[In the following article, White recounts Dewey's response to White's book Social Thought in America, and White's answer to Dewey's charges.]

Shortly after his ninetieth birthday, John Dewey¹ acknowledged receiving from me two publications in which I had criticized some of his views in ethics: my Social Thought in America, and my "Value and Obligation in Dewey and Lewis," both published in 1949.² Since I never heard anything more from Dewey about them, I surmised that he had probably not read them or that, if he had, he did not think it worth bothering to discuss my criticisms. I was therefore very surprised when I read in the final volume of his Collected Works that he had paid attention to them in a piece entitled "Comment on Recent Criticisms of Some Points in Moral and Logical Theory,"³ published in 1990, along with other posthumous writings of his which testify to his remarkable intellectual vigor in old age. His editors indicate that Dewey's reply was probably written as late as 1950, and this is confirmed by his reference in it to another criticism of my views by Sidney Hook in his article, "The Desirable and Emotive in Dewey's Ethics,"⁴ published in that year.

Despite the fact that all of this discussion took place more than forty years ago, I think that an examination of Dewey's reply is warranted today, partly because Dewey said there that the subject of the discussion occupied "a central position in the theory of method underlying [his] views on all philosophical topics" (op. cit., p. 480); and partly because Alan Ryan, in his recent widely praised study, John Dewey and the High Tide of Liberalism,⁵ has referred to my differences with Dewey on the relation between the desirable and the desired in a way that needs correction. By publishing a rejoinder to Dewey at this late date, I may make my exchange with him one of the longest-running ones in the history of philosophy as well as one in which he unfortunately cannot have the last word; but I think the current revival of interest in his thought fully justifies the attention I give here to ideas that Dewey regarded as fundamental in his philosophy. I am bound to say, however, that his reply does not convince me that my criticism of his view was unjustified.

1. TWO DIFFERENT CRITICISMS OF DEWEY'S ANALYSIS OF DESIRABILITY

At one point, Dewey writes: "The specific criticisms made by Dr. White already have been so amply and so adequately dealt with by Dr. Sidney Hook on the base [sic] of an extensive and critically accurate knowledge of my writings that I have only as concerns the particular views criticized to refer interested persons to Dr. Hook's discussion and express my deep and grateful appreciation" (op. cit., pp. 480-81). Dewey then goes on to comment on what he, following Hook, calls "one variant" of my criticism of his views on the relation between 'X is desired' and 'X is desirable'. The variants of which Hook and Dewey speak are two different criticisms by me of what I regarded as two different versions of Dewey's view of the relationship between what is desired and what is desirable, when the latter is understood as what we ought to desire or have a moral duty to desire. I dealt only with the first of these versions in the article mentioned earlier, but in the book, which appeared a little later in 1949, I dealt with it along with the second version.

It seemed to me that, in the first version, Dewey regarded moral statements of the form 'X ought to be desired' as synonymous with 'If X is considered by a normal person under normal conditions, X will be desired by that person', following the pattern used by some philosophers—notably C. S. Peirce—when analyzing the nonmoral statement This is objectively red' as synonymous with 'If this is looked at by a normal person under normal conditions, it will appear red to that person'. In so analyzing 'desirable', I said Dewey had tried unsuccessfully to generate a moral de jure proposition by performing a logical operation on two merely de facto propositions. I said that Dewey had linked a de facto proposition such as 'Adam considers this apple under normal conditions' and the de facto proposition 'Adam desires this apple' as antecedent and consequent, respectively, of the conditional proposition 'If Adam considers this apple under normal conditions, then Adam desires this apple', and then asserted that this conditional proposition was synonymous with the moral proposition 'This apple ought to be desired by Adam'. I thought that joining these two de facto propositions in a conditional proposition did not yield the equivalent of a moral proposition any more than joining them in a conjunction would, and said that, if it did, it would also generate a related embarrassment for Dewey's view. I pointed out that, if the moral proposition 'This apple ought to be desired by Adam' were synonymous with the conditional 'If Adam considers this apple under normal conditions, Adam desires this apple', then by parity of reasoning the conditional 'If Adam looks at this apple under normal conditions, then this apple appears red to Adam' should be synonymous with 'This apple ought to appear red to Adam'. This, I said, would make 'This apple is objectively red' synonymous with a statement that is false or nonsensical, since 'ought to appear red' contains the moral 'ought' of 'ought to be desired' and not the predictive 'ought' of a sentence such as 'The sun ought to set in a minute'.⁶

I added in Social Thought in America that Dewey seemed to defend a second version of the relationship between 'desirable' and 'desired' in which a thing is said to be desirable just in case we know the causal antecedents and consequences of desiring it (op. cit., pp. 216-17). In commenting on this reference to causal antecedents, I repeated what I had said about the first version, that knowing the causal antecedent of a desire not to smoke opium would not establish the moral desirability of not smoking it, for knowing that my desire not to smoke opium came about after I had considered not smoking it while in a normal state and under normal conditions would not make that desire morally obligatory. And merely knowing in addition the causal consequences of my desire if acted upon, such as preservation of my health, would not contribute to showing that my desire was obligatory. Therefore, I wondered whether Dewey might say in reply that I ought to desire not to smoke the opium just in case I knew the causal antecedents of my desire not to smoke it and also knew that the consequences of this desire would be .. . , where the dots were to be filled in by Dewey himself. First I asked whether Dewey could complete the task by putting the word 'desirable' in place of the dots. No, I said, because that would produce a circular definition or analysis of 'desirable'. And when I asked whether Dewey should put 'desired' in place of the dots, I made some observations that I now expand in order to facilitate later discussion of my exchange with Dewey and Ryan's account of it.

If Dewey were to put 'desired' in place of the dots, then saying 'Not smoking opium is desirable' would be false if one did not desire the consequences of desiring not to smoke. If I said that not smoking opium was desirable and learned that my desire not to smoke it had consequences that I did not desire, I would have to recant my statement. To my mind, this raised a question about putting 'desired' in place of the dots in the definiens of 'desirable', because doing so required that the consequences of a desire for X be desired and not that they be desired by a normal person under normal conditions or be desirable. I wondered, therefore, whether Dewey would be willing to say that knowledge of this raw desire, so to speak, for the consequences of the original desire could play this role in his analysis of the desirable. On the other hand, as we have seen, if Dewey were to say that the consequences had to be desirable, he would be involved in the circle I mentioned earlier.

II. DEWEY'S VIEW AND MILL'S COMPARED AND CONTRASTED

In order to discuss Ryan's views about my differences with Dewey in an intelligible way, I must also repeat what I said in my paper⁷ concerning the relations between J. S. Mill's views and those of Dewey. I said that although these views were more alike than is usually supposed, Mill held that we can prove that something is desirable by showing it is desired, whereas Dewey avoided this view in the first version of his definition of the desirable since it is obvious that a thing may be desired without being desirable, according to that version of his definition, if the desire were not that of a normal person under normal conditions. After pointing this out, I made some remarks on a difference between the so-called dispositional terms 'soluble' and 'objectively red' which I thought would illuminate the relationship between Dewey's first version of desirability and Mill's view. I said that we can infer 'a is soluble' from 'a is dissolving now' whereas we cannot infer a is objectively red' from 'a appears red now', because 'a is soluble' means the same as the statement that there are conditions under which a dissolves whereas 'a is objectively red' does not mean the same as the statement that there are conditions under which a appears red, but rather is synonymous with the statement that a appears red under specific conditions, namely, to a normal observer who looks at a in white light. That is why Dewey's view that the analysis of 'desirable' resembles that of 'objectively red' does not license the inference of 'a is desirable' from 'a is desired now'. Mill, I held, thought that 'desirable' resembled 'soluble', whereas Dewey avoided Mill's error by saying that 'desirable' resembled 'objectively red'. Nevertheless, I said, Dewey's first version was defective for the reasons I have given earlier, even though Dewey had not been misled by the 'ble' at the end of 'desirable' into thinking that it was a dispositional predicate like 'soluble', but instead had made the more subtle error of thinking that 'desirable' was like 'objectively red'.

All of this makes it comparatively easy to see a misunderstanding in what Ryan says about my exchange with Dewey. Ryan writes: "Dewey, said White, had committed the same mistake as Mill and had confused the desired with the desirable. Dewey thought the 'data' of ethics were our likings and wishings and longings and seemed to be reducing ethical argument to the discussion of what we desired rather than what we ought to desire" (op. cit., p. 337). This, I submit, is a very inadequate summary of my view. In my discussion of the first version of Dewey's view, I objected to Dewey's effort to assimilate 'desirable' to 'objectively red', while stating explicitly that Dewey did not make Mill's error of inferring 'a is desirable' from 'a is desired' or that of confusing them. Nor in my examination of the second version did I accuse Dewey of doing this, for there I merely asked whether Dewey wished his definition or analysis of desirable to require that the consequences of desiring a thing be (1) desirable or (2) desired, and I pointed to difficulties in both answers. I must therefore protest against Ryan's defense of Dewey's view as "infinitely far from collapsing the desirable into the desired" (op. cit., p. 337), as if to imply that I had accused Dewey of confusing the desirable with the desired or of collapsing one into the other.

III. TWO INEFFECTUAL CRITICISMS OF MY VIEW BY DEWEY

Now I turn to Dewey's own effort to answer my criticisms. He begins by referring to what he calls a fundamental methodological difference between my view and his, but in stating that difference he unfortunately relies too much on what Hook claims I say. This leads Dewey to misunderstand my statement that the second version of his definition of 'desirable' might be circular. Because he accepted Hook's misleading summary of my view, Dewey wrongly took me to hold that "knowledge of the causes and consequences of our desire and of what is desired does not make the desired desirable unless .. . we can get back to some rock-bottom desirable in itself (op. cit., p. 481). In other words, Dewey mistakenly thought I held that the expression 'desirable in itself must appear in the definiens of 'desirable'. Having got this wrong impression, Dewey then associated me with what he probably regarded as two of the most objectionable ideas in the history of philosophy, saying that my alleged "dependence upon that which is 'desirable in itself, that is, in complete independence of and isolation from investigation of the existential context of 'conditions and consequences', involves the assumption of what has been known in ethics as the method of Intuition and in epistemology as the necessity of the a priori to warrant the validity of statements made on empirical grounds" (op. cit., p. 481). In my book and article, however, I never appealed to the notion of desirable in itself. Indeed, if the phrase means something different from 'desirable'—as it seems to for Dewey—I would not have complained that its introduction into the definiens of 'desirable' created a circle. Furthermore, I did not defend ethical intuitionism and was probably less of an advocate of the necessity of the a priori than Dewey himself was.⁸ So much, then, for one of Dewey's ineffectual criticisms of my views.

I turn now to another such criticism that emerges in Dewey's response to my claim that he had generated a normative or de jure proposition by performing a suitable operation on merely de facto propositions. He admits that such generation would constitute "a variety of intellectual magic" (op. cit., p. 482) but does not admit that his view is open to this criticism. Instead, he uses my remark about performing an operation on propositions as an occasion for complaining that in using the word 'operation' in this way I completely neglect a view of his according to which an operation is performed not on propositions but with them (op. cit., p. 482). I knew, of course, that Dewey had used the word 'operation' differently from the way in which I used it, but this hardly affected the point I made while using 'operation' as I used it. Therefore, Dewey's remark about his operating with propositions rather than on them fails to rebut anything I said in criticism of his views.

IV. DEWEY'S MORE SERIOUS ARGUMENTS

After Dewey ironically says that I have the "personal right to take" a position which is a "heritage from times when scientific method as now practiced did not exist and when the rational as distinct from the observational had to be invoked to guarantee the validity of beliefs and statements," he adds that the method he "employed in making the distinction between the de facto desired and the de jure desirable is simply the method pursued in all sciences which conduct inquiries intended to find out what is fact—'objective' fact to indulge in a pleonasm—and in distinction from what is taken to be fact apart from inquiry into 'conditions and consequences'" (op. cit., p. 482).

Dewey also says: "Dr. White's reduction of the desirable to that which is desired 'under normal conditions' is wholly satisfactory to me provided the literally terrible ambivalence in normal is cleared away—which I do not find he even tries to do" (op. cit., p. 483). I shall postpone dealing with Dewey's view of the word 'normal', but I wish to emphasize here that he accepts the wording of the reduction or analysis of 'desirable' which I attribute to him as well as the similarly worded analysis of 'objectively red' or 'really red' which I attribute to him. But since I maintained that his analysis of 'This apple is desirable'—where 'desirable' is synonymous with 'ought to be desired'—is not acceptable whereas the parallel analysis of 'This apple is really red' is acceptable, Dewey tries to show that they are both acceptable because both reflect the method used in science "to find out what is fact... in distinction from what is taken to be fact apart from inquiry into 'conditions and consequences'" (op. cit., p. 482). Unlike Dewey, however, I think that the distinction between what is taken to be fact and what is fact is involved when we distinguish between the statements This appears to be red' and 'This is really red*, but that it is not involved when we distinguish between the statements This is desired' and This is desirable, that is, ought to be desired'. Indeed, I think Dewey's main error is to identify the distinction between the de facto This is desired' and the de jure This is desirable, that is, ought to be desired' with the distinction between what is taken to be fact and what is fact. Therefore, one way of stating my criticism of Dewey in traditional philosophical language might be to say that the distinction between the de facto and the de jure is not the same as that between the apparent and the real. I elected to say in less traditional philosophical language, however, that though the statement This apple is really red' may be analyzed (following Peirce) as synonymous with the statement 'If this apple is put before a normal eye in daylight (normal light), it looks red', the statement This apple is desirable, that is, ought to be desired' is not synonymous with the statement 'If this apple is considered by a normal person under normal circumstances, it is desired'.

In arguing that the de facto/de jure distinction is not the same as the apparent/real distinction, I might have said that the expression 'de facto' does not mean the same as 'apparent' and that the expression 'de jure' does not mean the same as 'real'. But I chose another way of arguing the point because I did not wish to use terms that some readers might find very obscure, and so I argued instead that, if the correct analysis of the moral statement

(a) This apple ought to be desired (is desirable).
is given by
(b) If this apple is considered by a normal person under normal circumstances, it is desired.
then the statement
(c) If this apple is put before a normal eye in normal light, it looks red.
should be the analysis of
(d) This apple ought to be seen as red (ought to look red).
But since (c) by hypothesis gives the analysis of
(e) This apple is really (objectively) red.

then (e) should also be synonymous with (d). In that case, a statement of the real color of an object would become synonymous with an ethical statement, which (d) is because its 'ought' is, as I have said, the 'ought' of (a). This would be absurd, however, and would lead us to say that all attributions to objects of their real or objective colors are made in statements synonymous with ethical statements and therefore to the unacceptable conclusion that many obviously nonethical descriptive scientific truths are ethical. Here, I was not arguing against ethical naturalism construed as asserting that all ethical statements are synonymous with scientific statements of behavioral science; I was protesting against a view that had as a logical consequence the view that This apple is really red' is an ethical statement.

Parenthetically, I suggest that the statement 'If this apple is considered by a normal person under normal circumstances, it is desired' might be regarded as the analysis of the statement This apple is really desired' by those who seek an analogue involving desire for Peirce's analysis of This apple is really red'. One of these analogous analyses would say that the real color of an apple is the color it appears to have in the eye, so to speak, of a normal beholder under normal circumstances; the other would say that the real attitude toward the apple is the attitude the normal reactor to it has under normal circumstances. I merely suggest this, as I say; I do not advocate it. But I do continue to maintain that although the Peirceian approach connects the real color of an object with the way in which the object looks or appears, it does not connect the attitude we ought to have toward an object with the attitude we do have toward it.

I turn now to the difficult and crucial concept of normativeness that figures in Dewey's reduction of the desirable to what is desired under normal conditions, for his reflections on it contain the heart of his argument against my view. In my opinion, only the outlines of that argument may be made out with confidence. Dewey thinks that the word 'normal is used normatively when it appears in the conditional statements (b) and (c) above, and from this he concludes that (a) and (e) are also normative in a sense that allegedly undermines my contention that (b) does not give the correct analysis of moral statement (a). Dewey never answers my complaint that his view leads to the absurd conclusion that statements about real colors are ethical, but since he insists that (a), (b), (c), and (e) are all scientific and therefore all normative, I shall try now to penetrate the outlines of his view in order to look more closely at his idea of the normal as normative and to see whether it can sustain the great philosophical load he puts on it.

Dewey tells us that 'normal' in the sense of what happens usually or on the average is "certainly de facto" and therefore not normative. But then, as I have already pointed out, he says that a statement like (e) of the form 'X is objectively (really) red' is normative and that the expression 'normal' in (c) has what he calls normative force. He also says in a crucial passage that the normal conditions referred to in statements like (c) are "not those of a majority or even the total number of cases in which X appears red. They are conditions instituted by continued experimental inquiries conducted for a definite endin-view" (op. cit., p. 483). Dewey seems to imply here that Peirce's condition of being before a normal eye in normal light applies neither to the majority of cases in which X appears red nor to the total number of such cases. I doubt whether he is correct in saying that the number of cases in which X is before a normal eye in daylight does not constitute the majority of cases in which X appears red, but I shall not quarrel about that since I do not know how to settle the issue. Dewey needs to say, however, that some cases in which X looks red are cases in which X is not before a normal eye in daylight, since that will allow him to say that an experimenter must put X before a normal eye in normal light in order to discover whether a statement such as (c) is warranted. Dewey seems to think that the condition 'is before a normal eye in normal light' has normative force because it expresses a norm or standard to be met which is not always met, and therefore the scientist must institute such a condition in order to test (c) and its synonym (e). Dewey's point seems to be that because objects that appear red are not always before a normal eye in normal light, the scientist puts them there to see whether (c) and therefore (e) are true.

Once it is recognized that normal conditions are normative in his sense, Dewey says he welcomes "the formal or methodological identification of the statement 'X is desirable' with 'X is objectively red', for if 'objective' has any distinctive relevant sense in the latter proposition, that sense, like that of 'normal' as having any relevance to the point at issue in the phrase 'normal conditions', is itself intrinsically normative or de jure" (op. cit., p. 483). Furthermore, Dewey says he welcomes the "identification when its direction is completely reversed," which means, I think, that Dewey assimilates the analysis of 'X is objectively red' to that of 'X is desirable' just as he does in the reverse direction. In short, he seems to hold that (a) and (e) are both normative because they are respectively synonymous with (b) and (c). These two singular statements seem to be normative for him because the generalizations that support them, respectively, are the following allegedly normative scientific statements:

(b') Whenever this apple is considered by a normal person under normal circumstances, it is desired.

and

(c) Whenever this apple is put before a normal eye in normal light, it looks red.

And these generalizations seem to be normative because their antecedent statement forms are said to express conditions that require institution by an experimenter, for a reason that Dewey conveys at the end of his "Comment": "Unfortunately, or perhaps fortunately, 'normal' conditions for and of an experiment that yields a warranted conclusion do not lie around nor force themselves upon us. They are obtained by undertaking the kind of activities which the best available knowledge at the time informs us should be tried in order to find out their specific consequences in and for further knowing" (op. cit., p. 484). Dewey's use of the phrase 'should be tried' conveys his idea that the activity of putting an apple before a normal eye in daylight is dictated by a normative judgment. But surely it is not a moral judgment, one is tempted to add.

Having said as much as I can in elucidation of Dewey's view as to why (a), (b), (c), and (e) are normative, I now ask: Does all of this show what he thinks it shows? In other words, does he succeed in showing that all scientific propositions are normative in a way that undermines my claim that his analysis or reduction of 'ought to be desired' cannot be modeled on the Peirceian analysis of 'really red' or 'objectively red'? I am afraid not. Even if one accepts what Dewey says, it is hard to see how it makes statement (a) a statement of moral duty. Moreover, according to a strict application of Dewey's own criterion for being normative, statement forms of the antecedents in statements such as (b') and (c') are normative in his terminology, since his word 'normative' strictly applies only to antecedent conditions or statement forms like 'X is considered by a normal person under normal circumstances' and 'X is put before a normal eye in normal light'. But how does it follow from the fact that these antecedent conditions or statement forms are normative in Dewey's sense and that the tester of (b') and (c') must institute those conditions so that he can perform an experiment to discover the truth of (b') and (c'), that statement (b') expresses a moral duty? And if it does follow, why should it not follow for the very same reason that (c) expresses a moral duty and is synonymous with (e)—a consequence I have said is absurd? So far as I can see, Dewey has no answer to these crucial questions.

Furthermore, I am not convinced that Dewey's characterization of all scientific generalizations is adequate. 1 am not sure that he is right when he says that all statements of the form 'For every X, if X is A, then X is B' are scientific if and only if the experimenter must put things in normative condition A in order to see whether they are also in condition B. 'All bodies attract each other' is a scientific generalization even though there is no verb like 'put' in its antecedent when it is transformed into 'For every X and Y, if X and Y are bodies, X and Y attract each other'. This is a scientific generalization which contains no experimental directions, and I cannot imagine an experimenter saying 'Let us put X and Y into the condition of being bodies to see whether they attract each other'. This is not to say that the law of universal attraction is not tested empirically by the use of scientific method, but it is to say that not every scientific generalization contains antecedents that contain experimental directions like those which Dewey regards as essential components of the antecedents of scientific generalizations. I want to emphasize, however, that, even if he had characterized all scientific generalizations correctly, his view of science as normative would not support his answer to my criticism of his view of moral duty. Even if all scientific generalizations were to contain antecedent conditions that are normative in Dewey's sense, that would not show that they support statements saying that we have a moral duty to desire certain things—that they are desirable. And if they did show that, they would also support statements that say absurdly that we have a moral duty to see certain things as red.

v. ETHICAL NATURALISM: REDUCTIVE AND HOLISTIC

Throughout my discussion of Dewey's ideas in my earlier writings, I assumed that his views on what is objectively red and what ought to be desired were analytic or reductive. I based this on Dewey's⁹ favorable quotation of a passage in which Peirce asks: "of the myriads of the forms into which a proposition may be translated, which is that one which is to be called its very meaning"; to which he replies: "It is, according to the pragmaticist, that form in which the proposition becomes applicable to human conduct" (ibid., p. 303). And I also had in mind passages like the following one in Peirce 10 : "to say that a Jacqueminot rose really is red means, and can mean, nothing but that if such a rose is put before a normal eye, in the daylight, it will look red" (ibid., p. 194). Therefore, I feel vindicated in my interpretation of Dewey when he says in his reply: "Dr. White's reduction of the desirable to that which is desired 'under normal conditions' is wholly satisfactory to me," even though Dewey goes on to say things about 'normal' that I do not fully accept or even understand. But I am sure that readers of my later work might see some irony in my saying this because in What Is and What Ought To Be Done and The Question of Free Will: A Holistic View¹¹—especially in the first—I urge that we abandon the reductive search for the meanings of ethical terms partly because of my increasing doubts about analyticity and analysis as expressed in my Toward Reunion in Philosophy.¹² I therefore came to focus instead on the epistemology rather than the semantics of ethics, which I later approached holistically under the influence of Pierre Duhem, W. V. Quine, and Alfred Tarski.¹³

When I approached ethics in this non-Mooreian way, I came to believe that the purpose of ethical thinking was to organize a heterogeneous flux that contained not only sensory experiences but also moral feelings, and that some heterogeneous systems or bodies of statements that are used to organize such a flux contain not only logical statements and those of descriptive science but also moral statements. Just as Quine's holism and Tarski's are used by them to argue against a sharp epistemological distinction between the method of testing logical statements and that of testing statements of natural science, I use my holism—which I label corporatism merely because I include moral feelings among experiences and moral statements in some of the systems we test—to argue against a sharp epistemological distinction between the methods of testing statements such as This is desirable' and This is objectively red'. Descriptive and ethical statements, I argue in my later work, appear in conjunctions of statements that organize heterogeneous experiences in a way that permits no sharp distinction between the methods used in testing such statements. By regarding ethical argument in this way, I have come to a view like that of Dewey insofar as the aim of his analysis of This is desirable' was to erase a sharp epistemological distinction between the method of testing it and the method of testing This is objectively red'. Because my corporatism requires Duhemian testing of conjunctions of statements by experience, I call myself an ethical naturalist but not a naturalistic reductionist like Dewey, who seems to seek synonyms for statements asserting moral duties and for statements attributing objective colors while engaging in what he calls reduction in his reply to me. And because I think scientists test extended hypothetico-deductive systems rather than isolated statements, Dewey's view that all "scientifically grounded propositions" contain or need experimental directions for testing them individually strikes me as excessively isolationist, so to speak. I say this because he regards each scientific generalization or law as experimental in a way that I criticized earlier, and because I do not think that such a criticism may be justly made when a scientific theory is viewed as a conjunction of statements that is confirmed as a whole by experience and not by separate experimental testings of each generalization contained in that whole. I believe that there is an important affinity between my present rejection of a sharp distinction between testing moral statements and testing descriptive statements, and what I took to be the antidualistic, naturalistic motive behind Dewey's effort in The Quest for Certainty¹⁴ and elsewhere. Like him, I do not think that the moral judge has any data except sensory and emotional experiences against which to test his theories, but I do think that Dewey's ethical naturalism was reductive and therefore subject to criticism of the kind I have made earlier.

In concluding, I want to say something about Dewey that Bertrand Russell¹⁵ once said so felicitously about William James after criticizing his views. I wish "to express, what in the course of controversial writings does not adequately appear, the profound respect and personal esteem which I felt for him, as did all who knew him. . . . For readers trained in philosophy, no such assurance was required; but for those unaccustomed to the tone of a subject in which agreement is necessarily rarer than esteem, it seemed desirable to record what to others would be a matter of course" (ibid., p. v). I say this about Dewey on the chance that even readers trained in philosophy might need such assurance while reading the present article, and I wish to add that I owe him a great intellectual debt as well. He was a towering figure in philosophy, one from whom I learned an enormous amount even when I disagreed with him.

NOTES

¹ Dewey wrote me on November 6, 1949: "Many thanks for a copy of your Social Thought in America which I've just received and [am] looking forward to reading soon. I want also to thank you for your reprint from the Philosophical Review. I had been waiting till I have had a chance to read it carefully, as I have been swamped and am only now beginning to get my head above water. I haven't even yet been able to study it but shall soon." This was the last letter I had from Dewey, the end of a cordial correspondence that began in 1941, when I was writing The Origin of Dewey's Instrumentalism (New York: Columbia, 1943). It saddens me to think that he might not have written me again because he did not wish to say in a letter what he says in the piece to which I reply here.

²Social Thought in America (New York: Viking, 1949); same pagination in paperback editions: (Boston: Beacon, 1957) and (New York: Oxford, 1976). "Value and Obligation in Dewey and Lewis," The Philosophical Review, LVIII (July 1949): 321-29; this article is reprinted with some alterations in my Pragmatism and the American Mind (New York: Oxford, 1973), pp. 155-67; the altered version also appears in J. E. Tiles, ed., John Dewey: Critical Assessments, Volume III (New York: Routledge, 1992), pp. 28-36.

³ In Jo Ann Boydston, ed., The Later Works: 1925-1953, Volume XVII (Carbondale: Southern Illinois UP, 1990), pp. 480-84.

⁴ In Hook, ed., John Dewey: Philosopher of Science and Freedom (New York: Dial, 1950), pp. 194-216.

⁵ New York: Norton, 1995.

⁶ "Value and Obligation in Dewey and Lewis," p. 326.

⁷ "Value and Obligation in Dewey and Lewis," pp. 324-25. For a more extended discussion of the relation between Dewey's views and Mill's, see the altered version in Pragmatism and the American Mind and the reprint thereof in Tiles.

⁸ This reference to the a priori should be read with the following in mind. I had contributed my essay, "The Analytic and the Synthetic: An Untenable Dualism," to John Dewey: Philosopher of Science and Freedom, the same volume to which Hook had contributed his essay, "The Desirable and Emotive in Dewey's Ethics" (op. cit., pp. 194-216). Had Dewey read mine, he would have known that I was sympathetic to his views on the dualism I was considering. Indeed, Hook wrote me upon receiving my manuscript on November 5, 1949, that it was "squarely in the context of [Dewey's] thinking." When I say that I was probably less of an advocate of the necessity of the a priori than Dewey, I have in mind some views I set forth in "Experiment and Necessity in Dewey's Philosophy," Antioch Review, XIX (Fall 1959): 329-44; reprinted, as "The Analytic and the Synthetic," in my Pragmatism and the American Mind.

⁹ "The Pragmatism of Peirce," in Peirce, Chance, Love, and Logic, M. R. Cohen, ed. (New York: Harcourt, Brace, 1923).

¹⁰Collected Papers, A. W. Burks, ed. (Cambridge: Harvard, 1958).

¹¹ (New York: Oxford, 1981) and (Princeton: University Press, 1993).

¹² Cambridge: Harvard, 1956.

¹³ Duhem, La Théorie physique: son objet, sa structure (Paris: 1914, 2nd ed.), pp. 278-89; Quine, "Two Dogmas of Empiricism," Philosophical Review, LX (1951): 20-43; and my "A Philosophical Letter of Alfred Tarski," this JOURNAL, Lxxxiv, 1 (January 1987): 28-32.

¹⁴ New York: Putnam, 1960.

¹⁵Philosophical Essays (London: Longmans, 1910).