| Voice marking | Proper marker | yes | Antipassives (suffix-derived) (Miyaoka 2012: 917).
As a bivalent, an agentive stem may occur with transitive or intransitive inflection. The former indexes P and A arguments. The latter, i.e. de-transitivized
without any suffixal modification, is antipassive, with P being demoted to the ablative-modalis (§ 25.2.1; hence no indexing in inflection) and A being promoted to S (absolutive status). The demoted P (as parenthesized below) is semantically an object but not syntactically (Miyaykoa 2012: 903). |
| Voice marking | Lookalike marker | no | |
| Voice marking | Synthetic marker | yes | Antipassive: -gi-, -uc-, -kenge- (Miyaoka 2015: 1194; also Miyaoka 2021: 918).
Derivational detransitivizers -(u)te-, -(g) (Mithun 2000: 96).
|
| Voice marking | Analytical marker | no | |
| Flagging | S-argument flagging | no | As a bivalent, an agentive stem may occur with transitive or intransitive inflection: the former indexes P and A arguments.
The latter: de-transitivized without any suffixal modification, is antipassive, with P being demoted to the ablative-modalis (hence no indexing in inflection) together with A being promoted to S (absolutive status) (Miyaoka 2012: 903).
KS: Then NO S = ABS
KJ: The answer should be NO. The single core argument in a P demotion is left unflagged. S/P are left unflagged. |
| Flagging | P-oblique flagging | yes | The semantic patient appears as an ablative (oblique) (Mithun 2000: 89).
The ablative-modalis status of object in the antipassive (Miyaoka 2012: 844).
|
| Flagging | P-oblique unflagging | no | |
| Flagging | P-oblique flagging variation | no | |
| Indexation | S-argument indexed | yes | The de-transitivized without any suffixal modification is antipassive with P being demoted to the ablative-modalis (hence no indexing in inflection) (...) (Miyaoka 2012: 903). |
| Indexation | S-argument indexation conditioned | no | |
| P-individuation properties | Incorporated P is generic (non-specific) | n/a | |
| P-individuation properties | Incorporated P is indefinite (non-specific) | n/a | |
| P-individuation properties | Incorporated P can be referential | n/a | |
| P-individuation properties | Oblique is generic (non-specific) | yes | @KJ: ‘The woman likes children.’ Miyaoka 2012: 918). |
| P-individuation properties | Oblique is indefinite (non-specific) | yes | Mithun (2000: 94) ex. (29): ""And they would hear people."" = Unidentifiable (indefinite) semantic patient.
‘They (bears) attack people.’ (Miyaoka 2012: 1115).
|
| P-individuation properties | Oblique can be referential | yes | (11a) ‘The man is eating the fish.’
(11b) ' The man is eating a fish.’
antipassive (“indefinite” P demotion) (...). The resultant abs ~ abm alternation yields contrast in the definiteness of ‘fish’ (Miyaoka 2012: 1177).
@yes, but see ex. 9. The absolutive object is definite in (a), but the ablative-modalis may be indefinite though not necessarily (may be either a or the). A demoted NP can be a person name or a possessed noun (Miyaoka 2012: 903, 754).
As for the difference in the nominal case between (9a), (10a) vs. (9b), (10b) and in the glosses the vs. a/the, the definiteness of the fish eaten and the whereabouts of the house to be bought are possibly known in the case of the transitive (a), but not necessarily in the intransitive or antipassive (b). The same kind of contrast between transitives and (suffix-derived) antipassives will be seen to obtain for patientive stems as well (Miyaoka 2012: 904).
The same semantic difference in definiteness of the P argument obtains between a transitive form (definite; the) and an antipassive (neutral in definiteness; a/the) (Miyaoka 2012: 918).
FYI: A particularly strict relationship between antipassivization and referentiality is found in Eskimo languages. For example, Central Alaskan Yupik does not have articles, but the P term of the transitive construction of Central Alaskan Yupik can only be interpreted as definite. Some Central Alaskan Yupik verbs are A-ambitransitive, and allow for the expression of indefinite P arguments as obliques in an intransitive construction without any change in their stem, as in (5b), but other verbs are P-ambitransitive verbs with which a morphologically unmarked intransitive construction can only have a decausative or passive interpretation (5d). With P-ambitransitive verbs, antipassive derivation with the voice marker -i- is required to express the indefiniteness of the P argument, as in (5e). (Creissels 2021: 348).
|
| P-individuation properties | Eliminated P is generic (non-specific) | no | |
| P-individuation properties | Eliminated P is indefinite (non-specific) | yes | (...) speakers would want to use patientive ambitransitive verbs (...) with indefinite semantic patients as well, to say such things as ‘she’s helping out’ or ‘she cut things’. Such expressions are made possible by two derivational detransitivizers, -(u)te- and -(g)i- (Mithun 2000: 96). @Other examples: 'The child passed by.'
|
| P-individuation properties | Eliminated P can be referential | no | |
| Oblique affectedness | Less affected oblique | yes | In the following intransitive construction (antipassives), unlike the corresponding transitive one given for comparison, the action of the subject is in focus and the demoted P/T, which is backgrounded, is indefinite or only partly affected (Miyaoka 2012: 754). |
| P-constraining properties | Animacy constrains oblique demotion | no | |
| P-constraining properties | Person constrains oblique demotion | no | |
| P-constraining properties | Number constrains oblique demotion | no | |