Four Kinds of Nuclear Length in English

FOUR KINDS OF NUCLEAR LENGTH
IN ENGLISH

© 2001 by Orchid Land Publications

C.-J. N. Bailey

[20010101, updated 20020208]

 Many varieties of English have four degrees of length--(A) one suprasegmental, (B1) one segmental-phonetological, and two (C) phonetic:

   A.     Certain aspects of intonation like emphasis.

   B.1.  Stress (including that of compounded words).  While modern Greek and Spanish stress involves a raised tone (the Spanish accent as well as the consonant shift resembles, and must be due to, that of the Hellenistic Greek spoken in Iberia during Byzantine rule--just as an Arabic sound was later taken over as Spanish jota and then passed on in more general contexts to Dutch during the periods of Spanish rule in the Netherlands.   English and Italian stress is (segmental) lengthening, as studies have shown.   Note that when vowel length disappeared in Hellenistic Greek, the tonal change heard in the circumflex accent inevitably got lost and was heard as an acute accent.   For the evidence from cerebral analysis concerning which sides of the brain hosts segmental length and which side hosts suprasegmental length, see Ross, E, Edmonson, J. A., and Seibert,. G. B., "The effect of affect on various acoustic measures of prosody in tone and non-tone languages" (Journal of phonetics, 14 [1986], 283-302.)   Edmondson and his medical colleagues at the University of Texas, Arlington, long ago showed that English stress resides in the left side of the brain (for right-handed people), whereas intonation resides in the right hemisphere.   One might predict that Spanish and modern Greek accent would reside in the right hemisphere. 

   C.1.  Nuclear length, where distinctive in a dialect--e.g., when due to the "dropping" of //r// in older Southern States and African American English (contrast par and pa [where not like paw] or bar and bah, as well as card : cod, tort : taught, and curd : cud!).  (Scottish English follows a different principle, lengthening vowels before tautosyllabic voiced fricatives [the Scottish "r" counts as one].)  In all non-Keltic varieties of English (other than in recent creole regions), a given vowel is longer before a tautosyllabic light (underlying "voiced") consonant than before a similar tautosyllabic heavy (underlying "voiceless") consonant.  For the way the three phonetic lengths realize two distinctive lengths, see my English phonetic transcription (SIL, Univ. of Tex. Arlington, 19 85), pp. 74-81.  

   C.2a.  In a given environment, the lower the vowel, the longer it has got to be in order for it to be heard as equally long with progressively higher vowels.  The vowel in seek is much shorter than the one indeck, and the latter is shorter than the vowel in sack.  

   C.2b.  Clause final or pre-pausal vowels are longer than the same vowels in internal environments.  In some varieties, a vowel preceding a single # is (claimed to be) longer than the same vowel preceding the plus boundary.  This is not so in varieties the writer has investigated.

     Vowel-length is generally un-affected by the segments preceding it.

      Analysts with confused views of the real syllabization principles of English will not know for sure when many internal consonants are tautosyllabic or not--and this can change gradiently in clusters when spoken with increasing tempos--as I have shown in various writings.  The examples of a phonetic analysis going astray through failure to understand the distinctions above are legion.  One department head in a Southern States university has written the  "i" in ride, etc., long.  There is a nuclear-length difference as well as a segment difference between wiper and diaper  but only a length difference between brineand Brian or libel and liable in conversational styles.  Had the analyst just referred to only realized it, he was claiming a three-way length-difference in the English in question--short, long, overlong--something patently absurd.  (The difference between right and riot is for many Southerners more than one of vowel length, though that is also involved.   Labov's principle of the effect of the observer on the observed vitiates the observations of a number of his own disciples.)   One well-known analyst, apparently unaware of D. Bolinger's showing that an intonational core or accent always falls on a stressed syllable (but not on every stressed syllable), claimed that tone is part of English stress in some instances.  The error was due to confusing stress with a co-incident intonational core.  

     On the other hand, the distinctions between bad and bared, fad and fared, etc., are real---where the form with //r// shows length and in-gliding in place of underlying //r// in various lects.  Also real is the difference between the first syllables of dia(#)phragm and bi(#)phrasal  and between the first syllables of lible and liable.  

     N. Chomsky has pointed out a segmental distinction between matter and mad#der in his speech.  This is not due to the # but to the marked rule ordering:  
 (i) Change of //æ//.
(ii) Change of //t// to [d] between vowels when the following vowel is unstressed.  Where Chomsky would think of this as virtual time, it is real time in my framework; e.g. one predicts that if the tempos increases significantly during a conversation, the unmarked [maximal] rule-ordering would be realized and  Chomsky would pronounce matter like madder.)  While many speakers in the North-East and even mid-western areas like Chicago distinguish  bad (with a raised and in-glided nucleus) from bat (with a lowered vowel that may not be in-glided), some speakers' are claimed to go so far as to distinguish badand bade (or  red and read [past of read]) (as though //bad#d//and /read#d// like bad and bared- (see above).  

     Note also the tide : ti#ed difference in Tidewater States English.  The single # boundary obviously can play a rôle in some lects, though older Southerners rhyme weary with beer#y and Harry with hair#y.  Non-Southerners rhyme berry and hair#y in a different manner--viz.  by pronouncing the first item with //r//-gemination (also in weary and Harry) as well as in the forms with #, where gemination would be expected.  Older Southerners and Easterners have typically lacked //r//-gemination in berry and carry, and hence do not sound them like hair#y.  (The # keeps the non-satellite vowel in they're apart from the vowel in their, which in turn is different from the lower vowel in there in the Southern States.) 

      There prevails such a lack of sophistication in many aspects of analysis.  Even phonetic manuals do not notice the positions in which the English underlying fricatives become stops--or in which stops become fricatives.  (See the volume cited above.)   It would be valuable for language-learners to point out that in emphatic that and (non-emphatic) synthetic, the "th" sounds of English are pretty much likedental [d] and [t] in most European languages other than Greek and Spanish.