepts Are Tools, Not Artifacts
Cross-post from: https://crispychicken.cc/2021/02/12/concepts-are-tools-not-artifacts/
The delightful John Nerst points out that we have more of a problem organizing and moving around knowledge than we do creating it. I think we need a Conceptual Logistics that tells us what kind of patterns people have observed over the centuries, how they relate to each other, and where we can find evidence for them.
The simple way to say it is: we need ways of organizing information about the patterns observed about the world and what we believe the scope of this pattern is. What makes patterns complicated to "scope" is not so much about figuring out some perimeter where this knowledge applies and where it does not, because the shape of that perimeter is fractal-like and unknowable. Since this is infeasible, we are left to reason by "positive spaces" we believe concepts can be usefully applied. Most important is the kind of metonymies we will accept for a given thesis: what parts of a statement can be replaced without distorting it too deeply?
If I say "Birds fly." to a child it is not long before they discover the Penguin and the Ostrich and look at me with scorn. Yet, most children quickly learn that this is the precision they can expect.
If I say, "Too much individualism is killing us!" then who would you believe "us" refers to? Probably I'm talking about some subset of what might be called the "West", despite the fact that many "Eastern" countries are experiencing increasing individualism as we live through the 21st century. That's because my choice of English as a language and my choice of words recall similar statements that I am assumed to have some relationship, statements about the decadence of Western society and how it is falling apart.
Again, I would like to return to the problem of teleology. We tend to view things as being for uses, even though things just exist. In this same way, we tend to view the history of a field as a stepping-stone path to the place we have come to today. It may be the path that was taken, but for most fields the knowledge that exists today should be verifiable by other means that the fact that an important person wrote about it.
Looking at historical documents is important for many reasons. The three main ones I feel strongly about are:
- Summaries of other people's ideas are distorted and censored, consciously and unconsciously. The real thing is whatever the real thing was allowed to be at the time, which is as good as you're going to get.
- Documents that have survived for a long time and are important to some group usually tell you interesting things about that group and often reveals changes in historical circumstances.
- Facing the mistakes of past thinkers is the most important exercise for being a truly self-critical thinker.
However, I believe that none of these necessitate introducing ideas as products of history. Don't get me wrong: ideas are products of history. And I believe the reason we teach them as having "arisen" in history is (a) because that's kind-of-sort-of how it happened and (b) because it makes it clear that they come from somewhere, they're not God-given. And yet, teaching and storing and writing about ideas history-first has led to three things I think are really awful:
(1) Cargo Culting ideas of a given thinker, because the thinker is often smarter in some certain way than anyone else seems to be. Many of these historical figures really are that insightful, but I really don't care for anything that you can't give me grounding I can go and find in the world, even if that "world" is the archives of historical documents. Marx comes to mind.
(2) Believing that different "parts" of the idea have to work together in a certain way, because literature that describes them describes the historical idea's anatomy very closely. Ideas are anything you can conceive of and we can change them in subtle ways that are hard to come-up with if we're taught in terms of boundaries. For instance, natural selection occurs in lots of phenomena, like rocks, not just things with genetics. However, I have trouble explaining natural selection of anything non-genetic or non-meme related to most audiences because people can't "see" the natural selection there. They were taught the boundary more than the mechanism.
(3) The most painful issue, for me, is that people seem not to believe they have the right to believe things from the ground-up. "Didn't X big thinker say Y, which is basically the same?" Newsflash: "basically" isn't good enough. We need conceptual tools that work for us today and usually we want to take apart ideas into components and see how robust these parts are. I call this their metonymic robustness, because it generally describes how much different parts of a statement stand-in for more general categories.
Teaching, describing, studying, and adding to fields as if they are historical scrolls you have caught the bottom of has many uses, but I think there is a space for a new kind of conceptual logistics, one that tries to isolate the usable tools from history. We will leave a paper trail to the original thinkers in our footnotes, but we need modern thinkers that see how the assumption of Chekhov's Gun operates in people's readings of GPT-3 outputs, convincing them of the importance of every quirk.
There will be no single "conceptual logistics", instead it will be a style, defined by the description of theses as tools that have dynamics that are well-understood in certain environments and still need to be studied in others. This is the problem of learning how to use guns in space; what components of gun use on earth and physical laws can we extract predictive power from that will minimize astronaut casualties?
The Inexact Sciences has a unique need for such a style, because too much of human behavioral studies are clouded by assumptions and styles of the time. The pipeline was never filtered because this murkiness became the culture of these fields, even as it mutated. It is time we became serious about concepts as tools, not artifacts.
- 8 回复
As you know, big fan of this frame. We should get Nerst in here too.
That said, I think we could learn and clear a lot up by grounding this out in particulars and case studies. eg How should we think about LessWrong if we take this frame seriously?
For one, criticisms that LW reinvents existing concepts, and isn't especially novel, gets deflated. If we take away the premium from novelty, and place it instead on clarity of explanation, we end up in a different place. And taking LW's goal of rationalizing psychology at face value, it's not clear that you need many of the deep theoretical frames that the "original" thinkers they pilfer contain. What you need or want is a practical, memorable system of debugging your own thinking, and a sequence entry like Taboo Your Words does a far better job of this than its Game-A rival, Chalmers' "Verbal Disputes." Indeed, Chalmers' own account is arguably a rehash of James's intro to Pragmatism (the famous squirrel anecdote), except he puts a name on it and explores some of its nuances. In either case, saying "this is just James" would be both true and missing the point.
More pedantry—recategorized as Maps Meta because it strikes me as a map of maps more than about academia itself per se. The collection style of forums makes me at least want to give categorization a proper try—metonymically, the parts should unveil the whole (category) and vice versa.
I'd also be interested in talking potential lessons for the surrogation project, but you know that already.
Elliot, "East Coker":
So here I am, in the middle way, having had twenty years—
Twenty years largely wasted, the years of l'entre deux guerres
Trying to use words, and every attempt
Is a wholly new start, and a different kind of failure
Because one has only learnt to get the better of words
For the thing one no longer has to say, or the way in which
One is no longer disposed to say it. And so each venture
Is a new beginning, a raid on the inarticulate
With shabby equipment always deteriorating
In the general mess of imprecision of feeling,
Undisciplined squads of emotion. And what there is to conquer
By strength and submission, has already been discovered
Once or twice, or several times, by men whom one cannot hope
To emulate—but there is no competition—
There is only the fight to recover what has been lost
And found and lost again and again: and now, under conditions
That seem unpropitious. But perhaps neither gain nor loss.
For us, there is only the trying. The rest is not our business.
Yep, I agree with all of the above. The one critique I will leverage at LW is that I think it tends to Cargo Cult itself and believe it's identifying some underlying territory of thought optimization, when there are lots of ways these things can fall together. So I guess my main issue that by creating new "concept ceremonies" (e.g. "epistemic status", which I like) but still claiming its just going from foundations each time it's pulling a sleight of hand that leans into some of the cultishness it's accused of.
“We shed as we pick up, like travellers who must carry everything in their arms, and what we let fall will be picked up by those behind. The procession is very long and life is very short. We die on the march. But there is nothing outside the march so nothing can be lost to it. The missing plays of Sophocles will turn up piece by piece, or be written again in another language. Ancient cures for diseases will reveal themselves once more. Mathematical discoveries glimpsed and lost to view will have their time again. You do not suppose, my lady, that if all of Archimedes had been hiding in the great library of Alexandria, we would be at a loss for a corkscrew?”
― Tom Stoppard, Arcadia
"It's all in the carving"/"there are lots of ways these things can fall together"—almost certainly true for most of the conceptual domains we're discussing. Maybe true all the way down:
In philosophy, categories in the world that are absolute, intrinsic, or objective are called “natural kinds.” Pretty much the only available examples of natural kinds of eggplant-sized objects were biological species. Philosophers have reluctantly more-or-less accepted the recent factual discovery that species are necessarily nebulous, and that this makes the “natural kind” concept obsolete.
But the fact that boundaries are fuzzy, that there's no pure natural kind, is very different from saying "there are no cross-culturally appreciable or intrinsically prominent features of clusters in thingspace" (if you take "intrinsic" to mean, to generic human needs/perspectives). And I want to actually figure out the extent to which frames can vary for given objects of scrutiny. There's a book I've been wanting/meaning to read, to hopefully shed light on these issues—Brent Berlin's Ethnobiological Classification: Principles of Categorization of Plants and Animals in Traditional Societies.
A founder of and leading thinker in the field of modern ethnobiology looks at the widespread regularities in the classification and naming of plants and animals among peoples of traditional, nonliterate societies--regularities that persist across local environments, cultures, societies, and languages. Brent Berlin maintains that these patterns can best be explained by the similarity of human beings' largely unconscious appreciation of the natural affinities among groupings of plants and animals: people recognize and name a grouping of organisms quite independently of its actual or potential usefulness or symbolic significance in human society. Berlin's claims challenge those anthropologists who see reality as a "set of culturally constructed, often unique and idiosyncratic images, little constrained by the parameters of an outside world."
I agree; species are the most "natural kind" it gets, and they're far from a central exhibit for social constructionists. That said, by the time you get to cross-cultural attitudes to emotions (e.g.), frames, lumpings, splittings are all over the place. So I think working our way up from species, to maybe common landscape features/inanimate objects, into concepts of self and society, seems like the best or only way to get a foothold on this.
Yeah, I love this idea of "building-up" from things that have cross-cultural taxonomic robustness.
What I mean when I say "here are lots of ways these things can fall together" is definitely not that there isn't any underlying reality, but that you can make a "motor" in many, many different ways. It's still going to use some sort of energy source to produce semi-reliable rotation. The importance, I think, is in how we think of the parts of the motor once built—those things have teleological significance.
We see this with species too—taxonomies were good before we understood evolution, and the elegance of the "tree of life" model made it very popular in the long run. However, since then lots of scientists have shown that species are less clear-cut than we originally thought, making the question a bit more difficult than it used to be. Most of us have DNA that can be traced to Neanderthal's (on average 3% of our genetic code, I've heard) and the Neanderthals are gone, this is a kind of convergence of close-by branches that complicates the story. Of course the tree of life still seems to be at least 90% of the story, much like relativity didn't break Newtonian physics, but it made us thing differently about how things fall into components and at what scales we see certain effects.
I'm 100% on-board with this species idea and I'd like to develop it further. I'd also like to explore a slightly different angle, looking at some specific cases where different words are described some object of heavy use (e.g. the same or very similar algorithm often exists in computer architectures, operating systems, and compilers with different names in each subfield) and start picking apart if there's any real difference, how much things converge, etc. I don't imagine it will be nearly as good for getting evidence from starting from species, and I agree we should definitely do that, but I can't help but feel I don't even have the right words to describe this and I want to find some basic examples that we can all become familiar with, ones that refer to objects most of know about—like the cultural knowledge of what constitutes a romantic relationship.
cleanliness request please put the source when you quote something 🥺 I usually go look at it just a bit to get the feel for where it's situated. In this case googling was enough but sometimes I waste a bunch of time googling before realizing it's something one of us wrote that's not indexed.