Revision and Generalization of Cruddy Folk Concepts: An Introduction

Preamble

I think we sometimes think in concepts that are much too coarse-grained, open-textured, and underdeveloped, and we end up getting badly confused as a result. Let me start with an example to give you a sense of what I have in mind. Then, I will generalize the proposal. I suggest that we start thinking with new concepts, concepts that are more fine-grained, definite-textured, and generalized. I don't exactly suggest that we get rid of the old concepts. We can keep them as special cases of the more general concepts, and think in terms of them when we don't have a use for more refinement. But the upshot here is that we do have good reason for more refinement, and hence have a reason to revise and generalize cruddy concepts.

The Crush: A Cruddy Concept and Some Refinements

First of all, the classifications for desires that we have are way too coarse-grained, insufficiently general, and underdeveloped. In short, they're pretty cruddy, although they're onto something importantly real. Suppose that one has an intense desire to interact with some person, and that it has a certain qualitative feel that most of us would call a "crush". This concept is insufficiently general. When we say that someone has a crush on someone else, we acknowledge the following implication: "If the other person had the same feeling, and both recognized this, then they may very well want to have a romantic/intimate relationship together." But why assume that this desire+crushy-feeling we have is specifically about a romantic relationship? Couldn't it be about something like that, but not quite that? Well, I guess one might insist: "Oh, if it were only about something like that, it wouldn't really be a crush, you silly goose."

I think the person who says this is the silly goose. Why don't we generalize the concept of a crush? Introduce some more general concept that picks out a desire with a certain insistent, crushy qualitative feel that is directed at engaging in a certain kind of interaction with that person. Instead of talking simply about crushes on people, we could talk about crushes on people for a particular kind of relationship. We could introduce a three-place relation crush*(x, person, type of relationship). Why stop there? Why not generalize a little more? A person can certainly have a crush on another person. But why can't a person have a crush on a cat, a tree, a book, a painting, or a city? I guess one could again insist: "No, goofball, you're really missing the point. You can't have crushes on those kinds of things." I say that this person is the goofball. This person is missing out on a better concept. We can replace "person" with "interactive entity", and introduce a yet more general three-place relation crush**(x, interactive entity, type of relationship).

This relation subsumes the old notion of a crush as a special case. X crushes Y in the old sense iff crush**(X, Y, romantic partnership) & Y is a person. But we could have other crushes. How about crush**(me, Shane, intellectual partnership)? I certainly had a desire+feeling that feels a lot like the crushy desire+feeling that I've had for potential romantic partners when I first got to know my friend Shane. Why not crush**(me, New York City, intellecto-spiritual partnership)? I've certainly had a crushy desire+feeling a lot like the desire+feeling that I've had for potential romantic partners when I first got to know Gotham City. Why can't I call these crushes, and just generalize a little bit? Someone who tells me that I can't do this is missing out, and is limiting himself in a sad way. Why? Because there are important similarities between crushes in the old sense and a whole class of crushy desire+feeling states that we can pick out with crush**. We have a concept for crush**(X, Y, Z) when X = a person, Y = a person, and Z = a romantic partnership. I think we'd be missing out on a more general phenomenon if we stiffened the categories and refused to call something a crush unless X = a person, Y = a person and Z = a romantic partnership when crush**(X, Y, Z).

I think we miss out in other ways. When we refuse to generalize, we get really confused. Suppose that two youngsters have crushy desire+feelings directed at each other and come to a mutual recognition of this. They immediately jump to the conclusion: "Oh, I guess we're starting to fall in love. This could be romantic intimacy!" This is a bit of a parody. But it's not far from the truth. I don't mean to say to these youngsters: "No, it isn't going to be romantic intimacy." I rather mean to say: "The mere fact that you crush** one another doesn't entail that Z = romantic partnership. Sure, it might be. That would be cool, I concede. But Z could be a zillion things. Take a second to think more carefully." Why take a second to think more carefully? Because you probably don't really know what it is that you're wanting until you think more about it; your wants are not as luminous as you'd think, if you think about it. You know that you crush** this person. But do you really know much more than that? Without reflection, I doubt it. And if you don't dig a little deeper, you might end up being disappointed! Suppose that, in your case, it was really that Z = music-based intellectual/romantic partnership, and that, in the other person's case, it was really that Z = literature-based intellectual/romantic partnership. You might miss out on these fine-grained details. And you might realize later on what the details were, and that this is why things didn't work out. Take a second to think a bit about how precise the value for Z is. Do it all over the place, not just for your crushes** about people, but also for your crushes on graduate schools, professors, paintings, books, and so on.

In fact, we could generalize even further. I think when we have crushes on stuff, we in fact have crushes** on stuff for several types of relationship Z_i to various degrees. We could introduce a further generalization that has slots for the several types, and weights for the various degrees. Call it crush***. Let Z_i = Z₁, Z₂, Z₃, and so on, where these are the types of relationship desired. Let n_i = n₁, n₂, n₃, and so on, where these are respective weights on the Z_i. Measure the weights in the interval [0, 1] so that they add to 1. Crush*** is a four-place relation that has a slot for the ordered set (Z_i) and the ordered set (n_i), which are the weights on the various types of relationship chopped up in the unit interval. Suppose that Bob has a crush*** on you. What might that amount to? It might amount to this. Crush***(Bob, you, (Z_4­), (n₄)), where Z₁ = literary partnership, Z₂ = sexual partnership, Z₃ = musical partnership, Z₄ = food-based partnership, and n₁ = .2, n₂ = .4, n₃ = .1, and n₄ = .3. If we were being realistic, we probably wouldn't really want pointy values for the n_i, but rather possibly overlapping sub-intervals of [0, 1], since people's degrees of crush probably aren't fine-grained to the point of wholesale pointiness. But I think this is a fair generalization.

Let me explain. One might plead: "How on earth could our crushes be that precise?" I say: "You aren't looking carefully enough." Suppose that there are three near duplicates of some girl, A, B and C. Suppose A has slightly shorter red hair, is a little more into German novels, likes classical music a little bit more than pop music, and likes fancy Co-Op food more than ordinary food. Suppose B has slightly longer darker red hair, is a little more into Roman poetry, likes pop music a little bit more than classical music, and likes ordinary food a little more than fancy Co-Op food. Suppose C has in-between lighter red hair, is a little more into late 19^th century English novels, likes classical and pop equally but also likes jazz, and likes fancy and ordinary food equally well. Apart from these factors, A, B and C are qualitative duplicates; surely this is possible -- although you might expect qualities to cluster, there are no necessary connections between distinct existences. Might you not have a preference ranking over them? I think you might. Let m₁ - m₃ be your degrees of crush on A, B, and C. If faced with the choice of pursuing a crush on one of them, it might very well be the case that m₁ > m₂ > m₃, or whatever. Our tastes are very picky. That's why I think we need something like crush***, not just crush**, crush*, or the really cruddy, plain and simple crush. We need to be able to think about these fine-grained differences -- not always, but sometimes. And so we have a good use for crush***. I don't say that crush*** is always the best. If your options are limited, it might suffice to think in terms of crush*. But if you're faced with a choice between thirteen graduate programs and you've got a generalized crush on all of them, I think you might very well want to think in terms of crush***. So would you if you were choosing to pursue a crush on one of a pair of identical twins. And so on, and so on, and so on, et cet., ad inf.

The General Point

The point I've been making generalizes way beyond the concept of a crush. And it is familiar to people who've done a bit of formal philosophy, although philosophers haven't pursued the method of generalization and refinement far enough. As philosophers know, the simple notion of all-or-nothing belief isn't very useful. It's more useful to think about degrees of belief, and to introduce a generalization of the concept of a belief that works like a probability function. Rational degrees of belief are values of a function that at the very least obeys the probability calculus. Just as we might want to replace belief with degrees of belief -- with credence functions -- so we might want to replace the concept of a crush with crush*, crush**, or even crush***.

We can generalize other cruddy concepts. I say that we get rid of the current concept of gender. We can represent the current concept of gender as a function from persons to the values M and F. This concept is not very useful. The function is undefined in certain cases. Let G be the ordinary function. Sometimes G(person) = undefined -- say, because G is a hermaphrodite. I say that we generalize even further. Stop thinking about categorizing people in terms of bits of flesh, and instead start thinking about people as collections of dispositions. Certain dispositions are more masculine, certain are more feminine. If we wanted to generalize just a little, we could let the gender function be a function from persons to ordered pairs of values in the unit interval that add to 1. The first value represents degree of masculinity, the second represents degree of femininity. Let the new gender function be G*. G*(Kurt) = (m, f), where m and f are numbers in [0, 1] such that m + f = 1. I think we need to generalize further. My gender degrees change from time to time. Today m_Kurt = .3 and f_Kurt = .7. Yesterday, it was the reverse. So, we need a binary function from people and times to ordered pairs. Call this binary function G**. We can then say: G**(Kurt, t) = (m, f). We could generalize even further. Get rid of masculinity and femininity, and replace them with an arbitrarily large number of distinctive sets of dispositions. Or, don't get rid of masculinity and femininity, but add some new stuff. Then, introduce G***. G*** is a function that maps person-time pairs to ordered n-tuples representing degrees of possession of distinctive, joint-carving sets of dispositions. Masculinity and femininity might be among these. So might a lot of other stuff.

In fact, we can use this idea to vindicate certain aspects of postmodern thinking without going relativist. The problem was that our concepts were too coarse-grained, too open-textured, and very, very insufficiently general. It's not that the old concepts didn't pick out anything. It's just that they didn't pick out anything very useful. We can get more useful concepts that are theoretical successors of these. That's what we should do. We shouldn't get rid of objective truth and joints in nature just because our concepts sucked. Instead, we should get new concepts, ones, I might indeed insist, that are closer to representing things accurately, and that hit at joints in nature better. The problem wasn't that there was no objective truth or there were no joints in nature. There is objective truth, and there are joints in the world. The problem was that our concepts were badly out of touch with both of them, or, if they were in touch, they were only in touch in a shallow, shallow way. So, I suggest we subject all of our concepts to the treatment of refinement and generalization, although I certainly don't suggest that we talk in terms of the more refined concepts. This would be progress in philosophy, and it would be progress in ordinary life, where we need better concepts.