Frank Jackson

The sentence ‘x is square’ might have had different truth conditions from those it in fact has. It might have had no truth conditions at all. Its having truth conditions and its having the ones it has rest on empirical facts about our use of ‘x is square’. What empirical facts? Any answer that goes into detail is inevitably highly controversial, but we think that there is a rough answer that is, by philosophers’ standards, relatively uncontroversial. It goes back to Locke 1689 and beyond, and is best known to contemporary philosophers through the work of Grice 1957 and Lewis 1969. It is that we (usually implicitly) agreed, as a matter of contingent fact, to use ‘x is square’ as a way of conveying our taking it to be the case that x is square.

ONE way t0 defend narrow content is to produce a sentence 0f the form ‘S believes that P’, and show that this sentence is true 0f S if and 0nly if it is true 0f any duplicate from the skin in, any doppclgangcr, of S. N0toriously, this is hard to d0. Twin Earth examples are pervasivc.1 Another way to defend narrow content; is t0 show that Only 2. narrow notion can play thc causal explanatory r01c we require 0f contcnt in 2. properly scicntiicm psychology 0r cognitive science. Notoriously, this is hard t0 d0. The considerations—mcthod010gicaI solipsism, the principle 0f autonomy, 0r what:cvcr—invokcd to show that a broad notion 0f content cannot..