On 12 April 2011 01:00, Raul Miller <[email protected]> wrote: > I cannot imagine any interpretations which conflict with my statement > which do not also conflict with the dictionary entry. > > But perhaps you can show me one?
You said: > Unless 0=#y (which is explicitly treated), u/y necessarily has _1+#y > instances of u. This follows immediately from the first sentence at > http://www.jsoftware.com/help/dictionary/d420.htm and the sentence you are referring to is `u/y applies the dyad u between the items of y'. If you choose to maintain that, somehow, `0 applications between 1 elements' has definite meaning, then it seems to me that, to be self-consistent, you should go all the way through and also consider `-1 applications between 0 elements' (in which case you would also have to explain how -1 applications of a verb yield that verb's neutral element). I prefer to stick by common sense and consider the DoJ statement as applying only to 1<#y, since only then the word `between' is meaningful. Thus, both 1=#y and 0=#y require a separate explanation. On 12 April 2011 01:15, Brian Schott <[email protected]> wrote: > Actually, then (u/k{.y) u (u/k}.y) produces the correct domain error > when y=:'!', doesn't it? Yes, it is correct to produce an error. But this error invalidates the identity u/y ↔ (u/k{.y) u (u/k}.y), because its l.h.s. produces a normal value. And the identity does not generally hold for several other reasons as well. What I mean is that it is not correct to use a generally false identity as the above, in order to infer the value (meaning) of u/y for 1=#y (as I believe that was what you intended -- or am I mistaken?). ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
