FESTIVAL NUIT D’HIVER #10 // 12 AU 21 DÉCEMBRE 2012 // MARSEILLE // GRIM Penny Green-Shard couvre le festival NUIT D’HIVER #10 en tant qu’intermittente. Création d'une bande-annonce du Festival ave...
// 19 Live video
ALEX GRILLO, ALEXANDER SCHELLOW, ALEXANDRE RÉGIS, ÉTIENNE JAUMET, BERNARD BLOCH, BRIGITTE SYLVESTRE, CHRISTIAN WOLFF,CHRISTOPHE MODICA, FREDDY EICHELBERGER, GASTON SYLVESTRE, GHOTUL,GRIM CONCERT ROCK EXPERIMENTAL, JEAN-MARC MONTERA, JEAN-YVES BOSSEUR, JOHN CAGE, JOHN CAGE PROJECT, LA TERRE TREMBLE!!!, LE CABARET CONTEMPORAIN, LUCIE ROCHE, PAUL ELWOOD,SARA HAEFELI, SÉVERINE BALLON, SUBSPECIES, TAKUMI FUKUSHIMA, TALWEG, ZA!
((Okay so for my Phonology class I have to summarize a series of articles, and I'm using Tumblr to help me collect my thoughts on them. If you do actually care to read any of this keep in mind that this is not my opinion or even necessarily good science, simply a summarization of whatever the original author wrote))
Bloch begins by noting that while many phonologists have noted and struggled with the sheer diversity in allophones that may be attributed to a single given phoneme there has been almost no academic endeavor whatsoever undertaken to answer the question 'can allophones of different phonemes overlap and what does it mean for the state of a phonetic analysis if they do?"
Bloch defines two different forms of phonemic overlapping: partial and complete. Partial overlap is when a given sound "x" exists as an allophone of phoneme "A" under certain conditions and also as an allophone of phoneme "B" under different conditions. Complete overlap is when a given sound "x" exists as an allophone of "A" under certain conditions and as an allophone of "B" under the same conditions. In other words partial overlap is when two or more phonemes share an allophone but only exist as that allophone under different conditions (and therefore are never ambiguous), while in complete overlap a certain phonetic environment will cause two or more phonemes to express themselves as identical allophones (and thereby causing ambiguity).
To clarify these types Bloch provides several examples in English of each.
Example 1 for partial overlap: The alveolar plosives [t] and [d] become expressed as the tap [ɾ] intervocalically after a stressed vowel. [ɾ] also occurs as an allophone of [r] (in some varieties of English) following interdental fricatives. Thus [ɾ] exists as an allophone of both [t] and [r] but not under the same circumstances, fitting Bloch's definition of partial overlap. (Interesting to note is that Bloch does not seem concerned that both [t] and [d] become [ɾ] under the same circumstances which would be an example of complete overlap. This seems to be because Bloch does not consider them different phonemes).
Example 2 for partial overlap: [k] is expressed as labial [kʷ] before labial sounds (compare "keel" and "cool"). This labialization naturally also occurs before the labial glide [w]. However in rapid speech words beginning [kw] (like "queen") often completely lose the [w] altogether, the labialization of the preceding [kʷ] being the only evidence of its existence. In such a situation Bloch proposes that the labialization of [kʷ] is an allophone of [w] and therefore [kʷ] is both an allophone of [k] (found when preceding labial sounds) and an allophone of [w] (in rapid speech when following [k]). As these are the same sound but two different circumstance for the two phonemes they fit Bloch's definition of partial overlap.
Example 3 for partial overlap: In final consonant clusters consisting of a nasal followed by a fricative ("warmth", "dense") an epenthetic consonant forms between them, homorganic with the nasal and matching the fricative in voicing (thus [mθ] becomes [mpθ] and [ns] becomes [nts]). When this fricative is a [ʃ] or [ʒ] the ensuing clusters are nigh identical to the affricates [t͡s] and [d͡ʒ], respectively. Bloch proposes that because of this the cluster [ts] and [dʒ] exist as allophones of the respective affricates [t͡s] and [d͡ʒ] (differing in that normally these sounds are a single phonetic unit as implied by the tie-bar but in this position must be analyzed as a cluster in order to fit Bloch's structuralist paradigm) and of the respective fricatives [ʃ] and [ʒ] (when following a nasal in a final consonant cluster). As these same sounds occur as expressions of different phonemes in different phonetic environments the fit Bloch's definition of partial overlap.
Example 1 for complete overlap: English contains the vowel [ə] only in unstressed positions (ex. about, confess). Bloch identifies it as the unstressed form of [ʌ]. However [ə] also appears as an allophone of many other vowels when moved from a stressed position to an unstressed one ("at" is expressed with [æ] when stressed in the phrase "where at" but with [ə] when unstressed in the phrase "she's at home"). As all these different vowel phonemes can all be expressed as the same allophone ([ə]) under the same condition (unstressed) this fits Bloch's definition for complete overlap.
Example 2 for complete overlap: in English vowels lengthen predictably before voiced consonants (compare "pot" to "pod") making vowel length an allophonic difference. However in Bloch's Chicago dialect words like "bomb", "bother", and "sorry" are also distinguished from words like "balm", "father", and "starry" by vowel length (the first set have short vowels, the second have the same respective vowels but long). As these pairs lack the predictable alternation of "pot" and "pod" Bloch proposes that in his dialect vowel length is also phonemic. Another words with a long vowel in Bloch's dialect is "pa". When plugged into the sentence "Pa'd go if he could" the utterance "pa'd" is phonetically identical to "pod" and yet the long vowel in "pa'd" is a phonemic [ɑː] while the long vowel of "pod" is a lengthened allophone of [ɑ]. Therefore there are two different phonemes being expressed as the same allophone under the same conditions which fits Bloch's definition for complete overlap.
Complete overlap raises questions from Bloch; if a single sound can be attributed to two or more different phonemes in a single given environment then how is one to ever know what the correct underlying phoneme is? Bloch believes this ambiguity to be inadmissible; the phonemic analyses leading to the conclusions must be incorrect, as Bloch says, "even though we have proceeded on both sides of the intersection according to sound principles ans usually valid methods". While Bloch does not attempt to reconcile the nature of unstressed English vowels to fit his "no complete overlap allowed" theory he does attempt to reanalyze the vowel lengths of his dialect.
Because "pa'd" and "pod" cannot, according to this theory, contain different phonemes one of them must be grouped with the other, the proper grouping to be determined by which analysis forms the more realistic phonological. Option 1 is that "pa'd" has the lengthened allophone of "pod", but this would mean that all of Bloch's dialect words ("balm" and so on) would all have to be treated as having that same lengthened vowel even when the conditions for lengthening (a subsequent voiced segment) are not met. So that option is unlikely. Option 2 is that the long vowel of "pod" is not a lengthened allophone but in fact the phonemic long vowel of "pa'd" (the same found in "balm" and so on). Therefore all length must be phonemic, meaning that pairs like "pot" and "pod" do not have the same vowels, but in fact contrasting short and long vowels respectively. Bloch admits that this analysis may seem poorly patterned, perhaps even contrary to common sense and speaker perception, but nevertheless must be accurate as it is the only analysis acceptable to his model.
