Might should we consider this?: Patterns of double modal inversion in Southern United States English

The Southern United States English (SUSE) double modal construction is of structural interest since clauses containing iterative modals are impossible in standard varieties of English. Existing analyses disagree regarding the interaction of double modals with syntactic processes, including subject-auxiliary inversion. Experiments testing double modal inversion patterns have yielded mixed results and differ in methodology and regional scope, rendering comparisons infeasible. To address these issues, an acceptability judgment task study was conducted with speakers of Kentucky, Tennessee, and Texas Englishes. The results show that joint inversion (e.g. Might should we...?) is preferred in some SUSE varieties; however, at least in Tennessee, second modal inversion (e.g. Should we might...?) is also acceptable. This variation is not immediately reconciled with previous analyses, but is captured under a novel, Tree Adjoining Grammar based model. By furthering empirical understanding of English syntactic variation, these results facilitate evaluation of theoretical analyses for SUSE double modals.