That Ayer’s interpretation is defective because Ayer does not grasp the underlying ‘key tenet’ is shown also by Ayer’s account of the Tractatus where Ayer also does not grasp the importance of the ‘key tenet’ to understanding that work. Crucially, Ayer does not appreciate how the ‘key tenet’ provides an answer to the fundamental question “how a sentence could at one and the same time express a pseudo-proposition and an unassailable truth” [p.113]? Ayer writes “I did not see, and still do not see, how” this could be the case. But, as Ayer fails to appreciate the ‘key tenet’, Ayer fails to consider W’s putative answer that it can be the case because a pseudo-proposition may show “an unassailable truth” though it says nothing with sense. It might be a contradiction of an obviously untenable sort to claim both that ‘what p says is nonsense’ yet ‘what p says is the truth’. But it is not the same sort of obviously untenable contradiction to claim both ‘what p says is nonsense’ yet ‘what p shows is the truth’. It is not simply that Ayer does not grasp the ‘key tenet’. There is more than a grain of truth in the view that W – both early, middle and later – is a ‘conventionalist’ of sorts in his view of knowledge. But of what sort? Both the early and later W, for example, regard the propositions of logic and mathematics as (roughly) holding in virtue of conventions as to their use and not because there are external ‘objects’ to which they correspond. And, both early and later, W’s view of ‘induction’ and of ‘science’ might be (best) described as a form of conventionalism. Yet we need not enter into what sort of ‘conventionalism’ might be attributed to W to doubt W is a conventionalist of the sort described by Ayer. Pace Ayer, W is surely not a conventionalist of the ‘we-make-it-up-at-every-step-of-the-way’ school. Though W does perhaps want to suggest that ‘we-could-make-it-up-differently’ and that there is no independent backing for these conventions such that they could not be otherwise, and though the later W may even want to suggest that ‘we-could-make-it-up-differently-at-any-given-point’, it does not follow that the conventions as they stand have no import but stand in need of continual ‘ratification’ for their sense. To acknowledge a “rule” may be changeable at any point is not to say there is no “rule” at any point: equally, to say that the sense of a ‘what-is-said’ may be changed at any point is not to say there is no sense to a ‘what- is-said’ at any point. Even in W’s philosophy of mathematics, the kind of constructivism that suggests that working out an acceptable proof is itself establishing the conventions that govern the expansion of its terms from conventions established by prior proofs (as opposed to theview that such expansions are logically determined by prior proofs or by independently existing ‘mathematical objects’), only allows that ‘we-could-make-it-up-differently-at-any-given-point’ in this process: it does not therefore mean ‘we-need-to-make-it-up-at-every-step-of-the-way’.**** Yet perhaps Ayer does not see this last distinction. Ayer suggests it is W’s view “that we are free not only to choose our rules but also to decide what counts as following them. What other philosophers represent as the logical consequences of conventions, thereby seeming to grant some independent power of constraint to logic, [W] treats as the application of further of conventions. Such radicalism is heroic, but it is hard not to feel that it grants us more liberty than we actually possess.” This way of putting it fails to distinguish the thesis ‘we-could-make-it-up-differently-at-any-given-point’ from the thesis ‘we-must-make-it-up-at-every-step-of-the-way’. If we grant only the first thesis that ‘we-could-make-it-up-differently-at-any-given-point’, we may also accept that ‘unless-we-go-differently-at-some-given-point’ then the direction may be set by the sense of ‘conventions’ as they stand**** – and admitting this is perfectly compatible with the ‘key tenet’, which would emphasise that the sense of conventions as they stand may only be shown and is never said by those conventions. In the light of the ‘key tenet’, W is ‘saying’ no such ‘conventionalism’ such as Ayer suggests, but is rather showing that the sense of “the rule” is not said in its statement – and following out some of the implications of this, both in PI and in his philosophy of mathematics. To return to what Ayer says on W’s ‘rule-following considerations’: pace Ayer, it is not W’s point to claim that the so-called “eccentric” is not ‘mistaken’ – certainly, for W, the “eccentric” would be mistaken if he thought the sense he was giving “the rule” was the same as our sense. W’s point is that whatever kind of ‘mistake’ it may be to take the sense as the “eccentric” does, it is not a mistake as to ‘what is said’ [simpliciter] – and (which is W’s fundamental point) this shows that the sense of ‘what is said’ is not said in ‘what is said’. Failure to give this 'key tenet' its role leads to a mistaken view of W's work. Donal Slum landlords sing ‘Stop the ‘key tenet’ we want to get off…’ * PI-223: “One does not feel that one has always got to wait upon the nod (the whisper) of the rule. On the contrary, we are not on tenterhooks about what it will tell us next, but it always tells us the same, and we do what it tells us.” ** See, for example, PI-172 onwards where W remarks on “the experience of being guided”: at no point does W suggest there is no such experience or that “being guided” involves ratification of the sense of the guidance at every step. What may sometimes happen is that there is a point where we do not ‘know how to go on’, at least until something further is shown that appears to resolve the impasse, after which it may then be correct to say “Now I know how to go on.” *** PI-219: “When I obey a rule, I do not choose. I obey the rule blindly.” [W’s emphasis.] This clear statement is not offset by any words elsewhere that might suggest that, for W, when I obey a “rule” I must (consciously) decide (or decide with others) what is obeying it at every step. ****PI-238: “The rule can only seem to me to produce all its consequences in advance if I draw them as a matter of course. As much as it is a matter of course for me to call this colour "blue". (Criteria for the fact that something is 'a matter of course' for me.)” [W’s emphasis.]