Zero-derivation of intransitives in ergative languages

[dɮ the phoneme]

In English, we have lots of intransitives which are zero-derived from transitives, some of which are unaccusative, but most of which seem to be unergative. In a primarily ergative language, would the analogous situation be the reverse: perhaps a few unergatives, but mostly unaccusatives? Honestly I'm not even sure if my assessment of English here is right to begin with. Maybe we have more unaccusatives than I think? Basically, is this sort of thing known to be correlated with ergativity (in terms of case/agreement marking)?

[MacAnDàil]

What's up with zero derivation anyway? Can we not just say that words are polysemic, with some having both transitive and intransitive meanings.

[KathTheDragon]

Certainly in the case of ambitransitivity, there's no synchronic derivation going on - it's purely a property of a specific class of verb.

[dɮ the phoneme]

Certainly in the case of ambitransitivity, there's no synchronic derivation going on - it's purely a property of a specific class of verb.

Ok, rephrasing the question in these terms: in an ergative language, when an ambitransitive verb is used intransitively, presumably the omitted argument is the ergative one. Is it typical for the semantic role of the absolutive to be the same in the transitive and intransitive usages, or might the absolutive argument have its semantic role reassigned to that of 'agent' when the verb is used intransitively?

[Travis B.]

The key question is that borrowed or coined words by default have both transitive and intransitive usages in a consistent fashion; if yes, then this is zero derivation, if no, then there is a closed set of ambitransitive words.

[Vardelm]

Ok, rephrasing the question in these terms: in an ergative language, when an ambitransitive verb is used intransitively, presumably the omitted argument is the ergative one.

Based on what I have read, yes, this is correct.

Is it typical for the semantic role of the absolutive to be the same in the transitive and intransitive usages, or might the absolutive argument have its semantic role reassigned to that of 'agent' when the verb is used intransitively?

I think it's typical that they stay the same. This is where you get into morphological vs. syntactic vs. semantic ergativity. (I use those terms as a result of referencing Typology of Ergativity by William McGregor).

If you have an example of a "pure" ergative language, which has all 3 types of ergativity listed above, then I would say that, at least in the majority of cases for that lang, the semantic role would stay the same. That is, the intransitive absolutive would usually (> 50%) be the more patient-like argument from the transitive. That's essentially the definition of ergative: the sole intransitive argument and the transitive object share some similarity. In the case of semantic ergativity, they would share semantic role.

My understanding of "ergative" languages that are actually morphologically ergative but syntactically accusative (such as Basque) is that they tend to still be semantically ergative. They will still have the same semantic role for intransitive argument and transitive object. It's just that the ergative argument may be in the same position as the intransitive argument, pivots may be accusative oriented, etc. I don't have data on this, so maybe it doesn't hold up.

I think the article by McGregor might cover the origin of different types of ergative languages. It's been a while since I read it. It would be interesting to see if morphological or semantic ergativity tend to develop 1st, leading to the other one. I assume syntactic tends to develop later since less langauges are syntactically ergative and syntax tends to change slower than morphology. Perhaps morphology & semantics tend to develop in tandem. Not sure.