CHAPTER III

THE INFLUENCE OF ACTORS[271]

The dithyramb and the comus, together with their derivatives, early tragedy and early comedy, were entirely choral. Actors were first developed in tragedy (see pp. 16 and 48, n. 1, above). Inasmuch as the early dithyramb and early tragedy were devoted to the worship of Dionysus and since their choreutae were his attendant sprites (satyrs or sileni), it followed that their songs would mostly take the form of prayers addressed to him, hymns in his honor, or odes descriptive of his adventures, sufferings, etc. A lyric duet between the coryphaeus and the other choreutae was also possible. Such performances bore much the same relationship to later tragedy that the modern oratorio bears to a sacred opera. That is to say, the choreutae were not differentiated in character, and there was no dramatic impersonation (μίμησις); despite their costumes the chorus sang as human worshipers of Dionysus, not in accordance with their character as sileni. From the duet between the coryphaeus and the other choreutae it was only a step, but a highly important one, no longer to think of the coryphaeus as one silenus among his fellows but as Dionysus himself in the midst of his followers, and then to set him off by himself as an actor in contradistinction to the choreutae and their (new) coryphaeus. This innovation was the work of Thespis, and however long the name “tragedy” may already have been applied to the previous performances this step marked the first beginning of tragedy in the modern sense (see p. 16 f., above). Now that the new actor had to impersonate Dionysus, the necessity rested likewise upon the sileni in the chorus to live up to their own, previously neglected, character. It was not long until by a change of mask and costume the actor was enabled to represent other personages as well as Dionysus himself. This practice made possible a much more involved type of drama than the limited resources would at first glance seem to permit.

[Illustration: FIG. 66

IVORY STATUETTE OF A TRAGIC ACTOR

See p. 162, n. 1]

Aeschylus’ earliest extant play, the _Suppliants_, belongs to the two-actor period, but employs the second actor so sparingly as to afford a very good idea of the possibilities of the one-actor play. Omitting the choral odes, the action runs as follows: The fifty daughters of Danaus (the chorus) seek sanctuary near Argos to escape the unwelcome suit of their cousins. At vs. 176 Danaus begins to admonish his daughters and a dialogue (vss. 204-33) ensues between them. At vs. 234 the king of Argos enters and engages with the chorus in a dialogue and a lyric duet (vss. 234-417). During this scene Danaus is present, silent, inactive, and all but unnoticed; cf. vs. 318. Of course in a one-actor play this character must have been removed so that the single actor might reappear as the king. But that could easily have been managed and would affect the present piece in no essential way. After an ode the dialogue between the king and the chorus is resumed (vss. 438-523), broken in upon only by a brief conversation between the king and Danaus (vss. 480-503). The former instructs Danaus how to supplicate the citizens in the town and, upon the latter’s request for protection, orders attendants to accompany him. Here for the first time are the two actors simultaneously employed, but their words serve no more important purpose than to motivate the exit of one of them. At vs. 523 the king likewise withdraws. At vs. 600 Danaus reappears and with but a slight interruption on the part of his daughters (vss. 602-4) informs them that the Argives have decided to shield them (vss. 600 f., 605-24). At vs. 710 Danaus descries the suitors’ fleet in the distance and declares, “I will return with helpers and defenders” (vs. 726). Nevertheless, the scene is continued until vs. 775, when Danaus departs to spread the alarm, incidentally releasing this actor to play the part of the suitors’ herald. At vs. 836 the herald enters and to the accompaniment of a lyric duet between himself and the chorus tries to drag the Danaids away. At vs. 907 this attempt at violence is brought to a standstill by the king’s return. The following altercation between the herald and the king (vss. 907-53) provides the only bit of genuine dramatic conflict, visually represented, in the play and the only instance of both actors being fully made use of together. In a one-actor play such a passage would have been impossible but could have been presented indirectly by means of a messenger’s narrative. At vs. 953 the herald withdraws, discomforted, and the king turns to the chorus (vss. 954-65). In reply the chorus ask that their father be returned to them (vss. 966 ff.). The interval having been sufficient to enable the actor to shift from the mask and costume of the herald to those of Danaus, the latter re-enters at vs. 980 and converses with his daughters until the final ode. Of all the extant plays of Aeschylus the _Suppliants_ probably makes the slightest appeal to the modern student. Its principal value for us lies in the fact that it could readily be revamped for presentation by one actor and in the light which it thus sheds upon the character of one-actor drama.

Several times in this play, as appears from the foregoing outline, an actor participates in a dialogue with the chorus. It was not the practice for the choral part in such dialogues to be spoken by all the choreutae in unison, but by the chorus leader alone. Thus, though a sharp distinction was drawn between actors and chorus, the former being furnished by the state and the latter by private means (cf. pp. 270 f., below), yet the coryphaeus served as a bond of connection between the two. We have seen how the first actor was developed from the chorus leader; doubtless the successive additions to the number of actors were suggested in each case by the advantages arising from this quasi-histrionic function of the coryphaeus. Thus in addition to the regular actors, at each stage of development the tragic poet always had at his disposal also one quasi-actor for carrying on his dialogues. And the comic poet always had two such quasi-actors, since the leaders of the two semi-choruses could be used in this way (see p. 44, above). In the one-actor period this quasi-histrionic function of the coryphaeus resulted in a convention which continued long after the necessity for it had passed away. It is obvious that at that juncture the single actor could converse with no one but the chorus. This practice became so stereotyped that in the two-actor period whenever a character came into the presence of the chorus and another actor he directed his remarks to the chorus before turning to the other character. Of course oftentimes this was the natural thing to do. But the force of tradition is seen in the fact that the principle was sometimes observed under unfavorable conditions. Thus, as we have already observed, in the _Suppliants_ the king enters at vs. 234 and at once begins a dialogue with the chorus, ignoring their father until vs. 480. Greek respect for age and partiality for the masculine sex make this arrangement in a Greek play very unnatural. Again, in the _Persians_ a messenger from Greece ignores his queen (vss. 249 ff.) and reports the Persian disaster to the chorus of elders. Not until vs. 290 does Atossa address him, and in typical Greek fashion Aeschylus strives to make her words gloss over the unreality of his characters’ compliance with convention. “For a long time have I kept silence,” she begins, “dumbfounded by catastrophe. This ill exceeds my power to tell or ask our woes.” The same convention persisted even into the three-actor period. Clytemnestra’s husband has been gone ten years or more, yet she must excuse herself to the chorus (Aeschylus’ _Agamemnon_, vss. 855-78) before greeting her lord (see p. 155, above). Another instance occurs in Euripides’ _Children of Heracles_, vss. 120 ff. Moreover the coryphaeus sometimes exercises an important influence upon the plot. For example, in Aeschylus’ _Libation-Bearers_, vss. 766 ff., it is the coryphaeus who induces the servant to alter the wording of the summons with which she is sent to Aegisthus. By this device he comes unescorted and falls an easy victim to the conspirators.

In view of the normal employment of the coryphaeus as a quasi-actor, Aeschylus took an easy and obvious step, or rather half-step, in advance when he introduced the second actor. We have seen that the _deuteragonist_ was already made use of, though sparingly, in the _Suppliants_. Also the _Persians_, the _Seven against Thebes_ (except possibly the closing scene; see p. 175, below), and the _Prometheus Bound_ require but two actors for presentation. The great advantage accruing from the second actor is manifest. Instead of being compelled to resort to a messenger’s report of an altercation or dialogue between two personages, the playwright was now enabled to bring the characters face to face in person upon his stage. On the other hand, so limited a number of actors often seriously embarrassed the dramatist in the economy of his play. Perhaps the best example of this is afforded by Aeschylus’ _Prometheus_. In the opening scene Cratos and Bia (Strength and Force) drag Prometheus to a remote spot in Scythia and Hephaestus nails him to a crag. How can these four characters be presented by two actors? In the first place Bia has no speaking part, and mutes were freely employed in addition to the regular actors. In the second place Prometheus was represented by a wooden figure. This explains how it was possible for a nail to be driven right through his breast (vss. 64 f.). It explains also why so great emphasis is laid upon the fastening process; first the hands are pinned down (vs. 55), then the arms (vs. 60), the breast (vs. 65) and sides (vs. 71), and finally the legs (vs. 74). Thus the immobility and lifelessness of the supposed Prometheus are accounted for. Neither Hephaestus’ sympathy nor Cratos’ insults elicit a single word of reply from his lips. Although this silence arises naturally from the Titan’s unyielding disposition, yet the real reason lies in the use of a dummy. At vs. 81 Hephaestus retires, and after six lines of further insults Cratos follows him. A slight pause would naturally ensue, so that Prometheus might be sure that his enemy had passed beyond the sound of his voice. These intervals enabled the former actor to take his place at some crack or opening behind the lay figure and break Prometheus’ speechlessness (vs. 88). The other actor reappears in a succession of rôles throughout the play, as Oceanus (vs. 284), Io (vs. 561), and Hermes (vs. 944); but these shifts were easily managed.

Soon after Sophocles’ first appearance (468 B.C. or possibly 471 B.C.)[272] he introduced the third actor. First of all this innovation permitted a larger number of characters to be presented. In Aeschylus’ two-actor plays the characters number three in the _Suppliants_ in addition to the chorus and coryphaeus, four in the _Persians_, six in the _Prometheus_, and five in _Seven against Thebes_. In the three-actor plays Aeschylus’ characters range from five to seven, Sophocles’ from five to nine, and Euripides’ from seven to eleven, except that Euripides’ satyr-play, the _Cyclops_, has but three characters. Secondly, a third actor allowed greater flexibility in handling entrances and exits. An artificial pause, more or less improbably motived, to enable an actor to change his mask and costume before appearing in another rôle would now be less frequently required (see further, p. 231, below). Thirdly, it allowed three personages to appear side by side in the same scene, whereby in turn a certain aesthetic effect became possible. I refer to the varied emotions which one actor’s statements or conduct sometimes produce in two other characters. An excellent illustration is afforded by the scene with the Corinthian messenger in Sophocles’ _Oedipus the King_, vss. 924 ff. As the awful conviction is brought home to Jocaste that Oedipus is her son as well as her husband, she rushes from the stage to hang herself; but Oedipus, on the contrary, still lacking the fatal clue, becomes elated at the prospect of discovering his parents’ identity. Similarly in the same playwright’s _Electra_, vss. 660 ff., the false report of Orestes’ death cheers his mother with the assurance that her murder of Agamemnon must now remain unavenged, but plunges Electra into the desperation of despair. Such situations would have been impossible in the two-actor drama. Finally, the introduction of a third actor contributed to the decay of the chorus. We have already noted in the last chapter how the importance of the chorus steadily declined, especially in comedy. But this change was quantitative as well as qualitative. In the prehistrionic period the chorus and its coryphaeus, from the nature of the case, monopolized every line. After Thespis had brought in the first actor the chorus yielded but a small place to its rival. Even in the two-actor period in our earliest extant play, the _Suppliants_, the chorus sang five hundred and sixty-five verses out of a total of a thousand and seventy-four, and in addition to this the coryphaeus spoke ninety verses. In six of Aeschylus’ seven extant pieces the choral element varies from three-fifths to about one-half of the whole play. The _Prometheus_, for special reasons, is exceptional, the fraction being only one-sixth. The effect of the third actor is seen in the fact that in Sophocles the proportion varies from one-fourth to one-seventh and in Euripides from one-fourth to one-eighth.

The question naturally arises, Why were the Greek dramatists so slow in increasing the number of actors? This was due partly to a paucity of histrionic talent and partly to difficulty in mastering the dramatic technique of the dialogue.

In the dithyramb and the prehistrionic drama the poet was his own coryphaeus. Accordingly when Thespis introduced the first actor he served in that capacity himself, appointing another as coryphaeus. So did Phrynichus, Aeschylus, and the other dramatists of that period. Since there were then no retired actors and no opportunity to serve an apprenticeship, it is obvious that these early poets had to teach themselves how to act. At this stage it was not possible for anyone except a playwright to become an actor, and actors must have been correspondingly scarce. The situation improved somewhat after Aeschylus introduced the second actor, for though the poets still carried the major rôles it now became possible for men with natural histrionic ability to develop it and gain experience in minor parts. By the time of Sophocles, actors had become so plentiful, relatively speaking, that he could increase the number employed by each poet from two to three and could retire from personal participation in the public presentation of his works. His weak voice is said to have been responsible for this second innovation; but he occasionally appeared in scenes where this weakness was no great hindrance, e.g., as a harp player in _Thamyris_ and as an expert ball player in _Nausicaa_. By 449 B.C. the profession was so large and its standing so well recognized that a contest of tragic actors was made an annual event in the program of the City Dionysia. This course of development reveals one reason for the long duration of the one- and two-actor stages in Greek drama.

We shall now pass to the second reason. In the prehistrionic period a series of lyric questions and answers between chorus and coryphaeus was the nearest approach to a dialogue that was possible (see p. 10, above). With the invention of the first actor this interplay of question and answer, still lyrical in form, could be carried on by the actor and the chorus (including the coryphaeus). Such a duet, which came to be known as a _commus_, continued in use, especially for dirges, as long as the chorus lasted. Side by side with this, however, there quickly developed a non-lyric interchange of spoken lines between actor and coryphaeus. But not until the second actor was added did true dialogue in the modern sense become possible. Yet the poets could not at once make full use of even these simple resources. Our analysis of Aeschylus’ _Suppliants_ (pp. 163 f., above) shows that in two instances Danaus stood silent and unaddressed during a conversation between the other actor and the coryphaeus. Moreover, priority of usage constrained the playwrights to give the actor-coryphaeus dialogue precedence over actor-actor dialogue (cf. pp. 165 f., above). They seemed unable to weld the two types together with a technique which would employ all three persons at once. In the three-actor period the embarrassment of riches made their helplessness the more striking. “A” might engage in a dialogue with “B” while “C” remained inactive; then with “C” while “B” was silent; and finally “B” and “C” might converse, with “A” remaining passive. Often the transitions are marked or the longer speeches set off by a few more or less perfunctory verses (usually two) spoken by the coryphaeus. The type is not frequently worked out as completely as I have just indicated, but the principle is illustrated on a lesser scale in almost every play. Compare, for example, Euripides’ _Helen_, vss. 1186-1300, and _Andromache_, vss. 547-766. Such an arrangement, needless to say, falls far short of a genuine trialogue or tetralogue. Yet we must not be unfair in condemning this practice. The Greek poets were feeling their way and could not immediately attain to every refinement. Even in Shakespeare and the modern drama, despite centuries of continuous experimentation and the numerous examples of superior technique, the tandem arrangement of dialogue is still not uncommon.

A half-step in advance consisted in the silent actor interrupting the dialogue with some electrifying utterance. For example, in Aeschylus’ _Libation-Bearers_ (458 B.C.), Clytemnestra’s appeal to Orestes on the score of her motherhood stays his hand in the very act of murdering her, and he weakly turns to his trusted friend, Pylades, for guidance. The latter’s ringing response,

Wilt thou abjure half Loxias’ behest, The word of Pytho, and thy sacred troth? Hold all the world thy foe rather than Heaven

[vss. 900-903, Warr’s translation],

is as effective as if uttered by the god in person, and urges Orestes on to the deadly deed. These are the only words that Pylades utters in the whole tragedy. In another play belonging to the same trilogy, the _Eumenides_, Aeschylus rose to the full possibilities of his histrionic resources—Orestes, the coryphaeus, Apollo, and Athena all participating in the conversation between vss. 746 and 753. Similarly, in Sophocles’ _Oedipus at Colonus_, Antigone, Oedipus, Ismene, and the coryphaeus all speak between vss. 494 and 506, and in Euripides’ _Suppliants_ the herald, the coryphaeus, Adrastus, and Theseus divide four lines among them (vss. 510-13). But after all, such instances are comparatively rare and seldom extend over a very long passage.

In contradistinction to tragic practice Aristophanes in the last quarter of the fifth century employed not merely three but occasionally even four comic actors in ensemble scenes. For example, in the _Lysistrata_, vss. 78-246, Calonice, Myrrhina, Lysistrata, and Lampito engage in a running fire of conversation quite in the modern manner. Again, in the _Frogs_, vss. 1411 ff., Dionysus, Aeschylus, Euripides, and Pluto all have speaking parts, although the last two do not address one another. In the same play (vs. 555) Dionysus utters three words while three other

## participants in the dialogue are present. Under similar circumstances

Pseudartabus interposes two verses (100 and 104) in the _Acharnians_, and Triballus parts of five verses (1615, 1628 f., and 1678 f.) in the _Birds_. In these passages the comic coryphaei have no speaking parts. Trialogues are not so rare in Old Comedy as to justify an enumeration of the instances, and they are sometimes embellished by the participation of the coryphaei. Nevertheless, the old tandem arrangement is still the more common one when three characters are present.

We thus pass from one problem to another: Why this disparity between the technique of tragedy and comedy? Must we suppose that the comic dramatists were more clever artists than their tragic confrères? By no means. Comedy was more mobile and reacted more quickly to the actual conditions of contemporaneous life; tragedy was more conventional, never could free itself entirely from the power of tradition, and could only slowly modify that tradition. The situation is clearly revealed in the field of meter. In the iambic trimeters written by Aeschylus a trisyllabic substitution (tribrach, anapaest, or dactyl) for the pure disyllabic iambus occurs only once in about twenty-five verses. In the earliest plays of Euripides such resolutions appear once in sixteen verses but gradually increase to a maximum of one in every alternate verse.[273] On the contrary, in the comedies of Aristophanes they are found in almost every line. Now we are not to suppose that Euripides required a lifetime in order to learn how to use resolutions with freedom or that he was never able to gain the facility of Aristophanes. Nor are we to suppose that Sophocles, whose iambics resemble those of Aeschylus, was never able to master this expedient. In both cases we see merely the power which convention and tradition exercised over tragedy. And the same influences made themselves felt in the comparatively archaic technique of tragic dialogue and tended to keep the tragic playwrights from making full use of their resources.

But were the resources of the tragic writers as great as those of the comedians? We have seen how the first, second, and third actors were added to Greek tragedy. Is there reason to believe that the tragedians of Athens ever followed the comedians in employing a larger number? Until recently a negative reply to this has been accepted without serious question, but in 1908 Professor Rees challenged the tradition. Three years later the old view was defended by Dr. Kaffenberger. Although neither has been able fully to establish his contentions, yet the discussion has helped to clear the air, defined the issues more sharply, and really settled certain important points. For one thing, since 1844 it has generally been taken for granted that three actors were the maximum for Old Comedy as well as for tragedy. But the passages just cited from Aristophanes would seem to be decisive against this view, and all the objections to the presentation of Greek tragedy by only three actors apply with still greater force to Old Comedy. Even Dr. Kaffenberger (_op. cit._, pp. 9 f.) accepts this conclusion, and it is an invaluable result of Professor Rees’s investigations that he has banished this phase of the subject from the field of controversy. Moreover, they are both agreed[274] that a fourth actor seems sometimes to be required also for New Comedy. It must be added, however, that Dr. Graf (_op. cit._, pp. 29 ff.) dissents. But in any case the question has been restricted, so far as the fifth century is concerned, to the practice in tragedy.

It can be said at once that if we are willing to grant that the Greeks made use of certain desperate expedients it is physically possible to stage all the extant tragedies with three actors. But these expedients are so offensive to modern feeling as to be tolerable only as a last resort. It will be best to begin at a point where comparative agreement is possible, viz., with Aeschylus’ earlier plays, which nearly everyone would admit were intended for two actors alone. Do they reveal any indication of this limitation?

In the analysis of Aeschylus’ _Suppliants_ on p. 164, the reader will remember that Danaus, having declared “I will return with helpers and defenders,” took his departure at vs. 775; after an ode, the suitors’ herald arrived on the scene (vs. 836) but was balked by the entrance of the Argive king (vs. 907). One would surely expect Danaus to accompany the king, but as a matter of fact he does not reappear until vs. 980. The reason for this is plain—Danaus and the herald are played by the same actor, and consequently the former can return only after the latter’s departure at vs. 953. Moreover, Aeschylus sought to gloss over the blemish by having Danaus refer in advance to the possibility of his being slow in spreading the alarm (vs. 730) and by having the chorus request the king to send their father back to them (vss. 968 ff.), as if his absence had been perfectly natural. This incident teaches us four things: (1) A single actor could carry several rôles; the simplicity and sameness of ancient costumes and the ease of slipping them off and on, together with the use of masks by the actors, made this practice more feasible than it is with us. Overzealous classicists have not merely asked us to tolerate this practice but even to admire its results. Thus, when one character returns to report the death of another the spectators are supposed to have been doubly moved if they could penetrate the messenger’s disguise and from the identity of stature, build, and voice recognize the ghost, as it were, of the departed visibly before them (!).[275] (2) This practice oftentimes necessitated the arbitrary withdrawal of a character from the scene of action and his enforced absence when he would naturally be present. (3) By inventing an inner reason for this the poet strove to conceal or gloss over his yielding to external need. (4) The intervals between the withdrawal of Danaus and the entrance of the herald (vss. 776-836) and vice versa (vss. 953-80) afford an inkling as to the length of time required for such shifts in rôles.

Further information is derived from Aeschylus’ _Prometheus Bound_ (see pp. 166 f., above). (5) Supernumeraries may be employed for silent parts, e.g., that of Bia. (6) A part may be divided between a lay figure and an actor, as in the case of Prometheus himself. From the nature of things, this expedient would not be frequently employed; but an analogous device (6_a_) is common, viz., to give the silent portions of a rôle to a mute and the speaking portions to an actor. (7) The stubborn silence of the mutes and supernumeraries employed according to principles (5) and (6_a_) is sometimes extremely embarrassing and difficult to motivate. (4_a_) The interval required for a “lightning” change from one character to another was much shorter than the _Suppliants_ led us to suppose. Six verses and a slight pause in the action enabled the actor impersonating Hephaestus to withdraw by the side entrance after vs. 81 and to get in position to speak from behind the wooden figure of Prometheus at vs. 88. This conclusion is confirmed by certain evidence in Plautus’ translation of Greek comedies, which indicates that about thirteen lines would suffice.[276]

Still other principles are derivable from Aeschylus’ _Persians_. The ghost of Darius having requested his widow to meet their son Xerxes with a change of raiment, Atossa replies (vss. 849 ff.): “I shall endeavor to meet my son ... and,” turning to the chorus, “if he comes hither before me, do you comfort him and escort him to his palace.” These words are clearly intended to prepare us for her failure to appear in the dénouement, and in fact she does not appear. But since one of the two actors is disengaged in the final scene, at first glance there seems to be no external reason for her absence. It is evident that Aeschylus valued the parts of Atossa and Xerxes so highly that he wanted them both played by the better of his two actors, the _protagonist_. If Atossa had appeared with her son, she must have been impersonated by a different actor than in the opening scenes. The poet preferred to sacrifice verisimilitude somewhat rather than to “split” Atossa’s rôle in this fashion. Hence, we must conclude (8) that at any cost star parts were reserved for the leading actor, (9) that split rôles were to be avoided, and (10) that sometimes for purely technical reasons the dramatist would unnaturally keep a character off the stage entirely in certain scenes.

If we could be sure that the final scene of Aeschylus’ _Seven against Thebes_ is genuine, it would be possible to deduce a final principle. The main support for the charge of interpolation is that this scene in a two-actor play apparently requires three actors. From vs. 961 to vs. 1004 Antigone and Ismene engage in a lyric duet; at vs. 1005 a herald enters and converses with Antigone. From this scene, which I am inclined to accept as genuine (see p. 283, below), we must concede either that a supernumerary could occasionally bear a brief singing (or speaking) part or preferably that the herald, standing in the side entrance concealed from the spectators and already dressed for his own rôle, sang Ismene’s share of the duet while a mute went through the dumb show of her part before the audience; at the conclusion of the duet he promptly appeared _in propria persona_. Though the latter alternative is offensive to present-day taste, it is not unparalleled in the annals of the modern stage.[277] In any case one of these alternatives is the last principle (11) to be drawn from the two-actor drama.

Now these eleven principles are so manifestly operative in the other Greek tragedies as to raise an irresistible presumption that some restriction (to three or at most to four actors) applied also to them. It would obviously be out of place to pass every play in review here; I must content myself with a few typical illustrations and then consider the crucial cases.

In order to avenge his daughter, Menelaus is on the point of murdering her rival (Andromache) and the latter’s son when he is interrupted by the arrival of Peleus, Hermione’s father-in-law. There is no reason why Menelaus should fear the old man’s blusterings; nevertheless he suddenly leaves Hermione in the lurch and takes his departure with the words:

Now, seeing that my leisure serveth not, Home will I go; for not from Sparta far Some certain town there is, our friend, time was, But now our foe: against her will I march, Leading mine host, and bow her ’neath my sway. Soon as things there be ordered to my mind, I will return, etc.

[Euripides _Andromache_, vss. 732 ff., Way’s translation]

Surely no excuse was ever less convincing than this! No wonder Professor Verrall’s ingenuity has built up a whole reinterpretation of the play around it.[278] The real reason for the sudden leave-taking is only too apparent—Orestes is presently to make his appearance (vs. 881) and Menelaus’ actor is required for his rôle. This exemplifies principles (1), (2), and (3).

Again, in Sophocles’ _Maidens of Trachis_, Lichas, Deianira, and a messenger are on the scene when Deianira spies Iole in a throng of captives and questions her (vss. 307 ff.). Iole makes no reply whatsoever. Lichas explains her refusal to answer by stating that from grief and weeping she has not uttered a word since leaving her fatherland (vss. 322 ff.). Since the three actors are already occupied in this

## scene it is evident that Iole is played by a mute and cannot speak. This

illustrates principles (5) and (7).

Still again, up to vs. 1245 of Euripides’ _Orestes_, when he enters the palace, Pylades speaks freely. At vs. 1554 Menelaus, Orestes, Hermione, and Pylades enter the scene. The last two are now played by mutes, the third actor appearing as Apollo at vs. 1625. Orestes threatens to kill Hermione; and after vainly striving to deter him Menelaus turns to Pylades with the query (vs. 1591): “Do you, also, share in this murder, Pylades?” What is a mute to do under such circumstances? Orestes relieved the situation by saying: “His silence gives consent; my word will suffice.” There can be no doubt that the playwright intended Menelaus’ question to create the illusion that Pylades could have spoken had he so desired, principles (6_a_) and (7).

Euripides avoided an awkward silence of this sort in the _Ion_ by leaving Xuthus unrepresented in the final scene, where the three actors speak in other rôles. Xuthus takes his final departure at vs. 675, intending to celebrate for his new-found son a public feast from which the host himself is most strangely absent. The poet prepares us in advance for this contingency by means of Xuthus’ words to his son, as reported by a servant at vss. 1130 ff.: “If I tarry in sacrificing to the Birth-gods,” a thin pretext, “place the feast before the friends assembled there,” principles (1), (2), (3), and (10).

Finally, for the presentation of his _Phoenician Maids_, Euripides must have had a leading actor of great musical attainments. For such a performer the rôles of Jocaste and Antigone were especially adapted, and he seems to have played them both, principle (8). The piece opens with a soliloquy by Jocaste, who withdraws at vs. 87. Immediately a servant appears on the palace roof and tells Antigone to tarry upon the stairs until he can assure himself that there is no one near to see her and to spread scandalous reports of her indiscretion. Thus, Antigone’s appearance is delayed for fifteen verses (vss. 88-102), which is sufficient to enable Jocaste’s actor to shift to the new rôle, principle (4_a_). The protagonist continues to play both parts without difficulty, except at vss. 1264 ff. Here Jocaste summons her daughter from the palace and both are present during vss. 1270-82, the latter speaking some six verses. Obviously Antigone’s lines in this brief scene must have been delivered by one of the subordinate players, though such splitting of a rôle violates Aeschylean practice, see principle (9). Perhaps the procedure in this case was condoned by the fact that Antigone’s part previously and (for the most part) subsequently was entirely lyric, while her few words here are in plain iambics. The difference between the singing and the speaking voice would help to conceal the temporary substitution of another actor. It is true that by assigning Jocaste’s and Antigone’s rôles to different actors throughout it is possible to distribute the parts in this play among three actors without any difficulty whatever. But this would require us to ignore the peculiar technique of the opening scenes, the true inwardness of which was recognized by ancient commentators.[279]

These examples are by no means exhaustive, but it is high time that we turn to the passages which are of crucial importance to the three-actor theory. In Aeschylus’ _Libation-Bearers_ a servant has just informed Clytemnestra that her paramour is slain, and she cries out: “Let some one quickly give me an ax to slay a man withal” (vs. 889). We are to suppose that the slave at once makes his exit to comply with her command. She speaks two lines more and Orestes enters. They divide seven more lines between them, and Orestes’ purpose is beginning to waver when he catches sight of Pylades entering and asks: “Pylades, what shall I do? Shrink from killing my mother?” Pylades’ electrifying response has already been quoted (vss. 900-902; see p. 170, above). Here we have four speaking characters between vss. 886 and 900 and consequently four actors, unless the servant can be transformed into Pylades within the space of nine lines, vss. 891-99. This would be a “lightning” change indeed (4_a_), and it is not surprising that it has been challenged. Yet the ancient scholiast accepts it and I do not believe we are warranted in pronouncing it impossible, especially since the shift is merely from one male character to another.

Another sort of difficulty is presented by Euripides’ _Andromache_. Menelaus, Andromache, and her son, Molossus, all have speaking (or singing) parts just before the entrance of Peleus at vs. 547. Since none of the earlier speakers has withdrawn and since Peleus at once begins to talk, it would seem at first glance that we had four actors indisputably before us. Not so, answer the defenders of the traditional view, for it is significant that Molossus becomes utterly dumb after Peleus enters. Therefore we are asked to believe that Molossus was played by a mute throughout, and the actor who is presently to appear as Peleus delivered from behind the scenes the words which belong to Molossus, the mute furnishing only the gestures. We have already found support for this kind of thing in a suspected scene of Aeschylus’ _Seven against Thebes_, principle (11), second alternative (pp. 175 f.). But we are asked to go further and believe that this was always the practice when children seemed to sing or speak upon the Greek stage;[280] and in confirmation of this it is pointed out that whenever children have a part, as in Euripides’ _Alcestis_, vss. 393 ff. and _Medea_, vss. 1271 ff., one of the actors is always off the scene and available for this purpose. The most difficult example of this problem has recently come to light in the fragments of Euripides’ _Hypsipyle_, vss. 1579 ff.[281] The heroine and Amphiaraus converse from the beginning of the fragment to vs. 1589, where the latter makes his exit. Two lines of farewell (vss. 1590 f.) are addressed to him and are assigned by the papyrus to “the children of Hypsipyle.” Moreover, they are of such a nature that one line must have been spoken by each of the two youths. Next, _one_ of them converses with his mother until Thoas, who also has a speaking part, appears at vs. 1632. Here, then, if the children’s parts are taken by actors we have four actors required in two successive scenes. The only alternative lies in supposing that mutes impersonated the boys and that Thoas’ actor, already dressed for his introit at vs. 1632, spoke their lines from behind the scenes. This would include twelve lines for one youth and one line, _in a different voice_, for the other.

[Illustration: FIG. 67.—Distribution of Rôles to Actors in Sophocles’ _Oedipus at Colonus_]

But the most intractable play of all is Sophocles’ _Oedipus at Colonus_. Antigone and Oedipus are on the stage continuously for the first eight hundred and forty-seven verses (the latter until vs. 1555), while the third actor appears successively as a stranger, Ismene, Theseus, and Creon (Fig. 67). So far there is no difficulty; but at this point Creon hopes to bring Oedipus to time by announcing that his guards have already seized Ismene (off-scene) and by having them now drag Antigone away. Creon threatens to carry off Oedipus as well, but at vs. 887 Theseus reappears and prevents further outrage. Note, however, that if only three actors were available Theseus must now be impersonated by Antigone’s actor, whereas previously he was represented by the actor who is now playing Creon’s part. Such splitting of a rôle is directly contrary to Aeschylean practice, principle (9), and has not in this instance the justification which Euripides had for splitting Antigone’s

## part in the _Phoenician Maids_ (p. 178, above). For Theseus’ second actor

## participates in the dialogue more extensively than did hers and his

lines are prose throughout, while hers were entirely prose for one actor and (almost) entirely lyric for the other. But there are still other obstacles ahead. At vs. 1043 Creon and Theseus withdraw; after a choral ode Antigone, Theseus, and Ismene rejoin Oedipus (vs. 1099). Inasmuch as Ismene now has no speaking part she is evidently played by a mute, principle (6_a_). Presumably the other two are represented by the same actors as at the beginning, although this second transfer in Theseus’ rôle doubles the chances of the audience noticing the shift. The only alternative, however, is to split also Antigone’s rôle at this point. Theseus retires at vs. 1210 and reappears at vs. 1500, his actor having impersonated Polynices in the interval (vss. 1254-1446). At vs. 1555 all the characters exeunt. In the final act a messenger is on the stage from vs. 1578 to vs. 1669. Since Antigone and Ismene enter immediately thereafter (vs. 1670), it is necessary to suppose that they are played by the same actors as at the beginning and that Oedipus has become the messenger. At vs. 1751 Theseus makes his final entrance, represented this time by Oedipus’ actor, so that this important rôle is played in turn by each of the three actors! This means splitting Theseus’ rôle twice. It is also possible to split his rôle and Ismene’s (or Antigone’s) once each, or to split his rôle once and to have the final actor in this part sing from behind the scenes the few words which fall to Ismene just before Theseus’ last entrance, principle (11). On the other hand, though a fourth actor would obviate all these difficulties we should then have no explanation for the complicated system of entrances and exits and for the strange silence of Ismene during vss. 1099-1555, especially during vss. 1457-99 (see p. 187, below).

I do not consider it warrantable to draw a categorical conclusion from the data considered in the last fifteen paragraphs. But in my opinion the technique of almost every tragedy is explicable only on the assumption that the regular actors were restricted to three; and, as I stated at the beginning, it is physically possible to stage every play with that number. In the case of a few pieces, however, this limitation imposes practices which so outrage the modern aesthetic sense that we instinctively long for some manner of escape. According to late and unreliable evidence an extra performer was called a _parachoregema_. This name would indicate that he was an extra expense to the man who financed each poet’s plays (the choregus, see pp. 186 and 270 f., below), and consequently that his employment would be determined by the wealth or liberality of the latter. But whether it was in fact possible for the tragic playwrights occasionally to have the services of such an extra, and, if so, under what conditions and how, are questions which in the present state of our knowledge can receive only hypothetical answers. It must be recognized, however, that the paucity of actors in the early days resulted, as we have just seen, in conventions of staging which perhaps were afterward accepted as part of the tradition, however unnecessary they may in the meanwhile have become. The technique of composition also makes it clear in my opinion that extra performers, if such were in fact engaged, were not on a par with the other three nor employed freely throughout the whole play but merely recited or sang a very few lines at those crises in the dramatic economy which were occasioned by the limitation in the number of regular actors.

We have now discovered why the dialogue technique of tragedy was more restricted than that of comedy, but there still remains a further question. Why was the number of actors in tragedy usually or always restricted to three, while four actors were not uncommon in comedy? So long as the poets did their own acting, there was no occasion for the state to interfere in the selection of actors. And this situation would naturally continue for some time after the plays were presented largely or wholly by actors alone—the poets would still have the matter in their charge. In fact there is no reason to suppose that the state interposed its authority before the establishment of the contest for tragic actors at the City Dionysia in 449 B.C. This supposition affords the best explanation for certain ancient notices. For example, Aeschylus is said to have used Cleander as his first actor and afterward to have associated Mynniscus with him, and Sophocles to have employed Tlepolemus continuously. Whatever truth or error may lie back of these statements they imply that in the first half of the fifth century the choice of actors rested solely with the poets. The same implication is inherent in the fact that the second and third actors were introduced by Aeschylus and Sophocles respectively. The poets must have made these additions upon their own initiative. For the state could not have shown partiality by providing Sophocles, for example, with more actors than were furnished the other dramatists in the same contest; and if they were all alike given an increased number, there would be no reason for crediting any one of them with the innovation. The state must have assumed supervision of the histrionic features of the dramatic contests at the same time that it established a prize for actors, viz., in 449 B.C. And since the tragedies of this period were presented by three actors, this number became crystallized, and so was never thereafter, so far as the state was concerned, exceeded in tragedy. Tragedies were added to the Lenaean program and a prize for tragic actors established for that festival simultaneously, about 433 B.C. Naturally the conventional number of tragic actors would be transferred from the older contest to the newer. In comedy, however, the development and tradition were entirely different (see pp. 52 f., above). Primitive comedies in Attica were performed by a double chorus of choreutae, who constituted an undifferentiated crowd and assumed no individual rôles, but sang (or spoke) singly, antiphonally, or in unison. Shortly before 450 B.C. regular actors were introduced in contradistinction to the choreutae; and Cratinus, imitating contemporaneous tragedy, set their number at three. Yet the choreutae did not for a long time entirely give up their old license and self-assertiveness. Consequently, it is not surprising that the number of performers did not remain at the tragic norm. The fact that a contest of comic actors was not established at the Lenaea until about 442 B.C. (at the City Dionysia not until about 325 B.C.) allowed a slight interval for this reaction to assert itself before usage became legalized. Such, then, are the reasons for the number of actors being less restricted in comedy than in tragedy.

For about a century, beginning with 449 B.C., the state annually engaged three tragic protagonists to be assigned by lot to the three poets who were about to compete with plays. Each protagonist seems to have hired his own subordinate actors (deuteragonist and tritagonist) and with their assistance presented all the plays (at the City Dionysia three tragedies and one satyric drama) which his poet had composed for the occasion. The victorious actor in each year’s contest was automatically entitled to appear the following year. The other two protagonists were perhaps selected by means of a preliminary contest, such as is mentioned for comic actors on the last day of the Anthesteria. These regulations applied, _mutatis mutandis_, also to the contest of comic actors and to the tragic and comic contests at the Lenaea. Thus at the Lenaea of 418 B.C. Callippides acted in the two tragedies of Callistratus, and Lysicrates in the other dramatist’s two plays. And it should be noted that, whereas Callippides won the prize for acting, Callistratus was defeated in the competition of tragedies. This must have been a point of considerable difficulty, for an actor’s chances must have been greatly hampered by his being required to present a poor series of plays; and a poet, likewise, must have suffered by reason of an inferior presentation of his dramas. But sometime in the fourth century, when the playwrights were no longer required to write satyr-plays (see p. 199, below), a more equitable system was introduced. Each of the protagonists in turn now acted one of the three tragedies of each poet, the histrionic talent at the disposal of each dramatist being thus made exactly the same. For example, at the City Dionysia of 341 B.C. (Fig. 76) Astydamas was the victorious playwright; his _Achilles_ was played by Thettalus, his _Athamas_ by Neoptolemus, and his _Antigone_ by Athenodorus. The same actors likewise presented the three tragedies of Evaretus and those of the third dramatist. On this occasion Neoptolemus won; a year later, under similar conditions, he was defeated by Thettalus.

We have seen how slow was the rise of actors into a profession distinct from the poets. At a later time, however, they were strongly organized into guilds under the name of “Dionysiac artists” (οἱ ἀμφὶ τὸν Διόνυσον τεχνῖται). Their strongest “union” (κοινόν or σύνοδος) was centered at Athens and it was also the earliest (fourth century B.C.). Others were situated at Thebes, Argos, Teos, Ptolemais, Cyprus, and in all parts of the Greek-speaking world. Now already in the fifth century traveling troupes had presented at the country festivals plays which had won popular acclaim in Athens. For economic reasons it was to the advantage both of the players who had to divide their emoluments and of the communities which hired them to make these traveling companies as small as possible and consequently to restrict their repertoire to plays capable of being performed by a minimum of actors. With the organization of guilds the presentation of dramas “in the provinces” or even at important festivals would be taken over by them; and the same economic causes as before would operate to restrict the number of players in a company. There is reason to believe that a normal troupe in the time of the _technitae_ consisted of three actors.[282] Inscriptions for the Soteric festival at Delphi for the years 272-269 B.C. inclusive contain the names of ten companies of tragic actors and twelve of comic actors. These performers belonged to the Athenian guild and in every case there are three names to a company. There is no reason to doubt that this number was customary also in the wandering troupes of the pre-technitae period. Some maintain that already in the fifth century a fourth actor was called a parachoregema, as being an extra burden upon the choregus (cf. p. 182, above). But Professor Rees has made it seem very probable that the term took its rise in the time of the technitae. For in later usage _choregein_ (χορηγεῖν) in most cases no longer meant “to defray the expense of the chorus,” “to act as choregus,” but simply “to furnish” without any reference to the choregic system at all. Parachoregema, therefore, would signify “that which is furnished in supplement,” “an extra.” In other words, if the officials of a city contracted with the union for one or more troupes for a dramatic festival they would be provided with three-actor companies; but if they desired to witness some four-actor play or to avoid the infelicities arising from the splitting or ill-assorted doubling of rôles (see pp. 191 f., below) they might _at extra expense_ secure a parachoregema in the form of a fourth actor and so gratify their wishes. According to either interpretation, therefore, the term may refer, _inter alia_, to a fourth actor, but there is a wide difference as to the theory of the circumstances and situation which produced this meaning.

Since our extant plays belong exclusively to the fifth and fourth centuries, the size of the troupes furnished by the guilds could have exerted no influence upon them. But it is quite possible that the dramatists of later times deliberately adapted their technique to the needs of subsequent presentation by such companies. For example, the number of characters who can have a speaking part in a dialogue naturally cannot exceed the number of actors at the poet’s disposal. Whatever may have been the situation previously, in the technitae period this would be three. Therefore if the technitae did not give rise to, they at least fixed the so-called aesthetic law that if a fourth character is present at a conversation between three others he must keep silent. This rule is expressed by Horace[283] in the words: “Let no fourth character strive to speak,” and it is often mentioned by writers of the Alexandrian and Roman periods. The scholiasts belong to this time and their comments frequently reveal an attempt to foist the aesthetic law upon the fifth-century dramas. The difficulty which the fifth-century writers encountered in mastering even the three-part dialogue (see p. 170) lends to such an attempt a misleading facility. In tragedy the normal restriction of actors to three makes the task especially easy, but even here the law is only superficially observed. For the coryphaeus often participates so freely in a conversation between actors (see pp. 164 f. and 169 f., above) that only by courtesy can it be called a three-part dialogue. In Seneca’s Roman tragedies, on the contrary, the coryphaeus never speaks if more than one actor is present.[284] Now Professor Rees would trace the aesthetic law back to fifth-century times, but Dr. Kaffenberger (_op. cit._, pp. 22 f.) rightly demurs. He points out that in Sophocles’ _Oedipus at Colonus_, vss. 1099-1555, Oedipus, Antigone, and Ismene are continuously present but that Ismene says never a word. What is the cause of this silence? During vss. 1099-1210 and vss. 1500-1555 Theseus is also present and during vss. 1249-1446 Polynices is present. In these scenes, therefore, it is possible to explain Ismene’s silence on the basis of the aesthetic law. But during vss. 1447-99 Oedipus and his two daughters are left alone, and Ismene still remains silent. Consequently the aesthetic explanation breaks down at this point and we must stand by our earlier conclusion (see pp. 181 f., above) that throughout these scenes Ismene is impersonated by a mute. Moreover, since Oedipus forbids his daughters sharing his final moments with him, why does the poet not let him take leave of them on the stage instead of resorting to a messenger’s narrative (cf. vss. 1611 ff.)? The answer is obvious. In such a situation Ismene simply _must_ have spoken and this a mute could not have done for her. Moreover, there is no aesthetic reason why the law should not be as binding in comedy as in tragedy. Nevertheless, fifth-century comedies indisputably violated it and possibly fourth-century comedies did also (see pp. 171-73, above). Therefore, if tragedy was more scrupulous it must have been because its actors were less numerous. But in truth it was not until the period of the technitae and their three-actor troupes that a hard-and-fast rule was established. Notwithstanding, the grammarians as a result of their closet study of Attic drama seized upon the observance of the law in fifth-century tragedy and usually in New Comedy, which was greatly influenced by Euripides, as a justification for tracing the practice back to an earlier origin. Except in one scene Seneca always observed the law.[285] But when Plautus and Terence attempted to transplant New Comedy to Italian soil, they encountered a difficulty. It was the use of masks which enabled the Greek playwrights to shift their actors from one rôle to another with lightning speed. But masks are said not to have been employed on the Roman stage until the next century. Therefore, even if the Greek comedies had been translated without modification it would have been quite impossible to present them at Rome with only three or four (maskless) actors. Accordingly, Plautus and Terence seem to have employed five or six performers and occasionally even more, and then proceeded to make further use of them so as to gratify the Roman desire for spectacular effects. By combining Greek plays into one Latin version (by “contaminating” them, to use the technical term) and by altering them freely they produced many scenes in which four or five persons participate in the same dialogue.

The fact that women’s parts in Elizabethan drama were played by boys has been used to explain the fondness of Elizabethan heroines for masquerading in masculine attire. Now the Greek theater, likewise, knew no actresses—all parts, regardless of sex, were presented by men. Can any effect of this practice be traced in the extant plays? In the first place Greek drama also was not unacquainted with the spectacle of masculine performers impersonating women who were disguised as men; cf. the rôle of Mnesilochus in Aristophanes’ _Women at the Thesmophoria_, and the chorus and several characters in the same author’s _Women in Council_. But in the Greek theater this occurrence was too rare to be significant. Secondly, it has frequently been observed that the heroines of Greek tragedy are as a rule lacking in feminine tenderness and diffidence and are prone to such masculine traits as boldness, initiative, and self-reliance. On the other hand the women who have speaking parts in comedy are usually either impaired in reputation or disagreeable in character—courtesans, ravished maidens, shrews, scolds, jealous wives, intriguing mothers-in-law, etc.[286] Now these facts are doubtless the resultant of many factors. For example, tragedy has little direct use for the modest violet type of woman, and the sharp demarcation between dramatic genres (see p. 201, below) tended to prevent their indirect employment in scenes meant merely to relieve the tragic intensity of the main plot. Likewise, social conditions must have had a great deal to do with the exclusion of women of unblemished reputation and attractive years from the comic stage (see pp. 277-79, below). Nevertheless when all is said I consider it quite possible that the representation of women by men actors was partially responsible for such a choice and for the delineation of female rôles. At least male performers must have found such types of women much easier to impersonate. Finally, if children were represented only in pantomime and their words spoken by a grown actor from behind the scenes (see pp. 179 f., above) we can understand why girls never have a speaking part and one reason why the words put in boys’ mouths are often too old for them. A competent critic has declared: “Euripides’ children do not sing what is appropriate to children in the circumstances supposed but what the poet felt for the children and for the situations. In particular the song of the boy over the dead body of his mother in the _Alcestis_ is one of his grossest errors in delineation.”[287] This situation, also, is capable of several explanations, but who will deny that the practice of having children’s parts declaimed by adults belongs among them?

In France the court compelled actors to furnish amusement and the church damned them for complying. In Rome the actors were slaves or freedmen and belonged to the dregs of society. Only in Greece did no stigma rest upon the histrionic profession. As we have seen (pp. 131 f., above) the actors were active participants in a religious service and during the festival performances their persons were quasi-sacrosanct. As such, they were entitled to and received the highest respect, and their occupation was considered an honorable one. Consequently, they were often the confidants and associates of royalty and wielded no mean influence in the politics of their native lands. In particular as they traveled from court to court they often acted as intermediaries in diplomatic negotiations. Thus Aeschines, an ex-actor, was almost as influential in the Athenian faction which favored the Macedonians as was Demosthenes in that which opposed them. And though the latter in his speeches indulged in frequent sneers at Aeschines’ theatrical career, this was not on account of his profession per se but because Demosthenes claimed he had been a failure at it. Aeschines and Aristodemus, another actor, twice went as ambassadors from Athens to Philip, king of Macedonia, with whom the latter was _persona gratissima_. Thettalus was an especial favorite of Alexander the Great, who sent him as an emissary to arrange his marriage with a Carian satrap’s daughter. When Thettalus was defeated by Athenodorus at Tyre in 332 B.C. Alexander said that he would rather have lost a part of his kingdom than to have seen Thettalus defeated. These men were contemporaries of Aristotle, who declared in his _Rhetoric_ that in his day actors counted for more in the dramatic contests than the poets.[288] The huge fees that they received are often mentioned. In view of all this it is not surprising that they arrogated to themselves many liberties. Aristotle states that Theodorus always insisted upon being the first actor to appear in a play, doubtless on a principle analagous to that which Mr. William Archer[289] mentions: “Where it is desired to give to one character a special prominence and predominance, it ought, if possible, to be the first figure on which the eye of the audience falls.... The solitary entrance of Richard III throws his figure into a relief which could by no other means have been attained.” This anecdote may mean merely that Theodorus assumed the rôle of the first character, however insignificant, in order to appear first upon the scene. But some have thought that he actually had the plays modified so that the character which he was to enact might appear first. Even upon the first hypothesis, however, slight alterations might sometimes have been necessary. For example, if he wished to impersonate Antigone in such a play as Euripides’ _Phoenician Maids_ and if no passage were provided like vss. 88-102 to enable the actor to shift from Jocaste, who opens the tragedy, to Antigone (see pp. 177 f., above), then perhaps the simplest solution would have been to interpolate a few such lines for this purpose. But however this may have been in Theodorus’ case there can be little doubt that the actors did sometimes take such liberties with their dramatic vehicles. To correct this abuse Lycurgus, who was finance minister of Athens in the last third of the fourth century B.C. and “completed” the theater (see p. 69, above), is said to have had state copies of old plays provided from which the actors were not allowed to deviate; and Lycon was fined ten talents, which Alexander paid, for having interpolated one line in a comedy.

Naturally most actors were peculiarly adapted to certain types of characters. Thus Nicostratus was most successful as a messenger, Theodorus in female rôles, etc. The interesting significance of the parts borne by Apollogenes, an actor of the third century, has only recently been recognized. At Argos he impersonated Heracles and Alexander, at Delphi, Heracles and Antaeus, at Dodona, Achilles, etc., in addition to winning a victory in boxing at Alexandria. Evidently this actor was a pugilist for whom rôles and plays were carefully chosen which would display his physique and strength to the best advantage. Now these special predilections and accomplishments of the actors, as well as their physical qualities, must often have run afoul of the constant doubling and the occasional splitting of rôles as required by the restricted number of players. Professor Rees makes good use of such points in arguing against the three-actor limitation in fifth-century tragedy.[290] But in such matters custom is all-important; we cannot be sure to what extent the Greeks were offended by infelicities of this nature. In my opinion such considerations are not strong enough to break down the arguments drawn from dramatic technique (see pp. 173-82, above).

I ought not to conclude this chapter without a few words concerning the manner in which act divisions arose from the alternation of choral odes and histrionic passages in ancient drama. The earliest tragedies, such as Aeschylus’ _Suppliants_ and _Persians_, began with the entrance song of the chorus, which is called the _parodus_. In later plays it was customary for one or more actors to appear before the choral parodus in a so-called _prologue_. The first instance of this which is known to us occurred in Phrynichus’ _Phoenician Women_ (476 B.C.). After the parodus came an alternation of histrionic scenes (_episodes_) and choral odes (_stasima_), concluding with a histrionic _exodus_. These are nontechnical definitions and do not cover every variation from type, but they will suffice for present purposes. Thus Aeschylus’ _Prometheus Bound_ falls into the following divisions: prologue, vss. 1-127; parodus, vss. 128-92; first episode, vss. 193-396; first stasimon, vss. 397-435; second episode, vss. 436-525; second stasimon, vss. 526-60; third episode, vss. 561-886; third stasimon, vss. 887-906; exodus, vss. 907-1093. Though the number of stasima (and of episodes) was more usually three, as in this case, there was originally no hard-and-fast rule on the subject. In several plays there were four stasima and four episodes, and in Sophocles’ _Antigone_ five of each. Therefore in a normal tragedy like the _Prometheus_ the number of histrionic divisions would be five—prologue, three episodes, and exodus. In the early plays which had no prologue the histrionic divisions fell to four—three episodes and an exodus. In several of the later plays, on the other hand, they rose to six, and in the _Antigone_ to seven. As the lack of connection between chorus and plot increased and the size and importance of choral odes diminished (see pp. 126 f., 136-49, and 168, above) there was the more excuse for ignoring the choral elements and for concentrating attention upon the histrionic divisions. The development of comedy led to similar results. The composition of an Old Comedy has already been discussed (see pp. 40 f., above). So long as the agon and the parabasis persisted, the structural differences between tragedy and comedy were unmistakable; but with the disappearance of these features early in the fourth century (see pp. 42 f., above) the assimilation of the two genres rapidly proceeded. Moreover, as the activity of the comic chorus was confined to _entr’actes_ and as their entertainment became so foreign to the plot as no longer to be written in the manuscripts but merely to be indicated by ΧΟΡΟΥ (see pp. 147 f.), this tendency to ignore the choral element in favor of the histrionic became pronounced. Now the number of histrionic divisions in Old Comedy and in New Comedy was limited to five even less frequently than in tragedy. And in either literary genre there was no more reason for such a restriction, whether on historical or technical grounds, than there would be in modern drama. In every period such a detail depends, or ought to be left to depend, entirely upon the requirements of the story chosen for dramatic presentation. Nevertheless, since the histrionic divisions in tragedy were more usually five and since comedy fell more and more under the domination of tragedy, the rigid principle was at last set up for both tragedy and comedy that each play should contain five acts, no more, no less; cf. Horace’s pronunciamento: “Let a play neither fall short of nor extend beyond a fifth act.”

It should be observed, however, that our English word “act” conveys a misleading impression in this connection. The Greek word was simply “part” (μέρος) and denoted merely a division of the play as determined by choral _divertissement_, whether written or interpolated. These “parts,” therefore, depended upon the more or less accidental and haphazard

## activity of the chorus and often two or three of them would be required

to make up an act in the modern sense. In other words the modern notion of an act as an integral part of the story, marking a definite stage in the unfolding of the plot, was for the most part yet to be developed, especially in comedy.

The leveling effect of the five-act rule is seen in the modern editions of Plautus and Terence. It is certain that neither four nor any other fixed number of pauses was employed at the premier performances of these dramatists’ works. In some cases they seem to have been given continuous representation with neither choral intermezzi nor pauses at the points where the Greek originals had had _entr’actes_. From this, however, we must not infer that Plautus and Terence did not know where the acts or the “parts” began and closed. If for no other reason, the recurrence of ΧΟΡΟΥ in at least most of the Greek comedies which they were translating and adapting would not have permitted them to be ignorant on this point, for in my opinion, so far as pauses were inserted in the Roman performances, they coincided with the corresponding points of division in the Greek plays. But by this I do not mean that the Latin divisions were always as numerous as the Greek; in my judgment, owing to contamination and other modifying influences they were uniformly fewer. Moreover, when these comedies were first published for the use of a reading public, it seems that the manuscripts contained no indication of act divisions. Within a century of Terence’s death, however, partisans of the five-act dogma were already attempting to force their Procrustean theory upon his works. A later effort of this sort is preserved to us in the commentary of Donatus (fourth century A.D.) and passed into the printed editions, with some modifications, about 1496 A.D. Likewise, the Renaissance scholars, obsessed by the tradition of what had come to be considered an inviolable law, proceeded to divide each of Plautus’ twenty plays into five acts; cf. Pius’ edition of 1500 A.D. The divisions in both poets rest upon no adequate authority and are easily shown to be incorrect. Yet, unfortunately, it is now impossible to re-establish the acts as known to their Latin authors. If we revert to the Greek terminology, however, somewhat more definite results may be obtained, though, even so, agreement is not possible in every case. Technical criteria now at our disposal would indicate that the original “parts” (μέρη) in these comedies ranged from a minimum of two or three to a maximum of seven or eight.

But Aristophanes was at the same time a dramatist contending for a prize, and had no wish to alienate the greater part of his audience.—T. G. TUCKER.

##