January 16, 1997
AN INTRODUCTION TO DEFLATIONISM
Abstract: My dissertation project centers on an inquiry into the nature of truth and the role it seems to play in certain explanations. I am looking at a particular kind of analysis of truth, and I am using considerations of truth's apparent explanatory role to test the adequacy of this kind of response to the question, "What is truth?"
Background: So, what is the analysis of truth I am investigating? Let me start by providing a little contemporary social context for the view. You will see by the end that what this really amounts to is a bit of a cheat in terms of providing a launch pad for the kind of view I'm after, but I think it will be useful rhetorically. So here goes: It is not uncommon around most humanities departments these days to hear the following sort of claims
There is no such thing as objective truth, or
There is no such thing as Truth (with a capital "T").
Often the matter is left at that, as if what this claim amounts to is clear. But there are actually at least a couple of different directions one could go from this starting point in developing this basic idea in more detail. The analysis of truth I examine in my dissertation is one that tries to make good on one of these directions. This kind of analysis is typically called deflationism about truth in the philosophical literature because it explains what it means to claim there is no such thing as truth by deflating the notion. What this means will become clearer below (I hope).
What I want to do today is introduce you to deflationism, give you a sense of what it claims, of what is radical about it (and what isn't), explain what motivates it, and finally hint at how this kind of analysis impacts the issue of how truth functions in certain explanations.
Truth as a Property: A moment ago I mentioned deflating truth as a way of making good on the claim that there is no such thing (as objective Truth). To understand what I mean by "deflating" it is helpful to give a sense of the kind of view deflationism is supposed to contrast with. Consider the following sort of claim:
(1) "Snow is white" is true.
Ok, compare claim (1) with
(2) "Snow is white" is three words long.
What is going on in (2)? In a very general sense, it is a lot like
(3) The Earth is (roughly) spherical.
In (2) and (3) what you are doing is picking out some thing and claiming that that thing has a particular property. In (2) you pick out a sentence (by putting quotation marks around it) and ascribe a property to it--the property of being three words long. So in (2) we are talking about the sentence; we are saying that it has this property.
Now look back at (1). It has the same form as (2)--it appears to be talking about a sentence and attributing a property to it, in this case the property of being true. This is undoubtedly the most natural reading of (1). On this reading, the question "What is truth?" becomes a question about the nature of this property. Traditionally, the response has been that it is a kind of relation: a sentence is true if it bears a certain relation to reality. One often sees claims like:
Truth is correspondence to reality
A sentence is true if it corresponds to the facts
A sentence is true if it accurately represents the world.
These claims are usually considered platitudes--claims that seem true almost by definition. The problem is that they all involve notions that are just as mysterious as truth--"correspondence", "the facts", "accurately represents"--and no one has ever been able to do a satisfactory job of explaining how the correspondence works, or what facts are, or what something representing something else involves. In fact, these notions are themselves most easily explained in terms of truth, as in
A sentence corresponds to the facts if it is true (or what it says is true).
That's why the traditional platitudes about truth are platitudes, and therefore unilluminating.
The resistance the traditional approach has shown to further clarification of the kind of property it claims truth is can motivate different sorts of responses. Before I move on to the deflationary response, let me quickly point out a kind of reaction that is not deflationary. This different analysis agrees with the traditional approach that truth is a kind of property, but it disagrees about what kind of property it is. This view sees truth as something like a special way of being believed. Versions of this view claim, for instance,
A sentence is true if we all believe it
A sentence is true if most of us (or the most powerful of us) believe it
A sentence is true if it is what the available evidence suggests
or something of this sort. Being true is still a property on this kind of view--it is a complex property tied to the beliefs held by certain people, or to what evidence is available. To get a sense of the kind of property I mean here, look again at
(3) The Earth is (roughly) spherical.
It has the property of being believed (I assume) by all of us. On the kind of view under consideration now, truth is claimed to be a property of a similar sort.
This kind of analysis of truth as a property has its own special problems. For instance, on this sort of view, what is true--and I mean what is true, not merely what is (perhaps mistakenly) taken to be true--but what is true at one time (or in one place) might not be true at (or in) another. I am not talking about obviously context-sensitive claims (e.g., an indexical claim like "It is cold here now"), but claims that intuitively have a stable truth-value. Consider
(4) The Earth revolves around the Sun.
This claim was not believed by all, or by most, or by the most powerful among people six hundred years ago, nor (one might argue) was it suggested by the available evidence. [If anyone wants to argue that it was, then consider the claim that the planets revolve around the sun in elliptical orbits. The evidence for this claim was not available before Tycho Brahe made his detailed astronomical observations in the late 16th and early 17th centuries.] On this kind of view of truth, then, (4) was not true six hundred years ago, although it is true now. But if (4) was not true six hundred years ago, then the Earth did not revolve around the Sun six hundred years ago, and if (4) is true now, then the Earth does revolve around the Sun now. Many people consider this consequence unacceptable. They argue that it is hard to believe that in the past six hundred years there have been radical astronomical changes in such things as the relative sizes of the Earth and the Sun, and the distance between them.
Another problem with this kind of view is that it allows for circumstances in which it is impossible for certain people to be mistaken in what they believe. For instance, if being true is just being believed by everyone, then it is impossible for everyone to share an erroneous belief. But surely this is possible--in fact it is how I would want to describe the situation if at some point in history everyone (women included) believed that women were mentally inferior to men--I would want to say that they were all wrong. It seems to be part of the concept of belief that it is possible for us, even all (or most, or the most powerful) of us, to be wrong in what we believe.
Deflationism: We have looked at two ways of analyzing truth as a kind of property, and neither has turned out to be very satisfactory. What deflationism says at this point is that we should get the hint. Truth, according to deflationism, is not a property or a relation of any kind. The grammar of
(1) "Snow is white" is true
is misleading--it tricks us into thinking truth is a property the way the grammar of
(5) It is snowing
might trick you into thinking there must be some object that is doing the snowing. Deflationism claims that the right reaction to the problems generated by the traditional attempt to analyze truth as some relational property of correspondence is not to make the same kind of mistake all over again, but with a different analysis of what the property is like. What we need to do is deflate the notion of truth (you've probably been wondering when I'd get back to this); we need to stop thinking of truth as a property at all. So how should we think of it then? What should we say about truth?
The first thing a deflationist will say is that all there is to truth are our uses of the word "true". The question to consider is what this word is used for. The traditional view of truth says that the word "true" is used to ascribe a property--truth--to sentences. But deflationism denies there is any such property. To figure out what to say about truth on this view we need to look at the pragmatic and grammatical or logical uses of the word "true". I'll just mention a couple.
(I) Indirect Endorsement: a pragmatic use
Consider the following conversation.
Bob: It's been a cold week.
Mary: That's true.
Here Mary endorses what Bob says. In effect, she repeats his claim while simultaneously indicating that the claim did not originate with her. Alternatively, the conversation could have gone like this:
Bob: It's been a cold week.
Mary: It's been a cold week. [or, more realistically,--It has been a cold week.]
So it looks like we don't absolutely need to use "true" in this case. But consider
(6) What Bob said is true.
By using a claim like (6) I can commit myself to what Bob said, I can say the same thing indirectly, without specifying exactly what it was he said. I may not even know exactly what he said (maybe he mumbled), but I might want to "say" the same thing anyway (for instance, maybe I have complete faith in Bob's judgements and I want to support him against someone else's doubt). Using the word "true" as it is used in (6) gives me a way to do this.
(II) Expressing Infinite Claims: a logical use
Let's say I want to express a commitment to the thesis of papal infallibility to someone who doesn't understand what "papal infallibility" means. One way I might try to do this is to claim the following
(7) If the Pope says that snow is white, then snow is white, and if the Pope says that the moon is blue, then the moon is blue, and if the Pope says that the Earth is the center of the universe, then the Earth is the center of the universe, and if the Pope says that evolution occurs, then evolution occurs, and...
The problem is that to express the thesis fully I need to include a sentence of the form
If the Pope says that p, then p
for everything that the Pope could say. Even covering everything the Pope actually ends up saying in his lifetime isn't enough; I need p to cover everything it is possible for him to say. That's a lot of things; in fact, it's an infinite number of them. So really I couldn't possibly express what I want to say directly as begun in (7) because my life is finite.
So how do I say it? This is where the truth-predicate, the word "true", comes in handy. We use it to express what it would take an infinite claim to express like so
(8) Whatever the Pope says is true.
The word "true" thus lets you express infinite claims in a finite, even quick, way. That's a powerful expressive ability for a language to have.
Here, then, is the radical claim deflationism makes. What we should say about the word "true" is not that it attributes some kind of property--truth. Rather what we should say is that the word "true" is just a linguistic or grammatical device. Truth is just part of the grammar or logic of our language; it isn't a property out in the world (the way, say, being liquid is). It has the same metaphysical status as other logical devices such as the word "not". The word "not" is just a logical device that lets us transform a claim into its denial. For instance, I use "not" to transform "The sky is bright blue" into "The sky is not bright blue"--there is no property of "notness" I attributed to the original claim in doing this. Similarly, the word "true" is just a logical device that lets us (among other things) express infinite claims. There's no property of truth.
According to deflationism, both of the attempts to analyze the essence of truth failed because truth has no essence. What has always been controversial about truth is how to characterize it's deep underlying nature. Deflationism dissolves this controversy by claiming that truth has no underlying nature. All there is to truth are the surface grammatical or logical features of the word "true"; they can account for all of the uses we make of the notion of truth. Those features are completely uncontroversial and out in the open. In the metaphysical sense, truth is a light-weight notion. But this is not to say it is useless. We have already seen just how useful it can be.
So now look back at
(1) "Snow is white" is true.
What does deflationism say is going on here? There is no property of truth being attributed, so in that sense (1) is unlike
(2) "Snow is white" is three words long.
Claim (1) is not a claim about a sentence. It does not say there is a thing that has a certain property. So what does (1) say? According to deflationism, all it says is what is said by
(9) Snow is white.
Adding the predicate "is true" adds no content to a claim. In that sense it is like a double use of "not"--as in
(10) It is not the case that it is not the case that snow is white.
In a sense, this also has the same content as (9)--it just says that snow is white.
What the "...is true" part of (1) does is cancel out the force of the quotation marks around "Snow is white". What the truth-predicate does when it is applied directly to a sentence as in (1) is turn a mentioned sentence (one placed inside quotation marks so that we can talk about it) into a used sentence (one being employed to say something about the world). This aspect of the truth-predicate partially satisfies the intuitions behind the traditional view of truth--that it somehow involves corresponding to reality (to the facts). The correspondence involved in (1) on a deflationary analysis of truth is just that if the world is a certain way, then snow is white, that is, the sentence "snow is white" is true--since all this latter claim says is that snow is white.
What follows from this is that while deflationism claims that there is no such thing as truth (no such property), it does not claim that there are no sentences that are true, and might be so objectively. There may be no such thing as (objective) Truth, but if whether snow is white or whether the Earth revolves around the sun does not depend upon what we decide (beyond the trivial sense in which it depends on facts like that we have decided to use the word "snow" to pick out snow rather than to pick out coal), but rather depends on how the world is, then both (4) and (9) are still objective truths. That is the sense in which deflationism is not radical.
Explanation: Let me end by briefly indicating some problems that arise for deflationism when we turn to the task of accommodating the role that truth appears to play in certain explanations. If there is no property of truth, then how could truth be used to explain anything? It seems like an understanding of truth as light-weight or thin as that put forward by deflationism couldn't be explanatory. After all, according to deflationism, to say that something is true is not to say anything about it. But consider such explanatory claims as
(11) True beliefs tend to facilitate successful behavior
(12) The practical success of science is due to its tendency to produce (more and more approximately) true theories
It certainly looks as if truth is doing some explanatory work in these claims. Can deflationism make sense of truth's apparent explanatory role, or will accommodating this use of the notion force us to accept an analysis of truth that takes it to be a property of some kind? That is the main question facing deflationism about truth.