5-1236msg1.txt

运用贝叶斯网编的一个识别垃圾邮件的程序

Subject: summary : tap / flap

tap / flap : summary : not long ago i ask about dialectal variation in english tap ( sometime call flap ) for < t > . many thanks to the thirty or so respondent , who i will name only where i cite them . * environment * there be general agreement that the trager description be not ( or , possibly , be no longer ) accurate , although one respondent suggest that his description may fit southern us accent better than northern one . flap / tap of a word-final < t > before a stress syllable be possible for most of those who respond . some variety be report to be lenit the flap further , possibly to zero . various speaker , however , have note flap / tap in other people 's dialect which sound odd to them , eg in word like _ button _ and _ eighteen _ . both of these seem to be ok here ( at least on iambic-reversal 18 ) . the most thorough suggestion come from ian mackay : > an intervocalic t or d be flap as long as the follow syllable ( ie , > the syllable contain the second of the two vowel that the t or d be > " inter " ) do not carry primary stress , and that the second vowel do > not carry significantly more stress than the first vowel , and with > the proviso that word boundary may block the phenomenon ( but do n't > ask me to specify the nature of those word boundary ! ! ! ) . one ( british ) respondent say that in their dialect , tap / flap seem to be lexically determine , not determine by phonetic environment , so that a flap / tap be possible in _ witty _ but not _ pretty _ , in _ british _ but not in _ skittish _ . and don churma bring in the spectre of morphological condition , which may also account for the example above : > in fact , there be still further detail that most folks > seem to be unaware of , like the relevance of " level 2 " morpheme > boundary ( cf . my flap / t / before the secondarily stress / i / in > ( deadjectival ) " elitism " v . aspirate / t / in ( denominal ) " magnetism " ) . at least all these different aspect of the problem make it clear why it be so hard to describe properly . * syllable - final , syllable-initial or ambisyllabic ? * alouse turk report that her experiment suggest that an ambisyllabic analysis be the most explanatory . this be also report as the solution use by kahn . other respondent argue for syllable-initial ( on the grounds that for them syllable-final be alway glottalise ) or syllable-final ( on the grounds that voiceless tap / flap also occur , but only in clear syllable-final position or that clearly syllable-initial / t / s be alway aspirate ) . as john harri remark : > as to the coda-v - onset issue , i ' m sure you 'd agree , the competition be not > go to be resolve by simply inspect the datum ( or even listen to it ) , > since coda and onset be not present in them . it 's a matter of compare > the two theory . i could n't agree more , especially since i be aware of conflict definition of ambisyllabicity in the literature on dependency phonology . * reference : * harri , john 1990 phonology 7 , harri john & kaye 1990 the linguistic review harri , john 1994 english sound structure ( blackwell ) chapter 4 jensen , john english phonology , kahn , daniel , 1976 syllable - base generalization in english . dissertation . olive , j . p . , a . greenwoord , & j . coleman . 1993 . acoustic of american english speech . springer - verlag picard , marc ( 1984 ) " english aspiration and flap revisit " . canadian journal of linguistics 29 : 42-57 . turk , a . ( 1993 ) effect of position - in - syllable and stress on consonant articulation . cornell ph . d . dissertation zue , v . & m . laferriere . 1979 . acoustic study of medial / t , d / in american english . j . acoust . soc . am . 66 : 1039-1050 . laurie . bauer @ vuw . ac . nz department of linguistic , victorium university , po box 600 , wellington , new zealand ph : + 64 4 472 1000 x 8800 fax : + 64 4 471 2070