G
gender
One of the features of a noun phrase. In English, gender is only marked in third-person singular pronouns and associated words. The possible values of the gender feature are masculine, feminine, and neuter.
type masculine feminine neuter example
pronoun (nominative) he she it he hit the ball.
pronoun (accusative) him her it Frank hit him.
pronoun(possessive adjective) his her its Frank hit his arm.
pronoun(possessive) his hers its The ball is his.
pronoun(reflexive) himself herself itself Frank hurt himself.
generalized phrasestructure grammar (GPSG)
An alternativegrammatical formalism, in which, among other things, the non-terminal symbols of context-free grammars are replaced by sets of features, and the grammar rules show the relationships between these objects much as context-free rules show the relationships between grammar symbols in a CFG.
generate
If there is a derivation of a sentence from a grammar, then the grammar is said to generate the sentence.
GPSG
= generalized phrasestructure grammar
grammar
A system for describing a language, the rules of a language.
A formalsystem for describing the syntax of a language. In COMP9414, we are principallyconcerned with context-free grammars, sometimes augmented.
See also Chomsky hierarchy.
grammar rule
See Chomsky hierarchy and context-free grammars.
H
HDPSG
= head-driven phrasestructure grammar
head feature
A head feature is one for which the feature value on a parent category must be the same as the value on the head subconstituent. Each phrasal category has associated with it a head subconstituent - N, NAME or PRO or CNP for NPs, VP for S, V for VP, P (= PREP) for PP.
For example, var is a head feature for a range of phrasal categories, including S. This means that an S gets its var feature by copying the var feature of its head subconstituent, namely its VP.
Head features are discussed on page 94-96 of Allen.
head subconstituent
See head feature.
head-driven phrasestructure grammar
A grammatical formalism. See Allen p. 154.
hidden Markov model
A Hidden Markov Model, for our purposes in COMP9414, is a set of states (lexical categories in our case) with directed edges (cf. directed graphs) labelled with transition probabilities that indicate the probability of moving to the state at the end of the directed edge, given that one is now in the state at the start of the edge. The states are also labelled with a function which indicates the probabilities of outputting different symbols if in that state (while in a state, one outputs a single symbol before moving to the next state). In our case, the symboloutput from a state/lexical category is a word belonging to that lexical category. Here is an example:
HMM diagram
Using this model, the probability of generating "dogcatchers catch old red fish" can be calculated as follows: first work out the probability of the lexical categorysequence → N → V → ADJ → ADJ → N, which is 1 × 1 × 0.5 × 0.1 × 0.9 = 0.045, and then multiply this by the product of the output probabilities of the words, i.e. by 0.3 × 0.2 × 0.6 × 0.2 × 0.5 = 0.0036, for a final probability of 0.000162.
history list
This is the list of discourse entities mentioned in recent sentences, ordered from most recent to least recent. Some versions also include the syntactic and semantic analyses of the previoussentence (or previousclause of a compound sentence). Some versions keep only the last few sentences worth of discourse entities, others keep all the discourse entities since the start of the discourse.
HMM
= Hidden Markov model
I
ill-formed text
Much "naturally occurring" text contains some or many typographical errors or other errors. Industrial-strength parsers have to be able to deal with these, just as people can deal with typos and ungrammaticality. Such a parser is called a robust parser.
Here is a list of 300+ ill-formed sentences.
imperative
An imperativesentence is one that expresses a command, as opposed to a question or a statement. See also WH-question, Y/N-question, indicative, subjunctive, and mood.
independence (statistical)
Two events A and B are said to be statistically independent if Pr(B | A) = Pr(B) - i.e. whether or not A is a fact has no effect on the probability that B will occur. Using Bayes' rule, this can be reformulated as Pr(A and B) = Pr(A) × Pr(B).
indicative
An indicativesentence is one that makes a statement, as opposed to a question or a command. See also WH-question, Y/N-question, imperative, subjunctive, and mood.
infinitive
A form of verbs. In English, this form is introduced by the word "to" - the infinitive particle. Examples: to go, to rather prefer, to have owned. Verb phrases that have an infinitive verb construction in them are referred to as infinitive verb phrases. Constructions that are not infinitive are referred to as infinitive verb phrases. See vp:inf and np_vp:inf in the article on subcategorization.
See also here for the distinction between "infinite" and "infinitive".
inflection
An inflection is a type of bound morpheme, with a grammatical function. For example, the suffix "-ing" is an inflection which, when attached to a participle form of the verb. Other inflections in English form the other parts of verbs (such as the past tense and past participle forms), and the plural of nouns.
Some words inflect regularly, and some inflect irregularly, like the plural form "children" of "child", and the past tense and past participle forms "broke" and "broken" of the verb "break".
instrument
A semantic case, frequently appearing as a prepositional phrase introduced by the preposition "with". For example, in "Mary ate the pizza with her fingers", the prepositional phrase "with her fingers" indicates the instrument used in the action described by the sentence.
intensifier
A kind of adverb, used to indicate the level or intensity of an adjective or another adverb. Examples include "very", "slightly", "rather", "somewhat" and "extremely". An example of use with an adjective: "Steve was somewhat tired". An example of use with an adverb: "Mary ran very quickly".
INTERJ
symbol used in grammar rules for an interjection.
interjection
Interjection is often abbreviated to INTERJ.
INTERJ is a lexical grammatical category. It usually appears as a single word utterance, indicating some strong emotion or reaction to something. Examples include: "Oh!", "Ouch!", "No!", "Hurray!" and a range of blasphemies and obscenities, starting with "Damn!".
intransitive
A verb that can take no syntactic object, like laugh, as in "He laughed loudly", or "She laughed at his remark". Contrast ditransitive and transitive. See also subcategorization.