Word-Order Transference Between Latin and Greek

Word-order Transference between Latin and Greek
Our research focuses on a feature of Greek and Latin style that was not readily apparent before the corpus of ancient texts was digitized and analyzable by various text processing programs.

Applied linguists have observed that learners of second language often transfer features of their first language to the language they are studying (cf. S. Gass and L. Selinker 1983, Andersen and Shirai 1994). These features may range from the phonological to the morphological or syntactic. In applied linguistics, the reason for studying such effects (sometimes called "input bias") is pragmatic: by understanding the kinds of errors that typically arise in language learning, it is hoped to streamline the process of language acquisition.

In this paper, we study one example of such transference for Greek and Latin concerning an aspect of word order in Greek and Latin: the placement of the direct object with respect to the main verb. First we establish a significant difference in the Greek and Latin distributions. Whereas Greek writers tend to place the direct object before and after the main verb with more or less equal frequency, Roman writers have a distinct tendency to place the direct object before the main verb. Evidence for this is presented from the following writers: for Greek, Diodorus Siculus, Diogenes Laertius , Eunapius, Eusebius, Herodotus, Philostratus, Polybius, Thucydides, Xenophon; for Latin, Caesar, Cicero, Historia Augusta, Livy, Tacitus. Searches of relevant recent bibliographies (Janse, 1994; Werner, 1994) and personal communications from Classical linguists suggest that this difference in the Greek and Latin distributions has not yet been noted. Preliminary data suggest that the patterns found for the direct object also hold for the indirect object.

Next we show that when Greek authors write in Latin, they usually preserve their native Greek word-order pattern; and this is also seen to happen when Latin authors write in Greek. For Greek, we use the examples of the Roman historian Dio Cassius, the emperor-philosopher Marcus Aurelius, and Clemens, bishop of Rome ; for Latin, the example of the Greek historian, Ammianus Marcellinus.

Our theory to account for these observations is as follows. Since both Latin and Greek had relatively free word order, there was no definite "right" and "wrong" position of the direct object with respect to the main verb. In each language, the direct object can either precede or follow the main verb. Thus, the second of Andersen's two conditions for the transfer to somewhere principle (TTS) is fulfilled: "(1) natural acquisitional principles are consistent with the native language structure or (2) there already exists within the second language input the potential for (mis-)generalization from the input to produce the same form or structure" (Andersen 1983; Andersen 1990). In learning the other language, Latin and Greek speakers were probably not corrected for their DO placement in any individual utterance. That the overall distributions in the two languages were quite different is not something that was ever consciously observed in antiquity; nor was it readily apparent before the application of computing to Classical philology through the digitization of the entire corpora of Latin and Greek and the application of statistical analysis to stylistic features readily retrieved from the corpora. Indeed, the ancient grammarians stressed the congruence and even consanguinity of the two languages, sometimes viewing Latin as derived from Greek or even as a form of the Aeolic dialect (cf. R. Giomini 1953; E. Gabba, 1963; K. Schoepsdau 1992).

Having seen that the author's native word-order pattern biases his style when he writes in the other language, we turn to a related, but unexpected, analogous case of input-bias: when a writer of one of the languages writes in his native language using a source in the other language, his word-order preferences begin to shift to the word order typical of the source language. This kind of input -bias we call "cross-influence." The most extreme cases occur in translations ( see, e.g., A. Debrunner and A. Scherer 1969, for Latinisms in Greek translation s generally); but it is also often operative when an author writing in one language simply uses a source from the other. Examples discussed for translation include: Aulus Gellius (cf. P. Steinmetz 1992), Cicero (cf. C. Mueller-Goldingen 1992), and the Res Gestae Divi Augusti. For the input-bias of sources we discuss: the Greek and Roman lives of Cornelius Nepos (a Latin writer) and Plutarch (a Greek writer). Cluster analysis is used to show that the Roman biographies of Plutarch (though written in Greek) have distributions resembling the Roman lives of Nepos; whereas the Greek lives of Nepos (though written in Latin) have distributions approximating those of the Greek biographies of Plutarch. Our claim is purely empirical: not that cross-influence must occur whenever an author uses a source in the other language, but simply that such input-bias can be documented in the cases we have studied.

Our theory for explaining cross-influence again is based on Andersen's theory of input-bias. In this case, however, the input comes not from the writer's own native Sprachgefuehl but from the syntax of his source, which evidently exerts an unconscious, attractive force on the author of the derivative text, who imitates not only the content but also the style of the source-text. (An interesting analogous case of the subtle influence of English word order on Eskimo is discussed by M. Fortescue, 1993.)

Once these transference effects have been established, we suggest that they can be utilized--not for instructional purposes, as has been the case with Applied Linguistics and living languages--but for the solution of some philological problems. For example, we study the distributions of Books I and II of Cicero's De Officiis (which we definitely know had a Greek source) and compare them to what is found in Book III (where the question of a Greek source is controversial; see Dyck 1996). We use a similar approach to determine the likelihood that several ancient authors of unknown origin were Greeks or Romans (e.g., Greeks (?) writing in Latin: Euanthius and Charisius; see P. L. Schmidt 1989; Roman/Greek (?) writing in Greek: Plotinus).

References
Andersen, R., 1983: "Transfer to Somewhere," Language Transfer in Language Learning, edited by Susan Gass & Larry Selinker, Boston: Newbury House Publishers, pp. 177-201.

Andersen, R., 1990: "Models, Processes, Principles & Strategies: Second Language Acquisition Inside and Outside the Classroom," in Second Language Acquisition/Foreign Language Learning, ed. B. VanPatten an d J. F. Lee, Clevedon and Philadelphia, pp. 45-68.

Andersen, R. and Shirai, H., 1994: "Discourse Motivations for Some Cognitive Acquisition Principles," Studies in Second Language Acquisition 16.2, pp. 133-1 56.

Debrunner, A. and Scherer, A., 1969: Geschichte der griechischen Sprache, vol . 2, Berlin.

Dyck, A., 1996: Cicero's De Officiis. A Commentary, Ann Arbor.

Fortescue, M., 1993: "Eskimo Word Order Variation and Its Contact-Induced Perturbation," Journal of Linguistics 29, pp. 267-289.

Gabba, E., 1963: "Il latino come dialetto greco," in Miscellanea di studi alessandrini in memoria di Augusto Rostagni, Turin, pp. 188-194.

Gass, S. and Selinker, L., eds., 1983: Language Transfer in Language Learning, Rowley, Mass., London, Tokyo.

Giomini, R., 1953: "Il grammatico Filosseno e la derivazione del latino dall'eolico," La Parola del Passato 8, pp. 365-376.

Janse, M., 1994: "L'ordre des mots dans les langues classiques. Bibliographie des annèes 1939-1993," Tema. Techniques et Mèthodologies modernes appliquèes a' l'Antiquitè1, pp. 187-211.

Mueller-Goldingen, C. 1992: "Cicero als Uebersetzer Platons," Zum Umgang mit fremden Sprachen in der griechisch-roemischen Antike, Palingenesia 36, 173-188.

Schmidt, P. L., 1989: "Grammatik und Rhetorik," in Restauration und Erneuerung . Die lateinische Literatur von 284 bis 374 n. Chr., ed. R. Herzog, Munich, pp . 101-214.

Schoepsdau, K., 1992: "Vergleiche zwischen Lateinisch und Griechisch in der antiken Sprachwissenschaft," in Zum Umgang mit fremden Sprachen in der griechisch-roemischen Antike, Palingenesia 36, pp. 115-136.

Steinmetz, P., 1992: "Gellius als Uebersetzer," Zum Umgang mit fremden Sprachen in der griechisch-roemischen Antike, Palingenesia 36, pp. 201-212.

Werner, J., 1992: "Bibliographie zur Problematik der Fremdsprachlichkeit in der griechisch-roemischen Antike," Zum Umgang mit fremden Sprachen in der griechisch-roemischen Antike, Palingenesia 36, pp. 233-252.