ࡱ> gifb Vjbjb{x{x $r2Oi $ 8@4t,L" , R?#:!:[" tvJ0tJ :,   q0R##  ddWILLIAMSONS ORDINARISM vs. INFERENTIALISM In his book Philosophy of Philosophy, in chapters 3 and 4, T. Williamson argues against philosophical position he calls inferentialism. In those chapters titled Metaphysical conception of Analyticity and Epistemological Conception of Analyticity Williamson sets fort and criticize the doctrine that analytical propositions has special status in philosophy arguing (in chapter 3) that their truth conditions are basically the same as in the case of synthetic propositions and (in chapter 4) that their grasping or understanding conditions are not special as well. I will call Williamsons position ordinarizm, referring to the claim that there is no special status for analytic propositions. In this paper I will focus mostly on views bestowed in chapter 4. Among philosophers that advocate inferentialism are prominent names as Boghossian, Peacocke, Jackson and Brandom. In his criticism of inferentialism Williamson particularly addresses Boghossian and his position. Lets start with a characterization of inferentialism. The characterization fitting well with Williamsons as well as with Boghossians view could be that Inferentialism is a doctrine claiming that grasping the concept or understanding the word is determined by willingness to infer according to the inference patterns characteristic for the concept or word in question. In a somewhat more formal way, let C be the concept, let C be the word and P the inference pattern of which C is a constitutive part. One who is willing to infer in the appropriate way, satisfies possession conditions for the concept or term. In contrast, one who is not willing to infer according to the right pattern does not understand the word (does not manage to grasp the concept). This schema could be applied to different kinds of concepts. In the case of logical constants, one satisfies the possession conditions of the connective end if one is willing to infer according to (or, assent to) the introduction and elimination rule for the conjunction. The same holds for other logical constants. In the case of analytic sentences, one understands the term vixen (grasp the concept vixen) if one is willing to infer according to the inference pattern If x is Vixen Then x is a female fox. Accordingly, one understands the word vixen if one assents to the sentence Vixen is a female fox. Employ T. Williamson attacks inferentialism claiming that its basic and clearest statement does not hold. This target statement is: Necessarily, whoever understands the sentence S, assents to it. Williamsons claim is: Understanding the sentence (grasping the thought) is independent of assent to it. Taking the analytic sentence as an example we have: Necessarily, whoever understands the sentence Every vixen is a female fox, assents to it. T. Williamson formulates it in a most basic way, as an elementary logical truth: Every vixen is a vixen And thus we have Necessarily, whoever understands the sentence Every vixen is a vixen, assents to it. To put forward the claim he offers examples that compellingly demonstrate the possibility of a deviant speaker who denies to assent to the inference pattern P (or the sentence containing the concept/word) and nevertheless understands the word (or grasps the concept). Among the prominent examples, two of them directly refer to the above formulation and are depicted as imaginary cases of Peter and Stephan. I am concerned here with Peters example only. Two further examples have to do with logical principles (connectives). One of them mentions the logician Van McGee who, finding some instances of modus ponens invalid, dissents from modus ponens rule, while the other example takes his imaginary student who, although unreflectively, dissents from modus ponens naively trusting his teacher. I will tackle the two latter examples in the due course of this paper. Williamsons examples being so meticulously designed are extremely hard to object to. Peter example consists of two elements. The first one is Peters meta-linguistic theory and the second is his actual belief about vixens. Regarding the first one, Peter holds that 1. Every vixen is a vixen logically entails: 2. There is at least one vixen. Hence, the truth of There is at least one F is a necessary condition for the truth of Every F is a G, and the falsity of There is at least one F is a sufficient condition for the falsity of Every F is a G. In this way, he takes universal quantification to be existentially committing PoP, 86. Peter also has a wired belief that (2) is false, due to his enchantment with conspiracy theories. Accordingly, he does not assent to Every vixen is a vixen. The argument is that, nevertheless, he understands the words (grasps the concept) in the sentence he dissents from. To perceive the frame of the discussion in a somewhat broader way, let me remind of an earlier dispute between Williamson and Boghossian. In Blind Reasoning Williamson characterises the crux of the inferentialist position as follows: They (inferentialist) seek a conceptual shibboleth, an inference or principle acceptance of which is necessary and sufficient for knowing what a word means or possessing a given concept. 276 Contrary to this, Williamson claims: Whatever local test is proposed, someone could fail it and still do well enough elsewhere with the word to be counted as knowing what it means. 276. In PoP TW expressed the same idea claiming that Peter and Stephan understand the sentence Every vixen is a vixen without assenting to it and would be able to fit in well enough with the rest of the linguistic community, to engage smoothly in useful communication and adjust to their differences with other speakers in order not to attract to much attention. 99 But, such a characterisation of the equality of understanding the word by deviant and by ordinary speakers amount to a clearly pragmatic requirement. I agree that a deviant speaker can successfully communicate despite his differences in opinion, but nevertheless, I do not believe that referring to the pragmatic dimension exhausts the issue. Therefore, Ill try to argue that a) deviant thinkers, though fitting well enough with the rest of the linguistic community, have different concepts from other community member, and that b) necessary condition for thinkers understanding the word or grasping the concept is tacit knowledge of basic inference rules founded in sub personal mechanisms. I am not claiming that tacit knowledge of basic inference rules is sufficient for understanding. The claim that dissents from a sentence s, when s is an analytic sentence, implies possession of a different concept, is to be discussed first. TW argues that Peter, dissenting to Every vixen is a vixen, and an ordinary speaker, assenting to the same sentence, understand the word vixen equally well. Nevertheless, Williamson admits that there might be some point marking out the difference in content. He locates this point in understanding of the quantifier every and in next passage offers the solution to the problem dissolving the differences in the pragmatic context of the linguistic community. The best candidate for the word or mode of composition of (1) that Peter and Stephan misunderstand is every. Peters not uncommon conception of the existential commitments of universal quantification makes little difference in practice, for when sentences of the form Every G is F occurs in conversation, There is at least one F tends to be a common ground among the participants anyway. 88. But, although Peter and an ordinary speaker can reach the common pragmatic and semantic ground, they might have (and I am arguing that they actually have) different concepts. Williamson of course envisages such a move and in BR offers an account arguing that such a position is not very promising. It is hardly possible to present the position better than he did: The proposal might be that different thinkers can understand the same unambiguous word in virtue of using it to express different concepts, which are equated with different inferential roles. For example, neo-Fregeans such as Gareth Evans (1982: 40) sometimes claim that different speakers can achieve linguistic competence with the same proper name by associating it with different concepts (modes of presentation) of the same object. The distinctions between inferential roles would thus cut finer than the distinctions between linguistic meanings. On this view, when our unorthodox thinker refuses to infer any longer according to P, she associates the word C with a new concept, because she changes the associated inferential role, while retaining the same meaning for the same word. BR, 271. Tims answer to this manoeuvre is as follows Since there are likely to be at least minor differences between any two speakers in their willingness to accept inferences that involve the word, some principle is needed to distinguish those patterns of inference that are essential to the concept from those that are not. The meaning of the word supplies no such principle, on the envisaged view of the concept. Without such a principle, the discussion of concepts becomes dangerously unconstrained. BR 272 TW is here absolutely right, and what is said is perfectly applicable to the situations where different (possibly endless) modes of presentation of the unique referent are available. The paradigmatic situation might be that where one speaker grasps the meaning of word Venus under the mode of presentation the shiniest star on the morning sky while the other grasps it under the mode of presentation the shiniest star in the evening sky. In this case, there are two different inferential roles none of which is essential to the concept Venus, so that it seems that the meaning of the word does not supply the principle in question. However, the two roles yield the sufficient contrast of senses. Nevertheless, I am inclined to say, this is even clearer in the case of analytic sentences and basic logical truths which Williamsons examples are designed to refer to. In the cases of analytic sentences and basic logical truths, unlike the situations in which any two speakers might disagree in any minor differences, it is supposed that whoever understands sentence s associates with it the same inferential pattern P and assent to it. My claim is that Evans description applied to such situations shows that confronting thinkers do have different concepts and that the word they apparently understand in the same way can supply the principles essential for each concept. As Williamson says, where Peter and an ordinary speaker disagree is the universal quantification. Peters meta-semantic theory commits him to claim that only concepts with non-empty extension can figure in the sentence (thought), which is to be accepted. He also believes that vixens do not exist. Since he takes universal quantification to be existentially committed he coherently dissents from the sentence every vixen is a vixen. If it is a case, the very existence of the instance of the concept plays a role in Peters grasping the concept. In his meta-semantic theory existence as a predicate clearly cuts the concepts to those with extension and to those with empty extension. Thus, grasping the concept one believes to have non-empty extension is quite a different thing than grasping the concept one believes to have empty extension. Even if Peter believes that vixens exist, his meta-semantic theory commits him to associate his vixen concept with different inferential role than it is the case with ordinary speaker having a different semantic theory. Depicting the story in Fregean terminology, Peter, understanding the word vixen, does not assent to the sentence Every vixen is a vixen, but certainly assents to the sentence If there are vixens, then every vixen is a vixen. This different sentence, as a different mode of presentation, provides him with a vixen concept different from that the ordinary speaker has. The principle that bestows essential inferential roles to each of the concepts is existence as a predicate. Lets see what happens if Peter, holding, of course, his preferred meta-semantic theory, does not believe that vixens exists. Let examine Peters meta-linguistic theory a bit more closely. The model for Peter (regarding his meta-semantic theory) might be Gareth Evans interpretation of Freges position before Frege made distinction between sense and Meaning. According to this interpretation Frege was more Russellian then Fregean, embracing the view that someone who utters a sentence containing an empty singular term would fail to say anything, in a sense that he would fail to express a thought (G E, 12). One might say the vixen case is not, strictly speaking, a singular term, but it is analogous with this kind of term playing a role of a quasi-singular term. Peter, as we are told, dessents from the elementary logical truth such as Every vixen is a vixen because he does not believe that vixens exists. Even more, he declines to accept any sentence containing empty quasi-singular term vixen. If he did not, a Fregean from Evans interpretation would say that Peter fails to express the thought. But, whatever interpretation might be the case, given that Peter dissents from any possible sentence containing the word vixen, how could one say that Peter has a vixen concept after all? At the end of this part, I would express a final worry. What makes the difference between Peter and an ordinary speaker, apart from their meta-semantics, is the fact that Peter believes that vixens do not exist. What matters here is Peters belief about existence, not actual existence. His beliefs about existence plus his meta-semantics make him to dissent from the elementary logical truth Every vixen is a vixen or from the analytic sentence Every vixen is a female fox. But, those sentences for Peter are not analytic any more, because, tacitly involving the existence assertion, they presuppose the empirical question, namely, whether vixens exist or not. After all, this turns out to be equivalent to the conjunction of the sentences: Vixens are vixens and Vixens exist. Since the second conjunct is empirical, the whole conjunction is. In this way, possible change in propositional attitudes (beliefs about vixens) change the status of the sentence, shifting it from analytic to empirical and vice versa. But, my interpretation might be completely on the wrong track. Maybe discussion is not about understanding of the word vixen(grasping the concept vixen) in the sentence every vixen is a vixen but about the understanding or grasping of the copula is, instead. If this is the case, the difference between an ordinary speaker and Peter is much clearer. For the firs one, copula expresses identity, while for Peter it expresses a mixture of identity and existence. The object of understanding the word or grasping the concept might be either vixen or is. If the discussion is about grasping the concept vixen, then concepts may be individuated in two ways. According to one way, concepts may be strictly or timely individuated. On the other way, concepts may be broadly individuated. Trilema My second point has to do with other two Williamsons examples. It is McGees example and the example of his non-reflective student that has to be objected to on somewhat different grounds than the previous one Lets expose the example of Van McGee, a distinguished logician. This example convincingly shows the possibility of perfect understanding of the inference rule of modus ponens and dissenting from some of its instances. As we said before, inferentialists claim that whoever understands the connective if assents to constitutive inference rules for the conditional, modus ponens and conditional proof. To assent to those rules of inference is a necessary condition for understanding if. McGee has published papers containing counterexamples to modus ponens and consequently dissents from those instances of modus ponens that are proved, or seem to be proved as invalid. Does it mean that McGee does not understand the word if? Of course not, because McGee is an expert logician who understands the word if and the constitutive inferential rules for the conditional, if anyone does. Before I scrutinize the example and its consequences, let me pose some preliminary remarks. Williamson admits that referentialists can put our formulation: Necessarily, whoever understands the sentence S, assents to it in a somewhat weaker form (which actually is one of the Boghossians formulations): Necessarily, whoever understands the sentence S, has a disposition to assent to it. Whatever dispositions might be, there are two ways in which they can be articulated. In can be at the personal level or the sub-personal level. I am interested here in sub-personal dispositions only. As Williamson says, at the sub-personal level, the postulated dispositions are grounded in something like an unconscious reasoning module 100. The module must include rules for deduction, since that is the kind of reasoning relevant to (1). A prima facie attractive conjecture is that the deductive rules would include analogues for natural language connectives of the introduction and elimination rules in a Gentzen-style system of natural deduction. According to the dispositional account, even in the Peter and Stephans cases, the default outcome of the underlining competence is to assent to (1) even if stable dispositions from others sources irreversibly override that default. In the example illustrated by McGees case, sub-personal dispositions are even more at work. While Peter on reflection dissents from Every vixen is a vixen, McGee dissents from only a small proportion of the instances of modus ponens. In the rest of the area he will use the rule of modus ponens. If, for instance, he would for some reason decide to convince me, the ignorant, that some instances of modus ponens are invalid he would most probably in his proving procedure keep use of modus ponens. Therefore, he has a default natural disposition, as anyone else, to use the rule, although this default disposition can be partly overridden by his reflection on it. If this would be the end of the story, the disposition account would be vindicated. But, it is not. The discussion going on examining whether psychological and cognitive theories at hand can offer any evidence that the data of human normal performance can support the hypotheses of a set of deductive rules unconsciously employed Ws diagnosis is rather pessimistic: There is a little sign of anything modular that contains formal rules to sub-serve conscious deduction, whether conceived as part of a language module or as part of reasoning module. 104 However, the answer one can give ob this issue significantly depends on the kind of the theory one relies on. The recent theoretical scene is divided into those theories that explore the experimental data that document human cognitive errors and irrationality and those that try to reinterpret those data or offer some of the opposite kind. The interpretation of data, as Stanovich and West emphases, has been the subject of contentious debate for almost two decades (S&W, 645). Among the authors that recently have offered theories that strongly support some variant of the reasoning module thesis I will mention Braine & Obrien and Lens Rips. On the other hand, celebrated Wason selection task that has been usually considered as the evidence for human poor deductive performance is recently challenged by the number of authors. I will mention here only Sperber and Girottos article from 2002. They convincingly argue that the value of the selection task as a tool for studying human inference has been grossly overestimated., 277. Although I agree that the recent state of the debate cannot decisively confirm the optimistic stance supporting evidence for the reasoning module, it is also true that it cannot prove the pessimistic thesis.  I am owing the term ordinarizam from Nenad Mia evi who employed it in a series of reacent articals. PAGE  PAGE 1 '/3@ !9m.ymmyh0chImH nH uhImH nH uh lhI>*mH nH uhthI6mH nH uh lhI6mH nH uh*hI6mH nH uhthImH nH u$jh lhI0JUmH nH uhhImH nH uhhI6mH nH uh lhImH nH u*+,( 4]d"9m[`gdIgdI2UVV[>[Cu E"#$D%E%f(( & 0` P@1$7$8$H$gdI`gdIgdI./2CDXY !ln    !!!!E"i#j###$$E%F%%%?&@&&&&&O'P'''((޿޿ޯ~~~~~~~h lhI^J_H mH nH u'h lhIOJQJ^J_H mH nH uhI^J_H mH nH uh lhI^J_H mH nH uh lhI6mH nH u h lhIB*mH nH phuh lhImH nH uh0chImH nH uhImH nH u1((())p)q)**a*b*++++,,, -:-<-......Q1R1110212223333442575T5X55567ﾊ~h lhImH nH u#hbhI^J_H mH nH sH uhhI^J_H mH nH u"h lhI6^J_H mH nH u'h lhIOJQJ^J_H mH nH uhI^J_H mH nH uh lhI^J_H mH nH uh lhI^J_H mH nH u0(a*-//3568::2;>@@.ABBBBBCCDpFF`gdI & 0` P@1$7$8$H$gdI778888889 9P9R9::::=;A;;;;a<j<<<W=[===??AAAAABBkBmBzB~BBBBBBBBBB&C̛̛̛̭̿̿̿̿h?EhImH nH uhXMVhImH nH u"h lhI6^J_H mH nH u"h*hI6^J_H mH nH uhI^J_H mH nH uh lhI^J_H mH nH uh lhI6mH nH uh lhImH nH uhImH nH u3&C'CTCUCDDQDUDDDDDDDDDE%E>E?EEEEENFRFFFFFFGNGOGQGSGjGoGuGwGGGGGHHIJJJ$KKKKK&L'LLLMMNNNܾh*hImH nH u h lhIB*mH nH phu#h*hI6B*mH nH phuhXMVhImH nH uh*hI6mH nH uh lhImH nH uhImH nH u?FFGOGGGRIIJyKpMN{NcOOFPR[T\T,U-U.U/U0U1U`gdIgdI & 0` P@1$7$8$H$gdINNP PPPPPQQdQeQhSrSsS2U3UVVVVVVVVVVVVVVÿ hI0JjhI0JUhIjhI0JU hGshIB*mH nH phuhhImH nH uh lhImH nH uhImH nH u1U2UVVVVVVVVVV &`#$gdI & 0` P@1$7$8$H$gdI ":pI|. A!"#$%H@HNormalCJ_HaJmHnHsH tH uDA@D Default Paragraph FontRiR  Table Normal4 l4a (k(No List4 @4.yFooter  !.)@. .y Page Number6@6 Footnote Text@&@!@ Footnote ReferenceH*OhkhhhhhhkOr + + + + + + + CW&4A.OOcu9Z+,(4] d " 9  m [>[CuEDEf""a$'))-/0244258::.;<<<<<C=>p@@@AOAAARCCDyEpGH{HcIIFJL[N\N,O-O.O/O0O1O2OOOOOOOOOO0|00ŀ0000000@P0@P000@P0@P00ŀ00ŀ000@P00|000000@P00P000@P0@P00@P0 00ŀ00P00@P00P00000@P0|0@P000p0 0@P0p0@P0 00 0@P00@P0@P0`0@P0@P0@P0@P0@P0@P0000000000@P0000000000000000000000@P0@P0@P0@P0@P0@P0@P0@P0@P0@0@0@00A@0 @0|+,(4] d " 9  m [CuEEf"a$-/024@@AOAAAO00ŀ00ŀ@00@00ŀ@00ŀ@00ŀ@00ŀ@00ŀ@00ŀ@00@00 @0B@0B@0 @0@00ŀ@0 @0 @0M900 ̢M900@0@0@;`c@;`c00ŀ@00ŀ@00ŀ@00ŀ@00ŀ@00ŀ@00ŀ@00@00@0000@00@00@00@00O90+0 `ϢO90+0M90+0@0@;`c@0@00ŀ@00ŀ@00ŀ@00@00ŀ .(7&CNV-12457[(F1UV.0368V/ !! 2OOOf o FM]"e"''11d3q3C=O=OAQALMSNUN2OOO:::::::::::'[Ǣi%-ky^`o(. ^`hH. pLp^p`LhH. @ @ ^@ `hH. ^`hH. L^`LhH. ^`hH. ^`hH. PLP^P`LhH.^`o() ^`hH. pLp^p`LhH. @ @ ^@ `hH. ^`hH. L^`LhH. ^`hH. ^`hH. PLP^P`LhH.%-k'[                  @09LMZOOP @PVUnknownUTimes New RomanSymbol5Symbol3 ArialK Times-RomanTimes;Helvetica 1hKL䦐ئk A! J%4dOW 2QW WILLIAMSON vsNenad SmokroviNenad Smokrovi   Oh+'0! $ D P \ ht|WILLIAMSON vsNenad SmokrovićNormalNenad Smokrović3Microsoft Word 11.5.6@F#@}('@g[@|k AG@ PICT 8Z HHZ ZHHZZ         ]kZwwo{o{wwkZo{so{wswkZwwwkZo{so{wg9wwwo{wwe+ZkZVcRVg9VcRVcRc^ZZwZkZF1g9cg9kZZg9^Z^ZkZR^kZVZg9  ` wwwwo{wwwsswwwswswwkZkZw?@ABCDEFGHIJKMNOPQRSTUVWXYZ[\]_`abcdehRoot Entry F/Ij1Table:#WordDocument$rSummaryInformation(L "DocumentSummaryInformation8^CompObjXObjectPool/I/I FMicrosoft Word DocumentNB6WWord.Document.8