Troilus and Criseyde is a story about human beings. In contrast to most of Chaucer’s earlier poems, which are dream visions populated by animals and allegorical figures, Troilus follows Filostrato in emphasizing character over plot and in describing these characters as fully developed, naturalistic men and women. For all the skill of the historical background created by Chaucer, the epic events of the Trojan War occur largely off-stage; instead of battle we are shown in detail the private lives and feelings of a small number of individuals: the two lovers especially and, to a lesser extent, Pandarus. Troilus also contains more minor characters than its source and expands the roles of others, though their primary function is to support the principal actors. Character has always been at the centre of criticism of Troilus. Chaucerians as diverse as Aers, Bishop, McAlpine and Wetherbee - like Donaldson, Muscatine and Robertson before them, and Kittredge and Lewis before them - pay special attention to the people in Chaucer’s poem.
Although ordinary readers tend to respond more immediately and more strongly to character than to any other aspect of fiction, theoretical work on the topic has languished for decades. In the 1960s W. J. Harvey complained that ‘modern criticism, by and large, has relegated the treatment of character to the periphery of its attention’ (192). Things have not improved much since, despite periodic calls to action: in the 1970s Rawdon Wilson identified characterization as ‘the least successfully treated of all literary concepts’ (191), and Thomas Leitch has recently noted that ‘character has become unfashionable in postwar criticism’ (148).1 Indifference to literary characterization is something shared by the old ‘New Critics’, with their interest in formal textual analysis and persona, and by modern Structuralists and Post-Structuralists (Hochman, 20 ff.; Rimmon-Kenan, 29–36). Many important contemporary theorists, such as Hélène Cixous, even question whether the idea of a coherent human subject, in life or in literature, is anything more than a passing bourgeois illusion (Leitch, 148; Hochman, 26).
The traditional critical debate about literary character, which goes back at least to Aristotle, is whether such figures are merely plot functions, whose being is expressed entirely in what they do, or whether they can achieve an independent existence beyond the text and can be said somehow to resemble living human beings (Leitch, 148–65; Rimmon-Kenan, 31–6). The distinction here is something like E. M. Forster’s famous division of characters into ‘round’ and ‘flat’, which, despite its impressionism and inadequacy, continues to be a starting-place for contemporary discussion.2 According to Forster, flat characters, who are sometimes called types, ‘are constructed round a single idea or quality’ (67), whereas the truly round character is ‘capable of surprising’ because it has the ‘incalculability of life about it - life within the pages of a novel’ (78).
The characters in Troilus have been treated by Chaucerians as both round and flat. Their realism and individuality were insisted upon by early critics such as Kittredge, who called the poem ‘a masterpiece of psychological fiction’ and the ‘first novel’ (Poetry, 109), and Patch, who asserted that the characters ‘truly live’ and can be analysed with ‘direct reference to life’ (Rereading, 65). Naïve as these statements may seem today, they announce a belief in the humanity of Chaucer’s people that has led some critics to claim to have ‘fallen in love’ with Criseyde or to feel justified in analysing such things as the psycho-sexual maladjustments of Pandarus.
Other Chaucerians, however, from a variety of critical schools, deny that the characters should be treated as real people. The most influential statement of this position is by Mizener, who adopts the Aristotelian view that ‘Chaucer’s chief interest was in the action rather than in the characters’ and that the latter are therefore static figures incapable of change or development (67). Shepherd also doubts the psychological coherence of even Troilus and Criseyde, seeing each instead as a ‘function of the plot’ (‘Troilus’, 78–80), and exegetical interpreters inspired by D. W. Robertson argue that Chaucer’s people are typological moral constructs (Troilus is said to re-enact the story of Adam) with none of the depth, autonomy or individuality of post-Romantic creations. Like Robertson, Payne notes the strength of such fixed characterization, which allows Chaucer to make statements of general significance ‘without the chance inconsistencies and non sequiturs of actual existence’ (Key, 223).
Both approaches to characterization in Troilus and Criseyde are justified. Certainly the poem contains many flat or type characters. Hector, for instance, possesses almost no individuality, but always appears as the perfect model of a gentil and generous knight. Diomede is equally and oppositely representational. We never see him as anything but a self-interested pragmatist. Neither surprises us for an instant. Even the three main characters can be viewed as types, or at least as characters playing clearly defined, traditional roles, which Miskimin identifies as lover (Troilus), lady (Criseyde) and go-between (Pandarus) (198). Most English poets in the late Middle Ages and Renaissance treat the characters thus and are content to turn Pandarus into a common noun and see the lovers as ‘true’ Troilus’ and ‘false’ Criseyde.
But neither ‘type’ nor ‘plot function’ explains the range of Chaucer’s characterization in Troilus and Criseyde. Even figures we would expect to appear as conventional caricatures, like Helen of Troy, prove elusive and surprising, as we saw in the previous chapter. Certainly none of the major actors can be satisfactorily reduced to a single idea or quality, despite a recent attempt to provide a ‘keyword’ to ‘help readers get started in forming judgments about each character’ (Shoaf, xxi). The words proposed include changeable for Criseyde, idealist for Troilus and expedient for Pandarus; but, while each is relevant, none captures, even preliminarily, the distinctiveness of Chaucer’s creations. Expedient is the least inadequate, but does not suggest Pandarus’ humour or differentiate him from Diomede; idealist is much too general for Troilus (it would work better for either Hector or Antigone) and ignores his sensuality and complicity in deceit; changeable is appropriate for almost every other representation of Criseyde in literature, but Chaucer repeatedly shows the limitation of this traditional anti-feminist label.
The achievement of Chaucer’s major figures, as some critics have noted, is that they successfully combine qualities of both flat and round characterization. Being ‘conventional and original at once’ (Ginsberg, 135), they are recognizably familiar as well as capable of surprise. They unite the clarity and general application of medieval allegorical portraits with a new immediacy and realism that was beginning to dominate Western art (Rowe, 57–8).3 By drawing on the strengths of both abstract and mimetic characterization, Chaucer is able to create figures richer and more powerful than those found in either tradition alone (Scholes and Kellogg, 89–98). Troilus, for instance, is somewhat more conventional in Chaucer than in Boccaccio - he becomes the very archetype of the courtly lover - while, at the same time, he is given more individual qualities, like his shyness. Despite the apparent clarity of their roles, Chaucer’s major characters are indeed capable of surprising the reader, for they are what one theorist calls ‘open constructs’, who are capable of existing outside the text: ‘The character may haunt us for days or years as we try to account for discrepancies or lacunae in terms of our changing and growing insight into ourselves and our fellow human beings’ (Chatman, 132–3).
That the principal actors in Troilus and Criseyde can be treated as both believable human beings and artificial literary types suggests that, while Chaucer learned from Boccaccio’s emphasis on character over plot, he nevertheless radically transformed the Italian text. In Filostrato Troiolo, Criseida, Pandaro and even Diomede are about the same age and share essentially the same worldly, if elegant, values and desires. Any apparent dissimilarities, such as Criseida’s discretion or Pandaro’s deference, are the result of social position or narrative role and are not innate. If he were to fall in love, Boccaccio’s Pandaro would probably act much like Troiolo, as Boccaccio’s Diomede in fact does. The characters in Troilus and Criseyde, however, are described with an increased thickness and variety of detail that distinguishes one from the other. Chaucer’s Pandarus is made older than his Italian model and given unparalleled verbal skills; Troilus is less experienced in love and more of a chivalric knight; Criseyde is wittier, more courtly, and less directly passionate than Boccaccio’s heroine. Many of these added qualities appear to be in conflict with one another; thus Troilus’ greater activity as a warrior contrasts with his passivity as a lover, just as Pandarus seems to care more deeply about the lovers even as he shamelessly manipulates them.
Even more significant than the individual qualities given to each character in Troilus are the radically different ways in which each is constructed. Troilus, Criseyde and Pandarus each has his or her own particular mixture of type and individuality, and each is so drawn that the reader is permitted to observe the figure only from a certain clearly defined perspective, as we shall see below. By presenting his characters so artificially, Chaucer reminds us that, however real they may become to us, each exists first as a text. Pandarus, for instance, is the most like a type: a vivid but shallow figure almost always viewed in action from the outside. In contrast, we are allowed to see deeply into the heart and mind of Troilus, whose emotions we can directly apprehend. The presentation of Criseyde is different again and more demanding still because at crucial moments we are denied access to her complex inner life and have to imagine what she is really thinking and feeling. Chaucer’s Criseyde is not the conventional emblem of female instability that she is elsewhere; instead she is an opaque but generative figure who must be created by each reader.
Pandarus is the most superficial of the three main characters in Troilus and Criseyde, as well as being the liveliest and most entertaining. Although he is the easiest to sum up in a single sentence or even a word (his own name), he also shows how vivid a type figure can be. Boccaccio, who invented the character of Pandarus as we know him (though a warrior of that name appears in Homer), uses him primarily as a bland plot device. A pale copy of Troiolo with little that makes him individual, Pandaro is what Henry James called a ‘ficelle’ (Harvey, 58), a minor character whose only real function is to advance the story. Although Pandarus continues to function as a go-between in Troilus, his role and characterization are greatly expanded and complicated by Chaucer. He is so central to the action that the English poem might plausibly include his name in the title with Troilus and Criseyde. In contrast to Filostrato, whose lovers arrange the consummation by themselves with Pandaro as little more than a messenger, it is hard to believe that Troilus and Criseyde would have managed to come together in Troilus without the relentless pressure of Pandarus, who at last succeeds in getting his passive friend and reluctant niece into bed. He is the driving force behind their affair, and his efforts allow them both to act with more delicacy than their Italian models.
Pandarus is one of those striking figures, intermediate between a protagonist and a background character, whom Harvey has labelled ‘the Card, the character who is a “character”’ (58). These exuberant, ‘larger than life’ type figures, common in the novels of Dickens, tend to be comic rather than heroic, though their comedy is frequently mixed with the pathetic or the sinister. ‘Part of the joy of these characters lies in their immunity to the knocks and buffets doled out to them, in their ultimate reassertion of their own nature’ (61). The advantage of a Card is that the writer can release through him ‘a vividness, an energy, an abundance that would submerge and obscure the more intricate contours of the protagonist’ (62). Chaucer’s Pandarus exemplifies this single-minded intensity to such a degree that he sometimes threatens to overshadow the lovers themselves.
Pandarus has been the object of extreme judgements by readers of Troilus. Early in the century one major Chaucerian called him ‘the most charming of companions’ (Root, Troilus edition, xxxiii), whereas another found him a ‘corrupter of virtue’ whose nastiness is only tolerable because of Chaucer’s ridicule (Legouis, 128). Critical opinion has continued to be sharply divided, although the debate is not so much about what Pandarus actually does as about the standards that are appropriate to interpret those actions. For example, both C. S. Lewis and D. W. Robertson agree that Pandarus is responsible for the sexual union of Troilus and Criseyde, but whereas Lewis sees this as the action of ‘a friend according to the old, high code of friendship’ and ‘a convinced servant of the god of Love’ (Allegory, 191), behaviour quite appropriate to one whom a more recent critic calls a ‘priest of Cupid and Venus’ (Bishop, 37), Robertson denounces Pandarus’ efforts as the sinful work of ‘a priest of Satan’ (Preface, 479). If Troilus and Criseyde is read as a Boethian poem, Pandarus must be condemned, but if it is read as a courtly love romance, he will be excused, even admired. Pandarus makes a successful considerate ami, but a wholly inadequate Lady Philosophy. Both interpretations are valid as far as they go, but because of their generic presuppositions each responds to only a few of the many facets that Chaucer has given this character.
Although the English Pandarus never abandons the role of pander (indeed, he is more active in that office than his Italian predecessor), Chaucer nevertheless adds a rich variety of competing traits to the portrait he inherited from Boccaccio. As we have seen, he makes Pandarus the uncle of Criseyde rather than her cousin. Thus Pandarus is presumably older than the lovers and has some justification for his attempts to arrange things for them. He is also given greater social status, as indicated by the mention that he is an adviser to King Priam (V.284). This more dignified Pandarus is also, at the same time, made more subversive of the established order: one who has no hesitation in deceiving the Trojan royal family with the false story about Poliphete, and one who repeatedly urges Troilus to prevent by force Criseyde’s departure to the Greeks regardless of the consequences to civic order (see IV.621–30). Further complexity results from Pandarus’ increased involvement with the lovers (to the point of voyeurism), even as the English poet shows him to be more manipulative, gives him an arsenal of verbal skills of which there is no hint in Boccaccio, and makes him funnier.
Chaucer’s Pandarus is one of the great comic characters in English literature, whom even Shakespeare could not improve upon when he came to portray him in his play. He is the focus and source of much of what is most amusing in Troilus and Criseyde, virtually all of which is original. He is never afraid to make a joke at his own expense, and he plays the clown even during the consummation scene. We laugh with him as well as at him, though his humour often has a leering cynical element (III.1557–61). In contrast, Troilus’ laments and tears leave little room for humour (though Pandarus is able to raise a smile from him at II.1639), and Criseyde laughs only when she is with her uncle. The cleverness of Pandarus’ many schemes and his irrepressible energy in promoting them are also delightful for the reader, as long as we are not offended by their amorality. He tricks Deiphebus into holding and Criseyde into attending a dinner-party at which his intricate manoeuvres and outright lies permit the lovers to be alone together without any of the other guests suspecting, and later he puts Troilus into Criseyde’s bed with equally bold and hilarious stratagems, which involve a rainstorm, an invented rival, a trap-door and the hero’s swoon.
Pandarus is more comic in Troilus than in Filostrato, but he is also given other, competing traits, such as his expressions of extreme emotion. We have already noted his affection towards Criseyde, but his concern for Troilus is especially intense. When he first comes upon his miserable friend lying prostrate and alone, Pandarus is described as having ‘neigh malt for wo and routhe’ (I.582), just as in response to the prince’s later sufferings he ‘Wex wel neigh ded for routhe, sooth to seyne’ (II.1356). When he learns that the lovers will be parted, Pandarus ‘Gan wel neigh wood out of his wit to breyde,/So that for wo he nyste what he mente’ (IV.348–9). None of these extravagant expressions appears in Filostrato. Yet Pandarus’ increased emotionalism is no simple counter to his wit. Although his sympathy is prompt and dramatic, we are never able to measure its depth or be sure that it is totally separate from his gamesmanship. The passages quoted above, which describe him as being on the point of death or madness, can be read as either sincere empathy or melodramatic playacting, and Chaucer gives us little reason to choose one interpretation over the other.4 In the face of Troilus’ ‘manly sorwe’ when first in the presence of Criseyde, ‘Pandare wep as he to water wolde,/And poked evere his nece new and newe’ (III.115–16). Pandarus’ warm sympathy is inextricably mixed in with his cool scheming - he weeps even as he pokes.
As suggested by the above, a further complicating trait added to the English Pandarus is his virtuosity at manipulating others, which hardly exists in Filostrato. Both Muscatine (French, 136) and Spearing (Troilus, 43) call him a ‘fixer’. Although professing to serve Troilus and Criseyde, Pandarus is incapable of being honest with either. Despite the deep sympathy we are told he feels when he comes upon the love-stricken Troilus, the first words out of his mouth, which attribute his friend’s distress to fear of the Greeks, are insincere but meant only to rouse the prince: ‘Thise wordes seyde he for the nones alle’ (I.551–61). Throughout the rest of the poem, we continue to observe his trickiness and lack of frankness in the pursuit of some immediate end. Professing openness, he carefully finds the most artful way to tell Criseyde of Troilus’ passion (II.255–73); in so doing, he uses a technical rhetorical vocabulary that, as David Burnley has shown, reveals a ‘deliberate display of duplicity’ towards one he claims to love and respect (Guide, 173–4). Later he creates the fictions of Poliphete and Horaste to further the love-affair, and, at the end, when he knows Criseyde will not return, he still cannot bring himself to be candid with Troilus (V.505–11). Pandarus’ motto, expressed at Deiphebus’ house, always seems to be: ‘While folk is blent, lo, al the tyme is wonne’ (II.1743).
Larger questions of social or moral ethics do not seem relevant to Pandarus. He tells Troilus he will help him pursue his passion even if it is incestuous (‘Be what she be, and love hire as the liste!’, I.679), and Criseyde indignantly charges that he, who as uncle and friend ought to prevent her from loving, is instead its principal advocate (II.409–27). His help for the lovesick Troilus is inventive but relentlessly practical. He scorns his friend’s passive suffering - repeatedly labelling it the behaviour of a fool (I.618, I.705, I.762) - and instead looks to what can be done. Far from leading Troilus to the summit of amicitia,5 as some moralistic critics think he should, Pandarus appears not to take seriously even the higher reaches of romantic love. From the first, his clear goal is the limited one of bringing his friend and niece to bed. When it is known that Criseyde will be sent to the Greeks, Pandarus recommends, whether joking or not, that Troilus find a new woman (IV.400–6) or that he ravish Criseyde (IV.530). Despite occasional talk of service and devotion, Pandarus appears to regard love as no more than ‘casuel plesaunce’ (IV.419).
Chaucer characterizes Pandarus not only by giving him a number of different traits whose wide range is often ignored, but also by constructing him in a certain way that is unlike the method used for the other two principal characters. Pandarus is more superficially drawn than either Troilus and Criseyde and is given no more depth than his own fictional creations. Despite his feverish activity, his motives and inner being remain absent, and the few times we are allowed into his mind all that we find there are the mechanisms of plots and schemes, as when he leaves Troilus’ palace after having promised to speak to Criseyde:
And went his wey, thenkyng on this matere,
And how he best myghte hire biseche of grace,
And fynde a tyme therto, and a place.
For everi wight that hath an hous to founde
Ne renneth naught the werk for to bygynne
With rakel hond, but he wol bide a stounde,
And sende his hertes line out fro withinne
Aldirfirst his purpos for to wynne.
Al this Pandare in his herte thoughte,
And caste his werk ful wisly or he wroughte.
This passage, derived ultimately from Geoffrey of Vinsauf’s Poetria Nova, suggests that here as elsewhere (II.266–73) what is meant by Pandarus’ inner self or ‘herte’ is less human feelings and desires than literary and rhetorical theory.
Pandarus is essential in getting Troilus and Criseyde together, but he becomes increasingly irrelevant as the story continues and deepens. He is an impresario rather than an actor - a brilliant initiator of events, but not a full participant. He shares neither the lovers’ profound satisfaction nor their tragedy. The larger forces at work in the story are beyond Pandarus. As a practical man of the world, he derides Troilus’ fatalism, believing that Fortune helps those who help themselves (IV.600–2), and he scorns belief in dreams and auguries as unworthy of ‘so noble a creature/As is a man’ (V.384–5). If he is a kind of humanist, as some have suggested (Howard and Dean, Troilus edition, xxvii-viii), that worldly philosophy proves inadequate to the desperate course of events. In the last two books, Pandarus is able to accomplish almost nothing, though he is increasingly scornful of Troilus’ hopeless faithfulness. He begins his final speech with ‘I may do the namore’ (V.1731), and he concludes it with ‘I kan namore seye’ (V.1743). Because Pandarus realizes himself almost entirely through practical action and language meant to influence others, his announcement that both are lost means his extinction as a character.
The attractiveness and limitation of Pandarus is that he seems to turn everything, including the love-affair, into an occasion for play. In response to Troilus’ blushes when asked about his lover’s name, Pandarus crows: ‘Here bygynneth game’ (I.868). Because of his inventiveness, Pandarus has often been seen as a figure of the artist and thus compared to Chaucer himself (Fyler; Bloomfield, ‘Distance’, 26n14; Donaldson, ‘Three’, 282). As we have seen, there is much truth in this: his stories of Poliphete and Horaste, which are crucial to advancing the love-affair, are sheer fiction, yet Pandarus manages to convince others of their reality. He is similarly skilful in creating a heroic picture of Troilus on the battlefield to impress Criseyde (II.190–203) and may have invented the garden scene during which he says he first learned about Troilus’ love (II.505–53; Donaldson, Chaucer’s Poetry, 972).
For all his imaginative skills, however, Pandarus is limited as an artist. He is a manipulative storyteller, like Nicholas in ‘The Miller’s Tale’ (Bishop, 40), rather than a great poet. Pandarus’ fictions, for all their wit, are shrewdly crafted to produce a predetermined result. As he explains to Criseyde, although some men ‘delite’ in narrating their stories with ‘subtyl art’, yet ‘in hire entencioun/Hire tale is al for som conclusioun’ (II.256–9). He cannot imagine a fiction that is not utilitarian. The hostility of moralistic critics like Robertson against Pandarus may reflect their awareness that his view of art is quite similar to their own, though in the service of different values. In Pandarus’ creations there is no post-Romantic nonsense about ambiguity or complexity of meaning; instead they are deliberately closed works whose every effect is tightly controlled by the author for a specific practical end - to make Deiphebus give a dinnerparty, for instance, or to make Criseyde receive Troilus into her bedchamber. Pandarus’ fictions are calculated indoctrination, a kind of art all too common in medieval (and modern) literature, but which some writers, like Chaucer himself, were able to transcend.
Although Pandarus is the most superficially drawn of the major characters in Troilus and Criseyde, Chaucer adds a few passages with no precedent in Filostrato that hint that even he has an independent private life, which we might have been told more about if the poet had so chosen. Both Troilus and Criseyde make brief reference to Pandarus’ amorous misadventures as if they were well known, and book II opens with mention of his having felt ‘loves shotes keene’ (II.58). The mysterious mistress is never named, however, and her reality remains shadowy. It must be some other aspect in the characterization of Pandarus that has led critics to treat him, not always successfully, as if he had a genuine psyche. We may be helped in understanding this supplementary quality in Chaucer’s Pandarus by Baruch Hochman’s observation that even a stylized character like the silly Mrs Bennet in Pride and Prejudice ought not to be classified as a simple type because her admittedly limited number of traits are so arranged that they implicitly ‘form a dynamic system of stresses that suggest greater complexity - that is, a high degree of inner tension and self-contradiction in enacting herself (124). Pandarus may be another such character. The variety of original, sometimes conflicting traits he is given in Troilus suggests a dynamic system of stresses that we sense even if we cannot fully explain. Something must be fuelling the manic energy with which Pandarus promotes, even invents, the love-affair, but we can only guess, to use Harvey’s terms from the beginning of this section, whether Pandarus the comic Card may also be pathetic or sinister.
The possible existence of such inner tensions in Pandarus is implied most clearly by an original and now notorious scene in which he visits his niece’s bedroom on the morning following her first night with Troilus (III. 1555–82). Pandarus coarsely jokes that he hopes the rain has not kept Criseyde awake and asks how she is. When she calls him a fox and hides under the sheet, we are told that Pandarus ‘gan under for to prie’, urges her to cut off his head with a sword if he is guilty, and ‘his arm al sodeynly he thriste/Under hire nekke, and at the laste hire kyste’. After these melodramatic acts, the poet blandly states, ‘I passe all that which chargeth nought to seye,’ notes that Criseyde ‘with here uncle gan to pleye’, and concludes by asserting that ‘Pandarus hath fully his entente’.
The explanatory note in the Riverside Chaucer to this odd scene is uncharacteristically dogmatic: ‘The now widespread view that Pandarus here seduces or rapes Criseyde, or that Chaucer hints at such an action, is baseless and absurd’ (1043). Those critics who find actual incest here are obviously claiming something that goes far beyond anything in the text; but, in contrast to the apparent innocence of the events themselves (nothing much actually happens with respect to action), the violent, even sexual language and images (‘prie’, the phallic sword, ‘thriste’, ‘kyste’ and ‘pleye’) provoke us to wonder about the motives that are driving Pandarus, which the narrator so deliberately refuses to discuss. We are left to guess whether some powerful inner needs (pathetic or sinister) are fulfilled by Pandarus’ close involvement in the love-affair that he has now brought to success. By calling attention to such questions without resolving them, by such a ‘presence of absence’, Troilus occasionally implies that even Pandarus, the most superficial of the major characters, has a self more profound than that revealed by his delightful japes and amoral schemes.
Despite the efforts of some to establish him as the hero of the poem, Troilus is usually slighted by critics (David, ‘Hero’). Tatlock scorned him as ‘the least lifelike’ of the principal figures (‘People’, 335), and current students of both sexes tend to find him weak and foolish. The recent Riverside Chaucer cites intense critical debate over the characters of Criseyde and Pandarus, but does not consider discussions of Troilus worth recording (1022–3). He has not always been overlooked, however. The Trojan prince was regularly celebrated as an exemplary knightly lover by English poets in the fifteenth and sixteenth centuries, and his emotional sensitivity and lack of male assertiveness may make him an ideal once again. Troilus, of course, may also be regarded as little more than a type character. Just as Pandarus never transcends his role of pander, so Troilus, once he sees Criseyde, always plays the lover. He has even been accused of overplaying this role, and Muscatine calls him ‘too perfect a courtly lover’ (French, 137, original emphasis).
Chaucer’s Troilus is not the rounded, psychologically believable character expected of modern realistic fiction, but he is far from simple. Although far closer to Boccaccio’s original than any of the other major characters in Troilus and Criseyde, he, like Pandarus, has been given a variety of new and contradictory traits. For instance, as I suggested earlier, he is more active as a warrior while being more passive as a lover. Early in the poem Chaucer adds a scene, reminiscent of the medieval histories of Troy, that describes the victorious prince riding wounded from battle (II.610–44), and near the end he supplements Filostrato with a formal portrait that reminds us again of Troilus’ masculine energy: ‘Yong, fressh, strong, and hardy as lyoun’ (V.830). If the prince is a lion on the battlefield, he is a lamb in the bedroom. In contrast to the amorous experience and competence of Boccaccio’s hero, Chaucer’s Troilus is innocent in love and hesitant in the pursuit of Criseyde. Not long before the consummation, both Pandarus and Criseyde are so dismayed by his supine inertness that they separately ask him ‘is this a mannes herte?’ (III.1098) and ‘is this a mannes game?’ (III.1126).
Chaucer gives other contrasting qualities to the character he inherited from Boccaccio. Troilus’ falling in love is treated with greater respect by means of a lengthy exploration of his physical and psychological responses, as we shall see in the next chapter; but, at the same time, moral criticism is stimulated by associating him with an original series of bestial images (Rowe, 74). Moreover, though Troilus’ conception of love is more idealistic and delicate in the English poem (service means at least as much to him as sex), he nevertheless participates eagerly in Pandarus’ schemes to deceive Criseyde, almost all of which are original in Troilus.
Given the range of traits that Chaucer has added to his Troilus, it is no surprise that the prince has been judged by critics as everything from a lovesick boy to a perfect lover, from a noble pagan deserving of heaven to an archetypal sinner. Early in the nineteenth century, in an echo of earlier views, William Godwin found him ‘the model of a true, a constant and a loyal lover’ (1.459); in our time, Donaldson has called him ‘the only unequivocally worthwhile person in the poem’ (Chaucer’s Poetry, 974), and Robert Burlin believes that he is an intellectual who merits a celestial seat (120–1). The prosecution, as we might expect, has been led by Robertson, who insists that the ‘heroic potentialities’ of Troilus are ‘undercut whenever they appear’ (Preface, 285) and who argues that his tragedy is ‘in an extreme form, the tragedy of every mortal sinner’ (‘Tragedy’, 36; see also Wood, Elements). Others have attacked Troilus more personally. Patch thinks that he is not masculine enough to satisfy Criseyde (Rereading, 89), Delany says that his ‘self-pity, self-deception and passivity become more and more prominent’ (84), and Stanley calls him ‘a poor fish’ (102).
Such disparate conclusions result from the preconceptions of individual critics and from their granting special privilege to one or two of the several elements that Chaucer has added to the characterization of Troilus, while ignoring or ironizing the others. None of these judgements is without some justification but, as so often with Troilus and Criseyde, each is limited. Such selective interpretation results in part from the pressure most academic readers feel to arrive at a definitive and total assessment of Troilus. Is he the model lover or a pattern of sin? Is he noble, comic or crazy? Is he to be admired or blamed? Even the few critics who find complexity in Troilus usually settle for some clear binary opposition, such as the contrast Wenzel develops between Troilus’ moral goodness in reference to chivalry as opposed to his moral blindness in relation to divine philosophy. Although Troilus, like the other major characters in Chaucer’s poem, has often been reduced to an ethical exemplum - or, rather, to a number of conflicting exempla - his greatest achievement is not in any lesson he can teach us, but in the depth and power of his experience.
The reader is allowed full access to Troilus’ psyche. In contrast to what is found in Filostrato, which tells us what all of its major figures are thinking and feeling, Troilus is the only character in Troilus and Criseyde whose consciousness is truly open. The extraordinary entry given to Troilus’ inner being is in keeping with the Chaucerian practice noted above of reconstructing each of Boccaccio’s similar major characters in a particular way. Unlike the deliberately restricted accounts of the hearts and minds of Pandarus and Criseyde (the former because he is relatively shallow and the latter because she is so deep), Chaucer allows us unrestricted admission to his hero’s mental and emotional life so that we always know his response to both suffering and joy. When the lovers first meet at Deiphebus’ house, we hear the public words of Pandarus and Criseyde, but the private musings of only Troilus (III.50–8). Later at Pandarus’ house the motives and sincerity behind Criseyde’s speech on jealousy are hard to determine (see especially III.1051–7), whereas the forces within Troilus that produce his answering swoon are explicated at length (III.1065–92).
Although rightly praised for his constancy, Troilus undergoes a radical change at the beginning of the poem from prince to lover, which the reader is able to follow closely. When we first see Troilus, he is a conventional member of a ruling élite, ‘kyng Priamus sone of Troye’ (I.2). Born to command, he leads his knights around the temple of Pallas (I.183–5), uttering words that are confident and smug in their criticism of any man who shows the slightest interest in women (I.194–203). As befits one so completely in control, he smiles, indulges in scornful irony, and is very pleased with himself: ‘And with that word he gan caste up the browe,/Ascaunces, “Loo! is this naught wisely spoken?”’ (1.204–5).
Once Troilus sees Criseyde, however, he is utterly transformed. Instead of an independent leader, the prince becomes ‘subgit unto love’ (1.231) and a ‘thralle’ (1.235, 1.439), like other strong, worthy and high-born folk who have been ‘overcome’ by love (1.243–4). Attacked by Cupid, Troilus retreats from the temple ‘thorugh-shoten and thorugh-darted’, as though he had been defeated on the battlefield (1.325). The change in his speech is especially revealing. He attempts to conceal his passion under cover of more blame of lovers (1.330–50), but he cannot sustain his old confident superiority and when he reaches the privacy of his bedroom we hear a new note: the ‘Canticus Troili’, which Chaucer adapted from a sonnet of Petrarch (I.400–20). Troilus moves from scorn to song (1.386–9), from a language of command and control to a questioning language of confusion and paradox: ‘If no love is, O God, what fele I so?/And if love is, what thing and which is he?’ (I.400–1). The oxymorons we now hear (‘O quike deth, O swete harm so queynte’, I.411) are repeated when Troilus first describes Criseyde to Pandarus: ‘my swete fo’ (1.874). He has experienced something overwhelming and indecipherable.
Troilus exchanges his public role as prince of Troy for the uncertainties of a lover. He enters a new and secret world when he decides ‘loves craft to suwe’ (1.379). Troilus remains a knight, but his real combat is now inward - not with the Greeks but with love. The proem to book I speaks of his ‘aventures’ in loving (I.3) and of his ‘unsely aventure’ (I.35), just as after first seeing Criseyde he privately concludes it was ‘to hym a right good aventure/To love swich oon’ (I.368–9). ‘Aventure’ is often glossed as simply good or bad ‘fortune’ or ‘chance’, which is its obvious meaning at 1.568 and elsewhere, but I would suggest the word also carries its full romance significance: ‘adventures’ are those extraordinary events by which a chivalric hero defines and realizes his high destiny, which is how Chaucer uses the word in The House of Fame (463) and ‘The Squire’s Tale’ (659). When Troilus commits himself to loving Criseyde, he embarks on a great adventure, which, as the Petrarchan song informs us, is as arduous as any war and equally associated with death and suffering. The riddling paradoxes of the song suggest that Troilus’ adventure in love will be one of exploration, a quest with no certain paths or easy answers. In contrast to the assured military commander we first saw, Troilus is now, after seeing Criseyde, ‘al sterelees withinne a boot’ (1.416). The familiar stilnovistic image might remind us of the lordless solitary of Old English poetry as Stephen Knight argues (43), but it could also suggest the more positive daring of those, like Constance in ‘The Man of Law’s Tale’, who commit themselves entirely to the will of God.6 Troilus’ God is Cupid, of course, and not the Christian Lord, but the knight’s worship is absolute when, at the very end of book I, he ‘dryeth forth his aventure’ (1.1092).7
The extent to which Troilus’ devotion leads him to reject all other values (I.463–8) is sometimes not sufficiently acknowledged. Patricia Kean declares that Troilus does not try to stop parliament from exchanging Criseyde because ‘he [puts] the commonweal before his private passion’ (Making, 130), and Donaldson similarly argues that Troüus’ integrity and ‘trouthe’ will not allow him ‘to substitute anarchy for law’ (Chaucer’s Poetry, 974). Criseyde will make such civicly responsible arguments, but never Troilus. He cares nothing for king, family, fellow-citizens, law or knightly reputation - only for his beloved.8 Hearing in the Trojan parliament that Criseyde will be exchanged for Antenor, Troilus wants to act, but first he thinks of her honour (IV.159). Then, in a passage added to Boccaccio, he vows that not his but her will be done: ‘And whan that she hadde seyd hym hire entente/Therafter wolde he werken also blyve,/Theigh al the world ayeyn it wolde stryve’ (IV.173–5). His principal reason for not accepting Pandarus’ advice to steal her away is the injury it might do to her good name: ‘As nolde God but if I sholde have/Hire honour levere than my lif to save!’ (IV.566–7). Although he is always ready to sacrifice Troy and his own honour, not to mention his life, Troilus will not risk Criseyde’s slightest disapproval. When he reluctantly accepts the exchange, it is only because she wishes it so.
Our access to the inner drama of Troilus’ love-adventure and awareness of its intensity do not preclude the possibility of different interpretations. For some his unshaken constancy in love makes him a romance ideal, and the English poets of the centuries immediately after Chaucer often referred to him alliteratively as ‘true’ or ‘trusty’ Troilus (Benson, ‘True Troilus’, 159). His sometimes painful sincerity is in sharp contrast to the artful pragmatism of Pandarus and Diomede. For example, although Troilus carefully rehearses what he plans to say to Criseyde at their first private meeting, he instantly forgets his planned speech once she actually arrives (III.78–84). He is incapable of carrying out an Ovidian plan of calculated seduction, for his is a genuine devotion. As a true believer, he dedicates his spirit to the God of Love, whom, along with Criseyde, he wishes to ‘serve’ (1.426, 1.430, 1.458). Some readers, however, attack the very nature of Troilus’ faith. Moralistic critics, for instance, find his ‘religion of love’ blasphemous at best and note its resemblance to Adam’s Original Sin (Wood, Elements, 99–128; Robertson, ‘Tragedy’).
Whether we think that the suffering and death associated with Troilus’ worship are signs of his nobility or his immorality, we must acknowledge that Troilus practises no comforting creed. Even when Pandarus announces that all is arranged for the consummation, Troilus’ immediate response, in a scene original with Chaucer, is ‘drede’ (III.707). His subsequent prayer to an array of pagan deities, during which he solicits their help in the name of their own unhappy love-affairs, prompts Pandarus’ disgusted response: ‘Thow wrecched mouses herte’ (III.736). This wonderful comic insult suggests another possible response to Troilus’ fervent religion of love: its extremism permits scepticism and even amusement. The reader may also have smiled earlier in the poem when Pandarus, in a parody of sacramental confession, urges Troilus to repent his past sins to the God of Love (‘Now bet thi brest, and sey to God of Love,/Thy grace, lord, for now I me repente’), though Troilus takes the advice very seriously indeed (I.932–8).
Although the intensity of Troilus’ love-adventure is undeniable, the prince evades unitary definition because Chaucer’s wide range of additions encourages us to interpret his character in different ways. If Pandarus is an author of texts, Troilus himself becomes a kind of text, even a textbook, of the lover, as he himself recognizes. Looking back over his life, Troilus addresses his god and declares:
‘O blisful lord Cupide,
Whan I the proces have in my memorie
How thow me hast wereyed on every syde,
Men myght a book make of it, lik a storie.’
As these lines suggest, the heart of the Book of Troilus is his transformation from knight to lover. His story is not about military conflict but about being ‘wereyed on every syde’ by Cupid: it is a narrative less of public chivalric deeds than of private amatory experiences. But if we know the subject of the text of Troilus, its literary genre remains in flux and so we are allowed to read it in many different ways.
Once the exchange for Criseyde is announced, destroying the happy ending that Troilus had seemingly achieved in his love- adventure (‘And Troilus in lust and in quiete/Is with Criseyde, his owen herte swete’, III.1819–20), the reader is offered a number of possible literary genres by which to understand the Book of Troilus. In keeping with the religion of love, the prince initially suggests that the story of his life be treated as hagiography: he imagines himself as dead and buried in a sepulcre that fellow-lovers come to visit (IV.327–9). The vision of himself as a martyred saint of love almost comes to pass. During the lovers’ final night together, the distraught Criseyde faints, and Troilus, assuming she is dead, draws his sword and resolves to join her: ‘Shal nevere lovere seyn that Troilus/Dar nat for fere with his lady dye’ (IV.1200–1). Had he fulfilled his intent at this point, Troilus would have indeed inscribed himself on the roll of those who died for love along with such as Pyramus, and we would be satisfied about the kind of story we were reading. Such a clear resolution is denied him (and us), however, as the genre of his life suddenly switches to something more like sentimental melodrama or farce. Criseyde regains consciousness, and his sword is put away amidst a flood of tears on both sides, leaving Troilus to act out a more problematic story.
Pandarus, as befits a maker of fictions, offers a number of genres that might be appropriate for the Book of Troilus. One of his suggestions, that Troilus find another woman, is reminiscent of Ovid’s love-manuals or of the fabliaux, whereas another, that Criseyde be stolen away, evokes the direct action of epic or romance, such as Paris’s rape of Helen. A more ominous genre for Troilus’ life is provided by his sister Cassandra: not romance of any kind, but de casibus tragedy. When asked to interpret Troüus’ dream of the boar, she tells him ‘a fewe of olde stories’ about how ‘Fortune overthrowe/Hath lordes olde’ (V.1459–61), the final victim of the wheel of Fortune being himself: ‘This Diomede is inne, and thow art oute’ (V.1519).9 The end of Troilus and Criseyde suggests at least two other literary forms that the Book of Troilus might take: military history and a sermon in contempt of the world. Having tried to be Pyramus, Troilus finally returns again to being the son of Priamus. Changed back into ‘this ilke noble knyght’ and no longer a lover, his life is now taken up with ‘many cruel bataille’ that ‘men may in thise olde bokes rede’ (V.1751–3). After Troilus’ death and ascent to the eighth sphere, Troilus is associated with a final genre. From his celestial perch he looks down with disgust and in the accents of a de contemptu preacher damns the ‘blynde lust’ of this world (V.1821–5). Although many in the last generation have accepted these final words as the correct way to understand the character of Troilus, they need not be given special precedence over any of the other literary genres we have noted. There are many ways to read the Book of Troilus.
Readers of Troilus and Criseyde have usually found Criseyde to be its most fascinating character. Henryson chose her as the protagonist for his powerful continuation of Chaucer’s story, and David Daiches has called her ‘the first truly complex heroine in post-classical European literature’ (100). She is complex even on the level of plot. Whereas Pandarus always plays the pander and Troilus, after the first few lines, the lover, she moves from reluctant object of desire to false betrayer after an intermediate period of amorous fulfilment. The English poets of the sixteenth century understood some of her possibilities: although they most commonly portray her as a type of the inconstant whore, she also appears as both an ideal sweetheart and a pitiable victim (Rollins; Benson, ‘True Troilus’). Modern Chaucerians also acknowledge the intricacy of Criseyde’s characterization, but their frequent practice, as we have just seen with Troilus, is to emphasize one or two dominant traits in order to produce a coherent interpretation. Few agree on which traits are most significant, however, with the result that the critical tradition offers us a number of radically different Criseydes.
The ‘false Cresseid’ suggested by Henryson’s Testament continues to be popular. In our time, Robertson has denounced her ‘pride and self-love’, though he finds her less guilty (because less serious) than Troilus (Preface, 487). Burnley notes that her refusal to accept responsibility and ‘a pervasive moral weakness in her character’ echo standard medieval attitudes towards women (‘Heart’, 32; cf. 38). Other critics stress her positive qualities. Tatlock called her ‘sweet, loving, and essentially good’ (Mind, 46), and more recently Mark Lambert says that she reveals the charms of the unheroic and is the Chaucerian character most like the poet himself (107, 125; see also David, ‘Comedy’). Even the most influential studies tend to deal with only a few of Criseyde’s qualities. In a brilliant analysis, C. S. Lewis argues (Allegory, 185) that her ruling passion is fear (‘fear of loneliness, of old age, of death, of love, and of hostility; of everything, indeed, that can be feared’); but, even if we admit some pervasive fearfulness in her character, it is not hard to find Chaucer’s Criseyde in more comfortable moods, including her lively curiosity about Troilus as a lover (II.498–504), appreciation of her own desirability (II.743–9), and self-confidence in her ability to trick her father and return to Troy (book IV).
The most satisfying analyses of Chaucer’s Criseyde are those that insist on how much more various her character is than that of Boccaccio’s heroine. Muscatine shows that she uses one idiom to speak with Troilus and another to speak with Pandarus (French, 153ff.), Lanham refers to her ‘multiple self and stresses the different ways she responds to different situations (‘Game’, 21–2), and Frank notes that Chaucer in the first three books has made her at once more sensitively emotional and calculatingly controlled (163). Her most sympathetic interpreter in this century, E. T. Donaldson, aptly calls her ‘a mystery’, which each reader will solve differently, but I would maintain that she is even more elusive than the ‘paradox’ he defines.10
Criseyde is not so much a collection of discrete traits, however diverse, as an endlessly protean figure who must be created anew with each reading. Troilus and Pandarus are complex characters because their thoughts and actions can be interpreted in different ways, but with Criseyde we must first decide what it is that she in fact thinks and does. Although we are constantly made aware, as is only rarely true with Pandarus, that Criseyde has psychological depth, the exact nature of this inner self, in contrast to the portrayal of Troilus, is often tantalizingly hidden from the reader. Many male critics claim that they have fallen ‘in love’ with Criseyde, but her real fascination is less erotic than literary. Her textuality dominates her sexuality. Criseyde may have no children, but with the reader’s help she is endlessly generative of fictions about herself.
Book I of Troilus suggests how we must read its heroine by repeatedly portraying her through the eyes of others who create their own individual Criseydes. Her public appearance and actions are fully available to these observers, but her inner being, whose depth is repeatedly insisted upon, remains largely hidden from view. The very first mention of Criseyde in Troilus adds a few lines to the account in Filostrato of her ‘drede’ and indecision after Calchas’s desertion (I.94–8, I.108), as if to assure us of the existence of her psychological life, but the stanza-length portrait that follows is drawn entirely from the outside, as revealed in the phrases ‘as to my doom’ and ‘seemed’ (I.99–105). The Criseyde thus seen or, rather, created by the narrator is a divine being (‘like a thing immortal’) of ‘aungelik’ beauty whose transcendent artificiality is stressed by the claim that she was ‘sent in scornynge of nature’.
In the following stanza we are shown another, more earthly Criseyde - this time from the perspective of Hector:
In widewes habit large of samyt broun,
On knees she fil biforn Ector adown
With pitous vois, and tendrely wepynge,
His mercy bad, hirselven excusynge.
Her dress and actions (kneeling, weeping) are emblematic rather than individual or personal and again observed only from afar. She resembles the very model of an innocent young widow, and this is how Hector envisions her: ‘[He] saugh that she was sorwfully bigon,/And that she was so fair a creature’ (I.114–15).
Our next view of the heroine is through the eyes of Troilus (‘His eye percede, and so depe it wente,/Til on Criseyde it smot’, I.272–3); this third male observer perceives a third Criseyde: not an angel or widow but a physical if courtly being.
She nas nat with the leste of hire stature,
But alle hire lymes so wel answerynge
Weren to wommanhod, that creature
Was nevere lasse mannyssh in semynge;
And ek the pure wise of hire mevynge
Shewed wel that men myght in hire gesse
Honour, estat, and wommanly noblesse,
The distant point of view, which is generalized to include that of any man (1.286), is insisted upon with words like ‘semynge’, ‘shewed’ and ‘gesse’. The passage is silent about Criseyde’s own feelings, and her single public action in this scene, a sideways glance, is tentatively interpreted by the narrator (‘ascaunces’) in a following stanza that begins and ends with accounts of Troilus’s emotions (I.288–94). In contrast to our full knowledge of the prince’s feelings, Criseyde’s inner being, here and elsewhere in book I, is largely inaccessible. Lacking a fixed human identity, she appears as a beautiful distant object, whose true character must be created by her readers. Like the three male observers within the text, each of us must construct our own Criseyde from the deliberately limited exterior information provided in the poem.
Criseyde’s lack of identity and her dependence on others for definition may be a result of her gender. In a pioneering article, David Aers sees her as a victim of male oppression and argues that the inferior condition of women in Trojan society denies her any chance to become a genuine individual. Such a feminist analysis, which continues to be developed in current criticism (Diamond; Dinshaw), is persuasive. Criseyde can be seen as a victim in both Troy and the Greek camp. Her father selfishly leaves her behind when he goes over to the Greeks, Pandarus dishonestly manipulates her into a relationship with his friend, the Trojan parliament treats her like a prisoner, and Diomede callously takes advantage of her vulnerability for his own selfish pleasure. Criseyde seems to achieve some measure of security only during the love-affair with Troilus, which occurs during the absence of her father. Once Calchas demands her back, however (his paternal rights in no way abrogated by his long neglect), Criseyde is once again described as her father’s daughter (IV.92, IV.663), as she has not been since early in book I (1.94), signalling both a renewed dependence, seen when she returns to in the presence of Calchas (‘And stood forth muwet, milde, and mansuete’, V.194), and her coming victimization by the predatory Diomede.
Feminist readings of Troilus emphasize the difficulty of Criseyde’s life in a patriarchal society, though their understandable political indignation need not obcure her triumph as a literary character. Chaucer seems sensitive to the social marginality of women, but he uses that historical condition to create a fictional heroine of extraordinary power and creativity. In contrast to Boccaccio’s Criseida, whose consciousness is fully as transparent as that of his other principal characters, Chaucer’s heroine is largely hidden from the reader. Her silence and lack of self-assertion, which accurately represent the traditional political state of women, have a positive artistic result because they create a space for readers to produce their own Criseydes. With respect to Chaucer’s characterization of Criseyde, less is certainly more. As Leo Braudy notes, this is also true for the characters in film, who, though apparently more limited and distant than characters in the standard realistic novel, are not necessarily inferior. More easily than novelistic characters, ‘film characters can leave their plots and inhabit our dreams, so free because they are so elusive’ (Braudy, 184). The elusiveness of Criseyde’s inner self makes her equally free and haunting.
My claim that the reader is frequently excluded from Criseyde’s psyche is not the dominant critical view. Many of the best Chaucerians have argued that we are able to see clearly into the heroine’s mind and heart, especially during her long inner soliloquy in book II. For instance, Jill Mann writes that Criseyde is unique in Troilus and Criseyde because ‘we are introduced to the minute-by-minute workings of her mind, to a complex notion of her psychological processes’ (Mann, Estates, 199).11 I would argue instead that we are only sporadically allowed into Criseyde’s mind and that even her private soliloquy, to be discussed further in the next chapter, provides little more than the illusion of interiority: we hear some conventional calculations rather than see her genuine feelings. A few critics have noted the lack of access to Criseyde’s inner being without discussing it fully, as in Robert Payne’s shrewd observation that ‘we are always moved outside her at critical moments’ (Key, 201; cf. Donaldson, Chaucer’s Poetry, 969). Again we may note the unique construction of Chaucer’s principal characters. His Criseyde differs from both Troilus, whose consciousness is continually on display, and from Pandarus, whose intentions are usually clear if rarely more than superficial. In keeping with our experience of most people in real life, Criseyde’s deepest self is deliberately withheld from us.
Troilus and Criseyde repeatedly incites us to wonder what its heroine is thinking and feeling while preventing us from certain knowledge. Such passages, almost invariably added to Filostrato, oblige readers to fill in the gaps for themselves and by so doing create their own heroine. For instance, soon after Criseyde is first told that she is loved by Troilus, Pandarus looks to a future time: ‘Whan ye ben his al hool as he is youre;/Ther myghty God graunte us see that houre!’ (II.587–8). The poem then gives Criseyde’s provocative spoken words in response (‘“Nay, therof spak I nought, ha, ha!” quod she;/“As helpe me God, ye shenden every deel!”’, II.589–90), but we are told nothing about her private thoughts. At least one critic, however, is confident that he knows what Criseyde is thinking: she now fully understands that Pandarus’ intent is to bring the proposed affair to a physical union and is both pleased and offended by that knowledge (Markland, ‘Pilgrims’, 74). This is a plausible interpretation, to be sure, but by no means the only one. The text incites interpretation but prevents certainty.
Other important scenes also keep readers at a distance from Criseyde’s private thoughts. When Pandarus brings his niece to speak with Troilus for the first time, he begs her to act boldly and secretly to cure her lover’s pain (II.1730–50). Immediately before her uncle’s words, we are told that Criseyde is ‘al innocent of Pandarus entente’ (II.1723), but we are not permitted to observe her response once that intent has been made clear. When Pandarus a few lines later puts the lovers on notice that he will soon summon them for a longer private meeting at his house, Troilus eagerly responds, ‘How longe shal I dwelle,/Er this be don?’ (III.201–2), but the text again provides not the slightest indication of Criseyde’s reaction to her uncle’s promise.
The most famous instance in which we are kept out of Criseyde’s mind at a crucial moment in the story occurs just after Pandarus finally does invite Criseyde to his house, though ostensibly only for dinner. When Criseyde asks whether Troilus will be there, Pandarus says that he will not be, but adds that she would have nothing to fear even if he were (III.568–74). The narrator then addresses us directly:
Nought list myn auctour fully to declare
What that she thoughte whan he seyde so,
That Troilus was out of towne yfare,
As if he seyde therof soth or no;
But that, withowten await, with hym to go,
She graunted hym, sith he hire that bisoughte,
And, as his nece, obeyed as hire oughte.
Many have assumed that they knew precisely what ‘she thoughte’. In 1913, William Dodd suggested that ‘on the whole, the poet subtly makes us feel that Pandarus’s reassurances are sufficient to allay the heroine’s suspicions, and that she went to his house in innocence’ (170), but during the same decade Kittredge expressed what has become the majority view: ‘Pandarus lies, of course, but it is perfectly clear that she does not believe his protestations’ (Poetry, 132). The lines themselves support neither opinion; or, rather, they are capable of supporting either and several others as well. The only interpretation offered in the text itself is purely conventional (if slyly humorous) and exterior: Criseyde obeyed her uncle as a niece should. Yet, even though the passage keeps us in ignorance, we cannot easily dismiss the problem of what Criseyde really thought because, here and elsewhere, the lines provoke us to seek an answer. The ‘author’ invoked at the beginning of the stanza is no help, for the entire scene is original with Chaucer. Once again it is the reader, each individual reader, who is left ‘fully to declare’ what lies behind Criseyde’s public words and behaviour.12
Our distance from Criseyde is especially pronounced during her betrayal of Troilus in book V, which is narrated largely from outside. Although Chaucer provides some of her public words and actions, he continually makes us guess what is occurring in her mind and heart at this crucial point in the narrative. For example, soon after her arrival in the Greek camp, Criseyde states her resolve to return: ‘To Troie I wole, as for conclusioun’ (V.765). The narrator then informs us that before two months were over she was very far from such an intention and had let both Troy and Troilus ‘knotteles thorughout hire herte slide’ (V.766–70), but he never shows how and why that change occurred.13 The closest he comes to an explanation is a single stanza describing her response after Diomede has both predicted the utter destruction of Troy and offered to serve her in love:
Retornyng in hire soule ay up and down
The wordes of this sodeyn Diomede,
His grete estat, and perel of the town,
And that she was allone and hadde nede
Of frendes help; and thus bygan to brede
The cause whi, the sothe for to telle,
That she took fully purpos for to dwelle.
The stanza is more of a list than an analysis, in which it is difficult to determine the relative significance of the forces that are noted. For instance, is Criseyde’s fear of the fall of Troy more or less important than the attractions of the friendship of the forceful high-born Diomede? Furthermore, even this rare insight into the heroine’s consciousness is severely limited. The thoughts are only an early stage in her change of resolve: they only ‘bygan to brede’ the reason that she decided to dwell among the Greeks, and they say nothing directly about how she comes to love Diomede, if, in fact, she ever does.
The narrator earnestly tries to makes sense of and mitigate blame for Criseyde’s actions in the Greek camp, but his confusion and bafflement only underline the elusiveness of her character. He reminds us that his knowledge of his heroine comes from the writings (often incomplete) of others, and he is thus reduced to interpreting a variety of public symbols:
And after this the storie telleth us
That she hym yaf the faire baye stede
The which he ones wan of Troilus;
And ek a broche - and that was litel nede -
That Troilus was, she yaf this Diomede.
Like a good semiologist, the narrator examines the signs provided by other accounts (her giving Diomede a sleeve as a pennon and weeping over his wounds), but he admits that he has no authoritative access to her deepest feelings: ‘Men seyn - I not - that she yaf hym hire herte’ (V.1050). Whatever others say, he does not know what she did with her heart. Soon after this, Criseyde herself disappears from the narrative except for one last appearance: Chaucer gives verbatim a final letter to Troilus (it is only summarized in Boccaccio), which is so cold and written in such apparent bad faith that we realize that we no longer have any idea what Criseyde is doing in the Greek camp, let alone what she is really feeling and thinking (V.1590–1631). She has become almost totally opaque.
On one of the few occasions that we are allowed any access into Criseyde’s mind during book V, we discover that she imagines herself becoming a literary text. After her betrayal of Troilus, Chaucer adds lines to Boccaccio based on Benoît’s Roman de Troie in which Criseyde predicts her metamorphosis into a lesson of falseness in any number of forms:
‘Allas, of me, unto the worldes ende,
Shal neyther ben ywriten nor ysonge
No good word, for thise bokes wol me shende.
O, rolled shal I ben on many a tonge!..
We will remember that Troilus also imagines his life becoming a text in book V (V.583–5). But, whereas Troilus proudly declares that his amorous sufferings are worthy of being enshrined in writing, Criseyde is terrified of what books will do to her. She fears that she will be made into a exemplum that will only present her negatively: everything said about her until the end of the world will be only bad (‘no good word’). This is an accurate description of how her character appears in most literary works, but not in Troilus and Criseyde, where she has proved capable of generating stories ‘on many a tonge’ that range throughout the spectrum from good to bad.
Chaucer’s Criseyde herself is what recent literary theory would call an open text - in absolute contrast to the closed fictions of Pandarus that are designed to produce a single practical end. The same lack of consistent identity (‘slydynge of corage’, V.825) that Criseyde rightly fears will bring her moral condemnation is used in Troilus to produce a brilliant literary success. Using the familiar, if overly rigid (Kermode), terminology of Roland Barthes, we might see Chaucer’s Criseyde as a supreme example of a scriptible rather than a lisible text. She does not represent a unified or even complex authorial statement whose meaning we must passively accept, but is instead a stimulating figure that challenges each reader to make her new.
My claim that Chaucer’s Criseyde is an open character who is capable of generating multiple texts may seem inappropriately modern, but there seems some evidence for it in the work that first told of her love for Troilus: Benoît’s twelfth-century Roman de Troie. In the long final speech of his heroine, Benoît has her utter a series of contradictory statements, which he makes no attempt to resolve: she says nothing good will ever be said about her, that she acted wrongly and stupidly in abandoning Troilus, that she was ruined by listening to Diomede’s speeches, that she has brought disgrace upon womankind, that she had the best lover a woman ever had, that there is no reason to repent, that she should be true to the valiant and worthy Diomede, that she never would have deceived Troilus if she had remained in Troy, that she has conquered her lonely state in the enemy camp and achieved contentment, that one should never suffer because of what people might say, that she is now both happy and sad, that she wishes Troilus well, that she must grant Diomede all that he wishes to keep him in love with her, and finally that she hopes to achieve joy and happiness (lines 20, 237–340).
The speech in Benoît does not characterize a single person; rather, it sketches an anthology of diverse attitudes to what has happened (fearful, hopeful, repentant, triumphant, opportunistic, satisfied, deceived, conflicted, nostalgic, scheming and proud), each one of which has the potential to be developed into the portrait of a different kind of woman. In addition to supplementing Boccaccio with individual lines from this powerful speech, which is not reproduced in Filostrato, Chaucer seems to have been inspired by its variety to create an even more open Criseyde in Troilus and Criseyde. Although Benoît eventually draws the standard moral about female unfaithfulness from the action of the episode, the words he gives to his heroine suggest an infinite number of narratives. Chaucer does not himself write these stories (what single author could?), but he empowers each reader to attempt such rewriting.
The narrator, a figure who has often been considered a fourth major character in Troilus and Criseyde, must also be discussed, however briefly, in this chapter, if only to show that he does not really belong. Although his many comments on the action provide some justification for linking him with Troilus, Criseyde and Pandarus, he never becomes a fully developed, independent human character. He plays no part in the action itself, despite his emotional involvement, and it is often impossible to distinguish his statements from those of the poet. The narrator in Troilus is better seen as a flexible literary voice than as a human personality - a rhetorical element of the text that Chaucer uses to create a number of different effects. The most important function of this voice may be to reveal the silences and uncertainties of the poem and thus encourage the interpretive role of each reader.
Despite the autobiographical claims of Boccaccio’s Filostrato, its narrative unfolds with remarkable objectivity, as discussed in chapter 2. Not so Troilus and Criseyde, to which Chaucer has added an intricate layer of narrative commentary. Substantial invocations precede each book except the last, and the reader is frequently addressed in casual asides: after a familiar proverb, for instance (‘This, trowe I, knoweth al this compaignye’, 1.450), or to mark a change of scene (‘Now lat hire siepe, and we oure tales holde/Of Troilus…’, II.932–3). We are also frequently told about the process of composition: the narrator discusses his source (a certain Lollius), announces his approach (the love-story rather than the war), admits gaps in his information (even when, as in respect to Criseyde’s children, Filostrato is perfectly clear) and justifies his abridgements. Even more striking are direct expressions of enthusiasm about the progress of the narrative and of the love-affair. The narrator urges on Pandarus’ first visit to Criseyde (‘Now Janus, god of entree, thow hym gyde!’, 11.77) and wishes he had experienced such a night of amatory joy (‘Why nad I swich oon with my soule ybought’, III.1319). Criseyde is an object of special concern; to choose two famous examples among many, the narrator defends her against possible accusations of falling in love too quickly (II.666–79) and is sympathetic even after her betrayal (V.1093–9).
For a long time, when they thought about it at all, critics assumed that such first-person comment, some of which is obviously comic, was in the poet’s own voice. Donaldson seems to have been the first to treat the Chaucerian narrator as a separate character. In an influential study of the Canterbury Tales (‘Chaucer the Pilgrim’), he argued that the portraits in the General Prologue were drawn not by the poet himself but by an independent persona, who was enthusiastic, naïve, and not wholly to be trusted. In subsequent articles and in his edition of Chaucer, Donaldson posited an equally independent and unreliable narrator for Troilus and Criseyde, whose sentimental opinions, especially in defence of Criseyde, we are meant to question. In order to make this point, Donaldson was forced to posit a fully fictionalized character, who is described in psychological metaphors that endow him with will and emotions. In one article the narrator, whose masculinity Donaldson insists upon, is portrayed as acting ‘irritably’, ‘knowing’ something, and wanting the audience to ‘share his enthusiasm’ (‘Masculine’, 54–5); in another as experiencing ‘one of his tenderest moods’ (‘Criseida’, 71); and in a third as suffering ‘internal warfare’ that results in ‘a kind of nervous breakdown in poetry’ (‘Ending’, 91).
At about the same time Robert Jordan also argued that the narrator in Troilus plays a ‘role… central to the life of the poem’ (‘Narrator’, 237). Jordan’s narrator is as fully personalized as Donaldson’s and similarly unreliable. A good storyteller, performer, reporter of facts and dispenser of commonplaces, he is finally ‘a man of no wisdom’: ‘Although warm hearted and ingratiating, he is remarkably obtuse, completely imperceptive of the esthetic and moral grandeur of his own creation’ (254). In response to the objections of Bertrand Bronson, Jordan later denied that he meant the narrator to be taken as a genuine character of the same order as Troilus, Criseyde and Pandarus (Shape, 67), but many subsequent critics have conceived of him in just this way.14 Often linked with Pandarus, the narrator is treated by some as a disturbed personality suffering from voyeurism or prurience.15 More positive interpretations also see him as an independent and coherent being. While discussing several different views, Ida Gordon, a critic in the Robertsonian tradition, approvingly cites Muscatine, a decided non-Robertsonian, to support an interpretation of the Troilus-narrator ‘as a persona distinct from the poet’ who, like Boethius, is ‘brought gradually to a clearer vision as the story proceeds to its inevitable end’ (Double, 61). Wetherbee also sees Troilus as the autobiography of the narrator, and a recent article by Carolyn Dinshaw using the insights of feminism and contemporary theory continues to accept the narrator of Troilus as a discrete and consistent consciousness.16
Despite the emotions and opinions he expresses, the narrator of Troilus and Criseyde is a character or independent persona only in a very limited sense. Unlike the narrators of the Divine Comedy, Pearl, Piers Plowman or Chaucer’s earlier dream visions, whose experiences truly are central to their respective works, the voice we hear in Troilus is never given physical shape and does not participate directly in the events of the poem. He remains, like us, only a reader, isolated in time and space from the genuine characters in the poem, whom he cannot affect and who remain oblivious to him.
A more radical problem with the narrator is the difficulty of defining the extent of his presence in Troilus. Is he there only when directly addressing us in the first person or is he responsible for the entire exposition of the poem?17 William Provost attempts to distinguish five narrative modes in Troilus (direct narrative, summary narrative, description, narrator’s comment, and invocation), but he is forced to admit that these ‘five modes, alas, are not always as distinct as we might wish’ (56–8). That is surely the point. How can we hope to distinguish between neutral description and subjective comment in Troilus? It is not difficult to identify an unreliable narrator in the various defences of Criseyde, but is the summary of the Trojan War (I.57–98) equally suspect (it contains strong opinions about Calchas) and what about the following portrait of Criseyde that describes her as angelic and heavenly (I.99–105)? It is hard to know where to draw the line. The ending of Troilus has been a particular problem in defining narrative presence. Whereas many have seen a retreat into conventional moralizing in the conclusion, and Donaldson imagines the narrator undergoing a nervous breakdown (‘Ending’), others hear the voice of the poet himself at last. And there is further disagreement among those who hold this last position: Gordon finds the ‘mature, humane poet himself taking over from the naïve narrator as early as V.1093 (Double, 87), whereas Jordan argues that it is not until the last twelve stanzas that Chaucer speaks in his own voice (‘Narrator’, 253).
Mehl suggests that little is gained, and much lost, by anachronistic attempts to separate clearly poet from naïve narrator (‘Audience’, 180), a rigid division called ‘more convenient than true’ by Salter (‘Poet’, 282), who warns that the identification of an unreliable narrator as the source of the poem’s quandaries may fail to recognize ‘what may be the poet, making his own statements, tentative as they may sometimes be, about the problematic background to his artistic decisions and procedures’ (286). The narrator is certainly not very consistent. He never becomes a familiar companion, like Conrad’s Marlow; but, rather, in Payne’s phrase, offers a ‘multiplicity of perspectives’ (Chaucer, 85). As Bloomfield first demonstrated (‘Distance’), the narrative voice moves between close involvement and historical distance, nowhere more abruptly than in the three formal portraits in book V (799–840); and the voice is much more prominent in the first three books of Troilus than in the last two. If the narrator sometimes sounds sentimental and naïve, he can also be authoritative, especially in the proems and in such summary judgements as ‘And thus Fortune a tyme ledde in joie/Criseyde and ek this kynges sone’ (III.1714–15). Other first-person passages, such as his comment during the consummation that ‘Resoun wol nought that I speke of slep,/For it acordeth nought to my matere’ (III.1408–9), are comically self-conscious and highly sophisticated.
David Lawton has recently proposed that we regard the narrator of Troilus not as a consistent persona but as a variable rhetorical device used to emphasize different moments in the text. Even the shift in mood and value at the very end of the poem is not a shift in voice: ‘It does what the first-person narratorial voice has done throughout: it responds appropriately to the particular stage of the work’s unfolding’ (Narrators, 82). Lawton rejects the idea that this device is a fourth major character; instead he sees it as the neutral ‘voice of performance”: ‘almost the voice of the poem itself speaking from the time and continuum of its own performance’ (89). Lawton’s formulation, which somewhat resembles Wayne Booth’s conception of the ‘implied author’, may minimize the extent to which the narrative commentary complicates as well as supports other aspects of the poem, but he seems quite right to regard the voice as textual rather than personal.
If the narrator never becomes a distinct and coherent character in Troilus and Criseyde, the effects of his various comments are important and multiple, especially for the reader. The first-person passages that celebrate the love-affair and empathize with both the joys and sorrows of the lovers increase the emotionalism of Boccaccio’s story. These invitations to empathy are a genuine element in the poem, which cannot be dismissed as merely ironic, though ironies of various kinds may also be present. The sympathy extended to Criseyde after she decides not to try to return to Troilus is especially poignant. The regretful pity expressed for one both reviled and miserable is neither naïve nor sentimental, but mixes pathos with tragedy:
Ne me ne list this sely womman chyde
Forther than the storye wol devyse.
Hire name, alias, is publysshed so wide
That for hire gilt it oughte ynough suffise.
And if I myghte excuse hire any wise,
For she so sory was for hire untrouthe,
Iwis, I wolde excuse hire yet for routhe.
In addition to stirring our emotions, the narrative voice also tests our judgement. Its comments force us to question not only the ultimate meaning of the story but also its very telling. Did Criseyde have any children and how closely does the poem follow Lollius? Admissions of ignorance about particular facts and the inability to describe the full joy of the lovers remind us that we are reading a fiction that we must interpret. Recognition of the narrator’s unreliability does not make our job any easier. We are right to be suspicious of narrative claims that defend the pace of Criseyde’s wooing or that certify the characters’ good intentions, but these suspicions do not automatically point to a specific ‘right’ answer, although this is what critics in the tradition of both Donaldson and Robertson often suggest.18 The narrative voice in Troilus creates openness rather than certainty. No deconstructionist is needed to point out the gaps, subjectivity and contradictions in the text as long as the narrator insists on them himself. Through this voice Chaucer renounces any claims to authorial omniscience and empowers the reader. The central experience of the poem is not the narrator’s but ours.