Matthew J. C. Scarborough
University of Copenhagen
The DOI to cite for this paper is https://doi.org/10.71160/SZAT4356
§66. Introduction: The problem of Classical Greek dialect geography
§66.1. A longstanding problem of Ancient Greek dialectology is the geographical distribution of the dialects as they are first attested in the Aegean in the Classical period (Figure 1). Since the ground-breaking studies of the dialect geography by Porzig (1954) and Risch (1955), most mainstream accounts of the dialects tend to admit the existence of four main subgroupings consisting of Attic-Ionic, Arcado-Cypriot, Northwest Greek / Doric, and Aeolic. In Porzig and Risch’s analyses these four groups may further fit more broadly into West Greek (Risch: North Greek) and East Greek (Risch: South Greek), although the exact alignment remains a perennial topic of debate.[14] A peculiar feature of Ancient Greek dialect geography is that it is characterized by two major dialect continua, namely those of Northwest Greek / Doric and Attic-Ionic, accompanied by four major dialect enclaves represented by Boeotian, Thessalian, Lesbian, and Arcadian.[15]
Figure 1: The Four Major Dialectal subgroupings of Ancient Greek in the Aegean[16]
§66.2. The main conundrum posed by the dialect geography here are the sharp discontinuities of these dialect enclaves. There is a need for a realistic historical hypothesis that can best explain the linguistic data in the geographic space that they are attested. The discontinuities are strongly suggestive of some kind of more recent disruption from a previous geographical distribution which would have preceded it in the late second millennium and early first millennium BCE, likely in connection with the general social and political disruptions in the Aegean following the collapse of the Bronze Age palatial economies. The lack of clear evidence for dialects other than Mycenaean and the apparent dialectal uniformity of Mycenaean Greek across sites, unfortunately, does not provide much to elucidate this picture.[17] To return to the picture of the first millennium BCE, some early literature attempted to use the Greeks’ own historiographic traditions about migrations at the end of the Bronze Age as an explanatory device for the attested dialect geography,[18] but the invocation of these traditions have since, quite rightly, been criticized, since much of the narratives from ancient authors claiming earlier migrations are closely bound up with the Classical Greeks’ constructions of ethnic identity that they were projecting onto mythic prehistory.[19] That said, it is nevertheless clear that the collapse of the Mycenaean palatial economies did entail great social upheavals in mainland Greece, and the dialect geography of the Classical Greek dialects may in itself provide some small clues towards broader historical processes.
§66.3. For the remainder of this article, I would like to focus on the evolution of a single linguistic feature which may provide us an interesting case study to probe the limits of what the dialect geography of the early first millennium BCE may be able to tell us about the late second millennium, namely the evolution of the Proto-Greek labiovelar stops. The first reason for focusing on the labiovelars is that their loss is common to all Greek dialects of the first millennium BCE and demonstrably a late, post-Mycenaean development in most, if not all the first-millennium dialects. Secondly, the palatalization processes affecting the labiovelars appear to be spread throughout the dialects as an areal sound change irrespective of previous divergent evolution, and any discrepancies in the outcomes of the labiovelars may give indications of alternative relative chronologies in their development and have implications for the early history of the dialects. To this end, I will examine the evidence for labiovelar developments on the Greek mainland and propose that the developments in the dialects as attested in the first millennium BCE suggest two alternative relative chronologies of development: the first of these was a common Greek set of developments that affected most of the dialects, which the second chronology affected specifically the Aeolic dialects (Boeotian, Thessalian, and Lesbian). I will further argue that the first-millennium dialect geography is strongly suggestive that the Aeolic relative chronology applied earlier than those that affected the rest of the dialects. This has historical implications for the Aeolic dialect group as a whole and is of significance for interpreting some of the first-millennium discontinuities in Ancient Greek dialect geography.
§67. The labiovelars as an Ancient Greek dialectal isogloss
§67.1. To recapitulate the basic facts as can be found in the reference handbooks, Proto-Greek inherited a series of velar consonants from Proto-Indo-European with labial coarticulation (Proto-Greek *kʷ, *gʷ, *kʷʰ from PIE *kʷ, *gʷ, *gʷʰand secondarily labialized sequences of *k̑u̯, *g̑u̯, *g̑ʰu̯) conventionally called labiovelar stops.[20] It was realized early on in the comparative investigation of the Greek dialects that the labiovelars as reconstructed for Proto-Indo-European were retained in Proto-Greek on the basis of their differing sound correspondences among the attested Classical Greek dialects, for example:[21]
- PIE *kʷetu̯r̥- ‘four’ > Attic τέτταρες, Ionic τέσσερες, Doric τέτορες, but Thessalian and Boeotian πέτταρες(cf. Latin quattuor, Sanskrit catvā́raḥ, Old Lithuanian keturì, etc.)[22]
- PIE *gʷelbʰ– ‘womb, belly’ > Phocian and Attic-Ionic Δελφοί ‘Delphi’, etc., as a composition-member in Ionic ἀδελφεός < *sm̥-gʷelbʰ-ei̯o- ‘brother from the same mother’ (cf. Sanskrit sagarbhya- ‘brother of the same mother and father’), but cf. Boeotian Βελφοῖς ‘Delphi’ (dative plural at IG VII 2418.23, etc.), Thessalian Βελφαίο̄ ‘the Delphaian’ (genitive singular at IG IX,2 257.10),[23] Lesbian Βελφούς ‘Delphi’ (accusative plural, Etymologicum Magnum 200, 27)[24]
- PIE *g̑ʰu̯eh₁r- ‘wild animal’ > PGk. *kʷʰe̞ːr- > Attic θήρ, but Lesbian φῆρα (accusative singular at Alcaeus fr. 286b.3 Voigt); as a lexical Aeolism in Homeric φηρσίν ‘wild beasts; centaurs’ (dative plural at 1.268, etc.); cf. also the verbal derivative attested via Thessalian πεφειράκοντες (perfect participle active, nominative singular three times in IG IX,2 536, cf. Attic τεθηράκωτες)[25]
In addition to this, as Michael Weiss points out, there is also strong evidence from East Ionic to consider that the labiovelars were retained in East Ionic at the time of the colonization of Asia Minor, to judge from the basis of a sound law restricted to East Ionic which appears to have dissimilated labiovelar stops when occurring between two back vowels, e.g. Ionic ὅκως ‘how’ (cf. Attic πῶς), and additionally there is an outcome of a labiovelar which can be found the loanword πάλμυς (e.g. accusative singular πάλμυν in Hipponax fr. 3 West) < Early Ionic *kʷálmu- ultimately borrowed from Lydian qaλmλu ‘king’ (Weiss 2020:100, cf. also Hawkins 2013:188–190). In total, the internal evidence from the Classical dialects points to the retention of the labiovelars, being lost only relatively recently by the time of their earliest attestation.
§67.2. By the mid-twentieth century, the decipherment of Linear B as Mycenaean Greek through the cumulative efforts of Emmett Bennett Jr., Alice Kober, Michael Ventris, and John Chadwick led to the discovery that the labiovelars remained intact in the Mycenaean dialect and were represented using the q-series syllabograms, for example:
- PIE *gʷou̯-kʷól(h₁)o- ‘cowherd’ : Mycenaean qo-u-ko-ro /gʷou̯-kólos/ (TI Ef 2, etc.), Attic βουκόλος (cf. the nearly perfect cognate match to be found in Middle Irish búachaill, Middle Welsh bugail < Proto-Celtic *bou̯-koli-);[26] but
- PIE *h₂m̥bʰí-kʷol(h₁)o- ‘attendant’ : Myc. a-pi-qo-ro /ampʰí-kʷolos/ (TH Of 34.1, etc.), Attic ἀμφίπολος (Lat. anculus ‘servant’, cf. Skt. °cārá- in pari-cārá- ‘servant’, etc.)[27]
We may also note that while the labiovelars are intact here in early Greek, already one sound change has already occurred affecting them, since the labiovelar present in the compositional member *-kʷól(h₁)o- has regularly become a plain velar in the environment of a preceding *u or *u̯.[28] The discovery that the labiovelars were preserved almost completely intact in Mycenaean still in the late second millennium BCE (with the exception of this early dissimilation) may also corroborate the earlier adduced evidence that the loss of the labiovelar stops was relatively recent in the Classical Greek dialects; however, I would urge exercising some caution since we cannot assume that the state of affairs as attested in Mycenaean is the same in other second-millennium dialects that are unattested but we can assume to have existed on the basis of the first-millennium dialect geography.
§67.3. The main distribution of the outcomes of the labiovelar stops in Attic-Ionic and in West Greek which overwhelmingly appear to be the ‘normal’ outcomes are summarized in Table 1 below:[29]
Table 1: The Reflexes of the Proto-Greek Labiovelars in Attic-Ionic and West Greek Dialects
Before *e(ː) the labiovelars become dentals with phonetic features corresponding to their original voicing and aspiration:[30]
- *-kʷe ‘and’ > τε ‘and’ (Latin -que, Sanskrit ca, etc.)[31]
- *sm̥-gʷelbʰ– in ἀ-δελφ-ός ‘brother’ (cf. Sanskrit sagarbhya- ‘brother of the same mother and father’)[32]
- *gʷʰer-mo- > θερμός ‘warm’ (Armenian ǰarm ‘warm’, cf. Sanskrit gharmá- ‘heat’, Latin formus ‘warm’, Proto-Germanic *warma- < *gʷʰor-mó-)[33]
Before *i(ː) the distribution is split: The voiceless labiovelar *kʷ also has a dental reflex /t/ as before *e(ː), for example in the case of the relative-interrogative pronoun:
- *kʷis, *kʷid > τίς, τί ‘who, what’ (Hittite kuiš, kuit, Latin quis, quid, etc.)[34]
Exceptionally, however, voiceless aspirated *kʷʰ and voiced unaspirated *gʷ have labial outcomes:
Typically, the dental reflexes that are found before *e(ː) and *i(ː) are explained as the result of a typologically common process of palatalization occurring before front vowels. In a typical palatalization process the place of articulation of a preceding consonant is moved forward in the mouth as a result of it being affected by the articulatory properties of a following mid- or high-front vowel.[37] The distribution before high-front *i(ː) in the majority of the Classical Greek dialects is split with the reflex of the voiceless labiovelar exhibiting a voiceless dental /t/, while the voiced and voiceless-aspirated labiovelars failed to undergo the same palatalization process. The split distribution of the labiovelars before *i(ː) is a separate problem which will be further addressed at §70 and §71 below. For now I will turn to the outcomes in Arcadian in §68 and in Aeolic in §69, since I will further argue below that the process of this unusual distribution in Attic-Ionic and West Greek can be better understood through comparison with the outcomes in these dialects and the nature of how the palatalization of the labiovelars was spread.
§68. The special case of Arcadian <Ͷ> (and other unusual spellings)
§68.1. It is well known that archaic Arcadian inscriptions occasionally attest a special character tsan, which I have transcribed here as <Ͷ>, and other unusual spellings whose precise phonological values remain unclear, but appear to represent an intermediary state of development of sounds that reflect labiovelars etymologically.[38] These are most famously attested in a fifth-century inscription from Mantinea recording judgements of persons guilty of sacrileges to the temple of Athena Alea (IG V,2 262, see Figure 2), but now with certainty four additional times in a new festival calendar from Arcadia (Heinrichs 2015, Carbon and Clackson 2016).[39]
Figure 2: Detail of IG V,2 262, col. ii, l. 24–36 (after Comparetti 1914, Plate 1)
To make my general point I will restrict myself to discussing the evidence of <Ͷ> for the palatalization of the labiovelars in archaic Arcadian inscriptions but see especially Dubois (1986a:64–70) and Duhoux (2006) for more extensive discussions of these forms and other unexpected spellings of Arcadian forms using <ζ> epigraphically and in glosses.[40]
- Examples of *kʷi- > Ͷι-
- Ͷι[ν]’ (τινά) (IG V,2 262.23)
- Ͷις (τις) (IG V,2 262.25, 27)
- Examples of *kʷe- > Ͷε-
- εἴͶ’ (εἴτε) (IG V,2 262.26)
- εἴͶε (εἴτε) (IG V,2 262.26, 27, 28, 29, 31)
- ὁͶέο̄ι (ὁτέωι) (IG V,2 262.14)
- Examples of *gʷe- > Ͷε-
- ὀͶελο̄́ ‘two obols’ (nom.-acc.du.) (Carbon and Clackson 2016:122, l.13) < Proto-Greek *ogʷeló- (cf. Attic ὀβολός < *ogʷoló-)
- ὀͶελόν ‘obol’ or ‘spit’ (contextual sense unclear) (acc.sg.) (Carbonand Clackson 2016:122, l.19)
- Other examples of <Ͷ>
- ἀπυͶεδομίν[ος](ἀποδεδομένους) (IG V,2 262.19)[41]
- ͶεσͶάρο̄ν̣ ‘four’ (gen.pl.) < *kʷetu̯r̥- (Carbon and Clackson 2016:122, l.11)
On the basis of these examples, we may observe that <Ͷ> is writing a sound that is intermediate and show an ongoing process towards the dental outcomes that are found in other Arcadian dialect inscriptions. It is notable that tsan is very likely being used primarily to distinguish what is probably a feature of affrication which is not phonemically identical in all cases, since the attested examples are used regardless of etymologically expected voicing contrasts (cf. εἴͶε (εἴτε) ‘and if’ versus ὀͶελόν ‘obol; spit’ above) or even specifically the affricates specific to the outcome of palatalized labiovelars (cf. <σͶ> in the spelling of ͶεσͶάρο̄ν̣ ‘four’ reflecting the intermediate outcome of *-tu̯-, possibly via *-stˢ-).[42]
§68.2. Later Arcadian from the fourth century onwards attests dental forms in these environments congruent with the outcomes found in Attic-Ionic and West Greek, for example:[43]
- Examples of *kʷi- > τι-
- τι ‘what’ (IG V,2 3.5), etc.
- τιμασίαν (Dubois 1986b:61–63, Té 4 l.17), abstract derivative of Arcadian *τιμά < *kʷi-meh₂[44]
- Examples of *kʷe- > τε-
- Full-grade derivatives of PIE 3.*kʷei̯- ‘to pay’ in ἀπυτεισάτω (3.sg.aor.imptv., IG V,2 6.32), etc.
- πέντε ‘five’ (IGV,2 3.1) < *pénkʷe
- In εἴτε ‘and if’ (IGV,2 343.10), etc. < *-kʷe
- τετόρταυ ‘fourth’ (gen.sg.f. IG V,2 6.104), cf. Attic τετάρτης from stem *kʷetu̯r̥-to-
- Examples of *gʷe- > δε-
- ὀδελός ‘obol’ (IGV,2 3.19, 24), cf. (14) above.
- ἐσδέλλοντες ‘throwing’ < as though continuing *eks-gʷelh₁-i̯e/o- (IG V,2 6.49), cf. Attic ἐκ-βάλλοντες[45]
§68.3. Individual problems in interpreting the granular phonetic data aside, we may nevertheless assume that the spellings using tsan represent an intermediary step to what also happened in Attic-Ionic and West Greek, given that the reflexes in later Arcadian are congruent with all other Greek dialects with the exception of Boeotian, Thessalian, and Lesbian.[46] The further implicature of this, taking the attested dialect geography into consideration, is that the sound changes associated with the palatalization of the labiovelars passed through the dialect geography as an areal sound change, regardless of existing isogloss boundaries or other linguistically divergent evolution. The fact that Attic-Ionic and West Greek agree on these outcomes was already suggestive of this, but taking the Arcadian evidence where the sound change only came to completion in the mountainous interior of the Peloponnese by the fifth century BCE is further strong evidence of the nature of the propagation of these developments.
§69. The outcomes of the labiovelars in Boeotian, Thessalian, and Lesbian
§69.1. As for the outcomes in the Aeolic dialects, there is a mixture of dental and labial reflexes between front vowels that need to be sifted and sorted. Elsewhere I have collected and analyzed the attested examples (Scarborough 2023:68–88, with full references from documentary sources). The main synopsis of the attested forms may be found in Table 2 and Table 3.[47]
Table 2: Synopsis of Labial Reflexes of Labiovelars
Regarding the labial reflexes of the labiovelars in the Aeolic dialects, I note that there are only two potential examples attesting a labial reflex for the labiovelars before high-front *i(ː), namely in Boeotian ἀποπισάτω ‘let him pay’ (SEG22:407.33, 3rd c. BCE, = Attic ἀποτεισάτω) and ποταποπιάτω ‘let him pay besides’ (IG VII 3172.85, ca. 222–200 BCE, = Attic προσαποτεισάτω). However, these two examples are not secure because they postdate the Boeotian narrowing of high-mid /eː/ to high /iː/,[48] and on the basis of comparison with other dialects the continuation of a full-grade form *ἀποπεισάτω is expected here.[49] Thessalian only attests a single counterexample ἀπ<π>ισαι (IG IX,2 1202.5, undated, archaic script), but in several other places the expected full-grade aorist forms are attested: infinitives ἀππε[ῖσ|α]ι (IG IX,2 1226.10–11, 5th c. BCE) and ἀπ<π>εῖσαι (Giannopoulos 1934–1935, No.1.11, 6th c. BCE?), and the third singular aorist imperative ἀππεισάτου (IG IX,2 1229.28, early 2nd c. BCE), so it is more likely that ἀπ<π>ισαι of IG IX,2 1202 is a one-off misspelling. Consequently, there are no clear examples for labial reflexes of original *kʷi- > πι- in Thessalian or Boeotian. In the remaining environments before high-front *i(ː) dental reflexes are found.
Table 3: Synopsis of Dental Reflexes of Labiovelars
While the outcomes of labiovelars in the Aeolic dialects before front vowels are largely in favor of labial outcomes (except *kʷi- and *gʷʰi- where there are no clear examples for labials), a small handful of dental reflexes remain. The most prominent examples before high-front *i(ː) are the relative-interrogative τίς, τί < *kʷis, *kʷid, and Lesbian τίμα, Boeotian and Thessalian τιμά ‘compensation, payment’ < *kʷi-meh₂ (Attic τιμή), never attested as †πίς, †πί or †πίμα/πιμά. Additionally difficult is the regular form of the enclitic conjunction τε < *kʷe, never surfacing as †πε. The occasional reflexes of PIE *penkʷe ‘five’ and derivatives of PIE *gʷelbʰ– ‘womb, belly’ have counterexamples with the more marked labial outcomes elsewhere. There is also a single early example for *gʷʰe attested in θ̣έρμαν (Alc. 143.10 Voigt), albeit in a fragmentary context.[50]
§69.2. To assess the comparative value of labial vs. non-labial reflexes of labiovelars in the Aeolic dialects, it is clear that the labialized examples before *e(ː) strongly outweighs the evidence for dental reflexes in the same environment. The enclitic conjunction τε < *kʷe is the most puzzling counterexample, and the most economical hypothesis to account for it would be to assume an early loss in these dialects, and then a re-borrowing from the literary or Homeric oral-poetic register. Short of setting up a restricted sound law that applies only to this environment for clitics (cf., e.g. Stephens and Woodard 1986:147, Lejeune 1972:45), the hypothesis of supradialectal borrowing seems to me to be the most plausible solution, given that it seems methodologically difficult to justify an entire sound law for which this is the only example, in addition to the observation that cross-linguistically connectors are the grammatical structures most susceptible to borrowing.[51] The Aeolic treatment of *kʷi- > τι- in Boeotian and Thessalian τιμά, Lesbian τίμα and especially in the relative-interrogative pronoun τίς, τί, which is far less likely to be a borrowing,[52]strongly points to the regularity of this sound correspondence in this environment.[53]
§70. Remarks on the observed distribution of the labiovelars
§70.1 To simply consider the attested sound correspondences, in addition to the ‘normal’ outcomes summarized in Table 1 above, we may add the synopsis of the observed ‘Aeolic’ outcomes in Table 4 below:
Table 4: Observed Reflexes of the Labiovelars in Boeotian, Thessalian, and Lesbian
As discussed at §69.1 and §69.2 above, the main distribution of the labiovelars in the Aeolic dialects tends towards complete labialization, but a small number of difficult exceptions remain. Most of these can be attributed to a concession to broader supradialectal tendencies and isolated dialectal loanwords, but the cases of the relative-interrogative *kʷis > τίς and *kʷi-meh₂ > Boeotian, Thessalian τιμά, Lesbian τίμα are much more difficult to dismissfor *kʷ developing to the voiceless dental /t/ before high-front /i/. Meanwhile, the ‘normal’ outcomes as found in Attic-Ionic, West Greek, and later Arcadian exhibit the outcomes described at §75.3. above, where the labiovelars are regularly palatalized to dentals before *e(ː), and the voiceless labiovelar *kʷ is palatalized to a dental before *i(ː). The first observation I make from this distribution is that there seems to be something special about the restricted environment of the voiceless labiovelar *kʷ before high-front /i/, where this appears to be not only the regular outcome in most of the dialects, but also in Aeolic where labialization appears to be regular in all other environments.
§70.2. The second observation that I would make following upon the previous one, is that the split distribution before *i where *kʷ palatalizes while *gʷ and *kʷʰ do not, despite regularly palatalizing before *e, is typologically unusual if one is expecting a simple palatalization process. To put it another way, the relative ability of a vowel to palatalize a consonant is on a gradient on the front-high axis of the vowel space, since this is the area of the vowel space where the tongue moves closer to the hard and soft palate of the mouth. Because of this the high front vowel /i/ gives the narrowest constriction to the palate, and so is the vowel that is most likely to palatalize a preceding consonant, and this effect becomes less likely the further back and low the articulation of the vowel is (see Figure 3).[54]
Figure 3: A Reconstruction of the Proto-Greek Vowel Space with a Typology of Typically Palatalizing Environments Overlaid
With this held in mind the crux of the matter is: if /i/ is the vowel that is prototypically more susceptible to palatalizing consonants, why do *gʷi-, *kʷʰi- escape palatalization while the labiovelars become palatalized before /e/ in the non-Aeolic dialects? Both observations from §70.1 and §70.2 appear to point to a much more complicated story in Aeolic and the other Greek dialects in the Aegean.[55]
§71. A proposed solution for Attic-Ionic, West Greek, and Arcadian
§71.1 The most economical solution to deal with the problem of the split reflex of the labiovelars before /i/, to my knowledge, was a proposal made by H. N. Parker (2013) which is the solution I adopt here and will outline in brief. Parker, however, has earlier argued against the historical unity of the Aeolic dialectal subgrouping and so the broader historical implications of his analysis in view of the dialect geography and the non-arboreal propagation of the palatalization of the labiovelars before /e/ in most of the dialects may be productively further expanded upon.[56]
§71.2 The first part of these developments addresses the first observation made in §70.1, namely that in all dialects, even Aeolic, the voiceless unaspirated labiovelar stop *kʷ has palatalized to a dental. The first stage of these developments therefore ought to be an early palatalization of only the voiceless unaspirated labiovelar *kʷ specifically restricted to the typologically most palatalizing environment preceding long or short *i (19).
- *kʷ > *tˢ > t / _i(ː)
A similar early palatalization of this specific environment was also proposed by Stephens and Woodard (1986:133) where the sequences of *kʷʰi and *gʷi may have been less susceptible to palatalization because of their additional phonetic properties (cf. Stephens and Woodard 1986:143–145 with typological parallels). In effect, this restricted sound change uniquely removes the voiceless unaspirated labiovelar *kʷ and sends it down an alternative sequence of developments from the other two labiovelars in this environment from the very beginning.
Table 5: The ‘normal’ reflexes after the Stage 1 developments
§71.3. The crucial intermediary step is the beginning of a palatalization process of the remaining labiovelars before i(ː) and e(ː). In these environments the labial element of the labiovelar stops would become labiopalatal in articulation:
- *Kʷ > *Kᶣ / _{e, i}(ː)
But crucially, following the argument of Parker (2013: 223), before long or short /i/ the labiopalatal element [ɥ] did not fully delabialize in the sequence [ᶣi] because it would have created a cross-linguistically disfavored sequence [ʲi].[57]In this environment new learners could have reanalyzed these retained sequences of [ᶣi] as allophones of the labiovelars that were still independent phonemes in the phonological system, and so in the specific environment before *i(ː) the labiovelars were restored.[58]
The remainder of the labiopalatals which continued to exist before *e(ː) would lose the labial feature and proceed down a familiar path of palatalization, eventually merging with the affricates and developing with them.
§71.4. Once the labiopalatals were restored to labiovelars before /i/ the remainder of the labiovelars would have lost their velar co-articulation in all environments giving the remainder of the labial outcomes. Cross-linguistically, at least among the early Indo-European languages this is a common sound change with parallel developments in the Sabellic languages and in some varieties of Celtic.[59] This is likely to have occurred in most dialects prior to the resolution of the palatalizations which took place as described at §71.3, since this wave of labialization is already present in the early Arcadian inscriptions which attest tsan as a grapheme, e.g. φονε̄́ς ‘murderer’ < *gʷʰon-es- (IG V,2 262.26, 30; cf. Attic φονεύς), βο̴ν ‘ox’ < acc.sg. *gʷōm (Carbon and Clackson 2016:122, l.7), etc. (see Table 6).
Table 6: The approximate state of affairs as attested in early Arcadian
Following the state which we have attested in early Arcadian, the only remaining change necessary to arrive at the pan-Greek ‘normal’ outcomes of the labiovelars is the resolution of the phonemes represented in early Arcadian through tsan (via de-affrication?) as dental stops, completing the palatalization process and is congruent with the observed outcomes as outlined in Table 1 at §75.3. above.
§71.5. An approximate relative chronology of the entirety of the developments concerning the palatalization and labialization of the labiovelars in Attic-Ionic, West Greek, and Arcadian is schematized in Figure 4.
Figure 4: A Schematized Relative Chronology for the Labiovelars in Attic-Ionic, West Greek, and Arcadian
§72. The relative chronology of the labiovelars in Aeolic versus other dialects
§72.1. At §69.2 and §70.1 I argued that the observed outcomes of the labiovelars in Boeotian, Thessalian, and Lesbian appears to regularly reflect *kʷi > τι, with full labialization in all other environments. The simplest solution to account for this distribution would be an assumption of a different rule ordering first assuming a pan-Greek palatalization of the voiceless unaspirated labiovelar *kʷ in the restricted environment when followed by high-front *i (§71.2), followed by a subsequent labialization of all the labiovelars in all remaining environments (§71.4), bypassing the process of the second palatalization before front vowels entirely (§71.3).
§72.2. If this is the best explanation for the linguistic facts alone, then the dialect geography must also be addressed, particularly because it appears evident that the second palatalization (§71.3) was an areally spread sound change (cf. §76). Thessalian and Boeotian are centrally placed in the dialect geography of the Classical dialects, and Boeotia in particular stands at the crossroads of Northern and Northwest Greece, the Peloponnese, and Attica. It seems reasonable to ask, given the centrality of Boeotia in mainland Greece, why the second palatalization did not apply in Boeotian (and to a certain extent also in Thessalian). While alternative explanations may be possible, the lack of the second palatalization of the labiovelars in Boeotian is suggestive that the full labialization occurred earlier and independentlyfrom the labialization which occurred in other dialects after the beginning of the second palatalization of the labiovelars began elsewhere. Essentially, the environments that would have been affected were lost before the process of the second palatalization could apply. Parallel to the developments schematized at §71.5, I offer a relative chronology for Boeotian, Thessalian, and Lesbian in Figure 5 below.
Figure 5: A Schematized Relative Chronology for the Labiovelars in Aeolic
§73. Conclusions
§73.1. To conclude, the dialectal distribution of the observed reflexes of the Proto-Greek labiovelar stops in the Aegean may be best reconciled by assuming two separate relative chronologies for their development. There is the first one, described in §71 which applied in the majority of dialects, and the separate developments which applied in Boeotian, Thessalian, and Lesbian described in §72. Given that the developments of the second palatalization (§71.3) were areal, the attested dialect geography has implications for their relative interpretation, and I have argued that, as a consequence of this, the full labialization of the labiovelars in Boeotian, Thessalian, and Lesbian was likely to have been earlier and independent of the remaining Greek dialects. While loss of labial coarticulation is a typologically not uncommon way for labiovelars to be lost in other attested Indo-European languages (§71.4), it is especially notable that the shared relative chronology for the elimination of the labiovelars in Boeotian, Thessalian, and Lesbian prior to the areally spread second palatalization is a stronger argument for the labialization found in the Aeolic dialects being a non-trivial sound change. This is because the opposite hypothesis would require the assumption that all three dialects underwent an alternative series of sound changes in an order different from the rest of the Greek dialects two or three times independently, and crucially the intersection of the attested dialect geography and the areal nature of the spread of the second palatalization of the labiovelars allows a more nuanced view of this isogloss to be possible. Finally, if this scenario is correct and it points to an earlier unity of the Aeolic dialects, it will be necessary to come up with a concrete historical hypothesis that can explain the discontinuous dialect geography of Boeotian and Thessalian around the Malian Gulf, but for now I will leave that question as a matter for further research.
§74. Acknowledgement
This was written while employed on the research project “Connecting the Dots: Reconfiguring the Indo-European Family Tree” (Independent Research Fund Denmark, grant no. 9037–00086B). I thank Thomas Olander for allowing me to use project time to write up this article.
For the discussion following this paper, please return to §75 in the MASt Spring 2023 report page, here: https://continuum.fas.harvard.edu/mast-reports-spring-2023-friday-may-19/
Bibliography
Allen, W. S. 1958. “Some Problems of Palatalisation in Greek.” Lingua 7:113–133.
Bartoněk, A. 1987a. “On the Prehistory of Ancient Greek.” Studi micenei ed egeo-anatolici 26:7–21.
Bartoněk, A. 1987b. Prehistorie a protohistorie řeckých dialektů. Brno.
Beekes, R. S. P. and M. de Vaan. 2011. Comparative Indo-European Linguistics: An Introduction. 2nd ed. Amsterdam and Philadelphia.
Béquignon, Y. 1935. “Études Thessaliennes: VII. Inscriptions de Thessalie.” Bulletin de Correspondance Hellénique59:36–77.
Bile, M., C. Brixhe, and R. Hodot. 1984. “Les dialectes grecs, ces inconnus.” Bulletin de la Société de Linguistique de Paris 79:155–203.
Brixhe, C., and G. Vottéro, ed. 2006. Peuplements et genèses dialectales dans la Grèce antique. Études anciennes 31. Nancy.
Buck, C. D. 1933. Comparative Grammar of Greek and Latin. Chicago. Available at: https://archive.org/details/BuckComparativeGrammer/mode/2up
Buck, C. D. 1955. The Greek Dialects: Grammar, Selected Inscriptions, Glossary. Chicago and London.
Carbon, J.-M., and J. P. T. Clackson. 2016. “Arms and the Boy: On the New Festival Calendar from Arcadia.” Kernos29:119–158. Available at: https://journals.openedition.org/kernos/2393#quotation
Chadwick, J. 1976. “Who Were the Dorians?” Parola del Passato 31:103–117.
Chambers, J. K. and P. Trudgill. 1998 Dialectology. 2nd ed. Cambridge.
Comparetti, D. 1914. “La iscrizione arcaica di Mantinea.” Annuario della Regia Scuola Archeologica di Atene 1:1–17.
Cowgill, W. 1966. “Ancient Greek Dialectology in the Light of Mycenaean.” In Ancient Indo-European Dialects. Proceedings of the Conference on Indo-European Linguistics Held at the University of California, Los Angeles, April 25–27, 1963, ed. H. Birnbaum and J. Puhvel, 77–95. Berkeley. Reprinted 2006 in The Collected Writings of Warren Cowgill, ed. J. S. Klein, 171–184. Ann Arbor and New York.
de Lamberterie, C. 2012. “L’apport du mycénien à l’étymologie grecque.” In Études mycéniennes 2010. Actes du XIIIe colloque international sur les textes égéens, Sèvres, Paris, Nanterre, 20-23 septembre 2010, ed. P. Carlier, C. deLamberterie, M. Egetmeyer, N. Guilleux, F. Rougemon, J. Zurbach, 489–509. Biblioteca di Pasiphae X. Pisa and Rome.
DELG = Chantraine, P. 1968–1980. Dictionnaire étymologique de la langue grecque. Histoire des mots. Paris.
de Saussure, F. 1889. “βουκόλος.” Mémoires de la Société de linguistique de Paris 6:161–162.
Dubois, L. 1986a. Recherches sur le dialecte arcadien. I: Grammaire. Bibliothèque des Cahiers de l’Institut de Linguistique de Louvain 33. Louvain-la-Neuve.
Dubois, L. 1986b. Recherches sur le dialecte arcadien. II: Corpus dialectal. Bibliothèque des Cahiers de l’Institut de Linguistique de Louvain 34. Louvain-la-Neuve.
Dubois, L. 1986c. Recherches sur le dialecte arcadien. III: Notes – index – bibliographie. Bibliothèque des Cahiers de l’Institut de Linguistique de Louvain 35. Louvain-la-Neuve.
Duhoux, Y. 2006. “Le lettre Ͷ et quelques problèmes connexes en arcadien archaïque (IG V 2.262).” Kadmos 45:20–68.
EDG = Beekes, R. S. P. 2010. Etymological Dictionary of Greek. Leiden Indo-European Etymological Dictionary Series 10. Leiden and Boston.
Egetmeyer, M. 2010. Le dialecte grec ancien de Chypre. Tome I: Grammaire. Berlin and New York.
Gallis, K. 1975. “Αρχαιότητες και μνημεία Θεσσαλίας.” Αρχαιολογικόν Δελτίον 26:291–304.
García Ramón, J. L. 2010. “On the Genetic Classification of Ancient Greek Dialects: Comparative Reconstruction versus Hypercriticism and Atomism at Work.” Studies in Greek Linguistics 30:219–235. Available at: http://ins.web.auth.gr/images/MEG_PLIRI/MEG_30_219_236.pdf
García Ramón, J. L. 2018. “Ancient Greek Dialectology: Old and New Questions, Recent Developments.” In Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, ed. G. K. Giannakis, E. Crespo, and P. Filos, 29–106. Trends in Classics – Supplementary Volumes 49. Berlin and Boston.
García Ramón, J. L. 2020. “Change and Grammatical and Lexical Structures in Postclassical Greek: Local Dialects and Supradialectal Tendencies.” In Postclassical Greek: Contemporary Approaches to Philology and Linguistics, ed. D. Rafiyenko and I. A. Seržant, 303–336. Trends in Linguistics. Studies and Monographs 335. Berlin and New York.
GEW = Frisk, H. 1960–1972. Griechisches etymologisches Wörterbuch. Volumes 1–3. Heidelberg. Available at: https://archive.org/details/hjalmar/mode/2up
Giannopoulos, N. I. 1934–1935. “Ἐπιγραφαὶ ἐκ Θεσσαλίας.” Ἀρχαιολογική Ἐφημερίς 73:140–150. Available at: https://ir.lib.uth.gr/xmlui/bitstream/handle/11615/9078/article.pdf?sequence=1&isAllowed=n
Hajnal, I. 2007. “Die Vorgeschichte der griechischen Dialekte: ein methodischer Rück- und Ausblick.” In Die altgriechischen Dialekte: Wesen und Werden. Akten des Kolloquiums Freie Universität Berlin 19.–22. September 2001, ed. I. Hajnal, 131–156. Innsbrucker Beiträge zur Sprachwissenschaft 126. Innsbruck.
Hall, J. M. 1995. “The Role of Language in Greek Ethnicities.” Proceedings of the Cambridge Philological Society41:83–100.
Hall, J. M. 1997. Ethnic Identity in Greek Antiquity. Cambridge.
Hansen = Hansen, P. A. 2005. Hesychii Alexandrini Lexicon. Volumen III: Π–Σ. Sammlung griechischer und lateinischer Grammatiker 11(3). Berlin and New York.
Harðarson, J. A. 1993. Studien zum urindogermanischen Wurzelaorist und dessen Vertretung im Indoiranischen und Griechischen. Innsbrucker Beiträge zur Sprachwissenschaft 74. Innsbruck.
Hawkins, S. 2013. Studies in the Language of Hipponax. Münchener Forschungen zur historischen Sprachwissenschaft 14. Bremen.
Heinrichs, J. 2015. “Military Integration in Late Archaic Arkadia: New Evidence from a Bronze Pinax (ca. 500 BC) of the Lykaion.” In The Many Faces of War in the Ancient World, ed. W. Heckel, S. Müller, and G. Wrightson, 1–89. Newcastle-upon-Tyne.
Hock, H. H. 2021. Principles of Historical Linguistics. 3rd edition. Trends in Linguistics. Studies and Monographs 34. Berlin and Boston.
Hoffmann, O. 1891–1898. Die griechischen Dialekte in ihrem historischen Zusammenhänge. Volumes 1–3. Göttingen.
Horrocks, G. 2006. Ελληνικά: Ιστορία της γλώσσας και των ομιλητών της. Athens.
Horrocks, G. 2010. Greek: A History of the Language and its Speakers. 2nd edition. Chichester and Malden, MA. Available at: https://archive.org/details/horrocks-2/mode/2up
ICS = Masson, O. 1983. Les inscriptions chypriotes syllabiques. Recueil critique et commenté, réimpression augmentée. Paris.
IG = Inscriptiones Graecae.
Janko, R. 2012. “πρῶτόν τε καὶ ὕστατον αἰὲν ἀείδειν: Relative Chronology and the Literary History of the Early Greek Epos.” In Relative Chronology in Early Greek Epic Poetry, ed. Ø. Andersen and D. Haug, 20–43. Cambridge. Available at: https://www-personal.umich.edu/~rjanko/Relative%20chronology.pdf
Jones, B. 2012. “Relative Chronology and an ‘Aeolic Phase’ of Epic.” In Relative Chronology in Early Greek Epic, ed. Ø. Andersen and D. Haug, 44–64. Cambridge.
Jeffery, L. H. 1990. The Local Scripts of Archaic Greece: A Study of the Origin of the Greek Alphabet and its Development from the Eighth to the Fifth Centuries B.C. Revised edition with a Supplement by A. W. Johnston. Oxford.
Klein, J. S. 1988. “Proto-Indo-European *gʷiH₃- ‘live’ and Related Problems of Laryngeals in Greek.” In Die Laryngaltheorie und die Rekonstruktion des indogermanischen Laut- und Formensystems, ed. A. Bammesberger, 257–279. Heidelberg.
Kretschmer, P. 1909. “Zur Geschichte der griechischen Dialekte.” Glotta 1:9–59.
Kümmel, M. J. 2007. Konsonantenwandel. Bausteine zu einer Typologie des Lautwandels und ihre Konzequenzen für die vergleichende Rekonstruktion. Wiesbaden.
Latte-Cunningham = Latte, K. and I. C. Cunningham. 2020. Hesychii Alexandrini Lexicon. Volumen IIa: Ε–Ι. Sammlung griechischer und lateinischer Grammatiker 11(2a). Berlin and New York.
Lejeune, M. 1972. Phonétique historique du mycénien et du grec ancien. Tradition de l’humanisme IX. Paris.
LIPP = Dunkel, G. E. 2014. Lexikon der indogermanischen Partikeln und Pronominalstämme. Band I: Einleitung, Terminologie, Lautgesetze, Adverbialbildungen, Nominalsuffixe, Anhänge und Indices. Band II: Lexikon. Heidelberg.
LIV² = Rix, H., M. Kümmel, T. Zehnder, R. Lipp, and B. Schirmer, ed. 2001. Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstammbildungen. Wiesbaden.
Matasović, R. 2009. Etymological Dictionary of Proto-Celtic. Leiden Indo-European Etymological Dictionary Series 9. Leiden and Boston. Available at: https://archive.org/details/EtymologicalDictionaryOfProtoCeltic/mode/2up
Matras, Y. 2007. “The Borrowability of Structural Categories.” In Grammatical Borrowing in Cross-Linguistic Perspective, ed. Y. Matras and J. Sakel, 31–73. Empirical Approaches to Language Typology 38. Berlin and New York.
McCone, K. 1996. Towards a Relative Chronology of Ancient and Medieval Celtic Sound Change. Maynooth Studies in Celtic Linguistics 1. Maynooth.
Meillet, A. 1930. Aperçu d’une histoire de la langue grecque. 3rd edition. Paris.
Meiser, G. 1986. Lautgeschichte der umbrischen Sprache. Innsbrucker Beiträge zur Sprachwissenschaft 51. Innsbruck.
Méndez Dosuna, J. 2007. “The Aeolic Dialects.” In A History of Ancient Greek: From the Beginnings to Late Antiquity, ed. A. F. Christidis, 460–474. Cambridge.
Nagy, G. 1968. “On Dialectal Anomalies in Pylian Texts.” In Atti e Memorie del Primo Congresso Internazionale di Micenologia, Roma 27 Settembre–3 Ottobre 1967, 663–679. Incunabula Graeca 25(2). Rome.
Ohala, J. 1994. “Hierarchies of Environments for Sound Variation; Plus Implications for ‘Neutral’ Vowels in Vowel Harmony.” Acta Linguistica Hafniensia 27(1):371–382.
Padgett, J. 2001. “Contrast Dispersion and Russian Palatalization.” In The Role of Speech Perception in Phonology, ed. E. Hume and K. Johnson, 187–218. San Diego. Available at: https://roa.rutgers.edu/files/504-0302/504-0302-PADGETT-0-0.PDF
Palaima, T. G. 1998–1999. “Special vs. Normal Mycenaean: Hand 24 and Writing in the Service of the King?” Minos33–34:205–221. Available at: https://gredos.usal.es/bitstream/10366/133912/1/Special_Vs_Normal_Mycenaean_Revisited.pdf
Palaima, T. G. 2011. “Scribes, Scribal Hands, and Palaeography.” In A Companion to Linear B: Mycenaean Greek Texts and Their World. Volume II, ed. Y. Duhoux and A. Morpurgo Davies, 33–136. Bibliothèque des Cahiers de l’Institut de Linguistique de Louvain 127. Louvain-la-Neuve.
Parker, H. N. 2008. “The Linguistic Case of the Aiolian Migration Reconsidered.” Hesperia 77(3):431–464.
Parker, H. N. 2013. “Palatalization of the Labiovelars in Greek.” In Multi Nominis Grammaticus: Studies in Classical and Indo-European Linguistics in honor of Alan J. Nussbaum, ed. A. I. Cooper, J. Rau, and M. Weiss, 214–227. Ann Arbor and New York.
Porzig, W. 1954. “Sprachgeographische Untersuchungen zu den altgriechischen Dialekten.” Indogermanische Forschungen 61:147–169.
Ringe, D. 2016. “The Nature of the South Greek Dialect Group.” In Tavet Tat Satyam: Studies in Honor of Jared S. Klein on the Occasion of his Seventieth Birthday, ed. A. M. Byrd, J. DeLisi, and M. Wenthe, 278–283. Ann Arbor and New York.
Risch, E. 1955. “Die Gliederung der griechischen Dialekte in neuer Sicht.” Museum Helveticum 12(2):61–76.
Risch, E. 1966. “Les différences dialectales dans le mycénien” In Proceedings of the Cambridge Colloquium on Mycenaean Studies, ed. L. R. Palmer and J. Chadwick, 150–157. Cambridge.
Rix, H. 1976. Historische Grammatik des Griechischen: Laut- und Formenlehre. Darmstadt.
Scarborough, M. 2023. The Aeolic Dialects of Ancient Greek: A Study in Historical Dialectology and Linguistic Classification. Brill’s Studies in Indo-European Languages & Linguistics 26. Leiden.
Schwyzer, E. 1939. Griechische Grammatik auf der Grundlage von Karl Brugmanns Griechischer Grammatik. 1. Band: Allgemeiner Teil, Lautlehre, Wortbildung, Flexion. Munich.
SEG = Supplementum Epigraphicum Graecum.
Sihler, A. L. 1995. New Comparative Grammar of Greek and Latin. Oxford. Available at: https://archive.org/details/sihler-andrew-new-comparative-grammar-of-greek-and-latin/mode/2up
Schumacher, S. 2004. Die keltischen Primärverben. Ein vergleichendes etymologisches und morphologisches Lexikon. Innsbrucker Beiträge zur Sprachwissenschaft 110. Innsbruck.
Stephens, L. D., and R. D. Woodard. 1986. “The Palatalization of the Labiovelars in Greek: A Reassessment in Typological Perspective.” Indogermanische Forschungen 91:129–154.
Tadmor, U. 2009. “Loanwords in the World’s Languages: Findings and Results.” In Loanwords in the World’s Languages: A Comparative Handbook, ed. M. Haspelmath and U. Tadmor, 55–75. Berlin and New York.
Thompson, R. J. E. 1996–1997. “Dialects in Mycenaean and Mycenaean among the Dialects.” Minos 31–32:313–333.
Thompson, R. J. E. 2002–2003. “Special vs. Normal Mycenaean Revisited.” Minos 37–38:337–369. Available at: https://gredos.usal.es/bitstream/10366/133912/1/Special_Vs_Normal_Mycenaean_Revisited.pdf
Thumb, A., and A. Scherer. 1959. Handbuch der griechischen Dialekte. Zweiter Teil. Heidelberg.
Vendryes, J. 1959. Lexique étymologique de l’Irlandais ancien: A. Dublin and Paris.
Voigt = Voigt, E.-M. 1971. Sappho et Alcaeus: Fragmenta. Amsterdam.
Weiss, M. 1994. “Life Everlasting: Latin iūgis “everflowing”, Greek ὑγιής “healthy”, Gothic ajukdūϸs “eternity” and Avestan yauuaēǰī- “living forever”.” Münchener Studien zur Sprachwissenschaft 55:131–156. Available at: https://canvas.uw.edu/files/81542594/download?download_frd=1&verifier=sgaP1dNDHNuXVdiC5VxLGxPGpZ63sjEmO3UPH4nY
Weiss, M. 2020. Outline of Historical and Comparative Grammar of Latin. 2nd edition. Ann Arbor.
Woodard, R. D. 1986. “Dialectal Differences at Knossos.” Kadmos 25:49–74.
Woodard, R. D. 2008. “Greek dialects.” In The Ancient Languages of Europe, ed. R. D. Woodard, 50–72. Cambridge.
Woodard, R. D. 2012. “Labiovelar Development in Greek and an Alphabetic Repercussion.” In The Sound of Indo-European 2: Papers on Indo-European Phonetics, Phonemics and Morphophonemics, ed. R. Sukač and O. Šefčík, 299–311. Munich.
Woodard, R. D. 2021. “Aeolian Origins, and Other Mycenaean Matters (book draft).” Classical Continuum. Available at: https://continuum.fas.harvard.edu/a-draft-of-the-forthcoming-book-aeolian-origins-and-other-mycenaean-matters/
Woodard, R. D. forthcoming. Aeolic: Origins of an Ancient Greek Language Community. Cambridge.
Wyatt, W. F. 1970. “The Prehistory of the Greek Dialects.” Transactions and Proceedings of the American Philological Association 101:557–632.
Footnotes
[14] See Cowgill (1966), Wyatt (1970), Bartoněk (1987a; 1987b), Hajnal (2007), among others. Cf. also the critical approach of Bile, Brixhe, and Hodot (1984) and the collection of papers contained in Brixhe and Vottéro (2006). For a recent critical re-assessment of the nature of South Greek (i.e. East Greek), see Ringe (2016). For an iconoclastic attempt to cast doubt on the unity of the Aeolic group, see Parker 2008, against which see García Ramón (2010), and extensively Scarborough (2023). For a critical overview of the state of the field of Ancient Greek dialectology in general, see García Ramón (2018). For a general survey of the Greek dialects see also Woodard (2008).
[15] For the notions of geographical dialect continua and dialect geography in general, cf. Chambers and Trudgill (1998:5–7, 13–31).
[16] Public domain map via Wikimedia Commons, after Horrocks (2006:71), in turn after Hall (1995:100) https://commons.wikimedia.org/wiki/File:AncientGreekDialects_(Woodard)_en.svg#/media/File:Ancient_greek_dialects-en.svg. Cf. Horrocks (2010:14).
_en.svg#/media/File:Ancient_greek_dialects-en.svg){kind=link}
[17] Much ink has been spilled on the old question of linguistic variation within Mycenaean, particularly with reference to the concepts of “mycénien normal” vs. “mycénien special” first distinguished by Risch (1966). See also Nagy (1968), Woodard (1986), Thompson (1996–1997), Palaima (1998–1999), Thompson (2002–2003). There is no widely agreed scholarly consensus on these phenomena currently. Some early studies attempted to invoke the at the time growing field of sociolinguistics to explain the differences as upper vs. lower class varieties of second-millennium Greek (Chadwick 1976, see also Wyatt 1970), but with a complete lack of evidence for non-Mycenaean Greek dialects in the second millennium BCE, a hypothesis of sociolinguistic class registers necessarily must remain speculative. One possible explanation, following an idea proposed by Palaima (2011:124–125), may be connected with the suggestion that the profession of Linear B tablet-writing was treated like other craft skills and passed on from parent to child, possibly of ‘Minoan’ ethnic origin. In such a case, one could imagine a dialectal continuity among the tablet-writers as they came to serve at different sites. On the other hand, if the tablet-writers spoke a different dialect from other people around the regional palatial centers, one would expect to see a lot more variation from other dialects creeping in over generations. I do not believe the question of the apparent uniformity of the Mycenaean dialect can be satisfactorily resolved at this time.
[18] See, e.g., the approaches of Hoffmann (1891–1898), Kretschmer (1909), among others.
[19] Cf. Hall (1997:34–66) and passim.
[20] Cf. Lejeune (1972:43–50), Rix (1976:85–88), Sihler (1995:160–165). See also Woodard (2012).
[21] As a representative of late nineteenth and early twentieth century scholarship in this area, cf. Meillet (1930:28–29), Schwyzer (1939:293–296).
[22] Cf. GEW, DELG, EDG s.v. τέσσαρες.
[23] The Attic equivalent would be *Δελφαίου. In the context of the inscription this is a pars pro toto referring to the temple of Apollo. Apollo Delphaios is also known from a third century dedicatory inscription from Larisa reported by Gallis (1975: 303): Ἀπλουνι Δελφαίου(i.e. Ἀπόλλωνι Δελφαίῳ).
[24] Cf. GEW, DELG, EDG s.v. Δελφοί, ἀδελφεός.
[25] Cf. GEW, DELG, EDG s.v. θήρ.
[26] Cf. GEW, DELG, EDG s.v. βουκόλος; Matasović (2009) s.v. *bow-koli-.
[27] Cf. GEW, DELG, EDG s.v. ἀμφίπολος.
[28] This discrepancy between the compositional member -πολος versus -κόλος was explained in this way already by de Saussure (1889:161–162). This dissimilation likely predated Proto-Greek itself, if it was not already of Proto-Indo-European date; see Weiss (1994), Beekes and de Vaan (2011:62). The compound formation found in Proto-Celtic *bou̯-koli- via Middle Irish búachaill and Middle Welsh bugail ‘shepherd’ shows that this sound change also occurred in the prehistory of Celtic (cf. Matasović 2009:72).
[29] Because I am primarily concerned in this article about the spread of areal sound changes and their interaction within contiguous dialect geography, for our present purposes I omit discussion of the developments in Cypriot, for which see the thorough treatment of Egetmeyer (2010:205–213). The outcomes in Cypriot may potentially be different from the rest of Greek in this environment because, e.g., it attests si-se/sís/ (ICS 217.10, etc.), instead of Pan-Greek τίς. See also Hesych. σ 570 (Hansen) σί βόλε· τί θέλεις. Κύπριοι.
[30] For these and more textbook examples, cf. Lejeune (1972:43–50), Rix (1976:85–88), Sihler (1995:160–165). Note that these examples ignore the cases where labials are found for reasons of paradigmatic levelling, e.g. 1.sg. ἕπομαι ‘I follow’ < *sekʷ-o-, but 3.sg. ἕπεται ‘s/he, it follows’ instead of †ἕτεται expected from the strict application of the sound laws < *sekʷ-e-, cf. Sihler (1995:164).
[31] Cf. GEW, DELG, EDG s.v. τε, LIPP s.v. 1. *-(s)kʷe ‘und’.
[32] Cf. GEW, DELG, EDG s.v. ἀδελφεός.
[33] Cf. GEW, DELG, EDG s.v. θερμός.
[34] Cf. GEW, DELG, EDG s.v. τίς, LIPP s.v. 1. *kʷó-, *kʷí- ‘wer?; irgendwer’.
[35] Cf. GEW, DELG, EDG s.v. βίος. Regarding the nominal formation of βίος see Klein (1988:268–269).
[36] Cf. GEW, DELG, EDG s.v. ὄφις. This is possibly the only certain example of this environment in Greek. Due to scarcity of attestation the dialectal distribution of this reflex is not entirely certain outside of Attic-Ionic but it is positively attested in the West Greek dialect used at Epidauros (ὄφιος IG IV²,1 121.113, ὄφιν IG IV²,1 122.78) indicating that West Greek and Attic-Ionic agree in this outcome.
[37] Cf. Hock (2021:82–87), Kümmel (2007:250–266).
[38] I follow Carbon and Clackson (2016:140) in referring to this letter as tsan. There is not currently a dedicated Unicode code point for this letter. I have transcribed it using <Ͷ> Greek Capital Letter Pamphylian Digamma (U+0376) as an approximation to its appearance in IG V,2 262. Carbon and Clackson (2016:140) transcribe with <ϟ> Greek Small Letter Koppa (U+03DF) which is a closer approximation to the epigraphic appearance of the character as appears in the new festival calendar. IG V,2 and Buck (1955:198) transcribe using a sigma with a combining macron below <σ̱>. See also the discussion of Jeffery (1990:212–213).
[39] The secondary provenance of the bronze tablet remains very uncertain. In the initial pre-publication by Heinrichs (2015) it is only given as “Allegedly it was in English private ownership and acquired around 1965 in a London flea-market for a moderate sum” (Heinrichs 2015:3n7).
[40] See especially at Pheneos (Dubois 1986b:195–202, Phé 1): ὅζις for ὅτις (l.4), ζ᾽ τ᾽ (l.4) possibly to be interpreted as an elided τε, and ζτε̄́ραιον possibly reflecting *θήραιον (l.1) and if so a rare example of a reflex of Proto-Greek *kʷʰ. For this inscription see Dubois (1986b:195–202, Phé 1) with detailed linguistic commentary and references to previous literature, cf. also Buck (1955:196–197, No.16), Jeffery (1990:208–209, Pl. 40 No.2). In addition to these, there is possibly a mixed form τζετρακάτιαι (Attic τετρακάσιοι) in the Arcado-Laconian Bronze from Tegea (IG V,2 159.10) where the first compound element τζετρα- shows an initial reflex potentially comparable to the spellings found at Pheneos, but the non-assibilated ‘hundreds’ morpheme proper to West Greek (cf. Dubois 1986a:67).
[41] The use of tsan in the form ἀπυͶεδομίν[ος] is difficult to explain since etymologically δε- is expected from the reduplicated syllable of the perfect stem. This may well be a hypercorrection where the letter-cutter might not be so familiar with the use of this unusual character but see Dubois (1986a:69–70), Dubois (1986c:24n435), and Duhoux (2006:33–36) for discussion of different possibilities of interpretation for this form.
[42] Cf. Carbon and Clackson (2016:141). On the outcomes of the cluster *-tu̯- in Greek, cf. Lejeune (1972:82–83), Rix (1976:93), Sihler (1995:186). Later Arcadian regularly shows -σσ- < *-tu̯-, cf. Dubois (1986a:78–79).
[43] For a fuller discussion of examples with full epigraphic references, cf. Dubois (1986a:65).
[44] Cf. GEW, DELG, EDG s.v. τιμή, τίω.
[45] The actual etymology of Arcadian -δέλλοντες is somewhat more complicated. It is generally accepted that the Arcadian variant reflects an initial labiovelar on the basis of this form, which is usually considered to be back formed from a full-grade thematic aorist *(h₁e)-gʷélh₁-twhich was secondarily thematized (LIV² 208n4 s.v. *gʷelh₁–, cf. Harðarson 1993:180, EDG s.v. βάλλω), where the original strong-stem thematized root aorists are only circumstantially attested via three unattributed glosses from the lexicon of Hesychius: ἔζελον· ἔβαλον (ε 597 Latte-Cunningham), ζέλλειν· βάλλειν (ζ 106 Latte-Cunningham), and κάζελε· κατέβαλε (κ 73 Latte-Cunningham). The present stem βάλλωattested elsewhere either reflects a secondary thematization of the weak-stem nasal present *gʷl̥-n-h₁– > *gʷalne- with analogical retention of -n- and later assimilation to -λλ- (LIV² 208n4 s.v. *gʷelh₁– following Harðarson 1993:161n69), or simply a Proto-Greek zero-grade yod-present *gʷal-i̯e/o- (cf. EDG s.v. βάλλω). Unfortunately, the only external comparison from outside of Greek to decide between the possibilities for the original derived present stem (Old Irish atbaill ‘dies’) remains uncertain (LIV² s.v. *gʷelh₁–, EDG s.v. βάλλω, cf. also Schumacher 2004:211–212 s.v. Proto-Celtic *bal-ni-, pace Matasović 2009:53 and Vendryes 1959:98). Regardless of these broader etymological issues, the assumption for root-initial *gʷ and secondary full-grade root vocalism in the Arcadian present stem is still likely to be the best explanation for ἐσδέλλοντες.
[46] See also Dubois (1986a:65).
[47] For a survey of the linguistic features characteristic of the Aeolic subgrouping see Méndez Dosuna (2007:460–472), Woodard (2021:287–337), and with a focus on the innovations Scarborough (2023:60–129). Woodard (2021) passim additionally discusses a substantial amount of comparative cultural and archaeological evidence arguing for a westward spread of the Aeolic dialects from Asia Minor. While this hypothesis differs from my own views on the spread of the Aeolic dialects from mainland Greece (Scarborough 2023 passim, based on relative linguistic chronology), a westward spread of the Aeolic dialects from Asia Minor would still be compatible with the linguistic argument for the Aeolic dialects labializing the labiovelars earlier and independently of the other Greek dialects which I will advance further in this paper.
[48] Cf. Buck (1955:153–154). Thumb and Scherer (1959:21–22).
[49] See also Arcadian ἀπυτεισάτω in exx. (17) above.
[50] There are additional attestations of the same root, predominantly in reference to the sanctuary of Artemis Thermia at Mytilene in later Roman period dialect-revival inscriptions. See Scarborough (2023:86) and Scarborough (2023:86n134) for attestations and further discussion.
[51] Matras (2007:54–56). Stephens and Woodard (1986:147) consider a loan of τε to be unlikely and assume that the specific treatment of τεin Aeolic was due to its status as a clitic but note the observation of Parker (2013:134n23) that the dialects do vary in their use of connectives and that Mycenaean exclusively uses -qe to the exclusion of καί. It would not be unthinkable to imagine a possibility where the clitic conjunction fell out of use in favor of another coordinating conjunction. See also recently García Ramón (2020:306) arguing for supradialectal influence in the use of τε to the exclusion of unattested †πε.
[52] In the results of the Loanword Typology Project carried out at the former Department of Linguistics of the Max Planck Institute for Evolutionary Anthropology the semantic concept ‘who?’ was found to be within the 100 most borrowing-resistant concepts in the meaning list used in the in the overall study (Tadmor 2009:65–68), cf. The World Loanword Database (WOLD) at https://wold.clld.org/.
[53] Observed already by Lejeune (1972:50), see also Stephens and Woodard (1986:147). Note that East Thessalian shows κίς, κί which is usually interpreted as a secondary dissimilation in proximity to a round-back vowel in collocations like *οὐ-κʷίς, cf. Lejeune (1972:45), Rix (1976:86). Elsewhere in Thessaly τίς is attested, e.g. in IG IX,2 257 (5th c. BCE, Thessaliotis), SEG 36:548 (3rd c. BCE, Hestiaiotis, Metropolis), SEG 31:572 (2nd c. BCE, Pelasgiotis, Krannon), and Béquignon (1935: 37–51, No.1) (2nd c. BCE, Pelasgiotis, Krannon).
[54] For a typology of palatalization processes, cf. Kümmel (2007:250–266).
[55] Cf. Sihler (1995:164) “There is no obvious explanation for the discrepancy between the behavior of *kʷ and the other two labiovelars” echoing the earlier statement of Buck (1933:129) “There is no satisfactory explanation of this divergence from the development of qʷ to τbefore ι in τίς, τῑμή, etc.”
[56] For Parker’s attempt to dismember the Aeolic group, see Parker (2008), against which see especially Scarborough (2023:40–49) and passim.
[57] Note that later Attic Greek did not permit phonetic sequences *Cʷu or *Cʲi (Ohala 1994:372, Allen 1958). Cf. Padgett (2001) discussing similar processes in palatalizing environments in Russian and Irish.
[58] So persuasively argued for by Ohala (1994:377).
[59] For the details of Sabellic and Celtic, see respectively Meiser (1986:79–92), McCone (1996:67–70). For a typology of labialized labiovelars, cf. Kümmel (2007:274–275).