The focus of this review is what the translator himself sees as his significant offering to Cavafy studies: its reform of Cavafy's poetic economy by rearrangement of the poems, in a bid to ‘help us think about Cavafy in a different way’ (307). Jones has rejected the standard editorial practice following G. P. Savidis, which established the standard presentation of the Cavafian ‘canon’ (those poems published by the poet in his lifetime). This canon had first been published in 1935 by Rika Sengopoulou in chronological order. In his 1963 edition, Savidis writes that he preferred Cavafy's own ‘thematic, not chronological’ arrangement of the published poems.Footnote ¹ This was ‘thematic’ in the sense that it reproduced the ordering of the poems which Cavafy had chosen for the two last collections he distributed himself. Most translations have more or less deferred to Savidis's arrangement. The problem with this approach is that Cavafy continually renewed and edited the collections until his death. The posthumous editor of a collected volume therefore has the choice of arranging the poems either by some editorially imposed order (e.g. chronologically) or by following Cavafy's own groupings. The latter route (taken by Savidis) is beset by choices between earlier and later circulated collections where there is some overlap and particular omissions. The classic editions of Savidis have therefore been criticized for concealing the editor's decisions by presenting them as the objective determination of the poems by the poet himself. Put simply, to edit Cavafy is to make particular choices of our own. Where then to ‘freeze’ Cavafy's evolving process of self-curation?

J. enters confidently into this zone of indeterminacy. For J., ‘[t]here is a sense … that Cavafy's poems were in flux during his lifetime’ and ‘[w]hen he died, he left them in flux’ (311). J. has taken up the challenge of rescuing the poems from Savidis. In line with J.'s self-identification as a ‘poet-translator’ (cf ποιητὴς ἱστορικός) (319), his edition takes a more creative turn than the philologist Savidis. As for the text, it often feels that J. is at pains to distinguish himself from earlier translators at the cost of more natural choices. J. has chosen a thematic arrangement, but one merely inspired by Cavafy rather than claiming (like Savidis) to be determined by him. J. has arranged the poems into ten thematic sets, based on a ‘thematic catalogue’ devised by the poet himself and published in a revelatory article in 1996 by S. Ekdawi and A. Hirst.Footnote ² J.'s own sections are titled ‘Ancient Days’, ‘Hellenistic Kings’, ‘Romans’, ‘The Beginning of Christianity’, ‘Byzantines’, ‘Passions’, ‘Fleeting Years’, ‘Prisons’, ‘Our Art’, ‘Portraits and Memorials’. Six of these are identical to Cavafy's, while J. has taken some liberties with the remaining sections. The contents mistakenly list all the Roman and Christian poems under the ‘Byzantines’ section, but this error is not repeated in the text on pp. 81–102. J. has broadly followed the incomplete allocation of certain poems by Cavafy to his corresponding sections, but the great majority he has had to place himself. J. has not included every known poem, and notably he has excluded certain ‘juvenile’ early poems from the so-called ‘repudiated’ works, and snubbed thirty ‘unfinished’ poems he views as mere ‘drafts’ (317). These are defensible choices for volume addressed to the general reader. The significance of the translation lies in its creative arrangement of the poems.

Since these problems of arrangement have been weaponized to support particular readings, they are not simply philological technicalities. There are two important questions for J.'s new arrangement: how it supports or resists the conventional separation of elements in the work (historical, philosophical, erotic), and how it portrays Cavafy's development as an artist. As J. observes in his afterword, the chronological arrangement of Cavafy's poems supports a linear conception of Cavafy's artistic development (311). Therefore the placement of the earliest poems is perhaps most revealing, since these have frequently been seen as immature trash at the start of a direct path of growth. J.'s arrangement reveals instead a more complex web of continuities and ruptures. The placement of some early poems alongside many of the 1910s–20s in ‘Portraits and Monuments’ reveals the poet's ongoing obsession with legacy and memory. Two poems which had been identified by Hirst as representing opposite approaches to eroticism are placed together in the set titled ‘Passions’: Ἐπιθυμίες (127) and Θυμήσου, Σῶμα (140). For Hirst, these works show ‘how difficult it would have been to combine all the published poems in a satisfactory single thematic sequence, for the two poems present irreconcilable points of view’.Footnote ³ There is something boldly Cavafian about J.'s decision to let the poems speak for themselves here — to let them confront each other. Even more confronting is the contrast between the ‘forbidden’ nature of the erotic in Ἡ ἀρχή των (135) which literally faces the aesthetic decadence of Μακρυά on the opposite page (134). Here Cavafy is a poet of contrasts, who resists the imposition of simple models of artistic growth.

The section ‘Passions’ makes plain the conflicted evolution of the poet's approach to eroticism. The important question which remains, however, is whether the apparent isolation of the ‘passionate’ poems from the rest merely reinforces the traditional schism between the historical, philosophical, and the erotic. Margaret Alexiou has asked, ‘how “objective” is the traditional distinction between the “didactic”, “mythical”, “historical” and “erotic” (or “personal”!) poems?’Footnote ⁴ The poems chosen for J.'s ‘Passions’ follow the ostensible criterion of being centrally concerned with the state of desire. Yet there are erotic poems with more varied themes distributed throughout J.'s ten sets. The fact has been long recognized — and, in some quarters, suppressed — that the erotic element in Cavafy is not merely confined to these poems, but deeply penetrates even the so-called historical poems. Thus we are pleased to find among J.'s ‘Hellenistic Kings’ the sensual poem Τεχνουργὸς κρατήρων (34), and the lament Μύρης⋅ Ἀλεξάνδρεια τοῦ 340 μ.Χ. (93–5) in the ‘The Beginnings of Christianity’, among many more examples in later chapters. J. has not tried to confine the erotic to one corner. It is odd, however, that ‘Passions’ begins with the poem Κρυμμένα (119), which appears to emphasize the self-censorship of Cavafy's approach to eroticism. The extent of Cavafy's inhibitions has long been questioned: the erotic is hardly less present in the published work than in the ‘hidden’ poems. That so many of the more explicitly erotic poems do not bear the stamp ‘hidden’ (which J. has affixed to each of the unpublished works) demonstrates the difficulty of pinning down Cavafy's poetic practice.

What Cavafy wanted to hide, and how desperately he wanted to suppress certain poems, remains to be reconstructed from his archived scribblings. The uncertainty which attaches to whether we will ever truly know Cavafy's intentions is not merely a curse for scholarship but a creative opportunity for poet-translators like J. When Cavafy's body was buried in 1933, he was still not dead in the sense that Barthes meant when he proclaimed ‘the death of the Author’. We have been taught by Savidis and his successors to follow Cavafy's own determination of his poems. But Barthes reminds us that whenever we are told to defer to the author, we are more often obeying their critics. In choosing to press mute on all this Cavafological noise, J. has begun a process of reimagining the unity of Cavafy's poems. This might just allow us to know Cavafy, not as the silent pupils of his scholiasts, but as engaged readers in our own right.