¹ Kenny, Anthony, Wittgenstein (Allen Lane: The Penguin Press, 1973), Ch. 9, p. 167Google Scholar. My purpose is not exegetical, and I do not intend to suggest that Kenny would agree with what follows.

² I am aware of the pressure at this point to move to a discussion of ‘effects characterizable in terms internal to (created by) a game’ and ‘effects characterizable in terms external to (not created by) a game’. However, the new terminology is not an obvious improvement upon ‘direct effects’ and ‘indirect effects’; it produces problems of its own; and above all it cannot assist with the metaphysical aspect of the remark which is the subject of this paper.

³ Dreams sometimes include conceptual impossibilities, and so not quite everything which happens in dreams could actually happen. This qualification does not affect the argument. It should be said, too, that nothing in the argument either helps or hinders the Cartesian sceptic, for it makes no proposals about the boundary commonly drawn between dreams and reality, or about grounds for distinguishing between the two in particular and in general.

⁴ Subject, of course, to the qualification mentioned in the preceding footnote.

⁵ In view of their mimicking relationship with the world, it is also fair to describe this class of games as degenerate. Characteristic of a degenerate game is that it is devised to satisfy in fantasy what is denied the player in actuality. The solemn reader will note the special conceptual problems such games throw up.

⁶ ‘Rules and Practices’, Philosophical Review, Vol. LXXVUI, 1969Google Scholar. I do not suggest that Schwyzer would approve of the use to which I put his chess example, or indeed of anything else in this paper.

⁷ Op. cit., pp. 463–4.

⁸ I attempt at some length to induce such uneasiness in ‘Games and Language’, Mind, 84, pp. 528–47.