What is the structure of the L.A. Chicano English vowel system? How does LACE compare to other dialects of English phonologically? What are the phonemes, the available contrasts? What are the word-classes associated with these classes and how do these compare with other dialects? This section answers these questions.
As in other chapters, I will use the lexical sets defined by J.C. Wells (1982) for comparison, pointing out the groups of lexical sets which are not phonologically distinguished in LACE. These are labelled ``merged sets''.9.7 Splits will not be discussed, since I found no evidence for them.
Many American dialects do not distinguish the word classes NORTH and FORCE (though Southern dialects like that of Anniston, Alabama, do keep them separate). I assume that LACE is among these. Like other American Englishes, LACE is a ``flat-BATH'' dialect. That is, it classes the BATH set with the TRAP set rather than with the PALM set. A third issue is whether or not Chicano English distinguishes the low-back vowels //(THOUGHT, CLOTH) and //(LOT). This historical merger is progressing rapidly on the West Coast of the U.S.9.8 I will examine the question of whether or not LACE makes this distinction, but for now I assume it does not. Hence, for my purposes Los Angeles Chicano English does not distinguish the following lexical sets of J.C. Wells:
Because of phonetic similarity and complementary distribution, stressed and unstressed // (NURSE, LETTER) are the same phonological class (unlike in Jamaican, where the unstressed vowel in LETTER is not rhotic and is classified as /a/, while the stressed rhotic vowel in NURSE remains phonologically //). Similarly, stressed and unstressed high-front-peripheral vowels (FLEECE, HAPPY) are classified together as /i:/ (unlike in older RP (cultivated Southern British), where HAPPY ends with the vowel in KIT).9.9
The phonological structure assumed here is presented graphically in Table .9.109.11
This structure is somewhat different from the Reference American structure (Table ), primarily due to mergers that have occurred in the historical development of LACE. These mergers include the merger(s) of RA //, /a:/, and /:/ (Wells' LOT, PALM, and THOUGHT sets, traditionally called ``short O'', ``broad A'', and ``long open O'', respectively). While the merger of broad-A with the other vowels may have occurred some time ago, the ``Low-Back Merger'' of Labov (1991) between Reference American // and /:/ appears to be currently in progress in California Anglo English (see section ). The resulting phonological class, here written as //, must be analyzed as a phonologically long vowel, given the basic English principle that short vowels cannot be stressed without a following consonant in the syllable.9.12 LA Chicano // occurs in such words as law, spa, etc. If this general English principle applies to this dialect, this sound cannot be analysed as a short vowel. Nor can it be analysed as a vowel with a glide. First, its low, non-front nucleus is identical with that of /w, r/ and possibly also /y/. Second, it contains no phonetic glide. Thus it cannot be structurally analysed as occurring within the Vr, Vy, or Vw subsystems.
The remaining alternative is to place // within the V: long vowel subsystem. If this were not the only alternative, then the other long vowels, /i:, e:, o:, u:/9.13 might be analysed phonologically as elements of the Vy and Vw systems, thus eliminating the V: subsystem, as in Alabama English (page ). But there is no such alternative (barring modification of the fundamental English principle that stressed rhymes branch), and so the proposed structure (Table ) of the LA Chicano vowel system locates /i:, e:, o:, u:/ and // in the V: long vowel subsystem, and provides separate subsystems for the distinctive classes of glides that occur with //: /y, w, r/. The large number of gaps in this system is due to the various mergers that have eliminated vowel classes from the structure, and to the need to posit Vy, Vw subsystems, which itself derives from the apparent identity of /, y, w, r/.
The non-high front vowels before intervocalic /r/ are presumably merged in this dialect (as in the local Anglo dialect and in Chicago, but not in Philadelphia, and various Eastern dialects). That is, Mary, merry, marry are pronounced identically. This phonological collapse has two simplifying effects. First, it eliminates a rather tenuous distinction based on syllable structure rather than segmental features: Mary and merry are elsewhere distinguished phonologically as /me:$ri/ and /mer$i/.9.14Second, by eliminating the /ærV/ class by merger with /rV/, the last instance within LACE of a front-back contrast among low vowels is eliminated. There is only one low vowel within each of the four subsystems: /æ, , ar, ay, aw/ are the sole low vowels for the V, V:, Vr, Vy, and Vw subsystems, respectively. Since base-5 systems (like Jamaican Creole) containing one low vowel but two mid and two high vowels are phonetically simpler (though phonologically less symmetric) due to the triangular shape of phonetic vowel space, this elimination of the front-back contrast9.15 has simplified an aspect of the system. The proposal above could be modified so as to have a base-5 structure, as in Jamaican Creole, at the cost of an unnatural rule that fronts the nucleus of the low short vowel to /æ/. No perfect solution is apparent, and for now I will accept Table as the surface-phonological structure of the Los Angeles Chicano English stressed vowel system. The same structure might also be applicable to the Anglo community.