Chapter VI · METRE AND RHYTHM IN GREEK TRAGEDY

§ I. Introduction

Poetry is illuminating utterance consisting of words the successive sounds of which are arranged according to a recurrent pattern. The soul of poetry is this illumination, its body this recurrent pattern of sounds; and it is with the body that we are now to deal. At the outset we must distinguish carefully between rhythm and metre. Rhythm is the recurrence just mentioned—the structure; metre is the gathering together of sounds into masses upon which rhythm shall do its work. Metre, so to put it, makes the bricks, while rhythm makes the arch.

Greek metre is based, not upon stress-accent,[858] but upon quantity—the length of time needed for the pronunciation of a syllable. In English the line

is “scanned” (that is to say, marked off into “feet”—the metrical units) as a series of five iambi; the iambus being a foot which consists of an unaccented, followed by an accented, syllable. The word “bosoms” can stand where it does because the stress of the voice naturally falls upon the first syllable of “bosom”; to begin a line with “my seréne bosom” would clearly be wrong. The length of the syllables has no effect on the scansion. That “sits” needs as long a time for its utterance as the first syllable of “lightly” does not alter the fact that “sits light-” is an accentual iambus.

Greek words, on the other hand, as metrical material, are considered only from the quantitative point of view, not the accentual. The voice-stress in the word λόγους rests upon the first syllable, but the word is an iambus, a “short” followed by a “long” (marked respectively thus ⏑–). Whereas an English blank verse consists of five accentual iambi, e.g.

To ént|ertaín | divíne | Zenócr|até,

the corresponding verse of all the Greek dramatists is composed of six feet each of which is theoretically a quantitative iambus, and most of which actually are such. Thus Andromache, v. 241 is to be scanned

⏑ – ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – τι δ ου | γυναιξ|ι ταυτ|α πρωτ|α παντ|αχου.

When is a syllable long and when short? A few rules will settle all but a minority. All syllables are long—

(i) Which contain a necessarily long vowel (η or ω), e.g. μη̄ν, τω̄ν.

(ii) Which contain a diphthong or iota subscript, e.g. ο̅ι̅νος, α̅ι̅νο̅υ̅μεν, ρᾳ̅διως; save that the first syllable of ποιῶ and τοιοῦτος (and their parts) is often short.

(iii) Which end with a double consonant (ζ, ξ, ψ), e.g. ο̄ζος, ε̄ξω, ε̄ψαυσα.

(iv) Which have the circumflex accent, e.g. υμῖ̅ν, μῦ̅ς.

Most syllables are long the vowel of which is followed by two consonants. But there is some difficulty about this very frequent case. It can arise in three ways:—

(a) Both consonants may be in the same word as the vowel. Then the syllable is long, save when the consonants are (i) a voiced stop (β, γ, δ) followed by ρ; or (ii) a voiceless stop or spirant (κ, π, τ; θ, φ, χ) followed by a liquid or nasal (λ, ρ, μ, ν)—in both of which cases the syllable can be counted long or short at pleasure. Thus ε̄σμεν, μο̄ρφη, ᾱνδρος; but the first syllables of ιδρις, τεκνον, ποτμος are “doubtful”—they can be either long or short as suits the poet.

(b) One of the consonants may end its word and the other begin the next. Such syllables are all long. Thus, τηκτο̄ς μολυβδος, ανδρε̄ς σοφοι, although both these long syllables are “short by nature” (see below).

(c) Both consonants may occur at the beginning of the second word. If the vowel is naturally short, the syllable is almost always short, though such scansions as σε̄ κτενω are occasionally found. But if the second word begins with a double consonant or σ followed by another consonant, the syllable is always long. Thus ο̄ ξενος, τῑ ζητεις, ταυτᾱ σκοπουμεν.

A vowel, naturally short, when thus lengthened is said to be “lengthened by position.”

The following types of syllable are always short:—

(i) Those containing a naturally short vowel (ε or ο) not lengthened by position, e.g. ε̆κων, ο̆λος.

(ii) Final α of the third declension neuter singular (σωμᾰ), third declension accusative singular (ελπιδᾰ, δρασαντᾰ), and all neuters plural (τᾰ, σωματᾰ, τοιαυτᾰ).

(iii) Final ι (e.g. εστῐ, τῐ), save, of course, when it is part of a diphthong.

(iv) The accusative -ας of the third declension (ανδρᾰς, πονουντᾰς). But μουσᾱς (first declension). The quantity in both cases is that of the corresponding nominative.

Hiatus is practically unknown. That is, a word ending in a vowel is not to be followed by a word beginning with a vowel, unless one vowel or the other disappears. Almost always it is the first vowel which is thus cut off, the process being called “elision.” In verse one would not write πάντα εἶπε, but πάντ’ εἶπε; not ἔτι εἶναι, but ἔτ’ εἶναι. When the first vowel is long and the second short, the latter is cut off by “prodelision,” a much rarer occurrence. Thus τούτῳ ἀνεῖπε would become τούτῳ ’νεῖπε. Two long vowels, as in καλὴ ἡμέρα, are not used together at all. But the rule as to hiatus does not normally apply at the end of a verse; usually one can end a verse with an unelided vowel and begin the next with a vowel. If in any metrical scheme this liberty is not allowed, it is said that “synapheia[859] prevails.”

We are now in a position to discuss the various metres to be found in Greek Tragedy.

§ II. The Iambic Metre

Practically all the dialogue and speeches are written in this metre. The student would do well to grow thoroughly accustomed to reading these aloud with correct quantities before he attempts the others.

The iambic line consists of six feet, any one of which may be an iambus. But a “pure” iambic line, one in which every foot is an iambus, as in Andromache, v. 241 (see above), is very rare. A speech written solely in such feet would be highly monotonous and far too rapid. Other feet are therefore allowed, under restrictions, to take the place of the iambus.

By far the commonest of these is the spondee, which consists of two long syllables (λο̄γχη̄, πᾱντω̄ν). This can occur in the first, third, or fifth places—one, two, or all three. Thus:—

– – ⏑– ⏑ – ⏑ – – – ⏑ – δησαι | βιᾳ | φαραγγ|ι προς | δυσχειμ|ερῳ (Prom. Vinctus, 15).

– – ⏑ – – – ⏑ – ⏑– ⏑ – ω τεκν|α Καδμ|ου του | παλαι | vεα | τροφη (Œd. Tyr., I).

Next, the lightness and variety is often greatly increased by the use of “resolved”[860] (or broken-up) feet. Each long syllable being regarded as equal to two “shorts,” it follows that the iambus can be “resolved” into ⏑⏑⏑, the spondee into –⏑⏑, ⏑⏑– (and ⏑⏑⏑⏑, but this last is not employed in iambics).

Of these three the tribrach (⏑⏑⏑) is much the most frequent. As it corresponds to the iambus, it can occur in any place, save the sixth; it is exceedingly rare in the fifth place:—

– – ⏑ ⏑ ⏑ – – ⏑ – – – ⏑ – φαιδρωπ|ον εδιδ|ου τοισ|ιν Αιγ|ισθου | φιλοις (Orestes, 894).

⏑ – ⏑ – ⏑ – ⏑ ⏑ ⏑ – – ⏑– περιξ | εγω | καλυψ|α βοτρυ|ωδει | χλοῃ (Bacchæ, 12).

The dactyl (–⏑⏑) is allowed in those places to which the spondee is admitted, save the fifth (just as the tribrach is excluded from the sixth). Thus:—

– – ⏑ – – ⏑ ⏑ ⏑ – – – ⏑ – ου φασ|ι πρωτ|ον Δανα|ον Αιγ|υπτῳ | δικας (Orestes, 872).

⏑ – ⏑ – – ⏑ ⏑ ⏑ – – – ⏑ – λογους | ελισσ|ων οτι | καθιστ|αιη | νομους (Ibid., 892).

It is rare in the first foot.

Least common of all is the anapæst (⏑⏑–), which appears only in the first foot, unless it is contained entirely in a proper name, when it can occur in any place save the sixth. This license is due to necessity: such a name as Ἀντῐγόνη could not otherwise be introduced into iambics at all. Examples:—

⏑ ⏑ – ⏑– ⏑ – ⏑ – – – ⏑ – στεφανους | δρυος | τε μιλ|ακος τ|ανθεσφ|ορου (Bacchæ, 703).

– – ⏑ – ⏑ – ⏑ – ⏑⏑ – ⏑ – δεσποιν|α γαρ | κατ οικ|ον Ερμ|ιονην | λεγω (Androm., 804).

Occasionally a line is to be found with two or even three resolved feet:—

– – ⏑ ⏑ ⏑ – ⏑ ⏑ ⏑ – ⏑ – ⏑ – λουτροισ|ιν αλοχ|ου περι|πεσων | πανυστ|ατοις (Orestes, 367).

– ⏑ ⏑ ⏑ – – ⏑ ⏑ ⏑ – ⏑ – ⏑ – μητερα | το σωφρ|ον τ ελαβ|εν αντ|ι συμφ|ορας (Ibid., 502).

⏑ ⏑ – ⏑ ⏑ ⏑ – ⏑ ⏑ ⏑ – ⏑ – ⏑ – αναδελφ|ος απατ|ωρ αφιλ|ος ει | δε σοι | δοκει (Ibid., 310).

Two licenses should be noted. The last syllable of the line may be short; no doubt the pause[861] at the end was felt to help it out. Lines of this kind are innumerable, e.g.:—

⏑⏑ Κρατος Βια τε σφῳν μεν εντολη | Διος (Prom. Vinctus, 12)

(which is followed by a vowel—ἔχει). It matters little whether such syllables are marked as short, as long, or with the sign of doubtful quantity ( ᷋). Next, synizesis (συνίζησις, “collapse”) occurs now and then—two syllables coalesce and are scanned as one, e.g. μ̅η̅ ̅ο̅υ̅, πολε̅ω̅ς:—

– – ⏑ – – – ⏑ – – – ⏑ – αλλ εα | με και | την εξ | εμου | δυσβουλ|ιαν (Antigone, 95).

– ⏑ – – ⏑ ⏑ ⏑ – ⏑ – ⏑ – ως μ̅η̅ ̅ε̅ι̅δ̅|οθ ητ|ις μ ετεκ|εν εξ | οτου τ|εφυν (Ion, 313).

– – ⏑ – – – ⏑ – ⏑ – ⏑ ⏑ σφαζ αιμ|ατου | θεας βωμ|ον η | μετεισ|ι σε (Andromache, 260).

(Synizesis is specially common in the various cases of θεός and θεά.)

Finally, two important rules of rhythm remain to be stated.

First, there must be a “cæsura”[862] in either the third or the fourth foot. A cæsura is a gap between words in the middle of a foot. Either the third foot, then, or the fourth must consist partly of one word, partly of another. It is indicated in scansion by the sign ‖. Many verses have this necessary cæsura in the third foot only, e.g.:—

⏑ – ⏑ – – – ⏑ – – – ⏑ ⏑ απανθ | ο μακρ|ος ‖ καν|αριθμ|ητος | χρονος (Ajax, 646).

Many show it in the fourth only:—

– – ⏑ – ⏑ – ⏑ – – – ⏑ ⏑ προς τησδ|ε της | γυναικ|ος ‖ οικτ|ειρω | δε νιν (Ibid., 652).

A still larger number have cæsura in both places:—

– – ⏑ – – – ⏑ – – – ⏑ ⏑ φρουρας | ετει|ας ‖ μηκ|ος ‖ ην | κοιμωμ|ενος (Agamemnon, 2).

This usage is essential to rhythm. It is of course possible for every foot in the line to exhibit a cæsura, but one in the midst is necessary to prevent the line from falling into pieces. That coextension of word and foot which is naturally frequent must at one point be emphatically excluded, so that the whole line may be felt as a single rhythmical whole. Such “lines” as

– – ⏑ – ⏑ – ⏑ – – – ⏑ ⏑ ταυτην αναξ λεγει καλην ειναι πολιν,

or

⏑ – ⏑– – – ⏑ – ⏑ – ⏑ ⏑ Οδυσσεως δουλοι μαχουμενοι ταχα,

are utterly impossible.[863] The first falls into six scraps, and the second into two mere lumps, of equal length. If a breach of the rule ever occurs, it is for a special reason. When Sophocles (Œd. Tyr., 738) writes

the dragging rhythm well represents the dawning dread of Œdipus. But the main cæsura may be dispensed with if the third foot ends with an elision, apparently because, if the word could be written in full, the fourth foot would be divided between two words. Thus:—

The other rule is that generally called “the rule of the Final Cretic”.[864] It is most simply stated thus: if there is a cæsura in the fifth foot, that foot must be an iambus, e.g.:—

⏑ – ⏑ – μη με̄ στυγησῃς· ουχ εκων | γαρ ‖ αγγ|ελω (Troades, 710).

⏑ – ⏑ – τον τουδε νεκρον ουκ αθαπτ|ον ‖ αν | λιποις (Ibid., 738).

This rule does not exclude from the first half of the foot long monosyllables which are in meaning and syntax closely connected with the “cretic” word or words. Thus τῶν σωμάτων is a quite correct ending, but not τούτων σωμάτων.

Subjoined is a scheme of the iambic verse as written by the tragedians. The writers of comedy allowed themselves licenses with which we are not here concerned. Euripides is much fonder of resolved feet than Æschylus or Sophocles.

1 2 3 4 5 6 ⏑– ⏑– ⏑‖– ⏑‖– ⏑– ⏑– ⏑⏑⏑ ⏑⏑⏑ ⏑‖⏑⏑ ⏑‖⏑⏑ ⏑⏑⏑ ⏑⏑ – – [⏑⏑–] –‖– [⏑⏑–] – – –⏑⏑ –‖⏑⏑ [⏑⏑–] ⏑⏑– [⏑⏑–]

§ III. The Trochaic Tetrameter

Under this head we shall deal only with trochees as used in dialogue. Originally all dialogue was written in this metre,[865] and they sometimes appear in extant plays when the situation is too hurried or excited for iambics though not agitated enough for lyrical dialogue. These passages are not usually long, and it is interesting to note that the longest are found in the two most melodramatic plays, Orestes and Iphigenia at Aulis.[866] The metre is always the trochaic tetrameter catalectic[867] (sometimes called the trochaic octonarius), that is, a line consisting of eight feet, mostly trochees, with “catalexis”. Catalexis occurs when the last foot of a line has not its full number of syllables, the remainder being filled by a pause in delivery.

Pure trochaic verses are occasionally to be found:—

– ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – κατα | πως αφ|ικομ|εσθα | δευρο | ταυτ αμ|ηχαν|ωꞈ| (Ion, 548). 1 2 3 4 5 6 7 8

The mark ꞈ means that there is a pause equivalent in length to a short syllable. It is often found in the scansion of lyrics, and there one also at times uses -̭ ⏗̭ ⏘̭, which mean pauses equivalent to two, three, and four short syllables respectively. As in iambics, the last syllable may be short by nature:—

– ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – ⏑ ⏑– ουχι | σωφρον|ειν γ επ|εμψε | δευρο σ | η Δι|ος δαμ|αρꞈ (Heracles, 857).

This metre is plainly analogous to Tennyson’s

But such purely trochaic lines are rare. Other feet are usually admitted, especially the spondee:—

– ⏑ – – – ⏑ – – – ⏑ – – – ⏑ – βλεψον | εις ημ|ας ιν | αρχας | των λογ|ων ταυτ|ας λαβ|ωꞈ (Iph. Aul., 320).

Spondees may occur only in the second, fourth, or sixth foot.

The tribrach also is often employed by Euripides:—

– ⏑ – – – ⏑ – – ⏑ ⏑ ⏑ – – – ⏑ – ψηφον | αμφ ημ|ων πολ|ιτας | επι φον|ῳ θεσθ|αι χρε|ωνꞈ (Orestes, 756).

– ⏑ – – – ⏑ ⏑ ⏑ ⏑ – ⏑ – ⏑ – ⏑ – ευτυχ|εις δ ημ|εις εσ|ομεθα | ταλλα δ | ου λεγ|ουσ ομ|ωςꞈ (Iph. Taur. 1232).

The fifth foot is the favourite place for the tribrach, and next to that the first:—

⏑ ⏑⏑ – ⏑ – ⏑ – ⏑ ⏑ ⏑ ⏑ – – – ⏑ – χρονιος | αλλ ομ|ως ταχ|ιστα | κακος εφ|ωραθ|η φιλ|οιςꞈ (Orestes, 740).

Euripides, late in his career, introduced a good deal of license, here as elsewhere. Firstly, tribrachs become far more frequent and occur in unusual places:—

⏑ ⏑ ⏑ – ⏑ – ⏑ – – ⏑ ⏑ ⏑ – ⏑ ⏑ ⏑ ⏑ ᷋ ανοσι|ος πεφ|υκας | αλλ ου | πατριδος | ως συ | πολεμι|οςꞈ (Phœnissæ, 609).

To place a resolved foot practically at the end of the line is bold—the metre is shaken almost to pieces. Here, as in other respects, Euripides points forward to the conversational manner of the New Comedy. But he goes further, and allows feet hitherto not found in trochaics: the anapæst and the dactyl. The latter, however, is extremely rare[868] and employed only with proper names:—

– ⏑ – ⏑ – ⏑ ⏑ – ⏑ – ⏑ – – – ⏑ – συγγον|ον τ εμ|ην Πυλαδ|ην τε | τον ταδ|ε ξυνδρ|ωντα | μοιꞈ (Orestes, 1535).

– ⏑ – ⏑ ⏑ – ⏑ ⏑ ⏑ – – ⏑ – ⏑ – ⏑ ᷋ εις αρ | Ιφιγεν|ειαν | Ελενης | νοστος | ην πεπρ|ωμεν|οςꞈ (Iph. Aul., 882).

The anapæst is commoner (there is a proper-name instance in the line just quoted):—

– ⏑ ⏑ ⏑ – – ⏑ – – ⏑ ⏑ ⏑ – ⏑ – ⏑ – ως νιν | ικετευσ|ω με | σωσαι | το γε δικ|αιον | ωδ εχ|ειꞈ (Orestes, 797).

– ⏑ – ⏑ ⏑ ⏑ ⏑ – – – ⏑ ⏑ ⏑ – – ⏑ ᷋ και συ | μητερ | αθεμιτ|ον σοι | μητρος | ονομαζ|ειν καρ|αꞈ (Phœn., 612).

There is no rule as to cæsura. The end of the fourth foot regularly coincides with the end of a word; such an arrangement is named diæresis.[869] In all extant tragedy only one certain exception to this rule is found:—

– ⏑ – – – ⏑ – – – ⏑ – – – ⏑ ᷋ ει δοκ|ει στειχ|ωμεν | ω γενν|αιον | ειρηκ|ως επ|οςꞈ (Philoctetes, 1402).

Since diæresis is practically always found in so many hundreds of lines, being preserved even in the loosest writing of Euripides, why should we regard the recognized trochaic verse as an unity? Why not write, e.g.:—

If the line falls into two clearly marked halves, why not show this to the eye? There is no unanswerable objection to doing so—the passage above corresponds exactly in rhythmical form to much English verse, e.g.:—

The practice in English is to break up the long trochaic “line” into two when the words at the diæresis rhyme (as in the above passage from Longfellow), but not to do so when the only rhymes occur at the catalectic foot. We print the opening of another poem by Longfellow thus:—

In Greek there is, of course, no rhyme-scheme to settle this, but the regular catalexis is felt to mark off separate units. The entire question depends upon personal fancy,[870] though the instance from the Philoctetes shows that Sophocles at any rate regarded the whole octonarius as the unit.

Subjoined is the scheme:—

1 2 3 4 5 6 7 8 –⏑ –⏑ –⏑ –⏑ –⏑ –⏑ –⏑ –ꞈ ⏑⏑⏑ – – ⏑⏑⏑ – – ⏑⏑⏑ – – ⏑⏑⏑ ⏑⏑⏑ [–⏑⏑] ⏑⏑⏑ ⏑⏑⏑ ⏑⏑– ⏑⏑– ⏑⏑– [–⏑⏑]

§ IV. The Anapæstic Metre

Whereas iambics and trochaics were declaimed by the actors, anapæsts were used mostly by the chorus, and were chanted in recitative. They are found when the chorus move into the orchestra, or salute the entrance of a new character. Most tragedies end with a brief anapæstic system, executed by the singers as they depart.

The most usual line is a tetrapody—that is, a verse of four feet:—

⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ – τι συ προς | μελαθροις | τι συ τῃδ|ε πολεις; (Alcestis, 29).

But lines consisting of anapæsts alone are very uncommon. The spondee is often found:—

– – ⏑ ⏑ – ⏑ ⏑ – – – ουκ ηρκ|εσε σοι | μορον Αδμ|ητου (Alcestis, 32).

Dactyls also are frequent:—

– – ⏑ ⏑ – – ⏑ ⏑ – – σφηλαντ|ι τεχνῃ | νυν δ επι | τῃδ αυ (Ibid., 34).

No other foot is admitted, but each of these three may occur at any place in the line.

Besides the tetrapody, we find now and then a dipody, or verse of two feet.

Anapæstic systems are invariably closed by a catalectic verse:—

– – ⏑ ⏑ – ⏑ ⏑– – αυτη | προθανειν | Πελιου | παιςꞈ̄ (Ibid., 37).

In systems of considerable length such lines occur at intervals. They are called “parœmiacs”.[871]

§ V. Lyrics

The metres of Greek songs form a difficult and complicated study. So long as we do not know the music composed for them, the scansion of lyrics must remain a more difficult and doubtful question than that of the iambics, episodic trochaics, and anapæsts.

The best preparation for their study is the habit of reading iambics and trochaics with correct quantities and natural emphasis. Let us, so prepared, address ourselves to the following passage[872] from the Agamemnon (975 sqq.):—

It soon becomes plain that the passage is, at any rate in the main, trochaic. The first two lines scan easily, ending with a catalectic foot. We note that the third seems to drag at the end:—

– ⏑ – ⏑ – ⏑ – ⏑ – – καρδι|ας τερ|ασκοπ|ου ποτ|αται,

for we remember that in the trochaic octonarius the last complete foot is never a spondee. But in the fourth line we are quite baffled:—

– ⏑ ⏑ – ⏑ ⏑ – μαντιπ|ολει δ | ακελευστ|ος...?

Anapæsts are very rare in trochaics, iambi unknown. That the iambus should never replace the trochee is quite natural. It would be hideous rhythm, in the first line of Locksley Hall, instead of “Cómrades, leave me here a little ...,” to write “Dragoóns, leave me ...”. The foot ο̆λεῑ cannot be right. The line seems hopeless; or rather, if we have any knowledge of Homeric and Virgilian metre, we recognize something like the dactylic hexameter:—

– ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ – – μαντιπολ|ει δ ακελ|ευστος αμ|ισθος α|οιδα.

But is such a passage possible in a trochaic passage written for Greek music? It is known that in Greek music the notes corresponded closely to the syllables; music composed for trochees will certainly be in three-eighths time, for dactyls in four-eighths time. All these feet should have three beats, not four.

The next two lines are plainly similar to the first and second. In the seventh line we first wonder why, though we are in the midst of a grammatical sentence, the words should begin farther to the right than is usual, as if for a new paragraph. When we try to scan, we find once more the iambus-difficulty:—

– ⏑ – – ⏑ – θαρσος | ευπειθ|ες ιζ....

If we work backwards from the end, -ος φιλον θρονον gives the familiar trochaic-octonarius ending, –⏑|–⏑|–ꞈ. But the middle of the line has fallen to pieces, and for the present we leave it.

The eighth line seems at first more familiar. Is it not the ordinary iambic senarius of § II? But where is the cæsura? And can we suddenly insert an iambic line into a trochaic system? Is it then possible after all to scan it as some kind of trochaics? Begin at the end. ... ε̅μ̅βο̆λα̅ι̅ςꞈ suits excellently; and if we work backwards we soon find that the whole would fall readily into trochaics if only we could ignore the first syllable:—

– ⏑ – – – ⏑ – ⏑ – ⏑ – χρονος δε | τοι προμν|ησι|ων ξυν|εμβο|λαιςꞈ

But why should we ignore it? And why does the line begin farther to the left?

The ninth line again offers perplexity in the first half, clearness in the second:—

– ⏑ – ⏑ – ⏑ ᷋ ηκεν | ευθ υπ | Ιλι|ονꞈ.

Grown by this time bolder, we attack the first half in detail, working backwards. ᾱς βε̆ is easy. Then ἐξ ακτ ... may be either –⏑ or – –, both of which are admissible. We are left with ψ̅α̅μμῐᾱς. Reading the whole line over slowly, marking the trochees carefully, we find ourselves somehow dwelling on the last syllable of ψαμμίας. Why should we? If that syllable were only –⏑, all would be well; but it is not. Finally, the tenth and last line is quite easy:—

– ⏑ – ⏑ – ⏑ ⏑– ωρτο | ναυβατ|ας στρατ|οςꞈ.

The whole passage then is trochaic; but we have met four difficulties: (i) the necessity to dwell upon certain syllables, (ii) the irrational presence of dactyls, (iii) the temptation to ignore the first syllable of χρόνος, (iv) the insetting of θάρσος. Understanding of these four facts will carry us a long way. We take them in order.

Our first point indicates that we must revise that division of all syllables into “longs” of equal value and “shorts” of equal value (each “long” being exactly equivalent to two “shorts”) which obtains in iambics. The lyric metres recognize syllables of greater length than ⏑⏑. Most frequent is the length ⏗, equal to ⏑⏑⏑. A syllable of this length is therefore admitted in lyric trochaic systems as a whole foot, and investigations, such as we have practised above, will generally show where such a foot is to be postulated. We can now scan certain portions which we found troublesome:—

– ⏑ ⏗ – ⏑ ⏗ θαρσος | ευπ | ειθες | ιζ | ...

– ⏑ ⏗ – ⏑ – ψαμμι|ας | εξ ακτ | ας....

Moreover, as we were suspicious of the final spondee (replacing the expected trochee) in the third line, we obtain at any rate a quasi-trochee by scanning thus:—

– ⏑ – ⏑ – ⏑ – ⏑ ⏗ – καρδι|ας τερ|ασκοπ|ου ποτ|ατ|αιꞈ.

This prolongation of a syllable is called τονή (“stretching”). Such a syllable may fill a foot, as in trochaics, and this rhythm is said to be syncopated.[873]

Next comes the dactylic fourth line, which introduces another vital rule. Trochaic systems admit, not genuine dactyls, but “cyclic” dactyls. To the “long” of each foot and to the first “short” is given less than their usual length: the rhythm is accelerated, so that –⏑ is equivalent to –, and the whole cyclic dactyl, marked ᷋⏑, is equivalent to a trochee.[874] Whenever we see a number of apparent dactyls, we must examine the whole passage to find whether it is trochaic or not. Trochaic systems which contain cyclic dactyls are called “logaoedic”.[875] The present line, then, being trochaic, we feel the same doubt of the final spondee (which would equal ⏑⏑⏑⏑, not ⏑⏑⏑, as it should) which we felt in the third line, and scan the whole:—

– ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ ⏗ – μαντιπολ|ει δ ακελ|ευστος αμ|ισθος α|οιδ|αꞈ.

Our third question touched the first syllable of χρόνος in the sixth line. It is, as a fact, to be regarded as standing outside the metrical line—a kind of prelude, called “anacrusis”.[876] It is plain that neglect of anacrusis will often throw our scansion out completely. A useful rule can be given: in almost[877] any line, whatever comes before the first long syllable forms an anacrusis. The reason is that the first syllable of a foot must have an “ictus” (see below) or stress-accent, and the foot-ictus normally falls on long syllables. It becomes natural then to pronounce the first short or shorts (if any) quickly, and to give the first long the ictus; in this way the short is felt as a mere preliminary to the line. The anacrusis, however, can be of three forms, ⏑, ⏑⏑, –. Its length must be that of the second part of the characteristic foot, ⏑ for trochees, – or ⏑⏑ for dactyls, and so forth. It is marked off from the first foot by the sign ⁝.

The fourth point was the insetting of θάρσος. It happens in the middle of a grammatical sentence, so that there can be no question of an ordinary paragraph. But if it does not point to a break in sense, its only reference can be rhythmical. The whole passage must fall into two distinct rhythmical paragraphs. Let us scan them separately and endeavour to find a reason for this break. Take the first, scanning, marking, and numbering the feet:—

– ⏑ – ⏑ – ⏑ – τιπτε | μοι τοδ | εμπεδ|ονꞈ‖ 1 2 3 4 – ⏑ – ⏑ – ⏑ – δειμα | προστατ|ηρι|ονꞈ‖ 1 2 3 4 – ⏑ – ⏑ – ⏑ – ⏑ ⏗ – καρδι|ας τερ|ασκοπ|ου ποτ|ατ|αιꞈ‖ 1 2 3 4 5 6 – ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ – ⏑ ⏑ ⏗ – μαντιπολ|ει δ ακελ|ευστος αμ|ισθος α|οιδ|αꞈ‖ 1 2 3 4 5 6 – ⏑ – ⏑ – ⏑ – ουδ απ|οπτυσ|αι δικ|ανꞈ‖ 1 2 3 4 – ⏑ – ⏑ – ⏑ – δυσκριτ|ων ον|ειρατ|ωνꞈ‖ 1 2 3 4

If we examine this to find structural unity, it soon appears. The first pair of lines answers to the last, and line three to line four, in the number of their feet 4 + 4, 6, 6, 4 + 4. The correspondence is indicated thus:—

⎰ 4 ⎛ ⎱ 4 ⎜ ⎛ 6 ⎜ ⎝ 6 ⎝ ⎰ 4 ⎱ 4

Each of these masses, it will be noticed, is marked off by the sign ‖. Such a mass is named a “sentence” or “colon” (κῶλον, “limb”), and such a balanced structure of cola is named a “period” (περίοδος, “circuit”). It happens that in the passage just examined the “sentence” division always occurs at the end of a word, but this is not invariably so. We proceed now with the second paragraph[878]—the second period as we shall now call it.

– ⏑ ⏗ – ⏑ ⏗ – ⏑ – ⏑ – ⏑ – θαρσος | ευπ|ειθες | ιζ‖ει φρεν|ος φιλ|ον φρον|ονꞈ‖ 1 2 3 4 1 2 3 4 ⏑ – ⏑ – ˃ – ⏑ – ⏑ – ⏑ – χρον⁝ος δε | τοι πρυμν|ησι|ων ξυν | εμβολ|αιςꞈ‖ 1 2 3 4 5 6 – ⏑ ⏗ – ⏑ – ⏑ – ⏑ – ⏑ – ⏑ – ψαμμι|ας | εξ ακτ|ας βεβ‖ηκεν | ευθ υπ | Ιλι|ονꞈ‖ 1 2 3 4 1 2 3 4 – ⏑ – ⏑ – ⏑ – ωρτο | ναυβατ|ας στρατ|οςꞈ‖ 1 2 3 4

That is: 4 + 4, 6, 4 + 4, 4. This would be an obviously well-balanced structure but for the last colon, to which nothing corresponds. Such an extra sentence is called a “postlude” (ἐπῳδικόν). Non-corresponding sentences like this are far from rare.[879] They may occur at the beginning of the period (“prelude,” προῳδικόν), in the middle (“mesode,” μεσωδικόν), or at the end. This very period supplies an example of a mesode as well as of a postlude. The scheme is:—

⎰ 4 ⎛ ⎱ 4 ⎜ 6—mesode. ⎝ ⎰ 4 ⎱ 4 4—postlude.

The whole passage, then, consists of two periods connected by meaning and grammar, but—for us—by no more intimate musical bond than the common use of trochees. But the dance and music which accompanied the whole would clearly demonstrate its unity. The end of a period is indicated by ‖.

It is necessary now to consider briefly the passage which immediately follows (vv. 988 sqq.):—

This is an exact counterpart in syllables, feet, cola, and periods, of the first passage. The first is called the “strophe” (στροφή, “turn”), the second the “antistrophe” (ἀντιστροφή, “counter-turn”). The chorus, while singing the one, performed various evolutions about the orchestra, and these were repeated exactly, but in reversed order, while they sang the antistrophe. All these lyrics are so constructed; the normal tragic “chorus” consists of one or more such pairs, though occasionally the antistrophe is followed by a passage called an “epode”.[881] The epodes correspond to each other, not to the strophes. This equivalence of strophe and antistrophe is often of value in determining the quantities or the text in one of them.

We have now gained some insight into the nature of a Greek choric song. But before proceeding further it will be well to deepen our impression by taking from the Agamemnon (vv. 160 sqq.) another, and a simpler, pair of strophes:—

⏗|⏗|–⏑|–⏑|–⏑|⏗‖–⏑|–⏑|–⏑|–ꞈ‖ –⏑|–⏑|–⏑|–ꞈ‖ –⏑|–⏑|–⏑|⏗‖–⏑|–⏑|–⏑|–ꞈ‖ –⏑⏑|–⏑⏑|–⏑⏑|–⏑⏑|⏗|–ꞈ‖ –⏑|–⏑|–⏑|–ꞈ‖

⎰ 6 ⎛ ⎱ 4 ⎜ ⎛ 4 ⎜ ⎜ 4—mesode. ⎜ ⎝ 4 ⎝ ⎰ 6 ⎱ 4

The chief interest of this subject is the art wherewith the Greek masters accompanied variations of emotion and the like with variations of rhythm. This passage affords a simple and stately example. The heavy opening (⏗⏗) is followed by the more confident trochees till, at the last line but one, religious rapture (in the strophe) and the ardour of triumph (in the antistrophe) burst forth with the leaping cyclic dactyls.

We have now become acquainted with three rhythmical masses: the colon, the period, the strophe. Are there others? What is a “verse” in lyrics? There is no such thing.[882] One must, of course, distinguish between a “line” and a “verse”. Lines there must be—that is an affair of the scribe and the printer; verses are rhythmical units, and there is no rhythmical mass in Greek lyrics between the colon and the period. How then are we to arrange our periods, there being no verse-division? The most obvious way is to write each colon as a separate line. The difficulty is that we shall often be compelled to break words:—

Another method is to let each line run on until we reach a colon-ending which coincides with a word-ending. Here is no new rhythmical rule: it is purely a question of convenience for the eye. Next, shall we ever write lines of (say) two cola the first of which does close with a word-ending? It is natural so to do when to the two cola in question there correspond (whether periodically or strophically) two cola which must on this system fill one line only. For instance, in Æsch., Supplices, 656, we shall write—

– ⏑ ⏑ – ⏑ ⏗ –⏑ ⏑ – ⏑ ⏗ – ⏑ ⏑ – ⏑ – ⏑– και γαρ υποσκιων ‖ νυν στοματων ποτασθ‖ω φιλοτιμος ευχα ‖,

though the first colon ends at the end of ὑποσκίων, because the corresponding passage of the antistrophe runs—

– ⏑ ⏑ – ⏑ ⏗ – ⏑ ⏑ – ⏑ ⏗ –⏑ ⏑ – ⏑ – ᷋ και γεραροι δε πρεσβ‖υτοδοκοι γεμοντ‖ων θυμελαι φλεγοντων ‖,

where the first colon ends inside a word. It is purely a matter of taste whether we give a line to each colon, in which case the drawback is the breaking of words, or continue our line till breaking of words is excluded, the trouble about which method is the reader’s difficulty in seeing where some of the cola begin.

We must now consider the most vital and difficult portion of our subject. How are we to determine the cola? The colon is the very soul of the rhythm. The period is generally too long for the ear to receive it as one artistic impression. The foot is too short; moreover, the mere foot too often tends to play one false: irrational syllables and τονή are against us. But the colon is neither too long nor too short. The colon-division serves the same purpose as non-commissioned officers in a regiment, or the determination of watersheds in geography—it gives a sense both of grouping and of control.

What precisely is a colon? It is as much of a strophe as can be uttered without making a new start. It is the embodiment of rhythm, as the foot is the embodiment of metre. In other words, it is a series of feet bound into a rhythmical unity by the presence of one main ictus. Three questions, then, arise. (i) What is an ictus? (ii) Which is the main ictus of a series? (iii) Can we with certainty determine the beginning and end of a colon when we have identified the main ictus?

(i) Ictus is stress-accent. The ictus of any single word is usually obvious. In the word “maritime” it falls upon the first syllable, in “dragoon” upon the second, in “cultivation” upon the third. In πάντων, λυσαμένοις, and κατάπαστος, it falls upon the first, second, and third respectively. Greek metre is based upon quantity, but Greek rhythm (like all other rhythm) is based upon ictus. A strophe can, and must, be scanned foot by foot on quantity alone; but when we go beyond the foot-division to exhibit the structure of the whole, we must refer to ictus and nothing but ictus—for structure is an affair of cola, and the colon is created by the main ictus.

(ii) Among the many word-ictuses of a considerable passage, a few will be found which are heavier than the rest. These are simply the ictuses of the most important words. Each of these prominent ictuses gathers the neighbouring minor ictuses into a group round itself. We should begin then by fixing some obvious example, one (that is) where the main ictus is unmistakable, and on this basis attempt, by the help of the correspondences which we expect, to determine other main ictuses. The strophe will thus gradually fall into cola. This leads us at once to our third question.

(iii) Can we with certainty determine the extent of each colon? Unfortunately no simple invariable rule can be given for the settlement of this vital point. But certain useful principles may be mentioned.

(a) A well-trained ear is the chief guide. Intelligent and careful reading aloud of an English prose-passage will show this. Take first (the best-known version of) a famous sentence of John Bright:—

The Angel of Death is abroad in the land: you may almost hear the beating of his wings.

It is plain that this falls into two rhythmical parts, though we shall not expect them to correspond, since this is prose, not verse. If we set a dash for each syllable and mark the ictuses by one or more dots according to their strength, we find this scheme:—

⁚ ⁝ ⁚ · - - - - - - - - - - - ‖ · ⁚ ⁝ ⁚ - - - - - - - - - - - ‖

(It will be noticed that in this superb passage the two periods do, as it happens, correspond in length.)

Who hath believed our report? And to whom is the arm of the Lord revealed? (Isaiah liii. 1).

· ⁚ ⁝ · ⁚ ⁝ ⁚ - - - - - - - ‖ - - - - - - - - - - - ‖

So with longer passages, where, however, we shall find at times that our voice quite naturally makes a colon-ending in the midst of a grammatical sentence.

Therefore let us also, ‖ seeing we are compassed about ‖ with so great a cloud of witnesses, ‖ lay aside every weight, ‖ and the sin which doth so easily beset us, ‖ and let us run with patience ‖ the race that is set before us ‖ (Hebrews xii. 1, R.V.).

· ⁚ · - - - - - - · ⁚ · - - - - - - - - · ⁚ ⁝ - - - - - - - - - · ⁚ ⁝ - - - - - - - ⁚ ⁝ ⁚ - - - - - - - - - - - - · ⁚ ⁝ - - - - - - - ⁝ ⁚ · - - - - - - - -

(Observe how, in the last two cola, first the mounting and then the declining emphasis provide a splendid close.)

Let us now attempt so to catch the rhythm of a passage from Sophocles (Antigone, 582 sqq.) if set out as prose.

εὐδαίμονες οἷσι κακῶν ἄγευστος αἰών. οἷς γὰρ ἂν σεισθῇ θεόθεν δόμος, ἄτας οὐδὲν ἐλλείπει, γενεᾶς ἐπὶ πλῆθος ἕρπον· ὅμοιον ὥστε ποντίαις οἶδμα δυσπνόοις ὅταν θρῄσσαισιν ἔρεβος ὕφαλον ἐπιδράμῃ πνοαῖς, κυλίνδει βυσσόθεν κελαινὰν θῖνα, καὶ δυσάνεμοι στόνῳ βρέμουσιν ἀντιπλῆγες ἀκταί. ἀρχαῖα τὰ Λαβδακιδᾶν οἴκων ὁρῶμαι πήματα φθιτῶν ἐπὶ πήμασι πίπτοντ’, οὐδ’ ἀπαλλάσσει γενεὰν γένος, ἀλλ’ ἐρείπει θεῶν τις, οὐδ’ ἔχει λύσιν. νῦν γὰρ ἐσχάτας ὑπὲρ ῥίζας ὃ τέτατο φάος ἐν Οἰδίπου δόμοις, κατ’ αὖ νιν φοινία θεῶν τῶν νερτέρων ἀμᾷ κονίς, λόγου τ’ ἄνοια καὶ φρενῶν Ἐρινύς.

If we first mark the quantities (ignoring, as we must at first, the possibility of ⏗ and ⏘) and go over the whole carefully, we soon find that it falls into two corresponding portions: εὐδαίμονες ... ἀκταί is the strophe, ἀρχαῖα ... Ἐρινύς the antistrophe. Next we look for rhythmical units. On the one hand, there is the great difficulty that, since we must have both periodic and strophic equivalence, certain cola may take in words not belonging to the same sense-groups or grammatical clauses. On the other hand, the fact that we have two great masses which correspond exactly will help us. First, then, we note that εὐδαίμονες ... αἰών looks promising, and observing that this points to ἀρχαῖα ... ὁρῶμαι as a colon also, and that this is in itself likely, we mark off both these groups. Conversely, at the end of the antistrophe, λόγου ... Ἐρινύς attracts us, and this is supported by the naturalness of στόνῳ ... ἀκταί at the end of the strophe. Working backwards, and seeing a pause in the punctuation at precisely the same place in both halves, namely, after πνοαῖς and δόμοις, we assume that κυλίνδει ... δυσάνεμοι and κατ’ αὖ ... κονίς are correspondent masses. But each is too long—sixteen syllables—to be pronounced as a unit. We soon perceive that κυλίνδει ... κελαινάν, θῖνα ... δυσάνεμοι, κατ’ αὖ ... τῶν, νερτέρων ... κονίς, are all separate cola. Going backwards again, we find that ἐπιδράμῃ[883] πνοαῖς and Οἰδίπου δόμοις, ὕφαλον ... πνοαῖς and φάος ... δόμοις, Θρῄσσαισιν ... πνοαῖς and ῥίζας ... δόμοις, and indeed longer masses still, all give a metrical correspondence. Which pair are we to select? οἶδμα ... πνοαῖς (= νῦν ... δόμοις) is too long; ἐπιδράμῃ πνοαῖς (= Οἰδίπου δόμοις) is too short. For we seek the longest unit which is convenient. We therefore mark off οἶδμα ... ὅταν, νῦν ... ὑπέρ, Θρῄσσαισιν ... πνοαῖς, ῥίζας ... δόμοις as cola. The same method will give us ὅμοιον ... ποντίαις and θεῶν ... λύσιν. Then we find ourselves left with οἷς γάρ ... ἕρπον and πήματα ... ἐρείπει, which we divide after ἄτας and πίπτοντ’.

At last we can set out the passage according to its structure. The strophe runs thus:—

˃ – ⏑ ⏑ – ⏑ ⏑ – ⏑ – ⏑ ⏗ – ευ⁝δαιμονες | οισι κακ|ων α|γευστος | αι|ωνꞈ‖ – ⏑ – ˃ – ⏑ – ⏑ ⏑ ⏗ – οις γαρ | αν σεισθ|ῃ θεο|θεν δομος | ατ|αςꞈ‖ – ⏑ – ˃ – ⏑ ⏑ – ⏑ ⏑ – ⏑ – ⏑ ουδεν | ελλειπ|ει γενε|ας επι | πληθος | ερπον ‖ ⏑ – ⏑ – ⏑ – ⏑ – ο⁝μοιον | ωστε | ποντι|αιςꞈ‖ – ⏑ – ⏑ – ⏑ – οιδμα | δυσπνο|οις οτ|ανꞈ‖ ˃ – ⏑ ⏑ ⏑ ⏑ ⏑ ⏑ ⏑ ⏑ ⏑ ⏑ – ⏑ – Θρησσ⁝αισιν | ερεβος | υφαλον | επιδραμ|ῃ πνο|αιςꞈ‖ ⏑ – ˃ – ⏑ – ⏑ – ˃ κυ⁝λινδει | βυσσο|θεν κελ|αιναν ‖ – ⏑ – ⏑ – ⏑ – θινα | και δυσ|ανεμ|οιꞈ‖ ⏑ – ⏑ – ⏑ – ⏑ – ⏑ ⏗ – στον⁝ῳ βρεμ|ουσιν | αντιπλ|ηγες | ακτ|αιꞈ‖

There are two periods:—

I II ⎛ 6 ⎛ 6 ⎝ 6 ⎜ ⎰ 4 ⎜ ⎛ ⎱ 4 ⎜ ⎜ 6—mesode. ⎜ ⎜ ⎰ 4 ⎜ ⎝ ⎱ 4 ⎝ 6

To this the antistrophe of course corresponds, though here and there an irrational long corresponds to a short (e.g. -εῑπε͐ι to ε̄ρπο̆ν); the last syllable of πήματα is lengthened by the following φθ.

It should be noted that this scheme differs somewhat from that given in Jebb’s edition of the Antigone (pp. lxi. sq.). One reader’s ear differs from that of another: hence the frequent divergencies to be observed between editors in the arrangement of many lyrics.

(b) The ancient writer Aristoxenus gives certain rules as to the maximum length of cola. They may be stated as follows:—

(i) There are three types of colon, the equal, the unequal, and the quinquepartite. The equal cola are the dipody of 1 + 1 feet, the tetrapody of 2 + 2; the unequal are the tripody of 2 + 1, and the hexapody of 4 + 2; the quinquepartite is the pentapody of 3 + 2.

(ii) Equal cola must not be of greater length than sixteen “shorts”. Therefore we may have a dipody of any foot, and a tetrapody of any save those of more than four shorts in value; that is (e.g.) a dipody of cretics (–⏑–) is allowed, but not a tetrapody of that foot, which would give 5 × 4 = 20 “shorts”.

(iii) Unequal cola may have the length of eighteen “shorts”. A tripody, therefore, of any foot is allowed, but a hexapody of trochees only: a hexapody of spondees would give 4 × 6 = 24 “shorts”.

(iv) Quinquepartite cola may extend to the value of twenty-five “shorts”. Pentapodies are therefore possible of trochees, dactyls, spondees and five-time feet.

(c) Certain detailed hints may be added:—

(i) The tetrapody is the most frequent length, the pentapody the rarest.

(ii) The end of a colon is often indicated in dactyls by a spondee, in trochees by a single long syllable (whether ⏗ or –ꞈ).

(iii) In any one period there is a tendency to conformity in length. If 6 + 5 + 4 and 6 + 6 + 4 are prima facie equally possible, the latter is as a rule to be preferred. In spite of the difference in sum-total (6 + 6 + 4 = 16; 6 + 5 + 4 = 15), this question often arises, because of the possibility of τονή. It has to be decided[884] whether (e.g.) παντός at the close of a colon is to be scanned as two feet or one:

⏗ – παντ|οςꞈ‖

or

– ⏑ | παντος ‖.

It is now time to offer an account of the various feet used in lyrics.

(a) Trochees.—With these we are now familiar. This foot is often called a choree, chorees with anacrusis being iambi,[885] without anacrusis trochees. The trochee is the most frequent foot in lyrics. Such systems express ordinary strong interest. Whenever more definite emotion is to be conveyed, either cyclic dactyls are introduced, or a change is made to some other metre:—

–⏑|–⏑|–⏑|–⏑‖ –⏑|–⏑|–⏑|–⏑‖.

So in English:—

Resolution into tribrachs is frequent:—

⏑ ⏑ ⏑ – ⏑ – ⏑ – ⏑ Αραβι|ας τ αρ|ειον | ανθος ‖ (Prom. Vinctus, 420).

Anacrusis is common.

(b) Dactyls.—These are found pure, or mingled with spondees or quasi-trochees (⏗⏑). They are often employed to express excitement and awe:—

–⏑⏑|–⏑⏑|–⏑⏑‖–⏑⏑|–⏑⏑|– –‖ –⁝⏗⏑|⏗⏑|⏗⏑|–ꞈ̄‖

Anacrusis is found, as in the second line above and in Medea, 635:—

– – ⏑ ⏑ – ⏑ ⏑ – – ⏗ – – – ⏗ ⏑ – στεργ⁝οι δε με|σωφροσυν|α δωρ‖ημα | καλλιστ|ον θε|ωνꞈ̄‖.

The tetrapody without spondees or catalexis gives an exquisite heaving effect in Soph. Electra, 147-9:—

Ariel’s lines in The Tempest (V. i.):—

are dactylic tetrapodies with catalexis.

(c) Spondees.—It is not certain that these are used as a base, though as a variant in anapæstic and dactylic metre they are common. Iph. Taur., 123-5, may be taken as spondees:—

– –|– –| – –|– –|– –|– –| – –|– –|– –|– –|

But they may be quasi-anapæsts, the whole passage which they introduce being an anapæstic entrance-march, though heavily spondaic. Ion, 125-7:

is scanned by Dr. J. H. H. Schmidt as molossi, a molossus being – – –.

Spondaic systems are scarcely to be found in English.[887]

(d) Cretics.—This foot (–⏑–) is rare; it generally expresses piteous agitation:—

–⏑–|–⏑–|–⏑–‖ –⏑–|–⏑–‖ –⏑⏑⏑|–⏑–‖ –⏑⏑⏑|–⏑–‖ –⏑–|–⏑–‖

Few cretics are found in English, though Tennyson’s brief poem The Oak is written entirely in this metre, e.g.:—

Most English verse of cretic appearance is shown by the context to be trochaic with alternate τονή. So in A Midsummer Nights Dream, II. i.:—

which is followed by

We are forbidden to view the Greek cretics given above in the same way, by the resolved feet. If we scan φρόντισον καὶ γενοῦ πανδίκως as –⏑|⏗|–⏑|⏗|–⏑|–ꞈ‖, this method will give us in the fourth line –⏑|⏑⏑|–⏑|–‖, where the second foot is impossible. ⏑⏑ can take the place of –, but never of ⏗.

(e) Bacchiacs.—This curious foot consists of – –⏑, the system being invariably introduced by anacrusis. Bacchiacs are regularly associated with dochmiacs (see below). They express strong emotion, generally mingled with perplexity or vacillation; resolved feet are therefore often found:—

⏑⁝– –⏑|– –ꞈ‖ ⏑⁝– –⏑|– –ꞈ‖.

⏑⁝– –⏑|– –⏑‖– –⏑|– –ꞈ‖ ⏑⁝⏑⏑–⏑|– –ꞈ‖.

⏑⁝– –⏑|– –ꞈ‖ ⏑⁝– –⏑|–⏗̭‖ ⏑⁝– –⏑|– –ꞈ‖ ⏑⁝– –⏑|–⏗̭‖.

But it should be noted that, though bacchiac scansion seems soundest for the above—“storm-winds” for instance has two ictuses—the poet probably meant the lines for dactylic dipodies with anacrusis: “storm-winds of” thus would be an accentual dactyl. But that would slur “winds” unduly.

(f) Ionics.—These are formed by – –⏑⏑. When anacrusis is found—the usual form—the foot is often called Ionicus a minore (i.e. ⏑⏑– –); otherwise it is called Ionicus a maiore:—

⏑⏑⁝– –⏑⏑|– –⏑⏑‖– –⏑⏑|– –ꞈ̄‖.

A strange variant is ⏑–⏑–; the variation is called “anaclasis” (“breaking-up”). Thus the above passage proceeds—

⏑⏑⁝– –⏑⏑|– –⏑⏑‖–⏑–⏑|–ꞈ̄‖.

Ionics are employed to express strong excitement governed by confident courage. The first lyric of the Persæ begins with a splendid example. It is sung by the Persian counsellors in expectation of Xerxes’ triumph, and makes a strong contrast with the piteous rhythms of the close. This poem should be studied carefully in comparison with another in the same metre—the opening of the first chorus in the Bacchæ (vv. 64 sqq.):—

⏑⏑⁝⏘⏑⏑|– –ꞈ̄‖ ⏑⏑⁝– –⏑⏑|– –⏑⏑|– –ꞈ̄‖ ⏑⏑⁝⏘⏑⏑|– –ꞈ̄〛 ⏑⏑⁝– –⏑⏑|– –⏑⏑|– –⏑⏑|⏘ꞈ̄〛 ⏑⏑⁝⏘⏑⏑|– –⏑⏑|– –⏑⏑|– –ꞈ̄‖ ⏑⏑⁝– –⏑⏑|– –⏑⏑|– –ꞈ̄‖ ⏑⏑⁝– –⏑⏑|– –⏑⏑|–⏑–˃|⏘ꞈ̄〛

I ⎛ 2 ⎜ 3—mesode. ⎝ 2 II ⎛ 2 ⎝ 2 III ⎰ 2 ⎛ ⎱ 2 ⎜ 3—mesode ⎝ ⎰ 2 ⎱ 2

This song of the Bacchantes, like that of the Persians, expresses both excitement and confidence; both are magnificent, and the metre is the same. But the difference is unmistakable; it lies in the rhythm. In Æschylus the practically unvaried rhythm and the gorgeous language give to such a passage as πολύχειρ καὶ πολυναύτας Σύριόν θ’ ἅρμα διώκων an almost intolerable weight and austere pomp. Euripides, by use of the doubly-lengthened syllable, by varying the extent of his cola, and by the irrationality of the penultimate foot, has, within the limits of the same metre, produced a sense of exotic beauty and urgency, a thrill of wildness as well as of awe.

(g) Choriambics.—These consist of –⏑⏑–. Anacrusis is not found:—

–⏑⏑–|–⏑⏑–‖–⏑⏑–|–⏑⏑–‖ –⏑⏑–|–⏑⏑–‖–⏑⏑–|–⏑⏑–‖

This measure expresses great agitation and perplexity. In the passage just cited they pass into ionics, which indicate a gradual comparative calming of mind. For example, the antistrophe reads:—

–⏑⏑–|–⏑⏑–‖–⏑⏑–|–⏑⏑–‖ –⏑⏑–|–⏑⏑–‖–⏑⏑–|–⏑⏑–‖ ⏑⏑⁝– –⏑⏑|– –⏑⏑‖– –⏑⏑|⏘ꞈ̄‖ ⏑⏑⁝– –⏑⏑|⏘ꞈ̄‖.

The late Rupert Brooke left some exquisite Experiments in this metre, e.g.:—

That is—

⏗⏗|–⏑⏑–|–⏑⏑–‖–⏑⏑–|⏗⏗‖

(h) Dochmiacs.—It is convenient to discuss these here, though the dochmius is not a foot, but a colon. The rule both of metre and music is that all feet or bars should have the same time-value; a trochaic colon may contain ⏗ or ᷋⏑ as well as –⏑, but not –⏑⏑. Dochmiacs are generally regarded as an exception to this rule. The dochmius is a colon of which the simplest form[888] is ⏑– –⏑–, to be divided ⏑⁝– –⏑|–ꞈ||, e.g. κακορρημόνων. The dochmius is always catalectic, but the anacrusis of one serves to complete the trochee of the preceding colon:—

⏑⁝– –⏑|–⏑‖– –⏑|–ꞈ‖.

But this simplest form is not the most frequent, and a considerable sequence is rare. Resolution of one or more long syllables is very common. The favourite form is ⏑⁝⏑⏑–⏑|–ꞈ‖:—

⏑⁝⏑⏑–⏑|–⏑‖⏑⏑–⏑|–ꞈ‖

This metre is frequent in passages of lamentation, and as these are extremely numerous the dochmiac measure is one of the most important. It is also perhaps the most difficult, because of the many varieties admitted. In all, twenty-two[889] forms are said to be found, though several of these are rare; this great number is due to resolution and irrational long syllables. Thus—

⏑⁝⏘⏑|–ꞈ‖ ⏑⁝⏑⏑⏑⏑⏑|⏑⏑⏑‖⏑⏑⏑⏑⏑|⏑⏑ꞈ‖ ⏑⁝⏑⏑–⏑|–⏑‖– –⏑|–ꞈ‖.

The second line of course would by itself have no rhythm at all, being so completely broken to pieces, in order to express the extreme limit of agitation possible in articulate speech. But it gains rhythm from the clearer lines of the context. The antistrophe shows a further variety—an irrational syllable in the last line:—

(κηδευ͐ων). Evidently it is important to accustom one’s ear thoroughly to the basic form ⏑– –⏑– and to ⏑⏑⏑–⏑–. Another instance may be of use:—

˃⁝⏑⏑–˃|–˃‖⏑⏑–⏑|⏑⏑ꞈ‖ ⏑⁝⏑⏑–⏑|–˃‖⏑⏑–⏑|–ꞈ‖ ˅⁝⏑⏑–˃|–ꞈ‖.

The last division of our subject is the different types of period, the various ways in which cola are combined and correspond. It should be noted that a colon with anacrusis can correspond to one without; so of catalexis and τονή.

(i) The simplest form is the stichic (στίχος “a row”), in which the cola are of the same length. The scheme

⎧ a ⎧ 2 ⎧ 4 ⎧ 6 ⎧ a is ⎨ —that is ⎨ or ⎨ or ⎨ , etc.—or ⎨ a and so forth: ⎩ a ⎩ 2 ⎩ 4 ⎩ 6 ⎩ a

–⏑|⏗|–⏑|–ꞈ‖ –⏑|–⏑|–⏑|–ꞈ〛.

Where correspondence is indicated by

⎛ 4 ⎝ 4.

It makes no difference that ⏗ is answered by –⏑ in the second foot.

(ii) To the stichic corresponds the palinodic period (παλινῳδία, “repetition”), in which not a single colon but a group of cola is repeated so far as length is concerned:—

–⁝⏗⏑|– –|⏗⏑|– –‖–⏑⏑|– –‖ –⁝⏗⏑|⏘|–⏑⏑|– –‖⏗⏑|–ꞈ̄‖

⎰ 4 ⎞ ⎛ ⎱ 2 ⎟ ⎞ ⎝ ⎰ 4 ⎠ ⎟ ⎱ 2 ⎠

This type of period is frequent in English poetry, where the use of rhyme and the absence of τονή make the cola perfectly plain, e.g.:—

⎰ 4 ⎞ ⎛ ⎱ 3 ⎟ ⎞ ⎝ ⎰ 4 ⎠ ⎟ ⎱ 3 ⎠

(iii) Antithetic periods are formed by the inverted repetition whether of different cola or of different groups of cola.

(a) The simplest type is that in which a series of ungrouped cola is repeated in inverse order:—

⏑⁝⏗|⏗|–⏑|–⏑|–⏑|–ꞈ‖ –⏑|–⏑|–⏑|–ꞈ‖ ⏑⁝⏗|⏗|–⏑|–ꞈ‖ ⏑⁝⏗|⏗|–⏑|–⏑|–⏑⏑|–˃〛.

6 ⎞ 4 ⎞ ⎟ 4 ⎠ ⎟ 6 ⎠

(b) In a similar manner groups may be repeated antithetically. Each group retains its internal order; hence such periods are called “palinodic-antithetic”:—

⏑⁝⏗|⏗|–⏑|–⏑|⏗|–ꞈ‖ –˃|–⏑⏑|⏗|–ꞈ‖ ⏑⁝–⏑|–⏑|–⏑|⏗‖–⏑|–⏑⏑|⏗|–ꞈ‖ –⏑|–⏑⏑|–⏑|⏗‖–⏑|–⏑⏑|⏗|–ꞈ‖ ⏑⁝–⏑|–˃|–⏑|–⏑|–⏑|–ꞈ‖ –⏑⏑|–⏑|⏗|–ꞈ〛.

⎰ 6 ⎞ ⎛ ⎱ 4 ⎟ ⎞ ⎜ ⎰ 4 ⎞ ⎟ ⎟ ⎜ ⎛ ⎱ 4 ⎟ ⎞ ⎟ ⎟ ⎜ ⎝ ⎰ 4 ⎠ ⎟ ⎟ ⎟ ⎜ ⎱ 4 ⎠ ⎟ ⎟ ⎝ ⎰ 6 ⎠ ⎟ ⎱ 4 ⎠

(iv) Any of the three periods just described, the stichic, the palinodic, the antithetic (whether simple or palinodic-antithetic) may be “mesodic,” that is, it may be grouped round a central colon (the mesode), to which no colon corresponds, save of course the mesode of the other strophe. The schemes, then, are:—

(a) Stichic-mesodic.

a ⎞ x ⎟ a ⎠

(b) Palinodic-mesodic.

⎰ a ⎞ ⎛ ⎱ b ⎟ ⎞ ⎜ x ⎟ ⎟ ⎝ ⎰ a ⎠ ⎟ ⎱ b ⎠

(c) Antithetic-mesodic.

a ⎞ b ⎞ ⎟ x ⎟ ⎟ b ⎠ ⎟ a ⎠

(d) Palinodic-antithetic-mesodic.

⎰ a ⎞ ⎛ ⎱ b ⎟ ⎞ ⎜ ⎰ c ⎞ ⎟ ⎟ ⎜ ⎛ ⎱ d ⎟ ⎞ ⎟ ⎟ ⎜ ⎜ x ⎟ ⎟ ⎟ ⎟ ⎜ ⎝ ⎰ c ⎠ ⎟ ⎟ ⎟ ⎜ ⎱ d ⎠ ⎟ ⎟ ⎝ ⎰ a ⎠ ⎟ ⎱ b ⎠

(a) The stichic-mesodic:—

⏑⁝–⏑|⏗|–⏑|–⏑|⏗|–ꞈ‖ –⏑⏑|–⏑|⏗|–ꞈ‖ ⏑⁝–⏑|⏗|–⏑|–⏑|⏗|–ꞈ‖.

6 ⎞ 4 ⎟ 6 ⎠

(b) The palinodic-mesodic:—

⏑⁝⏗|⏗|–⏑|–ꞈ‖ ⏑⁝⏗|⏗|–⏑|–⏑|⏗|–ꞈ‖ –˃|–⏑⏑|–⏑|–˃‖ ⏑⁝⏗|–⏑|–⏑⏑|–ꞈ‖ ⏑⁝⏑⏑⏑|⏑⏑⏑|–⏑|–⏑|⏗|–ꞈ‖.

⎰ 4 ⎞ ⎛ ⎱ 6 ⎞ ⎟ ⎜ 4 ⎟ ⎟ ⎝ ⎰ 4 ⎟ ⎠ ⎱ 6 ⎠

(c) The antithetic-mesodic:—

⏑⁝–⏑|–⏑|–⏑|–⏑‖ –⏑|⏑⏑⏑|–ꞈ‖ ˃⁝⏑⏑⏑|–⏑|–ꞈ‖ –⏑⏑|–⏑|–ꞈ‖ ˃⁝⏑⏑⏑|–⏑|⏗|–ꞈ‖.

4 ⎞ 3 ⎞ ⎟ 3 ⎟ ⎟ 3 ⎠ ⎟ 4 ⎠

(d) The palinodic-antithetic-mesodic:—

–˃|–⏑⏑|–ꞈ‖ ⏑⁝–⏑|–⏑|–⏑|–⏑|–⏑|–ꞈ‖ –⏑|–⏑|–⏑|–⏑‖–⏑|–⏑|–⏑|–ꞈ‖ ω⁝–⏑⏑|–⏑|–⏑⏑|–ꞈ‖ –˃|–˃|–⏑⏑|–⏑‖–⏑|–⏑|–⏑|–ꞈ‖ –˃|–˃|–ꞈ‖ ⏗|⏗|–⏑|–⏑|–⏑|–ꞈ〛.

⎰ 3 ⎞ ⎛ ⎱ 6 ⎟ ⎞ ⎜ ⎛ ⎰ 4 ⎞ ⎟ ⎟ ⎜ ⎜ ⎱ 4 ⎟ ⎞ ⎟ ⎟ ⎜ ⎜ 4 ⎟ ⎟ ⎟ ⎟ ⎜ ⎜ ⎰ 4 ⎠ ⎟ ⎟ ⎟ ⎜ ⎝ ⎱ 4 ⎠ ⎟ ⎟ ⎜ ⎰ 3 ⎠ ⎟ ⎝ ⎱ 6 ⎠

Most of the periodic structures which have been described are by no means obvious to the ear. A trained sense of rhythm, attention to quantity, and careful practice, will reduce the difficulties. But in any case Greek periods are far less easy to grasp than English. Their variety and length, the frequent occurrence of prolongation, resolution, and irrational syllables, the possibility of preludes or postludes—all these are formidable to modern students, who lack the help of the music. We may perhaps work out the period with ease on paper, but our ear often cannot appreciate the balance and contour of the whole as it can in English lyrics, where we have the immense assistance of a rhyme-scheme. But it is no sound deduction that the study of Greek lyric metre and rhythm is therefore useless. We cannot always hear the period—that is a question of music; but we can always hear the colon—that is a question of language. To utter the cola correctly is easy after a little practice; and it is these “sentences” which, by their own internal rhythmical nature and by the identities or contrasts existing between them, reinforce and more pungently articulate the sense of the words wherefrom they are moulded.