SOURCE: J. C. B. Lowe, "Aspects of Plautus' Originality in the Asinaria," in Classical Quarterly, Vol. 42, No. 1, 1992, pp. 152-75.

[In the following essay, Lowe compares Plautus's Asinaria to its Greek model Onagos and identifies several aspects of Plautus's comedy which are perhaps Plautine innovations rather than further derivations from Greek materials.]

Introduction

The problem of Plautus' originality

That the palliatae of Plautus and Terence, besides purporting to depict Greek life, were in general adaptations of Greek plays has always been known. Statements in the prologues of the Latin plays and by other ancient authors left no room for doubt about this, while allowing the possibility of some exceptions.¹ The question of the relationship of the Latin plays to their Greek models was first seriously addressed in the nineteenth century, mainly by German scholars, under the stimulus of Romantic criticism which attached paramount importance to originality in art.² Since then the question has been constantly debated, often with acrimony, and to this day very different answers to it continue to be given. Yet the question is obviously important, both for those who would measure the artistic achievement of the Latin dramatists and for these who would use the plays to document aspects of Greek or Roman life. It is not disputed that Plautus' plays contain many Roman allusions and Latin puns which cannot have been derived from any Greek model and must be attributed to the Roman adapter.³ What is disputed is whether this overt Romanization is merely a superficial veneer overlaid on fundamentally Greek structures or whether Plautus made more radical changes to the structure as well as the spirit of his models.

In the earlier part of the nineteenth century it was the originality of the Latin dramatists that was emphasized, in response to the charge that they were no more than slavish translators.⁴ Attention was drawn to the obvious Roman allusions in Plautus and to the great difference between his comedies and those of Terence. Whereas Terence was held to have followed his Greek models relatively faithfully, it was argued that Plautus must have treated his with considerable freedom. Confirmation of this was found in Terence's reference to the neglegentia of Naevius, Plautus and Ennius as a precedent for the liberties he took with his Greek models (And. 15-21). Terence admits incorporating into three of his plays material from a second Greek model (And. 9-14, H.T. 16-21, Eun. 30-3, Ad. 6-11), and makes clear that his critics objected to this on the ground that contaminari non decere fabulas. From these passages it was inferred that what Terence did was not new but had already been practised by earlier writers of palliatae, and that contaminare did not here have its normal sense of 'spoil' but a technical sense of 'combine'. If Plautus treated his models more freely than Terence, it seemed likely that he would have practised 'contamination' even more than Terence. 'Contamination' was seen as a mark of the Latin dramatists' originality; and the possibility was acknowledged that Plautus' neglegentia might also have included other kinds of change to his Greek model.

Throughout the rest of the nineteenth and into the twentieth century Plautus' plays were subjected to increasingly intensive analysis, pioneered by Ladewig, with the aim of discovering how Plautus 'contaminated'. Many inconsistencies were found in the plays and seen as indications of 'contamination', when they were not attributed to post-Plautine retractation.⁵ At the same time there developed a tendency to emphasize the Greek element in Plautus and correspondingly to devalue his originality, no doubt under the continuing influence of Romantic philhellenism. Before the first papyrus discoveries brought to light substantial fragments of Menander's plays, parallels were collected from extant Greek literature for many aspects of Plautus' subject matter, thought and dramatic technique, notably by F. Leo in the masterly chapter on 'Piautus und seine Originale' in Plautinische Forschungen.⁶ Leo recognized that not all apparently Greek elements in Plautus were necessarily derived from his Greek model (P.F.² p. 103), but in practice he was rather too ready to assume that they were. Like other German scholars of this period, Leo was grudging in his estimate of Plautus' ability to invent new dramatic action, although he paid generous tribute to his creative genius in the sphere of language and metre; he described him as essentially a translator, even if a very free translator, and his plays as inferior to their Greek models.⁷ He supposed that Plautus cut much from his models and sometimes drastically truncated a Greek plot, as in the case of the Casina (P.F.² pp. 167f.). He also attributed to Plautus the invention of numerous short passages and some rather longer ones, such as Ergasilus' auction of his logi in Stick. [Stichus] 193-235 (P.F.² p. 169). He supposed that Plautus sometimes drastically restructured his main model but only by 'contamination', now seen in a negative light as a practice adopted by playwrights incapable of original invention.⁸

Plautine scholarship became obsessed with 'contamination'; the result was a mass of theories but little agreement.⁹ It was easy to criticize many of the theories as based on flimsy arguments. Many of the supposed flaws in Plautus' plays used as arguments for 'contamination' were trivial or conventional features of ancient drama or could be shown to have some dramatic purpose. Moreover the nineteenth-century interpretation of contaminare in Terence's prologues as meaning 'combine' was refuted and the inference that Plautus had done exactly the same as Terence was questioned.¹⁰ Inevitably a reaction set in against the excesses of structural analysis and the associated 'contamination' theories, but this took different forms. Many scholars rejected the analytical method more or less entirely, regarding it as fundamentally flawed. As usual the pendulum swung too far. Almost any explanation of an inconsistency in Plautus was preferred to that of dual authorship. 'Psychological' explanations were put forward which confused drama with real life. Moreover the Unitarians divided into those who saw Plautus' plots as largely his creation and those who saw them as largely taken over from the Greek originals.

One party is inclined to believe that, although Plautus drew most of his material from Greek comedy, he used it with great freedom to produce plots which bore little resemblance to any Greek play.¹¹ This approach has flourished particularly outside Germany, in reaction to the low estimation of Plautus' originality by the German analysts. Its proponents can only support it, however, with general and inconclusive arguments, and tend to concentrate on denouncing and sometimes misrepresenting the 'contaminationists'.¹² Although they pay lip-service to the known fact that Plautus adapted Greek plays, they often ignore it in practice. The other party regards Plautus as usually faithful to his Greek models at least in the structure of his plays, and in general tends to regard as Greek anything in the Latin plays which is not obviously Roman, especially if something remotely similar can be found in Greek literature. This party naturally includes Hellenists whose primary concern is with Greek drama, but also others.¹³ They attribute the differences between Plautus and Terence to a large extent to their use of different Greek models. They hold that many of the inconsistencies in Plautus were already present in the Greek originals, some resulting from a sort of 'contamination' practised by the Greek dramatists themselves, that is the combination of heterogeneous motifs from earlier plays to make new plots.

Others, more fruitfully, have continued the analytical tradition but attempted to improve on the methods of their predecessors. A major advance was made by E. Fraenkel.¹⁴ Consciously reacting against his teacher Leo, Fraenkel set out to show that the Roman element in Plautus' plays was substantially greater than Leo had allowed.¹⁵ He started by collecting a number of figures of thought and style which he showed to be characteristic of Plautus, such as the grotesque comparison of a slave with a hero of myth or history (e.g. Bacch. [Bacchides] 925ff.) or the comic personification of an inanimate object (e.g. Stich. 191 eo … verbo lumbos diffractos velim). By a comparison of the Plautine plays known to be based on Menandrian models (Bacch., Cist. [Cistellaria], Stich.) with Terence's Menandrian adaptations and the newly discovered papyrus texts of Menander Fraenkel established a strong probability that most, at least, of these Plautine figures were his invention and not taken from his Greek models, even if there was room for argument about individual cases. In some cases supporting evidence of various kinds could be used to identify with probability the limits of a short Plautine addition, e.g. a repeated phrase at the beginning and end of the insertion.¹⁶ Whereas demonstrably Roman subject matter had previously seemed the only reliable criterion of Plautine authorship, Fraenkel was able to expand the range of criteria to include many characteristically Plautine features of thought and style. In particular he showed that many more apparently Greek motifs than Leo had supposed were probably in fact Plautus' invention, including comic allusions to Greek myth. Fraenkel further showed that these characteristically Plautine motifs were particularly associated with certain characters, especially the scheming slave, and convincingly argued that by such superficial changes, together with his metrical innovations, Plautus could substantially modify the characterization, and thus the whole emphasis of a play. Most scholars have accepted these positive conclusions of Fraenkel as largely right, even if not his interpretation of every individual passage.¹⁷ With regard to the possibility that Plautus made structural alterations to his Greek models, however, Fraenkel came to negative conclusions; he moved beyond Leo, but not very far. He supposed less 'contamination', and practically dispensed with the hypothesis of major 'contamination' involving the intricate fusion of two or more Greek plots. He showed that the inconsistencies which had given rise to theories of 'contamination' could often be explained by supposing numerous small Plautine alterations; for the rest he posited for certain plays minor 'contamination' as practised by Terence, namely the insertion of relatively limited material from another Greek play into the plot of his main model. Although, however, he credited Plautus with the invention of some new visual effects in his small-scale insertions as well as purely verbal ones,¹⁸ he denied that Plautus was capable of 'true dramatic invention and the creation of dialogue which advances the action'.¹⁹

Since Fraenkel the trend amongst analysts has been against the hypothesis of 'contamination' and towards that of more extensive Plautine invention, bringing them closer to the position of those who wish to maximize Plautus' originality. Already G. Jachmann²⁰ and Drexler²¹ supposed for the Rudens considerable structural changes by Plautus which did not involve a second Greek model. W. E. J. Kuiper²² and E. Lefevre²³ have argued for massive Plautus invention in the restructuring of some of his plots, even if many of their arguments have convinced few. Very recently a detailed analysis of the Menaechmi has led E. Stärk to the conclusion that the plot is entirely Plautus' invention.²⁴ In most at least of the places where Fraenkel supposed 'contamination' others now find Plautine invention more likely.²⁵ A. S. Gratwick supposes 'contamination' in the Poenulus in a very different sense from the traditional one, involving the borrowing from another Greek play of no more than the basic idea for an insertion.²⁶ Analysts are still prone, it is true, to going beyond the evidence and indulging in unsupported speculation of the kind that leads opponents of the method impatiently to dismiss their arguments out of hand. The discovery, however, since 1958 of the first complete play of Menander, substantial fragments of others and in particular fragments of the Dis Exapaton, the original of the Bacchides, has introduced a major new factor into the debate.

This new material has taught us a great deal about Menander's dramatic technique and provided reliable information on Plautus' methods of work. It justifies and greatly facilitates the attempt to separate Greek and Roman elements in the Latin plays. So far as they go, the Dis Exapaton fragments show Plautus treating the details of his Greek model with great freedom, but on the other hand following the Greek plot and sometimes translating very closely.²⁷ Better understanding of Menander's use of the conventions of five-act structure and the three-speaker rule provides us with a standard against which to measure the dramatic technique of the Latin plays and more objective criteria for identifying Roman originality than were previously available.²⁸ In recent years it has been possible to recognize ever more clearly certain types of change made by Plautus (and Terence) that go beyond what Fraenkel was willing to allow. Leo established the principle that it is necessary to collect groups of similar phenomena.²⁹ Fraenkel achieved this at the level of language and style, but it is now possible to progress to the level of dramatic structure and identify with considerable probability certain types of structural change made by the Latin adapters to their Greek models. One change is now well established, that Plautus and Terence, having no singing and dancing chorus at their disposal, sometimes bridged the act-divisions of their models; in itself this was a minor change, but it could involve altering characters' movements and could be linked to other changes. The Dis Exapaton fragments now document a case in which the omission of a choral entr'acte by Plautus was connected with the omission of the two flanking scenes and hence inconsistencies in Nicobulus' offstage movements.

The researches of several scholars, including the present writer, have shown it to be probable that both Plautus and Terence not infrequently inserted an extra character into a pre-existing scene of his Greek model, in Terence's case sometimes a character borrowed from another Greek play, traditionally 'contamination', but also sometimes an invented character,³⁰ sometimes a character who appeared in other scenes of the Greek original;³¹ in many cases this must be regarded as the probable explanation of the involvement of more than three speaking characters in a scene of the Latin play, contrary to Menander's practice and to Horace's dictum nee quarta loqui persona laboret (AP 192). The greater part of this article will seek to demonstrate further instances of this technique from the Asinaria, arguing that whereas Philaenium's first appearance in III 1 derives from Plautus' Greek model, her subsequent appearances in III 3 and V 2 were added by Plautus.

The Asinaria and its critics

The Asinaria has been subjected to harsh criticism, though not usually as extreme as F. Ritschl's description of it as 'devoid of any artistic merit'.³² Apart from objections to the play on moral grounds, it has been criticized especially for incoherence of plot and inconsistencies of action and characterization; and various theories have been put forward to explain these features. The play can serve as a not untypical example to illustrate some past trends in Plautine scholarship and to point the way to future progress.

The plot in outline is simple enough. An impecunious young man, Argyrippus, is in love with a meretrix, Philaenium, who returns his love (542). Their relationship is threatened when Philaenium's mercenary mother, the lena Cleareta, tired of mere promises (524-34), delivers an ultimatum; unless Argyrippus can at once produce 20 minae as the price of Philaenium's exclusive services for a year, she will find someone else who can (229-31). In the opening scene, a dialogue between Argyrippus' father Demaenetus and his slave Libanus, Demaenetus reveals that Argyrippus has sought his help and that he would like to give it but has no money of his own, being dependent on his well dowered and domineering wife Artemona, who keeps a tight control over all domestic expenditure through her steward Saurea. He therefore encourages Libanus to obtain the money for Argyrippus through some trick. The opportunity for such a trick is provided by a chance encounter of Libanus' fellow-slave Leonida with a stranger bringing 20 minae from a merchant of Pella to Saurea as payment for some Arcadian asses (333-7). Leonida impersonates Saurea and, not without difficulty, succeeds in obtaining the money from the stranger with Demaenetus' help (579-84). Meanwhile Argyrippus is in despair, having learnt that a rival, Diabolus, has promised to pay Cleareta 20 minae that very day (633-5), when the slaves bring him the money. The lovers celebrate with a party in Cleareta's house, joined by Demaenetus, who claims the right to share Philaenium with Argyrippus for one night as a reward for his help. Diabolus arrives, accompanied by his parasite, all prepared to sign a contract with Cleareta, but finds that he is too late. In his anger he sends the parasite to inform Artemona that Demaenetus is in the lena's house; and this leads to a farcical finale in which Artemona bursts in on the party and drags Demaenetus home in disgrace.

It is obvious that this plot contains numerous stock motifs, which can easily be paralleled from other plays of Greek New Comedy or its Latin adaptations, the love of a young man for a meretrix, his inability to pay the money demanded by a lena, the opposition of a parent, the rival, the slave's scheme to raise the money by a trick, the deception of a stranger by impersonation, the henpecked husband who attempts to indulge himself but is punished by his uxor dotata. More often the obstacle to a young man's amours is an unsympathetic, tightfisted father, who has to be tricked by a wily slave into providing the necessary cash; in this case the stock theme is varied in that it is the uxor dotata who is the unwilling source of funds for her son. The plot fits comfortably in the context of Greek New Comedy. It is more than adequate as the basis of a comedy and no more implausible than many other plots of Plautus, Terence and Menander. So far as the main framework of the plot is concerned there is no incoherence but a neatly integrated combination of traditional motifs. When the prologue informs us that the Latin play is adapted from Demophilus' Onagos (10f.), we are justified in taking it to be highly probable that Plautus took over from his Greek model at least the basic framework of the plot. The title Onagos, 'The ass-driver', if it presents certain problems, fits the plot well enough.³³

Many inconsistencies and loose ends have been noted in the play. They are more or less superficial, however, and do not destroy the essential unity of the plot, although they sometimes obscure it. Critics speak of the play as consisting of a series of loosely connected scenes. Yet every scene serves in some way to advance the plot; none is dispensable. A truer judgement is that there are inconsistencies in the connexions between scenes, or at least a lack of clarity about the action supposed to have taken place off stage, and that certain scenes are developed into comic set-pieces which retard the main action and by their disproportionate length tend to obscure the underlying plot. Various theories have been proposed to explain these inconsistencies, broadly speaking divided into those which find the explanation in Plautus' changes to his Greek model and those which find it in post-Plautine alterations to the text. Most often inconsistencies have been noted between I 2-3 and other scenes, and to explain these it was suggested by G. Goetz and G. Loewe that I 2-3 were 'contaminated' from another Greek play;³⁴ this theory gained the support of Leo,³⁵ but has had no recent adherents. L. Havet on the other hand argued that the cause of the inconsistencies was an error in the transmitted text of I 2-3; whereas the MSS. identify the lover who appears in these scenes as Argyrippus, Havet reattributed the rôle to Diabolus, and this change has won widespread support.³⁶ Havet also supposed the loss in transmission of a scene from the Latin play, but this theory has been generally ignored.³⁷ More extensive 'contamination' involving other scenes as well as I 2-3 has been supposed by M. Brasse,³⁸ and more recently by J. N. Hough³⁹ and G. Rambelli;⁴⁰ these theories have gained little support and their improbability has been argued by P. Ahrens,⁴¹ Munari⁴² and Bertini.⁴³ Finally, the omission by Plautus of a scene of his Greek model has been supposed by Webster.⁴⁴ The fact that all these theories have been advanced suggests that the inconsistencies are sufficiently serious to require some explanation and that they should not be brushed aside as trivial. I shall argue that, with the benefit of our increased knowledge, it is now possible to put forward a theory which explains them more satisfactorily than any previous one. I suppose substantial Plautine rewriting in I 2-3, where I believe L. Gestri was on the right lines but did not go far enough;⁴⁵ in III 3, where Fraenkel and others have recognized Plautine expansion,⁴⁶ I go farther and argue that Plautus introduced Philaenium to the scene, altering the action of his model. Again in the final scene of the play there are certain inconsistencies in connexion with the cena which takes place partly in and partly outside Cleareta's house. The solution of O. Ribbeck was to suppose the loss from Plautus' text of a scene before 828;⁴⁷ other scholars have found the solution to the problem in the deletion of two lines of text, 828f.⁴⁸ I shall argue that the difficulty results from a substantial rewriting of the scene by Plautus, and that, although it is impossible to identify all the changes Plautus made, these include the introduction of Philaenium once again and the transfer of the drinking party on stage from behind the scene.

Argyrippus and The Lena in I 2-3

Of the various theories advanced to account for inconsistencies between I 2-3 and other parts of the play, that of Havet at present holds the critical field. I concentrate on refuting it. Ahrens attempted a comprehensive refutation but, although he made some good points, not all of his arguments were sound.⁴⁹ I do not discuss the 'contamination' theories in detail but only note objections to them in passing. Instead I offer different explanations of the inconsistencies.

According to the MSS. the young man (133) who delivers a monologue (I 2) and then takes part in a dialogue with Cleareta (I 3) is Argyrippus, although he is not named in the text. Havet argued that the lover of these scenes is really Argyrippus' rival, Diabolus, who appears later in IV 1-2.⁵⁰ As a few scholars have recognized, however, Havet's theory is open to fatal objections. After I 1, of which the principal theme is Argyrippus' predicament, an audience would inevitably assume that the unnamed, impoverished lover of I 2-3 was Argyrippus.⁵¹ Furthermore the theory itself leads to new difficulties. The lover of I 2-3 has no money now, even if it is suggested that he did have in the past. Yet later in the play Diabolus appears to have no lack of funds; he is able and willing to enter into a contract that very day. Finally, the lover of I 2-3 is highly emotional; yet, unlike Argyrippus, Diabolus in IV 1-2 shows little sign of being emotionally involved with Philaenium.⁵²

Havet's case rests essentially on certain inconsistencies in the text, some more serious than others. He demands, however, a higher standard of consistency than is to be expected of Plautus. At least some of the inconsistencies are such as we should not expect to find in Menander, but that Plautus was tolerant of inconsistency to a high degree should be beyond dispute.⁵³ The inconsistencies in this case are of two kinds; some are structural, in that they concern the central action of the play, others are more superficial. The former relate to Argyrippus' next appearance in III 3. The lover who is banished by Cleareta in I 2-3 makes his exit at 245 towards the forum, whereas Argyrippus is described by Libanus as intus at 329, at his next appearance comes out of Cleareta's house together with Philaenium and says nothing of having been ejected before. These inconsistencies, however, can be explained satisfactorily by the hypothesis which I shall put forward of Plautine changes in III 3 without either invoking Havet's theory or postulating 'contamination'. I discuss them in detail below.

The superficial inconsistencies on the other hand can plausibly be explained by the hypothesis of Plautine rewriting in I 2-3. Whereas it is often stated or implied in these scenes that the lover has spent a lot of money on Philaenium, in III 1 Cleareta states that Argyrippus has never given anything except promises. To some extent the discrepancy could be attributed to exaggeration on both sides.⁵⁴ Still, Cleareta's statements in 524-31, and in particular the clear implication that Argyrippus has promised to pay when his mother dies (529), agree with the situation revealed in I 1 : Argyrippus' mother Artemona holds the purse-strings and neither Argyrippus nor his father have any resources of their own. We may accept that Argyrippus' claims in I 2-3 are inconsistent with this situation;⁵⁵ and we might well be reluctant to attribute this inconsistency to a Greek dramatist. It can be explained, however, by supposing minor Plautine additions and modifications such as were a normal part of the process of vortere. So far as the development of the plot is concerned, the essential point is that the lover has no money now, when Cleareta has lost patience and issued an ultimatum (cf. 534); on this these scenes are in full accord with Argyrippus' situation. They depict in fact a stock situation, that in which a lover has spent all his money on an insatiable lena, who now turns him away; Argyrippus' situation differed from this in that he never had any money to give, but only promises. It is not difficult to believe that Plautus has here distorted a sentimental picture of a young man's relationship with a hetaera by the introduction of conventional motifs associated with the stock figures of the mercenary lena and meretrix (cf. Ter. H.T. 39, Eun. 37). There is little doubt that he did similar things elsewhere, for example in Poen. I 2, where Adelphasium and Anterastilis are depicted as experienced meretrices, although this is inconsistent with their subsequent rôle in the plot.⁵⁶

There are other grounds for supposing Plautine rewriting in these scenes. I 2 consists of a 26-line entrance monologue, a canticum in which the speaker's emotional excitement is reflected in the metres, first cretics, then trochaic septenarii; Plautine expansion is as likely here as anywhere.⁵⁷ The reference to Roman tresviri in 131 (cf. Truc. [Truculentus] 761) is obviously Plautine, and most, if not all, of this verbose tirade against lena and meretrix could well be Plautine invention. There is a small inconsistency between 151 atque eccam inlecebra exit tandem, implying that Cleareta has not previously been on stage, and 153-5, which suggest that she has heard everything the lover said; it is a plausible guess that in the Onagos lover and lena entered at the same time in conversation.

In I 3 also there are indications of substantial Plautine rewriting. In a detailed study of this scene Gestri showed that the characterization of Cleareta in the scene has two contrasting aspects.⁵⁸ At times she is presented as giving a reasoned defence of her behaviour: she is running a business (186) and must look after her own interests (177), Argyrippus unreasonably expects something for nothing and resorts to abuse only because he has no money (188f.), she needs money for her own expenses and, like any other tradesman, expects payment for goods supplied (198-201), and her profession has special expenses which she must in the end recoup (215-18). At other times she is presented as a conventionally rapacious lena: like Ballio in the Pseudolus, she positively delights in being abused (153-5) and in living up to the bad reputation of her profession (173-5). Even if not all Gestri's subtle arguments command assent, he has made a persuasive case for supposing that Plautus transformed a more reasonable lena of the Onagos into a typical improba lena (Ov. Am. 1.15.17), as part of an extensive reworking of the scene.⁵⁹ The necessary corollary of depicting Cleareta as rapacious was that Argyrippus had spent a lot of money on her. In support of this hypothesis Gestri drew attention to abrupt transitions of thought in the dialogue (e.g. 156, 178, 186) and passages in characteristically Plautine 'cantante' style, in which rhythm and sound effects predominate over logic (179-85, 204-14). A clear example of Plautine expansion is the extended comparison with fowling in 215-25. The first part of the simile, 215-18, probably derives from the Onagos; it neatly makes the point that the lena legitimately expects a return on her professional investment. In the development of the simile in 219-25, however, this point, which coheres with Cleareta's other arguments justifying her treatment of Argyrippus, is dropped in favour of a characteristically Plautine list of particular points of comparison between the professions of lena and fowler and picturesque details of how the lena catches her prey.⁶⁰ The extended comparison with fishing in True. 35-45 is similar, one of a number of similarities, in both thought and language, in the depiction of the quaestus meretricius in these two plays;⁶¹ these led Leo to posit a connexion between the Greek models of the two plays, but they are better ascribed to the common Latin adapter.

A similar answer can be given to Havet's assertion that the violent threats and lack of any expressions of tender feeling for Philaenium in I 2-3 are inappropriate to the Argyrippus described by Cleareta as largus lacrumarum (533) and depicted in 585ff. Even when allowance is made for the speaker's anger,⁶² the fierce ness of his invective, and particularly his outburst against Philaenium as well as Cleareta in 13If. ibo ego ad trisviros vostraque ibi nomina faxo erunt, capitî te perdam ego et filiam, no doubt go beyond realism.⁶³ That is not a sufficient reason, however, to deny that Plautus could have put this diatribe in the mouth of Argyrippus. The lover of I 2-3 is depicted as emotional, no less than Argyrippus is later; in both cases strong emotion is expressed in exaggerated language. We should not expect a greater consistency of characterization from Plautus than that; he was concerned above all to make his audience laugh and for this purpose stereotypes were more useful than subtly drawn characters.⁶⁴ Argyrippus' threats and abuse cohere with the conception of Cleareta and Philaenium as pessumae in this scene and will equally be Plautine additions, as is confirmed by the Roman allusion in 131; Gestri has pointed to traces in the scene of a humbler Argyrippus (190, 229).

Havet's other arguments are nugatory and can be briefly answered. Taking 135a in mari repperi to imply that the speaker had engaged in profitable commerce, which Argyrippus could not have done,⁶⁵ he presses too hard the logical implications of a piece of Plautine rhetoric. The comparison of the women with the sea coheres with the conception of the voracious lena swallowing the lover's wealth, which dominates I 2-3 (cf. Truc. 568f.). It culminates in 135b hic elavi bonis, which plays with the idiomatic meaning of the verb, 'to be cleaned out'.⁶⁶ The idea of the sea as a source of wealth in 135a provides a convenient antithesis; it should not be accorded a significance extending beyond its immediate context. That Diabolus later brings a draft syngraphus in 746ff. is no reason to suppose Cleareta's words in 238 syngraphum facito adferas addressed to him rather than to Argyrippus. Cleareta was evidently willing to enter into a contract with anyone, provided he could produce 20 minae (230f.). It is no real inconsistency that, whereas in 89 Demaenetus knows about Argyrippus' need for 20 minae, Argyrippus should ask in 229f. die, quid me aequom censes pro ilia tibi dare, annum hune ne cum quiquam alio sit? and receive Cleareta's answer viginti minas. Argyrippus must be supposed to have heard Cleareta's terms before, but it is natural enough that he should seek final confirmation, if only as a delaying tactic (cf. mane, mane); this leads to a last desperate claim to be able to raise the money somehow (233f.).⁶⁷ Nor is it surprising that Argyrippus should in 245-8 announce his intention of trying to borrow money from his friends and say nothing of the possibility of enlisting his father's help; the opening scene made it clear that he had already approached him (74f.) but could have little hope in that quarter.⁶⁸ Finally, it is not significant that in III 1 Cleareta does not mention to Philaenium that Argyrippus still hoped to produce the money demanded (234); she had heard such expressions of hope before and was not now disposed to believe them.⁶⁹

Havet's change is therefore definitely to be rejected. Rather we should recognize that substantial Plautine rewriting has in these scenes depicted Argyrippus' relationship with Philaenium in a way that is inconsistent with what is implied elsewhere in the play.

Philaenium and The Horseplay Of III 3

Anomalies explained if Philaenium added by Plautus

Philaenium first appears in III 1 with her mother Cleareta. In this short scene she is depicted in an attractive light. She shows herself genuinely fond of Argyrippus and eager to continue a relationship with him, although he has no money. Her romantic attitude is contrasted with the hard-headed realism of her mother. There is no reason to doubt that in essentials this scene goes back to the Onagos. Its tone is serious and it follows naturally from I 3, in which Cleareta banishes Argyrippus from her house unless he can produce 20 minae (228-30 ~ 532f.).⁷⁰ Philaenium laments the loss of her lover (515 illo quem amo prohibeor and Cleareta lectures her on her filial duties.

The scene in which Philaenium next appears, however, III 3, shows a number of features suggesting modifications by Plautus to his Greek model. First, the scene involves four speaking characters. There is now very substantial evidence that this was at least not normal in Greek New Comedy; and in many of the scenes of Plautus and Terence which involve more than three speaking characters there is cause to believe that one or more of these were added by the Latin adapter.⁷¹

Secondly, as already noted, it is anomalous that at 591 Argyrippus enters with Philaenium from Cleareta's house, whereas his last exit was in the direction of the forum (245 pergam ad forum).⁷² It is usual in Plautus and Terence, regular in the extant Menander and probably a general rule of New Comedy, that a character who makes an exit in the direction of some place supposed to be situated off stage, whether into the interior of a house visible on stage or towards marketplace, harbour or country, on his next appearance enters as if returning from that place.⁷³ That it was a regular convention of New Comedy is supported by the fact that exceptions to it are occasionally explained in the text by a statement that a character will move from one place to another off stage by a 'back street' or 'garden' out of sight of the audience (e.g. 741-3).⁷⁴ Other exceptions without such an explanation can plausibly be attributed to the Latin adapter,⁷⁵ and one case is certainly due to Plautus; we now know that inconsistencies in the movements of Nicobulus in the Bacchides result from Plautus' omission of scenes of Menander's Dis Exapaton.⁷⁶ In the Asinaria it is nowhere explained how Argyrippus returned to Cleareta's house.

It is true there are some indications that as early as II 2 Argyrippus is again in Cleareta's house, but they do not stand up to critical examination; they look less like Greek dramatic technique than makeshift attempts by Plautus to cover up an inconsistency of his own making. When in II 2 in response to Leonida's inquiry in 328 ubinam est erus? Libanus replies maior apud forumst, minor hic est intus, an uncritical spectator might take this at face value and thereby be prepared for Argyrippus' next entrance from Cleareta's house. But how does Libanus have this information? If he is relying on knowledge that when he himself left for the forum at 117 (cf. 108), Argyrippus was still in Cleareta's house, he is wrong; and although he could have learnt of Argyrippus' ejection and subsequent return if the two met in the forum, it is nowhere stated that they did.⁷⁷ A more critical spectator might well hesitate to accept Libanus' statement as true, when it conflicted with the evidence of his own eyes. Again in III 1 it seems likely that Cleareta's words in 533 Me … trudetur … foras, whether or not in their context they necessarily imply it, are intended to suggest that Argyrippus is at that moment in Cleareta's house. This is contradicted, however, by the whole tenor of the scene, which strongly suggests that Argyrippus is not now there.⁷⁸ Is it likely that the two women would come out leaving him inside? Should the dialogue not have run differently if they had? When Cleareta says to Philaenium in 532f. nisi mi hue argenti adfert viginti minas, ne Me ecastor hinc trudetur … foras, this is hardly compatible with her having allowed Argyrippus to return empty-handed after she gave him his marching orders in 228-31. Rather she seems to be merely reaffirming her earlier ultimatum, 'Unless he brings 20 minae, he will be thrown out (if he succeeds in entering the house)'. Only the phrase trudetur foras, rather than excludetur, suggests that Argyrippus is at present in the house; the sentence as a whole suggests the opposite.

Thirdly, after Argyrippus has been thrown out of the house by Cleareta at 127 (161 eicis domo; cf. 152) and told to return only if he brings 20 minae (228-31), it is improbable that he should have attempted to enter the house again without any money (631) and that, if he did so, he should have been allowed to get as far as seeing Philaenium.⁷⁹ In fact Argyrippus' pathetic emergence from the house in III 3, as if only now finally ejected by Cleareta (cf. 596, 632), can be seen as an unneccessary and unrealistic duplication of his earlier ejection.

All these anomalies can be economically explained by the hypothesis that in the Onagos Argyrippus here entered alone from the direction of the market-place, no doubt expressing despondency about his situation in an entrance monologue, and that it was Plautus who made him come from Cleareta's house together with Philaenium in order to exploit the dramatic potential of the resulting situation. Philaenium's participation in the scene is not essential and serves primarily to enhance its comic effect. The phrases in 329 and 533 suggesting that Argyrippus is again in Cleareta's house will on this hypothesis be consequential Plautine additions preparing for III 3. A detailed analysis of the scene will provide further support for the hypothesis.

The slaves' delaying tactics suggest Plautine expansion

The scene contains a kernel of dramatic action which is essential to the development of the plot. The two slaves hand over to Argyrippus the money obtained through Leonida's impersonation of Saurea, in order that Argyrippus may have the means of continuing his relationship with Philaenium. Thus is fulfilled the task enjoined upon Libanus by Demaenetus in the opening scene of the play. Whereas in real life, however, the slaves might have been expected to hand over the money as soon as Argyrippus appeared, Leonida and Libanus deliberately delay doing so for a considerable time, so that the actual hand-over does not take place until 732ff., 140 lines after Argyrippus' entrance. Their delaying tactics make possible a variety of comic effects, and that is obviously their dramatic function. In this long scene, which delays the progress of the plot for the sake of comic effects, at least some Plautine expansion is intrinsically very likely.

The eavesdropping episode, 591-618

The first delaying device consists of an eavesdropping episode, 591-618, in which the slaves Libanus and Leonida, on stage from the previous scene, listen to and comment on the dialogue of the new entrants, Argyrippus and Philaenium. This eavesdropping is prepared by the end of the previous scene, 585-90, where the slaves observe the entrance of Argyrippus and Philaenium and stand back with the intention of eavesdropping. Now the eavesdropping convention, and the withdrawal which prepares for it, undoubtedly originate in Greek dramatic technique.⁸⁰ That does not mean, however, that in every case where these conventional devices are used by a Roman dramatist they must derive directly from his Greek model. In fact there is good reason to believe that Plautus developed the eavesdropping convention, as also other conventions of New Comedy, and used it in insertions of his own invention.⁸¹ In particular the technique whereby two characters eavesdrop on the dialogue of two others and themselves carry on an aside dialogue, necessarily involving four speaking characters, has no known Greek parallels and seems a Roman development.⁸² This eavesdropping episode and its preparation show a number of characteristically Plautine features. In 585-90 the assonance of 587 lacrumantem lacinia tenet lacrumans deserves mention; but much more significant is the way in which the idea of the eavesdroppers' silence (586 opprime os, 588b taciti auscultemus) is developed into a brilliant conceit, in riddle form, in which the money derived from the sale of some asses is identified with the asses themselves, who in the world of Plautine fantasy come to life and threaten to betray the eavesdroppers by their braying.⁸³ As to the eavesdropping itself, we may first note that it does not serve to advance the action by giving the eavesdroppers any new information; its dramatic justification lies in the momentary effect of the lovers' lamentations and in the sardonic comments of the slaves in the background. From the beginning the word-play on vale and salve in 592f. prevents us from taking seriously the pathos of the lovers' farewell.⁸⁴ Their extreme sentimentality, culminating in reciprocal threats of suicide, is surely meant to seem exaggerated and ridiculous. The slaves' comments and 'mocking presence in the background' also serve to undercut any pathos.⁸⁵ Libanus' verdict (616f.) is that an unhappy lover is a great deal more fortunate than a flogged slave! In 588-605 the slaves' aside exchanges extend to an unrealistic length characteristic of Plautus but not Menander or Terence; and these asides are prompted by Argyrippus' words in 597 nox, si voles, manebo, which are totally inappropriate to the dramatic situation but make possible the comic conceit of the nocturnal profligate and day-time 'Solon', which Fraenkel recognized as Plautine.⁸⁶ Both the form and most of the content of this eavesdropping episode can very plausibly be attributed to Plautus, although he may have derived one or two ideas for it from somewhere in the Onagos.

The meeting and preliminary skirmishing, 619-28

After the eavesdropping comes the meeting of the pair of slaves with the pair of lovers. Again the slaves are in no hurry to impart their good news. Instead, there follows a passage of badinage, in which conventional greetings are mixed with jokes and verbal skirmishing, before the dialogue turns to the business in hand (619-28). The eavesdropping is brought to an end with unnaturalistic stage movements in keeping with the rollicking iambic septenarii in which the whole scene is written; as the lovers embrace (615, 619), the slaves approach them from opposite sides (618 circumsistamus, alter hinc, hinc alter appellemus). This stage business, which we have to imagine as best we can, is clearly connected with a certain symmetry in the following lines, in which first one slave addresses Argyrippus (619), then the other addresses Philaenium (623).⁸⁷ One is reminded of the antiphonal jesting of the Fescennina iocatio (Catull, 61.120) traditional at Roman weddings, amongst other occasions.⁸⁸ Argyrippus' tears prompt a typically Plautine joke, in which Philaenium is comically identified, in the form of a riddle, with smoke (619f.)⁸⁹ Then Leonida indulges in a rather feeble equivocation on the meaning of perdere (62If.). Libanus' greeting addressed to Philaenium receives a conventional reply, dabunt di quae velitis vobis;⁹⁰ whereupon Libanus, taking literally the stereotyped phrase, a common Plautine device,⁹¹ makes the grossly impertinent retort noctem tuam et vini cadum … (623f). Argyrippus threatens Libanus, Libanus corrects himself⁹² and switches to a gratuitous attack on Leonida, so that the passage ends with the two slaves exchanging abuse on the theme of verberare. All this is typical Plautine slave-talk. Philaenium's rôle in the passage is to provoke a joke from one slave and an obscene suggestion from the other.

Argyrippus in despair, 629-38

629-38 contain a more serious piece of dialogue between Argyrippus and Libanus, in which Argyrippus laments the situation which drives him to contemplate suicide, his lack of money and ejection by Cleareta. The change of tone from the preceding passage suggests that Plautus is here following his Greek model more closely. This receives some confirmation in the fact that 633-5 introduce a crucial aspect of Argyrippus' situation, the existence of a rival, Diabolus, who has actually promised to pay 20 minae for Philaenium's services for a year. It is true that Cleareta had already offered a contract for this sum to Argyrippus (229-40) and hinted at the possibility of a rival by her statement that she was willing to accept 20 minae from the first comer (231; cf. 195). The first mention of a definite rival, however, and of Diabolus' name, is in 634. It is possible that Plautus cut some earlier reference to him, but the natural presumption is that Argyrippus only learnt of Diabolus' offer when he went to the market-place at 248, although Plautus does not make this explicit.⁹³ 633-7 then mark a new development in the action, which is likely to derive from the Onagos.⁹⁴ Libanus' response in 638 is significant; he inquires whether Diabolus has already given Cleareta the money, and on being told that he has not, encourages Argyrippus not to worry. The audience, knowing that the slaves have obtained money for Argyrippus' benefit, would expect them now to reveal the fact; but this expectation is disappointed.

The ludificatio planned in an aside dialogue, 639-48

Leonida suddenly draws Libanus aside and the two slaves must be imagined as putting their heads together in a conspiratorial huddle (639). For Argyrippus suggests that they might as well embrace while talking, a sally not at all in keeping with his present situation and mood (640). Libanus (or Leonida) replies that they do not fancy each other and makes the counter-suggestion that the lovers should embrace (641-4); this they willingly do and are thus kept occupied while the slaves, in a brief aside dialogue, plot to make fun of Argyrippus (645-8). Thus 639-48 explicitly prepare for a ludificatio of the lovers,⁹⁵ of which 619-28 provided a mild foretaste⁹⁶ but which will turn into slapstick farce and extend to 731. Leonida gives it as his specific aim to get Philaenium to embrace him in the presence of her lover (647); and this foreshadows the major rôle which Philaenium plays, as we shall see, in the actual ludificatio. Now there are indications that this little episode, clearly defined by the movement of the slaves away from the lovers (639 secede hue, 646 concedite istuc) and back again (648 sequere had), is a Plautine insertion.⁹⁷ It is a self-contained episode which abruptly cuts short the dialogue between Argyrippus and Libanus in 629-38⁹⁸ and separates Libanus' reassuring words in 638 from their natural continuation, the revelation that Leonida has obtained the 20 minae Argyrippus needs, which in fact follows in 649-56;⁹⁹ and it is unrealistic in the circumstances that Argyrippus should be so tolerant of the delay. Moreover the dramatic device of withdrawal in order to devise a scheme out of hearing of the intended victim, a specialized development of the aside convention, is one which occurs elsewhere in Plautus in a context where there is reason to suspect Plautine originality. In Pers. 833-43, in the boisterous finale of the play, involving five speaking characters, Toxilus persuades a reluctant Lemniselenis, in an unrealistically long aside dialogue, to ludificare Dordalus (833, 843), causing Dordalus to comment certo illi homines mihi nescioquid mali consulunt (844); the inconsistency of Toxilus' crude bullying of Lemniseienis with his love for her and the extended play with the idea of clientela there suggest the hand of Plautus.¹⁰⁰ In both passages the planning is rather belated, serving merely to make explicit a ludificatio which is in fact already in progress; and this explicitness is characteristic of the stagecraft of Plautus in contrast to the greater naturalism of Menander and Terence.¹⁰¹ Finally, the withdrawal itself gives rise to the joke about embracing slaves, which, with its connotation of homosexuality, is very Plautine.¹⁰²

The slaves deliver their good news, 649-56

It seems probable that in the Onagos the slaves proceeded to impart their good news to Argyrippus without delay, in other words that 649-56 represent the original continuation of the exchanges between Argyrippus and Libanus in 638. Even here allowance must be made for the distortion of Plautine vortere and we cannot hope to reconstruct the Greek dialogue. Leonida's claim to be rewarded with his freedom (649-52) could, in embryo, derive from the Onagos; but when he caps Argyrippus' libertos with non patronos? (652), the reversal of real-life rôles, whereby the resourceful slave is conceived as patronus, the helpless lover as libertus is characteristically Plautine.¹⁰³

The ludificatio, 657-731

In 657-731 massive Plautine expansion has been suspected before.¹⁰⁴ In this passage occurs the slaves' prolonged ludificatio of the lovers. That this is at least in large part the work of Plautus is likely on general grounds. The ludificatio delays the progress of the main action of the play very considerably, to great comic effect but at the cost of any pretensions to realism. That the preparation for it in 639-48 has been shown to be a probable Plautine insertion reinforces the case for supposing the ludificatio itself at least largely Plautine. Analysis of the whole passage provides further arguments in support of the hypothesis. It is possible to distinguish four sections: (a) 657-61, (b) 662-97, (c) 698-710, (d) 711-31. With these sections are associated three leading themes, (a) and (c) revolve around the idea of carrying, whether the purse full of money or Libanus; we may label this the 'burden' motif, (b) contains an exoratio directed by the lovers first at Leonida, then at Libanus. In (d) the slaves in turn demand that they should be addressed as gods, Salus and Fortuna respectively.

For the present argument (b) is the most important. Only in this section does Philaenium play a major and essential rôle, a rôle which was foreshadowed in 647 in the planning of the ludificatio; each slave in turn duly demands an embrace from her as well as verbal endearments.¹⁰⁵ There are grounds for supposing this section entirely Plautine invention. That it separates (a) and (c), which are thematically linked by the 'burden' motif, is at least suggestive. Then the parallelism, whereby the exoratio of Leonida balances that of Libanus (662-79 ~ 680-97), looks Roman.¹⁰⁶ Finally, the slaves' grotesque parody of lovers' endearments (666-8 ~ 693f.) is very Plautine and paralleled by Poen. [Poenulus] 355-96, another exoratio which is most probably Plautine invention.¹⁰⁷

It is probable that (d) also is purely Plautine. The identification of Libanus and Leonida with Salus and Fortuna and their implied deification takes the exaltation of the salve to its ultimate extreme. This is characteristic of Plautus but not very likely in Greek New Comedy.¹⁰⁸

Only in (a) and (c) is it likely that something derives from the Onagos. How much, can hardly be determined. It is at least not unlikely that the 'burden' motif itself, as introduced in (a) in a relatively sober form, derives from the Greek (cf. Ar. Frogs 1-32). Less likely, but not impossible, is that the farcical development of the motif in (c) also had some counterpart there (cf. Men. Dysk. 890ff.). Typically Plautine, one the other hand, is the humiliation of the free citizen at the hands of his slave (702 sic istic soient superbi subdomari),¹⁰⁹ as well as the reference to pederasty in 703 asta igitur, ut consuetus es puer olim.¹¹⁰

Conclusion

To sum up, it can hardly be doubted that Plautus greatly expanded III 3. There are substantial grounds for regarding all the parts of the scene which involve Philaenium as Plautine additions.¹¹¹ This supports the hypothesis that Philaenium was not present in the corresponding scene of the Onagos, a hypothesis which explains inconsistencies between this and earlier scenes. The convergence of a number of arguments is sufficient to give the hypothesis a high degree of probability.

Philaenium and The Cena of V 1-2

Finale with five speaking characters

Philaenium last appears in V 1-2, the final scenes of the play. That five speaking characters are involved in these scenes is alone sufficient reason for supposing some Plautine alterations to his model. 880-910 form an eavesdropping episode, in which, as in 591-618, a pair of eavesdroppers carries on a dialogue commenting on another dialogue; here the parasite and Artemona eavesdrop on a dialogue between Argyrippus, Demaenetus and Philaenium. The only dramatic purpose of the eavesdropping is the comic effect of Artemona's indignant reactions to Demaenetus' crude remarks about her. In 920-41, after the exit of the parasite, the remaining four characters converse together. There is no known Greek parallel for this dramatic technique. On the other hand there is evidence that especially in the finale of a play the Latin adapters not infrequently increased the number of speaking characters.¹¹²

Characterization of Philaenium

Furthermore, as has often been noticed, Philaenium behaves in V 2 with a pertness which is inconsistent with her characterization in III 1.¹¹³ The hypothesis that her rôle in this scene is a Plautine addition provides a good explanation of this inconsistency. That Plautus should here depict her as a typical meretrix and use her for momentary comic effects is entirely characteristic and in keeping with his treatment of her in III 3 and of Cleareta in I 3. This hypothesis is supported by other features of these scenes.

Drinking parties on stage introduced by Plautus

It is abnormal that a drinking party actually takes place on stage. Although New Comedy plots often involve some sort of feast, especially at the end of the play, this normally takes place behind the scene.¹¹⁴ Only in four plays of Plautus does eating or drinking definitely take place on stage. In all the scenes in question there is at least some reason to believe that Plautus has transferred on to the stage a party which in the Greek original took place behind the scene. Gaiser has very plausibly suggested that this was the case in Most. [Mostellaria] 308ff, another scene involving five speaking characters, one purpose of the change being to bridge the act-division of the Greek play.¹¹⁵ In Pers. [Persa] 753ff, another five-speaker finale and an elaborate polymetric canticum, substantial Plautine rewriting is probable, for reasons which cannot be discussed in detail here.¹¹⁶ That Plautus created this on-stage symposium is suggested by the fact that it is introduced in 757 as the traditional feast with which a Roman general entertained his victorious troops,¹¹⁷ and that an important rôle is played in it by the puer delicatus Paegnium (as cupbearer and as dancing cinaedus), a probably wholly Plautine character.¹¹⁸ Although it is now generally held that a nucleus of the Stichus finale derives from Plautus' model,¹¹⁹ some Plautine rewriting is nevertheless probable, and it is at least not unlikely, though difficult to prove, that Plautus is responsible for the scenario of reclining slaves (696, 703, 750, 752) drinking on stage in a parody of a symposium; the cena behind the scene (662-5, 678-81) no doubt goes back to Menander's Adelphoi I, but its finale could have presented merely some postprandial merriment similar to that at the end of the Dyskolos.

Peculiarities of the Asinaria drinking party

Asinaria V 1-2 display several peculiar features which do not look like Greek dramatic technique and for which a good explanation is provided by the hypothesis that Plautus brought on stage a drinking party which in the Onagos took place behind the scene.¹²⁰ Of course the theatrical conventions of ancient Greece and Rome required that many activities had to be depicted as taking place out of doors which in real life would more naturally have taken place indoors.¹²¹ Nevertheless the situation in Asinaria V 1-2 is particularly unrealistic. When Demaenetus, Argyrippus and Philaenium, garlanded (879), come out of the house at 829 and recline (829, 83If., 878, 921-5), presumably on couches, beside a mensa brought out by slaves (829), to take part in a symposium, they are transferring into the street an activity which Diabolus had already witnessed taking place in the house (825f. cum suo sibi gnato unam ad amicam de die potare; cf. 851).¹²²

Moreover this drinking scene on stage involves some dramatic awkwardness. Artemona enters with the parasite at 851 but not until 880 does she become aware of the party taking place on stage; on the other hand the dialogue of Demaenetus and Argyrippus is suspended during the 30-line dialogue of the parasite and Artemona, of which they take no notice. Granted that it was an established Greek convention that two characters could be on stage together but unaware of each other's presence, one may well feel that the convention is here stretched more than a Greek dramatist would have allowed.¹²³ This awkwardness can plausibly be explained as the result of Plautus' having brought the symposium on stage in V 1 ; and the nature of V 1 itself supports this hypothesis. The dialogue between Demaenetus and Argyrippus in V 1 does not advance the plot but elaborates and presents dramatically an idea already adumbrated in 738f., Argyrippus' feelings of jealousy at watching his father fondle Philaenium.¹²⁴ The theme is developed amusingly with Argyrippus' forced attempts to put on a smiling face (837-41). Philaenium is silent throughout this scene but has an important visual rôle as the object of Demaenetus' lascivious attentions; and the very fact that she makes no protest at having to play this rôle is hardly in keeping with the way she is depicted in III 1, although consistent with her behaviour in V 2. It is a plausible hypothesis that at most the basic idea of V 1 derives from the Onagos and that its scenic realization is the work of Plautus.¹²⁵

876-80 would be more appropriate as preparation for the revelation of a scene inside the house, the parasite's instruction in 876 sequere hac me modo, iamfaxo ipsum hominem manufesto opprimas serving to bring Artemona towards the door of the house, as is the situation in Bacch. 831 sequere hac me, faxo iam scies. In Plautus this manoeuvre further underlines the artificiality of Artemona's slowness in observing the symposium taking place on stage. If in the Onagos the drinking party remained inside the house, something of Demaenetus' behaviour and perhaps of his conversation could have been indicated indirectly through the comments of the onlookers, as in Bacch. 834ff. Artemona would naturally watch for a few moments before bursting in. The nature of the eavesdropping episode, 880-910, however, suggests that it is largely Plautine invention.¹²⁶

Perhaps the strongest argument, however, for supposing that in the Onagos the drinking party took place behind the scene, not on stage, is to be found in the parasite's monologue, 911-19. The parasite's exit, motivated by such an exit monologue, is natural enough, now that he is no longer needed. His exit monologue, however, awkwardly interrupts the meeting between Artemona and the revellers, which is already taking place. The preceding dialogue has prepared the audience well for a confrontation between Artemona and her husband. In 909f. Artemona, her patience exhausted and encouraged by the parasite, addresses a threat directly at Demaenetus, loudly enough to draw attention to her presence; in 911 Argyrippus greets her mater, salve and she replies curtly sat salutis. At this point she must be supposed to be at least starting to move towards the revellers and the centre of the audience's attention. Yet at this critical moment the audience is expected to switch its attention to the parasite. This is strange dramaturgy. The peculiarity is explicable, however, if the parasite's monologue is a relic of the situation in the Onagos; it would have been well suited to fill the gap while Artemona fetched Demaenetus from the house.

Conclusion

In this article a few scenes of one play of Plautus have been analysed in detail; but analysis of other plays would produce similar results. The picture of Plautus' practice which emerges is of free adaptation but within the framework of the original Greek plot. Whether Plautus is to be accorded the title of 'dramatist' or 'poet' is in part a matter of terminology. Certainly his drama is very different from that of Menander; in spirit it is probably closer to the improvised popular farces which existed in Italy before the introduction of the palliata. Plautus shows himself more concerned with the dramatic effectiveness of the individual scene than with the overall structure of the play. Moreover he is more dependent on his models than were Shakespeare or Molière. It is right to recognize the limits of his dramatic invention. There is no disputing his originality however. To call him a 'translator' is, to say the least, misleading, given the present-day implications of the word. Very few have ever denied that in his highly individual style he was a creative artist of a kind. His originality went beyond that, however, and beyond what Fraenkel was willing to allow. Fraenkel's isolation of discrete Plautine insertions was a necessary step in the discovery of the true extent of Plautus' originality, but it must be recognized that there are scenes, such as Asinaria I 3, which Plautus has rewritten so thoroughly that a neat separation of his additions from original Greek material is impossible.¹²⁷ Such rewriting can lead to inconsistencies with other parts of the play, especially in characterization; in particular we have noted a tendency to depict certain characters in brighter colours and to substitute comically exaggerated stereotypes, such as the greedy meretrix, for realistic characterization. Furthermore it must be recognized that Plautus has made structural changes and substantive additions to his Greek models which are clearly in some sense 'dramatic'. We have found reason to see him as a master of stagecraft in the creation of new action, if on a relatively limited scale, as well as new verbal comedy, in the addition of an extra character in Asinaria III 3 and the introduction of a symposium on stage in V 1-2, in the development of the Greek eavesdropping and aside conventions for novel effects. This fits well with the tradition that he came to playwriting from practical experience of the theatre.¹²⁸

Any attempt to separate by analysis Greek and Roman elements in his plays must take into account this enlarged conception of Plautus' capabilities. One important consequence of crediting Plautus with greater inventive capabilities is that there is less need to postulate 'contamination'; indeed the hypothesis of 'contamination' is rendered rather implausible, at least in its traditional sense of the more or less mechanical stitching together of material from different Greek models. It cannot be ruled out a priori but in fact it has nowhere been demonstrated beyond dispute. Even if it could be demonstrated, from the point of view of the modern critic it would not differ in principle from other modifications by Plautus to his principal model which did not involve borrowing from a second Greek play. It has sometimes been suggested that the scope of the word 'contamination' should be widened to include all such Plautine changes.¹²⁹ Better rather to reduce to the minimum the use of a word with an unhappy history and to define it in a narrow sense. It is indeed very probable that when he made changes to his primary model Plautus often used traditional Greek motifs and sometimes had a particular Greek play in mind; but it is not helpful to use a crude general term without discrimination to cover varying degrees of Greek influence, if it can be established, or none.

If this picture of Plautus as a free adapter of his Greek models is in general correct, it has implications for historians. The Hellenist needs to exercise extreme caution in using Plautus' plays as evidence for New Comedy. Greek thought and dramatic technique are undoubtedly embedded in the Latin plays but have often been given a new twist by Plautus. Where an aspect of social life or law forms an integral part of the basic plot, the Greek historian may with some confidence take it as reflecting Greek life in the New Comedy period, but he must be constantly aware of the possibility of Plautine distortion in detail and will be wise to look for confirmation from other sources. The Roman historian on the other hand may be encouraged to see reflections of contemporary Roman society in passages which structural analysis shows to be probably Plautine insertions. He may legitimately look for possible political factors behind Plautus' changes; and in any case he is entitled to see topical significance in what Plautus deliberately left unchanged. The problem in using Plautus' plays as evidence for Roman life is a different one, namely that they do not purport to depict Roman but Greek life;¹³⁰ and in fact the unashamed insertion of Roman allusions into plays with a Greek setting, together with Plautus' penchant for grotesque exaggeration, creates a Saturnalian fantasy world, an anti-Rome, which to a considerable extent turns the real world upside down.¹³¹ To use such texts as historical evidence requires nice judgement. Finally Plautus provides a case-study of the reception of Greek culture in Rome at the beginning of the second century. Master of racy Latin, although by birth an Umbrian from Sarsina, if we believe tradition, he is evidently thoroughly familiar with Greek New Comedy and its conventions but confidently adapts his Greek models to produce a very different kind of drama, one which proved a great success with Roman audiences. He is no fumbling imitator. He may lack the sophistication of the Augustan poets, but in his plays we can see already the characteristic Roman ability to transform borrowed Greek goods.

Notes

¹ E. g. Asin. 10f., Cas. 32 4; Merc. 9f., Trin. 18f, Ter. Ad. 6f., Gell. NA 2.23; cf. G. Michaut, Histoire de la comédie romaine: Plaute, ii (Paris. 1920), pp. 204-8. R. Perna, L'originalità di Plauto (Bari. 1955), pp. 6f.

² According to A. W. Schlegel. Vorlesungen über dramatische Kunst und Litcratur, ed. G. V. Amoretti, i (Leipzig, 1923), pp. 17If. Plautus and Terence could not be regarded as creative artists. He allowed that the Latin adapters made some changes to their Greek models, but only for the worse. See M. Barchiesi, Maia 9 (1957), 201-3, B. A. Kes, Die Rezeption der Komödien des Plautus und Terenz im 19. Jahrhundert (Amsterdam, 1988), pp. 59-65.

³ Michaut, Plaute, ii. 223-38.

⁴ W. H. Grauert, Historische und philologis che Analekten (Munster, 1833). pp. 116-207 'Über das Contaminiren der Lateinischen Komiker', W. A. Becker, De comicis Romanorum fabulis maxime Plautinis quaestiones (Leipzig, 1937), Th. Ladewig, Über den Kanon des Volcatius Sedigitus (Neustrelitz, 1842), id. in A. Pauly, Real-Encyclopadie der Altertumswissenschaft (Stuttgart, 1848), v. 1728-39, G. Boissier, Quomodo graecos poetas Plautus transtulerit? (Paris, 1857).

⁵ O. Zwierlein, Zur Kritik und Exégèse des Plautus, I: Poenulus und Curculio (Abh. Ak. Mainz 1990/4, Stuttgart, 1990), again sees retractatio as a major cause of inconsistencies, but unconvincingly.

⁶ 2nd edn. Berlin, 1912, pp. 87-187.

⁷P.F.² p. 87 'Seine Komödien sind nicht sein, und sie waren schöner und besser ehe er sie zu eigen machte', p. 185 'Plautus hat, neben einer so hoch gesteigerten Kunst des stilmässigen Ausdrucks, die eigentliche dramatische Fähigkeit … nicht entwickeln können'; cf. Th. Mommsen, Römische Geschichte¹¹, i (Berlin, 1912), p. 906 'Ohne Zweifel hat der Bearbeiter auch hierin mehr das Gelungene der Originale festgehalten als selbständig geschaffen'. According to A. Kiessling and his pupils Plautus often reproduced mechanically what he found in his Greek models without regard to whether his audience could understand it (Perna, L'orig. di PI. pp. llf.).

⁸Geschichte der römischen Literatur, i (Berlin, 1913), pp. 125-32.

⁹ Michaut, Plaute ii. 239-80.

¹⁰ K. Dziatzko and R. Kauer, Ausgewählte Komödien des P. Terentius Afer, II: Adelphoe (Leipzig, 1903), pp. 7f., W. Schwering, Neue Jahrb. 19 (1916), 167-85, A. Körte, Bed. phil. Woch. 36 (1916), 981, Gött. gel. Anz. 195 (1933), 355-61, W. Beare. The Roman Stage¹ (London, 1964), pp. 310-13. For a brief history of the theory of Plautine 'contamination' see L. Schaaf, Der Miles Gloriosus des Plautus und sein griechisches Original (Munich, 1977), pp. 11-14; on the meaning of contaminare see now G. Guastella, La contaminazione e il parassita (Pisa, 1988), pp. 11-80.

¹¹ E.g. P. Lejay, Plaute (Paris, 1925), p. 216 Ά raisonner seulement d'après les vraisemblances, à en juger d'après le tempérament vif et le faire rapide du poète, on peut supposer que Plaute a traitè les modèles grecs très librement, cousant à une intrigue une scène prise ici et une scène prise là, poursuivant une idèe comique qu'il a saisie dans quelques vers d'une troisième pièce, dèveloppant, raccourcissant, mèlant, et partout y mettant du sien'; cf. G. E. Duckworth, The Nature of Roman Comedy (Princeton, 1952), p. 385, Perna, L'orig. di PL p. 471, B.-A. Taladoire, Essai sur le comique de Plaute (Monaco, 1956), pp. 62f., E. Paratore, Storia del teatro latino (Milan, 1957), p. 168.

¹² Frequently repeated slogans are that the analytic method rests on the assumption that Greek New Comedy was perfect and without flaw, that its practitioners attribute anything good in the Latin play to the Greek original, any faults to the Latin adapter, and that they are interested only in reconstructing lost Greek plays—not entirely without truth but gross exaggerations. Since the analytic method was from the start closely associated with the theory of 'contamination', the two things have understandably not always been as clearly distinguished as they should be. For a vigorous defence of the method see H. Drexler, Gnomon 18 (1942), 28-30.

¹³ E.g. Ph. E. Legrand, The New Greek Comedy (London and New York, 1917), p. 43 Ί do not believe that a single essential element of a plot, a single important feature of a character in the plays of Plautus is fundamentally, necessarily, undeniably Roman', p. 283 Ί think that, as a rule, he and his rivals were content to be mere transcribers', p. 285 'It appears to me that the activity of the Latin transcribers was almost restricted to making omissions and to practising contamination', F. Marx, Plautus Rudens (Abh. Sachs. Akad. Wiss., phil.-hist. Kl. 38/5, Leipzig, 1928) Vorwort, 'Das mir vorschwebende Ziel war, bei jedem Vers des Plautus den Wortlaut des griechischen Vorbilds môglichst feststellen zu können'; cf. Β. Prehn, Quaestiones Plautinae (Breslau, 1916), p. 4, W. H. Friedrich, Euripides und Diphilos (Zetemata 5, Munich, 1953), pp. 259-61, T. B. L. Webster, Studies in Later Greek Comedy (Manchester, 1953), pp. 2f., Beare, Rom. Stage³, pp. 63-6, Zwierlein (n. 5), 5.

¹⁴Plautinisches im Plautus (Philol. Untersuch. 28, Berlin, 1922), Italian translation with Addenda Elementi plautini in Plauto (Florence, 1960).

¹⁵P. im P. pp. 3-5 = El. PI. pp. 3f.

¹⁶ E.g. Rud. 515 ~ 540; cf. P. im P. pp. 112f. = El. PI. pp. 106f.

¹⁷ W. G. Arnott, Menander, Plautus, Terence (Oxford, 1975), pp. 34-6. Sceptics include J. J. Tierney, Proc. R. Ir. Acad 50/c (1945), 21-61, H. W. Prescott, TAPA 63 (1952), 103-25, Ε. Csapo, CQ 39 (1989), 148-63.

¹⁸P. im P. pp. 391f. = El. PI. pp. 370f.

¹⁹P. im P. p. 406; cf. pp. 282, 320 = El. PI. p. 384; cf. pp. 271, 306.

²⁰Plautinisches und Attisches (Problemata 5, Berlin, 1931), pp. 3-104; cf. S. Prête. A. e R. 2 (1952), 145f.

²¹Die Komposition von Terenz ' Adelphen und Plautus ' Rudens (Philologus Supp. 26, 2, Leipzig, 1934).

²² E.g. The Greek Aulularia (Mnemosyne Supp. 2, Leyden, 1940).

²³ E.g. Hermes 112 (1984), 30-53 on Miles Gloriosus.

²⁴Die Menaechmi des Plautus und kein griechisches Original (ScriptOralia 11, Tübingen, 1989).

²⁵ See. for example, Phoenix 39 (1985), 19f. on Most. 690-713, BICS 35 (1988), 101-10 on Poen. I 2, H.-W. Nörenberg, Rh. Mus. 118 (1975), 285-310 on M.G. Ill 2, Gratwick, Mnem. 34 [1981], 331-5 on Trin. IV 3, A. Primmer, Handlungsgliederung in Nea und Palliata: Dis Exapaton und Bacchides (Öst. Ak. Wiss., phil.hist. Kl., Sitzungsber. 441, Vienna, 1984), pp. 84-8 on Bacch. IV 9.

²⁶Cambridge History of Classical Literature, ed. P. E. Easterling and E. J. Kenney, ii (Cambridge, 1982), pp. 99-103.

²⁷ E. W. Handley, Menander and Plautus. A Study in Comparison (London, 1968). For subsequent discussion see Arnott (n. 17), 38-40 with n. 49, J. A. Barsby, Plautus Bacchides (Warminster, 1986), pp. 139-45.

²⁸ On the five-act convention see Hermes III (1983), 442-4, on the three-speaker rule references in n. 71 below.

²⁹P.F² p. 110; cf. Barchiesi (n. 2), 185 n. 50.

³⁰ E.g. the cook in Cure. 251-370, a doublet of Palinurus (C. Ant. 4 [1985], 95-9), Pinacium in Most. 858-903; a doublet of Phaniscus (I. Weide, Hermes 89 [1961], 198-203). Dorias in Ter. Eun, 615-726, a doublet of Pythias (Webster, Studies in Menander [Manchester, 1950], p. 73); cf. Rh. Mus. 133 (1990), 292 η. 62.

³¹ E.g. Milphio and Agorastocles in Poen. 210-409 (BICS 35 [1988], 101-10), Palaestra and Ampelisca in Rud. 664-882, 1045-1128 (K. Gaiser, ANRW I. 2 [1972], 1075f., Antipho in Ter. Pho. 465-566, 606-712 (K. Büchner, Das Theater des Terenz [Heideiberg, 1974], pp. 330-5, 338-41, 347-50, 454-7, 482); cf. CQ 33 (1983), 428-44, 39 (1989), 390-9, Hermes III (1983), 431-52.

³² F. Bertini, Plauti Asinaria (Genoa, 1968), pp. 27-43, conveniently summarizes the judgements of earlier scholars.

³³ Cf. F. Delia Corte, Dioniso 35 (1961), 38-41 = Da Sarsina a Roma² (Florence, 1967), pp. 299-304.

³⁴T. Macci Plauti comoediae, Asinaria (Leipzig, 1881), praef. p. xxiv.

³⁵Plauti comoediae, i (Berlin, 1885), on 127.

³⁶Rev. Phil. 29 (1905), 94-7. His suggestion has been accepted in the editions of A. Ernout, Plaute, i (Paris, 1932), and Bertini, by Legrand, N.G.C. pp. 422f., G. Burckhardt, Gnomon 7 (1931), 421f., F. Munari, SIFC 22 (1947), 16-18, Delia Corte, Da Sarsina², p. 30, A. Traîna, Par. Pass. 9 (1954), 202, Fraenkel, El. PI. p. 434, R. L. Hunter, Mus. Helv. 37 (1980), 221 with n. 29, D. Konstan, Roman Comedy (Ithaca and London, 1983), p. 55 n. 7.

³⁷ Op. cit. (n. 36), 99; see below nn. 51, 93. So already A. Spengel, Die Akteinteilung der Komödien des Plautus (Munich, 1877), p. 47; cf. Bertini (n. 32), 48.

³⁸Quatenus in fabulis Plautinis et loci et temporis unitatibus species veritatis neglegatur (Diss. Breslau, 1914), pp. 80-2.

³⁹AJP 58 (1937), 19-37.

⁴⁰Dioniso 19 (1956), 46-57.

⁴¹De Plauti Asinaria (Diss. Jena, 1907), pp. 29-36.

⁴² Op. cit. (n. 36), 10-16.

⁴³ Op. cit. (n. 32), 51-3.

⁴⁴S.L.G.C. p. 235; see below n. 75.

⁴⁵SIFC 17 (1940), 181-214.

⁴⁶P. im P. p. 275 n. 1 = El. PI. p. 265 n. 1, Hough (n. 39), 32 n. 30. Webster, S.L.G.C. p. 236.

⁴⁷Rh. Mus. 37 (1882), 579.

⁴⁸ C. H. Weise, M. Accii Plauti comoediae¹, i (Leipzig, 1847), J. L. Ussing, T. Maccii Plauti comoediae, i (Copenhagen 1875), Leo ad loc. See n. 122 below.

⁴⁹ Op. cit. (n. 41), 13-23.

⁵⁰ Loc. cit. (n. 36).

⁵¹ Hough (η. 39), 24-6, Webster, S.L.G.C. p. 235. Munari (n. 36), 18 n. 1, following Burckhardt (n. 36), 422, supposes the loss of a reference to Diabolus from the prologue 'senza di che lo spettatore non poteva capire chi fosse l'adulescens di I, 2-3'. Hunter (n. 36), 221 admits 'any audience might, however, be forgiven for believing this young man to be Argyrippus in the light of the opening scene between the senex and his slave'. Hough (n. 39), 29 and Rambelli (n. 40), 53f. suppose for the Onagos an appearance of Argyrippus earlier than 585. Havet [n. 36], 99f., supposes for Plautus' play a lost Argyrippus/Cleareta scene.

⁵² Ahrens (n. 41), 21; cf. Munari (n. 36), 18 n. 1.

⁵³ There are collections of Plautine inconsistencies in P. Langen, Plautinische Studien (Berliner Studien 5, Berlin, 1886) and H. Marti, Untersuchungen zur dramatischen Technik bei Plautus und Terenz (Diss. Zurich, 1959).

⁵⁴ Ahrens (n. 41), 16f., Hough (n. 39), 25 n. 13.

⁵⁵ Cf. Hough (n. 39), 25, Rambelli (n. 40), 46-8, Delia Corte, Da Sarsina², p. 299.

⁵⁶BICS 35 (1988), 101-10; cf. Fraenkel, P. im P. pp. 147-52, 277f. = El. PI. pp. 140-5, 267, Gaiser (n. 31), 1083, Gratwick, C.H.C.L. ii. 105-10. I am not persuaded by the attempt of Zwierlein (n. 5), 149f., to explain away the inconsistency; and he himself admits (152) the possibility that Plautus here exaggerated the 'Hetären-Züge'.

⁵⁷ Cf. Fraenkel, P. im P. pp. 142-206 = El. PI. pp. 135-95.

⁵⁸ Op. cit. (n. 45).

⁵⁹ Gestri probably underestimates the extent of Plautus' rewriting. He does not address the inconsistency with III 1 of the references to Argyrippus' gifts in this scene. His attempt to reconstruct an outline of how the dialogue went in the Onagos is perhaps over-optimistic.

⁶⁰ Cf. Fraenkel, P. im P. p. 178 = El. PI. p. 169. Gestri (n. 45), 185-91 shows that the fish simile of 178-85 has probably been similarly expanded by Plautus; it is significant that in the development of the simile the lover's gifts are emphasized (18If.).

⁶¹ Leo on 127, P.F² p. 149, Munari (n. 36), 25 n. 1, Webster, S.L.G.C. p. 235.

⁶² Ahrens (n. 41), 17f., Hough (n. 39), 25 n. 13.

⁶³ Langen, PI. Stud. pp. 99f., Delia Corte, Da Sarsina², p. 299, Hunter (n. 36), 221 n. 29. In 209-14 also Argyrippus implies that Philaenium's attitude to him has changed and he addresses her and Cleareta as pessumae; cf. Gestri (n. 45), 20If.

⁶⁴ The monologue of Mnesilochus, Bacch. 500-25, in comparison with the corresponding monologues of Sostratos in Menander's Dis Exapaton, well illustrates how differently the two dramatists depict the emotions of a lover; cf. Handley, Men. and PI. pp. 14f., Barsby ad loc.

⁶⁵ Cf. Delia Corte, Da Sarsina¹, p. 299. Havet and A. Frété, Pseudo-Plaute, Le prix des ânes (Paris, 1925), p. 7, followed by Rambelli (n. 40), 50, unconvincingly take 191 aetatis atque honoris … tui, in conjunction with 135, as suggesting a middle-aged merchant (cf. Havet [n. 36], 102).

⁶⁶ Cf. Ussing ad he, Rud. 579.

⁶⁷ Ribbeck (η. 47), 55, Ahrens (η. 41), 14-16.

⁶⁸ Ahrens (η. 41), 18f.

⁶⁹ Ahrens (n. 41), 20, makes too much of the fact that Cleareta did not hear Argyrippus' soliloquy, 243-8.

⁷⁰ The theory of Rambelli (n. 40), 57-78, that III 1 and 3 are from a second Greek play rests on unconvincing arguments and is very improbable.

⁷¹ Diomedes, Gramtn. Lat. i. 490, Sad Aesch. Choeph. 899, Gaiser (n. 31), 1037f., 1073-9, F. H. Sandbach in Le monde grec—hommages à Claire Préaux, ed. J. Bingen, G. Cambier and G. Nachtergael (Brussels, 1975), pp. 197-204, Κ. Β. Frost, Exits and Entrances in Menander (Oxford, 1988), pp. 2-5. We need not here concern ourselves with the question whether a three-actor rule in the strict sense was valid for New Comedy.

⁷² Spengel, Akteinteilung, p. 47, Goetz-Loewe (n. 34), xxiv, Havet (n. 36), 96, Hough (n. 39), 25, Munari (n. 36), 18f, 21, Rambelli (n. 40), 50f.

⁷³ Duckworth, N.R.C. pp. 118-21. This was one of the ways the dramatist could control off-stage space, time and action so as to avoid glaring inconsistencies which might distract the audience; cf. Handley, Entretiens Hardt 16 (1970), 9f.

⁷⁴ M. Johnston, Exits and Entrances in Roman Comedy (Geneva, N.Y., 1933), pp. 137-43, Duckworth, N.R.C. p. 119 with n. 43, Beare, Rom. Stage³ p. 181.

⁷⁵ E.g. in Ter. Phormio it seems probable that Terentian changes are responsible for the inconsistent exit and re-entry of Phaedria at 310 and 484 (Hermes 111 [1983], 450) and of Demipho at 314 and 348 (Lefevre, Der Phormio ' des Terenz und der 'Epidikazomenos ' des Apollodor von Karystos [Munich, 1978], p. 17). The hypothesis of Webster, S.L.G.C. p. 235, that Plautus omitted a scene in which Argyrippus was seen to return, would explain this anomaly but not the others; Rambelli (n. 40), 53f, also supposes the omission of such a scene as part of an unconvincing complex of changes.

⁷⁶ Handley, Men. and PI. p. 20 n. 11.

⁷⁷ Cf. Hough (n. 39), 26f, with n. 17. It is somewhat awkward that both Libanus and Demaenetus go to the forum, their exits being separated only by Demaenetus' short monologue (618-26), and Libanus' trip achieves nothing (Langen. PI. Stud. p. 99, Hough [n. 39], 23f., Rambelli [n. 40], 52). There is little to commend Hough's theory that in the Onagos Libanus remained on stage for a scene which Plautus cut; more likely is Rambelli's suggestion that he went into Demaenetus' house (coming out again in the next act—the rest of Rambelli's reconstruction of his movements in the Onagos is unconvincing).

⁷⁸ Hough (n. 39), 31.

⁷⁹ Goetz-Loewe (n. 34), xxiv, Hough (n. 39), 25 n. 12, Perna, L'orig, di Pl. p. 245 n. 4, Rambelli (n. 40), 53.

⁸⁰ Cf. Men. Sam. 368…, Leo, Der Monolog im Drama (Abh. Gött. Ges. Wiss., phil.-hist. K1. N.F. 10, 5, Berlin, 1908), p. 68, Fraenkel, Beobachtungen zu Aristophanes (Rome, 1962), pp. 22-6.

⁸¹ On Most. 687-717 (687 hue concessero), introducing aflagitatio, see Phoenix 39 (1985), 20f.; on Pseud. 410-44 (414 hue concedam), bridging a Greek act-division, ibid. 20, n. 45; on Poen. 203-329 BICS 35 (1988), 105-9; on Pers. 548-74 (548 tacill contemplemus) CQ 39 (1989), 396f.; on Cas. 424-503 (434 concedam hue) Lefevre, Hermes 107 (1979), 321f.; on Trin. 516-68 (517 hue concede) G. Jahn, Hermes 60 (1925), 33-49 (wrongly supposing a transposition from later in the play; cf. Jachmann, Pl. und Att. pp. 230f.); on Trin. 1008-23 (1007 hue concessero), a servus currens entrance, Gratwick, Mnem. 34 (1981), 331-5; on Men. 571-601 (570 huc concedamus) Stärk, Menaechmi pp. 91f.; cf. Barsby on Bacch. 404.

⁸² Cf. Barsby on Bacch. 1149-65. In Ter. Eun. 1053-60 the aside dialogue is between Thraso and Gnatho, the two characters introduced by Terence, on his own admission (30-3), into his main Greek model; cf. B. Denzler, Der Monolog bei Terenz (Zurich, 1968), pp. 53-5. Zwierlein (n. 5), 153, asserts that the similarity of technique in a 'doppelter Zweierdialog' in Poenulus I 2 and V 4 'kann nicht auf Plautus zurückgehen'. Why not? Zwierlein does not discuss the other Latin parallels which suggest that it does.

⁸³ Plautus used variants of the same idea on several other occasions: Pers. 265 hominibus (bobus Ritschl, alii alia) domitis … ex crumina, 317 boves … in crumina, Truc. 654f. minas ovis in crumina, 956 pecua … in crumina. On Plautine riddles see Fraenkel, P. im P. pp. 48-50 = El. Pl. pp. 45f., on Belebung des Unbelebten P. im P. pp. 101-10 = El. Pl. pp. 95-104.

⁸⁴ Traîna, Par. Pass. 9 (1954), 187 'la beffa dell'amor patetico', Comoedia—antologia delta palliata² (Padua, 1966), p. 66, P. Flury, Liebe und Liebessprache bei Menander, Plautus und Terenz (Heidelberg, 1968), pp. 84f., 91f.; contra Munari (n. 36), 13f., Perna, L'orig. di Pl. p. 207. There is probably also a Roman joke in 594; mater supremam mihi tua dixit, domum ire iussit (suppremum codd.: corr. Turnebus) can be taken as referring to Argyrippus' impending death, but Fraenkel recognized an allusion to the praetor's formula for announcing the end of the day's session in the comitium (P. im P. p. 43 n. 4 = El. Pl. p. 40 n. 4). The use of juridical terminology in an amatory context is characteristic of Plautus (cf. 13If., 607, N. Zagagi, Tradition and Originality in Plautus [Hypomnemata 62, Göttingen, 1982], pp. 106-31).

⁸⁵ N. W. Slater, Plautus in Performance (Princeton, 1985), p. 63.

⁸⁶ Fraenkel, P. im P. pp. 215-17 = El. Pl. pp. 206-8; cf. Jachmann, Pl. und Att. pp. 35-7.

⁸⁷ I follow the distribution of speakers printed by recent editors, although this is not entirely in accordance with the evidence of the MSS, and is open to dispute (cf. Bertini's apparatus criticus). It is hardly possible to differentiate the characters of the two slaves or their roles in this scene so as to assign each speech to one or the other with confidence. This uncertainty does not affect my argument.

⁸⁸ Cf. Liv. 7.2.7, Hor. Ep. 2.1.145f., J. Blänsdorf in E. Lefevre, éd., Das römische Drama (Darmstadt, 1978), p. 96 with n. 14, Stärk, Menaechmi, p. 73 with n. 324. See Fraenkel, P. im P. pp. 40If. = El. Pl. pp. 379f. on Plautine altercationes. With 618 circumsistamus compare Pseud. 357 adsiste altrim secus introducing a Roman flagitatio; cf. H. Usrer, Rh. Mus. 56 (1901), 1-28 = Kl. Schr. iv.356-82.

⁸⁹ Fraenkel, P. im P. p. 49 = El. Pl pp. 45f.

⁹⁰ Cf. 45, M.G. 1038, Stich. 469, Trin. 436f., 1152, Hor. S. 2.8.75f.

⁹¹ Cf. 592f., P. J. Enk, Plauti Truculentus (Leyden, 1953), on True. 126, 259, Hough, AJP 66 (1945), 282-302.

⁹² Cf. 43, M.G. 1039f.

⁹³ Havet (n. 36), 99, implausibly supposes that Argyrippus learnt of Diabolus' offer in a lost Argyrippus/lena scene.

⁹⁴ Cf. Munari (n. 36), 15, against Hough (n. 39), 21f. This does not exclude the possibility of some superficial Plautine additions; the play with perdo/pereo may well be one (~ 243f.; cf. the play with perire in True. 45-50, 707, Flury, Liebe, pp. 81, 84f.), although it is not inconceivable that there was something similar in the Greek.

⁹⁵ Cf. 677 delusisti, 679 delude, 711 delusistis, 730 ludatis, 731 satî iam delusum.

⁹⁶ Cf. Bertini on 618.

⁹⁷ Rambelli (n. 40), 76 supposed this passage a Plautine insertion, but for inadequate reasons.

⁹⁸ Havet-Freté suppose a lacuna after 638.

⁹⁹ 648c ecquid est salutis? would follow well after 638 ne formida. In its Plautine context the phrase makes sense if Argyrippus supposes the purpose of the slaves to be to devise a scheme for his benefit (Ussing on 639, Havet-Freté on 638).

¹⁰⁰ Cf. Men. 571-87, Fraenkel, P. im P. pp. 159-62 = El. Pl. pp. 152-4, Stärk, Menaechmi, pp. 90f., S. J. Rosivach, Maia 35 (1983), 86f. with n. 11. Another scheming aside dialogue is Bacch. 1149-54, in a four-speaker finale (cf. Barsby ad loc).

¹⁰¹ Cf. Handley, Men. and Pl. pp. 14, 17.

¹⁰² Cf. S. Lilja, Arctos 16 (1982), 57-64, D. Hughes, Rh. Mus. 127 (1984), 51 η. 8. Bertini is wrong to follow the MSS. in distributing 643 between Argyrippus and Leonida, and in rejecting Pylades' correction of haec to hic. After 641 non omnia eadem aeque omnibus … suavia, 643 ego forms the necessary antithesis to 642 vobis and must be spoken by the same speaker; and in the context of Argyrippus' suggestion that the slaves embrace each other a reference to Philaenium is out of place.

¹⁰³ Cf. 689f., Cas. 739, Most. 406-8, Rud. 1265f, 1280, E. Segal, Roman Laughter (Harv. stud, in comp. lit. 29, Cambridge, Mass., 1968), p. 106 et passim.

¹⁰⁴ See references in n. 46 above.

¹⁰⁵ The view of Rambelli (n. 40), 63f., that Philaenium does not actually kiss Leonida seems refuted by 679 age sis tu in partem … amplexare hanc.

¹⁰⁶ Cf. p. 16 above with n. 88, Segal, Rom. Laughter, p. 105, Pers. 1-6 ~ 7-12 (Fraenkel, P. im P. pp. 227f. = El. Pl. pp. 218f.), 168-82 ~ 183-99 (Hughes, [n. 102], 54f.).

¹⁰⁷BICS 35 (1988), 104f., Gratwic k C.H.C.L. ii.110 n. 2; cf. Fraenkel, loc. cit. (n. 46). Amatory language is similarly parodied in Pseud. 64-73 (cf. 1259-61, Cas. 134-8, 837, Jachmann, Philologus 88 [1933], 451). See Perna, L'orig. di Pl. ch. 7 on 'la parodia dell'amore' in general. Another exoratio of a slave by a free man is Epid. 728-31 (cf. Segal, Rom. Laughter, pp. 109f., 122f., E. Fantham, Pap. Liv. Lat. Sem. 3[1981], 22f.).

¹⁰⁸ Fraenkel, P. im P. p. 116 = El. Pl. p. 110; cf. Segal, Rom. Laughter, pp. 108f.

¹⁰⁹ Cf. 670f., Epid. 728f., Pseud. 1285ff., Segal, Rom. Laughter, ch. 4.

¹¹⁰ Cf. Aul. 637, Capt. 867, Poen. 612, Rud. 1074, C. J. Mendelsohn, Studies in the Word-Play in Plautus (Univ. of Pennsylvania, Ser. in Phil, and Lit. XII. 2, Philadelphia, 1907), p. 84, W. Stockert, T. Maccius Plautus Aulularia (Stuttgart, 1983), on Aul. 637 and references in n. 102 above. I am not persuaded by suggestions that there are other obscene allusions in the passage (cf. Bertini on 702-10. Segal, Rom. Laughter, p. 203 n. 20. Mendelsohn, op. cit. 26, sees play with Argyrippus' name in the action of this scene; but, if the horseplay was in the Onagos, this could have inspired the Plautine character's name (cf. Ussing [n. 48], i.349).

¹¹¹ Rambelli (n. 40), 76f., noted that there is no sign of Philaenium's presence during the actual handing over of the money; but in a short passage this fact would by itself not necessarily be significant.

¹¹² See above p. 17 on Pers. 777ff., BICS 32 (1985), 83f. on M.G. 1394ff., Rh. Mus. 133 (1990), 287 on Poen. 1120ff., Barsby on Bacch. 1120ff., Sandbach (n. 71), 199-204 on Ter. And. 904ff., H.T. 1045ff., Eun. 1025ff., Ad. 958ff.; cf. Cas. 963ff., Curc. 599ff., Trin. 1125ff., Truc. 893ff.

¹¹³ Munari (n. 36), 20, Rambelli (n. 40), 64, and references in Bertini on 930.

¹¹⁴ Johnston, Exits, p. 144. It would be very rash to assume that drinking scenes on stage never occurred in Greek New Comedy, but those supposed by Webster, Stud. Men. p. 112 n. 1, are quite uncertain.

¹¹⁵ Op. cit. (n. 31), 1074f.

¹¹⁶ See above p. 17. 767 in summo (cf. 771^a) suggests the seating arrangements of a Roman triclinium (E. Woytek, T. Maccius Plautus Persa [Öst. Ak. Wiss., phil.-hist. Kl., Sitzungsber. 385, Vienna, 1982], ad loc).

¹¹⁷ Cf. Bacch. 1074, R-E vii.510 s.v. triumphus. Fraenkel, P. im P. pp. 234-40 = El. Pl. pp. 226-31, showed that in 753-7 Toxilus uses the traditional language of Roman triumphs.

¹¹⁸ Hughes (n. 102), 46-57.

¹¹⁹ Gaiser (n. 31), 1084; but cf. Fraenkel, El. Pl. p. 443.

¹²⁰ Weise on 828f., K. Kunst, Studien zur griechischrômischen Komödie (Vienna, 1919), p. 156 and Webster, S.L.G.C. p. 237, amongst others, suppose Plautus' scene played with the banqueters off stage and seen through a partly opened door, but nothing in the text justifies this interpretation (Johnston, Exits p. 144, Beare, Rom. Stage³ p. 179).

¹²¹ R. C. Flickinger, The Greek Theater and its Drama⁴ (Chicago, 1936), pp. 237-43.

¹²² Cf. Ussing on 821, Brasse (n. 38), 81, A. Thierfelder, De rationibus interpolationum Plautinarum (Leipzig, 1929), pp. 129f., Perna, L'orig. di Pl. p. 245 n. 4. Ussing and Leo follow Weise in deleting 828f., but this would not change the basic situation; moreover there are significant verbal parallels between these lines, Most. 308f. age accumbe igitur. cedo aquam manibus, puere, appone hie mensulam and Pers. 768-9^ahoc age, accumbe … date aquam manibus, apponite mensam.

¹²³ D. Bain, Actors and Audience (Oxford, 1977), pp. 162-71.

¹²⁴ There are also other echoes of earlier parts of the play: 834 merito tuo facere possum ~ 737 meritissumo eiius quae volet faciemus, 835f. nolo ego metui, amah mavolo, mi gnate, me aps te ~ 67 volo amari a meis (cf. 77).

¹²⁵ It seems not unlikely that the terms demanded by Demaenetus in return for handing over the 20 minae to Argyrippus, the enjoyment of Philaenium for one night, are a Plautine invention. This motif is introduced suddenly in 735f. and not prepared in I 1 (Langen, Pl. Stud. p. 104, Hough [n. 39], 22f., Munari [n. 36], 23, Webster, S.L.G.C. p. 234, Perna, L'orig, di Pl. p. 247). The fact that Demaenetus had aided and abetted the trick to misappropriate Artemona's money (cf. 814f.), aggravated by his participation with Argyrippus and Philaenium in a drinking party in Cleareta's house, would be a sufficient cause for Artemona to take revenge on him. That Artemona's revenge probably does derive from the Onagos (if not the Prügelmotif of 936 and 946; cf. E. Schuhmann, Philologus 121 [1977], 62f.) is implicit in my arguments above. After the scheme, which forms the main strand of the plot, has been brought to a successful conclusion and the young lovers have been reunited, the revenge of Diabolus and Artemona at Demaenetus' expense provides an amusing ending to the play; similarly in Terence's Phormio the parasite Phormio takes his revenge on Chremes by reporting him to his wife (Kunst [n. 120], 154, Webster, S.L.G.C. p. 237). It is appropriate that Demaenetus should be made to pay for his part in the deception of his wife; and it is well prepared by the characterization of Artemona in I 1.

¹²⁶ Schuhmann (n. 125), 55-64, observes that the expression by a husband of his distaste for sexual relations with his aging uxor dotata (894f.; cf. 872-4) and of his wish for her early death (901, 905; cf. 909) are motifs found elsewhere in Plautus (e.g. Most 703-7, Trin. 41, 51) but not in Menander or Terence, and shows that the characterization of the Plautine uxor dotata is in large measure to be attributed to the Latin adapter. Stark, Menaechmi, pp. 31-6, 47-59, plausibly sees the influence of the pre-literary Atellan farces in the crude comedy which Plautus regularly attaches to the figures of the senex amator and uxor dotata, noting that their typically Plautine features can more easily be paralleled in the literary Atellana than in Greek New Comedy; but some allowance must be made for the influence of the palliata on the literary Atellana. Gratwick (n. 81), 341 n. 4, believes the adventitious use of the palla motif in 884-6 (cf. 929f.) a Plautine borrowing from the Menaechmi, where the motif plays a central rôle in the plot (cf. Ε. Fantham, C. Ph. 63 [1963], 176f.); and he may be right. It seems likely, however, that the motif was a stock one in New Comedy (pace Stärk, Menaechmi, pp. 14f.), even if it is not actually attested in the Greek fragments; Plautus could have used it without having a specific model. The verbal similarities which Gratwick sees between this scene and the Menaechmi are hardly sufficient to prove the priority of the Menaechmi.

¹²⁷ Gestri (n. 45), 205; cf. Jachmann, Pl. und Att. p. 69.

¹²⁸ Gell. NA 3.3, Duckworth, N.R.C. pp. 50f.

¹²⁹ Grauert (n. 4), 205, Körte, Bed. phil. Woch. 36 (1916), 981, Schaaf, Miles, pp. 378-80.

¹³⁰Men. 7-9, Stich 446-8.

¹³¹ Leo, P.F² p. 111, Fraenkel, P. im P. pp. 400f. = El. Pl. pp. 378f, G. Williams, JRS 48 (1958), 18, P. P. Spranger, Historische Untersuchungen zu den Sklavenfiguren des Plautus und Terenz² (Stuttgart, 1985), p. 117, Segal, Rom. Laughter passim, Gaiser (n. 31), 1079, 1107.