This electronic thesis or dissertation has been downloaded from the King’s Research Portal at https://kclpure.kcl.ac.uk/portal/

Grammar, logic and philosophy of language : the Stoic legacy in fourth century Patristics.

Robertson, David Gordon

The copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without proper acknowledgement.

END USER LICENCE AGREEMENT

Unless another licence is stated on the immediately following page this work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ You are free to copy, distribute and transmit the work Under the following conditions:  Attribution: You must attribute the work in the manner specified by the author (but not in any way that suggests that they endorse you or your use of the work).  Non Commercial: You may not use this work for commercial purposes.  No Derivative Works - You may not alter, transform, or build upon this work.

Any of these conditions can be waived if you receive permission from the author. Your fair dealings and other rights are in no way affected by the above.

Take down policy

If you believe that this document breaches copyright please contact [email protected] providing details, and we will remove access to the work immediately and investigate your claim.

Download date: 23. Sep. 2021 73ý Iz

David Gordon: '2> tc PhD Tt- es_s Lei. artrr et of 'i ilosophy Xi s College London 2000

`, ý'ýýý

_ ,, ., Thesis Abstract

Issues of language are of central importance to the theological arguments of the fourth-century Patristic writer Basil of Caesarea. His work shows sustained philosophical interest in meaning: changes of the meaning of expressions, completion of meaning, determination of meaning, speaker's knowledge of meaning, correctness and purity of meaning. Patristics is shown to be particularly relevant to contemporary philosophical concerns on account of the Patristic preoccupation with linguistic matters, which previous work in Ancient Philosophy and Patristics has not taken into account. This linguistic expertise is inherited from the grammatical, logical, and rhetorical training of Late Antiquity, which received the philosophy of the Stoics as transmitted by the Greek grammarians and other sources. But often it is hard to find any secure philosophical precedent for these ideas. They seem to arise naturally in the context of trinitarian debate and exposition. The five dissertation chapters begin with an historical introduction, providing a historical basis for the other chapters by discussing fourth-century training in grammar, rhetoric, logic, and philosophy. The following chapters deal with four areas: proper names, prepositions and conjunctions, relatives, and language and thought. Most of these topics are heavily discussed in contemporary philosophy, linguistics, and related fields. The thesis attempts to build a conversation with recent work in philosophy and linguistics, even while searching out the fundamental contrasts between ancient and modern approaches. Theological issues are treated in order to achieve a complete understanding of the philosophical points made. The philosophical tradition that is rooted in the Stoic project of ordering and clarifying the relationship between words and meanings for dialectical purposes is transformed in the classic trinitarian controversies of Greek Patristics, in which Basil plays a leading role.

2 Contents

Introduction 4-6

0. Grammar, Rhetoric, and Philosophy 7-33

1. Proper Names 34-57

2. Prepositions and Conjunctions 58-93

3. Relatives 94-112

4. Language and Thought 113-142

Conclusion 143-144

Index of Abbreviations 145-146

Index Locorum 147-150

Bibliography 151-156

3 Introduction

The dissertation before you is meant to be a contribution to the study of philosophy of language in Late Antiquity, by adopting a new approach. This approach takes seriously the discipline, closely related to Ancient Philosophy, of Patristics. It is based on the insight that Patristic writers are philosophically interesting in a sense never before appreciated, a sense that does not depend on the idea that the Church Fathers are doing philosophy as 'Christian philosophers'. Patristics is relevant to contemporary philosophical concerns on account of the Patristic preoccupation with linguistic matters. With a few notable exceptions (e. g. Augustine), there has been remarkably little research on the massive corpus of Patristic writing on language. Work in Greek Patristics has not neglected matters such as the influence of Greek Philosophy on Patristics altogether. Much interesting material has been discussed, particularly in a few outstanding thinkers such as Origen and John Philoponus. However, the work to date in Patristic writers has been dominated by the philological approach. Research has not been very philosophical, although some recent work has acknowledged the strong links to the history of linguistics. ' To some degree, this is still necessary due to various

1Ineke Sluiter, Ancient Grammar in Context. Contributions to the Study of Ancient Linguistic Thought (Amsterdam: VU University Press, 1990): 168-171 discusses the importance of Patristic writers as sources for Ancient Grammar, pointing out that "the ancient Christian writers in the commentaries on and explanations of the Bible often have useful material to offer to the student of ancient grammatical theory. " Pagan grammatical training is fundamental to Christian exegetical techniques. And many Christian texts are extant, though largely ignored by scholars of the history of linguistics. "Nevertheless, this material does fill some of the chronological gaps in the more strictly grammatical material. " Sluiter notes the pioneering work in Patristics of Christoph Schäublin, Untersuchungen zu Methode tend Herkunft der antiochenischen Exegese (Köln and Bonn: Peter Hanstein Verlag, 1974), and his student Bernhard Neuschäfer, Origenes als Philologe, Schweizerische Beiträge zur Altertumswissenschaft (Basel: Friedrich Reinhardt Verlag, 1987), which examines the relevance of

4 that by the texts. I have focused research on philological problems are presented my influential fourth-century Greek theologian, Basil Caesarea. one prominent and of be figures But similar dissertations could easily written on other of similar stature Alexandria, Gregory Nazianzus, Theodore Mopsuestia. such as Origen of of and of It is important to explain why linguistic matters are so important to Patristics. I think that Patristic writers talk about language more than anything else. But work in Ancient Philosophy and Patristics has almost never taken this into account. The focus has been on areas such as theology, cosmology, psychology, anthropology, by and ethics. And of course many Christian writers address these topics. But the fourth century, with the biblical canon formed and the increasingly exegetical character of theological disputes, leading figures employ their philosophical and philological skills not only in theological debate but also in the related task of biblical exposition in homiletical and commentary form. This is the golden age of trinitarian controversy, before the debates moved on in subsequent centuries to christology. The five chapters that follow begin with a historical introduction. This first chapter has a different character than the other four. It does not deal with philosophical problems, rather it attempts to provide a historical basis for the other chapters by discussing fourth-century training in grammar, rhetoric, logic, and philosophy. But it is a crucial chapter. I hope that it shows how Greek learning is relevant to linguistic concerns in fourth-century Patristics, in a way that I think is quite important but has traditionally been ignored for the sake of well-worn topics like the Platonism of the Church Fathers. The following chapters deal with four separate themes, which are proper names, prepositions and conjunctions, relatives, and language and thought. The chapters are ordered with the two more philosophical chapters preceding the two more theological chapters. The extensive theological discussion of the latter two chapters is much reduced in the proper names and conjunctions chapters, because I felt that here the theological points were of lesser and the philosophical and linguistic points of greater interest. But all the chapters are meant to achieve three objectives, (1) to

pagan grammatical and rhetorical training in some leading Christian exegetes. But Sluiter complains of the lack of reciprocal interest in the study of Ancient Grammar.

5 show how theological arguments take on a philosophical character, (2) to situate the ideas within the philosophical and grammatical tradition, and (3) to demonstrate the relevance of the Patristic material to contemporary philosophical concerns. However, not every one of these three desiderata is satisfied with equal success by each of the four chapters of the body of the dissertation, due (I think) to various constraints inherent in the topics. Consequently, there are differences in the prominence of philosophical or theological discussion. The topics chosen for the chapters represent loosely related areas of logic and language which are all prominent in the various theological works of Basil of Caesarea. For the sake of focus and precision, material on other Patristic writers is kept to a minimum except where I find it is too relevant to omit. Finally, I need to make a few remarks on reading this dissertation. As mentioned above, this work spends considerable time with a wide range of philological matters, including dating, authorship, authenticity, textual criticism, interpretation and translation, and so on. Much of this material engages with previous scholarship. Some of this has been cut back, some has been relegated to footnotes, but much of it remains in the text. I hope that the reader will recognize what material is of genuine importance to the main argument of each chapter and will skip the remaining material, which I have tried to keep in the footnotes. ' A similar guideline would also be appropriate for my historical remarks on fourth- century theology. On the other hand, my style of exposition, particularly in the proper names and the conjunctions chapters, at times places some important subsidiary points in the footnotes or marks off what I will not discuss. I trust that the reader will spot these excursions of interest.

6 0. Grammar, Rhetoric, and Philosophy

The importance of the Stoic legacy to Greek Patristic thought might appear plausible solely on the basis of Stoic interest in rhetoric. It has commonly been observed by Patristic scholars that the most influential part of classical culture on Patristic writers, besides Platonism, was the study of rhetoric. There is promise for the study of rhetoric in the Church Fathers, and recent work in Patristics and later Greek rhetoric has proved quite valuable. Rhetoric had been a branch of philosophical inquiry before the time of Plato and Aristotle and was pursued by their successors. Among the early Stoics, there is ample evidence for a strong interest in grammar and rhetoric, continuing at least up to the time of Posidonius. 1 But for the study of later Greek philosophy, there is the problem that Stoic rhetorical theory seems to have attracted precious little of a-following from the Hellenistic period onwards. 2 Studies of the Cappadocian Fathers and their forerunners, among whom Origen is the most important, have for the most part not looked very closely at the significance of rhetoric and grammar in their writings. It is beyond dispute that the schools in the Imperial period were dominated by rhetoric, and this was doubtless the experience of Basil and his friend Gregory of Nazianzus, who studied together at Athens from about 351 until about 355. But this should not license neglect for the role that grammar and philology played in fourth-century rhetorical training. This chapter will examine the role of grammatical and dialectical studies in contemporary Athenian schooling, as well as under the tutelage of the leading sophist Libanius, with whom Basil studied for a limited

1Posidonius. Fragments and Commentary, 3 Vols., 2nd edition, eds. L. Edelstein and I. G. Kidd (Cambridge: Cambridge University Press, 1989): Kidd F192. 2Catherine Atherton, "Hand Over Fist: The Failure of Stoic Rhetoric, " Classical Quarterly 38 (1988): 392-427.

7 time in Constantinople. There will also be some consideration of the influence of Origen in the intellectual formation of Basil.

I

A good case can be made that Basil and Gregory invested considerable effort in the study of rhetoric, grammar, and related disciplines while at Athens 3 As for philosophy, Basil and Gregory probably disliked whatever philosophy was taught in Athens. What could endear them to currents such as Iamblichean Neoplatonism with its pagan religious emphasis (theurgy) derived from Syria and the East?4 But it is probable that Basil ingested his quotient of philosophy as he found it taught in his course of studies. There is not too much evidence for the development of the school at Athens in the third and early fourth century. The institution was re-organized under the emperor Marcus Aurelius (161-180), who provided for four chairs of philosophy and one chair of rhetoric 5 For the character of study at Athens during the later fourth century, one of the extant orations of Himerius has been considered important. Himerius was one of the two prominent teachers of rhetoric (the other was Prohaeresius) while Basil and Gregory were students. 6 Saffrey and Westerink think -that one particular section of the speech yields information about the general features of study at Athens, and in particular for representation of the four main philosophical schools? Not much weight should be placed on this text, it reads like a stock catalogue of 'the great philosophers' which might have been trundled

3There is a general discussion of Basil's studies in Athens by Phillip Rousseau, Basil of Caesarea (Berkeley: University of California Press, 1994): 31ff.. 4H. D. Saffrey and L. G. Westerink, Proclus. TheologiePlatonicienne (Paris: Les Belles Lettres, 1968): 1.xli-xlviii argue that from the middle of the fourth century onwards the influence of Syrian Neoplatonism was influential in Athens, and waxed over the course of the century. 5John Rist, "Basil's 'Neoplatonism': Its Background and Nature, " Basil of Caesarea: Christian, Humanist, Ascetic, ed. Paul J. Fedwick (Toronto: Pontifical Institute of Mediaeval Studies, 1981): 1.137-220 at 1.182-183 surmises that the philosopher who held the 'Platonic' chair would have been titled the 'Platonic Successor', but thinks that perhaps after the devastation of Athens in 267 the title was not observed. For the re-organisation of the school, Rist cites Dio Cassius 72.31. For a sketch of the teaching of philosophy and rhetoric in Athens in the late third and fourth centuries, Rousseau, Basil of Caesarea, 29ff. provides a good basic account. 6Rousseau, Basil of Caesarea, 31. T. D. Barnes, "Himerius and the Fourth Century, " Classical Philology 82 (1987): 206-225 at 212 and 224 argues for the arrival of 1-iimerius in Athens in the year 352.

8 out at appropriate times by Imperial sophists and rhetoricians. Himerius delivered this oration in the early 370s in honor of Hermogenes, proconsul of Greece, giving an account of his career. The relevant portion of the text is presented directly after mention of Hermogenes' enthusiasm for rhetorical studies. ovTC.o AA811Kai eni 'r v OEwpiav Tcovnpawµ 'aov 9pxcTal. Tpt%ý R Tijg närnnS $t%o0o4iaS vcva uj tevTS, Kai tii}S µEv E.i, S TäS npäýEtS, 'cvSSý nEpi ti v n'66LV, TTqR Tä veep ovpavöv ýýToI arýSTc xai npayµaTsvoµevrjS ' AptaioTEkov Sö Eni ...npcßßsüci µßv ovv Thv HWt uvos xai av xai Taig Tovtwv ptXoao4iatS eanovöaKEV" EMJF-t Sý Kai Tä Tovtwv 6vyypäµµata. oiScv öaa nepi uhv (yToäv EýtXo664rjaav Zi vcavaSTe Kai, KXeävOat, xep66vaS. ETi Sý Xpüßinnoi TE xai öaot Thv äa,oyov npaýty 7 öyc!) Koagr aavmq, 9ý kapaS EiS ETEpav nv ýavTwv ýO oßoýiav g6'[F-K6c ijcFav. oLSE Se xaI. TÖS xotV&q ' E? ctKoi you TE Kai OT11ioKpi'rou ivot SÖýaS Kai Öv Tp6nov £xe nepi Tfg 4'Ü sew; £4aVTaGOTj6aV' Kai hfl V ' AKaSTj'. Kai Texq tiaS anacaC E1tiGTaTal, KaL TT v EK AuKdIov µtxpt Ai(3in' c TE xai Kvpijvr S ao4iav äno8iWAua6av 8

So, at this time he advanced to the study of things of importance. Now, the entire field of Philosophy is divided into three parts: the part that concerns actions [i. e. Ethics], the part which deals with nature, and the part which seeks and is engaged in the things which beyond the heavens.. then he the doctrine Plato are .so revered of and Aristotle and was zealous with regard to the philosophies of these [philosophers]; and he became acquainted with their works. He gained knowledge of as many things as [the] Zenos and Cleantheses philosophised on the site of the porch [i. e. the Stoa], and further, [the] Chrysippuses, and those who by adorning irrational behavior with reasoning adapted their own philosophy from one school to another. And also he knew the common opinions of Epicurus and Democritus, and the way in which those [philosophers] formed mental images on the subject of nature; and what is more, he learned of all the Academies, and the wisdom which spread abroad from the Lyceum as far as Libya and Cyrene.

7Saffrey and Westerink, TheologiePlatonicienne, 1.xl 8Himerius, Or. 48.20-25 (Colonna). In both the Himerius text and the following Julian text, words supplied for the sense are enclosed in square brackets. There is a puzzle in the titles used for the last two philosophical schools. There is no question that rl Eic Avxeiou aopia indicates the Peripatetics, bearing the original link to Aristotle's school in the Lyceum, and presumably the parallel reference to the physical setting of the Old Academy yields the unorthodox title ai Axaöi fiat, 'the Academies'. Yet long before the fourth century 'the Academy' tended to denote the sceptical Academy, and the term 'Academic' was used for representatives of the New Academy. On the other hand, the term 'Platonist' referred to Platonic philosophers among the Middle Platonists and Neoplatonists, that is, from the second century forward, vd. John Dillon, "Self-Definition in Later Platonism, " Jewish and Christian Self-Definition, Vol. 3, eds. B. F. Meyer and E. P. Sanders (London: SCM Press, 1982): 60-75 at 61; 68-69. Probably, Himerius is referring to earlier Platonists such as Plutarch, Albinus, Atticus, and Numenius.

9 How much philosophy did Hermogenes learn? It would be easy to overstate one's guess. A similar list of philosophical highlights with perhaps a better claim to represent contemporary philosophical education appears in a fragment of one of Julian's letters, where philosophical studies are recommended for the sake of moral purity. Kai örro S Eiöevai ExrIS ö ßovXoµat ýpä(EtV, iEpoµevoS TtS n to ApXiXoxov ÖIVaytV(w(TKET6)j. ATE' IltltuSvaKTa RATE äXXov Ti.vä TOW Tä iv Totauta ypa46vTCov.... Kai icävTwS 1tpEnoi. S' äv ý ý ýO oao4ia µövrl, Kai 'rovTCwv of 6E0VS i yEµövaS irpo6TrjaäjEVOt TrrS EautUOv 7tatSEiaS, ömcp IluOayöpas Kai IlA, ä'o v Kai' Apu6To c9?ojS of TE äµßi Xpi cnmmov Kai Zi vwva. 7Lpo6EKTEOV JEV yap oiTE iwc tv OSTE ToiS nävTwv 56yµa6ty, ä?, EKEivotS µövov Kai EKEivwv, 6 Ga E'ÜaEßELaS E. tI rtotlITnKä Kai 8L86(6KEi iEpi OP-C0V.... nt' E1LLKo pEtoS Ei6LTo0 X6yo; LATE fluppwvstoS" r'jSij µ9v yap Ka? cw irotoüvTES of 060i Kai ävrJp1jKaaty, wa'E EittXEi2cEty Kai Tä zrXEi(YTa Tc,)v 3t 3? Icov. 6µ(0q of öev KwV)Et Tünou xäpwV EmtµviJGOjVat gv Kai TovTwv, örtoiwv xph iä? testa Toi icpeag ämexcßQai. Xöywv, Ei Sý ?,öywv, itoXv irpöTEpov Evvotwv. 9

And to the end that you are able to understand that which I would like to discuss, let no one who is consecrated read Archilochus [the iambic and elegaic poet] nor Hipponax [the iambic poet] nor any other [representative] of those who write such stuff.... But above all let philosophy alone be suited to us, and of [the philosophers], those who have promoted the gods as guides of their learning, which [is the approach of] Pythagoras, Plato, and Aristotle, and the disciples of Chrysippus and Zeno. In fact, one should neither attend to all [philosophers] nor to the dogmas of all [philosophers], but only to those [philosophers] and [the dogmas] of those [philosophers], which are conducive to piety and who teach about the Gods.... Don't let the Epicurean or the Pyrrhonian doctrine enter; seeing that already the gods have done well and have brought destruction, with the result that the greater part of the books have actually been wiped out. Nevertheless, for the sake of general knowledge there is nothing objectionable in taking thought of even these [books], which sort of discourses priests ought most to avoid, and if the discourses, much more so the thoughts. Here Julian almost sounds more Christian than the Christians. The passage is another standard list, yet surprisingly the Stoics receive honorable mention. Certainly the mention of Pythagoras, Plato, and Aristotle would be expected for an adherent of lamblichean Neoplatonism like Julian. The most straightforward explanation, in accordance with the moralistic tone of the passage, would be that the Stoics were known to Athenian students such as Julian for their ethical doctrines, which were learned in some predigested form. The highly interesting

9julian, Frag. Ep. 385.15-386.24(I-lertlein)

10 (and ironic) note of the destruction of Skeptic and Epicurean texts suggests that Julian had knowledge of collections of philosophical texts in the later fourth century; l° it is quite possible that philosophical texts were carried by Julian and his entourage during his tenure as emperor. In the case of Basil, studies in rhetoric were pursued in three phases: Before his arrival in Athens, Basil studied rhetoric in Cappadocian Caesarea and then, at least in part, under the famous sophist Libanius in Constantinople? 1 There followed a more extensive period in Athens from which there is some record of study with Prohaeresius and Himerius. Fortunately, there is a body of material, including the surviving orations and letters of Libanius, which yield some helpful information about the schools and instruction in rhetoric. On the Athenian masters, there is Eunapius, whose Lives of the Sophists emphasizes the eminence of Prohaeresius: His vita is discussed more than any other sophist, he was himself the teacher of Eunapius, and he was the most distinguished student of the famed Julianus. 12 There is no evidence of any particular interest in grammar for this sophist, although it is clear that he was considered highly competent in rhetoric for his time, and seems to have attracted the lion's share of students including Christians such as Basil and Gregory. 13 However, as will be shown below, the case of Libanius is quite different. In his Funeral Oration for Basil, Gregory of Nazianzus describes his friend's attainments in the disciplines which were represented in the education of young elites. Ilotov µ6V EiöoS ovK PETETI6c IEOR56ic e« g Tic µ6v prITOpudiv ... i 'roao'UToc, 'ChV 7['Up6q J.i£Voc JLVEloi aV, si, Kai to Oos airrc gh Kath pAiopac njv; Tic R ypauµatitKrjv, f yXwaaav eýEXXTJViýEtKai iatopiav

'°Simplicius, In Cat. 334.2-3 (Kalbfleisch), cited by Catherine Atherton, The Stoics on Ambiguity (Cambridge: Cambridge University Press, 1993): 176 records a similarly melancholy observation that by his time (the sixth century) most Stoic texts had been lost. 11Paul J. Fedwick, "A Chronology of the Life and Works of Basil of Caesarea, " Basil of Caesarea: Christian, Humanist, Ascetic, ed. Paul J. Fedwick (Toronto: Pontifical Institute of Mediaeval Studies, 1981): 3-19 at 5 places these initial rhetorical studies in Cappadocian Caesarea from 346- 348 and in Constantinople during the period 348-349. Fedwick (6) then dates the education in Athens from 349 or 350 to 355. Rousseau, Basil of Caesarea, 37 briefly presents the evidence that Basil actually was a student of Libanius in Constantinople, and that later in life Basil himself acknowledged his influence. I2Eunapius, Vit. Soph.9.2.1ff. (Giangrande) 13Robert J. Penella, Greek Philosophers and Sophists in the Fourth Century A. D. Studies in E1unnapiusof Sardis, ARCA 28 (Leeds: Francis Cairns, 1990): 85-94. Penella (92-93) discusses the question whether Prohaeresius was a Christian, which is challenged by Fritz Schemmel, "Die Schulzeit des Kaisers Julian, " Philologus 82 (1926): 455-66 at 461.

11 6uväyct Kai phipoig bntaTaT£i Kai voµo9£TEi ztoiiij a6ty; Tic ö- 4tXo6o4iav, Ti v övTCAS vTTJ AV TE Kai ävcO ßaivouaaV, 6011 TE rtpCCKTtK I Kai OEwprJ'r K, 60 -fl n£pi Täg XoyucäS änoSEiýEtc fi 6cvit8e6£tq EXEI Kai Tä itaXaißp. aTa, fv Sn ötaX£xtitid v övoµä(ouaty, hS päov Eivat XaßupivOouS 816ýC OEi fl k£ivou Xöycov äpxuS 5ta4uy£iv TovS I V iäS iwv EL ' ComoU 8E1j6£LEV;14

What learning did he Who area of not master? .... was so eminent in the rhetorical art, which exhales the power of fire, but without the bad character of orators? Who so [skilled] in the grammatical art, which hellenizes the tongue, composes narrative, attends to metrical' matters, and passes judgement on poetry? Who so [competent] in philosophy, [the knowledge] which is exalted and conducts upward, partly concerned with action and partly contemplative, and part of which involves contests in rational demonstration or rebuttal, which is called the dialectical art, with the result that it would be easier, if one should try, to find a path through the labyrinths than to escape the snares of his arguments? Then Gregory describes Basil's mastery of astronomy, geometry, arithmetic, medecine, and lastly, most importantly, ethics. It might appear that Gregory is merely ringing the changes of the standard 'circle of humanistic learning' (EyyiXia Tud8cocn.), which had by Late Antiquity become a traditional feature of education (as will also be seen below for Origen). 15 But it is highly unlikely that Basil's knowledge of grammar, rhetoric, and dialectic began and ended with the 'circle'. There is also the use of philosophical textbooks, handbooks, and compendia. The use of philosophical textbooks seems to have been a development of the Hellenistic age. A primary example of the codification of philosophical positions and doctrines in the rival Hellenistic schools is the composition of Stoic textbooks during the course of the third century BC, a natural step given the massive corpus of writings left from the

14Gr. Naz., Or. 43.23,6-24 (SC 384: 172-4) 150n Basil's early education in the 'circle of learning', Gregory speaks at Or. 43.12,14-17 (SC 384: 140): d( T'v 8yKÜKXI0V 1Lai8EU6nV JCatSEUÖI. IEVOS Kal 0eoaE0ctV Eýa6KOU11EVOS Kai, auvE2,6vTI Odval, rtpög Av 1JEXXO1aav tE?,Etöt1Ta Slä cwv El; äpXricµa0rlµ6. tü v ccyo}LEVOS.It must have still been the norm for young aspiring orators to undertake preliminary studies in grammar, followed by the other important fields of knowledge in the 'circle of disciplines', as described in the first century by Quintilian, 1.10.1. Theon of Alexandria, a contemporary of Quintilian in the first century AD (OCD, 3rd ed., s. v. ) or perhaps a second-century rhetorician, requires some philosophical studies before study of rhetoric, Prog. 59.1ff. (Spengel), and comments on the role of the 'circle': "But as things are, the great majority hardly attend to such logoi to such a degree, that not evendipping into any of the so-called circular disciplines they hasten towards oratory" (vvv äe Oi nXEiovS To60UT0V ÖEOUat T(UV Totob ov X6ycwv Eitatev, Ci6TE Ol1SE TCOV EyKUKXicOV Ka%ou}IEVCUV i'yEIV). ita011tt0 «oV OTLOUV ttETC0,0t106vOVTES aTTOU6lV E711 To

12 early Stoics. 16 In fact, it has been argued that much of an early compendium of Stoic ethics is preserved by later doxographers. 17 As for grammar and logic, it is possible that the Stoic-influenced work preserved under the name of Dionysius Thrax was the first systematic textbook of grammar to be written, an important document for the origins of technical grammar in Antiquity; 18there is testimony to Imperial use of introductions (Ei6ayc.'yai) and outlines (vnotivn(6aetq) in Stoic logic. 19 It is thus likely that such works dealing with 'the dialectical art' were studied by Basil and Gregory in the course of their studies. With respect to grammar, it is important to recognize two basic levels in the teaching of this literary art. Grammar in some contexts is just the learning of basic literacy, rudimentary skills in reading and writing Greek. But from the Hellenistic age onwards, talk of grammar is usually talk of the much more sophisticated Alexandrian tradition of philology, and an important part of this sort of grammatical study is the technical grammar represented by the famous treatise of Dionysius Thrax. It is this technical sort of grammar which corresponds most closely to the traditional grammar known to the history of linguistics. The idea is that both levels of grammar are relevant somehow at different stages of Basil's literary education 2°

16Manfred Fuhrmann, Das systematischeLehrbach (Göttingen: Vandenhoeck & Ruprecht, 1960): 154. 17Fuhrmann, Das systematische Lehrblech, 154 agrees with the suggestion of other scholars that Chrysippus himself had prepared such a work, titled' Tnoypaýh coü a.öyov toi [h0txov1 np6; GEÖrtopov, whose content is reflected to some degree in Diogenes Laertius 7.85-116 and Arius Didymus ap. Stobaeum 2.57.13ff. (Wachsmuth). 18Fuhrmann, Das systematische Lehrbuch, 145ff.. There have been serious challenges to the attribution of the c9Xv7jattributed to Dionysius Thrax, which is considered in most recent scholarship to have been written towards the end of the second century BC. However, there is general agreement that at least some initial portion of the treatise was composed by Dionysius the student of Aristarchus and fellow student along with the important grammarian Apollodorus of Athens. Richard Janko, "Crates of Mallos, Dionysius Thrax and the Tradition of Stoic Grammatical Theory, " The Passionate Intellect. Essays on the Transformation of Classical Traditions Presented to Prof. I. G. Kidd, Rutgers University Studies in Classical Humanities, ed. Lewis Ayres (New Brunswick and London: Transaction Publishers, 1995): 213-233 at 213-216 gives a brief but useful overview to the complicated questions involved. Recent studies on the work attributed to Dionysius recognize the influence of the Stoics as well as the purposes and methods of Alexandrian philology. 19jonathan Barnes, Logic and the Imperial Stoa (Leiden: E. J. Brill, 1997): 73-74 note 200 gives some references. There is additional information on Varro and other sources in Fuhrmann, Das systematische Lehrbuch, 166-168. 20Michael Frede, "The Origins of Traditional Grammar, " Historical and Philosophical Dimensions of Logic, Methodology, and Philosophy of Science, eds. R. Butts and J. K. K. Hintikka (Dordrecht: D. Reidel, 1997): 51-79 at 52.

13 Comparisons with Basil's intellectual and spiritual peers such as Gregory of Nazianzus and Diodorus of Tarsus add details but also confirmation to the picture as sketched. Gregory speaks of his own rather extensive course of studies in Athens, Berytos, and Alexandria, which sheds light on the higher levels of grammatical and rhetorical schooling experienced by the two friends. '92 näip-p, ev no6Ecw, ti Owv Kpäzoq, ävii vu narrow. Ka?, öv pEv pAcpric nupoev µ.Evog, Ev ti äyopljaty, "Ev 're 5tKaanAblut, Kalt C'Ü Ag tGt ÄÄyoic tv. Ka?,öv S ioiopbjq 4pýv CµnXEoq t,atiopir yap Euµtept i 6o4bj, noA.Xwv voo; ovK Wyov R Fpaµµattidi ýüouca X, yov, Kai ßäpßapov fxi'v, wßair Enixoupoq c E?cXäöo; eüyeveog y?, piGti-q- Kai A,oynKdq 'r9xvrlc -rä nakaI icna, ois vii öckijOE Kpvnmrat, ö tipt4 fti 8ý ?. yog 7Ecpi4avtov EOrlKEV"21

O father, for one thing I yearn, power of speech, before all other things. A noble thing is the fiery power of rhetoric, in deliberative, in forensic, and [in] epideictic speeches. And fair is the intellect skilled in narrative; since narrative is the keeper of wisdom, a knowledge of many things; yet no minor [pursuit] is the grammatical art which refines speech, and the foreign utterance, the peerless ally of the noble tongue of Greece; and the contests of the logical art, under which truth is concealed, which argument when practised presents in perfect clarity. Gregory mentions each of the three basic forms of oratory in Antiquity, and was presumably trained in each category. Grammar is said to be concerned with the purity, or rather the linguistic correctness of Greek, which corresponds to an understanding of grammar observed in Basil's theological arguments. 22 It might seem surprising that Gregory speaks in both texts of dialectical contests (iä itaXaiagata), because of the predominance of ethical concerns in philosophy of the Imperial period. But Gregory may provide some evidence that at least some aspects of logic were studied in the fourth century in the major centers of learning. Frederick Norris has recently shown convincingly that Gregory knew

21Gr. Naz., Cnrnt. 2.4 (PG 67,1510) 22The discussions in the following chapters, "Relatives" and "Prepositions and Conjunctions", will try to show this.

14 some philosophical rhetoric, involving Aristotelian enthymemes, Stoic logic, and perhaps even some Aristotelian syllogistic-23 Christoph Schäublin argues that another distinguished Christian who studied in Athens at roughly the same time, the Antiochene exegete Diodorus of Tarsus, benefited from instruction in philology, as can be seen from his exegetical works. 24 Schäublin rejects as one-sided the rather implausible view of previous scholarship that Diodorus was influenced primarily by the techniques and approach of the Aristotle-commentators in Athens. This would seem to be a misunderstanding of study in fourth-century Athens. Rather, he recognizes the importance of the EyK1K? 1. Os itaISsi cnS,compares with the studies of Gregory of Nazianzüs described in his poem, and argues that the time-honoured system of philological instruction known from the scholiasts to Dionysius Thrax still served as the basic outline for studies of literature in the fourth century. In his view, grammarians of the later Imperial period relied on commentaries which contained in some form much of the earlier Alexandrian learning, 25 so in this way the philological tradition was kept alive. This cautious approach can hardly be incorrect, but further research has supplemented and altered scholarly understanding of Imperial grammar. This traditional system of school-philology, which the scholiasts characterize as 'ancient', seems to have been used to structure the training of youths in the grammatically informed study of literary texts? 6 It divides

Frederick W. Norris, Faith Gives Fullness to Reasoning. The Five Theological Orations of Gregory of Nazianzus (Leiden: E. J. Brill, 1991): 17-39. Barnes, Logic and the Imperial Stoa, 58; 75 discusses rhetorical arguments and the study of enthymemes. Barnes contends (71) that instruction in logic in the Imperial period (up to Marcus Aurelius) normally included study of both Stoic hypothetical sylogistic and Peripatetic categorical syllogistic. 24Christoph Schäublin, Untersuchungen zu Methode und Herkunft der antiochenischenExegese (Köln and Bonn: Peter Hanstein Verlag, 1974): 43-44. For Diodorus' knowledge of various disciplines, Schäublin (44 notes 4-5) cites Photius, and also a letter of Julian, Ep. 90 (Bidez, 174.26ff.) = Julian Ep. ad Photinum ap. Facundum, Pro Defensione Triton Capitulorum 4.2 (PL 67,621A-C), which commends his learning. 25Schäublin, Untersuchungen, 31-36: "Indessen stützen sich die Grammatiker auf vorhandene Kommentare, und in diesen waren, trotz steter Verdünnung, immer noch ansehnliche Reste alexandrinischer Gelehrsamkeit gespeichert. " On the structure of later philological instruction, there is the influential contribution of Hermann Usener, "Ein altes Lehrgebäude der Philologie, " Kleine Schriften, Bd. 2 (Leipzig and Berlin: B. G. Teubner, 1913): 265-314. This scheme can be understood to present a systematic model for exposition in the main areas of application of philological skills, whatever the content and genre of the text. But as has been pointed out, in use it is present only in the loosest sense. 26Sch. in Dionys. Thr. 12.3ff.; 135.6ff. (Hilgard); cf. Varro ap. Diomedem, Ars Gramm. 426.21ff. (Keil)

15 grammar into four parts, in contrast to Dionysius Thrax who divides into six parts. These parts are comprised of the recital or reading of the text (ävayvw6'ci.x6v), the exposition or interpretation of the text which includes historical, cultural, chronological, geographical, and biological points (E 1 TJTtKv), the correction of the text i. e. textual criticism (&opO(oTtxöv),and the assessment of the text i. e. moral and literary criticism (xptitxdv or xpiatq

TLot1 ,th'r(ov) 27

The second and the fourth items are of particular interest. The scholiasts also talk about the four tools (öp-yava) of grammar, including the explanation of words (yXw6arjµatnxöv), meter (ieipucöv), historical and chronological matters (i.aioptKÖv), and 'technical grammar' i. e. grammar in the narrower sense (iexvtx6v). 28 Now it has been argued that these four grammatical tasks were standardly addressed by Imperial grammarians under the division of grammar that takes up the exposition of the text (ýýrjyryiuKÖv)29 Unfortunately there is no direct evidence for this attractive view. Research in Origen's bible-exegesis along the lines marked out by Schäublin shows an interesting avenue for reception of Stoic philosophy in the clarification of words (yX.o cntKOv)30 Careful examination of Origen's commentaries shows that he is heavily indebted to ancient lexicography from at least three basic sources, a Stoic doxographical lexicon of concepts, a lexicon of synonyms such as that of the grammarian Herennius Philo (second century AD), and a general dictionary like that of

27Fuhrmann, Das systematische Lehrbuch, 145-146 note 4 thinks that it is impossible to determine when this schema was introduced into Alexandrian philology, and is content to mention notices of several grammatical works earlier than the treatise attributed to Dionysius Thrax. Schäublin, Untersucltnngen, 35 suggests that this list of tasks does not necessarily represent instruction in the interpretation and criticism of texts, much like the theoretical definitions and divisions of grammar may at times only reflect the authority of Dionysius Thrax; rather, school-philology was the product of the traditional structure of philological training. But no evidence or further argument is given for this view. Perhaps the scheme is only a rough guide from which deviations and omissions were part of the practice of philology. 28Sch. in Dionys. Thr. 170.18-20 (Hilgard). The parts and the tools of grammar are not distinguished in the classification of Dionysius Thrax. 29Bernhard Neuschäfer, Origenes als Philologe, Schweizerische Beiträge zur Altertumswissenschaft (Basel: Friedrich Reinhardt Verlag, 1987): 139-140; cf. Usener, "Ein altes Lehrgebäude der Philologie, " 270-272. Neuschäfer supposes that the difference between the parts and the tools of grammar is simply between what grammar in theory involves with a view to gaining scientific knowledge and how grammar works in the practice of textual exegesis. In fact he uses the four instruments to structure his discussion of the Eýijyiiitxöv and its relevance to Origen's exegesis. Neuschäfer and many others note that there was much disagreement in the grammatical tradition over the divisions of grammar. 30Neuschäfer, Origenles als Philologe, 140ff.; 289

16 Diogenian. This insight could prove valuable for study of the homiletical works of Basil and Gregory. 31 The part of grammar which involves literary criticism is significant from a logical point of view. It was considered superior to the other parts of the art of grammar 32 and at least in the hands of the Stoic philosopher and grammarian Crates of Mallos there seems to be a strong relation to logic 33 Schäublin contends that judging by the increased Pergamene influence on the exegetical Scholia to Homer and occasional mentions of Crates, the Pergamene school provided some input to the practice of philology in Late Antiquity 34 Of course, Alexandrian philology was still dominant. Schäublin sensibly suggests that there was a reconciliation, in some of the leading grammar-schools of the Imperial period, of the rival Alexandrian and Pergamene factions headed in the 'golden age' of Hellenistic Homer-philology by Aristarchus and Crates respectively. There is even reason to believe that the school of Aristarchus at some point recognized the superiority in grammar of the Pergamenes, heavily conditioned by the Stoicism of founder Crates and his students. 35 By this route, the school of Crates may have contributed to later Imperial school-philology some of their

31Neuschäfer, Origenes als Philologe, 150 note 95 (406-407) even refers for comparison with Origen's Stoic-colored definition of 56 a to Basil of Caesarea, In les. 10.242 (PG 30,544C4-6), which is a work of disputed authorship. By his lights, Basil's definition "eindeutig von einem Synonymenlexikon abhängt. " On doxographical works as used by the Antiochene bible-exegetes of the fourth and fifth centuries, there is discussion by Schäublin, Untersuchungen, 48; 151ff.. 32Dionys. Thr., Ars Gramm. 6.2ff. (Uhlig); Sch. in Dionys. Thr. 12.13ff. (Hilgard) 33Crates was recognized in later grammar for his self-designation as 'critic' as distinguished from the mere 'grammarian', Sextus Empiricus Math. 1.79, for the 'critic' must be well-versed in the science of logic (Kai E6v µev KptctKÖv rtäans, Qnai, 5F--t Xoyucij; entatrjµnS eµrcetpov Elvat). Neuschäfer, Origenes als Philologe, 248-250 suggests that the importance of 'assessment' to Crates also involves the relevance of logic to the fourth part of grammar, if the designation 'critic' means the same kind of 'critical' activity as the 'criticism of literary works'. In this case, there would be a language-technical, logical aspect to Crates' 'criticism' which examines how language expresses logic (Xoytxä Ae(opitata). Whatever its nature, this logical emphasis is not found in the 'assessment' of Dionysius Thrax, but it is hard to tell from the slim evidence. 34Schäublin, Untersuchungen, 35-6 has several references for this, although none from the Antiochene bible-exegetes. His argument for the (Christian) Antiochenes is that they were heavily influenced by the professional school-philology of their day, which probably did not recognize much of a distinction between the rival Hellenistic factions; Schäublin never attempts to isolate a distinctively Pergamene contribution to the Antiochene school of biblical exegesis. Schäublin argues effectively against the view of some scholars that the Antiochene school emerged in reaction to the Alexandrian school of bible-exegesis, rather it grew out of the ashes of ancient philology. 35Rudolf Pfeiffer, History of Classical Scholarship(Oxford: Clarendon Press, 1968): 245; 270. Pfeiffer (237) says of Crates that "in contrast to the Alexandrian scholar poets he and his pupils approached the literary heritage as philosophers, and in particular as orthodox Stoics," although the influence of Alexandrian scholarship was strong in his school.

17 interests in logic and technical grammar. 36 But Schäublin concedes that this school was particularly susceptible to a rather banal moralism, and that their formal and aesthetic categories tended to be drawn from rhetoric as was characteristic of the Imperial period. The question of the Pergamene grammatical inheritance in third and fourth century schools is of interest to all of the studies in fourth-century Patristic texts which follow. In the following section on Origen, there will be an evaluation of Bernhard Neuschäfer's arguments regarding the dissemination of Stoic grammar in the Imperial period.

II

On the topic of philological training in the fourth century, there has been some very helpful work done recently, starting with the seminal researches of Christoph Schäublin on the Antiochene school of biblical exegesis 37 But most significant for the study of Basil and other Patristic writers of the later fourth century is the consideration of Origen, on account of the powerful influence of the third-century biblical commentator and exegete on his Greek successors.38 Fortunately, such studies have been greatly advanced by Bernhard Neuschäfer's seminal work, influenced by the contributions of Schäublin, on the philology of Origen's biblical exegesis39 Despite the closeness of their intellectual relationship and the interesting parallels in their careers, there is this fundamental difference: Prior to the ecclesiastical and theological phase of their careers, Origen taught as a grammarian in Alexandria, while Basil (probably) taught rhetoric in Caesarea before pursuing the ascetic life in Pontus. Now much can be written about what is known of their knowledge of each other's field of expertise. And as will be

36Schäublin, Unterstucluingen, 36: "Wohl aber werden die sonstigen Forschungsergebnisse der Pergamener, deren Interesse mehr den 'Sachen' als den 'Worten' galt, im Unterricht nachgewirkt haben. " On Crates and the far-reaching influence of the Pergamene school which he founded, there is Pfeiffer, History of Classical Scltolarsltip, 238ff.. There will be more on these points below in discussion of Origen. 37Footnote 7 contains the reference. 38Karl I-loll, Amphilochins von Ikonium in seittetti Verhältnis zu den grossen Kappadozieren dargestellt (Tübingen-Leipzig: 1904): 162-163 makes some interesting remarks on the greater influence of Origen on Gregory of Nazianzus than on Basil of Caesarea.

18 Basil learned far less grammar and logic than Origen. But it is seen, probably his Alexandrian's hard to deny that study of the great works amounts to a his knowledge of the tradition. considerable supplement of ancient grammatical So it is worthwhile to look at Origen's philological training, his apprenticeship in Imperial school-philology. Neuschäfer presents extensive argument that on the one hand Origen for provides valuable evidence grammar, rhetoric, and textual criticism in the later Imperial hand schools of the period, while on the other there is abundant testimony in the received Origenian corpus in confirmation of what is known from other sources regarding philological studies in Alexandria and other centers of learning such as Athens. These conclusions agree with the reports sketching Origen's extensive grammatical and philological training in the Alexandrian schools. Eusebius, the eminent fourth-century theologian, ecclesiastical historian, and confirmed Origenist, says of Origen's early studies that he inherited both a secular, classical education (EyK1 acc, O. A.rlvixä µaO1gcrra) as well as sacred learning (icpä nat8euµaia) ýQ As part of the standard educational curriculum enjoyed by wealthy youths, the 'cycle of humanistic learning' (ýyKÜKXta natöeia) was learned, based on the earlier Greek works of the Hellenistic period, in overview of a 'circle' of standard disciplines written up by a single author. Probably the best known example of such a presentation is Varro's influential Disciplinarnm Libri, which contained books devoted to Grammar, Dialectic, Rhetoric, Geometry, Arithmetic, Astronomy, Music, Medecine, and Architecture. 41

Now it is certain that Origen also learned at an early stage the elementary skills of reading and writing under a ypaµgattorY c, and then moved on to learning with the tutelage of a 'ypaµµaTix6S basic philological skills while acquiring a familiarity with Greek literature. Further study would then normally follow in rhetoric. Origen supports this basic sequence of instruction in one

39Bernhard Neuschäfer, Origenes als Philologe, Schweizerische Beiträge zur Altertumswissenschaft (Basel: Friedrich Reinhardt Verlag, 1987). 40Eusebius, Hist. 6.2.7-8 (Schwartz, 520.24-8). On the reliability of Eusebius' report, Neuschäfer, Origenes als Philologue, 32 note 153-154 (331) evaluates what is reported on the basis that this information is passed down by ecclesiastical tradition. 41Fuhrmann, Das systmiatische Lehrbuch, 163

19 homiletical text, which Neuschäfer thinks could be of Rufinian origin, giving a sketch of the studies of elite youths. Sunt multa et diversa studia litterarum in hoc mundo, tit videas quamplurimos incipientes a grammaticis ediscere carmine poetarum, commoediarum fabulas, tragoediarum vel commentitias vel honorificas narrationes, historiarum longs tic diversa volumina, tum deinde transire ad rhetoricam, atque ibi omnemfucum eloquentine quaerere, post haec venire ad philosophiam, perscrutari dialecticam, syllogismorum nexus inquirere, mensuras geometriae pertentare, astrorum leges ac stellarum cursus perscrutari, omittere quoque nec musicam; et sic per omnes istas eruditi tam diversas et varias disciplinas, in quibus nihil de Dei voluntate cognoverunt, multas quidem, sedpeccatorum divitias congregaverunt.42

There are many different studies of letters in this world, with the result that you see such a crowd starting to learn from grammarians the ballads of the poets, the tales of the comedies, the stories of the tragedies whether bringing shame or honor, the long and various books of histories, and then moving on to rhetoric, to seek there every deceitful trick of eloquence, after these studies arriving at philosophy, to be immersed in dialectic, to look into the constructions of syllogisms, to attend to the measures of geometry, to examine the laws of the movements of the stars and heavenly bodies, and also not to leave out music; so it is that those who are erudite in all those disciplines, so varied and multiform, have come to know nothing of the will of God. They have gathered wealth, indeed in great quantity, but only of vices. Origen supports here the general observation that in his day education started with elementary grammatical studies with its literary emphasis, which prepared the way for instruction in rhetoric. More specialized and less common would be the work in philosophy, with prominence given to the study of logic; the standard disciplines of geometry, astronomy, and music round out the catalogue. Here there is no tripartite structure of Physics, Logic, and Ethics standard in Late Antiquity in accordance with the earlier Stoic model. However, Origen certainly tips his hat to the study of logic, which is not at all surprising in view of the common prominence of dialectic in the ancient EyK KXLOSitat&eia or byK1KXia µaB1 atia43 by chance, it does not appear in the most important extant Origen

42Origen, Hom. 3 in Ps. 36 (PG 12,1341B-C). This text is adduced by Neuschäfer, Origenes als Pliilologue, 33 note 160 (332-333). 43H. Fuchs, "Enkyklios Paideia, " RAC 5 (1962): 365-398 at 372; 369; 377; 386; 392 notes that dialectic had a firm and eminent place in the standard works of the 'circle of disciplines', sometimes even termed 'philosophy itself'. Philo, Agr. 136-141 (Wendland, 122.6-123.11) = FDS 695 (part) preserves a quick run through some of the highlights of Stoic dialectic, mixed with the divisions of other disciplines.

20 passage on the topic in his Ep. ad Gregore1 44 In Origen's description of studies, the last three disciplines seem to be loosely tacked on to a plausible description of a (somewhat idealized) course of studies in an intellectual center such as Alexandria. Eusebius records that advanced studies in grammar and philology were undertaken by Origen, eventually leading to a successful career teaching grammar after his father's death as a martyr 45 1rpoaxOEic S' uno 'oü itaipöS Ev iotS' EX? vwv µaOijµaßiv EKOuµöTEpov 're [Kai] gerä tiýv Eisivou tiF-Äcut v irk nepi toi S XöyouS äc is et ÖXov Eittöoi S Eau'cöv, wS ieai. lapaaxsuýv ý7t1 'cä 'ypaµµatitKÜ µeTpiav Exetv....

And [Origen], having been been first guided by his father in the studies of the Greeks, devoted himself entirely after his death to the rigorous exercise of philology with greater zeal, with the result that he attained considerable distinction in the field of grammar.... Later, Origen gave up his secular career and moved into Christian catechetical teaching, as early as the year 211.46 At this time he sold his extensive personal library of classical literature, in a decisive shift of vocation which cast a long shadow over his later fourth-century admirers and students. 47 Given this background, Eusebius calls Origen the pioneer (mpc oq evpeiricl of grammatically skilled bible-exegesis 48 Yet as Neuschäfer points out, Origen's training in the schools of his day does not make him a direct product of the golden age of Hellenistic 'Homerphilologie'. "[Es] liegt vor allem darin, dass durch sie [i. e. the

44Origen, Ep. ad Gregorem = Phil. 13,1-18 (SC 148: 186-188). Origen names the disciplines of geometry, music, grammar, rhetoric, and astronomy. As representatives of the artes liberales, he names geometry and astronomy. Cf. Cels. 3.58. Neuschäfer, Origenes als Philologe, 156ff. takes up the role of the 'circle of humanistic learning' in his biblical exegesis, as well as Origen's view of the relation of Greek philosophy to Christian faith with its parallels in Clement of Alexandria and Philo. Perhaps Origen was influenced by Platonic and middle-Stoic notions of the 'circle of learning' as preparatory to study of philosophy, as Neuschäfer (157 note 131 (412)) suggests, cf. Philo, Congr. Erud. Grat. 11; 79 (Wendland, 74.9-13; 87.19-88.4), the first text of which ends a list of fields with c1-1iXX Xoytxrl 9ewpia ndmj, which cannot mean logic; or Origen (162 note 161-163) may have recognized the inferiority of the elementary 'circle' to the philosophical fields of metaphysics, physics, ethics, and logic, although Origen never gives notice of such views in the extant Greek texts. In general, Neuschäfer (202; note 455) is so impressed with Origen that he compares his polymathic achievements and influence to Posidonius in the practice of 'die historische und naturkundliche Sachexegese'. 45Eusebius, 159 Hist. 6.2.15-3.1 (Schwartz, 524.1-6). Neuschäfer, Origenes als Philologe, 32 note Ep. (332) and other scholars take this report to be informed by Origen himself, in part by the lost ad Alexandrtutt. 46Neuschäfer, Origenes als Philologe, 32 note 161 (333-334) 47Eusebius, Hist. 6.3.9 (Schwartz, 526.19-23) 48Eusebius, Hist. 5.28.13-19 (Schwartz, 504.11-506.19)

21 testimony of Eusebius] jene Einrichtung bekannt ist, durch deren Vermittlung Origenes mit den Traditionen und Methoden der paganen Philologie Bekanntschaft schloss: der kaiserzeitliche Schulunterricht. "49 This central position of Neuschäfer's work is helpful for understanding questions of language and grammar in Origen, as well as in his fourth-century admirers. The highlighting of dialectical studies in the text cited above can in part be explained by Origen's own interests: Some studies of Origen show that he possessed a fairly impressive knowledge of dialectic as he found it represented in Alexandria, highly unusual for Greek Patristic writers 50 Although he does not draw much on Origen, Jonathan Barnes has argued the bold thesis recently that the study of logic flourished at this time 51 One of the major theses of Neuschäfer's work is that in Origen there is abundant evidence of the use of Stoic technical terms and Stoic definitions, as well as evidence of knowledge of Stoic semantic theory in texts such as the Philokalin 52 Neuschäfer attempts to explain the Stoic inheritance of grammar in Origen from deeply-rooted trends in Alexandrian grammar. 53 These changes, he thinks, may have been generated or accelerated when Chairemon the Stoic, whose work 'On Comets' (IIepi xopyT(Ov) Origen mentions, was appointed head of the Alexandrian grammar-school. As several scholars have pointed out, in the first century BC the Alexandrian Dionysius Thrax (whose identity has been questioned) shows an openness to non-Alexandrian traditions as can be seen by the influence of the Stoic 'rexvii icpi. ýcovijSon his tExvn of grammar. The surviving grammatical texts seem to confirm

49Neuschäfer, Origenes als Philologe, 35 50John M. Rist, "The Importance of Stoic Logic in the Contra Celsum," 64-78 in Neoplatonism and Early Christian Thought. Essays in Honour of A. H. Artnstrong, eds. H. J. Blumenthal and R. A. Markus (London: Ashgate, 1981); "Beyond Stoic and Platonist: a Sample of Origen's Treatment of Philosophy (Contra Celsttm: 4.62-70), " 228-238 in Platonismus und Christentum. Festschrift ffir Heinrich Dörrie, hsgb. von H. -D. Blume und F. Mann (Miinster/Westfalen: Aschendorffssche Verlagsbuchhandlung, 1983). Louis W. Roberts, "Origen and Stoic Logic, " Transactions of the American Philological Society 101 (1970): 433-444. 51Jonathan Barnes, Logic and the Imperial Stoa, 126-127 might be pressing a useful point too far: "Nonetheless, it seems to me beyond doubt that logic engrossed men during this period in the history of philosophy as it has rarely engrossed men in any other period. Moreover, the logic which engrossed was not the logic of the elementary handbooks--that sort of thing, after all, is essentially unengrossing. It was more subtle, more advanced, more technical. If much of it was based on exegesis or commentary and hence on books, it was none the worse for that. " Apart from the Patristic sources, there is a scarcity of evidence; more research on the topic would be quite welcome. 52Cf. Origen, Phil. 14 (Robinson, 67.25-70.5) on levels of meaning. 53Neuschäfer, Origenes als Philologe, 215ff.

22 the shift in the use of Stoic technical terms, which Neuschäfer thinks was present at least a century before Origen's work. 5455 In his conclusions, Neuschäfer posits two basic grounds for what he finds in Origen: The first is the exposure in the Hellenistic period of the proud tradition of Alexandrian philology to Stoic grammatical theory; and it appears that this development was more advanced at the time of Origen in the early third century than when 'Dionysius Thrax' and Apollonius Dyscolus wrote their treatises on grammar. This sounds plausible enough, judging from the surviving texts. But Neuschäfer takes the additional step of claiming a special affinity of Origen to Stoic grammar and indeed to Stoic philosophy of language, beyond what was mediated and transmitted by the grammatical tradition in Alexandria 56 Obviously, this claim is more tenuous than the argument for Stoic grammar seeping into Alexandrian school-philology. Although Neuschäfer is probably correct that it is broad shifts in Alexandrian school-philology and not the dissemination of Pergamene Homer-exegesis that is directly relevant to understanding Stoic material in Origen, the Pergamene school may have still contributed to knowledge of grammar and logic in the major intellectual centers of the Empire 57

III

As for Basil over a hundred years later, the evidence for his studies and accomplishments places more emphasis on rhetoric, less on grammar and philology. In fact, there is some evidence that Basil taught rhetoric upon his

S4Neuschäfer, Origenes als Philologe, 216-217: "Das Eindringen stoisch-grammatischer Nomenklatur nicht nur in das Lehrbuch des Dionysius Thrax, sondern auch in die Abhandlungen des Apollonios Dyskolos verdeutlicht, dass ein Jahrhundert vor Origenes bestimmte 'termini technici' und ihre spezifische Bedeutung, aus dem ursprünglichen Zusammenhang gelöst, längst grammatikalisch-terminologisches Allgemeingut geworden waren.... Nun fällt allerdings folgendes auf: Origenes gebraucht mit grösster Selbstverständlichkeit stoische Termini, deren Inanspruchnahme bei Dionsius Thrax und Apollonios Dyscolos noch umstritten ist. " 55Neuschäfer, Origenies als Philologe, 219ff. argues for the growing use of Stoic grammar in Imperial Alexandrian philology in part by adducing the influence of Stoic rhetoric on Alexandrian Homer-exegesis from 100 to 300 AD, but this can hardly help his case. 56Neuschäfer, Origenes als Philologe, 217-218; 155 note 120 57Neuschäfer, Origenes als Philologe, 219ff.: "Dieser Bereitschaft zur Aufnahme stoischer Einflüsse im kaiserzeitlichen Alexandrien scheint zumindest nicht hinderlich gewesen zu sein, dass diese philosophische Richtung einst entscheidend die Auslegungsmethodik des pergamenischen Gegenlagers geprägt hatte. "

23 to Cappadocian Caesarea.58 But it is likely that much can be learned return about from Basil's knowledge of grammar and philology consideration of what remains from Origen. Since the question of Basil's knowledge of Origen is important, it is to discuss the dispute Basilian the worthwhile scholarly regarding editorship of pliilokalia. The Philokalia is a surviving collection of texts which deal with interpretation of the Bible, drawn primarily from the biblical commentaries but dogmatic the Alexandrian also from the and apologetical works of great exegete. It is valuable for the reliability of transmission of the texts included as well as for the material that it adds much to what is known of works which are not Latin The Philokalia otherwise available, or are only extant in translation. is clearly an anthology, in Gregory's Ep. 115 there is the characterization of the work as EKXoyai, but the title is a novelty and the work seems to inaugurate a new literary genre 59 Like all Gaul, it is divided into three parts. The first fourteen sections are hermeneutical, the following chapters 15-20 are apologetical, and parts 21-27 treat questions of God and human free will. As Harl and Neuschäfer have shown, the first part of the Philokalia testifies to Origen's knowledge of the philological tradition, with its traces of earlier grammarians, Homer-exegetes, and Stoic philosophers. Traditionally, the collaboration of Basil of Caesarea and Gregory of Nazianzus in the editing and composition of the Philokalia has been upheld, but recently Marguerite Harl and Eric Junod have challenged this story with some success 60

58Rousseau, Basil of Caesarea, 40-41 reviews the evidence for this claim, which Fedwick, "Chronology, " 6 considers well-established. It is doubted by Jean Gribomont, "Eustathe le Philosophe et les voyages du jeune Basile de Cesaree, RHE 54 (1959): 115-124 at 121ff. on less than conclusive grounds, and defended by Giuseppe Lazzati, "Basilio di Cesarea insegno' retorica?" Stttdi e materiali di storia delle religioni 38 (1967): 284-292. 59Eric Junod, "Remarques sur la composition de la 'Philocalie' d'Origene par Basile de Cesaree et Gregoire de Nazianze, " Revue d'Histoire Pltilosophique et Religeuses 52 (1972): 149-156 at 150-1. "Il s'agit de la premiere anthologie de textes en prose d'un auteur unique oü les extraits sont reproduits sans etre systematiquement abreges et oü les indications de provenance sont donnees avec precision. " 60Marguerite Harl, Origene. Philocalie 1-20. Stir les Ecritures, Sources Chretiennes 302 (Paris: Les Editions du Cerf, 1983). Despite the arguments of Harl, the traditional attribution is accepted by Neuschäfer, Origeizes, 43. Before Hart's attack, Jean Gribomont, "L'Origenisme de Saint Basile, " L'Hontnte Devant Dieu. MelangesOfferts aU Pere Henri de Liubac, Theologie 56 Tome I (Aubier: 1963): de 281-294 at 282-283 had presented a case for the traditional view: "On trouve le meme procedc composition dans les Regles morales, que Basile tint ä extraire du Nouveau Testament et qui datent du meme temps que laýPhilocalie, les annces oü avec Gregoire de Nazianze, Basile se recueillait Bans de la solitude, completant par la lecture sacree ses etudes atheniennes. L'examen ces notes

24 From the outset it is clear that the issue cannot be conclusively resolved, due fact Basil the Pliilokalia, indirectly, in partly to the that never mentions even his But have Gregory any of extant works. many scholars supposed that of Nazianzus speaks of its compilation in his Ep. 115. This letter, which both Harl date four Basil's death in 383, be and Junod to years posterior to would appear to the earliest mention of the Philokalia in the record. Gregory sends a copy of a work entitled the 'Philokalia of Origen' to the addressee, Theodore the bishop of Tyana. Gregory characterizes the work as "excerpts... useful to the erudite who are engaged in the study of the Scriptures" (ýKA.oyai... xpAmpot tio-q 40,0,6yotg). And there is mention of the late Basil of Caesarea: The collection of texts will be to Theodorus a 'memorial' (vnöµvqµa) of Gregory and 'the holy Basil' 61 Harl tries to cast doubt on the evidence provided by Ep. 115, pointing out that it does not support Basil's participation in the preparation of the anthology very well. 62 Initially Junod is more conservative and suggests, although he accepts the traditional attribution to Basil and Gregory, that the alleged collaboration in compiling and editing the Philokalia might have been quite minimal; the work may have been largely prepared by Gregory 63 But later, he changes his mind and strengthens his arguments. According to his mature view, Junod largely agrees with Harl, thinking it unlikely (although possible) from the wording of Ep. 115 that either of the Cappadocian theologians worked on the anthology. TMRather, both Gregory and Basil must have used the useful work in their studies together or separately, perhaps around the year 360. This would explain the fact that later Gregory regards the Philokalia as a souvenir (1bn6gvrjµa)

d'etudiant est fructueux; rien n'indique mieux les preoccupations, l'orientation, les intuitions de depart des Cappadociens. " 61The key bit of the letter, E_p.115,4-7 (SC 302: 170) reads as follows: "Iva59 2t Kai vn6gVIjga nap' ßjµ ov exi1c, T6 S' aütO Kai Toi' äyiou ßaßt7Rome: Institutum Patristicum 'Augustinianum': 1988): 349-360 at 357-360. Junod (355-6) also discusses the prosopography relating to the appearances of the name 'Theodore' in the letters of Gregory, as well as the dating and circumstances of the composition of Ep. 115. 621-Iarl, Pltilocalie, 20 note 5: "Le not vn6gv11µa renvoie souvent ä un ouvrage ecrit, mais on ne peat deduire de cela que Basile ait contribue ä son elaboration. " 63Junod, "Remarques, " 150: "La signature vent probablement dire que le discours a ete lu, approuve peat-eire retouche, par Basile. 11se pourrait fort bien, de meme, que la Philocalie a ete compilee par Gregoire pour etre ensuite soumise au jugement de Basile. "

25 of the late Basil, who is somehow associated with the codex (nuxTiov) of the book. And clearly the Cappadocian connection to this copy of the Philokalia was vital in securing its safe transmission to posterity. On the other hand, largely on the basis of Ep. 115 but also from Gregory's Ep. 6 (where there is mention of the study of the Scriptures, 6cicov Xoyi(ov ýiXonovia), scholars have dated the composition of the work much earlier to a period (358-360) of monastic retreat shared by the two friends in Pontus. Both Junod and Had have pounced on the most serious problem with this view, namely the implausibility of putting together such a work on slim bibliographic resources 65 The texts represented in the manuscripts of the Philokalia are drawn from 21 different works of Origen, encompassing great variety from the Alexandrian master's immense written corpus; this suggests an extensive familiarity with Origen, as well as the composition of the Philokalia in the proximity of a major library. In his later contribution, Junod brushes aside these difficulties too easily by suggesting that the two young men carried out their research for the compilation directly before they settled down in Pontus, in the course of some travels. 66 There are other facts which trouble the traditional position. As the early Junod argues, the contents of the anthology do not reflect an interest in the spiritual theology of Origen, and thus there is no obvious association with the project of the monastic life carried out so assiduously by Basil and Gregory. The information passed down by Sozomen regarding their studies in the Pontic retreat is hardly conclusive: uni titva Sý xpövov tiotq µa ftmai tiwv nap ' "EXXTI6i. 4aXoß6ýwv ývSta'rpiyravtcS Kai zäg ýý-nyij6Eic Twv iepwv Xöywv e. axptpwaavme ýK 'cwv ' S2pty6vovc Kai 'tcov npo aütov Kai µEt EKcivov F-v Tai, Ep n veiatq TCOV EKK?1.1 ataa'GLKcV hLßXiwv ei 0Mt UjaävTwv.... ' EKd EepoS yäp ävöpEiwS avviatccro zovtT Söyµau np6q tog to ' Apsiov 4povoüv'tag, öu X, wS i? Xa öpBwS 4povouvtaS to K2pty&Vou; Kai 6yXEV µntE iä .u ti&q' Sö aq, aL jth? testa EnijpsI8OVTO. 67

64junod, "Reexamen de la lettre 115 de Gregoire, " 357-360 651-Iarl,Philocalie, 22ff.; junod, "Remarques, " 154 66Eric Junod, "Reexamen de la lettre 115 de Gregoire, " 353: "Enfin leurs voyages leur auraient fourni la possibilite materielle de consulter la plus grande partie de l'eeuvre de l'Alexandrin et done de compiler cette anthologie. " 67Sozomen, HE 6.17.2-3 (Bidez); cf. Socrates HE 4.26 (Hansen)

26 And for a certain length of time they applied themselves to the studies of the Greek philosophers, and examined with rigor the exegesis of the holy Scriptures from the works of Origen and the exegetes who preceded and followed him in the interpretations of the sacred books of sound orthodoxy.... Both of them courageously advocated this dogma [i. e. Nicaea] against those in doctrinal camp of Arius, and confuted them on the grounds that they got plenty of other things wrong and in particular misunderstood the views of Origen, upon which they tried most of all to build their position. Harl blames this text for the general consensus of opinion, that upholds the collaboration of Basil and Gregory in their formative years directly after their formal studies were completed. Oddly, Harl finds it absurd that if Ep. 115 can be dated securely to 383, the composition of the Philokalia took place over twenty years earlier. Rather, it is clear that there is no incontrovertible evidence from the early period for activity which might have produced the Philokalia. Of course one wishes that the sources tell more about the philosophical studies of the Basil and Gregory living the contemplative life in Pontus. In view of these considerations, what can be said about the composition of this valuable anthology? It is certainly likely that Basil and Gregory studied Origen heavily early in their careers, but little can be safely claimed for Basil's involvement in the editing and composition of the Philokalia. Junod in his earlier article, still convinced that the Cappadocians produced the compendium, suspects that a later dating is correct due to the theological maturity detected in the work, probably in the period from 364 to 372. The place of composition might then be Cappadocian Caesarea, which had the best library in this part of Asia Minor. 68 One for be found in . motive composition might contemporary apologetical strategies against pagan opposition: "La tentation est Brande de regarder la Philocalie comme une reponse indirecte de Basile et Gregoire au rationalisme de Julien et de son entourage. " Julian was hardly an out-and-out rationalist, but he did embody a widespread pagan revival in the later fourth century, and there may have been sustained anti-Christian polemic even after his death. As Harl argues, the ecclesiastical tradition may well have originally produced the Philokalia, later attributing it to the distinguished and faultlessly orthodox Cappadocians. But Harl's case is damaged by another piece of evidence. In the manuscripts, there is an anonymous prologue to the anthology which attributes

68Jtuod,"Remarques, " 153-5

27 it to the work of Basil and Gregory 69 However, the anonymous writer only says that some scholars (itv6q) ascribe the compendium and its division into'chapters with their titles to the Cappadocian theologians, as well as the sending of the copy from Gregory to Theodore? ° He notes that the ancient manuscript before him establishes these points in a prologue of its own; 71 but one wonders on what evidence are these facts upheld. The only evidence which could be meant is Gregory's letter, originally sent with the copy of the Philokalia to Theodore of Tyana, and in any case the anonymous scholar accepts its authenticity72 What troubles the anonymous scholar is the unorthodoxy of much that he reads in the anthology, and so his prologue proceeds with the aim of exposing the misdeeds of the Origenists. So it is clear that not only does the anonymous scholar accept the editorship of Basil and Gregory, but also he takes pains to get around the problem of their responsibility for compiling an anthology of dubious orthodoxy; remarkably, he does not take the easier route of simply denying that they took part in the preparation of the Philokalia. At the end of his discussion, he concludes that the attribution to Basil and Gregory is indeed correct beyond doubt, but caveat lector. The Origenists have made impious additions to the anthology, which explains the objectionable bits that crop up in the collection. In summary, it appears that a case cannot be made against the traditional attribution of the Philokalia of Origen to Basil and Gregory strong enough to warrant its rejection completely. There is still the philological sophistication and learning evident in the selections from the vast body of Origenian commentaries and homilies to account for, and the Cappadocian theologians are some of the best candidates of the post-Eusebian era. At the risk of assuming what this section attempts to show, namely the extensive acquaintance of Basil with Origen's exegetical works, it is plausible to hold with Neuschäfer and (the early)

69Harl, Philocalie, 24ff. dates the prologue to the second half of the sixth century, on account of its anti-Origenist tone. This prologue is extant in two versions, the longer of which probably furnishes the original text. 70Origen, Phil. Prol. 3-8 (SC 302: 162) 71Origen, Phil. Prol. 9ff. (SC 302: 162) 72Later in the anonymous prologue, Phil. Prol. 21ff. (SC 302: 162), the author defends the authenticity of Ep. 115 on account of its similarity of expression to authentic letters of Gregory, and its non-controversial attribution by ecclesiastical tradition. The Cappadocians collected these Origenian texts only because they contain much that is useful and helpful for exegesis; like bees

28 Junod that even if Gregory was dominant in the preparation of the anthology, Basil would have been quite interested in such a project. At any rate the evidence is good that he was familiar with the texts represented in the compendium. The most difficult question in the case of non-Cappadocian provenance is when and where the Philokalia was put together; the agents of its compilation, as Junod briefly suggests, would probably be defenders of Origen. So the evidence of the Philokalia and its preparation supports the view that Origen accounts for at least some of the philological and grammatical points found in Basil.

IV

Now we will return to grammar and education. It appears that a case can be made that the study of grammar in the tradition of ancient philology was alive in some major cultural centers of the fourth century. But even more important for study of Basil is the activity of grammarians in association with rhetoricians and sophists. 73 As scholars have shown, this close attention to certain tasks of the grammatical art, above all the purity of speech (similar to what the Stoics called the virtue of Hellenism), was an important characteristic of the earliest stirrings of the Second Sophistic in the latter part of the second century. 74 In the poem of Gregory of Nazianzus discussed above, this primary task of grammar is evident. In the third century, there is evidence that some people taught both grammar and rhetoric, but later it seems that specialization prevailed. 75 Peter

they have chosen the good from what is mixed with floral dross, not including from Origen what is tainted by heresy. 73Peter Wolf, Vom Schulwesen der Spätantike. Studien zit Libanins (Baden-Baden: Verlag für Künst und Wissenschaft, 1952): 21 distinguishes the use of the term aoýtanc from the term i 'rcopin the works of Libanius. The former is used for an orator who on occasion delivers epideictic speeches and earns his living as an instructor of rhetoric; the latter is a less prestigious title, used for a rhetorically trained speaker or an advocate in the law-courts. 74Wolf, Studien, 37 note 71 following Wilhelm Schmid, Der Atticismus in seinen Hauptvertretern, Bd. 1 (Hildesheim: Georg Olms Verlagsbuchhandlung, 1887/1964): 192ff. describes the central role of Herodes Atticus in the formation of the Second Sophistic, in that Herodes studied grammar and Platonic philosophy with some seriousness for the purposes of rhetoric. But as Schmid shows, Herodes seems to have been particularly interested in the Attic purity of speech. 75Wolf, Studien, 37-9 claims of the era of Libanius that "Spezialisierung is ein Kennzeichen dieser Zeit", noting that Libanius himself did not teach grammar to his students. Wolf cites for such activity in the third century Aristodemus of Nyssa, who taught Strabo in Nyssa, Strabo 14.1.48 (Meineke, 908.8-11), teaching rhetoric in the morning and grammar in the afternoon. Wolf also adduces Longinus the Neoplatonist, grammarian and rhetor, who taught both to Porphyry, Eunapius, Vit. Sopl:. 4.1.4 (Giangrande, 6.21-23): nap' eKeivo f rrty ähpav Ercatöevcto icatöeiav, änpov c dicnep exetvoq, äýtKö Prttopucit But ypaµµaTtxtiq to eiq itdri;, LcvoS Kai . of course

29 Wolf shows that in the schools in which Libanius taught, there was considerable overlap between grammar and rhetoric: Grammarians tended to concentrate their attention on a generally recognized canon of classic works of Greek literature, and all students of rhetoric studied grammar as part of gaining familiarity with highly regarded models of (Attic) purity of language. 76 And for their part, grammarians seem to have been quite interested in rhetoric. 77 There is testimony to the standard place of grammar on a par with rhetoric in roughly the same period from a letter of Julian (no addressee is named), who was never a student of Libanius but corresponded with him: 8ý Eov ändv'ro Seiv ö Eni, A, ... noXv 212. v oiµat elvat tiotov'rouS of yots 'rotc v6otg 6uyyiyvovTat, icjiv raA, atwv 9ýiiyiyrai ytyvöµevot 6uyypaµµdiwv, cl'rc nItopac c''TC ypagµaitxoi, Kai Ezt 7X90V of ao4tGTai. ßovXovcat yap np6q'rotc äx?, otq oü Xaýewv µövov, i Ocwvö dvat 8t5wca? ot, Kai 'to Kath 6 äC, Eivai at 'n'v iLOAtTLKhV ýtXoc o4täv. 78

And I think that those involved in the ... above all else, who are oratorical training of youths, presenting themselves as exegetes of the ancient classics, whether rhetors or grammarians, and all the more the sophists, should be men of such a sort (i. e. reasonable and stable in their opinions). For they aspire, in addition to their other claims, to be teachers of morals and not of words alone, and they claim that what they have to offer is the philosophy of public affairs. Of particular interest is Julian's assumption that all three classes of teacher are really interpreters of a recognized canon of texts. One has the impression that for Julian, the three roles largely overlap, although the sophist enjoys the most prestige. As for Libanius' own education, it is remarkable that he studied extensively with a higher sort of grammarian, at least up to the age of twenty. 79 As Wolf points out, this was unusual but possible given the shared concerns of

Porphyry was an exceptional student who seems to have easily mastered everything that he studied. 76Wolf, Studien, 36-37: "In seinem Benähen um attizistische Reinheit der Sprache muß er oft den Rat des Grammatikers einholen. " There is a fine recent article on the aims and blind spots of the earlier grammatical tradition's concern with linguistic correctness by Catherine Atherton, "What Every Grammarian Knows? " Classical Quarterly 46.1 (1996): 239-260, who remarks (240ff. ) on the emergence of criteria of purity as one of the 'virtues of speech', the importance of Hellenism in Stoic grammar, and its development in 'technical' grammar. 77Wolf, Studien, 37-39. Wolf (32-33) also thinks that by the time of Libanius, grammarians may not have been called ypaµµanxoi, rather one finds in Libanius the terms 5t5d6KaXo5, JIatSEUTTýS,or ypaµµananic used for teachers of grammar. 78julian, Ep. 42 (Hertlein, 545.4-11)

30 grammar and rhetoric at this time 80 In a letter to a former pupil, Libanius speaks of an art (Tgxvr ), presumably of grammar, when he mentions the study of prosody, drawing a contrast between early grammar-schooling and later tutelage in rhetoric. 81 In the grammatical art, Libanius must have used textbooks of some sort, he may have known some logic, but the extent of his technical competence must be largely confined to rhetoric. A. F. Norman comments helpfully on the extent of his learning as shown by his frequent use of quotations and references: His was the culture of the commentator and the excerptor, and his stock in trade was largely that of paraphrase and anecdote. It is clear at the outset that Libanius had, besides basic texts, a large store of such sophistic material, one obvious item being the collection of literary and popular proverbs with which his works abound. Similarly, he had access to a collection of myth and fable both local interest.. to the technical of national and .not mention works necessary for the study and teaching of rhetoric. The learning that pervades his work is based as firmly on the encyclopaedia as on personal knowledge of texts 82 This assessment reflects widespread trends in the Imperial period. It is, unfortunately, indicative of the intellectual climate of the fourth century that a figure of the caliber of Libanius would rely so heavily on such materials. Accordingly it is easy to overestimate how much Libanius and his students actually read. As for philosophy, Libanius seems to have little use for such specialized studies apart from the works of Plato 83 The contribution of philosophy to his teaching seems to be largely ethical 84

79Libanius, Or. 1.9 (Martin); Ep. 12 (Fatouros and Krischer) = Ep. 727 (Förster) 80Wolf, Studien, 32-33. Wolf adduces evidence of learned exchanges in Libanius' inner circle at symposia in Constantinople, discussing rhetoric, poetry, philosophy, and grammar, Ep. 28 (Fatouros and Krischer) = Ep. 406 (Förster). He remarks (35): ''Offen bleibt also die Möglichkeit, daß zur Zeit des Libanius wenigstens einzelne Grammatiker auch den Elementarunterricht erteilten oder umgekehrt Elementarlehrer ihre Schüler in der höheren Grammatik unterrichteten. " BtLibanius, Ep. 10.4-5 (Fatouros and Krischer) = Ep. 255 (Förster): ov yäp ýglty of 46yot nepi rtpoaMou 'rrl; än6 v'1; 2exvnc... TouTi yap ov89v np6q TöS Moüaa; Eivat öoxe't. Wolf, Studien, 32 thinks that in the schools in which Libanius taught, grammarians usually treated poetic works, while rhetoricians dealth with prose works, perhaps swayed by this text. 82A. Norman, F. "The Library of Libanius, " Rheinisches Museum für Philologie 107 (1964): 158-175 at 160-161. 83Norman, "The Library of Libanius, " 159; 170 notes that Libanius never even mentions the Stoics, preferring to drop the names of Pythagoras, Socrates, and Plato. Plato is dear to Libanius, but largely a literary influence, 'from the point of view of form and diction than of content'. 84Fritz Schemmel, "Der Sophist Libanios als Schüler und Lehrer, " Neue jahrbiicher f ür das klassische Altertum, Geschichte und deutsche Literatur und für Pädagogik 20 (1907): 52-69. Reprinted in Libanios, Wege der Forschung Bd. 621, eds. G. Fatouros and T. Krischer (Darmstadt: Wissenschaftliche Buchgesellschaft, 1983): 3-25 at 21: "Er unterscheidet zwei Seiten der Erziehung: Xöyot Kai trpönot, die wissenschaftliche und sittliche Ausbildung. Die rhetorische Ausbildung auf philosophischer Grundlage hat das Ziel, einen guten Redner hervorzubringen, die sittliche, einen edlen

31 More significant than his philosophical interests is the likelihood that Libanius was familiar with the rhetoricians Hermogenes (second century AD) and Theon. 85 Of the two, the first-century (? ) Imperial rhetorician Theon of Alexandria is of greater interest: A handbook of exercises for beginning students of rhetoric (Tipoyvµvä6µa ca) survives from antiquity, and this text, in the section on narrative discussing obscurity of language, preserves a list of ambiguity types. 86 This treatment of ambiguity is ascribed to 'the dialecticians', but much of it is Stoic material, as is shown by the close kinship of Theon's list with an ambiguity classification in Galen's On Linguistic Soplzisms which is actually attributed to the Stoics. This affords reliable testimony to use of Stoic material by a leading figure in first-century (?) rhetoric. 87 Unfortunately, little is known of Theon's career, although (save his name) one would be tempted to associate him with the great library and cultural prominence of Pergamum, where Galen in the second century wrote several of his learned works. 88 Did Theon draw on Stoic grammar and rhetoric in his six other treatises that are found in the catalogues? If only the texts had survived. In any case, it is quite plausible that a significant part of Basil's own training in rhetoric was conducted under the shadow of this same Theon, as well as Hermogenes. V. H. Drecoll argues for the influence on Basil in style and presentation of arguments of these two late Greek orators, especially with respect to their (extant) introductory handbooks 89 The rhetorical

Menschen. Die Rhetorik ist der Abschluß der Bildung, die auf den Grundlagen der eyxvO, ta natSevµaia beruht. " 85Norman, "The Library of Libanius, " 173 thinks that knowledge of the rhetoricians Anaximenes and Menander is likely. Schemmel, "Der Sophist Libanios, " 14 argues more plausibly for a strong influence of the earlier rhetoricians Theon and Hermogenes on Libanius as a prime contributor to his teaching, especially with regard to the preliminary rhetorical exercises (npoyuµväaµata). 86Theon, Prog. 80.30-81.13 (Spengel). There are excellent discussions of this material in Theon by Atherton, The Stoics on Ambiguity, 175-214; 473-482; "Hand Over Fist, " 415ff.. 87Atherton, The Stoics on Ambiguity, 212ff. argues from internal evidence of the ambiguity classification that Theon's list probably stems from a Stoic treatise on style. Of course, Theon could have worked Stoic material into his handbook taken from mediating rhetorical sources. Atherton points out several times in her book that Theon's Stoic classification (preserved fairly intact) is quite unusual, for it is found nowhere else in non-philosophical texts, although (471-501) many grammarians and rhetoricians include bits that can be traced back to the Stoics. In fact, Atherton (495) thinks that "only Theon, probably Cicero and Quintilian, and perhaps Fortunatianus, seem to have had direct contact with Stoic material, whether classifications or more general treatments of the topic. " 88Pfeiffer, History of Classical Scholarship, 237 89V. H. Drecoll, Die Entwicklung der Triººitätslelire des Basilius von Cäsarea(Göttingen: Vandenhoeck & Ruprecht, 1996): 57. Drecoll's grounds external to Basil's works for the relevance of the progynuuºsmata are the widespread influence of Theon and Hermogenes on fourth century

32 structure presented by the progymnasmata of Theon and Hermogenes, Drecoll thinks, is reflected in the early treatise Contra Eunomium of Basil 90 This discussion must conclude with a sense of unfinished business, from be perhaps appropriate for an introduction. It is not clear, what can logic gathered of Basil's studies, whether his knowledge of grammar and is from his mainly drawn from the erudition of Origen's biblical commentaries or Athens. As rhetorical training in Cappadocian Caesarea, Constantinople, and be argued above, both sources are quite likely to be significant, although it may erroneous to assume extensive knowledge of Origen's voluminous exegetical corpus for Basil and indeed for Gregory of Nazianzus. However, the supposition of a strong influence from the great Alexandrian will serve as a working hypothesis for research, open to refinement and correction. It is more difficult to understand Basil in relation to the philological tradition, on account of the fact that for the most part Basil (and Gregory) did not write learned biblical commentaries like Origen and the Antiochenes.

rhetoricians, e. g. Apthonius and Nikolaos Sophistes, and the evidence for their influence in Athens. 90Drecoll, Die Entwicklung, 57 argues that on the one hand, the guidelines to writing a polemic against the previously formulated position of an opponent (dvaaxeutj, refutatio) form the background to the rhetorical structure of the Contra Eimomiutm, while on the other hand there is the input of 'die dialektische Argumentation'. Some elements of the named rhetorical structure are teased out fairly successfully (57-63), and similar arguments are briefly presented in the case of the later treatise De Spirihu Sancto (194-195). Drecoll's cursory discussion of the dialectical background to the earlier treatise is not as helpful. He cites Alcinous/Albinus for the divisions of dialectic (58), some of which (TÖ 8tatpETti: 6v and Tö 6ptßnx6v) certainly correspond to some arguments in Basil (59-61) which deal with language. But as will be seen in the following chapters, Drecoll in his examination of key Basilian texts either does not acknowledge or misunderstands much of the contribution of the. ancient dialectic and the grammatical tradition to arguments in Basil.

33 1. Proper Names

Few questions in the philosophy of language since Frege and Russell have attracted as much attention as the theory of proper names and definite descriptions. The ancient preoccupation with naming is rather different. Greek philosophers from at least the time of Plato and Aristotle are interested in the origin of substantives in language, adopting positions which develop the alternatives of naming by nature or by convention. In Hellenistic philosophy, the traditional concerns continue. But there is also some discussion of proper names, which were clearly distinguished from common names. Later still, the philosophical tradition sets up the later Patristic debate on names in trinitarian discourse in the latter stages of the Arian controversy. In the theory of names of the fourth-century theologian Basil of Caesarea there are some intriguing views on the signification of proper names. Although there are interesting similarities to ideas about proper names in earlier philosophy, Basil presents an original and interesting theory, which anticipates modern description theories. Remarkably, Basil is the first writer to make clear the importance of speakers' knowledge in a theory of naming.

I

Fortunately, there has been some recent work on Stoic proper names, which addresses some of the difficult logical and metaphysical questions-' As will be seen, the Stoics influenced later grammarians on the signification of names. The most reliable material for Stoic names is preserved in Diogenes Laertius'

Jacques Brunschwig, "Remarks on the Stoic Theory of the Proper Noun, " Papers in Hellenistic Philosophy (Cambridge: Cambridge University Press, 1994): 39-56. Michael Frede, "The Stoic Notion of a Grammatical Case, " BICS 39 (1994): 13-24; The Stoic Notion of a Lekton, " Language, Companions to Ancient Thought 3, ed. Stephen Everson (Cambridge: Cambridge University Press, 1994): 109-128.

34 account of Stoic dialectic, where he (or rather, his sources) presents a 'nutshell' theory of how names signify. Apart from this, there is not much good testimony to go by. It appears in the Stoic account of their canonical five parts of speech, where the Stoics talk about expressions, as distinguished from things signified. The testimony in Diogenes Laertius 7.58 comes in two flavors, each of which occupies a wholly separate part of speech. There will be closer consideration of some points from the Stoic definition of the proper name later, but here is a good place to present the Greek text with a translation. "E6Tt Sý npojyopia µßv Kath T6v Atoyaviiv µýpo; Xöyou uggc. vov xotvhv notö tia, otov "AvOpcono;, "Innos- övoga öE err g6pog Xöyov öfl? ovv iöiav not6vlTa, oiov Otoyevric, EcwxpäTrjS.

The appellative is, according to Diogenes, a part of speech which signifies a common quality, for example 'man', 'horse'; while a proper name is a part of speech indicating a peculiar quality, for example 'Diogenes', 'Socrates'. On the one hand they have the proper name (övoµa), while on the other hand they have the common name or appellative (nporn yopia). The designation of the proper name which 'indicates' is not a qualified individual but a particular quality ('Sia not6trLS).2 The appellative 'signifies' not a kind but a common quality (xotvý notö c). The examples of the proper name are the proper names Diogenes', 'Socrates' and for the common name the kind-expressions 'man', 'horse'.

2The translation of SriXovv here as 'indicating', not 'signifying' deserves special consideration. Catherine Atherton, The Stoics on Ambiguity (Cambridge: Cambridge University Press, 1993): 302-303 considers the possibility that the relation between proper names and individuals is distinguished by a technical sense of this term (Bilkoüv) as opposed to signifying (011taivEtv). The evidence shows that in other texts, not individuals but significations are 'indicated', as in the Stoic definition of ambiguity also in Diogenes Laertius (7.62). Stephen Menn, The Stoic Theory of Categories, " Oxford Studies in Ancient Philosophy (Oxford: Oxford University Press, 1999): 215-247 at 215-216 note 2 argues convincingly for the importance of key ideas in Plato (Timaeus 49A6-50C) in the Stoic account of substance and quality. This important observation supports the translation to indicate' or to denote' on account of the clear link between use of demonstrative pronouns and the term 81IXoüv (öaa öcucvi)vTES Tcw prjµau rw T65E Kai Toi-no npoaXpci tcvot SnXoüv i yovicOd Tt). On the other hand, Atherton points out that both the Stoic definition of appellative and of verb involve the same verb which is not öt}Xoüv but agµaivc v. Atherton suggests that different Stoics might employ different semantic terminology; she recommends caution in reading too much into such terminological differences. Richard Gaskin, "The Stoics on Cases, Predicates and the Unity of the Proposition, " Bulletin of the Institute of Classical Studies (Supplement) 68 (1997): 91-108 at 101 note 28 contends that Sext. Emp., Matli. 8.12 advocates the synonymy of the two terms in Stoic theory. "Sextus' assertion at AM 8.11-12 that linguistic items do not signify (senrninein) things in the world is... utterly general. Probably the Stoics did not have a precise term for what I am calling the denoting relation. " Thus Gaskin would claim that the Stoics (or some Stoics) do not mark off the denoting relation between language and individuals from the signifying relation between language and what is said.

35 The difficulty immediately arises as to why, according to this Stoic definition (probably traceable to Diogenes of Babylon), both types of noun designate ä quality and not a qualified object. Let us consider what is involved in the definition's recourse to particular qualities. There is solid testimony that the materialist Stoics consider qualities to be corporeal entities. And as Richard Sorabji contends, it is probably mistaken to suppose that somehow the Stoics consider qualities to be bodies distinct from the corporeal objects which have them 34 In fact it is likely that they think that qualities are a part of the objects which have them. Sextus Empiricus (Math. 11.24) provides testimony regarding Stoic qualities to the effect that they are a part of a qualified body, not identical with nor other than a body. (There is also evidence that some later Stoics such as Posidonius think that matter is in some sense a part of individuals, so the idea of such parts looks like it was widely upheld in the school. 5) Thus it is easy to see how qualities are not considered as bodies distinct from the individuals which have them. Also, it is plausible that the quality that is 'indicated' by proper names is something that Socrates has in some way. There is evidence to suggest that this particular quality of an individual plays an important role in Stoic philosophy. As is generally agreed in scholarly work on the Stoics, individuals such as Socrates have a particular quality, and this quality serves to mark an individual off from other individuals. 6 It is a special quality by virtue of which the Stoic sage (if only

3Richard Sorabji, Matter, Space and Motion (London: Duckworth, 1988): 88-93 argues that sometimes the Stoics think of qualities as merely bodies or matter disposed in a certain way. Sorabji explains this by invoking their third category (i«o; 9Xov) which, along with the other 'Stoic categories', is discussed in some sources. Hence Sorabji can show how qualities are not treated as something distinct from the individuals which have them. Menn, "The Stoic Theory of Categories, " 222 note 10 argues against Sorabji's view. He builds a strong point for his case by adducing Math. 11.24 (Mutschmann) for the idea that the quality is a part of the qualified body. Here is found the Stoic position that virtue is a part of the sage, as for example the human hand is a part of the whole man and is not the same as the whole man. In personal communication, Richard Sorabji has conceded that Menn is probably correct that Stoic qualities are not simply the qualified bodies disposed in different ways. 4As Menn, "The Stoic Theory of Categories, " 220 note 7 points out, (incorporeal) predicates are a different kettle of fish than (corporeal) qualities. Objects can't have predicates such as being- wise or being-temperate like objects have qualities such as wisdom and temperance, for predicates are incorporeal. This is testified in Stobaeus, Ecl. 2.97.15-2.98.6 (Wachsmuth). 5Sorabji, Matter, Space and Motion, 87; Eric Lewis, "The Stoics on Identity and Individuation, " Phronesis 40.1 (1995): 89-108 at 102ff.. The key text on this point is Stobaeus, Ecl. 1.177.21- 179.17 (Wachsmutti) = LS 28D. 6A. A. Long and D. N. Sedley, The Hellenistic Philosophers (Cambridge: Cambridge University Press, 1987): 1.174; Frede, "The Stoic Notion of a Grammatical Case, " 20; Eric Lewis, "The Stoics on Identity and Individuation, " Phronesis 40.1 (1995): 89-108 at 90ff..

36 the sage) picks out Socrates, by virtue of which identifying reference to Socrates is possible. The Stoics would seem to distinguish the particular quality of Socrates from the individual, Socrates, who bears this quality.? However, this by latter point is not accepted everyone .8 If we have the Stoic story right, these pieces of information present serious worries. Is the idea that a proper name designates a peculiar quality If supposed to be part of an explanation of uses of proper names? so, is the relevant individual quality of an individual only known and associated with the individual's name by the sage? A theory of names which has nothing to do with how ordinary speakers refer to individuals seems flawed. And further, it is hard to see from the texts that the Stoics are concerned with speakers' knowledge at all, at least for names. Maybe speakers do not have to know the appropriate particular quality to use the name successfully. Next we will take up the grammatical tradition in part two, before moving on to the Patristic theory of names in part three.

II

The grammarians bring to light some important aspects of the 'afterlife' of the Stoic theory which may be reflected in the later Patristic theory of names. To judge by the grammatical orthodoxy of Dionysius Thrax, the denotation of names is considered not a quality but a substance (oüaia). It is natural to suppose that in the case of proper names this simply amounts to holding that names

7Frede, "The Stoic Notion of a Grammatical Case, " 21ff. has a theory of Stoic particular qualities which carefully observes the distinction. The idea, which we cannot evaluate properly here, is that by virtue of particular qualities, proper names denote things of which there can be predication. Frede argues that for the Stoics, propositions satisfy truth-conditions in terms of the qualities which are obtained by objects, not simply in terms of the objects themselves. By his account, the first order of business in the determination of truth is the denotation of a particular quality and the match of an object to the denoted particular quality that it has. Next comes the application of the predicate to that which is so characterized, not to the quality. So the distinction between particular qualities and the individuals that have them is key to Frede's understanding of what conditions are met in order for propositions to be true. But Frede does not discuss here the evidence for the distinction in that the Stoic notion that the qualities of an individual are part of the individual, which is emphasized by Menn, "The Stoic Theory of Categories, " 222 note 10. 8Another scholar who does away with the distinction between individuals and the individual qualities which individuals have is Gaskin, "The Stoics on Cases, Predicates and the Unity of the Proposition, " 99. Gaskin assimilates individuals and the individual qualities designated by Stoic proper names. Socrates is identical with the individual quality of Socrates, he is not a particular which merely has this quality in some way. One problem for Gaskin would be to explain the testimony of Sextus Empiricus which seems to favor some sort of distinction.

37 denote individuals, and in fact in some grammatical texts which otherwise follow Dionysius closely, the signification of a substance is present for proper names but for The designation be discerned in missing common names .9 of particulars can both of the passages in which Dionysius Thrax says something about proper names. First, in his section 'On the Noun' he defines the noun (övoµ(x)as follows. p9pog Xöyov n'rcotucöv, 6coga il' npäyµa ai p.atvov, acwµapev otov XiOoq, npäyµa Sý otov natöEia, KoLVws TC Kai. 't iw; A,Cy6g6vov, xotvwq gEv oiov ävOpwnoS, innoS, i&i.ws Sý oiov EcoKp6ttT1c.io

[The noun is] an inflected part of speech, signifying a body or a thing, a body for example 'stone', and a thing for example 'learning'; commonly and properly said, commonly for example 'man', 'horse', properly for example 'Socrates'. Dionysius seems to be thinking along the lines of the basic Stoic distinction between the two types of name divided into two separate parts of speech." Further into Dionysius' discussion of nouns, there are traces of the Stoic inheritance in his catalogue of kinds (ei &l)of names, the first two items of which are the proper (KIptov) and the appellative (npoo-l'yopuc6v) names. In Dionysius' catalogue is found the 'classic' grammatical formulation of the signification of names in terms of qualified objects (bodies or things); The proper name signifies a 'particular substance' (iöia oüxia) while the appellative signifies a 'common substance' (xoi.vri oütia). Most of the examples here are identical to his examples when he first defines the noun in the text just cited above. 12 But of course here in the names catalogue he omits the 'body or a thing' (awµa, mcpäyµa)signification

9E. g. P. Yale 1.25, col. 1.6-9 (Wouters, 49) 10Dionys. Thr., Ars Gramm. 12 (Uhlig, 24.1-6) IlHowever, Dionysius presents here as an example the word 'learning', which would be a more abstract word than 'stone'. Could the Stoics have ever thought that nouns signify such entities? For the Stoics, is learning a quality which particulars have? Perhaps this is the case. The Stoics hold that each particular species of knowledge (enta'uui) is a quality which an individual has in such a way that he or she knows something by having a quality present in him or her. There is helpful discussion of knowledge and qualities in Menn, "The Stoic Theory of Categories, " 220; 230. Cf. Simplicius, hi Cat. 212.12-213.1 (Kalbfleisch) = LS 28N = FDS 852, where there is a Stoic threefold definition of 'qualified' (notöv). Only the third sense of 'qualified' is the one for which a quality corresponds to the way in which the qualified individual is qualified, and note the example used: The quality of being grammatical corresponds to the way in which the qualified individual (i. e. the grammarian) is qualified. In this sense, the Stoics say that the quality matches the qualified. Someone knows something by having some bit of knowledge, for instance Socrates' knowledge of arithmetic. This arithmetical knowledge in Socrates makes him arithmetical and thus Socrates knows something, he has knowledge. Perhaps the Stoics think of 'learning' or 'culture' (natocia) in a similar way. But even if they do, the division of signification in 'signifying a body or a thing' found here may not be inherited frone the Stoics. 12Dionys. Thr., Ars Granmm. 12 (Uhlig, 33.6-34.2)

38 and just has substances as significations. By his examples, it would appear that Dionysius understands 'particular substance' in the sense of an object such as the individual Socrates. So under the influence of Dionysius, many grammarians held things or substances to be the significations of names. But despite the influence of the treatise attributed to Dionysius Thrax, some grammarians still tended to follow the Stoics on the question of what names signify. Two texts from the Dionysian scholiasts say that proper names as well as appellatives signify 'a quality' (noi6'S). 13 The passages are almost identical, so they both must be drawing on the same material; in both the context is a presentation and critical review of the Stoic classification of proper names and appellatives as separate parts of speech. These grammarians seem to agree with the Stoics that names signify qualities. But even so, the grammarians seem to diverge from Stoic doctrine. This is suggested by the strong tendency to do away with the Stoic distinction between qualities and qualified objects. The scholiasts in question, who have qualities as significations, are commenting on the very passages of Dionysius Thrax which rework the Stoic signification of particular and common qualities into particular and common substances. And they certainly don't seem to think that they are contradicting Dionysius. This slide in the grammarians between qualities and qualified objects is even discussed openly at a much later stage of the tradition by the Byzantine grammarian Choeroboscus, 14 who attempts to explain why for some grammarians 'quality' appears in the definition instead of 'substance'. thy DEl SE 'ytV6)6KELV OTL TLVES, EcTLV 6 clXAicovo; Kai 6` P_wµaV6q 6 TO'UTOU 8t8d(6KaX05, ltotO 1Lta X1 youc v Ev r6 Öpw &vTl To'o o iav, olov övo 6c E6Tl jCpos Xöyov 7LTCUttKOV aKäatOV T(AV i)TtOKEtgEVGJV (YU)pa'rCOV 11 1Lpa7gdT(OV KOLVhV 7j ''(Slav o tTjta äitovE tov, EILEt8h Kai 1j 7LOL6T1js Ev o aIOc Oc(ApciTal' Kai L6TEOV ÖTl OÜala g9V Catty Tj avOun66Tato; vitapEtq, otov ävOpwrtoS ÜL?LOg Kai TCl Totauta, itotovjq Sc X8UKOv ýavOöv tXav 15 avT6 to 7Eol6v, oiov to Tö Tö . Kai Tä TolaiTa.

We must take into account that some [grammarians], among whom is Philoponus and Romanus his teacher, have 'quality' in the definition instead of 'substance', for example "the noun is an inflected part of speech which applies a common or an individual

13Sch.Marc. in Dionys. Thr. (ex Heliodoro) 357.22-26 (Hilgard) = FDS 567; Sch. Vat. in Dionys. Thr. (ex Stephano) 214.33-215.3 (Hilgard) = FDS 568 14Choeroboscusis dated to the ninth century by Robert A. Kaster, Guardians of Language: The Grammarian and Society in Late Antiquity (Berkeley and Los Angeles: University of California Press, 1988): 394-396. 15Georg. Choer., Prol. Scli. in Theodos. 106.3-10 (1-lilgard) = FDS 563

39 quality to each of the underlying bodies or things", since actually the quality is understood in substance; and we must bear in mind that on the one hand substance is the self-standing existent, for example a man, a horse, and such things, while on the other hand quality is the qualified thing itself, for example that which is white, that which is blonde, that which is black, and such things. According to Philoponus and Romanus, quality and substance are two sides of the same coin, for talk of qualities is really just talk of qualified objects insofar as they are qualified. This view, which seems quite well-represented in the grammatical tradition, might help to explain the dodge of the question of reference in the Patristic theory of names. This brief notice on Philoponus' grammar is interesting; the subject deserves further study which cannot be undertaken here. The justification ("the quality is understood in substance") of qualities in the definition of names, which at an earlier stage was Stoic orthodoxy, might be a window upon Philoponus' response to Porphyry and other commentators on the Categories: If a quality (e. g. rationality) is used as a differentia of a substance, then it belongs in the category of substance; in fact this is true of all defining qualities of a substance. 16 But as Schneider shows in his fragments of the lost Apollonian treatise 'On Names', the definition of the noun ascribed in the text to the later figures Romanus and Philoponus is actually that of Apollonius Dyscolus and his son Herodian. 17 So let us turn to Apollonius. The most important case of Stoic influence on the denotation of names in the grammarians is Apollonius' largely extant treatise on syntax. In several passages Apollonius affirms that the signification of proper names (xvptov övoµa) is the particular quality of objects. 18 But such brief glimpses do not exhaust all

16This point owes much to discussion of Porphyry with Richard Sorabji, who has supplied references for the notion that differentiae and other items which complete substance belong in the category of substance and should be classified as substances: Porphyry, In Cat. 94.17-28; 95.22-96.1 (Busse); cf. Simplicius, In Cat. 48.1-49.9 (Kalbfleisch). But Porphyry deals with the problem presented by Aristotle (Cat. 3a21-28) where is found a distinction of differentia from substance by arguing, In Cat. 95.17-20 (Busse); cf. Dexippus, In Cat. 49.4-6 (Busse); Simplicius, In Cat. 99.3-9 (Kalbfleisch), that the differentia is really neither a quality nor a substance but rather an intermediate item, i. e. a substantial quality. This latter view is in turn rejected by Ammonius, In Cat. 46.10-19 (Busse) and Philoponus himself, In Cat. 66.13-25 (Busse); ap. Ammonium In Cat. 64.22-66.12 (Busse), who take Aristotle to include differentiae as substances and only distinguish differentiae from composite substances, Ammonius, In Cat. 45.21-46.10 (Busse) or 'garden-variety' substances, Philoponus, In Cat. 67.1-3 (Busse). 17Sch. in Dionys. Thr. 524.8-12 (Hilgard) 18Ap. Dys., Synt. 66.11-12; 142.1ff.; 155.3-5; 159.7; 352.5-6 (Uhlig). However in one report of Apollonius and Herodian on cases, Sch. Lond. in Dionys. Thr. (ex Heliodoro) 546.5-14 (Hilgard) = FDS 780, they have substances as the signification of cases, "for [the nominative case] signifies the substance of the thing" (6p0cO; yap rtµaivet Týv ov6iav Tov npäyµaroc). A similar idea is found in some manuscripts of Ammonius, In Int. ad 43.19 (Busse) = FDS 776, so it is likely

40 He that he has to offer on the topic. says more about the meaning of names than transmitting his theory any other surviving grammarian, of names to posterity; 'character' in particular, his notion of the associated with a name is relevant to later 'classic' Patristic In understanding the theory of names. one passage which has been largely overlooked, Apollonius fleshes out his theory of proper names.

'H Tcwv övogdTcov BE6tq EnevoiiO i Ets notö'tT taq xotv&q ij iöiaS, wS ävOp(wTtoC, HX% T(Ov, Kai E1cEl o'ÜTE 616v ftiýEt Tä Totaüta o'Ü'CE äva4o pä, 7t6µno? oS ý uni ToüTwv OýGts E'yivctO, IV' ýKäcTOV Tö xapaKTrtpt6TtKOV 6(7tOVEiL13 Tr V ýKdGTOV notöuyra. (ov . ETpicoS y' ovv T&q notözryTaq enttapäTTovaty ai. OUVE11 IE6oucJat O cYEts EV TE 1tpocnyyoptKOtq Kai. K'Opiotq ÖVÖ jlauty, St' Ö Kai e6TEpETO TÖ FnL T(AV övoµtTwv vooi tevov icpÖ oolcov To'U Öpt(Tof Kai EVEKa Tov Totoütov i&tähaca 11 EüOEta KaO' EKaaioV .) %apaKn pa, ctnijTEt T&q itXayiaS ävT1KOÜ6acitp6g to i.Stov Tic Ev6Eias. ývTEVOEV ovv ävayKaiwS Kai Tä y6VII IcapEt66yETat, iva tETä T1jq not&gTog Toy %apaK'nlpoS auvötac teIX j Tö y&Voq.19

The imposition of names is conceived in terms of common or particular qualities, as for instance [the names] 'man', 'Plato', and since [names] of such sort are neither by pointing nor by back- reference, the imposition upon these entities is the most standard, to the end that the word-form of each [name] attributes the quality to each . (Of course the homonymous naming in the appellatives and proper names plays considerable havoc with the qualities, on account of which the person conceived by means of the names is deprived of determinacy. ) And for this reason the nominative, having established individuality according to the character of each [word], brings in its train the oblique cases which belong to the declensional pattern signalled by the nominative. Hence necessarily the genders are also introduced, in order that along with the quality, the gender of the character is distinguished. The translation provided will require some explanation and elucidation. Apollonius here discusses the signification of proper names in terms of individual qualities which things have, but does so in a different way than in any other that other grammarians hold this position. To support their claim that the nominative is also properly a case, Apollonius and Herodian introduce a doctrine of the 'generic name'. The idea is that the generic name includes every instance of the name. The cases 'fall under' the generic names, and the cases are word-forms Q,94tq dvöµaToc) used in discourse. Hence they are instantiations in use of the abstract entities of generic names; the abstract name is applied to form the particular case. Frede, The Stoic Notion of a Grammatical Case, " 24 thinks that this explanation is an answer to the question of what declined word-forms are cases of, what the terminology of 'cases' means; he suggests that the explanation is derived from the Stoics. There is examination of this idea by Atherton, The Stoics on Ambiguity, 282 note 51, who cites for comparison Sch. Lond. in Dionys. Thr. 547.18ff. and Georg. Choer., Prol. Sch. in Theodos. 109.24ff..

41 extant passage. In the course of explaining Apollonius on names, the conclusion for this text in the 'On Syntax' will be that the 'character' (%apaKt1jp)has to do with word forms of one sort or another, not with the semantics of names. Later, it will be seen that another use of 'character' in Apollonius does bear on the semantics of names in an interesting way. Only in this passage from the 'On Syntax' does Apollonius talk specifically on proper names while including the notion of 'character', but it is difficult to ascertain what Apollonius means by this term. The question of interest is what the 'character' has to do with his views on the signification of proper names. Apollonius' idea about nominal semantics is that what is signified by a proper noun, determined for the nominative case for the benefit of all the cases, involves particular qualities. But how does Apollonius understand the denotation of proper names to be qualities? Apollonius may be influenced by the Stoics, but he is no philosopher. He often does not articulate his ideas very adequately. Elsewhere, in his treatise 'On Pronouns', Apollonius' position seems to be that the particular quality of an object is part of what is signified by a proper name, when he says that pronouns signify (aiiµaIvety) substance (ov6ia) while names signify substance with quality (oix Ia µeiä notötntiog) 2° It looks like Apollonius has not entirely collapsed the conceptual distinction between qualities and qualified individuals. (It is plausible in this context that his substances amount to individuals. ) In this sense, he is in alignment with Stoic doctrine. Further, the texts from the 'On Syntax' and the 'On Pronouns' when taken together show that Apollonius has decomposed the signification of objects by names into two elements, namely substances and qualities. On the other hand, in the 'On Pronouns' Apollonius merely says that a pronoun such as the deictic pronoun 'this man' (oüios) signifies a substance, namely the very man being pointed at, while 'Socrates' signifies a substance as well as a quality which the substance has. All this cries out for some resolution. What does Apollonius think about the meaning of names? It is well-attested from the 'On Syntax' and the Choeroboscus (plus other fragments) testimony that Apollonius thinks that

19Ap. Dys., Synt. 142.1-143.3 (Uhlig). The cornered brackets enclose words that are conjectured by Uhlig, probably on the basis of standard forms of expression in Apollonius, e. g. Synt. 155.3-5 (tj iSia notötrlS tou vrtoxctµevou). 20Ap. Dys., Pro;:. 27.9-10 (Schneider)

42 'attribute' 'assign' (&noveµeiv) to bodies proper names or particular qualities and ) This is his talking the other (incorporeal? things. evidently way of of in 'On Syntax' On the hand, Apollonius signification of names the passage. other thinks that a proper name denotes a substance, or alternatively a substance as has. be help Well as a quality which the substance Perhaps it will of to consider denotation is more carefully that Apollonius often speaks of a as that which (vnoKEiµevov). It is subject to qualities, that which is 'underlying' a reasonable by language, he hardly be assumption that he just means substances this could frequency in charged with a developed matter-and-qualities theory. The which the word 'underlying' appears in similar contexts presumably motivated Uhlig to Most supply it with cornered brackets in the text of Apollonius cited above. importantly, it crops up in Apollonius' definition of names which is attested by a Dionysian scholiast and appears in the text of Choeroboscus cited above. It is easy to see that this Apollonian definition of name, which runs "the name is an inflected part of speech, attributing (änoveµov) a common or particular quality to each of the underlying (ürtoxstg9vcov) bodies or things, " involves the same idea of a substance which is the bearer of the quality in question. By this point, it should be emerging that for Apollonius, the quality is somehow associated with the substance by the use of the name to denote something. But what is the nature of the tie, and how is it involved in the use of a name? Apollonius only says that proper names 'attribute' or 'assign' individual qualities to (particular) substances. From his statements, it looks like he has in mind one and only one quality which is linked to an individual. He never lets on that such a particular quality could consist of a cluster of qualities or something like that. So the question is what sort of quality the quality which is part of what is signified by 'Socrates' could be. But this is really an impossible question to answer. As will be seen in the next section, the nature of particular qualities is very difficult to determine even for the Stoics. One possibility for Apollonius would be that this quality is the individual quality of Socrates, the property of being Socrates. On this view, no other property which Socrates has, e. g. being wise or being snub-nosed, would be as good a candidate for the property signified by the name 'Socrates' along with the substance of Socrates. In any case, Apollonius has a theory which reflects the original Stoic notion of quality as the signification in the semantics of names, but also has the idea that names mean individuals. He just doesn't tell his readers how the individual quality of Socrates

43 is 'assigned' to Socrates; presumably he goes into this in more detail in his lost treatise 'On Names'. Perhaps this is related to what Apollonius says in the 'On Syntax' passage about homonymous naming. Apollonius' parenthetical remark in this passage could be taken to say that when the qualities which are signified are free from homonymy, the entities (referents) designated by the names are definite on account of the unconfused signification. Only one referent is designated by the name. In the (rougher? ) language of his 'On Pronouns', a proper name signifies a particular substance, and it also signifies a quality of the substance. But these points do not license a conclusion very strongly. Apollonius may have some notion of the 'individuality' of significations, a way in which the qualities signified by names play a role in picking out the unique referents of names. But it is hard to see more than this from what is left of Apollonius. In summary, Apollonius shows how later grammarians could take the Stoic qualities and relate them to the substances which have them, all within a theory of how names signify. This survival is important for our discussion in the next section. But Apollonius' use of the term 'character' is also interesting as regards our Patristic theory of names. First, recall the notion of the 'character' of word-forms in the 'On Syntax' passage adduced and quoted above. Apollonius says that the nominative case establishes individuality according to the 'character' of expressions. Before encountering this passage, one would expect that any use of the term by Apollonius would be for lexical or formal characteristics, not having to do with semantics. Elsewhere, Apollonius uses 'character' for letters of the Greek alphabet (Synt. 41.1), or word-forms and inflections of words e. g. in the singular or the plural (Synt. 31.4-5; 146.7; 315.20). 2122 Now consider another difficult passage which comes soon after the text cited above from the 'On Syntax'23 Here it appears that the 'character' of a name

21This reflects the standard range of uses of the term in the grammarians. Cf. Ap. Dys., Conj. 247.30ff. (Schneider) = FDS 584 and references in the Dionysian scholiasts collected by Atherton, The Stoics on Ambiguity, 226 note 9, in which the term XapaKtjp is used for the written characters or letters as opposed to vocal elements (no1Xeia, EK(ýwVA6Etg). 22The cognate term in Synt. 142.1-143.3, Tö XapaKTi p yTLKÖV, seems in context to mean the distinguishing form of the name. This term does duty for the expression which 'attributes' qualities to objects. It is interesting that in a Dionysian scholium, Sch. in Dionys. Thr. 527.22-23 (1-Iilgard), the 'character' of words has to do with the individuality of an expression (Ka0' exc

44 denote can bear a semantic sense. Apollonius says that personal pronouns, only 'underlying substances' by means of their capacity of deictic reference. But he implicitly (EiAilnra) 'the adds that personal pronouns have things which (Tä This is supervene on the quality' ERtauµßEßr1K6Taijj noiöttt). terminology terribly obscure, but it is clear that Apollonius has objects insofar as they are qualified in mind. For Apollonius goes on to explain that such things are perceptible by sight, and qualified in some way; for example that which is white or black, large or small. For instance, the pronominal use of 'this man', pointing to Socrates, would implicitly 'have a qualified thing', it would involve Socrates' properties of being white, being snub-nosed, and so on. However, neither the soul nor its contents are understood even implicitly (evavvonta) when the pronoun is used; how could they fall under the pointing (SeiýiS) of the pronoun? Hence the pronouns have 'rejected', the pronouns don't involve, 'the many

characters' (TOVSztoXXovS XapaKTljpac). Presumably these are the 'characters' which nouns express but pronouns can't express, such as Socrates' being wise. It is hard to see how this use of 'character' could be taken to mean word-forms or something like that. Hence it is fairly clear that here in Apollonius is a semantic use of the term. In this passage, the 'character' means certain kinds of qualities or states which are signified by names.

III

Now we turn to the Patristic theory of names. In the latter stages of the Arian controversy, matters of language became disputed philosophical territory. The main theological opponent of Basil, Eunomius, has a philosophical account of names. Thus it was necessary to formulate a counter-account of the signification of names. 24 Not too much will be said here about Eunomius' theory of language and its philosophical antecedents, a topic which has proved attractive to Patristic scholarship. 25 But a few points from Eunomius will be discussed along the way.

24The important related issue of the origins of language is not treated here. 25Karl-Heinz Uthemann, "Die Sprache der Theologie nach Eunomiuts von Cyzicus, " ZKG 104 (1993): 143-175. Marina Silvia Troiano, "I Cappadoci e la questione dell' origine dei nomi nella polemica contro Eunomio, " Vetera C/hristinnorunz 17 (1980): 313-346. Behind the work of Uthemann and Troiano is the adventurous contribution of jean Danielou, "Eunome l'arien et Etudes l'exegese neo-platonicienne du Cratyle, " Revue des Greques 69 (1956): 412-432. Danielou advanced the view that Eunomius was influenced by the Neoplatonic tradition of commentary

45 wv Kaitot 'tic äv c(üXöycp rorrq aw4povv npäaoolto, öTl Tä övöµaTh äväyKTI; Eats &l 4 opa, 'TovTwv naplIXXäxOal Kai Idg oüaiag rkTpov yäp Kai IlanXov Kai änaýanX. ävOpwnwv nävTwv npooi yopial µßv 5lä4opol, oüaia Sý nävTwv Ilia. SlönEp Ev Toig nXcIaTols of aüTOi äxXi Xols E6u£v" ToiS öe iölwµaat µövotc Totg nepi 9KaaTOV 68wpovµEVotg ÜTEpoc ETepov SlcvvvöxaµeV. öOav xai aI. nporn}yopial oüxl TwV oüatGJV Eicl 67lJlavTixai, ädÄ, ä T(ýv i&lo'ri wV, cit TöV KüCO' EVa xapaKTTjpiýoucYtV. öTaV OUV äKOÜ6wg£V ITETpOV, Oü TA V O'ÜaiaV aütOÜ VOOügEv EK TOD övöµaTOS (oüGiav Sý XEy(ü vvv Tä üXLxöv vnoi i tevov, önEp oü6aµwS rn»µaival 'rovvoµa), äX.ß, ä ticýv I.bicogtTwv ä iepi aüTÖV OawpiiTal AV Evvolav EvtivnoüµeOa. e OüS yäp EK Til; 4wvijs TaüTIlS voou v Töv TOI) ' Iwvä, Töv ýx 'n l; Bii86ai. öä, Töv äöcX Ov' Avöpeov, Töv änö äXthwv ci; 'n' v Slaxoviav TIlS änoc 'no? ýS npo6K? T1OEVTa,Töv ölä Tres 7tiaTcwc ünEpoxrlv E4 ýavTÖV T'v OIKO6Op V TTg EKKXTjciaS öcýäµcVOV" wv oüSev . anv oüßia, Tj (hg ünöaTaats voovµhvrl. (iiaTE Tö övoµa Töv %apaKTýpa gEV ýµiv i opiýEt Töv IIETpov, aü'ni v SE oüSaµoü napia'Tjal Tijv oü6iav. näXty äKOV6av'ES fIa 5Xov, ETýpwv iölwµäicov 6uvöpoµýv £vvoi aµsv" Töv Tapaftc, Töv ' Eßpaiov, Töv KaTä vöµov caplaai. ov, Töv p.aOryti v FaµaA, lA;;, Töv xath ý1 Xov cStc6KT1V TWV ' EKK%Ti NUAV TOD OCOU, TÖV EK T7js 4oßspac ömtaoIaS ei; Týv ýmiyvwaiv ývaxOgvta, Töv änöaToXov Twv eovd v, TaüTa yäp nävTa EK µläs ýmfig Tip IlaüxoS ncplopiýBTat. Kairot ye, Eipep äxlloES ijv, önl wv Tä övöµaTa SlevAvoxev, EvavTiws gxovcty al oücial, Expiv öij nov Kai fIauXov Kai IZETpOV Kalt änaýan; ýw; 6v8pwnoug änavTaS ETSpoovaiouS a2,1A2,ol; Mal. 'End Se rouTO oi6eIS ovtü g äµaOýS Kai 'n ý; KOLVls ýüasws äVElniaKFm'OS w6T ' äV Einei. v npoaxOýval ('EK nrrXoi yap, 4rIai, ölrjptlaal av, wS Kai EycO O )6 V eTepov Toü X.öyov f Tö öµooüalov nävTwv ävOpwn(üv änoaij. Laivovto; ), weuSi g6 'rjj Sla4opä 'n v övog(iTwv Tö T1j; oüaiaS Slä4opov ene68at ao4l(6! EVOS. Oü yäp Toi. övöµaaly 11 Tcov npayµäTwv 4üctc 6KoXou8ei, äXX' vkTEpa ' acv npayg6Twv süpIlTal Ta övöµa'ra. 26

And yet who with their wits about them could subscribe to this account, that necessarily for things of which the names are different, the substances of these things are different? Of course the names of Peter and Paul and of absolutely all [other] men are different, but the substance of all [men] is the same. For this reason in most respects we are the same as one another, but each of us differs from the others solely by virtue of the characteristics which are observed concerning each individual. Hence the names [of particular men] are not actually signifiers of substances, but of the properties which characterize the individual. So then when we hear 'Peter', we do not from the name think of his substance (by 'substance' I mean now the material substrate, which the name in no way signifies), but we are imprinted with the notion of the characteristics which are observed concerning him. For on the Cratylus, as pursued by the followers of Iamblichus contemporary with Eunomius, who were the forerunners of Proclus from whom a Cratylus commentary is extant. 26Basil, Eun. 2.4,1-42 (SC 305: 18-22)

46 immediately from this utterance we think of the [son] of Jonah, the man from Bethsaida, the brother of Andrew, the man called forth from the fishermen into the service of the apostolate, the man pre- eminent through faith who received upon himself the edifice of the church; none of which [i. e. these characteristics] is substance, which is conceived as hypostasis. So the name marks out for us the character of Peter, but in no way stands for the substance itself. And again when we hear 'Paul', we think of a concurrence of different characteristics: the man of Tarsus, the Hebrew, the Pharisee according to law, the disciple of Gamaliel, the zealous persecutor of the churches of God, the man induced from the fearful vision into deeper insight, the apostle of the Gentiles. For all these [characteristics] are circumscribed from the single utterance of 'Paul'. However, if [Eunomius' theory] was true, that things for which the names differ also have opposite substances, it would have to be the case that Paul, Peter, and absolutely all [other] men are substances foreign in nature to one another. But since nobody is so untutored and ignorant of the common nature so as to be induced to say this (the [scriptural] statement "from clay, he says, you are fashioned, as also I [am]" [Job 33.6] signifies nothing other than the consubtantiality of all men), the man who claims sophistically that the difference of substance follows the difference of the names speaks falsely. For the nature of things is not based on the names, but the names are invented posterior to the [existence of] the things. Before entering the full discussion of the Patristic theory of names, it is important to cover just two preliminary points. A brief sketch of the Patristic debate will help to explain why for Basil, names don't mean substances. I think that he does not intend to deny that personal designations such as 'Peter' and 'Paul' refer to particular objects. As will be seen, Basil's theory is not explicit enough as a theory of reference. It deals with naming as a theory of meaning. The first point is about what names mean in the Patristic debate. In contrast to his opponent Basil, Eunomius thinks that names mean substances. Names different in meaning, at least when properly assigned to substances, perfectly and exhaustively express different substances. On the other hand, Basil argues that names, or at least proper names, don't mean substances. Rather, names distinguish between individuals by signifying properties. Of course, Basil doesn't just do away with the notion of substance altogether; he just takes up a notion of substance which is different from that of Eunomius. This redefined notion of substance serves his polemic better. Eunomius considers all objects, whether incorporeal or corporeal, to be substances. His argument starts from the idea that the term 'Ingenerate' is the name which is proper to God, only 'Ingenerate' expresses the substance of God. Other terms are applied to God improperly, by way of human (conjectural)

47 'conception' (ýmivota). Since 'Father' and 'Son' express substances and the names 'Father' and 'Son' don't have the same meaning, the substances are different. Against Eunomius, Basil's view is that names don't express substances. He presents this argument in the case of human individuals. Proper names signify the qualities which individual men have, and these very qualities distinguish men from each other. Different names, even for objects which share the same substance, do not all have the same meaning. For Eunomius, according to his notion of substance and his theory of language, presumably would hold that distinct word-forms which name the same substance are really identical in meaning. They would just be different forms of the same name. Elsewhere in his Contra Eunomium, Basil says that we don't have knowledge, by sense-perception or by revelation, of what substance is even in the case of material things. He uses the example of the substance (ov6ia) of the earth, which in the Genesis account is described as unseen and unformed. 27 However, sense-perception tells us what properties an object such as the earth has. Yet none of these properties, nor even an exhaustive catalogue of properties, is informative as to what the substance of the earth is.28 Basil's argument goes, how can we know what the substance of God is when we can't even know what the substance of the earth is. So his reasoning leads him to a position which doesn't say that we can't know anything at all about individuals, since we can know how individuals are characterized, but emphatically rules out knowledge of the substances of individuals. 29 Second, there is the problem of Basil's parenthetical remark about how by the word 'substance' the material substrate is meant, and how this underlying stuff is not expressed by the name. The idea is odd because elsewhere Basil talks

27Basil, Ettn. 1.12,30ff. (SC 299: 214ff. ) 28Basil, Eun. 1.13,13-14 (SC 299: 216): "None of these [properties listed above] could be said by them to be substance, not even if an exhaustive [catalogue of properties] could be readily supplied" (6v ovöev äv ovßiav eittotev oütot, ovSE ei n vta Oq&o); X yotcv). 29The epistemology here is geared to confute Eunomius' cardinal doctrine that "God himself is ingenerate substance" (a-6T6 dv Eid ovßi. a dy9vvriToc), together with Eunomius' supporting contentions at Ein. 1.13,36-37 (SC 299: 218) that information of what the correct name of God is (the Ingenerate) as well as of what God's substance is has been divinely revealed. Eunomius found Basil's view of the origin of names, that names are a product of the rational linguistic faculty of humans (enivota) which serves to communicate in a clear and unconfused way the concept of things, offensive to the work of divine providence. This can be seen in later reports of Eunomius in Gregory of Nyssa. The relevant references are assembled in Troiano, "I Cappadoci e la questione dell' origine dei nomi, " 320-321 note 37. Eunomius thinks that names are divinely imposed, but only according to the nature of things.

48 about a common 'account (Myoq) of substance' in terms of predicates unequivocally said of particulars. And in the beginning of the text cited, this would be a natural understanding of his idea of what men have in common. Presumably, Paul and Peter are equally characterized by predicating rationality, mortality, etc. of them 30 But in some respects, this story wants to avoid becoming too Aristotelian. Basil takes pains in his early writings such as the Contra Ennotnittnt to avoid a notion of substance which recalls too much of the secondary substance of the Categories; substance must be material in some sense31 Nor do his substances amount to something like the 'subject' familiar from the Categories. The citation of job 33.6 about men formed from clay looks like an illustration of what stuff is common to men. It is not a statement about defining characteristics. The metaphysical notion is that substance alone does not constitute particulars, for particulars are characterized by the sort of qualities which what is common to particulars cannot have. The linguistic notion is that the 'material substrate' (i tKöv ünoKEiµcvov) which lies under or in all men can't possibly be meant by uses of proper names. What is so remarkable about this Patristic theory of names? The main philosophical point about Basil on names is that it is suggestive of the modern descriptive theory of names, by which certain properties are associated with the names by speakers to determine reference. On the philosophical background of the Patristic naming theory, it is hard to pin things down. As argued in the first section of the chapter, it does not appear that the Stoics talk about names in terms of speaker's knowledge. The best information available is that (1) Stoic proper names designate particular qualities which particulars have, and (2) the Stoics distinguish in some way the particular quality from the individual who has the particular quality. In spite of our expectation that for the Stoics, individuals are designated by proper names, we gather from Diogenes Laertius 7.58 that what is designated by a proper name

30There is some examination of this point in David G. Robertson, "Stoic and Aristotelian Notions of Substance in Basil of Caesarea, " Vigiliae Christianae 52 (1998): 393-417 at 404-414. 310f course, much of this sounds like good basic Aristotelianism. Two important ways Aristotle talks about substance (Met. 7) are in terms of the substance of something (Met. 7.4-12), i. e. the essence which is stated by a definition, and in terms of the notion of subject (Met. 7.3) which is taken up with reference to matter. From the standpoint of Patristics, there is analysis of this difficult point by Robertson, "Stoic and Aristotelian Notions, " 407ff.. The insistence in some passages on the materiality of substance is part of what motivated previous scholars to suppose that Basil has a Stoic notion of substance. This supposition is discussed and attacked in the article.

49

s Lit[. i 0 is not apparently an individual but a particular quality which an individual has. Some Stoic ideas about names and individual qualities may have filtered down to the fourth century in the grammarians or in handbooks. Or perhaps Stoic ideas on the distinguishing qualities of individuals were developed and disseminated in Imperial philosophy, eventually being worked into doctrines of what names mean. There is a notion found in some later ancient philosophical texts of the particular quality as a combination of qualities, a 'concurrence' (auvöpo n) of qualities. This would bear an interesting similarity to the 'concurrence of features' (i8to)g 'ro v ßuvSpo n) in Basil, although Basil may not have a notion of the particular quality of an individual. The basic idea is that a set of properties does the job of characterizing an individual uniquely. Some scholars have attributed this 'concurrence of properties' doctrine to the Stoics, a doctrine that the peculiarity of an object consists in a unique complex of qualities, qualities which other objects taken separately have as well32 The report of the fourth-century Aristotelian commentator Dexippus provides the best information on this dubiously Stoic idea.

äA,ß. ' FA eJ.56, E6tL 'to Kath mXetövwv Kai. Sta4epövto vt äptOµc;ý ýv 'tc 'ti e 'rt KatrJyopovµevov, Tivi. Sta#pet ö ätioµoS Kai Eic toü ä'öµov Kai. EOS* Ev yäp äpi. O is sau Kai ovtos xäxsivoS. of µßv oüv V )OWE; tf v ärcopiav iaü"v Kath iö i8i oq notöv, TOUT ' ganv öii 6 µev 4epc ypuit iti ii fj x=06 reu fj äA,Xll 6uv8pop q noiozýzwv 6ý6)pißiai, äkkos SE alµöTrjtit fi ýAaxpö it ýj ykavxöirlzi, Kai ýäýLV iepoS ýTEpatS, ov xa? Lawsgot 8oxov6i Xev. 33

But if form is that which is predicated in the category of substance of a plurality of numerically different things, in what does single individual differ from single individual, seeing that each is numerically single? Those who solve this difficulty on the basis of the peculiarly qualified--that one individual is distinguished, say, by hookedness of the nose, by blondness, or by some other concurrence of qualities, another by snubness, baldness, or greyness of the eyes, and again another by other qualities--do not seem to me to solve it well 34 This is not the only evidence for the idea. Porphyry seems to be in agreement with what Dexippus reports here. Porphyry presents a notion, in order to

32David Sedley, The Stoic Criterion of Identity, " Phronesis 27 (1982): 255-275 at 261; 273 note 27. Sedley cites Porphyry, In Cat. 129.8-10 (Busse) = FDS 848; Dexippus, In Cat. 30.20-27 (Busse) = LS 28J = FDS 847; Simplicius, In Cat. 55.2-5 (Kalbfleisch) = FDS 848A; Simplicius, In Cat. 229.16-18 (Kalbfleisch) = FDS 848; Philoponus, In An. Pr. 167.17ff. (Wallies) = FDS 995. 33Dexippus, In Cat. 30.20-27 (Busse) = LS 28J = FDS 847 3'}The translation is taken from Long and Sedley, The Hellenistic Philosophers, 1.169.

50 explain Aristotle and not the Stoics, that Socrates differs from Plato by virtue of 'the particularity of the concurrence of qualities (i& ou ti, auvöpoµijS toi. otitjtcov) 35 Long and Sedley present a case for attribution of the idea to the Stoics, but there is room for doubt and it will not be considered securely Stoic here 36 As Long and Sedley concede, there is no mention of Stoic provenance in this or the other 'concurrence' texts, although the Stoics are attested in a passage of Simplicius often linked to particular qualities and 'concurrence' which uses the terminology of 'in a cluster' (&Opöcaq)37 The Simplicius passage is thought to show that the Stoic peculiar quality is a collection or bundle. Simplicius says that what in Stoic terms is the 'peculiarly qualified' is established in its entirety at the inception of an individual, persisting during the lifetime of the individual, and then totally undone at the termination point of the individual. But it is hard to tell whether the key term 60p6wS means here something like 'in a cluster' rather than, say, 'all at once' 38 And further, the passage thought to be parallel with Dexippus in Porphyry is hardly incontrovertible evidence for Stoic philosophy, although Simplicius says at the beginning of his Categories commentary that Porphyry's Categories commentary is significantly influenced by the Stoics 39 Why couldn't Dexippus be talking about Porphyry's discussion of Aristotle when he assigns the idea to philosophers "who solve this difficulty on the basis of the peculiarly qualified"? On the other hand, (1) the expression (iöi(S 2tot6v) is after

35Porphyry, In Cat. 129.8-10 (Busse) = FDS 848. Karlheinz Hülser, Die Fragmente zur Dialektik der Stoiker (Stuttgart-Bad Cannstatt: Frommann-Holzboog, 1988): ad loc. thinks that this bit of Porphyry's commentary is badly corrupted. In any case, this passage may not give us a Stoic idea about particular qualities, although it is certainly about a 'concurrence of qualities'. Simplicius, In Cat. 2.5-9 (Kalbfleisch) testifies that Porphyry borrows quite a lot from the Stoics in his longer commentary in seven books on the Categories. 36Long and Sedley, The Hellenistic Philosophers, 2.174 37The expression (ciOpöco;) is a paronym of a standard twin term of ßuvöpo tij, äOpotaµa. The passage is found in Simplicius, In An. 217.36-218.2 (Hayduck) = LS 281 = FDS 846: "... If in the case of compound entities there exists individuated form--with reference to which the Stoics speak of something peculiarly qualified, which both is gained, and lost again, in a cluster, (ö Kai ä9pöuo Eit yivc-rat Kai al') änoyivetat) and remains the same throughout the compound entity's life even though its constituent parts come to be and are destroyed at different times. " The translation is adapted from Long and Sedley. 38Richard Sorabji has pointed out in personal communication that Simplicius, a good reader of Aristotle, might well be using this term in a way which assimilates the two alternatives: In the passage Sens. 447a (Hett), Aristotle uses a cognate word when he speaks of the possibility of 'alteration' (evSE ctat yap ä6p6ov dUotoüßOat, Kai µr1 'cö ijµtau rtpö'epov) 'all at one go' with all parts changing at once; the pond might freeze over all together, the change does not necessarily occur at once or necessarily part-by-part. So in a sense, the notion of a cluster could be present in the 'all at once'. 39Simplicius, In Cat. 2.8-9 (Kalbfleisch)

51 all a Stoic expression, (2) Dexippus usually seems to be following Porphyry instead of reporting him, 40and (3) as Long and Sedley point out, the information preserved about the criterion of truth debates ascribes some kind of 'concurrence of impressions' theory to the Academic Carneades who elsewhere pinches Stoic ideas for skeptical purposes 41 So the question of who Dexippus is reporting is rather difficult to settle; there remains the possibility that some Stoics or Stoic- influenced philosophers are represented. Now we will bring identity considerations into the attribution dispute. Recall that for help on the question of what differentiates individuals from each other, on what it is about e. g. Socrates that determines him uniquely, the Stoics turn to the peculiar quality. Particulars, in Stoic terminology 'particularly qualifieds' (i&( ,)q Ttotöv), have a quality according to which individuals can be uniquely identified 42 So far, so good. But there is no explicit mention of particular qualities in what Dexippus offers, there is only the qualified individual (iöIo); Kot6v). 43 Whether or not the Stoics or some Stoics ever had such a notion of particular qualities, Sedley and Lewis point out how problematic such a notion might be» it just won't do the job of guaranteeing identity over a lifetime. The assumption of Lewis is that the Stoics are concerned with necessity as distinguished from contingency, that contingently possessing the relevant quality

40Thanks to Richard Sorabji for help on the relationship of Dexippus to Porphyry. 41Sext. Emp., Math. 7.176-184 (Mutschmann) = LS 69E 42Interpreters have upheld a relevance of Stoic particular qualities to the Hellenistic debate between Stoics and Skeptics on distinguishing individuals by means of the 'impression' or 'presentation' (¢avtaaia). As Lewis, "The Stoics on Identity and Individuation, " 91 puts it, "the subsequent debate between the skeptics and Stoics centers on the plausibility of such peculiar qualities [which peculiarly qualifieds have], and the cogency of theories of phantasia which preserve these unique marks. " The evidence for the leap from the 'impressions' debate to Stoic peculiar properties is not as solid as one might like. As adduced by Long and Sedley, The Hellenistic Philosophers, 1.251 the best evidence for this relevance seems to be Sext. Emp., Math. 7.247-252 (Mutschmann) = LS 40E. According to this report, the 'cognitive impression' (xatakilictud ¢avTacia) seems to grasp the 'properties' (Tä zcepi. avtoiS i&taiµara) of the object (vnoxeiµevov) which impresses. The 'properties' of 'impressors' or objects (Tä i& thpaTa tv Qavtaaicöv) are presented to the impression. But there is also the idea that some sort of 'peculiarity' (Tt cotoWrov i&icoµa) which the 'cognitive impression' has distinguishes it from other impressions. 43Perhaps this is one reason why Long and Sedley, The Hellenistic Philosophers, 1.174 caution that the Stoics may only have suggested that the particular quality of an object consists in some unique collection of common qualities. In his classic article on the Stoic criterion of identity (reference at footnote 32), Sedley says that he doesn't know how seriously to take Dexippus' report. 44Sedley, "The Stoic Criterion of Identity, " 261. The argument is expanded by Lewis, "The Stoics on Identity and Individuation, " 93-96.

52 over a lifetime isn't good enough 45 If these points about identity hold, there is another reason for doubt that Dexippus is presenting a Stoic view. According to the theory in Dexippus, the particular quality of an individual is constituted by a unique combination of (unessential, shared) qualities. 46 There is no mention of names or signification. The idea as introduced by Dexippus is about identity, about marking off individuals; it is not a theory of the meaning of proper names which is concerned with what speakers know in uses of names. And what Basil presents is an ancient forerunner of the modern description theory of names 47 All variants of the modern description theory of names (there are significant differences within the camp) take their starting point from the question, "How are proper names used to refer successfully to objects? ". According to John Searle, a 'classic' description theorist, the descriptive content associated with a name does the work of securing reference. In his first attempt to present his version of the theory, Searle argues that the use of proper names presupposes a set of characteristics, and that reference succeeds if and only if the object has some or all of these characteristics 48 Searle qualifies his position by the looseness of the criteria for applying proper names; he doesn't think, for example, that the set of characteristics constitutes a definition. But although he makes a distinction between d? finite descriptions and proper names based on the insight that the former refer by specifying what the object is while the latter don't refer by describing objects, he still claims that proper names are (loosely) associated with the characteristics of the objects to which they refer. Saul Kripke's famous critique of description theories brings in the notion of the rigid designation of names, nowadays widely accepted in philosophy 49 Kripke argues that certain intuitions show that proper names are rigid designators, and that failing to account for this is a shortcoming shared by adherents of the description theory of names. Reference is not achieved by the satisfaction of a

45Lewis, "The Stoics on Identity and Individuation, " 95 46There is no hint in the text of Dexippus as to whether the cluster, or any quality in the cluster, is essential. 470f course, Basil talks about properties in the text. But there is no problem in talking about properties and descriptions in a parallel way. In modern semantic theory, every property has a corresponding description, and any property can be formally correlated with a description. 48John Searle, "Proper Names, " Mind 67 (1958): 166-173. At least with respect to terms for kinds, Hilary Putnam, "The Analytic and the Synthetic, " Maid Langtrage and Reality, Philosophical Papers, vol. 2 (Cambridge: Cambridge University Press, 1975): 33-69 at 51 ff., looks like a description theorist similar to Searle.

53 descriptive condition or conditions associated by speakers with proper names. Once assigned to an individual by a baptism or dubbing, proper names refer to the same object in all possible worlds; if things were different than they are in the actual world, the proper name would still denote the same individual. Kripke also argues that there is an important distinction between a theory of meaning for names and a theory of reference. Giving the meaning of a designator (a proper name or a description) is not the same thing as fixing the reference of a designator. A description may be used as a synonym of a designator, or a description may be used to fix the reference of a designator, but these two uses should be carefully distinguished 50 The descriptivist theory of reference is by replaced by a 'causal chain' theory of names: Uses of names to refer succeed virtue of connections between names and name-bearers as recognized in a linguistic community. Names are initially assigned to individuals in various ways (typically for personal names by parental naming ceremony), then the reference information is passed on from speaker to speaker in a reference-preserving chain. The chain does the work of explaining successful reference, not any description satisfied by the bearer. In Basil's theory of names, what sort of properties are listed for the meaning of 'Peter' and 'Paul'? Do these properties when taken together constitute a unique combination of properties which are nevertheless shared by other individuals? Do these qualities only characterize Peter and Paul, or do other men have them? One might wonder whether any property in the list does the job of marking out Peter as well as any other, being simply the description which is associated by a speaker with the use of the name. At least on the surface

49Saul A. Kripke, "Naming and Necessity, " Semantics of Natural Language, eds. Donald Davidson and Gilbert Harman (Dordrecht: D. Reidel Publishing Company, 1972): 253-355. 50To support these claims, Kripke looks at what happens when a description or descriptions is used to yield the reference of a name. If the name for an object is equivalent in meaning to (is a kind of shorthand for) the description or descriptions associated with the name by speakers, as in the case of a definition, the name is not a rigid designator. For if things were different, some other object might satisfy the description and thus the name would not refer to its (actual) referent. Kripke argues that this offends our intuitions about naming, in that proper names refer to the individual designated in the actual world in every possible world in 'which the individual exists. Even if the description associated with the name is merely used to fix the referent of the name in an initial baptism, the object picked out by the description in the actual world will be rigidly designated by the name. And if an object is characterized by one or more essential properties, reference is secured in an initial baptism by some contingent mark of the object. So the proper name designates its (actual) referent whether the relevant individual in some possible world satisfies the reference-securing description or not.

54 of the text, no property signified by 'Peter' is set apart as more important to distinguishing Peter than any other. But on closer inspection, each list builds up to a property which is New strikingly unique to the individual, given what is said about him in the Testament. Moreover the combination of properties in the cases of Peter and Paul is actually unique, although most of the properties ascribed to them are actually shared with others. In the case of Paul, much fuss is made in his canonical correspondence about his unusual personal history and special, 'late' apostolic status which nobody else has. And the progression of the list of his properties seem to reflect this sense of uniqueness, capping it off with his (generally recognized) claim to be the 'apostle of the Gentiles'. Peter is pictured in a similarly unique way, for according to the Bible he and he alone is the rock upon the church is built. The last two items in each cluster also would save Basil from the charge of blatant circularity if not from some other form of circularity. For in both catalogues of properties, every description save the final two contain proper names. In the last two items of the Peter catalogue, 'the apostolate' (tiiiS änoatoXýq) doesn't look like a proper name (though beware the easy slide to 'the apostolate of Christ'); while 'the church' (%jq EiKXIJaiaS) might be a proper name (but is indeed liable in English and Patristic Greek to being equivalent in meaning to 'the Church of Jesus Christ'). And the Paul catalogue contains the term 'Gentiles' (ýOvwv), although this seems equivalent to 'non-Jews'. It appears that Basil sets things up in these examples to characterize his men distinctively. In this sense, he certainly has an important element for a theory of reference in place. His readers could hardly miss the way in which the referent could be picked out by the speaker or hearer by one or more or the qualities which Peter or Paul has and nobody else has. Of course, even these descriptions for Peter or Paul, taken as what the names mean, would not secure reference rigidly. One could think of alternative situations in which Peter and Paul would not satisfy any of these descriptions, and thus by this theory, the names 'Peter' and 'Paul' would not apply to them but to other people or perhaps nobody at all 51

51Before Kripke argued against the description theories, a modem descriptivist would have assumed that there is good support for a description theory of names in the special case of famous individuals for whom one description is particularly prominent, a description that only the referent actually satisfies. Take the description 'the man who broke Babe Ruth's home run record' to give the meaning and fix the reference of the designation 'Roger Maris'. With this

55 Basil says that proper names mean a cluster of properties for the hearer (or speaker), and these very properties mark out an individual from every other individual. 52 Not a class which has properties as members, just the properties. Basil does not appear to lay weight on a 'particular quality' which Peter or Paul might have in some special sense, by which Peter or Paul can be unerringly picked out. And this seems to be the case even though the referent could be identified in the actual world by one or more of the descriptions listed for Peter and Paul. However, he does talk about the 'character' of Peter or Paul marked out by their names. It is this 'character' which does the work of marking an individual off from other individuals. Finally, some closing remarks. The ideas from the Contra Eunoinium are suggestive of an early description theory. Basil runs into some of the same problems that trouble the modern descriptivists. In fact, the peril of circularity is of concern to the description theorist Strawson, who tries hard to find a way to avoid the circularity of using terms which designate individuals within propositions which in turn fix the reference of terms which designate individuals 53 It is also a serious concern for Kripke, who advocates a non- circularity condition for naming theories: The properties which correspond to a given name must not themselves involve the notion of reference in such a way that it is ultimately impossible to eliminate 54 Basil presumably thinks that names refer to objects, since 'Paul' is clearly the name of Paul; he says that when we hear 'Paul' we think of a suitable cluster of properties which Paul has. But he also needs to explain how Paul is referred to by the name 'Paul'. His idea is that the proper name designates the 'character' (Xap(XKT1jp)of Peter or Paul. It is possible that this idea derives from the

you have something similar to the 'Paul' example, if successful reference had been Basil's primary concern, by the single description 'apostle of the Gentiles' (abstracting away the others) which only Paul meets. And even Kripke, "Naming and Necessity, " 302 sets the names of famous people apart as a special case. 52He does not bother to explain whether the properties are corporeal or incorporeal. Of course in antiquity only the Stoics, apparently, were keen on corporeal qualities. 53P. F. Strawson, Individuals. An Essay in Descriptive Metaphysics (London: Methuen, 1959): 180ff.. Strawson (192-193) readily concedes that the statements upon which uses of proper names are often based contain expressions which designate particulars; there will thus be a regression. However, Strawson counts on heading off an infinite regression by eventually arriving at a certain type of expression which does not contain a part which designates a particular, even though the proposition as a whole designates a particular. The required type of sentence would take the form, for example, of "there is just one so-and-so there. " All this leads into his discussion (202ff. ) of 'feature-placing'. 5'}Kripke, "Naming and Necessity, " 283-286; 292-293

56 'character' of names in the grammarians. For as argued above, in at least one passage of Apollonius Dyscolus the term 'character' is used for certain kinds of qualities or states which are signified by names, such as the soul or its contents. The word is not always used solely to talk about letters or word-forms. So perhaps there is some grammatical background to the notion of 'character'. Basil does not explain very well how reference is secured, and what the descriptions which names mean have to do with that. Eunomius could reply that proper names refer to individuals, and all individuals are substances. Hence names refer to substances. Of course, Basil and Eunomius dispute the meaning of the term 'substance'. And Basil's account, however incomplete, has the advantage of recognizing the importance of what speakers and hearers know in the use of names. Basil addresses the question of the meaning of proper names as well as the question of how individuals are marked out by the descriptions associated with proper names. But he does not give us a developed theory of reference. Consequently, his theory of names still looks a bit removed from the modern theorists who focus on how names refer 55

55This chapter has benefited in many ways from the written and oral responses of Richard Sorabji, Keith Rossack, Anthony Savile, Jan Opsomer, and Theo Ebert among others.

57 2. Prepositions and Conjunctions

Modern work in philosophy of language and semantic theory has addressed the problem of the meaning of the expressions classed as 'logical constants', which typically include expressions such as 'if', 'and', 'or', 'not'. ' One issue, which should be carefully distinguished from the question of what is the meaning of the 'logical constants', has proved particularly interesting to philosophers and linguists? This issue is how does the literal or 'weak' meaning of the logical connectives relate to other uses of the expression which differ from the literal meaning. This question involves the highly contested relation of linguistic meaning and the pragmatics of natural languages, pioneered by Paul Grice.

In ancient philosophy, there is a notion of linguistic norms for uses of conjunctions, although allowance is made for deviations from the norms in ordinary speech. As part of grammatical discussion of the parts of speech, the Stoics and later grammarians paid special attention to conjunctions, words which are vital to linguistic structure by linking pieces of discourse together. This suggests the importance of the theory of prepositions and conjunctions, from its Stoic origins onwards. As will be seen, it appears that the Stoics talked quite a bit about conjunctions, among which they counted prepositions. For Chrysippus,

1There has also been extensive debate over delimiting a set of expressions as the 'logical constants'. There is a helpful overview of the debate by Mark Sainsbury, Logical Forms (Oxford: Blackwell, 1991): 312-325. An important problem that conjunctions pose for understanding Stoic theory is that is is hard to see what answers to conjunctions at the level of what is signified. By contrast it is easier, at least in general terms, to see from the evidence what corresponds to words such as names and verbs at the level of what is signified. Consideration of these points would lead too far off the course of this chapter, so the focus will remain on uses of prepositions and conjunctions.

58 conjunctions account for one of the five parts of speech which also included proper names, appellatives, verbs, and articles. For the Stoics, there would be a need for some treatment of the semantics of conjunctions just as for every other part of speech. Several scholars of Stoic grammar have noted the Stoic interest in conjunctions. Why do the Stoics pay particular attention to conjunctions? Probably because they play a central role in Stoic logic. For example, the various kinds of compound propositions are defined in terms of the conjunctions which are used in their formation. So the Stoics have technical terms for, say, hypothetical propositions or conditionals, but standardly flag them with the word 'if' when introducing them into discourse. It appears from the surviving evidence from the Stoics that there was deep concern with the irregularities of use by speakers (and writers) of Greek. In particular, it was noticed that at times a preposition is used in a sentence in a grammatical role that is normally occupied by a conjunction. The problem of whether Stoic conjunctions have signification has received some attention, but beyond this there has not been much discussion of the semantic properties of the Stoic conjunctions. And almost nothing has been said about what is most relevant to the Patristic writing on the topic, the theory of transferred uses of one part of speech for another (prepositions for conjunctions). The Stoic approach to the linguistic phenomena was inherited in altered form by the grammarians, who developed the Stoic grammatical notions into sets of observations with Homeric examples about the possible replacements for conjunctions. After describing the evidence for the original Stoic interest in certain features of the use of prepositions and conjunctions, there will be some remarks on the grammatical tradition followed by the Patristic approach to the semantic problems. It is impossible to tell how the Stoics dealt with the problem of explaining how the standard meanings of prepositions relate to their unusual, conjunctional uses. But at least the general interest in such uses, originating with the Stoics, seems to be inherited in a distant way by Patristic writers. At this late stage, points of linguistic usage for prepositions and conjunctions were taken up into exegetical and theological arguments of fourth-century Patristics. We will see how this Patristic interest relates to contemporary work in philosophy of

2Unfortunately, there is not enough space here discuss adequately how the meaning of conjunctions as a linguistic category is determined, although I think it is of importance for understanding Stoic theory.

59 language on the meaning of conjunctions.

I

First we will see how the Stoics got the ball rolling. The original Stoic concerns have deep roots in their philosophy of language. The explanation of changes in or preservation of meaning under substitutions of prepositions for conjunctions can be introduced in this way. We need to look at the best sources on Stoic conjunctions. An important piece of evidence for how the Stoics thought about the semantics of conjunctions and the conjunctional use of prepositions comes from a Posidonius, later Stoic from the first BC report of a and second centuries .3 Posidonius, a major and influential figure in the history of the school, is cited from his work 'On Conjunctions' (11F-piavvSgaµo)v). This passage is preserved in the treatise 'On Conjunctions' of the Imperial second-century grammarian Apollonius Dyscolus. 4 The text presents Posidonius' ideas about conjunctions. In the passage Apollonius is arguing that conjunctions and prepositions are different parts of speech and should be kept distinct, against the view championed by Posidonius. Ho6Et56)vtos Ev Tw IIEpi auvSEaµcov ävTtX9ywv 7tp6q Tons 4 th KovTag wS of 6v0eßµot ov & Xoi t µ9v Tt alk6 U µövov tf v ýpä6ty 6uv ix t, 411aiv StaýepEty wS TÖ EntSoivat Toi änoöoüvat, cbg Tö änatTEiv Toi npoaatTEiv Kai (Maq TtväS TotavTaq Guvth Etc, r bTj itunoi)pevog ÖTt Ev iEpog ?l you T TE np60Eatq Kai ö avv6Eßµoq, ýv yoüv Try avtrj 7rdxty Try Evavnwtäti npös Exeivouq ävttppi et 9KTi6ETat Tong Kath 4ü6ty 6uv6e6µouS, 4äßxwv wS Stä AV Süvaµtv... µ6 oaµev contx6v- 1 yap 4p6615 napi teito [Eis] Tö xaXELv a v&Eßµov. äß,A' ovittt ev Tw Iva e ogev äµýcw. ö pa µEV wS rjv, xpovtxöv ýnipp-Qµa ö0pa nEnoIO l q, 6ÜV8E6ROS änoteXEaTtK6q. Ti oüv xwXvet [Kai] Kath TovTOV Töv Xöyov Kai. Tä irnIv Svvaµty ävaSESEyµCva auvöE6µotq µöpta xaXety auvSEagouc 6µo4cwvovvtaq npo6Cacßty, of gý v, wS ESöKEt Tw IIo6Et6wviw, Kai Tä5

3Karlheinz Hülser, Die Fragmente zur Dialektik der Stoiker (Stuttgart-Bad Cannstatt: Frommann- Holzboog, 1988): 4.1844 sets for Posidonius of Apamea the dates ca. 135-51/50 BC. 4Ap. 1. lxiv-lxv Dys., Conj. 214.4-20 (Schneider) = FDS 583 = Edelstein and Kidd F45. Hülser, FDS, Edelstein thinks that this fragment as well as Ap. Dys., Sync. 487.11-488.4 (Uhlig) = FDS 589 = and Kidd F192 do not derive from Posidonius the Stoic from Apamea. As will be seen below, this 2nd position is not very sound. 1. G. Kidd and L. Edelstein, Posidonius. Fragments and Commentary, Apamea ed. (Cambridge: Cambridge University Press, 1989): 2(i). 200 argue for Posidonius of as the author on account of the strong Stoic context, the emphasis on meaning, and the philosophical character of Posidonius' argument.

60 [sc. 1tpo6£GEi. S] ýv ic-i Cotov-w 871%o1)µevwauv8k6µoug xa%Cty, EntSoüval, äito6oüvat 5

Posidonius, who in the 'On Conjunctions' argues against those who say that conjunctions don't signify anything but only bind together the expression, says that [conjunctions] differ [in meaning] as for example [the verbal expression] 'to give in contribution' is not equivalent in meaning to 'to give what is due', 'to require satisfaction' is not equivalent in meaning to 'to require in addition' and some other constructions of that sort, advocating the opinion that the preposition and the conjunction are one part of speech; in the same sharp polemic against those [opponents] he sets out the conjunctions according to nature, asserting that on account of the that hiria is locative for the expression meaning ... we say a adverb, forbids calling it a conjunction. But this is not the case with "in order that both of us know" [Il. 1.363]. "Now when it was daybreak" [Il. 8.66] is a temporal adverb, "to the end that you are persuaded" [Il. 1.524], is a final conjunction. So then what hinders us, by this argument, from calling the words which have taken on the same meaning as conjunctions, conjunctions which are homophonic with prepositions, yet not (as Posidonius would have it) also calling [prepositions in composition] 'to give in contribution', 'to give what is due', with such a signification conjunctions? Apollonius gives us Posidonius arguing against a rejection by some grammarians of the signification of conjunctions. The Stoic argues under the assumption that prepositions and conjunctions are a single part of speech, considering the former a subclass of the latter, reflecting the Stoic approach to the parts of speech of allocating kinds of words separately treated by Alexandrian grammarians into one of their five standard parts. 6 The examples presented (ei nöovvai, änoöoüvat) are meant to show that conjunctions are responsible for semantic differences in composition, and thus convey meaning by themselves. This initial argument seems to proceed as follows: Since prepositions have signification in

5Ap. Dys., Conj. 214.4-20 (Schneider) = Edelstein and Kidd F45 = FDS 583. Cornered brackets give Schneider's conjecture for the text where it picks up again after the lacuna, while as for the squared brackets, [eis] indicates Schneider's suspicion of the word, he thinks that it should perhaps be omitted, and [Kai] indicates Wachsmuth's bracketing which seems unnecessary. The squared brackets in the translation supply what is needed to clarify the meaning of the Greek. 6Kidd, Posidonius, 2(i). 201-203. There is some evidence besides the definition in Diogenes Laertius that the standard definitional approach in Stoic grammar centered on the syntactic aspects of conjunctions at the expense of semantic aspects, Ap. Dys., Syst. 436.10-437.4 (Uhlig) = FDS 590, where for the Stoics the 'words positioned in front' are named on the basis of their characteristic placement in constructions rather than signification: c Tä rcpOKeiJEVa p6pta Stä Iris eýatp. Xeya) övoµaaiav cutrJvEyxaTO. Ev0cv cucö Tou auvtäýc(, , tri npOOETtKflS, -njv yap Kai of iic Etoaq rcpo0c KOog exdXouv auvSea touS tdS ttpoOeaEts, ipetvov ijyrlaäµcvot a tö TTJseýatpeTOu o-uvTdýE(ocnjv 6VORaaiav 0EaOat ijrtEp dn6 tS SuvdttcU c, KaOdnEp Oi TE 6uVa7LTLKOIKai GUl tXEKTIKOt Kai of inn6kotnot. Cf. Sch. in Dionys. Thr. 519.26-520.15 (Hilgard) = FDS 592; Priscian, Inst. Gramm. 14.18 (Keil) = FDS 593. There is a thorough treatment of the classification of the Stoic parts of

61 composition, and since by assumption the class of prepositions is a subset of the class of conjunctions, it may be concluded that at least some conjunctions have signification. It looks like the text breaks off at the point where Posidonius expands his case for the signification of conjunctions into the conjunctions 'by nature', probably without drawing further from the differences in compositional signification of prepositions? Two points can be gathered from this missing argument: (1) Posidonius might have elaborated a distinction between the 'natural conjunctions' and other kinds of conjunctions such as prepositions, which the Stoics call 'preposed conjunctions' (7rpoOatuzoI crMeoEtot). This seems to agree with the statement of a Dionysian scholiast that the Stoics name prepositions as 'preposed conjunctions' but term other conjunctions the conjunctions 'without qualification' (änX(u) 8 (2) The conjunctions which are taken up are used in dialectic. If it is correct that Posidonius plunges into dialectical uses of conjunctions in the lacuna, he would perhaps be developing previous Stoic theory, said by Plutarch (in accordance with Chrysippus) to employ conjunctions most of all for the conjunctive and disjunctive Conjecture (2) is construction of compound propositions .9 also loosely supported by Apollonius' remarks directly before his report of Posidonius on the Stoic contributions to theory of conjunctions. 10 These remarks need to be examined a bit further. Some light may be shed speech with full references for prepositions, pronouns, adverbs, and participles by Catherine Atherton, The Stoics on Ambiguity (Cambridge: Cambridge University Press, 1993): 301. 7Schneider thinks that after the words ýdaKwv c 8tä Ay Süvagty (214.10), an entire leaf has been lost from the text of Apollonius. 8Sch. in Dionys. Thr. 519.26 (Hilgard) = FDS 592. This Stoic class of conjunctions must be the restricted one of Chrysippus, described in footnote 9. 9Plutarch, Plat. Qnaest. 10.4,1011a-b (Cherniss) = FDS 58& T6 SE 'coiÜ; &a%cKT zoÜ IjdXtata 6vv5eaµwv öcta8at ztpoS Täs cwv cc?;tco tdtov auva4äS Kai auµrrXoxäS Kai. St al; EV?;ets... ov µepoq %6you T6v o-üvSeaµov äß,A: öpyavöv a auv8ETtx6v äno4atvEt, Katärcep thvöµaatat, Kai auvEKTtK6v ov ndvTwv dXX, Twv oiix cutX6q XeyotEv ov. Michael Frede, Die stoische Logik (Göttingen: Vandenhoeck & Ruprecht, 1974): 73 argues convincingly that there are three types of compound proposition, two of which are mentioned here by Plutarch, recognized byChrysippus as the only types which are logically relevant. From these three Diogenes of Babylon expanded the list, as found in Diogenes Laertius 7.69-72. Hence it is likely that since the Stoics define compound propositions by the conjunctions used to form them, the Chrysippean conjunctions are just the implicative (conditional), copulative (conjunctive), and disjunctive conjunctions. Accordingly, they would be defined in one part of speech as binding together discourse. But these may not make up all of the conjunctions 'by nature' of Posidonius and certainly do not exhaust the conjunctions in the eyes of the Dionysian scholiast, as can be seen from his arguments (possibly drawn from Apollonius) against the Stoic classification of prepositions as prothetic conjunctions, Sch. in Dionys. Thr. 519.26-27 (Hilgard); 520.13-15 (Hilgard) = FDS 592.

62 on the Posidonius passage by looking at its placement near the beginning of Apollonius' short treatise on conjunctions. Apollonius is obviously mindful in his brief introduction of what other grammarians have said on the topic, some under the sway of the Stoics who apparently have written much on the topic. Such writers are said to use terms (6v6µa(Yiv) which are not among the terms needed for the purposes of the grammatical art; rather they reflect Stoic concerns, a tradition (map(iöoatq)which is of no use in the technical language of grammar (ypaµµattx1j zeyvo?oyia). As has often been pointed out, Apollonius must be talking about Stoic theory of conjunctions for the purposes of dialectic. It is easy to see how Apollonius would find such discussions to be not so helpful in the grammatical tasks of correcting, explaining, and preserving the received canon of classic works. But he intends to keep in mind what the Stoics have to say when he writes his short treatise. " Apollonius says that he wants to carry out his scholarly task while choosing what is useful from each of his predecessors, adding that his approach will proceed by "thinking matters through with the necessary clarity, not entirely setting aside the theory of the Stoics. "12 Immediately thereafter, Apollonius begins the report of Posidonius, so it is natural to suppose that this Posidonius is simply one of those who has been said to "introduce Stoic opinions" on the topic of conjunctions. Otherwise one must assume that some rather obscure grammarian named Posidonius is in question who prefers the Stoic definition and argues against standard Stoic views, not the famous first-century BC Stoic philosopher. 13 Now before the gap there is a forecast of what probably is some conjunctive (non-compositional) uses of prepositions, and when the text picks up again, we find Apollonius adducing some non-conjunctional along with conjunctional uses of adverbs. In view of this, it is plausible that in the lacuna Posidonius justifies lumping together prepositions and conjunctions with some

10Kidd, Posidonius, 2(i). 202. Kidd is confident that in the missing bit Posidonius talks about dialectic, partly on account of some arguments found in the text after Apollonius leaves aside the views of Posidonius. 11Especially when Apollonius turns to logical uses of conjunctions, Conj. 251.27ff. (Schneider). Apollonius finds the Stoic approach simply inadequate; he says that their treatment 'on vocal sound' (nepi. ¢(avng ?. yoS) by his lights "hardly scratches the surface", Conj. 213.8-14 (Schneider). 12Ap. Dys., Conj. 214.1-3 (Schneider): npös ois Kai airToi Tt emvoijaavTeq µF-TexTrjS Seoýý; aa¢rlveiaS napa&waoµcv, oUK EKTÖSytv6pEV0t Kath TÖ iravTEXEq 111; TCOV£T(AtK(AV 50411S. Clarity is cardinal among the Stoic virtues of speech, a point covered in Stoic rhetorical theory by Catherine Atherton, "Hand Over Fist: The Failure of Stoic Rhetoric, " Classical Quarterly 38 (1988): 392-427.

63 examples of conjunctional uses of prepositions. Somehow, Posidonius will show on the basis of such uses that prepositions can be termed conjunctions. 14 On the other hand, Apollonius thinks that conjunctions in prepositional (or adverbial) clothing, as in his example of the conjunctive use of the adverb öýpa in the sentence "to the end that you are persuaded" (öcpa ne7toIOi1), are not much of an argument for the joint classification. 15 Rather, conjunctive uses of prepositions can be considered cases of homonymy; in different uses, the same word can serve as a preposition or alternatively as a conjunction. 16 There must be a rich philosophical background to the position of Posidonius. As keen Homer-exegetes, the Stoics would surely have noticed the promiscuity of usage for some of the parts of speech which they distinguish. The oddity of such usages, generously illustrated in the Homeric texts, would have to be dealt with for purposes of dialectic, the faultless speech of the wise man. 17 Surviving testimony points to such concerns as part of the Stoic doctrine of 'anomaly', which involves discussion of irregularities between expression and meaning, not the irregularity of language in principle. 18 Such an approach can be traced securely back to Chrysippus, who is reported by Varro to have written a treatise entitled 'On Anomaly' (HEpi ävcwµa?daS), in which he sets out to show that "similar things are signified with dissimilar words, and dissimilar things with similar expressions. "19 Accordingly, the Stoics engaged in careful

13This view is shared by several commentators on the text, including Kidd (reference in footnote 4 above) and Atherton, The Stoics on Atnbigttity, 306 note 77. 14Cf. the Stoic argument for the classification of definite articles as pronouns, in which (Homeric) uses of the article in place of a pronoun are adduced, Sch. in Dionys. Thr. 518.33ff. (Hilgard) _ FDS 551; Ap. Dys., Pron. 5.13ff. (Schneider), noted by Atherton, The Stoics on Antbigttity, 370 note 130. 15lneke Sluiter, Ancient Grammar in Context. Contributions to the Study of Ancient Linguistic Thought (Amsterdam: VU University Press, 1990): 116 points out that the value of arguing from substitutions of one part of speech for another in questions of classification was probably in dispute before Apollonius, judging by Apollonius' citation of his predecessor Trypho at Conj. 220.7ff. (Schneider), who thinks that such criteria are far from decisive: The conjunction 'if' (ei) can be replaced by 'follows from' (äxo%ou0ct), even though these words are clearly not members of the same part of speech. 16Sluiter, Ancient Grammar in Context, 117 17There is discussion of the Stoic view of the rationality of language and its decadence from original correctness in Frede, "The Origins of Traditional Grammar, " Historical and Philosophical Dimensions of Logic, Methodology, and Philosophy of Science, eds. R. Butts and J. K. K. Hintikka (Dordrecht: D. Reidel, 1997): 51-79 at 68-69; Atherton, "Hand Over Fist, " 413 note 49; Atherton, The Stoics on Antbigttity, 92-93; David L. Blank, "Analogy, Anomaly and Apollonius Dyscolus, " Langtrage, ed. Stephen Everson (Cambridge: Cambridge University Press, 1994): 149-165. 18Blank, "Analogy, Anomaly and Apollonius Dyscolus, " 152-153 19Varro, Littg. Lat. 9.1 (Goetz and Schoell) = SVF 2.151

64 examination of substitutions of prepositions, adverbs, or anything else where a conjunction should be, proceeding on the assumption that original norms of language stood in a relation of correspondence which transparently reflected the meanings of propositions. 20 Thus the Stoics had a notion of proper, correct uses of language from which a deviation had occurred in the seas of usage, resulting in the presence of improper or transferred expressions in the language of ordinary speakers of Greek. But what gets said, what was signified by the perfect primaeval sentences, was not quite lost. Similarly, the Stoic story holds that what gets said cannot be ambiguous; only the utterance, articulate vocal sound (X6ý ) is ambiguous 21 Rather, the stand-in words and constructions could be understood to reflect however dimly the proper configuration of speech, and on this basis there could be examination of language from a stylistic or logical point of view. 22 So far, this account of certain features of the Stoic theory of language has dodged a very important point, the controversial theory of the 'what-is-said' or 'sayable' (XcKTÖv).Fortunately there is enough ancient testimony to assist general

20Victor Caston, "Something and Nothing: The Stoics on Concepts and Universals, " Oxford Studies in Ancient Philosophy 17 (1999): 145-213 at 195-204 has shown how important these logico- linguistic concerns are to defusing philosophical problems. Caston brilliantly explains how Chrysippus' linguistic approach in his rejection of Platonic Forms (i. e. that the original justification of the Forms was to explain conventions involving certain kinds of terms) involves a strategy for avoiding reference to generic objects. The strategy is to analyse misleading tendencies in ordinary speech by means of a logically transparent restatement, avoiding unnecessary existential commitments. Such an analysis of misleading conventions of language yields sentences which signify the same proposition (5uvdµet T6v avtOv övTa), but differ only in syntax' (Nftk1 auvth et), cf. Sext. Emp., Math. 11.8-11 (Mutschmann) = LS 301. It is interesting to note that in the example treated by Caston (Chrysippus' analysis of definitions as universalized conditionals), the switch from categorical to conditional form involves adding an 'implicative' conjunction (cf. footnote 9) where a conjunction should be (albeit where no preposition or adverb had been). Blank, "Analogy, Anomaly and Apollonius Dyscolus, " 164-165 discusses helpfully the Stoic 'translation' of substitutes with meaning intact as found taken up by Apollonius. The Stoic method is adapted by Apollonius from Stoic dialectic, where "the comparison of a construction to the construction from which it was 'translated' thus constitutes a sort of normalisation of the surface structure of the sentence... in logical analysis, normalisation of the sentence's surface structure enables the Stoic to read off the logical structure from the expression. " 21Diogenes Laertius 7.62; Atherton, The Stoics on Ambiguity, 136ff.. Atherton (235ff. ) also discusses at length the cluster of knotty questions having to do with the lack of a distinction in the Stoics between particular uses of words, phrases, and sentences on the one hand, and linguistic and semantic generalisations on the other. Modern linguistics has recourse to notions of tokens and types, or forms of lexemes and lexemes; the Stoics only seem to have utterances which are individuals, in fact material objects in the form of bits of sound. Hence for the Stoics it doesn't make sense to consider a XEýtq to be ambiguous, but uses of it are ambiguous or unequivocal as the case may be. 22Atherton, The Stoics on Ambiguity, 93. Many thanks to David Blank for help by personal communication on the Stoic approach to theory of language.

65 Stoics understanding of this semantic level whereby the early and some of the later Stoics talked about the meanings of sentences. It would be wrong to saddle the materialist Stoics with meanings which are 'beings' in the mind or the world. For the Stoics, anything ontologically robust enough to be a 'being' is a corporeal

object. Stoic doctrine holds that XEith are classed as a 'something' in the same ontological category as a few other 'somethings', namely place, time, and void. But they certainly hold that the 'what-is-said', which is reported to correspond to the 'rational impressions' or thoughts of the mind, is related to articulate speech. By their definition, Xexiä "subsist in accordance with a rational impression".

However, 7.exzä are not mind-dependent. The Stoics take seriously the linguistic character of human thoughts. The 'what-is-said' is the content of 'rational impressions'. Much more could be said about this topic which has enjoyed some very helpful discussion in recent work-23 But for present purposes the important point is that the 'what-is-said' is the semantic level which is used to explain how what is said by language is never lost. When something is said, the structure of meaning in language is considered as built into the 'what-is-said'. This linguistic structure cannot be destroyed, even though it may be obscured from the point of view of the language used to express the meaning of what is said. So it is helpful to regard the 'what-is-said' as the guardian of meanings, free from linguistic phenomena such as ambiguity. And in this sense, transferred forms and constructions can be evaluated with respect to an unchanging linguistic standard-24 From this standpoint it is easy to see how the Stoic sage would deal with abnormal uses of language to express something, for example a preposition used in a sentence where a conjunction might be expected. To borrow an example from Apollonius, the sage might hear the sentence "on account of it being day, it is light" (ötä to i iµ pav sivat q iS g(m). He might think to himself, "the speaker has expressed a subimplicative proposition (napaßuvy}µtEvov)25 The proposition might be more transparently expressed by the sentence, 'since it is day, it is light' (eirei hµhpa EßTi, 4c5Secmv). In this form, it is clear that the speaker has said that

23Atherton, The Stoics on Ambiguity, 44ff.; 140ff.; 250ff.; Michael Frede, "The Stoic Notion of a Lekton," Language,ed. Stephen Everson (Cambridge: Cambridge University Press, 1994): 109-128. 24Atherton, The Stoics on Ambiguity, 44 note 7 25At Diogenes Laertius 7.71-74 is found the definitions of compound propositions, among which is the 'subimplicative' proposition as defined by Crinis, which is closely related to the 'implicative' or hypothetical proposition.

66 'it is light' is consequent upon 'since it is day' and 'it is day' obtains. In truth- functional terms, what has been said is true if 'if it is day, it is light' is true and 'it is day' is true, and is false otherwise. The speaker's use of the preposition is a bit unusual, but the meaning is clear enough. " Thus the sage can re-express propositions to his liking; the sage might not always use quite the same words to express a given proposition, but surely variations would be minimal and above all what is signified by the sentence would be crystal clear. Now among the kinds of 'thing said', the Stoics were as keen dialecticians particularly interested in the proposition. Michael Frede was the first to provide a plausible explanation of the philosophical background to the origins of many traditional grammatical concepts and terminology which were built into their characterization of the elements (atotxeia) of the 'thing said', which were passed down from Antiquity. 26 In fact, he even claims that the modern term 'syntax' is the descendant of the Stoic use of the word for their syntax of 'things said'. As part of his seminal discussions of Stoic grammatical theory, he argues that expressions are related to meanings in some sort of correspondence, as sentential wholes but also with respect to the parts of speech. Thus whole sentences correspond to complete Xexioi, and the reports make clear that predicates correspond to incomplete Xeicrä, but it is obscure for some parts of speech (µepri ýe(oq)how they correspond to the elements (arotxcia) of the ý,exiöv. Crucially, Frede points out that it is hard to see how the Stoics would understand the correspondence between parts of speech and the elements of 'that which gets said' in the case of conjunctions and articles. 27 In the case of conjunctions, it is clear that the standard way to join together two complete propositions without relating the propositions to each other is to use a sentence which places a 'copulative' conjunction such as 'and' (Kai) between the two simple propositions. In the case of 'and', perhaps the Stoics would allow that we are close to the pristine form of expression, for no other kind of word can do this. However, other types of conjunctions, such as the 'subimplicative' conjunction Enei in the example above, might be replaced with another kind of word such as a

26Frede, "The Origins of Traditional Grammar, " 51-79; "Principles of Stoic Grammar, " The Stoics, ed. John Rist (Berkeley and Los Angeles: University of California Press, 1978): 27-75. 27It is useful to keep articles in mind when studying the Stoic conjunctions. Apollonius Dyscolus brackets off these two parts of speech together as words which must be used with other parts of speech (äci auaanµaivct), as in Synt. 13.8-14.4 (Uhlig). The examples illustrate that by themselves

67 preposition or an adverb to perform the same function. And in some cases, this can occur with no change in the meaning of the compound sentence. Hence the links between expression and meaning are maintained, since changes at the level 28 of the expression do not necessarily go along with changes in what is signified. Now to return to Posidonius, it looks plausible from the text of Apollonius that Posidonius talks about the ways that conjunctive signification can be achieved. However, it is likely that in Posidonius' lost treatise, 'natural' being conjunctions never return the semantic favor for prepositions by used in ways typical of prepositions, although both are classed together as conjunctions (we will return to this point for Apollonius in the next section). The conjunctions enjoy a sort of primacy in that the concern seems to be with what happens when other words replace them. There is also the question of which conjunctions are the conjunctions 'according to nature', which will be treated with attention to meaning. 29 Perhaps these 'natural' conjunctions are just the Chrysippean logical connectives, which join together two or more sentences to form a compound sentence. 30 But Posidonius probably has a more inclusive class of 'natural' conjunctions, as seems to be the case for Diogenes of Babylon and his followers. It appears that the conjunctions which are homophonic with prepositions are not considered members of the set of 'natural' conjunctions, similar to conjunctions that are homophonic with adverbs, in spite of the probable usages of prepositions in the premises of arguments. For in arguments one would of course expect a 'natural' conjunction to be employed. To support this view, there is a Dionysian scholium listing the Stoic parts of speech where it is said that under the fifth part, the preposition and the conjunction are grouped together 31 The Stoics term the former the 'prothetic conjunction' and the latter the 'hypotactic conjunction' (tiöv 'Ü7IOt¬LKItK6V

(YüvScßµov). It is not difficult to see why this term is used for the conjunctions

prepositions do not signify, only having meaning when construed with nominals. For conjunctions, there is a very interesting text making a similar point at Synt. 27.11-13 (Uhlig). 28Frede, "The Origins of Traditional Grammar, " 60 illustrates this theory with the moods of verbs. 29It should be useful to mention a point of lexicography, that the word that is preserved directly before the lacuna in the Posidonius text, Siivaµis, shows in its uses an important ambiguity between meaning and linguistic function. The senses are listed in LSJ as 'capacity, function' (II. 2. d) and 'force, meaning of words' (III). 3aThe logical connectives of Chrysippus are discussed in footnote 9. The co-ordinating conjunctions are the key elements in the Stoic theory of compound propositions, including copulative conjunctions such as 'and', disjunctive conjunctions such as 'or', implicative conjunctions such as 'if'.

68 apart from prepositions. Virtually all types of conjunctions could be gathered under the umbrella-property 'postpositive' 32 a distributional characteristic which would parallel nicely the marking of prepositions as 'placed before' the elements with which they are construed. Why is it unlikely that the designation 'hypotactic conjunction' has anything to do with relating sentences or phrases to the main clause of a sentence? Simply put, the conjunctions do not usually function in the subordination of segments of discourse, for some are co- ordinative while other roles appear as well. And there is unfortunately no Stoic distinction, familiar to modern grammar, between co-ordinating and subordinating conjunctions. In ancient grammar, the words that are today termed subordinating conjunctions of time, place, or manner were classified as adverbs, although as mentioned above conjunctive uses of adverbs were noted 33 Based on the arguments and examples of Apollonius on different uses of adverbs after the lacuna, in the missing text Posidonius probably argues for the joint classification of prepositions and conjunctions by adducing some examples of prepositions used as conjunctions. Somehow, Posidonius will show on the basis of such uses that prepositions are really just a kind of conjunction. In the text that follows his consideration of the views of Posidonius, Apollonius goes on to explain and illustrate the changes that take place under such substitutions. As some scholars have supposed, he may even be taking up here some examples of Posidonius, when he discusses a further argument that prepositions cannot be substituted for conjunctions. Ov XeXiiOc S' ßµäs xäKEivoq Xöyog, hS oüöý Kath Tö itavts?, wig auv&&a otg civOunäyovTat. Ei Sctvä 6 öpaKaS ävO ' wv Setvä SeöpaKQ iSov yap Kai äpOpou 7Tap6µc'rcoats. xai Sit ci ii pa ýaii, ýwg Eatii' Ste( Tö 1 LEpav Eivat 46)g £ßii" Kai näXty ltapetYeöu 'tö äpOpov, xai ävtii öpta'ctxcov ä1tape11ua'tamapEXi qBrj34

But also that argument has not escaped our notice, that [prepositions] are not substituted for conjunctions without any change at all. (1) "If you have done terrible things", (2) "in return for which you have done terrible things"; mark the addition of the article [i. e. hön]. And further, (3) "if it is day, it is light", (4) "on account of it being day it is light"; and again the article [i. e. to] has

31Sch.in Dionys. Thr. (ex Heliodoro) 356.7-16 (Hilgard) = FDS 542 32Cf the 'hypotactic article' (relative pronoun) and 'postpositive adverb', Ap. Dys., Synt. 9.3-5 (Uhlig). 33Fred W. Householder, The Syntax of Apollonius Dyscolus, Amsterdam Studies in the Theory and History of Linguistic Science (Amsterdam: John Benjamins B. V., 1981): 245. 34Ap. Dys., Conj. 214.20-25 (Schneider) = FDS 583

69 slipped in, and in place of the indicative [verbal form] the infinitive has been used. We have seen previously that Apollonius, who is opposed to the Stoic joint classification of prepositions and conjunctions, allows for the replacement of conjunctions by prepositions (conjunctive uses of prepositions can be considered cases of homonymy). Here he wants to show that such substitutions involve changes. Apollonius may reflect in this text Posidonian views on uses of prepositions, and the choice of examples suggest concern with conjunctive uses of prepositions in logic. But this bit of text is difficult. First, it is odd that only the changes in the constructions, grammatical and not semantic changes, are highlighted by Apollonius: the prepositional phrase in place of the conjunction, and the preposition construed with the articular infinitive in place of the conjunction arranged with the indicative verb. Second, Apollonius is using stock examples which ultimately derive from Stoic propositional logic. But as Kidd points out, it is hard to tell whether the examples were used in Posidonius' arguments, partly because further gaps plague the following portion of text 35 Third, it is uncertain whether the argument presented by Apollonius is meant to undermine his previous reasoning which accounts for the use of prepositions where conjunctions are expected 36 It does not appear, however, that the sentences adduced by Apollonius are equivalent in meaning, a notion of which he is quite fond on many other occasions, including times when various kinds of substitution are in question 37 Why doesn't Apollonius go into this? In the case of sentences (3) and (4), the semantic difference is clearly that between the conditional and causal import of the two antecedent clauses of the complex sentences.38 As for sentences (1) and (2), some differences are also present. For clarification, let us fill out (1) and (2) to create the compound sentences: (1') "If you have done terrible things, you will get the rewards. "

35Kidd, 2(i). 204 36This is what is claimed by Sluiter, Ancient Grammar in Context, 117, who thinks that even the key Apollonian point of the homonymy between prepositions and conjunctions is rejected in this passage; that is, the idea might be that the two parts of speech should be distinguished in every respect. 37Householder, The Syntax of Apolloriius Dyscolus, 11 38The example used above which has the Stoic sage confronted with unusual language comes in handy here. Presumably the Stoics would consider the semantics of what is loosely called here the 'causal import' for the antecedent of (4) in terms of their technical notion of the subimplicative proposition, which is defined by Crinis in a way that clearly marks it off from the implicative proposition of (3).

70 (2') "You will get the rewards, in return for which you have done terrible things. " Sentence (2') features a use of a prepositional phrase which is not conditional, but rather bears a sense of recompense, the 'in return for which'. Another difference is the fact that (1) and (2) are most naturally completed into sentences (1') and (2') in which the order of clauses is reversed. The 'in return for which' clause would seem to follow its antecedent in discourse, while the 'if' clause naturally comes before its consequent clause. The sentences do at least share a sense of causal connection between the clauses. But one wonders why such rough substitutions, the 'in return for which' for the 'if' and the 'on account of' + articular infinitive for the 'if', would be employed in a Posidonian (?) argument. At any rate it seems unlikely that Posidonius, if he is indeed in the sights of Apollonius here, posits the substitution sinipliciter (xatä zö iav'raXes). For it is clear that the meaning is not entirely preserved. Still, it is quite possible that after the loss of text, Apollonius preserves two Posidonian examples of substitutions of prepositions for conjunctions. The roughness of such substitutions, and whether they can be attributed to Posidonius or not, is an interesting point. It will be further discussed next with reference to Greek grammarians (including Apollonius) who draw from the Stoics.

II

At this point we will look at how later grammarians influenced by the Stoics understand the relation of the standard meanings of words to their non- standard uses. In the evidence on Stoic conjunctions, this issue arises in discussion of the exchanges in use of prepositions and conjunctions. Did the Stoics look at it the other way round, for prepositional uses of conjunctions? It has so far been argued that Posidonius may have adduced examples of conjunctive uses of prepositions. However, there is no good evidence that there was a reciprocal Stoic argument for prepositional uses of conjunctions. This should not seem surprising, given the particular Stoic interest in the conjunctions and perhaps also their semantic characteristics. On the other hand, the notion that conjunctions are used in place of prepositions is found in the important Imperial grammarian Apollonius Dyscolus. Hence perhaps it is best to approach the question in terms of how Apollonius is understood to take up the

71 tradition of Stoic grammar with respect to such prepositional uses. Not much definite has been said in the scholarly literature on Apollonius' relation to the Stoics despite its importance to the history of grammar and linguistics. The topic of prepositions and conjunctions might be an interesting case to study. As will be seen, the argument which the evidence can support is that Apollonius develops the grammatical ideas of the Stoics to his own liking. He develops what is handed down to him by close observation of prepositional uses of conjunctions and then by presenting a thesis of reciprocal substitutions for prepositions and conjunctions. First, something needs to be said about where the information available comes from. Much of the recent work in Stoic syntax and semantics has taken seriously the supposition that Apollonius is heavily influenced by the Stoics. The treatise that is found to be most useful for this work is Apollonius' 'On Syntax'. For example, David Blank in his book on Apollonius Dyscolus writes: In his syntactical theory Apollonius did not abandon the semantic approach of the Stoa, and he too located syntactic corectness in the direct, clear expression of the significatum. The significatum was considered to be logically sound [i. e. to be non-ambiguous and to reflect human rationality], and so for the expression to be correct it had to follow the same rational rule. Thus syntactical correctness for Apollonius was adherence to a rational rule 39 Blank thinks that this view matches what is known of the Stoic views on the rationality of language, and draws the conclusion that study of Apollonius is basically the only reliable way to reconstruct Stoic syntax and semantics. Later work in Stoic semantics by Catherine Atherton treats Apollonius in much the same way, as broadly and deeply under their sway in logic and semantics 40 Blank and Atherton also note significant disagreements with points of Stoic doctrine. But so far, not much is known about what Apollonius himself knew of Stoic treatises beyond his citations of the Stoics in the 'On Syntax' and his highly interesting discussion of Stoic work on conjunctions in the preface to his treatise 'On Conjunctions'. Why does he adopt an approach to language which reflects Stoic theory to the extent that it does? Hundreds of years of development in grammar and philology had elapsed in Alexandria and other centers of study such as Pergamum since the time of early Stoic influence upon Aristarchus, Apollodorus, and Dionysius Thrax. On the other hand, a much more proximate

39David Blank, Ancient Philosophy and Grammar. The Syntax of Apollonius Dyscolus (Chico, CA: 1982): 52. The square brackets are added to spell out the point.

72 predecessor frequently cited by Apollonius, Trypho, is also thought to be influenced by the Stoics. 41 Hence material found in Apollonius cannot be securely traced back to the Stoics without sufficient warrant, although it is often helpful to adduce material in Apollonius which is not attributed to the Stoics for comparison. Now we can see why consideration of Apollonius' relationship to the Stoics is so important to understanding how the grammatical tradition transforms Stoic doctrine. Apollonius, of course, is an impressive and influential figure among later Greek grammarians, and it is helpful to consider his achievement as partly a development of Stoic work. In one text Apollonius mentions the Stoics in close proximity to a discussion of conjunctive uses of prepositions. However, it is not plausible to hold, as some have assumed, that Stoic ideas are reported beyond the point of classification for prepositions. 42 12 µ9v ovv xai xatiä itvag aUag icapa696Etg at 1tpoOt6Etg auvöEaµtxtjs auvTäýEWg yiyvov'rat TcapEµ4aTtxai, XýXEKTat A uv. Eý wv xa1 ý' äýopµýj ci. prIiat tapä ETCOtxoi toi xaXctoOat & tä itpoOETtxovS (Y1)V5g6F10US' TÖ yap EVEKa TLVOS%1 nti ; xai Stä ti a,vit ý; Wis p Ea-ti, xai TÖ ýx 'r ý paOvµiag ýv iaw E. 'riv Tw Evexa tr paOvµiaS. ov Stj oüv änF-t4aivov xai iä npoiP-i tEva µöpta, ?, ycw Tö StäTt, xaOÖTt, 1rapategäIEVa Tag irpoOe6Etg 6vvöECrglK V ý1E xEty ßtivTaýty 43

We have already remarked that prepositions convey a conjunctional structure in other constructions as well, which is why the Stoics called prepositions 'prepositive conjunctions'. For instance, "For what reason are you grieving? " is equivalent in meaning to "On account of what are you grieving? ", and 'from laziness' is equivalent in meaning to 'on account of laziness'. Therefore, it is not at all surprising that these words, I mean dihoti and kathoti, should contain prepositions and have the constructions of conjunctions. 44 Apollonius classifies 'on account of' (Evcxa), an improper preposition, and ött as conjunctions. It is clear that although what is at issue is the conditions under which prepositions are used in conjunctive constructions, the examples illustrate this on the grounds of semantic equivalence. But the examples adduced do not strongly suggest a Stoic background; they could have nothing to do with uses of

40Atherton, The Stoics on Ambiguity, 489-495 41Atherton, The Stoics on Ambiguity, 492. A complete list of citations of grammarians in Apollonius is provided by Householder, The Syntax of Apollonius Dyscolus, 5. 42Sluiter, Ancient Grammar in Context, 116-117 takes this text to present a Stoic argument for their classification of prepositions on the basis of the equivalences in meaning, comparing to the passage from the beginning of Apollonius' treatise on conjunctions. 43Ap. Dys., Synt. 457.12-458.5 (Uhlig) = FDS 591

73 conjunctions in dialectic. On the other hand, it is interesting and perhaps relevant to note that in Apollonius' discussion of Siöu later in the 'On Syntax', there is a citation of material from 'Posidonius'. This Posidonius, which as some have assumed is arguably the Stoic philosopher from Apamea, is cited regarding the composition, analogous to the composition of Stdtit, of the causal (nap aßuvancuc6S) conjunction 'since' (Enei) from the conjunction 'if' (ei) and the preposition 'on' (Cmt)45 It may be that Apollonius is reworking some Posidonian ideas about the composition and uses of conjunctions. In this text, the substitutive uses of prepositions preserve meaning quite well, in contrast to the semantically rough substitutions which Apollonius in the first argument of his 'On Conjunctions' seems to present from Posidonius. For it can easily be observed from his numerous examples of substituting one part of speech for another in the 'On Syntax' that Apollonius often argues in this way, taking care to preserve the sense of the original expression. This fact strengthens the case for attributing the 'rough' substitutions of prepositions for conjunctions in dialectical uses to the lost arguments of Posidonius; Apollonius would perhaps only take them up for the sake of convenience. For Apollonius is usually careful to choose examples that preserve meaning when he talks about the uses of expressions in the roles usually reserved for other expressions. So Apollonius seems to be particularly concerned that meaning is preserved in transferred uses. Further, Apollonius obviously likes the idea that these kinds of changes can in some cases 'go the other way', as when conjunctions switch places with prepositions in order to express what would normally be expressed by prepositions. As noted above, there is no good evidence that the Stoics show interest in such prepositional uses, apart from what one finds argued by Apollonius in his 'On Syntax'. Apollonius argues in this work from the acceptance of the substitution to its converse for prepositions and conjunctions: napEairjcaµev M Kai ä? XaX60t ws iö oixov 59, OUuµmov SE, iä tioünotc öµota, Svo g6pi Myou xaOe(YTo)Ta Eig ßüviaýty giav EmpptjµaToq ltapaXaµßäveiat, Kai ýxeivo itpo6Stwsa4ý6avTeS, 6) ei irpoOacetq ävtii ßuvöeßµwv napa?, aµßävovzat, oü ßiatov Kai Tö ßvvSEßµov ävti npoOxßecwvävQunäyeGAat 46

44The translation is adapted from Householder, The Syntax of Apollonius Dyscolus, 236. 45Ap. Dys., Sytit. 487.9-488.4 (Uhlig) = Kidd F192 (Dubium). Kidd, Posidonius, 1.175 mentions another Posidonius, the reciter of Aristarchus', found in Eustathius, Comm. ad Hwn. Il. 6.511 (van der Valk), in connection with the possible attribution to the Stoic Posidonius. 46Ap. Dys., Synt. 481.3-8 (Uhlig)

74 We have shown elsewhere ['On Adverbs' 180.13-183.4 (Schneider)] that 'to the house', 'to Olympus', and similar expressions are each composed of two parts of speech in an adverbial construction, adding the. clarification that if prepositions can be sometimes used like conjunctions, it does no violence for a conjunction to be used in the role of a [postpositive] preposition 47 Fortunately, the short treatise 'On Adverbs' is extant and the full discussion of the phenomena by Apollonius can be examined. Apollonius undertakes in the 'On Adverbs' an explanation why phrases such as oixov 8ý have an adverbial function but not an adverbial classification, comparing to the adverbial functions of many prepositional phrases. The idea is that there are uses of the 'copulative' conjunction SE in roles usually reserved to (postpositive) prepositions, namely the locative relation expressed by some prepositional phrases. AX W Tontid 6xthtc E6Tiv ýK 6uVSE6gou Toi 59; "Otit Kai 1rpo9ETtKl1 ltapaOe6Et 6uvSEaµtK1j SvvaµtS Eyyive'rat, Slä Tö Aµepav Eivag 40)S Eazt, Stä L\.tovÜ6lov lcapeyav8To ' AnoA, 2,ci vtoS, EK vlj pc. 6uµiaS at Kax%at yivovtat, coget Evsxa Tilg paOuµiag. Ti ovv ßiatov Tö Kai avvSEa tov lrpo6ETtKfjv Svvaµty änotE2, E6at; 1tpög oiS ovöe 6VTO, cE a LTCÄ£KEL 6 59 6ÜV5EcYlloc, napakaµßävFTat SE Kai Kar' Egýaaty, o'Ü6E EXa? iiaE }.lot, ouö' £%1£aipElg TtatÖ( TE. Täxa Se Kai. sOMY0)S oito g ävOunAzoll ävti TontKrg allµaaiaS. Tö 59 ßuvexztxwzaTOV, ön e4a u cazo Toi dv 2apayu yijv, XýyUU TAV oIKaÖE, äypaSE" Kai ývteÜOEV, ov auvSewv Ev Tc oixov Se, T KOl1ET0 Co1ElKCAS48

But how is the locative relation obtained from the conjunction de? [One can observe] that the conjunctive function actually comes about in the prepositional phrase, "on account of it being day, it is light", "on account of Dionysius, Apollonius came along", "from (ek) sloth, vices come about", as though [one used] "on account of (heneka)sloth". What violence is done when a conjunction carries out a prepositional function? In addition, the conjunction de does not always join together [discourse], but is also used for stress, "he has not in fact spoken to me", "nor do you even take pity on the youth" [Il. 6.407-81. So perhaps quite reasonably it has been substituted with a locative signification in this way. But the most common [explanation] is that it has imitated the locative derivation, I mean the derivation of [the expressions] 'homewards', 'to the field'; and hence, since it does not bind anything together in the 'oikon de', it is understood locatively. In this account there is again an example familiar from Stoic dialectic, the sentence "on account of it being day, it is light". This sentence is argued in the 'On Conjunctions' to be not exactly equivalent to "if it is day, it is light" for grammatical, not semantic reasons (cf. the text in Apollonius cited above on uses

47The translation is adapted from Householder, The Syntax of Apollonius Dyscolus, 246.

75 of prepositions, which may derive from Posidonius). Besides the 'symmetry' argument from the case of conjunctive uses of prepositions, the other grounds for the prepositional use of the conjunction Se is the non-conjunctive use to stress the particle of negation in ov6E. The brief canvassing of an alternative explanation suggests that Apollonius knows of other accounts but forges an original account of the syntax, possibly drawing from previous observations of alternative functions and significations for particular conjunctions. However, his explanation suffers from the fact that post-posed prepositions are far less commonly used in Greek than in accordance with their normal distribution. Given what is known of the Stoic theory of conjunctions and what can be plausibly inferred from Posidonius, we can see how such arguments could be an Apollonian extension and development of what was formulated by Posidonius (and perhaps other Stoics) in their attention to conjunctive uses of prepositions 495° Alternatively, Apollonius might be drawing here on the grammatical tradition which developed Stoic ideas.

48Ap. Dys., Adv. 181.32-182.10(Schneider) 49 Blank, "Analogy, Anomaly and Apollonius Dyscolus", 165 50Like the Stoics, contemporary linguistics has said much about uses of the 'logical connectives'. There has even been discussion of prepositional uses of conjunctions as part of a description of conjunctions in English, which attempts to formulate rules for conjunction which employ techniques of phrase structure description to determine the grammaticality or ungrammaticality of sentences, as in the work of Lila R. Gleitman, "Coordinating Conjunctions in English, " Modern Studies in English. Readings in Transformational Grammar, eds. David A. Reibei and Sanford A. Schane (Englewood Cliffs, NJ: Prentice-Hall, 1969): 80-112. Tacked on to the end of Gleitman's description (111-112) is the consideration of nonconjunctional uses of conjunctions, which are not included in the rules for conjunction. One common example used by Gleitman and other linguists is the use of 'and' where a preposition (infinitival 'to') would normally be used, as in "Walt will try and fix it. " Clearly, this sentence can be rephrased with equivalence of meaning as "Walt will try to fix it". Gleitman argues for the impropriety of this use to justify bracketing it off from her description of conjunction on two counts of unsystematicity: First, this usage is unacceptable when the verb forms take on affixes, e. g. *"Walt tries and fix it". Second, iteration of conjunction is not always allowed, e. g. "Walt will look and try and fix it", *"Walter will try and fix and send it". In greater detail, the argument for the nonconjunctional use of 'and' has been presented by Susan F. Schmerling, "Asymmetric Conjunction and Rules of Conversation, " Syntax and Semantics. Speech Acts, vol. 3 eds. Peter Cole and Jerry L. Morgan (New York: Academic Press, 1975): 211-231 at 217-221. First, it is shown that such a sentence cannot be understood in terms of coordinate structures, and not to involve logical conjunction (logical conjunction is here defined as a sentence which is a reduced form of an underlying structure in which two complete sentences are conjoined). Then it is pointed out that the verb 'try' takes a complement somehow, and 'try' cannot be followed by a null anaphor in this example. Hence 'and fix it' must be construed with 'try' as its complement. That is, the second conjunct is the complement of the verb in the first conjunct. Since 'to fix it' preserves the sense of 'and fix it' in this sentence, the argument for a prepositional, non-conjunctive use of 'and' is complete. The goal for these theorists is understanding the indeterminacy of meaning of the conjunction 'and'.

76 III

Due to the importance of conjunctions in Stoic logic, prepositions occupy the special attention of the Stoics insofar as they are classified as a sub-group of the conjunctions. Thus it is plausible that Chrysippus and later Stoics such as Posidonius were concerned with some uses of prepositions where a conjunction would assume a role in the normative expression of what is said, the level of what is signified by language. Perhaps this is the sort of thing that Seneca is talking about in his remarks (Ep. 88.42) about the Stoic preoccupation with conjunctions, prepositions, and other subtleties of language. 51 Of course the Stoics may also have talked about adverbs, verbs, and participles in the conjunctive roles of logic. The Stoic concern with conjunctive uses of other words such as prepositions should cast doubt on one recently expressed view that the Stoics are distinguished from their predecessors the Dialecticians, in that the Stoics are concerned with the logico-linguistic norms of sentences containing conjunctions at the expense of observing the logical significance of uses of propositional connectives in ordinary speech. For the Stoics may indeed be focused on establishing linguistic forms which transparently and unerringly represent the logical forms of propositions, but they are still keenly interested in the many ways things are said by speakers of Greek 52

51A. D. Leeman, "Posidonius the Dialectician in Seneca's Letters, " Mnemosyne 7 (1954): 233-240 at 237. Leeman thinks that Seneca has Posidonius in mind in all the dialectical passages in the relevant letters, but Seneca also mentions Chrysippus in these letters, e. g. with Posidonius in Ep. 104.22. 52Theodor Ebert, "Dialecticians and Stoics on the Classification of Propositions, " Dialektiker und Stoiker, hrsg. Klaus Döring and Theodor Ebert (Stuttgart: Franz Steiner Verlag, 1993): 111-127 at 125-127. "The Dialecticians, it seems, took their dialectic to be a theory of the logical features embedded in our ordinary language. The Stoic logicians of the school of Chrysippus, however, deliberately ignored the great variety offered by ordinary language to express logical relations and set out to build a logically transparent ideal formal language, a language where the linguistic form is merely the mirror of the logical form. " In addition, part of Ebert's case is built on the controversial view that Sext. Emp., Math. 8.93-129 (Mutschmann) reports the logic of the Dialectical school, not the Stoics. But even if Ebert has detected another strand of logic besides what is represented in sources such as Diogenes Laertius, it is probably still Stoic albeit perhaps more directly influenced by the Dialecticians. Interestingly, Ebert notes that Chrysippean uses of conjunctions in propositional logic tend to be positioned in a way that indicates their scope, signalling the logical form of the complex proposition: "There is always an expression placed before the first conjunct (or disjunct) signalling the logical form of the complex proposition. These standardizations introduced by the Stoics have the obvious aim to have the first word of a instances proposition indicate its logical form. It is in agreement with this finding that in all the the for the other non-simple propositions-the connecting particle is invariably placed at the beginning of the proposition, despite the fact that for the conditional, the subconditional and causal propositon, the clause containing the connecting particle can be placed in second position without violating any grammatical or stylistic rules. "

77 The Stoics seem to have thought that prepositions are closely related to conjunctions. Prepositions might have been especially interesting to the Stoics on account of their relevance to the problem of ambiguity which tends to plague conjunctions 53 For prepositions, whose basic use is to "define the character of the verbal action and set forth the relations of an oblique case to the predicate with greater precision than is possible for the cases without a preposition, 's4 are indispensable to the elimination of the equivocation of phrases. 55 Among the parts of speech, prepositions are the most Protean words of all, used in almost every function of Ancient Greek except as the predicate (though it may serve as a verb in parallel clauses), on account of the flexibilities of Greek grammar 56 In fact, a remark is preserved by a Dionysian scholiast to the effect that taken by themselves, prepositions do not signify anything; they only signify in composition and in phrases 57 So perhaps it is not surprising that the Stoics consider them to be a special type of conjunction.

IV

Now we will see how the grammatical tradition with its Stoic roots influences some linguistic ideas about prepositions and conjunctions in the Patristic writers in the work of Basil of Caesarea. Discussion of literal or normative uses contrasted with non-standard, more or less informative uses of expressions plays a role in his theological polemics. The problem is how the normal meaning of the connectives relates to other uses which differ from the normal meaning. Basil deals with unusual uses of prepositions in standardly conjunctive roles. Basil argues his points on the meaning of prepositions and conjunctions in

53Cf. footnote 66 on the semantic indeterminacy of conjunctions. 54H. W. Smyth, Greek Grammar (Cambridge: Harvard University Press, 1956): 1637. 55Cf. Galen, PHP 2.5.25ff. (De Lacy, 132.25ff.), where Galen argues against Stoic arguments about speech, charging them with employing a sophistic argument which plays on the polyvalence of the preposition cued. Galen discusses the resolution of the ambiguity by means of expression of the relevant premise with vrtö or Ex. Atherton, The Stoics on Ambiguity, 450-456 discusses Galen's approach in detecting ambiguity and defusing the fallacy. 56Smyth, Greek Grammar, 1636-1674. Some of these uses are discussed by Apollonius, e. g. adverbial uses of prepositional phrases, Synt. 482.3ff. (Uhlig). 57Sch. Vat. in Dionys. Thr. 270.5-7 (Hilgard). Bede Rundle, Grammzar in Philosophy (Oxford: Clarendon Press, 1979): 390 suggests that, like conjunctions, the semantic contribution of prepositions must be seen in terms of larger linguistic units such as the prepositional phrase, in which the preposition plays a semantic role.

78 his treatise 'On the Holy Spirit', which is concerned with the uses of such expressions in the doxology of the church. Basil writes in response to local churchgoers unsettled by his unexpected switches in the doxological and liturgical use of words used to string together the names of the three persons of the Trinity, but also in response to the linguistic arguments of his theological opponents who consider the prepositions found in biblical and liturgical formulas to support their views on the degrees of divinity among Father, Son, and Holy Spirit. The theological argument of the De Spiritu Sancto boils down to two related points. First, there is no rigid, one-to-one link between particular prepositions and conjunctions used in Scripture and the formulas of the church, and particular theological claims regarding unity and equality of the persons of the Trinity. Second, in view of this flexibility of use, it is acceptable to use the prepositions 'with' (1ieträ or (Yvv) in place of 'in' (iv) and 'and' (Kai) in the doxology of the church in order to express doctrinal orthodoxy. These technical arguments of his opponents, the Neo-Arian theologian Aetius and his school, have to do with the interpretation of a few key prepositional phrases in the New Testament. For example, Aetius had argued that St. Paul's use in 1 Cor. 8.6 of 'from which' (e ov) with reference to God the Father and the use of 'through which' (Si ov) with reference to the Son marks a ranking of natures in support of subordinationist Neo-Arian doctrine. The argument is that the meaning of the phrases reflects a difference in the natures. Basil charges his opponents with drawing on technical uses from the philosophers, and doing so in an inexpert and inconsistent way 58 A broader view of biblical uses of prepositions does not support their view, since the uses to speak of divinity are inconsistent and not uniform. In one text (Rom. 11.36), three different prepositions are used when only the Son is under discussion (ex, WE,

Besides the grammatical discussions of 'On the Holy Spirit', ideas of 'prepositional metaphysics' in which metaphysical notions of divinity are associated with prepositional phrases are clearly present in some parts of the

58V. H. Drecoll, Die Entwicklung der Trinitntslehre des Basilius von Ciisarea (Göttingen: Vandenhoeck & Ruprecht, 1996): 212-218.

79 treatise. 59 Scholars interested in the Platonism of the Church Fathers have talked a lot about Basil and 'prepositional metaphysics'. 6° However, these ideas are not inherited from the Stoics or the grammatical tradition. 61 They have little to do with the philosophical analysis of language, which is our concern.

V

Let us first look at Basil on the meaning of the conjunction 'and' in contrast to 'with'. 62 Basil has a simple theory of the divergences from normative, root meanings for these words. In his discussions of prepositions and conjunctions he is sensitive to the semantic importance of the weak and strong meanings of conjunctions, which is at issue in the modern debates between Grice and his opponents. Basil has not worked his explanation for the phenomena out very

59Gregory E. Sterling, "Prepositional Metaphysics in Jewish Wisdom Speculation and Early Christian Liturgical Texts, " Wisdom and Logos. Studies in Jewish Thought in Honor of David Winston, eds. David T. Runia and Gregory E. Sterling (Atlanta: Scholars Press, 1997): 219-238 at 219-220. 60Basil charges his opponents with an appropriation of the Platonic theory of causes along with their use of the expressions characteristic of the standard catalogue of Platonic types of causes, Sp. Sanct. 3,5.6-8 (SC 17: 264); 4,6.23-26 (SC 17: 270). On 'die Prinzipienreihe der Platoniker' and how the Platonic tradition is taken up in the controversy of Basil with Aetius and Eunomius over the prepositions there are helpful accounts by Heinrich Dorrie, "Präpositionen und Metaphysik. Wechselwirkung zweier Prinzipienreihen, " Platonica Minora (München: Wilhelm Fink Verlag, 1976): 124-136 at 133-136; Willy Theiler, Vorbereitung des Neuplatonismus (Berlin: Weidmannsche Buchhandlung, 1930): 23-34. Theiler claims that Basil gives the best and fullest report available for this principle-series, a standard item in the school-Platonism of Late Antiquity. 61There is no good evidence, despite the claims of Dörrie, for such a Stoic doctrine. A more careful argument for such a Stoic position is laid out by Sterling, "Prepositional Metaphysics, " 222-225, but in the end all that really can be adduced for Stoic 'prepositional metaphysics' is Seneca, Ep. 65.2-4. There is not very much here in terms of prepositions, and the formulation could well be only due to Seneca's choice of expression of a Stoic doctrine of causes. Another text cited by Sterling from Sextus Empiricus, Math. 10.10 cannot be considered securely as a Stoic argument for place, while a passage used by Dörrie from the late Stoic Marcus Aurelius (4.23) sounds a bit like the 'prepositional metaphysics' found in other philosophical schools by applying three prepositional phrases to nature (4n nt ), but can hardly be considered a standard Stoic formulation. 62Origen discusses several prepositions and conjunctions, including 'and'. Bernhard Neuschäfer, Origenes als Philologe, Schweizerische Beiträge zur Altertumswissenschaft (Basel: Friedrich Reinhardt Verlag, 1987): 215 argues that the Stoic definition of the role of 'and' in the conjunctive proposition is present in one text from his commentary on Matthew's gospel. The relevant bit of text reads, In Mt. 10.24 (Klostermann, 33.20-21): "For hear in these [words] the conjunction 'and', joining and forming a copulative conjunction from different transgressions. " (äxouE yap ev romnot Tov xai auV56GA0v, T0,6KOvTOSxai au utenXEyµevov ttotovvTOS Ex 5iaý6pcov 6papTnAdiwv). Neuschäfer compares this syntactic, non-semantic definition with several passages which mention the 'copulative conjunctions', among which is listed icai, from Apollonius Dyscolus and Dionysius Thrax, Ap. Dys., Conj. 220.10ff.; 252.21-23; 253.12 (Schneider); Dionys. Thr., Ars 20.1 (Uhlig, 88.3-89.2). The example of the Stoic definition of 'and' (Kai) to form the conjunctive proposition rounds off an extended presentation of several examples of technical terms which Neuschäfer (206-216) argues are derived from Stoic grammar.

80 he both well, but shows some points of agreement with sides of the modern debate. One basic difference between the ancient and the modern approaches is influence The rooted in the ancient of the grammatical tradition: ancient linguistic approach to the non-standard uses of expressions always thinks of the phenomena in terms of taking on the meaning of another expression, as in the case of a use of a preposition where a conjunction is normally used. Based on the evidence available, the ancient assumption seems to be that every sense that prepositions and conjunctions have in some use is the standard meaning of some expression or other. This assumption is probably shared by Basil. 63 The theological argument is that both 'with' and 'and' are expressions equally acceptable in the doxology of the church, inasmuch as they are interchangeable in use and indifferently used in the writings of the New Testament 64 However, Basil argues that 'with' is preferable in religious use for theological reasons, even above the more scripturally warranted use of 'and'. Basil claims that the usage (xpij(nc) of 'our fathers' licenses ecclesiastical use of the preposition 'with' 65 Basil always terms 'and' as a conjunction (rniv&e(Yµoc). The connective 'and' features an almost total lack of literal semantic content 66 Basil's description of 'and' is semantic and not syntactic in character 67

63Sluiter, Ancient Grammar in Context, 111-117 notes that this technical notion in the grammatical tradition of 'translation' (tutd%il ) of meaning from one expression to another is a way to clarify the meaning of expressions. She also (115) is aware of the Christian inheritance of the technique: "Along with many other concepts from pagan grammar, the technique of PE-rdxrlynSwas taken over and applied in various contexts by the Christian writers. " 64Cf.Basil, Sp. Sanct. 5,12.1-30 (SC 17: 282-284); 25,58.5-14 (SC 17: 456) 65Basil claims to be firmly traditional in all points of language, as is apparent in his roll call of heroes pro verbo in Sp. Sand. 29,71.1-75.52 (SC 17: 500-518), and his rather forced interpretation of the intentions of the Nicene Fathers (always matching his own) on the issue of the distinction between the crucial terms ovaia and vn6crcaatq, as well as the proper understanding of the 6gooüatov, e. g. Ep. 125.1 (Courtonne, 2.30-32) and Ep. 52.1-3 (Courtonne, 1.134-136). Manlio Simonetti among others recognizes this tendency in Basil to rework Nicaea to suit his thought, "Genesi e sviluppo della dottrina trinitaria, " Basilio di Cesarea. La sua etfi, La sua opera e il Basilianesimo in Sicilia 1, Atti del Congresso Internazionale (Messina: Centro di Studi Umanistici, 1983): 169-197 at 184. 66This characteristic has been well observed by contemporary linguists and philosophers, for instance by Alice Davison, "Syntactic and Semantic Indeterminacy Resolved: A Mostly Pragmatic Analysis for the Hindi Conjunctive Participles, " Radical Pragmatics, ed. Peter Cole (New York and London: Academic Press, 1981): 101-128 at 123. Davison points out that 'and' is an unmarked conjunction, with a minimal semantic content compared with the conjunctions 'because', 'although', 'but', 'so'. Hence it is natural that the semantic content of the conjuncts and their position in discourse are often seen by linguists as grounds for explaining in a more precise way the relation between the clauses linked by the conjunction 'and'. Rundle, Grammar in Philosophy, 399-401 has a similar discussion of the properties of 'and'.

81 He marks off the meaning of the conjunction from the meaning of the preposition 'with' 68 The word 'and' never expresses what 'with' expresses, because the conjunction has a weaker meaning than the preposition. Basil finds a way to treat the (minimal) semantic content of 'and' by its semantic contribution to the meaning of a sentence with respect to the action, not by truth-functionality or by linguistic function.

Tij öý Stä 'coü Kai avv89aµov 6vµltXox11S '« v övop ätwv, nto%%ä ra ö ý_vEpysiag µapivpta. ... µev Kai aüv8E6µoq iö xotvöv 'rijs T[api(YM(SLV" öý 6üv itpöOeai, S ¶i1V xotvwviav nwS 6UVCV8EiKVUiat. Oiov, 916WWaV ei MaKc oviav IIaü?. oS Kai Ttf16ftoq- ä? CCKai. Tv%tKöS xai ' OvAotgog äite6t6Xilsav Ko%oac acücty" ex covTcov, ötit tv tiavtiöv evijp"aav, RE µaOrjxaµev. ' Eöcv S* dKovacogev öit ßuveitXeIXav Kai auvaitaGtä? aav, ötit Kai. µCti' äXXA%(Ov'trio 7cpäýLv ýRXi pcocav irpoas8186XOriµev.

That the conjoining of names [comes about] by means of the 'and', testimonies [in the conjunction there are many scriptures]. ... The conjunction 'and' conveys the common performance of the action, while the preposition 'with' co-signifies the sharing [of the action] in some sense. For example, "Paul and Timothy sailed to Macedonia", but also "Tychicus and Onesimus were sent out to Colossae". From these [sentences], we learn that they performed the same [action]. But if we hear that they sailed together and were sent out together, we are taught additionally that they performed the action with one another. Before moving on to some philosophical points, we need to work to clarify the ideas in this passage. There is an interesting point in that Basil talks about 'the conjoining of names (zuiv 6voµdtiwv)', for his idea about the meaning of the conjunction would be much better conveyed in terms of conjoining sentences. Basil clearly means us to understand two different senses for the two words. Presumably, each word has a distinctive standard sense, such as one might find nowadays as the central sense in a dictionary entry. And the meanings for the two expressions can only be clearly distinguished by interpreting the 'common performance of the action' as agents doing the same sort of thing but perhaps not in community. This way of determining meaning is perfectly acceptable as a possibility for sentences containing predicates like 'sailed to Macedonia' or 'took

67Fortunately, for the question at hand we can deal with the Greek use of rcai in the same way as the English 'and', although the compositional use of 'with' in Basil's examples presents a minor complication to grammatical analysis. 68Basil, Sp. Sanct. 25,59.1-55 (SC 17: 458-462). Basil also gives a summary statement of this argument at 27,68.28-32 (SC 17: 490). There is an important comparison to the passages, Sp. Sanct. 25,60.1-7 (SC 17: 462); 26,63.8-19 (SC 17: 472-474), where the meaning of 'in' is contrasted to with'.

82 an exam'. However it is not acceptable for sentences containing the expressions like 'met at the party'. Such problems aside, this sense only gives us Paul and Timothy on a voyage headed for Macedonia, perhaps even travelling on the same ship to Macedonia. The preposition is interpreted in sentences as not only performance of the same sort of thing but also doing it together. Paul and Timothy are on a mission to Macedonia. They sail together and wouldn't leave port without each other. The idea is that the preposition expresses a stronger sense than the conjunction. The conjunction takes a standard weak sense69 This is how Basil deals with the semantic content of 'and', relating it to a stronger sense which is also obtained by reference to the performance of the action. This 'thick' meaning, according to the standard uses of the expressions, is expressed by the preposition. It is not what is expressed by the standard use of the conjunction 7° There are two philosophical points to raise here.

69Modern linguists and philosophers recognize two main forms of conjunction, phrasal conjunction and sentential conjunction. This has been discussed by George Lakoff and Stanley Peters, "Phrasal Conjunction and Symmetric Predicates, " Modern Studies in English. Readings in Transformational Granunar, eds. David A. Reibei and Sanford A. Schane (Englewood Cliffs, NJ: Prentice-Hall, 1969): 113-142 at 115ff.. That is, 'and' is ambiguous between two distinct meanings which are syntactically represented in two different ways. Basil's ideas about the two meanings of the two expressions 'and' and 'with' cannot be compared with this ambiguity for 'and', because it is based on modern syntactic representations which are foreign to the ancient discussion. But the idea is similar in that Basil seems to think that (1) "Paul and Timothy sailed to Macedonia" is semantically equivalent to (2) "Paul sailed to Macedonia and Timothy sailed to Macedonia. " (2) is an acceptable restatement of sentential conjunction. Moreover, Basil seems to think (although he doesn't give full examples) that (3) "Paul sailed with Timothy to Macedonia" is equivalent in meaning to the sentence (4) "Paul and Timothy sailed together to Macedonia". Both (3) and (4) are acceptable restatements of the phrasal conjunction reading of (1), and both (3) and (4) adequately translate the verb compounded with the preposition 'with' (auvEnxeuaav). 70Simplicius argues in his commentary in the Categories against the Stoic use of the term 'synonym' for two names with the same denotation, In Cat. 36.8-12 (Kalbfleisch) = FDS 630 = SVF 2.150. In the wider context Simplicius is talking about synonyms and homonyms. Simplicius states that the Stoic use of the term is improper since their talk of 'synonyms' is really talk about polynyms. "The Stoics [apply the name 'synonyms'] to multiple names that hold together, as for example 'Paris' and 'Alexander' are [names which denote] the same [individual], [which] are actually just the [names] spoken of as 'polynyms'. For the preposition 'with', which signifies a sharing in the same thing, is not appropriately used for these [names], as [seen from the examples] 'to walk about together' and 'to go on military campaign together'. " (oi Evouzoi [sc. xeiX. aat auv63vuµa] Tä nokkdt äµa i ovta övöµaza, th; IIdptq Kai ' Aket avSpo; 6 aüids, Kai 6(0X6to zto?uwvuµa X.cy61EVa" [sc. Totq 6v6paaty1Tot rota yap ij of v 71p60E6tq ollx dpg6viet, too a&rov xotvwvfav miµaivouaa, 6S r6 (Yv}trteptnateiv Kai auotpatc ca0at. ) The idea seems to be that the Stoics shouldn't have called these names 'synonyms', since the names don't 'share in' the same thing, despite the identical denotation of the names 'Paris' and 'Alexander'. In the text, the preposition signifies as a constituent part of the word 'synonym'. It is interesting that Simplicius the thinks that 'with' conveys a sense of community, seeing that the same term is used to describe preposition's meaning (xotvwvia) in Basil's discussion of 'with' in contrast to 'and'. Basil and Simplicius may be drawing on similar ideas from the grammatical tradition.

83 The first point is that the recourse to how the action is carried out in determining the semantic content of sentences which contain 'and' in such a way as to mark it off from sentences which contain 'with' does not seem defensible. One problem would be the case of existential sentences, which cannot be treated in terms of how the agents do what they do, as in the sentences: (1) "Paul and Timothy exist. " (2) *"Paul exists with Timothy. " Sentence (1) makes good sense, but (2) does not, at least if we maintain a focus on the literal meaning of the sentence. Sentence (2) is odd, because the straightforward meaning of 'exist' does not give anything for Paul and Timothy to do in community. The presence of 'with' in (2) seems inappropriate, for the meaning of 'exist' in an existential sentence only seems to be the instantiation of an object. So maybe the right thing to say is that in this case, the meaning of 'and' cannot be compared to 'with'. (If we adopt the strategy of rephrasing (2), either we have a sentence that adds more information beyond an existential sentence such as "Paul and Timothy are together", or (2) collapses into (1) and we lose the distinction between the meaning of the conjunction and the preposition. ) But the problem remains that in this case, there is no action to talk about in order to determine the meaning of 'and'. The second philosophical point is best raised for 'and'. Here the seminal work of Grice on logic and conversation seems relevant. Basil thinks that the conjunction 'and' has a basic weak meaning. A stronger meaning would require, at least in the usual cases, the use of a preposition. Basil doesn't make explicit the stronger claim that 'and' cannot express the 'thick' meaning in some uses. On the basis of this weak standard meaning of 'and', the distinction from stronger meanings such as the standard meaning of 'with' can be made. This bears some interesting comparisons with Grice on the meaning of conjunctions71 Grice argues that the logical connectives 'not', 'and', 'if... then', 'or' never express anything different from their weak, truth-functional meanings. The weak meaning is the basic, standard meaning of the conjunction. In terms of what is said, the conjunction never has a stronger meaning; however, speakers may in a particular context use the conjunction in a way that gets a stronger meaning across. Grice explains uses of these words which diverge from their

71Paul Grice, Studies in the Way of Words (Cambridge, MA: Harvard University Press, 1989): 22- 143.

84 strictly logical use in terms of his theory of conversational implicature, based on certain rules of conversation which govern the communication of what is meant from speaker to hearer. The fundamental distinction is between what is said by an utterance and what is meant or communicated by an utterance. Grice lays down the all-encompassing rule which people in conversation observe, the Cooperative Principle (CP): Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged. For example, take the sentence (3) "Blake trimmed the hedge and Shawn raked up the cuttings. " The conjunctions are related by temporal sequence, so the sentence could be rephrased as (4) "Blake trimmed the hedge and then Shawn raked up the cuttings. " Grice would reject the supposition that in (3) 'and' assumes the sense of 'and then'. Rather, 'and' has its standard truth-functional meaning in (3), while the implication of temporal sequence (the use of 'and' with a stronger meaning) conveyed by (3) is due to the conversational presumption that speakers observe the maxim of orderliness, which falls under the category of Manner. (Grice has four categories which are associated with his CP, the categories Quantity, Quality, Relation, and Manner; conversational implicatures are deduced on the assumption of observance of CP and the obeying or failing to obey the associated maxims. 72) There is thus no need to ascribe another 'thick' sense to 'and' which involves the connection between conjunctions. The Gricean approach is meant to simplify the analysis of natural language and pave the way for formal representation. The recourse to various features of linguistic meaning is to be avoided whenever possible; Grice opposes the proliferation of senses. Now Grice obviously has a notion of the standard use of 'and', namely its logical use. His notion of the weak sense of 'and' based on its logical use is quite different than determining a weak meaning with respect to the action involved in the sentence. Grice thinks that this weak sense is what is said by 'and', even though speakers may communicate a strong meaning by an utterance in context.

72Grice, Studies in the Way of Words, 31 presents the form of how implicatures are worked out: "He has said that p; there is no reason to suppose that he is not observing the maxims, or at least the Cooperative Principle; he could not be doing this unless he thought that q; he knows (and knows that I know that he knows) that I can see that the supposition that he thinks that q is required; he has done nothing to stop me thinking that q; he intends me to think, or is at least willing to allow me to think, that q; and so he has implicated that q. "

85 Does Basil think that the 'thin' sense of the conjunction predominates in all uses? Perhaps he does. But in the passage which directly precedes the passage which distinguishes 'and' and 'with', Basil considers a use of the preposition which does not have the 'full-blooded' meaning but only the 'thin' meaning which the conjunction 'and' normally assumes. This use deletes the associative (joint activity) aspect of the linguistic meaning of 'with'. ` HjciS yap äµ4ozepag Ev 'rj 'r v ntßzwv xpýßet KaTAaµßävovtec täS PAaetq, äµcotEpatq Ka pAµeOa" tv µev Sö uv Tw Ilvsvµatit ögoi og ä4' ýKatiepaq icXiipoi a8at lent6icuK6tES' Tovs Sý KaKovpyovvtag tiijv & Ostau Ent6TOlUUE68at p &XXov 516 S 7EpoKEtµ6vric T,aýscoS, ýitg AV Süvaµty tiwv rpaýwv 7tapar?, -q6iav Exovßa, oÜK£'Ct E6'Lly öixoi oc 'tot avavtiiots e1 c1ctxsip11toc (9atit R aviri i ävitXcyoµ6vri vuv tapä ävzi eaµov "Iaovyäp . ,10-6, row) tioü Kai avv? napEtXrJggevT . aity EiitEW IlaiXoS Kai EtXovavöS Kai Tip6ft0S, Kai I1aiXoq 6vv TtpoOtw Kai EtXovavcq. H yap au tXoid 'r v övojt Tmv öt' ý,KaTEpaS öµoiuic 'ti c eK4wvrj6eo S acigszat. Ei roivvv, rov Kvpiov eiirövzo; Ilac9pa Kai Tiöv Kai äytov IIvsüga, av1 tio5 EilCotµt11 IlatiEpa Kai Yiöv aüv '[w äyiw l-Ivai tact, äXXo 'rt EipriK6q KaT& AV Süvaµty Eaoµat, 73

For when we examine the two expressions current in the speech of believers, we use both ['with' and 'and'], trusting that glory to the Spirit is rendered equally by means of either. But [we observe that] those who work against the Truth are hindered more through the word under discussion [i. e. 'with'], which is faithful to the sense of the Scriptures yet not so easily misused by opponents (this [expression] is the one presently denied by them) when pressed into service in place of the conjunction 'and'. For it is equivalent to say, 'Paul and Silvanus and Timothy', and 'Paul with Timothy and Silvanus', seeing that the conjoining of names is preserved equally by means of either expression. So then, although the Lord says 'Father and Son and Holy Spirit', if I myself should say 'Father and Son with the Holy Spirit', will I have said something different with respect to meaning? Basil here says that the 'conjoining of names', which normally involves the standard meaning of the conjunction 'and', is also achieved by a special weak use of the preposition 'with'. Presumably, 'with' is less prone to serve the theological arguments of his opponents on account of other stronger uses in scriptural or theological formulas (especially in church tradition). The idea is that 'with' is in general more conducive to expressing doctrinal orthodoxy, which Basil thinks is served by the idea that the three divine persons are inseparable in the religious language of the Church, hence worthy of equal honor, hence equal in divinity. In all his treatises, Basil argues in one way or another for the belief that the persons

73Basil, Sp. Sanct. 25,59.1-15 (SC 17: 458); cf. 5,12.1-26 (SC 17: 282-284)

86 of trinity are like and equal in nature. Since Basil talks about semantic equivalence, it might seem that he is actually saying the opposite, that 'and' has a use that takes on the 'thick' standard meaning of 'with'. This reading would perhaps be more philosophically interesting, on account of the unusual characteristics of 'and'. Basil would have a theory which recognizes a distinct strong meaning for 'and' in the spirit of the opponents of Grice, who reconceive or reject the Gricean pragmatic approach to explain the phenomena. 74 Among these are the theorists who adopt the 'semantic' approach against the 'pragmatic' approach of Grice. Most 'semantic' theorists develop analyses of different uses of words by positing a semantic ambiguity; two or more senses of the word are involved, so there is no need to invoke a Gricean theory of presuppositions in communication. However, the reading of this text where 'and' but not 'with' changes meaning is problematic. It then becomes hard to explain his talk of 'the conjoining of names', as well as his statement that the preposition is "pressed into service in place of the conjunction 'and"'. Rather, this Patristic semantic analysis licenses the theological preferences

74The Gricean approach for explaining various linguistic phenomena has been criticised by several philosophers and linguists. The pragmatic treatment of 'and' has been criticized powerfully by L. Jonathan Cohen, "Some Remarks on Grice's Views about the Logical Particles of Natural Language, " Pragmatics of Natural Languages, ed. Yehoshua Bar-Hillel (Dordrecht: D. Reidel, 1971): 50-68 at 54-59; Schmerling, "Asymmetric Conjunction and Rules of Conversation, " 211-231. There are different versions of anti-Gricean 'semantic' approach. Cohen (56) does not adopt the standard ambiguity treatment but assigns only one dictionary sense to 'and' which is a stronger sense than the weak truth-functional sense, "and allows a certain feature of this sense to be deleted on occasion, in the process of composing the meaning of a compound sentence out of the meanings of its constituent words and clauses. " Cohen has in mind a single complex lexical entry for 'and' which takes into account temporal and causal relations. Cohen would then have a standard strong sense for 'and' in contrast to Basil's weak standard sense for 'and'. Problems for Cohen's position are discussed by Roland Posner, "Semantics and Pragmatics of Sentence Connectives in Natural Language, " Speech Act Theory and Pragmatics, eds. J. R. Searle, F. Kiefer, and M. Bierwisch (Dordrecht: D. Reidel, 1980): 169-203. Still another response to Grice's work (and Cohen's) is the reworked pragmatic approach of Robyn Carston, "Implicature, Explicature, and Truth-Theoretic Semantics, " Mental Representations, ed. Ruth M. Kempson (Cambridge: Cambridge University Press, 1988): 155-181. Carston points out that the Gricean 'implicature' account for 'and' suffers from the problem that the extra connotations can fall within the scope of logical operators, which should not happen for genuine implicatures. Carston also argues (159- 162) that there are serious problems with the semantic ambiguity analysis of 'and'. The phenomena of temporal and causal connotations of conjoined expressions are best explained by a theory of the pragmatic determination of propositional content which involves 'explicatures' (the truth-conditional content of utterances) in addition to 'implicatures'. The former plays a role in accordance with the point that pragmatic principles are already involved in disambiguation, reference assignment, the resolution of vagueness, and the recovery of ellipsed or unexpressed material.

87 for the use of 'with'. 75 The discussion is about certain changes in the meaning of 'with'. In these remarks on the semantics of the preposition 'with', Basil would agree with points argued by one representative of the 'semantic' approach for analysing unusual uses. The modern debates over the analysis of definite descriptions, conjunctions, and other expressions tend to analyse the 'thick' uses of expressions which usually (or often) have a 'thin' meaning. But some philosophers have developed the 'semantic' approach against Grice in the analysis of a 'thin' use which deviates from a standardly 'thick' expression. L. J. Cohen carries out this program for the analysis of 'and', assigning 'and' a standard strong meaning which involves the thematic similarity of conjuncts and some kind of connection between the conjuncts, which is deleted on occasion by context or by explicit qualification. Cohen has the example: (5) The old king has died of a heart attack and a republic has been declared, but I don't know which of these two events preceded the other nor do I wish to suggest some connection tends to exist between two such events. The additional information that the use of 'and' might have conveyed is deleted in (5), hence the mere truth-functionality for the meaning of 'and'. Unlike some other 'semantic' theorists, Cohen does not need a strong as well as a weak sense for 'and'. He avoids the proliferation of senses in the 'semantic' approach. His explanation of the weak use in (5) is much the same analysis as he would offer for sentences like "That is a plastic flower", in which the normal meaning of the noun (part of a plant that grows) is cancelled by the adjective. We could say that this represents a weak sense of 'flower', but it is preferable to allow divergences from the root meaning. 76 Unfortunately, it is difficult to determine whether Basil posits a semantic ambiguity, supposing that 'with' has a weak as well as a strong sense, or (like Cohen) he thinks that 'with' has only one sense (the normative strong sense) which is sometimes cancelled. Basil certainly speaks as if 'with' has a standard meaning which is stronger than its use in the 'conjoining of names', although this

75Since the doxological disputes seem to have been preoccupied with the use of prepositions in phrases and not sentences, Basil does not bother to argue for the semantic equivalence of "Paul and Silvanus and Timothy sailed" and "Paul with Timothy and Silvanus sailed". This is odd in view of the sentential examples in the other bit of text, which illustrate the semantic content of 'and' in terms of the action carried out by the agents. Maybe he doesn't want to deal with the ambiguity of these sentences. 76Cohen, Some Remarks on Grice's Views, " 54-57

88 may not be thought by him incompatible with an ambiguity in the semantics of 'with'.

vi

In section one and it was argued that the Stoics, or at least Posidonius, talked about the exchanges in the meaning of prepositions and conjunctions, particularly conjunctive uses of prepositions: Then the story moved on in section two to later grammarians influenced by the Stoics, who discussed the relation of the standard meanings of words to their non-standard uses. The historical point in section two is that Apollonius developed the grammatical ideas of the Stoics by his close observation of prepositional uses of conjunctions, presenting a theory of reciprocal substitutions for prepositions and conjunctions. Sections four and five took up Patristic material on exchanges of meaning. In this section, some additional evidence will be briefly discussed for the relevance of the grammatical tradition to the Patristic discussions. Earlier in his influential treatise on theological language, Basil points out the importance of questions of grammar and meaning with regard to the observation of everyday use of the words. Ov µövov Sý uni -M-SOeoXoyiag ai xpi aF-tq vc v 4wv& EnctUdcroviat, ä?,X 'r 5iß imi npoc iä vn ' äXAij7. (ov ar)µatvöteva noa,XäKtS ävtt Oi6Taviat, öiav &TepaAV tS E'repas a'µaaiav &V,[tXaµß6v, 1.77

And not only in theology are the uses of the expressions interchanged, but already [in the Greek of native speakers] they are often interchanged with respect to the things signified by each other, whenever one [expression] takes on the meaning of another. This attention to interchanges in the signification of some parts of speech in the course of Basil's grammatical training. The general features of this approach can be detected in the surviving texts from Imperial grammarians such as Apollonius Dyscolus, which discuss various uses of individual prepositions and conjunctions. The grammatical training of Basil must have introduced him to the areas of ancient grammar which are most prominent in later grammatical sources,

77Basil, Sp. Sanct. 5,12.1-4 (SC 17: 282)

89 including the theory of the parts of speech. The surviving grammatical texts tend to follow a pattern: Each part of speech is first defined, then distributional and syntactic characteristics are treated, then (usually) the meaning is taken up with examples from the standard works of literature. However the parts of speech are divided, there is consideration of the indeclinable parts of speech. Among these indeclinable words, the later grammarians of the Imperial period, Latin and Greek, pay special attention to prepositions and conjunctions. They probably included treatment of the different significations and uses of the Greek prepositions, and perhaps all the conjunctions too. They are interested in the interchanges of prepositions for conjunctions. Let us start with Apollonius' treatise on conjunctions. In his prologue to the short treatise, Apollonius mentions some of the (in his opinion deficient) ways that grammarians have written on conjunctions, among which is listed those who "only catalogued the names of the conjunctions and the transformations (1ETa?,AWF-tS) from some conjunctions into other conjunctions, that is those conjunctions which are equivalent in meaning (ioüS i(Yo&uvaµoüv'aS äXXA?,.otic) to one another. "78 This approach seems to have been widely followed. For instance, the fourth-century Latin grammarian Charisius rounds off his section on conjunctions with a list of what can be correctly substituted for several of the conjunctions previously listed under the catalogue of types of conjunctions. 79 Each substitution is illustrated with examples and accompanied by a brief explanation for the use. It is remarkable that he starts this part with the substitution in use of a preposition, the 'relative' preposition ad, for the 'copulative' conjunction atttem. The reason given for the switch to another part of speech is considerations of euphony. At the end of the discussion, Charisius mentions his source on the topic of conjunction-substitutions, the third-century grammarian Julius Romanus' extensive work 'Origins' (' Aýopµai) under his section 'On Conjunction'. 8° Charisius also reports that a similar discussion of

78Ap. Dys., Conj. 213.3-5 (Schneider) 79Char., hist. Gramm. 229.9-230.2 (Keil) = 296.21-297.28 (Barwick) 800n Julius Romanus, who is cited along with Remmius Palaemon fairly often by Charisius to add more sophisticated material to his elementary school-grammar, there are helpful prosopographical notes by Robert A. Kaster, Guardians of Language: The Grammarian and Society in Late Antiquity (Berkeley: University of California Press, 1988): 424-426.

90 prepositions is presented in the same source81 It is likely that Julius Romanus draws heavily on Greek sources from the grammatical tradition, since even in Charisius there is frequent use of Greek expressions. Another influential treatment of the uses of prepositions in the grammatical tradition is testified from Apollonius' lost treatise 'On Prepositions' (rlEpi tpo8E acv). Among the fragments gathered by Schneider are several texts which state a few meanings for particular prepositions, usually defined in terms of the normative sense of other prepositions 8283 It is argued by Dyck that the material on various prepositions in the Byzantine 'Epimerisms on Homer' is drawn from the same work of Apollonius. 84 One of his points in favor of such a grammatical source is the focus, characteristic of Apollonius, on the sense of prepositions including their meaning in composition. Another point which Dyck does not bring to bear is the recognition of the employment of a preposition in the semantic role of another part of speech (as an adverb), which could easily derive from Apollonius 85 This non-standard use of prepositions is paralleled in a scholium on Dionysius Thrax, where there is notice of the changes in prosody and a technical term for the phenomenon of 'interchange' (ävacipo4i) S6 These glosses on individual prepositions are fuller in content than the material found in other sources. It is instructive to cite part of one gloss on the preposition icpös at Od. 15.109.

o-nµaivat Sý Ev iactq ßvvO . asat ji ktaia zijv ltpös u 1rpäypa EyyüiqTa, oiov 2cpo6eöpaµs, irpornjXoc, irpoaöpo n xai E{.tii öµotov 87

And in composition, it signifies in particular the proximity to something, for example 'he charged', 'he approached', 'charge', and

81Char., Inst. Gramm. 238.15-17 (Keil) = 310.26-311,2 (Barwick); cf. 236.16 (Keil) = 307.17-18 (Barwick) 82Ap. Dys., Fragmenta 137.35-138.10(Schneider). Cf. the similar approach to the meaning of Kai in terms of its semantic equality with Te, Ep. Horn. K 146 Kai (Dyck, 446.39-40). 83These lists of some of the basic senses of prepositions are loosely similar to the catalogue of the various senses of the preposition 'in' (iv) in 'On the Holy Spirit', all of which are said to say something different about the divine Spirit, Basil, Sp. Sanct. 26,61.1-64.23 (SC 17: 466-476). 84Andrew R. Dyck, ed., Epimerismi Homerici. Lexicon AIMQMEIN, Pars Altera, SGLG 5.2 (Berlin and New York: Walter de Gruyter, 1995): 36. This valuable work (ODB, s. 'Epimerisms') is an elementary commentary on Homer, as used in the Byzantine schools. It contains discussions of parsing, orthography, morphology, prosody, etymology, and semantics. Such works on Homer were first put together in Late Antiquity, and were used for training in grammar, exegesis, and philology. 85Ep. Horn. it 118 rtcpi (Dyck, 619.66-7). Cf. the quasi-example of a preposition for an (elided) verb, it 142 napcc (Dyck, 629.81-630.10). 86Sch.Vat. in Dionys. Thr. 271.11-18 (Hilgard) 87Ep.Horn. it 109 (Dyck, 615.39-48)

91 similar [expressions]. One sense that is common in the use of the preposition is selected. This meaning, a relational meaning, is signified as part of the sense of the compounded nominals or verbs. Perhaps such a sense of the preposition would feature prominently in an appropriate lexicographical entry, even if it is not a unique standard meaning. Also, from the grammatical sources there is particular interest in the conjunctive uses of prepositions. The grammatical sources talk frequently about this linguistic phenomenon, familiar enough from ancient writers of prose and poetry. One Dionysian scholiast seems to be influenced by Apollonius or his sources on conjunctive uses of prepositions. "Eau SE özE Kcal, ävii avvBEaµcov irap& aµ(3ävovtat, wS ýaiEv ex paOvµiaq 7EERovBaSävtii toi EvsxEv paOvµiag neirov8aq $8

It is possible for [prepositions] to be used at times instead of conjunctions, as for instance we say "you suffer from sloth", instead of "you suffer on account of sloth". The passage from Apollonius' 'On Syntax' in which the Stoic view of prepositions classed as conjunctions (Synt. 457.12-458.5) is mentioned and then the examples of semantic equivalence are adduced has the same example, though lacking the verbs which complete the sentences. The presence of such ideas in the grammatical scholia reinforces the impression that it had been passed down into the school-grammar of Late Antiquity. Origen, the influential theological predecessor of Basil, is now recognized by scholars as a key figure in the history of ancient linguistics 89 The great Alexandrian was highly trained in Imperial grammar and philology. On occasion, Origen talks about the meaning of prepositions and conjunctions in his exegetical works 90 But none of the preserved texts from Origen relate to the

88Sch. in Dionys. Thr. (Commetltaritus Heliodori) 93.7-9 (Hilgard); cf. Sch. Vat. in Dionys. Thr. 269.28-9 (Hilgard) 89The relation of Origen to Stoic grammar and Imperial school-philology is covered in the introductory historical chapter, "Grammar, Rhetoric, and Philosophy". 90Origen, as noted in footnote 62, has at least one preserved discussion of 'and'. In his commentary on Paul's letter to the Romans, he distinguishes the meaning of 'from' (Ex) and 'through' (8tä), In Rom. 170.10-172.11 (Scherer). Jean Scherer, Le Conunentaire d'OrigPne stir Rom. 111.5-V.7. Essai de reconstitution du texte et de la pensee des tomes Vet VI du "Commentaire stir I'Epitre aux Romritts", Bibliotheque d'etude 27 (Cairo: Institut francais d'archeologie orientale, 1957): 170- 172. Neuschäfer, Origenes als Philologe, 207 note 496 remarks on this text. His philosophical analysis wrongly assimilates Origen's prepositional semantics to the preposition-series of metaphysical principles which were built into the school-doctrines of the rival philosophical schools of Late Antiquity: "Den origenischen Argumentationshintergrund bilden die Praepositionsreihen der metaphysischen dpXai innerhalb des Platonismus (Theiler) und innerhalb

92 particular discussions of Basil on prepositions and conjunctions. The other sources for Imperial grammar testify to the more relevant characteristic of analysis. Grammarians such as Apollonius treat the meaning of prepositions and conjunctions in terms of one expression taking on the meaning normative to another expression. In conclusion, Basil's discussion of the semantics of the expressions 'with' and 'and' is in some ways similar to positions taken in the contemporary debate on the analysis of conjunctions. Basil handles the problem of how the normal meaning of prepositions relates to non-standard uses where the normal meaning is absent by the idea of semantic reduction to the weak meaning. This weak meaning, in accordance with the later Greek grammarians, is considered as an appropriation of the meaning of another word. Hence the weak use of the preposition is explained as a conjunctive role. In this way, the concerns of Stoic grammar relate to the Patristic contribution, which should be recognized in the history of linguistics 91

des Stoizismus/Gnostizismus (Dörrie). Es bedürfte einer exakten Analyse, um genau zu bestimmen, welchen Traditionen Origenes im einzelnen verpflichtet ist. " Cf. the classic article on ancient doctrines of prepositional metaphysics by Dörrie, "Präpositionen und Metaphysik, " 126- 127. The implausibility of Neuschäfer's 'prepositional metaphysics' analysis lies in the fact that Origen's semantic distinction between the two prepositions does not amount to an ontological distinction, which is what Neuschäfer claims. Neuschäfer's interpretation of 'prepositional metaphysics' for another Origen passage on prepositions is much more sound. In a text from Origen's commentary on Paul's letter to the Ephesians, In Epil. 1.1 (J. A. F. Gregg, "The Commentary of Origen upon the Epistle to the Ephesians, " JTIiS 3 (1902): 233-244 at 234.1-235.11), it is said that the preposition 'through' (Std) signifies that which is subservient (ütrjpectxöv). 91This chapter owes much to the comments and criticisms, through several drafts, of Richard Sorabji, David Blank, Jan Opsomer, and others in London philosophy.

93 3. Relatives

The logic of relatives attracts attention in the controversies between the Cappadocians and their most formidable opponent, Eunomius. In order to make the topic easier to deal with, the discussion will be limited to relatives in the first major theological treatise of Basil, the Contra Eunomiuni. 1 This narrowed scope will highlight the significance of notions inherited from grammarians and Stoics in Basil's theory of relatives, which can be compared with ideas about relatives found in his predecessors. Before taking a look at the most important texts on relatives, there is one point to mention. A distinction can be drawn between two kinds of terms in the early treatise Contra Eunomium. These are (1) the properties which mark out distinctions ('Si(oµata) for Father, Son, and Holy Spirit in a simple and undivided divine substance, and (2) words such as 'Father' and 'Son' which are relational,

1There is general scholarly consensus regarding the approximate dating of the Contra Eunontium between 360 and 366, as one of the first of Basil's theological works. Some scholars think that the work was composed directly before the Synod of Lampsacus (Autumn 364), including Paul J. Fedwick, "A Chronology of the Life and Works of Basil of Caesarea, " Basil of Caesarea: Christian, Humanist, Ascetic, ed. Paul J. Fedwick (Toronto: Pontifical Institute of Mediaeval Studies, 1981): 3-19 at 10-11 note 57, Bernard Sesboüe, Basile de Cesaree. Contre Editions Eunonte, T. 1, Sources Chretiennes 299 (Paris: Les du Cerf, 1982): 42ff., and Volker H. Drecoll, Die Entwicklung der Trinitntslehre des Basilius von Cdsarea (Göttingen: Vandenhoeck & Ruprecht, 1996), 45-46. This currently predominant view is largely based on Ep. 223 from Basil to Eustasthius of Sebaste, which mentions both the upcoming Synod and Basil's composition of an anti-heretical polemic. Lucian Turcescu concurs with this view after entertaining an earlier date (360-361), "Prosöpon and Hypostasis in Basil of Caesarea's Against Eunomius and the Epistles, " Vigiliae Christianae 51 (1997): 374-395 at 375-376. The earlier date for Euiz. (360 or early 361) is argued extensively by Thomas Kopecek, A History of Neo- Arianisnt (Cambridge, MA: The Philadelphia Patristics Foundation, 1979): 364-372 on the basis of evidence regarding a polemic against Eunomius in Ep. 20 which Kopecek dates earlier than the Synod; but also it is based on internal evidence from the treatise (notice of the activities of friends and enemies), as well as the alleged influence of Meletius as reflected in his Antioch sermon of 361.

94 fall (1) similar to 'master' and 'slave'. 2 It appears that the properties which under by Basil be are not understood to relational .3 The topic of relative terms is first mentioned in Eun. 1.5. In this text Basil points out the important distinction in meaning which can be made between 'Father' and 'Ingenerate': '5 rlatipoq gcovljq Svvaµavrl äycvv1jticp, tp6q r6 Kai tiiv ... t Gov Tw tEpi Tov Yioi Evvotav a viµt vcoq ýcnrc ötä Tllc 6xE68(og ßuvctßäyEty. 'O yap övico; rlcrc ýp Kai µövoq Eý ovösv6q E6tty Lr9pov" to SE Eý ov5sv6g Taviöv Labt tiw äyevvrytios. Ov ioivvv äyavvrytov 7tpoaayopcu't6ov p XX.ov i1p7ty f Ilaiapa....

the 'Father' is in to ...since expression equivalent meaning 'Ingenerate', while it also introduces the notion of the Son in conjunction with with itself through the relation. For the one who is truly and authoritatively Father is uniquely from no other; but that which is derived from no other means the same as 'Ingenerate'. Therefore, we should not term him 'Ingenerate' rather than 'Father'. It is clear that Basil understands the term 'Father' to be relational (i. e. 'being a father' is a relational property) in an important sense, in a sense in which 'Ingenerate' is not. It is said to be a relational property, and as such is intrinsically associated with its constitutive relation. This constitutive relation requires that, if God the Father does indeed belong to the class of things characterized by the relational property 'being a father', there is at least one Son such that the Father is the Father the Son Basil deny that the ingeneracy of .5 would not property of as here defined, and the property of fatherhood in the case of the one who is authoritatively a father, have the same extension. That is, the properties are applied to the same object. But his point is that 'Father' and 'Ingenerate' are

2The 'properties' (i&c6gata) in simple divine substance and the logical schemes in which they play a role are discussed by David G. Robertson, "Stoic and Aristotelian Notions of Substance in Basil of Caesarea, " Vigiliae Christianae 52.4 (1998): 393-417. 3For the sake of clarity they will be discussed separately, although as will be seen Basil himself speaks of the relational character of the properties of the divine persons (i&6Lata) in an important passage (Ettn. 2.26-29). Unfortunately, a full discussion of the two types cannot be attempted here, although this is an interesting issue that is relevant to Basil's theory of relatives. Sesboüe, Contre Eunonte, 1.81 assimilates the two kinds of terms when he takes up the subject of relatives in the introduction to his edition. 4Basil, Eun. 1.5,67-72 (SC 299: 176) 51n this chapter, relations are considered as classes, understanding them as classes of ordered pairs. This is the standard approach in modem logic as for example in W. V. 0. Quine, Word and Object (Cambridge, MA: The MIT Press, 1960): 105ff.; 209ff.; 257ff..

95 equivalent in meaning (ißov Svva i vric), although that is not the whole story. In short, Basil says here that 'Father' does the work of 'Ingenerate', plus a bit more. It is important to note in passing the basis for the partial equivalence of the two terms, which can be discerned in the words, 'for the one who is truly and authoritatively Father is uniquely from no other'. Here is evidence of a theological insight which takes its classic form in Athanasius, namely the thought that words used of God taken from ordinary language such as 'father' and 'master' take their proper meanings not from below but from above. That is, one understands what it truly is to be a father not from human experience but from the the divine father It be interesting to know how revelation of .6 would Basil would further characterize 'Ingenerate', since it is said to not be relational, at least not in the sense that 'Father' is relational. For instance, one might ask whether 'Ingenerate' is devoid of all relational attributes. Fortunately, there is more that Basil has to say in the same treatise on the issue of relative terms. In Eun. 2.9, Basil draws a distinction between relative and absolute names.? This distinction is set forth in the context of an argument for the view that names cannot fully express substances; since absolute names cannot adequately express substance, still less can relative names do so. Rather, absolute names only mark out properties (i,Stwµa-ca) which are considered in (nepI divine in substance aura v), whether substances earthly or are question .8

6For discussion of this point in Athanasius and Gregory of Nyssa, there is Catherine Osborne, "Literal or Metaphorical? Some Issues of Language in the Arian Controversy, " Christian Faith and Greek Philosophy in Late Antigttity, eds. Lionel R. Wickham and Caroline P. Bammel (Leiden: E. J. Brill, 1993): 148-170 at 158-161. It is important to recognize that of the two main sensesof dy6vvrltog in the Arian Controversy, namely not generated as creatures are generated and unbegotten as not deriving from any other, only one sense is primarily in view here. 7Basil, Ettn. 2.9,11ff. (SC 305: 36ff. ) 8Drecoll, Die Entwicklung, 65 and Sesboüe, Contre Eimoore, 2.38-39 note 1 both point out the significance of the portion of text which concludes Eitn. 2.9,2.9,24-27 (SC 305: 38): Kairot ye 9txpöv 41itpoaGev eSEixviTO trap' fiµcov ött xai 'rä ärtoke?, vµt va Twv övoµätwv, xäv Tä paxtnTa SoKf vnroxeiµevöv zt öriXovv, ovx avtily rtapiarrlßt ri v o1ýaiav, i& thiia ra Se Ttva rtEpi ct ri v d4opit; et. For Basil manifestly refers his readers back to 2.4,1-26 (SC 305: 18-20), where proper names (npocn yopicn) of human persons are said not to signify substance (ovaia), but rather the properties which characterize each person as individuals. (There is a full analysis of this theory of names in "Proper Names". ) Although Basil only sets the proper names of 2.4 and the relational names of 2.9 explicitly in the class of terms which do not denote ovaia or vrröazaßtS, Drecoll argues convincingly that in this respect the absolute names of 2.9 follow suit: "Er betont daher in AE 11,9 ausdrücklich (eingeführt durch xaizot), daß auch die erste Gruppe der von sich selbst ausgesagten (= dnoXelvµeva) Namen nicht die oüaia, sondern nur die iSuöµa'ra angeben (und das, obwohl die meisten von ihnen doch ein ünoxeipcvov zu bezeichnen scheinen. )"

96 Basil first cites a bit of Eunomius' Apology, which argues that if a name is said correctly of a substance, the name signifies the reality itself. 9 This text from Eunomius is representative of the theory of language of his Apology, which does not allow for a level besides words and things signified, such as can be found in Aristotle (thoughts, affections in the soul) or the Stoics Q eK Uov, the what-is-said or sayable). The text which contains Basil's exposition of the distinction between absolute and relative names reads as follows, in response to Eunomius. ' Ertei 'tic ovic oiösv kt tiwv övogd uov Tä µßv ätoXcX, vµavwS Kai xa6' ýautä irpo4Ep6geva, -To-)v vnoxetgEvcov aüiooc 7rpaw 'rwv E6TL cr tavtitxä" iä 89, itpöS kepa XEyöµeva, tvat nv µövrlv Eµ4aivst tihv iip6q ä ?,Eyszau; Otov, ävOpmuo; gEv Ica! innog xai. ßoUg aviö i xa6Tov tiwv 6voµa&oggvwv 1Lapimijaty, ui Sý, ij Soü? oq f ýiXos µövijq'j; 7tp6; tiö ßvvetevyµavov övoµa auva4cia; Ec'c öflX ottKd. '0 Toivuv äxoü6ag yevvAµato; ovx eni titva oüc iav tiý &avoia 4eps'cat, äi, 7' ÖTt E'cep(p E6tI 6uvan'r6jsvov Evvosi. Tö yäp y6vv-gga, rtvog %46'uat 'yEvv ta. "0 VEOL of 'Üico6tÖ eth; 'twVo; 9vvotav EgTotsi, äß,Xä µövov Av mpö; E'rEpov uXE6ty änocn aiv&, tiovno oü6iav Vivat voµoOa'ety, zcc, oü 'nj; ävwTäww napalrxijýiaq . ati; i°

Who does not know that among names, some are spoken absolutely and with reference to themselves, and mean the things which are substrates, while names of a different class are said in relation to other things? For example, 'man' and 'horse' and 'ox' are used to express each thing itself of the things named, but 'son' or 'slave' or 'friend' signify only the connection with the term which is paired with it. This is why the one who hears the word 'offspring' is not directed in thought to a particular substance, but conceives that the term in question is coupled with something else. For the word 'offspring' is spoken of as the offspring of someone. To lay it down that this name is substance, this name which involves no conception of a particular substance but only signifies the relation with something else, is this not the height of dementia? This text presents Basil's usual position as regards the signification of the names for the divine hypostases, that they signify as relational terms a relation with something else. It is interesting that the argument seems to be based on an appeal to the ordinary linguistic practices of people speaking and hearing in Greek, and a quite similar appeal also crops up in another important discussion in

9Basil, Eun. 2.9,3-6 (SC 305: 36); Eun., Ap. 12 (Vaggione, 48.7-10), a&Trty eivai. tf v vnöciaanv rjv arlµaivei Toiivoµa, Erca%rl6evoü" ovßia TqS 7Tpoanyopia5. It is impossible within the limits of this thesis to give an adequate account of Eunomius' philosophy of language. The most recent and detailed treatment of the question is found in Karl-Heinz Uthemann's detailed discussion of the Apology, "Die Sprache der Theologie nach Eunomius von Cyzicus, " ZKG 104 (1993): 143- 175 at 145-156. More accessible is Kopecek, A History of Neo-Arianism, 320ff.. 10Basil, Eun. 2.9,11-23 (SC 305: 36)

97 the Contra Eunomium of the semantics of the term 'Father' (Eun. 2.22ff. ). 11 What is more, Basil claims that the notion of the relation is suitable for all offspring, 12 whether earthly (icöv 'yp-vvriOevt(iv) or divine (ioü Movoyavovs), for divinity is not degraded by use of everyday language in theology. 13 Here Basil develops his notion of the relation. Basil undoubtedly has in mind Eunomius' theory of names when he rejects the association of 'offspring' with a certain nature (uni, utva ovsiav). For, according to Eunomius in the Apology, different names must signify divergent natures, at least if the names are properly assigned and have clearly distinct meanings. There is no confusion at least on Basil's part of word and object, for in saying that relational terms are signifiers only of 'the connection with the conjoined term (tio auveýcuyµevov övoµa)', Basil maintains that certain terms are related with one another; objects are not in question. Besides the point that 'offspring' is a relative name and does not signify an object (oüaia, vToKEiµevov npäyµa, vnoardutq), note what is said here about the relation constituted by the relative name 'offspring', parallel to the case of 'father' and 'son'. Basil says that 'offspring' is spoken of as the 'offspring' of someone, but he never says that the term 'offspring' is spoken of as the offspring of a progenitor. That is, Basil does not discuss the convertibility of the relation, and in fact the converse relation is never mentioned. This raises doubts about an Aristotelian influence on relatives in Basil. Before we go further into the philosophical background, we need a bit of the theological tradition to gain some perspective on what Basil says about relatives. There is the possibility that there is influence upon Basil's theory of relatives from other theologians known to him. First of all, one must consider

11In this respect Basil reflects Origen as biblical exegete. There is discussion of the appeal to usual language' (ouvA0sta) in Origen by Bernard Neuschäfer, Origenes als Philologe, Schweizerische Beiträge zur Altertumswissenschaft (Basel: Friedrich Reinhardt Verlag, 1987): 143ff.. Origen understands the term in the two primary senses of sermo cottidianus and the normal usage of the Scriptures. Neuschäfer gives copious references for auvrl0sta in Origen, and outlines the background of the term in pagan philology. 12In the case of 'offspring' discussed here, Basil says soon after this passage, Eun. 2.10,21 (SC 305: 38), that the conception or thought (Evvota) of the term is properly öTt nap' avToü yF-vvt rw eiq Tö swat napiiXBrl, which is simply an expanded description of the relative property 'being an offspring'. 13Basil, Din. 2.10,21-23 (SC 305: 38). Basil speaks in defense of his position against the counter- argument of degrading divinity with common language (KOtvý (YUVA0Eta, KOtVÖVTÖ Tqg axEaccoc), 2.10,23-27 (SC 305: 38-40): Kai ti ciq, Enctön Kotvöv Tö Tic GXEGF- xaOaipeaiv Ttva eivat Tr1S , d 86, r1SToi Movoysvovc voµtt; ettw. Oi) yäp Ev Tw np6q Ti 7t0; eXety i1 ötacopä up 1io np6; rtä 'ka, z?X Ev Trt iötöTTtt Tnq ovaias rl vnEpoxh Toü Oeov 7tp6q Tä Ovrltä 8taQaivcTat. It is interesting that Basil in this passage uses the terminology for the relation np6q Ti n Exety.

98 Basil's early theological nurture within the Homoiousian tradition. 14'Among the extant texts penned by this party and its leaders is the letter of George of Laodicea of 359, preserved by the fourth century heresiologist Epiphanius, who says that it was written by George together with Basil of Ancyra and 'those with him'. 15 In this letter can be found an interesting discussion of relations.

14The work of Drecoll, Die Entwicklung, 16-20; 143-146; 244-246; 332-340 clarifies in a convincing and careful manner the development of the Homoiousian party along with Basil's nascent theological strategy in relation to this development. In this connection, Drecoll draws heavily on some excellent recent work an the Homoians and Homoiousians, including the research of Hanns Christof Brennecke, Studien zur Geschichte der Hom6er (Eibingen: J. C. B. Mohr (Paul Siebeck), 1988); "Erwdgungen zu. den Anfdngen des Neunizdnismus, " 0ecumenica et Patristica. Festschrift ftir Wilhelm Schneemelcher zion 75. Geburtstag, hrsg. D. Papandreou, W. A. Bienert, K. Schaferdiek (Stuttgart: Kohlhammer, 1989): 241-257; Winrich A. L6hr, Die Entstehung der honidischen und homousianischen Kirchenparteien. Studien zur Synodalgeschichte des 4. jahrhunderts (Witterschlick/Bonn: Verlag Wehle, 1986). Accordingly, some crucial theological influences and positions argued against, such as that of Eustathius, in the De Spiritit Sancto are explained in terms of a ftmdarnental split among the Homoiousians; between 'Rechtshombusianer' and 'Linkshornbusianer'. The former group, whose cause the mature Basil solidly championed, came together in 363 and was characterized by (17) "nicht so sehr eine personelle Kontinuit5t zu den Hombusianem von 358, sondem zielt auf die inhaltliche Position: Die 'Rechtshombusianer' vertraten das 6Amog icaT ' o-býav und akzeptierten clas Nizdnum. " This group was mostly comprised of bishops who had pledged in 360 the Homoian position (including Meletius), but also included battle-tested 'old Homoiousians' such as George of Laodicea and Basil of Ancyra who had crafted the original Homoiousian position from 358 on. The latter faction (17 note 53; 143-146; 244-246) split off from the 'Rechtshom6usianer' in 364, whose views were definitively expressed by the second Antiochene formula. They, including Eustathius of Sebaste, did not espouse Nicaea outright, were like all Homoiousians adamantly opposed to Sabellian or Photinian thought, and were vague and probably subordinationist in their Pneumatology. At any rate, Basil supported Meletius' claim to the contested see of Antioch after his recall from exile under Julian, and was a keen admirer of Eustathius early on (emphasized by Jean Gribomont, "Eustathe le Philosophe et les voyages du jeune Basile cle C6sar6e," RHE 54 (1959): 115-124), but broke decisively with Eustathius in 373. As Gribomont affirms (120), Eustathius took part in the Council of Ancyra of 358, and was part of the delegation from the council to the court of Constantius at Sirmium. Philip Rousseau, Basil of Caesarea (Berkeley and Los Angeles: University of California Press: 1994): 99 note 11 points out that "no conclusion demands our believing, however, that Basil [of Caesarea] was present at the synod of Ancyra. " Of the Contra Eunoinizim, Rousseau (99) asserts that "it was to be a work written from a Homoiousian and Eustathian point of view. " Unfortunately, not as much can be gathered on Basil's relation to his namesake (bishop of Ancyra) in his formative years, beyond the important fact that the young Basil attended the Council of Constantinople (360) as an assistant to Basil, a leader of the Homoiousians; Philostorgius, Hist. 4.12 (Bidez and Winkelmann, 64.3-7), says of this: rlpoe(Yr(jKc0av 5ý v5v gýv icaO' oi)aiav 6gotov 7rpEaPF_,j6vro)v, Bacyiket6; cc icai EibmdOtoq: oltq (AXot zF-Kai BaoiXF-10S gTepogirapi-1v mvaaniý(ov, bta6vcov Ert TdýW EXCOW5I)V6EAF_t gýV T01)MYEIV 7COXA.CoV irP0#P(t)V* Vp K Tq; 'YV(6gTjqelOaPOEt, ITPOq'rO'L); K0IVouq wroat6%%6pP_voqetycova;. P. Maran, Vita S. Basilii 7.1 (PG 29, v-clxxvii at xxvi-xxvii) discusses young Basil's relation to the proceedings of Constantinople, concluding that young Basil was disappointed with his mentor's conduct as well as the outcome of the council. 151,6hr,Die Entstehung, 142 note 363 thinks that the author of this Homoiousian documentis actually Basil and not George, while the earlier document from the Synod of Ancyra (358)is the work of Georgeand others. As for the dating of the 359 document, L6hr (147) concurswith the received view placing it in the Summerof 359, after the Sirmian court-synod of May 359 and before Seleuciabeginning in late Septemberof 359.

99 ttp6; colkotq , ötit TO roü Maipo Kai utov ovoµa 'n v npoq it cY EGtV "µaivEt, StO Käv natepa µövov övoµä(wiev, Exogev 'r 6v6µait toy natp6g ßuvunaxouoµev11v ti iv 9vvotav Tov vioü (naidjp yap uiou na'j p ? 97F-Tat), xäv ui6v jiövov övop. äßwIEv, gxopEv Av Evvotav tiov natp6q, ÖTt ul6q na'tpOS %1£yE'Cat. gxctiat yap 9äiEpov OaTýpou Kai ov Stalcön'tETat 1 axe tc, äX7ä Kai µövov bcdTEpov a,Eyöµsvov ouvetaäyat Oatigpou'jv 9vvotav, Kai ov iö övoµa µövov, äß,X. ä µ6i6c toi 6v6µaioq Kai q ýü6eW tiiv oiuEtöuia. naiepa yap vooüvtCg 'c6v GF-6v OEOV voou v nacEpa Kai uiöv voovvmq OE0u OEÖv vooütcv E6v piOtvtia ui6v OEov Kai öµotov ýicivw KaE& Ehv 4üc äyEvvTy SE Xeyciat äcyevv-Qtiov tv ov voci. tiat ut6 . rov ov ycvviyrov ovöe ycvvTlTÖv äycvvýiou ycvvrjtiöv. l6

In addition, since the terms 'Father' and 'Son' signify the relation 1tp6q tit, even if we use the name 'Father' alone, we have, comprehended together with the word 'Father', the conception of the Son (for the word 'father' is used of the father of a son), [and] even if we use the name 'Son' alone, we have the conception of the Father, because he is spoken of as Son of the Father. For each is connected with the other and the relation is not sundered, but even if only one of them is mentioned, either presents the conception of the other, and not only the name, but with the name the proper belonging of the nature. For in thinking of God as Father, we think of the Father of God, and in thinking of God's Son we think of him who is said to be God's Son as God and by nature like him whose Son he is regarded as being. But 'ingenerate' is not spoken of as 'ingenerate of a generate', nor is 'generate' as 'generate of an ingenerate'. 17 The Homoiousians speak here in the first part of this passage in a way that is obviously loosely similar to the notion of relative terms in Eun. 2.9, although the emphasis is clearly on the relation together with its converse relation. 18 Löhr thinks that in this passage the Homoiousians have recourse to the Aristotelian 'lcp6q'ct' relation, in opposition to the Anhomoian theory of names which favors the non-relational terms 'ingenerate' and 'generate' over 'Father' and 'Son', but does not explain what is distinctly Aristotelian about their doctrine of relatives. 19

16Ep. Georg. Laod., ap. Epiph., Haer. 73.19.3-5 (Holl, 292.4-16). The corner brackets give the conjectures of Holl. 17The translation is adapted from Philip R. Amidon, The Panarion of St. Epiphanias, Bishop of Salamis (Oxford: Oxford University Press, 1990): 308. 18This feature of the Homoiousian doctrine of relatives in Ep. Georg. Laod. is not found in Ettn. 2.9, where the pairing with another term is expressed in the case of 'offspring', "for the word 'offspring' is spoken of as the offspring of someone." The surprise here is that Basil does not say, "for the word 'offspring' is spoken of as the offspring of a progenitor. " 19Löhr, Die Entstehung, 144 note 387. Löhr seems to draw a notion of co-temporal existence into this text from Aristotle, mentioning Arius' criticisms of the 'npö; Tt' relation in the dispute with Alexander of Alexandria on the grounds that the relation entails the co-eternity of the Father and the Son, Llrk. 6.4 (Opitz, 3.13.13).

100 Presumably he has in mind the language of 'is spoken of' (X. yetiat), and the convertibility of the relations that are discussed, but the claim should not be pressed too far. Much of this text is couched in non-Aristotelian language. Then the letter goes beyond the philosophy of names and relations, adding "and not only the name, but with the name the proper belonging of the nature ('r ý 4ü(Yccoqtijv oiKstoti Ta)" as well as "by nature like (öµotov) him". The Homoiousians argue that the Son is correctly addressed as God, with the same degree of validity that the Father is called God. For they add to the relational character of the terms, the point that "with the name the proper belonging of the nature" is brought in. Doubtless, Basil would agree, but he might feel that in Eun. 2.9 the point would muddy the waters of his distinction between absolute and relative terms. The Homoiousians also claim that everything that is said of the Father can also be said without equivocation of the Son.20 Now let us look more closely at possible sources of philosophical inspiration behind Eun. 2.9. It may be observed that Basil talks about relatives that are said (XEyctat). And this is must be one reason why Sesboüe and Drecoll assume that Aristotle's doctrine of relatives from the Categories is relevant to understanding this passage,21 referring to Aristotle's first definition of relatives (itp6q'CL), "we call relatives all such things which are said to be just what they are, of or than other things, or in some other way in relation to something else. "22

20The epistle does not actually say that everything that is predicated of the Father is also predicated of the Son without equivocation, but the idea is of being 'similar' as a son is to a his father involves a list of predicates said of the Son, e. g. in a passage before the bit on relatives, Haer. 73.18.8 (Holl, 291.20-25). 21Sesboüe, Contre Ettnonte, 1.78-86. Drecoll, Die Entwicklung, 65 note 56 claims without argument, "Den Hintergrund bildet die aristotelische Trennung von ovaia und rrpös rt, " citing Aristotle's first definition of relatives (Cat. 6a36-bl; b15-27) and his second, revised definition (Cat. 8a17-24). Drecoll then adds, "im 5. Buch der Topik über das i&tov unterscheidet Aristoteles dann xa6' ain6 - npö; ETEpov(Top. 128b16ff. ). " Drecoll (57) thinks that for the philosophical background of the Contra Ennontittttt, Aristotle is the predominant influence: "Neben den Schriften des aristotelischen Organons (besonders Kategorien und Topik) sind dabei Handbücher zu nennen, allen voran Alkinoos' Didaskalikos. In Frage kommt ganz besonders die Dialektik, genauer die Disputation. " 22Aristotle, Cat. 6a36-37: npös Tt & Tä Totavta XEyeTat, öaa ai'rä ärtcp east ETepwV civat XEyETat ij öncwaovv äXXow ETCpov. In his introduction, Sesboüi Contre Euttome, 1.81ff. itpös , speaks of "une inspiration aristotelicienne" on Basil's treatise in a way that seems to invoke the doctrine of relatives in the Categories of Aristotle. When he takes up Aristotelian influences in his introduction, he claims: "La theorie des noms relatifs donnee... est eile aussi de provenance aristotelicienne dans sa mise en eeuvre.... La distinction posee entre noms absolus et noms relatifs rappelle aussi ce texte des Categories (Cat. la16).... On retrouve de part et d' autre les memes exemples de I' esclave, de I' homme et du boeuf. Mais ces exemples peuvent titre les lieux communs des ecoles......

101 Certainly, this first definition of relatives is a definition as to what is said (X yc'rat). But this is also what one finds in the grammarians23 represented by the eighth century Greek grammarian Heliodorus. Heliodorus in the Scholia Marciana discusses the distinction in the Ars Granmiatica of Dionysius Thrax between relatives (itpöc it gxov) and quasi-relatives (hS irp6q rt gxov).24 It is likely that both this distinction of Dionysius Thrax and the commentary of his later scholiasts is inherited in some way from the Stoics.25 Certainly the case for Stoic provenance is strengthened by the fact that the bits from Heliodorus, as Schneider maintains in his edition of the second century AD grammarian Apollonius Dyscolus, can probably be traced back to Apollonius, who is undoubtedly influenced by Stoics.26

23Someof the scholiasts an relatives in Dionysius Thrax use the language (as does Basil) of speaking and hearing, e. g. Sch. in Dionys. Thr. 193.15-37 (Uhlig); Sch. in Dionys. Thr. (ex Heliodoro) 553.25-30 (Uhlig). 24Dionys. Thr., Ars Granun. 34.1-35.4 (Uhlig): np6;, rL 56 ý(; Ttv 9Xov 6; naAp -oi6q ýixo; ft-ýOq and then the other type, 6)qlEp6;, ct 9Xov 66 ý(YTtvCo; vi5ý hgýpa, 06vacog ý(Oý. 25This is accepted by Mario Mignucci, "The Stoic Notion of Relatives, " Matter and Metaphysics. Fourth Symposium Hellenisticurn, eds. Jonathan Barnes and Mario Mignucci (Naples: Bibliopolis, 1988): 129-222 at 189. It was first argued in detail by Karl Barwick, Renunitis Palaemon tind die rdinisclie Ars Grannnatica, Philologus Supplementband 15.2 (Leipzig: Dieterich'sche Verlagsbuchhandlung, 1922): 103-107. Both Barwick and Mignucci (200-202) note the differences between the relatives distinction in the Greek grammarians and the later Latin grammarians. Barwick sees some correspondence between the Roman grammarians and Dionysius Thrax on relatives, but although his broader thesis posits a grammatical lineage stemming from Diogenes of Babylon and the Alexandrians down to the Latin tradition, he denies that the Romans are dependent on Dionysius Thrax and the Greek grammarians; rather, wherever their grammatical doctrine diverges, the Roman tradition is generally closer to the original Stoic theory. That is, both strands are dependent on a Stoic source (,rEXvA 7rEpi Ocovýg),but the 'r6mische Schulgrammatik' is truer to the original Stoic scheme. The expositions of relative terms, he thinks, are particularly good examples of this: "Hier weicht die Terminologie des Dionysius Thrax (np6q ct iEXovund 6; irp6;, Et 9Xov) bei aller Ahnlichkeit doch in charakterischer Weise von der der r6mischen ars grammatica (,ce( np6qrt und, c6tnp6q ci nw, EXovicc)ab. " Barwick adds that the latter terms are more clearly Stoic, as they are the terms used in the important Simplicius text, which also makes a distinction between two types of relatives. This position is rejected by Max Pohlenz, "Die BegriAndung der abendlandischen Sprachlehre durch die Stoa, " Nacliricliten von der GeselIscitaft der Wissenschaften zu Gdttingen, Phil. Hist. KI., N. F., Bd. 3, N. 6 (Gbttingen: Vandenhoeck & Ruprecht, 1939): 151-198 at 186-188. But Mignucci himself sensibly sees no good reason to prefer Greek or Latin grammarians as representing the Stoics more closely. 26Schneider, Frag. Ap. Dys-, 49.22ff. posits two sections on relatives in a list of kinds of appellatives in the lost treatise of ApolloniuS, FlEpi 6vopdre)v fiTot 6vo1AaTtK6v. As Schneider points out, unfortunately Priscian follows Dionysius Thrax rather than Apollonius on the kinds of names, while the Dionysian scholiasts employ at times the doctrine of Apollonius in their commentary. In Priscian, the relatives distinction is found in his section De noinine, as two items in a list of eighteen kinds of appellatives (appellativoritin species), with an explanation of each kind attached. From the surviving corpus of Apollonius, Schneider adduces Synt. 367.2- 5 (Schneider) in support of 1 working notion of relatives -r& np6ý ct. Here relative nouns and

102 In fact, all the main elements in Basil's exposition of relative names in Etat. 2.9 agree fairly well with material from the grammarians. The distinction between absolute and relative names taken up by Basil is found in the grammarians as early as Dionysius Thrax 27 although in Dionysius the two types of relatives are discussed together while absolute and relative names are taken up in separate parts of his treatise. 28 In addition, the grammatical texts discuss the completion of relative terms. One of Heliodorus' comments on np6q it 9zov relatives says simply, "it can be noted that the relative is always constructed with the genitive, e. g. father of this man, son of this man, friend of this man. "29 Another scholium, from the Scholia Vaticana collection and also probably influenced by Apollonius Dyscolus, explains the icp6q ct Exov relative along these lines:

Exot/ TLp6q 'LL Ec'ri To Kath icäv itpög ETepav cx atV vooüµCVOV, KaO' aüt6 S' ävvn66Tatov, oi.ov lta'cr p viös Ltaipo; Tativa 66v &rr poig voei'rat, otov itctrip Twos, vi6gtiuvos.

A relative is that which is conceived in its entirety with reference to a relation with something else, but does not subsist in its own right, as for example in the case of father, son, and companion. For these

comparatives (,r& oruyicpvcticec)are mentioned as words which presuppose a complement which does not appear on the surface of the syntax. Priscian's explanation of terms ad aliqldd, Inst. 2.28 (Keil, 60.19-27), is also adduced along with Sch. Dionys. Thr. 235.9ff.; 387.8ff. (which includes the text on relatives discussed below); 553.27ff. (Uhlig). As forr& k np6qEt, Schneider notes that it is not found. in any extant text of Apollonius, which he attributes only to chance; Priscian and the scholiasts make up for this deficiency inasmuch as they appear to have drawn in their expositions from Apollonius, Priscian, Inst. 2.29 (Keil, 60.23ff. ); Sch. Dionys. Thr. 388.5ff. (Uhlig). With regard to both kinds of relative terms, the texts from Priscian are good but not conclusive evidence of Apollonius behind the passage of the explanation of the two kinds of relatives into the grammatical tradition; the case for this would be more secure if the quasi-relative could be found in the extant works of Apollonius. 27Dionys. Thr., Ars Graimn. 12.4 (Uhlig, 34.1-35.4); 12.23 (Uhlig, 44.6-7) 28The distinction between absolute and relative names taken up by Basil is also found in the grammarians, e. g. Dionysius Thrax, Ars Graiiini. 12.4 (Uhlig, 34.1-35.4) and 12.23 (Uhlig, 44.6- 7): *ATcoXe?, ug6vov 86 ý(ntv [sc. 6vopaj 8 icaO' ýa-L)T6voCit-rat, Oltov Oe6;, 2,6yog. But the two types of relatives are discussed together in Dionysius Thrax, while absolute and relative names are not. Theo Kobusch notes briefly the presence of the distinction in Basil as well as Gregory of Nyssa as an inheritance from the grammarians, "Name und Sein. Zu den sprachphilosophischen Grundlagen in der Schrift Contra Eunomium des Gregor von Nyssa, " El 'Coýdra Etmotidiiiii I' eii la Prodticcioiz Literaria de Gregorio de Nisa, eds. Lucas F. Mateo-Seco and Juan L. Bastero (Pamplona: University of Navarra, 1988): 247-268 at 266 note 74. 29Sch. in Dionys. Thr. (ex Heliodoro) 387.13-15 (Uhlig), (njpF_t(aT6ov R 6TL r6 7rp6; -it 9Xov dei 6Tou. Sch. Dionys. Thr. (ex YEVtKl, l GIM6GOEXCR, naTýP TOýTOD, Ult6g roi6-Eoij, ýiXog -co, Cf. in Heliodoro) 388.20-26 (Uhlig). Another similar Heliodoran scholium from the Scliolia Lotidottetisia, Sch. in Dionys. Thr. (ex Heliodoro) 553.31-35 (Uhlig), reads Ta 5ý 7rp6g rt gpvrcc fi EXEt 60TOK50TCEI, 6; 60T6 6MC(q 7TP6q YEVIKýV TCOV 7TP6q & TýV GX&; tV T6 alUOV CdTiCt TOIL) OiTICE-COIL).

103 are conceived with other things; for example, a father is a father of someone, and a son is a son of someone 30 This text seems to present the same notion of relative name and its distinction from names which "subsist in their own right" as is found in Basil's contrast of relative with absolute names. The shared approach for relative terms seeks terms which are conjoined with these relative terms. Some further comparisons with discussions of relative terms in the grammarians should cast more light on the sort of relation that Basil has in mind in Eun. 2.9. A key text for a notion of semantic relatives as opposed to contraries in the grammarians appears in a scholium also attributed to Heliodorus in his treatment of Dionysius' distinction between np6q tit Exov relatives and cbSitp6q TL Exov relatives. The text from the ScholiaMarciana reads as follows 31 wS Yäp SsýtÖS rcpöc äptaiepöv, Kai 1tatp icpös viöv, oviw Kai ij vii mpÖgi 1µpav SOK XayEaOat" tavir1 SE 51a4Epst, öit iö µav itpäG it Exot icpoc u taxovöjcvov iö ztvös" Eäv yap ýiýccwi iöS, 9 ýEt tits irävticog tiivog; Tö SM wS irpäg 'tt ov ioi, o 4aivctat Exov" ov yäp ßäv Einco vg, Xg&t 'nS tö 'tivoS; iva ýnayäyco g ýµepag, Eastöý ii vg iötÖv £6tit Stäaifµa, ov tpös ýIEpav OEwpoüµevov.

For the night appears to be said with respect to the day in the same way as the right is said with respect to the left and the father with respect to the son. But they differ in the sense that in the case of the relative, the 'of what? ' must be supplied in thought. For when I say 'son', in general the 'of whom? ' must be supplied to this expression. That does not appear to occur in the case of quasi-relatives. For when I say 'night', the 'of what? ' is not added to this expression, in the sense that I supply 'of the day', since the night is a proper interval, which is not seen with reference to the day. 32 The scholiast states that both kinds of relatives involve a cognitive dependence, but only in the case of the np6g it gxovra relatives does this amount to semantic dependence 33 Apparently, in Eun. 2.9 the relation conceived is this relation of semantic dependence: The terms in question, e. g. 'son', 'slave', 'friend', are paired with other terms. This is clear from the statement that the hearer

30Sch. in Dionys. Thr. 235.9-12 (Uhlig) = Ap. Dys., IIepi. 6voµätwv (Fragmenta) 49.40-50.10 (Schneider). Mignucci, "The Stoic Notion of Relatives, " 189 note 55 agrees with Schneider that Heliodorus' explanation of relatives probably goes back to Apollonius Dyscolus. Richard Schneider argues for the titles and works to which he attributes fragments of Apollonius in the Praefatio III-X of his edition, Grammatici Graeci 2.3 (Leipzig: Teubner, 1965). 31Sch. in Dionys. Thr. (ex Heliodoro) 388.20-26 (Uhlig) 32The translation is drawn from Mignucci, "The Stoic Notion of Relatives, " 194. Mignucci (194 note 60) gives further references in the grammarians for a similar idea. 33Mignucci, "The Stoic Notion of Relatives, " 195ff. points out that the grammarians never say exactly what the cognitive link between pairs of contraries is.

104 "conceives that the entity in question is coupled with something else." Basically, the semantics of such terms amounts to possessing a meaning which seeks completion with something else in order for the relative to make sense3435 There is another question which might be raised regarding the issue of cognitive entailment and ontological entailment. For the sake of his position against his theological opponents, Basil highlights the point that "this name...involves no conception of a certain substance." Perhaps for purposes of clarity, there is no mention in this text of the co-existence of paired objects. Nor is there any discussion of whether such a co-existence can hold apart from the cognitive entailment, or conversely whether the cognitive entailment can hold apart from the coexistence of the objects involved in the relation of cognitive dependence. But this issue claims considerable attention from the scholiasts on Dionysius Thrax. For example, Heliodorus apparently justifies the entailment in knowledge of the term corresponding to its yokefellow on the basis of the being or existence of the corresponding member of the relation. That is, "when I say 'father', I conceive also the son, for there is a father if and only if there is a if the the the father. "36 So, it child. .. son perishes, so also perishes name of seems reasonable to suppose that Basil's neglect of the coexistence of relative things in this text of the Contra Eunomi nn does little to preclude the possibility of both types of entailment working together in his thought elsewhere. Now let us look at still one more major treatment of the relative terms 'Father' and 'Son' in the Contra Eunomium. Basil takes a rather different tack when he attempts to show how these terms can be stripped of unwanted meanings, and their primary significations distinguished. The passage (Eurz. 2.22) sets forth this argument.

34Mignucci, "The Stoic Notion of Relatives, " 191 says of the notion of relativity which he primarily finds in Heliodorus, "roughly speaking, we can say that semantical relativity affects pairs of general terms which are such that the meaning of one of them is expressed by the other and is given by its representing the domain of a relation. " Mignucci (191-192) then gives formal expression to this notion of semantical relativity. 35This formulation raises many questions, such as what satisfies the basic meaning of a general term. However, no answer can be given here more sophisticated than that the cognition of a relative term such as 'father' and 'slave' cannot meet basic semantic requirements, such as avoiding ambiguity, without the completion of the meaning of the term with a suitable correlate. 36Sch. in Dionys. Thr. (ex Heliodoro) 387.8-12 (Uhlig). Mignucci, "The Stoic Notion of Relatives, " 190 remarks of this justification that it "only suffices in a very weak sense" and points out that "other scholiasts and even Heliodorus at 387.34ff. take [Mignucci's

105 ' AU Taira Xgyco, ý7j61, Thv Toi näüous EVVOIaV TT)VbK tijc Thu rlaTp6q c ovijS EyytvoµEViJv ExaXivaov. ' ExpT v öE ötjnov Töv yp- E'l)C3E3EtV ýyvwxöta, Tö änEp atvov vjs ýv Toi. pýµaßt TovTots ötavoias, EL Tt äpa Kai ývvlcrIp%E, 71apattiTl66(µcVOV, µrß näaav 68p6c)S Thv owvýv OFETCv, nr tiro npo46CaEt toü änpsnovg Kai TO ößov ä.n ' avTTIS xpAßtµov avvcxpäXXEty' äX ' änö iv TOW 'ranstvwv xai 6apKtKCÖV voijµä'cfv ýv 'roiq nEpi. Oeov ööyµaßt KaOapavEty, 'YEVVrJaty Sý Tý äytwßüv1 Kai TI änaOcia Toi O6oü 7tp6Eovaav ývvosi. v, Töv µßv Tpönov xa8 ' öv E'Evv'gc cV ö OF-öS 6); äpprJTOV Kalt al[EptVOII'COV äotavTa, 1rp6c Sý Thv xaT ' ovaiav 6µot6Tr Ta Ex Tllc Tov ycvvav npoorlyopiaq OE yo1 Evov. Kai'rotyE 4avEp6v 6KOnou. tEVQ) öTt ov aW 1attKwV naOwv xvpüos Kai 7rpc6twO Evvotav EµnotEty Tavti ne4uKa Tä övöµaTa, 6 nazi p, Xeyw, Kai 6 ui6q, äß,Xä Ka8' ftCvTä µev XEyöµcva, 't v np6g äXAii? a ax£aty Miixvutat µövTIv. IIaTrjp µev yäp EaTty 6 ET6pc Tou Evat KaTä ThV op.oLav Eautc OIc tv 'ri v äpxhv napaaxwv- uiOg SE, oE E' you To' ycvvijtwOEtvat 'ri v äpx1fV EGXIJKWs37

But I say these things, he [i. e. Eunomius] claims, in order to avoid the notion of passion which is involved with the word 'father'. It is the duty of the truly religious man to reject what is unacceptable in the meaning of these terms if something of this sort is found, not rashly to rule out use of the word altogether, nor to toss out what is beneficial from it under the pretext of inappropriateness; but rather to purify the mind from base and fleshly conceptions, and to conceive a generation which is fitting to the holiness and impassibility of God, leaving to one side the manner in which God generated, knowing that it is unspeakable and not to be grasped by human minds, while being guided towards the similarity according to substance by the use of the word 'to engender'. However, it is apparent upon examination that these names, that is 'father' and 'son', do not by nature bear a signification authoritatively and primarily of bodily passions, but rather are things which are said according to themselves, expressing only the relation of one thing to another. For a father is he who furnishes another with the principle of being in accordance with the nature like to himself, while a son is he who obtains from another the principle of being by generation. There are several points of interest in this passage. Basil is trying to rescue the term 'father' as applied properly to God from Eunornius' charge that it draws improper anthropomorphisms into the notion of God. Accordingly, the meaning of the term 'father' must be sifted and purified for theological use, while

formalization of cognitive entailment] and [Mignucci's formalization of ontological entailment] as just two ways to characterize to np6q'n exov." 37Basil, Eun. 2.22,33-51 (SC 305: 90-92)

106 the expression (cß(ový)of 'father' is preserved for the sake of what it tells the 6T 38 theologian about the nature of God (to öoov äßi Ct, Xp1 t}1OV) It is possible that here in Eun. 2.22 Basil is guided by the positions of the Homoiousian party. This doctrinal standpoint was constructed in the debates with the Neo-Arians of the years immediately preceding the composition of the Contra Eunondi ni. Basil's approach to the meaning of the disputed term 'father' appears quite close to the arguments of his Homoiousian predecessors, as is clear from a glance at the synodical letter from the Council of Ancyra (358)39 The synodical epistle states that our faith is in the Father, Son, and Holy Spirit; when these names are heard according to their natural sense (äcn6r6v 4vanx&), the name 'Father' is conceived as the cause of a like essence (aittov öµoiaq ainov o6aia5), and the name 'Son' as like (ö iotov) to the Father, whose Son he is 40 The Homoiousians are keenly aware that from the standpoint of everyday use, there are unwanted meanings attached to the term 'son', namely those of passion and lack of stability 41 Rightly conceived, the father is the father of the son, and only the notion of the likeness (gvvota toi öµoiou) remains. As for the problematic notion of begetting (TÖ ycvv&v), there is the same fear of contamination of an immaterial God from the 'passions' (mäOq)associated with earthly fathers and sons as one finds in the Contra Eunomium, specifically the fear of producing an

331naddition to the text presented above, a similar argument is found at Erin. 2.23,20-23 (SC 305: 94). There is a helpful discussion of these arguments in Osborne, "Literal or Metaphorical, " 167ff.. 39This important document is preserved by Epiphanius, who says that it was composed by Basil of Ancyra. The possible influence of the Homoiousians on Basil of Caesarea on these questions of meaning seems to have been missed by Drecoll, Die Entwicklung, 83 who simply points to the Eusebian tradition; Eusebius and his followers are cited by Drecoll in connection with the incomprehensibility of the Son's generation, not the relation. Lbhr, "Die Entstehung, " 75 notes the increased theological (and perhaps philosophical) sophistication of the Homoiousian document in the face of the resurgent Arian threat of the Anhomoian party with their dialectical sophistication. This general trend can be traced forward, and with greater intensity, to the treatises Contra Etinondunt of Basil and Gregory of Nyssa. 40Syn. Ep. Anc., ap. Epiph., Haer. 73.3.3 (Holl, 271.15-18). As Lbhr, "Die Entstehung, " 73 note 244 points out, in the hands of the Homoiousians the term 'similar' has a relational character, but the document only affirms that the Son is 'similar' to the Father, not that the Father is Isimilar' to the Son. For the Father is only the cause of a substance like to himself, who is the Father's image: "Die 'Ähnlichkeitsrelation' ist nicht symmetrisch aufgefaßt: der Sohn ist dem Vater 'ähnlich', nicht umgekehrt. Der Vater ist der Ursprung der 'Ähnlichkeit' des Sohnes. " 41Syn. Ep. Anc., ap. Epiph., Haer. 73.3.5 (Holl, 271-22-24). For this passage as weil as tile bit from Haer. 73.3.3, cf. Syn. Ep. Anc., ap. Epiph., Haer. 73.4.1-2 (Floll, 272.19-2713).

107 incomplete or imperfect (ätcXýq) offspring along with the entire apparatus of human reproduction. 42 It is also plausible that Basil's sifting out of the primary significations of 'father' and 'son' takes its cue from the arguments of the Homoiousians. This line of thought is present in the synodical epistle adduced above. In the letter Basil of Ancyra turns the tables on his opponents, writing that it is the only- begotten Son of the Father who is properly called 'son', while earthly sons are not properly called 'sons'. Accordingly, there is a catachresis (an 'abuse' of language) and homonymy (different senses of the same word) of earthly sons generated from earthly fathers. Kai oü Svvaiat i Eni üwv KaiaxpflYttK(j)q Kai öµwvüµcos Xsyoµevct)v viwv vvota 'tw µovoyevet äpµöýcty. 66nEp yap nvýiov µßv XEycTat Kvpicoc to EK nihov Ka'ceaKevaQtCVOV, KOLVÖTEpoV SE Kai Kaiaxprj6titK6s än EKEI,VOV nvýioV Kai T6 EK µoa,ivov Kai XaA,KOÜ Kai &XkllS T1vOS U119 ycyov6q [övo th ctai., of iu Kai vi6q OEov KvpiwS XEyEiat µövov ö EK Toi naTpOS YeY£vvrltCVOS, oi, SE äPAot KaTaxpTj6'LK(og. 143

And the concept applied to those called 'sons' loosely and equivocally does not fit the Only-Begotten. For just as the word pyxion [i. e. tablet] is applied properly to what is fashioned from pyxos [i. e. boxwood], but from that usage pyxion comes to be applied loosely, in common usage, to what is made of lead, bronze, and any other material, [so also the term 'Son of God' is properly applied only to the One begotten from the Father, but to the others in a loose sense. ]44 The Homoiousians say that only a proper notion of sonhood will do in the case of the Son of God. The transformation of this argument into Basil's purified notion of fatherhood and sonhood as what is signified primarily (icupi oq and mcpwicws),free from connotations of bodily passions, would be fairly natural. 45 Basil has a purely relational primary signification for the terms 'father' and 'son', namely "the relation of one thing to another". The two relational properties

42Syn. Ep. Anc., ap. Epiph., Haer. 73.4.5 (Holl, 273.15ff. ) 43Syn. Ep. Anc., op. Epiph., Haer. 73.5.5 (Holl, 275.1ff. ). The square brackets in the text mark a lacuna, where the conjecture of Holl is supplied. 44The translation is adapted from Amidon, The Panarion of St. Epiphanitis, 296. 45The importance of the meaning of the proper sense of names being known only in the case of God is not recognized by Kopecek, History of Neo-Arianism, 327-328, when he compares the Homoiousian position with Eunomius' attack in his Apology. The Homoiousians at times think of God as the one who is primarily Father, as can be seen from a passage in George of Laodicea's letter of 359, op. Epiph. Hoer. 73.12.1 (Holl, 284.14-16): The words spoken by God at Ex. 3.14 are spoken 6; eit tov nptto; vooup6vou natp6;, eý ov näßa narpia ev ovpavw Kai Eni yT ovo idl eTat.

108 expressed by these terms are then distinguished: The property of the former is "he who furnishes another with the principle (äpxij) of being", while the property of the latter is "he who obtains from another the principle of being by generation" 46 Basil even looks like he has here the "Homoiousian purified notion of fatherhood (evvota 'rov öµoiou) in what he tacks on to the relational property, the addition "in accordance with the nature like to himself"47 Before returning to Basil on relatives and the grammatical tradition, there is an additional remark to be made about a slight problem in Eun. 2.22. One possible source of confusion in the text is that the terms are designated 'things being said according to themselves' (xaO' : au-tä Xey6peva). But in spite of appearances, this cannot be talking about absolute terms. Rather, it simply states that the terms in question, when conceived correctly, are 'things said without recourse to any further meaning beyond the relation. Uppermost in Basil's mind as he says this must be that the terms signify the relation with one another (iriv np6q i? ijXu aXeaty µövriv), and not the underlying realities 48 On the other hand, Eunomius is charged with erroneous understanding, because he thinks that the terms are authoritatively applied to men49 and hence are used incorrectly (xaTaxpnc Tucwc)and metaphorically (ýx µETaýopäs)for God 50 By this Eunomius is simply supporting his view that 'father' is not properly or correctly used to designate the substance of God. To supplement the case presented here for the stamp of the grammatical tradition in the polemic against Eunomius, the presence of two kinds of relative terms in the Contra Eunomium will now be suggested. This should help to confirm that the distinction from the grammarians, which was argued above to be the context for the relevant discussions of relatives in the scholiasts, forms the background to the ideas about relatives employed in the Contra Eurnomium. At Doz. 2.26-29, Basil discusses the nature of the divine properties which distinguish

46Cf Syn. Ep. Anc., ap. Epiph., Haer. 73.3.3,73.4.1,73.9.6 (Holl) 47There is another interesting passage soon after, Eun. 2.24,1-4; 16-23 (SC 305: 98-100), which deals with what fathers and sons share. Prominent here is the important theological term 'kinship' (ob cct(h). But this notion of 'like in nature' is a general characteristic of fathers and sons as Basil understands them, not on the basis of the logic of the sort of relation which one finds exemplified by the pair father/son. 481n fact, this looks much like what one Dionysian scholiast, Sch. Vat. in Dionys. Thr. 235.9-12 (Uhlig), who probably reflects the views of Apollonius Dyscolus, says about the relative: "[the relative] does not subsist in its own right" (xa©' ai)tö S' (1vunößzaiov). 49Basil, Eun. 2.23,14ff. (SC 305: 94ff. )

109 Father, Son, and Holy Spirit, while taking up the question of what sort of alterity is conceived in the relation between Father and Son. He argues that they are not contraries 51 Basil undertakes thereby to refute the doctrine of Eunomius, who claims: ö6ov mapý?Xaxiat tiö yavvnTÖv 7tp6q rö äYEvv, ov, 'roGovtiov mapij? Xaxeat tiö 4coSmp6g c6 Kai chv ýcoi v npög chv (whv xai tir v Svvapty np6g 'rfiv Süvaµty.

To the degree that the generate diverges from the ingenerate, so light is different from light, and life from life, and power from power. In response, Basil argues that Eunomius does not understand the constitution of different kinds of relations by names (tiä övöµaia): Oüic Yap ýco v oütie Svvaµty nj ýwý xai nj Svvä t ävtitxsi66at ýairj 'r S äv, & 86vaiov Kai äSvvaµiav AV TE? cothtrJv ävtIOe nv ExitXijpovv.... "AmEp ovtoq... Kai 'Eo1Smv Xöywv TExVaa}laut tpöS tiö Evavtiiov tiw IIaipi tiiv 4üoty Toü Movoycvov; ättoaTjßa;, µövTv KaTaXeiTtct AV 8v4. qµiav iwv 6voµätiwv. 52

For someone would not say that life is opposed to life and power to power, but rather death and powerlessness make up the most complete opposition in these cases.... With respect to these very things he, having drawn by means of verbal trickery the nature of the only-begotten into the [relation of] opposition with the Father, leaves only the empty pretense of the names. Basil seems to charge his opponent with mistaking relative terms for contraries, drawing upon the distinction of two kinds of relatives that probably would have been familiar to him from the grammarians. However, there is nothing explicit said in the passage about the kind of relative which Dionysius Thrax in his Ars Gra, nmatica terms irpös ct g%ov, for which Dionysius gives the single example of 'father' and 'son'. Nevertheless, Basil provides a clue to his knowledge of the contrariety expressed by ws np65 t gxov relatives in his remark, "but rather death and powerlessness make up the most complete opposition. " There is little of the standard grammatical and logical terminology here of relatives, but the contraries in view are power (Svvagtq) and life (ýo ), which are relative in the same way that the grammarians' quasi-relatives night and day are relative. Basil

50Basil, Ettn. 2.23,41-43 (SC 305: 94ff. ) 51Basil, Eun. 2.26-29 (SC 305: 108-124) 52Basil, Eutt. 2.27,26-33 (SC 305: 114)

110 uses an example of Dionysius for wS np6; t Exov relatives, namely the pair life and death 53 In conclusion, Basil draws on the grammatical tradition in significant ways when he talks about relatives. Also, the notion of semantic relative such as one finds in the grammarians, and behind them the Stoics, contributes to the philosophical background to Lin. 2.9. And it is likely that much of what Basil says on the topic of relatives is anticipated and influenced by the work of his immediate predecessors in the leadership of the Homoiousian party, especially George of Laodicea and Basil of Ancyra, although the Homoiousians do not draw on the grammatical tradition in their discussions of relatives in the way that is apparent in Basil. But these grammatical influences are only discernible in Basil's early treatise, the Contra Ettnondion. This philosophical inheritance would be rooted in Basil's rhetorical and grammatical education 54 Although it is not possible to argue the point in this paper, it can be shown that the later treatise De Spirittt Sancto disappoints in this regard, for there is no clear trace of the semantic relative, and there is no technical treatment of relatives such as is found in the Contra Eunominm. In fact, ideas about relations found in the De Spiritu Sancto do not possess the logical and semantic character that they possess in the Contra Ennominm. I think that the turn away in the later treatise from relational terms which are paired with other terms is partly due to the predominant notion in the later treatise of divine beings ordered in a class, although as can be seen from Sp. Sanct. 6.14, Basil easily slips back into thinking of relational terms (Father and Son, not Son and Spirit) which are conceived with one another. At least in the Contra Eunomium, the attention to relatives helps to show how fine points of language are important to the formulation of theological arguments. At its basis, the role of relatives in Christian theology is simple, and

53Dionys. Thr., Ars Gramm. 12.4 (Uhlig, 34.1-35.4). The opposition of life and death is the example used in the summary of Stoic dialectic found in Diogenes Laertius (7.52-53), where there is a discussion of 'things which are thought' (Tä voovµeva), among which is listed 'things that are conceived by opposition' (iä at evaviicoßnv). The elaboration of this is given, Kai xat' evavTicuaty Ocvatog, that is, death is conceived in opposition to life. 54The evidence for grammatical and dialectical studies is evaluated in the preliminary chapter "Grammar, Rhetoric, and Philosophy". There is virtually no discussion of relatives, and certainly nothing of philosophical interest, in the rhetoricians commonly studied in the fourth century such as Theon and Hermogenes.

111 divine was no doubt dear to the Cappadocian theologians: The names 'Father' and 'Son' are inseparable in meaning 55

Thirteenth 55Several improvements were suggested by Richard Sorabji and participants in the 16-21 August International Conference on Patristics Studies, University of Oxford (Oxford, 1999), where the chapter was presented in reduced form.

112 4. Language and Thought

It has often been said that the fourth-century debates that followed the landmark Council of Nicaea in 325 were conducted in such a way that Greek philosophy had been finally purged from trinitarian thought, at least for the orthodox theologians. Perhaps there is something to this claim. But there are some surprising and important departures from this model. (One thinks, for example, of the brilliant Oratio Catecheticaof Gregory of Nyssa. ) We will try here to bring one important exception to light. Philosophical ideas about mind and language inherited from the Stoics inform the theological use of the distinction between inner speech (Xoyoq EvöuiOstoq) and language expressed (Xayoq npoýopi c6q) in the work of Basil of Caesarea.' The chapter is divided into four parts.

I

Although the Stoics were not the first philosophers to draw a distinction between inner and outer speech, they developed the notion as received from Plato (Soph. 263E3-9; 264A1-2; Phil. 38E; Theaet. 189E6-190A; 206D1ff. ) and Aristotle (An. Post. 76b24; Cat. 4b34), and gave it their stamp. 2 Following Plato

lFor recent recognition of the importance of the analogical arguments in Basil based on the distinction, there is Philip Rousseau, Basil of Caesarea (Berkeley and Los Angeles: University of California Press, 1994): 114-116 and Volker Henning Drecoll, Die Entwickl1ing der Trinitiffs1ehre des Basilius von Cdsarea (G6ttingen: Vandenhoeck and Ruprecht, 1996): 92-93 and 166. 20n the question of Stoic origins, there is a classic discussion by Max Pohlenz, "Die Begriinding der abendlindischen Sprachlehre durch die Stoa, " Nacliricliten von der Gesellscliaft der Wissenscliaften zit Gdttingen, Phil. Hist. KI., N. F., Bd. 3, N. 6 (Gbttingen: Vandenhoeck & Ruprecht, 1939): 151-198 at 196-197. Pohlenz strongly suspects that the terminology coined and the doctrine worked out in the second-century BC debate over the rationality of animals between Stoics and Academics, and perhaps was even first set forth by the Academics. For Plato on semantics and inner/outer speech, there is a summary of the evidence from the

113 and Aristotle, the wide dissemination of the Stoic Mpg distinction in Hellenistic and Later Greek philosophy must. also be taken into account. Of the many uses of the distinction, the Hellenistic debates on speech and human rationality in particular inform the theological uses of the distinction I will discuss. So I will introduce the Hellenistic debates by going into what is known about the Stoic position. The Stoics do not seem to have concentrated on thought and its precise relation to language as is now so prominent in philosophy, cognitive science, and related fields. At least we can see that the Stoics have a notion of inner speech or 'linguistic thought' which recalls modern cognitivist theories of language and its relation to thought. The distinction is key to marking off human beings from other animate creatures, in such a way that the relation between language and the rationality of human thought is made clear. It is safely attested (Diogenes Laertius, 7.55ff. ) that it plays an important role in the first part of Stoic dialectic, which later Stoics such as Diogenes of Babylon term the part dealing with vocal sound (itcpt ýwvrjg) and where the great logician Chrysippus deals with 'signifiers' (Diogenes Laertius 7.62).

The Imperial skeptic philosopher Sextus Empiricus reports the Stoics3 on the relevance of the distinction between inner and outer ?,oyo;: ävOpwnos ov iw mpo4opLxw Xöycu S&a#pEt iwv äXXwv ý6)wv (Kai yap KöpaKES Kai T'tTTaxot Kai. Kittat ýväpOpouS itpo4£poviat ýwvdS), &XXä 'rw Ev&a6Etw 4

A human being differs from other animals not by using speech which is uttered (for crows, parrots and jays pronounce articulated but inner words), rather speech.5 dialogues in Wolfram Ax, Laut, Stinnne tind Sprache (Gbttingen: Vandenhoeck & Ruprecht, 1986):105. 3Richard Janko thinks that Sextus here is referring to the important and influential 'second- generation' Stoic Diogenes of Babylon (floritit mid-second century BC in Athens, dates from about 230 to 150 BC) in his theory of 'vocal sound' (ýcový).But the text is only prefixed with an ascription to 'the dogmatists', which in context must be Stoics. Perhaps Janko is nevertheless correct about Math. 8.275,but his case has problems. The paper in questionis Richard janko, "Crates of Mallos, Dionysius Thrax and the Tradition of Stoic Grammatical Theory," Tile PassionateIntellect. Essaysoil the Transforinationof Classical Traditions, Presentedto Prof. 1. G. Kidd, Rutgers University Studies in Classical Humanities, ed. Lewis Ayres (New Brunswick and London: Transaction Publishers, 1995): 213-233. janko (215-216) discusses Diogenes' important contribution to Hellenistic grammatical theory. 4Sext.Emp., Math. 8.275(Mutschmann and Mau) = SVF 135 and 223 5For Diogenes, the difference between mere utterance and speech is also associated with the Stoic theory of the 'thing said', ?xK, [6v (Diogenes Laertius 7.57 = LS 33A, reporting Diogenesof Babylon's work FIF-piýwvi_jq)-. "Speaking differs from utterance, for soundsare uttered, while

114 This idea is loosely in accord with the view of Diogenes of Babylon on human (Diogenes Sextus rationality reported elsewhere Laertius 7.55-56), and thus may be drawing from Diogenes. For Diogenes of Babylon says here that the human from voice as distinguished from animal sounds is articulate and sent forth the mind (ävOpwnou S' gctity EvapOpoq xai äßö StavoiaS Exmeµntojtvrj), and thus in its semantic capacity becomes speech (?.oyog 89 ý(Ytit ýu of ungavin di änö ötavoiag ýx1cennoµtvn). And in fact for this definition of X6yo5, as well as 'articulate utterance' (XE4tg) and 'dialect' (&i ? sxiog), Diogenes is the representative of Stoic theory. Actually, the one sentence which is attributed to Diogenes (Diogenes Laertius 7.55) regarding the vital question of animal vocalization and human speech looks a bit different from Sextus' report of the Stoics. ýc ov µev 6 an ýcwvh th p ünö optic ncnX,rlyµhvos, ävüpchnov S' Ec my 9vapOpoS Kai 6(n6 StavoiaS ýxnSµnoµEVrj, (b ö AtoyavqS 4ijaiv, i nS änö ScKatE YYdpwv a'ro v 'reXEtouTai.

In the case of an animal, voice is air struck by impulse, while for a human being vocal sound is articulate and issued from the mind, as Diogenes says, which is fully-formed by the age of fourteen. As Wolfram Ax points out, according to this testimony Diogenes emphasizes the two key points of (1) the articulation of human speech as well as (2) the relation of speech to what is in the mind. (The evidence on Stoic Mpg EMtditToq consistently involves the Stoic word for thought or mind (5l,(jvota). )6

things, which are actually just sayables, are said. " 8ta#pet 5ý icai r6 X6ymv ro; irpo#pF-cFOat- Tcpo#povzat ýtý-v y&p a! ýcovai, X6yecat R icai ra irpayltaTa, ii 5h Kai ýxKc& 'r'uyXdvF-t. Janko (229- 231) tries to strengthen his case by drawing parallels between a text of Crates in the Herculaneum Papyri and Diogenes' theory of Ocovhon the point adducing birds, and argues (231) for the influence of Diogenes on Crates. "Now Diogenes' role in the evolution of linguistics tun-is out to have been fundamental.... Since Crates seems not to have been interested in animals, whereas Diogenes was not only concerned with their intellectual capacities but also exploited the biological teatises of Aristotle, who notes parrots' ability to speak, Crates seems more likely to depend on Diogenes than vice versa. His adoption of Diogenes' theory of sound (Ocov4) makes him more a Stoic than anything else. " Janko here draws on the work of Ax, Laid, Stilnille wid Sprache, 138-211 (especially 181ff. on animal speech and rationality in Diogenes), which is helpful but assumes far too readily an extensive influence of Aristotle on the Stoics, and Diogenes in particular. Ax does not assign the Sextus passage to Diogenes as does Janko, although it might seem plausible to do so. 61n a report in Galen (Plac. Hippocr. Plat. 2.5.9-13 = FDS 450) Diogenes talks about both points (articulation of human speech and relation to internal mental events), although Galen does not report anything about animal rationality and speech. There may be additional support for a Diogenian notion of inner and outer speech from his distinction between speech and utterance (Diogenes Laertius 7.57), for a bird utterance could hardly be connected with a 'sayable' (XEic-i6v), which for the Stoics is associated with a 'rational presentation'.

115 For the Stoics, the key characteristic of human rationality is the capacity for language.? Standard Stoic theory defines thoughts in terms of rational impressions (Xoynncaiýavtaaiat), which are internally ordered and articulated in sentences,but also have a certain correspondence to non-corporeal things which are said or can be said (XExi6). 8 In fact, the 'impressions' of rational creatures are spoken of as 'thoughts' or 'cognitive processes' (voi aei.S) 9 and it is clear that the Stoic notion of 'thought' (8tdvota) which follows 'presentation' or 'appearance' is directly associated with language. "The presentation precedes, and then thought, which possesses the capacity for speech, expresses with language that which is experienced by the agency of the presentation. "1o Accordingly, the Stoics have a sense in which they regard mental speech as unspoken, an internal phenomenon. The linguistic character of thought is important to the Stoics." In one argument attributed by Galen to Chrysippus for locating the ruling part of the soul (i ysµovtxöv) in the heart, what goes on in the mind is spoken of as inner discourse. ' Eµoi µEv & SoKEt Tw impi Tij ýcovrj Xöycp xai vvv itpoaXpijc Oat. Tsxµaipoµat SE Kai EK Twv ECt#poµývwv" ämö yap' is &avoias, 0716i, 5J X, ycty Kai ý-v Eavtcw X, yety ij 4cwvhv SLEýtEvat xai SLavoaicOat Ka! EV EauTOtq 1xovf V &EýtEVal Kai EKT6q &Irt gl[Ety. öµoXoyoütEVOv yap 'tt Xaµßävcov cbs TOD av'toü µopiou to X, yety eire xai 'rö Ev EauTý %1EyELV,Eiaa 7cpo6xaµß6vwv avTCp TT-ig xap&as 9pyov eivat 'tö ? yeiv,

7Catherine Atherton, The Stoics on Ambiguity (Cambridge: Cambridge University Press, 1993): 42ff. gives a fuller account of the link between language and rationality, as well as the XEK,16v and its significance in connection with the distinction of ?,6yo;. Cf. Max Miffil, "Der X6yo; ývStdft, coqund npoýoptOq von der 51teren Stoa bis zur Synode von Sirmium 351," Arclziv ffir Begriffsgeschiclite 7 (1962): 7-56 at 8-18. 8The Stoics as reported by Sextus Empiricus, Matli. 8.70 (Mutschmann) = LS 33C and Diogenes Laertius 7.63 = LS 33F define the XEKx6vas "that which subsists in accordance with a rational impression" QxKT6v eivat T6 icaT& ýavTaaiav Xoytichv i4tmdgevov), and define the rational impression as "that in which the content of the impression is presented by means of language" (KaO' ývc6 ýavTaaOýv 9(rrt ?.6ycp napaaVjuat). Sextus, Matit. 8.12 (Mutschmann) = LS 33B reports that for the Stoics, the true and the false have to do with 'what is signified' (IrEp! T6 a-qgatvop6vq)), and sets forth a three-level Stoic semantic theory. 9Diogenes Laertius, 7.51 = LS 39A6 1ODiogenes Laertius, 7.49 = LS 331): npoTlyetTat ydp h ýcEvTaaia, EIO' h 8idvota ýKXccXIjTtKh i)rcdpXo1jc; cc, 6 7rdaXet i)Tc6 m-q ý=Taataq, -roZTo ýKýgpet My(q. "Julia Annas, Hellenistic Philosopliy of Mind (Berkeley and Los Angeles: Uhiversity of California Press, 1992): 63ff. remarks helpfully that "the Stoics' detailed accounts of perception and action are explained in terms of the reception and interpretation of articulable content. Content is dominant in Stoic philosophy of mind. " Annas also notes that the work of the ýycpovtic& is relevant. "For the Stoics the hallmark of rationality is the ability to use language, and hence we find that the human lidgenionikon functions throughout in terms of a kind of language. "

116 ä t4 oiv 9xet 7tepatv6tevov ýv T11Kapfit 'yivEGOati Tö Ev £aviw

He [i. e. Chrysippus] seems to me to be making use in this argument of his work on vocal sound. I prove this on the basis of what he says in addition: "Then from the mind", he says, "must come speaking and speaking in oneself or producing vocal sound and thinking and producing vocal sound in oneself and sending outwards. " For he takes something agreed as his first premiss, namely that speaking and speaking in oneself belong to the same part [of the soul], then he takes as his second premiss that speech is a work of the heart, and from these two premisses he deduces the conclusion, that speech in oneself takes place in the heart. The snippet from Chrysippus contains obscurities, but there seems to be a notion that speakers articulate language within the mind which is then reflected in the utterance spoken to other (hearers, that is) speakers. The relation between inner and outer speech is not causal like the relation of speaker's mind (Siavoia) and speaker's utterance but rather is a relation of having the same character. 13

12Galen, Plac. Hippocr. Plat. 3.7.42-43 (De Lacy, 220.16-18) = FDS 451. Atherton, The Stoics on Ainbigiiity, 95-97 elucidates the Stoic recourse to "information provided by linguistic and paralinguistic phenomena" as well as other texts in Galen and other sources. 13Chrysippus is said by Galen to be drawing upon his work an vocal sound (TiEp! Tý; Ocovi_10, where presumably Chrysippus presents his treatment of the distinction between inner and'outer speech. Perhaps Galen refers to the title of a work by Chrysippus from which no other material is quoted in the reports, and is testified nowhere else. On the other hand treatises entitled IFIEp!Ti-1; Ocowl; are testified for both Diogenes of Babylon and Archedemus. Galen may be referring to a book in which, as perhaps in the lost treatise nEpi O(Ovrlýof Diogenes of Babylon, there is a treatment of the distinction between inner and outer speech in a Chrysippean discussion of vocal sound. But then why does the title of such an important work not appear in any list or citation of Chrysippus? One good possibility, warranted by the surviving evidence on the Stoic theory of vocal sound, is that at least one place where Chrysippus discusses voice is the second book of his treatise I& (D1XnK& (Diogenes Laertius 7.55). Here it is said that Stoic dialectic begins from a treatment of vocal sound. There follows immediately the Stoic definition of vocal sound from the rýXvq of Diogenes of Babylon, after which Diogenes is again adduced for the distinction of animal utterance from human voice. Then the Stoic view regarding the ontological status of vocal sound is reported, i. e. vocal sound is a body (aCoga). In support of this statement, several Stoics are cited: Archedemus' work rlepi ýcoVijc,Diogenes of Babylon, presumably again from his -T6XvYj,Antipater, whose treatise could perhaps be titled IlEpi ýcovi_j;or could be a treatise mentioned elsewhere as 'On the word and the things said' (rhpi ?,6ýF_o)g Kai Xeyopýva)v)as suspects Karlheinz Hfilser, Die Fraginente mr Dialektik der Stoiker (Stuttgart-Bad Cannstatt: Frommann-Holzboog, 1988): 4.1840-1, and lastly Chrysippus' treatise on physical theory. So there is apparently a parallel between the content of this text of Chrysippus and Diogenes, on a point of Stoic doctrine which might be included along with the argument on the location of the guiding part of the soul. In fact, before proceeding to X64tq and Myo;, the Stoic justification in Diogenes Laertius for a corporeal voice is briefly stated, that everything which acts is a body, vocal sound acts by advancing from speaker to hearer, and therefore (as unstated conclusion) voice is a body. This argument might well reflect some of the contents of Chrysippus' physical treatise, or perhaps this is just one of the places in his voluminous corpus which treats vocal sound as prefatory to dialectic.

117 Later Peripatetics were happy to appropriate the Stoic terminology. For our purposes, let us take up the influential and brilliant Stoicizing Peripatetic philosopher Porphyry, who wrote an entire (extant) treatise on animals. Porphyry says in his work On Abstinence from Animal Food that the definition which he gives of Xöyoq npoýopiK6q is not the sole property of any philosopher or philosophical school, but is very widely recognized and contains a notion inherent in the concept of X, yoq. 14 The Mpg which is speech articulated by means of the tongue is 'signifying of the inner and psychic affections' (a11µav'rud T(Ovi vSov Kai Kath wuxhv rcaOwv). Presumably, Porphyry thinks that the notion is found in Plato and Aristotle as well as the Stoics, and accordingly represents philosophical consensus. 15

II

Now we will turn to the relevance of the Stoic distinction between inner and outer speech or reason to the rationality debates of the Hellenistic philosophical schools. The legacy of these philosophical arguments will then be traced in fourth-century Patristics. Much of what is known today of the Hellenistic debates is preserved from the Skeptics. To start with, there is a valuable report of the third-century AD Pyrrhonist philosopher Sextus Empiricus (Hyp. 1.65ff. ), who attacks Stoic doctrine but only within the framework of internal and then expressed reason. Sextus is here presenting his ten modes (rpönot) of skepticism. The first rpömoghas to do with sense-perception; it is an argument which points to the physical variations among animals on account of which different animals do not receive the same impressions (ýavtaoiai, ) from the same objects. On these grounds Sextus argues (Hyp. 1.59) that human beings can only know how things appear to us, and must suspend judgement as to what they are like in their nature. Accordingly, human beings cannot determine whether the 'impressions' (4avtaGiaa) of animals are to be preferred to those of humans. Sextus provides strong evidence for the Stoic use of the Mpg distinction in this context of debate, for they are said (Hyp. 1.62-

14Porphyry, Abst. 3.3 (Nauck, 188.17-20): ei Sn npo4optx6s E6Tt Xöyos ¢covii St& yXthtuJS - ßrJµavukh TOWev8ov Kai Kath wuXiJv na0wv" KOtvotcc'nl yap ij äitö oatq aü" Kai aipeaeco, oiSeiao exopevij, äß.Xä µövov tu ; Tou ldyou evvoiaS.

118 65) to argue for a basic divide between irrational animals and rational humans. (He refers often to 'dogmatists' but picks out the Stoics by name, and his attack is clearly preoccupied with Stoic doctrines. ) oüxoüv bm1Töv Xöyov EX9coµev. -roüiov öa 6 µav Ea'ty ýv&täOc'coS 6 SE Trpo4optxög. i&wµsv ovv npÖTEpov 7CEpi.Tov ývötaO 'cou. oüTOSToivvv Kath tioüS µä?,t6Ta ýµiv ävTtSoýovvTag vvv Soyµaitxoüg, tioüg änä of g ZtoaS, eV TorTOtS ÜotKE 6aAEVEty, tiro aipecEt tiwv o' nP-Icovxai gvyrý tiwv &XX0'rpiwv, yvw6Et 'r(Ov ci ToiTO ßuvietvov6O)v TeevcOv, ävTa1jyrot iwv xath AV oixeiav ýüßty äpET(ýv zwv rtEpi Tä rcäo1.16

So then, let us move on to reason. There is reason which is internal and there is reason which is expressed in language. So let us first look at the former kind. According to the dogmatists who think in a way particularly opposed to us, those from the Stoa, inner reason is anchored in these things, in the choice of what is appropriate (with a view to the fulfillment of our rational nature) and avoidance of what is alien to it, in the knowledge of the skills which contribute to this end, in the grasp of the virtues in accordance with our nature and the virtues which concern the passions. Significantly, Sextus of all the animals chooses the dog as his example for discussion in the passage immediately preceding (Hyp. 1.64).17 To support his position, namely that the perceptions of animals are no less trustworthy than those of humans, Sextus argues that animals possess rationality to the same degree as humans. Thus he takes up 'internal reasoning' first (Hyp. 1.66ff.), 18in order to show that his representative animal, the dog, is 'perfectly equipped' (Th?Etoq) in every respect that humans are ti6?,F-toq: The dog chooses what is 'well-suited' (oiK6iov) to himself and avoids what is harmful, the dog has a 'technical skill' (zexvi) namely hunting with which to obtain the oi.xeiov, and the dog possess virtue (äpe cý).

15Besides the texts that are adduced here, Pohlenz, "Die BegrCindung," 191ff. and 197-198 includes references to Plotinus, Galen, and the Rhetoricians on the k6yo; distinction. 16Sext. Emp., Hyp- 1.65 (Mutschmann). There is obviously a serious difficulty regarding the reading of the text from v.1 &vEt?,ýyet onwards. Julia Annas and Jonathan Barnes, Sexhis Eiiipiricus: Outlities of PyrrhoWsm (Cambridge: Cambridge University Press, 1994): 19 propose the solution of leaving out and supplying after dipF-,r& the rather speculative , while bracketing c6v nepirdt n6011. 17Sextus Empiricus, Hyp. 1.64 (Mutschmann), says of the dog: 6Tt -Toivuv aioOýact 8taýgpF-t ý(; hp 50Y[L(ITIKOi 6crýpý(Yet Xov hg(ýv TO&TOT6 OV (ýV, Oi GI)VO[IOXOYO'L)(YtV'Kai yetp Tý p&?, civrt XaýIP6VE'Eal, -Ea g1l OP6gF-VCECLIýTý 011pia bt& 'taýnvlq 6vtXveýxtw rcal totq 6ý0a)qAoiq O&TTov 141COV TaZTa 6pý)V Kal III ewoý aia0av6pevo; 6ý&oý. 18At Hyp. 1.73 Sextus moves on to speaking nep! Tou-irpoýoptKoZ k6you. Obviously, SextUs has a problem with this branch. He takes recourse not in dogs but in animals which utter human sounds such as jays (KiUat), Cf. the speech of birds at Hyp. 1.77, and the likehood that humans cannot understand the utterances of animals.

119 Sextus'dog is also a dialectician. This is illustrated (Hyp. 1.69) by means of the famous 'dog syllogism' of the famous Stoic philosopher Chrysippus. Chrysippus is said to be 'particularly hostile' to irrational animals yet for some reason accords the dog a share in dialectic. 19 For the dog makes use of the fifth undemonstrable argument with several disjuncts when, tracking an animal, he or she arrives at a crossroads, and having sniffed at the two alternative roads, pursues the third without sniffing. 20 With regard to the Stoic view of animal reasoning in this case, Sextus tells us that "the old [philosopher] says that he reasons this out virtually (8uvdgF_t)." So to this counter-example, the Stoics would reply that no syllogism is carried out by the dog in the 'full' sense that is exemplified in human reasoning. As Richard Sorabji thinks, for Chrysippus the dog performs this in a way that is analogous yet inferior to the way in wl-dch human reason operates, for animals only possess a distorted image of complete rationality. 21 The behavior of the dog at the crossroads can be expressed by its as an argument, and the same conclusion reached. Sextus also addresses the question of rational humans and irrational animals in another passage in the course of his arguments that there is no sign (rnlµeiov).

of Soyµatitxoi... tiovvavtitov Se xaiarnccuäýEOVTES 4 wnv, öct ävOpcottos ovxt Tcwicpoýopi. x6p Xöycu SLaýEpet ic,)v äXöywv ýwcov (Kai yäp xöp(XxES xat yntTaxoi xat xiiiati EväpOpouS mpogepoviat ýcoväs), OAdc iq) EvSia6giw, ovSý µövov ýavia6ic (Eýaviaatovio yap xäxciva), äß,A, ä zý JETa jiatin1Tj Kai avvOcttKl3. SLÖmepäxoA. ovOiag 9vvotav EXciv 601')q scat orqµsiov v0r16iv XaµßävEL Stä CT,lv &KAOUOiavý Kai yap avTÖ '1ö aitEiöv Catii ioiovTOV Et Tööe, c68e.

19KaTex ý; 5ý XPUMIrTrov C6V 11fixt(YUC TCOXEPOýVTa TO-tq e(X6yOtq ý60tq KC(i Tq etot&pO-o 5MXF-KEtKý 11; PF-T6xct. 20There are several reports of the Stoic 'dog syllogism', conveniently listed by Hillser, FDS, 4.1154-1159. The argument can be represented as follows: A or B or C. Not A and not B. Therefore, C. Michael Frede, Die stoische Logik (G6ttingen: Vandenhoeck & Ruprecht, 104): 153-157 discusses the original Stoic argument which passed into later writers, along with some problems associated with the syllogism. 21Sorabji, Aidi? ial Mitids atid Hwiiwi Morals, 89: "Why then do animals not infer? The Stoics' best answer might be that they define reason (logos) as a collection of concepts, so reasoning (logizestizai) should involve the application of concepts, and they deny that animals have any. "

120 Eitetiat äpa S'6(Mt xai xataGKEU TävOpwnov Tö Kai rn1µsiov ünäpxciv. 22

The dogmatists hold is the that ...who a position which opposite, say man does not differ from the irrational animals by virtue of reason which is expressed (for in fact both ravens and parrots and jays pronounce articulate vocal sounds), but rather by virtue of reason which is internal; nor by virtue of the simple impression alone (for animals too receive impressions), but by virtue of the impression which involves inference and combination. On account of which the human being, when he conceives a notion of consequence, also in that very thought registers the cognition of a sign by virtue of the consequence; and the sign itself is of such a sort, "if this, then this". So then the existence of the sign hinges upon the nature and construction of the human being. Here the distinguishing mark of human rationality according to the Stoics is inference from signs. 23 On top of this, the Stoics would presumably add that the difference also lies in the meaningful quality of human utterances, a position that Porphyry labors to overcome in his treatise on animals. 24 As for 'impressions' and the inner-outer speech distinction, the Stoic position would spot the difference in the more sophisticated structuring of thoughts or 'rational impressions' (XOYtK(X'týavcauidt). As we might claim in a somewhat different context today, exclusively in the case of humans is a developed theory of meaning or a theory of inference for a (perhaps only fragment of) natural language appropriate. In fact, the Stoic philosopher Diogenes of Babylon is reported (Diogenes Laertius 7.55) as saying that animals produce vocal sound (ý(ovA), but only as air struck by impulse (6pgA), wl-dle human vocal sound is articulate (ývapOp6q) and issues from the mind (5tCCV0jCC).2526

22Sext. Emp., Math. 8.275-276 (Mutschmann) = FDS 529; 1031 = SVF 2.223 (part) = LS 53T. The translation and discussion here owes much to the suggestions of Richard Sorabji in personal correspondence and to the work of Myles Burnyeat, "The Origins of Non-Deductive Inference, " Science and Spectilation. Shidies in Hellenistic Theory and Practice, eds. Jonathan Barnes, Jacques Brunschwig, Myles Burnyeat, and Malcolm Schofield (Cambridge: Cambridge University Press, 1982): 193-238 at 206-207. 23Burnyeat, "Origins, " 209ff. addresses the problems involved in the Stoic standard account of what a sign is (Math. 8.245): "A sign is a proposition which forms the antecedent in a sound conditional, being revelatory of the consequent. " It is much more straightforward to observe, as he does, that Math. 8.275-276 should not be understood to imply that signs amount to conditional propositions. 24Porphyry, Abst. 3.4 (Nauck, 191.3-13) brings up the same three birds as Sextus (ravens, parrots, and jays), and in general argues along very similar lines for meaning-loaded utterances of animals. 25Sorabji, Aninial Minds and Hionan Morals, 17-29 argues with some plausibility against the orthodox view of most scholars. Sorabji claims that for the Stoics animal perception has a limited predicational content, and to this extent is propositional. Sorabji (81ff. ) compares

121 Porphyry in his treatise On Abstinence likewise builds his arguments for the rationality of animals along the lines of the distinction, echoing some of the same points covered by Sextus.27 Porphyry provides valuable testimony on Stoic theory in the course of his concise summaries of the opinions of 'the ancients' on reason and language. Before anyone else, he turns to the Stoics. 8vrrov Sij X, you Kath tog änö Ejg aioäg öviog, tiov µev Ev&ta6E'cov, Tov SE npo4optKoü, Kai nä?, ty Tov µev Ka'aopOwpEvov, rov SE ýµaptT vov, itoTEpov änoctEpov t iä ýcwa 8LapOpcO6at irpoafKOV.... Ensi 'roivuv &t'C6q AV, 8 µßv Ev t itpo4opä, 8 56 ýv 5ia8aaEt, äpýwpEO(x mp6Tepov änö Tov 7tpo4optKoü Kai ioi Katiä tifiv 4wvO 'etayµevov. EL &f 7TpoýoptK6S Lyn X yog 4o vh &a y%o')ut1js rn tavtuKII Twv Evöov Kai Kath yruy v TcaOwv" Kotvoi6T71 yap ij än6öo6tg aüiri Kai aipýaecoS ovöaitco E. oµtvij, 67e, ä µövov tifjs toi Xöyov Evvoia; -28

Seeing that reason is two-fold according to the Stoics, being internal in one respect, and expressed in another, and again being correct in one aspect, and incorrect in another, it is appropriate to distinguish which of the two [kinds] they deny to the animals.... Since then reason is two-fold, being present in speech, while being present in the (mental) disposition, let us take our starting point from the former, which is ordered according to the voice. So if reason expressed is vocal sound [uttered] through the tongue, semantic of

Chrysippus' denial that ravens, crows, and children can truly speak and Varro's view. that these lack syntax with modem debates on language, thought, and rationality. 26Asreported by Sextus, the 'dogmatists' hold that only humans have a notion of following or consequence(6EKokouOia), which amounts to the capacity for drawing inferences. (Sorabji, Animal Minds and Human Morals, 27 points out that the propositional character of the perceptual appearancesof animals is not here denied. Sorabji comparesand translates a very similar passagewhich lacks mention of the Stoics, Plutarch, E ap. Delph. 386F-387A, as well as another Sextan passage(Math. 8.270-271)which has Stoics assigning an understanding of conditionals and the signs associated with them to humans and not to animals. The Stoics define the sign in terms of the protasis of a sound conditional. ) Sextussays a bit more about this elsewhere at Math. 8.288(Mutschmann and Mau). Here the point is that human beings make inferences which have to do with temporal or logical relations between things; some inferences are based on temporal rules. Burnyeat, "Origins, " 209ff. has pointed out that the argument dependson the additional premiss which can be supplied from Math. 8.285-286,that human beings are equipped by providence with their distinctive nature and construction, which provides a nice soft spot for Sextus to attack. But also Sextus argues (Math. 8.287) that the external must be accompaniedby the internal k6yo; in the case of animals, for otherwise the speech which is expressedis deprived of reality (6Evt)n6cYracoq),which looks like a rather weak and obscure reply to the Stoic use of the distinction. 27Pohlenz, "Die BegrUndung, " 192, notes the agreement both in content and in presentation between Sextus and Porphyry, and on this basis posits a common source. But he then strangely uses this observation as leverage in order to compromise Porphyry's clear testimony to the Stoic origin of the doctrine: "So müssen wir durchaus mit der Möglichkeit rechnen, daß Sextus die Quelle reiner widerspiegelt und erst Porphyrios in ungenauem Bericht die Stoiker zu Begründem der Lehre gemacht hat. " 28Porphyry, Abst. 3.2 (Natick, 187.20-188.20). In the translation, the square brackets supply what is not in the Greek but serves to fill out the sentence.

122 the affections which are inward and of the soul [Porphyry breaks off the sentence here]; for this definition is the commonest and is not at all the property of a philosophical sect, but only is based on the notion of reason. Porphyry then defends animal speech according to this definition, insofar as animals have their own language, which they possess not by convention like humans but in conformity to the guidance of nature and the gods29 It so happens that humans cannot understand animal speech, the same line as Sextus takes (Hyp. 1.73ff. ) 3031 Hopefully, by now the relevance of the Stoic distinction to the rationality debates of Antiquity is clear. We are in a position to take up the Greek Christian tradition on mind and language. In Hex. 9 and Hom. 332of Basil the main strands

29Atherton, The Stoics on Ambigitity, 147 note 11 criticizes Porphyry's argument, noting that he seems to miss the important relation between syntax/ linguistic structure and rationality for the Stoics, although he has much to say about the dialectical capacity of animals. She also points out that Sextus' usage of the %6yo; distinction leaves to one side the issue of "where understanding of linguistic structure lies. " 30Porphyry, Abst. 3.6 (Nauck, 194.3-9). Soon after (Abst. 3.7), he turns to a discusson of inner reason (-t6v &T6; ainCovKai &Btdft-cov), but seems less interested in it than Sextus. 31Writing on the basis of the same tradition, the Imperial Platonist philosopher Plutarch's entertaining treatise On the Cleverness of Animals plays variations an some of the same philosophical themes although without bringing in the thought and speech distinction. Plutarch presents arguments in this anti-Stoic treatise in favor of considering animals as rational and indeed as dialecticians. First, he lists the definitions of several basic technical terms (icp60satq, ýntpo%ý, Tcapacwm)ý,and gwlgý), all of which are used for cognitive faculties and functions which could well have been found in introductory philosophical treatises (cicraycoyai) of the Stoics. Plutarch, Soll. Anim. 961C-D (Helmbold) claims that "there is not one of these [technical terms] which is not associated with rationality, and all these things belong to all animals. " Then we have the dialectical animals: Plutarch, Soll. Anim. 96817- 969A (Helmbold) argues that the fox is a dialectician when it is on the ice and listens for noise to test whether the ice is thin or thick, employing a 'deduction from perception' (ýý alooAacco; av?Aoyta1t6g). The form of the argument can be expressed as follows. If A, then B; if B, then C; if C, then D; if D, then E; therefore if A, then E. The role of perception here is particularly important. On the other hand, the dialectical dog, unlike the fox, is understood to be the creation of 'the dialecticians' at Soll. Anim. 969A-B (Helmbold) = FIDS 1159. The dialecticians themselves give the account of the dog's use of the multiple disjunctive argument while reasoning with himself' ((Yi)?,ý.oyiýEaOat np6; ýauT6v). Plutarch comments that according to them, only the minor premiss of the syllogism is supplied by perception, while reason supplies the major pren-dss and the conclusion. In any event, he denounces the whole Stoic story as false. The dog relies in this case on perception alone to track the wild beast, and not on disjunctive and copulative propositions. Other activities show the dog's rational capacities much better (Soll. Anfin. 969Cff. ). Plutarch's treatise defending the rationality of animals is of relevance to the material in Basil not only by virtue of its content, but also on account of its character as some type of school exercise. William Helmbold takes up the question of scho ol- traditions behind Plutarch in his Introduction to Plittarch's Moralia 12, eds. and trans. Harold Cherniss and William C. Helmbold, Loeb Classical Library (Cambridge, MA and London: Harvard University Press, 1957). Basil may have read this popular treatise. Plutarch shows how the Hellenistic debates are passed down through indirect channels. 32jean Bernardi, La Pridication des Nres Cappadociens (Marseille: Presses Universitaires de France, 1968): 67-68 dates Ho?n. 3 to the period of Basil's priesthood, a few years before the

123 of the philosophical tradition as sketched above are reworked into a Christian exposition of particularly impressive features of the created order, which show the wisdom of the creator. Basil discusses the untaught, innate nature (68t8diccoq ýA'Xytq)of all creatures, which bear the mark of their maker's wisdom. There are important distinctions to be made between humans and animals. Basil points to reason (Myoq) as the factor which differentiates humans from other animals: "In the case of irrational animals, the familial bonds between offspring and parents are untaught and unreflecting, by virtue of the fact that the God who formed them has made due allowance for the lack of reason with the excellence of the faculties of perception. "33 A slightly different view is presented shortly before this text from his homilies on creation, where Basil says that the difference between animals and humans has its basis in bodily posture, insofar as it reflects the soul. 34 Basil is confident that animals possess souls, but he thLinks that they are still without reason in spite of all that sense-perception does.35 The relationship of animal perception and rationality is elaborated a bit more in the following passage from Hex. 9, where Basil uses several animals as his examples, including the important case of the dog. He discusses the loyalty of dogs to their masters, but most impressive of all is the dialectical ability of the dog.

Aoyov µßv ägotpoq o xvcov, icoSvvaµovaav Sý öµwS c6) X, ycp n'v a6otrj6ty 9xF-t. "A yap of Kazä no?, % v 6xoXAv tiov 3iov Ka8sý6µsvot g6XiS EEsvpov of -rov Kö6µov aoýoi, roq rcov 6vXXoytcµ6v XEyw 70,oxäg, Taütza Ssixvviat naps Tilq ýüu o) ö Kücov nenatSEVµavoq. Tö yäp ixvoc tioi Oi piov StepauvthjEvoc, Enetöäv 65prg aütO noý, vipöncws oXtý6gevov, tidcS EKaßiaxov 4 cpov ag EKipon&q e.tt Owv, µovovovxi tv ai XXoytcntid v ýwvhv ä4ilj6t Si' wv npdaasu "H Ti v6c, 4'naiv, &rp6mil Tö Orrpiov fi ri v& rj Eni r68s zö pEpog & ph v oü'rs iijvft otte 'n v8F-, XEtnöµcvöv eatit tE 6)pµnjo8at avtÖ Kai oviw tj ävatpa(YEt tiwv iyrsu6wv E1)pIaKEt 'tö 6X116es. Ti neputakEpov

dating of Honz. 16 generally agreed upon in the literature (371-372). Moreover, Bernardi sees important points of similarity between Hoin. 3 and Hoin. 12 on the beginning of the book of Proverbs, which he dates (56) as a quite early homiletical work (about 364). In Bernardi's opinion, Hoin. 3 "sent plus le d6butant que Fhomme fort de son autorit6 et scir de son audience. " 33Basil, Hex. 9.4 (de Mendieta and Rudberg, 153.4-6): 'AliýXav6q LyTtv ý owpyh 'Ecýtq67,6yotq -u -T6Kwov im! yový(t)v np6g eikkilka, 5t61t 6 Bilgto-upyA(Yaq ain& E)F-6; Av co, k6you Wctytv 8t& T-qTI Tcov ai(309-nipiow irEpto-ouiaq napF-ftuOAaavTo. 34Basil, Hex. 9.2 (de Mendieta and Rudberg, 148.23-149.6): &%X& x6 oýpdvtov ý-U-r6v 6 dvOpconoq 6cyov, r6 oxA[tact Til; oa)gaTticil; 5taTr%dCFECOq,TOGOUTOV Kat T6 &ýtd)ga'Ct YUXýg 5IF-VAVOXF-V. The editors adduce for comparison texts from Plato fffiii. 90A-B) and Aristotle (Part. AlliIII. 2.10, 656alO-13 and 4.10,686a27-28). 35Basil, Hex. 9.3 (de Mendieta and Rudberg, 149.11-149.13)

124 lcoioüaty of Elri To-)vStaypaji tht v acµvwv xa6Eý61Evot Kai ci v xövwv Ka'taxapäaaovtiec, Tptwv itpoTäaewv ävatpovv'tES täS Svo xai EV Try Xstnoµýv q trö &%TjoýSýýsvpiaxovtES 36

The dog does not partake of reason, but nevertheless possesses a faculty of perception which accomplishes the work of reason. Seeing that the things which the wise men of this world have discovered with difficulty, working them out with much diligent study, I mean the construction of syllogisms, in these things nature instructs the dog. For the dog, while pursuing the track of a wild beast, when he finds that the track splits into several different directions, approaches each path which bears Off, by his actions all but emits the syllogistic utterance: Either this path, he says, the wild beast has taken, or this path, or in this direction; but neither this path nor this path, so the remaining option is that the beast has gone off in this final direction; and so by the denial of the false alternatives he finds the true one. What more do those men achieve who apply themselves to impressive diagrams and draw figures in the dust, by denying two of the three disjuncts and in the remaining disjunct deducing the true one? Basil remarks that perception (dioffilatq) which amounts to the work of nature (4ncnS) does the work of human reasoning (Xöyog). By contrast, Plutarch's account of Stoic canine syllogistic is finer-grained in that only the minor premiss is supplied by perception, while reason supplies the major premiss and the conclusion. The relevant point is that the notion of inner and outer reason would presumably be in the background of his distinction between the work of nature and reason. For, as sketched above, the Stoics apparently work Mpg £v&äOETos and Mpg npo4optK6Sinto their discussions of rationality. The dog 'all but utters the syllogistic speech' (µovovovxi tf v av7Xoyiattnd v ýcovfiv it is significant that Basil's report agrees with Sextus' inclusion of 'virtually' in his testimony of Stoic animal reasoning. That is, Basil thinks that the dog accomplishes, under the guidance of nature, what could be arrived at by means of an argument by a human being 37

36Basil, Hex. 9.4 (de Mendieta and Rudberg, 153.21-154.3) = FDS 1157, cf. Aelian, Nat. Aninz. 6.59 (Hercher, 166.10-28) 37Taking the idea about nature and perception a bit farther, Basil, Hex. 9.3 (de Mendieta and Rudberg, 150.20-151.10) accords a degree of rational understanding to animals with respect to the prognosis of things that are consequent upon other things: "And the foreknowledge of changes in the air, does it not reveal some sort of rational understanding? " (Ai R 7rpoyv6CTEtq TIjq n6pi [6v &ýPc( gETC(POX11;TEOictV Oi)Xi (316vecrtvXoytKO dnoicpýnTo-oatv; ). Cf. the Stoic theory of .signals* or 'natural signs' with a deictic reference to a fact presently perceived, treated by Charles Kahn, "Stoic Logic Stoic Logos, " AGPIz 51 (1969): 158-172 161; 164ff. Basil and at - adduces the examples of sheep, oxen, and bears, and then remarks on this testimony to the

125 The basic point is already in a similar discussion in an earlier homily (Hom. 3) where the point is that all that is needed with a view to what is good and suitable for irrational creatures is secured by the guidance or control of divinely instituted nature, while the same ends are served in the case of humans by means of reason. Diö xat 7jµiv 6 nathsüwv ý taS Osög 'tö tEya tiovro lrapäyyEXµa öE& csv, tva ö tep bbcivotg ýx 4v6swS, iov'ro hµty ýK 'r S 'toi Xöyov ßoiiOe ac itpocyavrjiat, xal ö7EEp KaTOpOovzati coiq äX, yotq ävEmta' tarn;, zovio nap' hµwv ant'tEAi tat &ä tires itpoaoMýq xai Crq avvExov; E6V ?,oyiaµwv imtßiäacwS 3s

So God, who is our educator, has issued this great injunction [i. e. to attend to oneself, 7tp6cexe aeuutw], to the end that whatever is to their benefit from nature, might be added to us from the saving assistance of reason, and that whatever sets irrational animals aright without their paying attention, is gained by us through the work of attention and the continual application of reasoning. It seems that this notion of human rationality is forecast in the introductory section of the homily (Horn. 3.1), with an idea of the relation of the mind to speech. The central idea is the revelation of what is hidden in the human mind through speech, in a way somewhat different from standard Stoic theory. Toi Xöyov tihv xpýcrty 96wxcv ýµiv 6 KTiOag ýias OF-6q, Iva rdq f3oiXäs t& xap&c v &X?i Xotg äTtoxa7tcoRav, Kai &t Tö Koivwvtxöv 'tfjS Svc a(o; ExaaioS 'c,) icXIlaiov µaTa&t&w tav, CM p Ex ztvuuv 'taµtaio v, 'cwv xpvTEi6v 'Ejg Kap6iaq 7tpo4epovieS 'cä POIAE taia.... bEP-Oý 5ý vicö icapatrc'cä tact ßapxi xaP,vnzoµhvrl i t6v i ivuxtj Viq Evvoias ýpyäýETat, prjµäiwv SaiTat Kai övoµäicov 7tp6s 'tö ST]µoatsüaty iä & tic 0666t Kaiµeva. ' Enat8&v ovv iroie Moil cat 4wvij T19 ýj Evvota ilµwv, woitep 7topOg6T'ttvi tic XöyuwEltoxovwEvrl, Staitepwaa Töv 69pa, Ex'oi 4 OaiyotEvov µEtaßaivet TrpOSTöv äxovovta 39

God, our creator, has given the use of reason to us, in order that we might reveal the deliberations of our hearts to one another, and each person might impart something to his or her companion by virtue of what is common to human nature, as if speaking forth the deliberations from some inner chambers, from the hidden regions of the heart.... But since the soul, hidden under the veil of the flesh,

providental governance of the cosmos in terms familiar to Stoicism (5iii ndwcov 81ýKmv, ýTctýtO,F-tav) as well as the 'perception of the future' (rou- ggUovTo; ciiaOTIatq) which irrational animals exercise. 3813asil, Hom. 3.2 (Rudberg, 25.12-16). This homily is not easy to date with much degree of confidence. Seeing that the Hex. is agreed to be one of his latest works, Fedwick, "Chronology, " 9 note 32 remarks sensibly of Hoim 3, "although probably earlier, this homily is a fitting complement to the unfinished Hex. 1-9. " 3913asil, Honz. 3.1 (Rudberg, 23.1-11). The concealed in thought as opposed to revealed in language theme is developed further, Hotiz. 3.1 (Rudberg, 24.17-25.7).

126 produces thoughts, it requires verbs and names with a view to making public the things stored in its depths. So then whenever our thinking takes to itself semantic vocal sound, as if being conveyed with speech as a kind of ferry, crossing through the air, it arrives from the speaker to the hearer. Here human reason as divinely instituted forms vocal utterances for the sake of expressing what is in the mind. Thought takes on language (which is termed Mpg) with a view to attaining a vehicle of expression. We might say that here is a notion of a 'language of thought' wl-dch, if strictly faithful to the Stoic legacy, has structural correspondencesto natural languages. However, the text is rather uninformative on mind and the content of mental states: Basil only says that thoughts from deep witl-dn the speaker emerge and hitch a ride on utterances. The utterance is encoded in natural language by the speaker and then on the other end is decoded by the hearer, suggestive of contemporary communicative theories of language and the mind. 40 There is certainly no notion here of inner speech conceived as a silent monologue or debate, for the soul is just occupied with private thoughts. On the other hand, distantly echoing an important Stoic idea, Providence endows humans with a distinctive nature which is equipped for rational inner thought and its expression through speech. These points on the legacy of the Hellenistic debates to Patristics can be rounded off with a few notes on Origen. In essentials, Basil and Origen agree on how to establish the demarcation between humans and animals. 41 In one text

40Asidefrom recent defensesof a 'language of thought' sparked by the classic work of Jerry A. Fodor, The Languageof Thought (London: The Harvester Press, 1976), the idea in the homily is suggestivefrom a semantic point of view of David Kaplan's reading, "Afterthoughts, " Themes From Kaplan, eds. Joseph Almog, John Perry, Howard Wettstein (Oxford: Oxford University Press, 1989): 565-614 at 599-601, of Frege and Russell, which posits an 'autonomy of apprehension'in a 'subjectivist semantics'. According to this view, propositions are accessible to thought and can be apprehended independently of language. That which is thinkable is epistemologically prior to language, and language serves as the encoding of propositions into appropriate expressions. "Language, of course,aids communication, and also makes it easier, perhaps even possible, to reason using very complex thoughts. " Users of language assign meaningsto the words used (meanings are not'downloaded' from words used by others in the community), while the linguistic community provides syntax to the user who supplies the semantics. As Basil has it, individuals express their thoughts by means of the divinely- instituted benefits of linguistic interchange, and these thoughts are in turn apprehended successfullyby hearers (decodersof utterances) on the basis of certain key characteiistics th at all people have. 41Thequestion of how well Basil knew Origen's immensecollection of works is bound up with the the difficult question of whether Basil and Gregory of Nazianzus jointly compiled biblical Philokalia, an extant collection of passages of exegetical interest from Origen's commentariesand other works. The issue is extensively treated in the introductorY chapter, "Grammar, Rhetoric, and Philosophy".

127 from the work On Prayer, Origen employs the notion of mental 'attention' as the mark of human rationality: Eäv fiE napaxoXovOIJ'Et Tj i,8ia xtvii et, Eit&. toüto &' aüiov xtvai Oat wvoµäaagev, äväyxrý Tovio eivat Roytx0 42

But if [the animal] pays attention to its own motion, since we have termed this [animal] to be moved 'by means of itself', this [animal] must be rational. In fact this text is widely regarded as good testimony to Stoic views of animal minds, hence Stoic ideas could perhaps reach Horn. 3 by this route. Another passage on the issue is from the debate of Origen with Celsus. ovS' Ev tovtotq SE EwpaKE, tivt 8La4apet tiä änö Xöyov 'Kai XOYMrq [4üaEwS] Envmkol)g8va c6v än ' ä?, yov 4üßECOS Kai Ka'CacKeuT ytvoµhvwv. c)v tjv ai'tiav ovösIS µßv ý-vunäpX v cotq nocovai. Xöyog ävaMxsiat (oW yap gxov6Lv a&TÖV), 6 irpe6ßviazog SE Kai via; µßv 'roü Osoü zräv'rwv SE 'rwv vIcoKetµgvccwv (3aatXc1')g 4üaty äxoyov irstoii Kc, poil0ovßav wS äxoyov 'roi, ovK äýuwOcýtcn Xöyov 43

Nor has he [i. e. Celsusl seen in these matters, in what respect the [creatures] which are perfected from reason and rational [nature] are distinguished from those which maintain their existence from irrational nature and mere construction. For those [creatures] for whom no reason inherent in them when they act is responsible (for it is Us that they lack), the most revered ruler of all things subject to him and son of God has created irrational nature, as an irrational assistanceto those [creatures which were not deemed worthy of reason. The view of both Origen (cf. Princ. 3.1.2-3 = SVF 2.988) and Basil is that animals perform actions under the guidance of nature in such a way that animal minds are 'hard-wired' in response to the environment. Animals set about obtaining what is serviceable to their well-being, but without the application of reasoning with a view to gaining their ends. These Patristic texts bear the marks of the Hellenistic debate on reason and animal rationality, which had employed the Stoic distinction. In Basil we see that inner, unspoken reason is involved in distinguishing human rationality. Rationality is conceived as a divine gift which has to do with language even in passages where Basil is not making clear that thought is of a linguistic character or has the character of 'inner speech' conceived in terms of structural similarities to natural languages. In the next part we will look at how the philosophical

420rigen, Orat. 6.1 (Koetschau, 311.8-10) = SVF 2.989; cf. Epictetus, Diss. 1.16.18-10 and 2.8-3-8 (Schenkl) 430rigen, Cels. 4.81 (351.7-13). Square brackets supply what is needed for the sense.

128 background relates to the discussions of inner and outer speech in selected homilies (Hor. 12, Hom. 16, Horns. in Psalmos).

III

Max Mühl has already shown in an influential article that the philosophical background to X6yoq ýv8tdftcoq and Mpg npoýoptic6q makes its mark on the theological tradition of the first few Christian centuries. Yet beyond his work, very little on this point has been said about the great Patristic writers of the later fourth century. I think that an important (if unexpected) use of the distinction appears in an important doctrinal homily of Basil (Honz. 16) on the opening sentence of John's gospel, ýv d(pXýTIv 6 A6,yo; "in the beginning was the Word. "44 Two basic problems present themselves. The first problem is why Basil takes up the distinction for theology, along with the surprising statement that the ?,6, yog issues from the mind (ýic rol) vou-) analogous to human speech, in a departure from his major theological predecessors and contemporaries:

44Basil, Hom. 16.3 (PG 31,475C-479A). This homily tends to be treated in scholarly literature together with Hom. 15, which bears the title "On the Faith". Paul J. Fedwick, "A Chronology of the Life and Words of Basil of Caesarea, " Basil of Caesarea: Christiati, Hilmanist, Ascetic, ed. Paul J. Fedwick (Toronto: Pontifical Institute of Mediaeval Studies, 1981): 3-19 at 9-10 thinks that Hom. 16 most likely can be dated to the same year (372) as Honz. 15 and yet earlier than Honi. 15, which bears the likely dating to 7 September (10 note 39). This dating would place Hoin. 16 securely before the composition of the major work on the Holy Spirit and around the beginning of Basil's separation from the views of Eustathius of Sebaste. Bernardi, Pr9dication, 86-87 also dates Hoin. 16 to the same period as Hom. 15. Bernardi argues that since the latter is more theologically advanced insofar as it treats the question of the Holy Spirit, it postdates the former and can presumably be dated to the early episcopate of Basil, prior to the purported September 372 dating of Honi. 15. The personal watershed of the break with Eustathius is discussed sensitively by Rousseau, Basil of Caesarea, 239ff.. If the chronological consensusfrom these scholars is correct, then the boldness and spiritedness of the homily may provide a useful indication regarding Basil's personal and theological outlook at this stage of his career. The increased pron-dnence of the Marcellan issue, so important for the theological background to Hoin. 16, loosely supports the view of Fedwick and Bernardi. As Rousseau, Basil of Caesarea, 246-247 aptly says of the development of Basil's views: "How quickly Basil developed more precise misgivings about Marcellus is difficult to judge. His willingness to identify with the somewhat different emphases of Basil of Ancyra [Marcellus' successor]might suggest an early date. His own more sharpened opposition, however, does not surface in his writings until as late, perhaps, as 371, when he began to make contact with Athanasius.... Very quickly the presumed heretic became a bete tioire pursued by Basil in letter after letter. Sermons also began to address the Sabellian issue. " Rousseau cites in connection with this I as t sentence three homilies, namely Hoi?i. 24, Hotiz. 27, and indeed Hotit. 16; the important and frequently cited Hoi?i. 24 is given by him an earlier date (around 372) than others, such as Bernardi (87-88), have suggested.

129 Eusebius, Athanasius, Meletius, and the leaders of the Homoiousian party. 45 We far cannot go too into this complicated question in this chapter, due to limitations Secondly, there interesting of space. are points in the relation of the ideas Hoin. philosophical tradition to the of 16.46 1 adopt the position in tl-ds that least human chapter at as regards rationality Basil represents a kind of but Stoicizing Christianity, only in a very weak sense. I mean by tl-ds the Christian Stoic transformation of ideas inherited from Hellenistic philosophy. from Platonism, Stoicism After all, aside was the dominant pl-dlosophical school Imperial Period, but logic. in the in ethics also in 47 I note in passing that traces of Stoic ideas can be seen here and there in the Eastern Origenist tradition which Basil inherited, particularly in Origen himself, but also in Dionysius of Alexandria and Gregory of Nazianzus. 48 Now we will see how the distinction works in the critical theological arguments of Horn. 16. ' EntýijTCi Sý hµwv h Stävota, tic i 1v Ev a'pjjj; '0 Aöyog, 4ßi. floioq Xöyog; ö ävOptTCtvos Mpg; ä%%' 6 Tccv äy Xwv Xöyog; Kai yap ý,viýaio ijµiv ö' AT1ö6ToXog, wS Kai tiwv äyy6Xwv i6iav ýxövtwv y%cw66av, Ei tthv- ' Eäv 'cats y%6aaai, s 'r(OVävOpciitwv ? a%(Oxai tiaiv äyy6Xwv. ' AX,%& xai 'tov 2,6yov &m? 'rtq Eaity gvvota. '0 µßv yäp 'riq Eaity 6 Stä 'rýS owvijg irpooEpö ievog, ov'ros 6 µ6,c6-c6 TrposvcxOývai zw

4-5Christopher Stead, "The Concept of Mind and the Concept of God in the Christian Fathers, " The Philosophical Frontiers of Christian Theology, eds. B. Hebblethwaite and S. R. Sutherland (Cambridge: Cambridge University Press, 1982): 39-54 at 39 claims that "in the fourth century Eusebius and Athanasius seem to be noticeably reluctant to characterize God as Mind. " Stead makes virtually no mention of Basil in the entire paper, despite discussing briefly the distinction in the concept of X670; (50) as obscuring the "very important distinction between intuitive and discursive thinking" which he finds in Plato. Marcellus is given a few sentences, and a positive use of the distinction is attributed to Athenagoras, Theophilus, Irenaeus, and Tertullian. Stead's contribution is heavily preoccupied with Plato and the Platonic tradition. 46We have already seen that talk about the philosophical roots of material in fourth-century Christian texts such as -these Patristic homilies is talk about the reworking of widely disseminated philosophical notions. For the 'Stoic' ?,6yo; &8td0ero; and Tcpoýoptic6;is passed down into the Imperial period by Platonists and Peripatetics, empiricist scientists and Stoic- influenced philosophers. 47The case for substantial later interest in logic in the Imperial Stoa, up to and including Marcus Aurelius, has recently been made against received opinion by Jonathan Barnes, Logic and t Ile Iniperial Stoa (Leiden: Brill, 1997). 480ne text from Gregory's poetry should be cited at this point, Carin. 2.1.12,267-269 (PG 37, 1185), which bears the title De se ipso et de episcopis. "EXEt y&p oiý,co-,; - 5trT6; hg-tv 7t&; ;Wyoq, / AgýEt; -re, xal vol-)g-ai pýv, 01ov gicroftv f'E(Y0qg', 6 8' 9v8ov crEpActýgýtF_cYA&ov. Also cf. Gr. Naz., Or. 28.13 (Barbel) and Gr. Nyss., Din. 1.539-542 (Jaeger, 182.17-183.15). The considerable evidence for traces of Stoic logic in Gregory of Nazianzus is introduced by Frederick W. Norris, Faith Gives Ftillness to Reasoning. The Five Theological Orations of Gregory of Nazianzen (Leiden: E. J. Brill, 1991): 27-31.

130 &ýpl ýVaTEOW)gevog. 6 66 liq ý(Y'Ctv 6 ýV6t6op-, 10q, ývlmdpxwv ýg(ov 'rcýtq Kap5iatq- 6 ývvoijgacuc6q. Kai (Moq, 6 rEXvtK6g X6yoq. "Opa AýITOT9 CFERapaKpO'6CFqTat T6 6A6)VA)AOVTýq 'AV dKoue uoz Myou Op-onpemýq. flF-p! yap roZ MovoyevoZq 8tccXey6gF-v6q (Tot, A6yov 617TEVOCUTOV. "9207CEP OA)VKat ý(O; gec' W"JOV Ep6t, Kai ý(OhV, Kai dCVdGTaGtV, Kai 015TE OC-0; 6KOýCraq, blt T6 ai(YOnT6V TOUTO Kai 6ýOuXpoýtq 6paT6vKaTa#pj Q, OikE ((OhV &KOý)Gaq, TýV KOtVhV Ta6MV VOJq, fiV Kai T& &XOYCEý1,1* Ol')T(O Kai A6yov &KOý)COV,ýIýXaýat RýTEOTE Tý Gý &(YOEV6iQt Týq 8t(XVO'tC(; np6g XagatýýXOI)q Kai 'IC(IEEtV&q 8taVOia; 1571EVEXOýq. A? X ýp6vccdv 5tdvotavcou p'ilgaToq. At& d A6yoq; "Iva 8F-txoý 6Tt ýK TOý VO1) npoýXOE. At& -Ti A6yog; "Ou &naO6g ý,yF-vvýOij. At&, ci A6-yog; "OTt 6iK6V'EOZ'YF-VVA(YaVTO;, UOV ý-V ýCCUT6561KVI)g T6V 'fP-VVA(3C(VT(X,0ý&V ýKEWP-V &nOgEpi(Yaq, Kai T6XEtOq lblt6(PXCOVKaO' ýal)T&* 6; Kai 6 ýggrspoq Myo; 6XljV llg6V dCICELKOViýEtAV 9VVOtaV. "A Yap KaT& K(Xp8iCIV &EVOýGagEV, 'T(XZ'Ca Tý OýgaUt 7EpOTlVgyKagF-v, Kai 9CFTt TOZ b Tý Kap5iý( voýgmoq duEctKowaga T6 XaXoiýptevov. 'EK Yap 'Cot) nF-pt(Y(YElUAa'COq '111; KC(p8iC(q 6 ?,6, yog npo#pcTc(t. Kai 9,uctv gýv Kap8ia ýýtCov olov Aic6q Etq bc Týq nijyý; -Taýmjq &wv. Tocroý'cov rIIII1 11 11A ouv TO alroppEov, To irpcomg 6vaýEp6gevow Kai otov T6 lVa KEKpuggývov, To(youTov-Ocyov Kat I T6 ýatv6gF-vov. A6yov Ol')V EITIF-V, 'EhV dcnaoý Got Yývvljatv '[01) rIaTp6q Rapaarý(Yl. j, Kai AV Texp-iav ý)Tcapýiv (101 TOý YiOý OEOXOYýGn Icat' AV &Xpovov auvd#tav rol) Yiol) np6q Ila, rgpa 5t& Tuftow &56ýTjwtt. Kai yap o ilpkepog X6yog Eo-o vo-u ,Ygvvllga, 6nao(ýq YEVV6)gcvog. 0ý, rs 'Yap r6gVF-,cc(l, oiý,rp- gF-piýp-,rat, OgEt' &W ý i5iq( Oiý'E6 ggVO)V* 6Xog 6 voýq & Ti, I 01)(YEdGEt, UOV T6V X6yov Kai dmilputagbov ýýtmqut- Kai npoeX06)v6 2,6yog,n&(: vccv r6 YgVVAGUV'EOg Vol) ThV 51ýVagtV F-V EaUTO) 7rEplgXF-1. "OGOV OA)V Si)GF-Pýq, MOVOYEVOI)q OF-OXOYtC(V TocY6'covXaOd)v np6q AV T6 ELK cTiq Tou Myou ý(ovijg, 6nep &v ciýpijq dumgýatvov Kat avapgocyTov ýatv6g6vov, 'Eol-)xo napai"cat, Kai 157r9pPijOtgijXavý ndorn 'Ev &pXý I'v 6 A6yog. Ei R P-tnev- 'Ev &pXý, 6T i6g, Tq- Yiol) amp-toýXOF-v dv crot 7"lv 7rpo(y-qyopiaI rol) 7rF-p!c6 TcdOoi)ggvvota. 'EnEt8h yap nap' hýCtvcdc yevv&, ra Xp6vq) ycvvý, Kalt ýgnaO(ýq yevvý, 8t& r6, co npoXap6)v, A6yov clice, RpO8tOPOOýgCVOq'EC'Cq O'CTEPP-TCEtqlbnOXýlVEtqjVC( CYOU ThV XVUXýV (ITP(OTOV 8ta(yC6(yjjTaL. 49

Does our n-dnd seek to know, who was in the beginning? 'The Logos',he says. What kind of logos? The language of men? Or rather, the speech of angels? In fact, the Apostle has indicated to us, that the angels speak in their own tongue, saying, "if I should speak with the tongues of men and angels." But also there are two meanings of 'logos' which can be distinguished. For one sense of the term means the logoswhich is expressed by means of vocal sound, the speech which dissipates in the air after being uttered; and another sense of the term means the speech which is inward, located in our hearts, the speechwhich is involved in thought. And there is another, the expertise that is employed in the arts. Observe the sensesof the term carefully, that the homonymy of the word

49Basil, Nom. 16.3 (PG 31,476C-480A)

131 does not mislead you.... But instead hear the Logos in a manner that God. is worthy of When he [i. e. John] speaks to you concerning the Only-Begotten, he calls him the Logos. So then just as he will call him a little later light, and life, and resurrection, and when you hear light, you are not carried down to that which is perceptible and visible to the eyes, and when you hear life, you do not think of this mundane existence, which even the irrational animals live; just so when you hear 'Logos', beware lest ever you are carried down by dint of the weakness of the mind to earthly and lowly thoughts. Rather, search out the meaning of the term. On what account does John use the term of logos? In order that it might be made plain, that he issued from the mind. Why logos? Because he was generated without passion. Why logos? Because he is the image of the one who generated him, making known the one who generated as a whole in himself, being no partition from him, although he exists in his own perfection. So also our logos represents our thought as a whole. That which we conceive in the heart, we express in speech, and that which is spoken is the image of the thought in the heart. For from the outflow of the heart, logos is brought forth. And our heart is like a source, while speech when spoken forth is like a stream, which flows from this source. So great is the outflow, so much also is that which is referred back to the origin; and of what is hidden, so great also is that which appears in the open. He uses the term 'Logos', in order that he might set the passionless generation of the Father before you, and speak theologically to you of the perfect reality of the Son, and through these things indicate to you the timeless union of the Son with the Father. In fact our speech is the offspring of the mind, begotten without passion, for it is not severed, nor is it divided, nor is it effluent; rather, the mind as a whole remains in its own nature, emitting speech which is whole and complete. And the logos which issues outwards contains in itself the meaning in its entirety of the mind which generates it. To the extent that is pious and reverent, appropriate from the term 'Logos' with a view to the theology of the Only-Begotten; whatever you find which appears alien and unfitting, this you must reject, and step beyond it by every means. "In the beginning was the Logos." If he had said, "in the beginning was the Son," the notion which involves passions would enter along with the term 'Son'. For since human generation takes place over time, and involves passion, for this reason he says 'Logos', correcting in advance the unsuitable conceptions, to. the end that he might keep your soul from harm. So here we see the analogy to human speech and the mind at work. When the semantic ambiguity is resolved into the distinct senses, Basil uses various forms of a verb (lrpo4Ep6µcvog,zrpo4 epc v) which is a cognate of the descriptive term in the philosophical tradition (itpoýoptxös). On the other hand, the term for inner

132 reason or speech (ýv& Ocioq) is of course identical to the term in the philosophical tradition 5° So the theological use of the distinction amounts to illustrating the relation of the divine X, yog to the inner nature of God, by the notion of speech which is expressed, originating from the mind of an individual who uses speech for communication. We miss here a notion of the theology of inner speech as divine thought, but despite the passing mention of irrational animals in the text we only hear about the divine mind. (Recall the other texts above in Basil where the idea that thought is constituted by language or has the character of 'inner speech' is passed over. ) The passage reworks the classic Greek theology of the image (Eixwv) of God in a striking way; care is taken to point out the perfection ('tOXetog)of the image, and the characteristic of permanence, of existing in its own right, balanced by the union of image to what is imaged. The supposition that Basil is working with standard ideas about the relationship of speech to thought represented in the philosophical traditions known to him will be shown to be partly incorrect in what follows. This is because there is surprising originality to the philosophy of mind and language behind the theological arguments. The most interesting philosophical points are (1) that speech is a kind of image of the speaker's mind and (2) the passionless, timeless utterance of speech. Let us now approach the first point in terms of the philosophical background. This idea of speech as the image or reflection of the mind of the

50Someremarks can be made regarding Basil's unexpected insertion in Hoin. 16.3 of a third sense of the term k6yoq, 'technical' reason (Teptic6q). What is the meaning of this third sense? The term is not particularly well-attested in later Greek philosophy, although there are several extant lists of the various senses of %6yoq. In his commentary In hit. 242.20 (Busse), Ammor-dus places the two alternative accounts -cF_Xvuc6;X6yog and 7rpoatpcrtK6; X6yo; ('account on the basis of free choice') next to one another, but does not explain what the former could mean. However, a passage in Eusebius of Caesarea helps to clear things up, Eccl. Theol. 2.13 (Klostermann, 114.10-21). Eusebius, while engrossed in an extended attack on an anthropomorphic use of %6yo; npoýoptK6; and 6v5td0eEo; in Marcellus' understanding of the Trinity (Eccl. Theol. 2.11ff. ), lists several senses of the term 'X6yoq'. He distinguishes a rational Mpg, a k6yog expressed in speech, a written X6yo;, an 'inseeded' Myog, and finally a 'technical' k6yoq: "And besides these, they standardly assign another sense to logos, which is the expert knowledge of a particular art or theoretical field, a grasp of all the pertinent theories, such as for example medicine or architecture or geometry. " This may be roughly the same understanding of Basil, as a passage from one of the (authentic, by Bernardi, Pridication, 22-23) homilies on the Psalms, Ho?ii. l?z Ps. 14 (15) (PG 29,249D-280C) suggests. In the context of a discussion of moral theory, Basil draws a distinction between a layman who performs medical help without knowledge of the medical art, and the same healing activity as performed by the qualified doctor, Holn. III Ps. 14 (15) (PG 29,256A); cf. Honz. 12 (PG 31,392B).

133 speaker must be understood in light of its prominence in the philosophy of the Imperial period. Ptolemy, a contemporary of Galen and a leading mathematician and astronomer of the Imperial period, wrote an extant short treatise entitled "On the Criterion and the Rational Part of the Soul". This work shows how certain features of the earlier debates between sceptics and dogmatists were passed down for general appropriation by philosophers and scientists, like Ptolemy himself. The treatise takes up various topics in the theory of language, including the traditional distinction between inner and outer speech.

Toi SE XoytKOU, Ka06 TÖ 'r v &v0p6)itcov i&tov 6ptatat, Stävota µav £atty Ö %1Äyoq o Ev&t Octog &9E oMq ttq oka Kai avaTc6?, Tj6tc Kai StÖKptatg tow [ivtjgovF-uOEVTCov, ötäXEKTog SE TÖ( ti1 4o vi S 6Üg o%a, SL JOV irpo4 pctat To1S RXI ioV Ta 8taVO1jOEVTa. Kai E tV ELKthv Ttq Ö µßv ý06yyoq a1 toü Tov vov, rug 8' bvoiag f4 wvh, Tres 5ý Stavoiaq öuäXEKTOS,Kai UO)q 6 rnpo4opt6q Xöyog Toi Ev&a0etov.

To the rational faculty, by which the special property of human beings is defined, belong thought and speech. Thought is internal logos,a kind of exposition and repetition and differentiation of what has been remembered. Speechis the vocal symbols through which thoughts are revealed to other people. The relationship is that of image to original. The voice is an image of the intellect, the vocal utterance an image of mental activity, and speech an image of thought. And generally, uttered logosis an image of internal logoS.51 Ptolemy's remarks on inner and outer speech in this passageemphasize speech as the image of the mind. So we discern in Ptolemy a theory of signs, echoing both Aristotle and the Stoics, reworked into a 'spoken image' doctrine of language. In Basil, on the other hand, there is a development of the 'spoken image' doctrine in the idea of the X6yoqas the image which expresses the thought of the n-dnd completely and unerringly. A Fregean take along similar lines might be that each thought can be apprehended by everyone and that whatever is expressed in language is in some prior sense an object of thought. 52 But the notion of Basil is stronger and less plausible in that all the contents of the mind are expressible in language and thus can be grasped by others. The idea seems to be that nothing is lost in translation into language. This would presumably be the point of the insistence on the commensurateness of what is inner and what is

51ptolemy, rlFpi KpjTjjpioL) Kai hYE110vtKob6.1-8 (Lammert). The translation is taken from "On the Kriterion and Hegemonikon, " Vie Criterioti of Truffi, 179-230 at 187. 52Gottlob Frege, "The Thought, " CbIlected Papers on Matlie? natics, Logic, and MilosopIty, ed. B. McGuinness (Oxford: Basil Blackwell, 1984): 351-372.

134 outer. Otherwise, there might be problems in the theological epistemology: Something is thus 'held back' from the expression of what God has in God's mind 53 Next we will consider (2), the passionless and timeless issue of Mpg. The principal motivation behind the analogical thinking of Honz.16 seems to be the freedom from passion (&TcaOCoq)which characterizes.the issue of human language from the mind. Basil says that understood correctly, John is particularly concerned to exclude all passion (ndOoq)from the eternal generation of the Son, and by the same coin the begetting of the Son must be timeless. But oddly enough, Basil in this passageavoids Son-language,as he does nowhere else in his received corpus of works. He claims that John rather cleverly passes up the sentence'in the beginning was the Son' in favor of the sentence 'in the beginning was the Word', with a view to eliminating harmful connotations of the term 'Son' 54 It is important to recognize that the passionless generation of the Son is a cornerstone of his trinitarian thought, for it is frequently upheld in his theological

53Anthony Kenny raised a subtle point of translation at Keele (footnote 68) about a few lines in the middle of Basil's text. Kenny questioned my translation of the lines which now read: "Why logos? Becausehe is the image of the one who generated him, making known the one who generated as a whole in himself, being no partition from him, although he exists in his own perfection. So also our logos representsour thought as a whole. " Kenny's point was that the Greek takes both adjectives in agreement with the accusative objects of what is represented within the n-dnd(divine and human), but the adjectives are not placed between article and nominal. The idea here would seem to be that what is represented is a non-partitioned whole, even though utterances are broken up into nominals, verbs, and so on. Certainly divine simplicity is a frequent concern in Basil's trinitarian thought. Kenny's point should be balanced by consideration of Basil's further statement in the text that "the mind as a whole remains in its own nature, emitting speech which is whole and complete" which speaks of both relata as whole' and is followed by a statement about logosusing the adjective 'all'. 54The idea of the passionless generation of word from mind is echoed by similar notions elsewhere in Basil. Other homiletic passagesspeak of the perfection of human reason in terms that evoke the simplicity of God, in accordancewith a strong notion of undeviating moral singleness. A good example of this can be found in Honz. 12 (PG 31,400C) on the prologue of Proverbs. (Bernardi, Pr6dication, 56 suspects that Hoin. 12 is the oldest of the surviving Basilian sermons, and should be dated to the early years of his priesthood, around 364.) "True reason,which issuesfrom a sound mind, is simple and undeviating. It always speaks the same words about the same things. But the reason which is many-wiled and geared to clever machinations, brimful of all that is twisted and clever, engineerscountless devices and turns on a dime times beyond telling, indulging the Prometheanwhims of the moment." Right thinking, as correct doctrine, is for Basil a static and straightforward rational exercise. Cf. the moral and spiritual discussion in Hoin. in Ps. 14 (15) (PG 29,253C-256B), on the words of the psalm, "who speaketh truth in his heart, on whose tongue is found no deceit". Here the idea of reasonin the hiddenness of the heart is present with the contrast between truth in the heart and straightness in the tongue. The theme of the simplicity of correct church doctrine is explored in detail by

135 works. For example, the idea is clearly present in one of Basil's major treatises, the Contra Ditiotid11111.55In the Cotitra Einimifitin there is an interesting passage which discusses the timelessness of the issuing of the Son from the Father, taking up the analogy of the human mind. voEtV SE [xp iv] ei va Tov 06ov Tov äopäzou, ov tcazä T&q Tgxvl lT6q zavtag vaTEpov änepyaß6Eißav up6q Tö äpxaTUnov, äX v7COaT7j6aVtt ,ä 6uvunäpxova0Ev Kal. TtapuýE6TTJKUtaV TG) Kai Up FtVat 'TO äpxEZUTtov oü(: Fav, O)K EKTUttwOEUYaV Stä µtgý GE(0S, 66iop Ev ü, a4payi& ztvt zijs 6XilS ýüc ewe zoiv IZaTp6q EvartoauµavOEianS ) Y1 Y si S& ßoü)Xet, ÖmoIa TWV TExvCUv ý 1Uir& Taltq EK T(ov öt&L YKÖVT(WV8X1l zoi; µaOfzsvog9vot; Eyytvoµevi ouTe Xcinov16q TtvoS Toi; SLöth KOV6t `H Kai 'IT]; TEXstwßewg wig P-KgavO6Cvou6t 7tpoaytvojEvllc. Tovzo µav OK äKptßýS tpög öµoicoßty Stä AV Toü xpövou iapäzaaty" oiKCtözepov & Eiir6v 6Tt öiroia Twv vorjµäzwv i' ýüats zotg zou you Ktvýµac tv äxpövws auvv4tßtagtVq56

Rather, [one ought to] conceive of the image of the unseen God, not in the manner of these artificial images which are fashioned later with respect to the archetype, but subsisting co-ordinately with the one who brought him into existence and who is the archetype, not impressed by imitation, but as in a seal the Son is inscribed with the whole and complete nature of the Father. And, if you will, that is what takes place in the substance of the arts when the entirety of what they are is passed from instructor to student, with nothing lacking to the teacher and the perfection of knowledge accruing to the learner. Certainly, this process is not exact in likeness due to the interposition of time. And it is more suitable to say that the nature of what is thought subsists together with the movements of the mind, apart from time. The idea of the Son as the perfect image or seat of the Father is fundamental as in Hoiii. 16, but missing is the notion of speech in relation to thought. Here Basil presents analogies which lead up to the analogy from human mental events: the seal, the passage of learning from teacher to pupil, and the n-dnd with its objects of thought. He is attracted to the learning analogy, but the problem is that it takes place over time in communication between individuals. There seems to be

Mario Girardi, "'SemplitA' e ortodossia nel dibattito antiariano di Basilio di Cesarea: Ia raffigUrazione dell' eretico, " Vetera Christianorunz 15 (1978): 51-74. 55Drecoll, Die Eiawicklung, 165 argues that Hoiii. 16 represents a step beyond the Colltra Eunotidum, for in the earlier treatise Basil had treated every word of johiz 1.1 up to but not including %6yoq,and "insofern sie die in AE gebrauchte Argumentation aufgreift und sie durch eine Erlduterung des Begriffes ?,6yoq ergdnzt. " Since Hotiz. 16 was probably delivered almost ten years after the composition of Euiz. (vd. footnote 41), it appears that Basil chose homiletic occasions later in his ecclesiastical and theological career to treat the important Christian concept of X6yog. 56Basil, Euti. 2.16,33-45 (SC 305: 64). The sqtjýre brackets supply what appears earlier and is understood in this bit of text.

136 a problem with the teaching of knowledge through language, which has nothing to do with positing 'affections' or 'disturbances' of the n-Lind in the process of communication. Thus, there is need for a further analogy which draws only from what happens Withiii the human mind. Perhaps his view is simply that thoughts are constitutive of mental events; as we might put it, agents think of objects of thought. Basil claims that mental events take place timelessly (&Xp6vcoq),which seems to be the same idea regarding utterances as in the horrdly (480A) when he justifies John's use of A6yoq instead of Yi6q on the grounds that "human generation takes place over time, and involves passion. " Now let us consider the bearing of the philosophical tradition on (2). The point about time and the mind is attested only once in the Stoics, for the post- Chrysippean Stoic Diogenes of Babylon. The view of Richard Janko has been outlined above (footnotes 3,5) that Diogenes employs the notion of X6yoq ýv&60e, coq and npoýoptic6q in his theory of vocal sound (O(ový) in an argument which is designed to show how human rationality operates beyond the pale of animal rationality. Galen preserves an argument of Diogenes for the location of the 'leading part of the soul' in the heart which talks about significant vocal sound being sent out from the n-dnd.57 The report says that the utterance is semantically imprinted by thoughts in the mind, and thus formed or stamped issues o utward in speech, being then extended in time. The implication seems to be that the thoughts of the mind are containe d in some inner, timeless state. However, it appears that the philosophical tradition does not dwell on the passionless generation of speech from the mind. None of the reports of Stoics say this, although it is clear that the Stoics consider humanity, and human speech in particular, as rational at its basis. For the Stoics, ordinary language reflects the order of the rational cosmos, and should be brought into perfect concord with it. Stoic orthodoxy conceives the passions as movements of the soul disobedient to reason; tl-ds reflects the Chrysippean psychology of a wholly rational soul without an irrational part. 58 But there is no notion of the speech of ordinary speakers having to do with anything like the passions of the soul; language is

57Galen, Plac. Hippocr. Plat. 2.5-9-13 (de Lacy, 130.7-19) = FIDS 450: Kai 6WO; 5ý MOav6v IýTC6 v5v ýVvotwv 9v(TEGIjjIaaPF-VovTaw EV T115taVOia Kai OtOV &TETV7M)A9VOVýK7rýJMEOOat T6V X6YOV Ka! TCaPEKTEiVF-GOatT(P Xp6v(p KaTd Eer6 8tavevoijaOat Kai Eýv KaT& T6 Xýyetv ývgpyCtav. 58Atherton, The Stoics on Ainbiguity, 119-122 presents texts (particularly from Galen), translations, and discussion of these points.

137 sent out from the rational part of the SOU1.59On the other hand, Aristotle, like the Stoics, presents a semantic theory with three levels (hit. 16a3-9; 24bl-2), and on the level of the mind he talks about the 'affections' of the soul. In the De Interpretatione Aristotle says that things which are spoken (,rdt ev vq ý(Ov@ are symbols of 'things that happen to the soul' (vý)v ýv rý xpXý nccOT1ýidruov),while 'things that happen to the soul' are likenesses (6gotd)gam) of objects, presumably by way of determining reference. (Aristotle also says that the utterances, wl-dch are signs, are priiiiarily signs (orqgCtcoof things that happen to the soul, and one might assume, only secondarily signs of objects. 60) Clearly, Basil's view of language and thought borrows nothing from the Peripatetic tradition, and any Stoic echoes are rather faint.

IV

Now let us round off the discussion with some remarks on the theological background of Hom. 16. Here we find two objections to the theological use of the analogy with mind and language. These reflections on the analogy are found in the Eastern Origenist theological tradition. Closest to the Basilian homilies in spirit is a famous homily of his contemporary Meletius, who was (briefly) bishop of Antioch in the early 360s and probably had substantial theological contact with Basil.61 In one passage

-59Bycontrast, Basil believes that the soul is immaterial, Hom. 3.7 (Rudberg, 35.16ff. ), and is divided into two parts, a rational (Xoytic6v, voEp6v) part and an irrational (TCaOTjrtiC6v,akoyov) part, Hoin. 3.7 (Rudberg, 35.6-8). 60The later Peripatetic tradition understands things that happen to the soul in this text to amount to thoughts (voAgaTa, cf. Int. 16a9ff. ), as shown by Jonathan Barnes, "Meaning, Saying and Thinking, " Dialektiker wid Stoiker, hrsg. Klaus D6ring und Theodor Ebert (Stuttgart: Franz Steiner Verlag, 1993): 47-61 at 50-53. The Aristotelian account counts nouns, verbs, and sentences all as signs of affections of the soul, but at Int. 24bl-2 sentences (spoken affirmations and negations) are the signs of affections in the soul. 61This homily, found in Epiphanius, Haer. 73.29-33, was delivered by Meletius at Antioch early in 361 as his inaugural speech as bishop of Antioch. Thomas A. Kopecek, A History of Neo-Arianisin, 2 vols. (Cambridge, MA: Philadelphia Patristic Foundation, 1979): 369ff- argues that the exegesis of Prov. 8.22 in this text represents "a clearly homoiousian exegesis of the text.... Meletius' explication of Prov. 8:22 can be said to be homoiousian because of its precise identity with the position taken by the Homoiousians in their Ancyran synodical letter of AD 358: the proper understanding of the Son's origin must combine the notion of begetting and the notion of creation. " But Hanns Christof Brennecke, Shidien zitr GescIlicilte der Honider (Ribingen: Mohr, 1988): 71ff. decisively refutes the viewpoint of Kopecek which is shared by several other scholars, pointing out that at the time that he took up the see at Antioch, Meletius was theologically in harmony with the Homoian leader Acacius, had signed the formula of Constantinople (360), and opposed the Homoiousians; his famous sermon offers no

138 from the homily, Meletius evaluates a theological notion very much like the 'outer speech', being far more cautious with the speech analogy than Basil. Kai 6Tt X6yoq gurt Te Koft MycTat <6> 1)i6q, oi) [thv ýa)vh Toý 0ýga ýOtUT6V 7raTp6q 0ý8ý VOCIT(Xt. 14ý(YTTjKF- 'Yd(P KOV KC(i ýVEPYCti, KC(i 5C ýV Oti)TOý Td( 7UdVTCC K(A CCýT(ý Td( RdCVTa' 6G7rcpKcci (Toýicc6 ai)T6; 6v MýgTjga 016K TOIL) TCaTPOq VOEtTat, OUTE KiVIJUIq TOIL) YF-[tOVtKOA) KCCi ýv6pyeta, &Wc y6vvilga 6got6v TE TOD7rCCTPO'q KOEi T6V XCCPOCKTýpOc TOI) ITC(Tp6q&KptpoýV. 62

And he is logosand is named the Son, yet he is not conceived as an utterance nor spoken word of the Father. For he subsists and acts in his own right, and through him are all things and in him are all things; just as Wisdom who is the same [i. e. logos] is not conceived as a thought of the Father, nor as a movement and activity of the intellect, but rather as offspring like to the Father and the perfect imprint of the Father. Developing his position, Meletius notes the limitations that must be observed with regard to human models for conceiving divinity. (At tl-ds point in the homily it is apparent that Meletius favors 'Wisdom' ((Yoýia) over X67oq, at least for some purposes. ) xai 6)g xaiä tiö itapd8StTµa 'js 6oýiaS Ti S ltEptEKtuzT1c Twv ävOpcunivo v ývOujqµäTwv iov mätipöc, oü ghv ävunöGtatiög TE Kai, ävvnapKTOS.... i.Kavöv SE EG Ct Kai TO tiTIg 6o4iaS övoµ. a Ktä(av gvvotav täOouS ältoKX8ißat. 63

And as in the analogy of the wisdom which is comprehensive of human thoughts, the Son is the Wisdom of the Father, but is not

good evidence for deviation from a fundamentally Homoian stance: "Sie bietet die vertraute 6stliche eusebianische Theologie." As for the personal and intellectual relationship of Meletius with Basil, Kopecek (371ff.) points out the basic agreement in their shared emphasis on the incomprehensibility of God, so apparent in Basil's Cotitra EtmoiWiniz,and surmisesthat while Meletius was bishop of Sebastein 360 he discussedtheological matters with Basil who was close at hand, staying on his family's property at Annesi. Thus, Kopecek thinks that " he and Basil developed their anti-Neo-Arian polemic in conjunction with one another." And surely this latter view is not damaged too much by Brennecke's arguments for a Homoian- minded Meletius in 360-361. By contrast, Rousseau,Basil of Caesarea,239ff. has nothing to say about any theological agreement or collaboration between the two leaders; he does however have much to say about their relationship later in Basil's career (after 372). As argued convincingly by Brennecke(220), the Synod of Tyana (366)convened by Eusebiusof Cappadocian Caesareawas indeed a Meletian synod, not just a Nicene-Homoiousian gathering. Since BasiI as presbyter played such a central role in this synod, the event provides support for his alleged theological links with Meletius and his followers. Brennecke (176-178)discusses helpfully the lack of ultimate theological reconciliation between Meletius and Athanasius, which legacy Basil of Caesarea (226ff.) later strove to overcomein his program of building a united front among Eastern Nicenes of every stripe. 62Hom. Mel., ap. Epiph., Haer. 73.30.6 (Holl, 305.19-27) 631-lom.Mel., ap. Epiph., Haer. 73.31.6-7(Holl, 306.28-307.4). The text in pointed brackets is what Holl conjectures.

139 insubstantial and non-existent.... While the term 'wisdom' is sufficient to exclude all notions of passion. The objection to the analogy is that the divine %6yo; subsists and is independent to a degree that the notion of spoken utterance does not bear. There is something insubstantial about 'speech expressed'. 64 On the other hand, there is a problem with'mental speech'. This is the understanding of 'Wisdom'in the mind of God, a notion along the lines of divine thought (ývffi)giiýta). Hence it looks like Meletius has the X6yoqdistinction in the back of his mind. Significantly, Meletius is convinced that our notion of speech or wisdom does not involve passion, which n-dght suggest that Meletius and Basil present notions of language from shared theological or philosophical influenceS.65 A second objection to the mind and language analogy, going back a bit farther in the Origenist tradition, is advanced by the influential Greek theologian of the early fourth century, Eusebius of Caesarea. Eusebius' late treatise On Chztrch Theology contains an extended polen-dc against the heightened anthropornorphisms of Marcellus, who has a remarkable doctrine of the X6yoq. Marcellus conceives the X6709of God as a rather thin and insubstantial divine 'person', the 'mere' reason (ytX6q Mpg) of a simple and undivided deity. Accordingly, God does not consist of a plurality of hypostases. Eusebius defends the sonship,of the second person of the Trinity with a vigor that left its mark on posterity. By his lights, a notion of divinity largely confined to CMogand 2,6-yogis not adequate.66

64Cf. Basil's description of X6yoq as speech which dissipates in the air after being uttered, Sp. Satict. 16,38.31-34 (SC 17: 378-380): Oý,r6 oU'v ?,6yoq, 66po; cim)atq m1gavnicý, 5tdt ý(I)vljwc& 6pydvct)výKopop6vq- oý)TEnv6ga, aT6gcrcoq&Tg6q, &, E(-j)vdEvcaEvF-u(YTtK& pcp6v ýý(oooýgEvoq. 65EIsewhere in the homily, Meletius uses language that more nearly approximates Basil's standpoint, as when he says of the Son of God that he is the "indescribable interpreter of the one who cannot be described, Word and Wisdom and Power", and at Hom. Mel., ap. Epiph., Haer. 73.30.5 (Holl, 305.12-19) that he is the "perfect and stable offspring, from the one who is perfect and stable in sameness, not effluent from the Father, nor severed and divided, but without passion and completely coming forth from the one who is bereft of nothing of what he is. " 66This area of agreement among Eastern conservatives can be traced with respect to the divided 2,6yo; in the Makrostich (344), sections V-VI, ap. Ath., Syn. 26 (Opitz, 252.34-253.7); also in one of the anathemas of the Ekthesis of Sirmium (351), ap. Ath., Syti. 27 (Opitz, 255.9). Mühl, "Der X6yoq ýv&dftwq urd Trpoýoptic6;," 52-56 remarks on the lack of constructive theological use of the concept in Origen, by contrast to earlier apologists. But as he then points out (53): "Blieb diese auch in der Folgezeit ein Flauptthemader kirchlichen Disputationen über die im g6ttlichen Wesen sich darstellende Zweiheit. " Besides the opposition to versions of the concept in Athanasius and Hilary of Poitiers, only Sirmium is brought in by MCIhl from the sphere of fourth century theological controversy.

140 ý56KF-t ýnei 5ý rjo dcv0p(oTciv(p X6ycp r6v xoý Osou- 7rupaodUctv MCEPKgXX(O, Kai n6cv(p PF-X'Ti(O ElVat ýýGOgEV, eiK6vt Xp6)lievov cco &vOp(o7tiv(q Myq), co,6, rq) [t&Xkov XpAaaa0at cý napa8dygan Kat ý6vat -coi-)icap' ilg-tv ?,6you nwcýpa F-1tvatc6v voýv, 9,cepov 6vca 7rap6c ,E6v k6yov. T6V ýIýV Yap VOUV, O(Y'Ctq TEOTE Kat 07TOlOg (OV 'EIIV OA)UtC(V d(VOP61r(()V ýV ýMiPXEL, 0ý56'tq R&[OTC 9YVC0, PUCTLXEýq 5' Ola &TEOPPA'COtq ETUCO TCýtq WkOý 'TajlEiOlq i8pVg&Oq T& 7EPOMTýa ýý d(HT(Ov PoUEý)sTat* 68' ýý ccin6 Myog 0'rM TECETP6; gL)X 1ýOV 7F-7F-vvll96vO9Tcýlg ýKT69 Tr&cytvKaOiCF'ca'c()ct'YV6)Pt9O9.... Ka*Ta T& albr6c 6 8ý, g&XXov 8' ýnýKctvoc Tc6"q eiK6vo; Te Kai Tcapa8FAygaEog, II nagomytk6mgOeoi-), rýkEtoq X6yog, ou KaEccu6v irpoýoptic& d(vOpc6n(,)v A.6yov &K mAXapow Kai OijR&rcovKai 6v%tdvr(,)v (yuyKP-t[tF-voq,ota 8ý govoyevhq, coý E)P-oýui6g (6v Kai l4em6g 7Ep6st(ytvjiýv Týq ncnpmýq OE6T7lT6g PaGtWaq 0601) 6), br6KEtVa TE Kai ... TOý 7raTP69 Kai TCOV (J)V 6viog &Xýncou re Kai dcXcopAroi) ccýtq Tc&atv, Kac& T6v &ýavý Kai dc6pa,rov vo&, nap' 6 Kai ý6q oiK O)V-, anpoatEOV Xý? LpKEat. 67

Seeing that Marcellus decided to compare the logos of God to human logos, we say so much the better for him Ji. e. the theologian, not Marcellus], if he uses the image of human logos, or rather uses tl-ds as the model and says that the mind is the father of our logos, being distinct from it. For the mind, what it is and how it is qualified with respect to its essence, none of humankind has ever known, as a king counsels within, as to what sort of things are to be done, in the secret storehouses of his soul; but the logos from him as it were a father, begotten from the innermost parts, is- made manifest to all who are without.... In the same way, or rather in a manner beyond every image and model, the perfect logos of God, the absolute king, not being constituted from syllables and verbs and nouns in the manner of the logos of men which is spoken forth, but inasmuch as the Only-Begotten Son of God who lives and subsists comes forth from the patriarchal divinity and kingship God Father is beyond is ...while the and who the universe incomprehensible and unattainable to all in the manner of the unseen and invisible mind, on account of which he is called 'the one who dwells in approachable light'. This second objection to the analogy seems to be that a notion of compositionality is not a characteristic of a divine nature. There is a subtle contrast here to the Basilian polemic against the ontological subordinationism of the Son to the Father, which seems to be the point of the 'so great' clauses and the emphasis on the exact, complete revelation of all that is hidden in the being of the Father (Hoiiz. 16.3). For Basil is concerned with the straightforward subordinationism of his theological opponents (the Neo-Arians), while Eusebius is concerned with safeguarding the transcendence of God whose perfect imprint is discerned in the X6yog(, cýXF_to; Mpg).

67Eusebius, Eccl. Theol. 2.17 (Klostermann, 120.33-121.18)

141 Besides the historical relationships between theological and philosophical ideas, we have seen the expressive power along with the worries generated by analogical arguments in Patristic theological discourse. There remains the unresolved question as to why Basil assumes so readily the passionless quality of speech. Perhaps by the fourth century this was commonly held. It might be expected that evidence from the Hellenistic rationality debates would shed some light on the question. Unfortunately, on this point there is little evidence to go by. Or perhaps the problem of begetting, once it was treated as distinct by Patristic writers from the philosophical notion of generation, took on philosophical assumptions of its own: It may be an entirely Christian idea. This view of n-dnd and language must have been a natural position to advocate, given the theological preoccupation with the timeless and passionless generation of the Son from the Father. Further, the timelessness of thought is taken up and developed in Medieval Philosophy. But consideration of the afterlife of this legacy of Greek Patristics must be left to other scholars. 68

68Special thanks are in order to Martin Stone for support and criticism of this work. Paul Helm and Anthony Kenny contributed valuable discussion and criticism in response to a paper based on this chapter at the Millennia] Conference of the British Society for the History of Philosophy, "Athens and Jerusalem: Christianity and the History of Philosophy", University of Keele 6-9 April 2000.

142 Conclusion

I hope that this work makes a contribution in two main respects. First, I hope that it demonstrates the promise of the relatively underdeveloped field of Ancient Philosophy and Patristics, by forging a new link with a linguistic approach. Greek Patristics is of philosophical interest, sometimes in unexpected ways. Second, I hope that it goes some way towards opening up a conversation with contemporary philosophers and linguists. The study of the history of linguistics is gathering steam and there is a need for more of this research in Ancient Philosophy, particularly in the growing field of Post-Aristotelian Philosophy. I think that the wide range and difficulty of the grammatical, rhetorical, and philosophical texts from this period of philosophy has resulted in the present situation, in which a large body of surviving ancient texts is still poorly understood and neglected. The bias of research tendencies to date is odd in view of the largely non-theological nature of much of the most pron-dsing material in Patristic writers such as Origen. Although I have tried to show where I think there is contact with Greek philosophy (particularly the Stoics) and the grammatical tradition, one result of work has been the surprising philosophical creativity of this Christian theologian. It seemsto me that this is most clear in the proper names chapter and also in the n-dnd and language chapter, which differs from the other discussions by focusing on homiletical works instead of theological treatises. It is hard to find any secure philosophical precedent for these ideas. They seem to arise naturally in the context of trinitarian debate and exposition. The topics that presented themselves from the Patristic texts, all dealing with language, are all (except relatives) heavily discussed in contemporary philosophy, linguistics, and related fields. But for the most part they are not topics which attract substantial attention in Ancient Philosophy. One major benefit of starting with Patristics to guide my choice of subject has been a

143 sense of creativity in what I looked for in the philosophical sources, the unexpected discovery of areas to explore which had hardly been touched previously. Perhaps the best example of this process was the conjunctions chapter, which led me into study of the Stoics and their successors on the semantics of the indeclinable parts of speech. The philosophical theme that persists through these explorations in language is the attention to meaning: changes of the meaning of expressions, completion of meaning, determination of meaning, speaker's knowledge of meaning, correctness and purity of meaning. The philosophical tradition that is rooted in the Stoic project of ordering and clarifying the relationship between words and meanings for dialectical purposes is transformed in the classic theological controversies of Greek Patristics.

144 Index of Abbreviations

Abst. De Abstinentia Adv. De Adverbiis Agr. De Agricultura An. Post. = Analytica Posteriora Ap. = Apologia Ars Gramin. = Ars Granzinatica Carm. = Carmina Cat. Categoriae Cels. Contra Celsum Comm. ad Hom. Il. = Commentarii ad Homeri Iliadem Congr. Erud. Grat. = De CongressitEruditionis Gratia Conj. = De Conjunctionibus Diss. = Dissertationes E ap. Delph. = De E apud Delphos Eccl. Theol. = De EcclesiasticaTheologia Ecl. Eclogiae Ep. Epistula Ep. Georg. Laod. = Epistula Georgii Laodicaei Ep. Hom. = Epimerismi Homerici Eun. = Contra Eunomium Frag. = Fragmenta HE = Historia Ecclesiastica Haer. = Adversus Haereses Hex. = Honziliae in Hexaemeron Hist. = Historia Ecclesiastica Hom. = Homilia Hoin. in Ps. = Homiliae in Psalmos Hom. Mel. = Homilia Melitii Hyp. = Pyrrhoniae Hypotyposes In An. = In Libros Aristotelis de Anfina In An. Pr. = In Aristotelis Analytica Priora In Cat. = In Aristotelis Categorias In Eph. = Fragmenta ex Commentariis in Epistulam ad Ephesios In Gen. = In Genesim In Ies. = Enarratio, in Prophetam Isaiam In Int. = In Aristotelis de Interpretatione

145 In Mt. = Commentarium in Evangelium Matthaei In Ronz.= Connnentarii in Romanos Inst. = Institutio Oratoria Inst. Gramm. = Institutio Grammatica Int. = De Interpretatione Ling. Lat. = De Lingua Latina Math. = AdversusMathematicos Met. = Metaphysica Nat. Anim. = De Natura Animalium Or. = Oratio Orat. = De Oratione Part. Anim. = De Partibus Anfinalium Phil. = Philebus Phil. = Philokalia Plac. Hippocr. Plat. = De Placitis Hippocratis et Platonis Plat. Quaest. = Platonicae Quaestiones Princ. = De Principiis Prog. Progymnasmata Prol. Prolegomena Prol. Sch. in Theodos. = Prolegomena et Scholia in Theodosii Alexandrini Canones Isagogicosde Flexione Nominum Pron. = De Pronominibus Sch. in Dionys. Thr. = Scholia in Dionysium Thracein Sch. Lond. in Dionys. Thr. Scholia Londiniana in Dionysium Thracem Sch. Marc. in Dionys. Thr. ScholiaMarciana in Dionysium Thracem Sch. Vat. in Dionys. Thr. = Scholia Vaticana in Dionysium Thracenz Sens.= De Sensit et Sensibilibus Soll. Anfin. = De Sollertia Animalium soph.= Sophist Sp. Sanct.= De Spiritu Sancto Syn. = De Synodis Syn. Ep. Anc. = EpistulaSynodica Synt. = De Constructione Theaet.= Theaetetus Tim. = Timaeus Top. = Topica Vit. Soph. = Vitae Sophistartun

146 Index Locorum

Aelian, Nat. Anim. 6.59 (Hercher, 166.10-28), 125 n. 36 Ammonius, In Cat. 45.21-46.10 (Busse), 40 n. 16; 46.10-19,40 n. 16; In Int. 242.20 (Busse), 133 n. 50 Apollonius Dyscolus, Adv. 180.13-183.4 (Schneider), 75; 181.32-182.10,76 n. 48; Conj. 213.3-5 (Schneider), 90 n. 78; 213.8-14,63 n. 11; 214.1-3,63 n. 12; 214.4-20,60 n. 4, 61 n. 5; 214.10,62 n. 7; 214.20-25,69 n. 34; 220.7ff., 64 n. 15; 220.10ff., 80 n. 62; 247.30ff., 44 n. 21; 251.27ff., 63 n. 11; 252.21-3,80 n. 62; 253.12,80 n. 62; flp-pi 6vogk(ov (Fragnienta) 49.40-50.10,104 n. 30; Fragmenta 137.35-138.10 (Schneider), 91 n. 82; Pron. 5.13ff. (Schneider), 64 n. 14; 27.9-10,42 n. 20; Synt. 9.3-5 (Uhlig), 69 n. 32; 13.8-14.4,67 n. 27; 27.11-13,68 n. 27; 31.4-5,44; 41.1,44; 66.11-12,40 n. 18; 142.1ff., 40 n. 18; 142.1-143.3,42 n. 19,44 n. 22; 143.9-144.4,44 n. 23; 146.7,44; 155.3-5,40 n. 18,42 n. 19; 159.7,40 n. 18; 315.20,44; 352.5-6,40 n. 18; 367.2-5,102 n. 26; 436.10-437.4,61 n. 6; 457.12-458.5,73 n. 43,92; 481.3-8,74 n. 46; 482.3ff., 78 n. 56; 487.3-488.4,7 n. 1; 487.9-488.4,74 n. 45; 487.11-488.4,60 n. 4 Aristotle, Ali. Post. 76b24,113; Cat. 1a16,101 n. 22; 3a21-28,40 n. 16; 4b34,113; 6a36- 37,101 n. 22; 6a36-bl, 101 n. 21; 6b15-27,101 n. 21; 8a17-24,101 n. 21; Int. 16a3-9, 138; 16a9ff., 138 n. 60; 24bl-2,138,138 n. 60; Met. 7,49 n. 30; 7.3,49 n. 30; 7.4-12, 49 n. 30; Part. Anim. 2.10,656alO-13,124 n. 34; 4.10,686a27-28), 124 n. 34; Sells. 447a (Hett), 51 n. 38; Top. 128bl6ff., 101 n. 21 Arius Didymus ap. Stobaeum 2.57.13ff. (Wachsmuth). 13 n. 17 Athanasius, Syn. 26 (Opitz, 252.34-253.7), 140 n. 66; 27 (Opitz, 255.9), 140 n. 66 Basil of Caesarea, Ep. 52.1-3 (Courtonne, 1.134-136), 81 n. 65; Ep. 125.1 (Courtonne, 2.30-32), 81 n. 65; Ep. 223,94 n. 1; Elm. 1.5,67-72 (SC 299: 176), 95,95 n. 4; 1.12, 30ff. (SC 299: 214ff. ), 48 n. 27; 1.13,13-14 (SC 299: 216), 48 n. 27; 1.13,36-37 (SC 299: 218), 48 n. 29; 2.4,1-26 (SC 305: 18-20), 96 n. 8; 2.4,1-42 (SC 305: 18-22), 46 n. 26; 2.9,3-6 (SC 305: 36), 97 n. 9; 2.9,11ff. (SC 305: 36ff. ), 96,96 n. 7,100,100 n. 18, 101,103,104,111; 2.9,11-23 (SC 305: 36), 97 n. 10; 2.9,24-27 (SC 305: 38), 96 n. 8; 2.10,21 (SC 305: 38), 98 n. 12; 2.10,21-23 (SC 305: 38), 98 n. 13; 2.10,23-27 (SC 305: 38-40), 98 n. 13; 2.16,33-45 (SC 305: 64), 136 n. 56; 2.22ff., 98; 2.22,33-51 (SC 305: 90-92), 105,106 n. 37,107,109; 2.23,14ff. (SC 305: 94ff. ), 109 n. 49; 2.23,20-23 (SC 305: 94), 107 n. 38; 2.23,41-43 (SC 305: 94ff. ), 110 n. 50; 2.24,1-4; 16-23 (SC 305: 98- 100), 109 n. 47; 2.26-29 (SC 305: 108-124), 95 n. 3,109,110 n. 51; 2.27,26-33 (SC 305: 114), 110 n. 52; Hex. 9,123,124; 9.2 (de Mendieta and Rudberg, 148.23-149.6), 124 n. 34; 9.3 (de Mendieta and Rudberg, 149.11-149.13), 124 n. 35; 9.3 (de Mendieta and Rudberg, 150.20-151.10), 125 n. 37; 9.4 (de Mendieta and Rudberg,

147 153.4-6), 124 n. 33; 9.4 (de Mendieta and Rudberg, 153.21-154.3), 125 n. 36; Ho"j. 3 (Rudberg), 123,123 n. 32,126,126 n. 38,128; 3.1 (Rudberg, 23.1-11), 126 n. 39; 3.1 (Rudberg, 24.17-25.7), 126 n. 39; Hom. 3.2 (Rtkdberg, 25-12-16), 126 n. 38;, 3.7 (Rudberg, 35.6-8), 138 n. 59; 3.7 (Rudberg, 35-16ff. ), 138 n. 59; Hom. 12 (PG 31, 392B), 129,133 50; Hom. 12 (PG 31,400C), 135 54; Hom 15 (PG 31), 129 n. n. - n. 44; Hom. 16 (PG 31), 123 n. 32,129,129 n. 44,130; 16.3 (PG 31,475C-479A), 129 n. 44; 16.3 (PG 31,476C-480A), 131 n. 49,133 n. 50,135,136 n. 55,138,141; 16.3 (PG 31, 480A), 137; Hom. 24 (PG 31), 129 n. 44; Hom. 27 (PG 31), 129 n. 44; Honz. In Ps. 14 (15) (PG 29,249D-280C), 133 n. 50; Hom. In Ps. 14 (15) (PG 29,256A), 133 n. 50; Hom. in Ps. 14 (15) (PG 29,253C-256B), 135 n. 54; In Ies. 10.242 (PG 30,544C4-6), 17 n. 31; Sp. Sanct. 3,5.6-8 (SC 17: 264), 80 n. 60; 4,6.23-26 (SC 17: 270), 80 n. 60; 5, 12.1-4 (SC 17: 282), 89 n. 77; 5,12.1-26 (SC 17: 282-284), 86 n. 73; 5,12.1-30 (SC 17: 282-284), 81 n. 64; 16,38.31-34 (SC 17: 378-380), 140 n. 64; 25,58.5-14 (SC 17: 456), 81 n. 64; 25,59.1-15 (SC 17: 458), 86 n. 73; 25,59.1-55 (SC 17: 458-462), 82 n. 68; 25, 60.1-7 (SC 17: 462), 82 n. 68; 26,61.1-64.23 (SC 17: 466-476), 91 n. 83; 26,63.8-19 (SC 17: 472-474), 82 n. 68; 27,68.28-32 (SC 17: 490), 82 n. 68; 29,71.1-75.52 (SC 17: 500-518), 81 n. 65 Charisius, Inst. Gramm. 229.9-230.2 (Keil) = 296.21-297.28 (Barwick), 90 n. 79; 236.16 (Keil) = 307.17-18 (Barwick), 90 n. 81; 238.15-17 (Keil) = 310.26-311,2 (Barwick) Dexippus, In Cat. 30.20-27 (Busse), 49 n. 32,49 n. 33; 49.4-6,40 n. 16 Diogenes Laertius 7.49,116 n. 10; 7.51,116 n. 9; 7.52-53,111 n. 53; 7.55,115,117 n. 13, 121; 7.55ff., 114; 7.55-56,115; 7.57,114 n. 5,115 n. 6; 7.58,35,49; 7.62,35 n. 2,65 n. 21,114; 7.69-72,62 n. 9; 7.71-74,66 n. 25; 7.85-116,13 n. 17 Dionysius Thrax, Ars Granim. 6.2ff. (Uhlig), 17 n. 32..24.1-6,38 n. 10; 33.6-34.2,38 n. 12; 34.1-35.4,102 n. 24,103 n. 27,103 n. 28,111 n. 53; 44.6-7,103 n. 27,103 n. 28; 88.3-89.2,80 n. 62 Epictetus, Diss. 1.16.8-10 (Schenkl), 128 n. 42; 2.8.3-8,128 n. 42 Ep. Georg. Laod., ap. Epiphanium, Haer. 73.18.8 (Holl, 291.20-25), 101 n. 10; 73.19.3-5 (Holl, 292.4-16), 100 n. 16 Ep. Hom. K 146 Kai (Dyck, 446.3.9-40),91 n. 82; it 109 (Dyck, 615.39-48), 92 n. 87; n 118 nF-pi (Dyck, 619.66-7), 91 n. 85; n 142 nap 6c(Dyck, 629.81-630.10),91 n. 85 Eunapius, Vit. Soph. 4.1.4 (Giangrande), 28 n. 75; 9-2-1ff., 11 n. 12 Eunon-dus, Ap. 12 (Vaggione, 48.7-10), 97 n. 9 Eusebius, Eccl. Theol. 2.11ff. (Klostermann), 133 n. 50; 2.13 (Klostermann, 114.10-21), 133 n. 50; 2.17 (Klostermann, 120.33-121.18), 141 n. 67; Hist. 5.28-13-19 (Schwartz, 504.11-506.19), 21 n. 48; 6.2.7-8 (Schwartz, 520.24-8), 19 n. 40; 6.2.15-3.1 (Schwartz, 524.1-6), 21 n. 45; 6.3.9 (Schwartz, 526.19-23), 21 n. 47 Eustathius, Conim. ad Hom. Il. 6.511 (van der Valk), 74 n. 45 Galen, Plac. Hippocr. Plat. 2.5.9-13,115 n. 6,137 n. 57; 2.5.25ff., 78 n. 55; 3.7.42-43,117 n. 12 Georgius Choeroboscus, Prol. Sch. in Theodos.106.3-10 (Hilgard), 39 n. 15; 109.24ff., 40 n. 18 Gregory of Nazianzus, Carm. 2.1.12,267-269 (PG 37,1185), 130 n. 48; 2.4 (PG 67, 1510), 14 n. 21; Ep. 6 (Gallay), 26; Ep. 115 (SC 302), 24,25,26,27 28 n. 72; Ep. 115, 4-7 (SC 302: 170); 25 n. 61; Or. 28.13 (Barbel), 130 n. 48; Or. 43.12,14-17 (SC 384: 140), 12 n. 15; Or. 43.23,6-24 (SC 384: 172-4), 12 n. 14 Gregory of Nyssa, Eim. 1.539-542 (Jaeger, 182.17-183-15),130 n. 48 Himerius, Or. 48.20-25 (Colonna), 9 n. 8

148 Julian, Ep. 42 (Hertlein, -545.4-11), 30 n. 78; Ep. 90 (Bidez, 174.26ff. ) = Julian Ep. ad Photintini ap. Facundum, Pro DefensioneTriton Capittilortan 4.2 (PL 67,621A-C), 15 n. 24; Frag. Ep. 385.15-386.24 (Hertlein), 10 n. 9 Libanius, Ep. 10.4-5 (Fatouros and Krischer) = Ep. 255 (Fbrster), 31 n. 81; Ep. 12 (Fatouros and Krischer) = Ep. 727 (Fbrster), 30 n. 79; Ep. 28 (Fatouros and Krischer) = Ep. 406 (Rrster), 31 n. 80; Or. 1.9 (Martin), 30 n. 79 Marcus Aurelius, Ad Se Ipsiwi 4.23 (Dalfen), 80 n. 61 Hom. Mel. ap. Epiphanium, Haer. 73.29-33,138 n. 61; 73.30.5 (Holl, 305.12-19), 140 n. 65; 73.30.6 (Holl, 305.19-27), 139 n. 62; 73.31.6-7 (Holl, 306.28-307.4), 139 n. 63 Origen, Cels. 3.58,21 n. 44; 4.81,128 n. 43; Ep. ad Gregorenz= Phil. 13,1-18 (SC 148: 186-188), 21 n. 44; Honz. 3 in Ps. 36 (PG 12,1341B-C), 20 n. 42; In Eph. 1.1 (Gregg, 234.1-235.11), 92 n. 90; In Gen. 3 (SC 302: 406-412) = Phil. 14,22 n. 52; In Mt. 10.24 (Klostermann, 33.20-21), 80 n. 62; In Roni. 170.10-172.11 (Scherer), 92 n. 90; Orat. 6.1 (Koetschau, 311.8-10), 128 n. 42; Phil. Prol. 3-8 (SC 302: 162), 28 n. 70; 9ff. (SC 302: 162), 28 n. 71; 21ff. (SC 302: 162), 28 n. 72; Princ. 3.1.2-3 (G6rgemanns and Karpp), 128 Philo, Agr. 136-141 (Wendland, 122.6-123.11), 20 n. 43; Congr. Enid. Grat. 11; 79 (Wendland, 74.9-13; 87.19-88.4), 21 n. 44 Philoponus, In An. Pr. 167.17ff. (Wallies), 49 n. 32; In Cat. 66.13-25 (Busse), 40 n. 16; 67.1-3,40 n. 16; ap. Ammonium In Cat. 64.22-66.12 (Busse), 40 n. 16 Philostorgius, Hist. 4.12 (Bidez and Winkelmann, 64.3-7), 99 n. 14 Plato, Phil. 38E, 113; Soph. 263E3-9,113; 264A1-2,113; Theaet. 189E6-190A, 113; 206DW., 113; Tim. 49A6-50C, 35 n. 2; 90A-B, 124 n. 34 Plutarch, E ap. Delph. 386F-387A, 122 n. 26; Plat. Qttaest. 10.4,1011a-b (Cherniss), 62 n. 9; Soll. Aninz. 961C-D (Helmbold), 123 n. 31; 968F-969A, 123 n. 31; 969A-B, 123 n. 31; 969Cff., 123 n. 31 Porphyry, Abst. 3.2 (Nauck, 187.20-188.20), 122 n. 28; 3.3 (Nauck, 188.17-20), 118 n. 14; 3.4 (Nauck, 191.3-13), 121 n. 24; 3.6 (Nauck, 194.3-9), 123 n. 30; 3.7,123 n. 30; In Cat. 94.17-28 (Busse), 40 n. 16; 95.17-20140 n. 16; 95.22-96.1,40 n. 16; 129.8-10, 49 n. 32,50 n. 35 Priscian, Inst. Granini. 14.18 (Keil), 61 n. 6; 60.19-27,103 n. 26; 60.23ff., 103 n. 26 Ptolemy, nepi icpvu'qpioi)Kai hyEgOVtKoi)6.1-8 (Lammert), 134 n. 51 Quintilian, Inst. 1.10.1,12 n. 15 Sch. in Dionys. Thr. 12.3ff. (Hilgard), 15 n. 26; 12.13ff., 17 n. 32; 93.7-9,92 n. 88; 135.6ff., 15 n. 26; 170.18-20,16 n. 28; 193-15-37,102 n. 23; 235.9-12,104 n. 30,109 n. 48; 235.9ff., 103 n. 26; 356.7-16,69 n. 31; 387.8ff., 103 n. 26; 387.8-12,105 n. 36; 387.13-15,103 n. 29; 387.34ff., 105 n. 36; 388.5ff., 103 n. 26; 388.20-26,103 n. 29; 518.33ff., 64 n. 14; 519.26,62 n. 8; 519.26-27,62 n. 9; 519.26-520.15,61 n. 6; 520.13- 15,62 n. 9; 524.8-12,40 n. 17; 527.22-23,44 n. 22; 553.25-30,102 n. 23; 553.27ff., 103 n. 26; Sch. Lond. in Dionys. Thr. 546.5-14,40 n. 18; 547.18ff., 40 n. 18; 553.31-35, 103 n. 29; Sch. Marc. in Dionys. Thr. 357.22-26,39 n. 13; 388.20-26,104 n. 31; Sch. Vat. in Dionys. Thr. 214.33-215.3,39 n. 13; 269.28-9,92 n. 88; 270.5-7,78 n. 57; 271.11-18,91 n. 86 Seneca, Ep. 65.2-4,80 n. 61; Ep. 88.42,77; Ep. 104.22,77 n. 51 Sextus Empiricus, Hyp. 1.59 (Mutschmann), 118; 1.62-5,118-119; 1.64,119,119 n. 17; 1.65,119 n. 16; 1.65ff., 118; 1.66ff., 119; 1.69,120; 1.73,119 n. 18; 1.73ff., 123; 1.77, 119 n. 18; Math. 1.79 (Mutschmann), 17 n. 33; 7.176-184,52 n. 41; 7.247-252,52 n. 41; 8.11-12,35 n. 2,116 n. 8; 8.70,116 n. 8; 8.93-129,77 n. 52; 8.245,121 n. 23; 8.270- 271,122 n. 26; 8.275,114 n. 3,114 n. 4; 8.275-276,121 n. 22; 121 n. 23; 8.285-286,

149 122 n. 26; 8.287,122 n. 26; 8.288,122 n. 26; 10.10,80 n. 61; 11.8-11,65 n. 20; 11.24, 36,36 n. 3 Simplicius, In An. 217.36-218.2 (Hayduck), 51 n. 37; In Cat. 2.5-9 (Kalbfleisch), 50 n. 35,51 n. 39; 36.8-12,83 n. 70; 48.1-49.9,40 n. 16; 55.2-5,49 n. 32; 99.3-9,40 n. 16; 212.12-213.1,38 n. 11; 229.16-18,49 n. 32; 334.2-3,11 n. 10 Socrates, HE 4.26 (Hansen), 26 n. 67 Sozomen, HE 6.17.2-3 (Bidez) 26 n. 67 Stobaeus, Ecl. 1.177.21-179.17(Wachsmuth), 36 n. 5; 2.97.15-2.98.6,36 n. 4 Strabo 14.1.48 (Meineke, 908.8-11), 28 n. 75 Syn. Ep. Anc., ap. Epiphanium, Haer. 73.3.3 (Holl, 271.15-18), 107 n. 40,109 n. 46; 73.3.5 (Holl, 271.22-24), 107 n. 41; 73.4.1,109 n. 46; 73.4.1-2 (Holl, 272.19-273.3), 107 n. 41; 73.4.5 (Holl, 273.15ff.), 108 n. 42; 73.5.5 (Holl, 275.lff. ), 108 n. 43; 73.9.6, 109 n. 46; 73.12.1 (Holl, 284.14-16), 108 n. 45 Theon, Prog. 59.1ff. (Spengel), 12 n. 15; 80.30-81.13,31 n. 86 Varro, Ling. Lat. 9.1 (Goetz and Schoell), 64 n. 19; ap. Diomedem, Ars Gramm. 426.21ff. (Keil), 15 n. 26

150 Bibliography

Philip R. Amidon, The Panarion of St. Epiphanfits, Bishop of Salands (Oxford: Oxford University Press, 1990). Julia Annas, Hellenistic Philosophy of Mind (Berkeley and Los Angeles: University of California Press, 1992). Julia Annas and Jonathan Barnes, Sexhis Enipiricits: Oittlines of Pyrrhollisill (Cambridge: Cambridge University Press, 1994). Catherine Atherton, "Hand Over Fist: The Failure of Stoic Rhetoric, " Classical Qttarterly 38 (1988): 392-427. Catherine Atherton, The Stoics oil Anibigitity (Cambridge: Cambridge University Press, 1993). Catherine Atherton, "What Every Grammarian Knows? " Classical Q11arterly 46.1 (1996): 239-260. Wolfram Ax, Latit, Stimme iind Sprache(G6ttingen: Vandenhoeck & Ruprecht, 1986). Jonathan Barnes, "Meaning, Saying and Thinking, " Dialektiker und Stoiker, hrsg. Klaus D6ring und Theodor Ebert (Stuttgart: Franz Steiner Verlag, 1993): 47-61. Jonathan Barnes, Logic in the Iniperial Stoa (Leiden: E. J. Brill, 1997). T. D. Barnes, "Himerius and the Fourth Century, " Classical Philology 82 (1987): 206- 225. Karl Barwick, Renindits Palaenion iind die rdmische Ars Granimatica, Philologus Supplementband 15.2 (Leipzig: Dieterich'sche Verlagsbuchhandlung, 1922). David Blank, Ancient Philosophy and Grainmar. The Syntax of Apollonizis Dyscollis (Chico, CA: 1982). David L. Blank, "Analogy, Anomaly and Apollonius Dyscolus, " Langilage. Coinpanions to Ancient Thotight 3, ed. Stephen Everson (Cambridge: Cambridge University Press, 1994): 149-165. Hanns Christof Brennecke, Shidien zitr Geschichteder Honider (Tilbingen: J. C. B. Mohr (Paul Siebeck), 1988). Hanns Christof Brennecke, "Erwdgungen zu den Ariffingen des Neuniz5nismus, " 0ectinzenica et Patristica. Festschriftffir Wilhelln Schneenielcherzitin 75. Gebitrtstag, hrsg. D. Papandreou, W. A. Bienert, K. Sch5ferdiek (Stuttgart: Kohlhammer, 1989): 241-257. Jacques Brunschwig, "Remarks on the Stoic Theory of the Proper Noun, " Papers in Hellenistic Philosophy (Cambridge: Cambridge University Press, 1994): 39-56. Myles Burnyeat, "The Origins of Non-Deductive Inference, " Scienceand Spectilation. Shidies in Hellenistic Theory and Practice, eds. Jonathan Barnes, Jacques

151 Brunschwig, Myles Burnyeat, and Malcolm Schofield (Cambridge: Cambridge University Press, 1982): 193-238. Robyn Carston, "Implicature, Explicature, and Truth-Theoretic Semantics, " Mental Representations,ed. Ruth M. Kempson (Cambridge: Cambridge University Press, 1988): 155-181. Victor Caston, "Something and Nothing: The Stoics on Concepts and Universals, " Oxford Shidies in Ancient Philosophy 17 (1999): 145-213. L. Jonathan Cohen, "Some Remarks on Grice's Views about the Logical Particles of Natural Language, " Pragmatics of Natitral Langitages, ed. Yehoshua Bar-Hillel (Dordrecht: D. Reidel, 1971): 50-68. jean Dani6lou, "Eunome I'arien et 1'ex6gýsen6o-platonicienne du Cratyle, " Revile des 1ýtiidesGreqiies 69 (1956): 412-432. Alice Davison, "Syntactic and Semantic Indeterminacy Resolved: A Mostly Pragmatic Analysis for the Hindi Conjunctive Participles, " Radical Pragmatics, ed. Peter Cole (New York and London: Academic Press, 1981): 101-128. John Dillon, "Self-Definition in Later Platonism, " Jewish and Christian Self-Definition, Vol. 3, eds. B. F. Meyer and E. P. Sanders (London: SCM Press, 1982): 60-75. Heinrich Dbrrie, "Pr5positionen und Metaphysik. Wechselwirkung zweier Prinzipienreihen, " Platonica Minora (Miinchen: Wilhelm Fink Verlag, 1976): 124- 136. V. H. Drecoll, Die Entwicklting der Trinitiitslehre des Basilhis von Ciisarea (G6ttingen: Vandenhoeck & Ruprecht, 1996). Andrew R. Dyck, ed., Epimerisini Homerici. Lexicon AIMS2AEIN, Pars Altera,, SGLG 5.2 (Berlin and New York: Walter de Gruyter, 1995). Theodor Ebert, "Dialecticians and Stoics on the Classification of Propositions, " Dialektiker tind Stoiker, hrsg. Klaus D6ring und Theodor Ebert (Stuttgart: Franz Steiner Verlag, 1993): 111-127. L. Edelstein and I. G. Kidd, eds., Posidonizis.Fragments and Commentary, 3 Vols., 2nd ed. (Cambridge: Cambridge University Press, 1989). Paul J. Fedwick, "A Chronology of the Life and Works of Basil of Caesarea," Basil of Caesarea: Christian, Hiananist, Ascetic, ed.. Paul J. Fedwick (Toronto: Pontifical Institue of Mediaeval Studies, 1981): 3-19. Jerry A. Fodor, The Langitage of Thonght (London: The Harvester Press, 1976). Michael Frede, Die stoischeLogik (Gbttingen: Vandenhoeck & Ruprecht, 1974). Michael Frede, "The Origins of Traditional Granunar, " Historical and Philosophical Dimensions of Logic, Methodology, and Philosophy of Science,eds. R. Butts and J. K. K. Hintikka (Dordrecht: D. Reidel, 1977): 51-79. Michael Frede, "Principles of Stoic Grammar, " The Stoics, ed. John Rist (Berkeley and Los Angeles: University of California Press, 1978): 27-75. Michael Frede, "The Stoic Notion of a Grammatical Case," BICS 39 (1994): 13-24. Michael Frede, "The Stoic Notion of a Lekton, " Langitage, ed. Stephen Everson (Cambridge: Cambridge University Press, 1994): 109-128. Gottlob Frege, "The Thought, " Collected Papers oil Mathematics, Logic, and Philosophy, ed. B. McGuinness (Oxford: Basil Blackwell, 1984): 351-372. H. Fuchs, "Enkyklios Paideia, " RAC 5 (1962): 365-398. Manfred Fuhrmann, Das systentatische Lelirbitch (G6ttingen: Vandenhoeck & Ruprecht, 1960). Richard Gaskin, "The Stoics on Cases, Predicates and the Unity of the Proposition, " Bidletin of the Instititte of ClassicalShidies (Supplement) 68 (1997): 91-108.

152 Mario Girardi, "'SemplicitA' e ortodossia nel dibattito antiariano di Basilio di Cesarea: la raffigurazione dell'eretico, " Vetera Christianortan 15 (1978): 51-74. Lila R. Gleitman, "Coordinating Conjunctions in English, " Modern Shidies in English. Readings in Transforniational Grannnar, eds. David A. Reibel and Sanford A. Schane (Englewood Cliffs, NJ: Prentice-Hall, 1969): 80-112. Jean Gribomont, "Eustathe le Philosophe et les voyages du jeune Basile de C6sar&, " RHE 54 (1959): 115-124. Jean Gribomont, "L'Orig6nisme de Saint Basile, " L'Honinze Devant Dim. M61anges Offerts ait Pýre Henri de Litbac, Thdologie 56 Toine 1 (Aubier: 1963): 281-294. Paul Grice, Shidies in the Way of Words (Cambridge, MA: Harvard University Press, 1989). Marguerite Harl, Origbie. Philocalie 1-20. Sitr les ýcriffires, Sources Chr6tiennes 302 (Paris: Les tditions du Cerf, 1983). Harold Cherniss and William C. Helmbold, eds., Phitarch's Moralia 12, Loeb Classical Library (Cambridge, MA and London: Harvard University Press, 1957). Karl Holl, Aniphilochiiis von Ikonizini in seinenzVerhiiltnis zzi den grossen Kappadozieren dargestellt (Tiibingen-Leipzig: 1904). Fred W. Householder, The Syntax of Apollonhis Dyscolits, Amsterdam Studies in the Theory and History of Linguistic Science (Amsterdam: John Benjamins B. V., 1981). Pamela Huby and Gordon Neal, eds., The Criterion of Tritth (Liverpool: Liverpool University Press, 1989). Karlheinz Hillser, Die Fragniente zitr Dialektik der Stoiker (Stuttgart-Bad Cannstatt: Frommann-Holzboog, 1988). Richard Janko, "Crates of Mallos, Dionysius Thrax and the Tradition of Stoic Grammatical Theory, " The Passionate Intellect. Essaiy s on the Transfortnation of Classical Traditions Presented to Prof. I. G. Kidd, Rutgers University Studies in Classical Humanities, ed. Lewis Ayres (New Brunswick and London: Transaction Publishers, 1995): 213-233. tric Junod, "Remarques sur la composition de la 'Philocalie' d'Orig6ne par Basile de C6sar6e et Gr6goire de Nazianze, " Revite d'Histoire Philosophiqite et Religeitses52 (1972): 149-156. Eric, Junod, "Basile de C6sarde et Gr6goire de Nazianze sont-ils les compilateurs de la Philocalie d'Orig6ne? R6examen de la lettre 115 de Gr6goire, " Mimorial Doni jean Gribomont, Studia Ephemeridis 'Augustinianum' 27 (Rome: Instituturn Patristicum'Augustinianum': 1988): 349-360. Charles Kahn, "Stoic Logic and Stoic Logos, " AGPh 51 (1969): 158-172. David Kaplan, "Afterthoughts, " TheinesFroin Kaplan, eds. Joseph Almog, John Perry, Howard Wettstein (Oxford: Oxford University Press, 1989): 565-614. Robert A. Kaster, Gitardians of Langitage: The Gratninarian and Society in Late Antiql1ity (Berkeley and Los Angeles: University of California Press, 1988). Theo Kobusch, "Name und Sein. Zu den sprachphilosophischen Grundlagen in der Schrift Contra Eunon-dum des Gregor von Nyssa, " El 'Contra Eitnoinhinz I' en la Proditccion Literaria de Gregorio de Nisa, eds. Lucas F. Mateo-Seco and Juan L. Bastero (Pamplona: University of Navarra, 1988): 247-268. Thomas Kopecek, A History of Neo-Arianisin (Cambridge, MA: The Philadelphia Patristics Foundation, 1979).

153 Saul A. Kripke, "Naming and Necessity, " Semanticsof Natitral Langliage, eds. Donald Davidson and Gilbert Harman (Dordrecht: D. Reidel Publishing Company, 1972): 253-355. George Lakoff and Stanley Peters, "Phrasal Conjunction and Symmetric Predicates, " Modern Shidies in English. Readings in Transformational Grammar, eds. David A. Reibel and Sanford A. Schane (Englewood Cliffs, NJ: Prentice-Hall, 1969): 113- 142. Giuseppe Lazzati, "Basilio di Cesarea insegno'retorica? " Shidi e materiali di storia delle religioni 38 (1967): 284-292. A. D. Leeman, "Posidonius the Dialectician in Seneca's Letters, " Mnemosyne 7 (1954): 233-240. Eric Lewis, "The Stoics on Identity and Individuation, " Phronesis 40.1 (1995): 89-108. A. A. Long and D. N. Sedley, The Hellenistic Philosophers (Cambridge: Cambridge University Press, 1987). Winrich A. L6hr, Die Entsteining der homdischeniind hom5tisianischen Kirchenparteien. Shidien mir Synodalgeschichte des 4. Jahrhiinderts (Witterschlick/ Bonn: Verlag Wehle, 1986). P. Maran, Vita S. Basilii (PG 29, v-clxxvii). Stephen Menn, "The Stoic Theory of Categories, " Oxford Stittlies;in Ancient Philosophy (Oxford: Oxford University Press, 1999): 215-247. Mario Mignucci, "The Stoic Notion of Relatives, " Matter and Metaphysics. Fourth Symposium Hellenisticum, eds. Jonathan Barnes and Mario Mignucci (Naples: Bibliopolis, 1988): 129-222. Max Miffil, "Der X6yog ýv&60euoq und npoýopuc6q von der 51teren Stoa bis zur Synode von Sirn-durn 351," Archivftir Begriffsgeschichte7 (1962): 7-56. Bernhard Neuschäfer, Origenes als Philologe, Schweizerische Beiträge zur Alterturnswissenschaft (Basel: Friedrich Reinhardt Verlag, 1987). A. F. Norman, "The Library of Libanius, " Rheinisches Mtisettm ffir Philologie 107 (1964): 158-175. Frederick W. Norris, Faith Gives Ftillness to Reasoning. The Five Theological Orations of Gregory of Nazianzits (Leiden: E. J. Brill, 1991). Catherine Osborne, "Literal or Metaphorical? Some Issues of Language in the Arian Controversy, " Christian Faith and Greek Philosophy in Late Antiqllity, eds. Lionel R. Wickham and Caroline P. Barnmel (Leiden: E. J. Brill, 1993): 148-170. Robert J. Penella, Greek Philosophersand Sophists in the Foitrth Centitry A. D. Stildies in Eitnaphis of Sardis, ARCA 28 (Leeds: Francis Cairns, 1990): 85-94. Rudolf Pfeiffer, History of Classical Scholarship(Oxford: Clarendon Press, 1968). Max Pohlenz, "Die Begriindung der abendldndischen Sprachlehre durch die Stoa," Nachrichtenvon der Gesellschaftder Wissenschaftenzit Gdttingen,Phil. Hist. KI., N. F., Bd. 3, N. 6 (Gbttingen: Vandenhoeck & Ruprecht, 1939):151-198. Roland Posner, "Semantics and Pragmatics of Sentence Connectives in Natural Language, " SpeechAct Theory and Pragmatics, eds. J. R. Searle, F. Kiefer, and M. Bierwisch (Dordrecht: D. Reidel, 1980): 169-203. Hilary Putnam, "The Analytic and the Synthetic, " Mind Langitage and Reality, Philosophical Papers, vol. 2 (Cambridge: Cambridge University Press, 1975): 33- 69. W. V. 0. Quine, Word and Object (Cambridge, MA: The MIT Press, 1960).

154 John Rist, "The Importance of Stoic Logic in the Contra Celsitm," 64-78 in Neoplatonisin and Early Christian Thought. Essays in Honoitr of A. H. Armstrong, eds. H. J. Blumenthal and R. A. Markus (London: Variorum Publications, 1981). John Rist, "Beyond Stoic and Platonist: a Sample of Origen's Treatment of Philosophy (Contra Celsitm: 4.62-70), " 228-238 in Platonismits iind Christentitm. Festschrift ffir Heinrich Ddrrie, hsgb. von H. -D. Blume und F. Mann (Miinster /Westfalen: Aschendorffssche Verlagsbuchhandlung, 1983). John Rist, "Basil's 'Neoplatonism': Its Background and Nature, " Basil of Caesarea: Christian, Humanist, Ascetic Vol. 1, ed. Paul J. Fedwick (Toronto: Pontifical Institute of Mediaeval Studies, 1981): 137-220. Louis W. Roberts, "Origen and Stoic Logic, " Transactions of the American Philological Society 101 (1970): 433-444. David G. Robertson, "Stoic and Aristotelian Notions of Substance in Basil of Caesarea," Vigiliae Christianae 52 (1998): 393-417. Phillip Rousseau, Basil of Caesarea(Berkeley: University of California Press, 1994). Bede Rundle, Grammar in Philosophy (Oxford: Clarendon Press, 1979). H. D. Saffrey and L. G. Westerink, Proclus. Thiologie Platonicienne (Paris: Les Belles Lettres, 1968). Mark Sainsbury, Logical Forms (Oxford: Blackwell, 1991). Christoph Schdublin, Untersuchtingen zu Methode und Herkunft der antiochenischen Exegese(K61n and Bonn: Peter Hanstein Verlag, 1974). Fritz Schernmel, "Der Sophist Libanios als Schiller und Lehrer, " Nette jahrbiicherflir dasklassische Altertum, Geschichteiind detitscheLiterahir zindflir Pddagogik20 (1907): 52-69. Fritz Schernmel, "Die Schulzeit des Kaisers Julian, " Philologits 82 (1926): 455-66. Susan F. Schmerling, "Asymmetric Conjunction and Rules of Conversation, " Syntax and Semantics. SpeechActs, vol. 3 eds. Peter Cole and Jerry L. Morgan (New York: Academic Press, 1975): 211-231. Wilhelm Schmid, Der Atticismits in seinen Haitptvertretern, Bd. 1 (Hildesheim: Georg Olms Verlagsbuchhandlung, 1887/1964). John Searle, "Proper Names, " Mind 67 (1958): 166-173. David Sedley, "The Stoic Criterion of Identity, " Phrohesis 27 (1982): 255-275. Bernard Sesboil6, Basile de Cisarie. Contre Eitnome, Sources Chr6tiennes 299 (Paris: Les ýditions du Cerf, 1982). Manlio Simonetti, "Genesi e sviluppo della dottrina trinitaria, " Basilio di Cesarea. La sita eth, La stia opera e il Basilianesitno in Sicilia 1, Atti del Congresso Internazionale (Messina: Centro di Studi Umanistici, 1983): 169-197. Ineke Sluiter, Ancient Grammar in Context. Contributions to the Study of Ancient Lingitistic Thought (Amsterdam: VU University Press, 1990). H. W. Smyth, GreekGrammar (Cambridge: Harvard University Press, 1956). Richard Sorabji, Matter, Spaceand Motion (London: Duckworth, 1988). Richard Sorabji, Animal Minds and Hitman Morals (London: Duckworth, 1993). Christopher Stead, "The Concept of Mind and the Concept of God in the Christian Fathers, " The Philosophical Frontiers of Christian Theology, eds. B. Hebblethwaite and S. R. Sutherland (Cambridge: Cambridge University Press, 1982): 39-54. Gregory E. Sterling, "Prepositional Metaphysics in Jewish Wisdom Speculation and Early Christian Liturgical Texts, " Wisdom and Logos. Shidies in Jewish,Tholight in Honor of David Winston, eds. David T. Runia and Gregory E. Sterling (Atlanta: Scholars Press, 1997): 219-238.

155 P. F. Strawson, Individitals. An Essay in Descriptive Metaphysics (London: Methuen, 1959). Willy Theiler, Vorbereitung des Neuplatonismus (Berlin: Weidmannsche Buchhandlung, 1930). Marina Silvia Troiano, I Cappadoci e la questione dell' origine dei nomi nella polemica contro Eunornio," VeteraChristianorum 17 (1980):313-346. Lucian Turcescu, "Prosdponand Hypostasis in Basil of Caesarea'sAgainst ElInomills and the Epistles," Vigiliae Christianae51 (1997):374-395. Herinann Usener, Tin altes Lehrgebäude der Philologie, " Kleine Schriften Bd. 2 (Leipzig and Berlin: B. G. Teubner, 1913: 265-314. Karl-Heinz Uthemann, "Die Sprache der Theologie nach Eunomius von Cyzicus, " ZKG 104 (1993): 143-175. Peter Wolf, Vom Scludwesender Spdtantike. Shidien zu Libanius (Baden-Baden: Verlag fiAr Kiinst und Wissenschaft, 1952).

156

r BIEL "S LONDON