Bryn Mawr Classical Review

Bryn Mawr Classical Review 2009.06.34

Andrew L. Sihler, New Comparative Grammar of Greek and Latin (Paperback reprint of 1995 edition).   Oxford/New York:  Oxford University Press, 2008.  Pp. xxii, 686.  ISBN 9780195373363.  $45.00 (pb).  

Reviewed by Zsolt Simon, Hungarian Academy of Sciences (
Word count: 1600 words

The hardback edition of the book being reviewed here appeared in 1995 and filled an existing gap at that time. Not only were comparative grammars of Greek and Latin written in English lacking since the publication of Buck's handbook,1 but there was also no up-to-date Latin comparative grammar. It was not only a technical problem that there were no textbooks for students, but also a scholarly one, since Buck's handbook included neither the laryngeals nor Mycenaean (the latter unknown at the time), and therefore his book was highly outdated. On the other hand, after such seminal works as Rix' book on Greek comparative grammar based on laryngeal theory or Schrijver's treatment of the Latin reflexes of the laryngeals,2 the time was ripe for writing a new comparative grammar of these languages.

Since Sihler's goal was a revision of Buck's handbook intended both as a reference work and as a handbook for students, its layout generally follows the structure of Buck's work except for the omission of derivational morphology. In other words, it has the well-known structure of comparative grammars (Introduction (pp. 1-34),3 Phonology (pp. 35-242), Declension (pp. 243-368), Pronouns (pp. 369-401), Numerals (pp. 402-441), Conjugation (pp. 443-629) and Indexes of cited forms (pp. 631-686)), where, however, Latin and Greek data are presented side by side within each chapter rather than separately from each other.

Sihler's attempt to rewrite Buck's handbook in light of contemporary knowledge has the welcome side effect of discussing the Proto-Indo-European (PIE) data and that of its descendants other than Greek and Latin (especially Vedic, Avestan, Gothic, Lithuanian, Old Church Slavic and Hittite) in more detail than earlier treatments, thereby providing insight for classicists into the evidence on which the comparative grammar of these languages has been built. Another welcome feature is that in most cases Sihler discusses in detail the meticulous investigations that lead from the data to particular theories and assumptions, illustrating very well the methodology of Indo-European linguistics, and he does this in an entertaining style that is unfortunately very foreign to such a technical topic,.

Sihler's new comparative grammar aroused much scholarly appreciation. All of the detailed reviews underlined its usefulness, though some of them called attention to the fact that the book is sometimes idiosyncratic (cf. below).4 A perplexing characteristic of the book was that it lacked any kind of reference or bibliography except for frequent formulae like "some authorities maintain".5 Though scholars of the field can recognize the underlying publications in most cases, students surely would not (and cannot be expected to); and this approach becomes even stranger in the case of such dubious statements as "Italic scripts might rather have the same source as the Greek" (p. 18, cf. p. 20 n. 1.). This sad fact, however, evidently leads to the conclusion that this book, despite its undoubted virtues, can serve neither as a handbook for students nor as a "vade-mecum" (p. viii) for Indo-European studies.

The book under review is the paperback edition of this volume, thirteen (!) years later, without any change in the text (including misprints).6 Unfortunately, this applies to the missing references too, though a paperback edition would have offered a good chance to add at least selected references or list of works "cited". Paperback editions are always welcome since they make the results of science more available to a wider audience including, especially, students, but the interval between the two editions is not immaterial: the more time passes between the two editions, the greater are the chances of producing an out-dated volume. In general, Latin and Greek are not languages the comparative grammars of which are in need of substantial revision; nevertheless, small but important changes from the research of the past thirteen years affect all parts of their comparative grammars. Let me cite some examples from the history of Latin (aside from etymological research):

First, the discovery of the Garigliano bowl now provides clear inscriptional evidence for dat. pl. of o-stems, MEOIS SOKIOIS (cf. p. 263) and for ESOM, the pre-classical form of sum widely held until then as a fabrication of Varro. This evidence not only serves to remove Sihler's asterisk before esom, but also renews the discussion of its history (cf. pp. 549-550).7

Second, recent investigations have shown that Proto-Italic maintained the Proto-Indo-European position of the accent.8 This observation should have a place in a comparative grammar (p. 239) and invalidates some of Sihler's suggestions (e. g. that the initial-syllable accent of Italic as a common innovation with Proto-Germanic and Proto-Celtic, p. 60).

Third, Sihler refers to some "recent investigations" suggesting that the Praenestine Fibula is a modern forgery (p. 15): thus he was forced to analyse the sign [F] in a peculiar way (p. 21) and to delete e.g. the forms NVMASIOI and FHEFHAKED as inscriptional evidence for the Archaic Latin dat. sg. of o-stems and final -d for secondary *-t among verbal endings, respectively (p. 58-59, 461). Meanwhile, more recent investigations have shown that neither the fibula nor the inscription are forgeries,9 and thus Sihler's analyses must be revised.

Finally, as with other comparative grammars, Sihler's work does not distinguish between sound laws that operated before and during the written phase of Latin, though our knowledge now allows us to present the sound laws according to various stages of the history of Latin, at least tentatively.10

These developments have widened the already existing divide between contemporary knowledge and Sihler's work.[[11] The reason for the original gap was the partial idiosyncrasy already called attention to by reviewers of the hardback edition. Such idiosyncratic points include, for example, the notion that satem languages constitute a real subbranch of Indo-European (p. 7), that all IE languages require only two tectal series of PIE consonants (p. 136, cf. 152-154),12 that *h2o becomes a in Greek and Latin (pp. 45-46) or that a Greek prothetic vowel does not automatically point to a PIE laryngeal, especially before resonants (p. 85-88)--all of these being assumptions that distort many of Sihler's PIE reconstructions. (For further, more minor, quibbles, see below.)

Despite these problems the usefulness of Sihler's book is beyond doubt. Nonetheless it would have been better to present a revised second edition rather than a paperback reprint; for one thing, until such a revision appears, the book can be used only with the help of an instructor who can explain the contemporary knowledge hidden behind the idiosyncrasies and numerous minor problems that would go unnoticed by students. Moreover, the book does not solve the problem of the lack of an up-to-date comparative grammar of the classical languages in English, though the case of Latin will hopefully be solved in the near future.13

Some minor quibbles not quoted by earlier reviewers:

p. 1-7. Phrygian is missing from the list of IE languages (as well as some other fragmentarily attested languages, like Messapic).

p. 2. n. 1. There is so far no linguistically established connection between the Hittites' name and the name of their capital, Hattusa. The Hittites did not call themselves "Nesis" but rather nesumenes 'men from the city of Nesa', or, in Anglicised form, "Nesites".

p. 4. The records of Albanian from the 16th and 17th c. are not meagre; this is true rather of the earlier records.14

p. 8-9. Sihler's PIE prehistory--with three Kurgan incursions coinciding with the collapse of "Old European" culture--would cause some raising of eyebrows even among followers of the Kurgan-hypothesis.15

p. 16. Among the ancient languages of Italy not only Latin and Oscan were still spoken by 100 B.C. but also Etruscan.

p. 16. That Latin would have been the official language of administration in the Eastern Empire is a serious oversimplification, and actually false for certain periods (cf. so-called bilateral unilingualism).16

p. 30. It is not only "hardly conceivable" that the Hittite spelling of the 1pl. ending -um-me-ni represents a long consonant (p. 30), in fact it does not represent a long consonant: the consistent Old and Middle Script spellings in Cu-me-(e)-ni show that this is not really long and reflects only a change in the phonetic value of the (C)Vm signs caused by a contemporary change in Akkadian pronunciation.17

p. 57. The statement "H drops in PIE after s" is contradicted by Hittite data with preserved -sH- clusters, e. g. isha- 'master, lord', ishai- 'to bind', ishamai- 'to sing', ishuwai- 'to throw, to pour'.18

p. 178. The loss of *w and the labial component of *kw before o could not have taken place "in prehistoric times" since they are attested in Archaic and Old Latin too, see e.g. DEIVOS on the Duenos Inscription, SVODALES on the Lapis Satricanus, and QVOIVS on the epitaph of Scipio Barbatus cited by Sihler himself (p. 387, cf. p. 388, 398-399). For the date of the change cf. the hypercorrect OQVOLTOD instead of *OC(C)OLTO in the Senatus Consultum de Bacchanalibus, 15.19

p. 299. The Hittite word for 'water' was never spelled ú-wa-a-tar, but only wa-a-tar. The former means 'inspection'.20

p. 343. To call nouns of variable declension "heteroclites" is an unfortunate use of the term, which refers to a regular type of declension.

The alleged forms of Hittite accented personal pronouns with -a extensions (ú-ug-ga, p. 370, tu-ga, p. 371) are in fact the pronoun + the clitics 'but' or 'and'.21

p. 405. n. 1. Farsi yak 'one' (recte yek) is not a borrowing from Turkic (cf. Turkish bir 'one'!), but a reflex of Old Iranian *aywa-ka- (cf. Middle Persian êwak).22

p. 430. Though Hittite duyanalli- is a hapax, its meaning is clearly not 'man of the 4th rank' but 'second in rank' on the basis of the context (IBoT 1.36 I 38 & 39).23


1.   C. D. Buck: Comparative Grammar of Greek and Latin. Chicago, 1933.
2.   H. Rix: Historische Grammatik des Griechischen. Laut- und Formenlehre. Darmstadt, 1976; P. Schrijver: The Reflexes of the Proto-Indo-European Laryngeals in Latin. Amsterdam; Atlanta, 1991.
3.   Including an overview of the Indo-European family of languages, a general history of Greek and Latin, their writing systems and some technical topics.
4.   W. F. Wyatt: Minos 29-30 (1994-1995) 389ff. (esp. on Greek data); E. C. Polomé: JIES 23 (1995) 463ff; J. Clackson: CR 46 (1996) 297ff. (esp. on minor quibbles); M. Weiss: AJPh 117 (1996) 670ff. (esp. on laryngeals); G. A. Sheets: CJ 93 (1997) 88ff. (esp. on morphology).
5.   Instead, there is a long list of scholars who influenced the book (and who, for the most part, were still alive at the time of writing), and a special emphasis (p. viii-ix) of the role of Warren Cowgill and Oswald Szemerényi whose ideas can often be detected throughout the book.
6.   E. g. "predominates in in mss." (p. 28), Enumakartidas (p. 87), largyngeals (p. 106), larynageals (p. 136), "9ayin" (p. 166), read 12 instead of 11 on p. 306 in the notes to the table; seeminlgy (p. 412). In general, however, the book is well proofread, despite the complicated text.
7.   For the latest treatment with detailed references see Zs. Simon: Zur Geschichte des lat. sum. In: G. Viré (ed.): Autour du lexique latin. Communications faites lors du XIIIe Colloque international de linguistique latine, Bruxelles, 4 au 9 avril 2005. Collection Latomus 316. Bruxelles, 2008, 267ff.
8.   B. Vine: On "Thurneysen-Havet's Law" in Latin and Italic. HS 119 (2006) 211ff. This paper refutes the view of Sihler that *ov sporadically becomes av (p. 44).
9.   E. Formigli: Indagini archeometriche sull'autenticità della Fibula Praenestina. MDAI(R) 99 (1992) 329ff; R. Wachter: Altlateinische Inschriften. Sprachliche und epigraphische Untersuchungen zu den Dokumenten bis 150 v. Chr. Bern, 1987, 55ff., resp., cf. M. Hartmann: Die frühlateinischen Inschriften und ihre Datierung. Eine linguistisch-archäologisch-paläographische Untersuchung. Bremen, 2005, 67ff.
10.   Cf. similarly the remark of Wyatt (n. 4) 390. See the recent attempt of M. de Vaan: Etymological Dictionary of Latin and the Other Italic Languages. Leiden; Boston, 2008, 5ff.
11.   Not to mention--to turn to Latin again--the abundance of recent handbooks of the history of Latin (e. g. G. Meiser: Historische Laut- und Formenlehre der latenischen Sprache. Darmstadt, 1998 or Hartmann (n. 9.)) whose views must be treated in some way in a handbook like this one.
12.   Dismissing the Anatolian (more precisely, Luwic) evidence for three tectal series by saying that "the crucial evidence in (...) Anatolian seems to hinge upon especially difficult or vague or otherwise dubious etymologies" (p. 154) is simply false, see the classic detailed argumentation of H. C. Melchert: PIE Velars in Luvian. In: C. Watkins (ed.): Studies in Memory of Warren Cowgill. Berlin; New York, 1987, 182ff; New Luvo-Lycian Isoglosses. HS 102 (1989) 23ff., a position that can now be supported even by Carian data (I. J. Adiego: The Carian Language. HdO 86. Leiden; Boston, 2007, 259).
13.   I refer to M. Weiss' forthcoming book.
14.   Cf. now the overview of J. Matzinger: Der Altalbanische Text Mbsuame e Krështerë (Dottrina cristiana) des Lekë Matrënga von 1592. Eine Einführung in die albanische Sprachwissenschaft. Dettelbach, 2006, 34ff.
15.   For an excellent critical overview written by one of the followers of this theory see J. P. Mallory: The homelands of the Indo-Europeans. In: R. Blench and M. Spriggs (eds.): Archaeology and Language I. Theoretical and Methodological Orientations. London--New York, 1997, 93ff.
16.   J. Kaimio: The Romans and the Greek Language. Helsinki, 1979, 319f. Cf. the illustrative graphic representations in B. Adamik: Zur Geschichte des offiziellen Gebrauchs der lateinischen Sprache. Justinians Reform. AAntHung 43 (2003) 230, 239 and Offizielles Kommunikationssystem und Romanisierung. In: C. Arias Abellan (ed.): Latin vulgaire--latin tardif VII. Actes du VIIème Colloque international sur le latin vulgaire et tardif, Séville, 2-6 septembre 2003. Sevilla, 2006, 17ff.
17.   H. C. Melchert: Anatolian Historical Phonology. Amsterdam; Atlanta, 1994, 24 with earlier references.
18.   For etymological treatments of these items see now A. Kloekhorst: The Etymological Dictionary of the Hittite Inherited Lexicon. Leiden, 2008, ss. vv.
19.   Meiser (n. 11.) 92.
20.   See now Kloekhorst (n. 18) ss. vv.
21.   See already P. H. J. Houwink ten Cate: The Particle -a and its Usage with Respect to the Personal Pronouns. In: E. Neu and C. Rüster (eds.): Festschrift Heinrich Otten. Wiesbaden, 1973, 119ff.
22.   See already P. Horn: Grundriss der neupersischen Etymologie. Strassburg, 1893, 252.
23.   See already H. G. Güterbock: Lexicographical notes. RHA 15 (1957) 2f.

Comment on this review in the BMCR blog
Read Latest
Index for 2009
Change Greek Display
Books Available for Review

HTML generated at 11:27:02, Wednesday, 24 June 2009