1

Hi I'm working on a software project for work that inflects english words into their various derived forms. e.g. work (verb) -> works, working, worked. My main problem at the moment is that I need to standardize some naming conventions or categories for each inflection type in my program, and then funnel scraped data from across the internet into these categories.

For nouns it was fairly easy since there is just plural and possessive (correct me at any point if there is an error).

For adjectives I have base form, superlative, and comparative.

For verbs the situation is more complicated. I have a mood -> tense -> person -> number hierarchy currently that was brought over from the Italian language system. I want to be clear that I do not need a category for every possible combination, and I do not need separate categories for conjugations that use auxiliary verbs, only those which inflect the verb's form. I want a minimum set of categories that will fully describe all possible flexed regular & irregular verb forms. For instance from what I can tell https://en.wiktionary.org/wiki/be#Conjugation "be" is the most irregular verb and has 8 different forms, so ideally I'd like to have at most 8 categories.

At the moment, when I say "category" I mean a single combination of multiple "aspects". Sorry if I am using jargon loosely or improperly, I'm learning things as I do research for this project.

So far I have the following jumble of aspects:

Moods: indicative, imperative, subjunctive, infinitive, participle

Tenses: present, past, preterite

Persons: first person, second person, third person

Number: singular, plural

An example of a category might be indicative, present, third person, singular for work -> works.

I do not need to keep this hierarchy, I can use any flat or nested structure necessary. However, I need to know how a standard conjugation table (from wiktionary for example) might map onto it.

Currently I'm most worried about moods and tenses. For instance is the preterite identical to the past form? I understand that it's used to describe a different tense but it seems like the base inflection is the same. Can I get rid of one, and if so which?

The moods were mostly just copied and pasted from Italian. Can I get rid of the imperative? Does it have the same inflection pattern (i.e. none at all) as the infinitive?

Thanks for any replies

Milo
  • 121
  • Possible duplicate: http://english.stackexchange.com/q/106629/24489. I disagree with Cascabel: with all verbs the imperative and infinitive forms are the same (to is just a 'marker' employed in some circumstances but not a piece of the form). The 'subjunctive' does not involve any additional forms; rather, it restricts the use of other forms. BE is anomalous, the only verb with an infinitive/imperative/'subjunctive' distinct from the base form and a distinct 1st person singular present form; and modal verbs are 'defective', with no participle or infinitive/imperative/'subjunctive' forms. – StoneyB on hiatus Mar 17 '17 at 18:39
  • 1
    You don't need past and preterite, just the latter will do. Re moods: Imperative and subjunctive are types of clause construction, not mood. Few talk of indicative nowadays, the term does no work and is not needed. Infinitive and participle are types of non-finite verb, not moods. – BillJ Mar 17 '17 at 18:47
  • @Cascabel Yes I'm trying to narrow down a complicated subject I know little about into something I can work with. Broadly speaking I need categories for verb inflections, and I need to be sure these categories capture all possible flexed forms for all verbs. – Milo Mar 17 '17 at 18:47
  • @StoneyB that link is useful. There it seems they use present, past, infinitive, past participle, and present participle. I could restructure the hierarchy to have [Present/Past -> Person -> Number], [Infinitive], [Past Participle], and [Present Participle] and combine moods and tenses into 1 aspect – Milo Mar 17 '17 at 18:56
  • Some POS-tagsets go into great detail. But be aware that there is a lack of agreement on how many parts of speech there are, never mind subclassifications and terminology. – Edwin Ashworth Mar 17 '17 at 19:02
  • @Milo. My biggest advice to you is not to listen to grammarians. This is something I've come to realize over the past 2-3 years I've been working in computational linguistics. Most of grammar is not actually understood, and they don't actually agree. If you try to program what grammarians tell you, you'll get no where. – William Mar 17 '17 at 19:37
  • 1
    @William Or nowhere. – deadrat Mar 17 '17 at 20:34

1 Answers1

2

I'm also a programmer that works in computation linguistics and have worked on this problem before.

Verbs in English only inflect for the following parameters:

non-finite forms: bare infinitive (base form), present participle, past participle

Person: first, second and third

Number: singular, plural

Tense: present, past

Mood: indicative, subjunctive, imperative

The bare infinitive and the participles are not moods but non-finite forms, that is to say, they do not function as verbs at all but belong to different word classes. This is an important distinction to pay close attention to - words also inflect to become different classes of words with different grammatical functions.

Other parameters, such as grammatical Aspect, Voice, and the future tense are constructed with auxiliaries.

William
  • 1,728
  • Probably you can do without mood, too. Imperative constructions are specific, and so are the two minor complement variations that are called "subjunctive" in traditional grammars. Irrealis or hypothesis is also mostly constructed with auxiliaries like modals. – John Lawler Mar 17 '17 at 19:40
  • @JohnLawler English has a morphological subjunctive and an imperative that are distinct from the indicative. The forms are mostly the same as the infinitive in regular verbs, but they differ enough that the parameter "Mood" is necessary. – William Mar 17 '17 at 19:45
  • It may be necessary for some, depending on how their model is coded. – John Lawler Mar 17 '17 at 19:50
  • @ John Lawler If you ignore Mood just because the forms are mostly the same, you'll create thousands of exceptions in all of the situations where they are distinct. – William Mar 17 '17 at 20:52
  • Again, that depends on how many other presuppositions you want to keep in your model. – John Lawler Mar 17 '17 at 20:56
  • 'The bare infinitive and the participles are ... non-finite forms, that is to say, they do not function as verbs at all but belong to different word classes.' is adopting a particular grammatical standpoint (see Aarts on gradience), so do you recommend we ignore what you have written? – Edwin Ashworth Mar 18 '17 at 08:50
  • There is no indicative, subjunctive or imperative mood in English today. "Indicative" is meaningless and the term does no work. Subjunctive and Imperative are not moods but types of clause construction headed by a plain form verb. Mood in Present-day English is analytic, rather than inflectional, and is marked by the presence or absence of the modal auxiliary verbs. – BillJ Mar 20 '17 at 11:26
  • @BillJ That's plainly not true. Verbs in all three moods are distinct. You cannot produce English if you ignore it. If you do, youll just create irregularities in all the places where the inflections are distinct from the indicative. – William Mar 20 '17 at 15:28
  • Can you give me an example of what you consider to be an inflectional mood form. – BillJ Mar 20 '17 at 15:31
  • For example: "I insisted that he eat dinner with us" in the subjuctive has different meaning than saying "I insisted that he eats dinner with us." The former is a demand indicated by the subjuctive and the latter example in the indicative is asserting that its true. – William Mar 20 '17 at 15:43
  • @BillJ or consider the follwing: "I demand that you are there tommorow" What's wrong with that sentence? That sounds horrible. The verb inflection is wrong. Its in the indicative when it should be in the subjunctive. – William Mar 20 '17 at 15:48
  • The difference is that you are quoting the trad approach, whereas many grammarians today treat the subjunctive not as a mood form, but as a type of clause construction headed by a plain form verb. The modern approach claims that historical change has more or less eliminated mood from the inflectional system so the main mood system is now analytic rather then inflectional. So, yes, the subjunctive is alive and kicking, but is called a kind of "clause construction" rather than a mood form. – BillJ Mar 20 '17 at 16:02
  • If you take a look at this link, you'll see a brief discussion of mood: link – BillJ Mar 20 '17 at 16:18
  • @BillJ I just gave two examples where the clause structure is the same but the form of the verb changes. Its easier to model that as a different verb inflection that carries a different meaning than as an irregular set phrase , idiom, or "analytic construction". It makes English more regular, which is easier to program. – William Mar 20 '17 at 16:18
  • The crucial point about the subjunctive is that there's no verb form reserved for it, not for any verb. It's just a content clause in which the verb is required to be in the plain form and the subordinator is "that", but no verb has a subjunctive part of its paradigm, not even the copula. Which is why it's best seen as a clause construction, not a mood. The other two constructions that require a plain form verb are the Imperative and the infinitival. – BillJ Mar 20 '17 at 16:40