CAUTION AND NECESSITY ∗

In this paper I examine Crispin Wright’s modal anti-realism as based on the availability of a certain attitude of Caution towards judgements of necessity. I think that Wright’s account should be attractive in several ways for modal theorists with an antirealist bend. However, the attitude of Caution to which it appeals has attracted some controversy. Wright himself has later come to doubt whether Caution is ultimately coherent. Here I first address Wright’s worries concerning the coherence of Caution and show that they are unfounded. But then I argue that although the attitude of Caution is coherent, it cannot provide a suitable basis for a non-eliminativist account of necessity. I offer two different objections against Caution. (1) I argue that Wright’s appeal to Caution, if successful, would show not only that modal judgement is nonobjective but also that it is dispensable. Thus, I claim that appeal to Caution would seem to serve more as a threat against a non-eliminativist account of necessity, rather than as a potential adequate basis for it. However, (2) I argue that Wright’s appeal to ∗ Many people have helped me to improve this paper. I would like to thank Axel Barceló, Maite Ezcurdia, Eduardo García Ramírez and María Ponte for their comments when I presented this paper at the II UNAM-ILCLI Workshop. Also, I would like to thank Bob Hale for his written and oral comments to a previous version of this paper. Special thanks go to John Divers, his help in developing the ideas that gave rise to this paper has been immense, and I thank him for allowing me to use some of these ideas here, since they have emerged as a result of a joint collaboration in other projects. Finally, I would like to thank the financial support given by the Research Project PAPIIT IA400412-2: Epistemología y metafísica de la modalidad, DGAPA (UNAM), which has been essential for developing the research from which this paper emerged.

Caution is unsuccessful, for there is no genuine Caution: Caution is a mere verbal attitude.
Keywords: Caution.Conventionalism.Necessity.Anti-Realism.Wright.In this paper I examine Crispin Wright's anti-realism about necessity as based on the availability of a certain attitude of "Caution" towards judgements of necessity (Wright 1980 and1986).As I explain in Sect. 1, Wright's account, in effect a form of conventionalism, has many advantages over other kinds of modal conventionalism and, so, should be attractive to many anti-realist theorists.However, the attitude of Caution on which the account relies may seem to give rise to some doubts.Wright (1989) himself has called into question the coherency of this attitude.However, I think that the prospects of showing that Caution is incoherent are not very promising, as I argue in Sect.2, where I show that Wright's worries are unfounded.So, I think that the problem with Caution is not that it is an incoherent attitude, but that, even if coherent, it cannot provide an adequate basis for a noneliminativist anti-realist account of necessity, as I argue in Sect.3.There I offer two different objections against Caution.(1) I argue that Wright's appeal to Caution, if successful, would show not only that modal judgement is non-objective but also that it is dispensable.Thus, I claim that appeal to Caution seems to pose a serious threat against a non-eliminativist account of necessity.However, (2) I argue that Wright's appeal to Caution is unsuccessful, for there is no genuine Caution: Caution is a mere verbal attitude.And once Caution is shown to be merely verbal, that will be sufficient to dispel the threat of dispensability of the modal that its availability seemed to bring. 1 1.Caution, decision and necessity 1.1 Wright's (1980 and1986) anti-realist project about necessity relies upon the availability of a certain attitude of Caution towards judgements of necessity.2Caution, as conceived by Wright, is an attitude of acknowledging all of the practical and phenomenological features associated with a judgement where we ordinarily deem that it is (absolutely and alethically) necessary that P, while refusing to acknowledge precisely that it is necessary that P (cf., Wright 1980, pp. 452-6).On Wright's view, these features involve crucially: (i) acknowledging the relevant non-modal facts that an ordinary modalizer acknowledges and (ii) the phenomenology of inconceivability which seems to accompany judgements of necessity. 3Thus, concerning any given judgement that it is necessary that P, the Cautious Man (CM) will typically accept P, grant also that he is not even able to conceive in any way that not-P, but still refuse to accept that it is necessary that P. 4Moreover, Caution is supposed to be comprehensive, i.e., directed not only to some specific judgements of necessity, but towards all such judgements, for it is the intention that Caution ranges over the very activity of modalizing, over the very activity of making and assenting to modal claims. 5n immediate reaction to Caution may be a suspicion that it involves some form of misunderstanding or misapprehension.However, Wright's claim is that this is not plausible in virtue of what CM acknowledges, i.e., (i) and (ii).But, then, what are CM's reasons for refusing to take the "step" to a judgement of necessity on the basis of (i) and (ii)?According to Wright, it is not that CM has any specific doubt about the epistemic reliability of the step, rather, Wright suggests that he may have two different kinds of reasons: (a) he feels unmotivated to modalize, unsure about what would be the point of accepting the necessitation of P over and above what he already acknowledges, and (b) he is unwilling to "project" conceiving limitations into "iron" necessities, he feels that such conclusion would be somehow "too strong". 6he role of Caution in Wright's anti-realist project consists in that, if Caution is coherent, and otherwise acceptable, its availability towards all judgements of necessity is supposed to show that such judgements fail to satisfy a necessary condition for objectivity, previously described as a condition for factuality (Wright 1980 and1986), but later characterized as a condition for "cognitive command" (Wright 1992): (CC) A statement or judgement is subject to cognitive command only if it is a priori that differences of opinion concerning it, if not due to vagueness, can be made intelligible only by discovering some material cognitive shortcoming in one or all of the disputants (some relevant mistake, ignorance or prejudice). 7  Clearly, (CC) embodies the idea that putative objective judgements are those with respect to which there cannot be faultless disagreement.So, Wright's crucial point about Caution is that its availability towards judgements of necessity shows that there can be faultless disagreement about them and that, for this reason, they fail to be objective.For, an ordinary modalizer and CM can, allegedly, disagree faultlessly about any ordinary judgement of necessity P*, by virtue of the modalizer holding P* while CM refusing to assert P*, given that CM is supposed to acknowledge all the relevant "modal psychology" notion of necessity] to the beliefs he acquires as the result of logicomathematical inquiry, or certain kinds of imaginative thought-experiment" (p.213).On (b), see (1986, p. 205): the CM sees "no cause to project aspects of our imaginative powers onto reality".If it is felt that (b) may involve, contrary to Wright's intention, some form of doubt about epistemic reliability, that will not be problematic, since (a) by itself may be enough to underpin Caution about necessity.
7 This formulation of (CC) is mine, but follows closely Wright (1986) p. 200.associated with a judgement of necessity.Thus, if Wright's appeal to Caution is successful, it looks like he will have a potentially knockdown argument against a certain kind of modal realism that sees modal judgement as cognitively commanded due to its capacity to track strong modal facts.8

1.2
Wright purports to build a positive anti-realist account of modal judgement based upon this argument for non-objectivity.According to Wright, modal judgement will be truth-apt but merely in a minimal sense, less than fully objective; its truth will lack the representational force of other judgements that we make, such as that the cat is on the mat.Wright's proposal is that judgements of necessity do not involve the recognition of a putative fact, which somehow impels us to assert them, because these judgements involve invariably an element of decision.That is, the step from P to it is necessary that P cannot be seen as one we are cognitively required to make because it is the outcome of a decision.However, the decision, Wright claims, is not simply arbitrary.On the contrary, the decision to modalize is supposed to be regulated and informed by a conventionally defined policy.The policy is delineated for Wright on the basis of what Caution acknowledges, i.e. empirical knowledge of non-modal matters and inconceivability.Hence, the proposal is, roughly, that to treat P as necessary is to engage in a conventional practice regulated by a policy which states in which cases that treatment is merited, namely, just in the cases where we accept that P and we are unable to conceive of not-P (Wright 1980, pp. 456-60 and1989, pp. 215-21).9 It appears that Wright's conventionalism would offer, if successful, some important advantages for modal theorists with antirealist sympathies.On the one hand, it promises not only to provide an account of the modal without buying into the substantive ontological and epistemological commitments associated with realism (those of a truth-making modal reality and a faculty adequate to detect it), but it actually offers to refute realism about necessity.On the other hand, Wright's brand of conventionalism avoids cleanly the objections that have been levelled against other forms of conventionalism about necessity, such as the "radical conventionalism" that is often associated with Wittgenstein or the more moderate form of linguistic conventionalism which was common among logical positivists.These objections concern, for the most part, (A) the capacity of these other brands of conventionalism to account for the necessity of all (potentially infinite) statements of necessity by means of particular (finite) conventions, or, even more strongly, in the case of linguistic conventionalism, (B) its capacity to account for the necessity of any statement of necessity by means of a linguistic convention (thus Quine (1936), Dummett (1959) and Hale (2002)).Yet, by locating the element of convention in the account at the level of the activity of modalizing and not (as the other forms of conventionalism) at the level of particular statements of necessity, Wright's account avoids objection (A).And by avoiding commitment to the idea that the necessity of particular statements has as its source facts about the meaning of those demarcating the relevant inconceivability, drawing on some ideas suggested by Craig (1985) and Blackburn (1986).Roughly, the idea is that the inconceivability which warrants the modal judgement that it is necessary that P is a certain state of incomprehension, of not being able to make anything of the thought that not-P, which is accompanied by awareness that our failure to make anything of not-P is not naturalistically explicable as just a fact about us.See Sect.3.2, where I return to the issue of the gap between inconceivability and necessity.statements it avoids objection (B).Wright's conventionalism avoids also the objection pressed against radical conventionalism to the effect that conventionalism makes our attributions of necessity wholly arbitrary, since, according to the latter, the ratification of every new statement of necessity involves an unconstrained decision.10On the contrary, according to Wright's account, the key decision associated with modalizing is that of engaging in the activity of modalizing, so no new decision is involved in the ratification of particular statements as necessary.Furthermore, by taking the standard phenomenology of inconceivability as the basis of its necessitating policy Wright's conventionalism avoids as well a frequent complaint against modal conventionalist accounts, namely, that they misrepresent the phenomenology of necessity, for it simply does not seem optional which statements appear to us as necessary (see Craig 1975, p. 1).In Wright's conventionalist account it would not seem optional indeed which particular statements strike us as necessary, once we have decided to embrace the convention of ratifying as necessary those statements whose opposite we find inconceivable.Nevertheless, despite these potential advantages promised by Wright's anti-realism, he himself has come to have some doubts on the viability of the account.In particular, Wright has come to doubt that the attitude of Caution upon which his account is based is coherent.However, in Sect. 2 I will argue that Wright's doubts are unfounded, that Caution is not an incoherent attitude.Yet, in Sect. 3 I will argue that this by no means implies that Caution may still provide a suitable basis for an anti-realist account.My main claim is that the problem with Caution is not that it is incoherent, but that it is a mere verbal attitude.

2.1
Wright came to doubt later that the attitude of (comprehensive) Caution, upon which his anti-realist project relied, was coherent (Wright 1989).The doubt emerged from a reconsideration of the result supposedly achieved by the argument Wright (1986) wielded against a Quinean holistic-pragmatic methodology.The conclusion of the anti-Quine argument was that it cannot be that, as Quine (1951) claimed, all judgements are subject to appraisal based only on global-pragmatic considerations.In any system of belief, Wright maintained, there are some judgements, some judgements of logical consequence, whose ratification requires a different explanation, namely, that they need to be appraised in a purely a priori fashion.I will not dwell on the intricacies of the anti-Quine argument, since that would be a diversion from my main concern here.What I will examine, instead, is, even if we assume that the anti-Quine argument is sound, does that result about apriority transmit to the case of necessity according to the strategy devised by Wright (1989)?For, if the transmission succeeded, comprehensive Caution about necessity would be incoherent.Wright's (1989) worry about the coherence of Caution starts from the idea that the result supposed to be achieved by his anti-Quine argument concerning the need for a priori judgements may transmit to the case of judgements of necessity.The worry is justified, Wright argues, because it seems that anyone, whether realist or anti-realist, has to admit that there is some sort of important connection between apriority and necessity.It appears, Wright claims, that the best explanation of why a statement can be ratified a priori is that it is necessary.So, it seems that the minimal commitment about the connection between apriority and necessity that one ought to assume is that there is some form of linguistic coordination between the two notions, to the effect that "prescinding from certain possible exceptions, the a priori ratification of a judgement is indeed the ratification of its necessity" (Wright 1989, p. 223).The possible exceptions Wright has in mind are, of course, the familiar Kripkean cases of contingent a priori statements (Kripke 1972).The coordination thesis amounts then to the following claim: (leaving aside Kripke's alleged counterexamples) if X ratifies a priori P, X ratifies P's necessity.
If the coordination thesis were correct, the result of the anti-Quine argument about apriority would, in fact, transmit to the case of necessity.For, the anti-Quine argument establishes that X has to ratify a priori some statement P concerning whether something is a logical consequence of other things; but, according to the coordination thesis, the ratification of P as a priori will be sufficient for actually ratifying P as necessary.So, X has to ratify some statement as necessary.But, if we all had to make at least some judgements of necessity, Cautious refusal to assent to no judgement of necessity would not be a coherent option.Therefore, if the transmission argument were sound, Caution would be incoherent (Wright 1989, pp. 222-5).

2.2
Although Wright takes the transmission argument very seriously, his immediate reaction is that, even if it were good, that result would not call for rejection of his anti-realist project about necessity, but only for modification.For, Wright claims, the argument would not establish that Caution is incoherent in every form it could take, but just in its comprehensive or global form, i.e., with respect to all judgements of necessity.According to Wright, there would still be the option of restricting Caution to range only locally over some judgements of necessity, even if not over all of them.As long as the argument maintains, as it seems to do, a very general character and does not state which judgements of necessity (of logical consequence) in particular have to be specifically ratified, it seems possible, Wright claims, that this kind of local Caution, which he calls "Eccentricity", can be directed to any judgement of necessity whatsoever, provided that some other (unspecified) judgements of necessity are ratified.Wright expresses as well some worries about whether Eccentricity is ultimately coherent (Wright 1989, pp. 229-30).I will not discuss here the option of Eccentricity because I think that Wright's concession is uncalled for.There is no need to move to local Caution because the transmission argument is unsound.So, comprehensive Caution seems to be, as far as the transmission argument is concerned, a coherent attitude. 11 As stated, the main question in the argument is whether the result about the necessity of making judgements a priori, assumed to be established by Wright's anti-Quine argument, transmits to the case of necessity in such a way that it entails the necessity of making judgements of necessity.The key premiss in the transmission argument is the coordination premiss that, if X ratifies a priori P, X ratifies P's necessity.I think that the transmission argument fails because this premiss should be rejected, at least in this context.There are two main, related, reasons, why the premiss fails.First, the crucial thing to remark here is that the premiss looks precisely like the kind of premiss that does not go through for someone affected by the Cautious attitude.However, this is not because this premiss is itself a modal statement, which, therefore, falls under the scope of Caution.Rather, the premiss seems to presuppose that there is a distinctive motivation to make judgements of necessity, something that is not granted by the Cautious Man (CM).
As we have seen, Caution is an attitude of refusal to assent to all statements of necessity which, essentially, derives from a lack of 11 There may be reasons to think, however, that Eccentricity is incoherent.Given that the Eccentric Man modalizes in some cases, it seems doubtful that his refusal to do so in other cases would be intelligible without an explanation.And, if an explanation were given, it would very likely disclose some important mistake or disagreement on non-modal matters on the part of the Eccentric Man.Hale (1997) raises this complaint against Eccentricity.motivation to assert those judgements.CM is a character who does not feel motivated to make modal judgements, who feels "out" of the practice of doing it because he does not see the point of modalizing.It may be helpful to have in view now the way Wright (1989) tries to describe more thoroughly the attitude of Caution, as a response to Hale's (1989) claim that it is some form of philosophical agnosticism.According to Wright, we have no difficulty in envisaging a demodalized version of our language, a language just like ours but with no explicit modal idioms.Now, the Cautious Man (or as Wright (1989) calls this character, "Hero") would be someone trained in this language who is nevertheless competent in all other relevant areas, like logic and mathematics.Now Wrights asks us to suppose that we, ordinary modalizers, interact with Hero and that, after a few moments, it becomes clear to Hero that we are using a concept that he lacks, since we classify certain judgements as necessary (typically judgements of logic and mathematics).Hero's reaction, according to Wright, is that of refusal to engage in this practice of modalizing, for although Hero understands what we mean by "necessary", he does not see the point of making these judgements.As Wright puts it: Suppose it had not occurred to Hero to think of these judgements as in any way set apart.No doubt he had conceived of certain of the truths of logic and mathematics as pretty obvious -even as, for practical purposes, completely certain; and no doubt he had recognized that the methods of logic and mathematics are distinctive.But it had never crossed his mind … to regard the methods of logic and mathematics as distinguished not just by their formal and reflective character but in the necessity of their products.And now, when it does cross his mind, because we bring it to his attention that we so view these matters, Hero finds himself without any sense of why we want to make such claims.Perhaps he has an inchoate philosophical worry about how anything which finite, rooted-in-the-actual human beings could accomplish, could amount to knowledge of such cosmologically impressive scope.But all he needs to feel -for our purposes-is 'out of it': a spectator on a practice with concepts whose basis remains obdurately mysterious to him.(Wright 1989, p. 211) Now, I think this fuller description of Caution makes it clearer why the coordination premiss fails, why it cannot provide a coordination, and hence a transition, between apriority and necessity, such that the a priori ratification of P will lead to the ratification of P as necessary.For this premiss can do this only if it is presupposed that we are bound to modalize.The claim that some modal judgements are naturally seen as ensuing from some other kinds of judgements that we make can work only if we are already taking for granted that there is a distinctive need or reason to modalize rather than to abstain from doing it.But, if the overall motivation to modalize is still sub judice, the claim by itself is then powerless.Thus, the coordination premiss seems to have any force only if it presupposes precisely what is in question for CM: the motivation to modalize.
Second, the previous reply needs to be supplemented, for the coherence of Caution is still threatened by the intended basis of the coordination premiss.The basis of the coordination premiss is Wright's thesis that the best explanation of someone judging a priori that P is that P is necessary.I have claimed that CM can legitimately reject the coordination of the notions of apriority and necessity.But in order to do this successfully, CM has to do two things: (i) explain how he can accommodate the result of the anti-Quine argument that he has to make some judgements a priori and (ii) provide a plausible alternative, non-modal, explanation of his making some judgements a priori.Recall the way Wright (1989) describes CM.According to this, CM is someone competent in logic and mathematics.As Wright claims, CM is aware that "the methods of logic and mathematics are distinctive"; for he has certainly "conceived of certain of the truths of logic and mathematics as pretty obvious -even as, for practical purposes, completely certain" (p.211).It is not clear whether Wright thinks that CM has the concept of the a priori before we interact with him.But let's suppose that he does not; that his regarding the truths of logic and mathematics as pretty obvious and as completely certain (for practical purposes) does not amount to regarding them as a priori true.Anyway, we may introduce him to the notion and to Wright's anti-Quine argument, so that he is able to see that he has to accept that at least some judgements of logical consequence have to be ratified a priori.I do not see any reason why he would not accept this.If so, he may very likely take those judgements of logic and mathematics that he found to be obvious as being in fact subject to a priori ratification.
But why is he able to accept these judgements as a priori but not as necessary, and what explanation can he give of a judgment's aprioriy that is at least as plausible as the explanation by means of the supposition that it is necessarily true?Well, I think CM has a plausible reply to this challenge.As we have seen, CM acknowledges the cognitive trappings of judgements of necessity concerning standard cases, judgements of logic and mathematics.So, for example, concerning the judgement that (A & B) → A, he accepts that it is the case that (A & B) → A, and that he is unable to conceive that (A & B) is true and A is false.But still he refuses to accept that it is necessary that ((A & B) → A).However, reference to precisely these cognitive trappings may provide all that is required for a plausible explanation of his making judgements a priori.In short, I take it that CM is entitled to hold that the right explanation of why he (and other people) believe a priori that P is that he (we) believe that P and he is (we are) unable to conceive of not-P.That is, CM will, in effect, acknowledge these cognitive features as the cognitive trappings of both judgements of necessity and a priori judgements.This suggestion concerning a priori judgements would need, of course, to be argued more in detail in order to be ultimately viable, but at least it looks prima facie plausible, since all the standard examples of statements concerning whose necessity Wright's Caution is directed are a priori (mainly true statements of logic and mathematics).And, moreover, CM may argue that there are, at least, two reasons why he takes "the step" from satisfaction of those cognitive trappings concerning P to accepting that P is ratified a priori, but does not likewise takes "the step" from satisfaction of them to accepting that P is necessary.First, given the anti-Quine argument he sees now what is the point of accepting some judgements a priori, but he does not see that there is any point in accepting some judgements as necessary.And second, CM may see no objection in moving from acknowledging these cognitive trappings to acknowledging that one thereby ratifies a priori that P, given that, by acknowledging apriority, he is only stating the way in which he believes that P, rather than granting the "stronger" claim that it is, in fact, necessarily true that P. In other words, a conclusion of apriority may not be for CM, unlike a conclusion of necessity, a conclusion of "such cosmologically impressive scope" (Wright 1989, p. 211).
So, I take it that CM has at least prima facie plausible resources that may allow him to accept coherently that he ratifies a priori some judgements without thereby being committed to making any judgements of necessity.Hence, the coordination premiss fails against Caution.And, therefore, Wright's transmission argument fails to transmit to the case of necessity the anti-Quine result about apriority and, so, fails as an argument against the coherence of comprehensive Caution.It seems that, in order to threaten the coherence of Caution, an argument would have to establish more directly why anyone has to be a modalizer. 1212 Peacocke (1999, pp. 187-88) argues that Caution is incoherent.According to Peacocke, CM has to say something about one of his central principles of possibility, the Modal Extension Principle (MEP), a principle which tells us "how contents are to be evaluated with respect to alternative possible circumstances" (p.187).Either CM gives an alternative to (MEP), which does not seem plausible, or he accepts (MEP).But if he accepts (MEP) then his attitude will not be coherent because CM will have to accept that there are some cases in which, once we have given semantic values to the logical constants of A, we have to accept not only that A is true in the actual world but, by (MEP), that A is true in any possible circumstance.However, Peacocke's objection fails for the same reason that Wright's fails: it 3. Two problems with Caution: dispensability and generalization.

3.1
In the face of the failure of the transmission argument Wright would, apparently, endorse the view that Caution could still provide an adequate route to an anti-realist account of necessity, as can be gathered from the following passage: We suppose (i) that a priori judgement will play a part in the operation of any coherent system of belief, and (ii) that non-cognitivism about necessity had probably better grant a role for judgements of necessity as co-ordinate to (some) a priori judgements.If supposition (i) is wrong, then global Caution about necessitated judgements is, after all, at the service of the non-cognitivist about necessity.(Wright 1989, p. 228) I disagree.In the previous section we saw that showing that Caution is an incoherent attitude is not an easy matter.However, I think that the main problem with Caution is not whether it is coherent or not.By focusing so much on the coherence of Caution, Wright overlooked that there are two more serious problems with this attitude.
The first problem is that Caution cannot provide an adequate basis for a non-eliminativist anti-realist account of necessity because, if Caution were coherent and otherwise acceptable, then its availability would show much more than Wright, probably, intended: it would show not only that modal judgement is non-objective, but, also, that modal judgement is dispensable.As we have seen, particularly in Sect.2, Caution involves a refusal to engage altogether in any modal judgement.The Cautious Man (CM) is a character who is able to speak presupposes what is in question for CM, the motivation to modalize.It presupposes that we are bound to make some modal judgements (about possible circumstances of evaluation, etc.).a non-modal language, and so, apparently, operate without any need of modal concepts and modal judgement.As Wright puts it, It would hardly be an inconvenience to speak such a language: science, mathematics, (non-modal) logic, psychology, literature and the artsalmost every area of human expression and inquiry could be prosecuted and discussed exactly as it is now.Only philosophy and modal logic would be exceptions.(1989, p. 211) So, even without making any modal judgement and speaking any modal word, all of our basic intellectual activities and practices may still be conducted without any difficulty.The only activities that will be affected are, obviously, philosophy and modal logic.But apart from these activities, it seems that, a Cautious Man, who abstains from making any modal judgement and of speaking any modal word, can live a perfectly effective theoretical and practical life, since it does not seem that it will have any crucial disadvantage by thinking and speaking nonmodally.The idea that seems to underlie, perhaps unintendedly, this characterization of Caution by Wright, is that modal judgement does not seem to have any essential function, purpose or benefit in our theoretical or practical lives.For that seems like a necessary requirement for the possibility of Caution: if modal judgement had an essential function to fulfil, then it seems that thinking and speaking completely non-modally as CM, allegedly is able to do coherently, would indeed bring a crucial disadvantage.So, it seems that if Caution is coherent then modal judgement turns out to be dispensable, in the sense that it is not a necessary activity for anything important that we do.If Caution is coherent, we may live without modal judgement, so it seems that modalizing is only an optional feature of our thought and language: something we may do if we wanted to do but that we do not need to do.
The dispensability of modalizing that seems to be entailed by Caution may be better appreciated if we compare Wright's view about the case of modal judgement with his view about other judgements, such as judgements about other minds, about the past or about the material world.Wright (1989) intends to characterize Caution less philosophically, as we saw in Sect.2, in order to claim that Caution is not available concerning the latter kinds of judgements.Thus, Wright asks: "can anything analogous be constructed for judgements concerning other minds, or the past, or the material world?Well, if so, it must be possible to characterise languages which Hero could practice during his innocence of the concepts of other minds, matter or the past.But what are they?"(p.212).He answers that it is not even prima facie plausible that this is possible, and so denies that someone can be Cautious concerning any of these kinds of judgements.And it is clear why this is so.Imagine a putative Cautious Man concerning judgements about the material world.This CM would have to abstain from speaking any word that purported to refer to the material world and, so, from making any judgement about the material world.But how could anyone manage to do this?And even if someone were able to do this, it is clear that abstention from making judgements about the material world would be a crushing disadvantage.Anyone who managed to do this would be deprived of a massive amount of knowledge and abilities to interact with other people and with her environment.Even the survival of such a creature would be threatened by her abstention from judgements about the external world.And it is clear that these disadvantages would stem from the fact that judgements about the material world have an essential function or role in our thought and language.Given this essential function, this kind of judgement is not dispensable, we cannot abstain from doing it.We need to do it in order to live an effective theoretical and practical life.So, this is the reason why genuine Caution is not possible concerning judgements about the material world: its possibility would require that such judgements be dispensable, an optional feature in our thought, not an activity that we need to perform.Thus, the possibility of Caution concerning a discourse D entails that D is dispensable, in the sense that it is not a necessary activity for anything important that we do.This connection between Caution and dispensability is, I think, a reason to consider Caution as a potential appropriate tool and basis only for a very radical version of modal antirealism, modal eliminativism, but not for other less radical, noneliminativist varieties, such as error-theoretic versions, like modal fictionalism, and "truth-theoretic" versions, like quasi-realism.Modal eliminativism is a very radical anti-realist variety which holds that we are never justified in making any modal judgement, either because all of them are systematically false or because they do not have any function for us, or because modal notions are simply unintelligible. 13liminativism holds further that given these deficiencies of modal judgement we should eliminate the practice of modalizing from our thought and talk.We should speak a non-modal language and think only non-modal thoughts.It is clear then how the possibility of genuine modal Caution would support such eliminativism: the possibility of Caution would show that such elimination of the modal can be done coherently, that nothing would be lost if we decided to abstain from modalizing.
However, modal eliminativism is an alternative to modal realism that most modal theorists consider as too radical and inappropriate (Wright (1992, p. 10) included).So, most modal anti-realists consider that alternatives which keep our modal language and thought in place are preferable in principle, since they tend to involve less departures from our ordinary way of thinking.It is clear that less radical antirealists are put in trouble by the eliminativist threat that Caution seems to bring, simply because they intend to keep our modal judgement in place.However, more specific commitments of each version of anti-realism show more appropriately how Caution is a threat to them.Take the case of an error-theoretic version, like modal fictionalism.The modal fictionalist accepts, in part, the negative thesis of the eliminativist, that we are never justified in believing any modal judgement, though the specific reason here is only epistemic, that such judgements are systematically false.However, it accompanies this negative verdict with a positive verdict: although modal judgement fails in tracking truth, it nevertheless has another subsidiary important function to perform for us, which is what leads us to keep the practice of modalizing.And the performance of this subsidiary function is supposed to be compatible with the falsity of the judgements. 14It is then the development of this positive element of the error-theoretic story that is threatened by the possibility of genuine Caution.For, if Caution is possible, the prospect that there is such a putative important function for modal judgement to deliver looks rather bleak.As I claimed, it is a necessary condition for the possibility of genuine Caution about modalizing that this activity is dispensable, that one can decide to avoid practicing, without there being any serious disadvantage.And this can only be the case if modalizing does not have an essential function or role in our thought and talk.Of course, the possibility of Caution is only a threat to error-theoretic anti-realism.It may be that there is a way of reconciling the dispensability of the modal with there being a function for modal judgement, which, although it is not strictly essential for us, it is certainly useful.However, my claim here is only that the dispensability of the modal that Caution seems to entail, puts error-theoretic anti-realism in a position of disadvantage, from which it is not clear how it may recover.
The case of truth-theoretic modal anti-realism is different, for here the eliminativist threat that Caution brings impacts more directly with essential components of the theory.For example, a truth-theoretic modal anti-realism like modal quasi-realism (Blackburn 1986) holds that we are justified, epistemically, in making some modal judgements.Some of these judgements are even true, although in a different way from realist truth, for quasi-realism intends to construct truth in a very modest way, out of the mental states that underpin modal judgement.Some of these judgements are warranted and some even amount to knowledge.Moreover, they perform for us an important function, in particular, the mental states that underpin our modal judgements are systematically related with other mental states that do not have overtly modal content.So, none of the potential sources of modal eliminativism have for the quasi-realist any ground: there is no systematic error in modalizing, modal notions are intelligible and modal judgement has an important function in our thought.It is, in particular, the view of quasi-realism on the function of modal judgement that makes Caution incompatible with quasi-realism.For, as we saw, if Caution is possible, modalizing is dispensable, devoid of any important function for us.But quasi-realism may be considered actually a function-driven approach in modalizing.The alternative that Blackburn proposes to what he calls "truth-conditional realism" is one that takes as its fundamental explanandum not the truth-conditions of modal judgement, but its function or role in our thought: "the alternative starts (and, I shall urge, ends) with our making of those utterances: the thing we intend by insisting upon a necessity or allowing a possibility.We could call it a 'conceptual role' or even a 'use' approach, but neither title is quite happy, for neither makes plain the contrast with truthconditional approaches that is needed" (1986, p. 54).Blackburn actually presents an argument that aims to show that making modal judgements, judgements of contingency in particular, has an essential function for our having any understanding of the world (1986, pp. 64-66).So, Caution seems like an entirely inappropriate tool for quasi-realism: if genuine Caution is possible, modal quasi-realism is out of business.
In this way, endorsement of Caution may serve to support an eliminativist objection which argues for the dispensability of modalizing, and, so, may be of any help only to the kind of modal antirealism known as modal eliminativism.Non-eliminativist modal theorists, whether anti-realists or realists, have then every reason to be anxious about Caution.

3.2
Blackburn (1986, p. 128) is concerned about Caution and, apparently, about the threat of dispensability it seems to bring.For this reason, even if he also shares sympathies with an anti-realist account of modalizing, he is anxious to show that genuine Caution is not possible.Blackburn claims that he finds no gap between what the Cautious Man (CM) is supposed to acknowledge and what he is supposed to refuse to acknowledge, so, according to him, CM is modalizing (ibid, pp. 133-34, 136-37).Yet, Blackburn offers no explanation of this claim or of its consequences.Here I attempt to develop an objection against Caution, inspired by Blackburn's, that appeals essentially, to CM's "modal psychology" and to a certain constraint that Caution ought not to generalize (NG), which I will explain now briefly.
It is agreed by all that it is not enough that Caution about necessity be coherent, it has to satisfy also a constraint of nongeneralization (NG).Hale (1989) has perspicuously formulated this constraint.According to Hale, it is essential that Caution, however construed, does not generalize, that is, is available unrestrictedly to all other kinds of non-modal judgements.For, otherwise, if appeal to Caution were to succeed in showing that modal discourse is nonobjective, then its general availability should be taken likewise to call into question the objectivity of all other non-modal discourses.But such generalization would amount, in effect, to a reductio of Wright's anti-realist argument, for the whole point of appealing to Caution is to show that modal judgement lacks the objectivity that other judgements that we make seem to possess.In what follows I will argue that Caution does not satisfy (NG).
Let us remember the fundamental psychological traits that, according to Wright, Caution has.Caution is supposed to be an attitude of refusing to judge that it is necessary that P, while at the same time acknowledging all of the practical and phenomenological features plausibly associated with our ordinary judgement that it is necessary that P. According to Wright, the latter involve (i) accepting that P, and, crucially, (ii) inconceivability of not-P.The reason why Caution needs to involve (i) and (ii) is because only in that way will it be strong enough to provide the means for showing that modal judgement does not satisfy the necessary condition for objectivity (CC).That is to say, only in that case will Caution be strong enough to provide the means for someone, CM, to disagree faultlessly with an ordinary modalizer about any judgement that it is necessary that P. But, perhaps it may be thought that Caution does not have to be as strong as Wright thought.Is weaker not better if we want to show that modal judgement does not satisfy (CC)?If Caution were a less demanding attitude, then it would seem that it would be easier to have, and so it would be easier to show that there is faultless disagreement concerning modal judgement.So imposing fewer constraints on the attitude would seem to be desirable.However, the crucial thing is not how easy Caution can be had, but how plausible it is that someone can be Cautious and still not be guilty of some cognitive shortcoming.For, suppose that Caution did not involve (i), in that case, the disagreement between an ordinary modalizer and CM could be easily explained as involving some sort of non-modal disagreement between them, about P, something that may very probably need to be explained in terms of some cognitive shortcoming on the part of CM.For example, suppose that CM says: "I refuse to accept that it is necessary that 2 + 2 = 4".Then we may ask him: "but do you at least accept that it is in fact the case that 2 + 2 = 4?".And if he responds, "no, I do not", then we may begin to think that he is simply incompetent in basic arithmetic.Alternatively, suppose that Caution involved (i) but did not involve (ii), so that CM claimed that, somehow, he is not unable to conceive of not-P.So, for example, CM accepts that it is in fact the case that 2 + 2 = 4, but he says that he is not really unable to conceive that it is false that 2 + 2 = 4, rather, he is quite able to conceive that, say, 2 + 2 = 5.That is, in effect, CM will not make any distinction between his appraisal of judgements that ordinary modalizers will take as contingent, such as that the cat is on the mat, and his appraisal of judgements that are standardly taken as necessary, such as that 2 + 2 = 4.In both cases he will grant that the relevant judgements are true and will hold that he is able to conceive that it is false that 2 + 2 = 4, just in the same way as he is able to conceive of the cat being somewhere else.However, that would be simply to lose the point of the attitude of Caution.CM was supposed to mirror the ordinary modalizer in her usual distinctions between necessary and contingent judgements without actually making any modal judgement.So, if CM does not grant inconceivability, which Wright has assumed is the standard basis for making those distinctions, two scenarios open.Either he cannot substantiate his claim of conceivability, in which case some kind of misconception or error on his part would sooner or later come up (e.g., he is merely conceiving that someone announces that top mathematicians have discovered that 2 + 2 = 5).Or, if he can establish his claim of conceivability concerning every P we deem to be necessary, then either we are, thereby, shown to be mistaken in believing any necessity (for he is showing us a way in which P might indeed be false) or, in the end, no one will be mistaken.The postulation of Caution about necessity would be anyway pointless according to either of these two outcomes.For, in the former case, we would lose the initial datum, the judgements of necessity of the ordinary modalizer; and, hence, the disagreement.While in the latter case, faultless disagreement about necessity may ensue, but Caution would not be necessary any more, since the result would be established more directly by confronting an ordinary modalizer who judges that it is necessary that P with an unorthodox modalizer who judges that it is not necessary that P. Thus, Caution has to involve both (i) and (ii), otherwise it would be too weak, or it would not be necessary, to provide a route to show that modal judgement does not satisfy (CC).Thus, Caution is going to be of any use in Wright's anti-realist project, it needs to make someone indistinguishable, in all relevant respects, from an ordinary modalizer who judges that it is necessary that P, save from the judgement that it is necessary that P (or any other modal judgement).
My claim is that, by strengthening the Cautious position in this way, Wright has made it, in fact, too strong.By making CM and the ordinary modalizer indiscernible in all the relevant respects he has caused the difference between them to be merely verbal, in the sense that it consists just in the lack of explicit use of modal words.That is, if Caution involves (i) and (ii) (and if these are all the practical and phenomenological features plausibly associated with the judgement that it is necessary that P, as Wright seems to suppose), then all that CM lacks in his pretended refusal to modalize is modal words, but not, crucially, modal judgement.For, in virtue of acknowledging (i) and (ii), Caution would involve internal psychological conditions that seem to be prima facie sufficient conditions for believing that it is necessary that P. Sufficient at least in a cognitive (and causal) sense: anyone who comes to satisfy conditions (i) and (ii) concerning P thereby comes to acquire belief that it is necessary that P. 15 15 A stronger thesis of sufficiency would be that of constitutive or metaphysical sufficiency: when X comes to satisfy (i) and (ii) concerning P, X thereby comes to believe that it is necessary that P because (i) and (ii) are constitutive of such belief.On this view, tokens of belief in necessity that P are Establishing this claim of sufficiency would require much more than I am able to do in this paper.But I take it that the claim is at least prima facie plausible.First, a thesis similar to this seems to be assumed implicitly by many philosophers, such as Kripke (1972).For those philosophers argue that when we come to genuinely conceive (in the relevant way) that P we thereby establish that P is possible and, so, come to believe that P is possible.So that it seems plausible to hold that in those cases in which we believe that P, and are unable to conceive of not-P, that is, have repeatedly failed to conceive of not-P and are confident that we will never do it, we have established, at least pro tem, that not-P is impossible and, so, come to believe, at least pro tem, that it is necessary that P.This is suggested also, from a different perspective, by the view held by Kripke (and even by Wright 2002), that a claim of necessity that P is defeated when we come to genuinely conceive that not-P, what Wright (2002) calls the "counterconceivability principle".When this happens we withdraw belief in the necessity that P, and this may only be because such belief essentially involves inconceivability of not-P.
Second, all those cases in which it is claimed that there may be a gap between inconceivability and necessity seem to be cases where what is involved is rather some form of sensory unimaginability or where the corresponding non-modal belief is missing.As an example of the first sort of case take cases like those discussed by Blackburn (1986), such as the case where someone claims that she fails to conceive/imagine that there is an extra primary colour (additional to those there actually are), but she does not thereby come to believe that it is impossible that there is an extra primary colour (and, so, that it is necessary that there is not an extra primary colour).The problem with this, as Blackburn correctly points out, is that this sort of inconceivability is some sort of failure of sensory imaginability.The identical to tokens of the conjunctive state of (i) and (ii).However, this stronger sufficiency thesis is not required by my objection against Caution.
subject is trying to imagine visually how would an extra primary colour look like.And she is, of course, unable to do that.But, as Blackburn points out, this does nothing to show that there is not a secure link between inconceivability and necessity.For, the reason why such imaginative or conceiving blocks do not give rise to a judgement or belief in necessity is because there is a clear (naturalistic) explanation of why we have such blocks: our failure of imagining/conceiving an extra primary colour is only to be expected given that what we can visually imagine is constrained by our previous visual experience.Thus, we can see that this imaginative block is only a fact about us.On the contrary, as in the cases that are relevant for the dialectic of Caution, such as those of truths of logic and mathematics, our failure to conceive that they are false, of making anything of the thought that they are false, is not naturalistically explicable as merely a fact about us.That is why, in these cases, inconceivability gives rise to belief in necessity.
As an example of the second kind of case, take the case of the strong Goldbach conjecture that every even integer greater than 2 is the sum of two prime numbers.We certainly are unable to conceive how the conjecture would be false.We are unable to make anything of the thought that there is an even integer greater than 2 that it is not the sum of two prime numbers (and this is not explicable as just a fact about us).However, we do not believe that it is necessary that every even integer greater than 2 is the sum of two prime numbers.But it is not difficult to see why this is so.The reason we do not believe that the conjecture is necessary is because we do not believe that it is true.That is, we only satisfy (ii) concerning the conjecture but fail to satisfy (i).If we came to believe that the conjecture is true, it seems plausible to say that we would thereby come to believe that it is necessary (given satisfaction of (ii)).
All this, of course, can only establish at most the prima facie plausibility of the claim that (i) and (ii) are sufficient for belief in necessity.Much more would have to be done in order to establish that this is really the case.However, this is enough to pose a serious challenge to Caution.And, moreover, I would like to emphasize that, as I will claim in a moment, my challenge to the attitude of Caution does not depend entirely on conditions (i) and (ii) being sufficient conditions for the acquisition of belief in necessity.The reader is invited to introduce her own conditions that she takes to be sufficient for acquiring belief in necessity as the antecedent conditions associated to Caution and, still, my challenge to Caution will remain.
So, I take it that it is at least prima facie plausible that by satisfying (i) and (ii), the Cautious Man (CM) has acquired belief in the necessity that P. But, in general, if X believes that P then X judges that P, even if X's judgement is not fully verbalized.Hence, Caution, as an alleged refusal to engage in modal judgement, seems to consist just in a mere refusal to verbalize a modal belief and, so, after all, a modal judgement.If this is correct, Caution should not be considered, in fact, as a different option apart from modalizing.The Cautious Man is modalizing, he just somehow thinks that something more is required for him to do this.Thus, genuine Caution is impossible, someone cannot, instead of modalizing, be "Cautious", for Caution is already modalizing but without verbal display.
It may be that Caution has seemed to be an option over and above modalizing, for Wright and for other authors, because they have failed to properly distinguish the verbal from the non-verbal, i.e., psychological, aspects of modal judgement.This seems to be actually reflected in Wright's (1989) description of Caution, where he seems to be constantly oscillating between talk of modal words and talk of modal concepts.For example: There does not seem to be great difficulty in envisaging a language in all respects like ours save that it is free of explicit modal idioms.It would hardly be an inconvenience to speak such a language … But now suppose that it becomes clear to [CM] that we are using a concept, to which he is not party, to classify certain judgements … and the general gist of the classification is that these are judgements which, if we have made no mistake, have to be true in all thinkable circumstances.(Wright 1989, p. 211 -the underlining is mine.)This failure of distinguishing properly the verbal dimension of modal judgement, i.e., (lack of) use of modal words from its non-verbal or psychological dimension, i.e., (lack of) use of modal concepts or (lack of) modal belief, is what seems to lead Wright to fail to distinguish between the verbal behaviour of CM and the internal conditions for modal belief that CM already instantiates.
However, Caution cannot be dialectically effective if it is merely verbal, for then it will fail to satisfy the previously defined constraint of non-generalization (NG) to other discourses where anti-realism is simply not an available option.For, merely verbal Caution, i.e., lack of use of the relevant words, while having the relevant beliefs, seems to be indisputably available concerning any discourse.Take as an example a statement about the material world: that the cat is on the mat.You can be verbally Cautious concerning it, when you satisfy conditions X and Y that are cognitively sufficient for believing that the cat is on the mat, and so believe that the cat is on the mat, but you refuse to say that the cat is on the mat.You just say that you only acknowledge that you satisfy X and Y but you do not want to say "the cat is on the mat".But this generalization of Caution will effectively trivialize the appeal to this attitude and, hence, the anti-realist argument for the non-objectivity of modal judgement mounted upon it.For the appeal to Caution was designed to show, precisely, that modal judgement compares unfavourably with respect to other kinds of judgement in terms of objectivity.
To be sure, someone who wished to appeal to Caution would want to deny that Caution is merely verbal, that is, deny that in virtue of involving (i), belief that P, and (ii), inconceivability of not-P, Caution already involves belief that it is necessary that P.But it is important to remark that the present objection does not depend, entirely, on (i) and (ii) being, in fact, sufficient conditions for acquiring belief in necessity.
The challenge presented here is quite general.A defender of Caution, in order to respond to the challenge, would have to identify a further plausible necessary condition for belief that it is necessary that P that CM somehow lacks, so that we can see that he does not really believe in necessity.But, the prospects of meeting this challenge are poor.The proposed additional necessary condition for belief in necessity would have to be a condition which could plausibly be attributed to us, ordinary modalizers, which did not trivially amount to belief in necessity, and which CM could plausibly lack.Thus, a proponent of Caution cannot simply appeal to any controversial condition, such as, for example, a state of "rational intuition that it is necessary that P".But, if the proponent of Caution could pinpoint a plausible additional necessary condition for belief in necessity that CM somehow lacked, then there would be the imminent risk of making Caution, thereby, too weak (in the sense explained above).For, then, CM may be too easily distinguishable from the ordinary modalizer and, hence, the possibility of faultless modal disagreement would vanish.This suggests that a proponent of Caution may be trapped in a fatal dilemma: either Caution is too weak, and so useless to establish faultless modal disagreement, or it is too strong, and so merely verbal.
However, I take it that any non-eliminativist modal theorist should welcome this challenge to Caution, for, if Caution is merely verbal, that will effectively dispel the threat of dispensability that this attitude seemed to bring.For, Caution should not be considered, then, as a coherent refusal to modalize, but, rather, just as a coherent refusal to verbalize modal beliefs.Its availability, therefore, does nothing to show that modal judgement is a dispensable feature of our thought.

Conclusion
In this paper I have examined and argued against Wright's antirealism about necessity based on Caution.I argued, first, that Wright's (1989) later worry about the coherence of Caution is unsound.But, later, I argued that, even if Caution is coherent, it cannot provide an adequate route to non-eliminativist anti-realism.I presented two main objections.The first one is that, if Caution were genuine and coherent, it would show not just that modal judgement is non-objective but that it is dispensable.I think that this is a crucial objection, for it gives reason to think that Caution can only constitute a threat against a noneliminativist theory of modality, rather than provide an adequate basis for it.The second objection is that Caution has a mere verbal character and, on account of this, it is not genuinely possible.