Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Cover
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Preface
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Introduction
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Chapter 1: The History of the "Mixed Aorist" Problem
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Chapter 2: Oisete
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Chapter 3: Duseto and Beseto
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Chapter 4: Orseo and Lexeo
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Chapter 5: Hixon
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - Chapter 6: Conclusions
Catharine P. Roth, "Mixed Aorists" in Homeric Greek - List of Works Cited
3. Δύσετο and Βήσετο
The "mixed aorists" are by no means a homogeneous group. If οἴσετε, ὄψεσθε, ἄξετε, and ἄξοντο were created by certain processes within the epic language, we cannot therefore assume, as Leumann does, that the same explanation is valid for ἐβήσετο, βήσεο, ἐδύσετο, δύσεο and their compounds.  The reading with thematic vowel was upheld by Aristarchus against the reading common in manuscripts with alpha (see above, chapter one). He was certainly correct, both because the thematic vowel is the lectio difficilior from a scribe's point of view, and because the sigmatic aorists of these verbs are regularly causative. For the causative middle one may cite Callimachus Hymn 5.65 πολλάκις ἁ δαίμων νιν ἑῷ ἐπεβάσατο δίφρω, "Often the goddess set her on her own chariot." The intransitive thematic forms are therefore to be retained in Homer.
Moreover, it is not sufficient to say, as Wahrmann does, that these forms are metrical variants of the athematic aorist.  Although one may cite, for example, B 35 ὣςἄρα φωνήσας ἀπεβήσετο, τὸν δ'ἔλιπ' αὐτοῦ as a metrical rearrangement of Z 116 etc.: ῾ὼς ἄρα φωνήσας ἀπεβη κορυθαίολος Ἕκτωρ· such parallel verses show only how the use of ἐβήσετο was extended, not how it arose in the first place. We must study all the occurrences of the forms in question to determine which are basic and which are derivative. The formulaic system of the epic language makes such a determination possible. Then it is from the basic usages, not from the secondary, that the original significance of the forms must be sought.
Both ἐδύσετο and ἐβήσετο are well attested in the epic formula-system. They occur in repeated phrases, often accompanied by archaic elements in the cadence (gloss-words or old case-forms), which help to prove the antiquity of the whole phrase. The imperatives δύσεο and βήσεο, on the other hand, occur in derivative phrases.  For δύσετο, the most common formula describes the setting of the sun (H 465, β 388, γ 487, 497, λ 12, ο 185, ο 296, ο 471, ζ 321, θ 417); δύσετο δ' (τ') ἠέλιος... A formula for diving into the sea occurs four times, with slight variations: δ 570 ὥς εἰπὼν ὑπὸ πόντον ἐδύσετο κυμαίνοντα, cf. δ 425, ε 352, λ 253, and Ξ 229 with ἐβήσετο. There probably was a formulaic expression for entering a crowd of people or a city, as in these verses: Φ 515 αὐτὰρ 'Απόλλων Φοῖβος ἐδύσετο Ἴλιον ἱρήν· Υ 379 ῾ὼς ἔφαθ', Ἕκτωρ δ'αὖτις ἐδύσετο οὐλαμὸν ἀνδρῶν. Notice the archaic cadence-formula οὐλαμὸν ἀνδρῶν preceded as always by hiatus representing an original digamma. A series of formulae describe a warrior putting on his armor: Γ 328 αὐτὰρ ὅ γ' ἀμφ' μοισιν ἐδύσετο τεύχεα καλὰ cf. H 103, N 241, Ψ 366. I 596 βῆ δ'ἰέναι, χροῖ δ'ἔντε' ἐδύσετο παμφανόωντα· cf. 0 120. λ 16 'Αργείους ἐν δ'αὐτὸς ἐδύσετο νώροπα χαλκόν. cf. B 578; another γ archaic cadence: the Homeric gloss νώροπα occurs only in this phrase with χαλκόν (also in the dative).
The imperative δύσεο occurs only four times, always in less standardized phrases, in each case derived from the formulae with ἐδύσετο. Thus Π 129 δύσεο τεύχεα θᾶσσον, ἐγὼ δέ κε λαὸν ἀγείρω reflects δύσετο τεύχεα καλὰ..,(N 241 etc.). In T 36: αἶψα μάλ' ἐς πόλεμον θωρήσσεο, δύσεο δ'ἀλκήν we have an extension of the arming-formulae, influenced by I 231: …εἰ μὴ σύ γε δύσεαι ἀλκήν. Likewise, Σ134 ἀλλὰ σὺ μὲν μή πω καταδύσεο μῶλον Ἄρηος is derived from verses like Y 379 ῾ὼς ἔφαθ', Ἕκτωρ δ'αὖτις ἐδύσετο οὐλαμὸν ἀνδρῶν. Notice the construction of μή with καταδύσεο, which implies that the imperative is not understood as aorist (the present imperative or aorist subjunctive is used in prohibitions, a rule to which this verse is cited as an exception).  A further extension of the formula for entering a crowd is seen in ρ 276: δύσεο δὲ μνηστῆρας, ἐγὼ δ'ὑπολείψομαι αὐτοῦ. Clearly, the use of the imperative follows no such pattern as that of the indicative.
Similarly, ἐβήσετο occurs in repeated phrases, from which the expressions with βήσεο are derived. One verse appears three times, varying only in the gender of the subject: Z 288 αὐτὴ δ'ἐς θάλαμον κατεβήσετο κηώεντα, cf. O 191, o 99. Here κηώεντα is a Homeric gloss. Another verse occurs twice, with a difference of preverb: α 330 κλίμακα δ'ὑψηλὴν κατεβήσετο οἷο δόμοιο, cf. φ 5 with προσεβήσετο. Notice the archaic genitive-ending in οίο δόμοιο and the observance of the initial digamma (original *sw-). These formulae may be modified, as in β 337: ῾ὼς φἀν ὁ δ' ὺψοροφον θάλαμον κατεβήσετο πατρός. Or they may be used to create variants of formulae with the athematic aorist, as the formula ... ὑπέρβη λάϊνον οὐδὸν (θ 80 etc.) is adapted in ν 6: ῾ὼς εἰπὼν ὑπὲρ οὐδὸν ἐβήσετο δῖος 'Ωδυσσεύς. cf. η 135, φ 43, H. Merc. 233. The expressions for entering a house can be applied to landing on a shore, as in δ 521: ἧ τοι ὁ μὲν χαίρων ἐπεβήσετο πατρίδος αἴης, cf. H. Apoll. 49, and in the second person H. Apollo 141: ἄλλοτε μέν τ'ἐπὶ Κύνθου ἐβήσεο παιπαλόεντος (codd. and edd. ἐΒήσαο). These are secondary developments, but verses like Z 288 and α 330 (cited above) sufficiently demonstrate the formulaic usage of ἐβήσετο in a domestic context. The imperative, on the other hand, does not occur in such a context.
A more complicated group of formulae describes a man mounting his chariot or dismounting from it. To see which expressions are basic and which are derived, we will depend upon two principles: that a phrase which appears in the cadence of the verse is likely to be older than one which appears elsewhere (the cadence being regularly a repository of archaisms); and that a phrase which contains a relatively late form of a word is likely to be more recent than a similar phrase with no late forms. The only version of the chariot-formula which is both located in the cadence and free from late forms is that which we see in Γ 262 = Γ 312: πὰρ δέ οἱ 'Αντήνωρ περικαλλέα βήσετο δίφρον and γ 481: 'ὰν δ'ἄρα Τηλέμαχος περικαλλέα βήσετο δίφρον. The fact that the same formula occurs in both epics helps to convince us that it is traditional. Further development of the chariot-formulae is found only in the Iliad, probably for reasons of content. There is a version of the formula with the irresolvable contracted genitive δίφρου at the end of the verse, presumably more recent than the version with the accusative; this comes either with the indicative, as in Θ 44 = N 26: χρυσείην εὔτυκτον, ἑοῦ δ'ἐπεβήσετο δίφρου cf. Ω 322, or with the imperative, as in E 109: ὄρσο πέπον Καπανηϊάδη, καταβήσεο δίφρου. A phrase occurring in the middle of the verse, hence to be considered derivative, uses both the indicative and the imperative: Λ 512 ἄγρει, σῶν ὀχέων ἐπιβήσεο, πὰρ δὲ Μαχάων βαινέτω... cf. E 221, Θ 105; Λ 517 αὐτίκα δ'ὧν ὀχέων ἐπεβήσετο, πὰρ δὲ Μαχάων βαιν', 'Ασκληπιοῦ υἱὸς ἀμύμονος ἰητῆρος. The enjambment of these verses and the disregard of digamma in Λ 517 support the contention that the phrases are relatively recent. Accordingly, if the imperative occurs not at all in the basic formula and most frequently in the latest derived expression, it follows that the imperative βήσεο must have been derived from the indicative ἐβήσετο and not vice versa: "nur die 3. P. Sing. ursprünglich erscheint ..." as Wahrmann says. 
The next question is, what is the tense of ἐδύσετο and ἐβήσετο as they are used in Homer? If they are sometimes indistinguishable in sense from the aorists ἔδυ and ἔβη, it does not follow that their basic usage is aorist; in fact it is more likely that assimilation to the sense of the aorist is a later development. Whatever traces there may be of a different sense are the best evidence for the original usage. To start with the most clearly formulaic verse, β 388 etc.: δύσετό τ'ἠέλιος σκιόωντό τε πᾶσαι ἀγυιαί· δύσετο is coupled with an imperfect describing the evolving situation: "the roads were growing dark." In H 465, also: δύσετό τ'ἠέλιος, τετέλεστο δὲ ἔργον Ἀχαιῶν, the pluperfect describes a state existing in the past: "the job was complete." In both cases, therefore, δύσετο ought to be some kind of descriptive imperfect. The sense of this imperfect may be understood more clearly by contrast with the aorist used to express the accomplished fact of nightfall: A 475 ἧμος δ'ἠέλιος κατέδυ καὶ ἐπὶ κνέφας ἧλθε, "when the sun had set and darkness had come," or Θ 487-8: Τρωσὶν μέν ῥ' ἀέκουσιν ἔδυ φάος, αὐτὰρ 'Αχαιοῖς ἀσπασίη τρίλλιστος ἐπήλυθε νὺξ ἐρεβεννή. “The sun set... night came." The future participle δυσόμενος will also help us to understand δύσετο. This participle appears in the Odyssey at α 24: οἱ μὲν δυσομένου 'Υπερίονος, οἱ δ'ἀνιόντος, "some where Hyperion descends toward setting, the others where it ascends," and in Hesiod Works and Days 383f: Πληιάδων 'Ατλαγενέων ἐπιτελλοηενάων ἄρχεσθ' ἀμήτου, ἀρότοιο δὲ δυσομενάων, "begin the harvest when the Pleiades are rising, begin plowing when they are going to set." To these we may add H. Merc. 197, if Voss is right in reading καταδυσομένοιο for καταδυομένοιο (on the basis of the long υ):  ταὶ δ'ἔβαν ἠελίοιο νέον καταδυσομένοιο "the cows went when the sun was just about to set." It would seem that sunset (or star-set) was conceived as an instantaneous event, expressed by the aoristic radical form ἔδυ. At a given moment, the sun was thought of as either still up or else already down. One evidently did not say, "the sun is setting" (δύεται) or "the sun was setting" ἐδύετο). The only kind of durative which seemed possible was, "the sun is about to set" (δύσεται). The participle of δύσεται is δυσόμενοσ "about to set," and the past tense is ἐδύσετο"was about to set.”  So the complete translation of β 388 etc. is this: "The sun was about to set, and all the roads were growing dark." The time of day is the same as in Plato's Phaedo l16e: καὶ ὁ Κρίτων, 'Αλλ' οἶμαι, ἔφη, ἔγωγε, ὧ Σώκρατες, ἔτι ἥλιον εἷναι ἐπὶ τοῖς ὄρεσιν καὶ οὔπω δεδυκέναι.
The arming-formulae present a less clear picture. Sometimes one cannot see any special sense in ἐδύσετο, though even in such cases there is no reason to consider it a replacement for the aorist rather than the imperfect. Compare N 241: δύσετο τεύχεα καλὰ περὶ χροῖ, γέντο δὲ δοῦρε, and N 25: χρυσὸν δ' αὐτὸς ἔδυνε περὶ χροῖ, γέντο δ'ἱμάσθλην. Regularly the formulae with ἐδύσετο serve to introduce arming scenes. Sometimes the description is carried no further; sometimes it is developed with considerable elaboration, as at Γ 328 and Λ 16. In these instances ἐδύσετο could have an inchoative or a prospective sense. Twice ἐδύσετο introduces an abortive arming-scene; these passages are particularly instructive. At H 103: ὣς ἄρα φωνήσας κατεδύσετο τεύχεα καλά· Menelaus has volunteered to fight Hector, but is then dissuaded by Agamemnon. At O 120: …αὐτὸς δ'ἔντε' ἐδύσετο παμφαμόωντα. Ares has expressed his intention to avenge his son's death, but then Athena persuades him not to risk Zeus' displeasure. Perhaps ἔντε' ἐδύσετο originally meant "he was about to put his armor on" or "he wanted to put his armor on" - what at a slightly later stage of the language would be δῦναι ἔμελλε or δῦναι ἤθελε. The future in Greek was originally a derivational category, the desiderative, which did not necessarily imply fulfillment. This remained true of the future participleAs the desiderative became an inflexional category, namely the future tense, the fulfillment of the action came to be assumed. Meanwhile, periphrastic constructions took over the function of expressing imminence or intention. When the future tense implied fulfillment, then ἐδύσετο was taken as meaning that the armor was actually put on. Accordingly, someone had to insert a mention of its subsequent removal (Η 122, 0 125).
That ἐδύσετο at one time had a desiderative sense is also suggested by the expressions which commonly precede it. Several times a person first commands others to do certain things and then ἐδύσετο his armor:
'Ατρείδης δ'ἐβόησεν ἰδὲ ζώννυσθαι ἄνωγεν 'Αργείους· ἐν δ'αὐτὸς ἐδύσετο νώροπα χαλκόν. Ω 119-120 ὣς φάτο, καὶ ῥ' ἵππους κέλετο Δεῖμόν τε Φόβον τε ζευγνύμεν, αὐτὸς δ'ἔντε' ἐδύσετο παμφανόωντα. Ψ361 - 366 σοὶ δέ, γύναι, τάδ'ἐπιτέλλω... ἧσθαι, μηδέ τινα προτιόσσεο μηδ'ἐρέεινε. ἦ ῥα καὶ ἀμφ' μοισιν ἐδύσετο τεύχεα καλά.
Λ 15-16.Perhaps ἐδύσετο in these examples should be translated "he intended to put on" (his armor): ἐδύσετο represents the subject's intention for himself, as the commands, direct or indirect, represent his intentions for others.
Similar phenomena can be observed in the case of ἐβήσετο. The formula for entering a chamber occurs each time preceded by a verb of commanding:
ὣς ἔφαθ', ἡ δὲ μολοῦσα ποτὶ μέγαρ' ἀμφιπόλοισι κέκλετο· ταὶ δ' ἄρ'ἀόλλισσαν κατὰ ἄστυ γεραιάς. αὐτὴ δ'ἐς θάλαμον κατεβήσετο κηώεντα. O 189-191 αὐτὰρ ὅ γ' υἷἷας ἄμαξαν ἐύτροχον ἡμιονείην ὁπλίσαι ἠνώγει, πείρινθα δὲ δῆσαι ἐπ' αὐτῆς. αὐτὸς δ' ἐς θάλαμον κατεβήσετο κηώεντα. O 97-99 τὸν πῦρ κῆαι ἄνωγε βοὴν ἀγαθὸς Μενέλαος ὀπτῆσαί τε κρεῶν· ὁ δ' ἄρ' οὐκ ἀπίθησεν ἀκούσας. αὐτὸς δ'ἐς θάλαμον κατεβήσετο κηώεντα.
Z 286-288One may infer that ἐβήσετο like ἐδύσετο had originally a desiderative sense. In κ 105ff: κούρη δὲ ξύμβληντο πρὸ ἄστεος ὑδρευούση, ... ἡ ηὲν ἄρ' ἐς κρήνην κατεβήσετο καλλιρέεθρον, an imperfect force of some kind is necessary to describe the situation of the girl when she is met. An imperfect expressing intention or imminence makes good sense: "she was intending to go" or "she was about to go." For the relation of tenses, we may compare η 18f, with the periphrastic expression of imminence in the past: ἀλλ' ὅτε δὴ ἄρ' ἔμελλε πόλιν δύσεσθαι ἐραννήν, ἔνθα οἱ ἀντεβόλησε θεὰ γλαυκῶπις 'Αθήνη.
The military formulae having to do with chariots present little evidence for the original sense of ἐβήσετο. As has been seen, most of the attested phrases are secondary developments; so it is not surprising if the original usage has been obscured. Like ἐδύσετο, ἐβήσετο can be considered imperfect as easily as aorist. Compare: Γ 261 ᾿ὰν δ'ἄρ᾿ ἔβη Πρίαμος, κατὰ δ' ἡνία τεῖνεν ὀπίσσω· Γ 311 ᾿ὰν δ'ἄρ' ἔβαιν' αὐτός, κατὰ δ' ἡνία τεῖνεν ὀπίσσω· Γ 262 = Γ 312 πὰρ δέ οἱ 'Αντήνωρ περικαλλέα βήσετο δίφρον. In general, the imperfect ἔβαινε is more common than the aorist ἔβη for mounting chariots (e.g. E 364, E 365, E 837, γ 483, γ 492, ο 145, ο 190; but aorist in Ψ 132, Ψ 352). Since ἔβαινε is so common, ἔβήσετο may well be admitted as an imperfect. Like ἐδύσετο, it stands in formal relationship to a future participle, seen in E 46 = Π 343: νύξ᾿ ἵππων ἐπιβησόμενον κατὰ δεξιὸν ὦμον· "He stabbed him in the right shoulder as he was about to mount his chariot." The same idea could once have been expressed with the finite verb ἐβήσετο in a temporal clause followed by νύξε or the equivalent in the main clause. But in the epics as we have them the original sense of ἐβήσετο in the chariot-formulae has been lost, and the form is preserved merely as a useful metrical variant of ἐβαινε.
The foregoing discussion has aimed to show that not only the form but also occasionally the Homeric usage supports the hypothesis that ἐβήσετο and ἐδύσετο are the secondary-tense forms of βήσεται and δύσεται. To us who have been raised on standard Attic Greek, the idea of a secondary tense of the future seems anomalous; but several considerations may make it seem less strange. In the first place, there are other Indo-European languages where the opposition of primary and secondary endings has been extended to the future system. Whitney's description of the Sanskrit conditional exactly fits ἐβήσετο and ἐδύσετο: "from the future-stem is made an augment-preterit, by prefixing the augment and adding the secondary endings, in precisely the same manner as an imperfect from a [thematic] present-stem...”  There is only one example of this formation in the Rig-Veda : ábharisyat "was going to carry off" (2.30.2). That is to say, the conditional is an innovation which had barely begun at this stage of the language, but which received a wider extension later. Parallel innovations also took place in Old Irish and in the Romance languages. In French, for example, as *cantare habeo gave the future je chanterai, so *cantare habebam gave the conditional je chanterais.  The same relationship may be seen in English between "I will sing," with the present tense of the auxiliary, and "I would sing," with the auxiliary in the past tense. In these languages the innovation was successful; whereas in Greek as we know it, it was abortive. Indeed, for all we know, it may have been created only in the epic language, and never have appeared in the natural language at all. Why did not the formation survive in Greek? A comparison with the other languages mentioned may give some clues. The secondary-future forms generally have two functions, derived from some original future-directed modality: the primary function of expressing futurity in the past, and a secondary modal function which gives the name "conditional." In modern French, the conditional is used mainly in the modal function, while in the function of futurity-in-the-past, periphrases tend to replace it i.e. j'allais chanter instead of je chanterais . Compare English "I was going to sing" instead of "I would sing." As usual, the newer forms take the primary function.  In Greek, the modal function was already occupied by the optative, and the function of futurity-in-the-past was renewed just as it has been in French and English. The regular construction of the periphrastic future is with the verb μέλλειν, in the present or the imperfect. We may show the renewal as follows:
δύσεται → δῦναι (δύσεσθαι) μέλλει
ἐδύσετο → δῦναι (δύσεσθαι) ἔμελλε.
The beginning of this functional replacement can be seen already in Homer: η 18 ἀλλ'ὅτε δὴ ἄρ'ἔμελλε πόλιν δύσεσθαι ἐραννήν. This represents a renewal of e.g. Φ 515 αὐτὰρ 'Απόλλων Φοῖβος ἐδύσετο Ἴλιον ἱρήν. Sometimes ἐδύσετο is replaced by expressions of desire, recalling the original desiderative character of the future: Λ 537 ...ὁ δὲ ἵετο δῦναι ὅμιλον ἀνδρόμεον ῥῆξαί τε μετάλμενος ... Υ 75 - 76 ... αὐτὰρ 'Αχιλλεὺς Ἕκτορος ἄντα μάλιστα λιλαίετο δῦναι ὅμιλον ο 328 εἰ δὴ μνηστήρων ἐθέλεις καταδῦναι ὅμιλον. These verses illustrate the original sense of Υ 379; ὣς ἔφαθ᾿, Ἕκτωρ δ'αὖτις ἐδύσετο οὐλαμὸν ἀνδρῶν. The periphrastic constructions make it clear that the Greeks could use an expression for futurity or intention in the past, and the parallels from other Indo-European languages provide support for the form of a secondary-future tense. When we add the evidence of Homeric usage, it seems necessary to conclude that ἐδύσετο and ἐβήσετο are in fact examples of a secondary future otherwise unattested in Greek. 
[ back ] 1. Leumann, Kleine Schriften 236ff.
[ back ] 2. Festschrift Kretschmer 309.
[ back ] 3. Wahrmann, Festschrift Kretschmer 310.
[ back ] 4. Wackernagel, Vor1esungen über Syntax 1.215.
[ back ] 5. Festschrift Kretschmer 310.
[ back ] 6. Schulze 316.
[ back ] 7. Magnien 2.2.
[ back ] 8. Whitney 334.
[ back ] 9. On which see Coleman, CQ 21 (1971) 215ff.
[ back ] 10. Kuryłowicz, Esquisses linguistiques 79.
[ back ] 11. Unless in Pamphylian: Schwyzer 1.788.