What is not so (E)strange about Greek as a Balkan Language

Back in the early 1980s, I was trying to raise some research funds for a project I had in mind involving Modern Greek, and I was looking at a Social Science Research Council (SSRC) brochure about their area studies grant programs. I saw that they had a program for “Eastern European countries” and one for “Western European countries”. I thought I had better check out both programs to see where my grant application belonged because Greece historically is both east and west, and could reasonably be considered as belonging in one or the other group. However, in looking at the list of Eastern European countries, I saw expected ones like Yugoslavia (then still intact), Albania, Poland, Czechoslovakia, and others, and in the list of Western European countries, there was France, Germany, Italy, and so on, but I could not find Greece on either list. Just to be sure, I telephoned1 SSRC to inquire into the status of Greece from their perspective and was told that I was indeed reading the brochure right, and that Greek and Greece were no place, so to speak, neither east nor west. But of course we know where Greece is: it is planted firmly in the Balkan peninsula that occupies most of what can be called “Southeastern Europe” and geographically speaking, it is to the east of “eastern” countries like Albania or the Czech Republic or Slovakia, and to the west of truly eastern countries like Russia. My SSRC experience is emblematic of an attitude about Greece and about Greek that pervades much of the way Greece and the Greek language are treated in the scholarly world, that is, they are seen as neither east nor west, located

in the Balkans but with no particular significance attached to the geography. As a reflection of this attitude, works on the Greek language typically act as if the fact that Greek is spoken in the Balkans is almost irrelevant to its history and development.
While such an attitude is understandable from certain points of view, it is especially curious because there are many linguistic characteristics that Greek has in common with the other "eastern" languages in the Balkans, specifically Albanian; the Slavic languages Bulgarian, Macedonian, and some parts (mostly southeast Serbian, the Torlak region) of the Bosnian-Croatian-Montenegrin-Serbian complex; the Romance languages Aromanian, Megleno-Romanian, (Daco-)Romanian, and Judezmo (also known as Judeo-Spanish); the Indic language Romani; and Turkish. In fact, the commonalities are so great that these languages are said to form a "Sprachbund", a term borrowed from German to signify a linguistic area where languages, through intense and sustained contact in a mutually multilingual society, have come to converge with one another structurally and lexically and to diverge from the form that they held previously.
To document and thus to begin to understand this view of Greece and especially of the Greek language, I first offer a brief historiography of the study of Greek in the Balkans. From such a starting point, I then document the status of Greek vis-à-vis its linguistic neighbors by way of building a case for why detaching the recent history of the Greek language from its Balkan element is a serious mistake, both methodologically and substantively.

THE HISTOR IOGR APHY OF GR EEK WITH R EGAR D TO THE BALK ANS
First let me offer a quick overview of what has been said about Greek visà-vis the Balkans in some relatively recent treatments of the history of Greek: − Horrocks ([1997] 2010): a scant 3 pages in a c.400-page book − Moleas ([1989] 2004): no mention at all (even when potentially relevant features are discussed) − Tonnet (1993): virtually nothing; some features that have been ascribed to Balkan influence, regarding the pluperfect in Medieval Greek, are said to be of French origin These three works are all by non-Greek scholars, but the same sort of treatment-or nontreatment as the case may be-can be said with regard to Greek linguists themselves, from somewhat more distant times. Jannaris (1897: 19), for instance, recognizes the possible relevance of Balkan languages for some structural aspects of northern dialects but makes it clear that he does not see much need to pay attention to it: We see then that, from a phonological point of view, the northern and southern groups, especially towards their extreme boundaries (e.g. Velvendos in Macedonia-Crete), exhibit a very marked difference of sonantism …. It is further evident that the geographical position of the several localities, their isolation or their vicinity to foreign races, their political and internal history, have, to a greater or less extent, conduced to shape the idioms at present spoken in the various Greek communities. That these various dialects have not the same historical value needs no special comment. Thus while northern speech has been influenced by alien (Albanian, etc.) phonology, the dialects of Pontos and South Italy bear unmistakable traits of Turkish and Italo-Venetian influence. Now as phonology in every language is intimately connected with morphology, it inevitably follows that the grammar of the above specified (northern, Pontic and Italian) dialects has been, within Neohellenic times, considerably affected by extraneous influences. At the same time, a careful examination of the southern group will show that, for various reasons, these dialects have withstood foreign influence with far greater success than the northern, and so preserved the ancient phonology, substantially also morphology and syntax … with such (chiefly morphological and syntactical) changes and vicissitudes only as would be inevitable from the nature of the case and the culture or spirit of the time. It is for these reasons that students of the post-classical and subsequent history of Greek, in looking for information in the present stage of modern Greek, should direct their attention not so much to the northern as to the southern group of Neohellenic dialects. This is an interesting perspective, and Jannaris is certainly right that based on what we know of the history of Greek, the southern dialects do preserve certain aspects of the ancient language, especially as to phonology, more faithfully than do northern dialects. Nonetheless, the northern dialects are part of the Greek-speaking world, and what has happened to them, one could argue, is part of the history of Greek, whether or not the changes are due to contact with "alien" influences; that is, the facts of their development should not be ignored.
Especially telling is the statement in Andriotis and Kourmoulis (1968: 30), where the authors say that the Balkan Sprachbund is "une fiction qui n' est perceptible que de très loin" and that the commonalities are "tout à fait inorganiques et superficielles. " Moreover, Balkanists, by which I mean scholars who look at the region as a whole and at the interactions between and among the various languages and who do not just look at one language in its Balkan context, have generally paid less attention to Greek than to other major languages in the region (that is, excluding those with far smaller numbers of speakers, such as Aromanian or Judezmo); Albanian, for instance, is quite the mysterious language, certainly the stepchild of Indo-European linguistics and thus less well-known and obscure, but that fact gives it a certain allure and attraction, so that there are numerous works that pertain to it in its Balkan aspect (mostly not by western scholars, however). The fact that it is spoken in six countries-Albania and Kosovo as the main ones, but also Montenegro, North Macedonia, and Italy, as well as … Greece, through both the now somewhat moribund but once quite vibrant communities of Arvanitika speakers who entered and settled in Greece some 500 or so years ago, and the more recent immigrant communities in Athens, Ioannina, and elsewhere-gives it a certain importance too (though the same could almost be said about Greek, inasmuch as it is spoken in Greece, in Albania, and in Italy, with enclaves too in Bulgaria and Turkey).
And, perhaps most importantly, most Balkanists (on the linguistic side) are by training Slavicists, lured into work on the Balkans by the intriguing parallels between several of the South Slavic languages and other non-Slavic Balkan languages, as well as the ways in which Balkan Slavic languages diverge from the rest of Slavic (e.g., regarding the system of cases in nouns). Indeed, from an historical point of view, it cannot be denied that most of the work done on the languages of the Balkans as a group has been by Slavicists; I have in mind early contributors like the Slovenian scholars Jernej Kopitar (1780-1844) and Franz Miklosich (1813-1891), as well as Roman Jakobson and Nikolai Trubetzkoy, in the 1920s, both of whom were Slavicists by training even if their interests were more general, and whose views on the Balkans was also important to understanding the linguistic situation there. Furthermore, the scholar who was the benefactor of the professorship I hold, 2 Kenneth E. Naylor , a South Slavic specialist who was also known as a Balkanist, should be added to this list. The Slavic orientation holds as well among Balkan linguistic scholars who are still living; I note, for instance, the following, listed roughly according to their age: − Helmut Schaller − Jack Feuillet − Ronelle Alexander − Petya Asenova − Victor Friedman − Grace Fielder − Andrey Sobolev as among those who began their scholarly lives as Slavicists and got into Balkan linguistics through Slavic; some, admittedly, especially Asenova, Fielder, and Sobolev, do give scholarly mention to matters Greek in some of their work.
There are some notable exceptions, most particularly Eric Hamp, sadly recently deceased (February 17, 2019) at the age of 98, whose interests are so broad that it is hard to say he got into Balkan studies just through one language, but whose dissertation (1954) was on the Albanian of southern Italy. Mention should be made here also of Christos Tzitzilis of Aristotle University of Thessaloniki, though he comes at Balkanistics from a Slavic orientation due to his studies in Bulgaria. A more relevant exception among 20th century scholars was the late Kostas Kazazis in that he was a Hellenist who extended himself into the other languages of the Balkans. And without wanting to seem self-promoting, I think it is fair to say that among current Balkanists, I am just about the only one who has come at the study of the Balkan languages from Greek (upwards, as it were, geographically, as opposed to downwards from Slavic). This is not to say that papers on Greek topics are not to be found in Balkanist conferences and Balkanistic journals, but that is because such venues allow within their ambit studies of individual languages, without requiring attention to the Sprachbund aspect of the Balkans.
Interestingly, looking back on Balkan linguistic historiography, it can be noted that it took a non-Slav, non-Slavicist, non-Greek, non-Hellenist scholar, Kristian Sandfeld, the Danish Romance scholar who was a specialist in the Classics and especially Romanian, to elevate the study of the Balkans from a linguistic standpoint to a high level. His 1926 work, in Danish Balkanfilologien but known mainly from the 1930 French translation, Linguistique balkanique: Problèmes et résultats, really focused attention on the Balkans as a linguistic area and contact zone with a large number of interesting shared traits that deserve particular mention and attention from scholars.
There are other factors that have played into the dominance of the Slavic line in Balkan linguistics, such as the fact that Romance scholars for the most part seem not to have cared much about Romanian over the years, in comparison to the intense interest in French, Spanish, and Italian. Moreover, the relative accessibility of Yugoslavia and even Bulgaria in the post-WWII era, before the fall of the Soviet Union, gave Slavicists a place to visit and to do research in where, given the nature of the differences between South Slavic and the rest of Slavic-differences largely due to Sprachbund-related language contact-they would often be drawn into Balkan linguistics, but again, from the Slavic perspective as their starting point.

THE GENESIS OF THIS ARTICLE
So, why do we find a general rejection of the Balkans by Greeks and a relative lack of interest in Greek by Balkanists? The latter may be due, as suggested above, to the fact of how it was that many Balkanists got into the field, i.e. coming from a Slavic perspective. For the former, however, one probably has to look, to a large extent, to ideology, especially as far as Greek linguistic scholarship is concerned (see Joseph 1985), 3

which mirrors the ways in which
Greek folklore studies and ethnography were affected, as discussed by Herzfeld (1982).
Nonetheless, some part of the answer may also come from insights to be gained from a lecture given at Princeton University in February 2013, sponsored by the Modern Greek Studies Program there. In particular, the renowned Greek historian, Professor Basil C. Gounaris of the Aristotle University of Thessaloniki, an Onassis Foundation Senior Visiting Scholar, 4 spoke on "Greece and the Balkans: A Story of a Troubled Relationship (19th-20th Centuries)". His abstract is worth quoting in its entirety to give an idea of his argument concerning the relationship of Greece to the Balkans: Before the ideas of Enlightenment and Hellas were infiltrated in the Balkan world, Balkan peoples shared a common mentality. Greek-and Vlach-speaking merchants topped the Christian social pyramid and it was their self-esteem and their economic prosperity which transformed enlightenment ideas into Greek nationalism. The glory of ancient Hellas gave a special meaning to their superiority. Through education it became increasingly clear that Greeks had absolutely no relation with the Slavs, formerly thought to be their brethren in God and in servitude to Islam. In other words Hellenisation could not be accomplished and turned into effective nationalism unless all links with the Balkan peoples were cut off. This paper argues that this process of estrangement was no easier or smoother than the transformation of the Greek-orthodox society itself into a Modern Greek nation. In fact the Balkan peoples and states became for the Greeks the convenient point of reference for evaluating social modernisation, politics, financial progress and irredentistic efforts. Furthermore it is argued that 3 The ideology also of Greek as "one language" diachronically and diatopically, as discussed in , may also have played a role in this phenomenon, since it would seem to deny the significance of dialectal variation and contact leading to divergence from Ancient Greek. 4 The publicity for the lecture described Professor Gounaris' considerable scholarly accomplishments as follows: Professor Gounaris' lecture afforded an ideal opportunity to explore the very interesting contrast between the "estrangement" of Greece and Greek society from the Balkans and the very profound influence the Balkans have had, and continue to have, on Greece from a linguistic standpoint. So, I take here this opportunity to carry out this exploration in print. 5 First My claim is that whereas recognizing the foreign, the alien (as Jannaris put it) in the development of the Greek language is not at all (e)strange-indeed the foreign has helped to shape Greek and to make the modern form of the language into what it is today, the southern varieties as well as the northern ones that Jannaris was so dismissive of-estrangement may have been necessary for the development of the Greek nation. That is, from a linguistic standpoint there is a longer history of engagement than of estrangement between Greek peoples and the Balkans. Interactions between Greek speakers and speakers of other languages in the Balkans have had profound effects on the Greek language that last to this very day.
Accordingly, I present here a side of Greece, namely the Greek language, that is not estranged from the Balkans, and explore the ways in which Greek has been affected by, and has influenced, other Balkan languages and the ways in which it can be considered to be a Balkan language.

LINGUISTIC PR ELIMINAR IES ON THE BALK ANS
To set the stage, I offer as a preliminary an overview of the languages in question here. The Balkans have been a hotbed of multilingualism and language contact since ancient times, 6 but given my focus here on the interactions Modern Greek has had with its neighboring languages, I concentrate just on the medieval and modern era, the periods during which the Balkan Sprachbund took shape.
There is an important distinction to be made between languages that are geographically in the Balkans, what can be called "languages of (or in) the Balkans", and languages that show significant convergence in structure and lexicon due to contact among their speakers, that is to say, languages that participate in and form the Balkan Sprachbund, what can be called "Balkan languages".

Languages of/in the Balkans
The following languages can be identified as the "languages of/in the Balkans", given here along with some brief notes as needed and as appropriate; omitted here are languages of very recent in-migrations, e.g. by Urdu speakers who have settled recently in Greece, and international languages in wide use such as English or

Balkan Languages
The following languages can be identified as the "Balkan languages", given here along with some brief notes as needed and as appropriate. They are a subset of the languages of/in the Balkans given in §4.1, and are those languages that participate to some significant extent in Balkan Sprachbund; varieties that are less involved in the Sprachbund are given in italics, though they differ considerably in degree of involvement: − Albanian (both major dialects, though to different degrees: Geg (North) and Tosk (South)) − Bulgarian − Greek (various dialects, including Tsakonian (but excluding Asia Minor dialects)) − Judezmo 7 − Macedonian − Romanian (actually more specifically Aromanian (spoken in Greece, North Macedonia, and Albania), and Meglenoromanian (spoken in a few villages in Greece and North Macedonia near the border between these two countries), less so Romanian (the national language of Romania and Moldova) and even less so Istro-Romanian) − Romani − Serbian (really only the Torlak dialects of the Southeastern Serbianspeaking area as most relevant; much less so Bosnian, Croatian, and Montenegrin) − Turkish (as in §4.1, not a "full" structural participant but crucial nonetheless) A useful terminological point that emerges from this listing is that Bulgarian, Macedonian, and Torlak Serbian can be said to constitute "Balkan Slavic" (i.e., that part of the Slavic group that is fully in the Balkan Sprachbund), and similarly, Aromanian, Meglenoromanian, and to some extent Romanian itself can be classified as constituting Balkan Romance. To follow up on this presentation of the Balkan languages, we can now turn to the features that characterize the Balkan languages, that is to say, the features on which significant convergence among the languages in §4.2 is found.

BALK AN CON VERGENT FEATUR ES
In order to see where Greek stands with respect to the Balkan Sprachbund, it is necessary to survey the features on which the Sprachbund languages converge, so-called "Balkanisms". Unfortunately, no definitive list can be easily compiled of all such features, due in part to the vastness of such an undertaking, as there are so many points of convergence, but also due to methodological issues that are hard to resolve, such as how many languages need to be in on a convergent feature for it to be significant. 8 I sidestep those issues here by giving a list of fifteen Balkanisms that have been discussed the most in the literature. These are but a small glimpse, in a sense, of the overall convergent picture but they are representative and have commanded the attention of analysts over the years. Moreover, they cover various levels of linguistic analysis: morphosyntax (a-g), semantics/pragmatics (h), syntax (i-j), and phonology (k-o); I add some lexical (and other more restricted) convergences in a later section ( §8). I give a description of each such feature, without giving details or a lot of the relevant data, but I illustrate each one with an example from Modern Greek, where possible, or from one other language, in order to give readers a sense of what is involved in each: (1) A selection of Balkan convergent structural features a. a reduction in the nominal case system, especially a falling together of genitive and dative cases, e.g. Greek του ανθρώπου 'of the man; to a man' (continuing earlier Greek genitive τοῦ ἀνθρώπου, dative τῷ ἀνθρώπῳ) b. the formation of a future tense based on a reduced, often invariant, form of the verb 'want' , e.g. Greek θα γράψω 'I will write' (from earlier θέλει να γράψω, literally "it-wants that I-write") c. the use of an enclitic (postposed) definite article, typically occurring after the first word in the noun phrase, e.g. Albanian njeri 'man' ~ njeriu 'the man' d. analytic adjectival comparative formations, e.g. Greek πιο όμορφος 'more beautiful' e. marking of personal direct objects with a preposition, e.g. Aromanian U vâdzuj pi Toma 'I see Toma' (literally "him I-see PREP Toma") f. double determination in deixis (= a demonstrative adjective with a definite article and a noun, e.g. Greek αυτός ο άνθρωπος 'this man' (literally "this the man")) g. possessive use of dative enclitic pronouns, e.g. Bulgarian knigata mi 'my book' (literally "book-the to-me") h. the use of verbal forms to distinguish actions on the basis of real or presumed information-source, commonly referred to as marking a witnessed/ reported distinction but also including nuances of surprise (admirative) and doubt (dubitative), e.g. Albanian qenka 'I allegedly am' i. the reduction in use of a nonfinite verbal complement ("infinitive") and its replacement by fully finite complement clauses, e.g. Greek πώς τολμάς να μου μιλάς έτσι 'How dare you speak to me like that' (literally "how you-dare that to-me you-speak thus"); cf. Ancient Greek εἰ … τολμήσεις … ἔγχος ἀεῖραι 'if you dare to raise (your) spear' (literally "if you-dare spear to-raise", Iliad 8.424) j. the pleonastic use of weak object pronominal forms together with full noun phrase direct or indirect objects ("object doubling"), e.g. Greek σε είδα εσένα 'you I saw' (literally "you I-saw you") k. the presence of a (stressed) mid-to-high central (thus, schwa-like) vowel, e.g. Albanian ë l. the presence of i-e-a-o-u in the vowel inventory without phonological contrasts in quantity, openness, or nasalization, e.g. Greek ι ε α ο ου m. voicing of voiceless stops after nasals (NT > ND), e.g. την τάση (pronounced [tin dasi]) 'the tendency' (accusative singular) n. presence of ð θ γ, as in Greek o. elimination of palatal affricates in favor of dentals, e.g. Greek τσιπς 'chips' (pronounced with dental [ts] even though from English chips (with palatal [t∫]) With this set of features established, the question of the position of Greek among the Balkan languages, i.e. whether it is part of the Balkan Sprachbund, and if so, to what extent, can be taken up.

DISTR IBUTION OF FEATUR ES
Crucial to an answer to the question of Greek as a Balkan language is the determination of which of the features listed in §5 occur in Greek. As already indicated by the fact that some of the features in §5 are exemplified by material from languages other than Greek, it is the case that not every feature is found in all of the Balkan languages. Accordingly, the distribution of these features is given here, where * signals partial or dialectal (as opposed to Standard language) realization, "Slavic" means the feature occurs generally across Balkan Slavic and "Romance" that it occurs generally across Balkan Romance. Given the focus herein on Greek, the fact of a feature being found in Greek is highlighted by the occurrence of "GRK" in bold capital letters, and those features which are not instantiated in Greek are specially marked by being given in italics. It must of course be noted that even if a feature occurs across the Balkans, it need not have arisen due to contact with other languages, as it could be an independent innovation in various languages; the matter of origins for the features is taken up in §7.
(2) The distribution of the features in (1) a. a reduction in the It is misleading to think of the Balkan Sprachbund as being determined in purely quantitative terms, judged by a scorecard of pluses and minuses with regard to a selection of linguistic features. Among other considerations, it is especially hard to quantify the cases of partial involvement and it is also the case that not all features necessarily count equally in terms of their effect on the overall structure of the language and how a language looks relative to the other languages; some of the phonological features, for instance, might affect only a relatively small number of morphemes in a given utterance.
Nonetheless, it is striking that 11 of the 15 features considered here find realization in Greek. Such a preponderance of representation of Balkan features in Greek intuitively gives a solid basis for considering Greek to be a true Balkan language and therefore a part of the Balkan Sprachbund.

THE DIACHRON Y OF THESE FEATUR ES IN GR EEK
Another dimension to the assessment of Greek as a Balkan language is the matter of how many of these features represent divergences from earlier stages of Greek-as noted in §1, with the convergence characteristic of the contact that creates the cluster of geographically connected languages referred to as a Sprachbund, there is typically divergence away from the structures and lexical forms that characterized these languages prior to the contact. This means that another index of the Balkan character of Greek is the extent to which the convergent features represent innovations away from the structures and vocabulary of earlier stages of Greek. In the case of Greek, we are fortunate in having the extensive documentary record of Ancient Greek, especially Greek of the Classical and post-Classical eras, so that it is possible to determine which features reflect changes that are candidates for Balkan contact-induced effects.
Four of the features under examination here are irrelevant for this question as they are not found in Greek at all: (3) Features from (2)  Of the remaining features, the ones that diverge from what is found in Ancient Greek are given in (4).
(4) Features from (1) that are innovative within Greek a. a reduction in the nominal case system b. the formation of a future tense based on 'want' d. analytic adjectival comparative formations i. the reduction in use of a nonfinite verbal complement ("infinitive") and its replacement by fully finite clauses j. the pleonastic use of weak object pronominal forms together with full noun phrase direct or indirect objects ("object doubling") l. the presence of i-e-a-o-u in the vowel inventory without phonological contrasts in quantity, openness, or nasalization m. voicing of voiceless stops after nasals n. presence of ð θ γ o. elimination of palatal affricates in favor of dentals Only features (f), double determination in deixis, and (g), possessive use of dative (genitive) enclitic pronouns, represent carry-overs from constructs found in Ancient Greek. Thus in 9 of the 11 features under consideration here that are found in Greek, we see structural changes on the way to Modern Greek.
Moreover, of these 9 features, it is possible to gauge how many are likely to be the result of or to have been enhanced by "alien" influence on Greek, i.e. due to contact with other languages-these are highlighted in bold below-as opposed to being a Greek-internal development, where the chronology often can tell us the extent to which contact was involved. For instance, a reduction of the case system, with the loss of the dative case, is evident in New Testament Greek and thus clearly predates Balkan contact. 9 This is admittedly a difficult determination to make definitively in some instances, in that some features show beginnings in pre-Balkan-contact times but accelerate in later Greek under conditions of contact; such is the case with the pleonastic use of weak object pronouns, for instance (see Janse 2008) and the developments with the infinitive (see Joseph 1983). Still, here is the list of features as run through this filter, again with (c), (e), (h), and (k) excluded, and now also (f) and (g), as they are irrelevant to this aspect of the assessment: (5) Innovative features in Greek likely due to language contact a. a reduction in the nominal case system b. the formation of a future tense based on 'want' d. analytic adjectival comparative formations i. the reduction in use of a nonfinite verbal complement ("infinitive") and its replacement by fully finite clauses j. the pleonastic use of weak object pronominal forms together with full noun phrase direct or indirect objects ("object doubling") l. the presence of i-e-a-o-u in the vowel inventory without phonological contrasts in quantity, openness, or nasalization m. voicing of voiceless stops after nasals n. presence of ð θ γ o. elimination of palatal affricates in favor of dentals This calculus suggests that Greek developed various Balkan-like features-just under half of those at issue here-on its own, or at least started down that path to showing such structures, prior to the period in medieval times of significant contact with other Balkan languages. Although the numbers here are not as clear-cut as the others reported on above, they do not vitiate the claim that Greek is fully Balkan in many respects.
There are several reasons for this assessment. First, the occurrence of some of the features in other languages may be due to contact with Greek, so that even if some features within Greek have a Greek-internal origin, Greek would be part and parcel of the overall convergence zone. As it happens, the origins of the Sprachbund is actually a rather complicated question that has been the subject of much discussion and cannot be resolved here; 10 still, it can be said that not all Balkanisms can be due to Greek influence-at the very least, since Greek does not have a postposed definite article, it could not have been the source of that feature in other languages. Second, even if a feature has a Greek-internal starting point, it could have gained scope within the language through contact, with influence from other languages enhancing the feature's viability within Greek. Third, it is not at all clear how many features are needed for a language to qualify as "Balkan"; as noted earlier, this judgment is not based simply on a scorecard of pluses and minuses-there has to be a qualitative dimension as well. Finally, even if of native/internal origin, the occurrence of a particular feature that is parallel to one found in another language in close contact gives a surface sameness between the languages, thus feeding the impression of a Sprachbund for bilingual speakers, regardless of the ultimate cause of the sameness.
Moreover, there are other features that can be considered, as the next section makes clear. I turn to those next.

ADDITIONAL FEATUR ES
As noted in §5, the features that have been considered in §6 and §7 are just a subset of the full scope of convergent features linking the Balkan languages to one another. Thus, there are others, actually many others, but in this section, a few additional features, of two types, are mentioned here. First, there are a few features that are quite restricted in Greek, in that they are found just in certain regional dialects (and thus not in the standard language) and not widely distributed across the entire Greek-speaking realm. Inasmuch as such features are not widespread across all of Greek, they might be viewed as being less significant for judging the "Balkanness" of the language. However, since overall, and for each feature even, the degree of involvement of a particular language can vary, these restricted features are not irrelevant. Moreover, they are no less real for the varieties in which they occur, and thus must be taken seriously. Second, there are features that are not structural in nature but rather involve lexical material.

Dialectally restricted features
The quote from Jannaris (1897) in §2 indicates that northern dialects of Greek show some effects of contact with other languages in the Balkans that are not found in other dialects. Two areas of grammar where such dialectally restricted features occur in Greek are phonology and morphosyntax, as detailed in the following subsections.

Phonology
One feature found in northern Greek dialects is the raising of unstressed mid-vowels ([+mid] > [+high]), this e > i and o > u. This raising is exemplified by forms such as άνθρουπους 'man' (nominative singular, vs. Standard Greek άνθρωπος) and πιρίμινι 'wait!' (imperative singular, vs. Standard Greek περίμενε!). This raising is found marginally in Albanian, in Judezmo (though under slightly different conditions so it may not be the same feature in a certain sense), and in Balkan Slavic. It is an innovation when compared with earlier stages of Greek, as reflected still in the standard language, based as it is on southern varieties (recall Jannaris's quote), and thus, given its geographic restriction, is plausibly to be attributed to language contact. In this way, therefore, northern Greek is brought in line phonologically with more centrally located Balkan languages.

Morphosyntax
In the realm of morphosyntax, there are two noteworthy features in northern dialects of Greek that show affinities with other Balkan languages.
First, in Thessalian Greek, as reported in Tzartanos (1909)-see (6a)-but also with a broader distribution in northern varieties, as reported in Thavoris (1977) and Ralli (2006)-see (6b)-an innovative placement of a weak indirect object pronoun occurs with plural imperatives. In particular, instead of the expected occurrence of the pronoun outside of (to the right of) the plural marker -τι (with raising of earlier -ε to -ι, as in §8.1.1), the pronoun is positioned inside of (i.e., to the left of) the plural marker; for instance, one finds (here and in (7), hyphens have been added to make the parsing of the morphemes more evident): 11 (6)  From a language contact perspective, this innovative placement is interesting because it mirrors exactly what is found in Albanian with plural imperatives (cf. Newmark et al. 1982, Rasmussen 1985, Joseph 2010): The geographic restriction of this phenomenon in Greek and the availability of a model from Albanian, spoken in some parts of central and northern Greece, makes a claim of language contact suggestive as a basis responsible for this innovation.
Second, in the dialect of the northern Greek prefecture of Kastoria, as described by Papadamou and Papanastassiou (2013), there occurs an impersonal use of the nonactive voice verb form together with an indirect object pronoun to indicate internal disposition, what can be translated as "feels like". For instance, they cite the following (showing northern raising of unstressed -ε/-αι to -ι, and accusative for genitive): These constructions are perfectly acceptable for these northern speakers, and are constructed as if standard Modern Greek, contrary to fact, allowed sentences like μου τρώγεται / μου πίνεται in the intended meaning.
What makes the sentences in (8) of particular interest in the Balkan context is the fact that other Balkan languages in the region, the same construction is found, with an impersonal nonactive verb and an indirect object personal pronoun, as in (9) The Balkan Slavic and Aromanian use of the reflexive pronoun with a 3rd person active verb form is the Slavic and Romance equivalent of the nonactive verb form in the Albanian and the Greek. This appears to be a Slavic construction in origin, as it is found in Slavic languages outside of the Balkans, so its occurrence in Kastoria Greek is clearly a contact-induced innovation, moving that variety in the direction of other Balkan languages it is in contact with.

Lexicon
The features discussed so far have been grammatical in nature, ranging over phonology, morphosyntax, syntax, and semantics, and it is certainly true that scholarly attention regarding the Sprachbund has long been on matters of grammatical convergence. However, there is an important lexical dimension to the Sprachbund as well, and the relevant evidence bears in important ways on the assessment of Greek as a Balkan language.
It is well documented that the lexicon is generally the first component in a language to be affected by contact, through the appearance of loanwords (borrowings) passing from one language into another. Not surprisingly, one can find numerous words that are shared across languages of the Balkans.
Greek is the source of many terms having to do with Orthodοx Christianity, for instance: 12 Moreover, there is another significant lexical group of wide distribution in the Balkans consisting of words of Turkish origin, especially administrative and Islamic terms and words associated with aspects of urban commercial life, a reflection of the fact that Turkish was the key language of Balkan urban areas during the period of Ottoman rule, but also covering terminology for food, names for items of material culture, and the like. Among such words of Turkish origin are the following, constituting a representative sample (meanings the same as the Turkish source; / separates variants within a given language): paça 'tripe, trotter': Alb paça, Aro pãce, Blg pača, Grk πατσάς, Mac pača tencere 'pot; cooker': Alb tenxhere, Aro tengire, BSl tendžere, Grk τεντζερές, Jud tenǧere, Rmn tingire In a certain sense, such culturally related loans represent a somewhat trivial sort of language contact effect, in that all they do is demonstrate that contact of some sort occurred, but they really say nothing about the nature of that contact. Even very casual contact can yield cultural loans of this sort. What is more telling than these regarding the Balkan lexicon is the penetration of a different class of elements into the lexicon of the various Balkan languages. According to Friedman and Joseph (2014;2021, Chap. 4) what is essential for understanding the Balkan Sprachbund is the recognition of a class of conversationally based loans that they refer to as "E.R.I.C. " loans. This label is an acronym for borrowings that are "Essentially Rooted In Conversation", 16 and their presence reveals something very interesting with regard to the nature of language contact in the Balkans. These loans go beyond the simple informational needs and the object/goal orientation that speakers of different languages who are interacting with one another have. Borrowing such words is not dictated by prestige or need, two of the most common motivations for loanwords; instead, E.R.I.C. loans are forms that can be exchanged only via direct conversational interaction, and they cover elements that include discourse particles, terms of address, greetings, exclamations, interjections, and the like, and therefore reflect a more human side of speaker interactions. Friedman and Joseph argue that the conditions of close and sustained contact that yield such lexical convergence, what they refer to as mutual multilateral multigenerational multilingualism, are also precisely the right type of social context in which Sprachbund-like structural convergence can emerge as well. Thus E.R.I.C loans point to conditions that are Sprachbund-conducive, as opposed to loans that take place under casual contact situations. E.R.I.C. loans are all over the Balkans, as documented extensively in Friedman and Joseph (2021, Chap. 4), and, significantly for the discussion here, such loans are found in Greek. Many are from Turkish, but their source is not limited to Turkish, and indeed some of the E.R.I.C. loans in the various languages have their origins in Greek. In (12), a very small sampling of such conversational loans is given:  Two very widespread conversational and discourse-related forms deserve special mention. The first is what Pring (1975, s.v.) calls an "unceremonious term of address", roughly 'hey you' but with many nuances of meaning and usage and a great many variant forms, almost all ultimately from Greek (cf. Joseph 1997): (13) Forms of an unceremonious address term in the Balkans Alb: o, ore, or, mor, more, moj, ori, mori, moré, mre, voré,  Greek here has forms such as μωρέ, μπρε, βρε, ρε, αρέ, μαρέ, μαρή, ωρέ, βορέ, etc., some 55 variants in all. The second is the various forms with an -m-nucleus meaning 'but' , of varied-and disputed-origins, and various uses (cf. Fielder 2008Fielder , 2009Fielder , 2015Fielder , 2019: (14) -m-based words for 'but' in the Balkans a. ama, ma, ami, mi (as discourse marker and conjunction) Aromanian Greek Bulgarian Macedonian Meglenoromanian b. ama, ma only (as discourse marker and conjunction) Albanian Judezmo Romani Turkish c. ama, ma (as discourse marker only) Romanian In some instances, it is not specific words that are borrowed but rather the semantic structure of a word or phrase, resulting in a calque or loan translation: (15) Some Balkan calques Trk kötek yemek 'get a beating' (literally "eat a blow"): Mac jade k´otek, Grk τρώγω ξύλο (literally "eat wood") Grk το ξέρω απ' έξω 'I know it by heart' (literally "it I-know from outside") = Agia Varvara Romani (Messing 1988: 61) dzanav-les avral (avral = 'from outside, from abroad') Relevant here too are various everyday expressions that match in the different languages but for which the directionality of borrowing is unclear; an example is the common greeting exchange in (16)  E.R.I.C. loans can also add color and affect to conversation; the highly expressive and mildly dismissive m-reduplication of Turkish, e.g. kitap mitap 'books (kitap) and such' , is an example of such an affective borrowing throughout the Balkans. Examples are given in (17): (17) m-Reduplication in the Balkans Blg knigi-migi 'books and such' Mac kal-mal 'mud or whatever' Alb cingra-mingra 'trivia' çikla-mikla 'tiny bits and pieces; crumbs; trivia' Grk τζάντζαλα-μάντζαλα 'this and that' ("rags and such"), πιπέρι-μιπέρι 'pepper and such᾽, καφέ-μαφέ 'coffee and such' , ιδού-μιδού 'see here, or whatever' 18 άρα μάρα 'who cares?' άρες μάρες (κουκουνάρες) 'nonsense' 19 Moreover, many ERIC loans are members of closed lexical classes, representing vocabulary domains that are generally held to be somewhat resistant to borrowing, and yet they are borrowed. These classes include kinship terms, pronouns, negatives, complementizers. Friedman and Joseph (2014;2021, Chap. 4) argue that the same intense and intimate conditions that yield the conversational borrowings offer the opportunity for the borrowing of these closedclass items. Some representative examples from these classes are given in (18): (18) Closed-class borrowed E.R.I.C. items Trk baba 'father': Alb baba, Aro baba, Grk μπαμπάς 'dad' Grk μου 'my': Aromanian -m (vs. native -n'i; from Latin mihi, presumably via *mnihi) Trk yok '(emphatic) no!': Grk γιοκ Grk ότι 'that': Mac oti 'that' Grk ό,τι 'for that reason': BSl oti 'because' E.R.I.C loans are thus found all over the Balkans and bespeak an intense sort of contact at a very human and personal level. In this way, therefore, even the lexicon provides some insight into the degree of Balkan integration that is seen in Greek. Moreover, the fact that Greek is both a donor and a recipient of E.R.I.C. loans means that Greek was a full participant in the contact that led to the Sprachbund, a relevant consideration when judging the degree of "Balkanness" that the language shows. 9. CONCLUSIONS geographically highly localized nature of the impersonal "feels-like" construction discussed in §8.1.2 suggests a relatively recent origin, inasmuch as it has not spread to other local varieties of Greek, and Lavidas and Tsimpli (2019) document the innovative omissibility of direct objects with specific reference in Modern West Thracian Greek, the local dialect of Evros, under conditions of contact with Turkish.
The answer, then, to the question implicit in the title of this piece is that there is nothing strange or estrange about considering Greek to be fully "Balkan" in all respects.

Brian Daniel Joseph
The Ohio State University joseph.1@osu.edu BIBLIOGR APHY