Getting Beyond Categories

by johnmccreery

It has been an immensely enjoyable and educational experience to lurk as the postmodern Christianity debate has developed. The thought I’d like to develop here is a tangent suggested by that debate. I propose for consideration the idea that, while postmodernism has successfully attacked theorizing in terms of categories, it has left largely unscathed theorizing whose logic subsumes but extend far beyond what categories have to offer. By “categories” I mean sets whose members uniformly share common properties. A set conceived in this way lends itself to definition in terms of classical scholastic logic and sets up debates in which there are only three possible ways to handle exceptions to definitions: (1) sharpen the definition to exclude the exception or (2) rewrite the definition to include it, or (3) reject the definition. Since (3) requires no further thought, it is far and away the most popular approach. Postmodernists observe, correctly I believe, that the big ideas in terms of which we humans define our identities are almost always ambiguous, polysemic, and shot through with contradictions that make definitions in terms of categories conceived in terms of classical scholastic logic impossible. Then, however, the usual approach is the simplistic overextension of (3), the assertion that no definition is possible.

That conclusion would be right if the definition of “definition” were restricted to the classical notion that a definition is a list of necessary and sufficient conditions. Allow definitions to include maps of relevant spaces and lists of conditions that affect whether traversing the map takes you to one place or another. Describe the conditions in terms of probabilities. Run simulations to test the range of outcomes. Now both pomo and categorical mo look increasingly obsolete.

The question we and our students now confront is how to learn, use and teach the knowledge necessary in a world where, more than finding a place to say, “Here I stand,” we need to learn how to stay upright (maybe even do some fancy tricks) on surfboards riding waves of change.

Advertisements

26 Comments to “Getting Beyond Categories”

  1. Carl: “Then, however, the usual approach is the simplistic overextension of (3), the assertion that no definition is possible.”

    Kvond: Please…WHO is it that you are referring to that makes the claims you object to? It would be much more informative, not to mention interesting, if we had a REAL example to consider, in the context of arguments made. Perhaps give us a definition (or even defining itself) that has been rejected, and we can examine what has been put in its place?

  2. Sorry, that was not Carl. JM feel free to answer my question as to WHO it is that claims x and y.

  3. Nope, not going to go there. I know that there are people for whom academic genealogies are the most important thing in life. I understand that we live in an era where all sorts of people worry about intellectual property. But me, I am more about understanding the world in which we live than fetishizing who said what.

    That said, the inspiration for the remark is Warren McCulloch, in the introduction to Embodiments of Mind, where the founder of automata theory says that his task is to build machines that simulate human behavior. He has noticed that, whenever his machines fail to do something that people do, there are those who say, “See, no machine will ever do what humans can do.” Then he goes off and builds a better machine.

    I also think that it’s pretty uncontroversial to assert that when people attack what they call essentializing, they are, logically speaking, asserting that definition, at least in terms of necessary and sufficient conditions, is impossible.

  4. We must distinguish between those categories that we construct via definition, and those categories for which we are seeking some definition.

    In the former case, objects qualify as members of the category if and only if they satisfy whatever conditions the definition entails, such as a set of necessary and sufficient conditions. As far as categories that involve necessary and sufficient conditions, this is essentially proof of their existence.

    In the latter case, we have categories that we, somehow, perceive or believe to exist, and that we might wish to find a definition that can be used to categorizes all positive instances positively, and all negative instances negatively. It needn’t be the case then that an object will belong to such a category if and only if it satisfies the definition- obviously enough because the definition might be inadequate to the task.

    When one encounters an exception, these fall into two kinds: those that satisfy the definition but don’t belong, and those that don’t satisfy the definition, but do belong. But here is the rub: how is it that these exceptions are being identified as such? There must exist *some* decision procedure by which the analyst or other human subject decides whether that instance belongs to the class or it does not. It doesn’t really matter how this decision procedure is implemented, or whether it is coherent, or if it can be described by a set of necessary and sufficient conditions. What matters is that it exists; therefor it is, in principle, implementable.

    Much more speculatively:

    In some sense then, whatever is so categorized by that decision procedure is what belongs to that category, at least at the time the decision is made. We see here that this second kind of category is actually closer to the first kind than we had thought. At least, it would be if there were only one such procedure.

    But the decision procedure needn’t itself be invariant over time or across individuals. Instead, we might imagine a category as a label for a family or related decision procedures, whose variants are distributed in a population. Moreover, individuals at different times might employ different variants, or tweak existing ones, depending for example, on how they’ve been primed prior to using it. And the category label itself, as another kind of object that is distributed, might mask or cover unlabeled concepts accounting for different variants of one sort or another.

  5. Sorry Jacob, I wasn’t sufficiently clear. The proposal is, at least as a thought experiment, to ditch categories altogether and see what happens if we conceive knowledge in other ways.

    For example, consider the word “dog.” If we assume that it refers to a classically defined category, we look for necessary and sufficient conditions. Suppose, instead, that we treat “dog” as a pointer to a set of links, a node in a web of associations. These may include the canines found in the classical categories used in biological taxonomies. They may also include other dogs, e.g., the ones referred to in calling a lecher a “dirty dog” or a piece of software we don’t like a “real dog.” Instead of the goal of analysis becoming a decision about what is or is not a dog, the goal becomes a map of possible associations, a network diagram that can then be examined for all sorts of properties: centrality, connectedness, betweenness, cyclic versus acyclic connections, clusters, components or bi-components, the list continues to grow. Historical or other research can explore how networks evolve.

    Sure, things can get pretty messy pretty quickly. But with this approach we aren’t stuck with endless A or not-A debates, or what I am calling the pomo gambit, “We don’t have a clue how to sort this out.”

  6. And this is why I love me some Wittgenstein. (and John, in that comment.) He gives us a fourth option: recognize that no one definition will work in all cases, and regard what seem to be exceptions as opportunities to explore the boundaries of the case. I’d argue that it also enables us to do some of what I think Jacob might be getting at, and, more specifically, enables us to question the power structures (for lack of a better phrase at the moment; i’m still embroiled in budget heck) that are coming into play. If we take “person” and “man” to be equivalent–consciously or unconsciously; structurally; economically; whateverly–then where does that leave women? What can “race” possibly mean? To whom and when? What does it mean to say someone is a “worker” or a “professional”? And so on.

  7. John, this is reminding me of the race and ethnicity course I taught this semester. I showed the students various cases and strategies to come to their own conclusions, but also nudged them toward at least considering the possibility that race does not have an essential definition or being, but rather a cluster of instances, uses and deployments sharing what Wittgenstein (and Kvond on the last thread) called a ‘family resemblance’.

    My sense is that the only work pomo does is to break down the category and find the usual play of difference. But if the relationships that gave rise to the category are still there they shrug and keep right on relating. Then we need a pragmatism to sort out how they work. And this is why I don’t think pragmatism (Peirce, James, Dewey, Mead, Wittgenstein, Rorty) is in the same family as postmodernism.

  8. I don’t care what you call it, just keep that thing from pissing on my sage plant every time it gets off the leash.

  9. If you’re plant’s so sage, how come it just sits there and lets itself get pissed on?

  10. Because it doesn’t know how to play language games with the pisser?

  11. [Speaking of language games I note a typo in #9 and apologize for calling you a vegetable in cartoon Southeast Asian pidgin.] … Maybe the plant is reveling in the postmodern condition of being pissed on and pissed off at the same time.

  12. @Narya

    And this is why I love me some Wittgenstein. (and John, in that comment.) He gives us a fourth option: recognize that no one definition will work in all cases, and regard what seem to be exceptions as opportunities to explore the boundaries of the case.

    Yes, indeed. That “opportunities to explore the boundaries of the case” is very nice. Does raise an interesting question, though, what if there are no boundaries? What if the possible uses of a term constitute a field that fades into the distance but never reaches an edge?

    @Carl

    My sense is that the only work pomo does is to break down the category and find the usual play of difference. But if the relationships that gave rise to the category are still there they shrug and keep right on relating. Then we need a pragmatism to sort out how they work. And this is why I don’t think pragmatism (Peirce, James, Dewey, Mead, Wittgenstein, Rorty) is in the same family as postmodernism.

    I am happy to count myself a pragmatist and turn for insight and inspiration to the thinkers you mention. My model, however, is Victor Turner, who had nothing but scorn for interpretations of myth and ritual unanchored in the nitty-gritty of social structure, politics and economy.

  13. LW talks about “boundaries for a purpose,” and that would probably serve here. I think part of LW’s point is that there are not intrinsic boundaries to anything, in some ways, that we utilize boundaries–in various ways, for various purposes–but no one boundary (or definition) will serve for all purposes.

    What I like about this is what Carl has been observing, that LW takes the pomo tendency to break everything down, but LW would not agree that one can break down everything simultaneously, nor would he (or, at least I) see any good reason to do so. What’s the purpose of that? To show that sands shift? We knew that. But you’re still standing on sand right now, which means both that it can/will shirt AND that it can be stood on, and that’s an important probably non-pomo point, a still point of a turning world, even if it’s not the same point, exactly, tomorrow.

  14. Right, Narya, that’s it exactly. To me pomo is like a trick that once you learn it gets old and a little silly real fast, Hey look gang, I can make make race/class/gender/pelicans disappear! And if you’re that guy who’s still doing that trick three parties later, the invites stop coming.

  15. The confusions which occupy us arise when language is like an engine idling, not when it is doing its work.” — Wittgenstein

    The moment has passed and the conversation has moved on, but for what it’s worth, my remark about the pissing dog was motivated by John M’s proposal “that we treat “dog” as a pointer to a set of links, a node in a web of associations,” combined with Carl’s that “we need a pragmatism to sort out how [categories] work.” Call me a a naive realist, but I’d say that I used the word “dog” to point to that thing in my garden. The word was embedded in a sentence that links “dog” to a pragmatic web of associations that include the sage plant, the pissing, the unsavoriness of eating pissed-on herbs, the irresponsibility of the thing’s controller. All of those associations unfolded out there in the world before they ever found their way into my linguistic construction.

    Dogs are among the very few creatures that, when you point at something, might actually follow the trajectory toward which you’re pointing rather than just looking at your finger. This ability to take another’s perspective toward the world and to understand the other’s communicative intent regarding the world is foundational to the acquisition and pragmatic use of language. Here’s the beginning of Michael Tomasello’s 2003 book Constructing a Language:

    “Nothing could seem less remarkable than a one-year-old child requesting “More juice” or commenting “Doggie gone.” But the remarkable fact is that even these baby utterances differ from the communicative activities of other animal species in a number of fundamental ways. For example, other animals do not refer one another’s attention to outside entities such as juice, they do not make disinterested comments to one another about missing doggies or the like, and they do not combine communicatively significant elements to create new meaning.”

    Tomasello, a developmental psychologist and ethologist, contends that there are two critical sets of abilities necessary for language acquisition. The first is intention-reading, which in its earliest developmental form includes prelinguistic acts like following the pointing finger. The second is pattern-finding, which forms the basis for categorization. Species that don’t use symbolic language engage their pattern-finding abilities for distinguishing food from predator from mate from something good to piss on. Linguistic categorization builds on this pragmatic animal ability to make distinctions, abstracting it into the context of shared intentionality. So if we’re standing around and I point out the window you might come over and see what I see. If I look out the window and say “that dog is pissing on the sage plant again,” it’s a complex pointing gesture that uses verbal symbols rather than gestures. The categorical labels “dog,” “pissing,” and “sage plant” do the pointing to things out in the world, and in understanding my sentence you follow in your mind’s eye the signifiers as they point to the things in the world that they signify.

    Now of course this post-and-thread might have nothing whatever to do with my comment or the pissing dog, which might mean that I’m not that skilled at following the pointing fingers.

  16. It’s already twenty years old (published in 1990), but my constant reference for this sort of thing is George Lakoff’s Women, Fire, and Dangerous Things , from which I take my critique of classical categories and a variety of ideas about how the mind works without them.

  17. I liked that book too, John, but it’s been awhile, so bear with me if I summarize to myself for awhile. Tomasello’s findings are generally compatible with Lakoff’s I’d say. Both are interested in describing empirically how humans actually discover patterns and define categories, rather than trying to fit them into an a priori rational logic. Lakoff characterizes this empirical approach as “experimental realism,” in contrast to the “objectivist realism” inherited from idealist philosophy. Ordinary human categorization schemes and the words used to name them do attempt to describe or point to a reality that exists outside the head and the community, even if these psychological constructs don’t accurately represent the actual entities and structures occupying that reality. I’m fine with that.

    Where I got thrown off was in your comment #5, in which you were proposing — rightly, I though — to set aside objectivist categorization. “Suppose, instead, that we treat “dog” as a pointer to a set of links, a node in a web of associations,” you said. This sounded to me like a coherence theory characteristic of structural linguists like Saussure, where the category “dog” gains its meaning only as it finds a place in the larger conceptual, linguistic, and cultural matrix of human-defined categories. I don’t think Lakoff has that in mind. The natural-language word “dog” still points to some fuzzily-defined set of creatures out there in the world, one of which pisses on my sage plant every chance it gets. However, my/our ordinary use of the “dog” label need not correspond precisely with what really (or “really”) distinguishes dogs from non-dogs independent of human perception and thought.

    Tomasello’s developmental work shows that infants learn to use language incrementally, starting with individual words (especially nouns) and only gradually incorporating them into more and more complex networks and frames and structures. In learning individual words like “dog,” children do try to match up the categorical boundaries of the word with the categorical boundaries of those 4-legged critters they see out there in the world.

    I’m in agreement with Lackoff, and no doubt with you too, that some categories are entirely human constructions and may vary widely between cultures — government is the example Lackoff uses. But these things too operate outside of the our minds and our language, so the ways in which humans categorize and name these human cultural artifacts isn’t fundamentally different from the ways in which humans categorize animals and other features of the natural environment.

    Do we have an accord here?

  18. Accord? Perhaps. We agree, at least in a loose sense, that humans categorize and that categories vary from one group to another. On the other hand, I am reminded of Howard Becker’s opening paragraph in Tricks of the Trade: How to Think About Your Research While You’re Doing It.

    Undergraduates at the University of Chicago, when I was a student there, learned to deal with all difficult conceptual questions by saying, authoritatively, “Well, it all depends on how you define your terms.” True enough, but it didn’t help us much, since we didn’t know anything special about how to do the defining.

    If I ask myself why I find Lakoff’s arguments in Women, Fire and Dangerous Things so appealing, I recall several moments in my intellectual trajectory when the challenge was, precisely, how learn something special about defining, and the classical search for essences, practically speaking, lists of properties shared uniformly by members of a set, was, if not useless, misleading. For example,

    1. A freshman class in symbolic logic, where the first sets of problems had to do with formalizing propositions extracted from ordinary language statements.

    2. A discussion of operational definitions in a philosophy of science class, the case in point being temperature and what happens to mercury thermometers near absolute zero or the surface of the sun.

    3. Calculus and mathematical statistics classes, where I learned about nominal, ordinal, integer and real number metrics, was taught how to think about things in motion, and introduced to the question of how to handle outliers (as errors, as negligible for purposes of the problem at hand, or, just possibly, clues to something needing closer investigation).

    4. Linguistics, from which I learned the method of minimal contrasts, which works pretty well for phonology and syntax but becomes a quagmire in semantics.

    5. A year as a research assistant in the Yale AI program, writing concept frames and scripts for a program optimistically titled FRUMP (=Fast Reading Understanding and Memory Program), and discovering just how much of the “news” pouring off the AP and UPI wires is incomprehensible without a grasp of metaphor.

    6. Then, of course, the languages: Bits of Latin, French and German in high school and college, Mandarin in graduate school and Hokkien for fieldwork in Taiwan, then Japanese, which after thirty years in Japan, is now my primary second language.

    7. Three decades making a living as a copywriter and translator in Japan, negotiating with clients who can turn to dictionaries to help them parse what I have written but are frequently clueless about the nuances of usage, irony, humor or trope on which anything but the most pedestrian uses of language depend.

    This is the experience which, over the years, has had me smiling as I read not only Lakoff but also, for example,

    J. L. Austin How to do things with words
    George Steiner Language and Silence: Essays on Language, Literature, and the Inhuman
    D. McCloskey The Rhetoric of Economics
    Stephen Owen Traditional Chinese Poetry and Poetics: Omen of the World

    It is also the ground in which my current fascination with social and semantic network analysis is rooted. Have you heard, for example, of Crawdad? Given several hundred pages of transcripts from UN debates, this program generates maps that show the terms most frequently used by delegates of different clusters of countries, highlighting not only overlaps but fundamental differences that stand in the way of productive policy making.

    The more I reflect on what I seem to have learned from all this, it isn’t the preposterous claim that classical categories are never useful. It is more along the lines of, yes, they are useful but only in a small set of cases. Thus, it seems to me, so many of the debates in which we engage both on line and elsewhere tend to founder on what we call “semantics,” i.e., the search for answers to the type of question to which Howard Becker refers in his opening paragraph. To me, even Narya’s suggestion that we, “recognize that no one definition will work in all cases, and regard what seem to be exceptions as opportunities to explore the boundaries of the case” still assumes that the classical category is the place to begin.

    I propose something both utterly pragmatic and more radical: The place to start is “What are we trying to do?” Sometimes the answer will be to define classical categories. Mostly, however, the answer will be something else, something like poetry, drama, physics, a bit of artful rhetoric or a well-crafted simulation that produces surprising results. To understand how that happens, “How do you define that?” is, I speculate, a lousy place to begin.

  19. Okie dokie, I hear you, and presumably we agree at least on what Lakoff meant by his experimental realism: we use words pragmatically to talk about stuff in the world.

  20. Yes, indeed. But also to tease, joke, shock, ignite passions, tell stories, negotiate, persuade…. all areas in which, when I look back on it, my formal education was notably lacking. I am still thinking about that.

  21. “To me, even Narya’s suggestion that we, ‘recognize that no one definition will work in all cases, and regard what seem to be exceptions as opportunities to explore the boundaries of the case’ still assumes that the classical category is the place to begin.”

    Sure, that’s right. Yet what is this but yet another hairsplitting gesture of categorization? How would her statement look different if we weren’t playing the old scholastic game of distinctions and instead were looking for nodes of affinity and alliance? Old habits die hard….

    “I propose something both utterly pragmatic and more radical: The place to start is ‘What are we trying to do?’ Sometimes the answer will be to define classical categories. Mostly, however, the answer will be something else, something like poetry, drama, physics, a bit of artful rhetoric or a well-crafted simulation that produces surprising results.”

    Yes, exactly. My feeling is that for the last little while – anywhere from 50 too 150 years, depending how you count – we in the high humanities have been talking around this question. And in the process that talk has gotten very, very cheap. I’m not exempting myself from this critique! but I am saying it’s past time for every programmatic statement about what we should be doing to be met with a firm ‘Fine. Get on with it.’ We need less telling and more showing around here.

  22. Speaking of personal lessons learned, I once submitted to a respected psychology journal a paper entitled “How to Do Things with Citations.” Scrupulously empirical, the paper concluded with a set of evidence-based pragmatic tips on how to assemble a set of citations if you wanted your own paper to be frequently cited in turn. The reviewer deemed my paper too cynical and rejected it for that reason. Maybe if I had cited Austin explicitly, the reviewer would have been more enthusiastic.

  23. Now that’s a paper I’d like to read.

  24. I am saying it’s past time for every programmatic statement about what we should be doing to be met with a firm ‘Fine. Get on with it.’

    Sounds good. See you around when there’s progress to report.

  25. I’m not sure what constitutes “the high humanities,” Carl (there’s that categorization thing again), but looking back on the pomo Christianity post, I’m a lot more enthusiastic about both Christianity and postmodernism when I regard them as genres of creative fiction than when I try to figure out whether they’re true or not.

  26. I like that phrase–nodes of affinity and alliance. It’s very Wittgeinsteinian, and also useful. It was partly how I chose the categories I chose for my dissertation–look for turns of phrase, oppositions, categories, that we (think we) take for granted, and see what struggles are taking place around them. Not for the purpose of saying that the categories don’t matter–they did, very much, to the people who were fighting about them defending (or trying to tear down or move) their boundaries, etc.–but to see what, really, the struggles are.

Leave a Reply!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: