Split subjects

"We don't nobody know how long we have."

(Montgomery and Hall 2004)

A split subject is a construction in which the subject of a sentence appears to consist of two parts that do not appear next to each other. Some examples are given in (1), with the two halves of the split subjects italicized:

1)  a. They didn't nobody like him.
         (Feagin 1979)

b. ... back in them days, they didn't nobody live up there.
(Feagin 1979)

c. They can't many people say that.
(Dante Oral History Project)

d. We don't nobody know how long we have.
(Montgomery and Hall 2004)

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

Split subjects is the term used by Zanuttini and Bernstein (2014), but these constructions are also called interposed pronouns, such as in Montgomery and Hall (2004).

Who says this?

Examples like those in (1) are attested in the following varieties: Alabama English in Anniston, Alabama (Feagin 1979), Appalachian English (Wolfram and Christian 1976; Montgomery and Hall 2004; Zanuttini and Bernstein 2014; Dante Oral History Project), and West Texas English (William Salmon, personal communication).

They were found in African American English in the past, as we know from ex-slave narratives, but are not found in contemporary African American English (Green 2002, 2011a, 2011b; White-Sustaíta 2010). The example in (2a) is an example from an ex-slave recording from 1944 whereas the example in (2b) is an example that is judged as unacceptable by a contemporary African American English speaker:

2) a. But they'd give me a note so there wouldn' nobody interfere with me.
        (FH, born 1848, recorded 1944, in Bailey et al. 1991)

b. *Dey didn't nobody see it.
(Weldon 1994)

Syntactic properties


The first component of a split subject is always a pronoun; it can never be a lexical noun phrase. Thus, whereas the examples in (3) are grammatical, their counterparts in (4) are not:

3) a. They didn't nobody like him.
        (Alabama English; Feagin 1979)

b. They won't nobody know you're gone.
(Alabama English; Feagin 1979)

4) a. *The children didn't nobody like him.
(Zanuttini and Bernstein 2014)

b. *The children won't nobody know you're gone.
(Zanuttini and Bernstein 2014)

The second component of a split subject is an indefinite quantificational element, like nobody, anyone, nothing, anything, or many people. It cannot be definite, whether quantificational or not, as shown by the unacceptability of the sentences in (5):

5) a. *They didn't all the children like him.
        (Zanuttini and Bernstein 2014)

b. *They didn't Fred and Mary like him.

Modals or finite auxiliaries required

Sentences with split subjects require the presence of a modal or finite auxiliary. In the absence of a modal or auxiliary, the co-occurrence of the pronoun and the quantifier is not attested; that is, one does not find examples of the type given in (6):

6) a. *They nobody like him.

b. *They nobody know you're gone.


Split subjects have an overwhelming tendency to appear in sentences in which the modal or finite auxiliary is negated. However, a few instances of sentences with split subjects that lack negation have been reported, as in (7):

7) There'd somebody come around with a truck once in a while.
     (Montgomery and Hall 2004: 551)

The negation on the modal or finite auxiliary is always the morpheme n't and never the morpheme not.

Expletives or referential pronouns?

In some of the varieties of English that allow split subject, like Appalachian English, the pronoun they can serve both as a referential pronoun, as in (8), and as an expletive pronoun (that is, the kind of pronoun that we find as the subject of an existential sentence), as in (9):

8) They asked four men.

9) a. They is something bad wrong with her.
(Appalachian English; Montgomery and Hall 2004: lxii)

b. I believe they is a cemetery there too, ain't there?
(Montgomery and Hall 2004:lxii)

c. They's nothin' to keep 'em from turnin'.
(Wolfram and Christian 2004:125)

d. Now, they's a difference in sayin' a fun ghost story and ...
(Wolfram and Christian 2004:125)

Both they and there are commonly found in sentences with split subjects in these varieties, as we see in (10):

10)  a. They won't nobody know you're gone.
           (Appalachian English; Feagin 1979)

b. ... there didn't anyone want to leave their church.
(Appalachian English; Feagin 1979)

c. ... there wouldn't nothing go down through there.
(Appalachian English; Feagin 1979)

With these observations in mind, a question arises: Is they an expletive in (10a), or is it a referential pronoun? Our preliminary investigation suggests that it is an expletive, as the interpretation of the sentence seems to be the same as that of Nobody will know you're gone—that is, the pronoun they does not appear to restrict the domain of the quantifier to a certain set of people relevant in the conversation.

Feagin (1979) also provides an example, shown in (11), of a split subject in which the pronoun is it:

11) No'm, it didn't very many of 'em dip snuff back then.
       (Appalachian English; Feagin 1979)

Note that it, like they, can also be found in existential sentences, as shown in (12):

12) It ain't nothin can stop it.
       (Appalachian English; Feagin 1979)

Related Phenomena

Belfast English

Split subjects similar to the ones described above are discussed for Belfast English in Henry and Cottell (2007), under the label transitive expletives (because their paper focuses on the occurrence of these subjects with transitive verbs). Some examples of transitive expletives from Henry and Cottell (2007) are given in (13):

13)  a. There shouldn't anybody say that.

b. There should lots of people have read this book.

c. There shouldn't anybody drink wine before dinner time.

Though these examples are similar to the ones we are describing in this page, we do not want to assimilate them into one construction (at least for the time being) because they also exhibit some differences. Four differences in particular stand out: (i) the Belfast English sentences are not restricted to negative contexts, whereas in the North American varieties they are predominantly found in negative clauses; (ii) the only pronominal element that is attested in initial position in Belfast English is there, whereas in the North American varieties we also find they and we; (iii) the quantificational part of the subject can occur in a range of positions, as discussed in Henry and Cottell's article, whereas it appears not to have the same range of possibilities in the North American varieties; and (iv) the presence of this type of subject correlates with the presence of negative inversion in the North American varieties, but not in Belfast English. (See our page on negative inversion for more discussion of that phenomenon.)

Two types of split subjects?

Montgomery and Hall (2004) group under the same label the type of sentences seen in (1) and sentences of the type shown in (14):

14) a. We don't any of us need anything.
           (Appalachian English; Montgomery and Hall 2004: p. 413)

b. They didn't any of them want to go out.
(Appalachian English; Montgomery and Hall 2004: 18)

c. But they won't any of them fight us square.
(Appalachian English; Montgomery and Hall 2004: 18)

We hesitate to group them together, however, as they exhibit different characteristics. First, the indefinite subject in the examples in (14) consists of any or none followed by a partitive phrase (e.g. of us), whereas in the previous examples of split subjects the indefinite components (nobody, anybody, nothing, etc.) have a nominal component and no partitive phrase.

In addition, the pronoun in sentence initial position in (14) seems to be the same as the pronoun present in the partitive phrase contained within the indefinite subject, though in the nominative case (we vs. us, they vs. them). In contrast, in the previous examples of split subjects the pronoun does not double a pronominal element present within the indefinite subject.

Finally, many speakers who do not accept sentences with split subjects like those in (1) do accept sentences with subjects like those in (14).

These differences suggest that the grammars that can generate sentences like those in (14) are not the same as the grammars that can generate sentences like those in (1).

Page contributed by Raffaella Zanuttini on November 4, 2013

Page updated by Tom McCoy on August 23, 2015


Bailey, Guy, Natalie Maynor and Patricia Cukor-Avila. 1991. The Emergence of Black English: Text and Commentary. Amsterdam: John Benjamins.

Dante Oral History Project. 2013. Available here.

Feagin, Crawford. 1979. Variation and Change in Alabama English: A Sociolinguistic Study of the White Community. Washington, DC: Georgetown University Press.

Green, Lisa. 2002. African American English: A Linguistic Introduction. Cambridge and New York: Cambridge University Press.

Green, Lisa. 2011a. Force, focus and negation in African American English. Paper presented at LSA Annual Meeting.

Green, Lisa. 2011b. Language and the African American child. Cambridge: Cambridge University Press.

Henry, Alison, and Siobhan Cottell. 2007. A new approach to transitive expletives: evidence from Belfast English. English Language and Linguistics 11 (2), 279-299.

Montgomery, Michael, and Joseph S. Hall. 2004. Dictionary of Smoky Mountain English. Knoxville: University of Tennessee Press.

White-Sustaíta, Jessica. 2010. Reconsidering the syntax of non-canonical negative inversion. English Language and Linguistics 14:429–455.

Weldon, Tracey. 1994. An HPSG account of Negative Inversion in African-American vernacular English. Manuscript, The Ohio State University.

Wolfram, Walt, and Donna Christian. 1976. Appalachian Speech. Arlington, VA: Center for Applied Linguistics.

Zanuttini, Raffaella and Judy B. Bernstein. 2014. Split Subjects in Appalachian English. In Raffaella Zanuttini and Laurence R. Horn [eds.] Micro-Syntactic Variation in North American English. Oxford University Press.