Appalachian English exhibits sentences that seem to contain two subjects, as in (1), which are not possible in standard English:

(1) a. We don’t nobody know how long we have. (Montgomery and Hall 2004)

b. ... they didn' nobody live up there. (our fieldwork)

c. There can't nobody ride him. (Montgomery and Hall 2004)

d. ... there wouldn' nothin' go down through there. (Feagin 1979, 238)

Our investigation of these types of (split subject) sentences leads us to discover that (a) they are restricted to having a pronoun in co-occurrence with a quantificational subject, (b) they obligatorily contain a modal or finite auxiliary, and (c) this modal or finite auxiliary is negated in the overwhelming majority of the cases.

We hypothesize that the pronominal element and the quantifier start out as a unit, and the pronoun then raises out of this constituent. This way of looking at things provides a straightforward answer to the puzzle of how we could have two subjects for a single predicate. It also provides independent support to the view that pronouns originate from within their associate, as has been suggested for the case of clitic doubling, of the relation between a pronoun and its antecedent (Hornstein 1999, Kayne 2002) and for expletives and their associates in existential clauses (Sabel 2000, Witkos 2004, Kayne 2008).

We further hypothesize that the raising of the pronoun is parasitic on the movement of the negated modal or finite auxiliary, which raises to a higher position in Appalachian English than in standard English. This is supported by the observation that Appalachian English exhibits so-called ‘negative inversion’, where a negated modal or finite auxiliary precedes the subject in a declarative clause, as shown in (2):

(2) Back in them days, didn' nobody live there but just a little while, did they? (Feagin 1979, 238)

We propose that, in Appalachian English, the marker of sentential negation raises to mark its scope overtly. This accounts for the distributional properties of negative inversion: it obligatorily involves a negated modal or finite auxiliary, and is restricted to quantificational subjects. Given the similarities in distribution, we conclude that sentences with negative inversion and with split subjects both involve the raising of the negative modal or finite auxiliary to a position higher than the quantificational subject; the latter also involves the raising of some of the features of the quantificational associate, which are spelled out as pronouns.