Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean E-Book
Erhalten Sie Zugang zu diesem und mehr als 300000 Büchern ab EUR 5,99 monatlich.
- Herausgeber: Narr Francke Attempto Verlag
- Kategorie: Geisteswissenschaft
- Serie: Sprachvergleich
- Sprache: Englisch
Latein und Griechisch werden in diesem Sammelband unter dem Aspekt des Sprachkontakts untersucht, ein Thema, das in unserer globalen und multiethnischen Gesellschaft besonders aktuell ist. Spezialist:innen verschiedener Universitäten und Länder nehmen in Ihren Beiträgen unter anderem die linguistische Variation der griechischen Dialekte, den griechisch-lateinischen Bilinguismus, den Sprachkontakt im alten Italien, Mittleren Osten und Mittelmeer sowie Übersetzungen und Glossen in den Blick. Landkarten und Bilder alter Inschriften und Manuskripte bereichern die Diskussion. Aus interdisziplinärer Perspektive wird außerdem die Linguistik des Lateinischen und des Griechischen in ihrem Zusammenhang mit Epigraphik, Philologie, Textkritik und grammatischer Theorie untersucht. Neben Latein und Griechisch werden Daten zahlreicher alter und moderner Sprachen mit einbezogen.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 1043
Veröffentlichungsjahr: 2024
Das E-Book (TTS) können Sie hören im Abo „Legimi Premium” in Legimi-Apps auf:
Ähnliche
Carlotta Viti (ed.)
Ancient Greek and Latin in the linguistic context of the Ancient Mediterranean
DOI: https://doi.org/10.24053/9783823395850
© 2024 • Narr Francke Attempto Verlag GmbH + Co. KGDischingerweg 5 • D-72070 Tübingen
Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetztes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen.
Alle Informationen in diesem Buch wurden mit großer Sorgfalt erstellt. Fehler können dennoch nicht völlig ausgeschlossen werden. Weder Verlag noch Autor:innen oder Herausgeber:innen übernehmen deshalb eine Gewährleistung für die Korrektheit des Inhaltes und haften nicht für fehlerhafte Angaben und deren Folgen. Diese Publikation enthält gegebenenfalls Links zu externen Inhalten Dritter, auf die weder Verlag noch Autor:innen oder Herausgeber:innen Einfluss haben. Für die Inhalte der verlinkten Seiten sind stets die jeweiligen Anbieter oder Betreibenden der Seiten verantwortlich.
Internet: www.narr.deeMail: [email protected]
ISSN 2569-2275
ISBN 978-3-8233-8585-1 (Print)
ISBN 978-3-8233-0521-7 (ePub)
Contents
Contributors
Mauro ARESU:
Research fellow in Latin Language and Literature at the University of Cagliari (Italy), Department of Literature, Languages and Cultural Heritage. Email: [email protected]
Monique BILE:
Associate Professor (Maître de Conférences) Emer. of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: [email protected]
Vit BUBENIK:
Professor Emer. of Linguistics at the Memorial University of Newfoundlands (Canada), Department of Linguistics. Email: [email protected]
Silvia CABRIOLU:
Post-Doc in Historical and General Linguistics at the University of Cagliari (Italy), Department of Literature, Languages and Cultural Heritage. Email: [email protected]
Gualtiero CALBOLI:
Professor Emer. of Latin Language and Literature at the University of Bologna Alma Mater Studiorum (Italy), Department of Classical and Italian Philology.
Email: [email protected]
Paola COTTICELLI-KURRAS:
Professor of Historical and General Linguistics at the University of Verona (Italy), Department of Cultures and Civilizations.
Email: [email protected]
Francesca COTUGNO:
Post-Doc in Historical and General Linguistics at the University of Verona (Italy), Department of Cultures and Civilizations.
Email: [email protected]
Filip DE DECKER:
Post-Doc holder of a Marie Skłodowska Curie Actions, European Fellowship, and research fellow at the University of Verona (Italy), Department of Cultures and Civilizations. Email: [email protected]
José Luis GARCĺA RAMÓN:
Universität zu Köln / Università Cattolica del Sacro Cuore di Milano (Italy), Dipartimento di Filologia Classica, Papirologia e Linguistica Storica. Email: [email protected]
Georgios K. GIANNAKIS:
Professor of Greek Linguistics at the Aristotle University of Thessaloniki & Center for the Greek Language (Thessaloniki, Greece), Department of Linguistics. Email: [email protected]
René HODOT:
Professor Emer. of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages Email: [email protected]
Stella MERLIN:
Post-Doc in Historical and General Linguistics at the University of Naples Federico II (Italy), Department of Human Studies. Email: [email protected]
Anna ORLANDINI:
Professor Emer. of Latin and Greek Linguistics, “Centre A. Ernout”, Paris-Sorbonne (France). Email: [email protected]
Valerio PISANIELLO:
Associate Professor of Historical and General Linguistics at the University of Chieti-Pescara (Italy), Department of Languages, Literatures and Modern Cultures. Email: [email protected]
Paolo POCCETTI:
Professor Emer. of Historical and General Linguistics at the University of Rome 2, Tor Vergata (Italy), Department of Literary and Philosophical Studies and of History of Art. Email: [email protected]
Luca RIGOBIANCO:
Researcher of Historical and General Linguistics at the University Ca’ Foscari of Venice (Italy), Department of Humanities. Email: [email protected]
Carlotta VITI:
Professor of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: [email protected]
Guy VOTTÉRO:
Professor Emer. of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: [email protected]
Rex WALLACE:
Professor Emer. of Classics at the University of Massachusetts Amherst (United States of America), Department of Classics. Email: [email protected]
Emmanuel WEISS:
Associate Professor (Maître de Conférences) of Greek and Latin Linguistics at the University of Lorraine (France), Department of History and Culture of Antiquity and of Middle Ages. Email: [email protected]
Michael WEISS:
Professor of Classics and Indo-European Linguistics at Cornell University, Ithaca, NY (United States of America), Department of Linguistics.
Email: [email protected]
Ryan WINDHEARN:
Research fellow in Indo-European Linguistics at Cornell University, Ithaca, NY (United States of America), Department of Linguistics.
Email: [email protected]
Introduction: Language contact, comparative linguistics, and comparative literature in their historical and cultural context1
1.Possible dialogues between language contact and historical linguistics
The popularity of linguistic research traditions often depends on factors external to language itself. In our global, multilingualMultilingualism and multicultural societies, a particular interest has recently developed in academia around the subject of language contactLanguage contact. In general linguistics, since Weinreich’s (1977) pioneering studies on this topic, language contact has been the subject of publications with data drawn from many different languages (cf. Thomason & Kaufman 1988; Aikhenvald & Dixon 2001a; Thomason 2001; Matras 2009; Ansaldo 2013; Bakker & Matras 2013; Grant 2020, etc.). For example, it has been often observed that nouns (especially nouns other than kinship terms, body part nouns, personal pronouns and low numerals) are usually borrowedBorrowing more easily than verbs or other parts of speech, and that lexical items in general are usually borrowed more easily than phonemes and morphemes (cf. Moravcsik 1978). But this also depends on the intensity of language contact. It is true that borrowing particularly prevails in situations of casual or superficial contact. In cases of intense contact such as substrateSubstrate interference, however, structural transfer has no constraints, and usually occurs earlier and more extensively than lexical transfer (cf. Thomason & Kaufman 1988: 83ff). For example, the Asia Minor Greek dialects studied by Dawkins (1916) adopt many features of Turkish grammar, ranging from the loss of gender and adjective-noun agreement to word order and even vowel harmony. In general, it is now recognized that manifestations of language contact depend more on the socioculturalSociolinguistics settings of a speech community than on purely linguistic factors. More specific predictions can be made by considering social variables such as the presence of an indigenous superordinate group, a migrant superordinate group, an indigenous subordinate group or a migrant subordinate group (cf. Thomason 2001: 23), as well as prestigePrestige factors, since the same linguistic phenomenon may have different outcomes according to different variables of language contact. There are various theories of language contact, which is also acknowledged as a major mechanism of language change. For example, Harris & Campbell (1995) consider language contact as one of the three main mechanisms of syntactic change, alongside reanalysis and extension.
By contrast, in the past, language contact was neglected. This neglect is even more pronounced in the field of historical linguistics and Indo-European (IE) studies. The reconstruction of Proto-Indo-European (PIE) is achieved by observing mechanisms of internal language change. This is not only evident in Internal Reconstruction, by definition, but also in the Comparative MethodComparative Method. It is traditionally assumed that two or more structures are cognates only once language contact, as well as other possible factors of formal correspondence such as chance and independent drift, are ruled out (cf. Meillet 1925). Accordingly, the establishment of a genetic relationship is “an argument by elimination” (Harrison 2003: 215). The Neogrammarians, among the founders of the Comparative Method, invoked language contact only to explain exceptions to sound lawsSound change. Latin rūfus “red, reddish; redhaired” (< PIE *h1rowdh-ó-), for example, presents a radical fricative f, instead of a plosive as in ruber “red” (< PIE *h1rudh-ró-, cf. gr. ἐρυθρóς), because it is borrowed from a Sabellic dialect, as the change PIE *dh > f is unconditioned in Sabellic. Exceptions to sound laws are usually unsystematic (unless they can be subsumed under another more specific sound law). Still in recent times, Watkins (2001) considers the Comparative MethodComparative Method to be the best heuristic tool to describe language relatedness, and dismisses alternative models which also take into account areal diffusion (e.g., Punctuated Equilibrium, cf. Dixon 1997) as unsatisfactory. Contact is still currently often seen as “a confounding factor” (Walkden 2013: 96) in the establishment of genetic inheritance. According to Dunn (2015: 190), for example, “all historical linguistics is phylogenetic, since phylogenetics encompasses the scientific investigation of the descent of organisms in general”, which in principle excludes the possibility of historical language contactLanguage contact. There have been studies of language contact in IE scholarship (e.g., Meid’s 2012 writings on the contacts between Celtic and Germanic, as well as on Cimbrian, an Upper Germanic variety spoken in some northeastern regions of Italy which has been strongly influenced by Italian). These studies, however, have set out to discuss specific cases of language contact, rather than to establish a theory of language contact, as we have in general linguistics. Moreover, they are rare in comparison with IE studies on other linguistic topics such as historical phonology, morphology and (to a lesser extent) syntax. As the IE languages are the most studied languages of the world, this approach has also affected the diachronic study of other language families.
Anybody trained in IE linguistics, myself included, agrees with Watkins (2001) that the Comparative MethodComparative Method remains the most valuable scientific instrument of linguistic reconstruction in IE and beyond (cf. also Baldi 1990; Polomé & Winter 1992; Weiss 2015). As Weiss (2015) rightly points out, one of the great merits of the Comparative Method is that it enables the reconstruction of much deeper linguistic stages than those of our earliest written records. Being the best method, however, or being a correct method, does not mean being perfect or complete – no scientific method can have such pretensions, as it always implies a certain simplification of reality. The Comparative Method devotes more attention to form than to function, since changes affecting forms are more regular and therefore more predictable than changes affecting meanings. Owing to this, linguistic reconstruction is performed by simplification, since there is apparently no need for two or more forms for (what seems to be) the same function: given a form x and a cognate form y, the Comparative Method reconstructs one single form for the proto-language, which may be x or y or a different form from which both x and y descend – it does not reconstruct both forms. As such, the Comparative Method cannot always adequately deal with linguistic variationLinguistic variation. This reduction of structural diversity is also a reason why language contact is traditionally excluded from the domain of reconstruction: language contact intrinsically implies language variation. Co-occurring structures, however, often show a functional competition, if function is not limited to the basic lexical-semantic level but also includes pragmatic, sociolinguistic, and dialectal factors. This functional competition, which may be captured by the framework of Construction Grammar (Goldberg 1995; 2006), for example, emerges even more clearly in situations of language contact. Moreover, if one accepts the uniformitarian assumption that languages of the past behaved in the same way as contemporary languages, functional competition and structural variation may be also reconstructed for proto-languages. The reconstructed PIE, for example, was neither simpler nor more regular than the attested daughter languages in morpho-syntax (cf. Viti 2015). Although a functional competition can be more easily identified in different morphological and syntactic structures than in different phonetic representations, phonetics is also amenable to a functional analysis. For example, in Kiezdeutsch (the variety of German spoken by young generations in multicultural suburban environments, originally in Berlin-Kreuzberg), Weirich et al. (2020) have observed that the phonetic alternation between the voiceless palatal fricative [ç] (ich-laut) and its coronal variant [ɕ] may express distinct sociolinguistic connotations. Although the pronunciation [ç] as in Hochdeutsch predictably has a higher prestige than its Kiezdeutsch realization [ɕ], not all listener groups behave alike (e.g., the non-canonical pronunciation is more stigmatized by older listeners than by young listeners, by in-groups as opposed to out-groups, etc.).
Similar considerations may be applied to the Family Tree model. The results of the Comparative Method make it possible to represent the relationship among the various IE languages in a Family Tree, which does not allow cross-branching. That is, after two or more branches are separated, they do not intersect anymore. In this, sub-branching is typically established on the basis of shared innovations. The Family Tree is also a cultural product of its time. Schleicher (1861), one of the creators of the Stammbaum, was inspired by the newly developed method of the stemma codicum of the philological research tradition. The latter aims to establish the relationship among different manuscriptsManuscripts of a certain text, and reduces the lectio transmitted by two or more codices to one original version. The stemma codicum is also presented as a cladogram (cf. Hoenigswald 1963; Fisiak 1990). We have to consider, however, that the Family Tree may not be equally satisfactory to describe cognateness in all language families. It works well for the IE languages, for which it was originally introduced, because these languages have spread across an extremely large area, ranging from Iceland and Ireland to the West to the Tarim Basin (in modern-day Xinjiang, China) to the East, and some of them have been isolated from each other for centuries. When, however, a language family has remained in a relatively more restricted area, the structure of a Family Tree is more controversial. In Semitic, for example, which has remained substantially limited to the Near East until the Muslim conquests in the early Middle Ages, there are different models of classification (cf. Huehnergard & Rubin 2012), and sub-branching is much more debatable than in IE. Similarly, a Family Tree does not adapt so well to situations of extensive bilingualism, as in the languages of South East Asia, where cross-branching may occur.
At their time, the concepts of sound lawsSound change and of the Family Tree were also challenged, especially by scholars working on dialectology and modern languages (cf. Schuchardt 1885), whose complex developments were not easily amenable to the regularity of the Neogrammarians’ Lautgesetze. Schmidt’s (1872) Wellentheorie, implying a gradual diffusion of linguistic features from its region of origin, was posited as an alternative to the Stammbaum to explain branching within the IE language family (cf. also Porzig 1954 for sub-branching and areal diffusion in IE). De Saussure’s (1916) attention to the social factors underlying linguistic change was also critical of the Neogrammarians’ reconstruction methodology. These alternative models (which are not incompatible with the Comparative Method in principle) were adopted by a minority of scholars. Otherwise, the Comparative Method has been tested by means of computational and statistical approaches, which are quite popular at the present time (cf. Ringe et al. 2002; McMahon & McMahon 2005, etc.; more recently, cf. Heggarty et al. 2023). The earliest statistical methods applied to historical linguistics, going back to Swadesh’s Glottochronology in the 1950s, assumed a constant rate of language change (which is clearly not true). Later models, e.g., the “character-based models of change”, continued to consider parsimony to be one of the most reliable criteria to account for branching, which is also controversial (cf. Dunn 2015: 196ff). The newest computational models, such as “likelihood methods”, accept a variable rate of change, but still assume that this variability can be measured by means of mathematical algorithms. Some models evaluate the statistical likelihood of branching by means of the Bayesian Monte Carlo Markov chain. Different computational methods exist, whose reconstructions do not always overlap (cf. Widmer 2018).
Most historical linguists, however, are sceptical of computational approaches to language change. This is not only due to the fact that most historical linguists are unfamiliar with complex mathematical methods (this problem would be easily solved by collaborating with a computer scientist), but also and especially because they reject a basic assumption of these methods, that is, the idea that language change and language affiliation can be statistically predicted. In historical linguistics, most scholars hold that language change is regular a posteriori but unpredictable a priori (or at least not predictable by mathematical methods), and characterized by variation between different outputs in the intervening time: A > A/B > B, where there is no way to predict that A will change to B. For instance, in sound changeSound change, we commonly have [s] > [h] > zero, but some languages retain [s] (e.g., Indic), others develop [h] (e.g., Iranian) and others proceed to the development of zero (e.g., Greek psilotic dialects). If we have to put forward hypotheses as regards the more or less likely occurrence of change, these hypotheses depend on a multitude of linguistic and extra-linguistic factors, which by definition defy considerations of parsimony, and which cannot be captured by a mathematical algorithm (Ockham’s razor usually does not work in language change, see below). Rural spaces, for example, are usually more conservative than urban spaces (cf. Janda & Joseph 2003: 62-63). In periods of political, economic and social stability, language change is slower than during periods of turmoil or war (but we cannot predict based on mere linguistic features whether a speech community will undergo social instability). There is a difference between variation in social sciences, which can be measured statistically, and the implementation of a language change. See Lazzeroni’s (1987) dated but still instructive observations against computational approaches to language change.2 Especially for lexical features, matters of a speaker’s choice play a crucial role (cf. Hagège 1993: 9ff). For ancient languages, we also have to take into consideration that the interpretation of a language change can also depend on text transmission, as some language varieties (especially those associated with socially higher registersRegisters) may have been better transmitted than others and may therefore not accurately reflect the language as it was really used. Typologists who know languages prima manu by fieldwork also realize that language change has no predictable rate and mainly depends on social and cultural factors. Consequently, they also do not appreciate computational approaches (cf. Aikhenvald & Dixon 2001b: 7 et passim).
The matter is not settled, however. As can be seen, the Comparative MethodComparative Method, as well as the Family Tree, are not a tenet established once and for all, but rather the result of a lively debate that has developed across two centuries (Hymes 1974; Fox 1995; Durie & Ross 1996). Although they are substantially correct, new findings from other language families or other geographic areas, as well as from dialectology and language contact, may be useful to elucidate certain linguistic phenomena concerning language variation which are not adequately captured by these theoretical tools.
The reluctance to contemplate contact factors in historical linguistics is based on two main assumptions, related to each other, which have been proven to be wrong in more recent literature. Firstly, it was assumed that internal and external language change are mutually exclusive, and secondly, that an internal explanation to language change is always to be preferred to an external one. Lass (1997: 201ff), for example, argues that explaining language change in terms of language contact must be the ultima ratio to be used only when an internal explanation (by means of analogy, reanalysis, etc.) is not available. This is because internal change always takes place, even without an external influence, while external change is not necessary. An explanation in terms of inheritance therefore seemed to be more parsimonious. This argument is substantially similar to that used by the Neogrammarians to account for sound laws. Terms such as “simple”, “economic”, “consistent” or “harmonious” also abound in structuralist studies, and are evaluated positively in comparison to complex, heavy or inconsistent systems. Although these beliefs continue to appear, especially in the generative research tradition, nowadays historical linguists commonly admit the possibility of multiple causes interacting in language change. Indeed, as Joseph (2013) observes, it is good practice in historical linguistics to take all possible factors into consideration, since a historical explanation has to be correct and complete, and not necessarily simple. Note that Joseph has worked intensively on Balkan languages – he understands the importance of language contact very well.
Dorian (1993) also contests previous assumptions whereby divergence and convergence in language change are simplistically associated to internal and external factors, respectively. She shows that language contact may bring about both divergence and convergence – it must be studied on a case-by-case basis. This may be seen as less economic, but it is more reliable. Moreover, a linguistic feature may develop by internal language change and at the same time be reinforced by the pressure of language contactLanguage contact. This is the logic of Heine & Kuteva’s (2003) “contact-induced grammaticalization”, that is, a kind of grammaticalization (i.e., an internal language change by which content words tend to become more grammatical with increasing formal erosion, fixed position, semantic bleaching, etc.) which also occurs in situations of language contact. Heine & Kuteva (p. 71-73) report the case of some Western Slavic languagesSlavic languages, such as Sorbian, Czech, and Slovenian, which at least in the spoken language have developed patterns of marking (in)definiteness in a similar way to the article. While Czech and Slovenian use some demonstrative elements to mark definiteness, Sorbian has grammaticalized the use of the numeral “one” as an indefinite marker. On the one hand, these represent the typical paths leading to the formation of articles in languages. On the other, the influence of German is undeniable. German, which possesses both definite and indefinite articles, is spoken in geographically close areas which have always had an intense commercial and cultural interaction with West Slavic speech communities.Slavic languages Also in this case, contact-induced grammaticalization is especially observable in morpho-syntactic change, but phonology is by no means excluded. In both cases, the change may affect both marked and unmarked features. For example, retroflex consonants, which represent phoneticallyPhonetics marked segments, could develop in the Satəm group of IE languages by internal language change already according to the so-called “Ruki” sound law. In Indic, however, retroflex consonants have a much wider extension than in ArmenianArmenian or Balto-Slavic. In Indic, retroflex consonants are presumably reinforced by a substrateSubstrate effect of the Dravidian languages, for which a retroflex articulation is especially characteristic. Retroflex sounds can be reconstructed for Proto-DravidianDravidian languages as well.3
Contact-induced grammaticalization may change past assumptions concerning language contact, as contact is not invoked anymore to explain the development of a seemingly unnatural change or of a language feature that appears to be inconsistent with other features of the system, but rather interacts with internal language change in the same direction. Note that the capacity to assess a directionality in language change is actually a strong point of the Comparative Method: a [p] changes into a [f] much more often than the other way round, so that, in the presence of a [p] in one language and of a corresponding [f] in another related language, we can also reconstruct [p] for their proto-language ceteris paribus. This directionality has been also extended from the domain of phonologyPhonetics, for which it was originally postulated, to morphology and syntaxSyntax. For example, it has been demonstrated that the change from postpositions to case markers is much more frequent than the opposite change (cf. Hagège 2010; Givón 2021, etc.), so that the proto-language of a language with postpositions and of another language with cognate case markers also has to be assigned postpositions. The same directionality may be identified in case of language contact. I have never had difficulties in accepting this concept as I saw it in operation in my own experience. The reader shall forgive this personal anecdote as it is functional to my argument. When I was working at the University of Zurich, German was my working language as well as the language of my daily life. When I had to express likes, I always said ich mag X “I like X” with canonical subject marking, instead of X gefällt mir with non-canonical subject marking. In most contexts both constructions are possible in German (apart from specific cases such as preferences concerning food, where we commonly use non-canonical subject marking with the verb schmecken “to taste”). Still, I consistently used mögen and not gefallen. Note that my mother language, Italian, only admits non-canonical subject marking with likes: mi piace has the same syntactic pattern as German es gefällt mir. (Italian is not like French, where the structure X me plaît exists but is less frequent than j’aime X, j’adore X. The latter options are not available in Italian for this function.) The same applies to my native dialect, Tuscan, which has a different verbal lexeme but the same syntax: mi garba (to.me it.pleases) “I like”. Thus, when speaking in German, I used the pattern which was different from the one used in my mother language. I asked my Italian friends who were living in Zurich, and they told me it was the same for them. This is understandable when we consider that diachronically, in internal language change, experiential predicates tend to acquire canonical subject marking. In Old English, the verb lician “to like” required a dative experiencer, listan “to desire” (German gelüsten) an accusative experiencer, and so on. With time, the English language has lost many impersonal constructions of experiential predicates (cf. van der Gaaf 1904; Allen 1995, etc.). The pattern where the experiencer is also the subject turns out to be preferred in internal language change as well as in situations of language contact.
All this confirms that the exclusion of language contact from the practice of linguistic reconstruction may seriously impinge upon an adequate understanding of language change, since no language has evolved in isolation. As LaPolla (2009: 227) pointed out, “language contact is a part of the development of all languages, and so we cannot treat internal language change independently from changes influenced by language contact”. Given the fundamentally communicative function of language, contact seems rather to be a natural condition of language, at both a microlevel (as in dialects and sociolects) and a macrolevel (involving different languages), and factors of multilingualism often have a profound effect on the development of a language. A scientific dialogue is therefore needed between scholars of historical linguistics and scholars of language contact, who so far have been usually working independently from each other, to compare and combine their techniques, so that some findings of language contact may be incorporated into the methodologies of linguistic reconstruction. This implies an interdisciplinary relationship of historical linguistics in general, and IE studies in particular, with other disciplines focusing on language interaction and on its extra-linguistic context, such as dialectology, sociolinguistics, anthropological linguistics, pragmatics, discourse analysis, contrastive linguistics, and translation studies.
2.Intellectual history of language contact in Indo-European studies
If in principle there is no incompatibility between language contactLanguage contact and the Comparative MethodComparative Method, one may wonder why language contact has been so long neglected in IE studies as compared to other linguistic topics. This is probably also due to ideological reasons. In the past, it was long assumed that language was directly connected to “race”, and this in turn discouraged the study of non-genetic linguistic relationships. We have to put the founding writings of IE linguistics in their historical context. Apart from more or less impressionistic statements about the possible relationships among various IE languages, the earliest serious endeavours to reconstruct the history of IE originated in a series of scholars mainly operating in 19th century’s Germany – a period in which issues of race and nationNational identities, in various forms, played a prominent role in scientific and literary discussions. The first generations among these scholars, that is, Wilhelm von Humboldt (1767-1835), Jacob Grimm (1785-1863) and Wilhelm Grimm (1786-1859), Franz Bopp (1791-1867) etc. – immense scholars – were deeply influenced by Romanticism. This movement, which developed between the end of the 18th century and the first half of the 19th century, was rooted in the idea of a “nation”, considered as a unity of language, religion, traditional customs etc. ultimately to be attributed to an ethnic group.4 Later, in the second half of the 19th century, Europe was rather influenced by Positivism as well as Darwinism, which are often presented as a reaction of rational thinking against Romantic ideas of feelings and subjective interpretations, but still shared the interest in genetic connections. Darwin’s On the origin of species, published in 1859, postulated a common descent of species through a branching pattern of evolution. In addition to philology, as we have seen in §1, Schleicher (1821-1868) was also deeply influenced by Darwin’s theory of evolution (cf. Schleicher 1863). Note that the basic terminology of the family tree (ancestor language, mother / daughter / sister languages, language family, cognate, etc.) is based on metaphorsMetaphor of genetic relations. The same metaphors recur in philology, which speaks about families of manuscripts, genealogy, generation, spurious etc. Latin spurius, for example, originally meant “illegitimate” – it was used in legal language to denote the child of an unknown father or a child born from incest (cf. OLD s.v. spurius; EM 645). The language of metaphors is based on experiences and ideas that are commonly shared in a speech community.
All this fits in with the pervasive climate of nationalismNational identities, that is, the idea that a state must coincide with (what is considered to be) a nation, an idea which was well established in the 19th century. Several movements and wars nurtured by nationalist ideas developed in this period. A long hostility existed between Germany and France, for example, as both wanted to extend their control over continental Europe. After the Franco-Prussian War (1870-1871), the Second French Empire lost Alsace and Lorraine, which were annexed to the North German Confederation led by the king of Prussia. The “young” generations of IE scholars, the Neogrammarians (Leskien, Brugmann, Osthoff, Delbrück, Braune, Behagel, Paul, Sievers, etc.) – also immense scholars – working between the second half of the 19th century and the first decades of the 20th century, lived in this political and social climate. It is understandable that they were more interested in studying genetic inheritance than contact in languages. The interests of linguists, as of any other people, are influenced by the ideas of their time.
We all know how these dangerous nationalist ideas ended up in the subsequent decades and there is no need to dwell further on those dramatic events. Under Nazism, which ruled Germany from 1933 to 1945, the very concept of Indogermanisch was associated to ideas of a pure “Aryan” race. Even the writings of early Indo-Europeanists were reinterpreted in the light of Nationalsozialismus – note the emphasis on “national” in this term. It is clear that language contact could not be an ideal topic of research at that time. Scholars interested in linguistics therefore continued to stress the indigenous material in a language rather than manifestations of language contactLanguage contact. The latter was considered as a sort of contamination of language. The fact that Ancient Egypt – in Africa – and the Ancient Near East – populated by Semitic people, among others – had advanced cultures which exerted a strong influence on the Greek world since Mycenaean times (see below) was virtually a taboo under the Nazis. But recall that these ideas circulated well beyond Germany and their allies. France and Great Britain had maintained vast colonial empires for centuries, again justified by ideas of racial superiority. In France, the diplomat Arthur de Gobineau, author of the Essai sur l’inégalité des races humaines (1853-1855), asserted the superiority of the Aryan race. In the British Empire, scholars of Sanskrit typically ignored or downplayed the linguistic and cultural contribution of DravidianDravidian languages and other indigenous languages of India to the development of Indic. It is enough to mention Beames (1872-1879), which remained one of the standard books on New Indic languages at least until Masica (1993). Beames often expressed the argument that the Indo-Aryans were more donors than recipients of linguistic features as they were, in his view, “superior morally as well as physically to the aborigines”, p. 10 et passim). The Jewel in the Crown did not obtain its independence until 1947. Again, language contact was hardly compatible with imperialist and racial arguments.
A few words are needed at this point on the Black Athena debate. As is well-known, Bernal (1987; 1991; 2006) considered the reception of Graeco-Roman civilization in Western tradition to be conditioned by racialist ideas, and proposed an alternative model in which certain Afro-Asiatic civilizations, notably Egyptians and Phoenicians, played a much more important role in the development of Greek language and culture. Relying on ancient Greek mythological and literary sources, he hypothesized that the ancient Egyptians and Phoenicians colonized parts of ancient Greece, which, in his view, has to be interpreted as a sort of periphery of the Levant. Moreover, he proposed numerous new explanations for Greek words with no etymology, or with a controversial etymologyEtymology, which he traces back to Afro-Asiatic expressions. In order to avoid misunderstandings, I wish to state clearly that I do not agree with Bernal’s linguistic reconstruction. Bernal is sympathetic with the idea of macro-families, such as Nostratic, and consequently does not follow the assumption of regular IE sound laws. In contrast, I do not subscribe to the concept of macro-families and follow the traditional procedures of the Comparative MethodComparative Method. In this, I agree with Jasanoff & Nussbaum (1996), who recall the method of finding correct etymologies, that is, by looking at regular sound changes, rather than superficial resemblance, etc.
However, while we share the same method of linguistic reconstruction, I disagree from Jasanoff & Nussbaum’s (1996) paper with regard to a couple of points.5 They maintain that Afro-Asiatic borrowingsBorrowing in Ancient Greek are “relatively few in number and – with some exceptions on the Semitic side – late in date” (p. 201). Relying on Masson (1967), a study on Semitic borrowings in Ancient Greek, they assign these terms to the domains of fabrics and items of clothing, commercial terms, vessels, and plants, e.g., Ancient Greek κύμινον n. “cumin”, attested since the Mycenaean age (ku-mi-no, ku-mi-na in a list of spices in Mycenae). But, in fact, Masson is much more open and cautious on this matter. Besides “sure” borrowings belonging to these semantic domains, she also discusses at length a series of “possible” borrowings, which are semantically much more heterogeneous.
Among these “mots dont l’origine sémitique est possible” (Masson 1967: 77ff), we find the name of certain animals, such as λέων / λĩς m. “lion” (p. 85-87). The former variant, λέων, is attested since Mycenaean (re-wo-pi, instr.pl.; re-wo-te-jo adj.). It is considered to be a borrowingBorrowing from Semitic (cf. AkkadianAkkadian lābu “lion”, Ugaritic lb’ “id.”, HebrewHebrew labī’ “id.”) or from an unknown Mediterranean language. The rarer variant λĩς has been connected with Hebrew laīš “lion”, but in this case as well the evidence is not conclusive, and it is possible that we are dealing with a mot voyageur of the MediterraneanMediterranean. Cf. also DELG 635; GEW 113; EDG 854 (for which a “Semitic origin is probable” especially for λέων – note that Beekes is not particularly keen to admit Semitisms). The lion is not native to Europe while it lived in various Near Eastern desertic regions for centuries (in the Near East, the last exemplar of an Asiatic lion was killed in Iraq in 1918). As such, the lion was commonly portrayed in Near Eastern art since antiquity, when it was a symbolic image of the king.
Despite the uncertainties about the ultimate source of Ancient Greek λέων and λĩς (in my view, they are more probably borrowed from an unknown Mediterranean language than from Semitic), what is sure is that these words are not IE. Still, some IE scholars have thought the contrary in the past. I briefly discuss this case not to revisit the etymology of λέων and λĩς, as both Masson (1967) and all lexica recognize that they are not of IE origin, but rather as an exercise of intellectual history of language contact. According to Thieme (1954: 32-37), λέων is etymologically connected with Old IndicVedic ruváti “lows” (sound of cattle) and with Homeric GreekHomer, Homeric Greekβουλυτός m. “evening” (the latter interpreted as “time when the cows come back from the pasture lowing”, rather than “time for unyoking oxen” (βου-λυ-τός), as it is commonly thought). Λĩς, instead, is connected by Thieme to Sanskrit līna- “lying or resting on, lurking, hiding’ as in Kālidāsa’s expression kuñjalīnān … siṃhān “lions hiding in the underbrush”. Clearly, all this is pure fantasy. But we may better understand the ideological ground behind this hypothesis if we know that it comes from a study, entitled Die Heimat der indogermanischen Gemeinsprache, where the author posits the original homeland of the Indo-Europeans in Northeastern Europe by means of lexical examples such as the name of the salmon (now obsolete for this argument). If then we know that Paul Thieme (1905-2001) was working in Germany during the Third Reich (he also served in the German army during the Second World War), then the picture is much clearer. With this, I do not want to imply that Thieme, one of the greatest Indologists, was sympathetic towards Nazism. On the contrary, we have evidence that he was hostile to the Nazis. For example, in his study Fremdling im Veda, published in 1938, Thieme studied the derivation chain of Vedic arí- “foreigner; enemy”, aryá- “related to the foreigner; kind, favorable”, and ā́rya-, the latter being a common endoethnonym of the Indians of IE origin, opposed to dásyu- or dāsá-). Although his interests rely rather on the morphology of these forms, he also explains that ā́rya- is to be interpreted in the sense of “hospitable” (zu den Gastlichen gehörig, wirtlich, p. 145), protector of foreigners. He argues that the meaning “lord” as well as the ethnic sense are only secondarily derived from a denotation of a master that is generous to his guests. This argument was dangerous at that time and yet corrected the Nazi’s abuse of the word “Arier”. Thieme later defected during the Cold War from the German Democratic Republic to West Germany (where he became Professor of IE studies at the University of Frankfurt), then moved to the US etc. Still, as everybody else, he reflected the interests of his own time.
My point is therefore that, when we study problems of language contact in the ancient IE languages, we must be very cautious because most authors of our sources lived in periods when language contact, as well as cultural contact, was refused or downplayed for ideological reasons related to racial and nationalistNational identities arguments – not necessarily their own personal ideas, but certainly ones that were widespread both in society and in academia. The relationship between linguistics and history is not limited to language change, as in historical linguistics, but also concerns the history of ideas – intellectual history. Ideas affect research interests and the interpretation of data. We have also to pay attention to the fact that etymologicalEtymology dictionaries and the secondary literature in general often repeat former sources without further elaborating the argument. For example, (apart from the Pre-Greek argument, see below), EDG is often a translated copy of GEW – which is understandable, as unfortunately Beekes died before having the time to complete his work.
It is also possible (albeit not certain) that the list of early Semitisms in Ancient Greek has to be enlarged with respect to what is commonly assumed to include certain theonyms. One finds insights on this in the works of Burkert, a great scholar of Greek religion and literature who had additionally been trained in IE linguistics and also studied AkkadianAkkadian. Burkert (2009: 36-37) suggests that the name Τηθύς, -ύος f. “Tethys”, denoting the wife of Oceanus and the mother-god of the river-gods and Oceanides, may be a borrowing from a variant of the name of the Akkadian deity Tiamat, which is also written as Tiamtu, Tâmtu, Tawatu etc. This is not argued just on the base of sound similarity, but rather because the Homeric passage at issue (Hom. Il. 14,200-201) has precise analogies in the Enūma Eliš, one of the most important creation myths of the Babylonian tradition. In the Iliad, Hera says to Aphrodite: εἶμιγὰρὀψομένηπολυφόρβουπείραταγαίης, / ὨκεανόντεθεῶνγένεσινκαὶμητέραΤηθύν “For I am going to see the boundaries of the all-nurturing earth, and Oceanus, the origin of the gods, and mother Tethys”. At the beginning of the Enūma Eliš epic, we read: “when on high no words was used for heaven / nor below was firm ground called by name, / Primeval Apsu was their progenitor, (Akkadian apsûm-(ma) rēštû zārûšun) / Mother Tiamatwas she who bore them all (mummutiāmtumuʾallidatgimrīšun)”.6 As Burkert (2009) observes, the similar formulation (Ὠκεανός / Apsu are described as progenitors, and Τηθύς / Tiamtu as mothers) suggests that the concepts of these gods are related. In the Akkadian theogony, Tiamtu plays an important role; for example, she is engaged in the decisive battle against Marduk (cf. Jacobsen 1968). Of Tethys, instead, we have no further information – note that we are speaking about Τηθύς “Tethys”, and not about Θέτις “Thetis”, the mother of Achilles (although some overlaps between these two marine goddesses is plausible).7 In Homer, Tethys is mentioned only in this incidental passage, within the story of the Διὸςἀπάτη, as Hera has to find an excuse to obtain Aphrodite’s girdle in order to seduce Zeus. It is therefore more probable that the representation of this Greek goddess derives from the AkkadianAkkadian one rather than the other way round. The name of Aphrodite herself is possibly borrowed from Semitic, although not directly from the name of her mythological Northwest Semitic correspondence Aštoret / Astarte (the relationship with ἀφρός “foam” is secondary by folk etymology), cf. GEW 196-197; EDG 179. This is because the cult of the goddess of love, sexuality, and fertility, as well as war and power, is certainly of Mesopotamian origin (cf. Sumerian Inanna, AkkadianAkkadian Ištar, etc.), and not IE. An oriental origin is also probable for Adonis, who in Greek mythology is consistently associated with Aphrodite. On the one hand, from the content point of view, the cult of Aphrodite and Adonis may be a continuation, adapted to the Greek pantheon, of the ancient Sumerian couple of Inanna and Dumuzid, which had been variously replicated in other religions of the ancient Near East, as in the Semitic figures of Baal and Tammuz (cf. West 1997: 57). On the other hand, from the formal point of view, the name of Ἄδωνις has been usually explained as a borrowing from the Canaanite form ’ādōn, which means “lord” (cf. DELG 21; GEW 22; EDG 23).8
The right track to identify new possible loanwords, in my view, is to search appropriate functional correspondences in mythologyMyth and literature. That is, 1) when a Greek word is without a plausible IE etymologyEtymology (this is the conditio sine qua non to start the research), and 2) this word has a similar form and meaning with respect to a word of an ancient Semitic language (as we are dealing with different language families, sound laws do not help in this case), then 3) we should search for similar contexts in which the forms at issue appear. Words are not transmitted in isolation, but rather in a context, and therefore may maintain their connotations after being borrowedBorrowing (as the association of Tethys with water). This implies interdisciplinary connections between IE linguistics and ancient literatures, religions, and cultures. A proviso in the use of the term “similar” is needed. As we have seen, the Comparative Method is based on regular correspondence and not on superficial similarity. We have therefore to ascertain that the rendition of a certain foreign phoneme is consistent with that of other borrowings presenting the same phoneme (although the same foreign phoneme may be rendered in different ways in different periods in which it enters the target language. According to whether the borrowing is more or less ancient, its phonetic structure will also be more or less integrated in the phonology of the borrowing language). The hypothesis of a Semitic borrowing becomes more plausible when the form has a triconsonantal and disyllabic root, as this is the typical root structure in Semitic, while in IE the typical root structure is biconsonantal and monosyllabic (CvC).9
But the matter is much more complicated, as we have to consider different kinds of borrowingBorrowing. Borrowing is not just a simple transfer of lexical material from a source language A to a target language B. Sometimes, the integration of an originally foreign source may bear some analogy to other lexemes of the target language. For example, Ancient Greek ἀδάμας, -αντος m., attested as a common name meaning “hardest metal, steel” since Hesiod (later also “diamond” or “fixed, unalterable”), is often considered, especially for semantic reasons, to be originally “a loanword that was adapted by folk etymology” to δάμνημι (EDG 19; cf. also GEW 19; Ayil 2024: §18 s.v. חַלָּמִישׁ – ḥallāmīs̆). A possible comparison has been suggested in this sense with Akkadian adamu “valuable stone” (HebrewHebrew), but this is not certain (DELG 18 considers ἀδάμας a native term). Thus, when we do etymologies, we often move in the realm of possibility, rather than of certainty, as regular sound laws may interfere with analogy, as the Neogrammarians had already stated. Interferences are even more complex to detect when we deal with calquesCalques, or semantic loanwords, i.e., when the phonetic material is indigenous but the semantic pattern is influenced by the meaning of a foreign source. Being possible does not mean being wrong, and various degrees of probability exist, which must be judged case by case.
3.Ancient Greek in its contact context
3.1.Mediterranean substrates
Greek borrowingsBorrowing from the East are not only from Semitic, of course – borrowings from Semitic are simply more interesting because in such cases we can often make textual parallels. Some Semitic languages, such as AkkadianAkkadian (Eastern Semitic), Ugaritic, Biblical HebrewHebrew, and Aramaic (all three belonging to Northwest Semitic), had great literary traditions which were widely influential in the ancient Near East. Akkadian, in particular, is attested from the middle of the 3rd millennium BC until the 1st century AD (although in the last centuries it was only used for liturgical or academic purposes). By contrast, other linguistic sources are not attested and often not even identifiable. Still, it is especially to these unknown MediterraneanMediterranean substrateSubstrate languages that Ancient Greek owes the largest part of its non-inherited vocabulary. Greek forms provided with suffixes such as -νθος (e.g., ὄλυνθος m. “edible fruit of the wild fig”) and -(σ)σος (e.g., κυπάρισσος f. “cypress”) are commonly related to a non-specified Anatolian language (cf. Meillet 1948: 65 et passim). Forms such as Greek μίνθη f. “mint” and Latin menta “id.”, or Greek ῥόδον n. “rose”, Latin rosa f. “id.”, Persian gul “id.” (< *wṛdi-), the latter borrowed into Armenian ward, are too similar to be due to chance but cannot be related to any regular sound correspondence between these languages. There are plenty of these forms, especially among names of plants, animals, and concrete instruments, as well as toponyms, which the Greeks had found in their migratory path through the Balkans to the Aegean.
García Ramón (2011) observes that we have to assume more than one substrate languages, and identifies three of them, that is, a heterogeneous Mediterranean substrateSubstrate language, a Minoan substrate language, and an Anatolian substrate language (Hittite or Luwian). In the latter case, borrowings can also proceed in the other direction, that is, from Greek to AnatolianAnatolian languages, or can represent independent developments of previous sources used in the area. Beekes (2014) calls one of these non-attested Mediterranean substrateSubstrate languages “Pre-Greek”. According to Beekes, a word deprived of a PIE etymology probably has a Pre-Greek origin 1) when it presents an anomalous phono-morphology from an IE point of view (marked consonant clusters such as κχ, e.g., Βάκχος, or τθ, e.g., Ἀτθίς, -ίδος, and rare suffixes such as -αμβ-, -ανδ-, -ανθ-, -αγγ-, -ινδ-, -ινθ-, -ιγγ-, -υμβ-, -υνδ-, -υγγ-), or 2) when the root presents an anomalous alternance with other words, an alternance which is not found in inherited PIE words. The most recurrent patterns are π / πτ (e.g., πόλεμος / πτόλεμος “war”), ξ / σσ (e.g., Ὀδυσσεύς (also Ὀλυττευς) / Οὐλίξης, Οὐλιξεύς (further borrowed through Western Greek dialects into Lat. Ulixes), ττ / σσ (e.g., θάλασσα / θάλαττα “sea” (cf. also δαλάγχαν· θάλασσαν Hesychius.), σ(σ) / στ (e.g., φαῦσιγξ / φαῦστιγξ “blister”), plosive / zero (e.g., κάνδαρος / ἄνραξ “charcoal”) etc. Along these lines, Beekes (2014) reconstructs the phonology and the morphology of Pre-Greek.
On the other hand, it may seem that Beekes (2010; 2014) connects words with different places and manners of articulation, as well as with zero, when it is convenient, as some forms have been explained differently in the literature and may also receive an IE etymologyEtymology. Eg. ὀφθαλμός “eye” (originally Pre-Greek according to Beekes 2014: 100-101) is commonly derived from the PIE root *h3ekw- “see, behold”, similarly to Latin oculus “eye” and Vedic ákṣi- “id.”, whose radical labiovelar is differently simplified in Ancient Greek, sometimes with irregular results (e.g., Boeotic ὄκταλλος “eye”), for taboo reasons related to the fear of the evil eye (cf. DELG 811-813; NIL 370-383). Moreover, not all alternations indicated by Beekes are clearly developed from the same form, especially when isolated glosses are put together with other more frequently attested forms with disparate meanings. E.g., μάργος “mad” and Hesychius’ glossesGlosses such as μαρικᾶς˙κίναιδος “catamite” and ἄβαρκνα˙λιμός “hunger, famine” (cf. Beekes 2010: 905) may also be etymologically separate. (Cf. also Colvin 2016 for a critical assessment of Beekes’ hypotheses). More study is needed to elucidate the complex substrateSubstrate vocabulary of Ancient Greek.
3.2.Eastern influences on Minoan and MycenaeanMycenaean civilizations
While the linguistic non-PIE influence on Greek vocabulary mainly comes from non-attested MediterraneanMediterranean languages, from the cultural point of view the massive influence of the ancient Near East is undeniable since the beginning of Greek civilization. Without going in details, I limit myself to recall here some of the most salient manifestations of this influence in art and material culture, as they have been reconstructed from archaeological findings and texts. Firstly, when the speakers of the IE variant that would become Greek descended into the Greek peninsula and the surrounding area of the Aegean, at the end of the 3rd millennium (ca. 2100 BC), they encountered a more advanced civilization, the Minoans, centred in Crete, which was at the crossroads of an intense commercial and cultural exchange with other palace societies of the Bronze Age. In Ancient Egypt, Mesopotamia, Anatolia, and the Levant, we also have in this period various centralized organizations where the palace, dominated by an absolute monarch, is the head of political, administrative, and economic powers. The monumental architecture symbolically represents the importance of the palace for the community. Artistic motifs are also shared among these civilizations. The bull, for example, often appears in Minoan paintings, and is the object of the story of the Minotaur and of several ancient Near Eastern myths and cults, such as the Egyptian Apis bull (ḥjpw), the Mesopotamian bull of heaven killed by Gilgamesh, the representation of Marduk as the bull of Utu (the sun god in Sumerian religion), etc.10
The Greeks took control of Crete and of the Minoan sphere of influence but continued this system of palace economy. Their MycenaeanMycenaean civilization (ca. 1600-1200 BC) is pervaded by Minoan cultural motifs. The Linear BLinear B, for example, used to record the Mycenaean language, is a syllabic writing system written on clay tablets similarly to the Linear A used for the Minoan language (although the Linear A is still undeciphered, and the Minoan language, still unknown, is certainly not IE). Different syllabic and cuneiform writing systems are also attested by other Mesopotamian and Anatolian languages of the 2nd millennium (within IE, HittiteHittite, (Cuneiform) Luwian, and Palaic). As the ancient societies of the bronze age had such a close interaction, it is understandable that they also came to a similar end. At the end of the 2nd millennium (from ca. 1250 to 1150 BC), a series of destructions occurred in the whole Eastern Mediterranean. Troy VIIa is destroyed in ca. 1220 BC, the Hittite Empire in ca. 1200 BC. Syria and Palestine experienced periods of instability. Ancient Egypt was attacked – ancient Egyptian sources speak of the “sea people”. The Mycenaean civilization was definitively destroyed in ca. 1050 BC – according to Greek mythologyMyth, this was due to the Dorian invasion and to the return of the Heraclids. The Mycenaean clay tablets were burned in the palaces, and have been preserved precisely because they were baked in these arson attacks. The same occurred e.g., to Hittite clay tablets. A type of Aegean syllabary will be longer in use in Cypriote, a dialect closely related to Mycenaean which has been deeply studied by Egetmeyer (2010).
3.3.Eastern influences on Archaic GreeceArchaic Greek
We do not know much about the subsequent Greek “dark ages”, usually posited between 1100 BC, at the beginning of the iron age, and 750 BC. The crisis of the Mediterranean may have temporarily reduced the contact between the Greeks and the East. This contact, however, resumed at the end of the dark ages with the migrationsMigrations (starting in the second quarter of the 8th century BC and lasting until the end of the 7th century BC), which led to the foundations of Greek emporiaEmporia and coloniesGreek colonies, Greek colonization around the whole MediterraneanMediterranean. Such migrations had different causes – the need for arable land, the search for necessary products for the survival of the community, rivalry between aristocratic families with consequent exile of one of the two parties etc. They were also enabled by previous experiences of the PhoeniciansPhoenicians (from Tyre, Sidon, Byblos, etc.), who in antiquity were considered to be the merchants par excellence. At the beginning of the 10th century BC already, the Phoenicians had settled down in Kition, in southern Cyprus, and in the subsequent centuries they had founded colonies in the Mediterranean up to Tartessos, in Southern Spain (8th century BC). Although Homer does not speak about the Greek migrations, he makes several mentions of the Phoenician people, described as clever but untruthful (e.g., πολυπαίπαλος “exceeding crafty” in Od.15,419; ἀπατήλιαεἰδώς “skilled in wiles” in Od. 14,288, etc.). Subsequent myths represent the Greek migrations as the journeys of Homeric heroes after the Trojan war.
A “colony” (ἀποικία)Greek colonies, Greek colonization and an “emporium” (ἐμπόριον)Emporia have different ties to the founding city. Although the colony is substantially an independent state (e.g., the city does not interfere with the political affairs of the colony), a linguistic and cultural relationship is always maintained between the two. Instead, an emporium has no political independence, it is just a place of exchange around a harbor, where different Greek and foreign groups may interact. The contact with non-Greek people is therefore more intensive in emporia. We can mention the emporium of Al Mina, in Syria (8th-7th century BC), and that of Naucratis in Egypt (7th-6th century BC). When we talk of exchange, we imply that the Greeks were both the receiver and the donor of material products and ideas. As anticipated, however, Near Eastern civilizations (especially Ancient Egypt and Mesopotamia) were more ancient and advanced than Greek civilization, as the Greeks always acknowledged (see below). The direction of cultural influence was therefore more from the Near East to Greece than the other way round at least until the Classical age. The most evident gift received from the East is alphabetic writingAlphabet. It is probably not by chance that the earliest Greek emporiumEmporia, Πιθηκοῦσαι “Ischia” in the Gulf of Naples (founded in 770 BC by Chalcis and Eretria of Euboea and home of a mixed population of Greeks, PhoeniciansPhoenicians, and EtruscansEtruscan) is also the place where we have one of the most ancient documentations of the Greek alphabeticAlphabet writing: Nestor’s cup (Νεστορος … ποτεριον). Since the 8th
