In this paper, I present a morphosyntactic phenomenon affecting past tense be in a moribund variety of American English, which has been studied extensively

In this paper, I present a morphosyntactic phenomenon affecting past tense be in a moribund variety of American English, which has been studied extensively. I then provide empirical results as evidence for a more general phenomenon than the one previously investigated.

A moribund variety of American English which has attracted the attention of sociolinguists in recent years is spoken in Smith Island, MD, a small group of marshy islands in the Chesapeake Bay. Smith Island English has been examined extensively as a case of language change in progress. That is, sociolinguists have studied the differences between generation groups on Smith Island and have used their observations to make claims about language change as it happens. The process whereby the speech of older and younger members in a community is taken to represent the facts of the language at various points of its acquisition is commonly referred to as the ‘apparent time construct’ (Bailey 2002, e.g.). The apparent time construct has been shown to be an adequate stand-in for real-time, longitudinal data (which is impractical, if not impossible, to acquire).

Previous investigators (e.g. Schilling-Estes 2000, Mittelstaedt & Parrott 2002) have studied various aspects of Smith Island English using apparent time, including a morphological process referred to as ‘weren’t’ leveling or regularization. Leveling is a process of language change whereby a particular form begins to replace other forms in a paradigm, resulting in partial or total syncretism within that paradigm. Leveling in the ‘be’ paradigm is quite common in English, but the Smith Island pattern is different for several reasons. In Smith Island English (SIE), the contracted forms of negated past be occur frequently as ‘weren’t’ regardless of the person and number context. Typically, leveling in English ‘be’ paradigms occurs with the singular form: leveling to ‘was’ and ‘wasn’t’ has been well-documented in many varieties (e.g. Poplack 2000 on African American Vernacular English; Schreier 2002 on Tristan da Cunha English; Cheshire 1982 on Reading English). In SIE, it is the plural form which is taking hold. The other unusual fact about ‘weren’t’ leveling is its lack of occurrence in positive contexts, as well as in contexts of negation realized as ‘not’. This pattern is found only with contracted ‘-n’t’ negation. Examples of attested and non-attested forms (the latter marked with an asterisk) are given in (1) and (2) on the following page. Schilling-Estes (2000) showed that the change to weren’t in first and third person singular contexts (where ‘wasn’t’ would occur in the variety of English probably spoken by the present audience) has nearly reached completion for younger generations of Smith Islanders. That is, younger speakers rarely, if ever, produce an instance of ‘wasn’t’.

The present study examines sociolinguistic interviews with thirty Smith Island community members across four generations to see whether similar leveling patterns take place elsewhere in the auxiliary verb system. As it turns out, non-past auxiliary ‘be’ and ‘have’ and non-past auxiliary ‘do’ are also leveling in this variety, to ‘ain’t’ and ‘don’t’. As with ‘weren’t’ leveling, it happens only in contexts of contracted ‘–n’t’ negation and in the direction of the plural form. The table in (3) on the following page shows the percentage of leveled forms (weren’t, ain’t, and don’t) out of the number of possible leveling contexts: first and third person singular contexts for weren’t; all person and number contexts for ain’t; and third person singular contexts for don’t. These are the environments in which my own variety of English predicts the various forms wasn’t; isn’t/aren’t/hasn’t/haven’t; and doesn’t, respectively. The empirical results show that the difference in occurrence of leveled forms between the oldest and youngest generations in this study is striking. Generation I produces leveled forms approximately two-thirds of the time, while Generation IV levels almost categorically: 98.6% of possible contexts. Based on these data, we can conclude that ‘weren’t’ leveling is not an isolated morphological phenomenon in this variety. Rather, it is part of a more general change in progress in the auxiliary verb system, in which past tense auxiliary ‘–n’t’ forms are no longer being analyzed as markers of negation and agreement. This, in turn, provides further evidence for the concentration model of language death (Schilling-Estes 1997). As this dialect approaches death over time due to environmental and socioeconomic factors, its characteristic features are becoming more frequent in occurrence and changes are reaching completion in as few as two generations. A comprehensive change across the auxiliary verb system can now be considered part of the general direction of change in Smith Island English.

(1) a. Ma weren’t doing no laughing.

b. *Ma were doing some laughing.

(2) a. I weren’t very old.

b. *I were not very old.

(3) Auxiliary and copula leveling in Smith Island English
Leveled to Leveled to Leveled to

Generation weren't ain't don't

Gen. I 27/42 5/10 16/20

b. 1923-1933 64.3% 50% 80%

Gen. IV 30/34 139/141 40/41

b. 1973-1987 88.2% 98.6% 97.6%

References
Bailey, G. (2002). Real and apparent time. In Chambers, J., Trudgill, P., & Schilling-

Estes, N. (Eds.), Handbook of language variation and change. Malden, MA:

Blackwell.

Cheshire, J. (1982). Variation in an English dialect. Cambridge: Cambridge University

Press.

Mittelstaedt, J., & Parrott, J. (2002). A Distributed Morphology account of weren't

leveling. Paper presented at NWAV 31, Stanford University.

The Oxford English Dictionary. (1989).

Poplack, S. (Ed.). (2000). The English History of African American English. Malden, MA:

Blackwell.

Schilling-Estes, N. (1997). Accommodation vs. concentration: Dialect death in two post-insular

communities. American Speech 72.1: 12-32.

Schilling-Estes, N. (2000). Exploring morphological change: The was/weren't pattern in Smith

Island English. Paper presented at NWAV 29, Michigan State University.

Schreier, D. (2002). Past be in Tristan da Cunha: the rise and fall of categoricality

in language change. American Speech 77 (1): 70-99.

Wolfram, W., & Schilling-Estes, N. (2003). Parallel development and alternative

restructuring: The case of weren’t intensification. In Cheshire, J., & Britain, D. (Eds.), Sociolinguistic dialectology: Papers in honour of Peter Trudgill. Philadelphia/Amsterdam: John Benjamins.