Necessary Truths

Alex Byrne and Ned Hall

Philosophical Analysis in the Twentieth Century, Volume 1: The Dawn of Analysis and Volume 2: The Age of Meaning
Scott Soames
Princeton University Press, $35 and $35 (cloth)

8 Philosophical Analysis in the Twentieth Century is a marvelous introduction to analytic philosophy. The two volumes unfold as a series of studies of some of the most important and influential philosophers in the analytic tradition, from its early 20th-century roots in the work of G.E. Moore and Bertrand Russell through Saul Kripke’s pioneering advances in metaphysics and the philosophy of language. (As Soames notes, there are some major, albeit excusable, omissions, notably the logician Gottlob Frege and the political philosopher John Rawls.) No place is made for Russell’s and A.J. Ayer’s womanizing, or for Ludwig Wittgenstein’s influence on literary theory or Gibert Ryle’s on the rock group The Police; the volumes are for those with a taste for philosophical ideas and, even more, for philosophical arguments, with explicit premises and conclusions and the essential apparatus of distinctions and qualifications. It is a philosopher’s history of analytic philosophy, with a careful and critical assessment of ideas about truth, morality, logic, mind, and meaning.

What is analytic philosophy? One familiar answer contrasts analytic philosophy with so-called Continental philosophy, whose major figures, in addition to Hegel, include Schopenhauer and Heidegger. But the labels probably hinder more than they help, not least because the study of Continental figures by card-carrying analytic philosophers is a thriving contemporary concern.¹ Another answer is that analytic philosophy is a body of theory or doctrine, but disagreement between analytic philosophers is widespread, and there is no substantial common core. If anything unites those who today would label themselves “analytic philosophers,” it is a method—one that stresses the importance of clarity and rigorous argument and sees as its end product truth and knowledge, not harmony with the universe or promotion of the good. (Analytic philosophers, as Soames remarks, have rarely given advice on how one should live; this reticence has probably been wise.)

Beyond this method, the analytic tradition in philosophy can be defined historically. It begins in the late 19th century with Gottlob Frege’s work on logic and the foundations of arithmetic and Russell and Moore’s reaction against “absolute idealism”—a doctrine derived from Kant via Hegel that was then prevalent in Britain (in the work of T.H. Green, Bernard Bosanquet, and F.H. Bradley). According to the absolute idealists, the world is one gigantic mind, or—to put it more impressively—the Absolute is experience.

As Soames observes in the introduction to the first volume, “Analytic philosophy is a trail of influence”: the history of analytic philosophy is the history of a continuing dialogue, in the course of which terms and distinctions crucial to philosophical debates become steadily sharper, standards of argument become steadily more exacting, and occasional imaginative breakthroughs transform the way philosophical problems are posed. Soames is a particularly appropriate historian: he has made important contributions to contemporary philosophy of language, and his own talent for clarity and rigor testifies to one important kind of progress analytic philosophy has made in the last hundred years.

Given that analytic philosophy is not distinguished by a body of received answers to philosophical questions, has it made progress of another kind? Soames offers the following suggestion:

To my mind the two most important achievements that have emerged from the analytic tradition in this period are (i) the recognition that philosophical speculation must be grounded in pre-philosophical thought, and (ii) the success achieved in understanding, and separating one from another, the fundamental methodological notions of logical consequence, logical truth, necessary truth, and apriori truth.

Since the analytic philosophy tent is now rather large, these claims are not entirely uncontroversial. (For what it’s worth, we think Soames has it more or less right.) The rest of this review elaborates on points (i) and (ii). The first can be illustrated by the beginning of Soames’s history: G.E. Moore’s epistemological writings. The second—and the first as well—can be illustrated by the end: Kripke’s book Naming and Necessity.

G.E. Moore is best known as the author of, in Virginia Woolf’s phrase, “the book that made us all so wise and good: Principia Ethica,” published in 1903. Principia Ethica did set the agenda for much contemporary work in ethics, but Moore had a more profound and lasting influence on the way analytic philosophy conceives of its methods and aims. When having tea in Russell’s rooms at Oxford, Moore heard J.M.E. McTaggart, one of the British Hegelians, propound his view that time is unreal. Moore was apparently horrified. As he recounts, “This must have seemed to me then (as it still does) a perfectly monstrous proposition, and I did my best to argue against it.” That was an early indication of Moore’s “recognition that philosophical speculation must be grounded in pre-philosophical thought.”

In his 1925 paper “A Defense of Common Sense,” Moore sets out, with excruciating clarity, what he calls the “Common Sense view of the world.” This comprises, first, truistic-sounding claims like “I have a human body,” “The earth has existed for many years past,” and, second, claims that we know claims of the first sort (“I know that I have a human body”). “I am,” he writes, “one of those philosophers who have held that the ‘Common Sense view of the world’ is, in certain fundamental features, wholly true.” And, Moore thinks, philosophers like McTaggart (tacitly) agree—the difference being that McTaggart’s overall view is inconsistent. In addition to holding that the Earth has existed for many years past, McTaggart holds a metaphysical doctrine that conflicts with this piece of common sense. Moore’s point is that if a philosopher produces an argument for wild conclusions—that time is unreal, that material bodies do not exist, or that we do not have any knowledge at all—then even though the argument might at first seem cogent, it must be flawed, and the philosopher’s responsibility is to figure out where things have gone off the rails.

As Soames emphasizes, this is a quite radical way of looking at philosophy. Many philosophers before Moore (and, in fact, many philosophers after him) recognize no such constraint on philosophical theorizing.² For example, according to Spinoza, there is exactly one thing, namely God; according to Hume, we have no reason to think that bread will nourish tomorrow (or that there will be any bread tomorrow, for that matter); according to Ayer, moral judgments are mere expressions of attitudes of approval and disapproval, and hence are neither true nor false. Common sense, these philosophers might say, is of course where theorizing of any sort starts, but there is no reason why it should remain securely in place when the theorizing ends—that the revolution begins in the ancien régime cannot stop all the aristocracy’s heads from rolling. A nice illustration of this attitude is in Russell’s The Problems of Philosophy, published in 1912, where it only takes him a few pages of argument to declare calmly that the question “Is there a real table at all?” is “very difficult.”

Philosophical skepticism—the claim that we do not know anything about the world around us—can serve to highlight the issue between Moore’s common-sense philosophy and the rival unconstrained kind. Consider an updated Matrix-style version of the skeptical scenarios that Descartes entertains in his First Meditation. Descartes worries that his ordinary beliefs about his environment and his own body might have been formed while dreaming, or produced by an evil, deceiving demon; the contemporary version asks us to imagine that we might be brains in vats, being stimulated by a computer in such a way as to give us exactly the same experiences that we now have. It seems to you right now that you are holding Boston Review, but perhaps that is because you are an appropriately stimulated brain in a vat who has never held a periodical in its life. How do you know that you are not in the vat scenario? After all, if you were in the vat scenario, you would still believe, as you now do, that you are holding Boston Review. If we stipulate that there are no familiar objects like shoes, string, sealing wax, and cabbages in the vat scenario, but that vat experiences are subjectively identical to non-vat experiences, then it appears you don’t know that there are any such objects. One might, then, demand some reassurance from expert philosophers that shoes and the like do exist. Descartes purported to provide it, but his solution—that a non-deceiving God ensures that our clear and distinct ideas correspond to the way the world is—did not carry much conviction. In the Critique of Pure Reason, Kant famously complained:

It still remains a scandal to philosophy . . . that the existence of things outside of us . . . must be accepted merely on faith, and that, if anyone thinks good to doubt their existence, we are unable to counter his doubts by any satisfactory proof.

Moore’s paper “Proof of an External World” starts with this quotation from Kant, and he delivers on his title by providing the following disarmingly simple proof. Moore holds up each of his hands, notes that they exist, and concludes (since his hands are, in the appropriate sense, external objects), that an external world exists. To this notorious “proof,” the near-universal reaction of the beginning student of epistemology is bewilderment: how could Moore possibly have thought that a skeptic’s profound doubts can be countered by a bit of hand-waving? After all, anyone who thinks that he might, for all he knows, be a brain in a vat will hardly accept Moore’s premises.

In Soames’s view, countering the skeptic’s doubts by proving him wrong was not Moore’s point at all. Rather, he is best read as arguing that no proof of the external world is required, so there is no such “scandal.” The skeptic’s insistence that the common-sense view is guilty until proved innocent rests on a mistaken assumption about the relation of philosophy to truisms about hands and such. The claim that one has hands is not something that needs to be supported by philosophical theories. On the contrary, if any philosophical theory—like McTaggart’s denial of the reality of time—conflicts with such obvious truths, then the only sensible conclusion is that something has gone badly wrong with the philosophical theory. If we still feel the pull of the theory, we should try to diagnose the temptation or locate the flaw in the argument—upending the apple cart of common sense is not an option to be taken seriously.

We now turn to the second kind of progress identified by Soames, the separation of certain “fundamental methodological notions,” and in particular the separation of three distinctions: between analytic and synthetic truths, necessary and contingent truths, and a priori and a posteriori truths.

The first distinction is couched in Kantian terminology, although the usual way of explaining the distinction is not Kant’s. An analytic truth is one that is “true solely in virtue of meaning”: any list of examples invariably includes “Bachelors are unmarried”; others might be “If Socrates drank hemlock quickly then he drank hemlock” and “Quadrangles have four sides.” Merely understanding these sentences puts one in a position to know that they are true. A truth is synthetic if and only if it is not analytic: “Ludwig Wittgenstein was a bachelor,” “Quadrangles are found in Oxford,” and “If Socrates drank hemlock then he died of hemlock poisoning” are all synthetic.

The second distinction is between necessary and contingent truths. A necessary truth is one that could not have been false, one that would have been true no matter how things had turned out. As Leibniz put it, a necessary truth is one that is “true in all possible worlds.” Plausible examples include “17 is prime,” “If Moore is a bachelor, he is unmarried,” and so on.

The third distinction is between truths knowable a priori and those knowable only a posteriori. An a priori truth is one that is knowable independently of experience, or without empirical evidence. Plausible examples include “9 = 32,” “Either Gilbert Ryle was a plumber or he wasn’t,” and the like. A truth is knowable only a posteriori (“on the basis of experience”) just in case it is knowable but not knowable a priori: “The number of planets is nine” or “Gilbert Ryle was a bachelor,” for instance.

Given the explanations sketched above, one might fairly suspect that the three pairs of distinctions coincide. That is: perhaps a truth is analytic if and only if it is necessary, and if and only if it is knowable a priori. And, in much of twentieth century analytic philosophy, either the equivalence of these distinctions was taken for granted or—worse—the distinctions were simply conflated.³ As Soames nicely brings out, a number of arguments from the period smuggle in the equivalence as a tacit premise. The essential point for the evolution of 20th-century philosophy is that once the equivalence is accepted, an extremely humble conception of philosophy itself is only a few steps away.

Here is one route to this humble conception. First, philosophical claims themselves are often taken to be both necessary and a priori: according to many philosophers, the business of philosophy is to deliver truths that describe what the world must be like and that are knowable by reason alone. Second, since analytic truths are “true in virtue of meaning,” and since what a word means is a conventional matter, it is natural to think that analytic truths are somehow “true by convention”—that the basis of their truth lies in our decisions or intentions concerning the use of words, rather than in the (extra-linguistic) world. As Ayer put it, analytic truths are true “simply because we never allow them to be anything else.” But now, if we assume the equivalence, then necessary and a priori truths are also true by convention, in which case they are not about the world either. And since the truths of philosophy are supposed to be necessary and a priori, this means that philosophy, contrary to the traditional advertisement, is not about the world, let alone the Ultimate Nature of Reality. “The propositions of philosophy are not factual,” Ayer announced, “but linguistic in character . . . they express definitions, or the formal consequences of definitions.” Philosophy is turned into the analysis of language: thus the title of Rudolf Carnap’s classic 1932 paper, “The Elimination of Metaphysics through Logical Analysis of Language.”

In 1970 Saul Kripke, then a young professor at Princeton University, gave a series of three long lectures that were later published as a book, Naming and Necessity. Here is Soames’s assessment:

In the philosophy of language, Naming and Necessity is among the most important works ever, ranking with the classical work of Frege in the late nineteenth century, and of Russell and Tarski in the first half of the twentieth century . . . Naming and Necessity played a large role in the implicit, but widespread, rejection of the view—so popular among ordinary language philosophers—that philosophy is nothing more than the analysis of language.

One of the specific contributions of Kripke’s book was to pull apart the three distinctions explained above—in particular, the distinctions between a posteriori and a priori truths and between contingent and necessary truths. Kripke begins by noting that these distinctions are couched in very different terms. The a priori–a posteriori distinction is an epistemological distinction, concerning how a truth may be known. By contrast, the necessary–contingent distinction is not explained in epistemological terms at all: it is about how things must or might be. So, Kripke says, it should not be taken for granted that the two distinctions come to the same thing. In fact, Kripke argues, they do not. Naming and Necessity announced a pair of ideas that represented a startling departure from conventional wisdom: that not every necessary truth is a priori (some are a posteriori) and that not every contingent truth is a posteriori (some are a priori).

Let us concentrate on the first claim. What sorts of truths are necessary but knowable only by observation and experiment? Kripke proposes two kinds of examples. The first might include Snoop Dogg = Calvin Broadus and Eminem = Marshall Mathers: true identities of the form a = b, where a and b are proper names. These truths are necessary: the man Snoop Dogg might not have been called Calvin Broadus (he might not have been called Snoop Dogg either), but he could not have failed to have been the very same individual as Calvin Broadus. Further, Kripke says, these sorts of truths are plainly knowable only a posteriori: one cannot know from the armchair that Snoop is Calvin; some empirical spadework—reading Rolling Stone, for instance—is required.

Another class of examples comprises true “theoretical identifications,” like “Light is a stream of photons” and “Gold is the element with atomic number 79.” These truths are necessary: although there could have been something with the look and feel of gold (ductile, yellow, etc.) whose atomic number was not 79, this would not have been gold. Furthermore, these truths are knowable only a posteriori—armchair chemistry is a mug’s game.

Soames agrees that the examples in the second category are genuine cases of the necessary a posteriori—of truths that can only be known on the basis of experience but that could not be false. But he disputes Kripke’s description of the first class.

To appreciate Soames’s point we need to introduce a distinction that we have so far deliberately fudged: one between true sentences and the truths expressed by such sentences (true propositions, in philosophers’ jargon). “Kripke is a philosopher” and “Kripke es un filósofo” are true sentences of, respectively, English and Spanish. They are different sentences, yet they express the very same true proposition, namely that Kripke is a philosopher. Things known are true propositions, not true sentences: thus a monolingual English speaker and a monolingual Spanish speaker might know the very same thing, namely that Kripke is a philosopher. Similarly, it is propositions that are necessary or contingent (sentences count as necessary or contingent only derivatively, according to whether they express necessary or contingent propositions).⁴

With that distinction made, note that the proposition that Kripke is a philosopher is not a proposition about language, although (of course) it is a proposition that can be expressed by sentences of various languages. In particular, the proposition that Kripke is a philosopher is not identical to the proposition that “Kripke is a philosopher” is a true sentence of English. So, for example, a monolingual Spanish speaker might know that Kripke is a philosopher without knowing that “Kripke is a philosopher” is a true sentence of English. For similar reasons, the proposition that Snoop Dogg is Calvin Broadus is not identical to the proposition that “Snoop Dogg” and “Calvin Broadus” are names of the same individual.

The second proposition, the one about the two names, is uncontroversially knowable only a posteriori. But it is contingent: it might have been that Snoop was content to be called simply “Calvin Broadus” while Wittgenstein’s parents dubbed him “Snoop Dogg.” Hence the second proposition is not an example of the necessary a posteriori. What about the proposition that Snoop Dogg is Calvin Broadus? As Kripke argued, it is necessary. But is it knowable only a posteriori? Soames argues that it is not and makes his case by drawing on another important theme in Naming and Necessity, that names are not abbreviated descriptions; for example, “Gilbert Ryle” is not short for “The pipe-smoking author of The Concept of Mind” or any other kind of description. (The “description theory of names,” in one more or less subtle form or another, can be found in Frege, Russell, and Wittgenstein.) Kripke’s attack on the description theory of names can be seen as supporting the view that the meaning of a name is simply its referent: the meaning of “Snoop Dogg” is simply the individual Snoop, and likewise for “Calvin Broadus.” Kripke himself did not go quite that far, but Soames argues that he should have. So, on this view, “Calvin Broadus” has the same meaning as “Snoop Dogg,” which is tantamount to saying that “Calvin Broadus is Calvin Broadus” expresses the same proposition as “Snoop Dogg is Calvin Broadus.” And since one may know a priori that Calvin Broadus is Calvin Broadus, it follows that one may know a priori that Snoop Dogg is Calvin Broadus—to know the one proposition is to know the other, because they are the same proposition. So, Soames thinks, Kripke was in error to claim that identities between names provided examples of the necessary a posteriori—but the necessary corrective can be found in Naming and Necessity itself.

In part by helping to lay to rest the aforementioned misconception of philosophical inquiry as an analysis of language—little more than specialized lexicography—Naming and Necessity played a large role in rehabilitating a more traditional conception of philosophy, where the philosopher fearlessly follows her argument no matter which garden path it leads down, and philosophical knowledge is about reality at large, not simply about linguistic conventions. While Carnap announced the elimination of metaphysics through logical analysis of language, Kripke announced its rehabilitation by rejecting the conception of philosophy as consisting entirely in such analysis. But at the same time, Kripke’s own philosophical methodology is decidedly Moorean, as Soames points out.An example of Kripke’s Mooreanism concerns the distinction between an object’s essential properties—those that the object could not possibly have lacked—and its other, accidental, properties. For example, one might think that Gilbert Ryle was essentially human but only accidentally a pipe smoker: although Ryle could not have been an alligator, he might have taken up gum chewing instead of pipe smoking. Some philosophers—most notably W.V. Quine—were highly suspicious of modal notions like possibility and necessity, and thought that this alleged distinction between the essential and the accidental amounted to metaphysical mystery-mongering. Kripke, on the other hand, thinks the distinction is a perfectly intuitive one that we recognize in ordinary life:

When you ask whether it is necessary or contingent that Nixon won the election, you are asking the intuitive question whether in some counterfactual situation, this man would in fact have lost the election. If someone thinks that the notion of a necessary or contingent property . . . is a philosopher’s notion with no intuitive content, he is wrong. Of course, some philosophers think that something’s having intuitive content is very inconclusive evidence in favor of it. I think it is very heavy evidence in favor of anything, myself. I really don’t know, in a way, what more conclusive evidence one can have about anything, ultimately speaking. But, in any event, people who think the notion of accidental property unintuitive have intuition reversed, I think.

The intelligibility of questions about whether this man could have won the election, Kripke might have said, is part of the “Common Sense view of the world.” While dismissing metaphysical ideas about possibility and necessity may sound sensible and down-to-earth, in fact it requires, Kripke says, rejecting claims that we all believe in everyday life. If a philosopher such as Quine purports to find such questions unintelligible, this is (almost) certainly to be traced to a misconception on Quine’s part.

As we observed earlier, there is no consensus on whether, or to what degree, philosophy should be subordinate to common sense. Further, there is a significant divide even among philosophers who accept a broadly Moorean outlook. On the ambitious view, philosophy can discover heady metaphysical truths that are consistent with common sense, but supplement it. On the modest view, little or no such supplementation is likely to be forthcoming.

An example of an ambitious Moorean is the late David Lewis—a giant of contemporary philosophy and an erstwhile colleague of Soames. Consider the claim that if the Supreme Court had not stopped the Florida recount, Al Gore would have won the 2000 election, and suppose that it is true. According to Lewis, what makes this claim true are facts about an existing alternative universe (or “concrete possible world”) in which a flesh-and-blood Gore doppelganger does win a presidential election. This universe, according to Lewis, is spatiotemporally disconnected from our own, but is no less real for it. Lewis is not denying anything that is plausibly part of common sense—rather, he is announcing some astonishing news about what reality has to be like, if various common-sense claims are true. In contrast, Kripke, Moore, and Soames himself are closer to the modest end of the spectrum. (Moore is actually a rather complicated case; at times he seems to consider claims that are arguably in conflict with common sense to be live philosophical options.)

Who is right? Philosophical Analysis in the Twentieth Century provides some reason for modesty. Ambitious philosophy is extremely hard. Philosophical arguments for really exciting conclusions invariably have some subtle flaw, as Soames demonstrates many times. That philosophers have become considerably better at diagnosing such flaws is progress, but at the same time this progress has exposed how high the bar really is.

It would not be wildly off the mark to take these two volumes as vindicating a remark of Wittgenstein’s from his Philosophical Investigations: that philosophy “leaves everything as it is.” To be sure, philosophy is not—as Wittgenstein also thought—“a battle against the bewitchment of our intelligence by means of language.” But to a large extent the analytic philosophy of the 20th century has been a battle against the bewitchment of our intelligence by other philosophers. <

Alex Byrne teaches philosophy at MIT. He has co-edited two collections of papers on color, Readings on Color, Volume 1: The Philosophy of Color and Volume 2: The Science of Color.

Ned Hall teaches philosophy at MIT. He has co-edited a collection of papers on causation, Causation and Counterfactuals.

Notes

¹ On the origins of the split between analytic and continental philosophy, see M. Friedman, A Parting of the Ways: Carnap, Cassirer, and Heidegger (Open Court, 2000).

² The 18th-century philosopher Thomas Reid, a founder of the Scottish “common sense” school and an important influence on Moore, is a notable exception. Reid’s adherence to common sense was not quite as pure as Moore’s, however.

³ Kant, incidentally, thought that some a priori and necessary truths were synthetic (for example, those of arithmetic). However, he did think that the a priori–a posteriori and necessary–contingent distinctions coincided.

⁴ The analytic–synthetic distinction, on the other hand, is best taken as applying primarily to sentences, and derivatively to propositions.

Originally published in the October/November 2004 issue of Boston Review.