I hated Roman comedy as an undergraduate. Not only was it hard Latin, but also it gave little promise to be worth the effort, as being apparently a stereotype-ridden exercise in lamentable literary secondariness. It was only many years later when I was forced to teach the genre that I began to see that there might be some fun in it. Perhaps I should apologise for being slow, but anecdotal research suggests that I am not alone in my early experience, and that, apart from a select group of experts, most people, even professional classicists, are not avid readers. The less-than-avid readers of Plautus and Terence are first among those for whom this book is written. If the experts also find something here to amuse, if not to inform, I shall be well pleased.

From a career beginning in Augustan elegy, my entry into Roman comedy does not follow what might be thought the traditional route from the former to the latter. That there has been some degree of connection between the two genres has been known since antiquity, although not greatly exploited by modern critics, but the connections made have in any case been largely through what one might call the fictional worlds of the two genres. Insofar as one believes the imagined world of elegy to be that of a ‘demi-monde’, where the main players are slaves and freedwoman-prostitutes interacting with citizen men, that world can be seen to relate to the fictional world of the fabula palliata.

What begins, must end. From Aristotelian formalists through New Critical aesthetes to post-modern deconstructionists, we all agree that the ending of a work of literature is not just a matter of neutral stopping but is one of the most (perhaps, the most) highly charged and semiotically important moments of interpretative activity for readers and audiences. How something ends affects how we finally imagine it to have been all along. Although at the performative level the comic ending is trumpeted by a grand array of signs which signal the end, at the level of plot the comic ending is frequently delayed, repeated and even denied. This playfulness about the possibility of closure is, at the same time, cut across by the Big Bang which almost arbitrarily imposes the end. As Charney says: ‘the conclusion of the comic action violates the feeling we have that comedy is dimensionless, nontemporal, infinitely extensible, and not amenable to finite solutions or resolutions.’ The comic coup which Plautus and Terence achieve in their different ways is the combination of boundlessness and the transcendence of limits with the powerfully end-stopped consummation, so often signified by the big party.

I begin with exposition of some closural signals which complete a performance, and then consider some of the emplotments and other tricks which cause the play to transcend its own boundaries.

The plot's the thing. It is not, of course, the only thing, for Aristotle perhaps means little more (or less) than a point about medium (that drama only comes into existence when some action happens), while Anglophone critics use the term ‘plot’ in different ways for different purposes, not all of which can automatically be said to reflect on each other. It is nonetheless possible, I suggest, to make a claim for a particular kind of plot activity as integral to the workings of Roman comedy: that internal plotting is a programmatic sign for comedy; that internal and external plotting (that is, the best laid plans of the characters in the play, and the plot of the play) are mutually reinforcing; and that instability of identity within a play slides into the precarious construction of identity in the production of a play. This means not just that there is a lot of deceit and a lot of instability of identity in Roman comedy, but specifically that these issues work together, and that they constitute a programmatic metaphor for the dramatic performance.

How does anything ever manage to begin? There is an enormous burden of proof placed on the opening of any cultural act, be it literary work, religious ritual, public performance, educational session or indeed work of literary criticism. This is in part simply because the initial impression is crucial to the reception of the act and because mistakes at this point are among the hardest to repair, but, beyond that technical difficulty of creating a good impression apparently out of nowhere, the opening must bear an additional burden, which is a performative one. The beginning has to act as if it is doing something, causing something to happen, changing the world – when everyone knows that no beginning is ever really The Beginning. As Said says (1975: 50), ‘a beginning might well be a necessary fiction’.

Yet for all this difficulty, it hardly needs saying that there is something very special about beginnings of all varieties, literary and social, and Roman culture is no exception to the pattern of strongly marking the opening, embodied especially in the cult of the god Janus. Our sense that every beginning is also an ending is symbolised by the very nature of the two-headed god, looking backwards even as he presides over novelty. Ovid, in the Fasti, has a discussion with Janus (Fasti 1.63–288) about the nature of the god and his jurisdiction over beginnings, in which the importance of beginning, and of starting as you mean to go on, is sanctified as essential not only to the poem thus initiated, but also to the Roman world and its relationship with the cosmos.

Bergson's tempting twilight zone where artifice rules is the domain of comedy.

Comedy is an art form which delights in its artificiality, glories in its artifice, and actualises its art – all with apparently effortless ease and studied avoidance of obvious studiousness. This study concerns the particular form of comedy which flourished as a performance genre in the late third and early–mid second centuries bc, the plays being first produced as part of festivals celebrating one of the gods, or on other religious, social, and political occasions, such as the dedication of a temple or funeral games for a great man like Lucius Aemilius Paullus. The audience was on holiday and expecting a play to match, but it was also, at least in part, a sophisticated theatre-going crowd with experience of different types of dramatic performance and of other forms of art. The plays of Plautus and Terence flourished also as a textual genre which manifested itself in a much wider range of reading practices, including the scholarship of Varro, the rhetoric of Cicero, the epitomisation of the grammarians and moralisers, the exemplification of the schoolroom, and the imitation of the Western comic tradition.