Insights on the Realization of Nominal Evaluative Morphology in the Modern Greek Dialect of Lesbos

Abstract

This paper investigates evaluative morphology in the modern dialect of Lesbos, focusing on the morphological strategies used for nominal evaluation, the range of meanings they express, and the characteristics that distinguish Modern Lesbian from other varieties. Special attention is given to borrowing and the integration of markers serving evaluative functions in the dialect. Dialectal data are drawn from available primary and secondary written sources. The analysis shows that Modern Greek dialects, Lesbian in particular, which have largely escaped the effects of diglossia and standardization, constitute an especially valuable resource for linguistic research, as they reveal prototypical tendencies of linguistic systems. These tendencies include the preference for specific gender values as defaults in the expression of diminution and augmentation, the overwhelming productivity of specific markers, closely linked to the local dialectal identity, the symmetrical distribution of suffixes and prefixoids in the realization of positive and negative evaluative meanings, and the creative adaptation of borrowed evaluative elements.

1. Introduction

Evaluative morphology is a favorite and relatively well studied topic in linguistic literature treated from different perspectives, all formal, semantic and pragmatic on a language specific, areal or cross-linguistic level. The term is used as an umbrella concept encompassing the full range of morphological mechanisms available within a language system for the formation of lexical items that express evaluation of the referent, either quantitatively (small/large; diminution, augmentation) or qualitatively (good/bad), including functions such as intensification, reduction, mitigation, approximation, the expression of positive or negative attitude, and other pragmatic meanings that signal deviation from a standard value (Grandi & Körtvélyessy, 2015, p. 10) which, in many languages may co-occur in the same morphological formation.

Although a considerable body of research has scrutinized individual evaluative strategies (e.g., prefixation, suffixation, reduplication), specific grammatical domains (e.g., nominal or verbal morphology), and particular evaluative meanings (readings) or functions (e.g., diminutives, augmentatives, intensifiers) in Standard Modern Greek, the dialectal diversity of Greek has not been investigated with comparable systematicity, let alone exhaustiveness. Studies are often limited to listing the most typical diminutive—and more rarely augmentative—suffixes of individual dialect areas (see, however, Koutita-Kaimaki, 1984; Melissaropoulou, 2006, 2007, 2009; Melissaropoulou & Ralli, 2008).

The aim of this paper is to provide a systematic overview of the available morphological strategies for the realization of the evaluative functions of Modern Lesbian with an emphasis on the particularities of the dialect, aiming at revealing how dialectal diversity can shed light on the prototypical tendencies of language systems, thus contributing to the study of language variation and change in general.

For the purposes of this paper, our data are drawn from all available written sources, both primary (e.g., journal articles, fairy tales, literary texts written in the dialect) (e.g., Dedekis, 2015; Psarianos, 2023) and secondary (grammatical descriptions, dictionaries, glossaries, scientific studies and articles) (among others, Kretschmer, 1905; Papadopoulos, 1926; Papanis & Papanis, 2004; Ralli, 2019; Sakkaris, 1940) as well as from the electronic atlas of Lesbos (Ralli, 2019; Alexelli, 2021).

The research questions we address for the purposes of this study are the following: a. What are the basic morphological strategies of nominal evaluation in the dialect in combination with the different functions/readings they realize? b. What are the particularities of Lesbian evaluative morphology? c. Are there any phenomena of transfer and adaptation of loan evaluative elements in the dialect and how do they function in comparison with other dialectal systems?

The article is structured as follows: Section 2 outlines the theoretical assumptions underlying the delimitation of evaluative morphology as a field of inquiry. Section 3 examines nominal diminutives on the basis of relevant dialectal data, elaborating on the theoretical generalizations that emerge from the analysis of the morphological strategies employed. In a similar vein, Section 4 emphasizes the morphological process of augmentation. Section 5 delves into the use of prefixoids in the realization of negative evaluation, as well as the adaptation of borrowed elements serving evaluative functions in the dialect, revealing further dialectal dynamics and inherent tendencies. The article concludes with a summary of the main findings and a list of references.

2. Background

Evaluative morphology and the specific evaluative functions realized through distinct morphological strategies have attracted considerable attention in contemporary linguistic research from a variety of theoretical perspectives. Some of the most commonly examined issues include the origin and development of evaluative markers, their morpho-semantic and morpho-pragmatic characteristics, the stacking and hierarchical arrangement of evaluative elements, and the place of evaluative morphology within the grammatical system—specifically, whether it should be treated as part of inflection, as a type of derivation, or as an intermediate or independent domain between these two morphological processes (for detailed discussion, see Grandi & Körtvélyessy, 2015; Manova et al., 2024 and references therein).

Building on existing scholarship on Greek (see Melissaropoulou, 2015 and references therein), we adopt the view that evaluative morphology belongs to the domain of word formation rather than to inflection or to an intermediate grammatical component (Melissaropoulou & Ralli, 2008). In Greek, evaluative meanings are primarily expressed through suffixation and prefixation, and to a lesser extent through compounding, depending on the grammatical category under consideration. More specifically, suffixation predominates in the nominal domain, whereas prefixation is more characteristic of the verbal domain (Melissaropoulou, 2015; Efthymiou, 2017, 2024). With respect to nouns, Greek conforms to the typological generalization that identifies suffixation as the preferred cross-linguistic strategy for diminutive formation (Štekauer et al., 2012, p. 325).

As in many other languages, a single suffix in Greek may simultaneously encode both quantitative and qualitative dimensions of evaluation. For instance, the diminutive form kuˈkl-aci ‘doll-dim’ may indicate not only small size but also the speaker’s positive stance toward the referent (namely, also a dear/sweet doll). Although certain evaluative suffixes may privilege quantitative over qualitative interpretations (or vice versa), the prevailing cross-linguistic tendency (Klimaszewska, 1983), also observable in Greek, is for diminutives to be more typically associated with positive evaluation and augmentatives with negative evaluation.

It is important to note that Modern Greek displays, alongside prototypical evaluative formations that express deviation from a default value along quantitative or qualitative dimensions, non-prototypical formations that combine evaluative meaning with additional semantic components, such as agentivity (e.g., faˈɣ-as ‘eater’ or that of a person with bearer of a property (e.g., aˈftças ‘a person with big ears’). Since such formations extend beyond purely evaluative functions, they fall outside the scope of the present study and warrant separate investigation. In the same spirit, semantically opaque formations—namely, those containing an evaluative morpheme without a clear and transparent semantic contribution—are treated as lexicalized and are likewise excluded from the analysis. Finally, given limitations of scope and the high productivity of evaluation, the present study focuses exclusively on nominal formations, beginning with diminutive formation and subsequently addressing augmentation, with particular emphasis on suffixation.

3. Diminution in Lesbian Dialect

3.1. Data

Modern Lesbian belongs to the group of Northern Greek dialects, thus it is subject to Northern vocalism, namely, the deletion of the unstressed /i/ and /u and the raising of the unstressed /e/ and /o/ to /i/ and /u/ respectively, plus the commonly referred phenomenon of strengthened palatalization, that is the realization of the sonorants /n/ and /l/ as /ɲ/ and /ʎ/ respectively, following high vowel deletion. For a more detailed description, the reader is referred to Vogiatzis (2026). The data under scrutiny throughout the paper is subject to the above-mentioned phonological changes. Focusing first on the process of nominal diminution in the dialect, the repertoire of diminutives in use is as follows:

The suffix -ur(i)1 forming neuter gender diminutive nouns (see example under (1) below), attaches mainly to inanimate nouns and displays limited productivity. In the dialectal atlas of Lesbos (Ralli, 2019), this particular suffix seems to show relatively higher acceptability/higher rates in South Lesbos (and especially in Agiasos, Polychnitos and Asomatos).

(1)

-ur(i)]N:

klaˈði

klaˈð-ur ‘branch-dim

The suffix -ar(i) also forms neuter diminutives, attaching to inanimate bases of mainly feminine and neuter gender and showing extremely limited productivity in all available sources, while the high-frequency formations in the texts that contain it are fully lexicalized (e.g., miˈtar ‘nose’ and puˈðar ‘leg’, kupiˈlar ‘boy’), displaying no signs of evaluative readings. The suffix -ar(i) also shows relatively higher acceptability rates in the southern part of the island (Ralli, 2019 on the geographical distribution of the suffix in the different areas of Lesbos). Some indicative examples can be seen under (2) below:

(2)	-ar(i)]N:	vno	vn-ar	‘mountain-dim’
		ˈpita	pt-ar	‘pie-dim’

The suffix -uð(i), also forming neuter nouns, attaches to nominal bases of all three gender values as well as adjectival bases. It displays higher acceptability rates in Northeast Lesbos (Ralli, 2019). By way of illustration, consider the examples under (3) below:

(3)	-uð(i)]N:	kuˈpela	kupiˈl-uð	‘girl-dim’
		fots	fuˈt-uð	‘proper name-dim’
		munaˈxos	munaˈx-uð	‘alone-dim
		skaˈʎi	skaˈl-uð	‘step-dim’

At the same time, the corresponding feminine diminutive suffix -uða is also used in the dialect, mainly by the enclaves of the descendants of Asia Minor refugees, which is associated only with feminine bases, both animate and inanimate. e.g.,:

(4)	-uða]F:	iˈʎa	iˈʎ-uða	‘olive-tree-dim’
		aˈvʎ i	aˈvl-uða	‘yard-dim’
		ak ʎiˈsa	akʎiˈs-uða	‘church-dim’

The suffix -ˈi attaches to non-human bases of all three gender values that are not stressed in the last syllable and is mainly attested in the northern and northeastern part of the island. E.g.,:

(5)	-i]N:	ˈporta	purˈt-eʎ	‘door-dim’
		ˈkaknus	kaˈkn-i	‘turkey-dim’

Lastly, the marker -eʎ (<-eli) is synchronically the most prevalent and productive diminutivizer in all different areas of Lesbos, attaching to all different types of bases subject to diminutivization, as can be seen in the examples under (6) below:

(6)	-eʎ]N (<-eli):
	a.	ˈðaskalus	ðaskaˈl-eʎ	‘teacher-dim’	(masculine)
	b.	kaˈreɣla	kariˈɣl-eʎ	‘chair-dim’	(feminine)
	c.	karˈpuz	karpuˈz-eʎ	‘water mellon’	(neuter)
	d.	ajvaˈʎ i	ajvaˈl-eʎ	‘Aivali-dim’	(place name)
	e.	ʝaɲs	ʝaˈneʎ	‘John-dim’	(proper name)
	f.	mkros	mkr-eʎ	‘small/young’	(adjective)
	g.	mnaˈres	mna-ˈreʎ	‘minaret’	(loan)
	h.	kuduˈɣuɲ	kuduˈɣneʎ	‘a kind of clothinɡ’	(compound)

Among all different diminutivizers, the suffix -eʎ could be characterized as one of the rare cases of absolute productivity in the field of evaluative morphology of Modern Lesbian, since any base susceptible to diminution can be combined with the specific diminutive marker.

3.2. Analysis

It should be noted that this suffix, considered to be of either medieval origin (-έλλιν < Hellenistic -έλλιον) or Latin origin (-ellum) (Andriotis, 1992; Chatzidakis, 1905–1907; Ralli, 2017, p. 100), underwent intra-Lesbian dialectal differentiation. This differentiation was associated with the settlement of Asia Minor refugees in 1922 from the opposite Asia Minor coast, specifically from Aivali and the Moschonisia.

More specifically, the suffix was used by the residents of Aivali, whereas the marker -aci was used by the residents of the Moschonisia. However, even in Moschonisia, at the beginning of the 20th century, -eʎ began to prevail due to the proximity of the two cities and the resulting linguistic contact (see Sakkaris, 1940). In the synchronic form of the Lesbian dialect, this element constitutes—through accommodation processes (Hickey, 2014; Niedzielski & Giles, 1996; Ruch & de Benito Moreno, 2023)—the most typical and dominant marker of diminution in the Modern Lesbian dialect, serving as a prominent feature or otherwise as an indicator of local dialectal identity.

In the relevant literature, linguistic adaptation is defined as the set of adjustments speakers make to become more or less linguistically similar to their interlocutor or to their local or social environment (Ruch & de Benito Moreno, 2023, p. 17) and is associated with specific mechanisms and constraints, both intra-linguistic and extra-linguistic. One of the most prominent linguistic mechanisms, especially in cases of long-term linguistic adaptation, is that of ‘salience’. In sociolinguistic research, salience is the ability to recognize a variable/feature (see Labov, 2001). Salience can be interpreted as “the property that allows a linguistic variable to be an indicator” (Rácz, 2013, p. 23). In the spirit of Labov’s work (2001), markers are recognizable by linguists and non-linguists, either consciously or unconsciously, and are thought of as salient and in terms of carrying social indexation. In cases where the differences are particularly prominent and easily recognizable, the variable, that is, a specific linguistic feature, may be considered as an extremely strong stereotype (‘strong stereotype’ Rácz, 2013, p. 23). Similarly, the significance of a geographical variable lies, as with social variables, in its function as an indicator of local geographical identity or origin.

The status of certain diminutives as indicators or local stereotypes, such as the suffix -eʎ in Lesbian, could serve as a supporting factor in favor of their preservation and diffusion within a clearly delimited area, as they function as cohesive indicators of local identity. This tendency is also observed in other dialect areas, as shown by the examples that follow under (7):

(7)
	-uɖɖi:	ˈanem(o)	>	aneˈm-uɖɖi	(Griko)
		‘wind’		‘wind-dim’
	-opon:	ˈelaf(o)	>	elaˈf-opon	(Pontus)
		‘deer’		‘deer-dim’
	-o(k)ko:	koritsi	>	koriˈts-okko	(Cappadocia)
		‘girl’		‘girl-dim’
	-uin:	ˈðondin	>	ðonˈd-uin	(Cyprus)
		‘tooth’		‘tooth-dim’
	-opulo:	peˈði	>	peˈð-opulo	(Achaia-Peloponnese)
		‘child’		‘child-dim’

Each of the above mentioned diminutive suffixes is closely associated in the mind of native speakers of Greek with the specific dialectal areas. However, the extent, the quantification and any possible limitations of this specific phenomenon have not been put under scrutiny using specific methodological tools and certainly call for further sociolinguistic or perceptual dialectological research. Taking into account that evaluative markers and diminutive suffixes in particular realize a number of evaluative functions in Greek and are directly intertwined with the attitude of speakers towards the objects of reference, it would be of particular interest to study diminutives as indicators of local identities and, from this perspective, as isoglosses for the delimitation of geographical linguistic areas, not necessarily of a categorical but mainly quantitative nature.

Summarizing the dominant trends governing the morphological process of diminution, we may note that in Lesbian, as in other Modern Greek dialects (see Melissaropoulou, 2006, 2007), a small repertoire of diminutive suffixes is used systematically. Nominal diminution is closely associated with the suffix -eʎ, which is completely productive and alternates with the other suffixes. As shown in the previous section, all suffixes, as is the case in Greek in general, have a derivational status, as they are subject to specific selectional restrictions (e.g., gender, +/− animate or human, stress properties, etc.) and function as heads of the derived formations, transmitting specific morphological characteristics (gender, inflectional class) to the derived forms (see Melissaropoulou & Ralli, 2008). Furthermore, at a semantic-pragmatic level, these elements carry both quantitative and qualitative (mainly positive) connotations. Negative evaluative readings and their realization in the nominal domain of the Lesbian dialect will be discussed in Section 5.

It is worth noting (see also Melissaropoulou, 2006, 2007) that no masculine diminutives occur in the dialect of Lesbos, in contrast to both other dialects (such as, for example, in the dialect of Mani ˈc/tsipos > c/tsiˈpakos_]M) but also to SMG (potaˈmos > potaˈmakos_]M, aðerˈfos > aðerˈfulis_]M). In Modern Lesbian, the neuter—the prototypical grammatical gender for the expression of diminution—is the dominant value. It seems that, especially for this dialect, the morphological function of diminutives has acquired the neuter as the prototypical gender value (Melissaropoulou, 2006). Further support for this view comes from the formation of hypocoristics of feminine proper names in several areas of Lesbos, especially in Plomari and Petra, where the grammatical gender conveyed by the suffixes is not only feminine but predominantly neuter (e.g., tu Kaˈtiɲku ‘Catherine-hypoc’ tu Liˈɲo ‘Helen-hypoc’, etc.), often with ironic or derogatory overtones (Sakkaris, 1940, p. 105).

The relationship between gender and diminution is among the most extensively discussed issues in the literature, both from a synchronic perspective—regarding dominant patterns of realization—and from a diachronic one. One of the well-established theses in the relevant literature asserts that the concept of ‘new X’ (where X = base) has been central to the development of diminutives across genetically and typologically unrelated languages (Corbett, 1991; Jurafsky, 1996; Magni, 1999; Creissels, 1999). From a typological standpoint, Grandi (2002, p. 133) and Grandi and Körtvélyessy (2015) observe that young animate nouns (e.g., referring to children) are systematically associated with the neuter gender, which functions prototypically as a classifier, since its use highlights a perceptual indeterminacy—namely, the inability or irrelevance of specifying biological gender. In this sense, the neuter represents the unmarked or prototypical gender value for diminution. As Daltas (1985, p. 74) notes for Standard Modern Greek, “the probability of a diminutive being of neuter gender is greater than the probability of being feminine, and the latter is greater than the probability of being masculine.”

The distribution of grammatical gender in Lesbian diminutives confirms these observations in its more prototypical realization; namely, the neuter is clearly dominant, the feminine is restricted to a single suffix, attaching only to feminine bases, and the masculine is completely absent. Finally, in line with my previous work on evaluative morphology in Greek dialectal variation (Melissaropoulou, 2006, 2007, 2009), I argue that the observed differentiation in Lesbian diminution may be interpreted as reflecting a broader tendency of linguistic systems—here, in a variety relatively free, due to its low-prestige status, from strong processes of purification and standardization—toward the establishment of morphological structures that realize prototypical, unmarked patterns, thereby contributing to the overall optimization of the grammatical system.

4. Augmentation

4.1. Data

Turning now to the process of nominal augmentation, the repository of augmentative markers in use in the dialect is as follows:

The marker -a forms feminine augmentatives, attaching to [-human] paroxytone neuter bases in -i. By way of illustration, consider the examples under (8):

(8)
	-a]F:	paˈputs	paˈputs-a	‘shoe-augm’
		aˈçiʎ	aˈçil-a	‘lip-augm’
		kvar	ˈkvar-a	‘skein-augm’

The suffix -ara, also forming feminine augmentatives, combines with [+/− animate] neuter and feminine bases. E.g.,:

(9)
	-ara]F:	mit	mˈt-ara	‘nose-augm’
		kuˈrits	kurˈts-ara	‘girl-augm’
		paˈtuna	patˈn-ara	‘sole-augm’

The marker -arus forms masculine augmentatives attaching to [+/− animate] bases of all three grammatical gender values. Some indicative examples follow under (10) below:

(10)
	-arus]M:	voð	ˈvoð-arus	‘ox-augm’
		aˈraps	aˈrap-arus	‘nigger-augm’
		mat	ˈmat-arus	‘eye-augm’

The suffix -ˈ_us attaches to non-oxytone trisyllabic bases forming masculine augmentative forms. e.g.,:

(11)
	-ˈ_us]M:	puˈtana	ˈputanus	‘hooker-augm’
		ciˈfaʎ	ˈcefalus	‘head-augm’

Lastly, the marker -ukla is of extremely limited productivity (confined mainly to the area of Agiasos, based on the available sources at least) and attaches to [+/− animate] bases to form feminine augmentative forms, carrying also negative connotations. e.g.,:

(12)
	- ukla]F:	laˈɣos	laˈγ-ukla	‘rabbit-augm’
		muˈro	muˈrukla	‘baby-augm’	(mainly in Agiasos)

4.2. Analysis

The repertoire and characteristics of augmentative formations in use in Modern Lesbian confirm and lend further support to the claim that neuter grammatical gender is not realized in the process of augmentation (see Melissaropoulou, 2007, 2009). Furthermore, contrary to what is thought to be the prevailing and thus characteristic (or in terms of Daltas, 1985 unmarked) gender value, value for augmentation—namely the feminine—the dominant gender value in Modern Lesbian appears to be masculine.

Indeed, in line with the arguments advanced in my earlier work on the dialect of Aivali and Moschonisi (see Melissaropoulou, 2007, 2009), there is a strong tendency in Lesbian as well for any base susceptible to augmentation to form a masculine augmentative (as well). In some cases, more than one masculine augmentative may even be available. Some illustrative examples can be seen under (13) below:

(13)
a.	maˈçer	maˈçer-a	maçiˈr-ara		maˈçer-ar(us)		‘knife-augm’
	kuˈmat	kuˈmat-a	kumaˈt-ara		kuˈmat-ar(us)		‘piece-augm’
	kaˈlaθ	kaˈlaθ-a	kalaˈθ-ara		kaˈlaθ-ar(us)		‘basket-augm’
b.	knup		knuˈp-ara		ˈknup-ar(us)		‘mosquito-augm’
	putˈk(os)		putˈk-ara		puˈtik-ar(us)		‘mouse-augm’
	fið		fˈð-ara		ˈfið-ar(us)		‘snake-augm’
c.	ˈʝneka	ʝneˈk-ara		ˈʝnek-ar(us)		‘woman-augm’
	cʎa	cl-ara		ˈcil-ar(us)		‘paunch-augm’
d.	ˈfurn(us)	ˈfurnar(us)		‘oven-augm’		not *furˈn-ara
	ˈpseft(s)	ˈpseft-ar(us).M		‘thief-augm’		not *psiˈft-ara.F
		(examples from Melissaropoulou, 2007, 2009)

A thorough look at the examples above reveals an interesting correlation between grammatical gender values and suffix alternation, involving not only inanimate nominal bases (13a) but also animate nouns that do not encode biological sex through grammatical gender. These nouns form augmentatives with the suffix -ara as well as with -arus (13b). Feminine nouns [±animate] may form both feminine and masculine augmentatives (13c), while masculine nouns [±animate] form exclusively, or predominantly, masculine augmentatives (13d).

Moreover, the alternation of augmentative suffixes on the same nominal base is neither free nor semantically void. On the contrary, the different suffixes usually encode distinct degrees of intensity, with the highest degree of augmentation realized in the masculine gender (via the suffix -arus). In other words, suffix alternation appears to contribute to the further intensification of augmentation, thereby offering additional support for the claim that augmentation is closely associated with the masculine grammatical gender value.

To summarize the main generalizations concerning the realization of diminution and augmentation in the dialect, it becomes evident that, with respect to the function of grammatical gender, the two principal aspects of quantitative evaluation display a striking symmetry in Modern Lesbian. This symmetry highlights the dominant role of grammatical gender and its prototypical values. Whereas in Common Modern Greek all three gender values are attested in evaluative formations—and the feminine is considered the most likely value for augmentation (Daltas, 1985)—in Lesbian there is a strong tendency for the masculine gender to prevail in augmentation. According to Grandi (2002, pp. 136, 137, 182), the masculine constitutes the typical exponent of the semantic feature [+human], which is, from a semantic viewpoint, more closely linked to augmentation. In the Lesbian dialect, therefore, the masculine emerges as the basic or prototypical gender value for augmentation, the neuter as the prototypical value for diminution (Melissaropoulou, 2006; Ralli & Melissaropoulou, 2008), while the feminine is employed in both evaluative functions, as illustrated in Figure 1 below.

5. Other Productive Morphological Strategies of Nominal Evaluation

5.1. Prefixoids for the Realization of Negative Evaluation

Diminution in the Lesbian dialect is strongly related, apart from the concept ‘small’, to positive evaluative readings (affection, endearment, etc.), in line with cross-linguistic generalizations and prototypical tendencies of diminution (Grandi, 2002; Grandi & Körtvélyessy, 2015). Negative evaluation in the nominal domain is more pervasively realized either through augmentation2, aligning with the cross-linguistic evolutionary tendency of augmentatives to shift to pejorative meaning (Mutz, 2015, p. 151) or mainly through compound constituents or prefix (-like) elements of both inherited and foreign origin.

The elements that are used to realize negative evaluative functions in Modern Lesbian are paʎu- vrumu-, skatu-, ðʝavulu- and ðimunu-3, as shown in the examples schematized in

Table 1. Elements expressing negative attitudes or depreciation.

Element	paʎu-	vrumu-	skatu-	ðʝavulu-/ðimunu-
Origin	<paˈʎos.Adj ‘old’	<ˈvroma.Noun ‘stink’	<skaˈta.Noun ‘shit’	<ˈðʝavulus.Noun/ ˈðemunas.Noun ‘devil’
Examples	paʎuˈðʎa ‘pej-work’	vrumu-ˈðʎa ‘pej-work’	skatu-ˈðʎa ‘pej-work’	ðʝavuˈlo-piða ‘pej-kids’
	paʎu-ˈsoʝ ‘pej-family’	vrumu-ˈsoʝ ‘pej-family’	skatu-ˈsoʝ ‘pej-family’	ðimuˈno-spurus ‘pej-seed’
	paˈʎ-aθripus ‘pej-(hu)man’	vruˈmo-mɣa	skatu-ˈkleftis ‘pej-thieves’	ðʝavuˈlo-fitra ‘pej-germ’

below. These elements appear on the left side of the stem and their status as first constituents, prefixes or prefixoids has not been thoroughly investigated in the Lesbian dialect (However, for the element ðʝavulu- see also the corresponding lemma in HDMG (1933–2021)).

Kallergi et al. (2023, p. 134) delving into the behavior and structural properties of some of these elements in Standard Modern Greek (and some additional ones, such as vlaxo- (<Vlach) and ʝifto- (<gypsy) that do not occur in Modern Lesbian), align with Schmidt’s (1987) thesis, arguing in favor of the existence of strict categories. In this spirit, they reject the existence of intermediate categories and the subsequent use of the term affix or prefix and treat most of these elements as first constituents. As regards the group of elements that are relevant for Lesbian, they exclude the element paʎo-4 which they consider a redundant diminutive element in Modern Greek expressing negative content (‘expletive pejorative’ that only contributes [negative content]), without, however, explicitly advocating its morphological status.

For the purposes of this analysis, we abstract away from this position and align with those scholars who stress the need to distinguish intermediate categories (see Stevens, 2000, 2005; Ralli, 2020, 2022) on the basis of specific criteria. The reasoning behind this choice lies in our belief that these types of categories have cognitive reality, constituting very strong evidence of an ongoing process of morphological change.

For the purposes of this paper, we adopt the criteria proposed by Stevens (2000, 2005) for the identification of affixoids, which are as follows:

• Affixoids in spoken languages are usually very productive;
• They exist alongside a formally identical, and usually free “parent” morph (formal similarity Ralli, 2020);
• The meaning of the affixoid is more generalized and abstract than the formally identical parent;
• There is a shift in the meaning of the relationship between the parent morph and the affixoid;
• An affixoid is in competition with or in complementary distribution with an affix.

Applying each of these criteria to the elements under consideration, we can see that:

• With regard to the first criterion, the specific elements are extremely productive and can be linked to many different nominal bases, realizing negative evaluation, as shown by the indicative examples in (14) above. Moreover, it should be noted that apart, perhaps, from the word klefts ‘thief’, which could be considered to carry a negative semantic content per se, the other bases are more or less semantically neutral and the negative readings result from their combination with the specific elements.
• They do indeed coexist with morphologically identical free parent morphemes, the paˈʎos.Adj ‘old’, vroma.Noun ‘stink’, skaˈta. Noun ‘shit’ devil, ˈðʝavolos.Noun/ˈðemunas.Noun ‘devil, demon’ (see, among others, Papanis & Papanis, 2004; Ralli, 2017; Anagnostopoulou, 2021). Some indicative examples5 can be seen under (14) below:

(14)
a.	aˈdamusa		tun	Panaˈʝot	ˈenan	ˈfilum	paˈʎo
meet.pst.1.sg the.acc.sg Panagioti.acc.sg a.acc.sg friend of mine.acc.sg old.m.acc.sg
	‘I ‘I met Panagiotis, an old friend of mine’					(Dedekis, 2015, p. 269)
b.	ˈvroma ðe tsi ðisuˈzmia
	stink.nom.sg and stench.nom.sg
	‘stink and stench’					(Dedekis, 2015, p. 126)
c.	mes	tu	ˈlaku	mi ta	skaˈta
	in	the.acc.pl	hole.acc.pl	with the.acc.pl	shit.acc.pl
	‘in the shit hole’					(Anagnostopoulou, 2021, p. 72)
d.	in		kaˈmomata	t	ðʝaˈvol
	be.prs.3.pl		action.nom.pl.	the.gen.sg	devil.gen.sg
	‘It’s devil’s actions’					(Dedekis, 2015, p. 220)
e.	ðiˈmon		piˈði
	devil.gen.sg		kid. nom.sg
	‘devil’s kid’					(p.c.)

3.

Their meaning is more general and abstract in relation to the parent morphemes. In the examples shown in Table 1 above, one can clearly see that the meaning of oldness has been lost completely from the element paʎo-, similarly the meaning of stink and shit or crap from the forms vromu- and skatu- respectively, as well as the meaning of devil, demon from ðʝavulu- and ðimunu- respectively; they all converge to pejoration, that is realizing the central meaning ‘bad’, (and/or the related ‘useless’, ‘annoying’, ‘problematic’, and ‘unpleasant’). On the contrary, the original meanings are clearly retained when used as independent words (cf. the examples under (14) above and the corresponding lemma of HDMG (1933–2021)).

4.

Consequently, there is a significant change in the semantic relationship between the elements of each one of the observed pairs.

5.

Regarding the fifth criterion, the peculiarity observed here has to do with the fact that the specific elements do not seem to be in competition or in complementary distribution with an affix but in alternation with other affixes that perform the same function. As illustrated by the examples in (14), these elements may alternate on the same nominal base (e.g., paʎuˈðʎa ‘pej-work’, vrumu-ˈðʎa ‘pej-work’, skatu-ˈðʎa ‘pej-work’), yielding comparable derogatory evaluative meanings that cluster around the semantic notion ‘bad’ (and the related ‘useless’, ‘annoying’, ‘problematic’, and ‘unpleasant’). Interestingly, this is also depicted in some dialectal lexicographic works, such as the one by Anagnostopoulou (2021, p. 211), where, by way of illustration, it is explicitly commented that all following terms, namely vrumu-ˈratsa, paʎu-ˈratsa, ðʝavulu-ˈratsa ‘pej-ˈratsa’ realize the same meaning, that is ‘bad family’.

Notably, these elements may attach to either semantically neutral bases or bases that are inherently derogatory (e.g., insulting or offensive). In the latter case, they do not merely add an evaluative nuance but rather reinforce or intensify the negative semantic content already encoded in the base (e.g., ˈputanus ‘whore-augm’, paʎu-ˈputanus ‘intens-whore-augm’).

On the basis of evidence drawn from evaluative morphology, we therefore propose a reformulation of Stevens’s (2000, 2005, p. 3) fifth criterion for the identification of affixoids. Specifically, the criterion could be modified as follows: “An (evaluative) affixoid stands in competition or in complementary distribution with another (evaluative) affix(oid).” This reformulation captures more accurately the paradigmatic relations observed among evaluative formatives in the dialect. The application of the above-mentioned criteria could thus be schematized in

Table 2. Criteria for the determination of affixoids (adapted by Stevens, 2000, 2005, p. 3).

Criteria		paʎu-	skatu-	vromu-	ðʝavulu-/ðimunu-
	Elements
Affixoids in spoken languages are usually very productive		√	√	√	√
They exist alongside a formally identical, and usually free “parent” morph (formal similarity Ralli, 2020)		√	√	√	√
The meaning of the affixoid is more generalized and abstract than the formally identical parent		√	√	√	√
There is a shift in the meaning of the relationship between the parent morph and the affixoid		√	√	√	√
An affixoid is in competition with or in complementary distribution with an affix or (an)other affixoid(s)		√	√	√	√

below:

In summary, the comparative application of the relevant criteria leads to the following conclusions. The elements that function as prototypical pejoratives of nominal bases in the modern dialect of Lesbos occupy a position within the morphological continuum of affixes-affixoids. Although they coexist with morphologically identical free (parent) morphemes, they have developed semantically opaque meanings relative to those parent forms. Their meanings are evaluative and largely revolve around the semantic notion ‘bad’ (and the related ‘useless’, annoying’, ‘problematic’, and ‘unpleasant’). Moreover, they display a considerable degree of productivity and may combine both with semantically neutral bases and with bases that are inherently derogatory (e.g., abusive or insulting), functioning in this latter case as intensifiers of the inherent lexical negative meaning.

Data in hand has not revealed specific selection restrictions or tendencies affecting their combinability in the nominal domain, such as animacy, concrete vs. abstract nouns or gender, since, as can be seen in the examples above, all different types of nominal bases are eligible. It is likely, however, that these elements do not all occupy precisely the same position within the morphological continuum, nor do they exhibit identical degrees of productivity. A more fine-grained assessment of their relative status requires further investigation into their distribution, combinatorial behavior, and structural properties across bases of different grammatical categories.

5.2. Adaptation of Loan Elements as Evaluative Markers

In the dialect, two additional nominal evaluative elements appear in use, which have been transferred from Turkish and were adapted to (mainly north-eastern) Lesbian. These are the elements kara- (15) and kudza- (16), attaching to nominal bases and tend to acquire or have already acquired a prefix-like character. The first element comes from the Turkish adjective kara, meaning ‘black’, while the second comes from the Turkish adjective koca(m) –kocaman, meaning ‘big/old’.

(15)
	kara-:	a. kara-biˈcars	‘intens-bachelor’
		b. kara-ruˈfçana	‘intens-bimp’
		c. kara-ˈmostra	‘intens-showcase

(16)
	kudza-:	a. kudza-biˈcars	‘intens-bachelor’
		b. kudza-ˈmanol	‘intens-proper name’
		c. kudzam-ˈʝneka	‘intens-woman’

These elements are of moderate productivity (or even low for kudza-) and seem to be fully adapted to the morphological system of Modern Lesbian, also combining with non-Turkish bases (15b,c and 16b,c, respectively), and functioning as intensifiers. It is worth noting that in the existing dialectal dictionaries, there is no consensus as regards their morphological status. Kakaronis (2020) lists them as independent lexical items, whereas Papanis and Papanis (2004) analyze kara- as a first compound constituent and treat kudzam as an inflected adjective. By contrast, Ralli (2017) classifies both elements as prefixoids. Such variation in their treatment is naturally related to the differing theoretical orientations, linguistic backgrounds, and aims of the respective authors, and is reflected in distinct lexicographic practices (e.g., whether the items are recorded as single or separate words, with or without a hyphen).

In his doctoral dissertation on the adaptation of Turkish elements into Standard Modern Greek and its dialects, Kyranoudis (2001, pp. 710, 716) observes that the element kara- in Ottoman Turkish already exhibited both a literal meaning (‘black’) and an intensifying function. Accordingly, its grammaticalization had already been initiated prior to its borrowing into Modern Greek (and the dialects), which inherited this developing, intensifying use. Kyranoudis further argues that the process of grammaticalization reached completion in Modern Greek, whereas in Turkish it remains in a transitional stage.

By contrast, kudza- does not appear to have undergone a comparable grammaticalization process in Turkish. Moreover, given its closer semantic transparency with respect to the original meaning of the parent morpheme and its comparatively limited productivity, kudza- does not seem—at least on the basis of the currently available data—to have reached the same degree of affixization as kara-. Nevertheless, both elements appear to have been productively and creatively integrated into the morphological system of the dialect.

In this respect, it is particularly instructive to compare their behavior in other dialectal systems, such as Cappadocian, which, although shaped under conditions of long-term and intense language contact, did not integrate these elements as bound forms (i.e., as left-hand constituents or prefixoids). Cappadocian, despite its prolonged and sustained contact with Turkish, borrowed kara and koca as fully lexical items (adjectives), preserving their original meanings rather than reanalyzing them as bound intensifying elements. E.g.,:

(17)
do	goˈdzan	do	neka
the.nom/acc.sg	old	the.nom/acc.sg	woman.nom/acc.sg
‘the old woman’			Ulaghatsh, Cappadocian
ɣaˈra	piˈper
black	piper.nom/acc.sg
‘black peper’			Sinassos, Cappadocian

Importantly, this contrast does not suggest that a higher degree of contact necessarily facilitates the structural adaptation and integration of borrowed elements into the recipient language. On the contrary, the available data indicate that the relatively less intense contact between the Lesbian dialect and Turkish did not exert a catalytic effect on the grammatical structure of the dialect. Instead, the integration of the loan elements proceeded through endogenous morphological mechanisms, ultimately leading to their adaptation as evaluative, prefix-like elements.

By contrast, in Cappadocian, although contact was both long-term and markedly intense, no comparable morphological integration took place. The borrowed elements remained independent lexical items and did not acquire a bound or functional status. Rather, the sociolinguistic conditions of regressive bilingualism appear to have triggered a different type of structural change in Cappadocian, contributing to a general weakening of prefixation processes (see Melissaropoulou, 2016).

It can therefore be concluded that, in the case of Modern Lesbian, the adaptation of these loan elements is fully compatible with—and indeed conforms to—the productive morphological mechanisms already available within the native system.

6. Conclusions

Summing up, this paper addressed the realization of nominal evaluation in the Modern Greek dialectal variety of Lesbos. Particular emphasis was placed on: (a) the main evaluative strategies and morphological processes involved, as well as the functions they serve; (b) the distinctive features of evaluative formation in Lesbian in comparison with other varieties of Modern Greek; and (c) the borrowing and integration of markers serving evaluative functions in the dialect. The analysis of the attested variation revealed that dialectal systems—Modern Lesbian in particular—unaffected by the history of diglossia and processes of standardization, constitute an invaluable source of evidence for identifying prototypical tendencies within linguistic systems. These tendencies include the preference for particular gender values as defaults in the expression of diminutive and augmentative meaning; the predominance of specific evaluative markers and their role as indices of local identity; the symmetrical distribution of positive and negative evaluative functions with the use of elements identified as prefixoids (elaborating on Stevens, 2000, 2005 criteria); and the creative adaptation of borrowed elements for the realization of evaluative meanings.