Hittite Historical Phonology. Institut Für Sprachwissenschaft, Innsbruck, 1999

209 BOEKBESPREKINGEN — HETTITOLOGIE 210

HETTITOLOGIE

KIMBALL, S.E. — Hittite Historical Phonology. Institut für Sprachwissenschaft, Innsbruck, 1999. (24 cm, 456). ISBN 3-85124-674-8. ATS 1.800,-. This much-heralded book (hereinafter referred to as HHP) lives up to expectations as regards content. Kimball independently undertook a project similar to Melchert’s Anatolian Historical Phonology (= AHP), limiting herself however to Hittite. Though as it is impossible to do any- thing with Hittite’s ancestor without considering the other Anatolian languages, K. cites them fairly often to illustrate Hittite forms. According to AHP’s introduction, the two authors were careful not to read each other’s manuscripts, although they of course consulted each other’s previous work. AHP came out in 1994, while this book carries the date 1999. HHP has however been rather long in the edit- ing (with commendable results), so that in reality the text stems from more or less the same time as AHP, which it cites on occasion. K. states (p. 35) that her work is aimed not only at students of Anatolian and IE phonology, but also at scholars of other branches of Hittitology and Ancient Near East studies. It therefore has to be readable and clear to non-specialists. It seems successful in this regard, in the presentation of indi- vidual entries, though a basic familiarity with Indo-European phonology and phonetic/phonological terminology is necessary to use it. K. starts with an extensive discussion of the Hittite writing system, and the problems in interpreting it, since it is hardly a clean reflection of how Hittite was actually pronounced. She establishes (p. 106), among other things, that the signs with S are to be pronounced with a dental/alveo- lar sibilant: in other words, they contained plain S, not the sound English writes as SH. It is devoutly to be hoped that this will have an effect on the field, since for some reason many scholars don’t accept this, and would still hail various Hittite kings as Shuppiluliumash, Murshilish, and Hat- tushilish, and the Hittite capital as Hattushash. There is nat- urally also extensive discussion of the representation of voiced and voiceless as against possibly geminate stops; and of the relatively recently established fact that plene writings indicate vowel length and (less clearly) accent/tone, much of the original work having been done by K. herself. 211 BIBLIOTHECA ORIENTALIS LVIII N° 1-2, januari-april 2001 212

K. then lists the reflexes of PIE phonemes, each in vari- just happens on — there is no other way to find them. Under ous combinations and environments, which are explicitly *nh2 (gemination section, p. 415 or p. 411), one finds no listed in the Table of Contents, and thus are easy to find for cross-ref. to the treatment of sanh- (I finally found it, under consultation. Under each entry she provides Hittite examples n, p. 334). walh- was even worse: after not finding it on p. for each combination, with cognates or PIE/PA reconstruc- 413 under *-lh2 (gemination again), I tried under l, having tions, and generally other scholars’ discussion, doubts and learned from sanh-. There (p. 352) it was barely cited, with suggestions, as appropriate. This discussion, as she explains, a good cross-ref. to the lowering of the vowel (p. 165), and revolves around the most recent work on phonology; older a derivation from *welh3-. *-lh3 doesn’t exist in the gemi- treatments are generally cited only when they have to be nation section, so I tried *h3, postconsonantal: walh- was modified or when they have been confirmed by further there (no cross-ref.s). Someone who doesn’t know how to research. look would probably get lost or give up. It is a pity, because K.’s treatment is almost universally painstaking and care- this sort of thing is tiring, and yet everything seems to be ful, and her arguments and writing are clear and easy to fol- there, somewhere. low.1) She organizes a vast and complex set of data and prob- At first blush, one might in fact get the impression that lems into a coherent and understandable system, in a masterly there are gaps in K.’s coverage — this turns out generally fashion. not to be the case.2) E.g., there is no section on final stops, The main problem with HHP is its arrangement, which is but then one discovers some treatment under the headings suited more to a linear presentation of a new theory, than to Loss and Devoicing (p. 300ff.). The unresolved argument a reference work that is likely to be used by different peo- about whether accented PIE short vowels lengthened only ple for different purposes. Too often one needs to read in open or also in closed syllables is not even mentioned around in the book to gain the proper perspective. This dif- in the evidence sections on those vowels (p. 129ff.), where ficulty is seriously aggravated by the omission of an index one might be most likely to look for the discussion. There, of Hittite forms treated, which would seem essential to a K. begins with her usual one-line synopsis of the relevant work of this kind. From any point of view, but certainly from sound-change rule, but this in fact represents her own view: that of the varied readership which K. wishes to attract, the one has to look back to p. 125ff. to learn that it is disputed lack of an index diminishes the usefulness, and certainly the (there the presentation of the argument is particularly user-friendliness, of the book. HHP is arranged essentially thorough). The “basic” changes illustrated in this chapter from the vantage point of PIE. So if one doesn’t know what are not all you need to know, by the way: it may come etymology K. favors, it is impossible to find a Hittite word, as a shock to discover that the analysis here has to be coor- and one cannot trace the genesis of a given Hittite sound. dinated with the next chapter (Ch. 3), which details all Since her ideas are not necessarily the same as those of other kinds of special but important changes regarding the scholars (which is what makes the book interesting), she may vowels. well have put a form where one does not expect it. Even if It might also have been nice to have a section on the rel- one does guess right, she may not have put it in the relevant ative chronology of various sound changes: the information section. And once in a while she is not consistent, and puts is dotted around in the book, but it would have been helpful one derivation in one place and another in a second. Some to gather it together somewhere. The book does not attempt examples below. Complications of this ilk could have been a total picture of Proto-Anatolian as distinct from PIE, though avoided with an index, even if it gave just the negative infor- some sections deal with this directly. mation that certain forms are not treated. The lack makes The field has clearly paused to take stock of where it is HHP to a great extent of use to a passive rather than an and where it has to go, with this book and AHP, which com- active reader. plement (and often contradict) each other. Great progress has Cross-references, which might partially alleviate this been made in our understanding of Hittite phonology in the problem, are erratic: sometimes they are good, sometimes last decades, but things are still quite unsettled: these com- missing. Examples: on p. 306, under n (alternating with d/l), pendia are valuable, probably even necessary guideposts for K. gives two possible derivations for le: *leh1 (she seems further study. to favor this), and the more standard *ne(i). The dissimila- Not so very long ago, most of the Hittite vocabulary was tion *n-> l- is then treated in another section altogether not thought to be IE at all. IE etymologies have slowly been (p. 336), without mention of a possible similar change with worked out for the bulk of the lexicon, but many forms are le, and without cross-reference. In the section on final laryn- still impenetrable, and the rest are mostly debated, some in geals (discussed below), the derivation from *leh1 alone is essence, some in details. Obviously cognates are not immedi- given, again without cross-ref. Experts may not have prob- ately recognizable. Data are generally not copious, with the lems, but an inexperienced reader will get a wrong impres- result that theories are often assumed on the basis of a hand- sion, if (s)he lands in a section where only one possible ful of sometimes isolated examples — that is, if one is lucky: derivation is mentioned. This sort of complication happens sometimes we have one lonely specimen. Understandably, dis- frequently in the book. Just to stick with examples from the agreements are systemic. It therefore is worth recalling that discussion at the end of this review: cf. at least four dis- even when there is general agreement, reconstructions are cussions of the -u of assu (p. 56, 123, 125, 426), which one often only plausible, not proven. The book tries to bring this out, but quite often sounds as if it is stating facts. E.g., the Dir.

1) Sometimes, by the way, there are strange parsings, as when she takes da-ah-hu-us-ma-as-ta of StBoT 8 IV 39 as the verb dahhun + -sma (direc- 2) She misses things like I-J. Adiego’s analysis of voicing of stops (Acts tive of the possessive pronoun!) + -asta (p. 333). This is most startling: she of the Colloquium of the Indogermanische Gesellschaft on Anatolian and got this from StBoT 26?! Such peculiarities seem however unusual in the Indo-European, Pavia, Sept. 1999), considered an advance on previous treat- book. ments: it came out when her book was already in press. 213 BOEKBESPREKINGEN — HETTITOLOGIE 214 ending is taken from *-o (p. 135, 138), which I guess is fairly prising confidence from *-eh2 (the (putative) ancestor of safe for PA, but since we don’t really know the genesis of the the neut. pl. ending). One might invoke the rule that unac- case, it seems odd to list even this as certain, without a reli- cented long vowels shorten. But assu, with plene -u, is able PIE construct. Any derivation from “perfects” or derived from *-uh2. The initial a- is probably the accented “aorists” is very much begging the question. Cf. a particularly syllable here. On p. 123, K. tries to reconcile all this (no volatile sector, the laryngeals, for some specific examples: cross-ref.! See also AHP p. 86). The whole discussion It is accepted that with *VRHV, the resonants geminate, seems created by one form, assu, the only remotely rea- 6 with loss of the laryngeal. Some forms however are mysti- sonable Hittite case of a long final vowel < *V+h2 ): uellu- fying. K. has one geminated example of *-nh2, sanna- (AHP as far as I know may be uelluwa, and even if it isn’t, it has p. 79 has a couple more). On this basis, she moves sanh- no plene. The ending of mekki (Laws I §47b) is taken from out of the “group” of full nasals, because the laryngeal is *-ih2; this however is a hapax, and the context makes it preserved (p. 334). She also favors taking walh- from *-lh3- rather doubtful that it is a plural. If not, it may not have for a similar reason: *-lh2- geminates. AHP proposes rather had *-h2 at all. that sanh- /walh- preserve *RH 3) before a stop and that their The argument seems to be that all the above forms could- intervocalic/final h is analogical. I could not find reference n’t have any other derivation. This is hardly probative. In to this interpretation in K.: hers sounded like the only one reality, no form above has an unambiguous origin (for possible. It raises doubts: since *-rh3- and *-nh3 geminate, sakuwa see below). it might be hard to see why *-lh3 wouldn’t (though it might Here and on p. 151, K. says there is no evidence for the not); and medial *h3 is not usually preserved. However, her quantity of the ending of the two “duals” cited above. On analysis is not the issue here, but rather the mechanics of p. 55ff. she observes in fact that plene writing is erratic the analysis. An actually pretty clear rule like this *VRHV (specifically mentioning assu), which with Hittite doesn’t > VRRV is arrived at as follows: If one can gather a body surprise one. She would probably like a number of words of say 8 or more derivations that point towards the rule, it to contain a long vowel, by etymology, but they have no is considered a good basis. Then one has to deal with strag- plene (e.g. ganut among other monophthongized diph- glers that don’t fit. Everyone has his own interpretation, thongs, p. 220ff.). Some forms give not even that much rea- even his own data sometimes; and not just the stragglers, son to suppose that they had a long vowel, except the expec- but some of the core evidence may be taken differently. It tations of the interpretation: huesu, dassu and idalu, e.g., is disturbing that questions are often debated on the basis of being neuters, might have been uncharacterized plurals. elzi one derivation, proposed by someone once in a paper (the and danhasti are also essentially reconstructed according to one K. chooses for walh- is a root aorist). One is essentially expectation. dealing in probabilities and reasonable doubts, not facts, To finish with the final laryngeals: in the hapax mek, the with all the attendant problems this implies. The laryngeals loss of final *-h2 is invoked to explain the long root vowel may be slightly more prone to complications than e.g. cer- (see good cross-ref. for interesting analysis). On this basis, tain aspects of the vowels, but in essence this atmosphere -wasta is rejected as a case of vocalized *-h2. These are the still pervades Hittite phonology. At this stage, a reader only forms cited with final *stop(s)+-h2, and as they are not requires all the differing analyses, and all the apparent quite comparable (one has *V+stop+h2, the other *V+sibi- exceptions, in one place and/or in a readily retrievable for- lant+ stop+h2), it’s not clear that one need be a guideline for mat. No one book does (or probably can) provide this. Nev- the other. ertheless, two books that complement each other could serve Finally, K. takes the final -e of pronom. pl.s from as well, if the information in them were fairly easy to access. *-eih2/h1: she means the neuter, as p. 214 has the usual It is in AHP, but the difficulties in using HHP make it hard derivation of the -e of animate ke, etc., from *-oi. How- to make the necessary comparisons among proposed analy- ever, both the animate and inanimate endings used to ses. be derived from *-oi. (AHP p. 201 takes Palaic neut. pl. Consider another difficult section, on final laryngeals (p. -e from *-oi). Can one derivation really be proved as 425ff.). against the other? Please note then that on p. 211, K. ana- First K. states: “Laryngeals, including *h2, were lost in lyzes the neuter -e differently: either < *-eih2 or < *- oih1. absolute final position” (no problem there). No cross-ref. GIS She lists two examples of the dual ending *-h1: elzi and There are obvious problems with this whole section. UZUdanhasti. But neither is convincing.4) Carruba mentioned5) How can one establish what happened to the laryngeals, that an elzi might have been a stadera, without geminate when often enough one is hardly sure they were even scales or plates. The “double” status of danhasti is expressed there? There are few controls, such as possible cognates, in the component dan: need it be a formal dual? At any rate, however uncertain, to fall back on here. In particular, *-h2 the data for duals in Hittite are obviously meagre and dubi- is assumed where one wonders how likely it was to be ous. appropriate. The ending -a occurs in geminate body parts Note that a *i lengthened from the loss of *-h1, if it ever in early Anatolian (van den Hout KZ 97, 1984). Watkins existed, would presumably have shortened, since neither mentioned (oral comm.) that, typologically, collectives are word has a plene at the end. Cf. the controversial sakuwa, not used for this sort of expression: Gebrüder, e.g., exists, also never with plene -a: K. derives its ending with sur- but not *Gehände or *Gefüsse, or equivalents. I’m cravenly sidestepping the whole quicksand of sakuwa! — though

3 ) He passes, on whether it is *h2 or *h3, with walh-. 4 6 ) On p. 192, she says “if” they contain *-h1. ) Palaic has anni wasu-ha, both of which could take their long final 5 ) In connection with M. Fritz’s paper on the dual (or lack of it) in Hit- vowel from *-h2; cf. however AHP p. 204 for doubts that the accent length- tite, given at the Hittite Congress in Würzburg last year. ened the -u of wasu. 215 BIBLIOTHECA ORIENTALIS LVIII N° 1-2, januari-april 2001 216 one wonders how likely it was to have *-h2 — but should ical distinction between “Personenklasse” and “Sachk- we not at least discuss the apparently unpalatable pos- lasse” for Old Hittite and even claims (p. 102) that its sibility that e.g. Luwian issara or pata contains a vocalized application to the dative extends to all periods of the lan- *-h1? guage. She honestly concedes numerous counterexamples K. gives all the relevant material here — one suspects she (citations 73, 88-90, 102, 257-259, etc.), but tries vainly to didn’t want to get involved in too much analysis — but one explain these away. Her claim on p. 101 that there are no would have been happier if she hadn’t made it sound pretty other counterexamples with da- ‘take’ is already contra- certain. Doubts may be particularly pressing in volatile sec- dicted by her own example 308 (p. 118). The use of the tors such as the above, but it is as well to keep a certain scep- dative-locative (sic!) to express place from which with ticism in mind when reading on any aspect of Hittite phonol- inanimate objects is already attested in Old Hittite and is ogy, at least at the moment. commonplace thereafter.1) Tjerkstra also seriously misuses Let all this serve as a word to the wise. To sum up, HHP the concept of “partitive apposition”. Hittite has genuine is a well-thought-out, painstaking and thorough handbook, examples of this construction, although present evidence though it is not as user-friendly as one would wish. It will no suggests that it is a post-OH innovation. True partitive doubt become a standard reference for the latest discoveries, apposition shows replacement of a dependent genitive disputes and difficulties in Hittite phonology. If K. could be expressing whole/possessor by a case form agreeing with persuaded to compile, and IBS to print, an index of the Hit- that of the part/thing possessed (e.g. annaz kartaz ‘from the tite forms treated, researchers would definitely derive more heart of (my) mother’). It is first of all false to follow benefit from the book. Starke in calling OH combinations such as É-ri andan ‘inside the house’ partitive apposition (in addition to the Old Saybrook, November 2000 Jacqueline BOLEY serious semantic and morphological problems, the reverse order andan É-ri falsifies this claim). It is even worse to deform the notion further with Tjerkstra by subsuming ** examples such as her 72: s dUTU …KAxU * takna =ma -i -az ‘from the mouth of the Sun-goddess of earth’. This is not partitive apposition at all, but simply the well-established TJERKSTRA, F. A. — Principles of the Relation between use of the dative-locative to express the person/thing Local Adverb, Verb and Sentence Particle in Hittite. affected by the action of the predicate (likewise examples (Cuneiform Monographs 15). Styx Publications, Gronin- 261, 302, and 312). Fortunately, neither of these errors seri- gen, 1999. (24 cm, XII, 204). ISBN 90 5693 028 1. ously affects the author’s basic thesis about local adverbs Nlg. 150,-. and particles. This work represents the published form of the author’s The same is not true for the most serious failing in her Amsterdam dissertation. It treats one of the most famous and analysis, her failure to distinguish an adverb katta(n)/katti- difficult problem complexes of Hittite morphosyntax. An ‘beside, next to, in(to) the presence of’, distinct from katta introduction presents the state of question and the method- ‘down’ and kattan ‘below’.2) In the first place, while she ology employed. Crucial to the analysis is the notion of translates most passages with the former adverb correctly “basic” and “derived predicates” based on valency or func- (25, 44-46, 124-125, 197-198, etc.), non-specialists will be tional grammar, and hence the status of local adverbs as quite perplexed by these translations, since they have been dependent or independent of the verb. The author also dis- told only about kattan ‘below’. In some cases, moreover, tinguishes carefully the relative chronology of Hittite texts the failure to recognize kattan ‘beside’ leads to genuine and manuscripts. Finally, she chooses to focus her study on confusion and mistranslations. The MH example 126 with the usage of local adverbs and particles with a highly kattan pai- as ‘go down to’ is an important innovation for restricted but particularly diagnostic set of core verbs. Chap- OH katta pai-, while the preceding 124-125 are merely ters on iya- ‘be on one’s way’, pai- ‘go’ and uwa- ‘come’, standard examples of kattan ‘in(to) the presence of’. Exam- and da- ‘take’ are followed by a conclusion that tries to sum- ple 182 with katti=ti arÌa can only be ‘(walk) past you’ not marize the results and point the way to further research, ‘beneath you’, and in 323 Kumarbi takes wisdom istanzani including a brief but useful comparison with the situation in kattan ‘into his mind’, not ‘down in’. Most seriously, the Ancient Greek. An index of passages cited completes the confusion of the two different kattan’s affects the overall book. Chief among the strengths of Tjerkstra’s work is the lucid- ity of the presentation. The theoretical apparatus is set forth 1) For this usage see my dissertation, Ablative and Instrumental in Hit- succinctly and clearly and applied relevantly and effec- tite (Harvard University, 1977) p. 188ff., p. 289, p. 309f. and p. 352. For tively—it is not mere “window-dressing”. The argumenta- Old Hittite see KBo 17.1 i 12-13: kardi=smi(y)=at=kan daÌÌu[n] tion is linear and progresses in a logical and consistent man- [(Ìars)]ani= smi(y)=at=kan daÌÌun ‘I have taken it from their heart(s). I have taken it from their head(s).’ This construction with inanimate refer- ner. The English prose flows well and is with very rare ents is a trivial analogy to the “dative of disadvantage” with animate ref- exceptions idiomatic (in note 65, p. 22, read ‘occasionally’ erents and confirms that the dative and locative have merged functionally for ‘incidentally’). There are some typographical errors, all as well as formally, contra Starke and Tjerkstra. For obvious pragmatic rea- self-correcting. The philological basis is sound, and the trans- sons, nouns with semantically animate referents are not often construed as locations, but there is absolutely no grammatical prohibition on such usage lations generally reliable, though inevitably one will not agree when the context calls for it. with every rendering in more than three hundred textual cita- 2) See Frank Starke, Die Funktionen der dimensionalen Kasus und tions. Adverbien im Althethitischen (Wiesbaden: 1977), p. 181ff., especially p. There are also some weaknesses. Tjerkstra unfortunately 185f. Whether the two synchronically distinct adverbs katta(n) are related prehistorically is a completely separate issue irrelevant to their proper analy- accepts uncritically Starke’s claim of an alleged grammat- sis in Hittite texts. 217 BOEKBESPREKINGEN — HETTITOLOGIE 218 analysis. On p. 137 Tjerkstra claims that the distinction standing of the complex interplay of adverb, verb, and par- between katta pai- and kattan pai- reflects the difference ticle in Hittite. between an animate and inanimate referent, but we are rather dealing with two quite separate adverbs ‘down’ ver- Chapel Hill, May, 2000 H. CRAIG MELCHERT sus ‘beside’. She also wrongly denies (p. 142) Starke’s cor- rect distinction of an OH katta that functions like kattan ‘beside’ from the more common katta ‘down’. KORTE AANKONDIGINGEN The self-imposed restrictions of the analysis are also double-edged in their effects. While they allow the author to clar- GINDIN, Leonid A. — Troja, Thrakien und die Völker ify a number of important points (see below), they also at Kleinasiens. Versuch einer historisch-philologischen times afford a too narrow perspective that leads to distortions. Untersuchung (Innsbrucker Beiträge zur Kulturwis- For example, Neo-Hittite (in examples like 42 and 94- andan senschaft, Sonderheft 104). Innsbruck, Institut für 97) is a postposition with an expressed locative indicating Sprachwissenschaft, 1999. (24 cm, 324). ISBN 3-85124- place to which with a motion verb. Its usage here is an inno- 189-4. ATS 760. vation that has nothing to do with the function of andan in Old Hittite, but represents rather a renewal of OH anda in The author of this book died in 1994, when the German similar contexts.3) Such changes in the usage of the local translation of his Russian book was in progress (W. Meid adverbs from Old to Neo-Hittite independent of combinations in his “Nachwort”, p. 322). This is the “philological” con- with particular verbs must be taken into account in any over- tinuation of his “linguistic” approach in an earlier book all analysis. (Sofia 1981). The Homeric poems are the main source. Despite the problems just cited, Tjerkstra clearly demon- “Nach meinen letzten Resultaten kämpften die Griechen strates her basic thesis. She has shown beyond all doubt der Antike (hethitisch AÌÌijawa) gegen ein Troja, das vor- that the function of the local adverbs cannot be determined wiegend von Thrakern (genau genommen Troern und Dar- by facts of word order. Nor can one simply operate in danern) sowie von Luwiern (den Lykiern aus Zeleia) terms of their mere cooccurrence (or not) with particular bewohnt war” (p. 10). The chapters are: I. “Die Bedeutung verbs. As she contends, one must take into account whether der thrakologischen Studien für die Probleme der Indoger- the adverbs are dependent or independent; i.e., do they or manistik”, II. “Die Besonderheit der homerischen Epen als do they not form a unit with the verb (what she terms a historische Quelle”, III. “Die Rekonstruktion der derived predicate). For example, she is able to account thrakischen Sprache aus arealer Sicht”, IV. “Die Thraker neatly for the contrast of example 35, where appan iya- is in Troja im Lichte der linguistischen Analyse: Thrakisch- a unit ‘go after, pursue’, versus those of 51&52, where the anatolische lexikalisch-onomastische isoglossen” (p. 54- combination means simply ‘walk [behind]’ (here the 195), V. “Die Luwier in Troja im Lichte der linguistisch- adverb is independent). Likewise istarna pai- ‘pass’ (of philologischen Analyse” (p. 196-223), VI. “Kèteioi (= time) in examples 165-167 versus ‘go [among X]’ in exam- Hethiter), Paiones (Protoarmenier) — Verbündete der ple 200. See also her excellent summary of the differing Troer”, VII. “Die thrakisch-luwische Areale und ethnolin- usage with da- ‘take’ on p. 128. In addition to validating guistische Symbiose”, VIII. “Der trojanische Krieg und das her principle claim, Tjerkstra also offers (p. 141) a very AÌÌijawa der hethitischen Keilschrifttexte” (p. 241-275), important corrective to Starke’s analysis of the function of IX. “Vergleich der Resultate der linguistisch-philologis- the “locatival” adverbs in Old Hittite. As she makes clear, chen Analyse mit den Daten der Archäologie. Einige Erwä- these can be dependent on the verb and indicate place to gungen hinsichtlich des protoethnischen Areals der which (one can add to her list the verb tiya- ‘step, station Indogermanen in der Epoche vor dem Sprachzerfall”. oneself’). As Tjerkstra herself acknowledges, the recognition of ** dependent versus independent adverbs (or if one prefers, of * derived predicates) is a necessary, but not sufficient crite- rion for determining the usage of the sentential particles. MARKOE, G.E. — Phoenicians. (Peoples of the Past). She cites the two different senses of the unit a ‘go par iya- British Museum Press, London, 2000. (25 cm, 224). on’ and ‘go/come out’ (p. 32f. on examples 31-33). Con- ISBN 0-7141-2095-2. £ 24.99. trary to her claim, in example 196 anda pai- also clearly is a unitary ‘go in’. Its differing syntax versus that of exam- This is a book for the general public, richly illustrated. ples such as 89-91 is based on the fact that it means merely The expert will appreciate this updated survey of Phoenician ‘reduce the interval’, with no stated or implied goal, studies and the Notes appended to it (p. 211-218). The book whereas the others mean ‘go into (some place)’. Such exam- does not begin in 1000 but in 1550 B.C. which is “a depar- ples could be multiplied. However, we cannot hope to make ture from earlier studies” (p. 11). The first chapter, “His- progress in sorting out such fine distinctions unless we first tory”, shows that wood was exported from the earliest time establish the more fundamental contrast of dependent and on (to Egypt), it discusses urban expansion in Galilee (p. 30) independent adverbs as advocated by Tjerkstra. In showing and the role of Sidon (31, 62). Tyre became important (33), both the importance and feasibility of making this basic Kition on Cyprus is discussed (53, 170). The chapter “The demarcation she has significantly advanced our under- city” describes the archaeological discoveries, like the Punic harbour (cothon, 70) and excavations in Carthage or Motya. In the chapter “Economy” much attention is paid to coinage (p. 98-101, 105-107). “Language and literature”; the spread 3) See Donna Salisbury, ‘anda and andan in Neo-Hittite’, to appear in Journal of Cuneiform Studies 51 (1999). of the alphabet. Among the topics in “religion” are gods like 219 BIBLIOTHECA ORIENTALIS LVIII N° 1-2, januari-april 2001 220

Melqart and Tanit, aniconism in art (125, 131), sanctuaries (open-air and roofed, 126-8), child sacrifice (133), care of the dead (137 ff.). The last two chapters are “Material cul- ture” (including art in ivory and metalwork, glass) and “Commercial expansion abroad” (Sicily, etc.). The Appen- dix gives a “Survey of cities in the Phoenician homeland” (p. 192-206).