David Lewis was born on 28 September 1941 in Oberlin, Ohio. He received a BA from Swarthmore College in 1962, and a PhD in philosophy from Harvard University in 1967, where he worked under W. V. Quine . He joined the philosophy department at the University of California at Los Angeles in 1966. He moved to Princeton in 1970, and later became the Class of 1943 University Professor of Philosophy, holding this position until his death. He was a fellow of the American Academy of Arts and Sciences and an honorary fellow of the Australian Academy of the Humanities. He received honorary degrees from the universities of York, Cambridge and Melbourne. Lewis died on 14 October 2001 in Princeton, New Jersey.
Lewis is widely thought of as one of the leading philosophers of the later twentieth century. He wrote extremely influential books and articles on virtually every central topic in analytic philosophy, doing a good deal to set the agenda in metaphysics, philosophy of mind, philosophy of language, philosophical decision theory, and epistemology. His impact on metaphysics was particularly profound. He came to dominate the field, and is regarded by many as the twentieth century’s most important systematic metaphysician. The body of his philosophical work consists of four monographs and over a hundred articles, many of which appear in five volumes of his collected papers. In addition to achieving pre-eminence in American philosophical circles, Lewis had a particularly significant impact on the development of philosophy in Australia, where he often lived during American summers.
Lewis’s doctoral thesis was reworked into Convention (1969). The book deepens our understanding of the platitude that language is conventional, by first providing a general account of what a convention is, and then explaining what the conventions of language are within that general framework. Lewis’s basic idea was that a convention is a special kind of regularity because it is arbitrary, since the purposes it serves could just as well be served by something else. On the other hand, it is self-sustaining, on account of the fact that expected conformity by others induces one to conform oneself. It is arbitrary whether we drive on the left or the right; but supposing our expectation is that others drive on the right, we are given decisive reason to do the same. In the case of linguistic convention, the key regularity according to Convention was truthfulness in a language. A particular language like English involves a distinctive pairing system from words to meanings. To be truthful in English is to utter sentences that are true relative to that pairing system. Lewis’s ideas on linguistic conventions were developed and revised in a subsequent paper, “Languages and Language,” contained in Philosophical Papers Volume 1. Convention contains much that is highly intuitive, and yet also makes liberal use of somewhat technical material – in this case both game theory and formal semantics – in the service of clarity. This combination was to remain a pattern throughout Lewis’s career.
The notion of a possible world makes an important appearance in Convention. Lewis’s basic notion of the meaning of a sentence has two factors: the mood of the sentence (interrogative, declarative, etc.), and the truth condition associated with the sentence, which is a set of possible worlds. Take a simple case: Snow is white. The truth condition is the set of possible worlds where snow is white. Lewis’s use of possible worlds to clarify important concepts and issues was hardly isolated to this case. It was a persistent theme throughout his career, nowhere better exemplified than in his second book Counterfactuals (1973), which investigates those ways of thinking and talking that report what would have happened if such and such had been the case. No better summary of that book can be provided than the one contained in its own opening paragraph: “‘If kangaroos had no tails, they would topple over’ seems to me to mean something like this: in any possible state of affairs in which kangaroos have no tails, and which resembles our actual state of affairs as much as kangaroos having no tails permits it to, the kangaroos topple over. I shall give a general analysis of counterfactual conditionals along these lines.” Suppose a general says, “If I had attacked at dawn, I would have won the battle.” Certainly, there are possible histories of the world in which the general attacks at dawn and wins the battle. And there are possible histories in which he attacks at dawn and does not win. What determines the truth of the general’s speech? Lewis enjoins us to (roughly speaking) consider those possible histories which are very much like the actual world up until dawn, but which depart from the actual course of events with respect to the matter of an attack by the general. If the general wins in all of those “nearby” possible histories, the counterfactual is true. Otherwise it is false. Lewis’s basic idea was not altogether new. But by all accounts, the systematic and rich analysis that he provided put the topic of counterfactuals onto center stage in analytic philosophy and made such turns of phrase as “at the closest possible world” part of the philosophical lexicon.
With an analysis of counterfactuals ready at hand, philosophers were encouraged to use them to shed light on myriad areas of philosophical puzzlement. An example is Lewis’s own counterfactual analysis of causation, which begins with the simple thought that event x causes event y just in case, were x not to have occurred, y would not have occurred either. The simple thought needed refinement and qualification. Suppose my shooting causes your death, but that if I hadn’t shot you, someone else on the sidelines would. Then it is not true that if hadn’t shot you, you wouldn’t have died. Lewis was sensitive to these and other concerns, but he remained convinced that a satisfying account of causation could be built upon the counterfactual idea. His papers “Causation,” “Postscripts to Causation,” (in Philosophical Papers Volume II, 1986), and his more recent “Causation as Influence” (2000) sparked a large body of literature, much of which has been concerned, like him, to find a suitable version of the counterfactual idea, rather than to overthrow it altogether.
Lewis frequently, and famously, appealed to possible worlds as a tool in philosophical analysis. But what are possible worlds? To this question, Lewis offered a distinctive (and for many, somewhat shocking) answer, his so-called “modal realism.” Possible worlds are concrete universes, just as concrete as this one, many of which contain flesh and blood individuals just like ourselves. What makes it true that there is a possible history in which donkeys talk – assuming that such a history is at least possible – is the existence of a concrete universe in which donkeys are talking. Such was Lewis’s preferred picture of what a possible history really is. For the most part though, Lewis framed his philosophical analyses in such a way that they did not especially rely on that picture. For that reason, the popularity of his work was not much impeded by the fact that modal realism only attracted a relatively small following.
Why did Lewis believe the thesis that possible histories are bona fide concrete universes that really exist? Part of the story involves, surprisingly, a kind of conservatism. Unlike Quine, and like G. E. Moore, Lewis had a profound respect for common sense. Since common sense embraces possibility and necessity, possible individuals as well as actual ones, so should philosophers. But why the particular conception of possible worlds as concrete? Some understanding of the matter can be achieved by noticing some converging themes in Lewis’s philosophical temperament: on the matter of basic ontology, a nominalistic tendency to prefer a metaphysics of concrete objects over a metaphysics of abstracta; and on the matter of basic ideology, a preference for simple first-order languages (of the sort one learns in a basic predicate logic class) as a framework with which to describe the world over languages with primitive operators (such as primitive tense operators or, what is especially pertinent here, primitive modal operators). But more importantly, Lewis saw many problems and difficulties in the details of alternative approaches to possibility and necessity, many of which are laid out and explained in his third book On The Plurality of Worlds (1986). The metaphysical posit of an infinite plenitude of concrete universes, one for each possible history, was part of a grand metaphysical system that Lewis developed and refined throughout his career.
Lewis is a modern proponent of Hume’s view that there are no necessary connections between distinct existences. What goes on at one time does not necessitate what goes on at earlier and later times. In Lewis’s hands, this idea gets articulated in the form of “a principle of recombination,” the view that any possible series of events can precede any other possible series of events and that any possible series of events can fail to precede any other possible series of events. He concludes that the laws of nature are not necessary: “Episodes of bread-eating are possible because actual; as are episodes of starvation. Juxtapose duplicates of the two, on the grounds that anything can follow anything; here is a possible world that violates the law that bread nourishes.” (On The Plurality of Worlds, p. 91)
Lewis advocates a sort of anti-holism. The goings-on in the actual world (and possible worlds like the actual world) are determined by extremely local goings-on. Describe what is local to each and every point in space and time and you will have, in essence, captured the whole truth about the actual world.
The fundamental facts about the actual world are physical facts: they are the sorts of things that an ideal physics would capture. No fundamental fact about actual reality would be “left out” by an ideal physics.
Things share characteristics: being red; being negatively charged; being within ten feet of the Eiffel Tower; being forlorn; being between 3 and 6 kilograms. Lewis thought that not all characteristics are on a par. Some mark metaphysically deep similarities between things. Others less so. Lewis posited a kind of metaphysical hierarchy – a “naturalness” ranking – running from similarities that are maximally deep metaphysically speaking, to ones that are shallow to varying degrees. He then put this naturalness ranking to work in explicating a variety of metaphysically important notions, such as laws of nature, intrinsicality, physicalism, and so on, and also in providing an account of the nature of thought itself.
A number of Lewis’s important papers were devoted to reconciling some phenomenon or other with his metaphysical system, though those papers were invariably written so as to be of interest to those philosophers who were not so interested in the larger system. One notable example is a topic of perennial philosophical interest: the relationship of matter and mind. Mental phenomena provide a pair of prima facie challenges to the physicalist in the form of consciousness (the existence of experience and feeling), and of intentionality (the fact of minds, and derivatively words, having content, being about something). Committed as he was to a metaphysical scheme according to which mind is nothing over and above the world described by physics, Lewis wrote a number of important papers that develop a physicalist conception of mentality. “An Argument for the Identity Theory,” “Mad Pain and Martian Pain,” (in Philosophical Papers Volume 1, 1983), “Psychophysical and Theoretical Identifications,” “Reduction of Mind,” and “What Experience Teaches” (in Papers in Metaphysics and Epistemology, 1999) together comprise one of the most developed physicalist treatments of consciousness and intentionality.
Another notable example is possible worlds, which as Lewis tells us, are concrete physical universes. A possible history is, in reality, another concrete universe. But is it not intuitively clear that I have many possible histories? How could this be if possible worlds are concrete universes, since I surely inhabit only one concrete universe – the actual one? The problem arises because on the one hand, Lewis says that individuals are “world-bound” – they inhabit only one possible world, and yet on the other, we think that there are many possible lives that an individual could have undergone. Once we suppose a possible life is another concrete history, it seems that no other possible life is a possible life of ours. Lewis was sensitive to this problem, and answered it with “counterpart theory”: an individual might have had a certain feature is made true, if it is true, by the existence of a distinct possible individual who is relevantly similar to the former individual. We do not literally inhabit other possible worlds. But the existence of other individuals in other universes that are relevantly similar to ourselves suffices to make it true that we might have done this or that.
Laws of nature would seem to provide a counter-example to localism. Lewis himself, as we have seen, insists that the laws of nature are contingent. But could not the laws of nature vary while the local matters of fact stay the same? (In the limiting case, we might imagine two empty voids with different laws of nature governing them). Sensitive to such worries, Lewis provided an account of laws of nature that rendered them fully compatible with localism. His picture – much in the spirit of David Hume – was that laws of nature are merely those true generalizations that “achieve an unexcelled combination of simplicity and strength” (Introduction to Philosophical Papers Volume II, p. xi). If the local facts remain constant across possibility, so too must the laws, since facts about which generalizations excel (in Lewis’s sense) cannot vary unless the local facts do too.
Lewis wrote influential articles on many more topics, including formal semantics, conditional probability in relation to conditionals, the meanings of theoretical terms, objective chance and subjective credence, and the foundations of set theory (the latter figuring as the topic of his last monograph, Parts of Classes, 1991). Only a select few topics can be mentioned here.
Beginning with “Survival and Identity” (in Philosophical Papers Volume 1), Lewis was one of the most important proponents of the view that, just as objects have spatial parts, they also have temporal parts. To take a simple example, Lewis’s theory maintained that there existed an instantaneous temporal part of Richard Nixon on 29 July, 2pm Greenwich Mean Time, 1954, that took up just as much space as Nixon, but which lasted only an instant. There is thus a deep metaphysical analogy between the way we take up space and the way we take up time. Just as we are spatially large things composed of vanishingly small things, so, according to Lewis’s favored view, a persisting concrete object is composed of temporally fleeting things (indeed, instantaneous things).
Is free will compatible with determinism? The incompatibilist answers no, arguing that since we are not free to change the past and not free to break the laws of nature, then if determinism is true, we are not free at all. Lewis held that free will and determinism are compatible, offering his own novel diagnosis of where the incompatibilist goes wrong. In his paper “Are We Free to Break the Laws,” Lewis argues that the incompatibilist illicitly trades on an ambiguity in the claim “We are not free to break the laws.”
Is time travel possible? Does not admitting such a possibility lead to insoluble paradoxes? For example, if I could go back in time, then it seems I should be able to go back in time and kill my own grandparents. But then how would I have come to exist in the first place? Lewis’s “The Paradoxes of Time Travel” is a sustained attempt to dispel illusions of paradox, arguing that there are no deep philosophical puzzles raised by the possibility of time travel. Do, then, the laws of nature actually permit it? Lewis says that this is not something for the philosopher to decide.
Common sense tells us that we know quite a lot about the world. The skeptic tells us that we do not know much. Lewis defends common sense without accusing the skeptic of making a mistake. In his “Elusive Knowledge” (in Papers in Metaphysics and Epistemology) he popularized and systematized a so-called contextualist approach to knowledge according to which the meaning of the verb “to know” does not stay fixed across conversational contexts. The upshot of his account is while ordinary people often speak the truth when they say “I know such and such,” it is also the case that when one engages in argumentation with the skeptic, the verb “know” acquires a meaning such that one should concede that the skeptic is right when she says “We know hardly anything at all.” Lewis’s work on knowledge applies some of the lessons provided by his more general picture of how meaning is influenced by conversational dynamics, a picture outlined in “Scorekeeping in a Language Game” (in Philosophical Papers Volume 1).