define SPlural ["+Sing":0 | "+Plural":[%^ s]];This is the component of the pipeline that analyzes Spanish words based on a
number of open-class word lists, present in the directory data/ES/OP/, and
also closed-class but variant words.
The tagset is multivalued, that is, every word can get more than one tag. Every word gets at least one tag, that of the part of speech. Open class words get the lexeme, while closed class words don’t. The extra tags for all words are different morphological pieces of information, encoded by the different morphemes. They are prepended to the part of speech in arbitrary ordering.
Common morpheme for plural with s
define SPlural ["+Sing":0 | "+Plural":[%^ s]];Variant morpheme for gender
define OAGen ["+Masc":o | "+Fem":a];Morpheme for “Masculine by default” words, the feminine gender is the morpheme a
define AGen ["+Masc":0 | "+Fem":[%^ a]];Common morpheme set for Spanish words: gender and number
define GenNum AGen SPlural;Consonants
define Cons [b|c|d|f|g|h|j|k|l|m|n|ñ|p|q|r|s|t|v|w|x|y|z];Vowels
define Vocal [a|e|i|o|u];Determiners, some of them are listed in the closed class dictionaries. Those with morphological information, even though closed class, are analyzed here to extact the corresponding tags. As most function words, they are highly irregular.
Articles can be definite or indefinite, but all have gender and number information.
Definite articles
define Determinado "+Masc":{el} "+Sing":0
| "+Masc":{lo} "+Plural":s
| "+Neut":{lo} "+Sing":0
| "+Fem":{la} SPlural
;Indefinite articles
define Indeterminado "+Masc":{un} ["+Sing":0 | "+Plural":{os}]
| "+Fem":{una} SPlural
;
define Articulo Determinado "+ArtD":0
| Indeterminado "+ArtI":0
;Demonstrative determinants are more regular, but still have the neuter gender inherited from latin and lost elsewhere in the language.
define DDemostrativo 0:{est} [ ["+Masc" "+Sing"]:e
| "+Fem":a SPlural
| "+Neut":o
| "+Masc":o "+Plural":s
] "+Dem":0;Numbers are defined here, and their letter-written ordinal variants are also loaded here, since they can have gender and number inflection.
define Digito [%0|1|2|3|4|5|6|7|8|9];
define Numero [Digito* [(%.|%,) Digito+]*] "+Numeral":0;
define Ordinal {primer} [OAGen SPlural | "+Sing":0]
| {tercer} [OAGen SPlural | "+Sing":0]
| @txt"data/ES/OP/Num.txt" GenNum
;The two contractions of spanish, de+el (del) and a+el (al) are listed here
with the determinants, but they have a special tag for identification.
define Determinante [ Articulo
| Numero
| DDemostrativo
| Ordinal "+Ord":0
] "+Det":0
| ["DE+EL"]:{del}
| ["A+EL"]:{al}
;Pronouns are a very important part of Spanish. They carry gender and number information, and also case information not present in other parts of the language morphology. They are bastardizations from latin pronouns, already irregular, so they are highly irregular.
Personal pronouns can have any of the Nominative, Accusative (ObjD) and Dative (ObjI) cases. They can also be of neuter gender. First and second person singular don’t encode gender.
define Personal ["+1a":0 "+Sing":0 ["+Nom":{yo} | "+ObjD":{mí} | "+ObjI":{me}]]
| ["+2a":0 "+Sing":0 ["+Nom":{tu} | "+ObjD":{tí} | "+ObjI":{te}]]
| ["+3a":0 "+Masc":0 "+Sing":0 [0:{él} | "+Obj":{le} SPlural]]
| ["+3a":0 "+Neut":0 "+Sing":0 [0:{ello} | "+Obj":{lo} SPlural]]
| ["+3a":0 "+Fem":0 [0:{ella} | "+Obj":{la}] SPlural]
| ["+1a":n | "+2a":v] 0:{osotr} OAGen "+Plural":s
| ["+1a":n | "+2a":0] "+Obj":{os}
| ["+3a" "+Masc" "+Plural"]:{ellos}
;Possesive pronouns don’t have gender information for the bound object, but they optionally add it for the modified one.
define Posesivo [ ["+1a" "+Sing"]:{mi} OAGen
| ["+2a" "+Sing"]:{tu} (0:y OAGen)
| "+3a":{su} (0:y OAGen)
| ["+1a":n | "+2a":v] 0:{uestr} OAGen
] ("+PlurObj":s);Demonstrative pronouns are hard to distinguish from their counterpart determinants with spelling only. In the past, they were distinguished by the ortographical accent (tilde) (é), but the Real Academia dropped that rule a few years back.
define PDemostrativo 0:[[é|e] s t] [ ["+Masc" "+Sing"]:e
| "+Fem":a SPlural
| "+Neut":o
| "+Masc":o "+Plural":s
] "+Dem":0;
define Pronombre [Personal "+Pers":0 | Posesivo "+Pos":0 | PDemostrativo] "+Pron":0;Open class variant words exhibit a more regular morphology than other parts of the language. The lexemes are listed in the dictionary, and they can take gender and number information. They are adjectives (Adjetivos) and nouns (Sustantivos).
Adjetivos
define Adjetivo @txt"data/ES/OP/Adj.txt" GenNum "+Adj":0;Sustantivos
define Sustantivo @txt"data/ES/OP/Sust.txt" GenNum "+Sust":0;Verbs in romance languages are quite complex, and in Spanish it is indeed so. They merit their own script, which is sourced here. See its documentation here.
source src/verbs.ES.foma
define Verbo;The lexicon is just the possible words that we are going to recognize, with the corresponding suffixes appended.
define Lexicon Determinante
| Pronombre
| Adjetivo
| Sustantivo
| Verbo
;The lexicon is then passed through ortographic rewrite rules, which take care of morpheme boundary phenomena. These are quite more regular in Spanish than in English. A few are listed here, but a linguist would be required to make an exhaustive list. Do note that verb rewrite rules are apart, in the corresponding script, since they are quite more complex.
regex Lexicon
.o. [..] -> e || Cons _ %^ s
.o. z -> c || _ %^ [e|i]
.o. o -> 0 || _ %^ a
.o. %^ -> 0;Diacritic flags are used in the verb subsystem, and here we compute the equivalent transducer without flags, so that lookup can use it.
eliminate flags