The Subjunctive

Book Nav


80. Athematic tense stems usually form the subjunctive by taking the thematic vowel, with the primary endings; except that when the thematic vowel enters into a diphthong, or is followed by two consonants, it becomes η or ω instead of ε or ο. 

Sing. Dual Plur.
Act. Mid. Act. Mid. Act. Mid.
1 -ομαι -ομεν -όμεθα
2 -ῃς -εαι  -ετον  -ησθον  -ετε -ησθε
3 -ῃ
-εται  -ετον  -ησθον -ωσι(ν)  -ωνται

The long η or ω, it will be seen, comes in place of ε or ο wherever it can do so without disturbing the meter.

Strong Aorists:

ἔ-φθη, subj. φθή-ῃ
ἔ-βη, subj. βή-ω (or βείω), ὑπερ-βή-ῃ, βή-ομεν (or βεί-ομεν)
ἔ-στη, subj. στή-ῃς, στή-ῃ, στή-ετον, στή-ομεν, στή-ωσι
ἔ-γνω, subj. γνώ-ω, γνώ-ομεν, γνώ-ωσι
ἔ-δυ, subj. δύω, δύ-ῃς, δύῃ
ἔ-βλη-το, subj. βλή-εται
ἔ-φθι-το, subj. φθί-εται, φθι-όμεσθα
ἄλ-το, subj. ἄλ-εται
stem θη-, subj. θεί-ω (or θή-ω), θή-ῃς, θεί-ομεν (or θή-ομεν), ἀπο-θεί-ομαι
stem ἡ-, subj. ἐφ-εί-ω, ἀν-ή-ῃ
stem δω-, subj. δώ-ῃ and δώ-ῃσι, δώ-ομεν, δώ-ωσι


εἰμί, subj. ἔ-ω (for ἔσ-ω), ἔ-ῃς, ἔ-ῃ and ἔ-ῃσι, ἔ-ωσι
εἶ-μι, subj. ἴ-ω, ἴ-ῃσθα, ἴ-ῃσι, ἴ-ομεν (ῑ̆)
φη-μί, subj. φή-ῃ
κιχῆ-ναι, subj. κιχεί-ω, κιχεί-ομεν (or κιχή-ω, κιχή-ομεν)

so ἐρεί-ομεν as if from *ἔρη-μι.

Passive Aorists:

ἐ-δάμη, subj. δαμεί-ω, δαμή-ῃς, δαμή-ετε

so δαεί-ω, ἁλώ-ω, ἁλώ-ῃ, σαπή-ῃ, φανή-ῃ, τραπεί-ομεν.

For δαινύῃ, 2 singular subjunctive middle (Od. 8.243, 19.328), we may read δαινύε᾿, i. e. δαινύ-ε-αι.


πέποιθα, subj. πεποίθ-ῃς, πεποίθ-ομεν
ἔρριγε, subj. ἐρρίγ-ῃσι
βέβηκε, subj. προ-βεβήκ-ῃ

so ἑστήκ-ῃ, ἀρήρ-ῃ, μεμήλ-ῃ, ὀλώλ-ῃ, ὀρώρ-ῃ, βεβρύχ-ῃ also ἱλήκῃσι (Od. 21.365)—unless we assume a present ἱλήκω (§ 45).

perf. mid. προσ-αρήρεται (Hes. Op. 431)
οἶδα, subj. εἰδέω, εἰδῇς, εἰδῇ, εἴδομεν, εἴδετε, εἰδῶσι

For εἰδέω, etc., Tyrannio wrote εἴδω, εἴδῃς, εἴδῃ, εἴδωσι (Schol. Od. 1.174), uniform with εἴδομεν, εἴδετε. Both forms may be accounted for: εἰδέω is subjunctive of ἐ-ϝείδεα (§ 68); εἴδω with the plural εἴδ-ο-μεν, εἴδ-ε-τε, is subjunctive of an athematic *ϝεῖδ-μι, Sanskrit ved-mi (M. U. iii. 18). The form ἰδέω, read by most MSS. in Il. 14.235, is a mere error for εἰδέω.

Aorists in -σᾰ:

ἐ-βήσα-μεν, subj. βήσ-ομεν
ἤγειρα, subj. ἀγείρ-ομεν
ἔ-τισα, subj. τίσ-ετε, τίσ-ωσι
ἠμείψα-το, subj. ἀμείψ-εται
ἠλεύα-το, subj. ἀλεύ-εται

and many more. These subjunctives properly belong to the older inflection of the sigmatic aorist without -ᾰ (§ 40).

To these should be added some forms used as futures.

ἔδ-ο-μαι, ἔδονται shall eat
(cp. Sanskr. ad-mi, Lat. est for ed-t)

δή-εις, δή-ομεν, δή-ετε shall find
with the strong stem answering to δᾰ(σ)- in δέδαεν, etc.

βεί-ο-μαι shall live, from the stem βίϝ-
Also in the form βέομαι. Evidently βείομαι : βιῶναι : : δήω : δαῆναι.

It will be found that the Homeric uses of these words are all such as can be referred to the subjunctive. On πίομαι and κείω see § 59. The form δήεις may be a trace of an older inflection, -ω, -εις, -ει, answering to -ομεν, -ετε.

It will be seen that the strong form of the stem is found in the subjunctive, as φή-ῃ, δώ-ομεν, ἑστήκ-ῃ. Apparent exceptions are

  1. The subjunctive of εἶμι—in which the ῑ of ἴομεν (for εἴ-ομεν) is unexplained, while the forms ἴ-ω, ἴ-ῃσι may be thematic, (as are opt. ἴοι, part ἰών); and
  2. The forms ἀφ-έ-ῃ (aor. of ἀφ-ίη-μι), μιγέ-ωσι, φθέ-ωσι, στέ-ωμεν, κτέ-ωμεν, φθέ-ωμεν, θέ-ωμεν, ἕ-ωμεν. These forms are the result of transference of quantity, στε-ω- for στη-ο-, etc., and it is important to notice that the last six are always scanned as disyllables, thus forming the transition to the contracted φθῶσι, στῶμεν, etc.

Anomalous lengthening is found in μετ-είω (Il. 23.47) for μετ-έ-ω.

81. Subjunctives with lengthened stem vowel. The formation of the subjunctive by means of the thematic vowel must have been confined originally to stems ending in a consonant, or in one of the vowels or u. The hiatus in such forms as φή-ῃ, στή-ομεν, γνώ-ομεν is enough to prove that they are not primitive. In Vedic Sanskrit, accordingly, while as-a-ti, han-a-ti are subjunctive of as-ti, han-ti, we find sthâ-ti, dâd-ti as the subjunctive answering to the aorists á-sthâ-t, á-dā-t. These would become in Homer στῆ-σι, δῶ-σι or (with the usual ι of the 3rd singular) στῇ-σι, δῷ-σι. Similarly we may infer an original plural στῆμεν, στῆτε, στῆντι (στῆσι); δῶμεν, δῶτε, δῶντι (δῶσι); and so on. The principle of the formation is that the stem ends in a simple long vowel—not one that has arisen from specifically Greek contraction.

Traces of this type of subjunctive are found in the Greek dialects.

δύνᾱ-μαι (for δύνωμαι)
προ-τίθηντι, etc. (Meyer, G. G. p. 502)

In Homer it may be recognized in the 3rd singular forms φῇσιν (Od. 1.168), φθῇσι (Il. 23.805), ᾗσι (Il. 15.359), μεθ-ίῃσι (Il. 13.234), δῷσι; perhaps in δῶ, δῷς, δῶμεν, δῶσι, περι-δώμεθον, ἐπι-δώμεθα; γνῷς, γνῶμεν, γνῶσι; ἐπι-βῆτον, πειρηθῆτον, etc.—which are usually regarded as contracted from the regular Homeric δώω, δώῃς, δώομεν, etc.—and in δύνη-ται, ἐπί-στηται (§ 87.3).

How then did the Homeric forms of the type of φή-ῃ, στή- ομεν, γνώ-ομεν arise? Doubtless by a new application of the process already familiar in ἴ-ο-μεν (εἶ-μι), φθί-ε-ται, χεύ-ε-ται, πεποίθ-ο-μεν, etc. We may compare the extension of the endings -ᾰται, -ᾰτο to the perfect βεβλή-αται, in imitation of κεκλί-αται, εἰρύ-αται (§ 5).

Contraction appears in the 3rd singular φῇ (Od. 19.122), στῇ (Od. 18.334), βῇ (Od. 2.358), φανῇ (Il. 9.707), γνῷ (Il. 1.411, 16.273)—unless we suppose that these are obtained by dropping the -σι of φῇ-σι, etc. on the analogy of the thematic -ῃ. Also in the 1st plural μεθ-ῶμεν (Il. 10.449), συν-ώμεθα (Il. 13.381), δαῶμεν (Il. 2.299), μεμν-ώμεθα (Od. 14.168); and the 3rd plural ὦσι (Il. 14.274, Od. 24.491), βῶσιν (Od. 14.86); but it is probably more correct to write these words with εω (like φθέωσι, ἕωμεν, etc.), except when a vowel precedes (as in δαῶμεν).

The two forms of the subjunctive present a certain analogy to the two kinds of derivative verbs—the Attic -αω, -εω, -οω, and the Aeolic -ᾱμι, -ημι, ωμι. Thus δύνᾱ-μαι, τίθη-ντι are related to δύνω-μαι, τιθέωσι nearly as φίλημεν, φίλεισι to φιλέομεν, φιλέουσι.

κεῖται occurs as a subjunctive in Il. 19.32, 24.554; Od. 2.102, 19.147. It has been explained as contracted from κεί-εται, the regular form answering to the athematic κεῖ-ται (Curt. Stud. vii. 100). The best MS. (Ven. A of the Iliad) gives κῆται. The true reading is probably κέεται (related to κείεται as τελέω to τελείω).

ζώννυνται, construed with ὅτε κεν (Od. 24.89) is regarded by Curtius as a subjunctive (Verb. ii. 67). But the example is uncertain; the clause refers to past time, so that ὅτε κεν with the subjunctive is quite irregular (§ 298).

σόῳ and σόῳς or σοῷs (Il. 9.424, 681) are probably optatives; see § 83.

82. Thematic tense stems form the subjunctive by changing ε into η and ο into ω.

The subjunctive of the thematic aorist and present frequently employs the personal endings -μι and -σι.

ἐθέλ-ωμι, ἐθέλ-ῃσι
εἴπωμι, εἴπῃσι
ἀγάγωμι, ἀγάγῃσι
τύχωμι, τύχῃσι
κάμῃσι, etc. (Bekker, H. B. i. 218).

These endings are also found (but rarely) with athematic stems.

Present: ἔ-ῃσι, ἴ-ῃσι (which however may be Thematic)

Aorist: δώ-ῃσι (Il. Il. 324)

Perfect: ἐρρίγ-ῃσι (Il. 3.353)

The 2nd singular sometimes takes -σθᾰ.

ἐθέλ-ῃσθα, εἴπ-ῃσθα, πίῃσθα, etc.

The subjunctive in -ωμι had almost disappeared at one time from the text of Homer, having been generally corrupted into -οιμι, sometimes -ωμαι. It was restored by Wolf, chiefly on the authority of the ancient grammarians. Some of the best MSS. (especially Ven. A) have occasionally preserved it.

It is interesting to observe the agreement in form between the thematic indicative and the athematic subjunctive; e. g. indicative ἄγω and subjunctive γνώ-ω, in contrast to subjunctive ἐθέλω-μι: just as ἄγο-μεν and γνώ-ομεν agree in contrast to ἄγω-μεν.

A few forms of the aorist in -σᾰ follow the analogy of the thematic stems.

ὄρσ-ωμεν (Il. 7.38)
ὄρσ-ητε (Il. 23.210)
δηλήσ-ηται (Il. 3.107),
μνησώμεθα (Il. 15.477, etc.)
παύσωμεν (Il. 7.29)
παυσώμεσθα (Il. 7.290, 21.467)
πέμψωμεν (Od. 20.383)
ἐνιπλήξωμεν (Il. 12.72)
φθίσωμεν (Od. 16.369)
περάσητε (Od. 15.453)
ἀντιάσητον (Il. 12.356)
τρώσητε (Od. 16.293, 19.12)
δείσητε (Il. 24.779)
βουλεύσωμεν (Od. 16.234)

In most of these instances the original reading is probably either a present subjunctive or an optative. Thus in Il. 21.467 the best MSS. have παυώμεσθα, and in Od. 20.383 there is good authority for πέμπωμεν (in Il. 15.72 the MSS. are divided between παύω and παύσω). Similarly we may read παύωμεν and ἐνιπλήσσωμεν. Again φθίσωμεν follows a past tense (§ 298), περάσητε an optative (§ 308.1.b): read φθίσαιμεν, περάσαιτε. For ἀντιάσητον we may have either the optative ἀντιάσαιτον or a present subjunctive ἀντιάητον. For τρώσητε we should perhaps read τρώητε (cp. the pres. ind. τρώει), and for βουλεύσωμεν βουλεύωμεν.

There are no clear instances of thematic stems forming the subjunctive with a short vowel (ε or ο).

The forms μίσγεαι, κατίσχεαι (Il. 2.232, 233), for μίσγηαι, κατίσχηαι, are like βέβληαι (Il. 11.380) in which the η forms a short syllable.

In Il. 14.484 τῷ καί κέ τις εὔχεται ἀνήρ κτλ. Hermannʼs conjecture καί τέ τις is found in two of La Rocheʼs MSS., and in any case the κε is unsuitable to the sense. The true reading is probably καί τίς τʼ (§ 332).

In Od. 4.672 ὡς ἆν ἐπισμυγερῶς ναυτίλλεται write ναυτίλεται, the aorist subjunctive. Three places remain to be mentioned:

Il. 1.66 αἴ κέν πως ἀρνῶν κνίσης αἰγῶν τε τελείων βούλεται ἀντιάσας ἡμῖν ἀπὸ λοιγὸν ἀμύνειν.

Curtius adopts the suggestion of Stier, βούλητʼ ἀντιάσας (Curt. Stud. ii. 138).

Il. 10.360 ὡς δʼ ὅτε καρχαρόδοντε δύω κύνε, εἰδότε θήρης,
ἢ κεμάδʼ ἠὲ λαγωὸν ἐπείγετον ἐμμενὲς αἰεὶ
χῶρον ἀνʼ ὑλήενθʼ, ὁ δέ τε προθέῃσι μεμηκώς.

Here ἐπείγετον is difficult because the subjunctive προθέῃσι is used in the next clause. Possibly the author of book 10 used the archaic form in -ῃσι as an indicative.

Il. 12.42 ὡς δʼ ὅτʼ ἂν ἔν τε κύνεσσι καὶ ἀνδράσι θηρευτῇσι κάπριος ἠὲ λέων στρέφεται.

The use of ὅτʼ ἄν in a simile is doubtful in Homer (see § 289). Should we read ὡς δʼ ὅτʼ ἔναντα? Cp. Il. 20.67.