- This was not particularly true in the case of the recently deciphered Greek inscriptions in the Linear B script, which have rather confirmed our suspicions about earlier stages of Greek than forced any fundamental re-evaluation.
- This kind of analysis of the functional and structural aspects of phonological change is most closely associated with the name of André Martinet; the most complete treatment of the theory and application of the approach is to be found in his Économie des changements phonétiques (Berne, 1955). For ealier works see also Alphonse Juilland, “A Bibliography of Diachronic Phonemics,” Word 9 (1953), 198–208. An important analysis of the vowels of Greek is to be found in Martin S. Ruipérez, “Esquisse d'une histoire du vocalisme grec,” Word 12 (1956), 67–81.
- One of the most thorough examples of this important and fruitful kind of analysis is that of A. C. Moorhouse, “Observations on Chronology in Sound-Changes in the Italic Dialects,” American Journal of Philology, 61 (1940), 307–329. For Greek, see the entry Chronologie Belatiue in the index to Michel Lejeune, Trailé de phonétique grecque2, (Paris 1955).
- This procedure is particularly well exemplified in Martinet's work on Italic (Économie, §§ 13.13–13.37, Affaiblissement et affermissement en italique) and in J. Fourquet, Les mulalions consonanliques du germanique (Paris, 1948).
- A study of the effects of the laryngeals on the morphophonemics of the Greek verb, now under way, should provide more information on the behavior of the laryngeals in Greek.
- The term “structure” is used here in the meaning of the paradigmatic interrelation of the phonemes in terms of their shared features, rather than in the Bloomfieldian sense of the pattern of their distribution in words.
- For matters of fact about which there is no dispute, reference is made to Carl Darling Buck, Comparative Grammar of Greek and Latin (Chicago, 1948); Michel Lejeune, Traité de phonétique grecque2 (Paris, 1955); and Eduard Schwyzer, Griechische Grammatik I (Munich, 1938). The three are in substantial agreement except on occasional minor details. In general the Traité of Lejeune, which offers the most detailed treatment of the various phonetic problems, will be used to represent the agreed opinion. There has now appeared an article by W. S. Allen, “Some Problems of Palatalization in Greek,” Lingua VII, 2 (1958), 113–133, that takes up some of the problems dealt with here, but from a somewhat different point of view.
- See, for example, A. Meillet, Le slave commun2 (Paris, 1934), §§ 102–119, Mouillures et amollissements.
- Op. cit.
- Traité de phonétique (Paris, 1933).
- “The Problem of Old Bulgarian št,” Word 11 (1955), 228–236.
- It should be noted that the treatment of voiceless stop plus y differs in no way from that of aspirate stop plus y. Cf. Lejeune, op. cit., §§ 62–63. E.g., kléplō ‘steal’ <*klepyō (cf. klopós ‘thief’), krúptō ‘hide’ < *kruphyō (cf. krúpha ‘secretly’). So also the dorsal and labio-velar orders are not differentiated in the product of palatalization. Cf. Lejeune, ibid., § 29. E.g., leússō ‘look’ <*leukyō (cf. leukós ‘bright’), péssō ‘boil’ <*pek yō (cf. Latin coquo ‘cook’). In dealing with problems of palatalization before y, no separate treatment will be accorded the aspirates and labio-velars.
- Op. cit., § 87. We shall discuss later the problems involved in *ts and *tw, which Lejeune includes here because they are involved in the eventual phonetic result of *ty and *kg.
- Ibid., § 62.
- Ibid., §§ 36–43, Premiers symptômes d'un relâchement de l'occlusion. This development will be discussed in greater detail below.
- Ibid., § 62, note 1.
- Ibid., §§ 92–93.
- The normal treatment of d+s is of course is, with assimilation of the d to the s; e.g., nominative singular pats ‘child’ < *pails < paid-s, genitive singular paid-ós.
- There is really no question that the phonetic value of the letter zeta was not zd in Attic and Ionic; see for example the arguments presented in Buck, Comparative Grammar of Greek and Latin, § 185, orin Schwyzer. Griechische Grammatik, pp. 329–332.
- The term survives even in Lejeune, but it is clear from his discussion of the process (“interversion par penetration”) that he does not use it in the older sense. Allen too (op. cit., fn. 73) discourages the use of the term.
- So Buck, The Greek Dialects (Chicago, 1955), § 74. Schwyzer (op. cit., p. 273) discusses both terms, and describes the process in a way that has much in common with the interpretation I shall give below: “… das i palatalisierte zunächst r, n usw., die palatale Farbung teilte sich dann aber auch dem vorhergehenden Vokal mit; schliesslich führte sie hier zu vollem i, während sie sich bei n r w verlor.”
- If this pattern represented a strictly phonemic analysis from a synchronic point of view one would have to deal with the problem of whether a palatalized geminate such as pp is distinctively both a geminate and palatalized, since there is no single palatalized labial to which it can be opposed. But in considering the evolution of a language it is important to emphasize the phonetic characteristics of the forms as well as their distinctive features. The voiceless aspirates have been left out of this representation; as was noted above (fn. 12) the voiceless stops present the same development. The clusters kw and gw are included on the assumption that they are, in effect, the geminates of the labiovelare kw and gw; kw> pp in hippos ‘horse’ as kw> p in hépomai ‘follow’. For gw there is no example in medial position to show the geminate, but for the initial position there is thér ‘wild beast’, Lesbian phér, Latin lerus, Lithuanian žverìs, the Lithuanian form showing clearly that there is an initial cluster, not a single labio-velar; cf. Lejeune, op. cit., § 66. As noted above (in. 16) there is no clear example of a reflex of by, nor is there one of my; cf. Lejeune, § 142. For the relatively rare gg (usually spelled kg, not gg) see Schwyzer, op. cit., p. 317.
- Économie, § 11, La lenition en cellique et les consonnes du roman occidental.
- For this term cf. Martinet, ibid., §§ 2.28–2.29.
- The word ‘second’ is not used here in a strictly chronological sense. Although this period lasted long after that of palatalization, it may of course have been contemporary with it also.
- Traité de grammaire comparée des langues classiques2 (Paris, 1948), § 51. Elsewhere (Introduction à l'étude comparative des langues indo-européennes (Paris, 1937), p. 34) Meillet describes how such a change would affect s: « Soit par exemple le phonème s, qui suppose une élévation de la langue près des dents, avec écoulement d'air constant, et qui est constitué par un sifflement: si la langue est relevée d'une manière insuffisante, il devient un simple souffle, c'est-à-dire h, le bruit du frottement de l'air entre la langue et les dents disparaissant ».
- Thus * py>pl (kléplō < *klepyo) in all dialects; *ty> tt in Boeotian and elsewhere, although because of special circumstances to be discussed below it became s in Attic, Ionic, and Arcadian (Boeotian méttos = Attic mésos ‘middle’); *ky> tt or ss (* phulakyo> Attic phulá;llō, Ionic phulássö); *dy and *gy> zd (spelled zeta) or dd (*medyön> Ionic mézōn, Boeotian méddonos ‘greater’).
- For theoretical remarks concerning the kinds of sounds that are likely to co-exist in a language, see Roman Jakobson, “Typological Studies and Their Contribution to Historical Comparative Linguistics,” Supplement to Reports for the Eighth International Congress of Linguists (Oslo, 1957).
- Cf. Buck, Comparative Grammar, § 129.
- Ibid., § 149.
- E.g., *s>h: hépomai = Latin sequor; *sss, *ts, and *ty>s: dative plural gènesi < * ge nessi (Homer has both genesi and génessi), sómasi < * sōmal-si, mésos< * methyos.
- Cf. A. A. Macdonell, Vedic Grammar (Strassburg, 1910), §§ 30 and 40; W. D. Whitney, Sanskrit Grammar2 (Cambridge, Mass., 1950), § 227.
- Économie, § 7.
- The labial order has been omitted; it will be discussed separately. The clusters ts and zd have been included in order to indicate their relevance to later developments.
- This preservation in the voiceless section of the pattern of a distinction that is lost in the voiced section is not without parallel elsewhere, where palatal articulations are involved. In Old Church Slavic the reflexes of the voiceless labio-velar before a front vowel and of the voiceless dorsal stop before y are kept apart, but the corresponding voiced sounds fall together: In the pairs reko, rečeši and pzsati, pišo there is a fourfold distinction between k, č, s, and š, but in lzgati, lžo and vezali, vežo there are only g, ž, and z; the expected *dž in *lzdžo has fallen together with ž. On this point see Meillet, Le slave commun, § 109. In Sanskrit there is a parallel development. The c of cakāra (<*kwekwor-) is distinguished from the ś of śuśrāva (<*kukrav-), but the j of jaghāna (<*gweghw on-) has merged with that of jajāna (<* gegon-). An experiment in the perception of affricates and hushing sibilants conducted by Louis J. Gerstman (Perceptual Dimensions for the Friction Portions of Certain Speech Sounds, New York University Dissertation, 1957) reveals that in this area voiced sounds are distinguished with less accuracy and consistency than are voiceless. See especially his Figure 2, Chapter 7.
- Économie, §§ 3.35–3.37.
- E. g. Boeotian méttos = Attic mésos (< *methyos); komittámenoi, aorist participle <*komitsamenoi; phuláttō < * phulakyō.
- E. g. Boeotian méddonos = Ionic mézōn (<* megyōn).
- Boeotian theózotos, for theósdotos, does not become *theóddotos.
- That the first element or a cluster is a weak positwn is suggested by developments in other languages too. In Avestan, the same kind of general relaxing of the force of articulation as in Greek is suggested by such changes as s>h, voiceless aspirates> voiceless spirants, and by the fact that word final i seems to have a spirantic pronunciation. A voiceless unaspirated stop is generally unaltered, but as first element of a cluster it becomes a spirant: hō ‘he’ = Sanskrit sas; kafəm ‘foam’ = Sanskrit kapham; katārō ‘which of two’ = Sanskrit kataras, but xratuš ‘wisdom’ = Sanskrit kratus. Similarly in Oscan and Umbrian, where the same process is at work (cf. Martinet, Économie, § 13), voiceless stops that are ordinarily preserved intact are spirantized as first elements of a cluster: Latin scriptus = Oscan scriftas = Umbrian screhto.
- E. g. Ionic phulàssō = Boeotian and Attic phulátlō.
- In an inscription from Halicarnassus (GDI 5726; Buck, Greek Dialects, 2) the spellings AlikarnaTeōn and Alikarnasseōn occur for the same form.
- From * holyos comes Homeric hóssos and hósos, but the ss of words of the type phulàssō is never simplified. Cf. Buck, Comparative Grammar, § § 182–183. On the pronunciation of this form, see Buck's note 6 to §.182, where he suggests it was a fricative of some kind.
- See the quotation from Lejeune above, p. 5.
- Arcadian mésos (s<*thy), edikásamen (s<.*ts), mézōn (z<*gy).
- This fact has recently been connected with the retention of the labio-velar in the Linear B texts (e.g. Buck, Greek Dialects, § 68, 3a), but that such an association should be made is perhaps not entirely warranted. The Linear B texts show the labio-velar preserved intact. That being the case it is difficult to see why the development in any one dialect should be regarded as having a peculiar connection with the earlier form. In reality, the only connection Arcadian and Linear B have in common, insofar as the labio-velars are concerned, is that they both differ from Attic-Ionic; but then Attic-Ionic and Linear B both differ from Arcadian.
- IG V. II, 262; Buck, Greek Dialects, 17.
- Cf. Lejeune, op. cit., § 72.
- E. g. śis = tis (atinL quis), eiśe = ette.
- E.g. zéllein for bállein in a gloss. For details see Lejeune, op. cit., § 34.
- Unfortunately this inscription from Mantinea does not happen to contain an example of the geminate; it would have been interesting to see how it would have been handled by so sensitive an engraver.
- Cf. Lejeune, §§ 27–35.
- Once we become aware of the existence of a palatal order at a fairly late stage, other apparently anomalous developments begin to be more understandable. It will be suggested below that there was a general east-west split (one is tempted to call it centum/satəm) among the Greek dialects, those in the east being characterized by the presence of a palatal order. Buck (Greek Dialects, § 61) notes that in spite of some confusion in distribution the change of ti to si and of tu to su (as in Doric didōti ‘he gives’, Attic didōsi; Doric tu ‘you’, Attic su) is on a distinct dialectal basis, “the retention of t being a notable characteristic of the West Greek dialects.” Should we connect the presence of an only partially illled palatal order with the palatalization of t before high front vowels (ti>ši>si)? Further, in the Aeolian dialect of Lesbos (whose pattern in this respect is identical with that of Ionic and is therefore safely “satəm”) there is a special development of d before i to a sound written zela, which in this case could not represent sd since the Lesbian alphabet indicated that cluster with the spelling sd: Lesbian zá = diá, kárza = kardia, but úsdos for ózos. Does this too represent a filling out of the palatal order?
- Cyprian sis = Arcadian śis = Attic lis. Cf. Buck, Greek Dialects, § 68, 3.
- For the classification Attic-Ionic see Buck, ibid., §§ 1–2.
- Attic mésos <.* thy-, meizōn <* -gy-, phulá;ttō <*-ky-.
- Euboean préttō ‘finish’ = Ionic préssō = Attic práttō.
- Cf. Buck, Greek Dialects, §§ 1,217.
- Aeolian includes Lesbian, Thessalian, and Boeotian.
- The term Doric is sued here in the more general sense of West Greek. The assumption that tt as well as dd was originally a general Doric feature is adopted here. The slight inscriptional evidence (cf. Schwyzer, op. cit., p. 318) is further supported by the appearance of ll-dd in Doricized Boeotian and by the extension to Attic and Euboean in terms of the present analysis. W. S. Allen (op. cit., p. 176) sees a partial “phonologisches Sprachbund” between Attic and Boeotian.
- That is, Arcadian, Attic, Ionic, and Aeolian. The exact relation among these dialects is still a matter for debate.
- Exceptions are found only in Cyprian and a single instance in Elean; see Buck, Greek Dialects, § 74b.
- The transfer of palatal quality follows very closely the examples of this phenomenon given for Old Irish by Martinet (Économie, § 7). This development provides an example of what Uriel Weinreich, in discussing the intersection of syntagmatic and paradigmatic factors, calls the unwrapping of a bundle of simultaneous features. Cf. his “On the Description of Phonic Interference”, Word 13 (1957), 1–11, esp. pp. 8–9.
- So also *ry> *ŕŕ> 'r, *kharyō> *khaíró> khaíró ‘rejoice’. For further examples see Lejeune, op. cit., § 142.
- In Attic the newly created intervocalic i seems to drop, giving *-oo, which contracts to a long omicron, spelled -o in earlier inscriptions, -ou later. The dropping of -i- here is comparable to the evolution of denominative s-stem verbs: *teles-yō> *teleśśō> *leleisō> Homeric teleiō, Attic teléō ‘complete’.
- In Old Irish, (or example, there was a distinction between labialized, palatalized, and neutral consonants. Cf. Rudolf Thurneysen, A Grammar of Old Irish (Dublin, 1946), §§ 85–88. Martinet finds traces of such a development in Balto-Slavic. Cf. Économie, § 13.55.
- Choose new content alerts to be informed about new research of interest to you
- Easy remote access to your institution's subscriptions on any device, from any location
- Save your searches and schedule alerts to send you new results
- Export your search results into a .csv file to support your research