Phonological Backdrop

  • François Dell
  • Mohamed Elmedlaoui
Part of the Kluwer International Handbooks of Linguistics book series (KIHL, volume 2)


In the first half of this chapter (sections 3.1 to 3.5) we review the properties of the geminates. In the remaining sections we present information about several phonological phenomena of Imdlawn Tashlhiyt. These phenomena are not central to the concerns of this book but their effects are seen in many examples and they come into play at one point or another of our discussion.


Sugar Assimilation Gall Posit Bark 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    V. Leben (1980) and McCarthy (1979, 1981). For recent overviews, v. Inkelas and Cho (1993), Kenstowicz (1994a), Broselow (1995) and Perlmutter (1995).Google Scholar
  2. 2.
    Concerning the internal structure of segments we shall adopt the version of ‘feature geometry’ advocated in Clements and Hume (1995), in which all the features defining a segment are organized in a tree structure. The Root node is the node dominating the whole tree.Google Scholar
  3. 3.
    Willms (1962) had already reached the same conclusion for Kabyle, basing himself on evidence about the phonotactics of Kabyle syllables.Google Scholar
  4. 4.
    Verb-final geminates are split by the ‘chameleon’, a vowel which is inserted to form certain imperfective stems, v. DE (1991: 96) and § 5.2 in this book.Google Scholar
  5. 5.
    On schwa epenthesis in Rifian Berber, v. § 6.5.Google Scholar
  6. 6.
    ’In general’, that is: unless the first consonant is released, which happens only in certain contexts, about which v. § 6.3.3.Google Scholar
  7. 7.
    There are a few exceptions to this statement. After a verb, for instance, a suffix or clitic comprised of one short consonant which immediately follows its long counterpart merges with it, e.g. /dd=d/ must be pronounced as though it were /dd/ in /s-bidd=d a-gždi/ (cau-stand=dir u(beam) ‘raise the beam on this side!’.Google Scholar
  8. 8.
    V. § for details.Google Scholar
  9. 9.
    Abdel-Massih (1968: 127ff) posits various instantiations of (7) for particular consonant sequences in Ait Ayache Tamazight.Google Scholar
  10. 10.
    Evidently, someone upholding the featural analysis would rather talk about a problem of tenseness.Google Scholar
  11. 11.
    The special fusion exemplified in note 7 does not weaken our argument against (7). In order to account for this fusion, (5) and (7) must be supplemented with devices of comparable complexity.Google Scholar
  12. 12.
    The formulation in (10) is only a first approximation, v. § 6.3.3.Google Scholar
  13. 13.
    After a vowel the geminated high vocoid created by assimilation is pronounced as a geminate glide, e.g. t-a-wada n u-yyul / tawada w w-yyul ‘the donkey’s gait’.Google Scholar
  14. 14.
    On Stray Erasure v. e.g. Marantz (1982: 446).Google Scholar
  15. 15.
    Ouakrim (1995: 102) claims that in Ihahan Tashlhiyt tense consonants (what we call geminates) cannot be ambisyllabic. Take t-nna’ she said’, which is disyllabic in Imdlawn Tashlhiyt ( If it is also disyllabic in Ihahan, the only syllabifications compatible with Ouakrim’s claim are t.nna and tnn.a.Google Scholar
  16. 16.
    In free variation with uKwkwim. In Imdlawn Tashlhiyt labialized consonants optionally delabialize when they occur to the right of a rounded vowel belonging to the same word. We only give the delabialized variants in order to avoid cluttering.Google Scholar
  17. 17.
    V. DE (1991: 96-99), DE (1992).Google Scholar
  18. 18.
    /ww/ is also realized as gwgw in certain morphological environments, v. § 7.4.Google Scholar
  19. 19.
    On labial dissimilation, v. Elmedlaoui (1995a: 43-78).Google Scholar
  20. 20.
    Hayes (1990) and Selkirk (1990) discuss other problems with that conception.Google Scholar
  21. 21.
    In chapter three, published as Saib (1977).Google Scholar
  22. 22.
    E.g. Chaker (1984: 84-85). Kossmann (1994: 59-60) poses problem (i) and discusses various relevant facts in Figuig, but in the end he leaves the question unanswered.Google Scholar
  23. 23.
    Published with slight revisions as Ouakrim (1995).Google Scholar
  24. 24.
    V. Elmedlaoui (1993) for some discussion.Google Scholar
  25. 25.
    In that article we also examine an argument by Ouakrim (1995) based on duration measurements in Ihahan Tashlhiyt. We suggest that the author’s conclusions probably rest on a faulty phonological analysis of the data on which the measurements were made.Google Scholar
  26. 26.
    On emphasis in Tashlhiyt, v. Elmedlaoui (1985), Boukous (1987a) and the references cited in these works.Google Scholar
  27. 27.
    In §3.6, emphasis is indicated by underlining or by ‘!’. The two are equivalent.Google Scholar
  28. 28.
    This sound is identical with the emphatic variant of /i/ in Moroccan Arabic, which various authors describe as [e], erroneously in our opinion.Google Scholar
  29. 29.
    /a/ is furthermore subject to a phenomenon of prepausal backing. Immediately before a pause it is realized more posterior than its nonprepausal counterpart, e.g. whereas /sala-n/ ‘they finished’ is realized as [sælæn], before a pause /sala/ ‘finish!’ sounds more or less like [sæla] (v. Heath (1987: 23) for a similar phenomenon in Moroccan Arabic). Prepausal backing also occurs in emphatic contexts, e.g. /a/ in prepausal /t-!bda/’ she shared’ is even more posterior than /a/ in /!bda-n/ ‘they shared’. Prepausal backing does not obliterate the difference between emphatic a and nonemphatic a; the vowel in /t-!bda/’ she shared’ and that in /t-bda/’ she began’ are both back a’s, but that in the first word sounds more back than that in the second. Prepausal backing and emphatisation have the same influence on aa as on a. The only audible difference between a and aa in the various contexts where they contrast is one in duration. On aa, see the next section.Google Scholar
  30. 30.
    We are unable to determine whether the long t which immediately precedes [!š:] is emphatic or not.Google Scholar
  31. 31.
    The existence of a marked audible difference between the plain variant of the pharyngeal fricative /ħ/ and its emphatic variant is evidence that the secondary articulation of Imdlawn Tashlhiyt which we are calling emphasis is not mere pharyngealization.Google Scholar
  32. 32.
    On the other hand the difference between [z] and [!z] is easy to perceive, even in the absence of an ajacent vowel.Google Scholar
  33. 33.
    Literally ‘eat him with a third’.Google Scholar
  34. 34.
    A few nouns have singular and plural kernels which do not agree with respect to emphasis, e.g. a-ydi ‘dog’ vs. !i-yda-n ‘dogs’. There are also a few nouns in which emphasis and the lack of it are in free variation, e.g. a-frux or !a-frux ‘boy’, t-i-frx-in or !t-i-frx-in ‘girls’.Google Scholar
  35. 35.
    We use the term ‘root’ merely as an expository convenience.Google Scholar
  36. 36.
    In !tizukatin the stem-final t is epenthetic, v. § 2.5, note 20. The corresponding ms noun is !a-zuka.Google Scholar
  37. 37.
    In ikid the degree of emphasis seems to increase gradually during the time course of the second vowel, which sounds like emphatic [i] near the end. The vowel gives the impression of beginning as a plain i, an impression which may in part be due to the fact that the preceding consonant is palatalized ([ky]). While /k/ and /g/ regularly palatalize before /i/ and /a/ in nonemphatic contexts, their palatalized variants never occur inside an emphasis span. For the speakers’ categorial perception, the semi-emphatic i in ikid counts as a plain vowel, like the plain i in the last syllable in i-zri=k=id (3ms-overtake=do2ms=dir) ‘he overtook you (coming hither)’.Google Scholar
  38. 38.
    The sentence has another acceptable pronunciation, in which the second emphatic span is rabukad.Google Scholar
  39. 39.
    Pronouncing iyismmidasit is not altogether inacceptable, but feels far less natural.Google Scholar
  40. 40.
    i before the suffix is epenthetic (v. note 36), which explains why it can be nonemphatic, in apparent contradiction with (31)b.Google Scholar
  41. 41.
    From more abstract /!bda=ax/, with a hiatus-breaking yod.Google Scholar
  42. 42.
    A possible exception is a-raam ‘camel’ (p i-raaman); this word has no close phonetic analogue in Arabic.Google Scholar
  43. 44.
    Hasnayʕ, the plural of !ssnaat, is the only exception to rule (35) we have encountered.Google Scholar
  44. 45.
    On Berber and Arabic plurals, v. § 2.5.Google Scholar
  45. 46.
    Violations of this restriction are extremely rare, e.g. t-i-mtd-in ‘loin (cut of meat)’. Some exceptions are only apparent, e.g. !ttd ‘coagulate’, whose underlying representation can be assumed to be /!dd-d/, on account of the fact that /!dd/ regularly surfaces as !tt. On ‘+’, v. § 6.4.1.Google Scholar
  46. 47.
    On geminate glides, see § 7.4.Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2002

Authors and Affiliations

  • François Dell
    • 1
  • Mohamed Elmedlaoui
    • 2
  1. 1.EHESS-CNRSParisFrance
  2. 2.Faculté des LettresOujdaMorocco

Personalised recommendations