This document describes the lexicon format for ACE 6.7.
Lexicons can be used in two ways. They can be used as a Clex ("common lexicon" or "compiled lexicon") which is compiled into the APE executable. Alternatively, they can be used as a Ulex ("user lexicon" or "uncompiled lexicon") that is loaded at run-time using the command line, the web interface, the web service, or any other interface to the parser.
User lexicons Ulex can be used independently, or can complement Clex in which case entries of Ulex take precedence over those of Clex.
The ACE lexicon describes English adverbs, adjectives, nouns, verbs, and prepositions. It does not describe function words like pronouns or quantifiers. There are 11 different word classes: adverbs, intransitive adjectives, transitive adjectives, countable nouns, mass nouns, measurement nouns, proper names, intransitive verbs, transitive verbs, ditransitive verbs, and prepositions.
The different forms of the words are represented by different Prolog facts. Each fact
minimally contains one entry that defines the word form which will appear in the ACE text, and
one entry that defines the symbol how it will appear in the logical representation. All entries
must be Prolog atoms. Furthermore, the entry for the word form (that will appear in the ACE text)
must contain nothing but lower and upper case letters (a-z,
A-Z), digits (0-9), hyphens (-), underscores (_),
dollar signs ($), and degree signs (°). The first character must not be
a digit or a hyphen. Blank spaces are not allowed. You should use hyphens in places where blank spaces
would be used in natural English. E.g. write 'credit-card' instead of 'credit card'.
Because the lexicon is written in Prolog, symbols special for Prolog (e.g. apostrophe) must be escaped. Capitalized words (e.g. proper names) must be in quotes, otherwise they would be considered variables by Prolog.
Adverbs (e.g. "fast", "quickly", "intensively") are represented by three different kinds of entries, defining the positive form, the comparative form, and the superlative form respectively:
adv(PositiveForm, LogicalSymbol). adv_comp(ComparativeForm, LogicalSymbol). adv_sup(SuperlativeForm, LogicalSymbol).
For regular comparative and superlative forms (starting with "more" or "most"), only the positive form needs to be defined. In order to use "more quickly", for example, it is sufficient to define the positive form "quickly". Some concrete examples:
adv(fast, fast). adv_comp(faster, fast). adv_sup(fastest, fast). adv(quickly, quickly). adv(intensively, intensively).
It is recommended to use the positive form also as the logical symbol. Furthermore, it is recommended to define a positive form for each comparative or superlative form.
Intransitive adjectives (e.g. "important", "large", "expensive") are represented by three different kinds of entries, defining the positive form, the comparative form, and the superlative form respectively:
adj_itr(PositiveForm, LogicalSymbol). adj_itr_comp(ComparativeForm, LogicalSymbol). adj_itr_sup(SuperlativeForm, LogicalSymbol).
Again, for regular comparative and superlative forms (starting with "more" or "most"), only the positive form needs to be defined. Some concrete examples:
adj_itr(important, important). adj_itr(large, large). adj_itr_comp(larger, large). adj_itr_sup(largest, large). adj_itr(expensive, expensive).
It is recommended to use the positive form also as the logical symbol. Furthermore, it is recommended to define a positive form for each comparative or superlative form.
Transitive adjectives (e.g. "valid-for", "fond-of", "pessimistic-about") are represented by three different kinds of entries, defining the positive form, the comparative form, and the superlative form respectively:
adj_tr(PositiveForm, LogicalSymbol, Preposition). adj_tr_comp(ComparativeForm, LogicalSymbol, Preposition). adj_tr_sup(SuperlativeForm, LogicalSymbol, Preposition).
Transitive adjectives should consist of an adjective and a preposition that is hyphenated to the adjective. Again, for regular comparative and superlative forms (starting with "more" or "most"), only the positive form needs to be defined. Preposition is the preposition part of the word, e.g. "for" in the case of "valid-for", "of" in the case of "fond-of", and "about" in the case of "pessimistic-about". It is only used for the comparative form in sentences where the preposition can occur on its own, like in "Mary is more fond-of Bill than of John". Some concrete examples:
adj_tr('valid-for', 'valid-for', for).
adj_tr('fond-of', 'fond-of', of).
adj_tr_comp('fonder-of', 'fond-of', of).
adj_tr_sup('fondest-of', 'fond-of', of).
adj_tr('pessimistic-about', 'pessimistic-about', about).
It is recommended to use the positive form also as the logical symbol. Furthermore, it is recommended to define a positive form for each comparative or superlative form.
Countable nouns (e.g. "woman", "credit-card", "month") are represented by two different kinds of entries, defining the singular form and the plural form respectively:
noun_sg(SingularForm, LogicalSymbol, Gender). noun_pl(PluralForm, LogicalSymbol, Gender).
Defining a countable noun "credit-card", for example, you can say "a credit-card", "every credit-card", or "at most 5 credit-cards". The plural form should usually be the singular form with an 's' at the end like "credit-cards" or "months". Some concrete examples:
noun_sg(woman, woman, fem).
noun_pl(women, woman, fem).
noun_sg('credit-card', 'credit-card', neutr).
noun_pl('credit-cards', 'credit-card', neutr).
noun_sg(month, month, neutr).
noun_pl(months, month, neutr).
Gender has to be one of {undef, neutr, human, masc, fem} and specifies the linguistic gender of the noun. undef is the top category having two direct subcategories neutr and human where human has again two subcategories masc and fem. Depending on the gender, different forms of pronouns are supported. The table below shows the concrete pronouns. Note that not only the pronouns listed for a particular gender but also all pronouns of all super- and subcategories can be used.
| Gender | Anaphoric Pronouns | Possessive Pronouns | Relative Pronouns |
| undef | they, them, themselves | their, their own | that, whose |
| neutr | it, itself | its, its own | which |
| human | he/she, she/he, him/her, her/him, himself/herself, herself/himself | his/her, her/his, his/her own, her/his own | who |
| masc | he, him, himself | his, his own | |
| fem | she, her, herself | her, her own |
It is recommended to use the singular form also as the logical symbol. Furthermore, it is recommended to define a singular form for each plural form, and vice versa.
Countable nouns (e.g. "water", "fear", "money") are represented by only one kind of entry:
noun_mass(WordForm, LogicalSymbol, Gender).
Mass nouns have only one word form. Defining a mass noun "water", for example, you can say "some water", "all water", or "5 kg of water". The gender of mass nouns is defined in the same way as for countable nouns (see previous section). Some concrete examples:
noun_mass(water, water, neutr). noun_mass(fear, fear, neutr). noun_mass(money, money, neutr).
It is recommended to use the word form also as the logical symbol.
Measurement nouns (e.g. "kg", "m", "°C") are represented by only one kind of entry:
mn_sg(SingularForm, LogicalSymbol). mn_pl(PluralForm, LogicalSymbol).
It is recommended to define measurement nouns in the form of measurement symbols (i.e. "m" instead of "meter"). In this case, the singular and plural forms should usually be the same. Some concrete examples:
mn_sg(kg, kg).
mn_pl(kg, kg).
mn_sg(m, m).
mn_pl(m, m).
mn_sg('°C', '°C').
mn_pl('°C', '°C').
It is recommended to use the singular form also as the logical symbol. Furthermore, it is recommended to define a singular form for each plural form, and vice versa.
Proper names come in two versions: one for those that require a preceeding definite article (e.g. "the Nile", "the United-Nations") and one for those that stand by itself (e.g. "John", "Nokia"). Proper names can also be declared in both of those categories (e.g. "(the) Mona-Lisa").
pn_sg(WordForm, LogicalSymbol, Gender). pn_pl(WordForm, LogicalSymbol, Gender). pndef_sg(WordForm, LogicalSymbol, Gender). pndef_pl(WordForm, LogicalSymbol, Gender).
Gender has the same meaning as for countable nouns. Some concrete examples:
pn_sg('John', 'John', masc).
pn_sg('Nokia', 'Nokia', neutr).
pndef_sg('Nile', 'Nile', neutr).
pndef_pl('United-Nations', 'United-Nations', neutr).
pn_sg('Mona-Lisa', 'Mona-Lisa', neutr).
pndef_sg('Mona-Lisa', 'Mona-Lisa', neutr).
It is not allowed to declare the same proper name as singular and plural at the same time. It is recommended to use the word form also as the logical symbol.
Intransitive verbs (e.g. "waits", "goes-away", "walks") are represented by two different kinds of entries, defining the third singular form and the infinitive form respectively:
iv_finsg(ThirdSgForm, LogicalSymbol). iv_infpl(InfForm, LogicalSymbol).
The infinitive form should usually be the singular form without the 's' at the end: "wait", "walk". Phrasal particles should be hyphenated to the verb, e.g. "goes-away". Some concrete examples:
iv_finsg(waits, wait).
iv_infpl(wait, wait).
iv_finsg('goes-away', 'go-away').
iv_infpl('go-away', 'go-away').
iv_finsg(walks, walk).
iv_infpl(walk, walk).
It is recommended to use the infinitive form also as the logical symbol. Furthermore, it is recommended to define a third singular form for each infinitive form, and vice versa.
Transitive verbs (e.g. "knows", "likes", "relates-to") are represented by three different kinds of entries, defining the third singular form, the infinitive form, and the past participle form respectively:
tv_finsg(ThirdSgForm, LogicalSymbol). tv_infpl(InfForm, LogicalSymbol). tv_pp(PastPartForm, LogicalSymbol).
Phrasal particles and direct prepositions should be hyphenated to the verb, e.g. "relates-to", "lives-in", and "comes-up-with". The past participle is the form that is used for passive voice: "known", "liked", "seen". If it is not defined then passive voice is not possible for the respective verb. Some concrete examples:
tv_finsg(knows, know).
tv_infpl(know, know).
tv_pp(known, know).
tv_finsg(likes, like).
tv_infpl(like, like).
tv_pp(liked, like).
tv_finsg('relates-to', 'relate-to').
tv_infpl('relate-to', 'relate-to').
tv_pp('related-to', 'relate-to').
It is recommended to use the infinitive form also as the logical symbol. It is also recommended to define a third singular form for each infinitive form, and vice versa. Furthermore, it is recommended to define third singular and infinitive forms for each past participle form.
Diransitive verbs (e.g. "shows", "forgives", "succeeds") are represented by three different kinds of entries, defining the third singular form, the infinitive form, and the past participle form respectively:
dv_finsg(ThirdSgForm, LogicalSymbol, Preposition). dv_infpl(InfForm, LogicalSymbol, Preposition). dv_pp(PastPartForm, LogicalSymbol, Preposition).
ACE allows two forms of ditransitive verbs:
In case a verb allows both forms (such as the verb "show"), each lexicon entry must be declared twice. Some concrete examples:
dv_finsg(shows, show, ''). dv_infpl(show, show, ''). dv_pp(shown, show, ''). dv_finsg(shows, show, to). dv_infpl(show, show, to). dv_pp(shown, show, to). dv_finsg(forgives, forgive, ''). dv_infpl(forgive, forgive, ''). dv_pp(forgiven, forgive, ''). dv_finsg(succeeds, succeed, as). dv_infpl(succeed, succeed, as). dv_pp(succeeded, succeed, as).
It is recommended to use the infinitive form also as the logical symbol. It is also recommended to define a third singular form for each infinitive form, and vice versa. Furthermore, it is recommended to define third singular and infinitive forms for each past participle form.
Prepositions (e.g. "in", "for", "during") are represented by only one kind of entry:
prep(WordForm, LogicalSymbol).
Prepositions have only one word form. Some concrete examples:
prep(in, in). prep(for, for). prep(during, during).
It is recommended to use the word form also as the logical symbol.
In order to define aliases, i.e. word forms that are parsed in exactly the same way, one has to create several entries with the same logical symbol. Some examples:
adj_itr(gray, gray).
adj_itr(grey, gray).
noun_sg(cactus, cactus, neutr).
noun_pl(cacti, cactus, neutr).
noun_pl(cactuses, cactus, neutr).
pndef_pl('United-States', 'USA', neutr).
pndef_pl('USA', 'USA', neutr).
Generally, word class intersections are allowed. E.g. the verb "eats" can be declared as an intransitive verb and as a transitive verb at the same time. But there are some intersections that are not allowed:
The following table illustrates these restrictions. Fields containing an "X" denote intersections that are not allowed. (The gray fields are a mirror image of the white ones.)
In addition, the following function words cannot be used as word forms in the lexicon:
null, zero, one, two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, dozen, there, and, or, not, that, than, of, if, then, such, be, provably, more, most, are, is, the, a, an, some, no, every, all, each, which, its, his, her, their, whose, it, he, she, they, him, them, itself, himself, herself, themselves, someone, somebody, something, nobody, nothing, everyone, everybody, everything, what, who, how, where, when
All other function words (e.g. "can", "at", "true") can be used.
2013-07-31