Translation and the paradox of analysis: a reflection on Wiredu's notion of tongue dependency

Abstract

Wiredu argues intriguingly that some philosophical questions only arise in certain linguistic settings. So philosophical questions are, on occasion, linguistically relative or, more vividly, Tongue Dependent. The phenomenon however does not rest on expressive differences between languages, or, better, on failures of translation. Though rejecting his example, I endorse the general possibility he constructs. I do so provided that there is a solution to the Paradox of Analysis. Indeed I point out that the possibility of Tongue Dependency is both necessary and sufficient for a solution to the Paradox of Analysis. I then offer a solution to the Paradox of Analysis in a Fregean setting, by distinguishing within the cognitive value of a claim between cognitive accomplishment and cognitive significance. The results of analysis involve cognitive accomplishment, though have no cognitive significance. I recommend my account as dealing well with a problem Salmon raised for the Church-Frege account. Finally, I direct my account on vocabularies rather than claims and note that, though possible, actual cases of Tongue Dependency within natural languages are highly doubtful. The phenomenon does not have the cultural significance Wiredu claims for it.

KEYWORDS:

• Translation
• analysis
• paradox of analysis
• Wiredu
• sense
• meaning
• truth

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 See his 2004.

2 The following section closely replicates the discussion in my 2019, 235–237.

3 Suitably, but not unproblematically, generalised.

4 Note that the claim is not that the problem will not occur to Akan speakers, though it will occur to English speakers. The claim is that there is a problem in English which fails to exist in Akan. So the observation is about the language's themselves; not the psychology of their speakers.

5 Compare with Hume's principle: The number of Fs is the same as the number of Gs iff the Fs can be placed in one-one correspondence with the Gs. Here the LHS and the RHS are often held to be synonymous. See, for instance, Wright (1983), where it is called N⁼, and the essays in Hale and Wright (2001), especially section 3.

6 My thanks to Richmond Kwesi.

7 I’ll interrogate assumptions about translation involved in this claim soon. My aim here is just to present what I take to be the general form of Wiredu's claim.

8 Imagine that both my twin and I employ the words ‘water’ and ‘twater’. I live in an environment containing only water; my twin, one in which there is both water and twin water. My terms are each introduced by reference to samples of water; my twin's term ‘water’ is introduced by reference to a sample of water and his term ‘twater’ by reference to a sample of twin water. Then: my terms ‘water’ and ‘twater’ are synonymous; but my twin's terms are not; the relation of synonymy is not transparent. (cf. Putnam, ‘The Meaning of “Meaning”’, essay 12 in his 1975)

9 Admittedly, this isn't wildly exciting.

10 See his 1892) 1952, 56.

11 Perhaps one might say, echoing Wittgenstein on logical tautologies(see his 1921, 6.12 and, for a suggestion about analysis, 6.121), that the analytic equivalence says nothing but, in virtue of doing so, shows (at least potentially) a relation between concepts. So what can be known by trivial means is itself trivial; but what is not trivial to know, may itself be trivial.

12 For example: where a, b and c are ‘the lines connecting the vertices of a triangle with the midpoints of the opposite sides’, the claim: ‘the point of intersection of a and b is the same as the point of intersection of b and c’. See Frege (1892) 1952, 57).

13 Please feel free to substitute any pair of sentences where you are happy to take the one to be the analysandum of the other. Salmon's pair is: ‘Holmes has an older brother’ and ‘Holmes has an older male sibling’.

14 See his 1993, 161 where he lays out the inconsistent triad.

15 Or: Tr([[P]]≠[[Q]])=[[M]]≠[[M]]; so a truth is translated by into a falsity.

16 Though he advertises his option as one of rejecting (3), I’m not clear that this is the best reading of it; rather he offers an account of the logic of an expression such as ‘the proposition that P’ which shows that ‘P’ is not a semantic constituent it. So the stricture on interpretation is not relevant.

17 Note that, though this retains the Fregean solution to the informativeness of the identity claim (and so the Frege-Church response to the Paradox of Analysis), the nature of premise (2) changes. We are not now concerned with a difference in indirect senses of ‘P’ and ‘Q’ but of the senses of the demonstrative propositional names formed from ‘P’ and ‘Q’.

18 See his 1995, 156–157.