Deriving Stative Verbs

[Terra]

Are there any languages that regularly derive stative verbs from eventive ones? (or vice-versa?) That is, like the following:
-- fall asleep -> sleep
-- put on -> wear (and likewise, take off -> not wear, I suppose)
-- pick up -> carry
-- grab -> hold

[vec]

A lot of old IE languages do similar things via perfectives.

[Sleinad Flar]

*perfects.

Indeed, Homeric Greek being the prime example.

Doesn't Arabic do this, e.g. with forms IX and XI-XV in this table http://en.wikipedia.org/wiki/Arabic_ver ... orms.22.29 ? Maybe someone more knowledgeable on Arabic than me can help.

[Yng]

No - Classical/Literary Arabic doesn't generally have much of a distinction between statives and what you might call 'inchoatives'. 'Become X' and 'be X' are usually expressed by the same verb. احمرّ iḥmarra for example, a form VIII verb, means both 'be red' and 'become red', though I would say in modern literature and in the classical literature I've encountered the 'become' interpretation is far more common because there are generally adjectives, or participles (which in Arabic are generally resultative, i.e. the participle of iḥmarra muḥmirr is 'having become red') with an unambiguously stative meaning. Forms XI-XV, which in any case are quite rare, are stative verbs but they are not derived from action verbs. There are some examples in Classical Arabic of form I (i.e. notionally 'underived') verbs which have both stative and non-stative counterparts - I can't actually think of any examples off the top of my head but there'll be two verbs identical in form except falling into different ablaut classes, one stative-ish/intransitive and one transitive. These are not productive though and it's difficult to say that the intransitive/stative is derived from the transitive as opposed to the other way around.

That said, the resultative participle arguably does this, though it's a morphological form rather than a derivational one and doesn't have entirely verbal characteristics - it can't be habitual, for example, and has to be combined with the verb 'to be' to express habitual meaning, like an English adjective. You have pairs in colloquial Arabic like هو نايم huwwe nāyim 'he's asleep [having gone to sleep]' < هو بينام huwwe binām 'he goes to sleep, falls asleep' and هو لابس huwwe lābis 'he's wearing [having put on]' and هو بيلبس huwwe byilbis 'he puts on, he wears...'. Persian is the same: نشسته‌است neshaste-e 'he is sitting [has sat down]' vs می‌نشیند mi-nešine 'he sits down'.

[vokzhen]

Maybe not quite what you're looking for, but Ingush (one end of a dialect continuum with Chechen) apparently has a set of adjective-inchoative-causative triplets. A few examples are shiila / shal-lu /shal-d.u (cold - become cold - make cold), d.weaxa / d.weax-lu / d.weax-.du (long - become long - make long) and ghoaza / ghoz-d.oal / ghoz-d.oaqq (happy / become happy / make happy), with d. marking noun class agreement. Most, but not all, have different ablaut grades between the adjective and verbs, and the endings -lu/-.du always deletes the final schwa found on the adjective and -d.oal/-d.oaqq sometimes does (which in the modern language is usually only present as the release of the last consonant anyways, but that's apparently a recent thing). The -lu and -d.u endings create inceptives and causatives when attached to verbs, same with -d.oal and d.oaqq, but I didn't have a chance to check if ablaut is a common result as well. All four endings are transparently derived from verbs: lu is give, d.u is make/do, d.oal is go, and d.oaqq is take.

[hwhatting]

*perfects.

Indeed, Homeric Greek being the prime example.

PIE had a stative suffix *(e)H₁ that was used for such operations in several early IE languages, e.g. Latin sede:re, Poto-Slavic sěděti "sit = "be seated") from PIE *sed- "sit (down)". This seems to have spread especially in languages where the perfect was lost (e.g. Slavic) or mostly lost the nuance of stative result and became a (perfective) past tense (Latin).

[Cedh]

Again, maybe not quite what you're looking for, but still relevant: Goemai, a West Chadic language spoken in Nigeria, has only very few stative verbs, and no stativizing derivational morphology either.

Goemai predominantly lexicalizes verbal concepts as inchoatives rather than statives, for example property concepts (b’ang ‘become red’ [Goemai does not have a word class of adjectives, and most adjectival concepts are expressed by means of inchoative verbs]), dispositions (k’oon ‘become face down’), and mental activities (man ‘get to know’). There are only very few unambiguously stative verbs in the language, including the form class of locative verbs (i.e. lang ‘hang/move’, t’ong ‘sit’, d’yem ‘stand’, t’o ‘lie’, d’e ‘exist’).

In order to allow these inchoative verbs to refer to a state, the language uses serial verb constructions (SVC):

In the absence of derivational morphology, serialization is a pervasive mechanism for inchoative verbs to occur in reference to a state. [...] In configurational SVC, a state-change verb occurs in reference to a state (i.e. ‘being in a certain configuration and position’).

wang k’oon t’ong k’a kuk sh’ep
pot become.face.down(SG) sit(SG) head(SG) stump wood
‘The pot sits face down on the tree stump.’

[...] the two verbs convey two complementary perspectives on the same event: a configuration and a location. And [...] the configurational SVC shows affinities to adverbial structures: speakers only ever rephrase it as an adverbial structure; furthermore, to assert two different configurations, speakers always combine an adverbialized state-change verb with the configurational SVC.

n-k’oon b’am lang sek gak
ADVZ-become.face.down(SG) become.stuck hang/move(SG) body wall
‘Being face down, (it) hangs stuck at the wall.’

(Source: Birgit Hellwig, "Serial Verb Constructions in Goemai", in: Aikhenvald & Dixon, "Serial Verb Constructions. A Cross-Linguistic Typology", Oxford University Press 2006, pp. 88-107)