Please join us this week for a meeting of the Language Variation & Change workshop, this Friday, February 28, from 3:30-5pm in Rosenwald 301.

Eliciting silence: Ellipsis as a constraint on morphosyntactic variation in African American English
Tracy Conner, University of California, Santa Barbara

Labov’s (1969) research documented that African American English (AAE) allows two realizations of the copula, an overt form, and a null or zero-marked form (ø), which are often interchangeable (1). This optionality plays a prominent role in understanding morphosyntactic variation in AAE. Often overlooked, however, is Labov’s important observational claim that zero-marking of the copula, unlike use of full forms (2a), is extremely rare in “absolute” position, i.e. in the environment of ellipsis as in (2b) below.

(1) Kayla is/’ø kind and brilliant.
(2) a. We also know that Andrea is kind and brilliant.
b. *We also know that Andrea ø kind and brilliant.

This observation elucidates the nature of morphosyntactic variation in this variety. Yet a major problem exists: no empirical investigations have been conducted to specifically evaluate Labov’s claim as these constructions are rarely produced in spontaneous speech, and thus cannot be examined in corpora or recorded spontaneous speech samples in the frequency necessary to make clear generalizations.

In this talk I present a novel experimental method that overcomes previous limitations, and whose results confirm Labov’s observational claim that the copula in AAE must be overt in the environment of ellipsis. Utilizing an elicitation experiment from psychology, i.e. a sentence repetition task adapted from Potter and Lombardi (1990), I collected production data from 33 AAE speakers from the Mississippi Delta. This task avoids the pitfalls usually associated with collecting data from speakers of socially stigmatized dialects, while also eliciting copula productions in rare environments. Analysis of 556 tokens revealed that zero-marking of the copula was virtually unattested phrase-finally, and results of a logistic mixed effects regression confirmed this preference for overt marking in that environment (p<.01). A follow-up experiment shows that this constraint on optionality also extends to possessive marking which is usually optional in AAE (Sam’ø book or Sam’s book), but is shown to be required in the environment of ellipsis (I like Claire’ø book, but you like Sam’s). These finding help better reveal the constraints on copula and possessive variation in AAE. Furthermore, this unique methodology is useful to those pursuing research on morphosyntactic variation or research with bi-dialectal communities. I ultimately make the case that the distribution of zero-copula and zero possessives is constrained due to requirements for ellipsis licensing. Namely, I assert that functional heads that license ellipsis must be morphosyntactically overt and provide evidence that this requirement holds cross-linguistically for verb phrase and noun phrase ellipsis.