1 ‘lfs and Cans', Proceedings of the British Academy, 1956, reprinted in Austin's Philosophical Papers, ed. Urmson and Warnock (Oxford, 1961). All page re· ferences are to this reprint.

2 ‘lfs and Cans', Canadian Journal of Philosophy 1 (December 1971) and (March 1972), reprinted in Questions in the Philosophy of Mind (Duckworth, 1975). All page references are to this reprint. Note that Pears finally elects to treat failure in the contrapositive test as both a necessary and sufficient condition of pseudoconditionality.

3 Ethics, (Oxford, 1912), Ch. VI.

4 I treat the noun ‘ability’ as being as versatile as the verb ‘can’ in its application to action, e.g. as being more generally applicable than ‘capacity’ or ‘opportunity', as having ‘all-in’ as well as more circumscribed uses.

5 Some of the objections to an unqualified affirmative to B have been listed by Roderick Chisholm in ‘J.L. Austin's Philosophical Papers'. Mind 73 (1964). My thesis presupposes that these and other objections rest on cases, which though common enough, can all be treated as exceptions. See this paper's final section.

6 We could make a comparison with the same purpose by negating (i), (ii) and the consequent of (iii), and by denegating the antecedent of (iv) in each list. This would illustrate the double failure equally well.

7 One could, of course, extend the use of the arrow-sign by representing (iii) and (iv) as respectively p → (p →q) and ∿ (p→q)→∿p. But then, what could that mean? What would it meand to say that someone's H-ing was a sufficient causal condition of his H-ing being a sufficient causal condition of disaster or that someone's H-ing not being a sufficient causal condition of disaster was a sufficient causal condition of his not H-ing? Statements of that kind might be given some plausible sense by some arbitrary stipulation, but such stipulations could not give the sense of our (iii)'s and (iv)'s in so far as these have their sense by standard convention.

8 That this is true in the case of L(iii) has been noticed by Donald Davidson. See his ‘Freedom to Act', Essays on Freedom of Action, ed. Honderich, Ted (London, 1973), p. 143.Google Scholar Davidson, however, does not make quite explicit where exactly the failure lies.

9 Compare with L‘(ii) and L‘(iii): ‘The cause of disaster would be his H-ing’ and 'The cause will be his H-ing if there is disaster?’ Note too that a number of specifically logical terms such as ‘material implication’ and ‘entailment’ can likewise acquire if-clauses as labels.

10 ‘An Empirical Disproof of Determinsm?’ Freedom and Determinism, ed. Lehrer, Keith (Random House, 1966), p. 196.Google Scholar

11 For other criticism of Lehrer's argument see Aune's, Bruce “Hypotheticals and 'Can': Another Look”, Analysis, 27 (June 1967);CrossRefGoogle Scholar and “Free Will, ‘Can', and Ethics: A Reply to Lehrer”, Analysis, 30 (January 1970); Donald Davidson, Freedom to Act', Essays on Freedom of Action, ed. Honderich, Ted (Routledge and Kegan Paul, 1973) p. 145;Google Scholar and Walton, Douglas ‘lfs and Cans: Pros and Cons', The Personalist, 56 (Summer 1975).Google Scholar

12 In refutation of Lehrer, Donald Davidson employs a similar comparison (loc cit), but not for the purpose of resolving the difficulty.

13 ‘lfs and Cans', p. 159.

14 ‘Performative Utterances', Philosophical Papers, p. 238.

15 Lecture VIII in How to do things with Words, ed. Urmson (Oxford 1962).

16 ‘lfs and Cans', p. 158.

17 Similarly the antecedent in the Geach example ('lfs-and Cans', p. 161). ‘I paid you back yesterday, if you remember’ is a ‘happiness’ condition of the consequent performing the illocutionary function of a reminder, and constitutes by its addition an explicit illocutionary formula for a nonconfrontational reminder. In both examples, however, an alternative interpretation is possible. We might regard the antecedents as specifying a background condition of the consequents not having the perlocutionary effect of seeming respectively presumptious and confrontational. On this alternative the antecedents would no longer label the consequents. They simply turn the latter into parts of a more refined illocutionary act. I owe this suggestion to my colleague Leonard Angel.

18 loc. cit. p. 145.

19 On my analysis, of course, labels are adverbial in a way as well. They qualify various kinds of operations, e.g. modal, conditional, and illocutionary, upon propositions, but not anything upon which these operations operate. Thus the antecedent in L(iii) qualifies the ‘can’ adverbially, but not the ‘X'.

20 This distinction of his becomes the subject of further refinement. Normally an S·factor will be distinct from an I-factor, but he suggests that the same factor may sometimes have both aspects. Then again, when they are distinct, how the distinction is drawn may vary from person to person according to context and concern. And further refinements have to be made as well.

21 loc. cit. p. 151.

22 For a psychologically more complete analysis of intentionality see my ‘The Non-Causal Self-Fulfillment of Intention', The American Philosophical Quarterly 9 (October 1972).

23 Quantitatively and qualitatively this paper has benefited from editorial advice on how to prune it.