STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. INORGANIC ROLES IN ROMAN COMEDY By Henry W. Prescott In introductory studies I have endeavored to apply a mild corrective to the views of Leo and his followers who, believing that Euripidean tragedy seriously affected the form and content of Hellenistic comedy, have been led by their opinion to differ- entiate as Roman many inartistic features of Latin comedy. 1 Among the inartistic features often selected by modern students of Roman plays to strengthen theories of contamination, or of Roman work- manship in general, are characters in the plays that seem to be loosely or mechanically attached to the action. So, for example, the roles of Ergasilus in the Captivi, of Lurcio in the Miles, of Syn- cerastus in the Poenulus, of Callipho and Charinus in the Pseudolus, of Gripus in the Rudens, of Gelasimus in the Stichus, have been used to confirm notions that these plays either include inlays of Roman material or are weakened in structure by poor dovetail- ing of material from two or more Greek plays. In my own opinion the method of segregating inartistic features in a given play as proof of Roman intervention is distinctly dangerous. Any inar- tistic feature may be the common property of all or many plays, not the peculiar possession of one or several comedies, and we may not safely estimate the value, for purposes of proving Roman craftsmanship, of a single inartistic feature in a given play until we have considered in the large the manifestation in all the twenty- six Roman plays of that special feature. In this paper, therefore, I undertake a synthetic study of inorganic r61es. Constant com- parison with Greek tragedy, though necessarily brief and sketchy, will indicate the relation to each other of the two dramatic types. If the results do not by any means settle the historical problems involved in a consideration of the precise extent and nature of Roman contribution, they may at least provide a convenient sum- mary of one interesting aspect of dramatic technique. iCf. Cla$s. Phil., XI, 125 ff.; XII, 405 ST.; XIII, 113 ff.; XIV, 108 ff.; and P. E. Sonnenburg, Woch. f. kl. Phil., XXXIV (1917), 623 ff.; Schwering, N. Jahrbb., XXXVII (1916), 167 ff. [Classical Philology XV, July, 1920] 245 246 Henry W. Prescott Precise definition may be necessary to successful argument but it is often a hindrance to helpful exposition. I do not feel obliged to define what I mean by inorganic roles, except in a some- what negative sense. Dramatic action in Roman comedy is carried on mainly by characters who continue through the play and are closely interwoven in the web of the plot. Apart from these there are a number of other characters who may in different degrees be called inorganic. Occasionally a fairly permanent character, in spite of his permanence, may be loosely attached to the main action, as, for example, Ergasilus in the Captivi. More obviously inor- ganic r61es are filled by temporary characters, who appear for a single scene or brief sequence of scenes. But temporary char- acters are not necessarily inorganic; if, for example, a plot involv- ing financial transactions makes the appearance of a banker desirable for a single scene, such a character is not loosely or mechanically attached to the action provided his appearance is properly pre- pared for and his departure at the end of the scene made as inevi- table as possible; under these conditions a temporary role becomes essentially organic. And any dramatist has to employ a number of minor roles, messengers, servants, and what not, who will be helpful in promoting dialogue or action and in serving various economic purposes but who will not always continue through the play. In brief, inorganic roles are likely to be temporary, but characters who appear temporarily are not necessarily inorganic. And contrariwise, occasionally even a fairly permanent rdle may be inorganic. I am interested in any rdles that seem to be in any degree loosely or mechanically attached to the main action. The standpoint from which such roles can be best appreciated seems to me to be that of the dramatist. He has problems of composition with which we must enter into sympathy if we are to understand the results. These problems differ at different stages of his work. At the outset he must expound a situation; after that expository stage he must initiate, develop, and solve the complications of dramatic action. These last three stages of dramatic composition have common difficulties to a large extent; but the solution sometimes makes special demands upon the play- wright's skill. Broadly speaking, the beginning and the end of Inorganic Roles in Roman Comedy 247 a play test the dramatist's art more than the middle; to this fact the prologist, the protatic role, the god from the machine, bear ample witness. I The inorganic prologist has been discussed by many others and need not detain us. The protatic role is a familiar feature of classi- cal drama. The term is applied to a character more or less mechani- cally attached to the opening scene, or scenes, of a play to assist in the exposition, but disappearing from the action as soon as this service is performed. In the half-dozen instances of the role in Roman comedy the protatic character seldom himself narrates and expounds. In the Epidicus, however, Thesprio does communicate essential facts, and in the Mercator Acanthio, more as messenger than as protatic r61e, reports the critical news that the father has seen his son's sweetheart. In the Phormio, Hecyra, and Andria, on the other hand, the protatic r61e simply provides ears to listen to the exposition by another character in the dialogue. In the Mostellaria the conversation in which the protatic role participates neatly reveals, not only the situation, but the sharply contrasted characters of Grumio and Tranio. And in the Miles the function of the protatic flatterer is to delineate the character of the soldier; the exposition is left to a subsequent prologue. But all these r61es are alike in so far as they are alien to the later action and more or less mechanically attached to advance the exposition or portray character. Thesprio, to be sure, in the Epidicus is momentarily rehabilitated later in the play (vss. 657-60), but without resum- ing an active part; on the other hand, in the contaminated exposi- tion 1 of the Andria, the protatic role is instructed to perform certain duties in the subsequent action (vss. 168-70) which are not repre- sented in the later scenes of the play. These expository r61es are filled by slaves except in the Miles and Hecyra, in which a flat- terer and courtezans, respectively, assume the part. The entrance and the exit of the protatic character are managed with varying degrees of skill, but no amount of skilful motivation can disguise the mechanical function of the rdle. 1 Cf. Donatus on And. prol. 10 and 14, and F. Scholl, "Menanders Perinthia in der Andria des Terenz," Sitzb. d. Heidelberg. Akad. (1912), No. 7, pp. 4 ff. 248 Henry W. Prescott The protatic role is substantially Greek. The technical term "protatic" is part of a Greek theory that labels the exposition "protasis"; the r61e appears, perhaps, in the extant scene of Menander's Hews; 1 and Aristophanic comedy and Greek tragedy abundantly foreshadow the protatic r61e of later comedy. The Knights, Peace, and Wasps of Aristophanes begin with a dialogue scene between two slaves, one of whom disappears from the action at the end of the scene. But Aristophanic comedy has no plot; there is a situation or theme. And in these three plays the situation is expounded in an address to the audience by one of the two slaves who is not protatic, his speech foreshadowing the later prologue of comedy. The protatic slave in Aristophanes, therefore, is hardly a device to assist in exposition, but rather a means of developing dialogue, and dialogue that covers not so much exposition as idle banter. It will be observed that in pro- moting conversation the protatic r61e in Aristophanes is similar to the corresponding slave in Roman comedy, and if there were a complicated plot to expound, doubtless his ears in Aristophanes would listen to exposition instead of to trivial jest, to which he not only listens but contributes. Greek tragedy before Euripides develops exposition in the form of either monologue or dialogue. In the monologue form the speaker is sometimes protatic in so far as he drops out of the action after his speech; so in the Agamemnon and Eumenides of Aeschylus; but as monologue his speech anticipates rather the later detachable prologue. In the dialogue exposition of the Prometheus there are characters alien to the subsequent action; the special conditions of the plot perhaps made it difficult to create organic characters who could naturally appear in the region of the Caucasus. In Sophocles the protatic role exists usually in a refined form. The priest who converses with Oedipus in the Oedipus Rex helps to reveal the situation, but he comes and goes with perfect naturalness, and he gives rather than receives information; his presence issues inevitably out of the distressing circumstances that he narrates. In the Ajax a protatic divinity gives Odysseus as well as the audi- ence valuable information. In the Antigone Ismene is not protatic; i Cf. Leo, Hermes, XLIII, 146. Inorganic Roles in Roman Comedy 249 she appears in a later scene to redeem herself; but it is interesting to observe that in the opening scene she sets in bold relief the char- acter of her stronger sister and provides ears to listen to Antigone's exposition. The servant and the nurse who participate in opening scenes of the Electra and the Trachiniae, respectively, return to the action later, as does Ismene, and both stir the other characters in the dialogue to action. Thus Sophocles hovers dangerously near the mechanism of the protatic r61e without descending to the lower grades of technique. Even in Euripides the mechanical protatic r61e is not profusely employed. Thanatos in the Alcestis is superficially protatic, but his quarrel with Apollo is a conflict of forces symbolizing the issues of the play. So, too, the reconciliation of the divine powers, Athena and Poseidon, in the Troades prepares us for the anti-Greek and pro-Trojan temper of the play. The effect in both the plays is of a supernatural and symbolic prelude to the action rather than of anything fundamentally expository. In the Andromache the maidservant conveys important information and is dispatched on an errand that promotes the action; she is, however, protatic in form. The servant in the scene following the prologue of the Hippolytus is also protatic, but his function is neither to give nor receive information but by his polite suggestion to bring out the "holier-than-thou" attitude of Hippolytus, assisting therefore in character portrayal as does the protatic role in comedy oftentimes, but emerging naturally from the retinue of huntsmen and casually conversing before his retirement into the palace. Even in the one case in Euripides in which the mechanism is most obvious, the protatic r61e gives rather than receives information; Teucer in the Helena, an exile from his native land, ostensibly seeks Theonoe to inquire from her the route to Cyprus, but his dialogue with Helen is a pathetic revelation to her of the ruin wrought by her wraith. The prominence of the hero in legend makes it harder for the modern reader to suffer his complete removal from the action, but the kind of information he has to give, and the necessity that it be authentic, mildly justify the choice of the character. Yet with all these apologies the mechanism is patent, and nothing saves him from being an exact prototype of the comic r61e except his 250 Henry W. Prescott function of giving information and his social status, superior as it is to that of the slaves who usually operate in comedy. In the brief page of his Plautinische Forschungen 1 which Leo devotes to the protatic r61e he finds in Teucer the only distinct foreshadowing of the technique of later comedy, though he admits that Aristophanes and tragedy faintly point the way. I should myself decline to argue much regarding the relation of Greek comedy to tragedy from the single phenomenon of the protatic r61e when it is only one of many inorganic roles that should be considered in their entirety. But in passing I discover nothing in the evidence that would prevent a view that the protatic role is a device of a rather crude sort which dramatists, comic or tragic, are likely to employ, quite independently of one another, and the crudity is mitigated or emphasized in proportion to the difficulties of exposi- tion and the skill of the dramatist. The conditions of the plot in Euripides' Helena make the exposition far from easy and the protatic role correspondingly mechanical. And the general com- plexity of comic plots of intrigue makes the range of the mechani- cal protatic role wider in comedy than in tragedy. II However closely akin Teucer in the Helena may be to the pro- tatic roles in comedy, tragedy parts company with comedy in the mechanism of exposition beyond the opening scene. Even beyond the first scene comedy does not hesitate to employ mechanical and temporary r61es for expounding situation and character and for initiating action, which Euripidean tragedy finds no occasion to use. The unconventional procedure of comedy is clearly illustrated in the Curculio. Here Palinurus is prominent for the whole of the first act and the first scene of the second act, only to disappear completely as soon as the parasite, Curculio, enters the action. In these four scenes at the beginning of the play Palinurus' r61e is that of confidential servant, a role which in other plots is likely to develop into that of arch-intriguer, or else in some lesser capacity to continue through the play. But his actual function is simply to * PL Forsch.*, pp. 243-44; cf. also Legrand, Daos, pp. 523 ff. Inorganic Roles in Roman Comedy 251 listen to long exposition and to participate in preliminary action. The mechanical nature of his role is made unpleasantly prominent in so far as the situation makes it clear that he is being informed of matters of which he is already pretty fully cognizant. In brief, Palinurus is a protatic role but with his passive part extended over the loose exposition of four scenes. 1 Euripides may similarly eliminate a character from the later action of the play, but hardly in so clumsy a fashion. Electra's bourgeois husband in the Electra, after speaking the prologue, hospitably receiving Orestes, and fetching the old servant of Aga- memnon, disappears from the action. But this peasant husband is distinctly contributing both to the action and to the pathos of the situation; his removal is justified by the circumstances of the story, and furthermore, though the poet does withdraw him as an active character, he is careful to restore him in an allusive reference at the end of the play (vs. 1311). In connection with Electra's husband and Palinurus it is interesting to observe the rather loose organic connection of Megaronides with the action of the Trinummus. In monologue and dialogue Megaronides, a self-appointed castigator, reproaches another old man, Callicles, in the two opening scenes of the play for supposed misdeeds; through the responses brought out by this castigation the situation is expounded at the beginning of the play. Megaronides might easily have been dropped at the end of the second scene at the conclusion of his main function as expository role. But he becomes a confidant and assistant (vss. 189-90), and when the critical matter of the dowry is under consideration Callicles resorts to him for advice (vs. 614), and in a subsequent dialogue scene (vss. 729 ff.) Mega- ronides offers definite suggestions toward the plan for securing the dowry. Thus he is active in only one scene beyond the expo- sition, but the dramatist neatly restores him at the end of the play, as Euripides does the peasant husband, by an allusive reference in verse 1147 which gives him full credit for his share in the opera- tions. If he were a servos rather than a senex, 2 we should expect ' Cf. Leo, PL ForschS, pp. 196-97. 2 The uxor in Menander's Perinthia could hardly have been a protatic r≤ cf. Donatus on And. prol. 14. 252 Henry W. Prescott him to disappear at the end of the expository scenes and to assume only a protatic role; and as it is, he is rather loosely attached to the plot. Most of the roles loosely attached to the exposition and initial action are slaves, 1 and as such, like the same characters in other parts of the plays, they excite little attention or adverse criticism. But such servants, especially the female servants, antillae, are often more mechanically used near the beginning of the play than else- where, and a few of them are worth noting as illustrations of the loose touch-and-go method of comedy in contrast with the stricter organism of tragedy. Nobody objects to the anus custos ianitrix, the bibulous Una of the Curculio, though she appears in only the second and third scenes of the play and is purely picturesque and ornamental; she does not expound the plot or listen to expo- sition, but provokes interest and amusement by the revelation of her own character, and suggests the entourage of Planesium. The author of the Greek original was not sophisticated, as his later introduction of the choragus attests. More often the inor- ganic servant near the beginning of the play contributes to the revelation of another's character. The Aulularia is a loosely woven play; 2 the miser and his pot of gold, the betrayal of the miser's daughter, are the two themes of the play; these are developed into a complicated situation by the proposal of the betrayer's uncle to marry the girl in ignorance of her relations with his nephew. Such a situation is too slight to provide very much incident, and the dramatist, particularly at the beginning of his play, is hard pressed to develop action ; he is driven to emphasis upon character, and in order to interest the audience in Euclio he employs the servant Staphyla for two scenes at the beginning of the play and for the third scene of the second act, briefly recurring to her in the sixth scene of the second act. In the dialogue of these four scenes in the first two acts the miser's character is portrayed through the conversation with the servant, and thereafter she disappears from the action without our missing her. Similarly, in a succes- sion of scenes that immediately follow in the same play, episodic 1 Daos, in Menander's Epilreponies, plays an important role in the arbitra- tion scene near the beginning of the play, and does not appear later. 2 Bonnet (Milanges L. Havet, pp. 17 ff.) suspects that the play is contaminated. Inorganic R6les in Roman Comedy 253 cooks are introduced mainly to elaborate further the character of Euclio, although they are helpful also in filling the time of Euclio's absence from the stage. Here the free employment of temporary and loosely attached r61es is largely due to the poverty of incident in the presuppositions of the plot. The Mostellaria, on the con- trary, is a play of action in which the complications are very neatly and compactly organized. But the exposition of the Mostellaria is very elaborate, constituting the whole of the first act, and includes a full portrayal of the character of four roles as well as a full expo- sition of the general situation. 1 In the course of this portrayal of character the dramatist brings on a servant, Scapha, in the third scene whose sole function is, through dialogue with her young mistress, to reveal by contrast the character of the mistress; having assisted thus in character portrayal Scapha disappears from the action completely. Her removal is made less conspicuous in this instance by the further fact that all the characters in these four expository scenes, whether prominent or insignificant, are with- drawn from the action of the subsequent scenes with the exception of Tranio and Callidamates; but this withdrawal is neatly made inevitable by the conditions of the plot. Slightly different in purpose is the use of Sophoclidisca, an ancilla, and Paegnium, a puer delicatus, in the second act of the Persa. Paegnium is rehabilitated in the last scene of the play, and to that extent is less casual than Sophoclidisca. In the second act, however, both are dispatched on very trivial errands for no other apparent purpose than to provide amusement in a brief series of dialogue scenes. These scenes seem to furnish relief of two sorts; as mainly song scenes they vary the form of delivery at this point; as a comic interlude they relieve the seriousness of the general situation. But the action is not advanced by their lively banter. The whole of this second act, in fact, with the exception of Saga- ristio's expository monody at verses 251 ff., is mere padding for musical and comic effects. With these loosely attached roles in the exposition and initial action it is interesting to compare a similar temporary servant r61e in a scene that is admitted to be the result of contamination. 1 Cf. Class. Phil., XI, 140. 254 Henky W. Prescott The prologue of Terence's Adelphoe states that the action in the beginning of the play in which a courtezan is stolen from a leno is taken from Diphilus' Sunapothneskontes and inserted by Terence in the action of Menander's Adelphoi. The precise limits of the material taken from Diphilus are a matter of dispute. 1 There is no doubt, however, that the first scene of the second act is essen- tially from Diphilus. Sannio, the leno, in this scene is active also in the three subsequent scenes of the same act; whether or not these later scenes were in Menander's play, it seems probable that a leno appeared in Menander's play, and in that case he was a tem- porary role but hardly inorganic. The slave Parmeno, on the other hand, who in the first scene administers punishment to the leno but speaks only two words, appears only in this scene; he is a slave of the household of Micio, and we should expect him to be active in behalf of Aeschinus and Ctesipho, as the slave Syrus actually is, in the rest of the play. It is, therefore, not unlikely that Parmeno is the slave who was active throughout the whole of Diphilus' play, while Syrus had the same function in Menander's play. And the fact that the temporary role of Parmeno is the result of contamination is manifested by his lack, in the present text of the Latin play, of any definite function such as is performed by the other servant r61es which we have just surveyed, and also by the circumstance that Syrus seems to duplicate the role of Par- meno for no justifiable reason. After this general consideration we are in a better position to estimate the value of inorganic r61es in the preliminary action adduced by modern scholars as evidence of contamination or of Roman modification of any sort. The Pseudolus contains two such r61es in the first two acts, and the scenes in which they appear have been used to support theories that the Latin play is contami- nated. 2 What I may say of these two roles does not tend to demol- ish modern theories but only to establish just what degree of strength the two r61es contribute to the modern speculation. Both are free citizens, not slaves, and as such they may properly excite 1 For details cf. Kauer's edition of the Adelph., p. 14, n. 1. 2 Cf. Bierma, Qu. de PI. Pseud., Groningen, 1897; Leo, Nachr. d. gdtt. Ges. (1903), pp. 347 ff.; Karsten, Mnemos., XXXI (1903), 130 ff.; Schmitt, De Pseud. PI. exemplo attico, Strassburg, 1909. Inorganic R6les in Roman Comedy 255 suspicion; for in the main the loosely attached roles in the pre- liminary action of comedy are slaves. But both have a definite function to perform, and to that extent they are above suspicion. In the fourth scene of the second act the arch-intriguer has arranged his plans. For their execution he requires a sum of money and an impostor. These two means of assistance cannot be furnished by the young man whom the intriguer is assisting because the conditions of the plot make it necessary that the young man should be without funds. It is true, however, that the young man might have obtained such assistance from a friend without the friend's appearing on the stage. But the dramatist has chosen to bring into the dramatic action this friend, Charinus, and from him the five minas and the clever impostor are obtained in the course of dialogue. Having provided the sinews of war Charinus drops out of the action. Such trivial preliminaries to the carrying out of the intrigue we should prefer to have dismissed in narrative of off-stage action, but the dramatist has presented it all to us in dramatic dialogue. Such a scene does not strengthen appreciably a theory of contamination, although I should admit the possibility of Roman handiwork of a different sort; for it is conceivable that the Greek original contained a monologue by the young lover, Calidorus, in which he narrated the fact that he had obtained the money and the impostor from his friend; Plautus may have been averse to this narrative, in monologue, of off-stage action and substituted dramatic dialogue; this change in the original would exactly correspond to the procedure of Terence at Eunuchus 539 ff., as we learn from Donatus' commentary, resulting in a temporary mechanical role like that of Antipho in the Eunuchus. 1 The other instance in the Pseudolus is of a different sort. In the fifth scene of the first act the arch-intriguer confronts his victim, and states plainly to his victim the double-barreled trick which he purposes to play. The dupe is an old man of stern and irascible disposition. The dramatist adds to the scene a third character, Callipho, another old man, but of a sweet and charitable nature, and the contrast in character is clearly one of the reasons, if not the main reason, for introducing him in this single scene. He 1 Cf. below, p. 272. 256 Henry W. Prescott serves also, however, as witness to the contract between the arch- intriguer and the dupe, and in verses 547 ff. he promises to stay in town and witness the course of the intrigue, and if necessary assure the fulfilment of the contract. In spite of this definite state- ment he never reappears in the later action. Certainly this specific promise to remain as witness of the later proceedings seems rather superfluous, and one may recall the similar superfluity in the open- ing scene of the Andria, in which Sosia is instructed to perform certain duties which are not realized in the subsequent action, remembering also that Sosia was not in Menander's Andria. But in so far as Callipho serves to throw into sharp relief the character of the other old man in the scene, he is performing the function which, for example, Scapha serves in the Mostellaria, and as a mani- festation of interest in character treatment for its own sake he is a feature of Greek technique rather than of Roman fumbling in contamination. It is, however, quite possible that he comes from one of two Greek plays in which he played more than a merely temporary r61e. Ill Callipho's function is only one instance of a general feature of Greek technique upon which I should like to enlarge at this point without regard to the part of the plot in which it appears. A recurrent phenomenon in the Roman plays is a balanced pair of characters, of the same sex or age or social status — brothers, sis- ters, old men, young men, wives, husbands, slaves. 1 These pairs are most familiar as organic r61es, each member of the pair equally active. Often they contribute to the manipulation of a double plot. Of such thoroughly organic pairs the old men and the young men of the Heautontimorumenos and the Phormio and the Adelphoe are sufficient examples. And it may be observed in passing that the balanced pairs of young men and of slaves in the Andria, though they are the result of contamination of the Andria and Perinthia of Menander, 2 can hardly serve to excite suspicion against the 1 For the contrast of character in such cases cf. Legrand, Daos, pp. 233-34, and Siess, "Ueber d. Charakterzeichnung in d. Kom. d. Terenz," Wiener Stud., XXVIII (1906), 229 ff., XXIX (1907), 81 ff. 2 This is the older view of the matter, but recently it has been seriously ques- tioned by Leo, Schanz, and Korte; for full discussion cf. Scholl, op. cit. Inorganic Roles in Roman Comedy 257 general feature. To a considerable extent the fixed stage setting of the plays contributes to the regularity of this phase of the plots; an immovable background of houses in a city street brings together neighboring families who naturally supply balanced pairs of domes- tic and servile r61es. Although in most cases both members of these pairs are equally active and organic, in some instances one member of the pair is hardly more than a foil for the other. A loosely attached member of a pair serves to set off by contrast the character of the other, who is more thoroughly an integral part of the action. Sometimes it is difficult to discover any enrichment of character treatment, and one member of a pair appears in such a case simply to increase the opportunity of dialogue, or of complex action, or of general enrichment of detail. The various degrees in which such characters are loosely attached to the action may be illustrated by a few examples. Of the two sisters in the Poenulus one is distinctly more prominent than the other, and the younger, like Ismene in Sophocles' Antigone, is hardly more than a figure to set in relief her stronger and more important sister, though the younger sister does form a link in the weak minor plot of the Latin play. Similarly in the Stichus there are two sisters and wives; these two wives and sisters bring with them two brothers and husbands, two rather than one, that the parasite of the play may be twice outwitted. But one member of each of the two pairs is distinctly in the background. With the two wives and husbands of the same play come two slaves of the two households, one of whom, Sangarinufc, is mere appendage to enrich the action of the convivial scene with which the play concludes. Only in the case of the wives does the weaker member of the pair appreciably enhance the portrayal of character. So, too, pairs of lovers bring pairs of sweethearts. The young men in the Heautontimorumenos are equally active and organic, but of the two sweethearts in the same play Antiphila appears in only a single scene and there simply to set in bold relief the mercenary courtezan, Bacchis, in contrast with her own refinement and sweetness. In the Bacchides, again, the two young men are equally active and organic. But of the two old men, their fathers, Nicobulus, against 258 Henry W. Prescott whom the main intrigue is developed, is so much more intimately connected with the main thread of action that the other old man, Philoxenus, seems at times to be little more than filling; he fills out the symmetry of two sisters, two young lovers, two old roues. In so doing he is, of course, available for economic purposes of dialogue, of incidental linking of minor chapters of action, of char- acter contrast, and of general enrichment. The two sisters, also, appear in only the opening and concluding scenes of the play, and of them one is a background figure like Anterastilis in the Poenulus. At the end of the elaborate exposition of the Mostellaria the drama- tist introduces a lyrical intermezzo to re-enforce the convivial atmosphere of the opening situation. In this convivial scene the young lover and his sweetheart, already elaborately presented in introductory scenes, are joined by two convivial chums, Callidamates and Delphium; they are a pair, matching the young hero and heroine of the piece. Callidamates appears later in the play, for comic effects in the scene immediately following and as a mechanical agent in the solution of difficulties at the end. But Delphium, except for slight activity in the next scene, concludes her role in this short convivial scene. Her function is simply to assist in the humorous effects of the drunken revel, to fill up the stage, and perhaps to contribute to the volume and variety of song. In the peculiar structure of the Mostellaria even the hero and heroine are completely removed from the action after the exposition, so that we accept very easily the removal of Delphium, a weak member of the pair that balances the more prominent couple of lovers. By grouping together these examples of a general feature of the technique of comedy I have intended to indicate the loose touch-and-go method of the literary type in respect to the various degrees of stability revealed in pairs and balanced pairs of char- acters. Whatever starting-point one may find for such procedure in an occasional case like that of Antigone and Ismene in Sophocles is, I think, soon lost sight of in the informal way in which comedy secures various effects by the loose attachment to the plot of at least one member of a pair. Among the economic purposes served by the device character treatment is not the least important, and the strength of Callipho in the Pseudolus as a link in the chain Inorganic Roles in Roman Comedy 259 of evidence which binds together the theory of contamination is weakened, though not destroyed, by considering him in this con- nection. But the important point is that in the use of loosely attached r61es to promote the portrayal of character 1 the Latin plays are only appropriating or continuing the technique of their Greek originals. IV In leaving the exposition and initial action of the plays and passing to the main action some general introductory statements may prevent misunderstanding. In this part of the drama par- ticularly, be it tragedy or comedy, temporary characters are not necessarily inorganic. I assume that nobody is disturbed if the prophet Teiresias comes and goes in a single scene of the Oedipus Rex, having as prophet performed his function, especially as the audience has been duly prepared for his appearance. Nor should Pheres' temporary activity in a single scene of the Alcestis of Euripi- des exemplify the use of inorganic r61es; for, again, earlier action prepares us for his coming, his behavior admirably contributes to the portrayal of Alcestis and Admetus by way of contrast, and his permanent disappearance is an inevitable result of his conduct. Even in the earliest form of tragedy, in which we might look for loose structure, temporary r61es are few in number and rarely mechanical. Io's appearance in the Prometheus of Aeschylus is somewhat opportune; as a mythical globe-trotter she may turn up anywhere, and as a fellow-sufferer with Prometheus at the hands of Zeus she appears appropriately if somewhat mechanically. Tra- gedy, of course, operates freely with servants and similar minor characters. Appearing infrequently and for short intervals they serve economic purposes, running errands, bringing news, in various ways linking minor paragraphs of the action. Such minor r61es may at any moment be mechanically employed; obviously so in the case of the servant in Euripides' Hecuba, who is sent for water to wash the body of Polyxena and opportunely discovers the corpse of Polydorus. In this instance we are immediately aware that the poet has somewhat arbitrarily bound together two unrelated chap- ters of the action which he unwisely chose to unite in a single play. 1 Cf . above, p. 247, for the protatic rdle as instrumental in portraying character. 260 Henry W. Prescott But even Euripides, in spite of a multitude of servants, heralds, messengers, is seldom very mechanical in his technique, although the tragic messenger is often arbitrarily employed. Euripides does, however, indulge in a few episodic r61es and scenes for some- what spectacular effect. Special conditions render them good of their sort, and they are as closely knit to the main action as such scenes admit: so, for instance, the spectacular appearance of Iris and Madness in the Hercules Furens, the Phrygian who escapes from the general slaughter in the Orestes, the Evadne and Iphis scene in the Supplices, which illustrates by a concrete example the general pathos of the larger situation. In these cases some may agree with the author of the hypothesis to the Phoenissae, who commenting on the teichoscopy near the beginning of the play asserts that it is not essential to the action. In general the organic steadiness of Greek tragedy is remarkable. And as a part of its general stability temporary r61es are very few, and seldom mechani- cally employed. The Latin comedies, on the contrary, use a very large number of temporary r61es; most of these r61es are organic; a few, however, are strikingly mechanical. The profusion of temporary r61es may be suggested by a brief survey of the material, with passing com- ment on those that are in various degrees mechanical. Comedy presents an array of servants, servi and ancillae. Even those who are active for only a small part of one scene, like Cro- cotium in the Stichus, do not disturb the most sophisticated; Dromo is slightly more active in the Heautontimorumenos; Cyamus is active in a single scene of the Truculentus; Syra in the Mercator passes on and off the stage in the fourth act; Dorias and Pythias, on the other hand, are almost permanent r61es in the Eunuchus. Seldom do such servants contribute much to the main action, though Mysis in the Andria is manipulated by the arch-intriguer into promoting his trick, and Halisca in the Cistellaria in her brief existence manages to lose opportunely the casket of trinkets, to this extent being distinctly a bit of the dramatist's wirepulling. We are prepared for Stephanium in the Stichus, but she is loosely appended to enliven the convivial party of slaves in the last act. The case of Sosia in Hecyra 415 ff. is unique; this slave suddenly Inorganic Roles in Roman Comedy 261 appears without warning, and disappears after fifteen verses of dialogue with Parmeno in which he discourses briefly on the incon- veniences of a sea voyage. Donatus, 1 sensitive to the irrelevance of the character and his utterances, remarks that your true poet considers, not only the plot, but the necessity of entertaining the audience, and he dignifies this part of the scene as a