Introduction

1As Claudine Tiercelin likes to remind, Peirce was so horrified by the kidnapping of his own ideas that he preferred to call it “pragmaticism”. However, if he had lived enough to know all the deformations of his views which have, since more one century occupied the field of ideas under the name of “pragmatism”, he would probably have withdrawn the term altogether. Pragmatism is a bride who has endured some many successive forced marriages that she may not be able to recognize whom she is married to. In Mon premier testament (1910), Julien Benda proposes that instead of studying the doctrines of philosophers through reading their original works one should rather study the deformations of their doctrines when they reach a larger public. When one reads in newspapers that what is distinctive of pragmatism that truth emerges out of experience and life through consensus, there are good reasons to think that pragmatism has become a kind of blob1.

2My intention here is not to do a history of these deformations, which has been done quite often. Claudine Tiercelin has had a hard time trying to show what pragmatism is not. She believes that there is a core doctrine, mostly Peirce’s, which represents the essence of pragmatism, and that is worth defending independently of its misrepresentations. For my part I have preferred to explore some pragmatists themes. These are : realism, truth, universals, the primacy of practice, belief, knowledge and inquiry, pragmatic encroachment, moral encroachment. On all these topics, I defend views which are, I believe, anti-pragmatist.

1. Realism

3Most pragmatists , from James to Rorty and Brandom and other latecomers, have stressed that pragmatism is not a form of realism: neither about truth (which has been either defined in verificationist terms or held to be undefinable), nor about universals (where the pragmatists defended various forms of nominalism or rejected ontology altogether), nor about scientific theories (where they have defended various forms of instrumentalism and anti-realism) and not about moral values (where they have been close to expressivism and conventionalism). Against such views, Claudine Tiercelin has long defended the view that pragmatism entails realism (Tiercelin 2017). She has held that Peirce’s Scotistic realism entails that there are natures or essences (Tiercelin 2018) and that one should embrace scientific realism (Tiercelin 2003). These positions have put her against the current of the pragmatist river, which according to James, it is impossible to resist: try to put your stick in the river and to alter its course, but ''round your obstacle flows the water and 'gets there just the same"' (James 1909, 55, quoted by Blackburn 1995).

4It would be wrong, however, to say that pragmatists have defended only various forms of anti-realism. They have actually defended the idea that one should reject altogether the distinction between realism and anti-realism. This is why they have been tempted by neutralist positions about truth and ontology such as Carnap’s distinction between external and internal questions, Wittgenstein’s quietism, Putnam’s anti-ontologist turn or ontological pluralism (Price 2011), and in general have held that the realism / and anti-realism distinction is spurious. Their overall message is that one should in general withdraw the distinction, as well as the issues, metaphysical, epistemological or foundational which surround it. Very often they have concluded that philosophy ought to be done without such issues and categories, if it is to be done at all (and quite often pragmatists have been tempted to describe their work as closer to sociology and anthropology).

5If, like Claudine Tiercelin, one intends to defend a pragmatist realism one has to remove two ambiguities. First Peirce’s doctrine may be called realism, but as she points out it can as well be called a form of “objective idealism” inspired by Berkeleyian idealism (Tiercelin 1998). Peirce is not a metaphysical realist in the sense of what Putnam calls external realism, or in the sense of Moore’s or Russell’s early philosophies. In some respects he is closer to a Hegelian objective idealism in Royce’s style than to an Australian realism in Armstrong’s sense. Second Peirce is not alone in defending a form of pragmatist realism. Ramsey himself claimed to be defending, especially with respect to laws and general statements, “the realist spirit”. Following Cora Diamond (1991), Steven Methven (2015) has used this phrase in the sense of an attention to the difference between fact and fantasy, and an attempt to reach the simplest theories. Both refer to the Berkeleyan spirit, but in Diamond’s hands, the “realism” at stake is of quietist sort, i.e anti-metaphysical. At this point if one does not want to fall into confusion, one has to look at the pragmatist doctrines on universals and truth.

6For my part I fully share Tiercelin’s commitment to realism. But I reject the view that realism has to be pragmatist. What follows is a very partial attempt to explain why. I shall leave aside many issues, such as scientific realism, and the pragmatist conception of ethics, on which Tiercelin has done important contributions.

2. Universals

7Claudine Tiercelin has given extensive commentaries of Peirce’s views on universals (Tiercelin 1991, 2002, 2004a, 2007). She has showed the originality of Peirce’s metaphysics, and spent some time distinguishing it from various forms of contemporary neo-Aristotelianism, such as Armstrong’s (Tiercelin 2004a). Peirce’s view of universals is unorthodox because it is framed into his theory of categories. Peirce’s answer consists, first of all, in a rejection of nominalism: there are real properties and essences. Second he also rejects all forms of Platonism: universals are not real in the sense of entities located in some intelligible world. Third he identifies, along with Duns Scotus, universals as “common natures” which are real in the sense of being “would be” or potentialities, not actualities, of things. The potentialities are real in the sense that they are laws, and laws connect real essences. They are dispositional. Neither individuals are things: they are indeterminate or vague. So Peirce rejects the Russellian distinction between particulars and universals. There are only universals. Could a Peircean pragmatist accept Ramsey’s famous argument that there is no real distinction between subject and predicate and between particulars and universals? Tiercelin, in her perspicuous essay on Ramsey’s pragmatism (2004b), has shown, against other commentators, that Ramsey does not evade the ontological problem of universals in order to reduce it to a logical distinction, and moreover to a spurious one. She compares Peirce's and Ramsey’s pragmatism, and claims that they both accept two theses:

1. With respect to the reality of abstract objects (and universals) there is no more to it than the truth of assertions about them. But this truth, according to the pragmatic maxim, rests on the practical consequences of it being true. So the problem of universals is transferred onto the problem of knowing the truth of statements about universals.
2. realism is not the “Platonic” (or nominalistic) view that universals (or particulars if one is a nominalist) are real but the view that universal or singular entities cannot be utterly independent of thought and signification: “The real is that which signifies something real” (5.320). Hence it is impossible to get out from language and representation, as if one could come to simple, ultimate, unanalyzable elements.

8Now, it has to be said that this is somewhat ambiguous. For if the question of the reality of abstract objects depends on the concepts that they express, and on the signification of the sentences and judgments that one makes about them, this sounds like a kind of conceptualism. Universals are not in the mind, nor in language, but they depend in some sense upon language and mind. If their reality furthermore depends upon the truth of these judgments, and if the truth of these judgments depends upon the practical consequences of their being true, this means that their reality depends on their practical consequences of assertions about them. This sounds like a kind of verificationism, unless one adopts a realist conception of their truth conditions. So pending an account of truth, the issue cannot be decided.

9But, Tiercelin insists, Peirce is a realist nevertheless, because, as a “Scotist”, he takes universals to be laws, and laws to have a fundamentum in things. They are actually dispositions, which are “would bes” and have an esse in futuro. But here we meet a dilemma: either the laws are real possibilities, in a modal realist sense, as necessary relations between universals (along lines similar to Armstrong or Dretske), but if so they cannot have the kind of indeterminacy that Peirce wants them to have; or the laws are true because they have practical consequences that we can assess, and then we are back to a kind of verificationism, perhaps an ideal form of it (see below). But if universals are laws and not “things”, either in heaven or on earth, then in what sense are there universals? They have reality, but not existence. If we accept a Quinean criterion of ontological commitment, we cannot quantify over them, and sentences such as “wisdom is a property of Socrates” do not make sense, unless a Peircean accepts second-order quantification over properties. But it is dubious that he could accept such a quantification, because that would mean, presumably, accepting some form of Platonism. Last but not least, Peirce’s view that individuals are vague, and not genuine Aristotelian substances, seems to entail that individuals do not really exist. So do we end up with a realism about universals without universals and without individuals, and universals which are not in the mind but depend on our thoughts and on the way they are verified?

10At this point, I should examine Tiercelin’s claim, according to which realism à la Peirce cannot be understood unless one attends his theory of vagueness (Tiercelin 2019). But the matter is so complex that I can only say the following. Peirce holds the view that vagueness is “real”. But by vagueness, he does not mean what we ordinarily mean, in contemporary philosophy, when we identify vagueness with the presence of sorites predicates such as “bald” or “heap”, which lead, according to the familiar analysis, to incoherence. By “vagueness” Peirce means generality (a feature of laws) and continuity (a feature of the physical world). The problem is that a realist theory of vagueness, according to usual characterisations, ought not to be an epistemic theory of vagueness. It ought to say that there is vagueness in the world, in the very nature of things. But most realists (e.g Russell, Williamson) defend an epistemic theory of vagueness.

11So I am at a loss trying to understand in what sense Peirce, and for that matter a pragmatist realist, could be a realist about universals. This, a Peircean will tell me, is that my spirits embued with Platonism. I must say that, getting old, I am more and more tempted by Platonism (in all quarters: ontology, epistemology, mathematics and morals). I accept second-order quantification and modal realism (with Williamson 2013). Another option, is to go Meinongian, and to say that universals (or properties) are not things because they carry no ontological commitment, although they have some form of reality: they are the truth conditions of subject-predicate sentences, which are unrealised states of affairs (Trueman 2021). Trueman (2021) calls this a “Fregean realism”. This entails either a form of possibilism (perhaps of the Meinongian sort) or a form of deflationism. But I doubt that this kind of view would be appealing for a pragmatist.

3. Truth and success

12James (1909) has said famously that « truth is only the expedient in the way of our thinking, just as “right” is only the expedient in the way of our behaving”. Elsewhere he said that “There can be no difference anywhere that doesn't make a difference elsewhere—no difference in abstract truth that doesn't express itself in a difference in concrete fact and in conduct consequent upon that fact”. Such pronouncements have been understood as a definition of truth as utility or as success in action, and have raised familiar objections, such as Russell’s (1905) claim that James has confounded a definition of truth and a criterion of truth. Ramsey in Truth (1990) has shown that James puts the cart before the horse: it is not because a proposition is useful that it is true, but because it is true that it is useful. The ambiguities of the pragmatist conception of truth lie in how one defines actions, in what one takes truth bearers to be (“conceptions”, beliefs, theories), in how one understands the relationship between belief and action, and in the conditions of success of an action (Tiercelin 2014).

13Ramsey seems to be faithful to the spirit of pragmatism when he adopts in “Facts and propositions” (1926) what has been called the redundantist theory of truth, which says that there is nothing more in “P” is true than the fact that one judges or asserts that P. But in Truth (1990) he holds against James a conception of truth as correspondence. He does not define truth as success, but comes close to what has later been called a “success semantics”: the meaning of a proposition or of a belief content, that is its truth conditions, consist in what guarantees the success of the actions based on that belief, whatever the underlying desires (Whyte 1990, Dokic & Engel 2002).

14What makes the success of an action according to Ramsey’s principle? The fact that a) one believes that if one performs this action such or such desired result will be produced and b) this belief is true. The truth of this belief is presupposed by the success of the action, not the other way round. This reverses the pragmatist view: truth cannot be defined by the consequences of our beliefs, on the contrary it’s the success of actions which depend on the truth of the instrumental belief.

15Typically the belief presupposed by the success of an action is a conditional belief. This raises the question of what it is to believe a conditional, a question which preoccupied Ramsey, who famously held that one can believe a conditional if one hypothetically accept its consequent under the supposition of the antecedent. This is often called “suppositional” view of conditionals and it is associated with the well-known “Ramsey test” for conditional. Edgington (2009) and Mellor (1998) have taken up the view and developed it. On this view conditionals do not have truth conditions, but only assertion conditions (it has been mostly discussed in terms of subjective probabilities, but we can leave this aside here). On Mellor’s view, to believe a conditional is to have a disposition to infer the consequent of the conditional from its antecedent. Now if this belief is a disposition, it has to include a causal element. Mellor (1993) and Mellor and Bradley (2022) claim that to believe a conditional is to be disposed to infer its consequent from the truth of its antecedent, and that the conditional is true if and only if the inference would not cause a true belief in the antecedent to cause a false belief in the consequent. This implies, as Bradley and Mellor show, that the conditional be safe.

16Mellor’s neo-Ramseyan conception of conditional beliefs goes some way towards a realist conception, as Dokic and I (2002) argued, but the fact that according to it conditionals do not have truth conditions means that, although success semantics is a theory of truth conditions, what determines the truth conditions are dispositions to infer, not states of affair or possible states of affair. Conditionals cannot have a truth-conditional semantics. This is somewhat undesirable consequence if conditionals are to be the basis of a realist semantics. Realist semantics for conditionals, accepting material implication, have been given by (Jackson 1987), (Lewis 1973), and more recently (Williamson 2021). So here again, pragmatism, even in its best version, parts company with realism.

17What kind of conception of truth can a pragmatist entertain? Famously, Peirce’s view has been interpreted as a kind of ideal verificationism. Again, in a masterly examination of these ideas, Tiercelin (2014) shows the misunderstandings which lie behind this view. She points out, against Putnam in particular, the realist aspects of Peirce’s view, such as, for instance the fact that he accepted that some unverifiable statements about the past are nevertheless true. But if Peirce’s considered view is that truth required only “that we can reach a state where no further evidence would disturb the belief that we have arrived at” and that “Real pragmatic truth is truth as can and ought to be used as a guide for conduct”, this looks like verificationism, ideal or not. Peirce might go some way towards realism, and certainly a longer way than most of other verificationists (such as James, Dewey, Dummett), but linking truth to the guidance of conduct and the assuaging of doubt is still to make it an epistemic property. One might think, however, that a genuinely realist conception of truth is one according to which truth has nothing whatsoever to do with any epistemic condition.

4. Belief and disposition

18Another great pragmatist thesis, central in Peirce and Ramsey, is that belief is a disposition to act. One can read it in a strictly behaviorist way, by trying to reduce beliefs to dispositions to behavior. But Ryle pointed out, the dispositional analysis can work only if dispositions to behavior are associated to other dispositions (to desire, to infer, to assert) and in connection to ascriptions of other mental terms. This led to the familiar functionalist view: a belief is a functional state, a causal relation between environmental inputs and behavioral outputs, modulo other mental states. This opens the gate to the holism of the mental: how many other mental states are needed to account for one given mental state and its content? Perhaps an open ended list (Fodor & Lepore 1991)? Peirce does not help us very much here: if “different beliefs are distinguished by the different modes of action to which they give rise” (CP 5.400, quoted by Tiercelin 2014, 23), his view falls in the same difficulties as the classical dispositional analysis. A number of interprets (e.g. Hookway 1980) have interpreted his view in functionalist terms. At this point success semantics might help, since it analyses the truth conditions of beliefs in terms of the successes of the action to which they lead. But success semantics should not be understood as a reductive analysis of belief contents (say in the style of teleological semantics) (Dokic & Engel 2002).

19Whatever theory of content we adopt, it seems to me that the dispositional analysis of belief fails. Belief is a functional state, associated to other mental states and to actions, but the mental states to which it is associated are not simply desires, as most belief-desire conceptions of action, but mental states which are more fine grained and which are not dispositions to act: assents (which are both mental and linguistic), acceptances, conjectures, suppositions, and a number of doxastic attitudes (Engel 2018) which all contain a voluntaristic element and involve kinds of mental actions. This involves stressing the conscious element in belief, against the habitual element stressed by pragmatists. It also involves a much more complex conception of the doxastic domain than the dispositional conception. This conception has at least to be refined, and one cannot tie all doxastic attitudes to their dispositional basis and their consequences for action. This means abandoning, one way or another, the Peircean conception of belief.

5. Knowledge, fallibilism and inquiry

20Pragmatists do not have much sympathy for the task of analysing the notion of knowledge. They hold that knowledge, if it is analysable at all, has to be analysed in terms of inquiry, which is, in Peirce's terms, the process of fixation of belief. This involves two commitments: a) that knowledge, if it is analysable at all, ought to be analysed in terms of beliefs, b) if beliefs can be justified, the justification process is dynamic: it entails revisions of our beliefs, removal of others, and change in beliefs. A further commitment, much emphasized in the pragmatic tradition, is that, given that our beliefs are fallible, and call for revision, if knowledge is defined in terms of inquiry, it has to be fallible too. Moreover, inquiry is not, like Cartesian mediation, a solitary process, but a collective one, which ultimately involves the whole community of inquirers. The features of the pragmatist definition of inquiry stand in opposition not only to the traditional definition of knowledge as justified true belief (JTB), but also to the conception, recently revived by Williamson but which probably goes back to Locke, according to which knowledge is a primitive notion not definable in terms of belief (“knowledge first”). Knowledge as inquiry goes against the JTB conception because the process of inquiry does not aim at justifying our beliefs, in the sense of grounding them or making them certain. The pragmatist conception takes the aim of inquiry to be truth, and true belief, rather than knowledge. In other words it is a belief first conception. Knowledge as inquiry goes against the knowledge first conception because it takes knowledge to be primitive and to be a (mental) state, not a process. If knowledge does not consist in belief plus some positive epistemic status, it cannot consist in true beliefs which have reached the stage of assuaging doubts.

21There are a number of other oppositions between a knowledge first conception and a pragmatist conception. On a knowledge first conception, knowledge consists in evidence and of conclusive reasons whereas an inquiry conception involves, in addition to evidence, pragmatic components in knowledge (on this, more below). In other words knowledge first epistemology is evidentalist and infallibilist, whereas an inquiry first conception is fallibilist and set limits to evidentialism. Moreover they do not give the same answer to the problem of scepticism: for the knowledge first view, radical scepticism is undermined because the scenario according to which there are equal reasons to believe that one is a brain in a vat and to believe that one is not a brain in a vat is incoherent. In other words, scepticism is false, not nonsensical.

22Now the knowledge first and the inquiry conceptions are at odds if one understands them as giving accounts (not definitions, since both reject the idea that knowledge can be defined in terms of necessary and sufficient conditions) of knowledge. They can be reconciled, at least in part if they are understood as complementary. On such a view, which I favor, knowledge and inquiry differ. It is one thing to say that an individual (or a group) possesses knowledge and another to say that the individual in question engages an inquiry. It is one thing to ask whether one knows and another to ask questions about what or how we should know, or to check whether one knows. The first is a state, the second is an activity. On the first view, “I know that P but P could be false” is contradictory, whereas it is not on the second. The two views may be in agreement if the definition of knowledge does not entail that our beliefs could be wrong, hence completely revisable, but only that our beliefs could not easily be wrong. On this view, knowledge is safe. This view can be, at least in part, reconciled with a conception of knowledge as inquiry, if the pragmatist relaxes the demands of fallibilism, and accept that we can reach certainty or conclusive reasons which we are not prepared to withdraw2.

23The question of how far one can make compatible the knowledge first (or, for that matter, the traditional notion of knowledge as JTB, since it also considers knowing as distinct from inquiring) and the inquiry conception will depend upon whether the knowledge first theorist is prepared to relax his evidentialism and accept that some practical factors can affect knowledge, and on whether the pragmatist is prepared to accept evidentialist and infallibilist components in his epistemology. But as things stand, they are in frontal opposition. In particular a pragmatist view will have to answer the following question: to what extent does the process of inquiry change our assessment of evidence? To this question I shall come back below.

24Claudine Tiercelin, in Le Doute en question (2005), has defended the inquiry conception, and confronted it with contemporary conceptions of knowledge such as Dretske’s. Her answer to the sceptical problem consists, basically, in deeming sceptical doubts irrelevant: one does not refute the sceptic, one puts his doubts aside. On a knowledge first view, the strategy is different: if there is such a thing as knowledge, and if it can be had for conclusive reasons, the sceptic is not only irrelevant, he is wrong. But she also gives (Tiercelin 2014) “constraints” on knowledge which include the safety conditions. So my question to her is how far she would be prepared to go towards a conception of knowledge as a state, and not as involving true beliefs through inquiry, i.e as a disposition to act within inquiry.

6. Knowledge, assertion and action

25The difference between the inquiry conception of knowledge and the knowledge first conception is equally manifest when it comes to the problem of assertion. Peirce is probably one of the philosophers who have first defined assertion as a kind of speech act committing the speaker to the truth of a proposition within a social context (Tiercelin 2014). Peirce distinguishes clearly the proposition asserted, the act of assent or judgment performed by the subject, and the force of the act within a social setting. An assertion is an act defined in terms of its consequences, and in terms of the responsibility that one takes for performing it. A number of commentators (Tuzet 2006, Marsili 2015) have pointed out that Peirce’s account of assertion is normative, in the sense that one who does not respect the rule of assertion will be corrected, or possibly sanctioned. Now, to say that assertion is governed by norms is ambiguous: by what kind of norm? And by how many norms? The norms may be associated with the content or with the force of an assertion. The content of an assertion is the proposition which is asserted, and it can be evaluated with respect to: (i) truth, (ii) knowledge (iii) some other epistemic position. The force of an assertion is evaluated with respect to different kinds of norms: cognitive, social, aesthetic, or other (e.g politeness). And the norms can compete with each other. Thus I may assert something true, i.e cognitively correct (e.g that the Queen’s hat is ridiculous), but impolite or offensive (one must pay respect to the Queen). Peirce grants this point:

An act of assertion supposes that, a proposition being formulated, a person performs an act which renders him liable to the penalties of the social law (or, at any rate, those of the moral law) in case it should not be true, unless he has a definite and sufficient excuse. (CP 2.315)

26So he accepts that assertion may be governed by different kinds of norms. But he has to grant what all contemporary speech act theorists claim: the primary norm of assertion is epistemic, and it is truth. If a proposition is uttered by a speaker, but without aiming at saying something true (for instance if it is ironical, or a joke) then it is not an assertion3. Now there are good reasons to defend the view that the primary, and unique, norm of assertion pertains to content, and not to the truth of the proposition asserted, but to the knowledge that the asserter has (or is in position to have) of this proposition. I am not going here to repeat the arguments in favor of this view (Williamson 2000, Engel 2008). But it is clearly in opposition with the pragmatist’s conception: if one presents a proposition as an assertion, one presents it as true, and as liable to be accepted or rejected and possibly withdrawn or revised within the process of inquiry. If speakers disagree, as it happens often in the course of inquiry, or if a speaker rejects his previous assertion, they disagree or revise their view on whether they know a proposition, not simply on whether they believe it to be true.

27The view that knowledge is the norm of assertion is not necessarily incompatible with the inquiring process, but it is crucial if, like the pragmatist, one intends to emphasize the intimate relationship between knowledge and action. For there are good reasons to hold that the standard of action is knowledge, rather than belief. If an action does not succeed and if the agent if blamable for it (for instance if a doctor gives the patient a drug which turns out to have disastrous effects), the natural reaction is to say: “But you should have known”. Practical reasoning is better understood as starting from the fact that the agent knows the premise (say: that such or such is dangerous), rather than believes it. So one can formulate the principle: reasons for actions ought to depend on what one knows, rather than on what one believes (one must act on what one knows). If this is correct, this reverses the usual pragmatist claim that knowledge depends on our actions, and is a kind of action or performance (inquiry is such a performance): for it’s not knowledge which depends on actions, but the reverse.

7. Pragmatism, rationalism and the a priori

28Pragmatism is, in all versions, a form of empiricism. So pragmatists usually have little sympathy, to say the least, for the notion of a priori truth. Most of them have rejected Kant’s notion of a priori, and his doctrine of categories. But they have, following the logical positivists, accepted various versions of the a priori, some semantical (identifying the a priori and the analytic), some epistemological (as propositions that one cannot easily discard, or which form the background of our inquiries (Lewis, Quine). But they have never accepted to consider that the basis, or at least a part, of our knowledge is a priori, and that there are hard necessities of thought that cannot be removed.

29The line which Peirce took on these issues is complex. He never defended the kind of view which Mach and Boltzmann held at the end of the nineteenth century, that the mind and its products, including logic and the categories of our thinking, are the long term products of evolution, and he never espoused the psychologism of many of his contemporaries about logic, although he flirted with the view (Tiercelin 2017). Neither did he defend the thesis, which was later the official one of the Vienna Circle, that a priori propositions are those which are true in virtue of linguistic meaning. Claudine Tiercelin has given what I take to be the best discussion of the idea of preformation of pure reason (Tiercelin 1993). She shows at least sympathy for the project of defending this view within the scope of an evolutionary conception of metaphysics and epistemology. I am not sure how far she is prepared to go in this direction, but for my part I am sceptical of any project of deriving logic from evolution.

30I have expressed sympathy for naturalism in metaphysics and epistemology. But I have always rejected this view concerning logic, and I have come to think that epistemology needs a robust notion of the a priori. I think, like Frege, that there are boundary stones of thought, which are norms of our thinking and of logic. These set limits to empiricism. I shall not detail here the precise form that they must take (Engel 2021). Now it may be that a Peircean outlook is not incompatible with a version of the a priori, but it has to be spelled out.

8. Pragmatic and moral encroachment

31The pragmatists have long argued, at least since James’ famous discussion of Clifford, that reasons to act in certain ways can override our epistemic reasons to believe. Since then not only the ethics of belief debate has been going on, under a number of versions, but also the discussion of the pragmatic factors which weight on reasoning, reasons, and knowledge. There are now many strands in these discussions, ranging from probabilistic versions of the will to believe doctrine, to claims about the nature of reasons and evidence. One the most discussed arguments are those which question the epistemological orthodoxy is that only truth-relevant factors, such as evidence and reliability of belief-forming processes, distinguish knowledge from mere true belief. It has been argued in particular that there are contextual factors affecting not only belief but knowledge which have to do with the practical interests and stakes of our ordinary attributions of knowledge and which can make a difference in determining whether or not a subject’s true belief counts as knowledge (Stanley 2005, Fantl & McGrath 2009). In my view, these arguments are moot. At best what they show is that in certain contexts our attributions of knowledge can vary depending on pragmatic factors, but they do not show that the core epistemic notions, such as evidence, reasons and knowledge are affected in their nature by these factors (Engel 2009). The precise workings of these effects both at the semantical and at the psychological level have been studied in a large literature, but they do not have any real impact on the core epistemological theses of evidentialism or reliabilism. In my view, these factors pertain to inquiry, i.e to the reasons that we may have, when we have a belief, or when we claim to know, to check whether our belief is well justified or if it is knowledge. They have to do with the formation, the revision, and the change of belief, not with justification, reasons, or knowledge. Nevertheless if a pragmatist thinker insists on identifying knowledge and inquiry, then it seems to me that he should welcome the pragmatic encroachment arguments.

32Pragmatic factors in knowledge attributions are not the only kind of encroachment of epistemological notions. Pragmatism can be understood, from its beginnings, as an attempt to criticize the classical conception of knowledge which became dominant with Descartes and the empiricists: the view that our beliefs are justified only by our personal evidence, and that we alone are responsible for the exercise of our reason. In opposition to Descartes, the pragmatists, and Peirce in the first place, have insisted on the communal and social, as opposed to the individual nature of knowledge, and they have emphasized the external, rather than the internal features of thought. The extent to which thought contents and the exercise of reason is internal or external is the focus of a huge debate within contemporary philosophy, and there are clearly various versions of externalism and weaker or stronger versions of this thesis. But a rather strong version is the one which emerges out of considerations about knowledge, testimony and the social context of knowers. Drawing both from virtue epistemology, which is in most versions, a form of reliabilism according to which justification is the outcome of a process pertaining to the cognitive formation and the good functioning of belief , and from views which define knowledge as what good informants can transmit within a social community, some philosophers have defended the idea that bad transmissions of testimony and other failures due to social injustice induce cases of “epistemic injustice” (Fricker 2007). The idea that what one knows, and the reasons that one can have for one’s beliefs are affected by the social context of knowing is constitutive of social epistemology, and this project is in many ways congenial to pragmatism. But to what extent are pragmatist thinkers ready to accept that the truth-relevant and evidence-relevant features of knowledge are not only affected, but swept out of this very notion? Peirce had a negative answer to this question, as well as the reliabilitist tradition of social epistemology associated to the names of Alvin Goldman and Ernest Sosa, but most pragmatists have been tempted by a positive answer. According to this line of thought, we should reject the very project of justifying beliefs not only at the internal level of justification and reasons, but also at the individual level: our reasons are external, in the sense that not only they come from our social involvement but are constituted by it. A first question, which I put to the pragmatists is: how far are they prepared to go in that direction?

33If one does go in that direction, there is another step that one can take, which is already implicit in the concept of epistemic injustice as well as in virtue epistemology. This is the idea that epistemic evaluation is not simply epistemic: one can be wrong in one’s beliefs not only because they are not evidence-supported or unreliable or in other respect function badly, but because one is blameworthy, morally, for them. Indeed one can say that one’s belief can be both epistemically and morally blameworthy, and that the two dimensions can compete although they are separate. But there is a harder line: moral evaluation of our beliefs, at the social level, comes first and “encroaches” our epistemic evaluation. This, in a sense, is one version of the ethics of belief, according to which practical and ethical matters determine what one ought to believe. But the twist that the social version adds it that what you believe ought to be determined by what your social community thinks that you ought to believe. This radically externalist line opens the way to a new form of moralism, where the community dictates not only what you think but also what you ought to think (see e.g. Srinivasan 2020) . This goes indeed against one of the main tenets of Enlightenments, that the individual alone is entitled to tell what he thinks and ought to think. Neo-pragmatists often call this view “purism” or “intellectualism”.

34My question to the pragmatists is: to what extent are they prepared to go in this direction, namely practical and moral encroachment on knowledge, given that it is in the spirit of their view? If they are not prepared to go in that direction, how can they accommodate their principles to avoid the most radical consequences of their views, and how can they secure – if they actually intend to do that – the autonomy of epistemic evaluation?

9. A last note: practical vs theoretical

35Pragmatism, in all versions, emphasizes the primacy of practice, and the necessity of basing thought on action. The pragmatist movement has often be celebrated, in particular by Dewey, as the advent of a new form of thinking, through which the traditional conception of philosophy as a theoretical enterprise is overcome by a conception of man as an acting and social animal. The Zeitgeist , during the last century and in this one, has been very receptive to this outlook. The Platonic ideal of the philosopher as looking up at the theoretical skies, and judging from above the sensible world, seems to be dead. I would even risk to say that it is not anymore intelligible to us, and I believe here that James is right that it is useless to put one’s stick against the current of the river.

36But there are a few caveats. The great pragmatists, Peirce, James, Ramsey, have been theoretical men: they practiced science (psychology, mathematics, logic, economy) in an absolutely disinterested manner, just in the same way as scientists of the past. They have never celebrated the prevalence of social over intellectual values. But they have, in my view, defended views, about belief, knowledge, science, metaphysics and reason which tie thought too much to action, and which can be resisted. The few points which have tried here to suggest are meant to be challenges to pragmatism, and possibly point towards an anti-pragmatist view of such matters.