Property talk:P9295

Documentation

transitivity
possible grammatical property of a verb accepting a object complement

Represents

Data type

Allowed values

pseudo-transitive verb (Q105933250), ergative verb (Q623554), transitive (Q116946936), intransitive (Q116946937), monotransitive (Q116946939), ditransitive (Q116946938), ambitransitive (Q116952633) or tritransitive (Q116953365)

Example

no label (L17690) → ergative verb (Q623554)
no label (L707490) → intransitive (Q116946937)
no label (L412870-S1) → transitive (Q116946936)
no label (L412870-S2) → intransitive (Q116946937)
no label (L309-S1) → transitive (Q116946936)

Which level?

Hi y'all,

Following @عُثمان: comment on Wikidata_talk:Lexicographical_data#Usage_examples_on_senses « for example on lexemes like vergessen (L412870) we might expect transitivity (P9295) to be on senses, while on lexemes like ਲੌਟਣ/لَوٹݨ (L1096159) we might expect to find this property only once at the top of the lexeme. »

I would like to discuss that point. For me the current placement on sense level on vergessen (L412870) seems a bit strange; for consistency (and easier adding, reusing, maintaining), I would rather put it on the lexeme level with qualifiers, in clear, replace : L412870-S1transitivity (P9295)transitive (Q116946936) and L412870-S2transitivity (P9295)intransitive (Q116946937) by vergessen (L412870)transitivity (P9295)transitive (Q116946936)subject sense (P6072)L412870-S1 vergessen (L412870)transitivity (P9295)intransitive (Q116946937)subject sense (P6072)L412870-S2 (@Nikki: what do you think?).

For the record, right now this property is used 571 times, 484 times on lexeme level and 87 on sense level : https://w.wiki/6e2n

Pinging the top user of this property @EconAmbiente, Vis M, Jon Harald Søby, Lepticed7, Balû: (without عُثمان the top one user but already pinged ). What do you all think ?

Cheers, VIGNERON (talk) 08:28, 29 April 2023 (UTC)[reply]

Hi. I think the important point is to ask ourselves if the property is added at lexeme level, what would it mean if it doesn’t have any qualifiers? Is the qualifier missing (and only some sense are concerned) or does the transitive property apply to all senses? Also, if the property is only for sense level, I think (I’m not sure though) that some values don’t have any meaning. For example, ergative verb (Q623554) looks only usable at lexeme level. Cheers, Lepticed7 (talk) 11:33, 29 April 2023 (UTC)[reply]
Full disclosure, when I started using this property and valency (P5526), there were not many uses of them to begin with, so they way I have been using them is not indicative of any precedent.

In Punjabi, intransitive and transitive verbs take different inflections and have different phonological properties. So it is a semantic, grammatical, and morphological property of the lexeme all at once. It is possible to deduce the way tone changes in the conjugation of a verb if we set them up this way without having to annotate each individual form.

Most Punjabi intransitive verbs can be made to take additional participants in the predicate as a feature of the inflectional paradigm, and can have them subtracted as well. However there is a "base" transitivity from which all these senses and forms relate to.

ਲੱਦਦਾ

laddadā - base form, intransitive, 1 participant/thematic relation

ਲਦਾਉਂਦਾ

ladaundā - additive form, transitive, 2 participants

ਲੱਦੀਦਾ

laddīdā - subtractive form, avalent, 0 participants

ਲਦਾਇਦਾ

ladāīdā - subtractive of additive form, intransitive, 1 participant

The fourth form here has the same valency as the first one, but they aren't used the same way and don't mean the same thing. When we consider these verbs as a whole they form one pattern or another depending on if the "base" form is intransitive or transitive.

All that is to say, I don't think German verbs work like this at all. I have taken steps to model verbs in Hindustani, Sindhi, Saraiki, and Hindko in the same way as Punjabi as their verbal systems form a continuum—it is desirable to have consistency where the languages are expected to be consistent with each other, but I don't think it makes sense to expect to be able to have conformity between German and Punjabi on the use of the transitivity property anymore than it would to start adding "weak verb" and "strong verb" to Punjabi ones.

I do not understand however why passive voice (P5560) is German-only. Punjabi could use this عُثمان (talk) 16:07, 29 April 2023 (UTC)[reply]

Hello VIGNERON , As the proponent of the property, I would like to know what is "a bit strange" according to you? That's my main question, because it's not clear. You cannot give a comparison of an amount of use in one type of entity with another type of entity, when the property gives the possibility of using it in these two types of entity (Sense and Lexeme).

(The use is certainly less present for the senses in French and @عُثمان: made intensive use of it in Lexeme. This is not conclusive and it depends on the language.)

Let's be clear, I am not against your proposal, but I would like to have your expert opinion on SPARQL. I think that if we must "simply" bring out the transitivity in a lexeme, we only get one result (without knowing that there are several transitivities). In my opinion, if the information is applicable for a sense, I think it makes sense to have it where transitivity applies, it's up to us to choose the right type of entity.

Hence my first question: strange? This information seems to me to be in the right place for vergessen (L412870). What are the issues affecting "adding, reusing, maintaining"? Cordially. @ArthurPSmith, Pamputt, Fnielsen: ―Eihel (talk) 16:52, 29 April 2023 (UTC)[reply]

Hello VIGNERON I see, what you mean: a sense couldn't be transitive, only a lexeme - but with different senses the lexeme could have different transistivity modes. After thinking about it, I share your opinion to put transitivity to the lexeme level. --Balû (talk) 06:00, 30 April 2023 (UTC)[reply]

I'm trying to answering everything (sorry if I forgot something).

@Lepticed7: very good point. If there is several senses (which is rarely the case right now, in most languages the languages is even below 1) and no qualifier, "apply to all sense" and "apply to some unspecified senses" are equally possible, I would tend to think that the first is more probable but there is an uncertainty, maybe there should always be qualifiers (if possible enforced with a complex constraint) but how often does a verb have multiple senses with differents transitivity ? Indeed obviously ergative verb (Q623554) can't go to the lexeme level (and it's not used like that right now, but if we stay with the current double model, then we should set up a constraint to check that).

@عُثمان: interresting, if I understand correctly, you have no reason to use it on the sens level? And indeed there is no reason for passive voice (P5560) to be limited to German.

@Eihel: strange precisely because of this possibility of using it on two levels, and even more strange that no clear rule for wich level to use in which case. It's basic quality process. Plus, I've look at some references, transitivity is almost always describe as a feature of verbs, not of senses. Then, if it's language dependant, why is it also used on two different level inside one language ? (French and Malayalam for instance). In SPARQL (and other query languages), it's way harder (and due to timeout, maybe often impossible) to make a query if you don't know where to look exactly (I'm not even sure how to do it bast). On the other hand, going through multiple values and look at the qualifier is pretty easy (for instance with COUNT: https://w.wiki/6eEu). For me, if the information is applicable for a sense, then if should be indicated at the lexeme level with the qualifier subject sense (P6072) « statement applies only to listed senses of statement's subject ». Especially as in most case (see my answer to Lepticed7 above), this qualifier may not be needed.

@Balû: thanks.

Cheers, VIGNERON (talk) 12:24, 30 April 2023 (UTC)[reply]

@VIGNERON Yes, I have no reason to use it on senses.

I would actually think English is a language for which transitivity may be better considered a property of senses. For example, the verb "rain" has a primary instransitive sense ("it rains") but it also has a figurative transitive sense ("it's raining cats and dogs"). A lot of English verbs are like this and it has no bearing on the morphology of the verb itself.

I do agree that statements within the same language should be consistent. However, I am less sure it is necessary for languages to conform with others given how differently verbs may function between them. عُثمان (talk) 19:37, 30 April 2023 (UTC)[reply]

Support ―Eihel (talk) 14:03, 30 April 2023 (UTC)[reply]

Property talk:P9295

Documentation

Which level?

Navigation menu

Search