Emma Scioli's book seeks to bring together texts and visual art by examining dreams in Roman elegy and forms of visual art such as landscape paintings (especially ‘Second Style’). While there is nothing unusual about setting these two modes of representation in parallel, and using one to illuminate the other, this approach has been used less often in analysis of Roman elegy, or has been limited to deliberating about what particular work of art might have provided inspiration. Two aspects of S.’s book distinguish it from previous scholarship. First, she concentrates on analysing dreams and dream-texts in Roman elegy, and on the parallels between that particular kind of vision experience and visual works of art. Secondly, and equally important, she pushes beyond the iconographic parallels to be found, stressing that we should not try to identify one particular painting in order to make these parallels work. Thus S. says (in regard to Propertius 3.3) that her intent is, ‘not to establish that works of art like these either did or could have influenced Propertius's poem — at least not directly. Instead, I suggest that considering works like these in conjunction with one another can help to reveal and illuminate mechanisms at work in the poem itself’ (135). It is not necessary that we posit the existence of a lost painting behind a dream-text in order for us to consider the compositional techniques in a work of art that make us appreciate elements of the poem such as illusion of space. Further, to insist that a poem is modelled on a certain work of art and visual composition limits the fluidity of imagination that mythological allusions demand.

On what basis does S. decide to elucidate the texts she chooses by using visual art as a lens? In the four poems she analyses in depth —Tibullus 1.5, Propertius 2.29a, Propertius 3.3 and the rape of Rhea Silvia in Ovid, Fasti 3 — vision, viewing and representation are the most important elements and the ones most analogous to the visual arts. The dreamer is a viewer, who tries to distinguish between real and representation, the internal and external worlds of the dream. The narration of the dream (on which S.’s attention is focused) removes the reader or narrator from the essential experience and creates a verbal image which we might call ekphrasis. The images in this ekphrasis correlate in some respect to elements in the non-dream world. The same tensions between representation and response, and between the primary experience and the telling of it, exist in visual art. Both feature a first-person meditation on an experience that tries to bridge the gap between introspection and exhibition.

S.’s choice of Roman elegy for her focal dream-texts seems an obvious one. Both Roman elegy and dream-texts are deeply centred on first-person viewing, narration and reactions that question illusion and reality, interiority and external representation, and the malleable rôle of the author/viewer/narrator. The same troubling and paradoxical issues bedevil our attempt to understand the two kinds of texts. Each is based on a first-person response to a deeply personal situation, but this response removes us from the immediacy and intensity of the original experience, distorting the experience in the telling of it. Further, elegy and dream-texts, as well as visual art, depend on the use of illusion, which deceives and confounds the categories of fake and real (on this, see the important book of Patricia Cox Miller, Dreams in Late Antiquity (1994)). Illusionism can easily lead us astray in its ability to disguise the very techniques it uses; the reader/viewer may not be aware that we have been led into this ‘twilight zone’ (Cox Miller 1994: 4).

Dreams/dream-texts/dreamscapes are not easy to pin down or define, and S. has some problems with this. Even her choice of texts calls definitions into question. She begins with Tibullus 1.5. Unable to call this a ‘dream’, S. resorts to ‘fantasy’ (or phantasia). Unlike Propertius and Ovid, the Tibullan text does not include any night-time dreams, but this Delian fantasy qualifies as a ‘dream’ in that it is an imagistic and creative act isolated from the framework of the poem. S. refers to this episode as ‘dreamlike’, appropriating the term ‘dream’ from those who would use it casually. The descriptor ‘dreaminess’ is often used of Tibullan poetry to define its elusive, illogical, disjointed nature, sometimes in a pejorative sense (56–7). The fact that S. must use the word dream in quotes illustrates the difficulty of defining exactly what it means. The passages examined in Propertius 2.29a and 3.3 and Ovid, Fasti 3 are more readily identified as dreams, each one dependent on a vocabulary of seeing (videre) and on translating the primary experience into a textual narrative that describes the experience.

S. has written a pivotal book that brings together the worlds of dreaming, Roman elegy and visual art, visual and textual points of reference that are rarely brought into the same orbit. She at times over-relies on the work of other scholars to help make her points (for example, Walde); at other times she might have made better use of important works such as Cox Miller (see above) or J. Amat, Songes et Visions (1985) — the latter is focused on later Latin texts but has valuable observations for the earlier material. In her ch. 3 on Propertius, where S. comments that Propertius ‘aligns himself with these heroines [Ariadne and Hero]’ (132), one wishes for more exploration of this topic and citation of more work on Propertius’ feminine side (for example, P. A. Miller, Subjecting Verses (2004), ch. 5). These quibbles aside, there is much to be learned from S.’s book, which brings together familiar texts and works of art that take on extended meaning when considered together in intersecting skeins.