Transcript Lecture 2 PP

Syntax
Lecture 2:
Categories and Subcategorisation
Recap
• All phrases are structurally the
same:
• All phrases differ in their lexical
content
• Lexical content determines the
syntactic differences between
phrases
– Particularly the lexical head
Differences in lexical heads
• So how syntactically different can lexical heads
be?
• There are two way heads differ
– They have different categories
– They differ in what complements they select
How many different categories are
there?
• Most syntactic descriptions work with a
relatively small number of different categories
•
•
•
•
•
•
Noun
Verb
Adjective
Adverb
Preposition
Pronoun
•
•
•
•
•
•
Determiner
Particle
Subordinator
Coordinator
Auxiliary verb
Degree adverb
• In fact, some of these collapse into single
categories
Collapsing categories: pronouns
and determiners
• Many determiners work as pronouns
– This book was banned
This was banned
– Some people are sad Some are sad
– Few aeroplanes crashFew crash
• Some pronouns work as determiners
– We humans
– Them rocks (dialectal)
– You lot
• It has been suggested that pronouns should be
analysed as determiners
Important Consequence
• If pronouns are determiners, then they head
DPs
• Pronouns have the same distribution as
phrases such as the man:
– The man/he drove to the shops
– I saw the man/him in the car
– I’ve heard about the man/him
• Therefore these phrases must also be DPs
Important Consequence
Collapsing categories:
subordinators and adverbs
• Some subordinating particles behave exactly
like adverbs
• Obviously, he had gone
• He, obviously, had gone
• He had gone, obviously
• However, he had gone
• He, however, had gone
• He had gone, however
• There is no reason not to analyse these as
adverbs
Collapsing categories:
subordinators and adverbs
• Some subordinating particles don’t behave
like adverbs
• Obviously, he had gone
• He, obviously, had gone
• He had gone, obviously
• ... that he had gone
• * ... he that had gone
• * ... he had gone that
• These are clearly of a different category
• We call them Complementisers
Collapsing Categories: adverbs and
adjectives
• Many adverbs and adjectives have the same
root:
– obvious: obviously fast: fast
great: greatly
• Adverbs and adjectives are in complementary
distribution
– Adverbs modify verbs, adjectives modify nouns
• They might be different subcategories of a
general category of ‘modifier’ (often called A)
Conclusion
• So it seems that the number of categories we
need to describe language is very small
• Why is that?
• Without a theory of categories, we can’t
explain this.
Different categories sometimes
have things in common
• Verbs and prepositions take ‘bare’ objects:
– visited London
– to London
saw the man
for the man
shot him
with him
• Nouns and Adjectives take objects with ‘of’:
– Picture of Mary
– Fond of Mary
growth of the trees
mindful of the trees
• How can we explain these facts if categories
are completely unconnected?
A theory of categories
• We know that all categories fall into one of
two main types
• Functional
•
•
•
•
Determiners
Auxiliary verbs
Complementisers
Etc.
• Thematic
•
•
•
•
Nouns
Verbs
A (modifiers)
Etc.
A theory of categories
• This suggests a ‘binary feature’ analysis
– (like distinctive features in phonology: ±voice,
±long)
• Suppose we assume a feature ±F
– +F = Functional categories
– -F = Thematic categories
A theory of categories
A theory of categories
• But this still isn’t very restrictive
• One way to restrict the system is to assume
that all categories are defined by binary
features
• This would also account for similarities
between different categories
– Distinct categories can share one or more features
A theory of categories
• How many more binary features do we need?
• Not too many!:
– 1 feature = 2 categories
– 2 features = 4 categories
– 3 features = 8 categories
– 4 features = 16 categories
– 10 features = 1024 categories
- not enough!
- not enough!
- perhaps
- too many!
- way too many!
A theory of categories
• Suppose we suggest two extra features:
– ±N
– ±V
(things which are ‘nounlike’)
(things which are ‘verblike’)
• Assuming nouns and verbs to be opposites to
each other we get:
– Noun = [-F, +N, -V]
– Verb = [-F, -N, +V]
• This is supported by the fact that nouns and
verbs share very little in common
-F categories
• There are two more –F categories:
– [-F, +N, +V]
– [-F, -N, -V]
• The first seems appropriate for A
– They modify both nouns and verbs
– Adjectives are often used as nouns
• The good, the bad and the ugly
– In some languages adjectives are used as verbs
– Nouns and adjectives don’t take bare objects
-F categories
• [-F, -N, -V] seems appropriate for prepositions:
– Prepositions have no morphological properties
• They can’t be tensed
• They can’t be plural
– Like verbs, they take bare objects (both are –N)
-F categories
• We predict the following possible categories
• Some
categories
have
some
things in
common
• We also predict that there are no other
thematic categories
+F categories
• The theory predicts four functional categories
• These are the functional equivalents to:
– Nouns
– Verbs
–A
– Prepositions
([+F, +N, -V])
([+F, -N, +V])
([+F, +N, +V])
([+F, -N, -V])
Functional equivalents
• The most obvious functional nominal is the
determiner
– Determiner = [+F, +N, -V]
• The most obvious functional verb is the auxiliary
– Auxiliary = [+F, -N, +V]
• Degree adverbs are similar to determiners in APs
(the man : so tall)
– Degree adverbs (Deg) = [+F, +N, +V]
• Complementisers are similar to prepositions
(both introduce arguments)
– Complementisers = [+F, -N, -V]
+F categories
• We predict the following possible categories
• We also predict that there are no other
thematic categories
A theory of categories
Subcategories
• The subcategories of a category are
determined by what follows them
– E.g. Verbs can be transitive (i.e. they are followed
by an object) or intransitive (i.e. they are not
followed by an object)
• In other words, subcategories are determined
by what appears in the complement position
Complement Position
Complement
The complement of functional
categories
• The functional categories do not usually have
subcategories – they almost always take the
same complements
– The complement of an auxiliary verb is always a
VP
• may [VP win the race]
– The complement of a complementiser is always a
sentence
• that [he may win the race]
The complement of functional
categories
– The complement of a degree adverb is always an
AP
• so [AP fond of chocolate]
– The complement of a determiner is usually an NP
• The [NP man from Brazil]
The complements of thematic
categories
• Thematic categories can take various types of
complement and so have a number of
subcategories
• Verbs can be followed by
– A DP
– A clause
– A PP
– An AP
see [DP the news]
think [ that he saw the news]
react [PP to the news]
feel [AP sorry]
The complements of thematic
categories
• Prepositions can have the same range of
complements as verbs, except for clauses
– DP
– PP
– AP
to [DP the west]
from [PP under the bed]
(range) from [AP heavy] to [AP medium]
The complements of thematic
categories
• Nouns can have the same complements as
verbs, except for DPs
– Clauses
– PP
– AP
belief [that he can fly]
reaction [PP to the news]
(his) feeling [AP ill]
The complements of thematic
categories
• ‘A’s can take clausal and PP complements
– Clause
– PP
likely [that he will fail]
keen [PP on ice hockey]
Conclusion
• Heads determine the category of phrases
– But there are only 8 categories of heads
– So there can only be 8 different categories of phrases
• Heads determine the category of their
complements
– Functional heads only take one type of complement
– Thematic heads take more types of complement, but
none of them are unrestricted
• We still have a restricted theory of phrases