威廉森論「宣稱」

行政院國家科學委員會專題研究計畫 成果報告

威廉森論「宣稱」

中 華 民 國 93 年 11 月 3 日

WILLIAMSON ON ASSERTION

Chin-mu Yang

威廉森(1995)論證，任何企圖將模態邏輯中的S4公設應用到一個素樸的「宣稱」概念上， 都將無法令人滿意地說明在(A4) Ap→ Aap中的「宣稱」概念。尤其是，就知識的原初性 與重要性而言，(A4)會造成一個無法接受的結果，就是有名的KK原則：Kp→ KKp，這 不是一個合理的知識概念所能允許的原則。威廉森(1996, 2000)進一步指出，宣稱的知

態規則就是如果主體S宣稱語句p，則S知道p，即(AK) Ap→Kp。 因此，如果我們接受(AK)

而拒絕(KK)，那我們就必須拒絕(A4)。而威廉森認為這是可以反駁反實在論之意義理論 的論證。 我首先論證，就威廉森對宣稱所做的知態說明，宣稱是一個強的規範性概念，(AKA) Ap→ Kap，也就是當一個主體宣稱p，主體不僅知道p，也知道他正在宣稱p。如同威廉 森(2000)指出，在知道與宣稱之間有一規範性關係，做宣稱時必須基於某些宣稱規則， 而這些規則是由知道概念來說明的。尤其是，宣稱是一個語言行為，是傳遞知識的載體： 我們藉由宣稱表達並傳遞我們的知識。(2000:238)這也說明了知道與宣稱間的強規範關 係。因為我們有某些型態的行為具有某些規範，即只要表現該行為，就知道自己在做什 麼。宣稱就是這樣型態的行為。因此，當‘Ap’成立時，不會有任何的理性主體會不知道 自己在宣稱p。這就證成了(AKA)。同時，(AKA)不會蘊含KK原則，雖然會蘊含弱的KK 原則，即(AKK) Ap→ (Kp→KKp)，但這並無損害。 甚者，因為宣稱的主要目的是傳遞知識──「我們需要宣稱來傳遞知識」──也因為 宣稱是語言行為，當主體宣稱P，其所傳遞的知識不僅是宣稱的內容，還包括宣稱這個 行為。因此我們可以說能夠成為被傳遞的知識一定是透過宣稱才行。因此，如果‘Ap’ 就是被傳遞的知識，‘AAp’就成立。我相信這樣對宣稱的說明可以說明(A4)中的‘A’。事 實上，我們可以證出，(AKA) 和 (AKK) 都是由( A4) 和 (AK)導出。因此，我們可以 給宣稱醫滿意的書名而不致導出錯誤的結果。

英文摘要

Williaimson (1995) argues that any attempt to apply the S4 axiom of normal modal logic to a naïve notion of assertibility would fail to find a satisfactory account of the notion of assertibility to serve as the intended interpretation of ‘A’ in the desired characteristic formula (A4) Ap→ AAp. In particular, granted the primitiveness and primacy of knowledge, (A4) would render some undesirable consequences, notably the KK principle: Kp→ KKp, which could hardly hold in any reasonable notion of knowing. Williamson (1996, 2000) further proposes the knowledge rule for assertion, a constitutive rule in character, according to which an agent S asserts that p only if S knows that p, in symbols, (AK) Ap→Kp. It follows that if

we accept (AK) and reject (KK), then we would be forced to reject (A4). And Williamson takes this as an argument against anti-realist’s programme for a theory of meaning.

*

This report was presented at, as a contributed paper to, The 12th International Congress of Logic, Methodology and Philosophy of Science (LMPS03) August 7-13, 2003 Oviedo (Spain),

Section C.5 Philosophy of Linguistics. A full version has been completed but some refinement is ongoing.

2

In this talk, I shall firstly argue, based on Williamson’s knowledge account of assertion, that assertion is a strong normative concept in that (AKA) Ap→ KAp holds, that is, when an agent S makes an assertion that p, S would not merely know that p, but also know that she/he is making an assertion that p. As Williamson (2000) rightly points out, there is a special normative relationship between knowing and asserting, in that the speech act of assertion must be based on certain rules of assertion, which are essentially constitutive and should be characterized in terms of the notion of knowing. In particular, assertion is a kind of speech act, which is to be taken as the vehicle for transmitting knowledge: ‘We express and communicate our knowledge by making assertions.’ (2000:238) This also suggests a strong normative relation between knowing and asserting as (AKA) shows. For we have some type of acts with a certain norm in that whenever one performs an act of this type, one knows clearly what one is doing. And strikingly, assertion, taken as overt speech act, belongs to such a type. Accordingly, when ‘Ap’ holds, it would never happen to any rational agent that she/he does not know that she/he is asserting that p. This would justify (AKA). Meanwhile, it can be argued that (AKA) would not imply the KK principle, though it will imply a weak version of the KK principle, i.e., (AKK) Ap→ (Kp→KKp), which is harmless.

Moreover, since the primary purpose of asserting is to communicate knowledge－‘We need assertion to transmit knowledge’ (2000:267; 1996:520), and since assertion is essentially a kind of overt speech act, it can be further argued that when an agent S asserts that p, not only the content of what she/he asserts (i.e., ‘p’) but also the making of this assertion (i.e.,‘Ap’) become transmitted knowledge. Now it would perfectly make sense to claim that to be a transmitted knowledge only if it is asserted. Therefore, if ‘Ap’ is understood as a transmitted knowledge, then ‘AAp’ holds. I believe that this account of assertibility would provide a sensible interpretation for ‘A’ in (A4). In fact, it can be shown that both (AKA) and (AKK) can be derived from (A4) and (AK). We then have a satisfactory account of assertibility without rendering fallacious consequence.

計畫緣由與目的

Williaimson (1995) argues that Dummett’s objections to truth-conditional theories of meaning also apply to any assertibility-conditional theory unless the following schema holds:

(A4) Ap→ AAp, (where ‘A’ intends to stand for the concept of ‘assertibility’)

which is essentially an application of the well-known S4 axiom in normal modal logic, namely,

(4) Mp→ MMp, (where M is a modal operator) to the notion of assertibility, a typical cognitive attitude.

But, he further points out that (A4) would fail to hold for any reasonable notion of assertibility. In particular, granted the primitiveness and primacy of knowledge, (A4) would imply the KK princple:

which could not hold in any reasonable notion of knowing.

Recently, in his book Knowledge and Its Limits Williamson further proposes a knowledge account of assertion:

(AK) Ap→ Kp

It is striking that if we accept (AK), and reject (KK), then it seems that we would be forced to reject (A4) as well.

The main objective of this research is to provide an acceptable notion of assertion to buttress (A4). But before we go into a detailed examination of Williamson’s knowledge account of assertion, a brief review of his program is called for.

結果與討論

The objective of this research is to provide a philosophical notion of assertion to show that (A4) can be reasonably accepted. In the meantime, I shall offer an argument for a weak version of the KK principle, namely,

(AKK) Ap→ (Kp→ KKp).

which states that if some agent S asserts that p, then if S knows that p then S knows that S knows that p. I shall then argue that the acceptance of (A4) will imply (AKK), rather than (KK), hence should be harmless.

In ordinary discourse, the primary purpose of making an assertion for some agent is to let the hearer (her/his audience) know of the content of one’s assertion. That is to say, asserting is a kind of linguistic performance, which intends to provide one’s audience with information through recognition that this is one’s intention. Clearly, this intention includes the intention that the audience should know that the speaker himself knows that he is asserting. This indicates that

(AKA) Ap→ KAp.

It is important to bear in mind that when ‘Ap’ holds, that is to say, when ‘p’ is asserted, it would become a communicated knowledge. Therefore, the knowledge rule (KA) not merely says that the speaker knows that p, but also indicates that the hearer(s) would know that the speaker knows that p. Accordingly, whenever Ap holds for some p, the (KK) principle would hold for p. The reason is that ‘Ap’ is objective in that it is a public linguistic practice, so it would never happen to any rational agent that she/he asserts that p without knowing that p. This implies (AKA). But then the knowledge rule, together with the distribution of knowledge to conjuncts, indicates that Ap→ KKp.

By contrast, it can be argued that based on the assumption that (A4) holds, we have a weak version of (KK) principle. The argument may run as follows:

Assume

4

Also, we take (KA) for granted, that is,

(KA) Ap→ Kp

Then, applying (KA) to ‘Ap’, we have

(1) AAp→ KAp

From (1) and (A4), by the transitivity of implication, it follows

(2) Ap→ KAp

Now, applying the distribution of conjunction of knowledge to its conjunts, from (KA), we have

(3) KAp→ KKp

And from (2), (3), it follows that (4) Ap→ KKp

Accordingly, from (4) we would have (5) Ap & Kp→ KKp

which is equivalenet to

(AKK) Ap→ (Kp→ KKp)

I shall take this as a weak version of the KK principle, which should holds for all p that it is asserted.

A satisfactory account of the notion of assertion will be proposed, which should be able to buttress (A4) without having any problematic consequence. This would provide a more solid setting for anti-realist theory of meaning.

參考文獻Reference

Bar-Elli, G. (1996), The Sense of Reference: Intentionality in Frege, Berlin: Water de Gruyter.

Brandom, R. B. (1983), ‘Asserting’, Noûs 17: 637-50.

Brueckner, A., and Fiocco, M. O.(2002), ‘Williamson’s anti-luminosity argument’,

Philosophical Studies 110: 285-293.

Carl, W. (1994), Frege’s Theory of Sense and Reference: Its Origins and Scope, Cambridge: Cambridge University Press.

Cozzo, C. (1994), ‘What can we learn from the paradox of knowability’, Topoi 13: 71-8. Dummett, M.A.E. (1978), Truth and Other Enigmas, London: Duckworth.

---, (1981), Frege: Philosophy of Language, 2nd edition, London: Duckworth. ---, (1991), The Logical Basis of Metaphysics, London: Duckworth.

---, (1993), The Seas of Language, Oxford: Oxford University Press. Dudman, V. (1992), ‘Probability and assertion’, Analysis 52: 204-11. Edgington, D., (1985), ‘The paradox of knowability’, Mind (94): 557-68. Frege, G. (1979), Posthumous Writings, Oxford: Basil Blackwell.

Geach, P. (1972), ‘Assertion’, in Logical Matters, Oxford: Basil Blackwell. Gettier, E. (1963), ‘Is justified true belief knowledge?’ Analysis, 23: 121-3. Green, K. (2001), Dummett: Philosophy of Language, Cambridge: Polity.

Gunson, D. (1998), Michael Dummett and the Theory of Meaning, Aldershot: Ashgate. Jones, O.R. (1991), ‘Moore’s paradox, assertion and knowledge’, Analysis, 51: 183-6.

McGuinness, B. and Oliverti, G., eds. (1994), The Philosophy of Michael Dummett, Dordrecht: Kluwer.

Picardi, E. (1994), ‘Convention and assertion’, in The Philosophy of Michael Dummett, B. McGuinness and G. Oliverti (eds.), Dordrecht: Kluwer, 1994, 59-78.

Price, H. (1987), ‘Truth and nature of assertion’, Mind (96): 202-220.

---, (1998), ‘Three norms of assertibility, or how the MOA become extinct’, in

Philosophical Perspectives 12, Language, Mind, and Ontology, 1998, edited by J. E.

Tomberlin, Oxford: Blackwell, 241-254.

Stalnaker, R. C. (1984), Inquiry, Cambridge, Mass.: MIT Press. ---, (1999), Context and Content, Oxford: Oxford University Press. Williamson, T. (1982), ‘Intuitionism disproved?’ Analysis 42: 203-7.

---, (1990) ‘Two incomplete anti-realist modal epistemic logics’, Journal of Symbolic Logic, 55:297-314.

---, (1995), ‘Does assertibility satisfy the S4 axiom?’ Critica 27: 3-22. ---, (1996), ‘Knowing and asserting’, Philosophical Review 105, 489-523. --- (2000), Knowledge and Its Limits, Oxford: Oxford University Press.

Wright, Crispin (1992), Truth and Objectivity, Cambridge, Mass.: Harvard University Press. --- (1997), ‘The indeterminacy of translation’, in A Companion to the Philosophy of