<<

chapter 8 Subtractive Imperative Forms in Bithynian Greek

Nikos Koutsoukos and Nikolaos Pantelidis

1 Introduction

Pursuing the goals of this volume, in this chapter we deal with Bithynian Greek. Bithynian Greek is not one of the well-studied varieties of Modern Greek, un- like Cappadocian or Pontic. However, it offers a number of interesting phe- nomena at the level of morphology. This variety was spoken in the area of and later on was preserved by immigrants who moved to various places in the mainland of Greece, such as Nea Kios (Νέα Κίος, Peloponnese) and Nea Triglia (Νέα Τρίγλια, Chalkidiki), after the exodus of the Greek population in 1922–1923. There are three ques- tions regarding the linguistic context of this area: (a) the relationship between the Bithynian sub-varieties and the hypothesis of a well-defined distinct Bithynian dialect group, (b) the relationship between the Bithynian varieties (or the possible dialect group) and other Modern Greek varieties, and (c) the relationship between Greek and other languages spoken in this region, espe- cially Turkish (see also Manolessou, this volume). These questions show that the study of Bithynian Greek is intriguing and deserves a closer look. Within this context, we describe the verbal system of this variety. In gen- eral, the variety follows the same patterns as most of the Modern Greek va- rieties with respect to the structure of the verbal system. In our description, we focus on the formation of the imperative. Imperatives in Bithynian Greek are characterized by subtraction in the perfective of the active in verbs with the formatives -ών(ω) [ono] and -άζ(ω) [azo]. Since subtraction has not been discussed with respect to Modern Greek varieties, we analyze the cases of subtraction in Bithynian Greek and we compare it with two other varieties, namely Standard Modern Greek (SMG) and the Peloponnesian variety. Subtraction as a morphological process per se also raises a number of the- oretical questions. For instance, an interesting question is what the limits of subtraction are and how it is differentiated from other processes which share an abbreviatory mechanism. Subtraction is also connected with the question of the relationship between form and meaning in morphological processes, since ‒ contrary to most of the common processes ‒ subtraction consists in the shortening of the shape of a word form and not in the addition of some

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004394506_010

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 255 13-Feb-19 6:10:26 PM 256 Koutsoukos and Pantelidis

material (e.g. affix) which signals the process and the new meaning. The analy- sis of subtractive forms in Bithynian Greek revolves around two basic issues: first, the examination of the morphological context of this change, namely why subtraction appears only in verbs which show these two formatives (-ών(ω) [ono] and -άζ(ω) [azo]); and second, the motives behind this process. To cover these issues, the structure of this chapter is as follows: after this in- troduction, we explore the linguistic context of the Bithynian dialect (section 2) and we describe the relevant data (section 3). In this section, we start with some preliminary notes on the verbal system of the dialect (section 3.1) and we focus on the subtracted verbal forms in comparison to similar data from other varieties (section 3.2). After that, we explore the of subtraction (section 4.1) and we analyze the subtracted forms in Bithynian (section 4.2). Conclusions are presented in the last part of this chapter (section 5).

2 Linguistic Context of Bithynia

2.1 Sociohistorical Notes on the Variety Geographically, Bithynia occupies the northwest of Asia Minor, on the south- ern shores of the Propontis (see map in Annex). In antiquity, it bordered on Mysia (to the west-southwest), Paphlagonia (to the northeast along the coast), and Phrygia (to the south-southeast towards the interior of Asia Minor). However, the borders of this region have not been very rigid (Sakellariou, 1990–1991). Modern Bithynia comprises ancient Mysia and the western part of the ancient region and Roman province of Bithynia.1 Several important cities were (re-)founded there in various epochs from the begin- ning of the Greek colonization (around 700 BC), among others: Kios (Κίος or Gemlik), Chryssoupolis (Χρυσούπολις or Üsküdar), Nicomedea (Νικομήδεια), Nicaea (Νίκαια or İznik), Bursa (Προύσα), (Χαλκηδών or Kadiköy). Bithynia was one of the linguistically most complex areas of Asia Minor. A process of Hellenization, which started with the conquest of the region by Alexander the Great in the late 4th century BC, led to the gradual extinction of the indigenous languages, among others, Phrygian and Mysian, and their replacement by Koine Greek. It is estimated that this process lasted several centuries, since some languages, such as Phrygian, seem to have survived in enclaves as late as the 7th century AD (Brixhe, 2008, p. 72; Brixhe, 2010; Hawkins, 2010, p. 215). Greek was the dominant language of western, central, and northern until at least the end of the 11th century AD (Vryonis, 1971, pp. 42–55). It has

1 See map in Brixhe (2010, p. 229 Fig. 16.1) and Manolessou (in this volume).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 256 13-Feb-19 6:10:27 PM SUBTRACTION IN BITHYNIAN 257

been in recession since the Turkish invasions and the beginning of Turkish settlement in Asia Minor (11th–14th centuries AD), but it survived in parts of western Asia Minor, in the north of Asia Minor, and in enclaves in Cappadocia and Lycaonia. During the time of the Ottoman Empire, Bithynia was ethnically and lin- guistically heterogeneous. Turks and Greeks made up the majority of the pop- ulation, while Jews, Armenians, Circassians, Slavs, and Tatars were also present in the same area. Parts of the Greek population were Turkish-speaking or Greek-Turkish bilinguals, a fact that seems to have had an impact on the Greek varieties of the respective settlements (Konstantinidou, 2005a, pp. 125–126; Konstantinidou, 2005b, pp. 337–338). Greek ceased to be spoken in Bithynia after the exodus of the Greek population in 1922–1923. Within this context, one of the most interesting issues is the relation be- tween Greek and other languages spoken in this region, especially Turkish. However, except for short discussions in Konstantinidou (2005a, p. 146; 2005b), this issue remains unexplored. In the pre-1923 period, Bithynia displayed a considerable degree of dialectal diversity, which was most probably connected with the fact that ‒ beside the older Greek-speaking population which can be traced back to Byzantine, Roman and Hellenistic times – up to the 19th cen- tury the region was a “melting pot” of Greek speakers from many Grecophone areas (including the Aegean islands, see Konstantinidou, 2005b, p. 337 and ref- erences therein).

2.2 General Linguistic Features Important questions raised in the research on the linguistic features of the Bithynian varieties involve (a) the interrelation between Bithynian (micro-) varieties and their classification into groups, (b) the question of whether there is a distinct, locally differentiated, Bithynian Modern Greek dialect group, and (c) the possible genetic links between Bithynian varieties (or the possible dia- lect group) and other Modern Greek varieties. Regarding the first question, according to Papadopoulou (2010, p. 21), Bithynian varieties can be classified into three groups: (a) varieties of the “northern Greek” type which are characterized by raising of unstressed mid vowels and deletion of original unstressed high vowels, for example, πουτάμ’ [putám] for ποτάμι [potámi] ‘river’, (b) varieties of the “southern Greek” type which are characterized by the absence of the above phenomena,2 (c) varieties of the “semi-northern Greek” type which stand in-between.

2 The phenomena typical of the “Northern” varieties are not completely absent from the “Southern” varieties. They may occur, but not in a systematic way; otherwise they are re- stricted to specific (phonological/morphological) contexts and lexical items.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 257 13-Feb-19 6:10:27 PM 258 Koutsoukos and Pantelidis

Due to the absence of a contrastive analysis of the different varieties of Bithynian Greek, a full listing of all their shared features is not feasible. However, below we list a number of important morphological phenomena, quoted in Danguitsis (1943) and Konstantinidou (2005a, b), that may consti- tute an interesting platform for discussion of this variety:3 (a) Leveling of the formal distinction between the masculine and feminine definite article in: – Nominative singular, η άντρας [i ádras] ‘man’ vs η γυναίκα [i ʝenéka] ‘woman’ (instead of ο άντρας [o ádras] vs η γυναίκα [i ʝenéka]) – Accusative plural τι αθρώπ’ [ti aθróp] ‘human beings’ vs τι μέρες [ti méres] ‘days’ It should also be mentioned that all three genders display the same form of the definite article in the genitive singular, e.g. τη άντρα [ti ádra] ­(masculine), τη γυναίκας [ti ʝenékas] (feminine), and τη παιδιού [ti peðʝú] (neuter). (b) Leveling of the formal distinctions in the singular of the masculine nouns which results in two forms, one for nominative and one for genitive/ac- cusative, e.g. η άθρωπος [i áθropos] nom.sg ‘man’ vs τη άθρωπο [ti áθropo] gen.sg, το(ν) άθρωπο [to(n) áθropo] acc.sg. (c) Masculine nouns concord with adjectives in the feminine form, e.g. ώρ’μες [órmes] adj.f.nom.pl ‘ripe’ χουρμάδες [xurmáðes] noun.m.nom. pl ‘date’ (instead of ώρ(ι)μοι χουρμάδες [ór(i)mi xurmáðes]). (d) Use of the suffix -εύ(ω) [evo]/(-εύγω) [evɣo] for the coining of verbs on the basis of Turkish nominals (nouns and adjectives), e.g. ζαbουνεύγω [zabunévɣo] ‘lose weight’ < zebun, zabun ‘slim’.4 (e) Addition of the suffix -ίσκ(ω) [isko] in some verb forms in the paradigm of the imperfective (present and imperfect tense). This suffix brings no se- mantic change, e.g. ομορφαινίσκω [omorfenísko] vs ομορφαίνω [omorféno] ‘to beautify’.5 Regarding the last question, a number of researchers have argued for a ge- netic link between the Bithynian varieties and the varieties of Eastern Thrace, proposing their classification into a Thraco-Bithynian group (Psaltes, 1905; Deligiannis, 1999; Tzitzilis in press). However, the heterogeneity of Thracian Greek itself ‒ note that in Eastern Thrace varieties of the Southern, Northern and “Semi-Northern” type were spoken ‒ poses an obstacle to the establish- ment of this link.

3 The examples come from Danguitsis (1943) with some necessary changes. 4 See also Ralli (2012 and in this volume) on this issue. 5 See also Pantelidis (2016) for a discussion of this phenomenon.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 258 13-Feb-19 6:10:27 PM SUBTRACTION IN BITHYNIAN 259

3 Subtracted Imperatives in the Dialect of Bithynia

3.1 Setting the Stage As already noted, the dialect of Bithynia is quite heterogeneous and a number of different sub-varieties can be identified. For the analysis of subtracted forms we draw on data from the variety of Triglia (see map in Annex), as described by Papadopoulou (2010; 2012). When necessary, we also note the differences between the variety of Triglia and the other varieties of the Thraco-Bithynian dialectal group. The Triglia Bithynian verbal system follows the basic patterns attested in Modern Greek varieties. Each verb form is formed by the combination of a (morphologically simple or complex) base and one or more inflectional suf- fixes which express the following morphosyntactic categories: person (first, second, third), number (singular, plural), voice (active, mediopassive), aspect (perfective, imperfective) and mood (imperative, non-imperative).6 The first question that we have to answer before establishing the verbal classes in the system of any Modern Greek variety concerns the criteria for the classification of the verbs. An important feature of Greek verbal stems is allo- morphy, namely the appearance of systematic form variants. The importance of allomorphy as an autonomous criterion for the description of verbal classes in Modern Greek has been introduced by Ralli (inter alia, 1988; 2005; 2013).7 A concrete example from SMG substantiates this formalization. In SMG, verb stems belonging to I(nflectional) C(lass) 2 display a systematic pattern X(a)~X(i/e/a) (note that X represents part of the stem and the vowel in pa- rentheses is the stem-final vowel), whereas the absence of this pattern char- acterizes verb formations belonging to IC1 (see 1). IC2 can be further divided into two subclasses IC2a and IC2b, according to the stem-final vowel (from Ralli 2005) (see 2).8

6 In the variety of Triglia, the verbal augment in the past tenses shows an interesting distribu- tion. Papadopoulou (2012, pp. 130–132) argues that the appearance of the augment is not sys- tematic. Hence we cannot consider it as an inflectional prefix. However, it seems that it also appears in non-inflected positions, in opposition to other varieties, such as SMG, in which it is considered a morphologically-conditioned phonological element (Ralli, 2003). 7 This criterion has replaced the traditional classification of the verbal system which was based on the distinction between contracted and non-contracted verbs (“oxytone” and “bary- tone” in older terminology, cf. Babiniotis, 1972 on this issue). Ralli (2005; 2013) has argued that Modern Greek verbs cannot be classified according to this criterion since this phonological rule is not active anymore. 8 It should be mentioned that there are several irregular verbs which do not conform to this pattern. Moreover, a further subclass of the IC2 verbs, which includes a small number of

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 259 13-Feb-19 6:10:27 PM 260 Koutsoukos and Pantelidis

(1) αγαπ(α)-ώ [aɣap(a)ó]9 ‘love’ ~ αγάπη-σ-α [aɣápisa]10 (IC2a) stem-infl stem-asp-infl διαιρ-ώ [ðieró] ‘divide’ ~ διαίρε-σ-α [ðiéresa] (IC2b) stem-infl stem-asp-infl τρέχ-ω [tréxo] ‘run’ ~ έ-τρεχ(>κ)-σ-α11 [étreksa] (IC1) stem-infl aug-stem-asp-infl

(2) X(a) ~ X(i/e/a) IC2a X ~ X(i/e) IC2b

The verbal allomorphs have a clear distribution. The allomorphs which are presented on the left-hand column in (1) are the stems used in the imperfective context (such as the stems aɣap(a)-, ðier-, trex-), while the allomorphs which are presented on the right-hand column are the stems used in the perfective context (such as the stem aɣapi-, ðiere- trek-) (see Ralli, 2003). In Triglia Bithynian, verbs follow the same pattern and can be classified on the basis of stem allomorphy. In Triglia Bithynian IC2 includes verb formations which display a systematic allomorphic pattern, whereas IC1 includes verbs with no systematic allomorphy. IC2 can be further divided into two subclasses according to the stem-final vowel, but the stem-final vowels may differ with respect to SMG:

(3) τιμ(ά)-ω [tim(a)ó] ‘honor’ ~ τίμη-σ-α [tímisa] (IC2a) stem-infl stem-asp-infl φορ(ά)-ω [for(a)ó] ‘wear’ ~ φόρε-σ-α [fóresa] (IC2b) stem-infl stem-asp-infl γράφ-ω [ɣráfo] ‘write’ ~ έ-γραφ(>π)-σ-α [éɣrapsa] (IC1) stem-infl aug-stem-asp-infl

verbs, such as κοιμάμαι [cimáme] ‘sleep’, θυμάμαι [θimáme] ‘remember’, φοβάμαι [fováme] ‘be afraid of’, and λυπάμαι [lipáme] ‘feel sad’, has been proposed by Pantelidis (2011). 9 We use broad phonetic transcription for the Greek data. All forms, unless otherwise men- tioned, are in the citation form, i.e. 1sg present for verbs and nominative singular for nouns. 10 Morphological analyses differ with respect to the morphological status of the aspectual marker -s- (cf. Ralli, 2003). 11 The changes from [x] into [k] in (1) and [f] into [p] in (1) and (3) are normal morphopho- nological processes in SMG.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 260 13-Feb-19 6:10:27 PM SUBTRACTION IN BITHYNIAN 261

(4) X(a) ~ X(i) IC2a X(e) ~ X(e) IC2b

3.2 Subtracted Imperative Forms 3.2.1 The Formation of the Imperative Forms The imperative is one of the most peculiar inflectional categories in Modern Greek. It inflects for 2sg/2pl, perfective/imperfective aspect,12 and active/me- diopassive voice. The formation of the perfective and imperfective imperative is based on the two different types of allomorphs which were described in the previous section, i.e. the stem used in the perfective context and the stem used in the imperfective context. In the following table, we present the formation of the imperative starting with some SMG examples:

table 8.1 Formation of imperatives in SMG

Active voice

Citation form Imperfective imperative Perfective imperative

IC1 χάνω χάνε [xáne] χάσε [xáse] [xáno] χάνετε [xánete] χάσετε/χάστε [xásete/ ‘lose’ xáste]

ετοιμάζω ετοίμαζε [etímaze] ετοίμασε [etímase] [etimázo] ετοιμάζετε [etimázete] ετοιμάστε [etimáste] ‘prepare’

καθαρίζω καθάριζε [kaθárize] καθάρισε [kaθárise] [kaθarízo] καθαρίζετε [kaθarízete] καθαρίστε [kaθaríste] ‘clean’

τρέχω τρέχε/τρέχα [tréçe/tréxa] τρέξε [trékse] [tréxo] τρέχετε/τρεχάτε [tréçete/ τρέξτε [trékste] ‘run’ trexáte]

12 It should be mentioned that the choice of the aspectual form to be used by the speakers in the imperative is influenced by many factors such as: (a) specificity, (b) politeness, and (c) morphological considerations (Mackridge, 1985, pp. 122–124).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 261 13-Feb-19 6:10:27 PM 262 Koutsoukos and Pantelidis

table 8.1 Formation of imperatives in SMG (cont)

Active voice

Citation form Imperfective imperative Perfective imperative

IC2a αγαπώ αγάπα [aɣápa] αγάπησε [aɣápise] [aɣapó] αγαπάτε [aɣapáte] αγαπήστε [aɣapíste] ‘love’

IC2b διαιρώ – διαίρεσε [ðiérese] [ðieró] διαιρείτε [ðieríte] διαιρέστε [ðieréste] ‘divide’

Mediopassive voice

IC1 χάνομαι [xánome] – χάσου [xásu] ‘get lost’ χαθείτε [xaθíte]

ετοιμάζομαι [etimázome] – ετοιμάσου [etimásu] ‘get ready’ ετοιμαστείτε [etimastíte]

καθαρίζομαι [kaθarízome] – καθαρίσου [kaθarísu] ‘clean myself’ καθαριστείτε [kaθaristíte]

IC2a αγαπιέμαι [aɣapçéme] – αγαπήσου [aɣapísu] ‘be loved’ αγαπηθείτε [aɣapiθíte]

IC2b διαιρούμαι [ðierúme] – διαιρέσου [ðierésu] ‘be divided’ διαιρεθείτε [ðiereθíte]

On the basis of these data, we can make some observations regarding the for- mation of imperatives in SMG: (a) In the mediopassive voice, the imperfective imperative form is defective. For example, the verb λούζομαι [lúzome] ‘to wash my hair’ could have the forms *λούζου [lúzu] 2sg and *λούζεστε [lúzeste] 2pl, which are well- formed, but not used. (b) Verbs of IC2b are defective in the 2sg of the imperfective imperative form of the active voice, for example *θεώρει [θeóri] for θεωρώ [θeoró] ‘consider’, *αφαίρει [aféri] for αφαιρώ [aferó] ‘remove’. The 2pl is used more often.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 262 13-Feb-19 6:10:27 PM SUBTRACTION IN BITHYNIAN 263

(c) Some verbs of IC1 have a parallel form in the 2sg of the imperfective im- perative of the active voice in -α [a]. For example, the verb τρέχω [tréxo] ‘run’ has two parallel forms, τρέχε [tréçe] and τρέχα [tréxa] ‘run!’. However, this parallel form may also have a perfective reading and it is not consid- ered polite. (d) The imperative is connected with retraction of stress; in the active voice, stress is placed on the penultimate syllable in disyllabic bases and on the antepenultimate in longer bases. (e) In the 2pl of the perfective imperative, the vowel -ε- [e] of the penulti- mate syllable is deleted. For example, διαιρέστε [ðieréste] for *διαιρέσετε [ðierésete] ‘divide!’. The deletion may be accompanied by some phono- logical changes in the stem-final consonant, such as in κοιτάξετε [citák- sete] ‘look!’ > κοιτάξτε [citákste] > *κοιτάκτε [citákte] > κοιτάχτε [citáxte] (2pl of the perfective imperative of the verb κοιτάζω [citázo] ‘look’) (Newton, 1972, p. 119). (f) In the mediopassive voice, stress is always placed on the penultimate syl- lable regardless of the syllabic structure of the verbal base. The formation of the imperative in Triglia Bithynian is illustrated in the fol- lowing table:13

table 8.2 Formation of imperatives in Triglia Bithynian

Active voice

Citation form Imperfective imperative Perfective imperative

γράφω [ɣráfo] γράφε [ɣráfe] γράψε [ɣrápse] IC1 ‘write’ γράφετε [ɣráfete] γράψ(ε)τε [ɣráps(e)te]

αρεύω [arévo] – άρεψε [árepse] ‘disperse’ – –

IC2a τιμώ [timó] τίμα [tíma] τίμησε [tímise] ‘honor’ τιμάτε [timáte] τιμήσ(ε)τε [timís(e)te]

IC2b φορώ [foró] φόρειε [fórie] φόρεσε [fórese] ‘wear, put on’ φορείτε [foríte] φορέσ(ε)τε [forésete]

13 In the table, we cite only the forms presented in Papadopoulou (2012, pp. 139–140, 149, 153, 156). When the exact form is not mentioned, but the author argues that it exists, we indicate it within parentheses.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 263 13-Feb-19 6:10:27 PM 264 Koutsoukos and Pantelidis

table 8.2 Formation of imperatives in Triglia Bithynian (cont.)

Mediopassive voice

Citation form Imperfective imperative Perfective imperative

γράφομαι γράφου [ɣráfu] γράψου [ɣrápsu] [ɣráfome] γράφεστε [ɣráfeste] - IC1 ‘be written’

χτενίζομαι χτενίζου [xtenízu] - [xtenízome] χτενίζεστε [xtenízeste] - ‘comb’

IC2a αγαπιέμαι - αγαπήσου [aɣapísu] [aɣapjéme] - αγαπηθείτε [aɣapiθíte] ‘be loved’

IC2b (διαιρούμαι) - διαιρέσου [ðierésu] [ðierúme] - διαιρεθείτε [ðiereθíte] ‘be divided’

The data in table 8.2 show that in the 2pl of the perfective imperative of the active voice, the vowel -ε- [e] of the penultimate syllable can be dropped, a fact which results in parallel forms, such as γράψετε [ɣrápsete] and γράψτε [ɣrápste] ‘write!’ Moreover, similarly to SMG, the imperative is also connected with specific stress patterns. In the 2sg of the active voice, stress is placed on the penultimate syllable in disyllabic forms (see the form γράψε [ɣrápse] ‘write!’) and on the antepenultimate in longer forms (see the example άρεψε [árepse] ‘disperse!’), while in the 2sg of the mediopassive voice (both perfective and imperfective), stress is always placed on the penultimate syllable regardless of the syllabic structure of the verbal base. Hatzidakis (1905, pp. 44–46) mentions that in Modern Greek varieties verbs of IC2b have two interchangeable forms in the 2sg of the imperfective impera- tive of the active voice: one with the suffix -ε [e],14 for example φόρειε [fórie] and φόρει [fóri] ‘put on!’ of the verb φορώ [foró] ‘put on’). In the variety of Triglia only forms with this suffix appear (Papadopoulou, 2012, p. 153): πορπάτειε

14 The epenthesis of this vowel may have occurred due to analogical pressure from IC1 (Hatzidakis, 1905, p. 46).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 264 13-Feb-19 6:10:27 PM SUBTRACTION IN BITHYNIAN 265

[porpátie] for πορπατώ [porpató] ‘walk’, κράτειε [krátie] for κρατώ [krató] ‘hold’, φόρειε [fórie] for φορώ [foró] ‘wear, put on’. In the 2sg of the perfective of the mediopassive voice, verbs can also be formed in -θησε [θise], as in λούθησε [lúθise] ‘wash your hair!’ (λούσου [lúsu] in SMG) for λούζομαι [lúzome] ‘wash my hair’, πλύθησε [plíθise] ‘wash yourself!’ for πλένομαι [plénome] ‘wash myself’ (πλύσου [plísu] in SMG).15

3.2.2 Subtracted Imperative Forms The formation of the imperative forms in Modern Greek is characterized by two phenomena, i.e. retraction of stress and subtraction of the formation, which are considered interrelated by many scholars (see, among others, Hatzidakis, 1907, p. 119). Although retraction of the stress has gained some at- tention in the literature (cf. Touratzidis and Ralli, 1992), subtraction is a rather understudied phenomenon in Modern Greek. In this section, we present the data from Bithynian Greek, which is the main topic of this chapter, but we also make a comparison with two other varieties of Modern Greek, namely SMG and Peloponnesian Greek.16 In SMG, subtraction appears in either the perfective or imperfective impera- tive of verbs of IC1, but it is not systematic. The following table presents some examples:

table 8.3 Subtracted imperatives-SMG

Active voice IC1

Citation form 2sg-2pl-Imperfective 2sg-2pl-Perfective imperative imperative

ακούω [akúo] άκουγε/άκου [ákuʝe/áku] άκουσε/άκου [ákuse/áku] ‘listen’ (?) ακούτε [akúte] ακούστε [akúste]

σωπαίνω σώπαινε/σώπα [sópene/sópa] σώπασε/σώπα [sópase/sópa] [sopéno] σωπαίνετε [sopénete] σωπάστε/σωπάτε ‘be quiet; become [sopáste/sopáte] quiet’

15 See also Hatzidakis (1905, p. 81) and Danguitsis (1943, pp. 103–104) for a discussion of these forms. 16 We would like to make clear that the decision to include SMG in our analysis stems from the fact that it displays some interesting data and not from our position that all Modern Greek dialects should be compared to this variety.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 265 13-Feb-19 6:10:27 PM 266 Koutsoukos and Pantelidis

table 8.3 Subtracted imperatives-SMG (cont.)

Active voice IC1

Citation form 2sg-2pl-Imperfective 2sg-2pl-Perfective imperative imperative

αρπάζω [arpázo] *άρπαζε/άρπα άρπαξε/άρπα [árpakse/árpa] ‘grab’ [árpaze]/[árpa] αρπάξτε>αρπάχτε/αρπάτε αρπάζετε [arpázete] [arpákste>arpáxte/arpáte]

κατεβαίνω κατέβαινε [katévene] κατέβα [katéva]a [katevéno] κατεβαίνετε [katevénete] κατεβείτε [katevíte] ‘go down, get off’

ανεβαίνω ανέβαινε [anévene] ανέβα [anéva] [anevéno] ανεβαίνετε [anevénete] ανεβείτε [anevíte] ‘go up, get on’

a This form is also attested in , cf. “ΦΙ. κατάβα κατάβα κατάβα ‹ κατάβα. › ΒΔ. καταβήσομαι.” ‘PH. get down, get down, get down BD. I will.’ (Ar. Wasps 979).

Ιn table 8.3, it can be observed that subtraction appears mainly in the 2sg of the imperative form of the active voice of IC1.17 The result of subtraction may be either a disyllabic or a trisyllabic form. The subtracted form may appear in free variation with the full form, while the distribution of subtracted and non- subtracted forms may be stylistically determined. At this point, we should also make an observation regarding the aspect of the imperative forms in general, and subtracted imperative forms in particu- lar. In SMG, the aspectual distinction in the imperative is neutralized and the two paradigms merge. For example, the verb περνώ/περνάω [pernó/pernáo] (IC2a) ‘pass, pass by’ forms the 2sg of the imperfective imperative of the ac- tive voice as πέρνα [pérna]. This form may have both an imperfective and a perfective reading; it can be used with a perfective sense instead of the form πέρασε (while in the 2pl the form περάστε is still used), and it can also have an imperfective reading. In the latter case, the morphologically more transparent

17 Very few exceptions display subtraction in the perfective imperative of the mediopassive voice. The most common example is the verb σηκώνομαι ‘get up’ which displays the forms σήκω 2sg [síko] (for *σηκώσου [sikósu]) and 2pl σηκωθείτε [sikoθíte].

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 266 13-Feb-19 6:10:27 PM SUBTRACTION IN BITHYNIAN 267

variant πέρναγε [pérnaʝe]18 may also be used instead of the form πέρνα,19 e.g. εσύ πέρναγέ τες τις κλωστές στις βελόνες [esí pérnaʝe tes tis klostés stis vélones] ‘keep putting the threads through the needles’. The same “mixture” of the two aspectual readings can also be observed in the subtracted verb imperfective forms άκου [áku] and σώπα [sópa] which also have a perfective reading. This is also evident by the fact that speakers use periphrastic forms (na+verb) and lexical means to express general exhortations, e.g. να τον ακούς πάντα/κάθε φορά [na ton akús páda/káθe forá] ‘you should listen to him always/every time’ (Bakker, 1965, p. 89). We would like to propose that the neutralization of the aspectual distinc- tion in SMG might be connected with the fact that some imperfective impera- tive forms may have an ingressive use, i.e. they may denote the beginning of an action or the performance of that action. Consider the following examples:

(5) a. Έλα φάε τώρα που είσαι νηστική από το πρωί. Να, το αυγό σου [éla fáe tóra pu íse nisticí apó to proí. na to avɣó su που είναι χτυπημένο. Ρίξε γάλα και τρώγε pu íne xtipiméno. ríkse ɣála ce tróʝe] ‘Come, eat first now. You haven’t had any breakfast yet. Here is your beaten egg; pour your milk now and eat (start eating)’. (from Bakker, 1965, p. 94)

b. -Έχω κάτι να σου πω [éxo káti na su po] ‘I have something to tell you’.

-Για λέγε [ʝa léʝe]. ‘Come on. Start speaking!’

c. Για ανέβαινε! [ʝa anévene] ‘Come on. Start ascending!’

18 The stem is perna- and the inflectional suffix is -e. The /ɣ/ is epenthesized for hiatus reso- lution, as in the form άκουγε 2sg [ákuʝe] in table 8.3. 19 It is also interesting that this form appears only when there is a complement, probably due to the homonymy with the 3sg of the imperfective indicative.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 267 13-Feb-19 6:10:28 PM 268 Koutsoukos and Pantelidis

In 5(i), the imperfective imperative form τρώγε [tróʝe] is interchangeable with the phrase άρχισε να τρως [árçise na tros] ‘start (eating)’ which shows that the imperative form has an ingressive interpretation. This ingressive interpre- tation can be justified from the fact that the imperfective imperative implies progressive or iterative action and is often used in ordering or forbidding ac- tion which is already in progress, or which seems to the speaker to be immi- nent (cf. Mackridge, 1985, p. 122). In other words, the speaker considers the action to be connected with the present situation and the action must neces- sarily result from it. In Triglia Bithynian, subtraction is quite systematic, as the following exam- ples indicate:

table 8.4 Subtracted imperatives-Triglia Bithynian

Citation form 2sg-Perfective imperative-active voice

γραπώνω [ɣrapóno] ‘grab’ γράπω [ɣrápo] (for γράπωσε [ɣrápose]) τσακώνω [tsakóno] ‘catch’ τσάκω [tsáko] (for τσάκωσε [tsákose]) δρολυκώνω [ðrolikóno] ‘devour’ δρολύκω [ðrolíko] (for δρολύκωσε [ðrolíkose]) σταυρώνω [stavróno] σταύρω [stávro] ‘bless with the sign of the cross’ (for σταύρωσε [stávrose]) γιομώνω [ʝomóno] ‘fill’ γιόμω [ʝómo] (for γιόμωσε [ʝómose]) μαλώνω [malóno] ‘scold’ μάλω [málo] (for μάλωσε [málose]) φωνάζω [fonázo] ‘shout’ φώνα [fóna] (for φώναξε [fónakse]) τινάζω [tinázo] ‘toss’ τίνα [tína] (for τίναξε [tínakse])

In Triglia Bithynian, subtraction occurs in the 2sg of the perfective impera- tive of the active voice of verbs with the formatives20 -ών(ω)21 [ono] and -άζ(ω)

20 Following Bauer (1983, pp. 16–17), a formative is defined as: “a distributional segment of a word-form independent of whether or not it is also a morph. (By a ‘distributional segment’ is understood an element which recurs in the morphological analysis of word- forms.). That is, all morphs are formatives, but not all formatives are morphs, so that the use of the term ‘formative’ can avoid the issue of whether a particular element of a word realizes a morpheme or not”. 21 For the suffixes, we indicate the inflectional part in parentheses.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 268 13-Feb-19 6:10:28 PM SUBTRACTION IN BITHYNIAN 269

[azo]. In this respect, subtraction in Triglia Bithynian is more systematic than in SMG. However, the fact that subtraction occurs in verbs which have a spe- cific morphological make-up invites us to focus more on this issue. The examples in table 8.4 do not all show the same morphological parsing. Among these verbs, we find two different groups of formations: (a) synchron- ically analyzable formations in which both the base and the formative are formally and semantically transparent, such as the verbs σταυρώνω [stavróno] ‘bless with the sign of the cross’ (< σταυρός [stavrós] ‘cross’ + the suffix -ών(ω) [ono]) or φωνάζω [fonázo] ‘shout’ (φωνή [foní] ‘voice’ + the suffix -άζ(ω)), and (b) verbs which cannot be analyzed.22 In the first group, the formatives -ών(ω) [ono] and -άζ(ω) [azo] are consid- ered formally and semantically verb-forming derivational suffixes ‒ falling under the rubric “verbalizers” (see Ralli, 1988; 2005) ‒ attaching to nominal, adjectival or adverbial bases to derive verbs. In the second group of formations, if one recognizes an autonomous status to the formatives -ών(ω) [ono] and -άζ(ω) [azo] for etymological reasons or due to phonological similarity, then we should consider them as – in the terms of Lass (1990, p. 82) – “functionally speaking, junk” at the synchronic level. Based on the distinction between the two kinds of formatives, i.e. formally and semantically transparent vs functionally empty -ών(ω) [ono], one could raise the question of whether subtraction started as a process of “amend- ment” of the verbal formations which display the functionally empty for- mative through the deletion of the “junk part”. A diachronic analysis of the process could shed light on this issue, i.e. whether the process started from the formations of the second group and then extended to the first group through analogical processes. However, due to the lack of sufficient data, we cannot safely argue for this hypothesis and we leave this issue open for future research. In the Peloponnesian variety, subtraction is even more systematic than in Triglia Bithynian. The data below are illustrative:

22 Verbs belonging to the second group are quite common in Modern Greek due to the diachronic development of the verbal system and the remodeling of some verbs (cf. Anastasiadi-Symeonidi, 1992, 2004; Koutsoukos and Ralli, 2013).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 269 13-Feb-19 6:10:28 PM 270 Koutsoukos and Pantelidis

Table 8.5 Subtracted imperatives-Peloponnesian variety

IC1-Active voice

Citation form 2sg/pl-Imperfective/perfective imperativea

χαμηλώνω [xamilóno] ‘lower’ χαμήλω [xamílo] (for χαμήλωσε [xamílose]) χαμηλώτε [xamilóte] (for χαμηλώστε [xamilóste])

μουτζώνω [mutzóno] μούτζω [mútzo] ‘to direct a moutza (for μούτζωσε [mútzose]) (insulting gesture)’ μουτζώτε [mutzóte] (for μουτζώστε [mutzóste])

ζυμώνω [zimóno] ‘knead’ ζύμω [zímo] (for ζύμωσε [zímose]) ζυμώτε [zimóte] (for ζυμώστε [zimóste])

φορτώνω [fortóno] φόρτω [fórto] (for φόρτωσε [fórtose]) ‘load, charge’ φορτώτε [fortóte] (for φορτώστε [fortóste])

σκοτώνω [skotóno] σκότω [skóto] (for σκότωσε [skótose]) ‘kill, murder’ σκοτώτε [skotóte] (for σκοτώστε [skotóste])

ξαπλώνω [ksaplóno] ξάπλω [ksáplo] ‘lie down, lay sb/sth down’ (for ξάπλωσε [ksáplose]) ξαπλώτε [ksaplóte] (for ξαπλώστε [ksaplóste])

απλώνω [aplóno] άπλω [áplo] (for άπλωσε [áplose]) ‘spread, stretch, hang’ απλώτε [aplóte] (for απλώστε [aplóste])

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 270 13-Feb-19 6:10:28 PM SUBTRACTION IN BITHYNIAN 271

Table 8.5 Subtracted imperatives-Peloponnesian variety (cont.)

IC1-Mediopassive voice

Citation form 2sg/pl-Imperfective/perfective imperativea

σηκώνομαι [sikónome] ‘get up’ σήκω [síko] (for *σηκώσου [sikósu]) σηκώτε [sikóte] (for σηκωθείτε [sikoθíte])

a In the manuscript collection n. 705 (West Corinth, p. 197) of the Research Center for Modern Greek Dialects-Academy of Athens, these formations are indicated as perfective.

In this variety, subtraction occurs in both the 2sg and 2pl of the active voice and mainly in verbs with the formative -ών(ω) [ono] (IC1). Similarly to SMG, subtraction is not very common in the mediopassive voice. The only attested example is the verb σηκώνομαι ‘get up’ which displays the subtracted forms σήκω in 2sg [síko] (for *σηκώσου [sikósu]) and σηκώτε in 2pl [sikóte] (for σηκωθείτε [sikoθíte]). Subtracted forms in this variety are neutralized for the distinction perfective/imperfective. Similarly to Triglia Bithynian, the examples in table 8.5 do not all show the same morphological make-up. Among these verbs, we find two different groups of formations: (a) synchronically analyzable formations in which both the base and the formative are formally and semantically transparent, such as the verb χαμηλώνω [xamilóno] ‘lower’ (< χαμηλός [xamilós] ‘low’ + the suf- fix -ών(ω) [ono]), and (b) verbs which cannot be analyzed, such as the verb σκοτώνω [skotóno] ‘kill, murder’ which is not synchronically analyzable.

4 Data Analysis

4.1 Delimiting the Notion of Subtraction The data described in the previous section show that there is a shortening in the 2sg of the active imperfective imperative form. Since shortening is a gener- ic term which describes various kinds of processes (cf. Heller and Macris, 1968; Kreidler, 1979), we need to delimit the phenomenon described in our paper. The kind of shortening that appears in our data can be analyzed as a case of subtraction. According to Dressler (2000, p. 582), prototypical subtraction can be defined as “a reductive grammatical morphological process which is

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 271 13-Feb-19 6:10:28 PM 272 Koutsoukos and Pantelidis

the only overt morphotactic signal of a morphological meaning and consists in reducing the base of the morphological process by one phoneme at its right edge”. Nida (1949, p. 75) argues that subtraction usually results from a historical process of sound change which becomes “morphologically meaningful”. Several examples in the literature that have been analyzed as cases of sub- traction raise a number of theoretical or empirical questions which mainly refer to the direction of derivation (directionality of the process)23 and the pho- nological material that is dropped (cf. Dressler, 2000). An example that seems to be one of the most convincing comes from Tohono O’odham (Arizona): per- fective verbs are formed from imperfective ones by deleting the last C or, in certain cases, VC, e.g., /hɨhɨm/ → /hɨhɨ/ ‘to laugh’, /hiːnk/ → /hiːn/ ‘to bark’, / ɡɨɡoʂid/ → /ɡɨɡoʂ/ ‘to feed’ (from Mugdan, 2015). Subtraction shows some characteristic formal properties. For instance, it prototypically deletes just one phoneme. Non-prototypical cases of subtrac- tion may involve deletion of the base-final CV sequence, rhyme deletion, or shortening of long vowels (or loss of a vowel mora) (Dressler, 2000, pp. 583– 584). However, subtraction never seems to target prosodic constituents larger than the syllable (Alber and Arndt-Lappe, 2012, p. 312). Subtraction can be classified on the basis of some criteria, such as the scope of the process, the type of base and the presence/absence of accompanying modification (Manova, 2011, p. 156). Prototypical subtraction is the only signal of a change in the meaning, but rarely, it is accompanied by another (co-)signal (Dressler, 2000, p. 584). Typical examples of subtraction may involve inflection- al or derivational morphology and nominal or verbal formations. Interestingly enough, cross-linguistically subtraction occurs in vocatives and imperatives (cf. Winter, 1969). Although subtraction is very rare, one typological generaliza- tion about its distribution seems to be feasible: it does not occur in agglutinat- ing or non-fusional languages (Dressler, 2000, p. 585). In derivational morphology, subtraction must be distinguished from phe- nomena that are sometimes confused with it, such as phonological shorten- ing, truncation, back-formation, and clipping (cf. Manova, 2011; Mugdan, 2015). What distinguishes subtraction from phonological shortening is the fact that the latter is “phonologically-conditioned”, i.e. phonological truncation rules are governed by phonological restrictions (Manova, 2011, p. 131). Moreover, subtraction-like phonological rules usually shorten various morphotactic seg- ments, a feature which is not typical for morphological changes (idem).

23 See for example the formation of male adjectives in French on the basis of the female form which has been a topic of considerable dispute (cf. Bloomfield, 1933; Nida, 1949; Dressler, 2000).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 272 13-Feb-19 6:10:28 PM SUBTRACTION IN BITHYNIAN 273

Subtraction is also different from the process of truncation. The morpho- logical term truncation was introduced by Aronoff (1976, p. 88) who defined it as an adjustment rule deleting a morpheme internal to an affix, e.g. nominate > *nomin-at(e)-ee > nominee. The nature of the targeted material in truncation has been a matter of debate. Some scholars have claimed that truncation al- ways involves morphological constituents (cf. Aronoff, 1976). Contrary to this view, a number of researchers have argued that truncation may be explained only by virtue of phonological and morphological constraints that interact (cf. Plag, 1998) or phonological constraints that are morphologically condi- tioned (cf. Bat-El, 2002). Subtraction should also be distinguished from back-formation. Back- formation (or back-derivation) refers to the process according to which words are extracted from longer words, which have the formal appearance of com- plex bases (Marchand, 1969, p. 391) For example, the English noun sculptor was borrowed in the 17th century without a base verb. Back-formation consisted in the creation of the verb sculpt. When this word entered the lexicon, sculptor was considered derived form on the basis of sculpt ‒ notwithstanding the fact that it existed before its “base” (Mugdan, 2015, p. 277). Hence the verb sculpt is an extraction from the noun sculptor. Back-formation should not be con- sidered a synchronic derivational process, but rather a historical connection, since the establishing of its direction always requires diachronic information (Marchand, 1969, p. 391; Manova, 2011, pp. 133–134).24 A frequently occurring type of back-formation may, however, eventually develop into a subtractive rule (Dressler, 2000, p. 583). Another interesting process with which subtraction may be confused is clipping. Clipping refers to a variety of word shortening processes that re- sult in new word forms; this cutting of a lexeme (one or more words) does not result in a change of meaning (Steinhauer, 2015, p. 352). For this reason clippings are usually defined as “extragrammatical morphology”25 and are ex- cluded from morphological description (Mattielo, 2013).26 The reduction may affect much more than one single phoneme. For example, phone is shortened for telephone and asap for as soon as possible (idem). Prototypically, clipping

24 Some linguists such as Bauer (1983, 64–65) understand back-formation as a synchronic rule. However, this view is criticized (cf. Mugdan, 2015, p. 277). 25 According to Mattiello (2013, p. 1), this term applies to a set of heterogeneous formations which are not the product of “classic” word formation processes, in that the processes through which they are obtained are not clearly identifiable and their input does not allow a prediction of a regular output. 26 Marchand (1969, p. 441) argues that clippings are not relevant to the linguistic system itself (“la langue”), but to the speech (“la parole”).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 273 13-Feb-19 6:10:28 PM 274 Koutsoukos and Pantelidis

does not involve cases where the input is an inflected form (cf. the relevant descriptions in Jespersen, 1942; Mattielo, 2013). However, Barić et al. (1997 cited in Manova, 2011, p. 139) point out a few examples of clipped verbs. The dividing line between subtraction and clipping is that the latter process does not appear systematically in a set of data and does not change the meaning (Marchand, 1969, p. 441).

4.1.1 Form and Meaning in Subtraction The process of subtraction poses a number of theoretical problems to morpho- logical description. One of them is the relationship between form and mean- ing in subtractive formations. Whereas morphological processes usually add morphological elements to the bases (such as affixes), the process of subtrac- tion consists in the shortening of the shape of a word form. This is the reason why in the framework of Natural Morphology subtractive morphological rules have been considered anti-diagrammatic or anti-iconic, and they represent a very “marked” (non-preferred, unnatural) and counter-intuitive phenome- non. Subtractive structures violate the most important semiotic principle, i.e. the iconicity principle (cf., among others, Dressler, 1987, p. 104; Mel’čuk, 1991, p. 283). The mapping of form and meaning in subtraction is also connected with the architecture of morphological theory. Subtraction has been cited as prime evidence against Item-and-Arrangement models (in the terms of Hockett, 1954) or “building-block models of morphology” (in the terms of Mugdan, 2015), which are based on the assumption that morphological processes (or operations) build a concatenation of morphemes which can be seen as mor- phological signs, i.e. form-meaning pairs. In this view, subtraction shows an asymmetry with the “additive morphemes” (i.e. classic affixes). American structuralists “solved” this problem by interpreting subtraction as a special kind of linguistic element, called “minus feature” (Bloomfield, 1933, p. 217), “minus morpheme” (Harris, 1942, p. 171), “subtraction morph” (Hockett, 1947, p. 340), or “subtractive morpheme” (Nida, 1949, p. 75), depending on the theoretical analysis. However, these counter-intuitive analyses were soon aban- doned in favor of the tenet that a morpheme, i.e. the expression of a linguis- tic sign, can be a process that manipulates such a string in Item-and-Process ­models (cf. Mugdan, 2015, p. 278).27 In the view of Anderson (1992, p. 66), if the “content” of a morpheme is deletion rather than the addition of phonological material, there is no way to

27 See also Mel’čuk (1991, p. 281) who argues that signs can be only additive and rejects the notion of the replacive/subtractive morpheme.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 274 13-Feb-19 6:10:28 PM SUBTRACTION IN BITHYNIAN 275

parse the surface form of a word containing this category so that some possibly discontinuous subpart of its structure constitutes the morph in question. It should, then, follow that the existence of such examples provides strong evi- dence in favor of a morphological theory that recognizes processes other than affixation. However, this type of analysis has also been criticized for being “too unrestricted” (Alber and Arndt-Lappe, 2012, p. 312).

4.2 Analysis of the Subtracted Forms in Bithynian The description of the data and the theoretical analysis of subtraction in the previous section lead us to argue that the phenomenon is a not a case of a pure phonological process, since it appears in a very specific context and not in the entire paradigm. If the change observed in the dialect of Bithynia is a case of subtraction, following Dressler’s definition of subtraction, it does not belong to the prototypical cases of subtraction, since more than one phoneme is deleted in order to create the new form. However, subtraction may be related to the prosodic patterns of the imperative form and in this respect related to prosody. In the next sections, we raise two questions, namely the morphological context in which subtraction occurs and the motives behind this change.

4.2.2 The Origin of Subtracted Forms The origin of the subtracted forms in Modern Greek dialects has not yet been extensively discussed. Thus, it is very difficult to examine the role of analogy in the pattern of subtraction. An old attestation of subtracted forms is found in Koumanoudis (1882, p. 370 fn2) who mentions the following:

[…] ἐν τῇ Δημοτικῇ ἀνθολογίᾳ τοῦ Μ. Λελέκου τῇ ἐκδοθείσῃ τῷ 1868, ἐνετύχομεν δὶς τῷ μάζο ἀντὶ τοῦ μάζωξε ἢ μάσε, ἐν σελ.104–5. Τοῦτο δὲ τὸ μάζο ἕνεκα τοῦ ο ἂς μὴ τὸ ὑπολάβῃ τις ὅτι εἶναι ὁ ἀποκεκομμένος κατὰ ἓν ν ἀρχαϊκὸς ἀόριστος ὁ εἰς ον, διότι οὗτος βεβαίως ἀπωλέσθη ἀνεπιστρεπτί.

[…] In the folklore anthology by M. Lelekos published in 1868, on pages 104–105, I found twice the form mazo instead of mazokse or mase. This form in -o should not be considered – due to the presence of -o – a sub- tracted archaic form in -on because this form has disappeared irrevocably.

To the best of our knowledge, the oldest attestations of subtracted forms in the Peloponnesian variety are to be found in Koukoules’ Oinountiaka (1908). The only subtracted form whose origin has been extensively discussed is that of the verb σήκω [síko] ‘get up’. Hatzidakis (1907, p. 119) argues that the

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 275 13-Feb-19 6:10:28 PM 276 Koutsoukos and Pantelidis

verbal form σήκω [síko] – instead of *σηκώσου [sikósu] – takes an ending in -ω [ο] by analogy with the adverbial endings in -ω, as in the cases κάτω [káto] ‘down’, πάνω [páno] ‘up’. This theory is based on the fact that adverbs such as απάνω [apáno] ‘up’ and κάτω [káto] ‘down’ can be used – with certain intona- tional contours – as imperatives. However, Hatzidakis does not take into ac- count that in these dialects the adverbs which might have served as models for σήκω [síko] are attested ending in -ου [u], for example απάνου [apánou] and κάτου [kátu], and not in -ω [o], as formed in SMG. Filintas (1927, pp. 84–86) argues that a form σήκο [síko] was coined after the contraction of the form σήκου [síku] with active second singular impera- tive ending -ε [e].28 Filintas’ explanation of the origin of the form seems quite complicated. First, it is difficult to explain why the active second singular im- perative ending -ε [e] occurs in the context of mediopassive. Second, the con- traction of /u/+e > /o/ is a sandhi phenomenon, that is, it occurs only between words and not word-internally.

4.2.3 Motives Behind the Change The motives behind this deletion may vary according to the theoretical analy- sis. In what follows, we examine the phonological motives behind subtraction, in general, and then we present how the Triglia Bithynian imperatives fit in this frame. As has already been described, subtraction concerns morphologically contextualized deletion or change. However, deletion is also connected with some phonological processes, a fact which leads us to examine the relation- ship between subtraction and prosody. Following Martin (1988, p. 234), we can recognize two kinds of interaction between prosody and subtraction: the first involves deletion of a fixed prosodic unit from a formation, and is generally called subtraction, while the second involves deletion of material of variable size from a formation, and is called templatic truncation. Similarly, Bat-El (2002, p. 653) distinguishes between two types of truncation that yield shortening of a morphological constituent, namely fake truncation and true truncation. True truncation is the deletion of segmental material directly required for the deri- vation of one lexical category from another, while fake truncation is the dele- tion of segmental material required by the imposition of a template. Subtraction and templatic truncation differ in many respects (Horwood, 2001); the aim of templatic truncation is the mapping of a morphological constituent onto a (smaller) prosodic template, resulting in some loss of seg- mental structure, while subtraction is the process by which a grammatically

28 For the verbal form σήκω [síko], Filintas gives an orthographic form in -ο [o], i.e. σήκο.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 276 13-Feb-19 6:10:28 PM SUBTRACTION IN BITHYNIAN 277

characterizable unit (typically prosodic) is truncated from the right (or left) periphery of some morphological unit and the portion “saved” may vary in length and shape. In general, templatic truncation is usually associated with the left edge of the word, while in subtraction the prosodic requirements are usually met by alignment to the right edge of the word (Martin, 1988, p. 234). From the description of the two processes, we can conclude that the aim of deletion in templatic truncation is to make a structure which meets the re- quirements of a template, while the reasons for the phonological deletion in subtraction are not always clear. Phonological deletion in subtraction has often been analyzed as a case of “prosodic circumscription” and subsequent deletion of prosodic constituents (Alber and Arndt-Lappe, 2012, p. 312). According to McCarthy and Prince (1996, p. 335) under prosodic circumscription a morphological operation is applied to a base that is a prosodically delimited substring within the grammatical cat- egory. In other words, circumscription places a cut in the form in a prosodically defined position, either before or after a prosodic unit and thus splitting the domain in two parts (Kager, 1999, p. 224). Examples of this kind of analysis can be found in Martin (1988) who has analyzed subtraction as rhyme “dis- sociation”, i.e., the delinking of the rhyme of the base-final syllable. A similar analysis can be found in Lombardi and McCarthy (1991). Kager (1999, p. 224) argues that this type of analysis still fails to answer the fundamental question of why a morphological operation should apply to the output of a prosodic operation and he proposes a model in which prosody and morphology apply “in parallel” and alignment plays a role.29 However, the an- swer to this question can be sought in view of the criterion of optimal word size. Subtractive derivation responds to the preference for bi- and tri-syllabicity that seems to hold cross-linguistically (Luschützky, 2015, p. 133). This also brings us closer to the pragmatic factors behind subtraction. According to Dressler (2000, p. 585), on a cross-linguistic basis vocatives and imperatives are often reductive. The reason is that both categories can repre- sent minimal sentences so that there may be an iconic relation of shortness between shortest syntactic and shortest morphological forms (cf. Winter, 1969, pp. 220–222; Hill, 2013). Regarding our data, as has already been presented, the imperative in Greek is connected with movement of stress. This stress movement can be regarded as either metrical information specified in the inflectional forma- tives (Touratzidis and Ralli, 1992, p. 441) or a condition imposed by a rule that

29 The presentation of the details of the approaches go beyond the scope of this paper, but an overview can be found in Alber and Arndt-Lappe (2012).

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 277 13-Feb-19 6:10:28 PM 278 Koutsoukos and Pantelidis

refers to these specific morphological categories (Philippaki-Warburton, 1976). In Triglia Bithynian, subtraction occurs in verbs with -άζ(ω) [azo] and -ών(ω) [ono] which have more than three syllables. In the 2SG of the perfective active voice of the imperative of these verbs, stress is placed on the antepenultimate syllable and subtraction deletes the last syllable of the word. Deletion results in a trochee, which according to Malikouti-Drachman and Drachman (1989) can be seen as the optimal prosodic pattern in Greek.30 Hence we could argue that subtraction applies in the domain after the prosodic unit of the trochee and deletes one syllable.

5 Conclusions

In this chapter, we have analyzed subtracted imperatives in Triglia Bithynian and have presented some data that have not been thoroughly discussed before. We have made a comparison between subtractive formations in Bithynian and similar formations in other Modern Greek varieties which points to a cline along which varieties can be placed according to their behavior towards sub- traction in imperative forms: SMG shows very limited and unsystematic sub- traction in imperatives, while the Peloponnesian variety is more systematic in this respect. The Triglia Bithynian variety is situated between the two ex- tremes, since subtraction is very systematic in a certain group of verbs, but only in the 2sg of the perfective imperative of the active voice (not in the en- tire paradigm). Within a general frame, subtractive imperatives in Bithynian offer significant evidence in favor of the view that subtraction should be considered a morpho- logical process in its own right. Despite the fact that it is sometimes difficult to delimit the process, Triglia Bithynian and its comparison to the other Greek varieties corroborates the claim that subtraction is cross-linguistically related to the category of imperative. It also shows that prosodic factors may motivate this type of change, since subtraction in Triglia Bithynian results in a trochee, which is considered the optimal prosodic form in Greek. Subtraction is also connected with the problem of the relationship between form and meaning in morphological formations, since its main feature is the

30 The prosodic pattern of the trochee seems to be quite strong in Modern Greek impera- tives, as evidenced by the chain reactions that occur to preserve it. For example, in some verbs there is an epenthesis of a vowel /e/ at the beginning of the base to form a trochee, as in the examples έμπα [éba] 2sg.pfv.imp (for μπες) of the verb μπαίνω [béno] ‘enter’, and έβγα [évɣa] 2sg.pfv.imp (for βγες) of the verb βγαίνω [vʝéno] ‘exit’.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 278 13-Feb-19 6:10:28 PM SUBTRACTION IN BITHYNIAN 279

shortening of the shape of a word form. However, the fact that subtraction is related to a specific prosodic pattern leads us to conclude that this “defect” in the form is remedied by prosody. The prosodic pattern gives cues for the “empty content” of the process and signals subtraction. With the analysis of this case, we hope to have opened up the discussion on similar phenomena in other Modern Greek varieties and to have given new impetus to further re- search on the process of subtraction.

References

Alber Birgit and Arndt-Lappe Sabine. 2012. Templatic and subtractive truncation. In Jochen Trommer (ed.), The morphology and phonology of exponence, 289–325. Oxford: Oxford University Press. Anastasiadi-Symeonidi, Anna. 1992. Η νεοελληνική παραγωγή κατά το μοντέλο της D. Corbin [Greek derivation according to D. Corbin’s model]. Studies in Greek Linguistics 12: 505–526. Anastasiadi-Symeonidi, Anna. 2004. Κλίση και παραγωγή: μύθος και αλήθεια [Inflection and derivation: myth and truth]. Studies in Greek Linguistics 24: 43–54. Anderson, Stephen R. 1992. A-Morphous morphology. Cambridge: Cambridge University Press. Aronoff, Mark. 1976. Word formation in generative grammar (Linguistic Inquiry mono- graphs 1). Cambridge, MA: MIT Press. Babiniotis, Georgios. 1972. Το ρήμα της ελληνικής: δομικαί εξελίξεις και συστηματοποίησις του ρήματος της ελληνικής (αρχαίας και νέας) [The verb in Greek: structural changes and systematization of the verb in Greek (Ancient Greek and Modern Greek)]. Αthens: Sophia Saripolou Library. Bakker, W. F. 1965. The aspect of the imperative in Modern Greek. Neophilologus 49: 89–103 & 203–210. Bat-El, Outi. 2002. True truncation in colloquial Hebrew imperatives. Language 78(4): 651–683. Bauer, Laurie. 1983. English word formation. Cambridge: Cambridge University Press. Bloomfield, Leonard. 1933. Language. New York: Holt. Brixhe, Claude. 2008. Phrygian. In Roger D. Woodard (ed.), The Ancient languages of Asia Minor, pp. 68–80. Cambridge: Cambridge University Press. Brixhe, Claude. 2010. Linguistic diversity in Asia Minor during the empire: Koine and non-Greek Languages. In Egbert J. Bakker (ed.), A companion to the Ancient , 228–252. Malden/Oxford: Wiley-Blackwell. Danguitsis Constantin. 1943. Étude descriptive du dialecte de Démirdési: (Brousse, Asie Mineure); dialectologie néo-hellénique. Paris: Maissonneuve.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 279 13-Feb-19 6:10:28 PM 280 Koutsoukos and Pantelidis

Deligiannis, Κyriakos. 1999. Κουβουκλιώτικα. Ένα μικρασιατικό γλωσσικό ιδίωμα [Kouvoukliotika: an Asia Minor linguistic variety]. Adelaide: Claxton Press. Dressler, Wolfgang U. 1987. Word formation as part of natural morphology. In Wolfgang U. Dressler (ed.), Leitmotifs in Natural Morphology, 99–126. Amsterdam/ : John Benjamins Publishing Company. Dressler, Wolfgang. 2000. Subtraction. In Geert Booij, Christian Lehmann and Joachim Mugdan (in collaboration with Wolfgang Kesselheim and Stavros Skopeteas), Morphologie. Morphology. Ein internationales Handbuch zur Flexion und Wortbildung. An international handbook on inflection and word formation. 1. Halbband/Volume 1, 581–587. Berlin: Walter de Gruyter. Filintas, Menos. 1927. Γλωσσογνωσία και γλωσσογραφία ελληνική [Study of the Greek lan- guage and Greek Glossography] Vol. 2. Αthens: Athina. Hatzidakis, Georgios. 1905. Μεσαιωνικά και Νέα Ελληνικά, τόμ. Α΄ [Medieval and Modern Greek, Volume 1]. Αthens: Sakellarios. Hatzidakis, Georgios. 1907. Μεσαιωνικά και Νέα Ελληνικά, τόμ. Β΄ [Medieval and Modern Greek, Volume 2]. Αthens: Sakellarios. Harris Zellig S. 1942. Morpheme alternants in linguistic analysis. Language 18 (3): 169–180. Hawkins, Shane. 2010. Greek and the languages of Asia Minor to the Classical Period. In Egbert J. Bakker (ed.), A companion to the Ancient Greek language 213–227. Malden/ Oxford: Wiley-Blackwell. Heller Louis G. and Macris James. 1968. A typology of shortening devices. American Speech 43 (3): 201–208. Hill, Virginia. 2013. Vocatives: How Syntax meets with Pragmatics. Leiden: Brill. Hockett Charles F. 1947. Problems of morphemic analysis. Language 23 (4): 321–343. Hockett, Charles F. 1954. Two models of grammatical description. Word 10: 210–234. Horwood, Graham. 2001. Anti-Faithfulness and subtractive morphology. Ms., Rutgers University, New Brunswick. Jespersen, Otto. 1942. A modern English grammar. On historical principles Part VI (writ- ten with the assistance of Paul Christophersen, Niels Haislund and Knud Schibsbye). Copenhagen: Einar Munksgaard. Kager, René. 1999. Optimality Theory. Cambridge: Cambridge University Press. Konstantinidou Magdalini. 2005a. Μικρασιατικά ιδιώματα περιοχής Βιθυνίας και Ιωνίας. Αποτελέσματα μιας γλωσσικής αποστολής [Linguistic varieties in the area of Bithynia and Ionia]. Leksikografikon Deltion Akadimias Athinon 25: 119–148. Konstantinidou, Μagdalini. 2005b. Το γλωσσικό ιδίωμα της Προύσας Βιθυνίας [The lin- guistic variety of Bursa]. Studies in Greek Linguistics 25: 324–335. Koukoules Phaedon. 1908. Οινουντιακά ή μελέτη περί της ιστορίας, των ηθών και των εθίμων και του γλωσσικού ιδιώματος του δήμου Οινούντος της επαρχίας Λακεδαίμονος [Oinountiaka

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 280 13-Feb-19 6:10:29 PM SUBTRACTION IN BITHYNIAN 281

or a study of the history, the laography and the dialect of Oinountas of the province of Lakedaimon]. Chania, Crete. Koumanoudis Stephanos. 1882. Νύξεις εἰς ἔρευναν τῶν ἰδιαζόντων τῇ νεωτέρᾳ ἑλληνικῇ γλώσσῃ [Observations on the study of the peculiar Modern Greek forms]. Πλάτων [Platon] Ε΄, 368–376. Koutsoukos, Nikos and Ralli Angela. 2013. Elements with ambiguous derivational sta- tus: The marker -idz(o) in Griko. Quaderns de Filologia 18: 13–23. Kreidler, Charles W. 1979. Creating new words by shortening. Journal of English Linguistics 13(1): 24–36. Lass Roger. 1990. How to do things with junk: exaptation in language evolution. Journal of Linguistics 26(1): 79–102. Lombardi Linda and McCarthy John. 1991. Prosodic circumscription in Choctaw mor- phology. Phonology 8(1): 37–72. Luschützky Hans Christian. 2015. Word-formation in Natural Morphology. In Peter O. Müller, Ingeborg Ohnheiser, Susan Olsen and Franz Rainer (eds.), Word-formation. An international handbook of the languages of Europe Volume 1, 123–144. Berlin/Boston: De Gruyter Mouton. Mackridge, Peter. 1985. The Modern Greek language: a descriptive analysis of Standard Modern Greek. Oxford: Oxford University Press. Malikouti-Drachman Angeliki and Drachman Gaberel. 1989. Τονισμός στα Ελληνικά [Stress in Greek]. Studies in Greek Linguistics 9: 127–143. Manova, Stella. 2011. Understanding morphological rules: with special emphasis on conver- sion and subtraction in Bulgarian, Russian and Serbo-Croatian. Dordrecht: Springer. Marchand, Hans. 1969. The categories and types of present-day English word forma- tion. A synchronic-diachronic approach (2nd edition). München: C.H. Beck’sche Verlagsbuchhandlung. Martin, Jack. 1988. Subtractive morphology as dissociation. WCCFL 7: 229–240. Mattiello Elisa. 2013. Extra-grammatical morphology in English. Abbreviations, blends, reduplicatives, and related phenomena. Berlin/Boston: De Gruyter Mouton. McCarthy, John and Prince, Alan. 1996. Prosodic morphology. In John A. Goldsmith (ed.), The Handbook of phonological theory, 318–366. Cambridge, Mass./Oxford: Blackwell Publishers Ltd. Mel’čuk, Igor. 1991. Subtraction in Natural Language. In Maciej Grochowski and Daniel Weiss (eds.), Words are physicians for an ailing mind. For Andrzej Bogusławski on the occasion of his 60th birthday, 279–293. München: Sagner. Mugdan, Joachim. 2015. Units of word-formation. In Peter O. Müller, Ingeborg Ohnheiser, Susan Olsen, and Franz Rainer (eds.), Word-Formation: An international handbook of the languages of Europe Volume 1, 235–301. Berlin/New York: de Gruyter Mouton.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 281 13-Feb-19 6:10:29 PM 282 Koutsoukos and Pantelidis

Newton, Brian. 1972. The generative interpretation of dialect: a study of Modern Greek phonology. Cambridge: Cambridge University Press. Nida, Eugene A. 1949. Morphology: The descriptive analysis of words (Second and com- pletely new edition). Ann Arbor: University of Michigan Press. Pantelidis, Nikolaos. 2011. Ρηματική κλιτική μορφολογία του μεγαρικού ιδιώματος [Verbal inflectional morphology in the dialect of Megara]. Patras Working Papers in Linguistics 2: 55–77. Pantelidis, Nikolaos. 2016. Το ρηματικό επίθημα -(ι)σκ- στην Καππαδοκική [The verbal suf- fix -(i)sk- in Cappadocian]. In Angela Ralli, Nikos Koutsoukos and Stavros Bompolas (eds.), Proceedings of the 6th Modern Greek Dialects and Linguistics Theory meeting, 134–143. Patras: University of Patras. Papadopoulou, Eleni. 2010. Το γλωσσικό ιδίωμα της Τρίγλιας Βιθυνίας [The linguistic vari- ety of Triglia]. PhD Dissertation, University of Ioannina. Papadopoulou, Eleni. 2012. Γλωσσογεωγραφική κατανομή των ιδιωμάτων της Βιθυνίας. Αναλογίες με τα Θρακικά ιδιώματα [Linguo-geographical distribution of the varieties of Bithynia. Similarities with the Thracian varieties]. In Zoe Gavriilidou, Angeliki Efthymiou, Evangelia Thomadaki and Penelope Kambakis-Vougiouklis (eds.), Selected papers of the 10th International Conference of Greek Linguistics, 1056–1062. Komotini: Democritus University of Thrace. Philippaki-Warburton Irene. 1976. On the boundaries of morphology and phonology: A case study from Modern Greek. Journal of Linguistics 12(2): 259–278. Plag, Ingo. 1998. Morphological haplology in a constraint-based morpho-phonology. In Wolfgang Kehrein and Richard Wiese (eds.), Phonology and morphology of the , 199–215. Tübingen: Niemeyer. Psaltes, Stamatis. 1905. Θρακικά ή μελέτη περί του γλωσσικού ιδιώματος της πόλεως των Σαράντα Εκκλησιών [Thracian or a study of the linguistic variety spoken in the city of Saranta Ekklisies]. Αthens. Ralli, Angela. 1988. Eléments de la morphologie du grec moderne. (Ph.D. dissertation, Université de Montréal. Ralli, Angela. 2003. Morphology in Greek linguistics. The state of the art. Journal of Greek Linguistics 4: 77–129. Ralli, Angela. 2005. Μορφολογία [Morphology]. Αθήνα: Εκδόσεις Πατάκη. Ralli, Angela. 2012. Verbal loanblends in Griko and Heptanesian: a case study of con- tact morphology. L’Italia Dialettale: Rivista di dialettologia italiana 73: 111–132. Ralli, Angela. 2013. Compounding in Modern Greek (Studies in Morphology 2). Dordrecht: Springer. Touratzidis Loudovikos and Ralli, Angela. 1992. A Computational treatment of stress in Greek inflected forms. Language and Speech 35 (4): 435–453. Sakellariou, M. V. 1990–1991. Τα όρια των χωρών και επαρχιών της Μικράς Ασίας [The bor- ders of countries and prefectures of Asia Minor]. Deltio Mikrasiatikon Spoudon 8: 207–220.

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 282 13-Feb-19 6:10:29 PM SUBTRACTION IN BITHYNIAN 283

Steinhauer, Anja. 2015. Clipping. In Peter O. Müller, Ingeborg Ohnheiser, Susan Olsen, and Franz Rainer (eds.), Word-Formation. An international handbook of the languag- es of Europe Volume 1, 352–363. Berlin/Boston: De Gruyter Mouton. Tzitzilis, Christos. In press. Η θρακοβιθυνιακή διάλεκτος [The Thraco-bithynian dia- lect]. In Christos Tzitzilis (ed.), The Modern Greek Dialects. Thessaloniki: Institute of Modern Greek Studies. Vryonis, Speros. 1971. The decline of Medieval Hellenism in Asia Minor and the process of Islamization from the eleventh to the fifteenth century. Berkeley: University of California Press. Winter, Werner. 1969. Vocative and imperative. In Jaan Puhvel (ed.), Substance and structure of language, 205–223. Berkeley/Los Angeles: University of California Press.

Annex: Map of Bithynia

map 8.1 Bithynia

255-283_Ralli_09-Koutsoukos-Pantelidis.indd 283 13-Feb-19 6:10:29 PM