Inflection

From UNL Wiki
Revision as of 19:47, 23 August 2012 by Vzakharov (Talk | contribs)
Jump to: navigation, search

In grammar, inflection is the modification of a word to express different grammatical categories such as tense, mood, voice, aspect, person, number, gender and case. Conjugation is the inflection of verbs; declension is the inflection of nouns, adjectives and pronouns.

Contents

Inflectional Paradigms and Inflectional Rules

In the UNLarium framework, inflection is indicated by a set of transformations carried over the base form. This set of transformations can be represented by:

  • inflectional paradigms, in case of regular behaviour (i.e., a set of transformations that is followed by several different words)
  • inflectional rules, in case of irregular behaviour (i.e., a set of transformation that is followed by very few words); or
  • inflectional paradigms and inflectional rules, in case of quasi-regular behaviour (i.e., when the word is mainly regular but has some inflectional particularities).

The difference between inflectional paradigms and inflectional rules is mainly a question of frequency. If a rule is applicable to several different words, it should be defined as a general inflectional paradigm; if it is applicable to a single word or to a very limited number of cases, it should be defined as an inflectional rule inside the very entry.

For instance, the plural of English nouns is considerably regular and can be treated, in most cases, by the general paradigm PLR:=0>"s"; (= add an "s" to the end of the word). This paradigm may be associated to a wide range of English words (such as table, book, boy, computer, etc). However, there are several special cases (such as man>men, child>children, foot>feet) that, being very limited, should be treated by inflectional rules instead of inflectional paradigms. The third case ("quasi-regular") is considerably rare in English, but may be found in French, for instance, where verbs such as acheter (= to buy) follows the conjugation of the second group in all cases but undergoes a small change in the root ("e" becomes "è") in some persons and tenses. To avoid the proliferation of paradigms, quasi-regular words are associated to general inflectional models, and only their corresponding exceptions are listed as inflectional rules.

Syntax

Inflectional paradigms and inflectional rules are expressed by A-rules, a special formalism for introducing prefixes, infixes and suffixes to the base form.

Inflections inside the dictionary

The UNLarium framework brings two main dictionary formats:

  • Generative, i.e., base forms and the corresponding lexical features and generation rules (to be used in natural language generation)
  • Enumerative, i.e., word forms, with the corresponding lexical features (to be used in natural language analysis)

A sample of each format of the English dictionary is presented below:

generative
[foot] {2883} "100284665" (POS=NOU,LST=WRD,NUM=SNG,PAR=M1,INFR=FLX(PLR:="feet";)) <eng,0,0>;
enumerative
[foot] {2883} "100284665" (POS=NOU,LST=WRD,NUM=SNG,PAR=M1) <eng,0,0>;
[feet] {2883} "100284665" (POS=NOU,LST=WRD,NUM=PLR,PAR=M1) <eng,0,0>;

The enumerative dictionary includes all inflected forms as entries and, therefore, does not bring any inflectional rule except for the inflectional paradigm (PAR=M1). The generative dictionary, which brings only base forms as entries, includes the information on the inflectional paradigm and on inflectional rules.
Inflections are represented in the dictionary entry structure as follows:

  • Inflectional paradigms are informed as PAR=Mxx, where "PAR" indicates "inflectional paradigm"; "M" is the fixed leading symbol for the paradigm; and "xx" is the number of the inflectional paradigm previously defined in the grammar.
  • Inflectional rules are informed as INFR=FLX(<A-RULE>) where INFR indicates "inflectional rule"; "FLX" is the fixed leading tag for the inflection rules set; and <A-RULE> is the affixation rule, or set of affixation rules, used to generate the inflections out of the base form.

Examples of dictionary entries with inflectional information

[kill]{5987}"200355177"(FRA=Y38,LST=WRD,MOR=ROO,PAR=M16,POS=VER,TRA=TST) <eng,70,80>;
[bring] {2345}"200767334"(FRA=Y38,LST=WRD,MOR=STE,PAR=M22,POS=VER,TRA=TST,INFR=FLX(PAS:="brought";PTP:="brought";))<eng,50,34>;

Software