Mergers and near mergers before liquids

The liquid consonants /ɹ/ and /l/ are well known for their tendency to influence the quality of adjacent vowels. A number of phonemic contrasts are neutralized in this environment. An example of this is the well established pattern in the West and Midwest whereby the distinctions among /æ/, / ε/, and /e/ are lost before /ɹ/ . The resulting vowel is typically closest to [ε] so that marry, merry, and Mary are all pronounced as [mεɹi].

The phoneme /l/ is also contributing to the reduction or loss of several phonemic contrasts across much of the US. Among the most important patterns for the region discussed here are conditioned mergers of /i/ and /ɪ/ , /u/ and /ʊ/ , and /e/ and /ε/ in the context of a following /l/. These mergers result in homophones for pairs such as feel and fill, fool and full, and fail and fell. The phonetic quality of the merged vowel approximates to the lax member of each pair; i.e., [ɪ] , [ʊ] , [ε] (Thomas 2001: 50).

Compared to the features described above, awareness of these mergers among dialectologists has come relatively recently. Labov, Yaeger, and Steiner (1972) identified mergers of /ul/ ~ /ʊl/,/il/ ~ /ɪl/ and /el/ ~ /εl/ among speakers from Albuquerque and Salt Lake City. Labov’s more recent investigations through the Telsur project show these mergers to be widespread across almost all of the US though they are distributed quite sparsely in many regions. Their geographical patterning among the Telsur respondents bears some resemblance to that seen with the fronting of /o/: they are relatively more common across the Midland and in southern regions of the West than in the Northwest and Upper Midwest. This similarity in regional distribution is not surprising given that the pre-L mergers, like the fronting of back vowels, are also common in the South.

The pre-L mergers appear to be a fairly recent development and moreover active changes in progress, at least in some areas. Thomas’ (2001) acoustic data suggest, for example, that /ul/ and /ʊl/ are merged for most younger Ohioans, those born after 1963, while older speakers maintain a clear separation in vowel space. Similar generational differences were found among Utahns by Di Paolo and Faber (1990). This latter study also established that these developments do not necessarily result in a complete merger of the vowels. Di Paolo and Faber found that even when the vowels overlap in phonetic space (as shown by acoustic measurements), speakers may preserve a distinction through phonation differences (e.g., creaky voice). One of the most intriguing aspects of these types of changes, which Labov (1994) labels ‘near mergers,’ is the finding that speakers may perceive no contrast between the sounds even when they consistently produce a distinction phonetically.