On Philosophy

April 3, 2007

Definition And Discovery

Filed under: Metaphilosophy — Peter @ 12:00 am

Is it possible to discover what a word means? Certainly it is possible to define a word into existence, for example the kind Q =df things that fly and weigh less than 10 grams. But it seems like to even use a word we must possess some definition of it, even if that definition is an unconscious one. Q, for example, couldn’t have been used meaningfully by us unless we had previously defined it. If we hadn’t an application of Q would be completely arbitrary, which is another way of saying that Q would be meaningless. Therefore if a word is meaningful we can conclude that there are some rules that govern its application, even if we aren’t able to explicitly state what those rules are. Which would seem to put an end to the possibility of philosophy to provide us with any new knowledge. For example, it wouldn’t make sense to investigate knowledge in the attempt to discover something new about it, or to better understand it, because we already possess a complete definition of what knowledge is, as evidenced by our ability to consistently apply the term [1], although we may be unaware of it. And thus the only reasonable investigation, in light of these facts, is to simply try to uncover what that definition is, perhaps by investigating our intuitions. [2]

But it is equally obvious that the above can’t be completely correct; although it may seem possible to deny that we can discover facts about what knowledge is, we cannot deny that we can discover new things about electrons. But certainly, since we use the word “electron” meaningfully we must have some criterion that tells us when something is or isn’t an electron. And another word for that criterion is the definition of what it means to be an electron. So it would seem that there is nothing we can discover about electrons either, since we possess a complete definition of what it is to be an electron. And so there seems to be a tension between understanding what a word means and being able to discover something about it. Because to discover something new implies that our definition was somehow incomplete, or wrong. But if that is the case then how do we know what we are discovering facts about? [3]

One possible solution to this problem is to argue that a discovery about electrons is not really a discovery at all, but rather a new definition, and that we are really dealing with two words, the word electron before the “discovery” and the word electron after the “discovery” that are simply spelled the same way. But this is a poor solution, because it creates more problems than it solves. For starters it would seem to make all scientific progress simply a matter of picking one language over another, almost arbitrarily. But it is pretty clear that the progress of science is not arbitrary, and that it is getting closer to the truth, as evidenced by the increasing accuracy of scientific predictions. Moreover, it makes the whole scientific process and scientific debate incomprehensible. How could we understand the point of a discussion between two physicists arguing about whether it is possible for particles to have factional charge, unless we recognize that they are both talking about the same things when they talk about particles, despite their giving to them different properties?

The real solution is to realize that not all definitions are categorical. A categorical definition simply picks out a class of thing in terms of properties. Our definition of Q, for example, is a categorical definition. In contrast we have what I call ostensive definitions, which are of the form “the thing that …” or “the kind of thing that …”. Both forms contain a list of properties, like a categorical definition. However, they both forms of the ostensive defintion appeal to some real object, not explicitly defined, that has those properties. And it is understood, as part of this kind of definition that the thing, or kind of thing, is natural division of the world, which in the case of a kind means that it consists of members sharing basically the same properties (except location), and in the case of a single thing that it is in some way separated from its environs. I know this may not be perfectly clear, so consider the following example that illustrates the difference between categorical and ostensive definitions. Obviously even before we had the atomic theory of matter the ancients had the concept of gold. Since they didn’t apply the word randomly we can assume they had some definition of the term. Consider the possibility that it was a categorical definition. If that was the case then gold =df a malleable, yellow, … substance. Now let us further assume that they came across what we call white gold. If they decided that it too was gold, for whatever reason, then it would be the case that their definition of gold changed, to gold =df a malleable, yellow or white, … substance. Obviously this is absurd. If they changed their definition to include white gold as one of the things that counts as gold not only do we have to assume that they all adopted a new definition of the word gold, but also that their decision to do so was basically arbitrary. Why not drop “malleable” and make pyrite one of the things that counts as gold? Thus gold is better thought of as having an ostensive definition, namely gold =df the kind of thing that is a malleable, yellow, … substance. Now in this case it turns out that the kind “pointed to” by this ostensive definition is gold atoms, and this means that white gold is really gold, since it too consists mostly of gold atoms, and that something that had many of the same macroscopic properties as gold, but which wasn’t made up of gold atoms would not in fact be gold, even if we couldn’t tell it apart from real gold under most circumstances.[4] And thus, most importantly, we can make sense of what it means to discover a fact about something defined ostensively, namely that we are discovering a fact about the thing or kind of thing that the definition “points” to.

Obviously this is the better analysis of scientific terms. And it is probably a better analysis for most everyday terms as well. For example “that chair in the corner” is probably best defined as “the thing that looks like a chair, feels solid, seems to be in the corner, …” instead of “the chair-looking, solid feeling thing in the corner”, because the first is compatible with a scientific worldview; even though “that thing” turns out to be a collection of particles that isn’t really solid, and doesn’t have many of the properties that we think of the chair as having (such as being a single thing), it still is what we are talking about, since it is what is responsible for the fact that it seems like there is a single solid object in the corner, and hence the ostensive definition “points” to it. In contrast, if we take the categorical approach then science seems to flat-out deny that there are chairs, or at the very least makes our talk of chairs seem like a serious confusion. In fact I think that we define very few things categorically, and when we do it is obvious that we are simply constructing the category from whole cloth, and not that we are trying to talk about some feature of the world. And thus it seems likely to me that the objects of philosophy are probably described ostensively as well. For example, knowledge might be defined as “the kind of statements that we should believe if we want to be right.” Since the definition is ostensive there can be an informative philosophical investigation of knowledge, into what kind of beliefs fit that definition, and what makes them that kind of beliefs. Which means that we can discover new facts about knowledge. And, in addition, that we can’t simply introspect on what knowledge means to us in order to reveal facts about it, because even if we can uncover our complete ostensive definition that definition only points to knowledge, it doesn’t tell us facts about it directly. And so there is real work for philosophy to do, the possibility for real progress to be made, and moreover that reflecting on our intuitions is a flawed philosophical method.


1. Of course different people may use the word knowledge inconsistently, but all that shows is that they are working with different definitions. A single person does not flip flop on whether a belief is properly called knowledge in a given description of a situation, although they may apply the term differently in very similar situations, or change their mind about its application when different facts are brought to their attention. But none of this undermines the idea that they have some criterion for applying the word knowledge.

2. Perhaps it is reasoning like this that encourages some philosophers to practice conceptual analysis.

3. You might think that we simply are adding to the definition of what an electron is when we claim to discover new facts about it. But consider electron =df a particle with positive charge. If we “discover” the fact that electrons have a charge of exactly +1 then we aren’t just adding to the definition, we are changing it, because now a theoretical particle with a charge of +2 is no longer an electron. Or, in other words, even adding new properties to a definition is a change, and not a refinement, because it excludes some things that previously were considered covered by that term. And of course it thus makes the idea that we can “discover” that electrons have +1 charge questionable. Even if all electrons have that charge the fact that we have discovered is only that all observed electrons have +1 charge, not that electrons must have a +1 charge by definition. To add that fact to the definition is really just to define a sub-class of the particles we previously called electrons. And even if all particles previously classified as electrons fall into this subclass we can’t simply equate the two. And so at best it would seem that we “discovered” the existence of a prevalent sub-class of electrons, not a new fact about electrons.

4. Obviously, because of imposters like this, an ostensive definition is possibly best phrased as: the kind of thing that is what we usually experience as …, but that is needlessly complicated for our purposes here.

Create a free website or blog at WordPress.com.