|Title:||Final Posting Re: issue 16-2712, Mysterious /s/|
|Description:||Many thanks to everyone who responded to my posting about the properties of /s/ with references, data and ideas. They were:
Rosemary Beam de Azcona, Pier Marco Bertinetto, Ed Burstynsky, Fred Cummins, Annette Fox, Dinha Gorgis, Claire Gurski, Damien Hall, Patrick Honeybone, Mark Jones, Arsalan Kahnemuyipour, Susannah Levi, Kristie McCrary Kambourakis, Bruce Moren, Faisal Al-Mohanna, Kathy Snow, Mark Southern, Nora Wiedenmann, Guido Vanden Wyngaerd and Bettina Zeisler.
I apologise that I have not had the chance to reply to all of you personally.
My original post can be found as item 16.2712 in the 2005 'DISC' section of
Respondents focussed on possible explanations of the unusual distribution of /s/, which in some languages occupies a position at the begining of word
initial consonant clusters denied to other C segments. I had offered, roughly, three alternatives based around articulatory constraints, universal grammar and language history.
Though most agreed that a simple explanation of the observed phenomena would continue to be elusive, the balance of opinion was tipped towards an
Mark Jones and Kristie McCrary Kambourakis pointed out that /s/ is extremely salient in acoustic terms. Jones noted that, in terms of its intensity, /s/ has a 'stand alone' status. Susannah Levi pointed out that ''fricatives have internal cues (to place and manner) that allow them to be perceived before a stop (whereas stop-stop clusters are more difficult to perceive since cues to their place and manner are largely found in their transitions with neighbouring sonorants)''.
Mark Jones noted that the /s/ in Spanish, in which word initial s+C is
prohibited, may be less intense than realisations of this segment in other
languages, and a possible effect of this may be the need to 'prop it up' with a prothetic vowel.
Patrick Honeybone recalled the tradition of treating s+C as single segments, in the same way as affricates. This view was echoed by Claire Gurski.
Several respondents pointed out that evidence from many languages (such as
definite article alternation in Italian) suggests that /s/ should be
syllabified in the coda. In this case the preceding nucleus is either realised or unrealised, resulting in the dichotomy between initial s+C and V+s+C languages. This is essentially the approached outlined by Jonathan Kaye in his 1992 paper 'Do you believe in Magic - the story of s+CC clusters' (SOAS Working Papers in Linguistics and Phonetics) and taken up by others working in the Government Phonology framework. It does not offer an explanation of a) why it is /s/ that behaves in this way and b) why the nucleus is realised or unrealised in various langauges. It makes an appeal to parameter settings which, as one respondent commented, is ''non-explanatory and hence a complete non-runner''.
Fred Cummins felt the it is the sonority-sequencing hypothesis as traditionally construed, which apears to be violated in these cases, that may be in need of adjustment.
Kathy Snow explained the variation between s+C and V+s+C in French (espace v. spacieux) in terms of language history and the relative age of these loans.
Bettina Zeisler gave examples of word intial s+C in Sino-Tibetan and
Damien Hall and Mark Jones noted the interesting questions posed by the alternation /s/ [h] zero in Spanish.
A number of respondents gave useful references which are listed (in the form they were given) at the end of this email.
In general it was interesting to see the all-too-familiar difference in approach of phoneticians and phonologists raise its head. Most of the former, as one would expect, saw the problem in terms of the physical properties of the segment /s/ and rather than any language internal (phonological) constraints.
Having had time to think a little more about this problem, and in view of the discussion, I would attribute the seemingly anomalous /s/ to its unique
acoustic properties. I think, though, that the essential problem of why some languages consistently have a prothetic vowel whereas others don't need one remains (though perhaps it will be cleared up when I get round to reading some of the references below!). I don't think it's just down to the phonetic properties of the segment /s/ in the language in question - there are many examples of languages that need an prothetic vowel other than Spanish, which may well have a phonetically 'weak' /s/. And positing a different ranking of constraints within Optimality theory doesn't really get any further than positing parameter settings: it's a way of formalising the facts, but it's not really explanatory.
References from respondents:
Pier Marco Bertinetto, 2004 ''On the undecidable syllabification of /sC/
clusters in Italian: Converging experimental evidence'' Italian Journal of Linguistics/Rivista di Linguistica 16.2
Harris, 1994, ''English Sound Structure''.
Iverson and Salmons, 1999 ''Glottal spreading bias in Germanic''. Linguistische Berichte.
Kristie McCrary Kambourakis, 2004, ''Reassessing the role of the syllable in Italian Phonology'', UCLA dissertation:
Frida Morelli, 1999 ''The phonotactics and Phonology of Obstruent Clusters in Optimality Theory'', University of Maryland at College Park dissertation.
Mark Southern, 1999 ''Subgrammatical Survival: S-mobile in Indo-European and Germanic'', Journal of IE Studies Monograph Series.
Richard Wright, UCLA dissertation on clusters in Tso
Richard Wright, ''Perceptual Cue Robustness and Phonotactic Constraints:
Rethinking Sonority'' in ''Phonetically Based Phonology'' Steriade, Kirchner, Hayes (eds)
Wenckje Yongstra, University of Toronto (Department of Linguistics) dissertation on acquisition of s+C clusters.