Summary ∵ The Local Dialect of Northwest Paphos : Phonological Description [ in Greek ] ( Aristotle University of Thessaloniki , 2015 )

In this contribution, I offer a summary of my 2015 Ph.D. dissertation from the Aristotle University of Thessaloniki on the phonological description of the local dialect of Northwest Paphos (Cyprus).


Introduction-Methodology
In this study, I provide a description of the local variety of Northwest Paphos,1 spoken on the west coast of Cyprus, with special focus on its phonological system.This is the first attempt at an in-depth description of a Cypriot Greek (cg) local variety and is mainly based on data collected by field research from 2008 to 2014 and data that I was allowed to transcribe from the Institute of Modern Greek Studies (Thessaloniki) and the Cyprus Research Center (Nicosia).I adopt the classic generative/derivational model (Chomsky & Halle 1968), assuming intermediate levels between an underlying representation (ur) and a surface form.Autosegmental representations are also of great importance throughout the thesis.I use an autosegmental model based on Revised Articulator Theory (Halle et al. 2000) for the phonological processes I describe.

2
Review of the Study

Aims
The main aim of this study is to thoroughly examine the phonological system of the local variety of nw Paphos.Apart from this, two additional aims are included: first, reconsideration of recent issues of cg that drew the attention of phonological theory, e.g.glide hardening, continuant dissimilation etc.Second, finding the phenomena that not only highlight the profile of this particular variety, but that can also be used as isoglosses for a classification of the geographical varieties spoken on the island.

2.2
Structure of the Thesis Apart from the introductory and concluding chapters (chapters 1 and 9 respectively), this thesis comprises six phonology chapters (chapters 3-8) and a chapter where I provide a concise grammar of the variety, as well as a description of other levels of analysis (chapter 2).Here, I first present the six phonology chapters and I then provide some general information about the other aspects of the study.Due to space limitations the description is confined only to the most important findings of each chapter.

2.2.1
Phonology Chapters In chapter 3, I present the vowel and consonantal inventory of the local dialect and discuss which distinctive features (dfs) describe each segment more efficiently.Some of the more important remarks of this chapter are: front vowels and palatal consonants (including glide /j/) are [dorsal] and [+front]; glottal [h] is [-sonorant]; affricates are monosegmental [-continuant], [+strident]; palatoalveolars are [coronal], [-anterior], but also [+front] (see review of chapter 8); all sonorants (nasals, rhotics and laterals) are [+continuant]; tap [ɾ] and trill [r] share the same dfs and differ only regarding their timing: [ɾ] consists of one timing unit (or x), whereas [r] consists of two timing units, i.e. it is a geminate.
In chapter 4, I briefly explain why syllable-based approaches of several phonological phenomena of cg (Malikouti-Drachman 1987, 2001, 2003;Drachman & Malikouti-Drachman 1997;Harris 1996;Christodoulou 2007, among others) are in fact inadequate.Some of the main arguments against syllabic approaches and in favour of a purely phonotactic approach are presented in the following paragraphs.
Chapter 5 concerns procedures related to consonant sequences.Two-consonant sequences are divided into two subgroups regarding their c 2 : c + sonorant and c + obstruent.In the latter category, I propose two templates regarding the df [continuant]: The first template covers the well-known continuant dissimilations (2a and 2b), sequences with /r/ + c (2c) and cases of segments that mutually change their [continuant] values in a stop + fricative sequence (2d).
( 2  Perhaps the most problematic issue emerging from syllabic accounts is that they fail to justify what consonant must be deleted in each case.For example, in a [stç] sequence (as shown above) there is no justification why c2 is deleted, since c1 deletion could have created a Coda + Onset sequence as well.Such cases suggest that the presence of phonotactics is necessary for dealing with simplification phenomena.
In chapter 7, I present processes affecting intervocalic or initial prevocalic consonants.Perhaps the most important finding during the examination of these processes is the clear tendency of fricatives towards the designated articulation [dorsal], another phenomenon that remains, so far, unreported in the literature.
( ' I propose that these three phenomena are connected through the df [+front]: palatal consonants (including glide /j/), palatoalveolars and front vowels form a natural class sharing this particular df.Combined with the df [dorsal], [+front] denotes consonants produced with the front of the tongue body touching the hard palate, i.e. palatals.On the other hand, combined with the df [coronal], [+front] denotes consonants produced with the blade touching the area behind the alveolar ridge, while the front of the tongue body approaches the hard palate, without touching it, i.e. palatoalveolars.8 Having these assumptions in mind, I consider velar palatalization as a case of assimilation, where [+front] is spread from a front vowel to its previous velar.During coronalization, a (dorsal) palatal changes into a (coronal) palatoalveolar, without losing its df [+front].Coronalization is not considered as an assimilation process, as it may occur before any vowel (e.g./paxjˈa/ → paçˈa → [paʃˈa] 'fatadj.fem.').Studies on similar phenomena in other languages show that coronalization is related to the special (articulatory, acoustic, perceptual and even typological) properties of palatal consonants, not the spreading of some df by an adjacent segment.Finally, alveolar palatalization is considered as a kind of fusion, as the output of a sequence alveolar + /j/ is a [+front]alveolar or palatal-segment.

2.2.2
Non-phonological Aspects of the Study In a special chapter (chapter 2), I provide a general account of the main typological characteristics of the local dialect, describing its morphology and syntax as well.Through this examination, I was able to identify several phenomena not reported in the literature to date, for instance, the special use of morpheme /-ˈeni/ denoting the main ingredient of a pie, e.g.[kolotʃˈeni] 'pumpkin pie' (cf.[kolokotˈi] / [kolokˈopitːʰa] in other cg varieties), and the use of /ˈien/ + (non-past) perfective periphrasis, often preceded by temporal con-junction /ˈaman/ 'when' .The whole periphrasis 'ˈaman ˈien + perfective' denotes a habitual action in the past (8a).Interestingly, the negative particle /mˈen/ and clitics always appear in-between the members of the periphrasis (8b).( 8 In each chapter, I also make observations regarding the geographical distribution of several cg phenomena.As regards nw Paphos, despite the fact that some phenomena occur exclusively in this area, the great number of characteristics being shared with other local dialects suggests that this particular variety belongs (or belonged) to a broader western zone within the island.

Conclusion-General Remarks
In this thesis, I provide a description of the phonological system of the nw Paphos local dialect.Through this examination, I was able to reveal some new aspects-not only phonological-of cg, reconsider some previous accounts on various phonological phenomena, and make some preliminary remarks on cg dialect geography.
In syllable-based approaches(Malikouti-Drachman 1987: 282; Drachman &  Malikouti-Drachman 1997: 59-60, among others), c + obstruent sequences are considered heterosyllabic, i.e. c1 takes the Coda position and c2 takes the Onset position.Moreover, segments situated in Coda must not be stronger than Onset segments.3Syllabicaccounts cannot adequately explain phonological processes involving strident consonants, as their outputs are not predicted by the aforementioned syllabic restrictions: according to those accounts an ur /kaluptʃˈis/ should have been realized as *[kaluftʃ ( ʰ ) ˈis], not [kalupʃ ˈis], and Turkish bohça should have been adjusted as *[poxtʃ ( ʰ ) ˈas], not [pokʃ ˈas].Syllabic accounts face similar problems that have to do with three-consonant sequences as well.