- Purchase/rental options available:
American Speech 76.4 (2001) 361-387
[Access article in PDF]
Tracing Native American Language History Through Consonant Cluster Reduction: the Case of Lumbee English
THE LUMBEE INDIANS of Robeson County, North Carolina, speak a unique variety of English, significantly distinct from that of their Anglo- and African American neighbors. Sociolinguistic studies of their speech conducted by the North Carolina Language and Life Project (NCLLP) (esp. Dannenberg 1996, 1998, 1999; Miller 1996; Wolfram 1996; Dannenberg and Wolfram 1998; Wolfram and Sellers 1999; Wolfram and Dannenberg 1999; Schilling-Estes 2000) have detailed and contextualized the special ethnolinguistic markers the Lumbee use to set themselves apart from other Robeson County residents. Until now, however, consonant cluster reduction (CCR) has not been examined in detail as a part of Lumbee English.
Syllable-coda CCR in English varieties has been, "in many respects, the paradigm case of systematic variability in variation analysis" (Wolfram, Childs, and Torbert 2000). 1 The Lumbee, as speakers who presumably transitioned from an ancestral language to English at some time in the past, reduce consonant clusters with greater frequency than neighboring Anglos. 2 Many studies have shown that extensive prevocalic CCR is typically associated with contact-based varieties. This paper will attempt to analyze whether prevocalic CCR might place Lumbee English in the context of substrate influence of language transfer.
CCR is a variable feature in which clusters (usually with shared voicing) that end with a stop may delete the second segment of the cluster. For example, CCR commonly occurs in the production of ol' for old, min' for mind, and lef' for left. While sociolinguistic analyses of this feature differ, the basic conditions of variation are clear. The relative incidence of reduction is governed by language-internal constraints, including the form of the following segment, the phonetic composition of the cluster, the morphological status of the final stop, and the prosodic status of the syllable. Preconsonantal environments favor deletion more than prevocalic environments [End Page 361] do, with prepausal environments falling in between. This gradation is intuitively plausible, as fully articulated preconsonantal clusters necessarily involve the production of three or more consonant sounds consecutively--quite difficult for English speakers to articulate. The phonetic composition of the cluster affects variability as well; clusters in which the first segment is a nasal are most likely to be reduced, followed by laterals, sibilants, and stops. Morphologically, reduction takes place in monomorphemic environments (ol' for old) more often than bimorphemic environments created through suffixation (walk' for walked), while a third category of clusters, characterized by word-internal vowel change and redundant suffixation (kep' for kept), falls in the middle. Logically, bimorphemic clusters would be less likely candidates for reduction, because they have a greater functional load than monomorphemic clusters. Prosodically, clusters in unstressed syllables are more prone to reduction than those in stressed syllables. The following is a summary of these internal constraints, synthesized from various sources and taken from Wolfram, Childs, and Torbert (2000) (CCR decreases from left to right):
preobstruent > presonorant > prevocalic
(e.g., [bEs kId] 'best kid' > [bEs nem] 'best name' > [bEs át] 'best at')
nasal > lateral > sibilant > stop
(e.g., [wIn] 'wind' > [waIl] 'wild' > [wEs] 'west' > [ák] 'act')
monomorphemic > redundant bimorphemic > bimorphemic
(e.g., [gEs] 'guest' > [slEp] 'slept' > [gEs] 'guessed')
[-stress] > [+stress]
(e.g., [kántræk] 'cóntract' > [kantræfik] 'contráct')
lower social status > higher social status
casual style > formal style
AAVE > Anglo vernacular varieties
Hispanicized Vernacular English > Anglo vernacular varieties
Vietnamese English > Anglo vernacular varieties
Since voicing agreement of cluster components is a necessary condition for reducing a cluster (d in land is reducible, but t in stunt may not be deleted), syllable-coda t, d, k, and p may variably be deleted in the clusters shown in table 1. Several studies have shown that variability in CCR is governed by these internal constraints, irrespective of external constraints [End Page 362] such as ethnicity and social class (Wolfram 1969; Fasold 1972...