Two decades after Albert Lord's death, oral-traditional performances remain essential for understanding Homeric poetry, despite the impression given by recent paleo-Analytic disquisitions. Three generations of comparatists have followed the author of The Singer of Tales and his mentor Milman Parry in fruitfully applying ethnopoetic approaches to problems from diction to dialect, motif to multiformity, and the intricacies of ‘textuality’, inter- and intra-. J.'s narrower concern is to extract from reports of others' fieldwork enough circumstantial detail to make plausible her long-held theses that the Iliad and Odyssey were dictated, once and for all, under the Pisistratids, and that their book divisions reflect that process. Her proposed textualisation succeeds no better than have predecessor silver-bullet solutions to the conundrum of how ‘Homer’ got created. But the attempt at least brings to a more general audience some of the best ethnographic work done in the last 40 years concerning the composition and performance of long mythico-heroic poems.

The book comprises two unequal segments. Chapters 9–11 represent a re-statement and expansion of the core of J.'s The Homeric Question and the Oral-Formulaic Theory (1980), together with several papers from the 1990s. She continues to argue from motive and opportunity: only tyrants would have enough wealth, power and desire for symbolic capital to undertake production of a full text; the circumstantial evidence of an Athenian phase in epic transmission, Athenian prominence in probable interpolations and the persistent lore about a ‘recension’ all fit the puzzle. As in her earlier book, J. ventures no literary interpretation, and makes no demonstration that the Iliad and Odyssey began as oral poems, contenting herself with showing that, assuming they did, a dictation-event provides the most reasonable scenario for how both became texts.

Hers is not the first such speculation (she acknowledges Janko and others), but it excels in painterly detail worthy of Alma-Tadema. The time is late August, 522 b.c., the place the Acropolis; Cynaethus of Chios, who has just been awarded first prize in the rhapsodic contest at the Greater Panathenaea, has been persuaded to dictate a longer poem (over 24 days) to a team of scribes, headed by the redoubtable Onomacritus. Inspired by this novel recording situation to expand on his winning selection (the Wrath of Achilles), the rhapsode pads out the original performance with re-doubled episodes, more duels, dialogue and battles, the Catalogue of Ships, retardation techniques, and some appropriate Athenian allusions – in short, all of the episodes we now call the Iliad (except, probably, Book 10). Anacreon, Simonides and Lasus of Hermione may have dropped in as the expert scribes elicited the ‘text’ from their star performer. Hipparchus, patron for the whole experiment, perhaps sat in for, say, Book 2, since that episode seems to feature a flattering statement about kingship. For some reason (J. is less sure) it is thought right to record an Odyssey at the same time, maybe from a competitor or house bard of the regime. If the shorter poem shows awareness of our Iliad, it is because the Odyssey composer most likely attended the earlier sessions. Finally, the crack scribal team was commemorated by a dedication of three statuettes that still grace the Acropolis Museum.

As with any historical fiction, one can no more deny that it all happened just so than one can prove that it did, especially since no facts stand in the way of a good story. As with all such speculations, two questions intrude: how are we to imagine pre-dictation Homeric poetry? And what difference does dictation make for the scope and quality of what eventually reached us by way of Alexandrian scholars?

On both scores, J.'s answers, or at least what can be understood of them from sometimes contradictory assertions, fail to convince. Before those August recordings, the Iliad and Odyssey simply did not exist; what was called ‘Homer’ covered a multifarious mass of heroic episodes that later trickled down as the Epic Cycle and other verse attributed to the master. The analogies from living traditions allowing us to imagine this situation are correct, but as old as Parry. Left unexplained are how and why this lesser material turned into (dictated?) texts and survived long enough for Aristotle to know reams of it. J. assumes that the Acropolis scribes were just better at eliciting uniquely masterful epics. She must tendentiously read information about the Panathenaic ‘rule’ (regulating sequential competitive recitation) as aimed at producing a vaguely chronological string of generic episodes, not a complete or even partial performance of Iliad or Odyssey. ‘Interpolations’ are also imagined as synchronic, last-minute personal adjustments by a rhapsode plunged into his new writerly environment.

Most difficult to credit is that a Cynaethus trained in shorter pre-Homeric performances would single-handedly expand one such episode to Iliadic proportions. Apart from the (unanswered) objection that the Wrath implies at least the skeleton of a proto-Iliad, the very evidence that J. trumpets does not support this scenario. Her single most important resource, the crowning work of the late Finnish folklorist Lauri Honko, reports the video recording of the Tulu-language Siri Epic from south-west India, which occupied six days in December 1990 and resulted in a text as long as the Iliad. (The same performer's painstakingly dictated version five years earlier was half that.) But Honko explicitly requested the fullest possible story (not an inflated episode) and the singer Gopala Naika obliged by telling a serial tale about three generations of Karnataka heroines – in other words, an entire Cycle.

The fate of Hipparchus' newly-minted Iliad and Odyssey? To lay unread for generations in a dusty archive on the Acropolis. J. insists that they had no further interaction with Greek oral traditions, and were not even read (except, perhaps, by the scholarly Thucydides) until introduced after 450 b.c. into the Athenian school curriculum. The Panathenaic competitors went on reciting ‘Homer’ but not the epics we know. This odd suggestion apparently is meant to insulate the pristine dictated text from any taint of multiformity: J. by turns mis-states and fails to engage with the detail of the main alternative approach, Nagy's long-term evolutionary crystallisation of the Homeric text, in which this sine qua non of oral tradition plays a role. Further puzzling arguments here include her insistence that the Acropolis scribes were expert at inducing masterpieces, smart enough to iron out non-metrical lines, good at spelling words correctly in the new Ionic alphabet (J. rejects metakharakterismos) and making a ‘clean’ copy when transcribing their wax-tablet notes onto papyrus – yet still tolerated entire episodes that contained slip-ups and contradictions, such as Nestor's wall-building instructions in Iliad 7.

The excuse for re-visiting J.'s old theories is that the current volume collects in its first segment (Chapters 1–8) confirmatory material from fieldwork. We are helpfully given a reasoned catalogue of many sorts of facts now available in the meticulous analyses by others (J. has not done autopsy) devoted to recent performances of epic from Egypt to Uzbekistan. In addition to prizing Honko's scholarship, J. rightly exploits richly documented studies on the Sirat Bani Hilal (Reynolds and Slyomovics); the Karakalpak Edige (Reichl); the Mwindo of the Congo (Biebuyck); the Talminadu Brothers' Epic (Beck); and the Telugu Palnadu (Roghair). But as anyone who has read these sources will realise, in reducing such careful accounts to the details that she thinks support her theses J. ignores their far more complex and suggestive thick-descriptions of audiences, patrons and singers; ritual and community; genre and gender; contestation, variation and living tradition. These were the phenomena that fascinated the scholars involved, and contemplating these will most aid future investigators of Homeric art.