67. The augment takes two forms, the syllabic and the temporal.
The syllabic augment is the prefix ἐ-, and is used for stems beginning with a consonant. The temporal augment is a simple lengthening of the initial vowel of a stem, the vowels ᾰ- and ε- becoming η-.
ἠλήλα-το (Pf. ἐλήλα-ται)
ᾤχετο (οἴχο- μαι)
So the imperfect ᾖα I went (Sanskrit âyama), from the stem εἰ (εἶ-μι); as to the form ἤϊα see § 12.
Many seeming exceptions are due to the loss of the original initial consonants, ϝ-, σ-, ι̯-. The loss of one of these consonants may generally be presumed whenever we find the syllabic instead of the temporal augment. Thus—
ϝ- has been lost in ἐ-άγη and ἔ-αξε (ἄγνυμι), ἐ-άλη (ϝελ-), ἔ-ειπε, ἑ-έσ-σα-το (ἕννυμι), εἶδον (for ἔ-ϊδο-ν), ἐ-ώθεο-ν; so perhaps, with contraction of εε- to ει-, εἴρυ-σα (ϝερυ-), and εἶλο-ν.
For ἔϊδον there is an Aeolic form εὔιδον (ἔ-ϝιδον, cp. εὔαδε), which should perhaps be restored in some at least of the numerous places where the present text of Homer has εἴσιδε (Nauck, Mél. gr.-rom. ii.407).
σ- in ἑ-έσσα-το (for ἐ-ἑσσα-, from σεδ-), and, with contraction, εἵπε-το (σεπ-), εἷσα-ν (σεδ-), εἶχο-ν (σεχ-), εἷρπο-ν (σερπ-). In these cases the -σ- passed into the rough breathing, which was then thrown back on the αugment: but εἶχον has the smooth breathing owing to the following -χ-. Also εἴα (ἐάω for σεϝάω).
ι̯- (or y) perhaps in ἕηκα (for ἐ-ι̯ηκα) and, with contraction, εἷμεν (ἐ-ἑ-μεν), and παρ-είθη (-ε-εθη), But see § 16.
Several Homeric forms have been supposed to point to a syllabic augment ἠ- (instead of ἐ-), One of these—ἤϊα went—has been already explained (§ 12). As to the others we have to note as follows.
- ἤειρεν (Il. 10.499) is not from εἴρω to join together (Lat. sero), but from ἀείρω: for, as Cobet has shown (Misc. Crit. p. 326), ἀείρω is a technical word in the sense required (cp. Il. 15.680 συναείρεται ἵππους, also the words ξυνωρίς, for ξυν-αορ-ίς, and παρ-ήορος).
- In several words (as usually written) the initial vowel of the stem is lengthened after ἐ-ϝ-
ἑήνδανε (for ἐ-῾ ϝανδανε)
ἀν-έ-ῳγεν, ἀν-έ-ῳξε (ἀνα-ϝοίγω)
also ἐ-ᾱ́γη (ϝάγ-νυμι), with ᾱ in one place (Il. 11.559), and the pluperfect forms
ἐώλπει (ἔολπα, ϝελπ-)
ἐώργει (ἔοργα, ϝεργ)
ἐώκει (ἔοικα, ϝῐκ-).
In some of these there may be merely confusion with the later use of the temporal augment: e. g. ἑήνδανε is doubtless due to the Attic ἥνδανε, a form which arose after the loss of ϝ-. Hence recent editors write ἑάνδανε, ἐοινοχόει, ἀνέοιγον, also ἐόλπει, ἐόργει, ἐοίκει.
Note— With ἑήνδανον compare the aorist form ἔηξα (for ἔαξα), preserved in the text of Zenodotus in Il. 13.166 (ξυνέηξε for ξυνέαξε) and 13.257 (κατεήξαμεν for κατεάξαμεν). In this case the change to η did not make its way into the vulgate—perhaps because the form ἦξα, which suggested it, was a rarer word than ἥνδανον.
- A different explanation is required for ἐάγη (ᾱ), supported as it is by Attic ἑώρων (ὁράω) and ἑᾱ́λων (ᾰ in ἁλῶναι, etc.).1 These point to an augment ἠ-, the combinations ηϝο-, ηϝᾰ- passing into εω-, εᾱ- (as in βασιλέως, -εᾱ for -ηϝος, -ηϝα). Such an augment is also found in ἠείδης, ἠείδει (plpf. of οἶδα), and ἤϊσκε. There is much probability in the suggestion of G. Meyer (G. G. p. 423) that this ἠ- is a temporal augment obtained from the prothetic ἐ- so often found before -ϝ-: e. g. in ἐ-εισάμενος (ϝειδ-). Thus ἤϊσκε would be the augmented form of ἐΐσκω, not of ἴσκω.
- The forms ἀνέῳγε, ἀνέῳξε are peculiarly difficult on account of the Homeric present ὀΐγ-νυμι, aorist ὤϊξα, and Lesbian ὀείγω (pres. inf. ὀείγην, Coll. 214, 43). We might read ἀν-όειγε, etc., but the ordinary forms οἴγω (Hes. Op. 817), ἀν-οίγω, etc., would still be unexplained.
Initial ρ- is nearly always doubled, initial λ-, μ-, ν-, σ- very often. This may often be explained as the assimilation of an original initial ϝ- or σ-: thus ἔρρηξα is for ἔ-ϝρηξα, and so ἔρρεξε (ϝεργ-) and ϝρεγ-), ἐρρίγησε (ϝρῑγ-). Again ἔρρεεν is for ἔ-σρεεν, ἔννεον for ἔ-σνεον, ἔλλαβε perhaps for ἔ-σλαβε (Joh. Schmidt, Pluralb. p. 434). So ἔδδεισεν (which Ar. wrote ἔδεισεν) is for ἔ-δϝεισεν: and ἔσσευα probably for ἐ-κι̯ευα (Sanskrit root çyu). So too in ἐ-γδούπησαν the -γ- reappears which is lost in the unaugmented δούπησεν.
There are instances, however, to which this explanation does not apply, as ἔμμαθε. These are probably due to the influence of forms such as those already mentioned upon the traditional poetic dialect (Curtius, Stud. iv. 479 ff.; for a different view see Hartel's Homerische Studien). Cp. § 371.
68. The Pluperfect. The perfect stem forms the corresponding historical or past tense—the pluperfect—in two ways:
- Simply, with the augment (often omitted) and the secondary personal endings. All middle forms of the tense are of this kind, as ἐ-τέτυκ-το, ἐφ-ῆπτο, τετά-σθην, ἠλήλα-το. In the active the examples are comparatively few
δείδιε (Il. 18.34)
ἀνήνοθεν (Il. 11.266)
ἐπ-ενήνοθε (Il. 2.219)
ἐ-γέγωνε (Il. 14.469)
in Hesiod ἐπέφυκον: see § 27.
- By composition, with the augment and the suffix -εα (probably for
-εσα), joined to the longer form of the stem: e.g. ἐ-τεθήπ-εα, πεποίθεα, ἠνώγ-εα. The 3rd singular usually has -εε(ν) contracted
The plural occurs only once in Homer, in ἐοίκ-εσαν (Il. 13. 102): the dual never.
To this group belongs ᾔδεα knew, 2nd singular ἠείδης (for ἐ-ϝείδεας), also ᾔδησθα, 3rd singular ἠείδει, ᾔδει (or, as Aristarchus read, ἠείδη, ᾔδη). As to the augment ἠ- see § 67. In respect of form ᾔδεα is a sigmatic aorist, standing for ἐ-ϝείδεσα, Sanskrit âvedisham, and is only a pluperfect because it is used as the past tense answering to οἶδα (M. U. iii. p. 16).
69. Loss of Augment.The augment is so often dropped in Homer that the augmented and the unaugmented forms are almost equally numerous. It has been observed however2 that the forms without the augment are comparatively rare in the speeches, the proportion of augmented to unaugmented forms (excluding speeches which mainly consist of narrative matter) being about 10 to 3, whereas in narrative it is about 5 to 7. It would appear therefore that the augment is chiefly omitted where the context shows that past time is meant; and this is confirmed by the remarkable fact that the iteratives, which are only used as historical tenses, do not take the augment. The only clear instance of an iterative form with the augment is ἐ-μισγέσ-κοντο (Od. 20.7). On the forms ἔ-φασκο-ν, ἔ-φασκε see § 49.3
- 1. ἥλω was taken (Od. 22.230 σῇ δʼ ἥλω βουλῇ κτλ.) should perhaps be written ἑάλω. The stem ῾ϝᾰλω- appears in the moods (ἀλώω, ἁλῴην, ἁλῶναι, ἁλούς), except in the form ἁλόντε (Il. 5.487), where the meter requires -ᾱ.
- 2. Konrad Koch, De Augemento apud Homerum omisso, Brunswick, 1868.
- 3. In an article on the augment in the Journal of Philology (xix. p.211 ff.), Mr. Arthur Platt has shown that, in the case of the aorist, the choice between the augmented and the unaugmented form is largely determined by the sense in which the tense is used. In the common historical or narrative use the augment is often wanting; but in the uses which we may call non-narrative—the use for the immediate past (§ 76), and the gnomic use (§ 78)—the augmented form prevails. With the gnomic use the rule appears to be especially strict. This is obviously a valuable extension and generalization of the facts observed by Koch. In the case of the imperfect there seems to be a preference for unaugmented forms in continuous narrative; but the difference is much less marked. Mr. Platt gives some good reasons for believing that the number of unaugmented forms was originally greater than it is in our text. In this we find a fresh example of the modernizing process to which the poems were subjected from a very early time.