Style=Call or something similar for INTJ specifically to get someone's attention

[AngledLuffa]

I see a few annotations for types of INTJ, especially Style=Expr for hmm although I think it's a little ill defined

One type we were looking to add to INTJ for Sindhi is that of getting a specific person's attention, such as Hey_INTJ AngledLuffa, ... There is no existing Style= feature for such an INTJ, and I don't see any other feature labels which would fill that role. @muteeurahman

Does this seem like a useful addition? Would it be something applicable to other languages as well? I personally could see it having a place in English and would be able to help annotate it there

[nschneid]

Is there a morphological class that has that meaning? Or is it just part of the lexical meaning?

I have always understood Style to be for non-default forms of words, where an alternate spelling indicates something about the speaker's attitude. I would say "hmm" is the default spelling (and lemma) and "hmmm" etc. is a lengthened form that could indicate expressiveness.

[nschneid]

(Opened UniversalDependencies/UD_English-EWT#554 regarding possible overuse of Style in English corpora)

[AngledLuffa]

Ok, that's fair. So your suggestion would be to simply not have any feature that indicates that? Or we could add an IntjType which may or may not cover the "hey you" usage in other languages as well

[dan-zeman]

In one of the Indian langs I've seen Case=Voc. It sounds a bit odd to me but I think in these languages vocative is traditionally recognized since Sanskrit, and it's often shown with an interjection (possibly in addition to a nominal suffix). I'm not sure I support it on INTJ though...

[nschneid]

If adding an IntjType feature, what would the other values be? Would the goal be to partition the set of all INTJ lexemes?

[amir-zeldes]

I agree this isn't so much a style as a function, and normally this would be the child of a deprel vocative token (though if the vocative is the root, that is no longer visible in the deprel). Making some kind of IntjType is an option, but I agree with Nathan, that would suggest looking at all INTJ types and making some kind of taxonomy.

[dan-zeman]

normally this would be the child of a deprel vocative token

I don't think so. If the sentence is something like

• Hey Joe, come over here

then we will have vocative(come, Joe), not vocative(come, Hey). I would either do discourse(come, Hey) as with other interjections, or, if we go for the interpretation that Hey brings the Case=Voc feature, we could do case(Joe, Hey). I now realize why it sounds familiar to me. I was recently revisiting the validation tests for DEPREL-UPOS compatibility, and I saw that INTJ was allowed for case, with this comment:

Interjection can also act as case marker for vocative, as in Sanskrit: भोः भगवन् / bhoḥ bhagavan / oh sir.

[AngledLuffa]

Excellent, thanks for the discussion. We'll remove the feature from the proposed Sindhi dataset.

[amir-zeldes]

then we will have vocative(come, Joe), not vocative(come, Hey).

I think my comment may have been hard to parse - by "this would be the child of a deprel vocative token" I meant that "hey" might be the child of a (different) token, whose deprel from its parent (the grandparent) is "vocative" (in your example, that grandparent is "come"). So I think we are saying the same thing 😅

[rueter]

normally this would be the child of a deprel vocative token

I don't think so. If the sentence is something like

* _Hey Joe, come over here_

then we will have vocative(come, Joe), not vocative(come, Hey). I would either do discourse(come, Hey) as with other interjections, or, if we go for the interpretation that Hey brings the Case=Voc feature, we could do case(Joe, Hey). I now realize why it sounds familiar to me. I was recently revisiting the validation tests for DEPREL-UPOS compatibility, and I saw that INTJ was allowed for case, with this comment:

Interjection can also act as case marker for vocative, as in Sanskrit: भोः भगवन् / bhoḥ bhagavan / oh sir.

In the definition of case, https://universaldependencies.org/u/dep/all.html#al-u-dep/case, it is stressed that Case-marking elements are treated as dependents of the noun they attach to or introduce. (Thus, contrary to SD, UD abandons treating a preposition as a mediator between a modified word and its object.) The case relation aims at providing a more uniform analysis of nominal elements, prepositions and case in morphologically rich languages

In light of this we can re-examine Hey, Joe, come over here
The word “Hey” is used to capture someone's attention -- at least in the Salish Sea region. More specifically, it is used to capture the attention of “Joe”, whom we subsequently tell to move himself in the direction of the speaker.

(1) Let us assume that the use of commas is a matter tradition in English punctuation rules, where the addressee is set off with commas;
(2) There is no other indication in the written source that the proper noun «Joe» has undergone form and become a vocative;
(3) The word «Hey», however, is collocated with the second person, addressee, or addressee-candidate, but not the first or third person, e.g., *hey, me..., *hey, us..., *hey, him..., *hey, her..., *hey, it..., *hey, them...

This function of activating or eliciting “addresseehood” would seem to parallel the function referenced by the term “vocative”.

One might consider the similarity of words, such as “hello” or “dear”.

• These two words, however, are salutary, whereas “Hey” is not.
• The words “hello” and “dear” can appear with first, second and third person, whereas the word “hey” avoids first and third person.
• The word “hey”, without an accompanying 2nd person pronoun or name, can also be use for the vocative function, which might be an instance of ellipsis -- vocative(come, Hey).

For these reasons, I would be inclined to align myself with the practice in Sanskrit.
vocative(come, Joe)
case(Joe, Hey)

Note: The Sanskrit, above, has a translation ‘oh, Sir’. For me, the use of the word “oh” introduces a falling tone in the word “Sir”, which is pretty much identical to the falling tone used when calling to someone way of in distance. The word “hey” does not seem to allow for a collocated tone like this.
It also seems that the word “oh” might collate with first, second and third person pronouns.

Hey, @Stormur, @ftyers, @jonorthwash ,@flammie, @OliverHellwig, @dan-zeman, @jasiewert, @garanes, @nikopartanen is this food for thought with regard to the vocative, or has it simply been easier to ignore the vocative?

[dan-zeman]

is this food for thought with regard to the vocative, or has it simply been easier to ignore the vocative?

It has been easier to ignore it, except for languages where it is a morphological case (because even in these languages its function seems to be very different from all other cases).

[amir-zeldes]

has it simply been easier to ignore the vocative?

I don't think we're ignoring it, it has a deprel after all. It's just not a case as such in English, because no word has a unique form for it (but pronouns do distinguish Nom vs. Acc vs. Gen in English), and because unlike normal case dependents in English (i.e. prepositions), words like "hey" can appear by themselves. Compare:

• Hey!
• * among

In other words, "hey" does not follow the valency structure of English case dependents, and seems to be a different word class, probably more like sentential adverbials or interjections, which is why it has the deprel discourse in the English TBs.

[Stormur]

Yes, probably the most telling diagnostics is that hey can be used independently, i.e. it is not bound (a clitic, in Haspelmathian terms), and no real ellipsis can be assumed. That is also why it is classified as INTJ, it is more of a cry than an actual lexical element.

In general, I would not put Case on adpositions or similar elements anyway if there is no change in the form (of the adposition). The function of the phrase is already managed by the relation it receives, in this case vocative. For example, in Latin we consider o (which is different from a surprise oh, compare modern Italian where the opening and length of the vowel change) a particle introducing a vocative, but we only label it with PartType=Emp(hatic), which is a lexical, not a morphosyntactic distinction.