An old theory known as localism holds that spatial relations, which are the framework in which we experience the world outside of us, form the basic underpinning of language. Hints of this theory, whose formal history begins in early nineteenth-century Germany, have been traced back to Maximus Planudes in the thirteenth century or even Apollonius Dyscoles in the second century CE; and it has always been most plausible when applied to cases and prepositions, whose spatial element, both in the classical and the modern European languages, is usually close to hand. Plausible though it be, however, the localist hypothesis has its opponents, and for those to whom it is not immediately obvious that government of the people, by the people, and for the people is a derivative expression whose origin is government off the people, near the people, and before the people it is not a simple matter to prove or disprove it.
As a hypothesis about synchronic linguistics, localism will have to wait; current ideas about the way languages structure meaning are so varied and fluid that it is not clear what we would have to demonstrate in order to know that the spatial meaning of a given preposition is its “basic” or “essential” meaning. Diachronically, however, it is clear what we would have to show: that the spatial meanings of prepositions appear before their abstract meanings. Even better would be to demonstrate the “unidirectionality hypothesis”, that development goes only in one direction: spatial prepositions develop abstract meanings, but abstract prepositions do not develop spatial meanings.
The Greek language, with the longest recorded history of any western language, is a good place to test these hypotheses. The tester will have to be someone familiar not only with ancient and modern Greek, but with the various Greeks that have gone between; an ability to recognize the influence of neighboring languages will also help. Not many classicists, and not many linguists, fit this description; and I was surprised to discover that Pietro Bortone, who has ventured on precisely this undertaking, has had the temerity to do it for his doctoral thesis at the very beginning of his career.
The results are now in, and they are both impressive and convincing. Eschewing the anecdotal, Bortone has taken four snapshots of the language, from the classical, Hellenistic, medieval, and modern periods, each of them analyzed by means of a representative corpus, though he occasionally resorts to other material (Mycenaean and Homeric Greek at the early end, neighboring languages for medieval and modern Greek) to illustrate or to test a point. His conclusion is:
“that in the history of the Greek prepositional system, whenever new sets of prepositions have come into use, they almost invariably expressed spatial meanings, and that those spatial meanings were contemporaneously lost by the older items (e.g. ὑπό > κάτω). The older (partial) synonyms, on the other hand, could survive, unless and until they fell out of use, carrying (often numerous) abstract meanings. They could even develop new meanings, but only non-spatial ones: meaning changed only from spatial to non-spatial. Of the old prepositions, only those which had no younger rivals were able to retain spatial senses. The new exclusively spatial prepositions, in time, developed a range of non- spatial meanings. There is now even marginal evidence that the new prepositions eventually lose their spatial sense completely, thus repeating entirely the life-cycle of their predecessors. (302-3)”
This is Bortone’s most important conclusion, and for a classicist it will be the medieval examples that show it most clearly. Many prepositions had lost their spatial meaning entirely: ἀνά for “up” had been replaced by ἐπάνω(θεν) or ἀπάνω(θεν), but was still used distributively; περί for “around” was almost entirely replaced by κύκλῳ(θεν) and (τρι) γύρῳ(θεν), but retained the meaning “about” in specifying a topic. Others were only partially replaced: ἐπί could still mean “on”, though that was more often expressed by ἐπάνω, but it mainly indicated the abstract grounds for a thing. Three prepositions for which no recent replacement was available—ἀπό, διά (now γιά, as it remains today), and εἰς (now σέ)—maintained their spatial meaning. These are a few examples from a rich and nuanced discussion; for unidirectionality a single example will have to suffice. In classical Greek, ἐκ, the most commonly used preposition for movement “away”, indicates movement “out of” (i.e., movement that began inside the object, ἀπό movement “from” (movement that began at or near the object), though this distinction is already fading where either might be used: cf., in adjacent paragraphs, τῶν ἐκ Πελοποννήσου νεῶν and τάς τε ἀπὸ τῆς Πελοποννήσου ναῦς (Thuc. 7.17.1-2) (164).1 In koine ἀπό, though still less common than ἐκ, is beginning to take over some of its functions: along with ἀνέβησαν ἐκ τοῦ ὕδατος (Acts 8:39) we find ἀνέβη ἀπὸ τοῦ ὕδατος (Matt. 3.16), which might have raised some classical eyebrows. Ἀπό, moreover, is moving into new abstract territory, taking over causality from ὑπό with increasing frequency (ἀπιστούντων αὐτῶν ἀπὸ τῆς χαρᾶς, Luke 24:41, “when for joy they did not believe it”) (185). By the middle ages ἐκ, now synonymous and in fact interchangeable with the now slightly more common ἀπό, has become something of an archaism: while ἀπό takes the accusative that has now become the “regular” case for prepositional objects, ἐκ still takes the genitive, and appears in more archaizing contexts. Its spatial uses are being usurped by the “improper” prepositions ἔξω(θεν) and ἐκτός(θεν) (228-30)—themselves wholly spatial prepositions in Homer that had developed some non-spatial uses as early as the classical period (170). In modern Greek ἐκ may be said to be on life-support, maintained by the continuing influence of the now disestablished katharevousa (254 n. 25). Ἐκ has gone through the full life-cycle of a Greek preposition (168 Fig. 4.3, 237 Fig. 6.1, 283 Fig. 7.8), from a presumed meaning that was originally entirely spatial, through the classical preposition that served both spatial and non-spatial meanings, through one that was predominantly non-spatial and finally to near-extinction.
Although Bortone is admirably cautious about the general application of his conclusion (“The recurring pattern of development identified here is not claimed to be cross-linguistically applicable, but it will need to be taken into account in future discussions about the historical validity of the ‘localistic hypothesis’”, 303), its demonstration is a major achievement. But it is only a small part of the insight into the long-term development of the Greek language that the reader will gain from this perspicacious book. A few other observations:
1. Most classicists will be aware that the case system of Greek has been in decline from early times to the present, but Bortone sketches a rich picture of what has been happening. Reinforcement of oblique cases by prepositions (ἐν Πύλῳ for Πύλῳ) was already increasing from Mycenaean to Homer,1 and that development continued throughout the classical period (σὺν ἐμοὶ ἠκολούθησαν, Xen. An. 7.5.3; μετ’ αὐτοῦ ἀκολουθήσαντα, Lys. 2.27; ἀκολουθεῖ κατόπιν ἀνθρώπου τυφλοῦ, Ar. Plut. 13, 155-6), and by the middle ages had become so regular that translators of classical texts added prepositions where they were absent (202-3).
2. As the case-system before it, so the prepositional system decayed in its turn. In classical Greek the case governed by a preposition was so important that it could change the meaning to an entirely different one (μετὰ τούτων/μετὰ ταῦτα, even an opposite one (παρ’ αὐτοῦ/παρ’ αὐτόν). But already in classical Greek these distinctions are getting smudged: along with the usual dative for παρά meaning “near”, we find such things as παρ’ ὑγρῶν Ἰσμηνοῦ ῥείθρων, “by the wet currents of the Ismenus” (Soph. Ant. 1123) and ἦν παρὰ τὴν ὁδὸν κρήνη, “there was a fountain by the side of the street” (Xen. An. 1.2.13) (156-7). The newer “improper” prepositions almost all take a single case, the genitive. In Hellenistic Greek the distinction fades yet further (183-4), and the accusative becomes almost the only case used with prepositions; using other cases is an archaism (203-5). By the modern period the prepositions themselves, like the cases before them, have become so multifunctional that they are usually reinforced by another preposition, producing an entire spectrum of compound prepositions: where classical Greek distinguished ἐν τῷ οἴκῳ (“in the house”) from παρὰ τῷ οἴκῳ (“next to the house”), the modern language (where σέ can mean “in”, “at”, and many other things) distinguishes μέσα στὸ (= σὲ τὸ) σπίτι (“inside at the house”) from μπροστὰ στὸ σπίτι (“before at the house”) (265).
3. Anyone who knows both ancient and modern Greek knows that the dative has effectively disappeared, but looking at the language from beginning to end one can see the progressive decline of the dative from Homer, where ὑπό + dative was common and μετά + dative was still possible, through Xenophon, where the dative is the least common case with prepositions (155-6), to koine, where its use with prepositions other than ἐν is becoming rare. The modern situation was a long time in the making.
4. The diglossia that has plagued the Greek language has resulted in a limited cross-fertilization between ancient and modern Greek, with the result that even “dead” prepositions “may surface in casual speech today, but in a more restricted range of expresssions” (292).
None of this will startle a student of Greek historical linguistics, but it is a major achievement to present it in a work that combines the broadest possible chronological framework with a firsthand detailed collection of examples that convincingly demonstrate the diachronic developments. It is presented in a carefully structured, though not overly rigid, theoretical framework, some of it quotable. This is Bortone’s judgment on “improper prepositions”, i.e., those that cannot be used as preverbs: “This is worse than a ‘terme … pas très heureux’ (Humbert 1960: 299): it is a classification that does not recognize as prepositions words which are purely prepositions, on the grounds that they are not something else” (118). If you find discussions of how language works and how it develops interesting, read this book. If you don’t—well, that is your loss.
1. Anna Morpurgo Davies, “Mycenaean and Greek Prepositions: o-pi, e-pi etc.”, in Alfred Heubeck and Günter Neumann, eds., Res Mycenaeae (Vandenhoeck and Ruprecht, Göttingen 1983), 288.