Amphitruo, Bacchae, and Metatheatre

The Amphitruo has always been considered a distinctive play within the Plautine corpus, and various, sometimes overlapping, attempts to explain---or explain away---its otherness have been made. Some have tried to find its origins in mythological travesty of the period of Middle Comedy. Others have emphasized the tragic character of the play's heroine, Alcumena. Interest in psychological approaches brought interpretations which celebrate the play as an enquiry into the nature of identity. Finally, work on the Saturnalian and festival background of Roman comedy has singled out the play for its uniquely successful adultery.1
All these explanations have added to our understanding of this delightful yet disturbing play, but none has to my mind done it full justice. Two problems, one purely literary, the other one of literary history, remain essentially unresolved. We might define the central literary problem of the Amphitruo as explaining the genre of this play, which Mercury in his prologue labels a tragicomoedia. We must determine whether the attempt to fuse the serious world of tragic myth and the farcical world of the Roman comedy tradition results in more than a burlesque alternating between the two. Its Roman comedy elements need no demonstration here; its transformation of tragedy may. The problem of literary history, to which I will offer only a tentative new answer, is of course the old one of source or sources. Unlike his practice in some other plays, here Plautus allows Mercury in the prologue to tell us nothing of the original of this play.
I propose to attempt answers to both the literary and historical problems by re-opening a question first raised by Zeph Stewart some thirty years ago in a penetrating article, in which he demonstrated a number of connections between the Amphitruo and the Bacchae of Euripides. He was able to trace several links of narrative and setting between the two plays, although the verbal similarities remain slight.2 He therefore did not suggest a direct line of influence but once more raised the possibility of an intermediary burlesque of the Bacchae to be ascribed to Rhinthon of Tarentum.3
Certain similarities between the Bacchae and the Amphitruo had been noted even before Stewart,4 but the assumption that a specific Greek comedy (or more than one) was the only possible source for a Roman comedy has until recently been virtually unassailable. K. M. Westaway's suggestion as long ago as 1917 that "a direct burlesque of Euripides by Plautus" might lie behind the Amphitruo simply found no hearers.5 Then just a few years ago, Eckard Lefèvre, employing the traditional analytic approach to Plautus, offered ample grounds for questioning the existence of any Greek New or Middle Comedy original behind the Amphitruo; he went on to suggest that Plautus was in large part parodying a recent Roman version of Euripides' Alcmena.6
Once we abandon the notion that only a previous comedy could have influenced the Amphitruo, however, a reconsideration of the Bacchae as a possible source, especially in light of recent work on its theatricality, is in order. Both Helene Foley and Charles Segal7 have contributed to our understanding of the self-conscious theatricality of the Bacchae. Foley brilliantly explicates the play as a theatrical demonstration of the theatrical means by which theatre's god, Dionysos, is both apprehended and worshipped. She sees the play ultimately making "a strong claim for art's ability to represent a reality inaccessible to ordinary human sight." Segal sees the play as breaking down the limits which had defined fifth-century tragedy and suggests Euripides "raises and explores the question of how the falsehood of (dramatic) fiction can bring us truth, how by surrendering ourselves and losing ourselves to the power of the imagination we can in some measure find ourselves...." Both emphasize the highly self-conscious nature of the play and its exploration of its own tragic form.8
In this light, the connections of the Bacchae and the Amphitruo grow clearer. The Amphitruo too is a play about the nature of theatre. Though Plautine theatre constantly concerns itself with roles, role-playing, and the nature of illusion, no play does so more explicitly than the Amphitruo with its role-playing gods enacting a Casina-like sex farce. Despite the vast superficial dissimilarities, both the Amphitruo and the Bacchae are plays pre-eminently conscious of their own theatricality. Both Foley and Segal have demonstrated this point and its implications for Euripides' play. Here I will examine both the theatricality of the Amphitruo and the nature of Plautus' contact with, and understanding of, Euripides' masterpiece.
I propose to unite existing insights and, I believe, solve many of the outstanding questions about the Amphitruo by interpreting it as a metaplay, an example of metatheatre. I define metatheatre as theatrically self-conscious theatre, theatre which is aware of its own nature as a medium and capable of exploiting its own conventions and devices for comic and occasionally pathetic effect.9 It is this self-consciousness (alien to Greek New Comedy) which, in light of the connections pointed out by Stewart, links the Amphitruo directly to the Bacchae of Euripides.10
My method will be performance criticism, an analysis of the Amphitruo in its natural environment, the theatre, and in the only way it existed, through performance. Therefore the approach will be strictly linear, as was the performance. The dynamics of staging and stage pictures also create meanings that may not be explicit in the text; these will be a constant concern as well.11


My prologue done, let us turn to that of the Amphitruo itself. Mercury appears, disguised as the human slave Sosia (1-152). His prologue opens with a sixteen-line contract between players and audience: he will help them in their business dealings, if they will give ear to his play (15, ita huic facietis fabulae silentium). There is far more here than bombastic parody of the Roman language of contract. It is an induction, a flatteringly phrased invitation into the world of the play which acknowledges the audience's part in creating that world. We might compare the Scrivener in Jonson's Bartholomew Fair who makes a contract between playwright and audience in the induction of that play. A play is not a play unless someone watches it; the audience, Mercury acknowledges, are as important as the actors. While Plautus often draws his audience in gradually, the explicit language of contract here suggests that Plautus is more than usually concerned with his audience's expectations of his play; like Jonson, he must re-make his audience to suit his play's unusual conception.12
Mercury now tells us that he is only the messenger of Jupiter, who is also flatteringly afraid of the audience's displeasure (26-27) and seeks to placate them. And why? This Jupiter is a human Jupiter (28), merely an actor; here Plautus begins to play with the very idea of roles and will continue to do so throughout the prologue. Some comment on roles and role-playing is typical of Plautus but here goes much further than usual. Jupiter is compared to a character in tragedies (41) and also called an architectus (45), which will remind any spectator familiar with Plautus of another Plautine role, that of the architectus doli in Plautus' many comedies with clever slaves.
Mercury himself first raises the question of genre (which will preoccupy us for some time), in a way that leaves us, the audience, confused as to what sort of play we shall see. He first announces a tragedy, then pretends to change the nature of the play in the face of a negative reaction from the crowd:
post argumentum hujus eloquar tragoediae.
quid? contraxistis frontem quia tragoediam
dixi futuram hanc? deu' sum, commutavero.
eandem hanc, si voltis, faciam (jam) ex tragoedia
comoedia ut sit omnibus isdem vorsibus.
The audience reaction is perfectly predictable; they came to laugh at a comedy by Plautus, not for more serious fare.13 What sort of play, though, can be tragedy or comedy with exactly the same verses?14 It is this curious property, more than any ironic scruples against gods appearing on the comic stage (note that Mercury is more offended at his own slave role than at seeing Jupiter disguise himself), which causes the play to be a tragicomoedia (59-63). What impact would the use of this term have had on Plautus' original audience? It is a typically humorous Plautine nonce word, of course, but it must have aroused curiosity as well. Indeed, the term once coined seems to be thrown away, and the play becomes a comedy again (88; 97), but audience expectations have nonetheless been unsettled by the discussion.15
Before we turn to the argumentum we should consider the impact of the first 96 lines in performance. We begin with a contract between stage and audience, the parties and their roles carefully defined. Then Mercury, in a bait-and-switch tactic, proposes to change the nature of the transaction from a comedy to a tragedy. Though he seems to return to the original bargain, we are more wary: our generic expectations as an audience have been subtly disturbed. We have had considerable discussion of actors and acting, establishing one of the themes of the play. All this might seem tedious foolery to a reader eager to get on with a narrative, but in the theatre, in performance, it has built a framework though which to interpret the rest of the play.
Mercury now narrates the background of the story, with only a few costume details (117; 142-47) to remind us the story is to be enacted. The theme of appearance and reality, so essential to the meaning of this play, is introduced by the use of the term imago (121; 124; 141) and developed by the description of Jupiter as vorsipellem(123), which we should not translate as simply "werewolf" but "skin-changer," for it implies not just a duality of shapes, but the polymorphous capability to put on any shape whatever. With these last few hints (Mercury's final word, histrioniam, reminds us of the fictive nature of the world we are about to enter) Mercury yields the stage to Sosia.
Now begins a scene often classified as an eavesdropping scene: Mercury in concealment listening (and commenting), Sosia soliloquizing. We need to understand the visual and theatrical dynamic of this scene before we even begin to examine its narrative content. As with any eavesdropping scene, the form is that of the play-within-the-play, here certainly in rudimentary fashion.16 Sosia is not really in direct communication with the audience in this soliloquy. Between him and the audience stands Mercury, whose comments aside interpret and color everything Sosia says.
Sosia's first few lines strike typical notes for the slave character: he boasts of courage (153) but soon displays fear of the nightwatch (155-63) and complains of harsh service under his master (166-75). Any sympathy we might feel for his condition is quickly turned to laughter by Mercury's aside:
satiust me queri illo modo servitutem:
hodie qui fuerim liber, eum nunc
potivit pater servitutis;
hic qui verna natust queritur.
The humor, of course, lies in Mercury's double role, god and slave.
Sosia's reply shows us something about the style of Plautine playing as well. He picks up the term verna, which Mercury has used opprobriously, and flaunts it like a badge of honor, at the same time wondering whether he should give thanks to the gods for his safe return:
sum vero verna verbero: numero mihi in mentem fuit
dis advenientem gratias promeritis agere atque adloqui?
In what sort of play world does this exchange take place? Realistically speaking Sosia has not overheard Mercury's aside; he displays no awareness of the other figure on stage until line 292. Nor is this a verbal irony of which the speaker, Sosia, is unaware. It is a tiny skirmish to regain theatrical advantage, a struggle for audience attention and sympathy, in which the actors, without regard for the overall illusion of the story, engage. It is a style of playing we may term non-illusory,17 in that the action does not take place within a plane of illusion, but appeals directly to the spectator for approval.
If nothing else (and the speech has many merits in its own right) Sosia's battle soliloquy demonstrates his ability to improvise. Note that Sosia himself was not present at the battle, having been too busy hiding (198-99). He must therefore rehearse a fictional account of the battle before he delivers the same to Alcumena. This "rehearsal speech" has been discussed as tragic parody before.18 We need only note how Sosia's preparations (200 simulabo, 201 fabularier) indicate that he himself approaches it as a theatrical fiction, one whose (almost accidental) truth Mercury later confirms in an aside (248-49). After Sosia's battle narrative concludes, Mercury announces his plan to delude Sosia (265, certum est hominem eludere) and conquer him with his own weapon, malitia (269). There is a short delay, though, while Sosia, through staring at the sky, introduces both the Nux Makra motif and the theme of drunkenness, which is his explanation for the phenomenon; the god Nocturnus has gotten drunk and is sleeping off his hangover.19
Sosia at last sees Mercury standing before the house (292), and the long-awaited confrontation is imminent. All of this time Mercury has in effect stood outside the play (which contains Sosia), remaining in close communication with the audience. Before stepping into the play world he hesitates long enough to acquaint us, the audience, with how he intends to play the scene: he will delude Sosia, of course (295, deludam), but by the seeming inversion of the scene just concluded. Mercury will pretend to be ignorant of Sosia's presence and play the victim's role in an eavesdropping scene:
clare advorsum fabulabor, auscultet hic quae loquar;
igitur magi' modum majorem in sese concipiet metum.
Mercury thereby seems to surrender his theatrical power over the scene, but in fact he increases it through changing places with Sosia. With his claim to be the real Sosia, the actor here must now maintain and make clear to the audience three superimposed roles: he is playing Mercury playing Sosia playing victim in an eavesdropping scene. On top of this, as he explicitly acknowledges in ironic parallel to the emphasis on fictional creation in Sosia's rehearsal speech (198 mendacium, 200 simulabo, 201 fabularier), Mercury himself will create a fiction through speech (300 fabulabor).20 Mercury's bombastic threats and eagerness for a fight have their intended effect on Sosia, who nonetheless manages to make a joke of most of them.21
The scene which follows is the locus classicus of the Doppelganger theme, but it need not occupy us in detail. Mercury not only by physical resemblance, but far more importantly by detailed knowledge of Sosia's life, succeeds in causing Sosia to doubt his own identity. The only metatheatrical irony in the scene is the exchange of oaths:
So. per Jovem juro med esse neque me falsum dicere.
Mer. at ego per Mercurium juro tibi Jovem non credere;
nam injurato scio plus credet mihi quam jurato tibi.
Sosia retreats in disarray. The comedy of the scene is undeniable, but the undertone of fear is very strong as well. There is real pathos in Sosia's use of tricolon in his parting speech:
abeo potius. di inmortales, opsecro vostram fidem,
ubi ego perii? ubi immutatus sum? ubi ego formam perdidi?
an egomet me illic reliqui, si forte oblitus fui?
Dreams and drunkenness are no longer a sufficient explanation for these phenomena; it is now "a mad world, my masters"---precisely Mercury's plans for both master and man in his ensuing soliloquy (470-71 erroris ambo ego illos et dementiae / complebo...).
As Jupiter and Alcumena now come out of the house, Mercury through his asides to the audience once again interprets and comments on the scene. His joke about Juno's jealousy (510-11) reminds us of Jupiter's double role as Amphitruo. When Mercury does choose to enter the scene, he changes his role from that of the slave (which we would expect) to parasite:
Mer. accedam atque hanc appellabo et subparasitabor patri.
Jupiter, however, does not welcome the interruption and rebuffs Mercury, who mournfully comments:
Mer. nequiter paene expedivit prima parasitatio. (521)
Having failed in the role of the parasite,22 Mercury will now have to return to his slave role.
Alcumena, somewhat placated by the gift of the golden cup which will figure so prominently in proofs of identity later in the play, departs, leaving the stage to Jupiter and Mercury. Jupiter commands an end to the long night and follows Mercury off.
Visually and verbally, Jupiter and Mercury's exit and Amphitruo and Sosia's entrance form a chiasmus. On the divine level it is the slave who is taking the lead and the master who follows (550 supsequar). On the human level, Amphitruo is in complete control at the scene's beginning, and it is the unfortunate Sosia who tags along behind (551 sequor, supsequor te). Amphitruo cannot believe the confused and confusing report of events at the house which Sosia has given him and suspects drink (574) or disease (581) as the cause of the disorder. Madness is also hinted at: 585 dictis delirantibus. At the heart of Amphitruo's rage lies the suspicion that Sosia is playing him for a fool.23
Apart from drink, madness, and disease, Amphitruo can now think of only one more natural explanation for Sosia's confusion: he must have dreamt it all. Sosia eloquently denies this:
Am. ibi forte istum si vidisses quendam in somnis Sosiam.
So. non soleo ego somniculose eri imperia persequi.
vigilans vidi, vigilans nunc (ut) video, vigilans fabulor,
vigilantem ille me jam dudum vigilans pugnis contudit.
Alcumena's soliloquy which follows is a study in miniature of tragicomedy. Visually almost grotesque in her pregnancy, she nonetheless conveys a touching affection for her husband. Once again the comments of an eavesdropper, here Amphitruo, help shape audience response: she is the virtuous, loving wife.24
Amphitruo's strange questions and manner when he appears make Alcumena suspect she is being made fun of (682 deridiculi gratia), and she playfully threatens to turn the tables on him (694 deludam contra lusorem meum). She is indeed too kind-hearted to do much of this, even if Amphitruo did not begin to bluster and confuse the issue even more. Once again Amphitruo offers the same naturalistic explanations: she is mad (696 deliramenta) or dreaming (726 in somnis fortasse). Sosia is the first to suggest a supernatural explanation with his implied comparison of Alcumena to a bacchant (703). Indeed, Sosia is often a step ahead of his master, whether with advice or explanation. He is the first mortal, for example, even to hint at the real state of affairs, when he suggests that all now have been doubled (786 omnes congeminavimus). Amphitruo's suspicions of his wife's chastity, aroused by her tale of a previous visit by himself, will admit of nothing but an examination by witnesses, and he goes off to seek his fellow voyager, Naucrates, as proof of his own version of events.25
Jupiter now returns for yet one more prologue. The plethora of prologues provided to explain (and excuse?) the action is perhaps indicative of the improvisatory nature of this play on the divine level. He begins with a playful restatement of who he is:
ego sum ille Amphitruo, quoii est servos Sosia,
idem Mercurius qui fit quando commodumst,
in superiore qui habito cenaculo,26
qui interdum fio Juppiter quando lubet;
huc autem quom extemplo adventum adporto, ilico
Amphitruo fio et vestitum immuto meum.
Jupiter announces that he will take up his guise as Amphitruo again27 in order to extricate Alcumena from her difficulties. His theatrical motivation (868 ne hance incohatam transigam comoediam) is equally important: like a typical divine prologue figure he looks at the play from the outside and shows his concern that the audience should find the play complete and satisfying.
Alcumena requires some soothing before she will respond to Jupiter's overtures. The king of the gods begs (932 obsecro) Alcumena to forgive him and excuses "his" behavior as a joke (916-17 joco illa dixeram dudum tibi, / ridiculi caussa; cf. 682). Even this is not enough (rightly so!), and Jupiter only saves himself through the rich irony of a metatheatrical curse, which the kind-hearted Alcumena rushes to avert:
Ju. ego si fallo, tum te, summe Juppiter,
quaeso, Amphitruoni ut semper iratus sies.
Al. a, propitius sit potius.
The real Sosia arrives and speaks with Jupiter, but no Comedy of Errors byplay ensues. He is merely dispatched to prepare for the sacrifice. Alcumena takes her departure (972), leaving Jupiter to address the unseen Mercury much as Prospero sometimes speaks to an unseen Ariel (974-83). Mercury is instructed to carry on with the play (980 volo deludi illunc), while Jupiter departs to sacrifice---to himself (983)! I think there is more here than wonderfully narcissistic blasphemy. It is virtually a metaphor for the play: Amphitruo's delusions and Alcumena's virtue are Jupiter's offering to himself through the medium of the play. As actor within the play he enjoys the beautiful Alcumena, while as playwright/director, standing outside the play, he creates a farce of mistaken identity for both his own and the audience's pleasure.
Mercury, in keeping with the usual multiplicity of leading slave roles now enters in the role of servus currens. Indeed, he explicitly compares himself to this theatrical figure:
concedite atque apscedite omnes, de via decedite,
nec quisquam tam au(i)dax fuat homo qui obviam opsistat mihi.
nam mihi quidem hercle qui minus liceat deo minitarier
populo, ni decedat mihi, quam servolo in comoediis?
Mercury reminds us that he has been called by Jupiter and ordered to delude (forms of deludo: 997, 998, 1005) Amphitruo. This whole speech is made possible only by an acute consciousness of role and role-playing within the context of traditional Roman plots. When Mercury on entering compares himself to the servus currens, he also reminds us of the stock situations in which the servus currens appears: to announce the safe arrival of a ship or warn of the approach of an angry senex (988). When he discusses his role as helper to the amorous Jupiter, he explicitly reminds us of his parasite role (993 subparasitor). Now he proposes to adopt the role (999 adsimulabo) of the drunken slave who abuses his master. Most importantly, he makes a direct appeal for audience involvement in the play. In his first explicit direct address to the audience since the original prologue, he promises them a good show:
nunc Amphitruonem volt deludi meu' pater: faxo probe
jam hic deludetur, spectatores, vobis inspectantibus.
Then he acknowledges the need for the audience's help; once again (as in the prologue), his role is to act, while theirs is to listen:
sed eccum Amphitruonem, advenit; jam ille hic deludetur probe,
siquidem vos voltis auscultando operam dare.
Amphitruo returns, having failed to find Naucrates, and finds his own doors barred against him. Moreover, Mercury, in the guise of Sosia, is guarding the gate and begins the promised scene of abuse. It is in this scene that a well-known but frustrating lacuna in the text occurs. Attempts have been made to fill this gap, but even the most detailed reconstruction of events will not give us a very clear picture of the staging.28 The fragments only show us that as Mercury and Jupiter resist Amphitruo's attempt to enter themes already established in the play (such as madness) recur.29
When the text resumes, Jupiter's abuse of Amphitruo is just ending, and Blepharo the pilot, to whom Amphitruo appealed for proof of his identity, is retiring in disorder. Amphitruo finally suspects a supernatural origin (1043 Thessalum veneficum) for the problems of his house, but in his rage at being deluded (1041, 1047), he can think only of immediate revenge and rushes the door.
It would be of the greatest value to know how the effect of Jupiter's thunderbolt was produced on the Roman stage, but we simply do not. The next scene opens with Amphitruo prostrate and the significantly-named Bromia30 bursting forth from the house with a report of the miraculous events within. She spots Amphitruo, revives him, and finishes her tale. Jupiter appears as deus ex machina to reassure Amphitruo and interestingly to dismiss from his mind the idea of consulting Tiresias or other soothsayers.31 On this rather abrupt note, the play simply ends.


At this point the connections between the Amphitruo and the Bacchae pointed out by Stewart should be clear.32 Both plays contain a god in disguise, and that disguise is repeatedly commented on. Verbal and visual ironies, flowing from this disguised god, permeate both plays. Both possess Theban heroes, who resist the invasion of their homes by what they characterize as foreign magicians. Finally and significantly, the language of both plays is dominated by themes of madness and drunkenness. Plautus has developed one other theme far more extensively: that of waking and dreaming. This theme is essentially confined to the inner, human play in the Amphitruo. While Mercury may threaten to drive the human protagonists mad (470-71) and play a drunken slave role himself (999 adsimulabo me esse ebrium), he and Jupiter never speak in the language of waking and dreaming. It is always the human characters who obsessively ask each other whether they are really awake (e.g., Amphitruo to Alcumena, 697) or protest their own wakefulness (most emphatically, Sosia, 623-24). Amphitruo in particular is eager to ascribe the supernatural events reported to him to the world of dreams (e.g., to Sosia, 621 si vidisses quendam in somnis Sosiam...). The use of this theme by Sosia and Mercury in the opening scene foreshadows its structural function in the rest of the play. Sosia fears that the unknown stranger before the house will put him to sleep with his fists (298 hic pugnis faciet hodie ut dormiam). Indeed that is just what Mercury proposes to do:
quid si ego illum tractim tangam, ut dormiat?
And in that sleep comes a dream of another man with his face, name, and memories. Here we see Abel's epigram exemplified: for the human characters of this play "the world is a stage, life is a dream."
The visual similarity of the two plays, an essential element of meaning apparent only in performance, reinforces and makes specific the thematic connections we have noted. The Bacchae takes place before the palace of Pentheus, the Amphitruo before the aedes of the title character.33 As Stewart has noted,34 the action of the Amphitruo, as in no other surviving Latin or Menandrean comedy, is split between two levels in the scenes where Mercury and Jupiter drive Amphitruo from his own door. Such confrontations are rare enough in tragedy, and given the other influences of the Euripidean model, the staging of this scene in the Amphitruo seems an obvious and brilliant comic variation on the appearance of Dionysos at the end of the Bacchae above the palace of Pentheus.35
The confrontations of doubled characters in the Amphitruo form powerful visual parallels to the Dionysos/Pentheus scenes in the Bacchae. As Foley and others have pointed out, Dionysos and Pentheus are visual doubles of each other, possibly even to the extent of themselves embodying the twin suns Pentheus says he sees through their costuming in saffron robes.36 Both wore youthful, beardless masks which, while not identical, would closely resemble each other and be set apart from other masks in the play. Certainly, there are many other twins in Plautus (and other Roman comedies, to judge from titles), but none are divine/human pairs for which the Bacchae offers so powerful a visual parallel.
These visual similarities are undergirt by parallels of metatheatrical structure between the Bacchae and the Amphitruo. We will first examine the functions of the prologues, then the roles of the gods as directors of the play being produced, then finally the genre and structure of those plays that they create.
Both prologues have an expository function, but that is not their first concern. The initial emphasis is on role and disguise and the audience's expectations based on those roles. Foley notes that Dionysos redundantly stresses his human disguise (4-5, 53-54) in the play,37 as though fearful that the audience will not accept his rupture of the stage conventions. Certainly the Greek audience expects that the gods will confine themselves to the theologeion and not share the skene with the mortal characters. Just so, Mercury seems deeply concerned with overcoming his audience's similar expectations, and repeatedly emphasizes his (and Jupiter's) human disguise.38 Where Dionysos stretches the genre of tragedy, Mercury must transform the genre in which he performs and his audience's expectations at the same time. This he accomplishes through the slippery nature of his contract with the audience, as we have seen above.
Next, the gods of both plays set about the creation of a play. Foley shows how Dionysos functions as stage director of the Bacchae.39 He creates a play-within-the-play, whose tragic hero is Pentheus, victim of the illusions the theatre god weaves. Note that we as audience, despite our deep involvement in the tragedy of Pentheus, stand outside the illusion of this play-within-the-play. We can keep separate the divine and human characters, whereas the human participants are trapped inside the illusion. An example of this division between internal and external audiences of the play's events is the destruction of Pentheus' palace in the play. As Foley has shown,40 this was not done with primitive special effects. In a bold stroke, Euripides allows his audience to stand outside the illusion and experience the event only through the description given by the chorus, for whom it is real.
In the Amphitruo the role of stage director has been split in the process of multiplying the comedy. If we wish to continue the modern analogies, we might think of Jupiter as the producer and Mercury as the stage director working under him.41 Jupiter provides the outline of the plot from his own desires and supplies the power whereby those desires are fulfilled (e.g., the prolongation of Night), but it is Mercury who oversees the details of stage business that create the play. One again, we as audience stand outside the illusion of this play-within-the-play as exemplified by the tassel and wings by which we can distinguish divine and human doubles. Unlike the unfortunate Alcumena, we are never confused as to which is Jupiter and which Amphitruo.42
Doubtless the play that Dionysos creates in the Bacchae is poetically richer. It is an explication through instruments, song, and dance of the theatrical nature of Dionysos worship.43 Plautus' play-within-the-play relies rather on the power of words and acting to create theatre. He celebrates the power of fiction in performance. Words spoken within the theatre space, in character, words that are in essence lies can create a new reality. Characters continually in this play accuse each other of lying (e.g., Alcumena on Sosia, 755 falsum dicere), yet the delicious irony is that the accused are nearly always speaking the "truth"---the truth of theatre. In one or another fictional world all these things are true. It is theatre itself, we are led to conclude, that is the lie---but a good lie, an entertaining lie, a comic and refreshing lie.
It is this comic affirmation of the fictive power of lies to protect one from reality that accounts for the diametrically opposed resolutions of the Amphitruo and the Bacchae. Pentheus at first sees and believes the fictions of Dionysos but is destroyed by the revelation of the truth. Amphitruo, Pentheus' "echter Bruder,"44 sees the truth all along but resolutely refuses to accept it. This resolution protects his sanity and makes possible the arbitrary happy ending---so long as Amphitruo does not look beyond the inner human play. Plautus has inverted the whole structure of the Bacchae in order to celebrate, not the god of theatre, but theatre as god, as self-creating divinity.
It remains for us to account historically not only for the similarities of narrative and setting of the Bacchae on the Amphitruo, but also their concern with the nature of illusion and its audiences.t The parallel but comically transformed usage of the play-within-the-play motif argues forcefully for a direct, not an indirect influence. Can we go beyond Stewart and argue that Plautus indeed knew the Bacchae directly?
Recent work has considerably illuminated the Hellenistic background from which the archaic Roman theatre emerged. The dominant feature in the landscape of the Hellenistic theatre is clearly the great actors' guilds, the Artists of Dionysos.45 The Artists and their performing tradition kept theatre alive in the Hellenistic period. It should not, therefore, be considered chance that, as A. S. Gratwick notes,46 Plautus' taste in Greek plays was dominated by the three chief writers in the Artists' repertoire, Menander, Diphilus, and Philemon. While most accounts of Plautus correctly portray him as a practical man of the theatre, few have taken this picture to its logical conclusion. The handsome papyrus codices of Menander, designed for a reading public, may lead us astray. One would not find these in the prop baskets of a travelling theatre troupe. Publishers and performers in the ancient theatre existed in a symbiotic relationship. Demand for texts was stimulated by performance, and the texts that met that demand depended on the performers' own copies of play texts.47 It seems likely that Plautus saw many of the plays he chose to "turn" performed by Artists' companies, whether in south Italy,48 or on tour in Rome. Whatever written texts he might have been able to obtain would have been heavily influenced by the Artists' repertoire. The Bacchae, one of the warhorses (along with most of Euripides' ouevre) of the Hellenistic theatre, ought to have been quite easy to obtain.
A much wider ranger of possibilities should replace our usual picture (unconscious though it may be) of Plautus with a tidy, complete Greek text in hand. The close adaptation of some of the Dis Exapaton in the Bacchides may suggest Plautus was working from a written script. But where had he obtained that script? It may have been the prompt book borrowed from one of the Greek touring companies or copied from such a prompt book. Alternately, like an Elizabethan publisher out to produce a "bad quarto" of Shakespeare, Plautus may have induced a Greek actor to recite as much of the script as he could remember to a scribe or to Plautus himself. Finally, Plautus could simply have attended a performance, liked what he saw and heard, and used as much or as little of what he remembered in his own work. None of these models is a priori impossible.49
The case for the direct influence of the Bacchae on the Amphitruo rests on the strong similarities of metatheatrical structure. If we no longer assume a specific Greek comedy behind every comoedia palliata,50 then the following sketch accounts for the Euripidean influence in both substance and style better than any intermediate Rhinthonic farce or otherwise unknown New Comedy. Plautus had numerous opportunities to see the Greek repertoire performed by the Artists of Dionysos in south Italy. His most likely source for written copies of Greek plays was, directly or indirectly, these same Artists. I think it likely that he saw a performance of the Bacchae, not only because it was one of the most popular and therefore most available plays, but also because of Plautus' unique use of two playing levels in the Amphitruo, which is so visually dependent on tragedy.51 Moreover, the dearth of direct verbal echoes of the Bacchae (and posssible parody of the Heracleidae) point to performance rather than text as the probable source. Plautus seems to have worked from his memory, rather than a written text, which, as the fragments of the Dis Exapaton have shown us, he could on occasion follow very closely.52 Plautus' treatment of the Amphitruo legend, then, is a comic response to the metatragic possibilities he saw in the Bacchae.
The virtue of seeing the Amphitruo as metatheare is that is preserves the artistic integrity of Plautus' play while at the same time acknowledging a debt to Greek performance. If we recreate a hypothetical tragedy of the Amphitruo myth, Plautus' work is just a farcical accretion on a serious narrative.53 The concept of metatheatre allows us to see how Plautus has re-imagined the situation of the Bacchae, where a divinity creates a play-within-the-play for his own pleasure and worship: in the Amphitruo an inner human adultery comedy replaces the Bacchae's tragic sparagmos. In the Bacchae outer and inner plays finally collapse into one; in the Amphitruo, they remain separate. The Amphitruo's inner play can then be in one sense a perfect comedy, in that it allows the adultery and yet re-unites husband and wife with no regrets.
Thus the Amphitruo is far more than a parody of the Bacchae.54 The Amphitruo is a resoundingly comic and healthy response to man's dilemma in the face of the caprices of the gods. Plautus takes the dark despair of the Bacchae and converts it into a celebration of the powers of comic theatre. Jupiter has not replaced Dionysos as the object of worship and awe. We admire Jupiter for his potency, not his omnipotence. The object of this joyous celebration is the traditional Roman theatre itself, with its adultery plots, clever slaves, and mass confusion. Plautus dethrones Dionysos and puts in his place the benevolent genius of comedy.55

Niall W. Slater
University of Southern California

1Middle Comedy relations: G. Duckworth, The Nature of Roman Comedy (Princeton 1952) 24; also T. B. L. Webster in Studies in Later Greek Comedy (Manchester 1953) 95-97, dating the Amphitruo late in the period of mythological travesty. Webster relies on the dubious strength of a supposed allusion to Alexander's cavalry victory over Darius in 331 BC in Sosia's description of Amphitruo's cavalry victory. If adaptations are any evidence, it was the nineteenth century which discovered tragic depths to Alcumena's character and situation: e.g., the discussion of Kleist's Amphitryon in C. D. N. Costa, "The Amphitryo Theme," Roman Drama, ed. T. A. Dorey and D. R. Dudley (New York 1965) 107-144. On the play as a study in identity, see for example Hazel E. Barnes, "The Case of Sosia Versus Sosia," CJ 53 (1957-58) 19-24. On the play as successful adultery comedy, see Erich Segal, "Perché Amphitruo," Dioniso 46 (1975) 247-67, which appears in English as pages 171-91 of the second edition of his Roman Laughter (Oxford 1987).

2Amphitruo 357 and Bacchae 968; Amphitruo 1066 and Bacchae 604-07. See A. Palmer, The Amphitruo of Plautus (London 1890) ad 1.1.203 and 5.1.14. I have cited the lines from the Bacchae in Dodds' numeration, the Amphitruo from Lindsay's OCT.

3Zeph Stewart, "The Amphitruo of Plautus and Euripides' Bacchae," TAPA 89 (1958) 348-373. He argues (364-68) against Vahlen's view, thitherto prevalent, which dismissed Rhinthon as a possible source. Stewart also differentiates between parody and hilarotragoidia, of which he sees Rhinthon as a possible originator. Cf. Gioachino Chiarini, "Compresenze e conflittualità dei generi nel teatro latino arcaico (per una rilettura dell'Amphitruo)," MD 5 (1980) 87-124, esp. 106-115. For a cautionary note, however, see G. K. Galinsky, "Scipionic Themes in Plautus' Amphitruo," TAPA 97 (1966) 207-08, who makes the cogent point that there is a major gap between Rhinthon's phlyax-style farces and the comedy of the Amphitruo.

4E.g., by W. H. Friedrich, Euripides und Diphilos (Munich 1953) 263-79.

5K. M. Westaway, The Original Element in Plautus (Cambridge 1917) 13-15, quoted in Eckard Lefèvre, Maccus Vortit Barbare: Vom tragischen Amphitryon zum tragikomischen Amphitruo (Wiesbaden 1982) 21.

6Lefèvre (above, n. 5). I obtained a copy of this work only after the argument of this paper was complete. Lefèvre begins by arguing that the character of Sosia is essentially a Plautine creation: "Sosia als komische Person ist insgesamt auf das Konto des römischen Dicters zu setzen" (16). He can then argue that all the comic scenes involving Jupiter are similarly Plautine. I agree with these conclusions but would not necessarily endorse the methods by which they are reached. The problem with the analyst approach is its assumption that, once everything Roman has been pared away, the residue will necessarily have a unity---that of the Greek source, which for Lefèvre is a tragedy about Alcmene. It is possible in this process to lose the unity of the Roman play itself, which I here endeavor to outline.

7H. P. Foley, "The Masque of Dionysos," TAPA 110 (1980) 107-33; a fuller version of this has now appeared in her Ritual Irony: Poetry and Sacrifice in Euripides (Cornell 1985) 205-58; C. Segal, Dionysiac Poetics and Euripides' Bacchae (Princeton 1982), "Metatragedy," 215-71. 8Foley (above, n. 7) 132. C. Segal (above, n. 7) 217.

9Cf. my Plautus in Performance: The Theatre of the Mind (Princeton 1985) 13-14, modifying the definition and position of Lionel Abel in his fundamental book on the subject, Metatheatre (New York 1963). Abel epigrammatically defines metatheatre by the proposition "The world is a stage, life is a dream." Abel asserts that metatheatre is a new genre which replaced tragedy during the Renaissance, when self-consciousness made tragedy impossible. The term metatheatre was first applied to Plautus by M. Barchiesi, "Plauto e il 'metateatro' antico," Il Verri 31 (1970) 113-30, but his analysis simply accepts Abel's definition. As E. Segal notes in his "Scholarship on Plautus 1965-1976," CW 74 (1981) 400-01, Abel's concept as such does not fit Plautus. Chiarini (above, n. 3) 98-99 has some further interesting observations on metatheatre as a "third" or mixed genre.
A more useful analysis is offered by F. Dupont, "Signification théatrale de double dans l'Amphitryon de Plaute," REL 54 (1976) 129-41. Dupont sees the essential elements of the Amphitruo's metatheatricality, as well as its thematic connections to the Bacchae, but continues to accept the existence of an intermediate parody by Rhinthon. We differ primarily in the interpretation of tone. She centers her interpretation around Sosia ("La figure du Sosie devient le mythe même de la théâtralité romaine." p. 129) and emphasizes the disorder of the world depicted. I see Jupiter rather as a virtually Aristophanic hero, recreating the theatrical world in his own image. We must beware the effect of the disappearance of the climactic scene between Jupiter and Amphitruo. In its absence the character and the concerns of the hapless Sosia bulk larger than they would in the complete play. We will see that the "Life is a dream" motif will be very prominent in the Amphitruo.

10On the lack of metatheatrical techniques in Greek New Comedy, see my "Play and Playwright References in Middle and New Comedy," LCM 10 (1985) 103-105. C. Segal (above, n. 7), though he once speaks of the Bacchae as metatheatre (270), prefers the terms "metatragedy" and "metadrama," following the terminology of J. L. Calderwood in Shakespearean Metadrama (Minneapolis 1971). Nonetheless Segal's definition, encountered only in the final stages of my own work on the Amphitruo, is strikingly similar to that I had chosen; Segal 216 says, "This self-conscious reflection by the dramatist on the theatricality and illusion-inducing power of his own work, on the range and limits of the truth that dramatic fiction can convey, I call the metatragic dimension of the play." Foley is more cautious in her analysis and suggests that "unlike many modern playwrights, (Euripides) was less interested in a self-conscious exploration of his own drama than in the way art interprets divine and human experience for the city" (108). The difference here is one of degree, not nature. Whether self-conscious theatricality is a principal or a subsidiary focus of interest in the Bacchae, it becomes the principal focus in the Amphitruo. Nor will a dividing line drawn between ancient and modern metatheatre seem anything other than arbitrary. The use of the play-within-the-play device is perhaps the most unambiguous feature of metatheatre. Nothing differentiates its use in the Bacchae (demonstrated by both Foley and Segal) and the Amphitruo (see below) from its re-appearance in as profoundly metatheatrical a play as the Tempest.

11I emphasize this, because the Roman comic stage is too often and too easily compared to the modern stage. The presence of the chorus in the orchestra in the Greek theatre, for example, is a constant reminder that a very different theatrical dynamic is at work. The stage of Plautus is, I suggest, equally but more subtly alien to the modern stage. The emphasis on linearity of interpretation in performance criticism resembles that in reader-response theory but is here even more important. Re-readings of written narrative are licit (though subordinate to the frame established by the first reading). Drama in performance allows no "re-readings;" any "meaning" of a play which is not apprehensible in the course of performance is not in fact a meaning of the play at all, but of an entirely separate entity, the script as reading text.

12On the induction scene in Jonson's Bartholomew Fair see Jonathan Haynes, "Festivity and the Dramatic Economy of Jonson's Bartholomew Fair," ELH 51 (1984) 645-668, esp. 657-662. Haynes shows how Jonson's decision to move the "fair" of the title from an open-air to an indoor venue and the tensions between popular and coterie theatre at the time created a need for Jonson to redefine his audience in the process. The function of most Plautine prologues is far more inductive than expository. They ask the audience's favor but also can attempt to change audience expectations. The closest parallel to Mercury's contract may be the edict of the Poenulus prologue. There the relations of performers to audience are mock-authoritarian: the prologue ingratiates himself by undercutting his own most peremptory commands. For example, he first asks the praeco to call for silence, then turns this into a joke by offering him double wages for being quiet (lines 11-15).

13Imagine a Broadway audience at the opening night of a new Neil Simon comedy being informed they are about to see Mourning Becomes Electra. As the prologue to the Captivi says, it is unfair to the audience to give them tragedy when they expect comedy. As Lefèvre notes (above, n. 5) 23, this toying with audience expectation only works in a Roman context, where the choice of plays lay with the organizers of the games; an Athenian audience knew by the day of the festival whether to expect comedy or tragedy.

14H. D. Jocelyn, The Tragedies of Ennius (Cambridge 1967) 25-28 discusses the normal Roman view that one story or argumentum could have several different versifications. He speaks of the "sharp ancient distinction between a work of art's substance and its verbal or material form; the one was constant and to an extent beyond time and space, the other was very much fixed in its creator's particular time and could not be reproduced in another's." (25-26). It would not then be surprising for the same story to be treated tragically or comically in different versions; that is the basis of mythological travesty. For exactly the same verses, word for word, to constitute a tragedy or a comedy at will is paradoxical and extraordinary. This must be a Plautine conceit: Greek comedy and tragedy differed in the trimeter itself, unlike Latin dramatic verse. This passage, however, militates against the challenging suggestion of G. A. Sheets, "Plautus and Early Roman Tragedy," ICS 8 (1983) 195-209 that tragic and comic language were not markedly different stylistically in Plautus' day: what then would Mercury's point be?

15One more bit of actor/audience business remains: a discussion of rivalries among actors and the use of claques (78, favitoribus). Not only does such a discussion continue to highlight the theatrical process; it also warns the audience against exceeding its own role in the play. Jupiter, god and actor, forbids this (83-88).

16Cf. J. L. Styan, Drama, Stage and Audience (Cambridge 1975) 213: "What is loosely called play-within-a-play can at one extreme be a rehearsal play like that of Pyramus and Thisbe in A Midsummer Night's Dream...and at another extreme the intense and urgent subject of an eavesdroppoing incident like that in which Othello overhears Cassio...: in this, the Moor provides the on-stage audience and the true object of our criticism (Othello, IV.i)." It is just this difference in focus which most analyses of eavesdropping and asides by type fail to show. David Bain, Actors and Audience (Oxford 1977) 154-84 collects most of the material from Roman comedy, though his purpose is to deduce Greek practice behind the types of aside in Plautus and Terence.

17See Styan (above, n. 16) 180-223 and Slater (above, n. 9) 147-167.

18This is one of the most-discussed passages in the play. P. Siewart, Plautus in Amphitruone fabula quomodo exemplar graecum transtulerit (Berlin 1894) 72-76, seems to have been the first to suggest that it was based on Euripides' Heracleidae. Fraenkel, Elementi Plautini in Plauto (Florence 1960) 333-35, thought it was based on a Greek parody of a tragic messenger's speech, but suggested that as it stood in the Amphitruo it had become thoroughly Romanized. L. Halkin goes further and suggests that the speech is a parody of a Roman general demanding a triumph ("La parodie d'une demande de triomphe dans l'Amphitryon de Plaute," AC 17 [1948] 297-304). Renato Oniga, "Il canticum di Sosia: forme stilistiche e modelli culturali," MD 14 (1985) 113-208 offers an exhaustive discussion, which I have seen only in the final stages of revising this article. Alluding to Chiarini's discussion (above, n. 3) of "mixed genre", he argues for a close connection to the style of epic and in particular Naevius. He discusses in great detail parallels to Attic messenger speeches as well.
The important question of whether the speech as it stands in the text of Plautus is intended to parody Roman tragedy (in addition to having Greek tragic predecessors) has been somewhat obscured in the discussion. Stewart (above, n. 3) 362-63 takes the speech quite seriously as "some of the finest Latin of (Plautus') time." He points out that certain features of the speech can be paralleled only in Pacuvius (Fraenkel had noted some of the parallels, but not their exclusivity). He seems to find no humor intended by this quotation of tragic style.
The heavy alliteration, the bombastic expressions, and parallels to Pacuvius lead me to conclude that, whatever the origins of the speech (and I shall speculate on these below), parody of Roman tragedy is clearly present in the speech as constituted (and is not excluded by echoes of epic). This verbal style stamps it as a "theatre piece," a differentiation from the style of speech surrounding it sufficient to mock both speech and speaker. Sosia's theatrical stance for delivery independently reinforces this conclusion.

19Zeph Stewart, "The God Nocturnus in Plautus' Amphitruo," JRS 50 (1960) 37-43, identifies Nocturnus with an epithet of Dionysos: Nyktelios. We shall see further relevance to this "Dionysiac" reference below.

20While fabulor can simply mean "to speak," I would suggest that this play's metatheatrical games can and do call to mind the notion of fabula as play or fiction contained within this verb. A similar and perhaps stronger case can be made for the verb ludificare in the play; see below, n. 23. 21E.g., 302-05; 313-14. Note that once again we have a non-illusory exchange of question and answer aside:
Mer. non feret quin vapulet.
So. quis homo?
Mer. quisquis homo huc profecto venerit, pugnos edet.
Sosia "feeds" Mercury a straight line, to which he responds, though Mercury will not in fact pretend to notice Sosia's presence until 332. We ought also to note Mercury's theatrical threat to metamorphose Sosia with his fists (316 alia forma esse oportet). Mercury's very presence on stage in the guise of Sosia will perform this alienation effect.

22See D. Guilbert, "Mercure-Sosia dans l'Amphitryon de Plaute: un rôle de parasite de comédie," LEC 31 (1963) 52-63. Guilbert argues that the role of Mercury is "parasitic" in the Middle Comedy tradition, because he assists Jupiter in his amours but does not seem to notice that his only explicit attempts at playing the parasite are total failures.

23In this highly theatrical play the Latin expression ludificare seems to mean "to make a play of or about." In this light Amphitruo's suspicions are in fact soundly grounded (565 ludificare, 571 ludos facis me, 585 ludificas). It is not Sosia, however, who is turning Amphitruo into the Amphitruo, but Jupiter. The suggestion of P. Grimal, "Le théâtre à Rome," Actes Bude IX (Paris 1975) 254-55 that Jupiter was the Roman god of theatre, which would have unique significance in this play, has apparently persuaded few.

24Lines 633-53 are essential to any "tragic" characterization of Alcumena: note for example that her sorrow in seeing Amphitruo go is greater than her joy was in receiving him home. Her earlier appearance was in dialogue with Jupiter; here she speaks directly to the audience, which must give her their undivided attention and judge her as an individual character. Jane E. Phillips in her "Alcumena in the Amphitruo of Plautus: A Pregnant Lady Joke," CJ 80 (1985) 121-126, has launched a frontal assault on this "tragic" view of Alcumena. She rightly insists we must visualize and hear the impact of this soliloquy spoken by a hugely padded, pregnant figure. A word such as voluptas, spoken by such a figure, has a strongly sexual connotation: voluptas is what got her into this mess. Chiarini (above, n. 3) 108-109 suggests this (apparently unique) use of a pregnant figure is a farcical element and an argument against its Attic origin. I doubt, however, that the sole purpose of this soliloquy is to make the "Roman audience howl" (Phillips, p. 125). Visual comedy and verbal pathos are here skillfully interwoven to show us Alcumena as the innocent bystander trapped in the games of gods and men. On the challenge posed by the soliloquy and its non-illusory function, see Styan (above, n. 16) 193. If Alcumena stood alone on stage, we might be more in doubt, but the admiring comments of the eavesdropping Amphitruo (654-58) tip the balance.

25It is certainly worth emphasizing the inconsistency of Amphitruo's behavior on one level here. If he is right that his wife was drunk, mad, or dreaming, then she has not in fact entertained any other bedfellow than himself. If she did, none of those explanations is valid. Amphitruo demands the worst of both worlds: he will insist that his wife is both insane and unchaste.

26Here Plautus makes a point of the (unparalleled in comedy) upper level. Lefèvre (above, n. 5) 16 finds the phrasing "eine Verfremdung," because he should say "I am that Jupiter who occasionally becomes Amphitruo." His deliberate confusion of the audience (is he the "real" Amphitruo? 861 suggests he is, while 862 denies it) emphasises his control over the play and its illusions.

27Why this redundant emphasis? He must still be wearing his distinguishing golden tassel. Perhaps it is simply the strength of the convention that a god in a comedy will normally appear only as prologue (whether intial or delayed) that leads him to reassure us that he is indeed the divine Amphitruo back for one more appearance.

28Elaine Fantham, "Towards a Dramatic Reconstruction of the Fourth Act of Plautus' Amphitruo," Philologus 117 (1973) 197-214.

29Fragments VI and VIII according to Lindsay's numeration both contain the expression larvatus. One fragment Lindsay ascribes to Mercury, the other to Alcumena. Whether this assignment is correct or not (I think it is), certainly the tables have been turned on Amphitruo, who has been so eager to suspect derangement in both Alcumena and Sosia: he in his turn is now described by others as "possessed."

30Cf. Lefèvre (above, n. 5) 36-37.

31The implication is that such soothsayers might reveal too much of the just-concluded events. Lefèvre (above, n. 5) 29 sees this as left over from the tragic version where Teiresias foretold the birth of Heracles and Iphicles---but this does not really explain why Plautus would bother to mention it at all. See below on how this protects the comedy of the inner, human play.

32See Stewart (above, n. 3) 350-56, for his summary of these points, especially for his lucid discussion of the madness and drunkenness themes.

33G. K. Galinsky (above, n. 3) 206-07, citing Duckworth, suggests that this aedes need not be a palace. The point, of course, is not the English word we choose, but the visual impact of whatever the text labels an aedes (which depends as much on the scene painter as the text). Most important is the disposition of the action before the facades; both make use of two levels.

34Stewart (above, n. 3) 371; cf. Chiarini (above, n. 3) 107-108.

35The use of two levels can be paralleled in Aristophanes as well. The current orthodoxy that his plays were not performed after his lifetime has been questioned by Oliver Taplin, "Phallology, Phlyakes, Iconography and Aristophanes," PCPS n.s. 33 (1987) 92-104. He argues that a vase painting in Würzburg (H5697) shows the Telephus parody scene from Thesmophorizusae, which therefore must have been performed in south Italy after Aristophanes' lifetime. I think he is right, but the performance must have been a rarity, its impact on New Comedy production doubtful. Tragedy certainly affected Menander; see S. M. Goldberg, The Making of Menander's Comedy (Berkeley and Los Angeles 1980) 16-20. We have no evidence, however, that any of Menander's divine prologues appeared on the theologeion.

36Foley (above, n. 7) 130 n. 38.

37Foley (above, n. 7) 126. Cf. lines 861-66.

38The theme is further emphasized by the constant discussions of acting, e.g., quasi vero novom / nunc proferatur Jovem facere histrioniam (89-90) and even more explicitly line 115, sed ita adsimulavit se, quasi Amphitruo siet, followed by further discussions of costuming.

39Foley (above, n. 7) 110.

40Foley (above, n. 7) 111. Cf. E. R. Dodds' commentary on the Bacchae 2nd ed. (Oxford 1960) 147-49.

41I am fully aware of the dangers of anachronism in this characterization. The organization of the ancient theatre, both Greek and Roman, is significantly different from the modern, which took on its present form only at the end of the nineteenth century. The Roman theatre seems more parallel to the Elizabethan model, with one dominant artistic figure in each company, such as the actor-manager-producer, L. Ambivius Turpio. Nonetheless, it will be more illuminating to speak of the split of power and function between Jupiter and Mercury as producer and director than to invent more abstract and even more arbitrary terms.

42Much has been written on the subject of whether masks were used in the original performance of the Amphitruo and whether those masks were identical for the doubled characters. Whether masks were used or not (and it seems likely to me that they were), they would not have been identical, only similar. Permit me a little anecdotal evidence. A number of years ago I directed a production of Rogers and Hart's The Boys from Syracuse, a descendant, via The Comedy of Errors, of Amphitruo. This play was chosen in part because two sets of twins were available to play the doubled parts. This brilliant casting stroke turned out to be quite the opposite, for preview audiences were hopelessly confused as to who was who. We finally had to sew gigantic letters on their costumes (S for Syracuse, E for Ephesus) to distinguish the twins and make the story intelligible to the audience. The moral: the humor of mistaken identity plots lies in the audience's superiority to the confusion on stage.

43See Foley (above, n. 7) part I, "Dionysos' Play," 109-16.

44Friedrich (above, n. 4) 278.

45For an account of these guilds see Sir A. W. Pickard-Cambridge, Dramatic Festivals of Athens, 2nd ed. by J. Gould and D. M. Lewis (Oxford 1968) 279-305.

46A. S. Gratwick, Cambridge History of Classical Literature, vol. II, ed. by E. J. Kenney and W. V. Clausen (Cambridge 1982) 97.

47See D. L. Page, Actors' Interpolations in Greek Tragedy (Oxford 1936) 106-15.

48Stewart (above, n. 3) 371-72, suggests that Plautus learned his craft in south Italy. See Gratwick (above, n. 46) 77-78 for the presence of the Artists at Rome.

49A recent book by Bruno Gentili suggests that the Artists' performing tradition had another, perhaps less welcome effect on the dissemination of tragedy: Bruno Gentili, Theatrical Performance in the Ancient World: Hellenistic and Early Roman Theatre (Amsterdam 1979) especially 15-41. The companies apparently took to performing excerpts from famous tragedies as virtuoso showpieces, and written collections of such excerpts, as indicated by some admittedly fragmentary papyrus evidence, seem to have gained some currency. This may in fact explain the influence of the messenger speech from the Heracleidae on Sosia's report of the battle. Such a messenger speech is an ideal candidate for solo recital and fits in perfectly with the idea of a collection of showpieces for the individual actor.

50S. M. Goldberg, "Plautus' Epidicus and the Case of the Missing Original," TAPA 108 (1978) 81-91, suggests that the Epidicus had no specific Greek model behind it. Cf. now E. Stärk, Die Menaechmi des Plautus und kein griechisches Original (Tübingen 1989).

51We cannot of course prove that no Middle or New Comedy in Greek ever used two playing levels, but we have no evidence for such. We ought therefore to hesitate to invent such a lost exemplar in order to spare Plautus the burden of originality.

52We have no satisfactory grounds for deciding where in the chain of events the Heracleidae speech was detached from its original context and parodied (see above, n. 18). Plautus may have seen the speech performed as a tragic excerpt or found in such a written collection (see Gentili, above, n. 49). Alternately, he may have adapted a speech from a Greek comedy, which was in turn based on the Heracleidae, though with each transformation the possibility of recognizing the original behind it grows ever dimmer. Sosia's speech in the Amphitruo is in any case an isolated showpiece.

53This is Lefèvre's approach, which praises the originality of Sosia, for example, precisely because it has nothing whatever to do with what he sees as the basic narrative.

54I have avoided the question of whether Plautus' audience would have recognized the indebtedness of the Amphitruo to the Bacchae precisely in order to avoid this implication. If Plautus could have seen the Bacchae in south Italy, so could his audience, especially as military service took many there. Stewart (above, n. 3) 359-64, considers the possibility that Plautus' audience may have seen a Roman version of the Pentheus story. He must reject Pacuvius' Pentheus on the grounds that a human attendant was substituted for the god in this version of the story and instead conjectures, on the basis of probability alone, that Ennius had produced a version of the story.
If we must have a Roman candidate, we ought at least to consider Naevius' Lycurgus. As D. F. Sutton has shown, "Aeschylus' Edonians," Fons Perennis (Torino 1971) 387-411, this play was based on the Edonians in Aechylus' Lycurgeia tetralogy. The Edonians as reconstructed by Sutton shows many parallels to the Bacchae, but much of this is common to the resistance myth. Sutton's assumption that Dionysos appeared as a mortal is not guaranteed by the fragments of either the Lycurgus or the Edonians. Without this key element, Naevius' version cannot supply the model for the Amphitruo, though all reference to the Lycurgus cannot be excluded. It may be barely possible that Amphitruo's suspicion that his house has been taken over by a Thessalian magician (1043 ego pol illum ulciscar hodie Thessalum veneficum) plays off the Lycurgus resistance story, which is set in Thessaly. I must emphasize again, though, that the essential parallels of the Bacchae and the Amphitruo center around the disguised god; otherwise we should not think of equating Jupiter and Dionysos.

55An NEH Summer Research Stipend generously supported the writing of the first version of this paper. I am indebted to James Tatum, Zeph Stewart, Sander Goldberg, and the editor for their helpful criticisms of its thesis, aliisque alio. The heterodoxies that remain must be laid to my own account.