wouldn't it make sense for "seem to read" to be considered a verb phrase, so that s(n(John)vP(seems ppP(to read))) ? Also, "seems" sounds like it can be replaced by the phrase "does seem", which allows "John does seem to read" to become "does John seem to read?". In this sense, does acts as an auxiliary verb for the verb phrase "seem to read" and, indeed, "John" is capable of "seeming X" whether "X" be a prepositional phrase, or adjectival phrase.
"To read" is not a PP. PPs take noun phrase (or determiner phrase) complements. Here, you're dealing with an infinitival clause, a type of embedded CP which in itself is the complement of 'seems'. In transformational grammar, this is called subject-subject raising.
I am wondering if you use Minimalist Grammar and still consider it Transformational? Also, you use English for examples; however, how about isolating and polysynthetic languages that include particles (e.g., evidentials, TAM, exclusion, inclusion, etc.)? Will we see you tackle these languages and if you do will we see the use of Role and Reference Grammar (which works well for those languages, while Chomsky does not)? I am trying to get an idea of the breadth of languages studied there.
To come back to my question : preposition belong to Verbal group or Nominal group ? As I said you, I believe preposition are antique verbs, and I feel they work as adverb (like out in look out).
+meriem drissi N' stands for N-bar (the bar on top of N is more difficult to write than N with an apostrophe). It is an intermediate phrasal category between NP and N.
Oh ! J'oubliai ! Sure etymon research can help. About "seems" you said tthe agent don't do the action and so, it is object of the verb. According to the etymon, it's not right. Seem means "substance" and in "semit etymon bank", it's about shape. So John seems = John wear an appearence > so he is actor, not object.
