Noun Phrases
Note that in previous examples, VPs have branches to the right.
Let's look at modifiers to the NPs.
E.g.,
Jemison landed the space shuttle.
This sentence can be diagrammed as below.
---------------------------------------------------------------
|
sentence |
/| |
/ | |
/ VP |
/ | |
NP | |
/ | |
/ | |
/ | |
/ - ______________ |
/ V \ |
/ | NP |
Jemison | | |
- | |
V | |
| __ - |
| / N |
| / | |
| / _ - |
| / / N |
| / / | |
| / / | |
| / | NOUN |
| / | | |
| the space shuttle |
| |
VERB |
| |
landed |
|
|
---------------------------------------------------------------
Where are the branches for an NP?
Phrases that are brought in from the right are called complements (and also adjuncts - which is one of many debated theories in the field of NLP.)
Phrases that are brought in from the left are called specifiers.
[Do you see why these terms are used?]
This could be a possibly satisfactory result except that some languages use nouns as verbs and vice versa (or specifically the same word can function as either).
To be universal it would be nice if NP's and VP's behaved symmetrically.
Consider the sentence (an example with PP specifiers) below.
Jemison landed the space shuttle precisely at eleven.
---------------------------------------------------------------
|
sentence |
/| |
/ | |
/ VP |
/ | |
NP - ___________________________ |
/ V \ |
/ | | |
/ | | |
/ - ______________ | |
/ V \ PP |
/ | NP / | |
Jemison | | / - __ |
- | / P \ |
V | Specifier | NP |
| __ - | | ^ |
| / N | | / \ |
| / | precisely | eleven |
| / _ - | |
| / / N at |
| / / | |
| / / | |
| / | NOUN |
| / | | |
| the space shuttle |
| |
VERB |
| |
landed |
|
|
---------------------------------------------------------------
Another example, this time with verb specifiers:
The astronauts all landed their space shuttles.
---------------------------------------------------------------
|
sentence |
/| |
/ | |
______/ | |
/ | |
NP | |
/\ | |
/ \ | |
/ \ | |
/ \ | |
/ \ | |
/ \ | |
the astronauts | |
| |
VP |
/| |
/ | |
/ | |
/ | |
/ | |
Specifier | |
| - ______________ |
| V \ |
| | NP |
all | ^ |
| / \ |
VERB / \ |
| their space shuttles |
landed |
|
|
---------------------------------------------------------------
The X-Bar Hypothesis says that all language structures can be recognized by binary tree phrase structures.
-------- the X-Bar Tree Structure -----------------------------
|
| |
XP |
/| |
/ | |
/ | |
/ | |
/ | |
Specifier | |
- ______________ |
X \ |
| Complement |
- ______________ |
X \ |
| Complement |
- ______________ |
X \ |
| Complement |
- ______________ |
X \ |
| Complement |
| |
X |
|
|
---------------------------------------------------------------
What this means to NLP.
It doesn't mean we will be able to use only grammars that generate these kinds of trees. That would be overwhelmingly complex; and the theory has not been shown to yield trees which match the underlying mechanisms of the brain.
It does mean that certain grammars are "better" than others. Those that try to match certain regularities inherent in the source language are preferred.