General Intro to NLP

Structural Analysis of Text
Structural vs Functional
Representation of meaning
Use of knowledge
Parts of Speech and Morphology
Phrase Structure / Dependency

Structural Analysis of Text

Infinite occurrences / finite processing.

Text is structured, but structure is not manifest. Therefore, problems of ambiguity (determining the intended structure). Local vs global ambiguity:

Global:
```
Flying planes made her duck
```
1. the airplanes made her change her position
2. the act of piloting made her change her position
3. piloting turned her into a duck
4. the airplanes caused her duck (the animal) to exist
5. the act of piloting made her duck exist
Local:
```
The company that bought [Scicon sold VIM].
```
Is [Scicon sold VIM] a proposition? Implications on IR.

Problem of parsing linear structure into constituent tree. The questions to address are:

Definition of constituent. When is a group a constituent?

   [John] [   [eats] [an apple]]
   [NP  ] [VP [V   ] [NP      ]]  (embedded)
   [NP  ]     [V   ] [NP      ]   (flat)

The definition of constituency depends on the goal of the analysis. Syntactic constituents, Informational constituents (given/new or theme/rheme constituents, intonation constituents. Questions raised:
1. Is there a connection between different structural dimensions? ie, Must a theme/rheme constituent also be a proper syntactic constituent?
2. How do you define empirical tests for the determination of constituents?
Are there discontinuous constituents?
```
   [Do] you [have] a hammer?
```
Are there limits to the complexity of the structure of constituents?

Why is syntactic structure important:

Structure determines meaning.
Structure impacts on form (for example, agreement).

Structural vs Functional

Map meanings to linguistic forms

What is the structural position of a constituent?
What is the function of a constituent within a sentence, the discourse?

Language is intentional - used for a purpose

Obtain info or services
Provide info or services
Just be in touch
Argument for a conclusion

Several parallel levels to language

Syntax: structure, grammar.
Semantics: meaning, propositional content.
Pragmatics: usage, speech-acts, presuppositions
Textual: contribute to extended coherent text.

Representation of meaning

Option1: meaning = procedure - set of instructions to achieve what the speaker wants. Simplest form:
- meaning of command = procedure to carry out action
- question = procedure to find answer
- statement = proc to update the hearer's beliefs
Problem: what is a procedure? opaque?
- if yes: how can explain refusal, reflection on meaning?
- if no: how different from regular representation?
Option2: network based representations. concept = nodes, links = relationships. facilitate certain type of inferences (inheritance, propagation). intuitive.
Option3: logic based representation. Compositionality principle (Frege 1890, Montague).

Use of knowledge

   "City officials refused the demonstrators a permit because"
   --- they feared violence
   --- they advocated violence

Difficult in terms of simple selectional markers. Need to reason about goals and plans of the participants to generate expectations about meaning.

   --- Excuse me, do you know if there's a bank hapoalim near here?
   --- Today is Monday!

Why is he telling me this? Regard utterances as actions given certain preconditions will achieve an Effect. Plan intended not marked in syntax:

   Can you pass the salt?
   It's rather cold!

Parts of Speech and Morphology

Syntactic analysis starts with the following questions:

What is a word? (basic terminal unit in the parse tree) Issues in tokenization.
What is a sentence? (top level unit in the parse tree) Sentence delimitation.
What are the basic properties of words

Linguists define groups of words which "behave in a similar manner" in syntactic contexts. The basic test is substitution:

  A {large, small, blue, enormous, ...} box is in the room.

Basic classes are: verb, noun, adjective. These are called parts of speech. For basic Parts of speech we can provide a semantic interpretation:

verb: action or state
noun: entity (people, concepts, things)
adjective: property

Distinguish:

Open-class (many members, new members often added)
Closed-class (few members, functional use): article, preposition...

Morphological features:

Features of words: number, gender, tense, person, case
Processes: Inflection, derivation, compounding.

Inflection is the process of modifying a root form by combining prefixes and suffixes to indicate the presence of morphological features.

Derivation is the process of creating a new lexical item from an existing, more basic one. For example, the derivation of the noun of action destruction from the verb destroy, or the chain of derivations luck/noun, lucky/adjective, luckily/adverb.

Compounding is the process of combining several lexical items into a new one, whose properties can be derived from the compounded elements, or can be independent. For example, in Hebrew beit sefer is derived from bayit and sefer. For more information on morphology, refer to this excellent Introduction to morphology.

Phrase Structure / Dependency

Constraints on word order. Words occur in groups, with dependencies among them. Constituent / Phrase.

Common Phrases:

Noun Phrase (NP)
Adjectival Phrase (AP)
Verb Phrase (VP)
Clause

Phrase Structure Grammars / (PSG)

Context Free Grammars (CFG)

Last modified Mar 18, 2007