The application to English has been carried out by the MULTEXT Group
at ISSCO (ISSCO 1994).
Note that this application has been carried out on the basis of the
attributes and the values as presented in the preceding version of
this deliverable (MULTEXT WP1.6 A2 version).
Notation:
Trailing place-holders have been omitted.
It should be borne in mind that the need for place-holders is an artefact of the linear representation of lexical descriptions. A number of extensions have been made to the various proposals circulated. It is not clear how language-specific they are, but they represent phenomena that are plausibly relevant for various text-processing tasks.
5.5.1 Nouns (N) --------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Type Common c Proper p - -------------- -------------- - 2 Number Singular s Plural p - -------------- -------------- - 3 Gender Masculine m Feminine f Neuter n = ============== ============== = Notes: Case is not relevant for English. Gender is probably unnecessary for most purposes. We can assume it may be of interest in constructions with pronouns. Many nouns are (or may be) unmarked for number: fish, sheep, aircraft. Examples: Ncsn house Ncpn houses Npsn Thames Nppn Alps Ncpf women Ncsm man Nc=n sheep 5.5.2 Verbs (V) --------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Type Main v Auxiliary a Modal m - -------------- -------------- - 2 Form Indicative i Form, not Mood Imperative m Subjunctive s Base b base, not infinitive Past Prt p Past Prt, not participle Present Prt g Present Prt added (not gerund) - -------------- -------------- - 3 Tense Present s Past d - -------------- -------------- - 4 Number Singular s Plural p - -------------- -------------- - 5 Person First 1 Second 2 Third 3 = ============== ============== = Notes: Voice is not lexical. Attributes have been reordered to minimize sequence lengths (assuming the proposal about trailing "-"s) - Tense only applies to Finite verbs, Number only to Present, and Person only to Singular. Modals have been included as a distinct subcategory. Finiteness as an attribute is redundant - predictable from verb-form and past/present. These tags do not attempt to represent distinctions found in the various compound verb-forms. These are composed of a sequence of auxiliary and non-finite verb as follows: future will/shall + base conditional would/should + base passive be + past participle perfect have + past participle past perfect have-past + past participle present continuous be + present participle infinitive to + base So there is no "aspect" attribute or "future" value, for example. Examples: go Vvm, Vvb, Vvs, Vvisp, Vviss1, Vviss2 goes Vviss3 going Vvg gone Vvp went Vvid have Vab, Vas, Vaip, Vaiss1, Vaiss2 has Vaiss3 had Vap Vaid having Vag be Vab, Vam am Vaiss1 are Vaiss2, Vaisp is Vaiss3 was Vaids1, Vaids3 were Vaids2, Vaidp been Vap being Vag will Vmi would Vmi
5.5.3 Adjectives (A) -------------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Degree Positive p Comparative c Superlative s - -------------- -------------- - 2 Position Attributive a Position added Predicative p = ============== ============== = Notes: Gender, number and case irrelevant for English. Attributive/predicative distinction reflects positional constraints - some adjectives ('mere', 'utter', etc.) only appear in prenominal position, while others ('awake', 'devoid', etc.) only appear in predicative position. Most can appear in either. Since many English comparatives & superlatives are formed with more/most, "positive" cannot be interpreted as "neither comparative nor superlative". See "Adverbs". Examples: big Ap bigger Ac biggest As more peculiar Dscn+Ap most remarkable Dssn+A awake App mere Apa 5.5.4 Pronoun (P) ----------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Pron.-Type General g General added Demonstrative d Possessive s Personal p Reflexive x - -------------- -------------- - 2 WH Not-WH n WH added Relative r Int q - -------------- -------------- - 3 Number Singular s Plural p - -------------- -------------- - 4 Person First 1 Second 2 Third 3 - -------------- -------------- - 5 Gender Masculine m Feminine f Neuter n - -------------- -------------- - 6 Case Nominative n Accusative a - -------------- -------------- - 7 Poss-Number Singular s Plural p - -------------- -------------- - 8 Poss-Person First 1 Poss-Person added Second 2 Third 3 - -------------- -------------- - 9 Poss-Gender Masculine m Poss-Gender added Feminine f Neuter n = ============== ============== =
Notes: "General" pronouns are those which are not personal, possessive, demonstrative or reflexive. The choice of these four categories is based on distributional facts, though at a rather high level of abstraction. They enter into anaphoric dependencies which are signalled morphosyntactically and are therefore (in principle) more amenable to automatic detection. Most general pronouns do not, although they too sometimes encode number information. "WH" attribute added to allow for combination of possessive and WH in "whose". Examples: Pgn some, all, ... Pgns each, something, nothing, everything, -body, ... Pgnp both, ... Pdns this, that Pdnp these, those Psn---s1 mine Psn----2 yours Psn---s3m his Psn---s3f hers Psn---s3n its Psn---p1 ours Psn---p3 theirs Psq whose Ppns1-n I Ppn-2 you Ppns3mn he Ppns3fn she Ppns3n it Ppnp1-n we Ppnp3-n they Ppns1-a me Ppns3ma him Ppns3fa her Ppnp1-a us Ppnp3-a them Prns1 myself Prns2 yourself Prns3m himself Prns3f herself Prns3n itself Prnp1 ourselves Prnp1 yourselves Prnp1 themselves Ppr which Ppq which, what 5.5.5 Articles/Determiners (R) ------------------------------ = ============== ============== = P ATT VAL C = ============== ============== = 1 Type Def-article t Indef-article a Demonstrative d Possessive s General g General added - -------------- -------------- - 2 WH Not-WH n WH added Relative r Int/Excl q - -------------- -------------- - 3 Number Singular s Plural p - -------------- -------------- - 4 Person First 1 Second 2 Third 3 - -------------- -------------- - 5 Gender Masculine m Feminine f Neuter n - -------------- -------------- - 6 Poss-Number Singular s Plural p - -------------- -------------- - 7 Poss-Person First 1 Poss-Person added Second 2 Third 3 - -------------- -------------- - 8 Poss-Gender Masculine m Poss-Gender added Feminine f Neuter n = ============== ============== =
Notes: Case not relevant to English. Definite and indefinite articles represented as values of "Type". All of these have been marked as 3rd person. This is redundant, since determiners are all 3rd person, if anything. Examples: Rtn-3 the Rans3 a/an Rdns3-ns this, that Rdnp3-np these, those Rsn-3-s1 my Rsn-3--2 your Rsn-3-s3m his Rsn-3-s3f her Rsn-3-s3n its Rsn-3-p1 our Rsn-3-p3 their Rsr-3 whose Rsq-3 whose Rgr-3 which Rgq-3 which, what, Rgns3 each, ... Rgnp3 all, both, certain, many, ... Rgn-3 some, ... 5.5.6 Adverbs (D) ----------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Degree Positive p Comparative c Superlative s - -------------- -------------- - 2 Function Specifier s Function added Modifier m - -------------- -------------- - 3 WH Yes q WH added No n = ============== ============== =
Notes: No distinction has been made between different types of "modifier" adverbs ("sentence-modifying", "VP-modifying", etc.), since their distributions overlap considerably. Examples: Dbsn so, too, very, as Dbsq how Dcsn more Dssn most Dbmn quickly, soon, here, then, now Dcmn better, worse Dsmn best, worst Dbmq where, when, how, why 5.5.7 Adpositions (S) --------------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Type Preposition e Postposition o = ============== ============== = Notes: Postpositions are rare in English. "possessive" 's and ' might be considered postpositions, especially if the alternative is to assign them to the unique membership class (where by definition they would be unrelated). Examples: Se in, near, behind,... So notwithstanding, ago 5.5.8 Conjunctions (C) ---------------------- = ============== ============== = P ATT VAL C = ============== ============== = 1 Type Coordinating c Subordinating s - -------------- -------------- - 2 Comp-Type Infinitive i Comp-Type added Finite f - -------------- -------------- - 3 Coord-Position Initial i Coord-Position added Non-initial n = ============== ============== =
Notes: Subordinating conjunctions are often identical to prepositions (before, after, since, ...). "Comp-Type" encodes information about the complement of subordinating conjunctions. Present-participle complements have not been allowed for here; they consist of bare VPs and are often treated as nominal constructions, the words that introduce them (by, after, etc.) being classified as prepositions. "Coord-Position" encodes the distinction between elements of a discontinuous coordination. "Initial" conjunctions are those that appear before the first conjunct, and "Non-initial" conjunctions are those that appear elsewhere. There is a dependency between initial and non-initial conjunctions ('both...and' but not 'either...and') which is not expressed in these attributes. Examples: Ccn and, or, but Cci either, neither, both Csi for Csf that, because, ... 5.5.9 Numerals (M) ------------------ = ============== ============== = P ATT VAL C = ============== ============== = 1 Type Cardinal c Ordinal o = ============== ============== = Notes: "Function" depends on syntactic context. These have not been subsumed under adjectives, pronouns, determiners, etc. because the internal structure of complex numerals is idiosyncratic. Examples: Mc six Mo sixth