BMCR 2014.01.26

Absolute Constructions in Early Indo-European. Cambridge classical studies

, Absolute Constructions in Early Indo-European. Cambridge classical studies. Cambridge; New York: Cambridge University Press, 2013. xiv, 255. ISBN 9780521767620. $99.00.


Absolute constructions (ACs) have been treated in monographs and articles since the 19th c. at least (cf. Wenzel 1828), and recent times have seen a fresh surge of research on this syntactic phenomenon (cf. e.g. Krisch 1988, Keydana 1997, Ziegler 2002, Maiocco 2005).1 As the book under review shows, the interest has not abated.

The book aims to give a definition of absolute constructions (ACs) and develop a theory of how they came into being in Proto-Indo-European (PIE) and developed in the daughter languages, mainly Greek, Latin and Sanskrit. Unfortunately, no other languages or language families having ACs are taken into account, such as Germanic and Slavonic. It would certainly have been worthwhile to dedicate some lines to the discussion of whether the constructions found in these branches are completely independent developments or in part (?) based on the models of Greek and Latin. The statement (p. 4) that ACs will be studied in three ancient IE languages “in which they are attested without doubt and in numbers that allow any meaningful study” would certainly also apply to other branches of Indo-European.

The book starts with a short history of research on ACs (p. 1-32) and then discusses the phenomenon in Greek (p. 33-81), Latin (p. 82-126) and Sanskrit (p. 127-171). The last chapter (5, p. 172-229) deals with the question of the possible origins of ACs in Indo-European.

In the review of the history of research on ACs, Ruppel criticizes earlier scholarship on ACs for not giving a precise definition of the phenomenon and for being misled by the term ʻabsoluteʼ into thinking that the defining characteristic of an AC is the fact that it does not attach to any other constituent of the phrase, but to the phrase as a whole—which may be said of adverbials, too. Keydana’s point (Keydana 1997: 24) that the case usage in ACs is deviant from the way the case is usually employed is rejected by Ruppel, since according to her, this is a matter of gradual rather than clear-cut differences (e.g. Greek τῆς ἐπιούσης νυκτός ʻin the following nightʼ beside ACs with temporal meaning). She agrees with Keydana and Ziegler on the fact that the “dominance” or obligatoriness of the participle is a defining feature of ACs. Her constant criticism of earlier scholarship for relying too much on the translation into the scholar’s mother tongue seems exaggerated, and the author herself does not always avoid impressionistic judgements, e.g. when claiming that (p. 25) Vedic sū́rye udyatí ʻat sunriseʼ “seem[s] equivalent more to noun phrases than to verbal clauses.” Her working definition of ACs (p. 30) is that ACs are temporal expressions with non-temporal nominal heads, i.e. heads not describing events, but things. The co-occurring participle (or adjective) serves to “temporalize” its head.

The chapter on Greek contains interesting results and points that merit further study, among them the fact that in Homer genitive absolutes (GA) are more frequent in direct speech than in narrative passages, which tend to be more conservative. Most of the GAs in Homer are restricted to the bare head noun + participle of the type ἠελίου ἀνίοντος, and it seems that the construction gradually became more complex by including adverbials and objects depending on the participle such as ( Il. 19.74f.) μῆνιν ἀπειπόντος μεγαθύμος Πηλεΐωνος. Accordingly, Ruppel tries to show that most transitive ACs in Homer are late or even post-Homeric. In some instances, this interpretation seems forced, though possible, e.g. in Od. 9.257ff. … ἡμῖν δ’ αὖτε κατεκλάσθη φίλον ἦτορ, / δεισάντων φθόγγον τε βαρὺν αὐτόν τε πέλωρον, which is taken by Ruppel to have an adnominal genitive dependent on ἦτορ, as in the frequent construction of a body-part noun followed by a genitive such as Il. 20.413f. τὸν βάλε … νῶτα παραΐσσοντος. Still, one wonders why it is not δείσασιν. In other instances, Ruppel’s close reading and comparison of similar passages is fruitful, e.g. Od. 24.535 πάντα δ’ ἐπὶ χθονὶ πῖπτε, θεᾶς ὄπα φωνησάσης (p. 54), which Ruppel shows to be a recycling of Il. 2.182 ὃ δὲ ξυνέηκε θεᾶς ὄπα φωνησάσης ʻhe heard the voice of the goddess as she spoke,ʼ reused as a GA, after φωνέω had been reinterpreted as governing ὄπα.

The minimal GA in Homer is not yet fully apt to supplant subordinate clauses in all their complexity. Since it mainly conveys temporal meanings it cannot step in for non-indicative subordinate clauses. GAs are proportionately more frequent in Hesiod than in Homer, but still mostly with subjects of “natural time” (day, night, month, season, year, etc.) though they tend to be more complex; and in 7th and 6th c. texts semantic restrictions no longer apply. The claim that the GA was a defunct syntactical device in post-Homeric Greek and was revived artificially by prose authors in the 5th c. (cf. e.g. Thesleff 1958) cannot be substantiated by these data, which rather speak for a continuous development. It is all the more regrettable that Classical Greek is not part of the study.

Latin: Since the Latin ablative goes back to the IE ablative, instrumental and locative cases, many of its adverbial uses are close to the ablative absolute (AA), which is less clearly delineated from other uses of the case than the GA in Greek, which only competes with the temporal genitive and the adnominal use. For example, it is difficult to differentiate the AA from the ablative “of attendant circumstance,” e.g. Pl. Am. 1093f. manibus puris capite operto (p. 91f.) and from the ablative of quality. As for nominal ACs of the type me vivo, Ruppel assumes that they may have arisen from AC with participles close in meaning to adjectives, since in Latin participles are more “nominal” than for instance in Greek, where they are derived from the various verbal stems. The transition may be seen in a case such as luna silenti beside luna silente ʻat new moon,ʼ with the adjectival ending in Cato, and the participial ending in Columella. As in Greek, most early ACs contain temporal expressions ( piro florente, vindemia facta). Due to the case syncretism, it is a priori possible that instrumentals and locatives as well may have played a role in the development of the AA. Ruppel argues this for the agent noun type me suasore and the type me praesente which may have developed from instrumental constructions where the head noun of the AC may originally have been involved in the action of the matrix clause. This may also apply to the type P. Muccio Cos. “when P.M. was consul.”

Sanskrit: After a tedious introduction to Sanskrit grammar and literature (p.127-137), Ruppel argues that since the locative is a fully semantic case in Sanskrit, the AC is even less clearly delineated against other uses of the locative than the Latin AA and the Greek GA: adhvaré prayatí ʻat the sacrifice, while it is/was going onʼ need not be an absolute construction, the participle may be left out: adhvaré ʻat/during the sacrificeʼ (cf. Ziegler 2002:80f. “Substantiv im Lokativ mit attributivem Partizip”). The past participle, however, as in avasite vacasi ʻwhen the speech has ended,ʼ cannot be left out, since vacasi alone means ʻduring / at the speech.ʼ If obligatoriness and semantic dominance of the modifier are taken as part of the definition of an AC, then this could be one, and a possible starting point. As she does for Greek and Latin, Ruppel claims that the seed of ACs in the Rig Veda are expressions of natural time (p. 153), e.g. ʻat sunrise/-setʼ, sū́rye udyatí (RV 8.27.19), where no spatial interpretation is possible, and where ʻsunʼ cannot stand alone meaningfully in this sense (*ʻin the sun/at the sunʼ). From such a starting point, constructions with past participle developed, sū́rye údite ʻwhen the sun has risenʼ and constructions with negation such as anudite sū́rye ʻwhen the sun has not yet risenʼ (first attested in the Brahmanas), a development she compares with that of me praesente and me absente in Latin: me praesente, frequent in Plautus and Terence, may rely on the comitative use of the instrumental ʻwith me present,ʼ from which me absente was derived, with a no longer physical, but only “situational” comitative. The development in Sanskrit is then assumed to have followed a path similar to the one described for Greek and Latin, i.e. from expressions of natural time to any temporalized head noun such as (Śatapatha-Brāhmaṇa) tāsu itāsu ʻwhen they had gone.ʼ In support of this, she argues (p. 159) that the native grammatical tradition on ACs in Sanskrit (notably Pāṇini) does not make an unequivocal special mention of the AC when discussing the locative—apparently, it was not seen as something sufficiently deviant from other uses of the locative to deserve a special entry in the grammar.

The final chapter dedicated to the possible IE origins of ACs (ch. 5) contains a number of banalities, repeating textbook knowledge (e.g. on p. 200 one can learn that “The nominative, for example, simply marks a noun as the subject of the clause.”), and some highly questionable statements that would have merited second thoughts, e.g. (p. 202): “We commonly reconstruct for PIE such a system where each formally distinct case represents one general idea or function.” (In the footnote, the instrumental is said to be an exception to this, since it has both instrumental and comitative function.) This is the weakest part of the book. Still, the basic assumptions made for the development of ACs seem reasonable: the starting point may have been expressions of natural time, such as “at dawn rising” (with a temporal head noun) in the locative, whence ʻat sun-riseʼ (with a non-temporal head noun), which was further extended to temporalize any noun phrase. In Latin (p. 213), there may have been a merger of absolute-like instrumentals and absolute locatives, if the types me praesente/absente and me suasore go back to the comitative instrumental. It seems more difficult to find a reason for why it was the genitive that came to be used in an AC in Greek (p. 219f.). Ruppel’s points, however,—the relative frequency of the temporal genitive, the fact that the dative nearly always allows the interpretation as a dative of respect, whereas a genitive is absolute if it is not ʻgoverned’ by the verb or adnominal—may at least give a hint as to why it was preferred to other cases.

To sum up: a useful study of the absolute constructions in Greek, Latin and Sanskrit with lamentable, yet understandable, lacunae. It will be of help for the further study of these and similar constructions.2


1. References: Krisch, Thomas. 1988. Zur semantischen Interpretation von absoluten Konstruktionen in altindogermanischen Sprachen. Innsbruck. Keydana, Götz. 1997. Absolute Konstruktionen in altindogermanischen Sprachen. Göttingen. Maiocco, Marco. 2005. Absolute Participial Constructions: A Contrastive Approach to the Syntax of Greek and Latin. Pisa. Thesleff, Holger. 1958. ʻOn the origin of the genitive absoluteʼ, Arctos 2: 187–207. Ziegler, Sabine. 2002. ʻZur Entstehung des locativus absolutus im Altindischenʼ, in Heinrich Hettrich (ed.), Indogermanische Syntax: Fragen und Perspektiven. Wiesbaden, 79–86. Wentzel, Eduard.1828. De genitivis et dativis linguae Graecae, quos absolutos vocant. Breslau.

2. A few typos and mistakes: p. 18 “What should make of such a weakening of grammatical ties (and therefore a syntactic reanalysis) unlikely?”
p. 84 “dependentibus” for “defendentibus”
p. 107: “perducendam” instead of “reducendum”
p. 118: “praesente” is discussed as qualifier in the phrase me suasore atque impulsore id factum audacter dicito (Pl. Mos. 916).
p. 182: “phase” instead of “phrase”
p. 230 “Wenzel (1828)” instead of “Wentzel (1828)” (given correctly on p. 249).
p. 230 (same author) “ De genetivis ” instead of “ De genitivis ” (also p. 249).