The pragmatics of style Edited by Leo Hickey (review)

BOOK NOTICES 403 scheme (75) should be revised per Hoffner, Kanissuwar (1986), and in 6.41 (85) there is no evidence that nu means 'now' in the sense implied (or any other). The statement that 'number concord between subject and verb is more according to meaning than to strict grammatical rules' (93) is incorrect, and (as noted on 94) is certainly not justified by the example cited: the grammatical rule that neuter plurals agree with singular verbs is not mentioned. The reading selections also include errors, e.g. the adoption of the reading HUL-lawahzi 'injures' on 11415 , lines 21-22, rather than the correct variant idalauwanni 'for harm' (which HS&G's translation follows); omitted are kuitki 'something' on 117, line 49, -as 'them' on 128, line 70, and -kan on 128, line 74. Also in the readings, the interlinear translation is often wrongly aligned with the Hittite; e.g., IR-s = a 'but a slave' (124, line 25) seems to be glossed 'that slave'. The glossary does not distinguish man 'if and the optative particle man. Omissions are inevitable from a grammatical sketch ofthis length. Still, some mention should have been made of the verbal suffix -annäi-, the 'ergative' in -anza, and the respectively abstract , collective, and factitive functions of the suffixes -atar (not '-tar' [24]), -essar (not '-sar' [24]), and -ahh-. The verbal noun in -war should be treated with the participles (it is mentioned only with nominal morphology [25]). Ex. 5.73d (73) presents an opportunity to discuss verb serialization , which is only alluded to in the glossary under pair (159) and uit (164-65). Some minor points are confusing. Although wa-al-ahmi ? attack' has (as HS&G state on 39) a typical monosyllabic root ending in two consonants, that fact is unclear from the form cited, and a student should be told that the root is walh- and why it is not *walah-. One might also suppose (wrongly) from the wording on 41 that the infix -nin- appears in other than velar-final roots. The distribution of the two infinitive endings could be clarified in one or two more sentences. The discussion of orthography is at times misleading . For example, after referring to a 'paucity of signs denoting e- plus a following consonant' (6), HS&G question the standard explanation of the initial V-C spelling of e-te-ir 'they ate': since 'there is no sign *et, the only other possibility' would be 'it' (9). In fact no such 'paucity' exists: while there are, for instance , separate EN and IN signs, many signs are polyvalent; in this case there is one sign with the two readings ET and IT, not an IT sign. The standard view correctly explains the nonexistence of a spelling *et-te-: orthographic -t-t- = IxI and -V-/- = IaI, and the root 'eat' has IdI. Elsewhere, orthographic variation is too willingly attributed to nonlinguistic factors; for example , 'vacillation' (8) explains the initial CVC vs. CV-VC spellings in ki-sa-at 'he became" (active) vs. ki-is-ha-ha-at ? became' (middle). In fact the latter is required by the consonantinitial ending -hhahat. Finally, even if HS&G do not accept the (common) Eichner/Kimball view of vowel length, more comment and upto -date references would be in order on pp. 89 ; the important phonological phenomenon of lenition should also be discussed. Because BH 'is a descriptive and not a diachronic grammar, no effort has been made to separately describe the three stages [sc. of Hittite ] in any detail' (11): paradigms mix linguistic features of different ages—e.g. 1st sg. près, -hé vs. -hi and demonstrative nom. pi. kè, apè vs. kits, apüs—and the sign list does not distinguish Old, Middle, and New Script variants. This procedure is inherited from all the standard texts, which predate most of the linguistic and paléographie dating techniques now known. In my view, however, these have become so essential to work in Hittite that a new text should introduce them at the beginning. The errors in this book warrant caution, and its omissions mean that it will not be useful long after a student finishes its reading selections. An up-to-date Hittite text is a desideratum, but...