A new series from the TLS, appraising the works and legacies of the great thinkers and philosophers

Willard Van Orman Quine is well known to the average analytic philosopher. True, he disdained much of the Pandora’s box of conceptual playthings often thought necessary for serious work in analytic philosophy nowadays, and his doctrines are thought to have been largely superseded. But he is treated by many – along with the earlier thinkers Gottlob Frege and Bertrand Russell – as an important intellectual ancestor, who developed many of the puzzles, paradoxes and paradigms that have dominated hardcore analytic philosophy for decades from the 1950s and 1960s.

Certain Quinean doctrines do continue to come up frequently:  the untenability of the traditional “Analytic–Synthetic” distinction (a distinction between statements the truth or falsity of which depends only on its words, as in “No bachelor is married”, and those that depend also on particular facts, as in “the cat is hungry” – a distinction that was thought by many to be the key to various problems in the theory of knowledge); the “web of belief” metaphor (non-metaphorically: that no statement, not even apparent certainties such as 2+2=4, is in principle immune to revision); the unavoidability of the assumption that in addition to tables and trees there really are abstract objects such as numbers (on pain of hamstringing science); the Indeterminacy of Translation (that in principle one correct translation of a sentence – any sentence, say one of Czech – might conflict with another correct translation of that same sentence, and they might even say the opposite of the other); that there is “no entity without identity” (that a scientific use of a term for a kind of thing requires a means of settling whether one of them is the same as another); that science has no serious need of modal logic (the logic of necessity and possibility); that “to be is to be the value of a bound variable” (which is rather too complicated to go into here).

It is somewhat less well-appreciated (though still appreciated) that Quine thought of these ideas as flowing from what he termed his naturalism. And that, many analytic philosophers think, is very much a mainstream position. Indeed all sensible philosophers are supposed to be naturalists. For surely the primary job of philosophy is to fit the findings of natural science together with those of philosophy, in a way that does due homage to all legitimate theoretical concerns and does not denigrate science from some fanciful metaphysical point of view in the manner of F. H. Bradley or Arthur Schopenhauer. The naturalist can even borrow scientific findings in the construction of philosophical theories.

But in fact this understanding of “naturalism” falls far short of capturing Quine’s thought.  According to a number of philosophers, it is the failure to appreciate this that is responsible for Quine’s having fundamentally been misunderstood, more so than any other twentieth-century philosopher (except perhaps Ludwig Wittgenstein). The key point, as Quine’s philosophical confidante Burton Dreben as well as Quine himself stressed, is to take seriously the idea that philosophy must begin in “the middle of things”. There is no viable “first-philosophy” whether of Cartesian, Kantian, or other stripe – no purely philosophical truths, no arguments of the philosopher that stand alone. We have to accept the apparent circular reasoning in the position. We are like sailors navigating a ship of knowledge on the open sea, as the Austrian Otto Neurath once trenchantly asserted, hoping to improve it without the luxury of docking at port.

Why? In the 1920s and 30s, the “Vienna Circle” of Logical Positivists aimed to modernize the discipline, in order, among other things, to lay to rest the old conceit that fundamental to knowledge are certain a priori philosophical principles that go beyond mere logic (and beyond tautological statements, such as “If inflation is increasing then inflation is increasing”). The old Kantian idea was that  “synthetic a priori” propositions provide a means of describing general features of the world for which there is no empirical evidence, and which therefore require a special faculty of a priori intuition. The Vienna Circle disagreed. All knowledge, according to the Circle is either empirical or analytic: the senses plus logic is sufficient.

Such a position went back to the work of David Hume, but with a crucial modern twist. Kant famously argued in the mid-1700s that Hume could not explain how we can have knowledge of certain truths that cannot be established by the senses or by logic. Not only the truths of arithmetic like 7+5=12, but the axioms of geometry, the fact that we live in a world of spatio-temporal mind-independent objects in mutual causal dependence, and other things, go far beyond the immediate deliveries of the senses, and equally beyond what the logician is professionally qualified to pronounce on. Unfortunately for Kant, however, science has a way of encroaching on the seemingly a priori. Geometry, it turns out, thanks to Gauss, Bolyai and Lobachevsky in the 1800s and then ultimately to Einstein, is not identical with Euclid’s geometry; it cannot be understood as a true description of actual space and knowable a priori. It must instead be understood as answerable to physics, which was by Einstein’s time an unimpeachably empirical science. And 7+5=12 and the truths of mathematics generally, thanks to the labours of Frege and then Russell, looked like logic after all. The idea that logic plus the senses is sufficient for human knowledge seemed viable again.

Enter Rudolf Carnap (1891–1970), the most influential of the Logical Positivists, or the Logical Empiricists, as he preferred.  To connect statements of the senses narrowly conceived – those that make no claim beyond what is happening with one’s senses – rigorously with ordinary statements such as “The cat is bristling”, which do claim that something is the case outside one’s senses, perhaps seems straightforward, but philosophers from René Descartes to Hume to George Berkeley to Russell struggled with it. Carnap was the first to sketch in exact detail how it might go.  And central to his scheme was the use of analytic statements such as “No bachelor is married”.  They functioned as a kind of mortar, connecting the one type of statement with the other. For example – put simply – “If a visual presentation has such-and-such a shape, coheres with other presentations, perseveres for such-and-such a time, then the cat is bristling”, might convey a sense of the idea. An enormous tangle of these analytic statements was supposed to connect statements pertaining purely to the sensory given not only with the common-sense world but ultimately with the higher flights of theoretical physics. Logical Empiricism shows the way, and philosophy is thus vindicated as having a distinctive place in what Carnap called the “Logic of Science”.  Philosophy is continuous with science in this view – there is no synthetic a priori as with Kant – but it has a unique role in the logical analysis of science.

Enter Quine. What exactly are these analytic statements, he wondered? In the 1930s and 40s, Quine, along with the logician Alfred Tarski, puzzled over this and pestered Carnap for a satisfactory clarification. The snag was that although Carnap could easily provide an “immanent” sorting of the statements of a particular language into the camps, which he labelled “analytic” and “non-analytic”, he couldn’t provide a general definition, a “transcendental” definition informatively telling us what it is for a statement of an arbitrary language to be analytic. In 1951 Quine went public with his criticisms in the landmark paper “Two Dogmas of Empiricism”, in which he argued that the statements of science are knitted together such that there is no way to distinguish sharply the analytic statements as special; in fact we cling tenaciously to some (special, putatively analytic) statements simply because the disruption of withdrawing them would be more comprehensive and Earth-shaking than finding some other statement (a narrower, more local statement) to revise, in the event of some conflict of “our theory” as a whole with observation. In effect, when we are confronted with a set of inconsistent propositions, it is easier to revise an apparently synthetic one, than an apparently analytic one.

This was one plank in the demotion of philosophy from the still-special role that Carnap envisaged.  Another, present in Quine’s work of the 1950s but made explicit in Word and Object (1960) and the paper “Epistemology Naturalized” (1968), was that the momentary entities held by the Logical Empiricists to be foundational for knowledge – “sense-data” – if they exist at  all, are too fragile, fleeting and thin for the job. The truth is that ordinary statements of perception take much for granted – one assumes that the world is generally thus and so, that one does not misclassify events, that what one perceives is a cat, complete with all its history of purring and yowling, not a momentary cat-shaped visual appearance, and so on. The whole project – buoyed by the idea of “first philosophy” – of building up the public world from a world of subjective appearances or sense-data is misconceived, Quine came to think. One has to begin in the middle of things.

Yet there is a viable project that looks much like the old project, but which does begin in the middle of things.  One can tell, from a developed third-person point of view with all the findings of science at one’s disposal, a story of how in fact a person’s evidence – a welter of sensory events as physiologically described – is related to the rest of a person’s beliefs or “theory”. Thus Quine described his notion of “observation sentences” – those sentences that are more or less well conditioned to events in the environment, such as “Kitty is purring!” – outlined how such sentences might be learned,  and defined the empirical content of one’s overall “theory” as the sum total of observation sentences that one accepts or has accepted.  Non-observation sentences, “theoretical sentences”, are related to observation sentences by logical inference (which involves the recurrence of theoretical terms in sentences generally). This project is conceived in terms of linguistic behaviour (potential assent to queried sentences), and more generally the project conceives the knowing subject and the subject’s statements in terms that are straightforwardly scientific, that is, as physical objects or events.

Through being in the middle of things, the Quinean accepts science as we have it, warts and all. There is no choice. Metaphysically, there is no higher or deeper standard of truth than the most well-confirmed science. Epistemologically, there is no firmer standard of justification than the one that emerges from centuries of hard struggle. Yet many metaphysicans – some of whom consider themselves to be naturalists – insist on finding answers to questions that are not raised by the sciences themselves.  Many epistemologists – some of whom consider themselves to be naturalists – insist on finding answers to the old epistemological questions, the ones that the Quinean will say fall away once the old first-philosophical dreams are given up.  Quine has fundamental criticisms of both types of philosophical work.  Not surprisingly this has led to his reputation as a critic and nay-sayer, even though such criticisms are motivated by this over-arching vision of science.

Philosophers are not thereby relegated to second-fiddle. There is a distinctive and some would say invaluable role to play in the extreme abstraction of their subject matter, in the “breadth of categories” they deal with, in striving to command the most general and exacting view of science.  Yet there is no telling where this ship of science and knowledge might sail.  To take a telling example, some Quinean disciples may look excitedly on the recent development of the many-worlds interpretation of quantum theory. For complicated reasons, Quine famously dismissed the logic of possibility and necessity, and did not accept David Lewis’s formulation of substantially the same ideas in terms of generalizations about other worlds taken as real but spatio-temporally discontinuous from our own. But a not inconsiderable and growing minority of physicists – Max Tegmark is the most visible – think that if the many worlds interpretation of quantum mechanics is true (Schrodinger’s cat is dead and its doppelgänger is alive, for example) then a surprising number of otherwise puzzling cosmic phenomena become explicable, notably the “Horizon Problem” (the inexplicable temperature-uniformity of the universe) and the “Flatness Problem” (the mathematically miraculous stability of the universe, that for example its density is just right for galaxies to form as they have).  If Tegmark’s side wins the argument, then these disciples of Quine will gladly revise their views of a certain dearly beloved subject branch of metaphysics called “ontology”, adding the other worlds to its domain.  Such is life, and such is philosophy.

Gary Kemp is a senior lecturer in Philosophy at the University of Glasgow