Language and Race: Theocritus and the
Koine Identity of Ptolemaic Egypt

by George Hinge


1. Race

Admittedly, the word ‘race’ is rather discomforting. Since I have nevertheless chosen to include it into the title of this article, I suspect I have to clarify my motives and my position first. It is radical to claim that any application of the concept of race must be, at the same time, racist. Yet, even if I shall not deny that man differs in physical appearance, the division into definite races presupposes that the accidental difference in phenotype has social implications as well. Therefore, any use of the word outside physical anthropology must be considered suspicious and perilous. In German and Danish, the corresponding words ‘Rasse’ and ‘race’ are in fact used only of physical differences. In English, on the other hand, the word covers at least three different meanings. Besides the continental denotation, it is also used synonymously with ‘species,’ when one speaks about, say, ‘the human race’ or even ‘extraterrestrial races,’ to my mind a dangerous extension of the word. Furthermore it designates a cultural unit sharing a common history, language, culture and origin. One may speak about ‘the Greek race’ for instance. Even though I consider the last-mentioned extension of the word no less dangerous than the first one, it is ethnicity and not physical race to which the title of my article alludes. The juxtaposition of the two words ‘language’ and ‘race’ was simply too catchy to resist.

It need not be said that the theory combining cultural and physical races experienced a well-deserved setback after the monstrous excesses of the Holocaust. Because the grotesque and so to speak gothic construction of the Aryan race was based partially on linguistic arguments, historical linguistics as such has been considered suspicious by post-war scholars, even though hardly any serious Indo-Europeanists would ever claim the existence of Indo-European genes (and only few of them did in fact subscribe to the Aryan hypothesis before the War).[1] Ironically enough, the marginalisation of classical historical linguistics has, at the same time, opened the way for a new genetic approach to the dispersion of language and culture.[2] This is, however, another story. So, the present article does not claim that language is correlated to physical race in any way. It will, however, be argued that language played a major role in the ancient concept of ethnicity; and linguistic differences and similarities were, to a certain extent, shaped by the ethnic divisions and groupings.

2. Linguistic identity in Classical Greece

A celebrated passage of Herodotus’ Histories sums up Greek ethnicity with the following words put into the mouths of the Athenians to a Laconian delegation as an argument of their alliance in the war against the Persians (8.144.2):

There are many reasons why we should not do this, even if wanted to: First and foremost, they have burnt and destroyed the statues and temples of our gods, and we are obliged to revenge them as far as possible rather than conclude a treaty with the offenders. Furthermore, there is the Hellenicity, consisting in the same blood and the same language, the common shrines of gods and cult and the same way of life, which the Athenians should not betray.

Πολλά τε γὰρ καὶ μεγάλα ἐστὶ τὰ διακωλύοντα ταῦτα μὴ ποιέειν, μηδ᾽ ἢν ἐθέλωμεν, πρῶτα μὲν καὶ μέγιστα τῶν θεῶν τὰ ἀγάλματα καὶ τὰ οἰκήματα ἐμπεπρησμένα τε καὶ συγκεχωσμένα, τοῖσι ἡμέας ἀναγκαίως ἔχει τιμωρέειν ἐς τὰ μέγιστα μᾶλλον ἤ περ ὁμολογέειν τοῖσι ταῦτα ἐργασαμένοισι· αὖτις δὲ τὸ Ἑλληνικόν ἐὸν ὅμαιμόν τε καὶ ὁμόγλωσσον καὶ θεῶν ἱδρύματά τε κοινὰ καὶ θυσίαι ἤθεά τε ὁμότροπα, τῶν προδότας γενέσθαι Ἀθηναίους οὐκ ἂν εὖ ἔχοι.

Accordingly, ethnicity can be defined in terms of four components: origin, language, cult, and culture. The last three of these components are in fact the main criteria for ethnicity in the ethnographical parts of Herodotus’ Histories. Thus, in the Scythian logos, different tribes are dissociated from the Scythians on the basis of their language, cult, and culture (cf. 4.23.2, 4.106, 4.108.2, 4.109, 4.117), and it is evident that Scythian language was to Herodotus an important prerequisite of Scythian ethnicity.[3] In the first book, Herodotus offers a short excursus on the origin of the Greek tribes (1.56-8). We are told that the Hellenic tribe (τὸ Ἑλληνικὸν ἔθνος) originally lived in Central and Northern Greece but later invaded the Peloponnese and was by then called Doric. The Athenian tribe (τὸ Ἀττικὸν ἔθνος), on the other hand, was autochthonous and originally Pelasgic. Herodotus concludes that, since the contemporary Pelasgians, i.e. those living in the northern Aegean, speak a non-Greek language (βάρβαρον γλῶσσαν), the Athenians were originally Barbarians, but after they went over to the Greeks (ἅμα τῇ μεταβολῇ τῇ ἐς Ἕλληνας), they learnt the Greek language as well. The historicity of this reconstruction is questionable.[4] What matters here is the line of argumentation. It supports once again that language is quintessential to Herodotus’ concept of ethnicity.

This fact may seem uncontroversial, and it has been so for some centuries. However, Jonathan Hall, who has written two significant books[5] about the construction of Greek ethnicity, claims that language played only a minor role. Instead the Greeks formulated their ethnicity in terms of fictitious genetic relations, mythical genealogies. Significantly enough, Hall considers Herodotus a possible exception to this rule, but he nevertheless underscores the role of language here too. Thus he says:[6]

... it is clear that the ‘Hellenic tongue’ (Hellas glossa) is not based on any empirically-derived observations referring to the isoglosses that separate the myriad of local Greek dialects from the non-Greek languages. It is rather an abstract reification that assumes the prior existence of an ‘imagined community’ defined according to other criteria.

Since no language can be separated from closely related languages 100 % empirically, one must say that Hall makes an unfair demand on Herodotus and the Greeks. Yet, even if it is impossible to draw an unambiguous line between language and dialect, there can be no doubt that Classical Greece encompassed a linguistic community which included a range of mutually intelligible dialects. There may have existed border cases. The Macedonian dialect is probably a good example of that and an example invoked by Hall. Both the ancients and the modern scholars disagree as to the ethnic affiliation of the Macedonians. The birth of the Slavic Republic of Macedonia in the 1990s has made the Hellenicity of the ancient Macedonians a matter of high political and scholarly priority. Thus, many scholars now conclude that the ancient Macedonians were in fact some kind of Dorians.[7] However, the glosses of the ancient lexicographers give the impression of a language that was closely related to Greek, but different in some fundamental aspects,[8] and the Macedonian dialect seems to have been unintelligible to the Greeks (cf. Curt. 6.11.4). To my mind the linguistic ambiguity of the Macedonian dialect is not only parallel, but also correlated to the ethnic ambiguity of the Macedonian people. At any rate, the Classical Greeks probably had an intuitive approach to the dialect division. They may not have known exactly where to put the finger, but they were certainly able to tell the provenience of a fellow-Greek by his accent alone.

Even if the concept of isoglosses (characteristic features dividing dialects of a language) is a complex issue, of which the ancients were perfectly unaware – as are most modern people –, they did play a role in the linguistic and social reality of the Greek world. Isoglosses may be described as fossils of both previous and present ethnic differences. If modern linguists detect a bundle of isoglosses between two areas, it is a clear sign that the two areas differ, or once differed, in their ethnic identities. Conversely, if the isoglosses are blurred through new common developments, it indicates that the areas in question have started to construct convergent identities. In the homeland it mattered what sort of Greek you were – a Spartan, an Argive, an Athenian etc. However, the great waves of colonisation in the eighth and seventh centuries BCE contributed to the construction of a Greek identity in opposition to the non-Greek natives in Cyprus, Egypt, Libya, Sicily, Italy or Scythia. The otherness of those “Barbarians” and the complete unintelligibility of their languages, which were frequently compared to the chirping of birds, made the existence of a specific Hellenic identity obvious. For instance, Sophocles contrasts Greece with the languageless people (Trach. 1060 οὔθ᾽ Ἑλλάς, οὔτ᾽ ἄγλωσσος).

3. Koine and Panhellenisation

It cannot, however, be stressed often enough that this identity is not natural per se, but a cultural construction. The linguistic, cultural and religious similarity of the future Greeks, which already existed as a consequence of the cultural convergence in the Mycenaean Age,[9] was an impetus to this construction. At the same time, the invention of a Hellenic identity that emphasised these parallels would naturally minimise the differences and maximise the similarities. The process that eventually led to the creation of a Koine is a clear evidence of the importance of language in Greek ethnicity. It will therefore not be superfluous to look closer at this process, and as we shall see, this path will in the end lead us to Alexandria.

The Finnish scholar Jaakko Frösén has suggested that Koine originated through a creolisation of a pidgin that was used as a means of communication between heterogeneous Greek dialects in the fifth century BCE.[10] There is, however, a fundamental difference between pidgin and koine. A pidgin bridges a language barrier between people speaking completely different languages. The grammar and the phonology are reduced to an absolute minimum. If the pidgin becomes a permanent language, it will eventually develop a complete grammar and phonology, a process known as creolisation.[11] The Greek dialects were, however, anything but mutually unintelligible in the Archaic and Classical Ages. The grammar and the vocabulary were largely identical, and nothing indicates that a Greek from one city would not readily understand a Greek from another city – on the contrary.

The Koineisation is therefore better described as a gradual convergence of related linguistic norms towards an imagined superregional standard (“dialect levelling”).[12] In the Archaic Age, the dialects spoken in the southern Balkan and in the western Asia Minor started to orientate themselves towards each other. The concept of a Hellenic ethnicity was introduced, and one of its main pillars was the realisation that the Hellenes shared a common language, i.e. they were able to understand each other even if their various idioms were different. We are dealing with the concept of Panhellenisation. Local deities were integrated into one and the same Olympian pantheon, and the hymns of the local cults were formulated in a new common linguistic and poetic form. As I have pointed out in my study of the language of Alcman[13], poetry composed in the Archaic Age was basically formulated in the same language (deep structure), even if the actual poems were performed with different accents (surface structure). Panhellenisation was relevant not only to the literary language, but also to the local dialects, which were now conceived not as independent entities, but as epichoric units of a Panhellenic whole. The Laconian, Boeotian and Athenian dialects were no longer autonomous systems, but parts of a linguistic collective. The dialects, which had been drifting apart in the Dark Ages, now started to converge. Since the Archaic Koineisation or Panhellenisation was primarily poetic and oral, it tended to ignore the phonetic differences of the dialects and worked in the phonological deep structure of the language. In the Classical Age, with the emergence of a written prose culture, people started to be more aware about the dialectal peculiarities of the surface (pronunciation and orthography), and the Koineisation took another direction, as it is evident from the fact that Ionic and Attic prose was written by men speaking other dialects.

The Hellenistic Age led to an intensification of the Koineisation. Greeks coming from different regions of the Greek world were integrated into new societies together with non-Greeks. An important example is the city of Alexandria. Let us therefore move our focus to the capital of Ptolemaic Egypt. The life of the city is nowhere put so vividly into scene as in Theocritus, writing at the court of Ptolemy II Philadelphus ca. 275 BCE. The most interesting piece in our context is the Fifteenth Idyll, where a linguistic and ethnic consciousness is formulated dissociating itself from the Doric ethnicity of Syracuse on the one hand and from the Egyptian ethnicity on the other.

The main characters of Theocritus’ Fifteenth Idyll are the ordinary housewives Praxinoa and Gorgo, who decide to go up to the palace of Ptolemy to see a festival of Adonis. As soon as they step out on the street, they meet the mob of Alexandria. The street is one chaos, and as they approach the palace, the situation gets even worse. They have to fight themselves through to the gate. People are jostling and trampling on each other’s feet. Finally they succeed in entering the palace and gaze at the extraordinary tapestries. All the way the two women have been chatting, and at last a man interrupts them and asks them to shut up; he is fed up with their broad accent (v. 88). He uses the uncommon verb πλατειάσδοισαι, i.e. “broadening.” In his work On the genres, Hermogenes comments that Theocritus refers to the Doric habit of pronouncing an alpha all the time; the same interpretation is given in the scholia as well.[14] There can be no doubt that the retention of the long ᾱ, where Attic-Ionic and Koine had a long η, was considered one of the most remarkable linguistic features in the speech of a Dorian.

Commenting on this very passage, Jonathan Hall[15] objects that the preservation of the long ᾱ was by no way restricted to Doric, which is certainly true since all dialects outside the Attic-Ionic branch shared this archaism. He concludes that the Greeks were not able to distinguish the single dialects sufficiently because they could not survey the complicated web of isoglosses which made up the Greek dialectological landscape; they simply described a dialect as Doric because it belonged to a city whose population was Doric according to a traditional mythological reconstruction and not on the basis of a linguistic analysis of the dialect in question. His criticism is, however, totally unfair. It reminds me of Socrates, who claims that a horse dealer knows nothing about horses just because he cannot define a horse scientifically. Just as one can have a practical approach to horses, there can, in my eyes, be no doubt that an average Greek was able to tell if a person came from, say, Sparta or Boeotia from his accent alone. There was a whole range of diagnostic features that made the identification obvious to a native Greek, even if he was not able to draw a dialectological map of these isoglosses or even verbalise his intuitive judgment. Similarly, every Dane can tell if a person comes from the islands of Bornholm or Funen, even if he is perfectly unaware of concepts like palatalisation or glottal stop.

4. The dialect of Theocritus’ idylls

Theocritus’ idylls are for the most part written in some kind of Doric themselves, even though it has remained a matter of debate which kind of Doric it is. Since Theocritus originally comes from Syracuse, it has been suggested that the Syracusan dialect was his primary model. However, the Doric idylls present some peculiarities which are not compatible with what we know about the Syracusan dialect, the primary sources of which are the fragmentary comedies of Epicharmus and Sophron. Therefore, the Dutch linguist Cornelis Ruijgh has conjectured that Theocritus attempted to imitate the dialect of the common Alexandrian population.[16] It is true that no extant epigraphic or papyrological sources support that Doric was ever spoken in Alexandria. However, since a large part of the population may have come from the neighbouring colony of Cyrene, as we know at least in the prominent case of Callimachus, Ruijgh considered it highly probable that the Cyrenaean Doric dialect was prevalent in Alexandria in the early Ptolemaic era.

Thus, Theocritus has in the feminine of the participle the ending -οισα instead of -ουσα, which is known in the Doric dialects only in Cyrenaean. In Theocritus’ Doric idylls, a regular ζ of Koine is regularly replaced by σδ. Ruijgh speculates that this pronunciation was retained in the Cyrenaean dialect. The inscriptions write ζ, but it cannot be ruled out that it was in fact pronounced [zd]. Ruijgh assumes that in the Fifteenth Idyll, σδ is written when the speaker uses the Alexandrian Cyrenaean dialect, whereas the text has ζ when the ladies switch back to their Syracusan dialect. Yet, considering the unstable transmission of the Theocritus corpus, it is extremely difficult to draw conclusions about the presence of a specific pronunciation in a specific word; the papyri and the manuscripts are very often at variance. In the oldest papyrus containing examples of this phoneme, P.Oxy. 2064, from the second century CE, ζ is written in nine instances, and we find no examples of σδ. The papyrus may have been anomalous, but the Theocritean corpus shows a tendency towards am increasingly Doric surface in the paradosis.

Both features are known from the transmitted fragments of the Archaic poet Alcman as well.[17] Ruijgh explains this agreement with reference to the theory of Ernst Risch[18] according to which the Alexandrian scholars have altered the text of Alcman to fit the model of the neighbouring Cyrenaean dialect, which was in their eyes the ideal form of Doric. However, this assumption is not only improbable a priori since the Alexandrian philologists would definitely have distinguished Old Laconian and contemporary Cyrenaean. It is also invalidated by the fact that the participle in -οισα is common not only in Alcman, but in practically all lyric poetry of the Archaic and Classical Ages. The manuscripts and papyri of Pindar usually have -οισα, and it is also relatively frequent in Archaic and Classical inscriptions, which were certainly free from Alexandrine alterations.[19] Since the puzzling digraph σδ is not attested in any epigraphic sources before the Roman Age, its presence in Alcman, Sappho, Alcaeus and Theocritus must be explained by a convention of the Alexandrian philologists. I am convinced that it is not what it looks like, i.e. a testimony for the hypothetical pronunciation [zd], but I shall not lay forth my arguments here.

The vocalism of Theocritus’ Doric idylls is particularly complicated: As a rule one can say that the secondary long ē (originating from compensatory lengthening or contraction) is written ει in Theocritus as it is in Attic-Ionic and Koine, whereas the secondary long ō, which Attic-Ionic and Koine represents with ου, normally appears as ω in Theocritus. The Syracusan dialect has ει and ου, whereas the Cyrenaean dialect has η and ω. In other words, Theocritus’ dialect cannot be reconciled with either Syracusan or Cyrenaean without further ado. Ruijgh assumes a partial influence from Koine, which is rather unsatisfying. Another Dutch scholar, Jelle Abbenes, has suggested that Theocritus’ choice of vocalism is influenced by the text of Alcman.[20] His examples from the Alcman corpus are, however, incomplete, and his conclusions therefore inaccurate.[21]

Teresa Molinos Tejada, who has written a thorough dissertation on the language of Theocritus, assumes that the Fifteenth Idyll was originally composed in Syracusan.[22] She bases her conclusion on the fact that the cases of irregular ου accumulate exactly in this idyll. The word ‘acumulación’ is, I must say, an exaggeration. The primary sources of the secondary long ō are found in the endings of the genitive singular and the accusative plural of the 2nd declension. The most complete papyrus fragment of the Fifteenth Idyll, the Antinoë Papyrus from ca. 500 CE, has twenty examples of the genitive singular and three examples of the accusative plural; respectively three and one of these are written with ου:





PHamb. 201 (1st cent. AD)





POxy. 1618 (5th cent. AD)





PParVind. (5th cent. AD)





PAnt. (ca. 500 AD)





In other words, the idea that certain poems or parts of poems imitate the habits of particular Doric dialects cannot be supported by the evidence. The Fifteenth Idyll is written neither in Syracusan nor in some hypothetical Ptolemaic Cyrenaean, but in an artificial colloquial Doric which combines linguistic features from different traditions. It may of course seem ironic that the man, who criticises the chatting Syracusan women, speaks the same Doric dialect as they do: He uses the “flat” Doric ᾱ (δύστανοι, ἀνάνυτοι) and the feminine of the participle in -οισα (κωτίλλοισαι, πλατειάσδοισαι). On the other hand, the dialect serves not as a parody of the persons, but lends a certain ethos to the scene as such. The linguistic and metric form marks the text as a hybrid of mime, lyric and epos.[23]

5. Linguistic prejudices in Ptolemaic and Roman Egypt

Even if we cannot use the Fifteenth Idyll as a testimony to the spoken language of the streets of Ptolemaic Alexandria, it is an important document of the attitudes towards language. The city is indeed pictured as a cultural melting-pot. The streets are crowded to the bursting point with both native Egyptians and Greeks from diverse parts of the world. In her book on the so-called “urban” mimes of Theocritus, Joan Burton argues that the Second, Fourteenth and Fifteenth Idylls describe the new cosmopolitan identity of Alexandria and at the same time contribute to the construction of that identity.[24]

In the Fifteenth Idyll, the journey of the two Syracusan ladies from their homes through the crowded streets to the royal palace symbolises the integration of the citizens into a new urban identity under the auspices of Ptolemy. The climax of the poem, and the goal of their journey, is a religious festival sponsored by Queen Arsinoë. It is tempting to analyse the development of the poem in terms of a rite of passage:

  • Rite of separation: the domestic scene (vv. 1-43).

  • Rite of liminality: the crowded streets, the urban jungle (vv. 44-77)

  • Rite of aggregation: the entrance into the palace and Queen Arsinoë’s Adonis chorus (vv. 78-149).

Through this rite of passage, the women change their identity, from Sicilian immigrants to Alexandrine citizens, from bumpkins to cosmopolites. As soon as they enter the palace, a stranger reproaches them with their broad Doric accent. In other words, the new identity as an Alexandrian cosmopolitan required the abandonment of the old epichoric dialects and the adoption of Koine since in the multiethnic milieu of Alexandria, the old local affinities have lost their meaning. Praxinoa argues that they are Syracusan women and therefore Corinthians. Consequently, they speak Peloponnesian, and she adds, “Dorians are in my opinion allowed to speak Doric” (vv. 92-93 Πελοποννασιστὶ λαλεῦμες, Δωρίσδειν δ᾽ ἔξεστι, δοκῶ, τοῖς Δωριέεσσι). She is not ready to give up the old linguistic discourse just like that. The anachronism of Praxinoa’s Doric identity is, however, emphasised by her invocation of the mythical figure Bellerophon (vv. 91-92 ὡς εἰδῇς καὶ τοῦτο, Κορίνθιαι εἰμὲς ἄνωθεν, ὡς καὶ ὁ Βελλεροφῶν). Furthermore, Praxinoa is pictured, though with great empathy, as a ridiculous narrow-minded housewife. She is unreasonable to her slave girl Eunoa and negligent of her infant boy Zopyrion. Her perseverance in the Doric dialect is yet another example of her stubbornness.

In modern western societies, underclass and middleclass women generally speak a variant which is closer to the standard than the variant spoken by their male peers, whereas Praxinoa and Gorgo allegedly spoke a more regional variant.[25] The difference may be explained by different gender roles in modern times and in antiquity: When modern working class women speak a middleclass dialect, it is an indication of their higher social and linguistic ambitions. In Ptolemaic Alexandria, women were apparently less ambitious, or that is at least the impression that Theocritus wants to give us.

Praxinoa is also a racist. When on the street, she cries out (vv. 44-50, transl. Gow 1950):

Heavens, what a crowd! How and when are we to get through this plague? They’re like ants - there’s no numbering or counting them. You’ve done us many a good turn, Ptolemy, since your father was in heaven. Nowadays no ruffian slips up to you in the street Egyptian-fashion and does you a mischief - the trick those packets of rascality used to play, one as bad as another with their nasty tricks, a cursed lot.

Andrew Gow elucidates in his commentary, with a somewhat racist malice, “the Egyptians had by long tradition a bad reputation in such matters.”[26] Praxinoa’s unjust prejudice may in fact have been shared by many Greeks in Alexandria. As pointed out earlier in this article, Greek ethnicity was after all to a great extent formed in direct opposition to the non-Greeks in the colonies.

The question is, however, how one qualified to be a Greek. Fluency in Greek was without doubt a prerequisite. Not being able to speak Greek was equivalent to being a barbarian. Herodotus states that the Egyptians made a similar distinction; they labelled all that did not speak their language as ‘barbarians’ (2.158.5 βαρβάρους δὲ πάντας οἱ Αἰγύπτιοι καλέουσι τοὺς μὴ σφίσι ὁμογλώσσους). On the other hand, it need not be an unchangeable status. We have already seen that Herodotus derived the Athenians from the Pelasgians, who spoke a non-Greek language; so, a barbarian might become a Hellene under certain circumstances. Dorothy Thompson has demonstrated that a targeted language policy was exacted in Ptolemaic and Roman Egypt.[27] Thus, not only were Greeks and Persians excepted from the obol tax, but teachers of Greek were also exempted for the heavier salt tax. She concludes that in Egypt those were Hellenes who had learnt Greek language at the gymnasium; from these Hellenes the corpse of bureaucrats and officers was recruited. This conclusion is of course extremely attractive, not only because it would support the importance of language in Greek ethnicity, but also because it would exonerate Ptolemaic Egypt of the charge of being an apartheid society. The assumption is, however, anything but certain. Herbert Youtie has argued that literacy was not demanded of a person holding a higher administrative post.[28] On the other hand, fluency in Greek is not the same as being able to read and write it fluently. At any rate, the considerable amount of documents in Greek speaks for itself. Even though the majority of the population in Upper Egypt remained Egyptian-speaking, the Greek language gained more and more importance throughout antiquity, whereas Demotic suffered a gradual loss of domain.

A widespread bilingualism would show in two ways: Egyptian interference in Greek and Greek interference in Egyptian. The Greek interference in Egyptian is minimal in the Demotic documents, but in the Coptic language it is rather impressive. Not only is the alphabet Greek (with some additions), but there is also a considerable amount of Greek loan words. That these loans are not learned, but parts of the colloquial language can be demonstrated by two examples: 1) Greek θάλασσα was borrowed into Coptic as thalassa, but subsequently, the aspirate was conceived as the definite article, t + an initial h; accordingly, the indefinite form is written halassa.[29] Such false word divisions are commonplace in the history of language, but it is at the same time a clear proof of the fact that the Greek loanwords did not belong to the learned and literate layer of the language. thalassa must have become an integrated part of the spoken language before it was analysed as halassa. 2) It is quite astonishing that Coptic has borrowed a large amount of common Greek conjunctions and particles such as alla, gar, de, kan, oute, hina, hos and many more. What is more, in the Coptic translations of the New Testament, these conjunctions are used not only where they are found in the Greek original. They are also used as translations of other Greek conjunctions. alla translates not only ἀλλά, but also δέ and καί; gar corresponds not only to γάρ, but occasionally also to ὅτι, etc. In other words, the Greek particles were not only leftovers from the translation, but part of the Coptic language as well.

The Egyptian interference in Greek is harder to detect. The documentary papyri of Egypt show several orthographical aberrancies from the norm of Koine. Frequently the voiceless stops π, τ, κ are written instead of the voiced stops β, δ, γ and vice versa. Francis Gignac attributes this fluctuation to a Coptic substratum.[30] In Coptic the distinction between voiced and voiceless stops is phonologically irrelevant (like in modern Danish, where the opposition p, t, k ~ b, d, g is a matter of aspiration, not voice). On the other hand, as Sven-Tage Teodorsson points out, most of the orthographic aberrancies are in accordance with the general development of the Greek language; the insecurity in the representation of the voiced and voiceless stops may just as well be a consequence of the reorganisation of the Greek phonological system, according to which the voiced stops were developing into fricatives, i.e. [β], [ð], [γ].[31] The substratum theory was introduced in the 19th century by the Italian scholar Graziadio Isaia Ascoli[32], who attributed the particularities of the different Romance dialects to the influence of pre-Latin, primarily Celtic, languages. Accordingly, the French [y] for Latin ū was allegedly caused by a pronunciation habit of the ancient Gauls.[33] Even though it is incontestable that one can never learn to speak a foreign language without an accent, it is questionable if such lack of competence would be hereditary as a matter of course. We have still not left the matter of the present article, the interdependence of language and race: The substratum theory assumes, overtly or tacitly, that language is a consequence of race; for some reason, the Celtic or the Egyptian “race” cannot pronounce a closed back vowel or a voiced stop properly, and this defect remains even after they have become fluent in their new languages.

On the other hand, one shall not disregard the importance of language as a marker of identity. When speaking a second language, one of course makes errors; the inaccuracies of my own English pronunciation are not a deliberate choice, but an inescapable condition. However, to become standard in a language community, such deviations must have become markers of a desirable identity.[34] There can be no doubt that the persons that wrote the Greek papyri aimed at a correct Koine. As a matter of fact, the spelling errors are an indication of a high competence in Greek rather than the opposite: They demonstrate that the writers have not learnt fixed formulas by heart, but mastered the language orally; they have not learnt Greek for writing purposes only, but must have practiced it in their everyday life. I therefore tend to agree with Teodorsson, who attributes the orthographical deviations to an inner-Greek development, originating in Classical Attic; the Egyptians that identified with the Greek Koine took over the substantial speech habits of the ethnic Greeks as well.

6. Conclusion

Language and race are both cultural constructions. They are expressions of man’s attempt to organise the world around him in a meaningful way. Thus, if we establish that a Laconian and an Argive speak Doric, it is equivalent to saying that they have orientated themselves in the same direction through the history of the dialects. They have shared an identity at least periodically. When the local dialects are replaced by Koine, it is not just a practical way of bridging a language barrier, but the symptom of a new identity, and not only a symptom, but also a most powerful contribution to that identity. In the Classical Age, inner-Greek ethnic affiliation was formulated mythologically in terms of genealogy and expressed linguistically through different manners of speech, regional dialects. In the Hellenistic Age, Greek identity was first of all a matter of paideia, to which the Museum and the Gymnasium of Alexandria were invaluable contributions. The timeless and superregional Koine was the natural vehicle of this common Hellenic identity. It is significant that the disappearance of the spoken epichoric dialects coincides with a growing interest in the literature transmitted in those dialects (Alcman, Sappho, Alcaeus) and the construction of a new literary pseudo-colloquial Doric dialect. It is no coincidence that this antiquarian interest in the old dialects is centred exactly in Ptolemaic Alexandria, where the cosmopolitan milieu deprived the classical linguistic identities of their meaning. Theocritus’ Fifteenth Idyll bears testimony to that development both in its linguistic form and its discourse.


[1] For an exhaustive discussion of the evidence regarding the phenotype of Early Indo-Europeans and its significance for reconstructing the prehistory of Indo-European see Day 2001.

[2] Cavalli-Sforza 2000, 133-72, points to a correspondence between genetic cladistics and linguistic cladistics. Undeniably, people who speak similar languages are often similar genetically as well; yet, at closer inspection there are many important exceptions (besides the well-known modern ones), for instance the Ethiopians who speak a language related to Arabic and Hebrew in spite of the large “racial” difference. Genetically, Iranians are closely related to the other people of the Middle East, but they speak an Indo-European language. The examples are legio.

[3] Hinge 2005.

[4] The alleged autochthony of the Athenians implies that Hellenicity is not something constant and natural. Cf. also Hnge 2005, 107-15.

[5] Hall 1997; 2002.

[6] Hall 2002, 191-2.

[7] E.g. Brixhe 1999; Kapetanopoulos 1999; O’Neil 2006; Panayotou 2007. Their main evidence is two inscriptions in the Doric dialect found in Macedonia, but they prove only that Greek was gaining ground in Macedonia in the Classical age already (later, under Philip II, the inscriptions were written in pure Attic Greek).

[8] The Macedonian personal names and the glosses of Hesychius exhibit IE *bh > β instead of Gk. φ, e.g. Βερενίκη = Φερενίκη. Here, Brixhe (1999, 51-62) assumes a retrograde development *bh > Gk. φ > Mac. β. However, it is to my mind more economical to relate this feature to the unanimous b of the neighbouring Indo-European dialects of Central and Eastern Europe (Phrygian, Thracian, Illyrian, Celtic, Balto-Slavic, Iranian etc.).

[9] Finkelberg 2005, 109-39.

[10] Frösén 1974.

[11] Thomason 2001, 157-95; Mufwene 2008 downplays the difference between creoles and “naturaæ languages”.

[12] Cf., for the Hellenistic and Roman periods, Bubeník 1989; 1993.

[13] Hinge 2006; cf. also Hinge 2009.

[14] Hermog., Id. 1.6: ὁ γὰρ Θεόκριτος ἀχθόμενόν τινα πεποίηκε δωριζούσαις γυναιξὶ διὰ τὸ πλατύνειν τῷ α τὰ πλεῖστα χρωμέναις τὴν φωνήν; Sch. Theocr. 15.87-88 τὸ γὰρ «πλατυάσδοισαι» τοιοῦτόν ἐστιν, ὅτι πλατυστομοῦσιν οἱ Δωριεῖς τὸ α πλεονάζοντες.

[15] Hall 1997, 176.

[16] Ruijgh 1984.

[17] Hinge 2006, 43-6, 91-9.

[18] Risch 1954.

[19] CEG 352 ε]ὐ̣μένοισα (Corinth, seventh cent. BCE), CEG 114 ]οισα (Boeotia, 479 BCE?); Page 1962, nos. 938(e) Μοῖσα (c. 480 BCE), 938(c) ἄγοισα (c. 450 BCE), Wachter 2001, nos. COR 36 Μοῖσαι (Corinth, sixth cent. BC), COR 96 Πνατομέδοισα (Corinth, sixth cent. BCE).

[20] Abbenes 1996.

[21] Abbenes claims that Alcman has no examples of η before original *rw; there are however, two possible cases: 3 fr. 11 πήρα[ (= πέρας, πεῖραρ) and 7 fr. 1.14 ἀπήρ[ι]τον (= ἀπείριτος). Cf. Hinge 2006, 25-30.

[22] Molinos Tejada 1990, 73, 377-8.

[23] Hunter 1996. He supposes that Theocritus does in fact imitate Alcman in the Eighteenth Idyll, but it is a generic mimesis (pp. 153-5).

[24] Burton 1995.

[25] Willi 2003, 157-07, surveys the concept of “female speech” in Classical Athens. For a critical discussion of language and gender see Eckert & McConnell-Ginet 2003, 266-304.

[26] Gow 1950.

[27] Thompson 1992; 1994.

[28] Youtie 1971.

[29] Rahlfs 1900.

[30] Gignac 1970.

[31] Teodorsson 1977. Bubeník 1989, 221-5, insists on “bilingual interference from the Coptic substrate”.

[32] Ascoli 1881.

[33] Ascoli argues that [y] for Latin ū occurs in French, Occitan and North Italian, i.e. in the areas inhabited by Celts in Antiquity. However, there is no proof that the kind of Celtic spoken in Gaul had the pronunciation habit. On the contrary, Gaulish ū is rendered with ου in Greek sources, even though Classical Greek did have the vowel ȳ (e.g. Λούγδουνον = Lyon). Welsh has [i:] for Celtic *ū, presumably by the way of *ȳ. However, in the Latin words borrowed into Welsh, ū is usually retained (e.g. mūrus > mwr); in other words, the substratum influence did not work in the only dialect in which we have positive evidence for ū = [y:].

[34] Mufwene 2008 describes contact changes and creoles with reference to the societal (“ecological”) function of the varieties in question.


Abbenes, J.G.J. 1996 “The Doric of Theocritus: a literary language” in: A. Harder, R.F. Regtuit & G.C. Wakker (eds.) Theocritus Groningen: Egbert Forsten, 1-17.

Ascoli, G.I. 1881 “Una lettera glottologica’ Rivista di filologia classica 10, 1-79 (German excerpt in R. Kontzi (ed.) Substrate und Superstrate in den romanischen Sprachen (Wege der Forschung 165) Darmstadt: Wissenschaftliche Buchgesellschaft 1982, 29-54).

Brixhe, C. 1999 “Un «nouveau» champ de la dialectologie grecque: le macédonien” in: A.C. Cassio (ed.) Katà diálekton Napoli: Istituto Universitario Orientale di Napoli, 41-71.

Bibeník, V. 1989 Hellenistc and Roman Greece as a Sociolinguistic Area Amsterdam-Philadelphia: John Benjamins.

Bubeník, V. 1993 “Dialect contact and koineization: the case of Hellenistic Greek” International Journal of the Sociology of Language 99, 9-23.

Burton, J.B. 1995 Theocritus’ Urban Mimes: mobility, gender, and patronage Berkeley London: University of California Press.

Cavalli-Sforza, L.L. 2000 Genes, Peoples and Languages New York: North Point Press.

Day, J.V. 2001 Indo-European Origins: The Anthropological Evidence Washington, DC: The Institute for the Study of Man.

Eckert, P. & McConnell-Ginet, S. 2003 Language and Gender Cambridge: Cambridge University Press.

Finkelberg, M. 2005 Greeks and Pre-Greeks Cambridge: Cambridge University Press.

Frösén, J. 1974 Prolegomena to a study of the Greek language in the first centuries A.D. Helsinki.

Gignac, F.T. 1970 “The pronunciation of Greek stops in the papyri” Transactions and Proceedings of the American Philological Association 101, 185-202.

Gow, A.S.F. 1950 Theocritus Cambridge: Cambridge University Press.

Hall, J.M. 1997 Ethnic Identity in Greek Antiquity Cambridge: Cambridge University Press.

Hall, J.M. 2002 Hellenicity. Between Ethnicity and Culture Chicago / London: The University of Chicago Press.

Hinge, G. 2005 “Herodot zur skythischen Sprache” Glotta 81, 86-115.

Hinge, G. 2006 Die Sprache Alkmans: Textgeschichte und Sprachgeschichte (Serta Graeca 24) Wiesbaden: Dr. Ludwig Reichert Verlag.

Hinge, G. 2009 “Cultic Persona and the Transmission of the Partheneions” in: J. Jensen, G. Hinge, P. Schultz & B. Wickkiser, Aspects of Ancient Greek Cult (Aarhus Studies in Mediterranean Antiquity 8) Aarhus: Aarhus University Press, 213-34.

Hunter, R.L. 1996 Theocritus and the Archaeology of Greek Poetry Cambridge: Cambridge University Press.

Kapetanopoulos, E. 1999 “Alexander’s patrius sermo in the Philotas affair” The ancient world 30, 117-28.

Molinos Tejada, T. 1990 Los dorismos del Corpus bucolicorum Amsterdam: Adolf M. Hakkert.

Mufwene, S.S. 2008 Language Evolution: Contact, Competition and Change London-New York: Continuum.

O’Neil, J.L. 2006 “Doric forms in Macedonian forms in Macedonian inscriptions” Glotta 82, 192-209.

Page, D.L. 1962 Poetae melici Graeci Oxford: The Clarendon Press.

Panayotou, A. 2007 “The position of the Macedonian dialect” in: A.-F. Christidis (ed.): A History of Ancient Greek: From the Beginnings to Late Antiquity Cambridge: Cambridge University Press, 433-44.

Rahlfs, A. 1900 “ θαλαccα im Koptischen” Zeitschrift für ägyptische Sprache und Altertumskunde 38, 152-3.

Risch, E. 1954 “Die Sprache Alkmans” Museum Helveticum 11, 20-37.

Ruijgh, C.J. 1984 “Le dorien de Théocrite: dialecte cyrénien d’Alexandrie et d’Égypte” Mnemosyne 37, 56-88.

Teodorsson, S.-T. 1977 The phonology of Ptolemaic Koine Göteborg: Acta Universitatis Gothoburgensis.

Thomason, S.G. 2001 Language Contact. An Introduction Edinburgh: Edinburgh University Press.

Thompson, D.J. 1992 “Language and literacy and early Hellenistic Egypt” in: P. Bilde, T. Engberg-Pedersen, L. Hannestad & J. Zahle (eds.) Ethnicity in Hellenistic Egypt Aarhus: Aarhus university Press, 39-52.

Thompson, D.J. 1994 “Literacy and power in Ptolemaic Egypt” in: A.K. Bowman & G. Woolf (eds.), Literacy and Power in the Ancient World Cambridge: Cambridge University Press, 67-83.

Willi, A. 2003 The Languages of Aristophanes Oxford: Oxford University Press.

Wachter, R. 2001 Non-Attic Greek Vase Inscriptions Oxford: Oxford University Press.

Youtie, H.C. 1971 “Ἀγράμματος: an aspect of Greek society in Egypt” Harvard Studies in Classical Philology 75, 161-76.