The Pre-Greek substrate (or Pre-Greek substratum) consists of the unknown pre-Indo-European[dubious ] language(s) spoken in prehistoric Greece before the coming of the Proto-Greek language in the Greek peninsula during the Bronze Age. It is possible that Greek acquired approximately 1,000 words from such a language or group of languages, because some of its vocabulary cannot be satisfactorily explained as deriving from Proto-Greek and a Proto-Indo-European reconstruction is almost certainly impossible for such terms.
Based upon toponymic and lexical evidence, it is generally assumed that one or several languages were once spoken in both the Greek peninsula and western Asia Minor before Mycenaean Greek and the attested Anatolian languages (Hittite and Luwian) became predominant in the region. Various explanations for this phenomenon have been given by scholars.
One substrate language, whose influence is observable on Ancient Greek and Anatolian languages, is taken by a number of scholars to be an Indo-European language related to the Anatolian Luwian language, and to be responsible for the widespread place-names ending in -ssa and -nda in western Asia Minor, and -ssos and -nthos in mainland Greece, respectively. For instance, the name of the mount Parnassos in Greece has been interpreted as the Luwian parna- ('house') attached to the possessive suffix -ssa-. Both Hittite and Luwian texts also attest a place-name Parnassa, which could be related. Philologist Martin L. West has proposed to name this unattested Anatolian language "Parnassian", and has argued for "a parallel movement down from Thrace by a branch of the same people as entered Anatolia, the people who were to appear 1,500 years later as the Luwians". From the distribution of the names, it appears that this language was spoken during the Early Helladic II period, which began around 2800 BC.
However, since many clusters of sounds are possible in both pre-Greek and Indo-European, it is difficult in most cases to distinguish between eventual "Parnassian" loanwords and shared pre-Indo-European substrate words. For instance, terms like τολύπη (tolúpē; 'clew, ball of wool ready for spinning') show typical pre-Greek features while being related to Anatolian words (in this case Luwian and Hittite taluppa/i- 'lump, clod') with no other attested Indo-European cognate, suggesting that they were borrowed from the same probably non-Indo-European source. Some scholars have thus proposed that at least part of the pre-Greek substrate was brought to Greece by pre-Indo-European settlers from Asia Minor, and that we should distinguish between different layers of loanwords coming successively or concurrently from different families of languages.
While the correlations between Anatolian and Greek placenames may be a strong indication of a common early phase of Indo-European – possibly Anatolian – influence in the area, some pre-Greek loanwords still remain incompatible with Indo-European phonology while showing certain recurrent patterns that set them apart from other languages. This likely indicates that "one language, or a group of closely related dialects or languages" was the source of another, possibly earlier, layer of pre-Indo-European loanwords in the region. Of the few words of secure Anatolian origin, most are cultural items or commodities which are likely the result of commercial exchange, not of a substratum. Some of the relevant vocabulary can also be explained as linguistic exchange between Greek and Anatolian languages across the Aegean Sea without necessarily originating from a change of language.
Estimates for the introduction of the Proto-Greek language into prehistoric Greece have changed over the course of the 20th century. Since the decipherment of Linear B, searches were made "for earlier breaks in the continuity of the material record that might represent the 'coming of the Greeks'".
The majority of scholars date the coming of Proto-Greek to the transition from Early Helladic II to Early Helladic III (c. 2400−2200/2100 BC), with the diversification into a southern and a northern group beginning by approximately 1700 BCE. This has been criticized by John E. Coleman, who argues that this estimate is based on stratigraphic discontinuities at Lerna that other archaeological excavations in Greece suggested were the product of chronological gaps or separate deposit-sequencing instead of cultural changes. Coleman estimates that the entry of Proto-Greek speakers into the Greek peninsula occurred during the late 4th millennium BC (c. 3200 BC) with pre-Greek spoken by the inhabitants of the Late Neolithic II period.
Although no written texts exist or have been identified as pre-Greek, the phonology and lexicon have been partially reconstructed via the considerable number of words that have been borrowed into Greek; such words often show a type of variation not found in inherited Indo-European Greek terms, and certain recurrent patterns that can be used to identify pre-Greek elements.
The pre-Greek language had a simple vowel system, with either three or five monophthongs. This system consisted of either /a/, /e/, /i/, /o/, /u/, or most likely just /a/, /i/, /u/, in which /a/ varied between /a/~/e/~/o/ as a result of palatalization for /e/ and labialization for /o/.
Additionally, it had at least one diphthong (/au/), and it may also have had /ou/, although this is also often explained as the sequence -arʷ- adapted in Greek as -ουρ-, since /ou/ is often seen with an /r/.
The phonology of pre-Greek likely featured a series of both labialized and palatalized consonants, as indicated by Mycenaean inscriptions in Linear B. These features were found not only in stops, but in resonants as well (presumably including even the rare modified approximants /jʷ/ and /wʲ/), which was different from Indo-European languages at the time and is generally considered a rare feature characteristic of pre-Greek. It is, however, unlikely that voicing or consonantal aspiration were distinctive features, as pre-Greek loanwords in Greek vary freely between plain, voiced and aspirated stops (e.g. ἀσφάραγος/ἀσπάραγος, aspháragos/aspáragos, 'asparagus'). The observation of such variants for a particular word is often a strong indication of substrate-derived etymology.
Furthermore, while the existence of word-initial approximants /w/ and /j/ can be safely inferred from common motifs in inherited words (e.g. the ἰα‑ from *ja- in ἴαμβος, Ἰάσων) or even retained in early and dialectal forms (e.g. *wa- in the cases of ἄναξ-ϝάναξ, Ὀαξός-ϝαξός, ὑάκινθος-ϝάκινθος), word-initial aspiration probably did not exist, with /h/ considered by Beekes a non-native phoneme in pre-Greek.
The initial consonant σ- /s/ or /z/ is very common in pre-Greek and characteristic when it shows up as an s-mobile.
Certain characteristic consonant clusters associated with pre-Greek phonology as reflected in words inherited into Greek, as listed by Beekes according to their frequency in the PIE language:
Not possible in PIE
Rare in PIE
Possible in PIE
There are different categories of words that have been suggested to be pre-Greek (or "Aegean") loanwords such as:
Possible Anatolian or "Parnassian" loanwords include:
Other explanations have been made for these substrate features. Some fringe theories ranging from the mild (e.g., Egyptian) to the extreme (e.g., Proto-Turkic) have been proposed but have been given little to no consideration from the broader academic community and as such are not developed in this article.
The existence of a Minoan (Eteocretan) substratum was the opinion of English archaeologist Arthur Evans who assumed widespread Minoan colonisation of the Aegean, policed by a Minoan thalassocracy.
Raymond A. Brown, after listing a number of words of pre-Greek origin from Crete, suggests a relation between Minoan, Eteocretan, Lemnian (Pelasgian), and Tyrsenian, inventing the name "Aegeo-Asianic" for the proposed language family.
However, many Minoan loanwords found in Mycenaean Greek (e.g., words for architecture, metals and metallurgy, music, use of domestic species, social institutions, weapons, weaving) have been asserted to be the result of socio-cultural and economic interactions between the Minoans and Mycenaeans during the Bronze Age, and may therefore be part of a linguistic adstrate in Greek rather than a substrate.
A Tyrsenian/Etruscan substratum was proposed on the basis of the Lemnos funerary stele: four pottery sherds inscribed in Etruscan that were found in 1885 at Ephestia in Lemnos.
However, the Lemnos funerary stele was written in a form of ancient Etruscan, which suggested that the author had emigrated from Etruria in Italy, rather than the Greek sphere, and the Homeric tradition makes no mention of a Tyrrhenian presence on Lemnos.
If Etruscan was spoken in Greece, it must have been effectively a language isolate, with no significant relationship to or interaction with speakers of pre-Greek or ancient Greek, since, in the words of C. De Simone, there are no Etruscan words that can be "etymologically traced back to a single, common ancestral form with a Greek equivalent".
In 1979, Edzard J. Furnée proposed a theory by which a pre-Greek substrate is associated with the Kartvelian languages.