By Stesichorus, 26.5, Helen of Troy is described as λιπεσ-άνωρ ‘husband-leaving’; by Aeschylus, Ag. 689, as ἕλ-ανδρος ‘man-destroying’, with etymological wordplay on her own name. Both of these words combine a verbal first constituent (FC) with a nominal second constituent (SC); English examples such as ‘cut-purse’ and ‘pick-pocket’ have a similar structure. The book under review here discusses the possibility of the Indo-European existence of such compounds based on comparative evidence, their development in the history of Greek, and their relationship with other kinds of Greek compound.
This is an all-together excellent book, and a much-needed addition to the scholarly literature on Greek word-formation.1 It is not a comprehensive handbook of Greek compounding (as Tribulato puts it: ‘for all its length, this book is only a “chapter” itself’, VII); yet the thoroughness of its coverage, its stimulating argument, and its rich data will make it a serviceable handbook, a reliable guide, and a call to arms for further research. A comprehensive study of Greek compounding is therefore still a desideratum, but Tribulato deserves our thanks for bringing this goal closer. The book will be an essential addition to any scholarly library for historically and synchronically oriented linguists interested in Greek, and a valuable research tool for commentators and literary scholars. It contains a wealth of philological detail, but it addresses a much wider audience than only philologists.
A short survey of the book foregrounds the reasons for studying verb-initial (or V1) compounds in Greek and Indo-European. They are typologically unusual, and contrast with the normal syntactic alignment of Greek; furthermore they exist alongside verb-second (V2) compounds, and the reasons for this coexistence needed to be explored. This section also introduces Tribulato’s dataset (4). This is derived from earlier published material rather than from a newly compiled set of texts. Hence it is not a ‘corpus’ study in the strict linguistic sense of the word; nevertheless, this is a wider dataset covering a longer chronological span (Mycenaean to Modern Greek) than used in the earlier studies of V1 compounds. The inclusion of personal names is also a boon.
The first chapter covers compounding from a modern theoretical perspective, the second Ancient Greek compounding in general against that background, assessing the different types of compound found in Greek. Different readers will need different parts of these chapters more. General linguists will find Chapter One mostly covers familiar problems in the classification of compounds (criteria for compounding, such as spelling, and accentuation; semantic and syntactic approaches to compound classification; headedness and exo- and endocentricity). The material in Chapter Two presents data familiar to most classicists, though within the terminology of Chapter One, which may not be familiar to all. It is in any case an accurate and detailed overview of the possibilities for compounding that exist in Ancient Greek; with some appropriate glossing of the technical vocabulary it would be an excellent thing for students to read to familiarise themselves with the workings of Greek compounds. Tribulato casts her net still wider: at 85-103 she surveys root nouns, adjectives in -ής, forms in -ος, agent nouns in -της, -τήρ, -τωρ, participles, forms in -τος, and some residual categories; though the focus is on compounds, in fact we read a full history of these suffix’ origins and synchronic role in Greek word-formation. This comprehensive approach extends throughout the range of Greek linguistics; the summary of the IE verbal system (186-190) is a case in point.
The third chapter looks at the history of scholarship on Greek compounds, particularly the V1 type. In fact, the whole book is attentive to the history of scholarship: it was interesting to find noted how few studies of the role of compounds in Greek poetic language there are (10). Some of the debates noted by Tribulato have contemporary relevance: the role of syntax in reconstruction; the possibility of morphological segmentation (e.g. whether or not a ‘stem’ is an accessible entity); the inventory of parts of speech in Indo- European (these themes are also explored in the rest of the book). Furthermore, Tribulato provides a sensitive account of the interactions between nineteenth-century linguists.
The rest of the book is devoted to the diachronic (Chapter 4) and the synchronic (Chapters 5 and 6) analysis of V1 compounds. Central to the diachronic analysis is the issue of what one is to compare to Greek V1 compounds in other branches of Indo-European, and what, therefore, one may reconstruct in the parent language. This has consequences for the ontology of the first member, which has been interpreted as a noun stem, an inflected verb (imperative or indicative), and a verbal stem. The analysis of the Mycenaean evidence is crucial. First Tribulato analyses elements that seem to have been productive within Greek, namely -o- (beginning from the thematic aorist) and -i- (with sensible rejection of Risch’s unmotivated reanalysis), both of which spread at the expense of inherited -e- (beginning from the thematic present). The types in -esi-, e:si-, -se- and -so- also require inner-Greek explanations; they arise partly from contamination between the φερέοικος and τερψίμβροτος types, and partly from the spread of sigmatic forms derived from verbal stems, whence they are grammaticalised and spread as markers of V1 morphology. Tribulato concludes that the V1 compounds in fact represent the convergence of two types: the compounds with a first member in -si- were derived from verbal stems formed to roots which had sigmatic forms, and reinforced by analogy on action nouns in -sis; and compounds with first members in -ti-, equally derived from verbal stems, but this time from roots which did not (yet) have a sigmatic aorist. The consequence of this is that the type in -si-, pace Schindler, is not consistently derived from the type in -ti- by sound change (the assibilation, or change of /t/ to /s/ before /i/), nor has Vedic lost a sigmatic V1 compound type; furthermore, Tribulato demonstrates that Mycenaean shows precisely the expected distribution between -ti- and -si- expected on this account. Given the importance and the productivity of the s-aorist in Greek, this means that the forms in -si- may be a Greek innovation, not to be pursued back into the parent language. Examples of compounds preserving -ti- (type βωτιάνειρα) are archaisms preserving their original consonantism because the relevant sigmatic aorist was lacking. These are important results which clarify several puzzles in diachronic Greek linguistics.
The synchronic analysis is divided into two parts, distinguishing V1 compounds without, and with, V2 counterparts (i.e. examples like μελλόγαμος (S. Ant. 628) versus pairs like ἀλγεσίθυμος and θυμαλγής). Those without a V2 counterpart are divided into three classes, according to whether the relevant verbs do (Class II) or do not (Class I) form agent nouns, or whether the verbs have V2 counterparts with significant semantic differences (Class III); Class IV is precisely those V1 compounds with V2 counterparts. Tribulato thus provides richly documented discussion of each verbal root and the kinds of formations that could be derived from it. In the case of Classes I-III, it emerges that the lack of an agent noun prevented the formation of compounds with V2 structure and required the exploitation of a different pattern; in some cases this is due to a verb’s rarity (thus obviating the need for a productive agent noun), in others because of phonological blocking (agent nouns in -ός cannot be formed to a non-ablauting root), in still others because of semantic blocking (e.g. in the case of ἀγαπάω which is a less common synonym of φιλέω). Class IV, finally, reveals stylistic variation in the use of compounds (explored in detail 336-344), as well as the specific role of V1 compounding in onomastics. Tribulato demonstrates an inverse relationship in compounding practice: highly productive V2 patterns predict very small sets of V1 compounds (though sometimes both are common, e.g. in the family of lexemes from ἔχω). The synchronic chapters steer an instructive course between drawing out trends in word formation while accounting for differences in individual cases; the creative potential in language is thus balanced against speakers’ needs for systematic principles.
The argument of this book is compelling, the philology is rigorous, and the sensitivity to the language and the texts analysed is considerable; this book is an altogether splendid achievement. Without wishing to be what I have learned from Tribulato (161) is called a cepidlaka in Serbian, the following paragraphs note minima, mistakes, disagreements, and the occasional further thought.
66: ‘pulse’ for ‘pulses’.
161, middle of last paragraph: ‘a a’ for ‘a’.
167: including βλαψι- in discussion of the form of the FC of the φερέοικος type is somewhat confusing. It will be part of Tribulato’s strategy to distinguish the behaviour of root and s-aorist; evidently then the issue is not simply that the aorist stem per se is preferentially selected. In addition, the other examples (δακε- from δάκνω and λαχε- from λαγχάνω) are unmarked (root) aorist forms, as opposed to βλαψι- which is related to the characterised (s-aorist) form (and also opposed to the more obviously thematic aorist FCs such as λιπο-).
168: what are the conditions for the analogical pressure of s-stem sounds such as μῖσος leading to FCs with a linking vowel -o-? Does this imply the loss of intervocalic /s/, and that this is therefore a post-Mycenaean development?
171: in what sense is ἀστεργής a V1 compound?
175: three of Tribulato’s list of sixteen verbs which do not form action nouns do in fact have action nouns: ἄρσις (first in Aristotle) to ἀείρω, ἕλξις (first in Plato) to ἕλκω, and ἕσις (more common as the second member of a compound, but found uncompounded at Pl. Crat. 411d) to ἵημι. They therefore belong to the second category, of verbs which produce an action noun later than the compounds. (Since Tribulato accepts the pair πῆξις and πηγεσίμαλλος in this category, the different ablaut grades are no barrier to her categorisation; in any case, it is precisely this mismatch which is crucial for her argument).
189 fn.74: I miss a reference to Jasanoff’s views on the s-aorist.
199, final line: the argument is surely not that -ti- is blocked, but that it is retained; presumably something has fallen out of the sentence (‘the
200: I am not convinced that ‘transparency of the root’ is pertinent in the development of personal names; rather, the lack of a shift from -ti- to -si- will be due to the lack of a competing formation (i.e. no available sequence of [root + s] to which a name might be assimilated).
276: Menander does not use the word ξενηδόχος. The passage referred to s.v. in LSJ is from the Menandrean Sententiae, Mon. 556 Jäkel (= 402 Meineke), where it is in any case an (otherwise unattested) emendation of the unmetrical MS reading ξενοδόκους. The second member -δοχος therefore cannot be dated; it cannot be cited as a counterexample to Tribulato’s argument.
404: it would be interesting to have Tribulato’s view on the readings π[ε]λεμαίγιδος vs. π[ο]λεμαίγιδος in Bacchyl. 17. 7. Are there linguistic criteria about which reading is more likely? In other words, could the verb πελεμίζω form a V1 compound like πελέμαιγις? I find no agent noun attested, which would put the verb in her Class I, so at least the morphology seems possible, especially since the verb is of elevated character.
1. The overall scope and quality of the book reminded this reviewer forcefully of T. Meissner, S-stem Nouns and Adjectives in Greek and Indo-European (2005), on which see D. Kölligan, BMCR 2007.02.05.