The Nomologicity Solution
Davidson holds the view that T-sentences that are derived from truth theory are offered as warranted empirical generalizations about the speakers of a language L and hence must be not only true but also law like.17 Here, one may ask if T-sentences are law like then can we respect the idea of conventionality of language. Secondly, T-sentences are law like means that they support counterfactuals. What would that mean? It would mean that if we have a T-sentence like "Susama is a singer" is true in English iff Susama is a singer, then necessarily in all the possible worlds where "Susama is a singer" is true are the worlds where Susama is a singer. But is there any reason to believe this? There may be a possible world where "Susama is a singer" as a T-sentence is true but in actuality it's meaning differs from as we understand in English language. It simply means in that possible world the T-sentence "Susama is a singer" means 'Cow is white'.
One answer is that the T-sentence in question which says "Susama is a singer" means that Susama is a singer in English, and if there is a language where "Susama is a singer" does not mean Susama is a singer then that just is not English. But this makes T-sentence not nomologically necessary but only conceptually necessary and also it makes the use of the expression in English in the T-sentence question begging.
| 17 | Fodor, J. & Lepore, E. (1992). Holism: A Shopper's Guide. Oxford: Blackwell , p.84. |