So in Ouhalla's original 1993 paper on Berber anti-agreement he says that basically all agreement features are blocked, but then in a 2005 papers he goes "oops turns out this varies with dialect and that original dialect was weird." So for most of the rest of Berber number and gender features are still expressed.Micamo wrote:I'll admit that all of my knowledge of AAE's comes from the literature on wh-movement, but I don't know of any examples of such effects blocking agreement with person but not number or gender. Could you cite some for me?
In Bantu, AAEs occur only with extracted Class 1 subjects, which includes pronouns. Extracted local pronouns level their agreement paradigm to the AAE marker for Class 1 when singular and then Class 2 (plural of class 1). So you get a leveling of the paradigm that basically looks like 3sg vs 3pl. If one takes 3rd person to be underspecified this basically gives only number agreement.
In Dinka, it looks like the extraction paradigm of person marking basically shows number. In Maasai the situation looks similar.
In Seereer, the language I work on, finite verbs take a prefix (sometimes suffix) to show person/number of the subject. In addition to this, many verbs undergo initial consonant mutation in the plural. I'll give you the basic declarative paradigm for ret `go':
Singualar
ret-aam 'I went'
ret-aa 'you went'
a-ret-a 'he/she went'
Plural
i-ndet-a 'we went'
nu-ndet-a 'you guys went'
a-ndet-a 'they went'
Focus and wh-questions involve fronting of the element in question and changing the final suffix of the verb to -u. When a subject is extracted, the paradigm levels, but consonant mutation stays around.
Singular Subject Focus
mi ret-u 'It's me who went'
wo' ret-u 'It's you who went'
ten ret-u 'It's him/her who went'
Plural Subject Focus
ino' ndet-u 'It's us who went'
nuun ndet-u 'It's you guys who went'
den ndet-u 'It's us who went'
So, all marking of person is completely erased, but the marking of number is not: the verb still shows that distinction. Seereer doesn't have class agreement on predicates so gender is irrelevant here. There are other examples of this kind of thing but I can't remember them off the top of my head. It's a striking trend, though.