Loading ...
Sorry, an error occurred while loading the content.

322Re: Palatalization and Syllabification in Quenya

Expand Messages
  • Pavel Iosad
    Feb 19, 2003
    • 0 Attachment
      David Kiltz has set forth a theory proposing how one can resolve the
      difficulty regarding the apparent violations of the syllable weight
      limitation. While I agree with the principle, I have to point out
      several issues that, in my opinion, need more dwelling upon.

      [...]
      > As the examples adduced by Pavel Iosad and Ales Bican show
      > combinations such as Cy and Ry, when deriving from C+y, R+y,
      > are biphonemic in Quenya and form long syllables.
      > As phonetics tell us, a syllable ideally starts with a higher
      > degree of consonanticity followed by a sound with a lower value.
      > A somewhat simplified consonanticity hierarchy may look like
      > this: Occlusives > Spirants (Sibilants, s having a special status) >
      > Nasals > L,R > "half-vowels". We may call these sounds then class
      > 1,2,3,4, and 5.

      While this is good, I feel compelled to point out that the special status
      of 's' appears to be a specific Indo-European feature, provoked, I'd
      suggest, by the special status of 's' in the PIE sound system, it being
      the only non-laryngeal fricative. This markedness of the sound and its
      important role in the root-building process (witness the 's-mobile')
      give it a certain special status. 's' is not always the most consonant
      sound. While I cannot think of any non-Indo-European examples off the
      top of my head (I'm not taking modern Slavic languages into account
      here, since they have killed the IE syllable structure, witness a word
      like Polish _pstra,g_ 'salmon'), but Ancient Greek, for instance allows
      initial combinations of the _muta cum liquida_ type (_prosêkei_, _khraô_
      etc. etc.), sC ones (_sphallô_) *and* Cs - _pseudomai_. Thus the
      consonanticity status of 's' is actually ambiguous. It is also
      demonstrated by the fact that Romance languages (and not only Romance)
      did develop a prosthetic vowel before Latin initial sT(R)- groups (Fr.
      _écrire_, Sp. _escrebo_, Welsh _ysgrifennu_), because a sT(R)- onset is
      somehow uncomfortable (compare this with the appearance of the
      prosthetic vowel before undoubtedly offending clusters, as in Russian
      _rzhanoj_ 'of rye', Byelorussian _arzhany_, Rus. _mgla_ 'mirk', Byel.
      _imgla_). It must also be noted that a Latin word like _magister_ is
      stressed _magíster_, not *_mágister_ as it would were the medial -st-
      tautosyllabic (i. e. fully identical to the _muta cum liquida_ group in
      _emígrô_).

      > Word-initially Quenya seems to allow for the following combinations:
      > 1 + 5 (cf. _tyulusse, tyálie_),
      > 2 + 5 (cf. _hyarin_ < SWAR-. As "h" probably represents [ç] here, it
      > should be treated as a spirant which, historically, it certainly is.),
      > 3 + 5 (cf. _nyello_),
      > 4 + 5 (doubtful, _lyenna_ ?).
      [...]
      > Given the attested Quenya words, I will argue that Quenya in principle
      > honours the biphonemic rule but may, under certain historical
      > circumstances, allow for a different syllabification.
      > Case 1) The obvious instance where e.g. _ty_ is primary and hence
      > monophonemic. Cf. _intya-_.

      Indeed, if the _Cy_'s are monophonemic, this is not much of a problem

      > Case 2) _máryat_. In my view this is not a violation of the biphonemic
      > rule but has to be syllabified as má (<ma3-) +ryat. The division is
      > due to morphological reasons and is phonetically permissible since a
      > syllable-initial cluster "ry" is in accord with the consonanticity
      > hierarchy for syllable-onsets outlined above.

      While it certainly is in accord with the hierarchy, it must be mentioned
      that Quenya does avoid clusters of that type, because, as David notes,
      the phonetic rules do require that a consonant cluster where the first
      element is more consonantic than the first one be tautosyllabic (i. e.
      the syllable boundary should not split it). VT42:26 says:

      '...the strong predilection which Quenya showed for the sequences of
      sonants: _m_, _n_, _ñ_; _l_, _r_ before stops, as against those in which
      the sonants followed. Transposition also occurs in Quenya in ancient
      forms of _tr_, _tl_, etc. > _rt_, _lt_'

      With this confer Quenya _alcar_ 'glory' from AKLA-R, cognate with
      Noldorin/Sindarin _aglar_ (V:348). The very point here is that the
      sequences, being of the 1 + 3 or 1 + 4 types (in David's notation),
      require to be regarded as parts of one syllable. But the reversal argues
      to support the point made in IX:417-8, that Quenya does not tolerate
      onsets of more than segment (confer the lack of the reversal in
      Sindarin, which does allow initial mutae cum liquidis), even if they
      comply with the consonanticity/sonority hierarchy rules.

      > Now for the case of _aistana-_.
      > In "The Etymologies" one can see that PQ (or PE) allowed for a wide
      > range of s + C clusters initially.

      However, if one suggests that Quenya did so as well, one may well wonder
      why the initial _st-_ was simplified to _s_. (I realise this is however
      weak as an argument, being _ex nihilo_).

      > Excursus: On the peculiarity of "s" in consonant clusters.

      [snip excursus on the special behaviour of _s_ in IE]

      > The special behaviour of "s" that can be seen e.g. in Indo-European
      > Also seems to feature in PQ (PE).

      Now there is also the question of what to allow as initial PQ clusters.
      There is no compelling need to suggest initial /MB ND ÑG/ are single
      phonemes in PQ. The African languages where such consonants are viewed
      as single phonemes do not allow any other initial clusters, which is not
      the case in PQ. There is also a solution in the lines of Modern Greek,
      where the /mp nt ng/ used to substitute voiced stops in loans, as in
      _mpar_ 'bar', but as opposed to the /mp/ type, /mb/ cluster are
      semiotically suspect (just as the traditional PIE reconstruction, which
      can be a strong argument *for*).

      > As I argued in the case of _máryat_, Quenya seems to allow certain
      > PQaic clusters at the onset of a syllable which it has otherwise
      > simplified word-initially.

      It can, apparently, true; but I am at a loss to reconcile this with the
      stress in _hiruvalye_.

      > Hence, I would suggest that this is the case for "st", too.
      > _Aistana-_ is therefore to be syllabified as _ai-stana_.

      ...and _Hrísto_ as _Hrí-sto_. This would also explain the syncope which
      must have given the form _aistana_ rather than *_aiastana_.

      Alternatively, one can suggest that in _aistana_ that _ai_ is not a
      diphthong, but rather two vowels on the model of _oïkta_ in Narqelion
      (see Christopher Gilson's article in VT40). The only problem with this
      is that Tolkien did not mark it as such. Only too bad.

      In spite of the above criticism, I think this is very much possible
      theoretically and even not unlikely (in fact, I implicitly suggested
      this in my latest post ('Or are we dealing with a special status of the
      _st_ group [...]?)), and can be reconciled with the explanation I have
      been propounding. One can then ascribe a special (near-monophonemic?)
      status to the _st_ group and describe the lack of bisegmental initial
      realization as a marked situation. This however raises several problems,
      such as lack of a _sC_ in words from SC-initial roots when that group is
      intervocalic (_Nurufantur_ rather than *_Nuruspantur_. Overall, this is
      a complex case (surprise, surprise...).

      Pavel
      --
      Pavel Iosad pavel_iosad@...

      Is mall a mharcaicheas am fear a bheachdaicheas
      --Scottish proverb
    • Show all 22 messages in this topic