Verbos

This script analyzes Spanish verbs morphologically. Spanish verbs carry person and number information along with tense and mood.

Note that apart from those analyzed here, there are additional possible tenses, which are formed periphrastically and thus cannot be analyzed by this script. They are always formed with the verb haber plus a non-finite form, so they would be easy to detect in a multiword tagger or in the next level, e.g. a syntactical analyzer.

Person and number information is collapsed into the same composite tag, where +1p means first person plural. Preson can be 1, 2 or 3, number singular (s) or plural (p).

The participle form has adjective-like characteristics, so we need the usual gender and number morphemes.

Vowels are an important part of the fusional verb morphology, so we need to identify them.

Lexeme

Here regular lexemes are loaded. However, many verbs in Spanish are irregular in that the root changes with different tenses, while sticking to the normal inflection paradigm (or sometimes jumping paradigm). This “regular irregularity” is encoded with a priority union, which bypasses VLex and gives the correct surface form for a given verb in that tense.

The root form of Spanish verbs is the infinitive. Here it is loaded from the dictionary file. Verbs in Spanish infinitive always end in a, e, i + r.

Flag diacritics are used to find out what paradigm a given verb belongs to. The root form is passed through this transducer, which adds a feature (CONJ) representing what “conjugación” (paradigm) should be used, according to the last vowel in the root.

Though very irregular, personal morphemes can be identified and are defined here. When used, they do not suffer boundary changes, but second person plural may add a “tilde” to the vowel, so it’s marked with a preceding %'.

Non-finite forms

A number of special markers are used, which indicate the type of morphological change that will occur when the suffix is added. A common one is %R, that discards the ‘r’ at the end of the root form.

The infinitive is the dictionary form, so no morpheme.

The gerund is formed with the suffix “ndo”, but transforms ‘i’ and ‘e’ in the last syllable into ‘ie’.

The participle ends in “do”, but ‘e’ transforms into ‘i’. Gender and number information can also be added.

The imperative only exists as a different form for the second person, singular and plural.

R-keeping forms

These forms come from the deformation of the root form + the verb “haber”, so they keep the last ‘r’ and are quite more regular.

The future tense is formed with the present tense of ‘haber’ attached (minus ‘h’).

The conditional is very regular, with a set morpheme plus the personal morpheme

“tener” has a different root in the conditional. Other verbs do too, and they would have to be added here.

Subjunctive past

Pretérito perfecto simple de subjuntivo

It’s a very regular tense, but different in the first “conjugación”. The flag diacritics allow each form only if the correct paradigm is matched.

As before, the verb “dar” has a different form in this tense. In fact, it also jumps paradigm, and we encode that re-setting the flag diacritic. In following tenses the same pattern of defining alternative lexemes is used.

Past continuous

Pretérito imperfecto

Again a regular tense, in the second and third paradigms the last root vowel has to be deleted %V.

Past perfect

Pretérito perfecto

The most irregular tense, it’s regular in its own way. The morphemes are almost the same but with slight variations in the first and other paradigms.

A number of verbs have different roots for this tense.

Present

Presente

The present is regular but a number of morpheme boundary merges have to be recorded.

In the subjunctive mood, the present is even more regular.

Final touches

Spanish verbs can have clitic pronouns affixed, which are appended here and noted with a +Clit tag.

Regular verb inflection is defined in this regex. First comes the root (VLex), which can be bypassed by specific roots for specific tenses for the “regular irregular” verbs.

Here completely irregular forms are listed, and bypass directly all regular inflection. There are a lot of examples in the language, and only a few are listed here. Adding more would just mean appending them to the list.

The final regex is composed here. Irregular bypasses with the priority union (.P.) the regular inflection, clitics are optionally added, and the part-of-speech tag is included. Finally, the rewrite rules process the special symbols that different suffixes way have included.

%R deletes an ‘r’.

%V deletes a vowel.

%IE turns i into e.

%EI turns e into i.

%EIIE turns e or i into ‘ie’.

%' adds a tilde to the vowel.

Unused symbols are cleaned up.