Assyriology has always had an uneasy relationship with textual criticism. In his early work Kramer usually identified a single manuscript that could serve as a matrix for the alignment of numerous smaller tablet fragments. The aligned fragments then served as secondary witnesses when the key text, invariably labelled manuscript A, was damaged or erroneous. The key manuscript for Kramer's edition of Inanna's Descent (Revue d'Assyriologie 34, 1937, 93–134) was Ni 368 + CBS 9800 (manuscript A) and the variants that Kramer recorded in the footnotes always come from manuscripts other than A, except when manuscript A omits something. The same pattern appears in the first half of Gilgamesh and the Huluppu-Tree (Chicago, 1938) and Lamentation over the Destruction of Ur (Chicago, 1940). Kramer never distinguished between his method of textual reconstruction and textual criticism in the strict sense of the term, and never seems to have said anything in print concerning textual criticism (see, however, his typology of variants in Lamentation over the Destruction of Ur). Thorkild Jacobsen's work on the Sumerian King List, which was sandwiched in between Kramer's editions of Gilgamesh and the Huluppu-Tree and Lamentation over the Destruction of Ur in the series known as Assyriological Studies, did quite the opposite: Jacobsen grouped sets of manuscripts together as manuscript families on the basis of shared errors and arranged these families into a proper stemma. Kramer and Jacobsen, each in their own way, were reacting to the problems of their respective corpora: Kramer was faced with a veritable deluge of more-or-less unreliable student exercises that happened to witness to the great bulk of Sumerian literature, while Jacobsen was dealing with the numerous manuscripts of a single chronographic text that was, in all likelihood, never part of the school curriculum. But if Kramer's methods of textual criticism were so profoundly unsound, particularly in opposition to Jacobsen's textbook use of Lachmannian method, why did they work so well?

This is not exactly the question that Paul Delnero's The Textual Criticism of Sumerian Literature seeks to answer, but he arrives at a solution to Kramer's conundrum nonetheless. The Textual Criticism of Sumerian Literature represents the distillation of a 2006 University of Pennsylvania dissertation (Variation in Sumerian Literary Compositions: A Case Study Based on the Decad), supervised by Steve Tinney. The most important innovations in Delnero's thinking between dissertation and book are twofold: the entire project has been reconceptualized as a case-study in Sumerian textual criticism; and the role of memorization, which is also the theme of a recent paper by Delnero (“Memorization and the transmission of Sumerian literature”, Journal of Near Eastern Studies 71, 2012, 189–208), has taken centre-stage in Delnero's thinking on variation in Sumerian literature. Delnero's discussion of the principles of textual criticism is somewhat compromised by his limited exploration of classical methods of textual criticism: this is particularly clear on pp. 8–9, where he alludes to a concept that is central to his entire project, namely “contamination”, but does not offer a proper definition. The lack of a clear description of contamination (viz. horizontal transfer of a reading from one manuscript family to another, thus presupposing the existence of manuscript families, even if difficult to identify) is troubling because, as Timpanaro noted from the grave: “… there are cases (many, as Pasquali noted) in which contamination and interpolation have acted so extensively and so early as to make it impossible to trace out any stemma at all (those cases in which Pasquali, referring only to contamination, speaks of as ‘total pretraditional contamination’)” (Timpanaro, The Genesis of Lachmann's Method (Chicago, 2005), p. 212, citing Pasquali, Storia della tradizione (Florence, 1934/1952), pp. 146–55, 177–80). The Sumerian literature recovered from the Old Babylonian schools would certainly qualify as a didactic literature that has suffered from extensive “contamination and interpolation” and in some sense this fact is the central empirical finding of Delnero's work.

The irony is that, even as Delnero categorically denies the relevance of traditional methods of textual criticism (most explicitly on p. 203), he actually carries out a textbook example of the isolation of a manuscript family in his discussion of manuscripts N_I1 and N_III12 on pp. 189–94. Amid the deafening variational noise of dozens of schoolboy copies, Delnero's identification of a manuscript family is unadulterated philological gold. Yet would-be emulators should probably root their analyses in the methods of its best practitioners (see Reeve's Manuscripts and Methods I, Rome, 2011) rather than the critics (McGann and Cerquiglini) on whom Delnero relies; for some background on Cerquiglini, see Varvaro, “The ‘new philology’ from an Italian perspective” (Text 12, 1999, 49–58, reference courtesy of L. Raggetti). But why does Kramer's method, from which Delnero's approach must be seen as an indirect descendant, work? Unlike Kramer, who simply privileged the largest and best-preserved manuscript available, Delnero assumes a principle of majority rules: which ever reading is most frequently attested is adopted as the reading of the text (ignoring the possibility that all the Nippur manuscripts could easily derive from a single [sub]archtype), and all other readings (including the odd Ur III witness and any significant non-Nippur tradition) are then defined as variants. In his discussion of line 66 from the royal hymn known as Šulgi A (pp. 182–6), the sixteen manuscripts include ten distinct variants, but there are only two minor variants among a very homogeneous set of Old Babylonian manuscripts from Nippur. Given the principle of majority rules and no consideration of eliminatio codicum descriptorum, these eight homogeneous manuscripts will invariably win the day (see Worthington's discussion of similar issues in his Principles of Akkadian Textual Criticism, De Gruyter, 2012, pp. 292–7). Delnero's method is particularly clear in his analysis of the variation between -ib₂- (TUM) and -ib₁- (IB) on p. 182: the four manuscripts from the Old Babylonian schools in Nippur that preserve the relevant part of the text all have -ib₁-, while the four witnesses that have -ib₂- either do not stem from the city of Nippur or come from an earlier phase in its history (the Ur III manuscript N_P2). Two occurrences of -ib₁- in unprovenanced manuscripts (X_I2 and X_I3) yield a majority for -ib₁- and consequently Delnero adopts -ib₁- rather than -ib₂- in the synthetic text. Although the alternation between these two renderings of /ib/ is in itself trivial and admittedly messy (see pp. 159–62), it suggests that Delnero's goal is the reconstruction of what we might call a textus receptus for Old Babylonian Nippur (see the methodological statement on p. 180), and it is this objective that lends the methods of Kramer and now Delnero their heuristic force. If Delnero had adhered to the conventions of Classical textual criticism, the presence of -ib₂- in an older manuscript tradition (N_P2) and its attestation in lateral areas (Ur and the unidentified sites behind X₁ and X₃) would have favoured the reconstruction of -ib₂- in the archtype.

The most valuable part of the book, however – the part that makes it an essential purchase for anyone working on Sumerian literature – is its extensive discussion of variations that are characteristic of particular cities (pp. 61–84). Owing to the ever-present difficulty of contamination in didactic environments, the most promising way forward vis-à-vis the literary materials recovered from the Old Babylonian schools is a renewed emphasis on geographical variation. In the absence of well-defined families of manuscripts, the stark differences that arose in particular scribal centres can serve as a useful stand-in for the traditional postulates of the stemmatic method, as Pasquali and his inheritors have often pointed out (see the discussion of “the norm of lateral areas” in Timpanaro, Genesis, 86–7 and 138 and Reeve's “Some applications of Pasquali's ‘Criterio Geografico’” in Manuscripts and Methods, pp. 221–8). Overall the volume represents a valuable contribution to the study of Sumerian literature and the author is to be congratulated for making a concise account of his work available to a broader audience.