The philosophies of language and mind, together with the theory of knowledge, have been at the very center of Western philosophy throughout this and the last century. Both their results and the techniques they employ have been widely adopted in every other area of philosophy. This essay concentrates on the philosophy of language. Later we shall see how this links up with thought, a basic theme in the philosophy of mind.
The philosophy of language might proceed in either of two ways. First, it is an approach to standard issues in philosophy through a study of the crucial linguistic expressions in which they are raised. This linguistically-oriented treatment of classical problems might also be called "linguistic philosophy". For example, we might approach moral philosophy by examining the implications of pre-philosophical occasions in which it is appropriate to predicate 'is good' of a person or 'is right' of an action Second, the philosophy of language may be construed as the systematic study of language itself—language as a philosophical topic in its own right. These applications intertwine, but I shall describe them separately for our purposes.
Inquiries into Beauty, Truth, Causation, Matter, Goodness and Authority are typical philosophical pursuits. But what sorts of creatures are these subject-matters? The classical answer is that they are concepts, but what is a concept? Ideally a concept has meaning and reference, and implies and is implied by other concepts. Philosophers have been converging on the view that concepts are suspiciously similar to bits of language. Thus, getting straight on the inguistic counterparts of concepts would be progress toward getting straight on the philosophical facts. Two warnings are in order. First, this need not be the end of the matter. Getting straight on the language in which our problems are presented may only allow us to pose better questions; but in some cases there may be profound issues that go beyond the linguistic facts. However, unclarity is no friend of progress, so it is imperative that at least we clarify the language in which we deliberate these questions. Second, although many of these concepts can be summed up in single words or phrases, it would be a mistake to suppose that the focus on language was a focus on a word in English or Vietnamese. Expressions for, say, truth (for example, the predicate 'is true') occur in many, perhaps all, natural languages. And we are interested only in what is common between the various languages that embody this concept. I illustrate with a few prominent examples.
Suppose we were to ask, "What is knowledge?" Philosophers have approached this question by taking an example of the use of "know" as it occurs in everyday speech. Take as our specimen an assertion of the sentence "Sam knows that Vietnam is in Asia". Let us then raise the question, "Under what conditions might this assertion be true?" The traditional answer (hinted at as early as Plato's Theaetetus) is, first, that the proposition—"that Vietnam is in Asia"—must be true (no one can actually know what is false), second, that Sam must accept it, and third, that Sam have some sort of justification for his acceptance of it. This account has been challenged, largely over the requirement that Sam needs to possess a justification. Counterexamples have been produced, and this has led to a lively debate. The details aren't important here. What matters is that the thesis and its rejection each turn on attempts to understand how we use sentences ascribing knowledge in everyday speech. These are inquiries into the linguistic facts.
Next, a philosophical tradition, dating at least to Rene Descartes (1596–1650), is that the knowledge an individual has of his own sensations and thoughts is the standard for all knowledge because it provides the least margin for error. While skeptics may doubt that we can attain certainty about the external world, we at least know the way things seem to us and what we are thinking. Thus, it was assumed, all attempts to construct knowledge about the world must start from one's own mental contents and work up to knowledge about the external world. In a work published in 1953 Ludwig Wittgenstein claimed that this implied that each individual had a private language. In this language an individual could name his own sensations and other mental contents, but the language in principle couldn't even be understood by anyone who didn't have direct access to those experiences. Wittgenstein argued that such a private language was impossible: language itself had to be governed by rules, and the existence of rules required the possibility that mistakes, and thus their corrections, were recognizable. In a private language, what seemed right would be right; the collapse of actually being right into merely seeming right meant that there was no rule guiding one's linguistic habits. Wittgenstein concluded that language could only be public; a public check on linguistic habits was needed for the existence of rules. Thus, the starting point in a private realm was the wrong approach to take to the study of knowledge. If correct, this overthrows several centuries of assumptions shared by epistemologists. As with the earlier example, Wittgenstein's conclusion has been vigorously debated, but the case illustrates how a basically linguistic approach to traditional issues has influenced current philosophy.
I turn now to the systematc study of language itself. Babylonian astronomers named the morning star "Phosphorus" and the evening star "Hesperus", only to discover later that they were the same—in fact, not a star at all, but the planet Venus. It was a significant astronomical discovery that Hesperus is Phosphorus. However, what it says is only that a certain celestial body is identical with itself, which is also stated by the trivial "Phosphorus is Phosphorus". Why is the one informative and the other wholly trite? In the late nineteenth century a German thinker, Gottlob Frege, attempted to answer this question by distinguishing between an expression's meaning (or sense) and its reference. While the words "Phosphorus" and "Hesperus" refer to the same thing, they present what they refer to under different modes (say, as the morning star and the evening star) and thus have different meanings. That difference accounts for the one between the important discovery and the triviality. Indeed, independently of this issue meaning is something we expect any genuine word, phrase, or sentence to possess. It distinguishes sounds or marks belonging to a language from mere noises or scribbles. For Frege, and his followers, meaning was something each user of a language could carry around in his head. It was a repository of knowledge about a word that enabled a speaker to use it. For example, if I know what the word "elephant" means, it gives me an ability to recognize elephants and distinguish them from, say, giraffes. It is crucial to this outlook that the meaning we carry in our heads determines the references of those terms. Although some philosophers continued to believe that meaning is no more than reference, Frege's distinction led to the development of a variety of theories of meaning and of reference, and to a robust literature about each.
Frege's distinction, and his application of it, are intuitively attractive. But they have been challenged in recent years by certain thought experiments—that is, imaginary, even if only remotely possible, scenarios in which we entertain circumstances that put a strain on our assumptions. Thus, we now know that water is a chemical compound of hydrogen and oxygen, H2O. Of course, we might be wrong, but for the sake of argument let us suppose that this is correct. Now imagine a twin-planet, Twin-Earth, which is exactly like Earth in almost every respect, and whose inhabitants speak languages that sound in almost every respect like our own languages. Just as we have a term "water" in English, there is a term in Twin-English for the liquid stuff that fills their lakes, rivers, and seas that sounds the same. But they have a radically different chemistry than we do, and what they call "water" has a radically different chemical composition, which we may abbreviate as XYZ. Thus, what they refer to by "water" is not what we refer to by our similar-sounding term. The upshot is that anything, in our language, that counts as water must be H2O, wherever in the universe it is found. No doubt, they would say something similar about us. But this doesn't depend on what is in our heads or theirs. That water is H2O is an empirical scientific discovery, not a matter of dictionary definition: perhaps most people in our world nowadays who are perfectly competent users of the term don't know that it is H2O. Certainly our ancestors who lived before the era of modern chemistry didn't know this, although we assume that they used these same terms as we do. (We make this assumption about their similar uses of words whenever we read historical works.) Thus, what we and the Twin-Earthians refer to with our use of the similar term "water" doesn't depend on what is in our heads, but in part on the environment in which we find ourselves. And if what is in the head doesn't determine reference, it doesn't serve the role that would make it (linguistic) meaning.
This too has resulted in a lively discussion between semantic internalists, who believe that reference is determined by what is in the heads of users of the language, and externalists, who believe that our physical and social environment determines how our words relate to the world.
The case illustrates how the study of language relates to that of thought. Typically thoughts are about, of, or that something. (Plausible exceptions are conscious sensations, such as a pain or itching.) Franz Brentano called this "Intentionality" and believed it was the definining feature of the mental. For example, one cannot believe without believing something, expect without expecting something, desire without desiring something, and so on. Call the things that thoughts are about their objects. Some of these objects, as in beliefs and expectations, will have the structure of sentences. But just as the sentence "Vietnam is in Asia" has a meaning, so also a belief that Vietnam is in Asia has a meaning (or a "content" as it is often called), and a different one from the belief that Hanoi is east of Philadelphia. If semantics is the study of the meaning (or content) and reference of words, phrases, and sentences, then we can call the study of mental objects psychosemantics. Taking belief as our model, philosophers typically called what is believed a thought or a proposition. Just as sentences are compounded from words, so also thoughts are compounds, and their constituents are typically called concepts. In this way, an externalist would say not only that the inhabitants of Twin-Earth mean something different when they utter the word "water", but that their concept is a different one. (Of course, an internalist would deny both claims.)
Thus, many of the same issues that get discussed in the philosophy of language translate smoothly into issues about the semantics of thoughts, which is a central topic in the philosophy of mind. In fact, some philosophers hold that linguistic meaning and reference is secondary, depending ultimately on the original, non-derivative content of objects of thought. This too, as almost everything in philosophy, has been an issue over which various philosophers disagree. It leads into a fascinating set of queries—such as whether animals other than humans have beliefs, concepts, and the like although they lack language.
I have repeatedly mentioned disputes between parties concerning all of these claims. They should not be regarded as an indication of the futility of philosophical inquiry. Quite the contrary, they demonstrate its robust health. The disagreements have been largely rational and moderate; they have led to fruitful clarification and greater precision about what is at issue. These, in turn, have yielded many insights of a sort that even competing theorists have been equally able to acknowledge and use to sharpen their views.