What's Wrong with Definitions?

This essay has a history. Part of the history is that it has been rewritten as material for students in various philosophy courses. I worked it up originally for upperdivision G.E. courses and then for seminars for philosophy majors, and finally for audiences with less background, Introduction to Philosophy students and other lower division courses. The references in the last lines are to Plato’s Symposium, to a speech given by Diotima which contains a classic example of a definitional account of love, and to the following and final speech of Alcibiades, who presents not a definition but a pure example. My main claim here, supported in different ways, is that definitions are hugely overrated. There are of course some uses for definitions, but every time that definitions play an important part in a philosophical investigation or a philosophical argument they are almost certainly worse than useless. Students who have been taught otherwise should sue their teachers. I’ll pass over most of the ways that definitions interfere with critical thinking and paper over problems with arguments in order to focus more on the arguments that definitions cannot be crucial in helping us understand any difficult philosophical problem or idea. Not much of this is original with me, though those who gave me these ideas haven't said them very concisely. I take it that definitions, especially philosophical analytical definitional accounts in terms of necessary and sufficient conditions, are often the target of Wittgenstein’s investigations, and that one thing he works toward is to get us to be suspicious of the very first philosophical impulse, which is an impulse to suppose that something like this, a definition or an explanation, is what we need in the face of some problem. Sometimes what’s needed instead is questioning the question, in particular excavation of the thinking that went into articulating the problem or supposing that there is one. Other sources are Plato, J. L. Austin, G. E. Moore, and Frank Ebersole.

              Our issue is in the context of some main and central problems of philosophy. It is, How important are definitions of philosophical terms? First of all comes a clue that something odd is going on, in that the words we are talking about as we ask whether we need definitions are not technical terms. Admittedly, they are adult, educated words, but when we ask whether we can justify our moral judgments, or what knowledge is, or whether life has a meaning, or what the relation is between appearances and reality, the terms in which we set our questions are not artificial or peculiar only to philosophy. Neither are they scientific or literary, nor do they belong to a specialization where they require explanation. Instead, we are talking about a fairly small range of terms which occur in our ordinary conversations outside philosophy, but which often are used by philosophers in philosophically fraught ways. These are terms like knowledge, truth, art, justice, beauty, culture, science, goodness, meaning, language. Another way to put this oddity is that the issue is about whether we need or will be helped by definitions of terms we generally use quite freely without definitions.

              Here are two fairly simple arguments against the idea that definitions of philosophical terms are important to our understanding. One is a kind of criterion argument, and the other is developed by Plato in his dialogue, Meno. The two are quite similar.

First, since it’s quick, here's the basic criterion argument.

1. It is possible to give a bad definition. By bad I mean a definition which misleads us as to which cases are the right cases. So, I define a fid as a pointed tool, and then someone identifies a fork as a fid (this is a cartoon of the mistake, but surely we could spell it out, or give other examples for higher stakes--think of President Clinton and how to define what having sex is.)

2. It is sometimes possible to tell that a definition is a bad definition. A fork is not a fid, having sex need not be restricted to copulation.

3. Definitions cannot tell you themselves that they are bad definitions or good. We have to catch them with their pants down ourselves.

4. Therefore, it is something else (e.g., the examples of the things being defined, or our separately understanding the term at issue) whichmust serve as the criterion of whether a definition is good or bad. The criterion or standard for whether the definition is good or bad cannot be provided by the definition, but must be something else.

5. Therefore, the crucial help to our understanding cannot be by way of definition. We have to be able to test the definition by some external standard.

              Another way in has the advantage of going more slowly and step by step. It is Meno's paradox, articulated in Plato’s dialogue Meno. The topic turns on our ability to recognize or to know the goal of our philosophical investigation. That is, if we are investigating what is real or the relation between appearances and reality or what is knowledge or what is art or what is justice or what is love, we will want to note how well we understand the thing we are investigating.

              I'll remark first, though, on the place of Meno's paradox in Plato and in philosophy, in order to emphasize that it's a big deal and in order to give a separate, parenthetical argument against definitions by laying out the role they play in Plato's dialogues.

              The basic pattern in Plato's dialogues shows this paradox but has the status of a separate argument against definitions. That pattern is in four or five steps, depending on how we count. The first step is that a question, usually unmotivated, crops up. What is X, really? where X could be love, knowledge, virtue, justice, piety, truth, meaning, statesmanship, perception, courage, art, and so on. The second step is that some poor sap of an interlocutor offers to answer the question by offering some examples of X. So in the Theaetetus, where the question is what is knowledge? there is a song-and-dance between Theodorus the math prof and Socrates which results in Theodorus offering up the bright young graduate student Theaetetus as a sacrificial lamb. Theaetetus offers as answer to the question, what is knowledge, that there is the knowledge of making and repairing shoes that a cobbler has, and also there is knowledge of geometry and other mathematics he is learning from Theodorus. The third step in the pattern is that Socrates squishes the examples-type answer like a noxious bug. --I asked after one thing and you have offered me instead a great many things, an endless list, and how shall we tell what makes these examples the right examples, or how shall we tell given this list what else belongs on the list and what does not? The line of thought goes like so: there must be a way we identify those given examples or the next examples not yet given, so let's articulate that and it will serve as a rule which illuminates the essence of the thing we are after. Definitions then are supposedly a path to insight in a way that examples are not. The fourth step of the pattern makes up the bulk of these dialogues: that is, the dialogues mainly consist of what comes next: proposals of definitions and then arguments testing them, looking for inconsistencies, absurd consequences, problems of many kinds, and then in the light of those revisions, revisions to the proposals or new proposals, and then more debate prompted by critical inquiry into possible problems. This fourth step is the model for a great deal of philosophy ever since, a model which is still one of the most widely used models for doing philosophy in most of the Western World. But there's a fifth step left out by the modern imitators–Plato’s dialogues end with an admission of defeat and an expression of humility. --We don't, Socrates says, really know what knowledge is, or piety, or justice; we have met defeat, and must acknowledge what we do not know. Not a single definitional account in Plato survives to the end of any dialogue.

              I think that the presence of this last step has received inadequate attention, and that it is regrettable that philosophers do not include it more often in their journal articles, and that perhaps Socrates' concept of wisdom as the sort of humility in which one admits what one does not know is still related to philosophy in some crucial way. That's too mild; I’ll say it more strongly: Philosophy is related to humility. We might learn that from Socrates; we might learn that from those Platonic dialogues which include the pattern outlined above.

              One might ask why Socrates and Plato continually march us through what looks like a quest for definitions, since the quest is always futile. One answer could be that it's hard to give a good definition, especially when we expect the definition to illuminate the heart of a thing and to carry the burden of guaranteeing understanding. Socrates keeps on, despite the difficulties, because he's hopeful and stubborn. Another possibility is that the agenda is ironic--that Socrates is, rather than hoping for success, quite open to the idea that we will fail and that the failure is important, and he is planning to rub our noses in it. Here's a hint: the right answer is b, the latter.

              Furthermore, Meno's paradox leads to the idea that that last step in those dialogues is inevitable. It goes like this: We philosophers often investigate something of philosophical interest and as part of our investigation we attempt to give an insightful definition of the thing to help our understanding what is at the heart of it. Socrates asks in the dialogue The Meno, “But how shall we decide when we have succeeded and when we have failed?” Consider the possibilities for whether we understand the thing we are investigating. It seems that either we understand at the beginning of our investigation the thing which we are investigating, and so then we will be able to understand when we have succeeded (that is, we will be able to recognize a good definition) but we will not then have added to our understanding, or else we do not understand the thing we are investigating, in which case we will not know whether we have gotten the definition right or wrong and cannot recognize whether a definition is good or bad, so that we are in danger of being misled. The investigation can either accomplish nothing (which is the best possible scenario) or we will be in danger of being misled by a bad definition.. Nothing can be learned by definitions of these terms. Definitions are either useless or they are dangerous.

              All these lines of thought lead toward a conclusion that trying to accomplish something of philosophical import by way of definitional accounts is doomed. That's probably too strong, but still it is the general tendency of these lines of thought. We have no examples before us, but all you need if you want examples is to study the history of philosophy, which in my bleaker moments seems to be a history of failed definitions, of war zones of battles over definitions, each war zone littered with bomb craters and schools of philosophers huddled inside artillery bunkers with their binoculars up, waiting for you or me to just ask a question, go ahead, just poke your head up and ask a question if you dare. What is good? What is meaning? What is knowledge? What is intention? What is being? What is art? Booom. Boooom. Look out, here comes a bombardment from the process philosophers. There's an attack from the pragmatists going on over in ethics. The existentialists have taken northern France again. Nagarjuna's troops have a beachhead in South Carolina. Rorty is rumored to be using philosophical nerve gas. The post-structuralists have turned our words against us. Derrida is drawing our fire but is a hologram. Keep your head down.

              This bleak version of philosophy at war over things which do not matter at best and are destructive of understanding at worst is one of my worries. I want to show some alternative to that.

              One alternative opens up if we can head off the temptation to think that definitions are crucial. That is, if there really is anything to Meno's paradox, then that's important, and we either can just keep butting our heads against the wall of inevitable ignorance (perhaps there’s an existentialist and macho answer--embrace absurdity and construct an answer based not on good arguments but on force of will) or even worse can just keep working with recovered memories which make our work unneeded, or we need to have somewhere else to go. Further, what did we think definitions were going to do for us anyway? Did we think that definitions were on our side in any substantial issues which have real stakes? That is, did we think that definitions of good or justice or moral or life would make an abortion permissible or impermissible? That a definition of knowledge would settle whether scientists can clone people or resolve the problem about stars being older than the universe? This (what I'm saying now) is not really an argument so much as a reminder that, after some defeats and looking at our history of defeats, and after considering the role of definitions in general, those goals we might have taken seriously begin to look less plausible. If we were really attached to those goals, then we might start to share that bleak view above, because those goals still seem noble and yet they recede to become as unreachable as some distant star.

              Here are some other reminders about possibilities. It is possible that in many debates, such as the debates over substantial issues like abortion or whether there are stars older than the universe, the arguments in the discipline or arguments relevant to the issue might still be relevant despite anything a philosopher could do with definitions.

              Another reminder is prompted by the question, which seems to tempt us back toward definitions, "But how do we know what we are talking about?" The reminder is about what some philosophers and linguists call genetic analysis, looking at how we acquire mastery of words: definitions, after all, are not how we learn most of what we know, and even when we do learn about what something is by way of a definition, there is after all an education and a competence in the language which comes before the definition and there is the trying the definition out, the education which follows hearing it. How do we acquire our understanding of what is just or fair? Think of little children learning that “It’s not fair!” means something different from “I don’t like it!” It may be that at some time articulating a definition will play a role, but understanding the force of that definition is more like understanding that we are justified in a certain tone of voice when we say, "That's not fair." We come to understand the word science or the word art or the word love not by way of a definition but by way of an education, and an education is a long, complex business which requires life outside of school as well as school. Sometimes, rather than a definition, we need to remind ourselves of how we came to be able to use the word properly. We learn some about love before we can say the word, and some more from examples of hatred, a lot from our first love and first breakup, and so on and on, all that time and all those adventures and disappointments and examples of love and of things which we mistake for love, giving us opportunities to deepen our understanding. What part is played in that by definitions? Very little.

              Another possibility: Socrates' impulse to squash examples, if we can keep our critical faculties in the face of his questions, might be a mistake. (There's even a line of thought that suggests that the last move in the pattern above and the pervasive use of irony in Plato might be evidence that the dialogues are trying to get us to see that mistake, and that reading Plato as a Platonist recommending definitional agendas has been a laughable mistake, a not-getting-the-joke type of mistake–in other words, Socrates is being ironic when he dismisses examples.) The idea that if we can identify examples then there must be a way we do the identifying and that way is what we want to articulate--well, that's a toughie. It's an argument that is very difficult to deal with, especially if we insist on thinking about it at an abstract level. But the criterion argument and Meno's paradox are also toughies and lead to the claim that possibly Socrates has reversed the logical priority of definitions and examples. That is, what is the test when examples and definitions do not agree? Granted, there are borderline cases and there are cases which don't seem to be either the thing defined or not the thing defined, and there are cases where we want to know more--but the thing to say about those cases is not to give in to the temptation to find a rule which says yes it is, really, or no it is not, but rather to say that it's a borderline case or a case which is problematic or a case where we want to know more. And given cases of knowledge or cases of science or cases of fairness or cases of love, where we know the cases are good cases (and there are such cases), or we know the cases are marginal or problematic, if the definition does not agree or confirm what we know about the cases, then it's the definition that goes in the crapper, not the examples. To the extent this is so, examples are logically prior to definitions. And attention to particular cases and to the issues which arise in them lead to a reduction in importance of these issues regarding definitions while we work through the relevant arguments and follow the good ones where they lead. (Is there a science of astrology? Is the new physics still science or is it now a kind of woowoo subjectivism gone bonkers? Can we now take Fermat's Last Theorem as proven? Do we know that the indigenous peoples of the Americas arrived by way of a land bridge from Siberia? Is the constancy of Hubble's Constant known? On what basis? --and so on.)

              Somewhere in here the question arises: Why are we enamored of definitions? What did we think they would do for us? My suggestion is that we have to watch out for the following possibility: We want a definition so that we don’t have to think so hard about the thing defined or its relation to cases. We use definitions as shortcuts. Perhaps this is not always what is going on, but a slight suspicion is enough to raise the issue, and when it looks like a request for a definition is a request to get out of thinking, just say no.

              And then here is a note about a rhetorical question: why couldn't we define love, or art, or science, or knowledge by giving a bunch of examples and then explaining them as we look hard at them?

              More and more it looks like the right answer to the rhetorical question is, well, maybe we could, and maybe we should. That would be more like the education by which we have come to use the word properly. The examples in Plato's dialogues may be more important than we had thought. Theoretical accounts by philosophers (or by philosophically-smitten artists) are less important than a grounding in art history, which takes years (and never really ends). Abstract characterizations of science have to be based on knowing a perspicuous array of varieties of scientific work. A philosophical definition of knowledge based on necessary and sufficient conditions will be tested by its consistency with examples in which people correctly say they know or do not know or ask about knowing or deny that they know different kinds of things.

                             d:\wpwin\wpdocs\courses\deflimn.s03