Zompist Bboard Again

Are there any natlangs where an auxiliary verb is uninflected while the main verb keeps its usual inflection?
As an example of what I'm thinking, from my conlang:

Perfective:
neam-ete tsu mièr.
[gloss]break-PFV 1sg bowl[/gloss]
"I broke the bowl."

Anticausative perfective:
keham neam-ete tsu mièr.
[gloss]be break-PFV bowl[/gloss]
"The bowl broke."

Here keham is an auxiliary verb marking the anticausative, but neam maintains the perfective suffix -ete.
I imagine it's much more common to have something like keham-ete neam, but are there any natlangs with the above structure, and if so, do we know anything about how that structure arose historically?

Question is why you want to call this an auxiliary if it doesn't show any verbal behaviour? Normally this would be called a grammatical marker / particle / modifier / adverb. Historically, these can develop from auxiliaries, like the future markers in some Balkan languages that go back to forms of "will / want" plus the conjunction "that":
E.g. Modern Greek tha grafo: / tha grafeis / tha grafei "I / you / he will write", going back to thelo: na grafo: / theleis na grafeis / thelei na grafei .

I suppose I don't need to call it an auxiliary then. In fact, trying to think of this as an auxiliary made this seem like much more of a puzzle than it otherwise should have been. Thanks!

I think it's common for auxiliaries to be used with "participle"-type verb forms where marking for tense/aspect/mood or voice might occur on the main verb (possibily number/gender marking as well in a language where that regularly appears on adjectives). In English "The bowl is broken" and "The bowl has broken", the auxiliary is used along with a main verb placed in a special non-finite form; in "The bowl is shattered" and "The bowl has shattered", that non-finite form happens to look the same as an inflected finite form of the verb.

In a structure involving an auxiliary, the category I would be most surprised to see marked by inflection on the main verb is person agreement. Although I would expect person agreement to appear on the auxiliary if it exists at all, it might not be marked on either part: some languages lack person agreement in any part of the verb system. English modal auxiliaries lack any person inflection, but are still commonly analyzed as verbs (I think some generative analyses avoid that analysis) because some of them can inflect for tense, and in terms of syntax they behave similarly to normal verbs in some ways, and even more similarly to the auxiliaries be and have, which do possess person/number inflection in the finite present-tense forms and which have a more filled out paradigm of characteristically verbal non-finite forms.

An auxiliary with invariant form might be analyzed as showing verbal behavior syntactically because of how it interacts with things like the order of other grammatical words. Where are pronouns, negative markers, and adverbs placed relative to keham and the main verb in your language?

There are actually lots of possibilities here; the familiar European type of auxiliary verb construction is just one among many found in natural languages, and it's certainly possible for the lexical verb to be inflected while the auxiliary is not. I recommend this Conlangery episode for an eye-opening overview.

If you want to justify calling keham an auxiliary verb rather than a particle I think you'd need to show that it has some verblike behavior, and that it's acting as the syntactic head of the clause. Word order can help with this, like Estav suggested; it would also help if the lexical verb was in a dependent form, but this doesn't have to be a non-finite participle or infinitive: it could be something like a subjunctive or irrealis mood that still allows for aspect marking.

This is partly a matter of definition, since some might argue that a verb is only an auxiliary if it inflects for the categories that other verbs inflect for and if it's the head of the construction, but not everyone takes that point of view. I have a book by Gregory Anderson called "Auxiliary Verb Constructions", and under his definition there are many patterns including Aux-headed (auxiliary = finite inflected verb), Lex-headed (non-aux verb = finite inflected verb), doubled (inflection on both verbs), and split (some categories on one verb, some on the other) exist.

Just going from the table of contents, he has sections to discuss the following patterns:

Lex-headed/inflected

Aux-headed/inflected

Doubled:
subject inflection
object inflection
TAM inflection
negation

Split:
negation on lex, subject/TAM on aux
subject/TAM on lex, negation on aux
object agreement on lex, subject on aux
subject agreement on lex, object on aux
TAM on lex, subject/object on aux
subject on lex, TAM on aux

Looking at the lex-headed section, many of the examples are what would typically be labelled "particles" in other sources. But regardless of how you label them, it is certainly true that they are not inflected, are not themselves inflectional affixes, and mark TAM categories.

Corumayas wrote: ↑Mon Dec 02, 2019 2:47 pm If you want to justify calling keham an auxiliary verb rather than a particle I think you'd need to show that it has some verblike behavior, and that it's acting as the syntactic head of the clause. Word order can help with this, like Estav suggested; it would also help if the lexical verb was in a dependent form, but this doesn't have to be a non-finite participle or infinitive: it could be something like a subjunctive or irrealis mood that still allows for aspect marking.

You could also follow one of the split/doubled patterns I mentioned in my previous post. If you mark some verbal category on the auxiliary then it looks a bit more verb-like.

Kilanie wrote: ↑Sun Dec 01, 2019 11:39 pm Are there any natlangs where an auxiliary verb is uninflected while the main verb keeps its usual inflection?

Taking the question literally, we have an example in English with the auxiliary 'must'. That auxiliary has no visible inflextion. while the main verb is marked to show whether it is progressive or not, perfect or not, and active or passive.

Estav wrote: ↑Mon Dec 02, 2019 9:26 amAn auxiliary with invariant form might be analyzed as showing verbal behavior syntactically because of how it interacts with things like the order of other grammatical words. Where are pronouns, negative markers, and adverbs placed relative to keham and the main verb in your language?

Corumayas wrote:If you want to justify calling keham an auxiliary verb rather than a particle I think you'd need to show that it has some verblike behavior, and that it's acting as the syntactic head of the clause.

I think the most distinguishing factors would have to be the placement of the negation word, as nouns don't come between the axiliary and main verb and there's no other verb inflection aside from the perfective. I think I'd permit adverbs to come between.

chris_notts wrote:Split:
negation on lex, subject/TAM on aux
subject/TAM on lex, negation on aux
object agreement on lex, subject on aux
subject agreement on lex, object on aux
TAM on lex, subject/object on aux
subject on lex, TAM on aux

Now this is a fun set of possibilities! I think in my case it'd make sense to let negation sit on the auxiliary. I'm torn between that and just letting it be led-headed, but that's for me to work out.

Richard W wrote:Taking the question literally, we have an example in English with the auxiliary 'must'. That auxiliary has no visible inflextion. while the main verb is marked to show whether it is progressive or not, perfect or not, and active or passive.

That's true; though I suppose in English once there's an auxiliary you can only get TAM from other auxiliaries, as there's no way to get tense or person marking on the main verb or on any other auxiliaries as the uninflected auxiliary 'blocks' this.

Hungarian also have auxiliary that is not conjugated. I think that it's pretty common for a voice to be marked on the main verb instead of auxiliary, but I might be wrong.

Xwtek wrote: ↑Wed Dec 04, 2019 1:27 am Hungarian also have auxiliary that is not conjugated. I think that it's pretty common for a voice to be marked on the main verb instead of auxiliary, but I might be wrong.

Yes, because auxiliaries often govern infinite forms, which in turn frequently are marked for voice.

Far as I can tell, verbs' inflectional work can be distributed any number of ways across auxiliary and core verb--the Finnish negative verb, for example, only marks person and number (and the imperative).

Of course, diachrony imposes its own cross-linguistic trends, if not hard-and-fast rules.

Meanwhile, the Irish preverbal particle is marked for mood (negative, declarative, interrogative) and tense (past vs non-past). Variations of the same particle can be used to achieve predication of nouns, which increases the argument for considering it an auxiliary. E.g.:

An múineann tú? "Do you teach?" (-PST,+INT,-NEG)
An múinteoir thu? "Are you a teacher?" (-PST,+INT,-NEG)
Nár mhúin tú bitheolaíocht do mo dheirfiúr? "Didn't you teach my sister biology?" (+PST,+INT,+NEG)
Nár mhúinteoir bitheolaíochta mo dheirféar thu? "Weren't you my sister's biology teacher?" (+PST,+INT,+NEG)

Richard W wrote: ↑Mon Dec 02, 2019 9:47 pm

Kilanie wrote: ↑Sun Dec 01, 2019 11:39 pm Are there any natlangs where an auxiliary verb is uninflected while the main verb keeps its usual inflection?

Taking the question literally, we have an example in English with the auxiliary 'must'. That auxiliary has no visible inflextion. while the main verb is marked to show whether it is progressive or not, perfect or not, and active or passive.

I wouldn't say the main verb "keeps its usual inflection" though; we can have another auxiliary plus a participle, yes, but we can't have tense marking or subject agreement (*I must played, *She must plays, *I must am playing etc.).