Background

So distraught were the daughters of the Sun, the Heliades, at Phaethon's death that they were turned into trees around his tomb and their tears into amber; while Cycnus (a grieving relative of Phaethon) was turned into a swan (Greek kuknos). The Sun was finally persuaded to return to his daily task, and Jupiter surveyed the damage to the world. While doing so, he had his way with the huntress Callisto, whom a furious Juno turned into a bear, but Jupiter re-transformed into a constellation. Various stories about gods' affairs, some told by a crow and a raven, ensue, and Book 2 ends with Jupiter, disguised as a bull, riding off with Europa.

Book 3 opens with Europa's father Agenor, who came from Phoenicia (Lebanon), ordering his son Cadmus to find Europa or go into exile. Cadmus chose the latter and consulted Apollo at Delphi, who told him to found a city (Thebes) in Greece, in Boeotia. Defeating a terrifying serpent there (which had killed all his companions), Cadmus was told by Athena/Minerva to sow its teeth in the ground. From these sprang armed warriors, who fought among themselves until the last five still standing agreed to stop and join Cadmus in founding the city. Cadmus married Harmonia, the daughter of Mars and Venus, and all seemed set fair for him. But Ovid goes on ‘Yet a man should await his final day, and no one be called happy until he dies and his last rites are paid.’

As Schwyzer pointed out in his article on expressions of the agent, the Attic tragedians provide many examples of non-standard agent markers (1943: 20–8). In fact, ὑπό is so rare in comparison to other prepositions, that, were these plays our only source of Ancient Greek, we would not at first glance be able to pinpoint it as the default agent marker. In iambic passages of the Oresteia, for instance, ὑπό only occurs in three PACs, as against seven with πρός+G, two with ἐκ, and one with παρά. Does this variety of agent marking mirror that in prose of the period, or is it simply a feature of poetic diction, conditioned by the meter? In order to answer this question, I will look at prepositional PACs in Aeschylus, Sophocles, and Euripides. To impose some uniformity on the data, I will only consider PACs in iambic passages. Because this limitation reduces the number of PACs under consideration, I will extend the study to those constructions where agent expressions occur with intransitive verbs, such as πάσχω or ἀποθνῄσκω, that act as suppletive passives to transitive counterparts like ποιέω and ἀποκτείνω. Such constructions are quite common in tragedy because of the frequent description of suffering in the genre.

In general, the pattern in prose whereby certain agent markers are associated with certain verbs does not hold true in poetry.

The analysis of PACs in Homer will be quite different from that of those found in later authors. To begin with, there are very few PACs in the Homeric corpus, as seen in the table on p. 40. The independent passive voice was a relatively recent development at that time, and Homer preferred to keep the verb voice in the active, even at the expense of having to change the sentence subject frequently. Additionally, the metrical, formulaic nature of the oral poetry skews the data: the apparent frequency of some PACs might simply result from their metrical utility or occurrence in a formula. Nevertheless, some agent markers can be singled out which seem to have been perceived as characteristic of Homer, to judge from their use in the late epic poets who imitated Homer’s language.

The PACs in Homer can, for convenience’s sake, be split into two groups: those in which the agent is marked solely by a case, and those in which a preposition marks the agent. Although the former category is said by some to include a genitive of agent, it is argued here that the only case in Homer that by itself can mark the agent is the dative. The PACs with prepositions as agent markers themselves break down into two types.

Most passive verbs in Greek express their agent by means of the preposition ὑπό with the genitive. The most common exception to this rule is that passive verbs in the perfect generally construe with an agent in the dative case:

Because of this anomaly, many scholars have denied that the dative found with perfect passives is an agent at all. Instead, they would describe this usage as a dative of interest. What is less clear, however, is the reason why the perfect is distinguished from the other aspects of the Greek verb in this way. The answer seems to lie in the stative nature of the perfect: if no dynamic action is being described, what place is there for an agent? Furthermore, while it is quite plain that the usage of the Greek perfect changed significantly over the period from Homer to Koine, the effects that this change had on the expression of the agent with the perfect passive have not been fully examined. As early as Herodotus, some perfect verbs have ὑπό+G rather than the dative marking the agent, notably when the subject (that is, patient) of the verb was animate. This use of the perfect passive with an animate patient becomes more frequent by the time of Polybius.

In the previous chapters, it was seen that several prepositions, primarily ablatival in sense, competed with ὑπό as means of marking the agent. While ὑπό was clearly the dominant preposition in Attic authors like Plato, Xenophon, or Demosthenes, its position was not particularly stable: especially in Demosthenes, παρά sometimes alternates with ὑπό in apparent free variation. These other prepositions were thus situated in a place from which they could potentially oust ὑπό from its role as the agent marker. And ὑπό did indeed fall out of use in the end. As can be seen from modern Greek, it was ἀπό that eventually replaced ὑπό as the grammaticalized means of marking the agent. Scholarship on this development is scanty. For the most part, it is summarily stated that ἀπό simply took over from ὑπό, primarily on the basis of its occurrence in the Septuagint (LXX) and New Testament (NT). A fuller examination of the evidence, however, suggests that the development was not so straightforward. Although ἀπό does occur suggestively early in the LXX and NT, it is not the dominant agentive preposition in the papyri of late antiquity or the lower-register texts of the Byzantine period. It appears rather that παρά+G was the immediate successor of ὑπό, and that ἀπό was not the primary agentive preposition until after the twelfth century AD. Because texts not influenced by classical language are so rare in the relevant period, the details of this change will remain obscure.

Beginning students of Ancient Greek soon learn that the agent of a passive verb is marked with the preposition ὑπό followed by the genitive. Then, of course, the exceptions come to light. The most common of these is the dative of agent, which, for the beginner at least, may be explained away as occurring with perfect passives and -τέος verbals. Later, however, one comes across other irregularities, notably the use of prepositions other than ὑπό+G. The conditions that motivate these apparently anomalous agent markers have not yet been satisfactorily explained. The aim of this book is to do so.

I begin with an introductory chapter that lays a theoretical foundation for the work and discusses the reasons why these passive-with-agent constructions (PACs) occur in the first place. In Chapter 2, I move on to Homer, as the Iliad and Odyssey represent the earliest texts that are syntactically complex enough to have PACs. Because the Homeric data are so different from the later evidence – in particular, these constructions are far less common in Homer – they are best dealt with separately. Next, in Chapter 3, I look at the dative of agent.

While ὑπό is by far the most common agent marker in classical Greek prose, other prepositions do occur in its place. In this chapter, I will explore the linguistic motivation for these non-standard agent markers. Generally speaking, there are two factors that affect the selection of agent markers. The more important is the semantic field to which the verb belongs: verbs of sending and giving often use an ablatival preposition like ἐκ or παρά+G; verbs of thinking often use a locatival preposition like πρός or παρά+D. The second, lesser factor concerns the syntax of the verb: participles sometimes take different agent markers than would finite verbs. The body of this chapter will examine the works of six major prose authors in turn (Herodotus, Thucydides, Xenophon, Lysias, Plato, and Demosthenes) to demonstrate how these two factors affect the expression of the agent.

Survey of earlier literature

Scholarship on this question has been surprisingly scarce. Aside from the section in Kühner–Gerth (1898) on the uses of prepositions generally, the chief existing studies are Schwyzer’s lengthy article on agent constructions (1943) and a brief article on agenthood by Luraghi (2000). Both Kühner–Gerth and Schwyzer suffer from a failure to address the conditions under which non-standard agent markers are used. They merely offer lists of examples, and vague statements as to their distribution in the various authors.