You are perhaps right, but perhaps that didn't need to be posted the the
whole mailing list. But, to avoid the appearance of nitpicking you might
have at least posted an LFG phrase structure rule or some kind of
description:
Analysis:
1. "what this hack does" is a sentential complement used as an argument
to the verb KNOW<(SUBJ)(SCOMP)>. For example:
A. I know [what this hack does]. (grammatical)
B. I know [what does this hack do]. (ungrammatical)
Not all sentential complements are embedded questions: I've marked
one here that isn't:
A. I believe [that Linus is cool]!
You can perhaps see that the phrase-structure rule in this case is fairly
simply:
S' -> "that" S (s' is kind of the node corresponding to SCOMP)
2. "What does this hack do?" is probably not a sentential complement, but
only a question. Notice that you insert an auxilliary very "does" after
the "what" and then move the tense information onto the auxilliary.
A. What this hack does? (ungrammatical)
B. What does this hack do? (grammatical)
3. In English, you must alter embedded quotes if you use a sentential
complement:
(Embedded SCOMPs marked with [])
Case A: (Embedded quote marked with (SURPRISE!) quote marks :)
I told him "You know [that Linus is cool]!"
Case B:
I told him [that he knew that Linus was cool].
Notice the changes:
you->he
know->knew
is->was
So, there. Perhaps I've done justice to the situation. I hope so...
But, perhaps, next time, we could keep such complicated topics off the
kernel mailing list by not nitpicking publicly or adding some ":)" or
something :)
-BenRI
P.S.
I also noticed:
I hit the person THAT did it. (OK)
I hit the person WHO did it. (OK)
I hit the thing THAT did it. (OK)
I hit the thing WHAT did it. (NOT OK!) Why? Because! :)
Is this perhaps a regional variation?
(I live in USA, California. Are you surprised? :)
Footnotes:
Lexical Functional Grammar: A formal System for Grammatical Representation
BY: Ronald M. Kaplan & Joan Bresnan
IN: Grammatical Representation of FUnctional Relations (a big volume
editted by bresnan)