Anyone who has ever taught or studied the Aeneid will be familiar with student gripes that the protagonist, Aeneas, does not meet their expectations of a hero: stolid, boring, wooden, uninspiring, lacking in emotional range. Likewise, students of Lucan's Civil War often find it hard to get a handle on the figure of Cato, and his hard-line heroics are usually met with a combination of disbelieving horror and ridicule. The important and deceptively simple suggestion of J. Mira Seo's new monograph is that such apparently two-dimensional and unsatisfactory ‘problem characters’ in Latin literature (19) are the result not of the failure of the ancient poets to depict their protagonists successfully, but rather of the different expectations that Romans held about literary characterization. Her book sets out to explore the possibility that Roman writers were not attempting to present characters who are psychologically ‘rounded’ in the way that we moderns expect, with our Cartesian approach and our high regard for radical individuality and subjectivity. Rather, she argues, Roman characterization was based on a distinctively Roman approach to self as ‘aemulatory, referential, and circumscribed by traditional expectations of society’ (15). For Seo, characterization is a literary technique (4) rather than mimetic of real people (5) and, like genre, characters in literature are established through reference to earlier material. Indeed, characterization is a form of allusion, and characters in literature are ‘nodes of intertextuality’ (4) created out of generic expectation and familiar schemata, and the significant and creative modification of these. This technique is often evident in ancient literature (the intertextuality of Virgil's depiction of Dido is well known); however Seo pursues its implications through close readings of five case studies: Virgil's Aeneas, created through the conflicting voices of fama, with effeminate Paris as his ghostly doppelganger; Cato as Lucan's lethal exemplum; Seneca's Oedipus, becoming ‘himself’ under the pressure of decorum and the literary tradition; and two of Statius' most stereotypical and over-determined characters, the archetypal ‘doomed beautiful youth’, exquisitely intensified in the figure of Parthenopaus, and the doomed prophet Ampharius. In her series of illuminating and insightful readings, Seo shows how such characters are built up through schematization, through articulation from a variety of perspectives in the texts, and through the evocation and skilful modification of familiar literary motifs. Although I am not sure she has entirely cracked the problem of Roman characterization, her book opens up a stimulating new approach to Roman poetry and characterization, which I hope will inspire others to take up the call for more research in this area.