>>52
Correct.
>>51
Whoever mentioned Chomsky was full of shit.
Maybe in
"[[it]NP [[is]V [difficult [to sleep]VP [while [reading]VP [that book]NP]CP]AP]V']VP"
the Japanese speaker has incorrectly attempted to front the NP "that book" and replace the PRO subject NP with it. This simple reordering could be possible in Japanese because verb argument order is free, however in English the order of arguments is fixed SVO* order. Trying to front that results in a subject NP "that book" which serves as the object argument to a transitive verb "to sleep" embedded in the "difficult" AP. That won't work in English, since "to sleep" is only intransitive. The construction would work with any transitive verb, however. Example: "that beer is difficult to drink while driving", wherein the NP "that beer" has been fronted from its object position of the verb "to drink".
BTW, all that crap with transformations is silly. More modern syntax has surpassed Chomsky's requirements for transformations to make up for deficiencies in grammatical processing. Instead of complex multipass top-down parsing like Chomsky requires, modern syntax theories use bottom-up parsing techniques and one-pass parses which obviate the need for transformations and the D-structure/S-structure dichotomy. Unfortunately, although these modern syntax theories are far more efficient and mathematically computable (transformations appear to be NP-complete), nobody is teaching them to undergraduate students, so they remain a feature of theoretical linguistics unavailable to the wider scientific community. Some CS people are familiar with them because the modern syntax theories, being computable, lend themselves towards research in NLP and language understanding, but in general most academics have no exposure to them, nor even knowledge that they exist. This will probably change within the decade, however, and will certainly change once Chomsky dies and stops with his incessant attacks on anyone who doesn't do things his way.