When, in “Was heißt ‚praktisches Wissen‘?” (there's a nice mess of typography for you ), Müller is trying to make sense of the "direct falsification" of an intention (or its report) by an act, the fourth possibility he canvasses seems to lose some of its features between its introduction and its discussion. Or perhaps: it is introduced with a collection of features that don't really belong together and when discussed only one of these is attended to. The example is that I say "now I press Button A" and then press Button B ; the first three possibilities Müller considers are that (1) between the intention and the act falls a changing of my mind; (2) I discover that I can't press Button A and press B instead, since they do the same thing; (3) between deciding what to do and doing it I forget what I was going to do. The fourth is: the possibility that "ich versehentlich („geistesabwesend“ oder indem ich „mich vergreife“) etwas anderes tue, als ich zu tun gedenke und behaupte". The gloss on "versehentlich" (whose definition here combines a few distinguishable things: "inadvertently", "erroneously", and "by mistake" aren't the same) as "absent-minded" is what gets picked up in the discussion later, where Müller likens what is going to the case of a driver who, told to go left, echoes "left" and then turns confidently to the right. Here, he says, possibly there is no right answer to the question of what the driver didn't know, what he said or what he did. And if the case really is one of absent-mindedness, that may well be right; but if it's one of proceeding erroneously, it needn't at all be the case. Consider: I am giving a narrative of my actions so as to explain them to you, my apprentice. Naturally in doing this I sometimes turn partially to face you and so I come to say "now I press Button A" but, the two buttons being very close to each other, press B instead. I don't do this absentmindedly and I don't think there should be any question as to whether I perhaps didn't know what I was saying. (Perhaps I do do it versehentlich.)
That Anscombe should choose the locution "now I press Button A" instead of "now I'm pressing Button A" is interesting; the former is much more suited for the imagined scenario of above, because it is more suited to describing what is to be done (what one does, what is habitually done, etc.) than it is to describing what is actually being done. I hesitate before the panel, uncertain how to proceed, and remind myself: I just did such-and-such, so "now I press Button A". Clumsily I press Button B anyway, but here what I was claiming to know wasn't what I was actually up to.
 The HTML entity for this symbol: „ is "„", that is, I guess, bottom double quotation mark. But for this symbol: ‚ it's "‚", that is, I continue to guess, single bottom quotation mark. Why the position of the "b" relative to the single-or-double determiner should vary is beyond human comprehension.
 Actually, the example Müller considers has the agent saying that he shall press Button B and then presses Button A, and he quotes the translation of Intention (Absicht, not Intention, which reminds me that if you look in the index for the first volume of Wittgenstein's Remarks on the Philosophy of Psychology there are a few entries for Intention but many more for Absicht, while in the second volume there are practically none for Absicht and many for Intention—how intriguing) giving the example just so. But in the English original it's the other way around. Bizarre.