The dramatic unities, three principles governing the structure of drama and supposedly derived from Aristotle’s Poetics, are a subject of perennial interest. They are known as the unities of time, place, and action, respectively, and require that “the action of a play should be represented as occurring in one place, within one day, and with nothing irrelevant to the plot.” The essential facts concerning them were recognized at least as long ago as the publication of Lessing’s Hamburgische Dramaturgie (1767). But so deep-rooted is the popular impression that the Greeks formulated these rules arbitrarily and observed them slavishly that no attempt to state the true situation can be superfluous. The current doctrine is based on the fact that the classic dramatists in France and Italy blindly obeyed the rules as a heritage of the past, without regard to the demands of the theater at their own disposal; and, consequently, the inference has been easily and naturally drawn that the ancient practice was equally irrational.

But in the Greek theater, where there was no drop curtain, no scenery to shift, and a chorus almost continuously present, a change of scene was difficult to indicate visually. Nevertheless Aristotle nowhere mentions the unity of place, and the Greek dramatists not infrequently violate it. The most familiar instances occur in Aeschylus’ Eumenides and Sophocles’ Ajax. The former play opens at the temple of Apollo in Delphi, whither the avenging Furies have pursued Orestes after his mother’s murder. During a momentary lapse from their watchfulness Orestes makes his escape, but the Furies soon awaken and take up the trail once more. The scene is thus left entirely vacant (vs. 234) and is supposed to change to Athens, where all parties presently appear for the famous trial before the Council of the Areopagus. The beginning of the latter play takes place before Ajax’ tent, and Sophocles wished to introduce the very unusual motive of having a scene of violence enacted before the audience. As the presence of the chorus was an insuperable obstacle to such a deed, Ajax was allowed to leave the scene and, suspicion being soon aroused, the chorus was sent in search of him (vs. 814). Thus, the scene is again entirely deserted by both actors and chorus, and Ajax returns, not to his tent, but to some lonely spot near the seashore (see pp. 129 and 244, above). This was by far the most natural and logical method of leading up to a change of scene, was infinitely superior to Shakespeare’s practice in King Henry V, where Chorus is introduced in the prologue of each act to acquaint the spectators with the scene of the succeeding action, but was so difficult to motivate that only some half a dozen examples are known to us in the whole Greek drama. On the other hand, such a technical device was usually not well adapted to represent considerable shifts of scene, since it would seem unnatural for so large a body of persons as the chorus always to accompany the dramatic characters to widely separated localities. To this general restriction, however, the Eumenides furnishes a brilliant exception, because it was the especial duty of the Furies to track the guilty Orestes wherever he might flee. In Old Comedy, ever fantastic and intentionally impossible, greater freedom was naturally allowed than in tragedy, so that in Aristophanes’ Frogs no less than five different scenes are successively required (see pp. 88-90, above).

At the same time the need of such scene-shifting was largely obviated by the arbitrary placing of almost all scenes before a building, by the exclusion of interior scenes, and by the various devices substituted therefor (see pp. 237-42, above). In particular the use of the messenger’s speech enabled dramatists to bring indirectly before their audiences events which had taken place, not merely in the scene-house interior, but at far distant spots. Very commonly the unity of place was observed by conventionally bringing together as close neighbors structures or localities which would actually be separated by considerable intervals. Thus the murderers of Agamemnon would not wish his grave to stare them in the face and to remind their subjects of their crime; nevertheless in Aeschylus’ Libation-Bearers palace and tomb stand side by side. Likewise in Euripides’ Helen, King Theoclymenus has buried his father in front of his palace. Now these arrangements are not to be interpreted in the light of the prehistoric custom of placing the dead within the house or before its threshold. It is purely a theatrical convention, and Euripides shows what he thought of it by deeming it necessary to put an excuse on the Egyptian king’s lips:

Many similar instances of incongruous juxtaposition in Greek drama can be cited, and those who remember the use of the continuous set in mediaeval theaters will feel no surprise.

Slightly different but no less efficacious is the method of procedure in the Persians. For dramatic effect Aeschylus wished to introduce the ghost of Darius. But according to ancient notions on the subject ghosts do not normally wander far from their tombs, and the real grave of Darius was at Persepolis. Furthermore, under the conditions supposed the Persian elders, the royal messenger, and Xerxes himself would not naturally resort thither. Consequently, without the slightest compunction, Aeschylus transferred the dead monarch’s tomb to Susa!

Sometimes the unity of place was observed by causing a character to come to a spot to which he would not naturally resort. The scene of Euripides’ Phoenician Maids, for instance, is laid in Thebes, and the poet wished to show a meeting of the Theban king and his brother. Since the latter is considered a traitor and the enemy of his country, is in banishment and at the head of an invading army, such a meeting in real life would almost inevitably be held between the hostile lines. Yet Polynices is forced to intrust his head to the lion’s jaws and enter the city. He expresses his own misgivings in vss. 261 ff., concluding:

At vss. 357 ff. he alludes to the matter once more.

Similarly, a character is oftentimes forced to remain upon the scene of action when he would not naturally do so. Thus, in Plautus’ Menaechmi, owing to a failure to distinguish Menaechmus I from his brother, his father-in-law and a physician consider him insane and make arrangements, in his hearing, for his apprehension. Notwithstanding, when they both leave the stage at vs. 956 he makes no attempt to escape—an act which would transfer the next two scenes elsewhere—but unconcernedly awaits developments.

Finally I may mention one especially amusing artifice. In Euripides’ Iphigenia among the Taurians, Orestes has left the scene and is now supposed to be some distance away. Notwithstanding, Athena addresses him and apologetically adds: “For, though absent, you can hear my voice, since I am a goddess” (vs. 1447). The same situation recurs, without apology, at vs. 1462 and in Euripides’ Helen, vss. 1662 ff.

Likewise, the unity of time arose, not from the whim of ancient writers, but from the same theatrical arrangements which resulted in the unity of place, viz., the absence of a drop curtain and the continuous presence of the chorus. Under these conditions an intermission for the imaginary lapse of time could be secured only by the withdrawal of the chorus, and without such intermissions the constant and long-continued presence of the same persons in the same place without food or slumber was in danger of becoming an absurdity. Now we have seen how difficult it was to invent motives for the successive reappearances of actors; to motivate the movements of a body of twelve (fifteen) tragic or twenty-four comic choreutae was naturally still more difficult (see pp. 229-33 and 150-52, above). Consequently the chorus is rarely removed from the stage during the action. Two instances have already been mentioned (p. 247, above). In the Ajax advantage is taken of the withdrawal to change the scene slightly; naturally a slight interval of time is also supposed to elapse, but in this instance this is negligible and without significance. In the Eumenides the case is different. Here the scene is not shifted a few rods merely but from Delphi clear to Athens. As the crow flies this was a distance of about eighty miles and, in view of the physical conditions and ancient methods of travel, would require two or three days to traverse. Accordingly a considerable lacuna in the dramatic time of the play must be assumed. What is still more remarkable is that, except for the empty stage, the spectators are given nothing to help “digest the abuse of distance.” At vs. 80 Apollo dispatches Orestes to the city of Pallas, at vs. 179 he begins to drive the chorus of Furies from his shrine, at vs. 234 he leaves the stage and the scene is empty. Up to this point we are still at Delphi. In the very next verse (235) Orestes rushes into the theater and exclaims, “O queen Athena, I come at the bidding of Loxias.” He has reached Athens! In Euripides’ Alcestis the chorus forms part of the queen’s funeral cortège and is absent during vss. 747-860. Although it is not usually so regarded I am inclined to think that there is a slight change of scene here (see p. 235, above); there is also a slight condensation of time, but neither constitutes a serious violation of these unities. This is one of the rare cases where the withdrawal of the chorus resulted naturally from the normal development of the plot. For if the choreutae had been present when Heracles announced his intention of rescuing Alcestis from death (vss. 840 ff.) the poet must have invented a reason for their not reporting this news to Admetus or have spoiled certain features of the finale. It was much simpler to avoid the difficulty by allowing the chorus to do the natural thing. In the following instances apparently no change of scene or undue compression of time is involved. In Euripides’ Helen (vs. 385) the chorus accompany their mistress inside the palace to consult the seeress Theonoe and re-enter at vs. 515. The only advantage that seems to accrue from this maneuver is to prolong Menelaus’ uncertainty as to the identity of his newly recovered wife. In Aristophanes’ Women in Council (vs. 311) the women of the chorus, disguised as men, leave for the assembly in order to vote the management of the state into their own hands, returning at vs. 478. Unless the playwright wished to have the assembly scene enacted before the audience he had to withdraw the chorus. As it is their doings are reported by a messenger (Chremes) in vss. 376 ff. In the pseudo-Euripidean Rhesus the chorus is absent during vss. 565-674, being sent in front of the camp to receive Dolon (cf. vss. 522 ff.). The presence of Trojan guards would have prevented the intervening scene between the Greek marauders, Odysseus and Diomedes. It will be noted how few are the instances of the withdrawal of the chorus in the extant plays and that the observance of the unities figures in just half of them. In New Comedy the chorus appeared only between acts (see p. 145) and it would have been feasible to assume a lacuna several times in each play. That this was not done was probably due to the fact that the other practice had become stereotyped and that concentration of action resulted in greater unity of plot. Sometimes the stage is left empty before the entrance of the chorus by the retirement of all the actors on the scene either between the prologue and the parodus or between monologues (or dialogues) in the prologue. Euripides’ Alcestis (vs. 77) furnishes an example of the former and his Iphigenia among the Taurians (vs. 66) of the latter. So far as I have observed such pauses are not made use of to accelerate the time unduly.

Since it was not often possible to suspend the audience’s sense of time by removing the chorus, the poets had recourse to the next best expedient, the choral odes. Inasmuch as several of these occurred in every play, this artifice was far more available than the other. In many respects the chorus moved upon a different plane from the actors, and we are now dealing with one of these differences. As Professor Butcher expressed it: “The interval covered by a choral ode is one whose value is just what the poet chooses to make it. While the time occupied by the dialogue has a relation more or less exact to real time, the choral lyrics suspend the outward action of the play and carry us still farther away from the world of reality. What happens in the interval cannot be measured by any ordinary reckoning; it is much or little as the needs of the piece demand. A change of place directly obtrudes itself on the senses, but time is only what it appears to the mind. The imagination travels easily over many hours; and in the Greek drama the time that elapses during the songs of the chorus is entirely idealized” (op. cit., p. 293). Thus the choral songs were roughly equivalent to the modern intermission, and after them the action is often farther advanced than the actual time required for chanting them would warrant. For example, during a single stasimon of Aeschylus’ Suppliants (vss. 524-99) the Argive king must leave the scene, summon his subjects to public assembly, state the object of the meeting, and allow discussion before the final vote—all in time for Danaus to report the people’s decision at the beginning of the following episode! An analogy to ancient practice occurs in Shakespeare’s The Winter’s Tale, where Time as Chorus announces the passage of sixteen years between Acts III and IV.

But at the same time that the chorus conferred this liberty it restricted it. The presence of such a body of performers at all the scenes of a play could seldom be entirely natural. Yet that the same persons should be found standing about, in the same place, at various intervals during the day is conceivable, though it does not often happen. But that they should be found there at every moment chosen for representation during weeks or months or years is inconceivable and ridiculous. Only by shortening the supposed action of the piece and the supposed lacunae in the plot could the convention be tolerated at all. However, Professor Verrall was lacking in historical imagination when he maintained that “the point at which the discrepancy between the facts presented and the natural facts began to be flagrant and intolerable was when the audience were told to pass in imagination from day to day. Night is the great natural interrupter of actions and changer of situations” (op. cit., p. 1). To the spoiled theatergoer of today this would seem to be true. But the ancient drama knew no lighting effects (see p. 224, above). On the stage day and night looked the same to them. Scenes at midday, in the darkness of night, in the gloom of Hades, were alike enacted in the glare of the sun. Ostensibly the entire action of the anonymous Rhesus and much of that in Euripides’ Cyclops fell within the hours of night, and characters frequently addressed the heavenly constellations in (actual) daylight. So far were the playwrights from avoiding the discrepancy involved in passing from one day to another that in Terence’s translation of Menander’s Self-Tormentor, when a night is supposed to elapse between Acts II and III, attention is deliberately called to it by Chremes’ words, “It is beginning to grow light here now” (vs. 410). In my opinion this play extends over about as much time as the conditions which obtained in ancient drama would normally allow; and it should be noted that it does not exceed the twenty-four hours permitted by the unity of time.

In the third place, perhaps it is unnecessary to point out that acceleration of time is possible in all drama quite apart from an empty stage or choral songs. Instances can be cited even from dramatists who owned no allegiance to the unities—note, for example, the striking of the half-hour every twenty or twenty-five lines at the close of Marlowe’s Doctor Faustus. In Aristophanes’ Plutus the blind god is escorted from the stage for a night’s treatment in the temple of Asclepius (vs. 626), the chorus remaining in its place but apparently not singing.[321] At the very next verse one of the escort returns to announce that Plutus has recovered his sight and to relate the events of the night! But here again, despite the transition from one day to another, the action does not exceed twenty-four hours. In the same writer’s Acharnians, Amphitheus goes from Athens to Sparta and returns again during the dialogue contained between vss. 133 and 174. There is no hint, however, that his reappearance is premature or that his trip would occupy more than the apparent space allotted it.

But neither the ordinary acceleration of time in drama nor the use of stasima nor yet the stage left empty by the retirement of chorus and actors tells the whole story of Greek practice. Nowadays the playbill clearly informs us how much time has elapsed between acts, and the piece is constructed accordingly. If a character in the third act has occasion to refer to something which occurred in the first act ten years or so ago he must not speak of it as if it happened yesterday. Not so in ancient drama. The Greek audiences had no playbills, and even the introductions to Greek plays prepared by Alexandrian scholars contained no such information as this. I fancy that the Greek dramatist never laid his finger upon a given line and said: “Here we must assume a lapse of several days, or months, or years.” The events of a drama, regardless of actualities, were conventionally treated as occupying no more than twenty-four hours. A like convention was customary in the Greek epic: when once a Homeric character was given a definite age or form he maintained each unchanged throughout.[322] For example, Telemachus is introduced in the first book of the Odyssey as a young man just reaching his majority, ready and anxious to assume the duties of manhood; but nine years before, when he could not have been more than twelve years of age, he is spoken of as just as old and as already a man among men (cf. Book xi, vss. 185 f. and 449). Again, in the third book of the Iliad, Helen is pictured in the prime of youth and beauty; ten years later and thirty years after her elopement with Paris she is likened to the same goddess as is the Maiden Nausicaa (cf. Odyssey iv. 121 f. and vi. 102 ff.). In Greek drama time relations are similarly ignored. At the opening of Aeschylus’ Agamemnon the watchman sights the signal fire which announces the capture of Troy, and within a few hundred lines Agamemnon has finished the sack, traversed the Aegean, and appeared before his palace! No hint is given, however, that there is anything unusual about all this; not a word[323] indicates that the action is disconnected at any point.

This is the most flagrant instance, and I conceive that it is to be interpreted as follows: The performance of Greek drama in the fifth century was continuous in the sense that with negligible exceptions (see pp. 250 f., above) actors or chorus or both were constantly before the audience. Notice that this is not the same as saying that the time of the plays was continuous. When critically examined it is found to have been interrupted by numerous gaps, as we have already seen and shall see again. But the continuity of performance gave a semblance of continuity also to the action. Therefore when a modern playwright like Pinero restricts his action to one day and represents the lapse of several hours by the fall of the curtain between acts, he does not thereby observe the unity of time in the Greek sense. The dramatic events were tacitly treated by the poets as if they occupied no more than a day and were so accepted by the public. By “tacitly” I mean that if such crowding involved a physical or moral impossibility the dramatists never stooped to apologize or explain but placed their events in juxtaposition just the same. In Plautus’ Captives, Philocrates travels from Aetolia (the scene of action) to Elis and back again between vss. 460 and 768. In real life such a trip would have required several days, but in the play it consumes less than one! Do we positively know this? Beyond the shadow of a doubt. A parasite is introduced at intervals during the play scheming to be invited to a meal. He is first seen at vs. 69 and does not get a satisfactory invitation until vs. 897. A more detailed statement would show conclusively that the same day’s meal is under discussion throughout. Moreover, this is no mere lapsus calami, such as a few phrases which are found in an opposite sense,[324] but is unmistakable in its import and is closely interwoven with the plot. If anyone feels amazed at so deliberate a contradiction he may console himself with a study of the use of “double time” in Shakespeare. It would be possible, but is quite unnecessary, to cite other plays in which restriction of time to a single day is indicated with sufficient exactness. Of course the Greek dramatists did not consistently introduce references to the precise date or to the time of day. In general they were wise enough to act upon the principle which Corneille[325] expressed as follows: “Above all I would leave the length of the action to the imagination of the hearers, and never determine the time, if the subject does not require it.... What need is there to mark at the opening of the play that the sun is rising, that it is noon at the third act, and sunset at the end of the last?”

It is somewhat remarkable that Professor Verrall, who fully recognized the dependence of this unity upon local conditions and published eminently sensible observations on the subject, nevertheless felt constrained to challenge the obvious interpretation of two plays in which a glaring violation of the unity of time occurs. In the Agamemnon he supposed the watchman and the populace (including the chorus) to be misinformed as to the meaning of the beacon and that it really served to Clytemnestra, Aegisthus, and their supporters as a warning of Agamemnon’s being close at hand! His elucidation of Euripides’ Andromache was still more ingenious and complicated.[326] But to bolster up such interpretations Mr. Verrall ought to have explained away all similar instances as well—to explain, for example, how in Euripides’ Suppliants an Attic army can march from Eleusis to the vicinity of Thebes and fight a battle there, and how tidings of the victory can be brought back to Eleusis, all between vss. 598 and 634, which, as Dryden[327] expressed it, “is not for every mile a verse.” Nevertheless not the slightest attention is paid to such patent impossibilities, and in every case the whole action is unmistakably supposed to fall within a day.

In view of the foregoing it is not surprising that Aristotle does mention the unity of time, though only incidentally. His exact language is: “Tragedy and epic differ, again, in their length: for tragedy endeavors, so far as possible, to keep within a single circuit of the sun (περίοδος ἡλίου), or but slightly to exceed this limit; whereas the epic action has no limits of time.”[328] “Endeavors” (πειρᾶται) was mistranslated as doit by some French writers. Aristotle rather commends the unity of time as a rough generalization which works out well in practice than enjoins it as an invariable rule. Actually the restriction was further reduced, in most cases, to the hours of daylight, and Dacier even maintained that περίοδος ἡλίου means no more than twelve hours. But Aristophanes’ Plutus and Terence’s Self-Tormentor (see pp. 253, above) furnish clear examples of dramatic action beginning in the late afternoon of one day and not concluding until the next day.

It remains to consider some of the expedients which the poets found useful in solving the difficulties (both of time and place) caused by local conditions. In the first place the practice of writing a series of three plays on the same general subject (see p. 198, above) often enabled the playwright to distribute his incidents in different places and time-spheres without loss of verisimilitude, for a whole trilogy was no longer than the average modern play, and each tragedy would thus correspond to a single act and, since the chorus was withdrawn at the close of each play in the trilogy and its place taken by another entirely different, changes of time and place between plays were absolutely without restriction. Thus Scythia of Aeschylus’ Prometheus Bound becomes Caucasus in the second piece in the trilogy, the Prometheus Unbound; in the former was shown the binding of the Titan and in the latter his release, and he is said to have been bound for 30,000 years. All but two days of this time elapses between plays! In Aeschylus’ Orestean trilogy the scene of the Agamemnon and Libation-Bearers is laid in Argos; that of the Eumenides in Delphi and Athens. Several years are supposed to pass by in the two interims.

But even Aeschylus did not always employ the trilogic form, and Sophocles and Euripides rarely did. When, therefore, the three or four plays in each series were severally devoted to utterly unrelated material, it sometimes became necessary to bring almost as many events within the scope of one play as would otherwise be dealt with in a whole trilogy. Inasmuch as a large fraction of these events could not possibly be conceived of as taking place in the same locality or within the same day, it was imperative either to exclude them or to include them in some indirect fashion. Now two striking peculiarities of Euripidean technique were admirably adapted to help solve these difficulties. His prologues regularly take the form of a monologue, which, with scant regard for dramatic illusion, rehearses the story of the myth up to the point where the play begins. Again, Euripides’ dramas frequently terminate with the epiphany of a deity. This device was the accustomed recourse of unskilful playwrights, when their plots had become complicated beyond the possibility of disentanglement by natural means, in order that a god’s fiat might resolve all difficulties. It has often been charged that this was also Euripides’ motive, but most unjustly (see pp. 293 ff., below). He rather “wished, by the help of a divine foreknowledge, to put before the spectators such future events or unknown circumstances as should settle their minds, satisfy all curiosity, and connect the subject of the piece with subsequent events or even with the times of living men.”[329] Thus in Euripides’ Andromache the complications of the plot are entirely solved before Thetis’ appearance at vs. 1231, and she merely gives directions for Neoptolemus’ burial and prophesies the future of Peleus, Andromache, and Molossus, and of the latter’s posterity. When these two pieces of technique were combined in the same play, the prologue, the body of the tragedy, and the epilogue sometimes corresponded roughly to the successive dramas of a whole trilogy. This appears most clearly in the case of Euripides’ Electra and Aeschylus’ Orestean trilogy. The opening monologue of the former (vss. 1-53) passes in rapid review the Greek expedition against Troy, the murder of Agamemnon, and the present fate of his children. With the exception of the last item, which is brought out in the second play of the Oresteia, these are the matters contained in the prologue, which naturally is comparatively short, and in the action of Aeschylus’ Agamemnon. The body of the Electra corresponds to the second tragedy in the trilogy, the Libation-Bearers. At the Electra’s conclusion (vs. 1238) Castor as deus ex machina forecasts among other things the acquittal of Orestes at Athens, which is the theme of Aeschylus’ Eumenides. Whatever other explanations, therefore, may be advanced for Euripides’ prologues and epilogues (see pp. 294 f. and 299 f., below) this consideration must also be allowed a certain weight, viz., that they permitted him to bring events of the most diverse nature within the scope of his piece without violating the unities of time and place.

A fourth device looking to the same ends consisted in setting conversations at times and places which would naturally be different. Even such a master of dramatic technique as Sophocles represented Orestes as communicating to his fellow-conspirators the result of his inquiry at Delphi only after they had reached Argos (Electra, vss. 32 ff.), and as waiting to formulate a definite plan of action until they were in the most unfavorable place in all the world for such a purpose—before Clytemnestra’s palace (vss. 15 ff.). The latter incongruity does not occur in Euripides’ version of the same story because the scene of his Electra is laid, not in the city of Argos, but before Electra’s hut in the country. The device under consideration was conveniently supplemented by the convention that if two or more characters enter the stage together no conversation is thought of as passing between them until they have come within the hearing of the audience (see p. 310, below). It will be seen that the passage just cited from Sophocles’ Electra conforms to this rule. Another instance occurs in Euripides’ Madness of Heracles, vss. 822 ff. Iris appears above Heracles’ palace with Madness, whom she orders to incite the hero to the murder of his children. Madness protests but is overborne and forced to perform her bidding. Though Iris and Madness must have come a considerable distance together, all discourse between them is apparently postponed until they reach their destination. Furthermore, these instructions would naturally have been given to Madness elsewhere and somewhat earlier. In that case the audience must have lost an effective scene. The device discussed in this paragraph enabled the poet to circumvent the unities and place the scene before his audience; and the convention which I have mentioned preserved it for them in its entirety.

We have seen that the unities of time and place are largely due to the striving for illusion in a theater comparatively bare of scenery and of facilities for scene-shifting. Conversely, their observance in the modern theater with its ample scenic provision would naturally militate against the scenic extravagance and actualism of which the present-day theatocracy is so enamored. Thus it would seem that the much-abused unities are not without a meaning and truly artistic tendency even today, for some of the most significant influences in contemporaneous staging are directed against excesses along these lines. Even a modern producer, Henry W. Savage, included the following in his advice to a young playwright: “Do not distribute your scenes so widely that you have one on an island, another at Herald Square, and a third at Chicago. Make the action of your play take place all in one day, if possible”[330]—in other words the unity of time expressly and an approximation to the unity of place. Ibsen surely retained no theatrical conventions merely because they were old; yet he usually observed the unities. A recent critic has written: “Though the unities of time and place were long ago exploded as binding principles—indeed, they never had any authority in English drama—yet it is true that a broken-backed action, whether in time or space, ought, so far as possible, to be avoided. An action with a gap of twenty years in it may be all very well in melodrama and romance, but scarcely in higher and more serious types of drama.”[331]

The unity of action is the only one that is universal, since it alone springs from the inmost nature of the drama. Yet even here local conditions make themselves felt. The modern playwright, free (if he pleases and has a producer complaisant enough) to change the scene ten times within a single act and with superior facilities for motivating entrances and exits, delights in shifting different sets of characters back and forth and thus secures an alternation of light and shade, an intermingling of comedy and tragedy quite beyond the ancient dramatist’s reach. The preceding discussion has shown the immobility of the ancient theater in these respects and, consequently, one reason why the Greeks ruthlessly excluded everything that was not strictly germane to their action (see also p. 201, above).

This unity, it is needless to say, plays an important part in Aristotle’s Poetics. He recognized that “plot is the first essential and soul of tragedy and that character comes second.”[332] The most lengthy statement runs as follows: “Let us now discuss the proper construction of the plot, as that is both the first and the most important thing in tragedy. We have laid it down that tragedy is an imitation of an action that is complete and whole, having a certain magnitude, for there is also a whole that is wanting in magnitude. Now a whole is that which has a beginning, a middle, and an end. A beginning is that which is not itself necessarily after anything else and after which something else naturally is or comes to be; an end, on the contrary, is that which itself naturally follows some other thing either as its necessary or usual consequent, and has nothing else after it; and a middle is that which is both itself after one thing and has some other thing after it. Accordingly, well-constructed plots must neither begin nor end at haphazard points, but must conform to the types just mentioned.”[333] These principles were excellently restated by Lowell:

Though it is now admitted on all sides that the unity of action is the sine qua non of dramatic composition, many fail to realize the meaning and extent of its limitation. Aristotle indicated a mistaken notion current in his day, and likewise in ours, in the following words: “The unity of a plot does not consist, as some suppose, in its having one man as its subject. An infinite multitude of things befall that one man, some of which it is impossible to reduce to unity, and so, too, there are many actions of one man which cannot be made to form one action. Hence, the error, as it appears, of all the poets who have composed a Heracleid, a Theseid, or similar poems. They suppose that, because Heracles was one man, the story also of Heracles must be one story.”[335] Freytag discussed the matter with keen discrimination and exemplified it by showing how Shakespeare remodeled the more or less chaotic story of Romeo and Juliet’s love into a unified plot whose incidents follow one another almost as inexorably as Fate. The passage is unfortunately too long for quotation here, but is highly instructive.[336]

The same reasoning reveals the shortcoming in Professor Lounsbury’s contention: “What, indeed, is the objection to this mixture of the serious and the comic in the same play? By it is certainly represented, as it is not in pure comedy or pure tragedy, the life we actually live and the mingled elements that compose it.... As there was no question that sadness and mirth were constantly intermixed in real life, it was impossible to maintain that the illegitimacy of this form of dramatic composition was due to its improbability.”[337] The word “pure” gives away the whole case. Aristotle would have to grant that Shakespeare’s plays are admirable, even sublime; but he could hardly admit that they were “pure” tragedies or “pure” comedies, however legitimate in other respects. They fall short in the quality which Mr. Albert H. Brown placed in the forefront of his definition: “A great drama is a clearly focused picture of human conditions.”

Aristotle also pointed out that epic poetry has an advantage in that it can present many events simultaneously transacted, while the drama is restricted to but one.[338] A curious violation of this self-evident principle occurred in a recent American play. Toward the end of Act II in Eugene Walter’s Paid in Full, Emma Brooks is disclosed making an appointment with Captain Williams over the telephone. In the next act we are transferred to Captain Williams’ quarters, and the dramatic clock has in the meanwhile been turned back some fifteen minutes, for presently the telephone bell rings and the same appointment is made over again. In other words, Act III partially overlaps Act II in time, but the scene is different. It can scarcely be denied that the dramatic situation has been enhanced by this device, but this gain has been secured at the sacrifice of verisimilitude and dramatic illusion. Such “cut-backs” may be all very well in moving pictures, but they hardly have a place in spoken drama.

Thus, the Greek masters were so far from evolving unities out of their inner consciousness or from observing them invariably that they constantly violated the unities of time and place in both letter and spirit. Their practice throughout simply reacted to theatrical conditions as they found them. It has remained for their successors, whose theater has for the most part been quite dissimilar, to observe the unities with a literalness and exactness such as never characterized the great dramatists of Greece. That both ancients and moderns have produced masterpieces under these restrictions is, of course, beyond dispute. In fact, some of our most impressive plays of recent date such as Kennedy’s Servant in the House, have conformed to them. That many modern plays would have been improved by observing them is doubtless also true. Even so uncompromising an admirer of Shakespeare as Professor Lounsbury[339] wrote:

But that the unities should be arbitrarily imposed upon every drama without exception is absurd, since the theatrical conditions that called them forth are no longer the same. That Aeschylus and Sophocles, if present with us in the flesh, would avail themselves of the greater flexibility and adaptability of the modern theater I cannot doubt. At any rate that restless spirit, Euripides, would certainly have gloried in its freedom.

As a cumulative result of the conditions already described the action of a Greek drama was restricted to the culmination alone, corresponding to the fifth act of most modern plays. Though we have seen that the Greek poets arbitrarily juxtaposed, as if within the confines of a sun’s circuit, events which were actually separated by considerable intervals, yet even the widest license would hardly permit a whole series of transactions, of sufficient dignity and importance to be chosen for tragic representation, to be compressed within a single day and limited to a single spot. As Dryden[340] expressed it, the ancient playwrights “set the audience, as it were, at the post where the race is to be concluded; and, saving them the tedious expectation of seeing the poet set out and ride the beginning of the course, they suffer you not to behold him, till he is in sight of the goal, and just upon you.” Thus in Aeschylus’ Suppliants we see nothing of the unwelcome suit of Aegyptus’ sons and of the events which led the daughters of Danaus to take refuge in flight. All this lies in the past and is brought before us indirectly. The action begins when the Danaids have reached another land and are on the point of being overtaken by their cousins. Similarly, in Euripides’ Alcestis we learn by hearsay the long story of Apollo’s servitude at the court of Admetus, of his providing a way of escape from death for the king, and of the latter’s disheartening search for a substitute. Only the final stage in the action, the day of the queen’s self-immolation and rescue, is chosen for actual representation. The same situation recurs in almost every piece. Of course in trilogies it was possible to select three different time-spheres and three different localities for the dramatic action. But here again only the crests of three crises in the story were put before the spectators’ eyes; all the rest was narrated. So invariable a method of attack would seem monotonous to us today, but its successful employment by Ibsen and many another in modern times proves that there is nothing blameworthy in the practice per se.

Finally, since the dramatic action was confined to a single day (however elastic) at the culmination of the story, it was rarely possible for the dramatis personae to experience any particular change or development of character during the course of the play. This fixity of type was not only a natural result of theatrical conditions in ancient times and of the use of masks but was also in thorough accord with Homeric conventions (see pp. 254 f., above). Moreover, it harmonized completely with the Greek fondness for schematization. Horace’s words in his Ars Poetica are entirely Hellenic in spirit: “Either follow tradition, or invent that which shall be self-consistent. In the former case, let Achilles be impatient, irascible, ruthless, keen...; let Medea be untamed and unconquerable, Ino tearful, Ixion treacherous, Io ever roving, and Orestes in sorry plight. In the latter case, keep the character to the end of the play as it was at the beginning and let it be consistent” (vss. 119 ff.). All this implies more than we would think desirable today. Not only was a positive development into a character seemingly inharmonious with that seen at first rarely possible, but the singleness of purpose in ancient plays, which has been called the unity of mood (see p. 201, above), crowded out incidents which might have revealed other phases, no matter how consistent, of a dramatic personage’s character. The taste of some critics objected to even the slight modifications in rôle which ancient conditions did permit. For example, to modern readers the manner in which Medea, in Euripides’ tragedy of that name, wavers between love for her children and the desire to punish her recreant husband by murdering them is esteemed one of the finest touches in ancient drama. But the Greek argument which is prefixed to this play reports that “they blame Euripides because he did not maintain Medea’s rôle but allowed her to burst into tears as she plotted against Jason and his second wife.” Again, so excellent a critic as Aristotle cites the title rôle in Euripides’ Iphigenia at Aulis as an example of inconsistency,[341] inasmuch as the Iphigenia who pleads for her life at vss. 1211 ff. in no wise resembles her later self, who willingly approaches the altar. To modern feeling, since the change is psychologically possible and is plausibly motived by the sudden realization that her death can serve her country, it seems entirely unobjectionable. But these two passages and the usual practice of the Greek stage reveal a discrepancy between the ancient and the modern points of view. The simplicity of character-drawing which resulted from Greek methods is strikingly described, in a different connection, by Mr. Cornford: