Laura Bafile

Segmental phonology and the foot: the case of proparoxytones in some Italian dialects
Laura Bafile
University of Ferrara
It is generally acknowledged that the foot is not simply a phonological unit dedicated to the
organisation of stress prominence, but is also a relevant domain for some segmental phenomena
concerning both phonotactics and phonological processes (cf. Nespor & Vogel 1986; Harris 1994;
Hammond 2012). As a domain of segmental phonology, the foot is not simply required by the fact
that stress may determine phonetic quality of vowels and consonants contained in stressed syllables;
it has been shown that a constituent wider than the syllable but smaller than the word is necessary to
account for different phenomena in different languages. For instance, Harris (2013 [2006]) shows
that, in some British varieties, t lenition as well as h and r deletion cannot be confined in the
syllable, finding their appropriate domain in the foot.
In Italian varieties, main stress falls in most cases on the penultimate syllable on the word.
However, since in a fair number of words stress comes on the antepenultimate syllable, the
definition of metrical constituency in those languages is somewhat controversial. While it is
generally assumed that bounded feet are maximally binary (cf. Hayes 1995) and, therefore, that
trochee is the foot corresponding to Italian stress pattern, some authors have proposed that metrical
structure should be enlarged to admit ternary constituents (cf. Nespor 1993, Thornton 1996).
However, the two positions do not make different predictions about stress, since in those varieties
stress placement is lexically determined. In fact, stress may fall on one of the last three syllables of
the word, and in few, exceptional cases, even on the fourth from last syllable, without provoking
metrical restructuring: e.g. péna ‘pain’, péttina ‘he combs’, péttinano ‘they comb’.
More interestingly, in some Italian dialects, the actual size of stress domain may have consequences
for the segmental structure of the word.
In this paper, I will take into account data from different Italian dialects, illustrating three kind of
phenomena affecting proparoxytones.
i) Syncope and epenthesis
In some dialects of Piedmont and Emilia-Romagna, non-final unstressed vowels delete both in the
pretonic and in the post-tonic section of the word. As for the post-tonic syncope, the deletion of the
non-final vowel is possible only if the final one is preserved; when the final is deleted, a default
vowel appears in the penultimate syllable:
(1)
védva védav (cf. Standard Italian védova ,védovo ‘widow (f., m.)’)
I will adopt here an analysis based on the idea that phonological structure may contain empty
constituents (cf. among others Kaye 1990); thus, the structure at the base of both forms in (1)
contains an empty nucleus in either final or penultimate position:
(2)
védØva
védavØ
Since the content of the last nucleus determines the content of the preceding one, the relevant
domain of syncope must be a foot that contains the stressed vowel and the unstressed ones including
the final, i.e. a trisyllabic foot.
ii) Vowel harmony
Some dialects of Central and Southern Italy show various kinds of vowel harmony (VH). In a kind
of regressive VH, the unstressed penultimate vowel becomes identical to the final vowel, e.g.
Servigliano (Marche)
ˈprɛdoko, ˈprediki, ˈprɛdaka ‘I/you/he preach/es’ (Maiden 1997). In the
dialect of Vallepietra (Lazio) analysed in Schirru (2012), a mid stressed vowel triggers the VH of
post-tonic vowels including the final, that becomes -o instead of -u: j ˈarberu ‘the tree’, but ju
ˈvedovo. In both cases, VH affects a section of the word that corresponds to the foot and contains
three syllables.
iii) ‘Trisyllabic shortening’
Many dialects, from Northern to Southern Italy, exhibit qualitative effects of quantity in vowels
occurring in open stressed syllables, often producing diphthongization. However, stressed vowels of
proparoxytones generally behave like vowels of closed syllables, e.g. ˈkrajtə ‘I believe’, ˈlɛŋgwa
‘tongue’, ˈkrɛdənə ‘they believe’ (Marotta&Savoia 1994). In the last form, shortening is not
conditioned by syllable structure, but rather by foot structure.
The paper will analyse data of the kind illustrated above and propose a formalisation of the
relationships among nuclei inside stress domains, in order to account for the special behaviour of
proparoxitones, as well as for locality requirement holding in formal models of phonological theory.
Hammond, M. (2012) “The Foot” in M. van Oostendorp, C. J. Ewen, E. Hume and K. Rice (eds.),
The Blackwell Companion to Phonology, vol II, ch. 40, Blackwell.
Harris, J. (2013 [2006]) “Wide-domain r-effects in English”, Journal of Linguistics 49.2: 329-365.
Harris, J. (1994) English sound structure, Oxford: Blackwell.
Hayes, B. (1995) Metrical stress theory: principles and case studies, Chicago: University of
Chicago Press.
Kaye, J. (1990) “Government in phonology. the case of Moroccan Arabic”, The Linguistic Review
6: 131-159.
Maiden, M. (1986) “Vowel systems” in Maiden, M. & Parry, M. (eds.) The dialects of Italy,
London: Routledge.
Marotta, G. & L.M. Savoia (1994) “Vowel properties and nuclear constituents: evidence from
Italian dialects” Probus 6: 43-79.
Nespor, M (1993) Fonologia, Bologna: Il Mulino.
Nespor, M. & I. Vogel (1986) Prosodic phonology, Dordrecht: Foris.
Schirru, G. (2012) “Osservazioni sull’armonia vocalica nei dialetti della Valle dell’Aniene e in
quelli dei Monti Aurunci”, in M. Loporcaro, V. Faraoni & P.A. Di Pretoro (eds.) Vicende
storiche della lingua di Roma, Alessandria: Edizioni dell’Orso.
Thornton, A.M. (1996) “On some phenomena of prosodic morphology in Italian: accorciamenti,
hypocoristics and prosodic delimitatio”, Probus 8: 81-112.