Introduction
Since most Greek aorist stems have an etymological zero-grade root, the aorist is an important source of reflexes of *rÌ¥. Although the extent to which thematic and reduplicated aorists reflect PIE formations may be debated, it is certain that reconstructions like *likw-e/oâ (λιÏεá¿Î½) and *drÌ¥k-e/oâ (δÏακεá¿Î½) pre-date the vocalization of the syllabic liquids.
8.1 The Evidence
Aorists with a root of the phonological shape /CraC-/ are presented in Table 23 (when attested in Homer, no first attestation is indicated) on p. 357.1 Aorist stems (attested in Homer, classical Attic or both) with a root of the phonological shape /CarC-/ are presented in Table 24 (on p. 358).
Of the verbs mentioned in these Tables, the following forms have no bearing on the issue of the double reflex of *r̥ in Ionic-Attic:
-
In six aorist stems, the reflex of *rÌ¥ may have been influenced by the full grade in a corresponding present or perfect stem: á¼Î´Ïαμον (âδέδÏομε), δÏαÏών (δÏá½³ÏÏ), á¼ÏαÏδον (Ïá½³Ïδομαι), á¼Ïá½±ÏÏην, ÏαÏÏώμεθα and ÏεÏá½±ÏÏεÏο (Ïá½³ÏÏομαι), á¼ÏÏαÏον (ÏÏá½³ÏÏ), á¼ÏÏá½±Ïην and á¼ÏÏαÏον (ÏÏá½³ÏÏ).
-
No conclusions regarding the regular Ionic-Attic vocalization of *rÌ¥ can be based on the form á¼ÏÏÎ±Î´ÎµÏ in Sophron, the 5th c. Syracusan poet who composed mimes in a form of literary Doric.2 The normal aorist of Attic Ïá½³Ïδομαι âto fartâ was á¼ÏαÏδον (mostly with preverb). On the other hand, for assessing the regular vocalization in Syracusan or Corinthian the form á¼ÏÏÎ±Î´ÎµÏ is highly relevant (see chapter 3).
-
The Homeric hapax á¼Î½á½³ÎºÏαγον (Od. 14.467) contains a secondary zero grade beside the full grade CRÄCâ in the pf. κέκÏᾱγα, which is either onomatopoeic or reflects a root *krh2gâ.3 The model for such a reshaping was provided by roots of the shape *Ceh2Câ, e.g. Ïá½µÎ³Î½Ï Î¼Î¹, á¼Ïάγην.4
-
Another secondary zero grade is á½ÏεÏÏάγη, which stands beside the root ῥηγâ < *u̯reh1ǵâ (LIV2 698).
-
The relation between the present stem ÏÏá½½Î³Ï and aorist stem á¼ÏÏαγον cannot be understood in terms of Indo-European morphophonology. Assuming Å-vocalism in a thematic root present would be entirely unmotivated. Hackstein (1995; taken over by LIV2: 647) reconstructs a PIE root *trh3gâ on the basis of a comparison with Toch. B treṣṣäṠâchewsâ; in this case, the zero-grade vocalism of á¼ÏÏαγον must be secondary. If one insists on reconstructing a Proto-Greek thematic aorist stem *trÌ¥g-e/oâ, the vowel slot of ÏÏαγâ may have been influenced by that of the present stem ÏÏώγÏ. However, the vocalism of á¼ÏÏαγον may also have been influenced by its near-synonym á¼Ïαγον.
-
Homeric á¼Î²ÏαÏε does not have a convincing etymology; it is therefore uncertain whether its pre-form contained *rÌ¥.
-
The Homeric aorist á¼ÏÏαον âdashed, attackedâ has no ascertained cognates either (its relation with ÏÏÎ±á½»Ï âto glance offâ is uncertain), and there is no indication that the pre-form contained *rÌ¥.
-
The same holds for Ïá½³ÏÏαδε, which belongs to a non-ablauting root ÏÏαδâ without established etymology.
-
The Homeric reduplicated aorist ÏεÏá½±ÏÏεÏο and the 1st plural subjunctive forms ÏÏαÏείομεν < *trÌ¥pÄ-o-men and ÏαÏÏώμεθα (all belonging to Ïá½³ÏÏομαι) have been discussed in section 6.8.5. It was shown there that ÏαÏÏώμεθα and ÏεÏá½±ÏÏεÏο may be analogical creations, and that ÏÏαÏείομεν has the reflex âÏαâ of Epic *rÌ¥.
This leaves us with three Homeric aorists with a root of the shape /CraC-/ (from *CrÌ¥C-e/oâ) that cannot be an analogical reflex: á¼Î´Ïαθον (no cognates with a full grade root), á¼Î´Ïακον (δέÏκομαι) and á¼ÏÏαθον (Ïá½³ÏθÏ). At first sight they seem to contradict our hypothesis that âαÏâ is the regular Proto-Ionic reflex of *rÌ¥.
Table 23
Aorist formations with âÏαâ in Greek
|
Aorist with âÏαâ |
Other attested formations |
|---|---|
|
á¼Î²ÏαÏε âresoundedâ |
no clear cognates |
|
καÏέδÏαθον âwent to sleepâ |
Att. καÏαδαÏθάνÏ, aor. καÏέδαÏθον |
|
á¼Î´Ïακον âbeheld, lookedâ |
pres. δέÏκομαι âlookâ, pf. δέδοÏκα |
|
á¼Î´Ïαμον âranâ |
pf. á¼Î½Î±â, á¼Ïι-δέδÏομε âruns up / overâ |
|
ptc. δÏαÏών (Pi.)5 |
pres. δÏá½³ÏÏ âreapâ |
|
á¼Î½á½³ÎºÏαγον âspoke upâ |
pf. κέκÏᾱγε âshrieksâ (post-Hom.) |
|
á¼ÏÏÎ±Î´ÎµÏ (Sophr.)6 |
pres. Ïá½³Ïδομαι âfartâ, aor. á¼ÏαÏδον |
|
á¼ÏÏαθον âpillagedâ |
pres. Ïá½³ÏÎ¸Ï âpillage, destroyâ |
|
(á½Ïâ)εÏÏάγη âbroke upâ |
pres. á¿¥á½µÎ³Î½Ï Î¼Î¹ âbreakâ, pf. á¼ÏÏÏγα |
|
á¼ÏÏαγον âateâ |
pres. ÏÏá½½Î³Ï âgnaw, chew, eatâ |
|
ÏÏαÏείομεν 1pl. subj. < *trÌ¥pÄ-o-men |
pres. Ïá½³ÏÏομαι âto enjoy oneselfâ |
|
á¼ÏÏαÏον âturnedâ (tr., intr.) á¼ÏÏαÏόμην âturnedâ (intr.) |
pres. ÏÏá½³ÏÏ âturn, directâ, pf. mid.-pass. Ïá½³ÏÏαÏÏαι |
|
á¼ÏÏαÏον âwas raised, grew upâ á¼ÏÏá½±Ïην âid.; got thick/fat, etc.â |
pres. ÏÏá½³ÏÏ ârearâ mid. ÏÏá½³Ïομαι âget thick/fat, etc.â |
|
Ïá½³ÏÏαδε âshowedâ |
pres. ÏÏá½±Î¶Ï âshowâ, ÏÏάζομαι ânoticeâ |
|
á¼ÏÏαον âdashed, attackedâ |
(?) pres. ÏÏÎ±á½»Ï âto glance offâ |
There are, however, serious reasons to doubt that âÏαâ is the vernacular outcome of *rÌ¥ in these three thematic aorists. First of all, as we will see it is difficult to give a convincing analogical explanation for the reflex âαÏâ in Attic καÏέδαÏθον and ἥμαÏÏον (Hom. ἠμαÏÏε). Secondly, and crucially, á¼Î´Ïαθον, á¼Î´Ïακον and á¼ÏÏαθον are typical Homeric forms, unattested in Classical prose. For these reasons, we must consider the possibility that these forms contain the reflex of Epic *rÌ¥, following the framework elaborated in chapter 6.7
I will now first argue that the forms καÏέδαÏθον and ἥμαÏÏον display the regular reflex of *rÌ¥ in Proto-Ionic (section 8.2), then discuss the actual evidence for âÏαâ < *rÌ¥ in the type á¼Î´Ïαθον (section 8.3), and finally explain how such forms may have come into being within Epic Greek (section 8.4).
Table 24
Aorist formations with âαÏâ in Greek
|
Aorist with âαÏâ |
Other attested formations |
|---|---|
|
ἠμαÏÏε, Att. ἥμαÏÏον âfailedâ |
pres. á¼Î¼Î±ÏÏάνÏ; νημεÏÏá½µÏ âunfailingâ |
|
Att. καÏέδαÏθον âsleptâ |
pres. καÏαδαÏÎ¸á½±Î½Ï (Att.) |
|
Att. á¼ÏαÏδον âfartedâ |
pres. Ïá½³Ïδομαι (Att.) |
|
á¼Ïá½±ÏÏην âgot satisfactionâ ÏαÏÏώμεθα âletâs satisfy ourselvesâ ÏεÏá½±ÏÏεÏο, ÏεÏαÏÏώμεÏθα |
pres. Ïá½³ÏÏομαι âto enjoy oneselfâ |
8.2 The Regular Development *rÌ¥ > âαÏâ in the Thematic Aorist
8.2.1 Homeric á¼Î´Ïαθον versus Attic καÏαδαÏθάνÏ, καÏέδαÏθον
In Homer, the thematic aorist á¼Î´Ïαθον âwent to sleep; sleptâ is attested once as a simplex (Od. 20.143), but otherwise only with preverb: καÏέδÏαθον (5â¯Ã), ÏαÏέδÏαθον âlay down besideâ (2â¯Ã).8 After Homer, the aorist stem δÏαθâ is found only in the epic poet Antimachus and later in Hellenistic poetry (Theoc., Call.). The only genuine Attic form, on the other hand, is καÏέδαÏθον âslept, fell asleepâ (attested in Attic prose and Aristophanes, but absent from Ionic prose). Thus, we have a genre distribution: á¼Î´Ïαθον is epic, whereas καÏέδαÏθον is the non-poetic classical form.
Let us first consider whether the different vocalizations can be explained by influence of a full grade root. The thematic aorist καÏέδαÏθον / καÏέδÏαθον has no direct cognates, neither in Greek nor in other languages, and is therefore primary, at least from a Greek perspective. It could be an inherited formation in view of the similar Indo-European roots *dremâ (cf. Lat. dormiÅ âsleepâ, CS drÄmati âdoze, slumberâ) and *dreHâ (Ved. opt. 3sg. ni-drÄyÄÌt âto sleep, slumberâ).9 Thus, for the Greek aorist we may start from a zero grade root *drÌ¥dhâ.10
Is it possible to determine the full grade slot of this root? In the LIV2, Kümmel mechanically reconstructs a root *derdhâ, adding the comment: âfür Vollstufe I spricht die analogische R(z) gr. att. δαÏθâ.â However, although δÏαθâ is attested earlier, it does not follow that δαÏθâ arose as a reshaping. On the contrary, given the full grade slot of the Indo-European root variants *dremâ and *dreHâ just mentioned, one could argue for an original full grade *dredhâ, in which case Attic δαÏθâ must be the regular outcome of *drÌ¥dhâ.
In any case, to invoke the influence of a hypothetical ablauting full grade form is unwarranted: the only old formation within Greek is the non-ablauting thematic aorist, PGr. *drÌ¥th-e/oâ. This aorist was used in suppletion with εá½Î´Ï (Homer), ÎºÎ±Î¸Îµá½»Î´Ï (Classical Attic); beside this stative present, the aorist has complexive value. As Kölligan (2007a: 172) notes, the first author to attest the paradigm καÏαδαÏÎ¸á½±Î½Ï : καÏέδαÏθον âto fall asleepâ is Plato, who uses the new present stem to specifically refer to catching sleep as an ongoing process.11 That is, καÏαδαÏÎ¸á½±Î½Ï was based on the ingressive reading of the aorist καÏέδαÏθον, and is therefore unlikely to be of high antiquity. The same holds, mutatis mutandis, for the intransitive aorist καÏαδαÏθá¿Î½Î±Î¹.
Thus, the only form reconstructible for Proto-Ionic is the thematic aorist. Even if καÏέδαÏθον does not occur before the fifth century, it must be the regular reflex of PGr. *âdrÌ¥th-e/oâ in the Attic vernacular.12 The variant á¼Î´Ïαθον, on the other hand, must have an artificial epic reflex (see section 8.4 below).
8.2.2 á¼Î¼Î±ÏÏάνÏ, aor. ἥμαÏÏον and Homeric ἤμβÏοÏον
The present á¼Î¼Î±ÏÏá½±Î½Ï and the thematic aorist ἥμαÏÏον (á¼Î¼Î±ÏÏεá¿Î½) are attested in Homer and Classical Ionic and Attic alike. Beside these forms, Homer also has an aorist ἤμβÏοÏον, with an o-colored reflex that cannot have originated in the Proto-Ionic vernacular.13
Most scholars have explained the vowel slot of ἥμαÏÏον as analogical, invoking analogy with the full grade attested in νημεÏÏá½µÏ âunfailingâ. This is an emergency solution, because νημεÏÏá½µÏ is a fossilized nominal compound, attested only in early Greek epic and three times in Aeschylus. Since a relic nominal form can hardly be expected to have influenced the shape of the verbal stem in the vernacular, an analogical explanation of ἥμαÏÏον would be feasible only if full grade forms of the verb were still in use when *rÌ¥ vocalized in Proto-Ionic.
Such a scenario has in fact been proposed by Ruijgh (1992: 91). Being unable to explain the vocalization to âαÏâ in δαÏθάνÏ, he assumed that the present secondarily acquired a zero grade root, replacing *δεÏθάνÏ. The alleged model for introducing this zero grade is the aorist á¼Î´Ïαθον, which had a zero grade root all along. In a similar vein, Ruijgh claims that á¼Î¼Î±ÏÏá½±Î½Ï is secondary for *á¼Î¼ÎµÏÏá½±Î½Ï after a hypothetical *á¼Î¼(β)ÏαÏον, a form which itself was supposedly superseded by ἥμαÏÏον (after á¼Î¼Î±ÏÏάνÏ).
This scenario cannot be upheld. First of all, as we have just seen, the present καÏαδαÏÎ¸á½±Î½Ï is probably a late creation based on the aorist καÏέδαÏθον. Similarly, á¼Î¼Î±ÏÏá½±Î½Ï beside the aorist ἥμαÏÏον follows a productive pattern and looks like a relatively recent formation. Secondly, a full grade root would be out of place in an inherited nasal present: wherever such a full grade nasal present is attested, it must have been influenced by the aorist (cf. Î´Îµá½·ÎºÎ½Ï Î¼Î¹ âpoint outâ beside á¼Î´ÎµÎ¹Î¾Î±, Ïá½³Ïνημι âsellâ beside á¼Ïá½³ÏαÏα).14 Thirdly, the assumed chain of analogical influences is too complicated to be credible: supposing the existence of an aorist *amrat-e/oâ, the a-vocalism would first have spread into the present stem, but maintaining a different vowel slot (á¼Î¼Î±ÏÏάνÏ); after that, the vowel slot of the present stem would have been introduced into the thematic aorist.
Clearly, scenarios like the one advocated by Ruijgh are developed only in order to maintain the claim that âαÏâ (in forms like á¼Î¼Î±ÏÏεá¿Î½) cannot be the regular reflex of *rÌ¥. In reality, the only straightforward way to explain δαÏÎ¸á½±Î½Ï and á¼Î¼Î±ÏÏá½±Î½Ï is to assume that these presents were created (or reshaped) on the basis of the corresponding thematic aorists after the vocalization *rÌ¥ > âαÏâ had taken place in Proto-Ionic. The vernacular form á¼Î¼Î±ÏÏá½±Î½Ï was then also introduced into Epic Greek.
It remains to explain the origin of epic ἤμβÏοÏον, the alternative thematic aorist form. It is generally assumed to be of Aeolic origin, because an infinitive αμβÏοÏην is attested in epigraphic Lesbian (see section 3.3.2). Indeed, thus far this has been the only way to make sense of the combined appearance in ἤμβÏοÏον of the reflex âÏοâ < *rÌ¥ and psilosis. In section 7.2.4, however, I pointed out that ἤμβÏοÏον can be analyzed as the regular epic reflex of an augmented pre-form *ÄmrÌ¥ton. This *ÄmrÌ¥ton was a traditional element of Epic Greek. In spoken Ionic-Attic, the same form vocalized with âαÏâ and a non-etymological initial aspiration was added, yielding á¼Î¼Î±ÏÏάνÏ, ἥμαÏÏον. These vernacular forms were introduced at some point into the epic tradition.15 When Epic *rÌ¥ developed to âÏοâ after labial consonants, the ensuing form *Ç£mroton > ἤμβÏοÏον no longer resembled the Ionic aorist, so there was no reason to introduce the aspiration.
Thus, the similarity between epic ἤμβÏοÏον and Lesbian á¼Î¼Î²ÏοÏε (αμβÏοÏην) might be accidental. The pair ἥμαÏÏον beside ἤμβÏοÏον is another Homeric doublet consisting of an Ionic vernacular form and an artificial epic form (compare καÏδίη : κÏαδίη, Ïá½³ÏαÏÏÎ¿Ï : Ïá½³ÏÏαÏοÏ, etc.).
8.3 The Pattern of Attestation of the Thematic Aorists with âÏαâ
If âαÏâ is the regular outcome of *rÌ¥ in á¼Î´Î±Ïθον and ἥμαÏÏον, the appearance of âÏαâ in a number of other aorists must be accounted for. The focus of this section will be on the three forms singled out earlier in this chapter, á¼Î´Ïακον, á¼Î´Ïαθον, and á¼ÏÏαθον, in which âÏαâ cannot be explained by a simple analogy with other verbal stems.
Before discussing these forms, let us first consider three other aorists where âÏαâ is the expected analogical reshaping: á¼Î´Ïαμον, á¼ÏÏαÏον, and á¼ÏÏαÏον. The first two regularly occur from Homer onwards, and must have been present in Proto-Ionic. The transitive active á¼ÏÏαÏον was replaced in Classical Greek by the sigmatic form á¼ÏÏεÏα, but the middle á¼ÏÏαÏόμην remained current as an intransitive counterpart denoting body motion.
The antiquity and provenance of the third form, á¼ÏÏαÏον âgrew up, was rearedâ (beside Hom. and Class. á¼ÏÏá½±Ïην âid.â), are less clear. West (1998: xxxvi) thinks the thematic form is old and pleads, following an old proposal by Buttmann, for restoring 3sg. ÏÏá½±Ïεν, 3pl. ÏÏá½±Ïον in place of 3sg. ÏÏá½±Ïη, 3pl. ÏÏá½±Ïεν in the Homeric text (in many places against the entire tradition). It is true that the thematic aorist looks old in the verse-end γενέÏθαι Ïε ÏÏαÏέμεν Ïε âto be born and raisedâ (2â¯Ã Il., 1â¯Ã Od.). However, it is difficult to see why and how the vernacular form á¼ÏÏá½±Ïην would have been created secondarily as a replacement for á¼ÏÏαÏον. It is conceivable that there were originally two aorist stems, perhaps reflecting a semantic difference: á¼ÏÏαÏον âwas raisedâ versus á¼ÏÏá½±Ïην âgot thickâ. Alternatively, á¼ÏÏαÏον could be a traditional form that came from a different dialect contributing to the epic language. Finally, á¼ÏÏαÏον could be viewed as an artificial creation of Epic Greek: the form is attested only there, and the normal vernacular form á¼ÏÏá½±Ïην was hard to use in the hexameter.16 We do find á¼ÏÏá½±Ïην in Epic Greek, but only before vowels in the 3pl. ÏÏá½±Ïεν, á¼ÏÏαÏεν and the 3sg. ÏÏá½±Ïη.17
8.3.1 á¼Î´Ïακον
It is customary to translate this Homeric verb with âto look, seeâ, but the situation is actually much more complex. Let us therefore consider the attestations and their semantics more closely.
In Homer, the present δέÏκομαι âto look, gaze; beholdâ (also with preverbs) denotes the volitional activity of directing oneâs eyes at something: δεινὸν δεÏκόμενοι (Il. 3.342, cf. 23.815) denotes the âfearsome gazeâ of warriors. The aorist appears mainly in combination with preverbs (á¼Î½Î±â, á¼Ïâ, διαâ) and has complexive or ingressive value, as will be illustrated below. The perfect δέδοÏκα is stative and, combined with adverbials, means âto have a certain appearanceâ: cf. ÏμεÏδαλέον δὲ δέδοÏκεν á¼Î»Î¹ÏÏá½¹Î¼ÎµÎ½Î¿Ï ÏεÏὶ Ïειῠâand it [a snake] looks terrible as it crawls around in its lairâ (Il. 22.95).18 While δέÏκομαι occasionally retains these meanings in later poetry, Classical Attic expresses the same types of events with βλέÏÏ âto look, gaze; behold; to look likeâ (see Kölligan 2007a: 273â274). Clearly, βλέÏÏ (a verb without a decent etymology) has replaced the poetic archaism δέÏκομαι.
Let us now consider the Homeric uses of the aorist more closely. There are two clear instances of an ingressive aorist:
-
á¼Î½á½³Î´Ïακον âlooked up againâ (Il. 14.436, of Hector who has just regained his conscience). The form is an archaism: á½Ïá½±Ï / εἶδον is not attested in combination with á¼Î½Î±â, and Classical Greek uses á¼Î½Î±Î²Î»á½³ÏÏ.19
-
á¼ÏέδÏακον á¼Î½Ïην âI looked [the god] in the faceâ (Il. 24.223).
The other four attestations are complexive aorists:
-
When the maid Eurycleia tells how she tried to make eye contact with Penelope, she phrases this as follows: ΠηνελόÏειαν á¼ÏέδÏακον á½Ïθαλμοá¿Ïι, ÏεÏÏαδέειν á¼Î¸á½³Î»Î¿Ï Ïα âI looked towards Penelope with my eyes, wishing to give her a signâ (Od. 19.476â477); this is a complexive use of the activity verb á¼ÏδέÏκομαι.
-
καÏνὸν ⦠á¼Î´Ïακον ⦠διὰ δÏÏ Î¼á½° ÏÏ ÎºÎ½á½° καὶ á½Î»Î·Î½ âI saw smoke through the thick forestâ (Od. 10.197), where discerning the smoke is the result of a volitional action.
-
á¼Î½Î¸â οὠÏÎ¹Ï Ïὴν νá¿Ïον á¼ÏέδÏακεν á½Ïθαλμοá¿Ïιν âAt that point, no one could see the island with his eyesâ (Od. 9.146); again the subjects are performing a volitional activity (trying to discern the island through a thick mist).
-
οá½Î´â á¼Î½ Î½á¿¶Ï Î´Î¹Î±Î´Ïάκοι á¼¨á½³Î»Î¹á½¹Ï ÏÎµÏ ânot even Helios could discern us (through the clouds)â (Il. 14.344).
There are indications that á¼Î´Ïακον and εἶδον âsawâ were occasionally used as metrical alternatives. Kölligan (2007a: 264â265) compares á¼ÏέδÏακον á¼Î½Ïην (Il. 24.223) with á¼Î½Ïην εἰÏιδέειν (Il. 19.15), and καÏνὸν ⦠á¼Î´Ïακον á½Ïθαλμοá¿Ïι (Od. 10.197) with καÏνὸν ⦠á½Ïῶμεν (Od. 10.99), á¼Ïεὶ ἴδον αἴθοÏα καÏνόν (Od. 10.152). It seems as if the old form á¼Î´Ïακον was retained when corresponding forms of εἶδον were metrically problematic. This would account for the formulaic phrases |T á¼Î½á½³Î´Ïακεν á½Ïθαλμοá¿Ïι and |T á¼ÏέδÏακεν á½Ïθαλμοá¿Ïι (2â¯Ã): they were metrically distinct from the corresponding compounds of εἶδον. Interestingly, the instrumental dative á½Ïθαλμοá¿Ïι is not redundant in these cases, whereas it often seems superfluous when preceded by ἴδον. It is therefore possible that certain instances of the ptc. ἰδών are replacements of δÏακών, which was metrically awkward as it required tautosyllabic scansion of δÏâ. This would explain why so few remnants of the aorist á¼Î´Ïακον are left in Homer.
After Homer, á¼Î´Ïακον remains rare: there is only one attestation in Pindar (καÏέδÏακεν âlooked downâ Nem. 4.23, again with preverb), one in Stesichorus (ptc. δ]Ïακοá¿Ïα fr. S135.9), and six cases in Aeschylus and Euripides.20 In addition, two alternative aorist formations are found: Pindar uses the participle δÏακένÏ- of the η-aorist; furthermore, á¼Î´á½³ÏÏθην âlooked atâ is attested seven times in Sophocles and the author of the Prometheus Vinctus.21 While the latter form is clearly an innovation on the basis of δέÏκομαι, the Pindaric form δÏακένÏâ has played a prominent part in reconstructions of the PIE verbal paradigm. Since Forssman (1964), it is usually analyzed as deriving from an archaic PIE root aorist ptc. *dr̥ḱ-éntâ. As I will argue below, however, δÏακένÏâ may have been created within Greek.
In sum, δέÏκομαι often occurred with preverbs, especially in the aorist (cf. á¼Î½á½³Î´Ïακον, á¼ÏέδÏακον, διαδÏάκοι) but also in the present stem (ÏοÏιδέÏκομαι Il. 16.10, Od. 17.518, 20.385). The present stem refers to a volitional activity; the aorist has ingressive or complexive value, and the perfect denotes a state. Contrary to what is usually stated, the present stem δέÏκομαι must be inherited from PIE. The restriction of the verb to poetic texts and the general paucity of attestations in post-Homeric Greek (even in poetry) are compatible with the assumption that á¼Î´Ïακον is an epicism. The low frequency of á¼Î´Ïακον in Epic Greek can be due to its ongoing replacement by εἶδον.
8.3.2 á¼ÏÏαθον and the Etymology of Ïá½³ÏθÏ
The verb Ïá½³ÏÎ¸Ï âto raze, pillageâ is a relic of Epic Greek and the poetic language; it is not attested in Classical prose.22 The normal verb derived from this root in Classical Attic, with the same meaning, is ÏοÏθέÏ. Given that its meaning is typical for heroic poetry, Ïá½³ÏÎ¸Ï may well be an epicism in authors like Pindar. I will therefore concentrate on the Homeric forms.23
The most frequent stems in Homer are the thematic aorist á¼ÏÏαθον (9â¯Ã, including prefixed forms) and especially the sigmatic stems aor. á¼ÏεÏÏα, fut. Ïá½³ÏÏÏ (35â¯Ã, including prefixed forms). The only genuine attestation of the present stem is the dual Ïá½³ÏθονÏε (Il. 18.342), a precious archaism.24 As in Classical prose and poetry, the productive present stem formation was ÏοÏÎ¸á½³Ï already in Homer (5â¯Ã, including prefixed forms): the 3pl. impf. á¼Ïá½¹Ïθεον (Il. 4.308, with synizesis of εο) was preferred over *á¼ÏεÏθον. The productive aorist stem ÏεÏÏαâ may well be an innovation (cf. also the future Ïá½³ÏÏÏ).25 Thus, the oldest paradigm seems to be pres. Ïá½³ÏθÏ, aor. á¼ÏÏαθον, fut. Ïá½³ÏÏÏ.
The etymology of Ïá½³ÏÎ¸Ï is not quite clear. Janda (2000: 229â240, followed by LIV2 s.v. *bherdhâ) reconstructed *bher-dhh1â âBeute machenâ > erbeuten (i.e. âto seize as booty, captureâ), where *bherâ would be the root of Ïá½³ÏÏ âto carryâ. Such an analysis is possible in theory, because several other Greek verbs (both presents and aorists) have an extension âθÏ.26 There are, however, no directly comparable formations in other IE languages that could confirm this idea.27 An important objection against Jandaâs analysis is that the object of Ïá½³ÏÎ¸Ï is (with one exception) always a city, rather than the booty contained in it; the synchronic Homeric meaning is therefore âto raze, pillageâ.28
In fact, a number of post-Homeric attestations suggest that the meaning âto razeâ may have developed from âto cut offâ, cf. especially κεÏαλάν á¼ÏÏαθε ÏαÏÎ³á½±Î½Î¿Ï á¼ÎºÎ¼á¾· â[when] he cut off the head with the edge of his swordâ (Pi. Pyth. 9.80â81), and καί μοι Î³ÎµÎ½Îµá½·Î¿Ï Ïá½³Ïθε Î»ÎµÏ Îºá½µÏη ÏÏá½·Ïα âcut the white hairs off my chinâ (A. Pers. 1056). For this semantic development we may compare for instance κείÏÏ, which according to LSJ (s.v.) occurs in the following meanings:
-
âcut short, shear, clip, esp. of hairâ;
-
âcut downâ (of trees), whence âravage a country, esp. by cutting down crops and fruit-treesâ (thus in Hdt., Th.);
-
âgenerally, destroy, consumeâ.
Thus, if the verbâs original meaning was âto cut offâ, especially of hairs and plants, the development of meaning to ârazeâ (of cities) is straightforward.29 In PIE terms, we would have to posit a verbal root *bherdh- meaning âto shear, lopâ (of hairs, crops, foliage). Interestingly, there is perhaps further evidence for such a root in Italic: Umbrian furfa- denotes an action carried out on sheep as a direct object, and has consequently been translated as âshearâ (cf. Meiser 1986: 101).30 What is more, the noun for âbeardâ, reflected only in European branches (e.g. OPr. bordus, Lith. barzdà , Ru. borodá, OHG bart), can also be derived from this root. This would semantically be very attractive: the original meaning would be something like âhaircutâ (cf. again Î³ÎµÎ½Îµá½·Î¿Ï Ïá½³Ïθε ⦠ÏÏá½·Ïα, just quoted from Aeschylus). The internal sibilant of Lithuanian barzdà complicates the reconstruction, but we may posit a root *bhersdh- and assume that Greek lost the sibilant in the e-graded root (*phersthâ > *pherthâ), after which the aorist *phrÌ¥sthâ may have been changed along to *phrÌ¥thâ.31
Irrespective of the etymology just proposed, the thematic aorist á¼ÏÏαθον is old within Greek and its zero grade reflex requires an explanation. In what follows, it will be of some importance that this form occurs either with a preverb (διαâ 6â¯Ã, á¼Î¾â 1â¯Ã) or in the phrase Ïόλιν á¼ÏÏαθον (2â¯Ã).
8.3.3 Conclusion
The aorist forms with âÏαâ in which Homer deviates from Classical Ionic and Attic are limited to poetry, and rarely attested outside of epic. The forms á¼Î´Ïαθον and ÏÏαÏείομεν are exclusively epic. After Homer and before the end of the Classical period, the aorist á¼ÏÏαθον is found only in Pindar (4â¯Ã) and Corinna (1â¯Ã), while á¼Î´Ïακον is attested only in Pindar, Stesichorus (each 1â¯Ã) and the tragedians (6â¯Ã).32 However, the last two forms also compete with other formations (á¼ÏεÏÏα, á¼Î´á½³ÏÏθην). It is therefore plausible that the thematic aorists with âÏαâ are epicisms.
8.4 Epic *r̥ in the Thematic Aorist?
No analogical account of the reflex âαÏâ in the prose forms καÏέδαÏθον (Attic) and ἥμαÏÏον (Ionic and Attic) seems within reach. On the other hand, ex hypothesi the forms á¼Î´Ïαθον, á¼Î´Ïακον, á¼ÏÏαθον cannot have the regular Ionic outcome of *rÌ¥. We must therefore consider the possibility that âÏαâ in the poetic forms á¼Î´Ïαθον, á¼Î´Ïακον, á¼ÏÏαθον is an artificial epic reflex, as in ÏÏαÏείομεν. This idea receives support from the distribution between δαÏθâ (Attic prose) and δÏαθâ (only epic). The pair ἥμαÏÏον (Attic, Ionic and epic) versus ἤμβÏοÏον (epic) can also be interpreted in this way.
I can see two basic ways to account for an artificial reflex âÏαâ in á¼Î´Ïαθον, á¼Î´Ïακον and á¼ÏÏαθον: (i) âÏαâ in these aorists reflects Epic *rÌ¥; (ii) it arose as a secondary reshaping of âÏοâ, which was the regular reflex of *rÌ¥ in varieties of Aeolic that contributed to the tradition at an early stage. Before investigating these possible scenarios in more detail, it is necessary to consider the distributions and rhythmical behavior of these aorist stems.
8.4.1 Distributions and Metrical Behavior of Thematic Aorists with âÏαâ
As argued in chapter 6, the one-time presence of Epic *rÌ¥ in a specific Homeric form can be assumed if it is plausible that the lexeme in question was absent from the vernacular at an early date (e.g. κÏαÏαιόÏ) or if there was a plausible motive for not introducing the vernacular form (e.g. metrically awkward καÏδίη). Since there would have been no motive for avoiding forms like καÏέδαÏθον, *á¼Î´Î±Ïκον and *á¼ÏαÏθον on the basis of their rhythmical structure, we must assume that these forms no longer existed in Ionic when other forms with âαÏâ became available for introduction into Epic Greek (such as the pre-forms of καÏδίη, καÏÏεÏá½¹Ï, ÏαÏÏέεÏ). This assumption is unproblematic for á¼Î´Ïακον and á¼ÏÏαθον which, as we have seen, are poetic relic forms.
More remarkable is the coexistence of Attic καÏέδαÏθον and Epic á¼Î´Ïαθον. The Attic form presupposes that Proto-Ionic preserved this word when *rÌ¥ developed into âαÏâ. Ionic prose, however, preserves no trace of this verb: Herodotus and the Hippocratic Corpus use the aorist καÏεκοιμήθην âto go to sleepâ, an innovative form that also occurs in Homer but is absent from Attic prose. It therefore seems that the vernaculars of Homer and his immediate Ionian predecessors had already lost καÏέδαÏθον (and replaced it by καÏεκοιμήθην), but also that the tradition resisted the introduction of Ionic forms at an earlier stage, when καÏέδαÏθον was still current in spoken Ionic. If this reasoning is correct, it shows that, at the stage when *rÌ¥ had just developed to âαÏâ in Proto-Ionic, Ionic was not yet the default language of epic poets.33
Turning to prosodic issues concerning the active thematic aorists with âÏαâ, the problems can be summarized as follows: the rhythmical behavior of these forms is at odds with a prolonged presence of *rÌ¥, but their phonological reflex âÏαâ (instead of expected âαÏâ) is explained most naturally as the regular outcome of Epic *rÌ¥.
One prosodic issue is that we find no traces of McL scansion in the thematic aorists with âÏαâ. This means that participle forms such as δÏακών, δÏακόνÏÎ¿Ï are unattested. Given the large number of attestations of these aorists, this is probably not due to chance: apparently the forms with McL were actively avoided, in line with the general tendency to avoid McL when possible (cf. chapters 6 and 7, especially concerning βÏοÏá½¹Ï and its derivatives).34 Nevertheless, one wonders why δÏακών and similar participle forms were apparently disallowed, while traditional epic forms like δÏάκÏν âsnakeâ and βÏοÏοá¿Ïι were tolerated.
This distribution between forms with McL and forms without McL would receive a natural explanation if we assume that the thematic aorists with âÏαâ arose from a reshaping of earlier Aeolic forms with âÏοâ, while δÏάκÏν âsnakeâ and βÏοÏοá¿Ïι continue pre-forms with Epic *rÌ¥. On the other hand, the distribution does not by itself exclude that the pre-forms of á¼Î´Ïαθον, á¼Î´Ïακον, á¼ÏÏαθον had Epic *rÌ¥. For one thing, it is conceivable that poets found a workaround for undesired McL scansions in the thematic aorists, but not in the case of δÏάκÏν âsnakeâ and βÏοÏοá¿Ïι. For instance, as we have just seen it is plausible that δÏακών was replaced by ἰδών. Another relevant factor is that word-internal McL (in augmented indicatives like á¼Î´Ïακον and in prefixed forms) was much more strongly avoided than word-initial McL (in δÏάκÏν and βÏοÏοá¿Ïι).
Another noteworthy prosodic fact concerns the opportunity to generate length by position. This option is widely used in all thematic aorists under discussion, especially in forms with augment and/or preverb such as καÏέδÏαθον, διαδÏάκοι, á¼ÏÏá½±ÏεÏ(ο). Again, this behavior seems to be at odds with the idea that pre-forms of these aorists had Epic *rÌ¥. The same issue is at play in κÏαÏεÏá½¹Ï âfierceâ, which seems to reflect a pre-form with Epic *rÌ¥ (given its reflex âÏαâ < *rÌ¥), but at the same time has an onset that often generates length by position. On the other hand, we have seen that κÏαδίη < *krÌ¥diÄâ and βÏοÏá½¹Ï < *mrÌ¥toâ were hardly used to generate length by position.35 Again, the high number of attestations of κÏαδίη, κÏαÏεÏá½¹Ï and βÏοÏá½¹Ï virtually excludes a coincidence.
Let us zoom in on κÏαδίη versus κÏαÏεÏá½¹Ï. Given the high number of attestations of both forms, the difference in their rhythmical behavior cannot be due to chance. As argued in chapter 6, the precursor of κÏαδίη was retained unaltered in the form *krÌ¥diÄâ until Epic *rÌ¥ developed into âÏαâ. As for *krÌ¥terós, I propose that its root vocalism may have been influenced by the related form κÏαÏá½»Ï, which had acquired âÏαâ already in Proto-Ionic by analogical leveling (see chapters 4 and 5). Introducing the root shape κÏαÏâ into *krÌ¥teróâ was highly attractive, as it enabled poets to use κÏαÏεÏá½¹Ï after words ending in a short vowel, including prepositions like καÏá½± and á¼Î½á½·. This development also led to a marginalization of κÏαÏá½»Ï, which is retained only in the name-epithet formula κÏαÏá½ºÏ á¼ÏγεÏÏόνÏηÏ. In *krÌ¥diÄâ, on the other hand, no model for an early introduction of âÏαâ was available.
We will now consider whether the metrical behavior of thematic aorists like á¼Î´Ïακον can be explained like that of κÏαÏεÏá½¹Ï, looking first at the reconstruction with Epic *rÌ¥, and then turning to the possibility of an Ionicized Aeolic form.
8.4.2 A Possible Origin of âÏαâ in á¼Î´Ïακον, á¼Î´Ïαθον, á¼ÏÏαθον
Most of the approximately 60 active thematic aorist stems have a light root syllable before the thematic vowel.36 In such cases, the structure of the stem is VCVC-e/oâ, CVC-e/oâ, or CCVC-e/oâ. Only the last two types are of interest here: at first, forms like *drÌ¥k-e/oâ had the structure CVC-e/oâ, whereas the Homeric outcome δÏακε/οâ had the structure CCVC-e/oâ. If we reconstruct the pre-forms of these thematic aorists with Epic *rÌ¥, the elimination of this sound would have changed the possibilities to use them in the epic hexameter quite drastically.
Table 25
Localization of thematic aorist forms with âÏαâ in Homer
|
Root in: |
4th thesis |
5th thesis |
Elsewhere |
Total |
||
|---|---|---|---|---|---|---|
|
+ prev |
â prev |
+ prev |
â prev |
|||
|
δÏαθε/ο- |
6 |
1 |
1 |
8 |
||
|
δÏακε/ο- |
4 |
1 |
1 |
6 |
||
|
ÏÏαθε/ο- |
3 |
2 |
4 |
9 |
||
|
δÏαμε/ο- |
20 |
2 |
5 |
2 |
29 |
|
|
ÏÏαÏε/ο- |
3 |
6 |
3 |
12 |
||
|
á¼ÏÏá½±ÏεÏο37 |
6 |
4 |
7 |
4 |
21 |
|
|
Total |
42 |
12 |
13 |
10 |
8 |
85 |
In order to show this, let us consider the token frequency and localization of the relevant thematic aorist forms, as summarized in Table 25. I have not included forms of á¼ÏÏαÏον in view of the numerous variant readings such as aor. á¼ÏÏá½±Ïη or impf. á¼ÏÏεÏεν (see above), nor forms of ÏÏαÏá½³Ïθαι with McL, of which the root syllable is placed in the 2nd half of the thesis. Thus, all forms included in the table have their root syllable in the 1st thesis syllable.
In á¼Î´Ïαμον and á¼ÏÏαÏον (active and middle), the Homeric stem is identical to that of the corresponding vernacular form. These two frequent stems account for 62 instances (72.9â¯%).38 Moreover, in both stems âÏαâ was also present in the Ionic vernacular as the analogically restored reflex of *rÌ¥. On the other hand, the stem of the other 23 instances (27.1â¯%) did not appear in the vernacular (á¼Î´Ïακον, á¼ÏÏαθον), or had a different shape there (á¼Î´Ïαθον vs. Att. καÏέδαÏθον).
The high number of forms occurring in the fourth foot (54â¯Ã, or 63.5â¯%) can be ascribed to two factors: a general tendency to put clause-final verb forms in the fourth foot (cf. Parry 1971: 41) and the high number of forms with preverb (42â¯Ã, or 49.4â¯%). I have listed such forms with preverb separately, as they strongly determine the localization: forms like á¼Î½á½³Î´Ïακον fit exactly between the third foot caesura and the bucolic dieresis, and they could hardly be used in other places.39 The augmented simplex forms (e.g. á¼Î´Ïαμον, á¼ÏÏαÏον), given their dactylic structure, naturally tend to occupy the 1st, 4th or 5th foot.40 Thus, the localization of the forms follows directly from their rhythmical structure. Note that only ÏÏαÏε/οâ is frequent as a simplex; the other four stems (δÏακε/οâ, ÏÏαθε/οâ, δÏαθε/οâ, δÏαμε/οâ) mainly occur with preverbs, with 8 exceptions.41
In order to test the hypothesis that these thematic aorists had Epic *r̥, we must now ask how the prehistoric forms *dr̥ke, *dr̥me, *dr̥the, *pr̥the and *tr̥kwe would have been used in epic verse. First of all, prefixed forms could not have been used in forms like *anedr̥ke or *epedr̥me, with four consecutive light syllables. Thus, if these stems were predominantly used with preverbs already at an early stage of the tradition, the question becomes how they could have been used at all.
At first sight, it seems that this problem could have been resolved by means of metrical lengthening of the second of four consecutive light syllables. Upon closer consideration, however, this idea appears to be problematic: there are no secure traces of a metrically lengthened augment with other roots of the structure *CVCâ, where the same problem would have occurred.42 On the contrary, a different traditional means of creating heavy prefixed syllables appears to have been used, involving the apocopated preverbs καÏâ and á¼Îºâ. For instance, in Homer κάλλιÏε, κάββαλε, and κάÏÏεÏε occur in the same meaning as á¼Î»Î¹Ïε, á¼Î²Î±Î»Îµ, á¼ÏεÏε. Likewise, á¼ÎºÏÏ Î³Îµ + acc. functions as a metrical alternative for Ïύγε âescapedâ.43
A second option, again merely theoretical, might have been to use alternative forms of the preverb in âαιâ. Thus, a preverb ÏαÏαι- (instead of ÏαÏαâ, ÏαÏâ) appears in ÏαÏαιÏεÏιθοῦÏα, ÏαÏαιÏá½±Î¼ÎµÎ½Î¿Ï and in the nominal derivative ÏαÏαίÏαÏιÏ. Thus, we might in theory account for ÏαÏέδÏαθεν (Od. 20.88) and ÏαÏαδÏαθέειν (Il. 14.163) by positing a pre-form *parai-drÌ¥the. However, in Homer the alternative form with âαιâ is practically limited to ÏαÏαιâ: the only exception is the compound καÏαιβαÏαί (Od. 13.110); the alternative form διαι- beside δια- first develops after Homer. Since the prefixed forms of δÏακâ (á¼Î½á½³Î´Ïακον, διαδÏάκοι, á¼ÏέδÏακον) and ÏÏαθâ (á¼Î¾ÎµÏÏάθομεν, διαÏÏαθέειν) never occur with ÏαÏαâ, they cannot reflect pre-forms in *âaiâ.
In view of these problems, let us consider how the problem of using preverbs was solved with other thematic aorist indicatives to roots of the structure *CVCâ. It is instructive to compare the aorist stem θοÏε/ο- âto jumpâ, because this is attested almost exclusively44 with preverb and semantically close to δÏαμε/οâ. Its indicative is used mainly in two ways:
-
with a prefixed preverb if this ended in a consonant: e.g. á½Ïá½³ÏθοÏον (Il. 9.476), ὠδâ á¼Ïâ á¼ÏθοÏε ÏÎ±á½·Î´Î¹Î¼Î¿Ï á¼ÎºÏÏÏ (Il. 12.462), ὠδâ á¼ÏθοÏε δαίμονι ἶÏÎ¿Ï (Il. 21.18), á¼Î½Î¸Î¿Ïε μέÏÏῳ (Il. 21.233), á¼Î½Î¸Î¿Ïâ á½Î¼á½·Î»á¿³ (Il. 15.623), á¼ÎºÎ¸Î¿Ïε δίÏÏÎ¿Ï (Il. 16.427);
-
with a preverb in tmesis: e.g. κὰδ δâ á¼Î¸Î¿Ïâ á¼Ï μέÏÏον (Il. 4.79), á¼Îº δâ á¼Î¸Î¿Ïε ÏÏομάÏÏν (Il. 15.573), á¼Îº δâ á¼Î¸Î¿Ïε κλá¿ÏÎ¿Ï ÎºÏ Î½á½³Î·Ï (Il. 7.182), á¼Îº δὲ κλá¿ÏÎ¿Ï Î¸á½¹Ïε (Il. 23.353), á¼Îº δâ á¼Î¸Î¿Ïε κλá¿ÏÎ¿Ï (Od. 10.207), á¼Îº δίÏÏοιο Ïαμαὶ θόÏε (Il. 8.320 = 23.509).
Thus, when *rÌ¥ was still current, one would expect to find dactylic forms of the type *katdrÌ¥the, *andrÌ¥ke, alongside forms with preverb in tmesis such as *ana ⦠drÌ¥me or *an dâ edrÌ¥me. There is, however, no evidence for such forms among the roots ÏÏαθâ, δÏακâ and δÏαθâ, apart from one instance of the 3rd dual καδδÏαθέÏην (Od. 15.494).
At first sight, this seems detrimental to the idea that these aorists reflect pre-forms with Epic *rÌ¥, for wouldnât one expect to find more instances of tmesis in Homer? On second thoughts, however, this lack of attestations could be due to the metrical convenience of compounded forms like á¼Î½á½³Î´Ïακον once these became available, and the metrical inconvenience of the tmesis construction.45 The main question is therefore: how could forms like á¼Î½á½³Î´Ïακον develop?
In Van Beek 2013, I proposed that âÏαâ was introduced into *drÌ¥ke/oâ, *drÌ¥the/oâ and *prÌ¥the/oâ from the vernacular forms á¼ÏÏαÏον and á¼Î´Ïαμον. Given the metrical convenience of the latter two forms (and compounds), one could expect that they ousted the traditional forms with *rÌ¥ from Epic Greek soon after they became available. Not only was á¼Î´Ïαμον the most frequent thematic aorist with âÏαâ in Homer, but it was used exclusively with preverb (á¼Î½Î±â, διαâ, á¼Ïιâ, á½Ïoâ, etc.). This means that pre-forms like *ana / dia ⦠drÌ¥me (with tmesis), which were difficult to use, were ousted by á¼Î½á½³Î´Ïαμε(ν), διέδÏαμε(ν) once these became available. The same holds for the replacement of *epi ⦠trÌ¥kwe with e.g. á¼Ïá½³ÏÏαÏε(ν). I therefore supposed that the two frequent stems δÏαμε/οâ and ÏÏαÏε/οâ dragged the other three forms *drÌ¥ke/oâ, *drÌ¥the/oâ, and *prÌ¥the/oâ along with them. That is, when the forms á¼Î´Ïαμε and á¼ÏÏαÏε had become available as alternatives for *drÌ¥me and *(e)trÌ¥kwe and were in the process of replacing them, the forms á¼Î´Ïακε, á¼Î´Ïαθε and á¼ÏÏαθε were artificially created as metrical alternatives for forms with preverb of *drÌ¥ke, *drÌ¥the and *prÌ¥the, which were inconvenient to use because the preverb usually had to be placed in tmesis.
If this possibility is granted, the introduction of âÏαâ also generated a problem that has already been discussed in the previous section. Participles and many subjunctive and optative forms of the aorists δÏαμε/οâ and ÏÏαÏε/οâ could not be used: at this early stage, McL was still out of the question. An artificial introduction of âÏαâ into *drÌ¥ke/oâ, *drÌ¥the/oâ and *prÌ¥the/oâ would have entailed that their participles and many modal forms could no longer be used. However, this probably would not have been detrimental: these verbs mainly occurred as compounds anyway, and compounded participle forms like *anadrÌ¥kontâ could not be used either. In reality, it is likely that alternative ways of expression had emerged already at earlier stages of the tradition: consider the ptc. (á¼Îº)Ïá½³ÏÏανÏâ of the sigmatic aorist, which occurs instead of the reflex of *prÌ¥thontâ; ἰδών in the meaning âlooking, glancing (at)â instead of the reflex of *drÌ¥kontâ; and μεÏάλμενοÏ, á¼Ïá½±Î»Î¼ÎµÎ½Î¿Ï âjumping among/atâ instead of unmetrical *μεÏαθοÏών, *á¼ÏιθοÏών.
It remains to explain why âÏαâ was not introduced in the thematic aorist ἤμβÏοÏον. If we assume that this form reflects *ÄmrÌ¥te/oâ with Epic *rÌ¥, the reasons are not difficult to find: the augmented pre-form *ÄmrÌ¥te already had a dactylic structure, so there was no clear motivation to introduce âÏαâ. Moreover, since the root structure of *amrÌ¥tâ was different from that of *drÌ¥mâ and other similar forms, there was no obvious model. At some point the Ionic vernacular form ἠμαÏÏε was introduced as a convenient metrical alternative to *ÄmrÌ¥te. We may assume that the latter form was preserved as such, and that it eventually took part in the regular vocalization of Epic *rÌ¥, yielding ἤμβÏοÏον.
8.4.3 Reconsidering the Possibility of Aeolisms
The form ἤμβÏοÏον, with its reflex âÏοâ < *rÌ¥, could also be an Aeolism. In view of the problems involved in reconstructing Epic *rÌ¥ for the thematic aorists with âÏαâ, let us reconsider whether these can be accounted for as Aeolisms. The simplest scenario would be that their stems, at an early pre-stage of Epic Greek, had the Aeolic vernacular reflex âÏοâ < *rÌ¥, and that their vocalism was changed into âÏαâ under influence of corresponding Ionic vernacular forms. This would indeed work well for á¼ÏÏαÏον and á¼Î´Ïαμον, cases where the aorist stem is attested both in Ionic and in Aeolic (cf. ÏÏá½¹Ïην Alc. fr. 70.9, á½Î½á½³ÏÏοÏε fr. 72.8, ÏεδέÏÏοÏ[ε fr. 75.11; δÏá½¹[μÏμεν Alc. fr. 6.8). Such a scenario would be the easiest way of accounting for the rhythmical behavior of the thematic aorists with âÏαâ in Epic Greek: as we have seen, their root-initial PL-onset is always heterosyllabic in Homer.
However, it is less trivial to view á¼Î´Ïακον, á¼ÏÏαθον and á¼Î´Ïαθον as Ionicized versions of original Aeolic forms with âÏοâ. Such forms are not attested in literary Lesbian or in inscriptions from the Aeolic dialects, and the aorists á¼Î´Ïακον and á¼ÏÏαθον are poetic relics. Furthermore, á¼Î´Ïαθον is unattested in this specific shape in Attic-Ionic: the verb is absent from Ionic, and found only in the form καÏέδαÏθον in Attic. If we assume that á¼Î´Î±Ïθον influenced the vocalism of an older Aeolic *á¼Î´Ïοθον (with a different vowel slot), we are confronted with several further questions. For instance, why wasnât ἤμβÏοÏον changed into *ἤμβÏαÏον under the influence of ἥμαÏÏον? Why wasnât á¼Î´Î±Ïθον introduced into Epic Greek?
In order to avoid such problems, I propose the following scenario: when á¼ÏÏοÏον, á¼Î´Ïομον were in the process of being replaced by their Ionic counterparts á¼ÏÏαÏον, á¼Î´Ïαμον in the epic tradition, the root vocalism of the other three thematic aorists *á¼Î´Ïοκον, *á¼ÏÏοθον and *á¼Î´Ïοθον (stemming from some early Aeolic dialect) was changed accordingly. Phrased differently, given the existence of two frequent thematic aorists with Ionic Ïα beside Aeolic Ïο, it is conceivable that Ionic poets extended the same equivalence to other thematic aorists with Ïο that were current in the tradition.
If this is accepted, we must note that a similar scenario will not account for the forms with Ïα discussed in chapter 6. There is no evidence that poetic relic forms like δÏάκÏν âsnakeâ, κÏαÏÎ±Î¹á½¹Ï âstrongâ and ÏÏαÏείομεν âlet us get satisfactionâ originally had o-vocalism (corresponding Aeolic forms with âÏοâ are non-existent), or that they owe their a-vocalism to a now-lost Ionic counterpart. Furthermore, an important reason for not reconstructing Aeolic predecessors with âÏοâ of these forms, but instead to posit pre-forms with Epic *rÌ¥, is their prosodic behavior (McL scansion) in Homer.
Thus, if this scenario is correct, the following picture emerges. For an Early Dark Age stage of the epic tradition, we must assume a coexistence of forms preserving *rÌ¥ (some of which were of Mycenaean origin) with forms showing the Aeolic vernacular reflex âÏοâ < *rÌ¥. At a later stage, Ionic forms with âαÏâ (and analogical âÏαâ) < *rÌ¥ were also introduced. In cases where an appropriate model existed, Aeolic forms with âÏοâ < *rÌ¥ had their vocalism changed into âÏαâ. In addition, new instances of âÏαâ and âÏοâ came into being when Epic *rÌ¥ was vocalized.
8.5 Pindaric δÏακένÏâ
It remains to explain the reflex âÏαâ in the Pindaric participle δÏακένÏâ.46 Traditionally (e.g. LSJ s.v. δέÏκομαι), this form has been interpreted as what it appears to be: a participle of the η-aorist. However, in a brief and influential contribution, Forssman (1964) argued that δÏακένÏâ reflects a relic ptc. form PIE *dr̥ḱ-éntâ of the athematic root aorist. In his view, the participle survived only in Pindar, whereas the indicative had been thematicized already in Homer.
This analysis has become widely accepted among Indo-Europeanists.47 If it is correct, δÏακένÏâ could directly continue PIE *dr̥ḱ-éntâ in non-Epic Greek and thus constitute a potential counterexample to the regular vernacular vocalization to âαÏâ defended here. However, it is not quite clear from which variety of Greek δÏακένÏâ stems. The origins of Pindarâs poetic language are notoriously difficult to pin down, and if âÏαâ was indeed the regular outcome in some West Greek dialect (cf. chapter 3), δÏακένÏâ might stem from there and thus lose its probative force for Ionic-Attic.
Moreover, I wonder whether the traditional interpretation of δÏακένÏâ as an η-aorist can really be rejected. Forssmanâs first objection against this interpretation is that δÏακένÏâ takes a direct object in all three attestations. In reality, this is true for two instances, á¼ÎºÏá¿Î½Î±Ï ⦠μαÏμαÏÏ Î¶Î¿á½·ÏÎ±Ï Î´ÏÎ±ÎºÎµá½·Ï (fr. 123.2â3) and οὠÏάοÏ, οὠμέλαιναν δÏακένÏÎµÏ Îµá½ÏÏόναν (Nem. 7.3), but not for the third: in δÏακεá¿Ïâ á¼ÏÏÎ±Î»á½³Ï (Pyth. 2.20) âlooking with a straight gazeâ, we find an internal accusative construction that is clearly old and frequent with this verb.48 However this may be, the direct object is not an argument for reconstructing an old root aorist, as the thematic aorist á¼Î´Ïακον is often transitive, too: thus, Homeric á¼Î´Ïακον may also have a direct object (κάÏνον ⦠á¼Î´Ïακον, Od. 10.196â197). Moreover, the construction of δέÏκομαι without preverb and with direct object is found mainly with the aorist stem and may well be a secondary syntactic development of this aspectual form.
Forssmanâs second argument against viewing δÏακένÏâ as an intransitive aorist is that the indicative *á¼Î´Ïάκην is not attested anywhere else in Greek. This argumentum e silentio is weak: δέÏκομαι has a low overall frequency in Greek, and the single attestation of the thematic aorist indicative καÏέδÏακεν in Pindar could easily be an epicism.49
Is it really excluded that δÏακένÏâ originated as a reshaping of δÏακόνÏ-? In my view, it is not. Another intransitive aorist á¼Î´á½³ÏÏθην is attested in the tragedians, always in the indicative, meaning âto look at, beholdâ. In fact, in five out of seven attestations, á¼Î´á½³ÏÏθην governs a direct object, thus completely undermining Forssmanâs argument concerning δÏακένÏâ.50 Its analogical creation (on the basis of the present) may be explained by the observation of Allan (2003: 159) that δέÏκομαι becomes non-volitional in Classical Greek; for this reason, its aorist could be aligned with other mental process middles that formed an intransitive aorist in Classical Greek, such as á¼ÏÏá½±Ïθην âI noticedâ to ÏÏάζομαι.
Allan, too, thinks that Pindaric δÏακένÏâ is a secondary creation. He compares the replacement of thematic á¼ÏÏαÏον (Hom.) with á¼ÏÏá½±Ïην (Hom.+), where only the latter is used by Pindar as a medio-passive aorist of ÏÏá½³ÏÏ.51 In my view, it is conceivable that Pindar viewed the archaism á¼Î´Ïακον as a typical epic form, like á¼ÏÏαÏον. Moreover, the creation of a ptc. δÏακένÏâ beside the Homeric indicative á¼Î´Ïακον may have been favored by the absence of the participle δÏακόνÏâ from Homer.52 A final point is that the same replacement seems to have occurred in the ptc. á¼ÏιÏών âcollapsingâ (Hom.) â á¼ÏιÏένÏ- (Pi.).53
In conclusion, I think that Forssman was wrong in viewing the Pindaric ptc. δÏακένÏâ as an archaism; and even if he were right, it cannot be excluded that δÏακένÏâ is a West Greek element of Pindarâs language.
8.6 Conclusions
In thematic aorist forms, Classical prose has *rÌ¥ > âαÏâ in the following forms:
-
καÏέδαÏθον âwent to sleepâ (only Attic);
-
á¼ÏαÏδον âbroke windâ (only Attic);
-
ἥμαÏÏον âmissed; erredâ (Att.-Ion., Hom.).
While á¼ÏαÏδον may be analogical beside Ïá½³Ïδομαι, this explanation is not available for καÏέδαÏθον and ἥμαÏÏον which are both, from a Greek perspective, primary thematic aorists. From this fact, I have concluded that καÏέδαÏθον and ἥμαÏÏον have the regular outcome of Proto-Greek *rÌ¥.
A number of aorist forms with âÏαâ can easily be explained as analogical. For instance, the vowel slot of á¼Î´Ïαμον âranâ corresponds to that of the perfect δέδÏομε (Hom.) and the noun δÏá½¹Î¼Î¿Ï âcourseâ, while the vowel slot of á¼ÏÏαÏον, á¼ÏÏαÏόμην âturnedâ and á¼ÏÏαÏον, á¼ÏÏá½±Ïην âwas rearedâ matches that of ÏÏá½³ÏÏ and ÏÏá½³ÏÏ.
The reflex âÏαâ in á¼Î´Ïαθον âsleptâ, á¼Î´Ïακον âlooked, sawâ and á¼ÏÏαθον ârazed, pillagedâ causes more difficulties. These forms are virtually restricted to Epic Greek. Moreover, there are good reasons to assume that á¼Î´Ïακον and á¼ÏÏαθον were restricted to poetry at an early date: this allows us to account for the absence of by-forms with âαÏâ in Homer. We have considered in detail two scenarios accounting for the reflex âÏαâ in á¼Î´Ïαθον, á¼Î´Ïακον and á¼ÏÏαθον:
-
to reconstruct pre-forms with Epic *rÌ¥ (archaisms of the epic tradition) whose root vocalism was adapted to Ïα at an early date under the influence of the thematic aorists á¼Î´Ïαμον âranâ and á¼ÏÏαÏον âturnedâ;
-
to reconstruct pre-forms with Ïο of Aeolic origin, which were turned into Ionic-sounding forms with Ïα under the influence of the correspondences á¼Î´Ïαμον ~ Aeol. á¼Î´Ïομον and á¼ÏÏαÏον ~ Aeol. á¼ÏÏοÏον.
In both scenarios it must be assumed that the reflex âÏαâ in á¼Î´Ïαθον, á¼Î´Ïακον and á¼ÏÏαθον is a secondary adaptation to the vocalism of á¼Î´Ïαμον and á¼ÏÏαÏον, verbs which are not only frequent in Homer but also remained current in spoken Ionic. In the first scenario, Ïα cannot be the regular outcome of Epic *rÌ¥, as this would fail to account for the prosodic behavior of á¼Î´Ïαθον, á¼Î´Ïακον and á¼ÏÏαθον (absence of McL; regular position length). The second scenario, Ionicized Aeolisms, is preferable because it accounts for this prosodic behavior without a problem.
The scenario sketched in this chapter allows us to view the vocalism of καÏέδαÏθον and ἥμαÏÏον as the regular vernacular reflex of *rÌ¥ in Proto-Ionic. At the same time we may account for the restriction of á¼Î´Ïακον and á¼ÏÏαθον to poetry, and for the distribution between á¼Î´Ïαθον (epic) and καÏέδαÏθον (Attic). I have also proposed an account of the prosodic behavior of κÏαÏεÏá½¹Ï (which normally makes position in Homer) as opposed to that of κÏαδίη (which hardly ever makes position). A traditional form *krÌ¥terós may have been changed into κÏαÏεÏá½¹Ï under the influence of κÏαÏá½»Ï (with an analogical vowel slot) well before the elimination of Epic *rÌ¥. A similar substitution was not possible in the isolated noun *krÌ¥diÄ. It was therefore retained in this form, and yielded κÏαδίη only later, when Epic *rÌ¥ was vocalized.
On the intr. aor. 3pl. βλάβεν, which may contain the regular reflex of *l̥ or have introduced the reflex of a vocalized nasal from the present βλάβομαι, see section 10.3.1.
Sophronâs fragment is known through Hesychius ν 734. In addition, the Suda has the forms á¼ÏÏάδει, á¼ÏÏάδειν (without dialect indication), which look like imperfects of a contract verb.
The reduplicated aorist stem κεκÏαγâ is frequent in Aristophanes. The present κÏá½±Î¶Ï âto screakâ occurs once in the same author and is probably a late formation, replacing the older âintensiveâ perfect; the compounded verb á¼Î½Î±ÎºÏá½±Î¶Ï occurs only as a thematic aorist (Hom., Pi., X., Ar., etc.).
For secondary vocalism in a thematic aorist, cf. also Hom. διέÏμαγον âI crossedâ, διέÏμαγεν âthey separatedâ beside pres. (á¼Ïοâ)ÏμήγÏ, aor. (διαâ)Ïμá¿Î¾Î±Î¹.
Cf. also Aeol. aor. subj. 3pl. δÏá½¹ÏÏÏι.
The LIV2 (s.v. *perdâ) reconstructs a PIE root aorist *perdâ / *prÌ¥dâ on the basis of YAv. pÉrÉδÉn and the Greek thematic aorist á¼ÏαÏδον. It is remarked there that Kellens analyzed the Avestan form as a present; cf. also the doubts of Allan (2003: 209 n. 362) concerning the reconstruction of the PIE aorist. However this may be, the inner-Greek variation between Att. á¼ÏαÏδον and the 2sg. aor. ind. á¼ÏÏÎ±Î´ÎµÏ (attested in Sophron fr. 136 K-A) guarantees the antiquity of a zero-grade thematic aorist *prÌ¥d-e/oâ within Greek.
Cf. in particular the table at the start of chapter 6.
For the Homeric semantics, see the discussion in Kölligan (2007a: 173â179), especially his remark that âder ⦠Bedeutungsansatz âeinschlafenâ lässt sich bei Homer nicht belegen. á¼Î´Ïαθον bedeutet entweder âsich schlafen legenâ oder fungiert als komplexiver Aorist zu εá½Î´Ï und bedeutet dann â(eine Zeit lang) schlafenâ.â (o.c. 174). In my view, the simplex á¼Î´Ïαθον Od. 20.143 could be analyzed as in tmesis with the preceding á¼Î½ á¼Î´ÎµÏá½µÏῳ βοέῠκαὶ κώεÏιν οἰῶν (line 142), cf. Ïλαá¿Î½Î±Î½ ⦠καὶ κώεα, Ïοá¿Ïιν á¼Î½Îµá¿¦Î´ÎµÎ½ (Od. 20.95), and also Od. 3.349â351, Od. 10.11â12. This means that á¼Î´Ïαθον is attested only as a compound.
CS drÄmati derives from a lengthened grade formation *drÄmâ (see EDSIL, q.v.).
Note that Attic âδαÏθâ excludes a reconstruction *drmÌ¥-dhâ for Hom. á¼Î´Ïαθον.
See Kölligan (2007a: 181â182): âGegenüber dem homerischen Zustand, in dem εá½Î´Ï und á¼Î´Ïαθον sowohl in der Bedeutung âschlafenâ als auch âsich schlafen legenâ belegt sind, findet sich seit klassischer Zeit bei á¼Î´Ïαθον [sic; the only classical form is καÏέδαÏθον] zusätzlich die Bedeutung âeinschlafenâ.â
For this reason, OâNeil (1971: 19) is mistaken when he asserts that the Attic aorist âδαÏθεá¿Î½ may have replaced the older form âδÏαθεá¿Î½ after the present âδαÏθάνÏ.
The hapax á¼Î²ÏοÏάξομεν has already been discussed in chapter 7 and will be left out of further consideration here.
Apart from ÎºÎµÏ Î¸á½±Î½Ï and ληθάνÏ, all thematic nasal presents cited by Ruijgh have a zero grade root. In fact, both ληθάνει (verse-initial in Od. 7.221) and á¼ÎºÎµá½»Î¸Î±Î½Î¿Î½ (Il. 3.453) are hapax legomena that look like artificial extensions of Î»á½µÎ¸Ï and κεύθÏ, respectively.
It is not straightforward to indicate the origin of this aspiration, which may have been adopted from a different lexeme. Even so, the relic adjective νημεÏÏá½µÏ âunfailingâ shows that the aspiration cannot be old (cf. Beekes 1969: 109).
A similar picture is shown by the intransitive aorist of βλάÏÏÏ, which is usually á¼Î²Î»á½±Î²Î·Î½ in Classical Ionic-Attic, but á¼Î²Î»á½±Ïθην in Homer (with the exception, again, of the 3pl. forms βλάβεν, á¼Î²Î»Î±Î²ÎµÎ½). Perhaps, the θη-form was avoided in the case of ÏÏá½³ÏÏ because it was too ambiguous (á¼ÏÏá½±Ïθην, á¼ÏÏá½³Ïθην could also be thought to belong to ÏÏá½³ÏÏ; cf. ÏÏαÏθá¿Î½Î±Î¹ at Od. 15.80). In the case of βλάÏÏÏ, on the other hand, creating a thematic aorist may have been avoided in view of possible confusion with the archaic thematic root present βλάβομαι.
At Il. 23.84 van Thiel prints á¼Î»Î»â á½Î¼Î¿á¿¦, á½¡Ï á¼ÏÏá½±Ïημεν á¼Î½ á½Î¼ÎµÏá½³ÏοιÏι δόμοιÏιν (with McL); there is also a well-attested variant reading á¼ÏÏá½±Ïην ÏÎµÏ (the reading of the vulgate). However, the problem is that Aeschines, Contra Timarchum cites this passage with two plus-verses after line 83, and has our line 84 start with á½¡Ï á½Î¼Î¿á¿¦ á¼ÏÏá½±Ïεμέν ÏÎµÏ (á¼ÏÏá½±Ïομέν ÏÎµÏ conj. Scaliger). West prints á¼Î»Î»â á½Î¼Î¿á¿¦, á½¡Ï ÏÏá½±Ïομέν ÏεÏ, a conjecture by Buttmann that was also accepted by La Roche. See for further discussion Richardson ad loc., with reference to Van der Valk.
The Homeric perfect can translated as âavoir telle ou telle expression dans le regardâ (Chantraine 1927: 11), or âeinen bestimmten Blick, Gesichtsausdruck habenâ (Kölligan 2007a: 260). After Homer, the perfect occurs absolutely in the meanings âto be visibleâ (e.g. ÎºÎ»á½³Î¿Ï â¦ Î´á½³Î´Î¿Ïκε, Pi. Ol. 1.93) and, in the tragedians, âto have eyesightâ (as opposed to being blind).
See Kölligan (2007a: 264â265).
ÏούÏÎ¿Ï Ïá½³Î³Î³Î¿Ï á¼¥Î´Î¹Î¿Î½ δÏακεá¿Î½ (A. Ag. 602), δεινὰ δâ á½Ïθαλμοá¿Ï δÏακεá¿Î½ (A. Eum. 34), Î³á¾¶Ï â¨Ïââ© á½Î¼Ïαλὸν ÏÏοÏδÏακεá¿Î½ (A. Eum. 166), á¼ÏÏ ÎºÏον á½Î¼Î¼Î± ÏÏοÏδÏάκοι ([A.] PV 903b), á¼Î»Î»Î¿Ï Îµá¼°Ï á¼Î»Î»Î¿Î½ δÏακών (E. Herc. 951), á¼Î´Ïακον (E. Or. 1456). The meaning is âto beholdâ, always of spectacular or horrid sights, or of eye-contact. Sophocles does not have δÏακεá¿Î½ but uses δεÏÏθá¿Î½Î±Î¹ instead.
δέÏÏθηÏ(ε) ([A.] PV 93 and 141), á¼Î´á½³ÏÏÎ¸Î·Ï (ibid. 547), δέÏÏθη (S. Aj. 425), δεÏÏθένÏÎµÏ (S. fr. 387.2), and two with preverb, ÏÏοÏδεÏÏθῠ([A.] PV 53), καÏαδεÏÏθá¿Î½Î±Î¹ (S. Tr. 999). Sophocles uses only δεÏÏθá¿Î½Î±Î¹, while Euripides and Aeschylus use only δÏακεá¿Î½ (except for the Prometheus Vinctus, of which Aeschylus was probably not the author).
It is âauch nach Homer nur dichtersprachlich überliefertâ (Forssman 1997: 42). LSJ translates âwaste, ravage, sack, in Homer only of citiesâ, but as we will see there are traces of an original meaning âcut offâ. Jandaâs proposal (2000: 229â240) to translate Ïá½³ÏÎ¸Ï with âerbeutenâ cannot be upheld: see below.
The thematic aorist ÏÏαθε/οâ occurs 4â¯Ã in Pindar, alongside the s-aorist ÏεÏÏαâ (also 4â¯Ã) and the pres. ptc. ÏεÏθόμενοι (1â¯Ã). The tragedians only use the sigmatic stem ÏεÏÏâ.
There are only three forms of the thematic stem ÏεÏθε/οâ: Ïá½³ÏθονÏε, Ïá½³ÏθεÏο, ÏεÏθομένη, the latter two with passive meaning. Meillet speculated that Ïá½³ÏθεÏο and ÏεÏθομένη recover older thematic aorists *Ïá½±ÏθεÏο, *ÏαÏθομένη (see Chantraine 1958: 384 and 389â390, with further references). As for the hapax Ïá½³ÏθεÏο, Forssman (1997) agrees that this is an aorist formation, and suggests a conceivable scenario for its artificial creation. In addition, he argues that the aor. inf. Ïá½³Ïθαι is artificial. The participle ÏεÏθομένη was probably an aorist, too, because it only occurs in conjunction with the aorist ptc. á¼Î»Î¿á¿¦Ïα. This leaves us with the dual form Ïá½³ÏθονÏε as the only ascertained attestation of the present Ïá½³ÏθÏ.
As for the motive to create the aorist stem ÏεÏÏ(α)â, it is conceivable that the ptc. Ïá½³ÏÏανÏÎµÏ was first made as an alternative for metrically problematic ÏÏαθόνÏεÏ.
Cf. e.g. ÏÎ»á½µÎ¸Ï âto be filledâ, αἰÏθέÏθαι âto perceiveâ < *h2eu̯is-dh(h1)âe/oâ, Î»á½µÎ¸Ï âto go unnoticed, be hiddenâ, aor. λαθεá¿Î½ âto escape noticeâ; in all such cases, âθâ has become part of the verbal stem.
Jandaâs comparison (2000: 240) between á¼ á½² Ïá½³Ïοιεν ἤ κεν á¼Î³Î¿Î¹ÎµÎ½ (Il. 5.484) and Ïὴν δὲ διεÏÏάθομέν Ïε καὶ ἤγομεν á¼Î½Î¸á½±Î´Îµ ÏάνÏα (Il. 1.367) does not prove anything: in the second phrase, the object of διεÏÏάθομεν is a city which is stripped of all its valuables, that of ἤγομεν the possessions contained in it. The single attestation of bháre dhÄâ in the Rigveda (Janda 2000: 241) does not prove anything either.
Cf. LSJ s.v. Ïá½³ÏθÏ: âin Hom. only of townsâ; s.v. διαÏá½³ÏθÏ: âalways of citiesâ. The only exception occurs with á¼ÎºÏá½³ÏθÏ: Ïá½° μὲν ÏολίÏν á¼Î¾ÎµÏÏάθομεν âthat [booty] which we took from citiesâ, Il. 1.125. Janda puts too much emphasis on this single instance: usually á¼ÎºÏá½³ÏÎ¸Ï (like á¼Î¾Î±Î»Î±Ïá½±Î¶Ï âid.â) governs an accusative object, which shows that the preverb á¼Îºâ has no spatial value.
Another semantic development undergone by Ïá½³ÏÎ¸Ï is to âdestroy, slayâ, cf. ΤιÏύνθιον á¼ÏεÏÏαν ⦠ÏÏÏαÏόν (Pi. Ol. 10.32), á¼ÏεÏÏεν á¼Î½Î¸Ïá½½ÏÎ¿Ï Ï (S. Aj. 1198).
Lat. forfex âtongs, pincers; shears, scissorsâ is semantically close, but its ârfâ cannot be the result of regular sound change in Latin, so it could be a borrowing (from a Sabellic language? See EDL s.v. forfex). De Vaan (l.c.) also compares the Umbrian form furfaâ to Greek Ïá½³ÏθÏ, but does not comment on the semantics, which makes the proposal gratuitous.
The phonology of Lat. barba âbeardâ (word-initial bâ; a-vocalism) is not well understood, and its analysis remains uncertain.
After the Classical period, á¼Î´Ïαθον and á¼Î´Ïακον are used exclusively in Hellenistic poetry, and á¼ÏÏαθον is no longer current.
This does not imply the existence of an Aeolic phrase. However, I am no longer certain of the claim (which I made in Van Beek 2013) that there was an uninterrupted Ionian tradition. In the wake of works like Hooker 1977 and Hoekstra 1981, I would now rather assume a conservative poetic language that was preserved from Mycenaean times, containing both Mycenaean and (continental) Aeolic elements. During the Dark Ages this traditional language became popular especially with Ionian poets, who introduced forms with Î±Ï < *rÌ¥ such as καÏÏεÏá½¹Ï, ÏαÏÏá½³ÎµÏ and Ïá½±Ïμη. At a certain point, probably in the 9th or 8th c. BCE, they also started to modernize or update the language more structurally. This would be a natural period in which to date the vocalization of Epic *rÌ¥.
The middle aorist ÏÏαÏá½³Ïθαι is used 7â¯Ã with McL (see section 6.8.9); note that this form may have been protected against replacement because it also remained current in the vernacular.
See section 6.1 and 7.2.1, respectively.
The only exceptions to this claim are ἦλÏον, ἦλθον (beside á¼¤Î»Ï Î¸Î¿Î½), εá½Ïον, á¼ÏÏαιÏμε, á¼ÎµÎ¹Ïον, á¼ÏÎ±Ï Ïεá¿Î½, á½Î»Î¹Ïθε(ν), á¼Î½ÎµÎ¹ÎºÎ±, and ἠμαÏÏε (see Risch 1974: 238â¯ff. for the Homeric forms). The only middle thematic aorist taken into consideration here is ÏÏá½±ÏεÏο âturnedâ. This is an exceptional case: whereas active thematic aorists normally have intransitive meaning, ÏÏá½±Ïε has a transitive meaning âturned, changed the direction ofâ, where the object is e.g. a horse or an enemy. There are no other middle thematic aorists of the same metrical structure.
Mostly á¼ÏÏá½±ÏεÏâ; the unelided middle á¼ÏÏá½±ÏεÏο occurs only 4â¯Ã.
I have counted both the active and middle forms of the thematic aorist ÏÏαÏε/οâ, but excluded the middle forms with McL scansion that were discussed in section 6.8.9. The transitive semantics of the active ÏÏαÏέειν and the alternation with an intransitive middle ÏÏαÏá½³Ïθαι are not paralleled in the other five thematic aorists discussed here.
In view of Hermannâs bridge, prefixed forms are used in the fifth foot only when they are vowel-initial and preceded by a prepositive word, as in καὶ á¼ÏέδÏακον á¼Î½Ïην (Il. 24.223).
A similar distribution is found for á¼ÏÏαον âattackedâ (á¼ÏÏαε 3â¯Ã in the 5th foot, once verse-initially; á¼Ïá½³ÏÏαον 3â¯Ã in the 4th foot).
Most of these cases are in tmesis; they include the phrase Ïόλιν á¼ÏÏαθον (2â¯Ã), á¼Î´Ïαθâ á¼Î½á½¶ ÏÏοδόμῳ (Od. 20.143, nowhere in later Greek), and á¼Î´Ïακον á½Ïθαλμοá¿Ïι (Od. 10.197), which looks like a truncated form of the second hemistichs á¼Î½á½³Î´Ïακον á½Ïθαλμοá¿Ïι and á¼ÏέδÏακον á½Ïθαλμοá¿Ïι.
Even the phenomenon of âresonant lengtheningâ cannot be adduced as a parallel. In á¼Î¼Î¼Î±Î¸ÎµÎ½ (Od. 17.226) and á¼Î¼Î¼Î±Î¸ÎµÏ (Od. 18.362) as against μάθον (Il. 6.444) the phenomenon is exceptional and clearly secondary; the same holds for the non-etymological gemination in á¼Î»Î»Î±Ïε (h. Cer. 86 and 87), against which Homer only has á¼Î»Î±Ïον or λάÏον without resonant lengthening. Contrast the metrical behavior of λιÏε/οâ âto leaveâ, where root-initial λâ generally counts as a single consonant (72â¯Ã λιÏε/oâ in the first thesis syllable, against only once á¼Î½á½¶ ÏÏá½¹Î»ÎµÏ Î»á½·ÏεÏâ á¼Î½á½µÏ Il. 24.707). In á¼Î»Î»Î±Î²Î¿Î½ âtookâ the geminate reflects etymological *hlâ < *slâ, and εá½Î±Î´Îµ âit was agreeableâ reflects *e-hu̯ad-e. See Chantraine (1958: 176â177), and Eben (2004) for a more extensive discussion of âresonant lengtheningâ in Homer.
This explains why Homer could use unaugmented κάÏθανε (Il. 9.320) as a gnomic aorist, instead of metrically difficult á¼Î¸Î±Î½Îµ (for the problem, see already Meister 1921: 35, in whose view κάÏθανε stands in for καÏέθανε).
The only exception is Ïαμᾶζε θοÏών (Il. 10.528).
Note that apocope of the preverb was not an available option for pre-forms with á¼Ïιâ (á¼ÏέδÏαμον), á¼Ïοâ (á¼ÏέδÏαμον), διαâ (διέÏÏαθον, διέδÏακον, διέδÏαμον), and ÏεÏιâ (ÏεÏίδÏαμον). The pre-forms with *rÌ¥ could only be combined with such preverbs by using the tmesis construction.
The isolated εá½Î´ÏÎ±Îºá½µÏ âseeing wellâ (only S. Phil. 846) is a deverbal compound derived from δÏακεá¿Î½ (cf. Meissner 2006: 216). It is distinguished in both form and meaning from old compounds like εá½Î´ÎµÏÎºá½µÏ âwell visibleâ, and clearly secondary.
Cf. e.g. Peters (2004); Willi (2018: 294).
Cf. especially Ïá¿¦Ï Î´â á½Ïθαλμοá¿Ïι δεδοÏÎºá½½Ï âlooking like fire with his eyesâ, i.e. âblazing fireâ (of a boar, Od. 19.446). See also δεινὸν δεÏÎºá½¹Î¼ÎµÎ½Î¿Ï âwith a fearsome gazeâ (Il. 3.342, 11.37, and 23.815), á½Î¼Î¼Î±Ïι δέÏκομαι λαμÏÏόν lit. âI look bright with my eyeâ (Pi. Nem. 7.66), and ÏμεÏδαλέον δὲ δέδοÏκεν â(the snake) looks terribleâ (Il. 22.95).
Note that the preservation of καÏέδÏακεν in hexameter verse would be well-motivated: *kateu̯iden would be unmetrical. The Pindaric passage goes: ÏίλοιÏι Î³á½°Ï Ïá½·Î»Î¿Ï á¼Î»Î¸á½¼Î½ ξένιον á¼ÏÏÏ ÎºÎ±ÏέδÏακεν ἩÏÎ±ÎºÎ»á½³Î¿Ï á½Î»Î²á½·Î±Î½ ÏÏá½¸Ï Î±á½Î»á½±Î½ (Nem. 4.22â24). The interpretation of these lines continues to present problems: Willcock (1995: 97) even states that âthe expression is awkwardâ. Taking ξένιον á¼ÏÏÏ with καÏέδÏακεν and á½Î»Î²á½·Î±Î½ ÏÏá½¸Ï Î±á½Î»á½±Î½ with á¼Î»Î¸á½½Î½ requires a heavy hyperbaton. In my view, the idea to emend αá½Î»á½±Î½ to a gen. pl. αá½Î»á¾¶Î½ is worthy of consideration, because this yields a natural interpretation of ÏÏá½¹Ï: âFor having come as a friend to friends, he looked down upon a hospitable city from the blessed halls of Heracles.â
E.g. á½¡Ï ÏÏιÏόλβιοι κεá¿Î½Î¿Î¹ βÏοÏῶν, οἳ ÏαῦÏα δεÏÏθένÏÎµÏ Ïέλη μόλÏÏâ á¼Ï á¼Î¹Î´Î¿Ï (S. fr. 387).
The form ÏÏá½±Ïε (Nem. 3.53) is best analyzed as an imperfect: cf. Slater (1969, s.v. ÏÏá½³ÏÏ).
Henry (2005: 33) suggests that âPindar may have used δÏÎ±ÎºÎµá½·Ï (etc.) rather than δÏακών (etc.) in order to avoid confusion with forms of the substantive δÏάκÏν, indistinguishable in strophic song from those of δÏακών. There was no danger of such a confusion outside the participle.â However, I fail to understand how δÏακών and δÏάκÏν, given their different accents, could ever be confused.
Forssman (1964: 18 n. 6) remarks: âδÏακένÏâ ist also nicht mit á¼ÏιÏένÏι (dat.) âstürzend, fallendâ Pi. Ol. II 43 auf eine Stufe zu stellen (â¦), das gegenüber hom. á¼ÏιÏών (zu ἤÏιÏε) auf á¼ÏιÏÎµá½·Ï weist: Hier handelt es sich um ein intransitives Verbum.â However, I see no reason why δÏακένÏâ and á¼ÏιÏένÏâ, attested more than two centuries after Homer, could not be replacements of older thematic aorist forms.