Pragmatism and Philosophy of Science: A Critical Survey

Abstract

After delineating the distinguishing features of pragmatism, and noting the resources that pragmatists have available to respond effectively as pragmatists to the two major objections to pragmatism, I examine and critically evaluate the various proposals that pragmatists have offered as a solution to the problem of induction, followed by a discussion of the pragmatic positions on the status of theoretical entities. Thereafter I discuss the pragmatic posture toward the nature of explanation in science. I conclude that pragmatism has (a) a generally compelling solution to Hume’s problem of induction; (b) no specific position on the status of theoretical entities, although something like the non‐realism of the sort developed by van Fraassen seems a defensible candidate for most pragmatists in general, even though there are non‐trivial objections to van Fraassen’s position; and (c) central to the pragmatic conception of scientific explanation is the abandonment of our common conception of truth as a necessary condition for sentences to provide adequate explanations, and a drift in the direction of a contextualist account of explanation.

Notes

[1] However unfortunate, there is an enduring tradition that consists in characterizing the epistemology of James as endorsing the view that the truth of a belief is fundamentally a matter of whether one attains psychological satisfaction in wilfully accepting any proposed belief as true, as if the requirements of scientific methodology could be generally ignored in determining whether propositions about the world are more of less worthy of rational acceptance in virtue their robustly confirmed predictive power. This tendency continues in Blackburn’s recent defence of Craig’s assault on irrational conviction, taking as a paradigm example of such irrational behaviour the position adopted by James (Blackburn 2005, 3–20). For the better interpretation of James, see Tiercelin (2005) and Almeder (1986; 1990). Doubtless, James is partly responsible for this traditional misinterpretation. But, as we shall soon see, James’s position was less a matter of endorsing wilful belief carte blanche than it was a matter of granting provisional permission to accept certain beliefs for which natural science could provide no systemically compelling evidence, either for or against, under induction, when and only when the effects of so believing tend to produce consequences that provide more good or even happiness than would otherwise occur if one had believed the denial of the proposition in question or if one had chosen to believe nothing at all. Why he adopted that position is partly explained, and defended, in Almeder (1986).

[2] When it comes to the epistemic justification of one’s beliefs, the evidence for determining whether the beliefs are successful because the beliefs in question in some way either assert, assume, or imply assertions that are more or less correct descriptions of what they purport to describe, is problematic. Some contemporary pragmatists will find good reasons to assert that the best available explanation for the long term success of any hypothesis or theory is that such hypotheses or theories have within them sentences or affirmations that succeed in correctly describing the external world, even if we have no way to determine which assertions in current theory are in fact the correct descriptions. Such pragmatists may characterize themselves as either structural realists or blind realists (Almeder 1990; Worrall 1989). Other pragmatists, as we will see, leave it an open question of whether successful theories are empirically adequate theories because they, in some important measure, successfully describe an external independently given world. There are many pragmatists who will, for example, simply urge that empirical adequacy or applicative success of our theories or beliefs will be enough to satisfy the fundamental goal of inquiry. For them, while a goal of science may be to search for the truth, and to accept only what we take to be true in the ordinary sense of ‘true’, we are perfectly content to declare our theories confirmed and acceptable when they satisfy the primary goal of allowing for precise prediction and control, and there is no fundamental need to assert that the utility of the theory is a function of its truth rather than say that the long‐term success of hypotheses and theories is simply a matter of natural selection. For them the goal of science is not so much a matter of providing true sentences about an external world as it is to provide systems of belief that shall not disappoint us by way of providing successful instruments for prediction and control, hence for biological adaptation. For examples of pragmatists who affirm as much, those whose works come to mind immediately are Reichenbach (1938), Carnap (1950, 1966), Quine (1953), Sellars (1963), Putnam (1978, 1981, 1982), van Fraassen (1980), Cartwright (1983), and Rescher (2001; 2003).

[3] The second most frequent objection, one also raised against verificationism, is that pragmatists, in defining truth in terms of confirmation conditions confuse truth with confirmation, or the concept of truth with the criterion for truth. Anti‐pragmatists and anti‐verificationists have not understood, however, that pragmatists are not trying to define truth as we ordinarily employ the concept. They are rather best viewed as offering a substitute notion on the grounds that truth as we ordinarily understand it is an empty concept relative to the attainable goal of inquiry because we have no reliable decision procedures for reliably determining which sentences in our language satisfy that alethic concept owing to the fact that we can never attain to anything more that a high probability of truth which is logically distinct from attaining the truth itself. It would have been nice, of course, if they had all made this a bit more explicit.

[4] For a restatement of this objection, see, for example, Blackburn (2005), 7–13.

[5] For a similar view see Worrall (1989, 99–124), Almeder (1992, chapter 4), Putnam (1978).

[6] For this sort of mistake see, for example, Bonjour (1998), chapter 5.

[7] Pierce is often conflicted on whether we seek the truth, with capital T, or just what we deem or sincerely think to be true, even if it is not true (Almeder 1980, chapter 1).

[8] For a treatment of the Cartesianism at the root of demanding a justification of induction and why the demand should be rejected, see Tiercelin (2005, section 1).

[9] See the interesting discussion of this argument by Feldman (2003, 136–137).

[10] For similarly expressed sentiments among pragmatists, see Dewey (1929), Sellars (1963), Carnap (1966), van Fraassen (1980), Cartwright (1983), and Mellor (1991). Mellor offers an especially spirited defence of this pragmatic vindication of induction, as did Reichenbach (1938) and Levi (1967).

[11] Elsewhere, incidentally, I have argued that Peirce finally came to believe that the final irreversible opinion on any answerable question will come, that he argued as much but that the arguments are not terribly persuasive, and finally that we can offer him compelling arguments for that same thesis (Almeder 1986, 1990, chapters 4 and 5). I characterized his position by calling it ‘blind utopian realism’. But a majority of pragmatists adopt the view that Truth, as Peirce understood it, was an ideal regulative concept approached by the scientific community asymptotically in the ideal limit of inquiry under the assumption that scientific inquiry will continue forever and progressively into the future. This same position is evident, for example, in the works of Sellars (1963), Putnam (1981), Ellis (1985), Jardine (1986), Rescher (2003), and Tiercelin (2006). For them, knowledge will be a matter of warranted assertibility relative to available evidence without truth as a necessary condition, since the latter cannot be attained rather than asymptotically approached. For them, as well as for those who, as Rorty once did, simply abandon truth as a meaningless concept, knowledge will be a matter of warranted assertibility relative to contemporary social standards of evidence. Neither position seem to avoid the pitfalls of cultural relativism and classical idealism. But for those who would adopt the Peircean position that truth, in its ordinary sense will be attained if inquiry will continue indefinitely long, such idealism and cultural relativism would not be such a problem. But it does have the serious problem of proving that scientific inquiry will continue indefinitely long answering more and more non‐trivial questions, and that the number of such questions is finitely many. Partly because of the difficulty involved in solving this latter problem it seems easier to most of these Peirceans just to say that truth can only function as a regulative concept. But for the cultural relativism or classical idealism implied in this view of Peirce, it might seem easier and attractive. The more disconcerting point, however, is that there are texts in Peirce’s writings that support both positions although I would argue that the latest opinion he had on the matter was that the final irreversible opinion will come.

[12] For reasons of space, I pass over without comment what has recently become known as Peirce’s abductive argument for realism. This is what Tiercelin (2005, section 6) has termed ‘abductive realism’, which she discusses and defends as Peirce’s best argument for the existence of the external world.

[13] On this item, it is again interesting to observe that Peirce argued strenuously that induction will lead to the truth more often than not because it is absurd to think that induction should fail as often as it succeeds in giving us the truth in the long run (Peirce 1931–58, 2.757n1, 2.758, 2.769).Whether this argument would support the position adopted by Davidson and Rorty to the effect that the vast majority of our beliefs must at any given time be true, is doubtful for the reasons that we have mentioned above.

[14] For Van Fraassen, the ‘empirical adequacy’ of a theory can be either weak or strong depending on whether or not the theory fits or explains all available present and past observable phenomena, or whether it fits all the present, past and future observable phenomenal facts. This distinction is important because he claims the goal of scientific theory is empirical adequacy in the strong sense (van Fraassen 1980), and a ticklish subject is whether anybody can show that strong empirical adequacy can be shown to obtain in any given case, the implication being that the condition is too strong and so only the weaker sense of ‘empirical adequacy’ can stand as a legitimate goal. This objection would sustain, without further comment, the view that van Fraassen’s instrumentalism in theoretical science is really not much better that the sort of anti‐realism offered by Rorty and other cultural relativists.

[15] This author has defended a weaker form of instrumentalism affirming the view that some of our beliefs about an external world, including the theoretical beliefs, must in fact be correct descriptions (even if incomplete) of an external world, although we have no reliable method for determining which of our beliefs are correct descriptions rather than very probably correct descriptions (Almeder 1992).

[16] See note 14 above for the distinction between strong and weak ‘empirical adequacy’ as it relates to this criticism.

[17] For a critical discussion, evaluation, and emendation of the D‐N model of explanation see, for example, Salmon (1984).

[18] Van Fraassen (1980, 125), reprinted in Papineau (1996, 82–92). See also Salmon’s discussion of the differences between his and Van Fraassen’s concept of explanation and why Salmon thought that in the end the pragmatic emphasis on context‐sensitivity of explanation is indefensible owing to the fact that complete theoretical explanations will always be required in science, and that in itself requires an understanding of all the causes and causal mechanisms involved in producing a certain event (Salmon 1984, 127ff).

[19] An historical example, incidentally, of what van Fraassen here advances relative to the context‐sensitivity of an explanation, and how its adequacy is to be determined by the goals or purposes of the inquirer, is one offered earlier by Russell in his famous BBC debate with Frederick Copleston on the existence of God. In fact, we may view Russell’s position as an earlier instance of what some contemporary pragmatists advance as an adequate explanation. Copleston had argued that we do not have a causal explanation for anything unless we admit to the existence of God. Russell replied that if he wanted to know why the tides, for example were higher or lower at different times, or why the one succeeded the other in a regular way, he only needed to know something about the law of gravity, and varying gravitational effects on different locations on the earth under various phases of the moon, in order to predict precisely the times of the high and low tides. Russell, insisted that the demands of the question required for an explanation nothing like a belief in the existence of a God. For the question at hand, an explanation is perfectly adequate as long as the events to be explained could have been predicted rather precisely by appeal to certain laws and conditions under which they apply. Copleston replied that for some purposes, some explanations will be more or less adequate or complete depending on one’s purposes, but one cannot presume to have explained completely why the tides rise in the way they do in different locations, without having an answer as to why the law of gravity exists or works the way it works, or why there is anything at all. The same position Copleston advanced is also offered by Swinburne (1979) in his defence of the existence of God and also in Burr and Goldinger (2004).

[20] Van Fraassen, incidentally, accepts as true some observational beliefs when there is no good reason to think anything else, because otherwise we could not confirm any proposed theory in terms of its deductive implications at the sensory level (van Fraassen 1980). He sees nothing particularly problematic in affirming the truth of some of our common sense beliefs while denying determinable truth associated with theoretical beliefs. More radical pragmatists, such as Sellars, who saw natural science as a matter of common sense gone systematic, would tend to see the claims of common sense and natural science as no different in kind, the alleged distinction between the theoretical and the observational a distinction without a real difference, and hence regard all claims about the external world, whether theoretical or observational, as revisable in the light of future changes in evidence or rules for interpreting the evidence.

Additional information

Notes on contributors

Robert Almeder

Robert Almeder is at the Department of Philosophy, Hamilton College.