Another backdrop Hookway repeatedly refers to as offering an alternative philosophical path to Quine's is Goodman's «worldmaking» approach. Hookway brings his book to a close (220) by suggesting that Goodman's account is more promising whereas `it is easy to conclude that it is only a physicalist or empiricist prejudice that justifies restricting our conception of the real in the way that Quine desires.' Hookway praises Quine (219) for having `worked through what is involved in a physicalist empiricism more thoroughly and rigorously than any other positivist philosopher,' suggesting that by so doing `he has helped us to see why it is unsatisfactory.'

The book's organization plan is somewhat peculiar, trying to strike a middle course between a diachronical tracking of Quine's philosophical development and a thematic account of what can be termed his system. Thus the first part is claimed to be devoted to Quine's early thought (in From a Logical Point of View), while parts two and three are of a more systematic purpose dealing with Quine's philosophy as put forward in his later works; finally part four is given over to a mild criticism of Quine's conclusions. However, the plan's effective implementation is not that clear or straightforward. As I said above, the leading thread is, from start, the 1975 paper «Five Milestones of Empiricism». Now, construing Quine's early thought through the hermeneutic keys provided by so a late paper is a daring exegetical enterprise which would call for some warranting argument.

The reviewer wonders if such a procedure is what leads to Hookway's conclusion that Quine has somehow changed his attitude concerning pragmatism -- at least in the sense of distancing himself from such a position. As I view it, Quine's thought has remained fairly stable through the years, oscillations and all. In fact the essays collected into Quine's Theories and Things -- one of which is «Five Milestones» -- evince that oscillation, esp. when the 2d. printing of the book is compared with the 1st one, with chapter 2, «Empirical Content», having undergone an important modification. While the former version -- which is the one Hookway quotes on p. 200 -- claimed that two empirically equivalent theories are equally true, or false, the recast version only says that they are equally warranted. However even in the last printings, Theories and Things is full of remarks which clearly tally with the nonrealistic view espoused in the former version. In that connection it is worth mentioning how Quine strove, in «On empirically equivalent systems of the world», 1975, to reconcile his nonrealistic leanings with his thesis of underdetermination of theory by experience. Such hesitations elicited Gibson's critical comments which in turn caused Quine to withdraw some of his nonrealistic claims. But even then Quine's retractation is far from complete or thoroughgoing. So, the oscillation is still with us. Another claim which is in that connection quite telling is that treating simplicity as a presumption of truth is not Quine's line (see «Reply to Gibson», in The Philosophy of W.V. Quine, ed. by L.E. Hahn & P.A. Schilpp, p. 155). If simplicity affords no presumption of truth, nothing does, since all scientific criteria -- as those proposed in The Web of Belief -- can be taken to be as many standards of simplicity in a broad sense.

Hookway is not even very clear on what realism amounts to. The already mentioned claim about how two observationally equivalent theories are equally true, or false, is alleged on p. 210 to be what allows Quine `to avoid being forced to acknowledge anti-realism or relativism explicitly' despite underdetermination of theory choice by evidence. Isn't it the other way round? Underdetermination does not compel us to espouse anti-realism unless we reject the idea of unwarranted truth, or, better, of a theory being truer than another even when it is not more warranted than the latter. Yes, saying that they are equally true enables one to keep short of confessing to anti-realism in the sense of the thesis that a scientific theory needn't be true, but paying the price of a strongly nonrealistic conception of truth, one according to which we couldn't be wrong and yet heed the evidence (a view Thomas Nagel has challenged: he is one of those who are not mentioned in the book).

Not that Hookway is not aware of persisting nonrealistic leanings in Quine's writings (on p. 57 he says that `the «pragmatist» claims from From a Logical Point of View seem not to have been left behind,' and again on p. 200 he rightly says that a number of statements to be found in Quine's later writings `suggest a view closer to anti-realism than to the scientific realism which Quine proclaims'). Yet I suspect that he has not fully realized how deep the tension is in Quine's thought between realism and nonrealism. (His book seems to have been written before the last mentioned collective volume on Quine appeared.)

In fact Hookway seems to be much more concerned over Quine's physicalism than over his empirism. As regards the latter, he claims (7) that `[w]e do not require a precise definition of «empiricism»: it is enough that empiricists take seriously the claims of the sciences to provide our best knowledge of reality, and hold that this knowledge is grounded in sensory experience.' A very vague claim. Hookway later (15) maintains that `[a]n empiricist who thinks that our grasp of our language involves linking it to experience need claim only that we understand an expression when we know how to test the sentences in which it occurs against experience.' Only? What then emerges as empiricism is a strongly verificationist theory of meaning. That trend is doubtless present in Quine's motivations, and one of the essential premises of his most cherished argument for indeterminacy of translation, IT. But as we have seen such a line of thought issues in abandoning or anyway warping the underdetermination thesis and more generally realism. There is an alternative argument for IT, that put forward in «On the Reasons for the Indeterminacy of Translation» (Journal of Philosophy 67 (1970), pp. 178-83), a paper which Hookway does not mention in his discussion of the IT thesis (in ch. 8, on pp. 127ff.), a discussion which nevertheless contains a lot of insightful remarks.

In my opinion the most regrettable limitation of this book -- a limitation which does not cancel its interest as a survey of a wide range of debates around Quine's several contentions -- is that it all but ignores Quine's emphasis on degrees. Once or twice (for instance on p. 42, while discussing moderate holism) the book does quote some of Quine's assertions to that effect, but it goes on taking but little notice of Quine's insistence. Now, if we heed Quine's urgency, we'll realize that many, perhaps most, philosophical questions are matters of degree. The trouble is that Quine deals with them both as matters of degree and yet also as yes/no questions. Is that an inconsequence? Or is it just a sort of way of speaking -- resorting to the yes/no façon de parler as a mere abbreviation? The reviewer finds hard to accept such an exegesis, but also to simply charge Quine with inconsequences on that point. The inconsequence seems to lie at a deeper level, namely that concerning logic. In fact there are two difficulties here. The first is that classical-logic treatments certainly cannot -- except through some sort of devious paraphrase stipulations -- account for a predicate to be used both as one-place, in categorical non-comparative assertions, and as two-place, in comparative sentences. The second difficulty is that, if having a certain property is a matter of degree, to what extent needs a given thing have the property for our statement that it has it to fit the facts, so to speak, i.e. to be truthfully assertable? (And, also, to what extent has the thing to lack the property for us to be allowed truthfully to assert that it fails to have it?) Quine has not been wholly able to solve that two fold riddle (although he has said something closely connected with it in «What Price Bivalence?»).

Now, even though Hookway quotes (203) Quine's assent to Smart's claim `that the physicist's language gives us a truer picture of the world than does the language of common sense,' he has nonetheless apparently remained blind to the very two-pronged difficulty, let alone its many branching intertwinements with problems as diverse as cross-world identity (brought up on pp. 117ff.), the notion of physical object (99ff), IT and facts of the matter (passim), certainty, justification, and so on, and particularly the issue concerning the bounds of [moderate] holism, or to what extent observation sentences are immune from challenge (on p. 42 Hookway's both quotes Quine's statement that the bounds of holism are a matter of degree and misconstrues it by saying that with such a statement `Quine stresses that observation sentences do have a separable empirical meaning,' which is not utterly false but deserves a qualification, observationality coming in degrees as Quine has insisted once and again).

Thus, Hookway's final reductio argument to the effect that Quine's beautiful enterprise proves that empiricist physicalism is wrong -- besides suffering from another weakness arising from lack of a clear definition of `empiricism' -- can be answered by considering such unsatisfactory consequences as apparently follow from Quine's premises either to be in need of a gradualistic construal which would lessen their unsatisfactoriness, or else to show us that some of the premises themselves deserve such a gradualistic treatment in virtue of which the inference would be blocked. If Quine's general gradualistic approach is to be taken seriously, the hypothesis cannot been ruled out that such is the lesson to learn, should it carry us beyond Quine's own use of gradualistic qualifications.

There are also some further points worth remarking. Thus, on pp. 91-3, «formal logic» is discussed in a vein which seems to the reviewer far away from Quine's own philosophy of logic, even though the author quotes a couple of Quinean statements. Anyway Quine's view is definitely not that logic describes the semantics of an artificial language, or that it provides an expression framework for science. (With such a view, logical truth vanishes.) The error bears on many other issues.

Despite such defects the book is a rather good introduction into the study of Quine's philosophy.