Greek-Canadian Koiné: The Emergence of a Koiné among Greek-Canadian Immigrants

Abstract

The present paper is a contribution to the study of a new Modern Greek variety that is spoken in Canada by first-generation immigrants who arrived in this country between 1945 and 1975. This variety displays features originating from: (a) A Common Modern Greek spoken in Greece around the middle of the 20th century, (b) mutually intelligible characteristics of the immigrants’ native dialectal varieties, mainly from the Peloponnese, (c) contact with English, (d) Standard Modern Greek. We present, discuss, and analyze data collected within the framework of the project “ImmiGrec: Stories of Greek immigration in Canada.” We focus on linguistic elements that could be considered indicative features of a Greek-Canadian Koiné, more particularly by investigating the borrowing and integration of English nouns and the variation in the use of the unstressed syllabic augment /e-/ and two different imperfective past suffixes.

1. Introduction

Human history is a history of migration and, as a result, every country in the world includes communities who speak different languages. The linguistic interaction between host and immigrant populations leads to contact-induced changes in both languages (although primarily in the languages of the immigrant communities). Understanding these changes, the linguistic constraints that shape them, and the social parameters that determine whether they take a hold in the language, or whether they are ephemeral, has proven to be a very fertile field of research for linguistics from both a theoretical and an applied perspective (Myers-Scotton 2002, 2006; Clyne 2003).

This article is concerned with the speech of first-generation Greek immigrants who arrived in Canada between 1945 and 1975. It examines how their Greek evolved in a language-contact situation, where English is the dominant donor language and Modern Greek (MG) is the recipient, but also in a situation where speakers of different MG varieties are living and interacting in the Greek communities in Canada. It discusses the various linguistic resources available to immigrants and highlights both the innovation and the maintenance of features that were already present in MG spoken in the homeland and their transfer within the Greek-Canadian community. Specifically, it examines: (a) The patterns of usage of the syllabic augment and two different imperfective past (i.e., imperfect tense) suffixes in the speech of Greek-Canadians from various dialectal backgrounds, (b) the integration of borrowed English nouns and the role played in this process by the morphology of the recipient language, (c) the types of linguistic practices prevalent in the Greek-Canadian communities and how they affect their Greek. Thus, the paper provides a description of the emergence of a new variety in the Greek communities of Canada, as well as a contribution to the study of immigrant speech in general.

Our evidence comes from the speech of 12 different communities across four different Canadian provinces: Vancouver, Victoria, Edmonton, Calgary, Regina, Ottawa, Gatineau, Montreal, Toronto, Markham, Winnipeg, and Mississauga. The data are drawn both from written and oral sources. We examined published work, such as Aravossitas (2016) and Maniakas (1990), and analyzed speech from interviews recorded for the purposes of the large-scale interdisciplinary research program, “ImmiGrec: Stories of Greek immigration in Canada” (2016–2018; https://immigrec.com/, accessed on 20 March 2022, Ralli 2019), funded by the Stavros Niarchos Foundation. The examples mentioned in this work, both words and phrases, are written with the characters of the Greek alphabet and in IPA transcription, while they are also given an English translation. For reasons of space, a repeated example is only used with the IPA transcription.

The article is organized as follows: Section 2 sketches the socio-historical background of Greek immigration in Canada and defines Canadian Greek (CG) as a Greek immigrant variety. Ιn Section 3, our methodology and data are presented. Section 4 provides an overview of the main characteristics of Common Modern Greek (CMG), which was spoken by first-generation Greek immigrants at the time of their arrival in Canada. It consists of features from the learned register, Athenian (the variety spoken in the capital of the Greek state), and several regional varieties. Section 5 discusses the usage patterns regarding the unstressed syllabic augment, the choice of imperfective past suffixes, as well as the integration of English loan nouns, particularly with respect to grammatical gender and inflection class (IC) assignment. We conclude with a review of our main findings and their theoretical implications.

2. Greek Immigration and Canadian Greek: An Overview

Greeks began immigrating to Canada during the age of great migration (1850–1913) (Vlassis 1953). In 1911, the total Greek population in Canada was 3614 people (Vlassis 1953), and before the Second World War (WWII) the Greeks in Canada were estimated at about 12,000 (Constantinidis 2004). In the period between 1945 and 1975, more than one million Greeks emigrated (mostly to Western Europe, the U.S., Canada, and Australia) because of economic and political reasons connected with the consequences of WWII, the 1946–1949 civil war, and the 1967–1974 period of the military junta (Kalogeropoulou 2015). More than 120,000 Greek immigrants arrived in Canada, the majority of them coming from rural or semirural areas from various dialectal regions of Greece (Georgogiannis 1996). The two main Canadian destinations for Greek immigrants were Ontario (50%) and Quebec (34%), while most newcomers (84%) chose to live in large cities (population of more than 500,000). The two sexes were almost equally represented, 53% male and 47% female.

The communities established by Greek immigrants have thrived over the past decades. While still mainly concentrated in the major urban centers of Canada (Toronto, Montreal, Calgary, Vancouver), there are Greek speakers in most areas of the country, even in small communities. Gavaki (1974) maintains that, from the beginning, Greeks in Canada tried to balance their efforts at integration in their new country and maintaining their Greek identity as many of them were hopeful that they would return to Greece within a decade. For this reason, they established churches, language schools, and many secular associations, where Greek is used to a larger or lesser degree. Of course, they also run Greek-related businesses, such as restaurants and grocery stores, where members of the community also congregate and socialize. As the original immigrant population entered retirement age, Greek-oriented nursing homes and retirement communities, or benevolent associations, have also been created. The population used to be served extensively by Greek language media (mostly newspapers, but also radio and, to a lesser extent, TV), but the prevalence of these has waned in the last two decades since the internet has made access to media sources from Greece more easily accessible. One interesting aspect of Canadian Greek media is that it features more dialectal features than mainstream media sources from Greece.

For various reasons (cf. Daskalaki and Pappas, forthcoming), most of the Greek immigrant families stayed in Canada, and for the original immigrants and their descendants, integration became the main goal. Participants of the ImmiGrec project speak proudly of succeeding in mainstream fields, such as medicine, the law, corporate business, and provincial or federal politics. This means more contact with the official Canadian languages, English and French, but mainly with English, even at home, as teenage children became dominant in that language.1

Today, the Greek of first-generation immigrants, that is, CG, can be defined as an immigrant language since its speakers were exposed to one of the dominant Canadian languages at some point in their adulthood. In the literature, immigrant languages are identified as those which are spoken in immigrant communities, the members of which have arrived recently in a country and keep their own native tongue as the principal language for communicative purposes (see, among others, Myers-Scotton 2006; Appel and Muysken 1987; Montrul 2008; Baran 2017).

As shown in the following sections, CG has been affected in many ways by English, resulting in the appearance of contact phenomena, which are mostly lexical in nature, and constitute the most frequent borrowing cases, as suggested in a number of works (see, among others, Thomason 2001; Matras 2009). In fact, as shown in Section 5.4, many English loans are morphologically integrated into CG, while very few remain unaltered. Interestingly, another case of language contact has also taken place in the frame of CG, due to the coexistence and daily interaction of immigrants speaking different MG varieties. In earlier studies, we have shown that Greek immigrants arriving in Canada in the period after WWII brought with them a vernacular Common Modern Greek, spoken by the majority of the Greek population in the first half of the 20th century, and features of their native MG dialectal variety (Tsolakidis et al. 2019).

To sum up, CG of first-generation immigrants seems to be the product of various language contact processes between: (1) MG in general and the dominant language, that is, English, (2) different MG dialects, and (3) vernacular varieties and Standard Modern Greek (SMG). Thus, it can be considered a new variety confined within the limits of Greek-Canadian Communities.

3. Materials and Methods

Our dataset is constructed from a corpus of interviews of 453 speakers, which were conducted as part of the interdisciplinary project ImmiGrec. The objective of this project was to record the oral history and document the second wave of Greek immigration to Canada (ca. 1945–1975), cataloging the experiences of first-generation Greek immigrants who live in 12 different communities in Canada. The participants were interviewed on the basis of a common questionnaire by four different research teams, which comprised linguists, historians, Ph.D. and M.A. students but also members of the local communities who were hired as research assistants. The interviews were between 30 and 90 min long and were structured around three central themes: Greek origins, settlement in Canada, and integration into Canadian society. All interview procedures followed the rules and regulations concerning personal data protection of the Ethics Committees of the four Universities which participated in the ImmiGrec project: McGill, York, Simon Fraser, and the University of Patras.

Most of the participants were born between 1940 and 1955 and moved to Canada between 1955 and 1970. The main reason for immigration was to escape poverty or to improve their standard of living. The political persecution of leftist sympathizers also played a very important role, especially in the early 1950s in Greece, right after the Greek civil war, and in the late 1960s, after the right-wing dictatorship of 1967. Family connections to people who had already established themselves in Canada, as well as arranged marriages to someone who had immigrated to Canada or to someone whose parents had moved to Canada in the previous wave of immigration, were other reasons mentioned in the interviews. The ratio of men to women is roughly 2 to 1 (271 men to 174 women), while 8 participants did not want their sex/gender to be listed. A large number of participants come from the Greek areas of the Peloponnese, the Ionian islands, and Crete. Interestingly, very few of them were residents of Athens and Thessaloniki, the capital and second-largest Greek city, respectively, although these cities did play a part in our participants’ journey from rural Greece to Canada. Typically, the average speaker had primary education, and most immigrants had very little knowledge about Canada or its official languages upon arrival. However, it was not uncommon for them, especially for those who arrived in the late 1960s, to attend language schools as part of the integration process.2

An important sociolinguistic observation is that all of these speakers left Greece before the prestigious Athenian-based SMG became universally accepted and adopted by the educational system in late 70’s (Frangoudaki 1992). Thus, a fair assessment of the sociolinguistic profile of the participants is that there is a great deal of variation in their speech according to several dimensions: The linguistic variety they grew up with,3 exposure to urban varieties (not yet fully standardized themselves), gender, and knowledge of the dominant languages.

In the next sections, we first present some background information on SMG and its predecessor CMG, which formed the basis for the new variety developed among Canadian Greeks (CG). Then, we examine specific phenomena that show the influence of both the dominant language, English, as well as the interaction between the different varieties of Greek spoken in these communities.

4. Common Modern Greek as Input for Canadian Greek

The phase of the Greek language that is called Modern Greek was shaped as a vernacular perhaps as early as the 13th c. AD and definitely by the early 1500s (Browning 1983; Horrocks 2010; Holton et al. 2019). As most vernaculars, it was characterized by a great amount of regional variation, which, however, did not become resolved in the form of a Standard language, SMG, until 1980 (Frangoudaki 1992). The reason for this delay was the dispute between, on the one hand, supporters of an atticizing form of the language that had been cultivated by the Orthodox church for centuries and, on the other, supporters of various vernacular forms. These two positions drove each other to extreme stances over the 19th and 20th centuries, and the final compromise was not achieved until several major political and social issues dividing the country were settled when Greece was accepted into the European Union (European Economic Community at the time) in 1979.

Despite the lack of a stable language policy, once Greece’s borders were established after World War I (WWI) and the failed Asia Minor campaign (1919–1922), the ensuing period of urbanization did lead to the development of a vernacular koiné which has been referred to as Common Modern Greek (Triantaphyllidis 1936). This variety was spoken by a big part of the Greek population during the first half of the 20th century, whereas after WWII, the concerted effort to raise the education standards across the country increased the influence of SMG. CMG varied from region to region, where the spoken dialects were increasingly becoming assimilated to it (Mirambel 1937; Horrocks 2010; Tzitzilis 2016) and was characterized by the parallel use of several phonological, morphological and syntactic features. According to Caratzas (1958), CMG in the 1950s was based on the colloquial Greek spoken mainly in Athens. It included dialectal features highlighting the competition between different forms from various dialects, which were introduced in the period when Athens became the dominant urban center of Greece.

In general, it could be argued that CMG developed as a koiné among internal immigrants through contact between the regional dialects4 or the koinés formed around them. These internal immigrants came into contact in the urban and industrialized centers to which they moved for work, primarily in the three largest Greek cities, Athens, Thessaloniki, and Patras. Once established, they reinforced their common features, favoring those of the variety of Athens as the largest and most economically successful community. Thus, these features gradually became part of the new vernacular, CMG, superseding features of the contributing varieties (cf. Siegel 1993). Nevertheless, CMG was characterized by a large degree of variation. In our data, the Greek language spoken by our participants shows this variation, as it is influenced by three factors: Learned forms, Athenian forms, and rural ones. These are illustrated in the following tables.

In

Table 1. Examples of learned form variation in the speech of Greek-Canadians.

	Learned		Vernacular CMG		Translation
a.	καύση	[ˈkafsi]	κάψη	[ˈkapsi]	‘burning vs. heat, fever’
b.	κυβέρνησις	[ciˈvernisis]	κυβέρνηση	[ciˈvernisi]	‘government.sg.nom’
c.	κυβερνήσεως	[civerˈniseos]	κυβέρνησης	[ciˈvernisis]	‘government.sg.gen’
d.	φοιτηταί	[fitiˈte]	φοιτητές	[fitiˈtes]	‘student.pl.nom’
e.	φοιτητάς	[fitiˈtas]	φοιτητές	[fitiˈtes]	‘student.pl.acc’
f.	διετήρησε	[ðieˈtirise]	διατήρησε	[ðiaˈtirise]	‘maintain.pst’
g.	αρχικώς	[arçiˈkos]	αρχικά	[arçiˈka]	‘initially’

, the first example shows an etymological doublet, where the learned form has a more formal interpretation (e.g., ‘cremation’) than the vernacular one. Examples (b) and (c) showcase the variation in the inflectional endings of nominative and genitive singular, respectively, for feminine nouns of inflection class IC4 with stems ending in /s/ (cf. Ralli 2000, [2005] 2022). The next two rows (d and e) do the same for the variation in the nominative and accusative plural endings of IC2 masculine nouns (cf. Krimbas 2018). The variation in the use of an internal unstressed augment (cf. Krimbas 2018) is exemplified in (f). This variation also leads to hypercorrection: We find forms such as προενέφερα [proeˈnefera] for SMG προανέφερα [proaˈnefera] both in learned and vernacular registers. Example (g) shows the variation between the learned and vernacular forms of the adverbial ending.

There are also two types of usage of the relative pronoun οποίο [oˈpiο] ‘which’, preceded by the definite article το [tο] (so literally ‘the which’), instead of που [pu] ‘which’. In the first type, the relative pronoun does not agree in gender, number, and case with the noun that it modifies but appears in the neuter, singular nominative form as in example (1). In the second case, the same form of the relative pronoun replaces [pu] in its use as a discourse particle to introduce a parenthetical (ex. 2). Both of these usages are influenced by the status of the relative pronoun as a more formal way to introduce a relative clause than [pu]. This development could be a type of hypercorrection or a structural reanalysis, but there are not enough data points to determine this question.5

• (1) Τους λέω κουβέντες το οποίο δεν τις ξέρει ούτε η κόρη μου.
  [tus ˈleo kuˈvendes to oˈpio ðen tis ˈkseri ˈute i ˈkori mu]
  ‘I tell them words that (‘(lit.) which’) even my daughter doesn’t know’.
• (2) Μικρά παιδιά που ήμασταν, ήτανε η προίκα, το οποίο εγώ δεν πίστευα σε προίκα.
  [miˈkra peˈðja pu ˈimastan, ˈitane i ˈprika, to oˈpio eˈɣo ðen ˈpisteva se ˈprika]
  ‘When we were young boys, there was the custom of dowry. That was something that
  (‘(lit.) which’) I did not believe in’.

In

Table 2. Examples of Athenian features in the speech of Greek-Canadians.

	Athenian		SMG		Translation
a.	πατεράδες	[pateˈraðes]	πατέρες	[paˈteres]	‘fathers’
b.	μανάδες	[maˈnaðes]	μάνες	[ˈmanes]	‘mothers’
c.	έχομε	[ˈexome]	έχουμε	[ˈexume]	‘we have’
d.	κοιμόνταν	[ciˈmodan]	κοιμόντουσαν	[ciˈmodusan]	‘sleep.3pl.ipfv.pst’

we provide a list of Athenian features which appear in the data. These are the use of plural forms of IC2 (ex. a) and IC3 (ex. b) nouns with an epenthetic /ð/, the formation of 1st plural endings of the present tense in /-ome/ instead of /-ume/ (ex. c), and variation between the /-odusan/ and /-odan(e)/ forms for the 3rd plural ending of the imperfective past, in the mediopassive voice (ex. d).

More (vernacular) CMG forms are shown in

Table 3. Examples of (vernacular) CMG features in the speech of Greek-Canadians.

	Vernacular CMG		SMG		Translation
a.	αχελώνα	[açiˈlona]	χελώνα	[çeˈlona]	‘turtle’
b.	απαρατήσει	[aparaˈtisi]	παρατήσει	[paraˈtisi]	‘leave.3sg.prf’
c.	λεφτά	[leˈfta]	λεπτά	[leˈpta]	‘minutes’
d.	πλερώνεις	[pleˈronis]	πληρώνεις	[pliˈronis]	‘pay.2sg.prs’
e.	αυτόνα	[aˈftona]	αυτόν	[aˈfton]	‘he.sg.acc’
f.	εμένανε	[eˈmenane]	εμένα	[eˈmena]	‘I.acc’
g.	μαγαζά	[maγaˈza]	μαγαζιά	[maγaˈzja]	‘shops’
h.	μάγειρου	[ˈmaʝiru]	μάγειρα	[ˈmaʝira]	‘cook.sg.gen’
i.	μαγείροι	[maˈʝiri]	μάγειρες	[ˈmaʝires]	‘cook.pl.nom/acc’
k.	άρχιξε	[ˈarçikse]	άρχισε	[ˈarçise]	‘begin.3sg.prf.pst’
l.	χάλαγα	[ˈχalaγa]	χαλούσα	[xaˈlusa]	‘destroy.1sg.ipfv.pst’
m.	εδούλευα	[eˈðuleva]	δούλευα	[ˈðuleva]	‘work.1sg. ipfv.pst’
n.	ήμαν	[ˈiman]	ήμουν	[ˈimun]	‘be.1sg. ipfv.pst’
o.	με δώσαν	[me ˈðosan]	μου δώσαν	[mu ˈðosan]	‘to-me give.3pl.prf.pst’

. We see the use of prothetic /a/ in both nouns (ex. a) and verbs (ex. b), continuant + stop instead of stop + stop consonant clusters (ex. c), and the use of /e/ instead of /i/ in example (d). In examples (e) and (f), there are cases of the extension of the accusative forms of pronouns and in (g) an example of palatal deletion in a consonant cluster. In (h) and (i), the change from IC2 to IC1 for masculine nouns in /as/ is showcased in both the genitive singular and the nominative/accusative plural forms. In (k), we see the substitution of /ks/ for /s/ in the perfective stem, while (l) is an example of the use of /ɣ/ to form the stem of the imperfective past of IC2 verbs. Example (m) shows the retention of the unstressed syllabic augment and (n) an alternate form of the 1st person singular of the imperfective past of ‘be’. Finally, in (o), we report the use of the accusative case of the personal pronoun, με [me], as an indirect object, instead of the genitive μου [mu] in the phrase ‘they gave me’.

5. Contact Phenomena

In this section, we discuss four particular phenomena that highlight the effect of contact among the different varieties of Greek spoken in Canada as well as the effect of contact between Greek and English, the dominant language in the Greek-Canadian communities. These phenomena are: (a) The use of the unstressed syllabic augment, (b) the variation between two different imperfective past suffixes for IC2 verbs, (c) the variation in the 3rd person plural of the imperfective past, (d) the integration of English loan nouns.

5.1. The Unstressed Syllabic Augment

The verbal augment in Classical Greek, either syllabic or temporal, was an inflectional prefix attached to the verbal stem and marked the past. The syllabic augment was an ε- /e/ prefix. For instance, Ancient Greek τοξεύω [toˈkseuoː] (present6) ‘I shoot with the bow’ had the imperfective past form ἐτόξευον [eˈtokseuon]. The temporal augment appeared with verbs beginning with a vowel, taking the form of a lengthening of the initial vowel (Chantraine 1961), as in the Ancient Greek example ἐθέλω [eˈtʰeloː] ‘I want’, whose perfective past form was ἠθέλησα [ɛːˈtʰelɛːsa].

Both of these augments were affected by changes over time. The use of the temporal augment had already declined in Hellenistic Greek (ca. 3rd c. BC—3rd c. AD) although some forms with /i/ were preserved in Medieval and Modern Greek because of their high frequency, e.g., ήθελα [ˈiθela] ‘I wanted’.7 Other forms arose through various analogical change schemata in Medieval Greek (Horrocks 2010), as is the case with ξέρω [ˈksero] (present) ‘I know’ whose perfective past form is ήξερα [ˈiksera]. It is also found in some MG dialects (MGD). For example, in Lesbian Greek, the imperfective past form of λέγου [ˈleγu], ‘I say’ is ήλιγα [ˈiliγa] (Kretschmer 1905), whereas SMG has έλεγα [ˈeleγa]. Lastly, the temporal augment is kept in SMG learned verbal forms which have been directly inherited from Ancient Greek, for example, ήλεγχα [ˈileŋxa] is the imperfective past of ελέγχω [eˈleŋxo] ‘I check’.

In SMG, the syllabic augment is still /e/, but its presence depends on whether or not it is stressed. This is the result of a more general process in which word-initial unstressed vowels were deleted (Horrocks 2010). Since SMG also has a stress assignment rule (Kaisse 1982) that only allows the ultimate, penultimate, or antepenultimate syllables to be stressed, the syllabic augment is frequently unstressed and is, thus, deleted even in the paradigm of the same verb. For example, παίζω [ˈpezo] ‘I play’, has the 1st person singular in the imperfective past έ-παιζα [ˈe-peza] ‘I was playing’, but the 1st person plural παίζαμε [ˈpezame] ‘we were playing’. In the latter, the syllabic augment is deleted because being the fourth syllable from the end, it cannot be stressed.

This process led to the loss of the clear morphemic status that the syllabic augment had in previous stages of Greek. In SMG, the syllabic augment is considered to be a morphophonological formative whose only function is to receive stress when it shifts outside the confines of the word (Babiniotis 1972). Ralli (1988) argued that this stress shift is due to the phonological properties of past tense inflectional endings.

At the same time, the SMG pattern of unstressed augment deletion does not appear in several MG dialects (Triantaphyllidis 1936), where the syllabic augment is obligatorily present. This is true of the dialects of Chios, Cyclades, Cyprus, Pontus, Smyrna, but also in some parts of Crete, the Peloponnese and the Ionian islands (Contosopoulos 1994; Tzitzilis 2016). For illustration, see the examples in

Table 4. Examples of augment retention in Modern Greek dialects.

Dialect	Present		Past8		Meaning
Cypriot	γράφoμεν	[ˈɣrafomen]	εγράφαμεν	[eˈɣrafamen]	‘write.1pl’
Chiot	παίρω	[ˈpero]	επήρα	[eˈpira]	‘take.1sg’
Ionian	κάθομαι	[ˈkaθome]	εκαθόμουνα	[ekaˈθomuna]	‘sit.1sg’
Peloponnesian	φιλώ	[fiˈlo]	εφίληγα	[eˈfiliɣa]	‘kiss.1sg’

:

For the varieties of these areas, Ralli ([2005] 2022) has argued that the syllabic augment still keeps its morphological status and that it is the first part of a discontinuous morpheme, the second member of which is the past tense ending.

In Tsolakidis et al. (2019), we examined the augment usage pattern of 131 speakers in our sample who grew up in several major dialect regions of Greece: Macedonia and the North Aegean islands, Thessaly, Central Greece (Sterea Ellada), the Peloponnese, the Ionian Islands, Crete, and the Southwest Aegean (Cycladic) islands. The constructed dataset of 2168 tokens revealed a general decline of the unstressed augment.9 The overall pattern of use for speakers based on their Greek region of origin is presented in

Table 5. Distribution of the deletion of the unstressed syllabic augment among CG speakers by region of origin (N = 2168).

Region	Expected	Deletions (N)	Deletions (%)	Total N
Athens	Deletion	152	87%	175
Central Greece	Deletion	35	45%	77
Macedonia	Deletion	50	82%	61
Ionian Islands	Augment	151	51%	296
North Aegean	Deletion	420	91%	462
Peloponnese	Both	297	65%	457
Thessaly	Deletion	97	51%	191
Western Greece	Deletion	173	81%	214
SW Aegean	Augment	112	72%	156
Crete	Augment	44	56%	79

, where the second column indicates the expected pattern.

Our first observation is that there is no area whose speakers use the unstressed syllabic augment categorically (above 80%). At most, there is variable usage, even for speakers from the Southern dialect region (cf. Crete). Overall, the results indicate that speakers are in the process of general convergence with respect to this dialectal feature. It is not simply the case that non-standard speakers are converging to the standard, but even speakers from standard areas (e.g., Athens or Macedonia) are using the non-standard variant to some extent. It is important to point out that there is variable usage of the unstressed syllabic augment, even for Greek-Canadians coming from regions whose dialect displays it obligatorily (cf. the Ionian islands). Moreover, it is also used by speakers from areas where its compulsory use is unknown, for instance, by those originating from the Aegean island of Lesbos. As noted by Bubenik (1993) and Trudgill (2008), the parallel use of morphological features is a common characteristic of koinés. Thus, variability in the usage of the syllabic augment can be considered a typical feature of this new Greek koiné, CG. Below, we present some representative CG utterances taken from the ImmiGrec project, where this variation is shown even within the same sentence (tokens in bold). In parentheses, we give the initials of the participants’ names, as well as the places of origin and settlement in Canada.10

• (3) Ionian (A.D., Cephalonia-Vancouver)
  Εδούλευα στο φέρυ στην Κεφαλλονιά. Πηγαίναμε Πάτρα-Κεφαλλονιά-Πάτρα. Στο Άγιος Γεράσιμος δούλευα.
  [eˈðuleva sto ˈferi, stin cefaloˈɲa. piˈʝename ˈpatra-cefaloˈɲa-ˈpatra. sto ˈaʝios ʝeˈrasimos ˈðuleva]
  ‘I was working on the ferry, on Cephalonia. We were going Patra-Cephalonia-Patra. On the St. Gerasimus I was working’.
• (4) Peloponnesian (H.K., Arcadia-Edmonton)
  Aπλώς έκανα μια επιβεβαίωση όταν πήγα εκεί, και επήγα.
  [aˈplos ˈekana mɲa epiveˈveosi ˈotan ˈpiγa eˈci, ce eˈpiγa]
  ‘I just confirmed it when I went there. So, I went.’
• (5) Athenian (I.K., Athens-Edmonton)
  Δούλευε, δούλευε. Πάρα πολύ εδούλευε.
  [ˈðuleve,ˈðuleve. ˈpara poˈli eˈðuleve]
  ‘He worked, he worked. He worked very much!’
• (6) Northern Greek (N.P. Karditsa-Vancouver)
  Έφυγα όταν ήμουνα, εφύγαμε δύο φορές, όταν φύγαμε, όταν ήμουνα τριών χρονών πήγαμε στο χωριό του πατέρα μου.
  [ˈefiγa ˈotan ˈimuna, eˈfiγame ˈðio foˈres, ˈotan ˈfiγame, ˈotan ˈimuna triˈon xroˈnon ˈpiγame sto xoˈrʝo tu paˈtera mu]
  ‘I left when I… we left twice, when we left, when I was 3 years old, we went to my father’s village.’

Our data clearly show that, in the speech of the participants, a great tendency is for them to employ features that belong either to their native dialectal variety of MG, or to a vernacular form of CMG (mentioned in Section 4), characterized by extensive variation. A typical characteristic of this variation is the parallel use of various verbal forms, such as the one we see in the case of the unstressed syllabic augment. Taking into account that all of our participants were born after 1920, we could reasonably assume that most of those people, who had lived in a multidialectal context before leaving Greece, used a form of CMG before their immigration to Canada. Thus, we suggest that they brought this vernacular form to Canada and kept using it after their settlement there.

We see that the dialectal pattern of the presence of the unstressed syllabic augment persists, as no speaker coming from an area with compulsory deletion of the augment entirely avoids it. For example, speakers from the Aegean island of Lesbos occasionally use an unstressed syllabic augment, although it is unknown not only in the Lesbian MG variety but also in the CMG used in Lesbos (Kretschmer 1905).

The extension of the variable pattern to dialectal groups could be attributed to the fact that the Peloponnesian MG variety, being close to SMG, and thus enjoying elevated prestige, displays a variable pattern itself, namely both the presence and the absence of unstressed syllabic augment.11 Since in many Greek communities in Canada, the majority of the Greeks are of Peloponnesian origin, we hypothesize that immigrants from other Greek areas started adopting the Peloponnesian pattern of speech variation. According to Trudgill (2001), members of small, stable, and tightly-knit communities tend to share more linguistic information among themselves in contrast to members of larger, more dynamic, and loosely-knit communities. In the case of our participants, it is likely that Greek immigrants coming from different dialectal regions adopted more variable patterns in order to downplay their differences and give prominence to common bonds. In this context, the parallel use of different verbal forms (with and without unstressed syllabic augment) even by the same speaker and/or even in the same sequence of sentences became, and is still, quite acceptable. There is a similar pattern with respect to the variable use of the active imperfective past suffixes, as we will see in the following section.

5.2. Active Imperfective Past Stem Suffixes for IC2 Verbs

In SMG, verbs are classified into two inflection classes (ICs), cf. Ralli (1988, [2005] 2022). In IC2 verbs, the first singular form of the present tense ends in stressed /o/, such as in the verbs μιλώ [miˈlo], ‘I speak’ and φιλώ [fiˈlo], ‘I kiss’. In Early Medieval Greek, the imperfective past of the active voice of IC2 verbs developed two different stem markers, -γ- /ɣ/ and -ουσ- /us/ (Holton et al. 2019).

The most likely scenario for the development of /ɣ/ is that it developed epenthetically as a hiatus-braker [j] between the stem vowel- ending form of the verbs and the past tense suffix /e/, in the 3rd person singular (for details, see among others, Joseph and Ralli 2019). For example, εμίλαε [eˈmila-e] ‘(s)he was speaking’ became [eˈmila-j-e]. The glide [j] developed into the fricative consonant [ʝ] and spread to the other forms of the imperfective past paradigm (e.g., μίλαγα [ˈmila-ɣ-a] ‘I was speaking’, μιλάγαμε [miˈla-ɣ-ame] ‘we were speaking’, etc.). Ιt was then reinterpreted and recategorized as the /ɣ/ suffix, marking the imperfective aspect for the past tense and filling a slot for the aspectual marking in the structure of Greek IC2 verbs. Note that the /ɣ/ imperfective past forms were originally used in the Peloponnese, the Cycladic islands and the Sporades but also occurred in Old Athenian, and some northern varieties of Central Greece (Sterea Ellada) and Thessaly (Contosopoulos 1994).

The second form, /us/, appears to have developed first in the 3rd person plural of the imperfective past, where the original ending /un/ (e.g., εμίλουν [eˈmilun] ‘they were talking’) was reanalyzed as /u/ + /n/ (Horrocks 2010). Then, the perfective past ending /san/ replaced the /n/ by analogical extension, and from there, another reanalysis yielded /us + an/, and the entire paradigm shifted to /us/ plus the imperfective past endings (e.g., μιλούσα [miˈlusa] ‘I was speaking’, μιλούσες [miˈluses] ‘you were speaking’, μιλούσαμε [miˈlusame] ‘we were speaking’, etc.). As in the case of /ɣ/, this change was completed in Early Medieval Greek and appears to be an innovation of the Northern Greek area, especially around the center of the Byzantine Empire, Constantinople (Istanbul).12 Horrocks (2010) suggests that the /us/ form became the standard after the influx of Ottoman Empire Greek speakers to the newly founded Greek state in the middle and late 19th century when Greece was still a very small country.

Even before the establishment of SMG, Newton (1972b) characterized the /ɣ/ variant as less formal than /us/, and it was considered to be non-standard till recently (cf. Iordanidou 1994; Setatos 1992). Today, in some areas (e.g., Central Thessaly), the /ɣ/ variants compete with those in /us/, the latter being typical of the northern varieties of Greece, that is, of those north of Central Thessaly, as noted by Chatzidakis (1905–1907), Papadopoulos (1926), and Pantelidis (2003).

In Ralli et al. (2019), we examined this pattern of variation in a sample of 116 speakers. Our dataset is constructed from 1989 tokens with the imperfective past active endings. Of these, 485 forms of μπορώ [boˈro] (the basic verb for denoting ability and permission in MG) were excluded because of their categorical pattern (95% /us/). For the remainder, the 1504 tokens were equally divided between /ɣ/ (N = 719, 48%) and /us/ (N = 785, 52%). The overall pattern shows that the variant /us/ is more widely used than expected, based on the region of origin of our participants, as seen in

Table 6. Distribution of the /us/ variant among CG speakers by region of origin (N = 1504).

Region	Expected	/us/ (N)	/us/ (%)	Total N
Athens	both	98	56%	175
Ionian islands	/ɣ/	136	46%	296
Lesbos	/us/	31	94%	33
Macedonia	/us/	39	64%	61
Peloponnese	/ɣ/	223	44%	506
Skopelos	/us/	32	84%	38
Thessaly	/ɣ/	118	77%	153
Western Greece	/ɣ/	52	42%	123
Central Greece	/ɣ/	32	54%	59
Skyros	/ɣ/	23	38%	60

.

We illustrate the usage of these forms with the following examples drawn from the recordings of the ImmiGrec project. Again, we note the initials of the participants’ names as well as the places of origin and settlement in Canada.

• (7) Ionian (A. T., Zante-Montreal)
  Συνενογιόμουνα μια χαρά. Τότε μιλούσανε πάρα πολύ Ελληνικά.
  [sinenoˈʝomuna ˈmɲa xaˈra. ˈtote miˈlusane ˈpara poˈli eliniˈka]
  ‘I was communicating just fine. At that time, they were speaking Greek a lot.’
• (8) Peloponnesian (I. H., Messinia-Toronto)
  Και έτσι βοηθούσες τη μάνα σου και τον πατέρα σου.
  [ce ˈetsi voiˈθuses ti ˈmana su ce ton paˈtera su]
  ‘And this way you were helping your mother and your father.’
• (9) Peloponnesian (I. H., Messinia-Toronto)
  Θυμάμαι χτυπάγαν οι καμπάνες στο χωριό.
  [θiˈmame xtiˈpaɣan i kaˈbanes sto xorʝo]
  ‘I remember the church bells chiming in the village.’
• (10) Athens (E. O-H., Athens-Toronto)
  Όποτε προσπαθούσανε δεν μπορούσανε να τον βρούνε.
  [ˈopote prospaˈθusane, ˈðen boˈrusane na ton ˈvrune]
  ‘Whenever they were trying to, they couldn’t find him.’
• (11) Peloponnesian (S. K., Laconia-Vancouver)
  Όταν ζούσαν οι γονείς μου, πήγαινα.
  [ˈotan ˈzusan i ɣoˈnis mu, ˈpiʝena]
  ‘As long as my parents were alive, I used to go (visit them in Greece).’

The data in Table 6 show a small but clear reduction in the usage of /ɣ/ by speakers who immigrated to Canada from regions where /ɣ/ is in use. The most unexpected result is the high preference for /us/ among speakers from Thessaly (77%), where /ɣ/ is the expected active imperfective past marker. For all the other /ɣ/ areas, the Peloponnese, Central and Western Greece, the Ionian islands and the island of Skyros (in the central Aegean), we see that there is an almost equal distribution between the two variants. Note that we did expect the almost equal distribution of the two suffixes in immigrants coming from Athens since it has been the meeting place for speakers of different dialectal backgrounds ever since Athens became the capital of the modern Greek state in 1834 and began attracting educated speakers from northern Greece and Constantinople (Istanbul) in particular. These speakers introduced /us/ as a prestige variant in a community of mostly /ɣ/ users from the Peloponnese and Central and Western Greece (cf. Horrocks 2010). We also see that there is some usage of /ɣ/ among speakers from Macedonia (36%), where no /ɣ/ was expected at all. In Macedonia, we expected a pattern more like the categorical ones seen in Lesbos (94%) and Skopelos (84%).

In an attempt to interpret these patterns of variation, that is, the increased usage of /us/ among the CG speakers who came from areas where /ɣ/ was prevalent, we suggest that the spread of the standard /us/ variant may have occurred in CMG before our speakers immigrated to Canada. Our reasoning is that since the majority of Greeks who immigrated to Canada are of Peloponnesian origin, it is unlikely that the adoption of the standard /us/ variant occurred after the move there, although some small degree of influence from the school system or media exposure (newspapers and radio) is also possible. On the other hand, the presence of the non-standard /ɣ/ variant in the speech of immigrants from Macedonia could be the result of contact with so many speakers who use that variant to a great extent.

5.3. Mediopassive 3rd Person Plural Forms of the Imperfective Past

The third feature we examine is the formation of the 3rd plural ending of the mediopassive voice in the imperfective past paradigm. It is well established that there are two competing variants in CMG: (a) The form /odusan/ (e.g., κοιμόντουσαν [ciˈmodusan] ‘they were sleeping’), which nowadays is mainly associated with Athenian,13 Peloponnesian Greek, and with some varieties of Central Greece and (b) the forms /odan/ or /otan/ (e.g., κοιμόνταν [ciˈmodan] or κοιμόταν [ciˈmotan]) which are common in Northern Greece14 (Horrocks 2010). According to Holton et al. (2019), the Northern Greek forms are considered to be more formal and standard, but the Athenian form is more widely used. The same observation is also made by Newton (1972a), whose data show that /odusan/ was the dominant suffix in CMG spoken in Athens in the ‘70s by 15–35 years old speakers, and Pantelidis (1999).

In the same dataset that we used for the examination of the formation of the imperfective past stem in Section 5.2, we also looked at the formation of this grammatical category. As this is a less common construction, we only found 250 tokens, whose distribution is shown in

Table 7. Distribution of the 3rd plural mediopassive imperfective past by region of origin (N = 250).

Region	Expected	/odusan/ (N)	/odusan/ (%)	Total N
Athens	/odusan/	29	97%	30
Peloponnese	/odusan/	40	48%	84
Ionian	/odusan/	24	52%	46
Thessaly	/odan/ or /otan/	4	14%	29
Central Greece	/odusan/	22	73%	30
Macedonia	/odan/ or /otan/	7	54%	13
Lesbos	/odan/ or /otan/	1	13%	8
Skyros	/odan/ or /otan/	3	30%	10

.

The pattern observed reflects the one we would expect, with a couple of important exceptions. First, in Macedonia, there is a high rate of usage of the variant /odusan/, but the overall number of tokens is small, so we are not certain how representative this is. Second, in the Peloponnese, the rate of usage for /odusan/ is only 48%, not because of competition from /odan/, but because other, more regional, forms are used, such as /osan/, /osade/, and /odusane/.15 Indicative examples drawn from the recordings of the ImmiGrec project are listed below.

• (12) Peloponnesian (A. P., Achaea-Calgary)
  Επικοινωνούσαμε με πολλούς από το Βανκούβερ, γιατί γινόσανται τα τούρναμεντς σε διάφορες πολιτείες.
  [epicinoˈnusame me poˈlus aˈpo to vanˈkuver, ʝaˈti ʝiˈnosade ta turnaments se ðiˈafores poliˈties]
  ‘We were in contact with a lot of people from Vancouver since the tournaments were taking place in various cities.’
• (13) Peloponnesian (H. K., Arcadia-Edmonton)
  Όλα γινόσαν από μάς. Όλα ό,τι χρειαζόταν τους το προσφέραμε.
  [ˈola ʝiˈnosan apo ˈmas. ˈola ˈoti xriaˈzotan tus to proˈsferame]
  ‘We did everything. We offered them whatever they needed.’
• (14) Peloponnesian (N. K., Corinth-Montreal)
  Έπρεπε νά΄χουνε προίκα για να παντρευτούνε. Δεν παντρευόσανε έτσι.
  [ˈeprepe ˈna xune ˈprika ʝa na pandreˈftune. ˈðen pandreˈvosane ˈetsi]
  ‘They should afford a dowry in order to get married. They were not getting married so simply (= without providing a dowry).’
• (15) Northern Sporades (P. K., Skopelos-Vancouver)
  Δίναμε πληροφορίες για τον πρόεδρο του Δήμου, για τη μετανάστευση που ερχόσαν στον Καναδά.
  [ˈðiname plirofoˈries ʝa ton ˈproeðro tu ˈðimu, ʝa ti metaˈnastefsi pu erˈxosan ston kanaˈða]
  ‘We were providing information about the president of the municipality, the immigration coming to Canada.’
• (16) Central Greece (Th. S., Meganisi, Lefkada-Vancouver)
  Τα βρήκε λίγο δύσκολα. Τα θυμόσανται και στην Ελλάδα.
  [ta ˈvrice ˈliɣoˈðiskola. ta θiˈmosade ce stin ˈelaða]
  ‘At first, he had some hard times. They remembered this, even in Greece.’
• (17) Peloponnesian (S. R., Arkadia-Edmonton)
  Ερχόντουσανε από την Ευρώπη εκείνη την εποχή.
  [erˈxodusane aˈpo tin eˈvropi eˈcini tin epoˈçi]
  ‘They were coming from Europe in that period.’

These forms, although currently stigmatized in Modern Greek, play an important role in our dataset because of the large number of speakers from the Peloponnese. As far as /odusane/ in (17) is concerned, it always appears in the speech of immigrants coming from Kastritochoria, a mountainous region in the Peloponnesian area of Arkadia, who have settled in the city of Regina (see St. Paul’s Greek Orthodox Community of Regina 2011). We could assume that /odusane/ is a characteristic of the dialectal speech proper to the Arcadian area and that the immigrants from Kastritochoria brought it to their new homeland and kept using it in everyday communication in the context of a relatively homogeneous and tightly-knit Greek-speaking community.

With respect to /osade/,16 it does not appear only in the speech of the Peloponnesian immigrants but also in the data recorded from an immigrant coming from Meganisi, a small island in the vicinity of the Ionian island of Lefkada (cf. 16).17 We also found a token of /osade/ (ex. 15) in the speech of P.K., a Greek born in Vancouver to a family of fishermen who migrated from the Aegean island of Skopelos and settled on Deas Island at the turn of the 20th century (Hellenic Community of Vancouver 1977). This exceptional use of /osade/ is indicative of the influence that Peloponnesian speakers exert on the Greek speech of non-Peloponnesian immigrants.

Overall, in all three grammatical features that we examined in this section, we have noticed the same type of effect of dialect contact. First, there is a very strong tendency to use the variant that has become the accepted one in SMG, which demonstrates that much of this process of contact between varieties had already begun in Greece before our participants immigrated. Secondly, we see that the preponderance of Peloponnesians in these communities has led speakers from other areas to adopt Peloponnesian variants, even when these have always been considered non-standard.

In the next section, we examine the effects of contact between Greek and Canadian English on the vocabulary of Greek-Canadians.

5.4. Contact with English: Loan Word Integration in Canadian Greek

In lexical borrowing, it is important to determine the type of transferred material and the degree of integration of loan words. For instance, it is usually admitted that verbs are more difficult, although not impossible, to borrow than nouns, because of the rich information they carry, for example, argument structure (see, among others, Matras 2009). Moreover, there are loans that remain unaltered, accepting only some phonological adjustments to the pronunciation of the recipient language, and loans that are fully or partly integrated.

Wohlgemuth (2009) has observed that words may be directly or indirectly borrowed or integrated into a frame using an auxiliary (if a verb). Direct borrowing means that there is very little modification of the source material in terms of phonology or inflection, while indirect borrowing involves the use of an integrating element, usually a derivational affix taken from the system of the receiving language.

In what follows, we extensively deal with English loan nouns in CG and their most typical characteristics. Note that English verbs are also borrowed in their infinitival form, undergo a phonological adjustment to the Greek pronunciation, and contrary to loan nouns that receive Greek inflection (see below), they maintain their invariant character. Greek verbal inflection marked for tense, person and number is carried by the Greek verb κάνω [kano] ‘to do’ which is used as an auxiliary. This verb always precedes the English loan and cannot be separated from it by any insertion. For illustration, see the periphrases κάνω φράι [ˈkano ˈfraɪ] ‘I fry’, κάνουμε γουέικ απ [ˈkanume ɣuˈeikap] ‘we wake up’ or έκανε γιουζ [ˈekane ˈʝuz] ‘(s)he used’ (see Maniakas 1991; Ralli 2016). According to Ralli (2016), the adoption of this strategy for integrating English verbs is triggered by the difference between Greek and English inflection because while Greek has a rich and overtly realized inflection, English is inflectionally poor. By keeping the invariant form of the verb and assigning inflection to the Greek auxiliary, the morphological properties of both languages are respected. A less frequent strategy in CG for integrating verbs is to use the derivational suffix /ar/, a loan itself from the Italo-Romance infinitival marker /ar(e)/ (e.g., Italian parl-are ‘to speak’). For Ralli (2016, 2021), /ar/ has changed status by allogenous exaptation, a cross-linguistic phenomenon that has been studied by Gardani (2016) in various contact settings. In MG, the inflectional status of /ar/ has been changed into derivational, and /ar/ is now specialized in creating loan verbs. The /ar/ strategy is very common in loan adaptation in MG, and its use in forms with /ar/ in CG (e.g., μουβάρω [muˈvaro] ‘I move’ from English move) is probably based on its use in CMG before they immigrated to Canada.

With respect to the borrowing of nouns, we find a rich system of accommodation. This is not surprising since, according to Thomason and Kaufman (1988), when a language has rich morphology, it tends to morphologically adapt loan words to its system (see also Melissaropoulou 2013 for contact phenomena specific to Greek). We find very few examples that are kept invariable in CG as far as their morphological form is concerned. One of them is κέτσοπ [ˈcetsəp] ‘ketchup’ and another is πάρτι [ˈparti] ‘party’. It may be that the first word was used mostly with English-speaking customers in restaurants, and so it was important to integrate it without any major changes. As for the second one, it already adheres to the phonotactic constraints of MG word endings (neuter nouns in -i, e.g., MG σπίτι [ˈspiti] ‘house’), and as a consequence, it remained invariable. At the same time though, we must note that the word cake is integrated as κέκι [ˈceci], that is, with the addition of an MG neuter noun ending.

In accordance with the native morphological nominal structures consisting of stems and inflectional suffixes, as proposed by Ralli ([2005] 2022), an adopted English noun in CG undergoes the following modifications: (a) A reanalysis as a stem, since MG word formation is stem-based, (b) allocation of a grammatical gender value and an inflectional marker responsible for classifying it to a specific IC. For pronunciation purposes, a slight phonological modification may also occur. For illustration, consider the examples in

Table 8. Loanwords in the speech of Greek-Canadians.

	Canadian-Greek		SMG		Source
a.	μπόσης	[ˈbosis]	αφεντικό	[afediˈko]	boss
b.	μπλόκος	[ˈblokos]	τετράγωνο	[teˈtraɣono]	(city) block
c.	ρολός	[roˈlos]	ψωμάκι	[psoˈmaki]	(dinner) roll
d.	μπάρα	[ˈbara]	μπαρ	[bar]	bar
e.	τζάρα	[ˈdzara]	βάζο	[ˈvazo]	jar
f.	κάρο	[ˈkaro]	αυτοκίνητο	[aftoˈcinito]	car
g.	μπόξι	[ˈboksi]	κουτί	[kuˈti]	box
h.	μπάσι	[ˈbasi]	λεωφορείο	[leofoˈrio]	bus
i.	φλόρι	[ˈflori]	όροφος	[ˈorofos]	floor
j.	μπασέρης	[baˈseris]	οδηγός	[oðiˈɣos]	bus driver
k.	μπαγκαδόρος	[baŋkaˈðoros]	τραπεζικός	[trapeziˈkos]	banker
l.	κρέντιτο	[ˈkredito]	πίστωση	[ˈpistosi]	credit
m.	κοκονότσι	[kokoˈnotsi]	καρύδα	[kaˈriða]	coconut
n.	μεσίνι	[meˈsini]	μηχανή	[mixaˈni]	machine
o.	μπάνκα	[ˈbaŋka]	τράπεζα	[ˈtrapeza]	bank
p.	μαρκέτα	[marˈceta]	αγορά	[aɣoˈra]	market
q.	σάινα	[ˈsaina]	πινακίδα	[pinaˈciða]	sign

, taken from Ralli and Makri (2020, p. 247), which display change into stem, addition of a vowel according to Greek stem-ending vowels, and Greek inflection.

One of the main points of difference between English and Greek that causes morphological material to be added is that, while English phonotactics permit almost any segment to appear word-finally, in MG only vowels and the consonants /n/ and /s/ are allowed. Thus, as Ralli and Makri (2020) observe, in many cases, the English word is adapted into Greek as a stem with the addition of a particular vowel. This stem then receives grammatical gender and the appropriate inflectional ending to be used as a word.18 For instance, car becomes [ˈkaro] with the addition of /o/ (ex. f) while box, bus, and floor become [ˈboksi], [ˈbasi] and [ˈflori] with the addition of /i/ (ex. g–i). Note that most neuter nouns receive a final /i/ since, as proposed by Christophidou (2003), this class constitutes the most productive one in SMG. In contrast, neuter nouns are built with /o/ (e.g., [ˈkaro]), when they correspond to a number of equivalent forms in SMG. Other loan words, such as bar and jar, become [ˈbara] and [ˈdzara] with the addition of /a/ (ex. d–e), while the endings /is/ and /os/ are attached to words such as boss, (city) block, and (dinner) roll, to produce [ˈbosis], [ˈblokos], and [roˈlos] (ex. a–c). Sometimes, a MG derivational affix is added to a loanword to allocate gender and a specific semantic value, for example, the agentive /eri/ is attached to bus to yield the meaning of ‘bus driver’ (cf. [baˈseris], ex. j).

With respect to the assignment of grammatical gender, an important determining factor is whether or not the noun denotes a human being. In accordance with a proposal put forward by Ralli (2002), the [+human] feature and biological sex regulate a specific value of the grammatical gender. Thus, male humans are masculine and female ones feminine, as shown in the difference between [ˈbosis] and [boˈsina] for male and female ‘boss’, respectively. We note that for some professions, masculine nouns, such as banker ([baŋkaˈðoros], ex. k) or bus driver ([baˈseris], ex. j), reflect the stereotypes of post-war Canada in terms of gender-appropriate jobs.

The majority of non–human loans are assigned neuter gender, which is the default gender value, when no other clear motivation exists or prevails (Ralli et al. 2015), as in examples (l)–(n). However, in some cases, the gender of the original Greek word appears to influence gender assignment in the loanword, as in examples (o)–(q).19 Lastly, homophony can also play a role in the assignment of gender. Ralli and Makri (2020) observe that some CG nouns are formed on the basis of existing homophonous nouns in MG. For example, the creation of [ˈkaro] for car (ex. f), must have been straightforward since SMG already had the same word for carriage. Overall, though, the processes by which gender is assigned to a borrowed stem and by which the inflectional suffix set is chosen appear to be determined by several factors, phonological, morphological, semantic, and pragmatic.

Our findings confirm the conclusion by Ralli and Makri (2020) that Canadian Greek does not seem to have suffered attrition in terms of the morphological complexity of its system. Attrition would be a change one might expect (Montrul and Yoon 2019), given that English, the source language, has a much simpler morphology without specific gender or case suffixes for nouns and limited person endings for verbs. Contrary to this expectation, speakers of CG integrate the majority of loan nouns, assigning gender and inflection, and tend to integrate even nouns that are left unchanged in MG. The word [ˈceci] ‘cake’ is a typical example of this tendency, which in MG, appears as κέικ [ˈceik].

6. Conclusions

We have examined the speech and linguistic practices of first-generation Greek immigrants who arrived in Canada between 1945 and 1975 and lived and worked in a contact situation among multiple language varieties (English, SMG, CMG, MG dialects). Concerning the loan integration in CG, it was shown that most English loan nouns are subject to complete integration into the MG nominal system as they are reanalyzed as stems, they are allotted grammatical gender, and they receive inflection according to the native inflection rules. For non-human loanwords, there is a tendency among immigrants to assign neuter gender to them, in line with the same tendency in MG, and an inflectional ending of the most productive in MG IC6. Loan noun integration, which closely follows the rules of Greek morphology, shows that CG of first-generation immigrants has not undergone morphological attrition.

In the case of the unstressed syllabic augment and the imperfective past suffixes, we observed the existence of common patterns of usage in the speech of Greek-Canadians from various dialectal backgrounds, in that the informants employ features that belong to either the dominant dialectal variety of MG, that is the Peloponnesian variety, or to the vernacular form of CMG. As shown in Section 4, CG is characterized by the maintenance of various elements of this CMG, which was brought from Greece, and was used in at least those urban settlements where people with different MG linguistic backgrounds were living, working, and generally acting together. Crucially, the variation detected in the speech of ImmiGrec project participants consists of linguistic features that belong to three sources: (a) A native MG dialectal variety, (b) the vernacular form of CMG, (c) a contact effect with the dominant language, that is, English, or with the dominant MG dialect in most Greek-Canadian communities, i.e., Peloponnesian. We would like to suggest that the most promising approach for understanding this variation would be a variationist analysis which is theoretically grounded in the studies about koinéization (cf., amongst others, Bubenik 1993, Britain and Trudgill 2005, Trudgill 2008). Britain and Trudgill (2005) have shown how the rapid increase in the population of the city of Norwich during the 19th century has resulted in the formation of a new dialect through the accommodation of the mutually intelligible varieties of speakers who immigrated from the rural areas around the city. In a similar sense, it is possible that, upon the establishment of Greek communities in Canada, a Canadian-Greek koiné began to develop, influenced by, but essentially independent from, CMG, SMG, and the dominant MG dialect in their social context, as well as under the heavy influence of Canadian English.

However, it would be difficult to give a simple answer to the basic question that could arise regarding the reasons that triggered the emergence of the new CG variety. On the one hand, our speakers are immigrants who came to Canada with non-prestigious, perhaps even stigmatized dialectal backgrounds, such as the ones in Regina, which was transformed through contact with CMG/SMG and other Greek dialectal varieties of immigrants who were members of the same Greek community. On the other hand, we observe that even speakers of Peloponnese, a variety that was considered to be basic for CMG/SMG, do not always use the expected patterns in terms of unstressed augment deletion and the use of the imperfective past suffixes but they seem to adopt, although to a lesser extent, features of other regional varieties. Trudgill (2001) argues that members of small, stable, tightly-knit societies are likely to share more linguistic information than members of larger, more dynamic, and loosely-knit communities. Thus, in the sociolinguistic situation of our participants, it is likely that Greek immigrants coming from various dialectal regions adopted more variable patterns in order to downplay their differences and emphasize their common bonds.