TUTORIAL MATERIAL: more advanced phrase structures

Here is a small, but quite complicated, grammar:

S ® NP VP
NP ® (Det) (AP) N (PP) (S¢)
VP ® V (NP)(PP)(AP)
AP ® (Intensifier) A
PP ® Prep NP
S¢ ® Comp S

Here are the lexical items that you are allowed to use with this grammar (you can ignore morphological variation, i.e., language is taken to include languages, use is taken to include uses, using, used, etc.):

V ® {be, use}
A ® {obvious, different}
Intensifier ® {very, quite, really}
N ® {language, nation, country, question, tool, people, fact, it}
Prep ® {of, from, in}
Det ® {the, a, all}
Comp ® {whether, that, for, of}

Can this grammar and lexicon produce trees that corresponds to each of the following sentences? And, in each case that it can, what are the Process, Participants and Circumstances and what is the Subject?

  1. All people use different languages.
  2. It is a fact.
  3. It is very obvious.
  4. It is a fact that people use different languages
  5. It is an obvious fact that people of different nations use different languages.

Does the grammar overgenerate?

Can the grammar produce trees for the following sentences? Can the grammar produce more than one tree for any of them?.

  1. It is a people that nations use tools.
  2. It is a fact for language is a tool.
  3. Languages are a tool of different nations very obvious.
  4. The question that people use languages is obvious.
  5. The fact that people using languages is obvious.

Here are some more lexical entries involving selection restrictions or subcategorization (as usual, something in brackets is optional):

fact, N: ___ (S¢ )		[i.e., ‘fact’ may be followed by an S¢ ]
people, N: ___ 		[i.e., people'e does not want anything after it!]
be, V: ___ NP		[i.e., ‘be’ wants an NP after it]
be, V: ___ AP		[i.e., ‘be’ can also, however, want an AP after it]
use, V: ___ NP		[i.e., ‘use’ wants an NP after it]

Does this change what the grammar can generate? How? Are all of the sentences 6—10 still possible? Are all of the sentences 1—5 still possible?

Here are some further lexical entries, but now using grammatical features:

for, Comp: ___ S [to-infinite]	[i.e., a ‘sentence’ with a ‘to infinitive’ for the verb]
whether, Comp: ___ S [finite]	[i.e., a full normal sentence with  a finite verb]
that, Comp: ___ S [finite]				ditto
of, Comp: ___ S [ing]		[i.e., a ‘sentence’ where the verb is in the –ing form]

Has this changed whether any of the sentences in 6—10 are possible are not? If so, show the old tree and indicate why it is not now possible.

If we change the lexical entry above for fact to be the following and add one for whether:

fact, N: ___ (S¢ [that])		[i.e., fact may be followed by an S¢ with a  feature ‘that’]
question, N: ___ (S¢ [whether])	[i.e., question may be followed by an S¢ using  ‘whether’]

Can we generate a sentence like: ‘The question whether nations using tools is obvious.’ ? If so, why? If not, why not? How does the information that Comp is the head of an S¢ phrase, and V is the head of a VP, and VP is a head of an S help, if at all? Think about what syntactic features can move where....