Prolegomena: A classical text of Homer in the making

Pⓢ1. The Homeric Koine

P§1 Homer the Classic centers on ancient concepts of Homer as the author of a body of poetry that we know as the Iliad and Odyssey. This body of poetry, this corpus, became a classical text, but it started as something else. That something, as I have argued in earlier projects, is oral poetry. In the present project, however, my aim is not to reassess the Homeric corpus as oral poetry. Rather, I aim to show how the Homeric Iliad and Odyssey became a classic in the days of the Athenian empire and later.

P§2 I will clarify what I mean by the concept of Athenian empire as my argumentation proceeds. I need to say from the start that I could have described the historical reality behind this concept in a variety of alternative ways, steering clear of the English word empire, but I prefer to use this particular word in the light of its derivation from the Roman idea that we translate as ‘empire’, imperium. As for the specific concept of an Athenian empire, I focus on another word, a Greek word:

koinē (κοινή) / plural koinai (κοιναί) ‘common, standard’

The plural form koinai was once used as a substantive in referring to texts of Homer that were ‘common, standard’. The concept of a ‘common’ or ‘standard’ text of Homer, to which I will refer hereafter simply as Koine, is central to my overall project. It comes closest to capturing what I mean by classic in the title Homer the Classic.

P§3 At first glance, my formulation seems anachronistic. Homer, we may be saying to ourselves, does not speak for the Athenian empire. By the time we reach the last chapter of this book, however, we will see that Homer does exactly that: he actually does speak for the Athenian empire – as far as Athenians in the fifth century BCE were concerned. In that age, Homer was {1|2} a classic, an Athenian classic, and Homer’s poetry gave meaning not only to civilization in general but to the Athenian empire in particular. The Athenian version of Homer is shaped by the idea of a unified text of Homer, a Homeric Koine. This idea of a Homeric Koine, a text that is both ‘common’ and ‘standard’, matches the idea of a society that is both democratic and imperial.

P§4 I plan to show how the multiple koinai or ‘common’ texts of Homer, as they were known in the city of Alexandria around the middle of the second century BCE, stem from a notionally singular Athenian Koine or ‘standard’ version of Homer that goes as far back as the era of the Athenian empire in the fifth century BCE.

P§5 Beyond the Athenian Koine, the Homeric tradition can be traced all the way back to a poetic lingua franca current already in the Bronze Age, in the second millennium BCE. Athenians who lived in the middle of the first millennium BCE imagined this remote age as the era of the so-called Minoan thalassocracy, the maritime empire of King Minos of Crete, who supposedly lived in the second millennium BCE. The very idea of a Homer who knew about such a remote age, however, is beyond reach without an intermediary. For me that intermediary is the idea of Homer in the age of the Athenian empire. That idea converges with what I am calling the Homeric Koine. To grasp that idea is the objective of this book, as reflected in the title Homer the Classic.

Pⓢ2. Twin books about six ages of Homeric reception

P§6 Homer the Classic (HC) is complemented by the twin book Homer the Preclassic (HPC), which covers the vast prehistoric era that led to the formation of the Homeric Koine. These two books, between the two of them, cover six ages of Homeric reception. These six ages correspond to six lectures I gave in the spring semester of 2002 at the University of California at Berkeley while I was teaching there as the Sather Professor for 2001/2. Here are the six ages, arranged in a sequence going backward in time:

HC ch. 1. Homer the Classic in the age of Virgil

HC ch. 2. Homer the Classic in the age of Callimachus

HC ch. 3. Homer the Classic in the age of Plato

HC ch. 4. Homer the Classic in the age of Pheidias

HPC Part I. A Preclassical Homer from the Dark Age

HPC Part II. A Preclassical Homer from the Bronze Age. {2|3}

P§7 The first four entries in this list correspond to the titles of the four chapters in Homer the Classic. The next two entries correspond to the titles I give to the two parts of the twin book Homer the Preclassic. Here I outline the rationale for all six ages. The first four “ages” are of course definable as historical periods, but the last two are prehistoric periods that can only be defined imprecisely and that cover much wider time-spans than the other four. None of these six ages, of and by themselves, has a direct relevance to Homer. They become relevant only if we think of each of them as windows through which we observe the history and prehistory of Homeric poetry. If we look through any one of these six windows, we can see one of six different views of Homer. Through four of the six windows, Homer can be viewed as a classic. Through the other two windows, Homer can be viewed as a classic in the making. Each of the six time-frames I have chosen is different to the extent that Homer and Homeric poetry will have evolved through time.

P§8 The idea of viewing Homer through several different time-frames is consistent with my overall approach to Homer, which goes beyond current debates concerning Homeric “orality” or “literacy.” [1] Such debates presuppose some alternative ideas that I do not share. In terms of such ideas, Homer was not really classical or even preclassical: he was primordial. Such a primordial Homer, whether or not his name was Homer, was some kind of primitive; if he was a genius, he was a primitive genius. [2]

P§9 By contrast, the Homer of Homer the Classic and Homer the Preclassic is more than just a hypothetical person. He is a historical concept. As a concept, Homer is a metonym for the text and the language attributed to Homer in historical times. By metonym I mean an expression of meaning by way of connecting something to something else, to be contrasted with metaphor, which I define for the moment as an expression of meaning by way of substituting something for something else. [3]

P§10 The Homer of Homeric poetry was not just prototypical. He was also definitive in the minds of Greek-speaking people who lived in each of the six time-frames I have chosen for viewing Homer. By the time Greek literature emerged as a historical reality in the sixth and fifth centuries BCE, Homer was already a standard of definition for this literature. In other words, the poetic figure of Homer was already a classic. {3|4}

Pⓢ3. An evolutionary model

P§11 Before we set out to explore what historical and prehistorical forces could have made Homer such a classic, we must first ask ourselves what it was that made him different in each of these six ages. My answer, in its simplest form, is that Homeric poetry kept evolving throughout the ages. I highlight the word evolving. Here I turn to my earlier work, in which I developed an evolutionary model for the making of Homeric poetry, positing a tentative descriptive framework of five periods, “Five Ages of Homer,” as it were. Unlike the six ages I have just listed, which go backward in time, the ages of this evolutionary model go forward in time. To avoid confusion with the six ages, I will from here on speak exclusively in terms of five periods of Homer, not five ages, whenever I refer to my evolutionary model. The details of that model have been worked out in two books I published in 1996. [4] Here I offer an outline, in its simplest form:

Period 1 of Homer was a relatively most fluid period, with no written texts, extending from the early second millennium BCE to the middle of the eighth century in the first millennium BCE. When I say that this is a most fluid period, I mean that epic was most susceptible to change in this period of its evolution.

Period 2 of Homer was a more formative or Panhellenic period, still without written texts, extending from the middle of the eighth century BCE to the middle of the sixth. [5]

Period 3 of Homer was a definitive period, centralized in Athens, with potential texts in the sense of transcripts, extending from the middle of the sixth century BCE to the later part of the fourth. Somewhere near the start of this period, there was a reform of Homeric performance traditions in Athens.

Period 4 of Homer was a standardizing period, with texts in the sense of transcripts or even scripts, extending from the later part of the fourth century BCE to the middle of the second. Somewhere near the start of this period, there was another reform of Homeric performance traditions in Athens. {4|5}

Period 5 of Homer was a relatively most rigid period, with texts as scripture, from the middle of the second century BCE onward. This period starts with the completion of the editorial work of Aristarchus of Samothrace on the Homeric texts, not long after 150 BCE or so.

P§12 In this sequence of five periods, we see the intervention of what I call written texts, starting at period 3. The term written text, however, is too imprecise for my purposes, and I introduce the more precise terms transcript, script, scripture. [6] By transcript I mean the broadest possible category of written text: a transcript can be a record of performance, even an aid for performance, but not the equivalent of performance. [7] We must distinguish a transcript from an inscription, which can traditionally refer to itself in the archaic period as just that, an equivalent of performance. [8] As for script, I mean a narrower category, where the written text is a prerequisite for performance. [9] By scripture I mean the narrowest category of them all, where the written text need not even presuppose performance. [10]

P§13 To keep the relative chronology in focus, I now offer a composite outline that correlates the four chapters of Homer the Classic and the two parts of Homer the Preclassic with my evolutionary model of five periods of Homer:

HC Chapter 1. Homer in the age of Virgil is to be situated within period 5 of Homer, which I just described as a relatively most rigid period. To repeat, when I say most rigid I mean that Homeric poetry was least susceptible to change in this period of its evolution.

HC Chapter 2. Homer in the age of Callimachus is to be situated in period 4 of Homer, which I have described as a standardizing period. The term standardizing is meant to convey the idea of a movement toward standardization, not the actual achievement of a standard version of Homer. In the course of this standardizing period, there is an exceptional sub-period when the Homeric tradition is found to be in flux, fluctuating between older and newer standards. That sub-period {5|6} is what I call the age of Callimachus. The newer standard is represented by the age of Aristarchus in period 5, as explored in HC Chapter 1, while the older standard is represented by the age of Plato in period 3, as explored in HC Chapter 3. The age of Callimachus, as explored in HC Chapter 2, falls between these two standards.

HC Chapter 3. Homer in the age of Plato is to be situated within the later phases of period 3 of Homer, a definitive period. The term definitive is meant to convey the idea of an older standard version of Homer. This idea corresponds to what I call in this book the Koine of Homer.

HC Chapter 4. Homer in the age of Pheidias is to be situated within the earlier phases of period 3 of Homer. In this age as well, the idea of a standard version of Homer corresponds to what I call the Koine of Homer. We will consider the idea of such a Koine in the historical context of Athenian cultural hegemony, with specific reference to the appropriation of Homer by the Athenians. As I will argue, the state of Athens appropriated Homer as an expression of its imperial ideology.

HPC Part 1. Homer in the Dark Age is to be correlated with period 2 and with the later phases of period 1 of Homer.

HPC Part 2. Homer in the Bronze Age is to be correlated with the earlier phases of period 1 of Homer, a relatively most fluid period. To repeat, when I say most fluid I mean that the antecedent of what later became Homeric poetry was most susceptible to change in this period of its evolution.

P§14 Having outlined these six ages of Homer in terms of my evolutionary model, one age after the next, I will now restate my goal in these terms. My goal is to explore what made Homer a classic in the course of these six ages.

Pⓢ4. Metaphors of rigidity and definitiveness

P§15 I am working with two metaphors in this composite descriptive scheme: rigidity and definitiveness. In terms of the first metaphor, rigidity, we will proceed, overall, from the most rigid to the most fluid period of Homer as we move backward in time. I could have said instead: from least fluid to least rigid. {6|7} From here on, however, I prefer to say that we are going from the relatively most rigid to the least rigid period of Homer. That is because I want to avoid formulations that oppose the terms rigid and fluid as polar opposites. Such a binary opposition would imply an absolute criterion that forces a choice between rigid and fluid. I need to make it clear that the metaphor of rigidity conveys a relative criterion, designed to measure varying degrees of rigidity or fluidity in the evolution of Homeric poetry. As we will see later on, moreover, there is one era that defies the general trend toward relatively more rigidity as we move forward in time: that era is what I am calling the age of Callimachus, which falls within the broader time frame of period 4. As for why I choose rigidity as the default term rather than fluidity, I will offer an explanation in Chapter 1.

P§16 More important, for the moment, is the second of my two operative metaphors, definitiveness. In this case, my general aim is to examine relative degrees of definitiveness in the status of Homer from one age to the next. I also have a specific aim, which is to trace the history of changes in the meaning of a Greek word that had once upon a time referred to the definitiveness of Homer. The word in question is the adjective koinos ‘common, standard’ and its derivatives, as applied to the transmission of Homer. In the chapters that follow, I will focus on the application of the feminine singular form, koinē, to the notionally unified text of Homer. The notion of such a text, such a Koine, will become increasingly relevant to my argumentation as we move historically backward in time from period 5 to period 3 of Homer.

P§17 In period 5, the significance of this word koinē was relatively simple, as we will see in Chapter 1 when we view a specific era within the broader time-frame of this period. The era is the age of Virgil. In this era, the singular koinē and the plural koinai were used to refer to ‘common’ manuscripts of the text of Homer. In the earlier period 3, by contrast, the significance of this same word koinē was far more complex. In the era of the Athenian empire, around the second half of the fifth century BCE, this same word koinē meant not only ‘common’ but ‘standard’, and these two meanings were not at all opposed to each other. As I hope to show in Chapter 4, this earlier usage reflects the combined cultural heritage of Athenian democracy and Athenian empire, in that the Koine version of Homer was once upon a time not only democratic but imperial.

P§18 I stop here for a moment to observe that the ideas of common and standard become opposites only if the standard loses its definitiveness. In the case of such a loss, the concept of common can become perceived as not standard but substandard, even vulgar. {7|8}

P§19 For the present, in any case, it is enough for me to stress that the term Koine conveys a relative concept as it applies to Homer. In other words, translations like ‘standard’ and ‘common’ are relative criteria, not absolute ones, when we apply them to Homer. Earlier, when I described the so-called period 3 of Homer as definitive and period 4 as standardizing, I intended these descriptions to be understood as relative terms. In order to make this intention of mine more explicit, from here on I will describe periods 3 and 4 simply in terms of relatively more or less definitiveness, and I will extend the description further to period 5, which can be viewed as the relatively most definitive period of them all.

P§20 Having stressed the relativity inherent in my criteria of definitiveness and rigidity, I now proceed to apply them to the concept of Homer as a classic. This time, I am saying a classic, not the classic, because the term classic can be viewed as an absolute only from the synchronic standpoint of each Homeric period taken separately; it needs to be viewed as a relative term, however, from the diachronic standpoint of all Homeric periods taken together. [11]

P§21 By now I have streamlined my evolutionary model to indicate consecutive Homeric periods of increasing definiteness, which can be correlated with increasing rigidity. The one exception to the trend of increasing definitiveness and rigidity, as we will see in some detail when we reach Chapter 2, is the age of Callimachus, situated in period 4.

P§22 Having explained my two criteria of definitiveness and rigidity, I am ready to start with Homer in the age of Virgil, situated in period 5, which I describe as the relatively most definitive and most rigid period of Homer.

P§23 At this point I have reached a crossroads, where I offer readers a choice between two pathways of reading what follows overall. The longer way is to continue reading these Prolegomena, where I proceed to test the criteria of definitiveness and rigidity. The shorter way is to stop here and to skip ahead to Chapter 1 on Homer in the age of Virgil.

P§24 For those readers who have not yet decided at this point whether to go ahead and read the rest of the Prolegomena, I will now say in one sentence what I aim to do here. Essentially, I will be exploring the historical background of the actual text of Homer as we have it, outlining that background by tracing the usage of the word koinē with reference to the textual history of Homer. Having read this sentence, some may now decide not to confront the historical {8|9} background of the text, or to confront it perhaps only after reading the overall argumentation of the four chapters that follow the Prolegomena. For those who have made such a decision, now is the time to stop and to skip ahead to Chapter 1.

Pⓢ5. Aristarchus and the Homeric Koine

P§25 The term Koine corresponds to the word koinē as it was used by Aristarchus of Samothrace, director of the Library of Alexandria in the middle of the second century BCE, with specific reference to the classical or ‘standard’ text of Homer. The evidence comes mostly from the usage of the Aristarchean scholar Didymus, who flourished in the first century BCE. [12]

P§26 The idea of a Homeric Koine is approximated by the reality of the base text used by Aristarchus for his editing of Homer. To put it another way, the Homeric Koine was an idea that reached its ultimate form through the efforts of Aristarchus in seeking to build – or, as he thought, rebuild – the best possible base text for his editing of Homer. By base text here I mean simply the text that serves as the basis of the editing process.

P§27 The idea of a Homeric Koine can be reconstructed on the basis of analyzing the use of the actual word koinē / koinai in ancient sources. As we will see from the Homeric scholia, the word koinē / koinai was used as a default term for Homeric texts. I will argue that this usage stems from criteria applied by Aristarchus in his hupomnēmata ‘commentaries’ on Homer. By Homer here I mean a definitive base text of Homer, as built or rebuilt by Aristarchus.

P§28 The use of the word koinē / koinai in the Homeric scholia is complicated and potentially confusing, reflecting different critical trends stemming from different times – as represented by Didymus and Aristarchus. I start by offering an overview of these differences.

P§29 By the time of Didymus, in the first century BCE, the word koinē must have lost some of its definitiveness: he must have understood it to mean simply ‘common’ – and no longer necessarily ‘standard’. Some sense of definitiveness was still felt, however, in the earlier time of Aristarchus, in the second century BCE. Faced with multiple koinai or ‘common’ texts of Homer, Aristarchus treated them as if they were all derived from a definitive and {9|10} notionally singular koinē or ‘standard’ version of Homer. His own base text of Homer was a reconstruction of such a koinē.

P§30 Before I proceed to examine specific applications of the term koinē / koinai in the Homeric scholia, I need to clarify further my terminology. I am saying that the ancient idea of a singular Koine of Homer, a core text, was tied to the reality of a base text that Aristarchus established for his editing of Homer. But the question remains: what exactly was this reality, as far as Aristarchus himself was concerned?

P§31 The reality of a base text as the editorial point of reference for Aristarchus is not to be confused with the modern idea of an edited text of Homer, where an editor reproduces, word for word, what Homer had supposedly created. We can see the essential difference between an ancient base text and a modern edited text of Homer when we review the methods developed by Aristarchus in confronting the problem of Homeric textual variants.

P§32 When modern editors of Homer are faced with a choice between two or more textual variants, they show in their edited text what they judge to be the correct variant, relegating the supposedly incorrect variant or variants to an apparatus criticus. By contrast, an ancient editor like Aristarchus displayed the sum total of his editorial judgments – that is, his diorthōsis ‘correction’ – only in his hupomnēmata ‘commentaries’, not in his base text.

P§33 But there is more to it. When it came to deciding which variant to show in his base text, Aristarchus allowed the external evidence of Homeric texts described as koinai ‘common’ to override the internal evidence – as he saw it – of Homeric usage. As a rule, he would show in his base text the variant that reflected what he understood to be the koinē or ‘standard’ version of Homer – even if he preferred another variant on the basis of internal evidence.

P§34 What I mean by external evidence is this: Aristarchus actually assembled and collated a wide variety of Homeric texts. This aspect of Aristarchean research is sometimes not recognized. [13] Moreover, as I am about to argue, Aristarchus relied on a consensus of koinai texts in order to build – or, as he thought, rebuild – a definitive base text, which for him was tantamount to what I am calling the Koine version of Homer. In other words, Aristarchus posited as his base text of Homer a notionally singular and unified Koine – a koinē or ‘standard’ version that needed to be restored by way of ‘correction’ or diorthōsis of texts. In order to build such a singular koinē, Aristarchus would have had to assemble and collate a number of koinai texts. {10|11}

Pⓢ6. Homeric koinai and khariesterai

P§35 As we see from the reportage of Didymus, mediated by the Homeric scholia, Aristarchus also assembled and collated a number of less common Homeric texts, described as khariesterai, that is, ‘having more kharis’ or ‘more graceful’ than the koinai texts. [14] The concept of kharis, which conveys the idea of beauty and the pleasure that goes with it, can be translated as ‘pleasurable beauty, grace, gracefulness’. Its adjectival form, kharieis ‘graceful’, was for Aristarchean critics a criterion for determining which version of the Homeric text was most likely to be genuinely Homeric. This criterion did not originate with the Aristarchean critics in the first or even in the second century BCE: as we will see later on, it can be traced as far back as the fourth century BCE. For now, however, I concentrate on the actual opposition between koinai and khariesterai.

P§36 I start with the khariesterai. The implied noun that goes with khariesterai is graphai in the sense of ‘texts’ (as we see in the explicit combinations ἐν δὲ ταῖς χαριεστέραις γραφαῖς in the scholia for Odyssey i 356 and αἱ χαριέστεραι γραφαί in the scholia for Odyssey xi 196). From the Homeric scholia, which reflect primarily the reportage of the Aristarchean scholar Didymus, we can see that there were basically two kinds of khariesterai texts of Homer:

(1) the editions of two pre-Aristarchean editors of Homer, namely, Zenodotus of Ephesus (third century BCE) and Aristophanes of Byzantium (second century BCE), as well as other texts derived from even earlier figures such as Rhianos of Crete (third century BCE) and Antimachus of Colophon (fifth/fourth centuries BCE) [15]

(2) the so-called politikai or ‘city editions’ stemming from Massalia (Marseille), Chios, Argos, Sinope, Cyprus, and Crete. [16] {11|12}

Thekhariesterai

texts were supposedly better than the texts known askoinai

‘common’ (as we see from the use ofkoinai

in the scholia A forIliad

IV 170). Thekoinai

texts were otherwise known asdēmōdeis

‘popular’ (scholia A forIliad V

881a1).

P§37 In the critical discourse of Aristarchean scholars, antigrapha ‘copies’ of texts judged to be khariestera were expected to contain written forms that were khariestera, as opposed to copies judged to be koina:

Pⓣ1 Scholia for Odyssey xvii 160

ἐν τοῖς χαριεστέροις οὗτοι μόνοι οἱ β' ἀθετοῦνται … ἐν δὲ τοῖς κοινοτέροις ἀπὸ τοῦ ὣς ἔφατο ἕως τοῦ ἐξ ἐμεῦ.

In the khariestera only these two lines are athetized [17] [= Odyssey xvii 160-161]. … But in the koinotera [there is athetesis] from ὣς ἔφατο [at the beginning of xvii 150] all the way to ἐξ ἐμεῦ [at the beginning of xvii 165].

P§38 Here the antigrapha ‘copies’ that are khariestera athetize [18] only two verses (xvii 160-161), while the antigrapha that are koina (= koinotera by comparison with the khariestera) athetize sixteen verses (xvii 150-165).

P§39 For Aristarchus, the existence of Homeric graphai ‘texts’ described as khariesterai and therefore judged to be better than other Homeric texts that were koinai ‘common’ did not rule out the possibility that even ordinary texts occasionally contained variants judged to be better. Conversely, variants judged to be ordinary could occur even in supposedly better texts. Examples include some variants that Aristarchus judged to be koinai even though he found them in the Homeric ekdosis ‘edition’ of Zenodotus, whose text he judged to be one of the khariesterai. Such variants stemming from the ekdosis of Zenodotus in the third century BCE were koinai in comparison to variants stemming from the ekdosis of Aristophanes in the second century, with whom Aristarchus agreed more often than with Zenodotus. Generally, the editions of both Zenodotus (scholia A for Iliad II 579; AT for Iliad VII 428a1) and Aristophanes (scholia A for Iliad II 53a1) adduced variants judged to be khariesterai, but Zenodotus was considered to be less consistent and less thorough than Aristophanes – if we apply an Aristarchean standard when it comes to finding khariesterai as alternatives to the koinai (scholia T for Iliad II 53a2). {12|13}

P§40 The editorial methods of Aristarchus resembled most closely those of Aristophanes. [19] Still, his own methods were superior, because he made a special effort to develop a broader data base of textual variants by way of assembling and collating a variety of texts (graphai) – not only khariesterai but also koinai.

P§41 Through the khariesterai, Aristarchus achieved a superior view of degrees of multiformity in the textual history of Homeric poetry. Through the koinai, he achieved a superior view of degrees of uniformity. The relative multiformity of the khariesterai is a generally recognized fact, but not so the relative uniformity of the koinai.

P§42 I will offer a variety of arguments to support what I just said about the koinai. In the next paragraph, I begin with an overall formulation, which is meant as an alternative to another formulation that has received wide currency. [20]

P§43 Aristarchus distinguished himself from his predecessors by demonstrating, on the basis of the textual evidence he assembled and collated, that the consensus of the koinai texts recovers a Koine text. This Koine was not the real Homer for Aristarchus, but it was the base text from which a real Homer could be reconstructed by way of extensive analysis and debate in his hupomnēmata ‘commentaries’. The text of the real Homer as he saw it was latent in the relative multiformity of the khariesterai texts, and this multiformity could be displayed only in the background, that is, only in his commentaries. By contrast, the text of the Koine was overt in the relative uniformity of the koinai texts, and this uniformity could be displayed in the foreground, that is, in the base text. The ultimately real Homer could take shape only through a process of further selection, emerging from a background of relative multiformity in the khariesterai texts, while the Koine text had already achieved its shape through a process of consensus, evident in the foreground of relative uniformity in the koinai texts. For Aristarchus, an accurate picture of this consensus was the basis for reconstructing the text of a genuine Homer that transcended this consensus. In other words, the Koine as a consensus of koinai texts was the basis for reconstructing this supposedly genuine Homer through the variants provided by the khariesterai texts.

P§44 Here I need to return to a point I made earlier about the base text of Aristarchus: it approximated such a Koine text. Aristarchus kept out of this base text the special forms he found in the khariesterai texts, privileging the {13|14} consensus emerging from the forms he found in the koinai texts. He would express his own editorial preferences only in his hupomnēmata ‘commentaries’. O nly where variations did not affect the content did Aristarchus make decisive choices in the external appearance of his base text. I will consider some of these choices later on, but for now the focus remains on the point I am making about the base text of Aristarchus. This base text approximated the Koine of Homer.

P§45 The rationale of Aristarchus in building or rebuilding such a Koine by way of collecting and collating koinai texts can be explained as his active response to two kinds of variation in Homeric textual transmission, (1) “vertical” and (2) “horizontal.” [21]

Pⓢ7. “Vertical” variation and the numerus versuum

P§46 Aristarchus included in his base text of Homer even those verses he suspected of being non-Homeric. That is, he included any such verses if the given verses were strongly attested. The term strong attestation here refers to situations where attested verses were well represented in the Homeric manuscripts that Aristarchus assembled and collated. [22] Verses that he admitted on these gr ounds but suspected on other grounds were athetized by him, that is, they were marked by a marginal sign. This sign, also used by his predecessor, Zenodotus, was the obelos (–). The obelos, placed in the margin to the left of an athetized verse, cross-referred to the relevant analysis of Aristarchus in his hupomnēmata ‘commentaries’, where he would give his reasons for rejecting the given verse as not genuinely Homeric. It is essential to stress, from the start, that verses athetized by Aristarchus were included, not excluded, in his text of Homer – that is, in his base text.

P§47 To be distinguished from the athetized verses are the excluded verses: Aristarchus excluded from his base text of Homer verses he judged to be non-Homeric – if the given verses were only weakly attested. The term weak attestation here refers to situations where attested verses were not so well represented in the Homeric manuscripts that Aristarchus assembled and collated. [23] These are the verses known today as the plus verses. [24] {14|15}

P§48 Occasionally, we learn from reportage in the Homeric scholia that a given plus verse had been explicitly rejected in the Homer edition of Zenodotus by way of deletion, not by mere athetesis. [25] Here we see an essential difference between the base texts of Zenodotus and Aristarchus. Zenodotus had developed an editorial system that allowed him to choose between either athetizing a suspected verse, by way of an obelos, or deleting it, by way of various deletion marks and brief annotations that accompanied these marks. [26]

P§49 The base text of Homer as edited by Zenodotus actually included not only the verses athetized by him but also the plus verses marked with signs indicating the editor’s deletion. [27] By contrast, the base text of Homer as edited by Aristarchus excluded the plus verses altogether, and this policy of exclusion eliminated the need for signs indicating deletion. [28] Also eliminated from Aristarchus’ base text were comments written into the margins, which were replaced by extensive hupomnēmata ‘commentaries’ written into separate scrolls. [29] The commentaries of Aristarchus were correlated with an expanding set of marginal signs in the base text, indicating a variety of editorial judgments. [30] These signs could help the reader navigate between the base text of Homer as displayed by Aristarchus and the separate text of his commentaries, where all his editorial judgments about all the textual variations could be fully discussed. [31] In short, the base text of Aristarchus was more streamlined than that of Zenodotus. As for the base text of Aristophanes, it was almost as streamlined as that of Aristarchus himself. [32] In what follows, I will explain why I qualify my statement.

P§50 The differences between the length of the base texts of Aristophanes and Aristarchus on the one hand and the length of the base texts of Zenodotus on the other hand are an example of what I call vertical variation.

P§51 The relatively exclusive base text of Aristarchus adhered to a notionally fixed and standard verse-count, and there seems to have been {15|16} a nearly identical verse-count in the base text used by Aristophanes. [33] I say nearly identical rather than identical because, as I will argue, Aristarchus must have surpassed Aristophanes in the assembling and collecting of Homeric texts. Such a fixed verse-count is the numerus versuum. [34]

P§52 To justify this working definition of the numerus versuum, I recall the fact that Aristarchus included in his base text the verses that he athetized by marking them with an obelos. To repeat, this base text was the sum total of non-athetized and athetized verses. Further, this base text represented the consensus of the koinai texts that Aristarchus assembled and collated. Such a consensus of many koinai texts was the closest approximation of a single koinē text, and this notionally single Koine showed a fixed verse-count. To repeat, such a fixed verse-count is the numerus versuum.

P§53 What Aristarchus confirmed by way of assembling and collating his samples of koinai texts was that they tended to conform to such a fixed verse-count, the numerus versuum, whereas the khariesterai texts did not (we see examples in the scholia for Odyssey i 356, xvii 160). The consistency of this fixed verse-count, the numerus versuum, became more and more of a reality in the actual process of assembling and collating more and more samples of koinai texts. For Aristarchus, the singularity of the numerus versuum emerged out of the plurality of the koinai texts. The singularity of this numerus versuum corresponded to the singularity of the Homeric Koine. In other words, the Koine that Aristarchus built or rebuilt from the koinai texts turned out to be a reality – something that had a fixed verse-count. That is the essence of the numerus versuum.

P§54 In short, my argument is that the numerus versuum of the base text of Aristarchus and, before him, of Aristophanes, was inherited from a preexisting Homeric Koine. The exactitude of Aristarchus in reconstructing this Koine would have surpassed the exactitude of Aristophanes for this simple reason: Aristarchus assembled and collated more texts than did his predecessor.

P§55 The standard verse-count of the Homeric Koine version of Homer – the numerus versuum – was not the same thing as the verse-count of the ostensibly real Homer as reconstructed by Aristarchus the editor. In his commentaries, Aristarchus systematically explained his reasons for rejecting as non-Homeric the verses that he athetized in his base text, marking these verses there with the obelos; to be contrasted are the plus verses, which he {16|17} simply excluded from his base text. [35] For Aristarchus, the real text o f the real Homer was whatever number of verses would be left over after he subtracted the number of verses that he athetized or deleted. Such a text would have a verse-count far lower than the standard verse-count of the standard Homer. That standard Homer was the Koine, and its standard verse-count was the numerus versuum.

P§56 The supposedly real Homer of Aristarchus was not the only Homeric text with a verse-count that was far lower than the standard numerus versuum of the Koine. If we could subtract the number of verses that Zenodotus had athetized from the total number of verses in his base text of Homer, the resulting verse-count would be even lower. But the essential difference between the verse-counts of Aristarchus and Zenodotus is to be sought elsewhere. As we will see in Chapter 2, the number of verses in the edition of Zenodotus, if we subtracted only the verses he excluded but not the verses he athetized, would have exceeded the number of verses produced by Aristarchus in his own edition, which was a number that corresponded to the numerus versuum of the Koine. In terms of the ongoing argumentation, this numerus versuum could be recovered only because Aristarchus actually had a working procedure for collating the Homeric texts available to him.

P§57 As for the verse-count of Aristophanes, it would have been based on an earlier form of the procedure that was later to be perfected by Aristarchus. In this respect and in many others as well, the Homer edition of Aristarchus was far closer to the edition of Aristophanes and much farther apart from the edition of Zenodotus. [36]

P§58 Crates of Mallos, who was director of the library of Pergamon during roughly the same period when Aristarchus was director of the library of Alexandria, was likewise engaged in the editorial practice of athetizing Homeric verses, that is, rejecting verses on the grounds that they are non-genuine, non-Homeric. [37] But the number of verses in the base text of Crates – as also of Zenodotus – exceeded the standard verse-count or numerus versuum of the Homeric Koine, whereas the base text of Aristarchus adhered to it. [38] In contrast to the base text of Aristarchus, which followed the numerus versuum, the augmented base text of Crates – as also of Zenodotus – corresponded to what I call the Homerus Auctus. I will explain in Chapter 2 what I mean by {17|18} this term. For now, I simply summarize the emerging patterns of distinctions between the editorial approaches of Crates and Aristarchus: [39]

A) Both men athetized verses, but both included in their texts the verses that they athetized.

B) Aristarchus athetized more verses than Crates.

C) Aristarchus also omitted verses that Crates included; or, to put it differently, Crates had access to texts containing some verses that Aristarchus could not or would not verify on the basis of the texts to which he had access. These extra verses correspond to what I have been calling plus verses.

D) Both men tracked variant readings within the verses, recording them and commenting on them in their formal commentaries (hupomnēmata / diorthōtika). Needless to say, the “same” verse containing different variant readings was a “different” verse as far as Aristarchus and Crates were concerned. [40]

E) Aristarchus used a system of signs, affixed at the left-hand margins of his text, to indicate his editorial differences with Crates – as also with his own Alexandrian predecessors, especially with Zenodotus and with Aristophanes of Byzantium. [41]

P§59 Earlier, I noted that the base text of Aristarchus included verses he personally judged to be non-Homeric – if such verses were strongly attested. Now I need to stress again that the very concept of a strong attestation was possible only because Aristarchus had assembled and collated texts that he judged to be koinai ‘common’ – not only the less common texts that he judged to be superior {18|19} and therefore khariesterai ‘having more kharis’ or ‘more graceful’. As I have argued, the base text of Aristarchus was built from a consensus of manuscript readings found in Homeric texts that Aristarchus judged to be koinai, not khariesterai. Such a consensus, as we are about to see, was needed as a control not only in the case of vertical variation. It was needed also in the case of horizontal variation.

Pⓢ8. “Horizontal” variation

P§60 The equivalence of the base text of Aristarchus with the singularity of what I call the Homeric Koine extends beyond patterns of consistency in verse-counts, that is, beyond patterns of conformity to the standard of a numerus versuum. Besides confronting the “vertical” textual variations of more or fewer verses in any given sequence of Homeric verses, Aristarchus in his hupomnēmata ‘commentaries’ also confronted the “horizontal” textual variations, that is, variant wordings within the frame of any given verse. [42] For his base text, Aristarchus sought a consensus among horizontal as well as vertical variations.

P§61 Wherever the horizontal variations of wording significantly affected the content of any given Homeric verse, Aristarchus retained in his base text the wording that best reflected the consensus of the koinai texts, while he adduced the variant wordings only in his commentaries. [43] That is, he retained in his base text the wording he found in the koinai texts – if that given wording was strongly attested. The term strong attestation here refers to the testimony of the majority of the Homeric manuscripts that Aristarchus assembled and collated. [44]

P§62 As I noted earlier, only where variations did not significantly affect the content did Aristarchus make his own choices in his base text. Many of these choices had to do with external appearances, as I will show later on. For now, however, I continue to concentrate on the idea of the base text of Aristarchus as a consensus of koinai texts. As we will see from the horizontal as well as the vertical dimension of textual variations, there is an equivalence between this base text of Aristarchus and the standard of what I am calling the Homeric Koine. {19|20}

P§63 In some cases, Aristarchus seems to have re-introduced into his base text a Koine reading that had earlier given way to a reading judged to be khariesteron ‘having more kharis’ in the base text of his predecessor Aristophanes (scholia A for Iliad II 53a1). In such cases, the re-introduction of the Koine reading could have been caused by the availability of more evidence from the koinai. What I noted in the case of vertical variations applies also here in the case of horizontal variations: the exactitude of Aristarchus in reconstructing the Koine would have surpassed the exactitude of Aristophanes – to the extent that Aristarchus assembled and collated more texts than did Aristophanes.

P§64 Although Aristarchus conformed to the standard of the Koine, later generations of Aristarcheans preferred a different standard, attributed to Aristarchus himself. In the case of vertical textual variations, there was no serious difference between the two standards, in that the number of verses in the base text of Aristarchus – which included the verses athetized by him but which excluded the plus verses – matched the numerus versuum of the Koine. Later generations of Aristarcheans may have agreed or disagreed with the decisions of athetesis made by Aristarchus, but, even if they disagreed, they would not exclude the athetized verses from consideration, just as Aristarchus did not exclude them from his base text. In the case of horizontal textual variations, on the other hand, the variant wordings as reported by Aristarchus in his commentaries could easily infiltrate the base text, actually ousting the wordings inherited by the Koine. Such is the state of affairs already in the time of Didymus. By his time, in the first century BCE, the authority of wordings found in the Koine had already given way to the authority of variant wordings preferred by Aristarchus himself – wordings originally confined to the master’s hupomnēmata ‘commentaries’. For Aristarcheans like Didymus, the preferred readings of Aristarchus became more significant than the received readings of the Koine.

P§65 Such a shift from the Koine standard to an Aristarchean standard is a source of confusion for editors of the Homeric scholia – and even for editors of the Homeric Iliad and Odyssey. When Aristarchus quoted a lēmma, that is, an ‘extract’ from the Homeric text that provided the headline for editorial discussion in the Aristarchean commentaries, that lēmma referred predictably to the wording of his base text, which reflected the Koine (we see an example in the scholia A for Iliad I 465b). When Didymus quotes a lēmma, on the other hand, we find that such a lēmma may refer to variant readings originally reported by Aristarchus in his hupomnēmata ‘commentaries’ (we see an example in the scholia A for Iliad XII 404a). The distinction between lēmmata derived from {20|21} the base text and lēmmata derived from the commentaries of Aristarchus thus becomes blurred in the Homeric scholia.

Pⓢ9. Homeric editions attributed to Aristarchus

P§66 The privileging of the Aristarchean standard in the Homeric scholia is linked with the fact that Didymus, one of the primary sources of the scholia, relied heavily on two manuscripts that featured base texts containing alterations derived from the hupomnēmata ‘commentaries’ of Aristarchus. Such alterations, reflecting the preferred readings of Aristarchus, displaced the received readings of the Koine. The two manuscripts in question, in the terminology of Didymus, were hē Aristarkheios ‘the Aristarchean’ and hē hetera Aristarkheios ‘the other Aristarchean’ (scholia A for Iliad II 579). The noun implied by these adjectives is ekdosis ‘edition’. Didymus refers to both of these ekdoseis ‘editions’ as khariesterai ‘having more kharis’ (scholia A for Iliad II 12a). The combined testimony of these two ekdoseis is described as khariesteron (scholia AT for Iliad VII 428a1). Within this exclusive category of two distinct Aristarchean ekdoseis ‘editions’, we find differences in content. There are occasions, for example, where a variant reading in one of these two ekdoseis is different from the variant reading in the other. [45] It also happens that the variant reading preferred in one of the two ekdoseis is described as khariestera in comparison to the variant reading preferred in the other (scholia A for Iliad II 579). It even happens that one of the two ekdoseis, in giving a variant reading that differs from the variant reading given in the other, is described as khariestera in comparison to the other (scholia for Odyssey iv 727).

P§67 The usage of Didymus concerning these two ekdoseis ‘editions’ that he attributes to Aristarchus is relevant to what we learn from the titles of two monographs produced by Aristarchus’ successor, Ammonius of Alexandria: (1) περὶ τοῦ μὴ γεγονέναι πλείονας ἐκδόσεις τῆς Ἀρισταρχείου διορθώσεως ‘About the fact that there did not exist more [pleiones] ekdoseis of the Aristarchean diorthōsis’ (via Didymus in scholia A for Iliad X 397-399a) and (2) περὶ τῆς ἐπεκδοθείσης διορθώσεως ‘About the diorthōsis that underwent an epekdosis’ (via Didymus in scholia A for Iliad XIX 365-368a). I offer comments about each of these two titles, in reverse order: {21|22}

(1) For a working translation, I use the noun ‘epekdosis’ in rendering the concept conveyed by the verb epekdidonai (ἐπεκδοθείσης) in the title attached to the second of the two monographs by Ammonius. We find this word epekdidonai attested nowhere else in extant Greek texts.

(2) There has been an ongoing debate over the meaning of the title attached to the first of the two monographs. The question is, did Ammonius claim that there did not exist ‘more [pleiones] ekdoseis beyond the existing single ekdosis’ or ‘more [pleiones] ekdoseis beyond the existing two ekdoseis’? If the second of these two interpretations were correct, then ἐπεκδοθείσης in the title of the second monograph would refer to the second of two ekdoseis ‘editions’ supposedly produced by Aristarchus himself. In that case, the coined word ‘epekdosis’ in my working translation of ἐπεκδοθείσης would be interpreted to mean something like a ‘subsequent edition’. [46] If the first interpretation is correct, however, then Ammonius had in mind simply an ongoing process of diorthōsis or ‘[editorial] correcting’, as opposed to two separate ekdoseis ‘editions’. [47] In terms of this alternative interpretation, the coined word ‘epekdosis’ in my working translation of ἐπεκδοθείσης would be interpreted to mean something like a ‘re-editing’ of an already existing ‘edition’ or ekdosis. [48] In what follows, I will support this interpretation, ‘re-editing’.

P§68 I start by noting that Didymus, who as we saw makes use of two ekdoseis that he thinks are Aristarchean, recognizes the divergent thinking of Ammonius, the successor of Aristarchus. In the scholia A for Iliad X 397-399a, we see that Didymus distances himself from Ammonius: εἴ τι χρὴ πιστεύειν Ἀμμωνίῳ τῷ διαδεξαμένῳ τὴν σχολήν, ἐν τῷ Περὶ τοῦ μὴ γεγονέναι πλείονας ἐκδόσεις τῆς Ἀρισταρχείου διορθώσεως τοῦτο φάσκοντι ‘… if one should to any extent believe Ammonius, the one who took over the School, in his volume entitled “About the fact that there did not exist more [pleiones] than one ekdosis of the Aristarchean diorthōsis,” when he says this …’.

P§69 Despite the divergence between Didymus and Ammonius in their respective views about the number of ekdoseis made by Aristarchus, we can also see an important convergence: Didymus and Ammonius have in common a sustained interest in the question of earlier and later phases in the editorial work of Aristarchus. In the same context that I just cited from scholia A {22|23} for Iliad X 397-399a, Didymus cites Ammonius as his authority in reporting that Aristarchus preferred one variant in an earlier phase of his editorial work and another variant in a later phase: Ἀμμώνιος δὲ ὁ Ἀριστάρχειος πρῶτον μὲν στιγμαῖς φησι τὸν Ἀρίσταρχον παρασημειώσασθαι αὐτούς, εἶτα δὲ καὶ τελέως ἐξελεῖν ‘Ammonius the Aristarchean says that Aristarchus first marked these verses with signs in the margins, but then, at a later point, he took them out completely’.

P§70 What we have just seen is the first of three examples I have found where Didymus consults Ammonius as an authority in determining earlier and later phases in the editorial work of Aristarchus. Two of the three examples, including this first one, involve vertical variation, while the third, as we will see later, involves horizontal variation.

P§71 Now we come to the second of the three examples. Here we find Didymus himself in the act of reporting on different variants found in earlier and later phases of the editorial work of Aristarchus, and in this case he says he finds no relevant report from Ammonius (scholia A for Iliad XIX 365-368a1): ὁ δὲ Σιδώνιος ἠθετηκέναι μὲν τὸ πρῶτόν φησιν αὐτοὺς τὸν Ἀρίσταρχον, ὕστερον δὲ περιελεῖν τοὺς ὀβελούς, ποιητικὸν νομίσαντα τὸ τοιοῦτο. ὁ μέντοι Ἀμμώνιος ἐν τῷ Περὶ τῆς ἐπεκδοθείσης διορθώσεως οὐδὲν τοιοῦτο λέγει ‘[Dionysius] the man from Sidon says that Aristarchus first athetized them [= the verses under consideration], but then, at a later point, removed the obeloi [= the marginal signs indicating athetesis], thinking that such a thing [= what he had athetized] was a poetic creation [of Homer]; Ammonius, on the other hand, in his treatise “About the re-edited [ep-ekdidonai] diorthōsis,” says no such thing’.

P§72 As for the third of the three examples, we find Didymus making a distinction between one horizontal variant (μέγα) found in ‘some of the commentaries’ of Aristarchus (scholia A for Iliad II 111b κατά τινα τῶν ὑπομνημάτων) and a second variant (μέγας) found in ‘one of his commentaries that has been corrected for greater accuracy’ (ἔν τινι τῶν ἠκριβωμένων ὑπομνημάτων); in this context, Didymus notes that Ammonius and another expert, Dionysodorus, attest the second variant but have nothing to say in addition to what Aristarchus had already said about it: καὶ τοὺς ἀπ’ αὐτοῦ δὲ χρωμένους ἔστιν εὑρεῖν τῇ γραφῇ Διονυσόδωρον καὶ Ἀμμώνιον τὸν Ἀλεξανδρέα. ἐπιλέγουσι δὲ οὐδέν· διὸ καὶ τὰς μαρτυρίας αὐτῶν οὐκ ἐγράψαμεν ‘and it is possible to ascertain that those who use the way of writing it [= that variant form] that came from him [= Aristarchus] include Dionysodorus and Ammonius of Alexandria, but they have nothing to say in addition, and so I did not write down their testimonies’. Τo summarize what is being said by {23|24} Didymus here (scholia A for Iliad II 111b): Ammonius and Dionysodorus use the second of two horizontal variants (μέγας vs. μέγα) adduced by Aristarchus, but they add no further information about this second variant, and so the wording of their testimony is not copied out.

P§73 In the third of these three examples we have just considered, I draw attention to a most significant additional detail. We see Didymus himself making a distinction here not only between earlier and later phases in the editorial work of his predecessor Aristarchus in general but also between earlier and later phases in the predecessor’s actual hupomnēmata ‘commentaries’ in particular. [49] This distinction needs to be correlated with another distinction made elsewhere by Didymus – between earlier and later ekdoseis corresponding to earlier and later phases of the Aristarchean hupomnēmata. According to Didymus, Aristarchus based the earlier phases of his hupomnēmata not on an ekdosis of his own but on the previous ekdosis of Aristophanes (scholia A for Iliad II 133a ἐν τοῖς κατ’ Ἀριστοφάνην ὑπομνήμασιν Ἀριστάρχου). [50] This statement of Didymus has led to the inference that Aristarchus went on to produce his own ekdosis only at a later point in his career, corresponding to a later point in his continuing editorial work as reflected in his hupomnēmata. [51] Such an ekdosis by Aristarchus would be tantamount to a ‘publication’ of his own diorthōsis ‘corrective editing’ of Homer. Relatively earlier and later sets of Aristarchean hupomnēmata, let us call them H1 and H2, could then be correlated with respectively earlier and later phases of Aristarchus’ diorthōsis, let us call them D1 and D2. According to one reconstruction, the sequence would be H1 D1 H2 D2. [52] In terms of such a sequence, the claims of Ammonius could be restated this way: he preferred to think of two distinct phases in the history of a single Aristarchean ekdosis of Homer, while others claimed – wrongly, according to Ammonius – that there were two distinct ekdoseis made by Aristarchus. An heir to such an alternative line of thinking was Didymus, {24|25} for whom there existed two ekdoseis, one of which was earlier than the other. In fact, Didymus explicitly refers to one of the two Aristarchean ekdoseis as hē protera ‘the previous one’ (scholia T for Iliad XIX 365-368a2). Moreover, Didymus makes this reference in the same context where he refers also to the work of Ammonius about the ‘epekdosis’ of the Aristarchean text. Didymus, then, seems to assume that the second of the two Aristarchean ekdoseis available to him is derived from the ‘epekdosis’ mentioned by Ammonius.

P§74 We have by now narrowed the differences between Ammonius and Didymus: one spoke of two phases of one ekdosis made by Aristarchus, while the other spoke of two ekdoseis stemming from two phases in the editorial career of Aristarchus. Clearly the idea of two ekdoseis was based on the reality of two distinct texts, and we have evidence from the Homeric scholia that Didymus did use two Aristarchean texts and, further, that these texts were in fact distinct from each other, with one showing earlier features of Aristarchus’ editorial work and the other showing later ones.

P§75 In retrospect, the positing of two ekdoseis that correspond to two different Aristarchean texts seems a most reasonable explanation. After all, it was not an unknown practice for an editor to publish more than one ekdosis in his lifetime, and in fact we have historical documentation of this practice in the era of Hellenistic scholarship.

P§76 A case in point is the information we have about Apollonius of Perga, who flourished in Alexandria during in the second half of the third century BCE. Apollonius produced two editions of portions of his treatise on cone sections, or Conica, and what we read about these two editions turns out to be directly comparable to what we read about the Aristarchean editions of Homer. I start with the remarks of Eutocius of Ascalon (who flourished in the first half of the sixth century CE) with reference to his own editorial work on the Conica of Apollonius:

Pⓣ2 Eutocius Commentaries on the Conica of Apollonius of Perga ed. Heiberg vol. 2 p. 176

πλειόνων δὲ οὐσῶν ἐκδόσεων, ὡς καὶ αὐτός φησιν ἐν τῇ ἐπιστολῇ, ἄμεινον ἡγησάμην συναγαγεῖν αὐτὰς ἐκ τῶν ἐμπιπτόντων τὰ σαφέστερα παρατιθέμενος ἐν τῷ ῥητῷ διὰ τὴν τῶν εἰσαγομένων εὐμάρειαν, ἔξωθεν δὲ ἐν τοῖς συντεταγμένοις σχολίοις ἐπισημαίνεσθαι τοὺς διαφόρους ὡς εἰκὸς τρόπους τῶν ἀποδείξεων. {25|26}

Since there were a number of ekdoseis [pleiones ekdoseis], as Apollonius himself says in his preface, [53] I thought it better to assemble them from whatever source was available and to juxtapose [paratithesthai] in the text [rhēton] the clearer things, in order to facilitate the understanding of beginners, and to indicate on the outside [= in the margin], in the scholia that have been put together, the evidently different variations of the arguments. [54]

P§77 As we have just seen from the statement of Eutocius, the existence of pleiones ekdoseis ‘a number of editions’ of Apollonius of Perga is known from the testimony of Apollonius. The relevant wording of Apollonius himself has actually survived:

Pⓣ3 Apollonius of Perga Conica Book 1, prologue lines 1-22 ed. Heiberg vol. 1

Ἀπολλώνιος Εὐδήμῳ χαίρειν. Εἰ τῷ τε σώματι εὖ ἐπανάγεις καὶ τὰ ἄλλα κατὰ γνώμην ἐστί σοι, καλῶς ἂν ἔχοι, μετρίως δὲ ἔχομεν καὶ αὐτοί. καθ’ ὃν δὲ καιρὸν ἤμην μετά σου ἐν {5} Περγάμῳ, ἐθεώρουν σε σπεύδοντα μετασχεῖν τῶν πεπραγμένων ἡμῖν κωνικῶν· πέπομφα οὖν σοι τὸ πρῶτον βιβλίον διορθωσάμενος, τὰ δὲ λοιπά, ὅταν εὐαρεστήσωμεν, ἐξαποστελοῦμεν· οὐκ ἀμνημονεῖν γὰρ οἴομαί σε παρ’ ἐμοῦ ἀκηκοότα, διότι τὴν περὶ ταῦτα ἔφοδον {10} ἐποιησάμην ἀξιωθεὶς ὑπὸ Ναυκράτους τοῦ γεωμέτρου, καθ’ ὃν καιρὸν ἐσχόλαζε παρ’ ἡμῖν παραγενηθεὶς εἰς Ἀλεξάνδρειαν, καὶ διότι πραγματεύσαντες αὐτὰ ἐν ὀκτὼ βιβλίοις ἐξ αὐτῆς μεταδεδώκαμεν αὐτὰ εἰς τὸ σπουδαιότερον διὰ τὸ πρὸς ἔκπλῳ αὐτὸν εἶναι οὐ {15} διακαθάραντες, ἀλλὰ πάντα τὰ ὑποπίπτοντα ἡμῖν θέντες ὡς ἔσχατον ἐπελευσόμενοι. ὅθεν καιρὸν νῦν λαβόντες ἀεὶ τὸ τυγχάνον διορθώσεως ἐκδίδομεν. καὶ ἐπεὶ συμβέβηκε καὶ ἄλλους τινὰς τῶν συμμεμιχότων ἡμῖν μετειληφέναι τὸ πρῶτον καὶ τὸ δεύτερον βιβλίον πρὶν {20} ἢ διορθωθῆναι, μὴ θαυμάσῃς, ἐὰν περιπίπτῃς αὐτοῖς ἑτέρως ἔχουσιν.

Apollonius sends his greetings to Eudemus. If you are in good health and if everything else is coming along according to plan, then things must be going quite well for you. As for me, I am doing {26|27} moderately well myself. Back when I was with you in {5} Pergamon, I observed that you were interested in becoming familiar with my treatise on cone sections [= the Conica]. So I have sent you the first volume, having made a diorthōsis of it. As for the remaining volumes, I will send them to you after I am satisfied with the results of my reworking. For I think you will remember having heard from me that I had produced a course [ephodos] {10} on this subject [= cone sections] at the request of Naucrates the expert in geometry back when he came to Alexandria and visited my school, and that, having treated this subject [= cone sections] in eight volumes and having transmitted them from my course [ἐξ αὐτῆς, where αὐτῆς refers to the ephodos] into written form under some pressure, since his ship was scheduled to set sail soon, I did not take the time {15} to clean up [diakathairein] the text thoroughly but simply put things together as they occurred to me, intending to come back to them when the time came. So that is why I now always take the opportunity and make an ekdosis [= make a publication] of whatever happens to be the current state of my diorthōsis. And since it has happened that some others of those who have contact with me had access to the first and the second volumes before {20} they were corrected [diorthoûn], do not be surprised, if you happen upon these volumes, that they are different [from the corresponding volumes of the second ekdosis].

P§78 In this case, it is made explicit that the author of the two ekdoseis intends for the second ekdosis to become a replacement of the first in its authoritativeness. Further, it is implied that the first ekdosis could potentially retain a claim to authority only if the second ekdosis lacked the authorization of the original editor.

P§79 Reflecting on this illustration of the mentality behind the production of a second ekdosis, I return to the citations by Didymus of the two monographs attributed to Ammonius: (1) περὶ τοῦ μὴ γεγονέναι πλείονας ἐκδόσεις τῆς Ἀρισταρχείου διορθώσεως ‘About the fact that there did not exist more [pleiones] ekdoseis of the Aristarchean diorthōsis’ (scholia A for Iliad X 397-399a) and (2) περὶ τῆς ἐπεκδοθείσης διορθώσεως ‘About the epekdosis of the diorthōsis’ (scholia A for Iliad XIX 365-368a). My question is, what would be the motive of Ammonius in denying the existence of a second edition made by Aristarchus himself? In order to formulate an answer, I offer three points for consideration: {27|28}

(1) Didymus, even while distancing himself from Ammonius, acknowledges that this distant predecessor of his had been the immediate successor to Aristarchus as director of the Library of Alexandria (scholia A for Iliad X 397-399a εἴ τι χρὴ πιστεύειν Ἀμμωνίῳ τῷ διαδεξαμένῳ τὴν σχολήν ‘if one is to believe Ammonius, who was the successor to the school [of Aristarchus]’).

(2) As director of the Library about a century after Ammonius, Didymus speaks of two Aristarchean texts of Homer, not one, as a basic working reality, and he refers to the two of them as two ekdoseis.

(3) Ammonius, in the title of the second of the two monographs in question, likewise speaks of the reality of two Aristarchean texts of Homer, not one, but he refers to the second of the two texts in terms of a diorthōsis that underwent an ‘epekdosis’: περὶ τῆς ἐπεκδοθείσης διορθώσεως.

P§80 Earlier, I described the verb ep-ekdidonai of ‘epekdosis’ in terms of a ‘re-editing’. Here I propose to take a closer look at the use of the prefix epi- in ep-ekdidonai. I have found a parallel in the usage of Galen (second century CE), specifically, in a work about his own works, both published and unpublished (De libris propriis liber). In this work, Galen speaks of hupomnēmata of his that he had not initially intended for ekdosis or ‘publication’ and that needed his editorial correction once he decided to go ahead with publication. This process of correction, made necessary because of alleged distortions resulting from unauthorized circulation of Galen’s unpublished hupomnēmata, is designated by the verb ep-anorthoûn and the noun ep-anorthōsis:

Pⓣ4 Galen De libris propriis liber ed. Kühn vol. 19 p. 10 line 4 to p. 12 line 8

φίλοις γὰρ ἢ μαθηταῖς ἐδίδοτο χωρὶς ἐπιγραφῆς ὡς ἂν οὐδὲν πρὸς ἔκδοσιν ἀλλ’ αὐτοῖς ἐκείνοις γεγονότα δεηθεῖσιν ὧν ἤκουσαν ἔχειν ὑπομνήματα. … φωραθέντων δ’ ἁπάντων τῷ χρόνῳ πολλοὶ τῶν αὖθις κτησαμένων ἐπεγράψαντ’ ἐμοῦ τοὔνομα καὶ διαφωνοῦντα τοῖς παρ’ ἄλλοις οὖσιν εὑρόντες ἐκόμισαν πρός με παρακαλέσαντες ἐπανορθώσασθαι. γεγραμμένων οὖν, ὡς ἔφην, οὐ πρὸς ἔκδοσιν αὐτῶν ἀλλὰ κατὰ τὴν τῶν δεηθέντων ἕξιν τε καὶ χρείαν εἰκὸς δήπου τὰ μὲν ἐκτετάσθαι, τὰ δὲ συνεστάλθαι καὶ τὴν ἑρμηνείαν αὐτήν τε τῶν θεωρημάτων τὴν διδασκαλίαν ἢ τελείαν ὑπάρχειν {p. 11} ἢ ἐλλιπῆ. τὰ γοῦν τοῖς εἰρημένοις γεγραμμένα πρόδηλον δήπου μήτε τὸ τέλειον τῆς διδασκαλίας ἔχειν μήτε τὸ διηκριβωμένον, ὡς ἂν οὔτε δεομένων αὐτῶν οὔτε δυναμένων ἀκριβῶς μανθάνειν πάντα, πρὶν {28|29} ἕξιν τινὰ σχεῖν ἐν τοῖς ἀναγκαίοις. … τὰ δ’ οὖν εἰς ἐμὲ κομισθέντα πρός τινων ἐπανορθώσεως ἕνεκεν ἠξίωσα ‘τοῖς εἰσαγομένοις’ ἐπιγεγράφθαι· … {p. 12} … ἐν αὐτῷ γὰρ αἱ κατὰ γένος ἀλλήλων αἱρέσεις διαφέρουσαι διδάσκονται· κατὰ γένος δ’ εἶπον, ἐπειδὴ καὶ διαφοραί τινες ἐν αὐταῖς εἰσι, καθ’ ἃς ὕστερον οἱ εἰσαχθέντες ἐπεκδιδάσκονται.

For my writings were being handed out without an inscribed epigraphē to friends or students not for the purpose of ekdosis but as writings done for those same people because they had asked to have hupomnēmata of the things they had heard about. … From among all the writings of mine that were misappropriated in the course of time, many of the people who re-acquired them would inscribe on them an epigraphē of my name and, upon discovering that these writings did not match the writings that others had, they brought these writings to me and asked me to re-do the anorthōsis [ep-anorthoûn]. So, since these writings had been done by me, as I said, not for an ekdosis [= of these writings] but in accordance with the preparation and needs of those who had asked for them, there was of course the likelihood that some parts were expanded and other parts were compressed and that the interpretive and explanatory power of my theories was left either in an integral or in a {11} deficient state. At any rate, the things that had been written for these aforesaid people are evidently not going to have the required perfection of explanatory power or accuracy [diakriboûn], since these people neither asked to learn everything with accuracy [akribōs] – nor could they if they wanted to before they had some measure of preparation in the basics. … In any case, I thought it fitting that some of the writings brought to me by some of them for the sake of re-doing the anorthōsis [ep-anorthōsis] should be inscribed with the epigraph For beginners. … {12} … For in this [= this introductory book] the haireseis are taught [didaskesthai] as they differ from each other, genos by genos. I just said “genos by genos” because there are differences in them [= in the haireseis], in terms of which at a later point the beginners are thoroughly taught all over again [ep-ekdidaskesthai].

P§81 As we see from the contexts of the verb ep-anorthoûn and of the noun ep-anorthōsis, the correcting that is being done by Galen is a matter of authoritatively ‘re-doing’ something that has already been done before. It is a {29|30} matter of reapplying a standard of exactness – of what is akribes ‘exact’. It was this standard of exactness that the editor had applied in the first place.

P§82 This concept of ep-anorthōsis is comparable to the concept of ‘ep-ekdosis’ as applied by Ammonius in the title of his monograph about the second text of Aristarchus. The term ‘ep-ekdosis’, I propose, implies that the original ekdosis of Homer by Aristarchus retains its claim to its original authority.

P§83 A moment ago, in reflecting on the mentality of Apollonius in describing his second ekdosis of portions of his Conica, I drew attention to his attitude about his first ekdosis. His wording makes it explicit that the second text replaced the first text as the authorized edition. By contrast, the wording of Ammonius implies that the second text – the ‘epekdosis’ – re-establishes the authorization of the editor who produced the first text – the original ekdosis.

P§84 The question remains, how do we reconcile the idea of an earlier ekdosis and a later ‘epekdosis’ of the Homeric text of Aristarchus, as understood by Ammonius in the second century BCE, with the idea of two Aristarchean ekdoseis of this Homeric text, as understood by Didymus in the first century?

P§85 In the case of the two Aristarchean ekdoseis of Homer as known to Didymus, both are treated as authoritative, but one of the two is considered better, khariestera, than the other (scholia for Odyssey iv 727). The situation is different in the case of the Aristarchean ekdosis and ‘epekdosis’ of Homer as known to Ammonius a century earlier. In that case, as we saw, Ammonius prefers to speak of relatively earlier and later phases in the overall editorial work or diorthōsis of Aristarchus, evidently as reflected in relatively earlier and later hupomnēmata. Further, there is only one ekdosis and, by implication, only one authorization by Aristarchus.

P§86 As we saw from the wording of Apollonius of Perga, quoted earlier, ekdosis is in and of itself an act, that is, the act of authorization: ὅθεν καιρὸν νῦν λαβόντες ἀεὶ τὸ τυγχάνον διορθώσεως ἐκδίδομεν ‘so that is why I now always take the opportunity and make an ekdosis [= make a publication] of whatever happens to be the current state of my diorthōsis’. A new ekdosis as made by Apollonius is a new authorization, presupposing an older authorization that can now be superseded.

P§87 In light of these considerations, I return to my earlier question: what would be the motive of Ammonius in denying the existence of a second edition made by Aristarchus? The answer, I propose, has to do with a crisis of transmission. In what follows, I attempt to reconstruct the circumstances of such a crisis. {30|31}

P§88 Let us begin with a simple fact. Whatever terminology we apply – second ekdosis or ‘epekdosis’ –the fact is that the existence of some kind of a second Aristarchean text of Homer was acknowledged not only by Didymus in the first century BCE but also already by Ammonius in the second century BCE. And this second text was a point of contention for Ammonius and his contemporaries.

P§89 In claiming that there was no second ekdosis produced by Aristarchus, Ammonius must have been contending with Aristarchean rivals who claimed that they possessed such a second text. In terms of this rival claim, the first text had been replaced by a second text that represented a new ekdosis by Aristarchus. As we saw from the passage I cited earlier from Apollonius of Perga, a second ekdosis implies the intervention of a living editor who simply replaces an earlier ekdosis with a later one. By contrast, if Ammonius denied the existence of a second ekdosis produced by Aristarchus himself, then it could be that Aristarchus was no longer living at the time when any kind of a second text was produced. In terms of this formulation, the death of Aristarchus led to divergent editions stemming from a single edition that he had once produced. Such an explanation has been attempted more than once in the history of modern scholarship, starting with Villoison (1788) in the prolegomena for his edition of the Iliad. [55] What still needs to be explained, however, is the basic contradiction between the claim of Ammonius about a single ekdosis and the counter-claim implied by the very concept of a second ekdosis. Why should there be two ekdoseis after Aristarchus died if there was only one ekdosis when he was still alive? In what follows, I formulate an explanation of this contradiction.

P§90 Ammonius could have claimed that there was only one ekdosis if he possessed a unique text of an original ekdosis made by Aristarchus himself. Conversely, any Aristarchean rivals of Ammonius could have claimed access to a second ekdosis – if they possessed a text supposedly re-edited by Aristarchus himself. In terms of this formulation, the rivals of Ammonius did not possess a unique original text – even if they claimed to possess a second text that was said to be a second ekdosis supposedly produced by Aristarchus. Ammonius could have rejected such a second text on the grounds that it had not been authorized by Aristarchus, and he could have produced an alternative text. {31|32} Such an alternative text, produced by Ammonius himself, could have been what he recognized as the true ‘epekdosis’ of Aristarchus.

P§91 In this formulation, I have reconstructed the circumstances of a crisis of Homeric transmission in the second century BCE, during the life and times of Ammonius. Even if some details of the reconstruction may be open to question, one detail is for sure: there really was a crisis, since the after-effects are clearly visible from the testimony of Didymus in the first century BCE. Didymus was faced with the reality of a split Aristarchean tradition, that is, with two divergent Aristarchean texts, two divergent Aristarkheioi ekdoseis.

P§92 It does not follow, however, that this reality in the era of Didymus is identical with the corresponding reality of two divergent Aristarchean texts in the era of Ammonius, a century earlier. As I will now argue, we cannot assume that the two Aristarkheioi ekdoseis used by Didymus were the same texts as (1) the ekdosis made by Aristarchus and (2) a later ‘epekdosis’ made by Ammonius on the basis of his immediate predecessor’s ongoing editorial work.

P§93 In the Homeric scholia, we find telling evidence for arguing against such an assumption. There we see references to variant readings found in the two Aristarkheioi ekdoseis used by Didymus that contradict the received readings transmitted in the Koine (I have already noted an example in scholia A for Iliad XII 404a). Earlier, I used the term Aristarchean standard with reference to the prioritizing of variant readings preferred by Aristarchus over the received readings of the Koine. As I argued, this standard applies not to what Aristarchus in his own time showed in his base text. In that earlier time, the base text of his own ekdosis did not privilege the Aristarchean standard but still followed the Koine standard, as we know from the overall testimony of lēmmata transmitted by the Homeric scholia. By contrast, the Aristarchean standard prevailed in the base texts of the two Aristarkheioi ekdoseis used by Didymus.

P§94 What, then, were these two Aristarkheioi ekdoseis used by Didymus? One possible explanation is that they were (1) a derivative of an alternative ekdosis made by Aristarchean rivals of Ammonius and (2) a derivative of the ‘epekdosis’ made by Ammonius. In other words, we may be dealing here not with rival claims about the same second text but rather with two different second texts that competed with each other.

P§95 In order to justify the second text, both sides would have claimed the authorization of Aristarchus himself. If Aristarchus had died before the publication of a second ekdosis, then those who would claim possession of such a second text would have needed to claim also the authorization of this text by Aristarchus himself. Likewise, Ammonius would have needed to claim that {32|33} his own second text, as an ‘epekdosis’ that he himself produced, was a faithful continuation of the editorial work accomplished by Aristarchus. The actual editor of the ‘epekdosis’ would have been Ammonius himself, by virtue of his claim to be the immediate successor of Aristarchus. Similarly, the actual editor of any rival second ekdosis would have been a rival of Ammonius.

P§96 The motive for the terminology of Ammonius could then be explained this way: as the editor of an Aristarchean ‘re-edition’ of Homer, Ammonius would have produced a de facto new ekdosis of Homer, but he chose to describe it instead as a re-edition or ‘epekdosis’, derived directly from the authority of his immediate predecessor, Aristarchus himself. We may compare the fact that Aristarchus had produced a de facto new ekdosis of Homer on the basis of his hupomnēmata linked to the previous ekdosis of Aristophanes. Just as a de facto new ekdosis of Homer as produced by Ammonius would have claimed the authorization of Aristarchus, so also any other second ekdosis would have been based on a similar but rival claim.

P§97 In the last three paragraphs, we have been considering the possibility that the two Aristarkheioi ekdoseis used by Didymus were based on (1) a second ekdosis that supposedly stemmed from Aristarchus and (2) an epekdosis that likewise stemmed from Aristarchus. This possibility has been offered as an alternative to another possibility that we first considered, that the two Aristarkheioi ekdoseis used by Didymus were the same texts as (1) the ekdosis made by Aristarchus and (2) a later ‘epekdosis’ made by Ammonius on the basis of his immediate predecessor’s ongoing editorial work. As we have seen, this other possibility must be rejected, since the evidence produced by the testimony of Didymus himself shows that both of the Aristarkheioi ekdoseis of Homer were in fact post-Aristarchean. That is, neither of these two texts as used by Didymus can be derived directly from the editorial work of Aristarchus himself. [56]

Pⓢ10. The base text of Aristarchus

P§98 I now turn to the question of a first ekdosis of Homer produced by Aristarchus. The reality of such a first text is presupposed, as I have argued, by both of the two rival second texts as they existed in the days of Ammonius. {33|34} Both the ‘epekdosis’ and the second ekdosis, as concepts, presupposed an original ekdosis. The question can be formulated this way: if two rival second texts as known to Ammonius became the ultimate sources of the two Aristarkheioi ekdoseis as used by Didymus, then what became of the first text, and who inherited it?

P§99 In seeking an answer to this question, I start by repeating my point that the two Aristarkheioi ekdoseis used by Didymus stemmed ultimately from an ongoing diorthōsis by Aristarchus – and that this word diorthōsis refers to the procedure of making editorial judgments without specifying the format of presenting these judgments. In other words, the basis for the diorthōsis of Aristarchus was what he said in the hupomnēmata ‘commentaries’, not what he showed in the base text he used as a point of reference for making the judgments recorded in his hupomnēmata. To that extent, even our use of the term base text can distract us from understanding the true basis of Aristarchean diorthōsis.

P§100 As I was arguing earlier, the base text of Aristarchus was based on the converging uniformity of the koinai texts of Homer, not on the diverging multiformity of the khariesterai texts. In other words, the base text used by Aristarchus followed a Koine standard, to be contrasted with the base texts of the two Aristarkheioi ekdoseis used by Didymus, which followed an Aristarchean standard.

P§101 To be contrasted with Didymus is the contemporary Aristarchean scholar Aristonicus. [57] Unlike Didymus, Aristonicus used a base text that followed the Koine standard. This text, I propose, goes back to the first ekdosis of Aristarchus.

P§102 Essential for this proposal is the fact that Aristonicus nowhere distinguishes between two ekdoseis of Aristarchus. One explanation that has been offered for this unitary approach of Aristonicus is that “he did not need to seek out the original manuscript or manuscripts of Aristarchus.” [58] According to this explanation, Aristonicus had no such need because any copy containing the marginal signs that Aristarchus had used for his editing the text of Homer would have been sufficient. [59] So, supposedly, the approach of Aristonicus to manuscripts of Homer should have been more simplistic than that of his {34|35} contemporary Didymus, “who should have pored over them constantly, if they were still extant.” [60] I offer an alternative explanation: that Aristonicus actually had direct access to the definitive base text of Aristarchus and to the definitive hupomnēmata ‘commentaries’ of Aristarchus, while Didymus had only indirect access.

P§103 In support of this explanation, I turn to the cumulative evidence provided by the Homeric scholia concerning the work of both Aristonicus and Didymus. I start with Aristonicus.

P§104 As we see from the scholia, the detailed information drawn by Aristonicus from the Aristarchean hupomnēmata is systematically correlated with his knowledgeable references to marginal signs that he finds in a base text that he equates implicitly but authoritatively with the edition of Aristarchus. The lēmmata that are linked with Aristonicus in the Homeric scholia regularly correspond to the Koine standard that Aristarchus himself had followed for setting up his own base text, not to the Aristarchean standard followed by the two post-Aristarchean ekdoseis used by Didymus in Alexandria (I have already noted an example in scholia A for Iliad XII 404a). [61]

P§105 In making this distinction between Aristonicus and Didymus, I find it relevant that the information derived from Didymus consistently omits references to the marginal signs of Aristarchus. I propose to explain this pattern of omission by arguing that Didymus, unlike Aristonicus, lacked direct access to the definitive base text that Aristarchus had marked up with marginal signs cross-referring to the master’s definitive hupomnēmata ‘commentaries’. I also find it relevant that Didymus was working in Alexandria while Aristonicus had been relocated from Alexandria to Rome (Suda π 3036), along with other Alexandrian scholars like Philoxenus and Seleucus. [62] I propose further that this relocation to Rome involved not only the three Alexandrian scholars themselves but also the definitive base text of Aristarchus, marked with the master’s marginal signs. [63] In the days of Aristonicus and Didymus – that is, in the age of Virgil – the most prized Aristarchean texts were to be found in Rome, not in Alexandria. {35|36}

P§106 Back in Alexandria, Didymus lacked direct access not only to the definitive base text but also to the definitive hupomnēmata of Aristarchus. That text too was being used by Aristonicus in Rome. As we noted before, the lēmmata that are linked with Aristonicus in the Homeric scholia regularly correspond to the Koine standard that Aristarchus himself had followed for setting up his own base text, not to the Aristarchean standard followed by the two post-Aristarchean ekdoseis used by Didymus in Alexandria (as we see in scholia A for Iliad XII 404a).

P§107 In making this negative argument about Didymus, I am not going so far as to say that he had no Aristarchean texts to consult: in fact the Homeric scholia show that Didymus knows of several different versions of Aristarchean hupomnēmata, and he even quotes extensively from these different versions. [64] I argue only that Didymus lacked direct access to the single most definitive version of the Aristarchean hupomnēmata – the version that was directly linked to the definitive base text. In support of this negative argument, I point to one particular situation where Didymus seems to have information from the hupomnēmata of Aristarchus and yet cannot verify what Aristarchus had actually said in these hupomnēmata. In this context, he cites as his only other available sources the two post-Aristarchean ekdoseis and some sungrammata ‘monographs’ of Aristarchus (scholia for Iliad III 406a1). On the basis of such patterns of citation, I infer that Didymus had access only to abridged versions of the final and definitive Aristarchean hupomnēmata. [65] Also, I raise the possibility that the two post-Aristarchean ekdoseis used by Didymus were texts that contained in their margins such abridged versions of the information that Aristarchus had provided in his final and definitive hupomnēmata. [66] If the two post-Aristarchean ekdoseis were formatted to feature in their margins the abridged information extracted from the definitive Aristarchean hupomnēmata, then there would have been no need for marginal signs that cross-refer to separate volumes containing such hupomnēmata. I see {36|37} here a possible explanation for the absence of references to the marginal signs of Aristarchus in the information provided by Didymus.

P§108 Despite its limitations, the work of Didymus preserved a vast repertoire of variants assembled by his predecessor Aristarchus, and the preservation of this repertoire ultimately helped transform the Koine standard of the Homeric base text into the Aristarchean standard as represented by the two post-Aristarchean ekdoseis used by Didymus. The Aristarchean standard, as we saw, tended to privilege readings that Aristarchus judged to be khariestera (hereafter I refer to the word simply in the neuter plural) over readings he judged to be koina (again, neuter plural).

Pⓢ11. The kharis of the “real ” Homer

P§109 Here I return to the Aristarchean criteria for establishing what is really Homeric and what is not. We saw that the Aristarchean term khariestera ‘having more kharis’, that is, ‘more graceful’, is applied to variant readings in the Homeric text that are more likely to be Homeric. By default, the meaning of koina – where the term is contrasted with khariestera – is ‘common, commonplace, ordinary’ – and therefore less likely to be Homeric.

P§110 The comparative form of the term khariestera stems from usage that goes back at least as far as the age of Plato in the fourth century BCE. [67] In Chapter 3, which concerns that era, I will highlight a context where the superlative form of the term, khariestata, implies a sophistic understanding of Homer that transcends and even challenges a purely civic understanding. For now, however, I continue to concentrate on the distinction between khariestera and koina as understood by Aristarchus in the second century BCE and by Aristarcheans like Didymus in the first century BCE.

P§111 For Aristarchus, the term khariestera as opposed to koina was not so much sophistic as it was idealistic – idealistic to the point of impracticability. Despite the implication that the khariestera characterized Homer, what must have happened in real practice was that the Aristarchean base text consistently defaulted to the consensus of the koina – or, notionally, to a singular Koine. So the base text was highlighting in the foreground the kind of wording that was less likely to be Homeric, while the wording of the supposedly real Homer, along with the variant wordings of any number of supposedly false Homers, was kept in the shade, that is, in the background provided by the hupomnēmata ‘commentaries’ of Aristarchus. {37|38}

P§112 In the days of Didymus, as I noted earlier, variants reported by Aristarchus in his hupomnēmata could easily infiltrate the base text, actually ousting the wordings inherited by the Koine. The authority of wordings found in the Koine had given way to the authority of variant wordings preferred by Aristarchus himself – wordings originally confined to his hupomnēmata. For Aristarcheans like Didymus, the preferred readings of Aristarchus were more significant than the received readings of the Koine. This shift is exemplified by what I have been calling the Aristarchean standard.

P§113 Whereas the term khariestera was a relative criterion for Aristarchus himself, it was approaching the status of an absolute criterion for Didymus. For this follower of the Aristarchean standard, khariestera meant not only ‘more Homeric’ while koina meant ‘less Homeric’. More directly, khariestera implied that the wording of Homer possesses kharis, while the wording of non-Homers – of false Homers or pseudo-Homers – must be without kharis. Such a dichotomy may have been implied by Aristarchus himself, though the rhetoric of litotes built into Aristarchean comments such as ouk akharis ‘not without kharis’ (scholia bT for Iliad XVI 313) and ouk akharitōs ‘executed not without kharis’ (scholia A^im for Iliad XX 188b1) suggests a lack of absolute certainty about kharis as a criterion for determining what is Homeric and what is non-Homeric.

P§114 By contrast, we find a sense of absolute certainty in the attitude of Aristarcheans like Didymus. The absolutism carries over even into the comparative degree, as we see in the usage of khariestera as reflected in the Homeric scholia. Besides the relativizing sense of the comparative degree – ‘having more kharis’ as opposed to ‘having less kharis’ – we can see signs of an absolutizing sense, indicating possession of a quality as opposed to not possessing it at all. In other words, the term khariestera can be used to exclude the koina. In this case, the possession involves ‘having kharis’ as opposed to ‘having no kharis’, instead of ‘having more kharis’ as opposed to having less kharis’. [68] As we see in the reportage of the Homeric scholia concerning the judgments of Didymus, hai pleious (scholia A for Iliad XII 382a1) and hai pasai (scholia T for Iliad II 196c2) can be used to mean respectively ‘the majority of’ and ‘all’ manuscripts (graphai) that are khariesterai – not all the versions taken together, which would have included the koinai. Along the same line of thinking, the {38|39} superlative degree, khariestatai, can be used to refer to the aggregate of khariesterai (scholia A for Iliad II 53a1).

P§115 Why is kharis, in the sense of a ‘favor’ achieved through pleasurable beauty, applied as a criterion for determining the authenticity of a given text attributed to a given author? There is no way to answer this question directly, but I find it useful to compare the use of this same word in an essay by a near-contemporary of Didymus, Dionysius of Halicarnassus, entitled On Lysias (sections 10-12). [69] At the summit of a catalogue of objective criteria for judging the authenticity or non-authenticity of texts attributed to Lysias, the author adds, as a pièce de résistance, a subjective criterion that he considers to be more important than all the other criteria he has adduced up to this point: if the given text has kharis, then the author must be Lysias; if it does not, then the author must be someone who has been mistaken for Lysias. In the context of such a critical choice, the criterion of kharis is absolutizing: the text either has or does not have kharis, and therefore the author is either Lysias or pseudo-Lysias:

Pⓣ5 Dionysius of Halicarnassus On Lysias 10

πολλὰ καὶ καλὰ λέγειν ἔχων περὶ τῆς Λυσίου λέξεως, ἣν λαμβάνων καὶ μιμούμενος ἄν τις ἀμείνων γένοιτο τὴν ἑρμηνείαν, τὰ μὲν ἄλλα τοῦ χρόνου στοχαζόμενος ἐάσω, μίαν δὲ ἀρετὴν ἔτι τοῦ ῥήτορος ἀποδείξομαι, κρίνας καλλίστην τε καὶ κυριωτάτην καὶ μόνην αὐτὴν μάλιστα τῶν ἄλλων τὸν Λυσίου χαρακτῆρα δυναμένην βεβαιῶσαι, ἣν ὑπερεβάλετο μὲν οὐδεὶς τῶν ὕστερον, ἐμιμήσαντο δὲ πολλοὶ καὶ παρ’ αὐτὸ τοῦτο κρείττους ἑτέρων ἔδοξαν εἶναι τὴν ἄλλην δύναμιν οὐθὲν διαφέροντες· ὑπὲρ ὧν, ἂν ἐγχωρῇ, κατὰ τὸν οἰκεῖον διαλέξομαι τόπον. τίς δ’ ἐστὶν ἥδε ἡ ἀρετή; ἡ [τις] πᾶσιν ἐπανθοῦσα τοῖς ὀνόμασι κἀπ’ ἴσης χάρις, πρᾶγμα παντὸς κρεῖττον λόγου καὶ θαυμασιώτερον. ῥᾷστον μὲν γάρ ἐστιν ὀφθῆναι καὶ παντὶ ὁμοίως ἰδιώτῃ τε καὶ τεχνίτῃ φανερόν, χαλεπώτατον δὲ λόγῳ δηλωθῆναι καὶ οὐδὲ τοῖς κράτιστα εἰπεῖν δυναμένοις εὔπορον.

I could mention many other fine qualities of Lysias’ style which would improve the expressive powers of anyone who adopted and imitated them. But I shall keep my eye on the time, and confine myself to mentioning one more, which I consider to be his finest and most important quality, and the one above all which enables us {39|40} to establish his peculiar character. None of his successors excelled him in it, but many of those who aspired to it were considered superior to their rivals on the strength of this alone, not because they had greater general ability. But I shall discuss these authors in their proper place, if I have the opportunity. What is this quality? It is his kharis, which blossoms forth in every word he writes, a quality which is beyond description and too wonderful for words. [70] It is very easy and plain for layman and expert alike to see, but to express it in words is very difficult, nor is it easy even for those with exceptional descriptive powers. [71]

Pⓣ6 Dionysius of Halicarnassus On Lysias 11

καὶ ὅταν διαπορῶ περί τινος τῶν ἀναφερομένων εἰς αὐτὸν λόγων καὶ μὴ ῥᾴδιον ᾖ μοι διὰ τῶν ἄλλων σημείων τἀληθὲς εὑρεῖν, ἐπὶ ταύτην καταφεύγω τὴν ἀρετὴν ὡς ἐπὶ ψῆφον ἐσχάτην. ἔπειτα ἂν μὲν αἱ χάριτες αἱ τῆς λέξεως ἐπικοσμεῖν δοκῶσί μοι τὴν γραφήν, τῆς Λυσίου ψυχῆς αὐτὴν τίθεμαι καὶ οὐδὲν ἔτι πορρωτέρῳ ταύτης σκοπεῖν ἀξιῶ. ἐὰν δὲ μηδεμίαν ἡδονὴν μηδὲ ἀφροδίτην ὁ τῆς λέξεως χαρακτὴρ ἔχῃ, δυσωπῶ καὶ ὑποπτεύω μήποτ’ οὐ Λυσίου ὁ λόγος…

Whenever I am uncertain as to the genuineness of any speech that is attributed to him, and find it difficult to arrive at the truth by means of the other available evidence, I resort to this criterion to cast the final vote. Then, if the writing seems to be graced with those additional qualities of kharis, I judge it to be a product of Lysias’s genius, and consider it unnecessary to investigate further. But if the style is devoid of pleasure [hēdonē] and sensuality [aphroditē], I view the speech with a jaundiced and suspicious eye, and conclude that it could never be by Lysias. [72]

P§116 I return to the criterion of khariestera as applied by Didymus to Homer. If we compare the criterion of kharis as applied by Dionysius of Halicarnassus to Lysias, we may be justified to extrapolate at least this much: Didymus seems to be saying that whatever Homeric variant is judged to ‘have more kharis’ is more likely to be really Homeric than whatever variant has less {40|41} kharis. To say too much more than that, however, is to press the comparison too far. After all, we cannot really be certain whether Didymus considered his criterion to be subjective – as was the criterion of Dionysius – or objective. Moreover, despite the sense of certainty and even absolutism that we may detect in some contexts where the criterion of Didymus is applied, other contexts show traces of relativism and even uncertainty about which variant is genuine and which one is false, what is Homeric and what is pseudo-Homeric.

P§117 In fact, from the standpoint of Aristarchus as a collector of Homeric textual variants, the criterion of khariestera must have been relative, not absolute. Homeric variants judged to be khariestera were relatively rare but mostly multiform whenever they did occur, whereas variants judged to be koina were relatively commonplace but at least mostly uniform. In other words, attestations of readings judged to be khariestera tended to involve more than one variant, allowing for more options in choosing the supposedly genuine variant, whereas attestations of readings judged to be koina allowed for fewer options – or for no option at all. Consequently, the idea of koinai as an aggregate of ‘common’ manuscripts sharing ‘common’ features must have reinforced the inference that they stemmed from a ‘common’ source. For Aristarchus, that source was in theory the Koine and in practice the base text of Homer. By contrast, the only thing that the khariesterai as an aggregate of special manuscripts would have had in common was their divergence from what was common. Thus the actual choosing of readings taken from texts that were khariesterai over readings taken from texts that were koinai involved an element of uncertainty for the editor: the more choices he had to make, the more likely it became that he could make a mistake in trying to isolate the supposedly single genuine variant from among a number of false variants. Choosing among khariesterai was a matter of relative uncertainty, whereas not choosing and staying with the consensus of the koinai was a matter of relative certainty.

P§118 As a rule, Aristarchus chose relative certainty. Here I need to stress again what I said earlier about the diorthōsis or editorial procedure of Aristarchus. As a rule, he would show in his base text the variant that reflected what he understood to be the koinē or ‘standard’ version of Homer – even if he occasionally preferred other variants as featured in the khariesterai. In his base text, he practiced ‘correction’ only to the extent of regularizing a single Koine version as he reconstructed it from an aggregate of koinai. As I have noted, he confined to his hupomnēmata ‘commentaries’ his information about variants judged to be khariesterai. The Koine version remained his base text. {41|42}

P§119 As I have also noted, Aristarchus allowed the external evidence of a strong manuscript attestation in koinai or ‘common’ Homeric texts to override the internal evidence – as he saw it – of Homeric diction. Aristarchus preferred the relative certainty of the Koine standard over the relative uncertainty of having to decide among the variants he assembled.

P§120 By contrast with Aristarchus, who preferred to base his editorial work on the relative certainty of the Koine standard, later generations of Aristarcheans preferred the standard of Aristarchus himself – as they understood it. The Aristarchean standard of later Aristarcheans like Didymus became a new certainty that tended to absolutize the master’s relativistic editorial criteria. The boundaries between Homeric and pseudo-Homeric forms could now be drawn with a sense of certainty. And yet, this certainty proved to be an illusion, based as it was on the relative uncertainty of Aristarchus himself whenever it came to deciding among the variants he assembled. The Aristarchean standard was itself a grand uncertainty, inviting ever new redrawings of boundaries for distinguishing the real Homer from the amorphous mass of competing pseudo-Homers.

Pⓢ12. The Koine standard reaffirmed

P§121 In the practical world of editing Homer, the base text had to remain the point of reference for Aristarchus as editor. To establish this base text as correctly as possible, the editor needed to have access to a representative aggregate of koinai or ‘common’ texts of Homer, derived from an authoritative but relatively ‘uncorrected’ textual source. The notion that many koinai presuppose a singular koinē is reflected even in the fluctuations that we occasionally see between plural and singular uses in our sources. For example, mentions of the plural koinai in the medieval texts of the Homeric scholia are occasionally matched by mentions of a singular koinē in annotations written in papyrus texts of Homer. [73] Such matches have led one expert to draw attention to “the severely reduced nature of the scholia.” [74] I should point out, however, that singular koinē is also attested in the medieval Homeric scholia (as for example in scholia T for Iliad V 461b, scholia A for XII 404a1). Such usages of the singular, relatively more common in the papyri and less common in the medieval scholia, can I think be traced back to the age of Aristarchus. This is not to say that Aristarchus posited a single surviving authoritative text of {42|43} Homer. It is only to say that his base text represented a single uniform koinē as reconstructed by ‘correcting’ a number of multiform koinai. [75] To say it another way, Aristarchus produced a standard text from the multiform koinai, and such a ‘standard’ was something he could quite easily have called the koinē.

P§122 According to an explanation that differs from the one offered here, the mentions of plural koinai in the Homeric scholia refer simply to “the early Ptolemaic papyri that we may see as specimens of the ‘common’ text(s).” [76] This is to assume, however, that plural koinai and the more distinctive singular koinē meant ‘common’ only in the eroded sense of ‘vulgar’. It is also to assume that koinē is merely a foil, an inferior copy. As I have been arguing, however, the koinai provide an authoritative point of departure for the process of scholarly diorthōsis ‘corrective editing’ that ostensibly leads to the edition of a superior text.

P§123 The authenticity of koinē readings, where the designation koinē is actually made explicit by the scholia, can be confirmed on the basis of two independent criteria: (1) comparative linguistics and (2) oral poetics. [77] But this is not to discredit the authenticity of non-koinē readings that the scholia attribute to the diorthōsis or ‘corrective editing’ of scholars like Aristarchus: in many instances, the variant readings attributed to Aristarchus or Aristophanes or Zenodotus can likewise be confirmed on the basis of those same two independent criteria of comparative linguistics and oral poetics. [78] Thus it is unjustified to assume, as some have done, that the variant readings resulting from the diorthōsis ‘corrective editing’ of Alexandrian critics are as a rule scholarly conjectures. Many of these variants stemming from the learned editions prove to be just as authentic, from the standpoint of oral poetics, as the variants stemming from the koinē texts in general. [79]

Pⓢ13. Reading out loud the Homeric Koine

P§124 The Koine standard was not only the textual basis for the diorthōsis ‘correction’ of Homer by Aristarchus. It was also the performative basis, as expressed by the concept of hē koinē anagnōsis ‘the Koine reading’. In the Homeric scholia, we find examples where this expression refers to a standard {43|44} way of reading out loud the Homeric verses. As we are about to see, what was being read out loud from the textual transmission in the days of Aristarchus reflects linguistic patterns that derive from performative transmission.

P§125 I start with Herodian, an Aristarchean scholar who flourished in the era of Marcus Aurelius (emperor from 161 to 180 CE). Most of the extant information reported by Aristarchus about traditional patterns of pronunciation in Homeric performance is mediated through Herodian, and the accuracy of this information, stemming ultimately from descriptive observations collected by Aristarchus himself, can be validated independently through comparative linguistics. [80] In the descriptions reported by Herodian, as mediated in the Homeric scholia, the expression hē koinē anagnōsis ‘the Koine reading’ refers to the way in which Aristarchus himself read Homeric poetry out loud:

Pⓣ7 Scholia A for Iliad II 662a1 (Herodian)

{φίλον μήτρωα} κατέκτα: Πτολεμαῖός φησιν ὁ Ἀσκαλωνίτης ᾿Αρίσταρχον ἀνεγνωκέναι ὁμοίως τῷ ἔκτα σὺν οὐλομένῃ ἀλόχῳ κατὰ συστολήν. Τυραννίων δὲ κατ’ ἔκτασιν. οἶμαι δὲ ἀκόλουθον εἶναι ἐκείνῃ τῇ γραφῇ τῇ κομιζομένῃ ὑπ’ Ἀριστάρχου ὣς ἔμεν <ὡς> ὅτε δῖον Ἐρευθαλίωνα κατέκτα. [81] ἡ μέντοι κοινὴ ἀνάγνωσις ἡ κατὰ συστολὴν ἀφορμὴν ἔσχε τὴν τῆς ἀποκοπῆς, ὁμοίως τῷ οὖτα κατὰ λαπάρην.

{φίλον μήτρωα} κατέκτα: Ptolemaios Askalonites [p. 42 B.] says that Aristarchus has read it [= the α of κατέκτα] in a way that parallels the expression ἔκτα σὺν οὐλομένῃ ἀλόχῳ [Odyssey xi 410], making the vowel [= the α of ἔκτα] short. Tyrannion [F 10 P] reads it as a long. But I think there is an analogy with the reading as preserved by Aristarchus: ὣς ἔμεν <ὡς> ὅτε δῖον Ἐρευθαλίωνα κατέκτα [Iliad IV 319]. The Koine reading [hē koinē anagnōsis], with a shortening of the vowel, had its point of departure from apocope, as also in the case of οὖτα κατὰ λαπάρην [Iliad VI 64].

P§126 According to Herodian, hē koinē anagnōsis ‘the Koine reading’ in this case matches what Aristarchus himself had ‘read’ (Ἀρίσταρχον ἀνεγνωκέναι), pronouncing the α of κατέκτα in the verse-final phrase φίλον μήτρωα κατέκτα {44|45} at Iliad II 662 as short rather than long. So this example makes it explicit that such criteria as hē koinē anagnōsis ‘the Koine reading’ stem ultimately not from Herodian but from Aristarchus himself: in other words, Herodian is simply reporting on descriptive facts already collected by Aristarchus. It is important to distinguish these facts of description from the theories of causation adduced by Herodian in his own right. Attempting to explain why the Koine reading features the pronouncing of short rather than long α in κατέκτα at Iliad II 662, he claims that the cause is an ‘apocope’ of -ε in a hypothetical form that would look like *κατέκταε, just as οὖτα supposedly results from an ‘apocope’ of -ε in a hypothetical form that would look like *οὔταε. In terms of comparative linguistics, such a scenario of causation is inadequate, even wrong: the short vowel α of κατέκτα is in fact a formulaic reshaping of morphologically archaic forms of the root κτα-, where the short α is an archaism, not an innovation; [82] similarly, the short vowel α of οὖτα results from a formulaic reshaping of morphologically archaic forms of οὔτα- featuring short α. [83] Still, the actual attestation of the pattern of pronouncing a short vowel α in κατέκτα as reported by Herodian is valuable in and of itself. It indicates that there did in fact exist such a pattern of pronunciation in Homeric poetry – as read aloud by Aristarchus himself. This attestation is all the more valuable because the question of deciding between short or long -α in verse-final κατέκτα does not affect the content or the interpretation of the given verse. The question here is purely a matter of pronunciation, not meaning.

P§127 Another such example of hē koinē anagnōsis ‘the Koine reading’ involves the sequence ΕΠΕΙΣΕ at Iliad VI 355. Herodian reports that ΕΠΕΙΣΕ is ‘nowadays’ read as ἐπεὶ σέ, featuring the non-enclitic pronoun σέ instead of the enclitic σε. He goes on to add, however, that ‘the Koine reading’ (ἡ κοινὴ ἀνάγνωσις) consistently ‘reads’ (ἀνέγνω) ΕΠΕΙΣΕ as ἐπεί σε, and he proceeds to adduce several Homeric passages to back up his statement:

Pⓣ8 Scholia A for Iliad VI 355a1 (Herodian)

τὴν δὲ σέ ἀντωνυμίαν ὀξυτονοῦσι, τουτέστιν ὀρθοτονοῦσιν, ἐπεὶ πρός τί ἐστιν. ἔστι μὲν οὖν ἀληθὲς ὅτι ἀντιδιασταλτική ἐστιν νῦν ἡ ἀντωνυμία· ἡ μέντοι κοινὴ ἀνάγνωσις ἀνέγνω ἐγκλιτικῶς ἀεὶ τὴν τοιαύτην σύνταξιν. ὃ δὲ λέγω, τοιοῦτόν ἐστι· τὸ ἐπεί σε εὑρέθη συνεχῶς οὕτως ἀνεγνωσμένον, ἐγκλιτικῶς ἀεί, μὴ ἐπιφερομένου συνδέσμου, ἐπεί σ’ εἴασεν Ἀχιλλεύς, ἐπεί σε πρῶτα κιχάνω, ἐπεί σε {45|46} φυγὼν ἱκέτευσα, ἐπεί σε λέοντα. οὕτως δὲ καὶ ἐπεί σε μάλιστα πόνος φρένας.

The pronoun σε is pronounced as oxytone, that is, with an acute accentuation, since it is emphatic. In any case, it is true that the pronoun is nowadays non-enclitic. But the Koine reading [hē koinē anagnōsis] always reads [anagignōskein] [84] this kind of syntax by way of enclitic accentuation. Here are some examples of what I am saying: it is found [85] that the combination ἐπεί σε is read this way consistently, always with enclitic accentuation, if there is no linking word that follows: ἐπεί σ' εἴασεν Ἀχιλλεύς [Iliad XXIV 684], ἐπεί σε πρῶτα κιχάνω [Odyssey xiii 228], ἐπεί σε φυγὼν ἱκέτευσα [Odyssey xv 277], ἐπεί σε λέοντα [Iliad XXI 483]. So also here: ἐπεί σε μάλιστα πόνος φρένας.

P§128 In the ensuing text of the scholia, the examples of ἐπεί σε are reinforced by further Homeric examples of ἐπεί με, featuring the first person enclitic με instead of the emphatic ἐμέ. As Herodian argues, those who ‘read in this way’ (οἱ οὕτως ἀνεγνωκότες) ΕΠΕΙΣΕ as ἐπεί σε at VI 355 are following a pattern:

Pⓣ9 Scholia A for Iliad VI 355a1 (Herodian, continued)

καί μοι δοκοῦσι τῷ πρώτῳ προσώπῳ ἀκολουθεῖν οἱ οὕτως ἀνεγνωκότες, πιθανῶς πάνυ· διὰ γὰρ τῆς φωνῆς τὸ πρῶτον πρόσωπον ἐπιδείκνυται τό τε ὀρθοτονούμενον καὶ τὸ ἐγκλιτικόν, εἴ γε ἡ ἐμέ αἰτιατική, ὅτε φυλάσσει τὸ ε, ὀρθοτονεῖται, εἰ δὲ ἀποβάλοι, ἐγκλιτική ἐστιν. εὑρέθη τοίνυν μετὰ τοῦ ἐπεί συνδέσμου παρὰ τῷ ποιητῇ κατὰ ταύτην τὴν σύνταξιν ἀποβάλλουσα τὸ ε· Ἕκτορ, ἐπεί με κατ’ αἶσαν, ἐπεί μ’ ἀφέλεσθέ γε δόντες. τούτῳ τοίνυν τῷ λόγῳ πιθανὸν ἂν εἴη κατακολουθήσαντας ἡμᾶς ἀναγινώσκειν ἐγκλιτικῶς ἐπεί σε μάλιστα.

I think that those who have read it this way are following the pattern of the first person, and reliably so. The first person – both the accented form and the enclitic form, is made manifest through the word itself. For if the word ἐμέ is accusative, when it keeps the ε it is {46|47} accented with an acute. But if it loses the ε, it is enclitic. In the poet [= Homer], this word is found [86] combined with the conjunction ἐπεί in this syntactical pattern, and it loses the ε: Ἕκτορ, ἐπεί με κατ’ αἶσαν [Iliad III 59], ἐπεί μ’ ἀφέλεσθέ γε δόντες [Iliad I 299]. Following this rationale, it would be reliable for us to read, with enclitic accentuation, ἐπεί σε μάλιστα.

P§129 In these examples of ἐπεί με, the form itself guarantees the enclitic construction, whereas in the examples of ἐπεί σε it was only ‘the Koine reading’ that justified the enclitic pronunciation – as opposed to the emphatic pronunciation corresponding to the way readers ‘nowadays’ pronounce the sequence ΕΠΕΙΣΕ on the basis of syntactical considerations alone. Herodian explains the Koine reading in terms of analogy: those who read enclitic σε instead of σέ in contexts involving the second person are doing so by analogy with the reading of με instead of ἐμέ in analogous contexts involving the first person. Once again, Herodian is explaining why the Koine reading is the way it is. In this case, he is attempting to explain why ἐπεί σε follows the distinct accentual pattern that he describes. In terms of comparative linguistics, the explanation is once again inadequate, even wrong. Still, the actual attestation of the pattern reported by Herodian is once again essential. The attestation shows that there did exist a pattern of pronouncing ἐπεί σε in Homeric poetry as it was read aloud, and that this pattern of reading ἐπεί σε was distinct from the pattern of pronunciation that was current in the Greek language as spoken ‘nowadays’, ἐπεὶ σέ. By implication, the reader whose Koine reading is reported here by Herodian is once again Aristarchus himself.

P§130 In an abridged version of Herodian’s testimony about the Koine reading ἐπεί σε at Iliad VI 355, the formulation is worded as follows:

Pⓣ10 Scholia b(BCE³)T for Iliad VI 355a2

<ἐπεί σε μάλιστα:> ἔστι μὲν νῦν ἀντιδιασταλτικὴ ἡ σέ ἀντωνυμία, καὶ ἐχρῆν αὐτὴν ὀρθοτονεῖσθαι. ἡ δὲ συνήθεια ἐγκλιτικῶς ἀνέγνω.

<ἐπεί σε μάλιστα:> The pronoun σέ is nowadays distinctive [= non-enclitic, as opposed to non-distinctive or enclitic]. And it should have been accented with an acute. But the customary way [sunētheia] reads [87] it as an enclitic. {47|48}

P§131 Here in the scholia b at Iliad VI 355a2, hē sunētheia ‘the customary way’ functions as the subject of the verb anegnō ‘reads’. I draw attention to the syntactical parallel with hē koinē anagnōsis ‘the Koine reading’, which functions as the subject of the same verb anegnō ‘reads’ in the scholia A for Iliad VI 355a1. As we are about to see, the methodology conveyed by the descriptive term sunēthēs ‘customary’ and its derivatives is relevant to the concept of hē koinē anagnōsis ‘the Koine reading’. In order to show this relevance, I start with the related concept of koinē khrēsis ‘Koine usage’:

Pⓣ11 Scholia b(BCE³E⁴)T^il for Iliad II 135b (exegetical scholia)

<καὶ δὴ δοῦρα σέσηπε νεῶν καὶ σπάρτα λέλυνται:> ἐν ἑνὶ στίχῳ ἔθηκε τὴν Ἀτθίδα καὶ κοινὴν χρῆσιν.

<καὶ δὴ δοῦρα σέσηπε νεῶν καὶ σπάρτα λέλυνται:> In the same single verse he [= Homer] placed the Attic and the Koine usage [khrēsis].

P§132 We see in this description a contrast between τὴν Ἀτθίδα (χρῆσιν) ‘Attic usage’ and (τὴν) κοινὴν χρῆσιν ‘Koine usage’. The point that is being made about this particular verse, Iliad II 135, is that it contains the ‘Attic’ usage of a singular verb with neuter plural subject as well as the ‘Koine’ usage of a plural verb with neuter plural subject. Moreover, it is claimed that ‘Homer’ was capable of both ‘Koine’ and ‘Attic’ usage, which happen to coexist in this Homeric verse.

P§133 With reference to this same verse, Aristarchus applies the term sunēthēs ‘customary’:

Pⓣ12 Scholia A^im for Iliad II 135a (Aristonicus)

<καὶ δὴ δοῦρα σέσηπε νεῶν καὶ σπάρτα λέλυνται:> ὅτι κατὰ τὸν αὐτὸν στίχον καὶ ἑαυτῷ καὶ ἡμῖν συνήθως ἐξενήνοχε τὸ λέλυνται καὶ σέσηπε.

<καὶ δὴ δοῦρα σέσηπε νεῶν καὶ σπάρτα λέλυνται:> [Aristarchus marks this verse in the margin] because he [= Homer] has produced in the same verse, as is customary [sunēthōs] both for him [= Homer] and for us, the λέλυνται and σέσηπε. [88] {48|49}

P§134 The comment that I just quoted is ostensibly saying the same thing that was said in the comment previous to this one. In that previous comment, the ‘Attic’ usage was being contrasted with the ‘Koine’ usage, and it was claimed that ‘Homer’ could accommodate both the ‘Attic’ and the ‘Koine’ usages. In the comment I just quoted, there is a similar contrast: what is spoken in the time of Homer sunēthōs heautōi ‘as is customary for him’ is supposedly both ‘Attic’ and ‘Koine’. In other words, Homer can accommodate both the ‘Attic’ and the ‘Koine’ usages. As for Greek usage in the time of Aristarchus, what is spoken sunēthōs hēmīn ‘as is customary for us’ is implicitly ‘Koine’, as we will see.

P§135 There is an essential clarification to be added to this description stemming from the ancient grammarians. Applying the methods of comparative linguistics, we can ascertain that the usage of a singular verb with neuter plural subject is the older syntactical pattern, while the substitution of a plural verb is an innovation; moreover, there is nothing intrinsically ‘Attic’ about the older pattern from the standpoint of Homeric grammar. The pattern seems distinctly ‘Attic’ only from the hindsight of Aristarchus, in whose own time the Attic literary heritage provided the primary evidence.

P§136 The application of the adverb sunēthōs ‘as is customary’ to ‘Koine’ as used in the time of Homer is comparable to the application of the adverb koinōs ‘in a way that is Koine’ in the following context:

Pⓣ13 Scholia H for Odyssey vii 132

τοῖ’ ἄρ’: τόσα ἔσαν, ὡς σπάρτα λέλυνται κοινῶς, τὸ δὲ Ἀττικὸν τοῖα

τοῖ’ ἄρ’: τόσα ἔσαν, as in the expression σπάρτα λέλυνται, in a way that is common [koinōs], while the Attic form [= as opposed to τόσα] is τοῖα.

P§137 The example that is being cited here by the commentator to illustrate ‘Koine’ usage, σπάρτα λέλυνται, comes from the same passage that we considered a moment ago (Iliad II 135). We saw from the scholia (at Iliad II 135b) that such a combination of plural verb with neuter plural subject is considered to be ‘Koine’ as opposed to ‘Attic’ usage. Here at Odyssey vii (132), the commentator draws attention not only to this ‘Koine’ usage but also to the ‘Attic’ usage represented by τοῖα. What is being highlighted is the coexistence of ‘Koine’ and ‘Attic’ usages within the same verse. As we saw earlier, the commentary at Iliad II (135) highlights a coexistence within that other verse between the ‘Koine’ syntax of a plural verb featuring a neuter plural subject and the ‘Attic’ {49|50} syntax of a singular verb featuring a neuter plural subject. The wording of Eustathius confirms this highlighting:

Pⓣ14 Eustathius Commentary 1.293.12-14 on Iliad II 135

Τὸ δέ δοῦρα σέσηπεν ᾿Αττικόν· Ἀθηναῖοι γὰρ πληθυντικοῖς οὐδετέροις ἑνικὰ ἐπάγουσι ῥήματα. τὸ μέντοι σπάρτα λέλυνται κοινόν. διὸ καί φασιν οἱ παλαιοί, ὅτι ἑνὶ στίχῳ ἔθηκε τὴν ᾿Ατθίδα καὶ τὴν κοινὴν χρῆσιν.

The expression δοῦρα σέσηπεν is Attic. For Athenians produce singular verbs with plural neuters. But the expression σπάρτα λέλυνται is the one that is common [koinon = the Koine usage]. That is why the ancients as well say that he [= Homer] put inside one single verse the Attic and the Koine usage [khrēsis].

P§138 In what we have seen so far, then, the term sunēthēs ‘customary’ refers to Greek usage not only in the time of Aristarchus but also in the time of Homer – that is, in the time of Homer as reconstructed by Aristarchus. The term is descriptive, not prescriptive, as far as Aristarchus is concerned. On the level of internal analysis, he is describing what is ‘customary’ in Homer by studying the language of the text as a system. On the level of comparative analysis, he is describing what is ‘customary’ in the Greek usage of his own time in order to shed light on what he understands to be the combined usage of ‘Koine’ and ‘Attic’ in the time of Homer.

P§139 A moment ago, we saw the use of the adverb sunēthōs ‘customary’ with the dative of the first person plural to indicate contemporary usage: καὶ ἑαυτῷ καὶ ἡμῖν συνήθως ‘in a way that is customary [sunēthōs] both for him [= Homer, in the past] and for us [in the present]’ (scholia A^im at Iliad II 135a). In general, the use of the adjective sunēthēs ‘customary’ with the dative of the first person plural indicates contemporary usage. Here is another example:

Pⓣ15 Scholia A^int for Iliad X 461c (Aristonicus)

<καὶ εὐχόμενος ἔπος ηὔδα:> ὅτι συνήθως ἡμῖν νῦν κέχρηται τῷ εὐχόμενος

<καὶ εὐχόμενος ἔπος ηὔδα:> [Aristarchus marks this verse in the margin] because he [= Homer] uses this form εὐχόμενος in accordance with the way that is customary [sunēthōs] for us nowadays. {50|51}

P§140 Besides such examples where contemporary usage happens to match ‘Homeric’ usage, Aristarchus also finds examples where ‘Homeric’ usage is clearly distinct from whatever is sunēthēs ‘customary’ for ‘us’:

Pⓣ16 Scholia A for Iliad II 807 (Aristonicus)

ἠγνοίησεν: … οὐ κεῖται δὲ συνήθως ἡμῖν τὸ ἠγνοίησεν ἀλλ’ ἀντὶ τοῦ οὐκ ἀπίθησεν

ἠγνοίησεν: … This form ἠγνοίησεν is attested here not in the way that is customary [= sunēthōs] for us but in the sense of an alternative way of saying οὐκ ἀπίθησεν.

Pⓣ17 Scholia A for Iliad III 206a (Aristonicus)

σεῦ ἕνεκ’ <ἀγγελίης>: ὅτι Ζηνόδοτος γράφει σῆς ἕνεκ’ ἀγγελίης. οὐ λέγει δὲ συνήθως ἡμῖν, τῆς σῆς ἀγγελίας χάριν, ἀλλ’ ἀγγελίης ἀντὶ τοῦ ἄγγελος.

σεῦ ἕνεκ’ <ἀγγελίης>: [Aristarchus marks this verse in the margin] because Zenodotus writes σῆς ἕνεκ’ ἀγγελίης. But he [= Homer] does not say it in the customary way [sunēthōs] as we say it, that is, ‘for the sake of your angeliē’, but, rather, he says it as an alternative to angelos [= nominative angeliēs]. [89]

Pⓣ18 Scholia A for Iliad III 99a

Ἀργείους καὶ Τρώας [sic:] ὅτι Ζηνόδοτος γράφει Ἀργεῖοι καὶ Τρῶες, ὡς ἀποστροφῆς τοῦ λόγου γεγονυίας πρὸς αὐτούς. ἔστι δὲ τὸ διακρινθῆναι διχῶς χωρισθῆναι. ὁ δὲ Ζηνόδοτος συνήθως ἡμῖν τέταχεν.

Ἀργείους καὶ Τρώας [sic:] [Aristarchus marks this verse in the margin] because Zenodotus writes Ἀργεῖοι καὶ Τρῶες, as if there had {51|52} taken place a direct address to them [= Argives and Trojans] in the discourse. The form διακρινθῆναι [= διακρινθήμεναι at Iliad III 98] is in the sense of ‘be divided’. Zenodotus has arranged the wording [= the vocative instead of the accusative] in the way that is customary [sunēthōs] for us.

P§141 In all three of these examples, the ‘Homeric’ usage fails to match the ‘customary’ usage of Greek as spoken in the time of Aristarchus. In the last two examples, we see Aristarchus arguing in favor of textual variants that correspond to the ‘Koine’ usage and against variants that correspond to current usage in the time of Aristarchus. In these examples, the variants found in the text of Zenodotus happen to be non-‘Koine’: as we see from the lēmmata, the supposedly ‘Homeric’ usage is represented by the ‘Koine’ reading, to be contrasted with readings that correspond to current usage.

P§142 For Aristarchus, the ‘Homeric’ reading is generally hē koinē anagnōsis ‘the Koine reading’ by default, as we see from this example:

Pⓣ19 Scholia A for Iliad I 465b1 (Herodian)

ἄρα τ’ ἄλλα: Πτολεμαῖος ὡς τἆργα τἆλλα. ἡ μέντοι κοινὴ ἀνάγνωσις παραπληρωματικὸν ἔλαβε τὸν τέ σύνδεσμον, ἐν ἐκείνῳ δὲ συνεσταλμένον τὸ α ἐφύλαξεν, καθότι ἤδη ἔθος ἐστὶ τῷ ποιητῇ ἐλλείπειν τοῖς ἄρθροις.

ἄρα τ’ ἄλλα: Ptolemaios [Askalonites, p. 41] reads τἆλλα, like τἆργα. But the Koine reading [hē koinē anagnōsis] has the conjunction τέ as a filler construction, keeping the α short within that construction. This is in accordance with the Poet’s [= Homer’s] custom [ethos] of leaving out the articles.

Pⓣ20 Scholia b(BE³E⁴)T for Iliad I 465b2

τὸ δὲ τἆλλα κατὰ συναλιφὴν ὡς τἆργα·

The form τἆλλα is by way of contraction, like τἆργα, …

Pⓣ21 Scholia b(BE³)T for Iliad I 465b2

ὁ δὲ Ἡρωδιανὸς τὸν τέ παραπληρωματικὸν ἀποδέχεται, καὶ λείπει τὸ ἄρθρον Ὁμηρικῷ ἔθει. {52|53}

… but Herodian [2.28.16] accepts [apodekhetai] the filler τέ and leaves out the article, in accordance with Homer’s custom [ethos].

P§143 The expression apodekhetai (ἀποδέχεται) ‘accepts’ in the last example is an indirect indication that Herodian here is following a procedure followed by Aristarchus himself. At a later point, we will encounter this same expression in a context where Aristarchus apodekhetai (ἀποδέχεται) ‘accepts’ as genuine a variant reading of a Homeric verse that is being literally read out loud to him by an anagnōstēs ‘reader’ (scholia A for Iliad XVII 75a). [90]

P§144 This stance of Herodian in tending to favor the ‘Koine’ usage is parallel to the stance of Aristarchus himself when he is faced with an actual choice between textual variants that he would describe as ‘Koine’ and ‘Attic’. In such situations, he tends to favor the ‘Koine’ variant as more ‘Homeric’ than the ‘Attic’ variant. In the scholia A for Iliad II 397b, for example, Didymus reports that both of the two Aristarchean ‘editions’ (hai Aristarkhou) show the plural rather than the singular verb in combination with the neuter plural subject, and that Aristarchus in his hupomnēmata prefers the variant showing the plural over the variant showing the singular. In the scholia A for Iliad II 397a, Aristonicus reports that a given variant showing the plural was Ὁμηρικώτερον ‘more Homeric’ for Aristarchus than another given variant showing the singular. Similarly in the scholia A for Iliad XIII 28b, with reference to another case where Aristarchus prefers the plural over the singular, Aristonicus reports: σύνηθες γὰρ Ὁμήρῳ οὕτως λέγειν ‘for it was customary [sunēthes] for Homer to say it this way’. [91]

P§145 The last example, where we see the concept of sunēthes ‘customary’ being equated with the concept of ‘Homeric’, brings me back to the concept of hē sunētheia ‘the customary way’, which as we saw is equated with the concept of hē koinē anagnōsis ‘the common reading’ with reference to ‘Homeric’ usage.

P§146 In the reportage of Aristarchus, as we saw from the examples I have just surveyed, the term sunēthēs ‘customary’ refers to linguistic usage as described by an ostensibly objective observer. An indication of the degree of objectivity sought by Aristarchus is the fact that his descriptions of Homer’s {53|54} own customary usage or sunētheia are regularly tested against descriptions of usage current in his own time. Such an ideal of objectivity is also what drives the methodology of the Aristarchean scholar Herodian. For him as well, the usage of ‘Homer’ needs to be compared to the usage that is current. For a telling example, I quote again the passage I quoted at the very beginning of this section:

Pⓣ22 Scholia A for Iliad VI 355a1 (Herodian)

τὴν δὲ σέ ἀντωνυμίαν ὀξυτονοῦσι, τουτέστιν ὀρθοτονοῦσιν, ἐπεὶ πρός τί ἐστιν. ἔστι μὲν οὖν ἀληθὲς ὅτι ἀντιδιασταλτική ἐστιν νῦν ἡ ἀντωνυμία· ἡ μέντοι κοινὴ ἀνάγνωσις ἀνέγνω ἐγκλιτικῶς ἀεὶ τὴν τοιαύτην σύνταξιν. ὃ δὲ λέγω, τοιοῦτόν ἐστι· τὸ ἐπεί σε εὑρέθη συνεχῶς οὕτως ἀνεγνωσμένον, ἐγκλιτικῶς ἀεί, μὴ ἐπιφερομένου συνδέσμου, ἐπεί σ’ εἴασεν Ἀχιλλεύς, ἐπεί σε πρῶτα κιχάνω, ἐπεί σε φυγὼν ἱκέτευσα, ἐπεί σε λέοντα. οὕτως δὲ καὶ ἐπεί σε μάλιστα πόνος φρένας.

The pronoun σε is pronounced as oxytone, that is, with an acute accentuation, since it is emphatic. In any case, it is true that the pronoun is nowadays non-enclitic. But the Koine reading [hē koinē anagnōsis] always reads [anagignōskein] this kind of syntax by way of enclitic accentuation. Here are some examples of what I am saying: it is found that the combination ἐπεί σε is read this way consistently, always with enclitic accentuation, if there is no linking word that follows: ἐπεί σ’ εἴασεν Ἀχιλλεύς [Iliad XXIV 684], ἐπεί σε πρῶτα κιχάνω [Odyssey xiii 228], ἐπεί σε φυγὼν ἱκέτευσα [Odyssey xv 277], ἐπεί σε λέοντα [Iliad XXI 483]. So also here: ἐπεί σε μάλιστα πόνος φρένας.

P§147 In order to make the point that current Greek usage does not match ‘Homeric’ usage in this particular case, Herodian cites examples of ‘Homeric’ usage as collected by Aristarchus. These examples show most clearly that ‘Homeric’ usage is being equated with the ‘Koine’ usage of Homer in the past, as opposed to current usage in the present. The ‘Koine’ way of reading ἐπεί σε at Iliad VI 355 is being equated with the sunētheia of Homer himself in the past, as opposed to the sunētheia of Greek speakers in the present: {54|55}

Pⓣ23 Scholia b(BCE³)T for Iliad VI 355a2

ἐπεί σε μάλιστα: ἔστι μὲν νῦν ἀντιδιασταλτικὴ ἡ σέ ἀντωνυμία, καὶ ἐχρῆν αὐτὴν ὀρθοτονεῖσθαι. ἡ δὲ συνήθεια ἐγκλιτικῶς ἀνέγνω.

ἐπεί σε μάλιστα: The pronoun σέ is nowadays distinctive [= non-enclitic, as opposed to non-distinctive or enclitic]. And it should have been accented with an acute. But the customary way [sunētheia] reads it as an enclitic.

P§148 So now I have come back full circle to the fact that the scholia b at Iliad VI 355a2 feature hē sunētheia ‘the customary way’ as a parallel to hē koinē anagnōsis ‘the Koine reading’ in the scholia A for Iliad VI 355a1. Both expressions function as the subject of the verb anegnō ‘reads’. In this case, to repeat, Herodian is reporting that the ‘Koine’ usage is different from contemporary usage. Like Aristarchus, he is making the point that the ‘Koine’ reading occasionally differs from ‘customary’ usage in his own time.

P§149 Although Aristarchus tends to prefer the ‘Koine’ variants whenever he is faced with a choice between ‘Koine’ usage and the ‘customary’ usage of his own time, there are examples where his preference seems to be headed in the other direction. Here is one such example:

Pⓣ24 Scholia b(CE³E⁴)T for Iliad II 53c (Herodian)

Ἀρίσταρχος τὸ ἷζε ἐκτείνει, ἡ δὲ κοινὴ συστέλλει

Aristarchus stretches out [the pronunciation of] ἷζε, but the Koine makes it short [= ἵζε].

P§150 In this case, the augmented form of the verb (ἷζε) is evidently the variant reflecting the ‘customary’ usage in the time of Aristarchus, and yet Aristarchus seems to be treating it as ‘Homeric’ despite the availability of the older unaugmented form of the verb (ἵζε), which is described as ‘Koine’. As we see from such examples, the observations of Aristarchus about any given reading are simply a matter of description, not prescription. Sometimes Aristarchus accepts the ‘Koine’ usages as ‘Homeric’, and sometimes he does not. Either way, whether or not he prefers the ‘Koine’ usages, he meticulously reports them. {55|56}

P§151 After the age of Aristarchus, however, his descriptions tend to be reinterpreted as prescriptions. Later generations of Aristarcheans can even compete with Aristarchus himself by prescribing what is or is not ‘customary’ for ‘Homer’. For example, where Aristarchus reports variant attestations of syntactical patterns featuring either the presence or absence of ‘articles’ (ὁ, ἡ, τό, etc.), Didymus prescribes the presence of ‘articles’ as a mark of ‘Homeric’ usage, as if Aristarchus had prescribed the opposite:

Pⓣ25 Scholia T for Iliad III 18b1 (Didymus)

αὐτὰρ ὁ δοῦρε δύω: ἄνευ τοῦ ἄρθρου τὸ αὐτὰρ δοῦρε ἐν πάσαις ταῖς χαριεστέραις. ἡ δὲ σὺν τῷ ἄρθρῳ γραφή, καίτοι μὴ οὖσα ᾿Αριστάρχου, ὅμως ἔχει ῾Ομηρικὴν συνήθειαν· ἔθος γὰρ αὐτῷ περὶ τοῦ αὐτοῦ διαλεγομένῳ μεσολαβεῖν τὸ ἄρθρον, ὡς περὶ ἄλλου λέγοντι, ὡς ἐπὶ τοῦ Σαρπηδόνος· αὐτοῦ μὲν ἤμβροτεν, ὁ δὲ Πήδασον οὔτα.

αὐτὰρ ὁ δοῦρε δύω: The sequence αὐτὰρ δοῦρε is without the article in all the khariesterai texts. But writing it with the article, even though it is not the way Aristarchus does it, nevertheless has a Homeric sunētheia about it, since it is a custom [ethos] for him [= Homer] to have an intervening article when he is talking about the same person who has been mentioned before, as in the passage about Sarpedon: αὐτοῦ μὲν ἤμβροτεν, ὁ δὲ Πήδασον οὔτα [cf. Iliad XVI 466-467].

Pⓣ26 Scholia b(BCE³) for Iliad III 18b2 (Didymus)

τὸ αὐτάρ παρὰ Ζηνοδότῳ καὶ τοῖς ἄλλοις οὐκ ἐπάγεται τὸ ἄρθρον. ὅμως οὖν ἐστι καὶ τοῦτο τῆς ῾Ομηρικῆς συνηθείας· ἔθος γὰρ αὐτῷ περὶ αὐτοῦ διαλεγομένῳ μεσολαβεῖν τὸ ἄρθρον, ὡς καὶ περὶ ἄλλου λέγων ποιεῖ· Σαρπηδὼν αὐτοῦ μὲν ἀπήμβροτεν, ὁ δὲ Πήδασον…

The αὐτάρ in the texts of Zenodotus and of others does not introduce the article. Nevertheless, even this [= the use of the article] belongs to the Homeric sunētheia. For it is his [= Homer’s] custom [ethos] when he is talking about the same person to have an article in the middle, as when he says in connection with another character: Σαρπηδὼν αὐτοῦ μὲν ἀπήμβροτεν, ὁ δὲ Πήδασον…

P§152 Whereas the prescription formulated by Didymus implies that the variant showing the presence of the ‘article’ is in this case typical of ‘Homeric’ {56|57} usage, Aristonicus describes it simply as an example of what is koinon ‘common’. Here are the reports of Didymus and Aristonicus, as juxtaposed in the Homeric scholia:

Pⓣ27 Scholia A for Iliad III 18a (Didymus | Aristonicus)

{καὶ ξίφος} αὐτὰρ ὁ δοῦρε: ᾿Αρίσταρχος ἄνευ τοῦ ἄρθρου, αὐτὰρ δοῦρε. οὕτως καὶ ἡ Ἀριστοφάνους καὶ ἡ Καλλιστράτου καὶ σχεδὸν οὕτως καὶ αἱ χαριέσταται· καὶ ὁ Ἰξίων ἐν τῷ πρώτῳ Πρὸς τὰς ἐξηγήσεις ὁμοίως προφέρεται. ἔχει δὲ τὸν ῾Ομηρικὸν χαρακτῆρα καὶ ἡ σὺν τῷ ἄρθρῳ γραφή, καίπερ οὐκ οὖσα ᾿Αριστάρχειος· σύνηθες γὰρ τῷ ποιητῇ ἐπὶ τῶν αὐτῶν μένοντι ὑπόνοιαν παρέχειν ὡς περὶ ἑτέρου διαλέγοιτο προσθέσει ἄρθρου καὶ μεταλλάξει τοῦ συνδέσμου, ὡς ἐπὶ τοῦ Κύκλωπος κὰδ δέ μιν ὕπνος | ᾕρει πανδαμάτωρ, ὁ δ' ἐρεύγετο οἰνοβαρείων ἀντὶ τοῦ καὶ ἐρεύγετο. {|} Ζηνόδοτος δὲ ἠθέτηκε τοῖς ἑξῆς οὐ συντιθέμενον. δεῖ δὲ κοινὸν παραλαβεῖν αὐτὰρ ὁ δοῦρε δύω ἔχων καὶ νοεῖν τὸ ἄρθρον ἐξ ἐπαναλήψεως παρειλημμένον ἐπὶ τοῦ αὐτοῦ προσώπου.

{καὶ ξίφος} αὐτὰρ ὁ δοῦρε: Aristarchus has it without the article, αὐτὰρ δοῦρε. So also the text of Aristophanes and the text of Callistratus and so also practically all the khariestatai texts. Likewise Ixion adduces it in the first volume of his Πρὸς τὰς ἐξηγήσεις. But writing it with the article also has a Homeric character, even though it is not Aristarchean. For it is customary [sunēthes] for the Poet when he is staying on the topic of the same [two] entities to provide a hint that he is talking about the other of the two entities by way of adding an article and shifting the conjunction, as in the case of the Cyclops: κὰδ δέ μιν ὕπνος | ᾕρει πανδαμάτωρ, ὁ δ’ ἐρεύγετο οἰνοβαρείων [cf. Odyssey ix 372-374], where we see ὁ δ’ ἐρεύγετο instead of καὶ ἐρεύγετο. [In the corresponding Greek text, I mark this point with the sign “{|}”: it is where the testimony of Didymus stops and the testimony of Aristonicus begins.] Zenodotus athetized this verse as not fitting the contiguous verses. But it is necessary to take it the way it is common [koinon], [92] αὐτὰρ ὁ δοῦρε δύω ἔχων, and to interpret the article as referring back to the same person to whom reference was made earlier. {57|58}

P§153 The description as reported by Aristonicus, indicating that the variant usage featuring the ‘article’ is koinon ‘common’, may actually stem from Aristarchus. In other words, it may be Aristarchus himself who is describing as koinon ‘common’ the usage featuring the ‘article’, that is, the Koine usage. But the question is, did Aristarchus actually find the Koine variant attested in the verse that we know as Iliad III 18? In this case, it seems that he did not: otherwise, he would have shown the Koine variant in his base text while conscientiously reporting the alternative variant as found in the khariesterai. It is not that Aristarchus removed the ‘article’ from Iliad III 18 in his base text; rather, it appears that he found no ‘article’ attested for this verse, and he expressed the opinion that Homeric usage had led him to expect an attestation.

P§154 In the era of Aristarchus, the descriptions of what is sunēthes ‘customary’ or koinon ‘common’ in Homer are based on the evidence of variant texts he has at his disposal, not only on the evidence of usage he extrapolates from the language of the texts. In the post-Aristarchean era, however, the opportunities for finding such new textual evidence become increasingly rare, and the various questions about what is sunēthes ‘customary’ or koinon ‘common’ in Homer become increasingly limited to the evidence of usage alone – primarily as described by Aristarchus himself. Accordingly, scholars in the post-Aristarchean era become increasingly dependent on the authority of Aristarchus concerning whatever he describes as sunēthes ‘customary’ or koinon ‘common’. Such descriptions tend to be reinterpreted by later Aristarcheans as prescriptions of whatever is ‘Homeric’ pure and simple, as if all the descriptions of Homeric usage were meant to be compared with current usage in what Aristarchus calls ‘our’ time. In the following passage, for example, the purpose clause implies that the Aristarchean ‘editions’ of Homer feature readings in the base text that conform to usage that is sunēthes ‘customary’ for ‘us’:

Pⓣ28 Scholia in Papyrus 12 (= Oxyrhynchus Papyri 221 second century CE) of Erbse’s edition of the Homeric scholia (vol. 5 1977) commenting on Iliad XXI 221 (Erbse vol. 5 p. 98, Col. XI line 15)

ἔασον: αἱ Ἀριστάρχ<ε>ιοι οὕτως, ἵνα τὸ σύνηθες ἡμῖν ἦι

ἔασον: This is the way [it is read] by the Aristarchean editions – so that the form may be the one that is customary [sunēthes] for us. [93] {58|59}

P§155 In this example, what is sunēthes ‘customary’ in the current usage is assumed to be the kind of variant that a good Aristarchean should prefer as the truly ‘Homeric’ usage – as if the criterion of current usage could somehow be applied consistently to a given variant that Aristarcheans prefer as more ‘Homeric’ than another given variant.

P§156 It all comes down to this: in the post-Aristarchean era, scholars no longer had access to the ‘Koine’ tradition that had been available to Aristarchus himself when he established his own base text. So they had no independent judgment about what is or is not ‘Koine’ whenever Aristarchus invokes the concept of hē koinē anagnōsis ‘the Koine reading’.

Pⓢ14. Reading Homer out loud to Aristarchus

P§157 In the era of Aristarchus, hē koinē anagnōsis ‘the Koine reading’ (ἡ … κοινὴ ἀνάγνωσις, as at scholia A for Iliad II 662a1 and elsewhere) was actually a matter of reading out loud to Aristarchus the Homeric base text for verification. That is, the Homeric base text was read out loud to Aristarchus as part of the editorial procedure of his diorthōsis. In the Homeric scholia, we find indirect references to such reading out loud.

P§158 A case in point is the use of the word anagnōstēs ‘reader’. [94] In the Homeric scholia, this word is applied in two references to a man called Posidonius, who is described as the anagnōstēs ‘reader’ for Aristarchus (Ποσειδώνιον τὸν Ἀριστάρχου ἀναγνώστην / Ποσειδώνιος … ὁ ἀναγνώστης Ἀριστάρχου in the scholia A for Iliad XVII 75a / VI 511a).

P§159 In one of the two references, where the meaning of a Homeric verse (Iliad XVII 75) depends on whether a given word in that verse is syntactically connected with what precedes or with what follows, Aristarchus decides on what he considers to be the correct reading by listening to the reading of ‘his’ anagnōstēs Posidonius:

Pⓣ29 Scholia A for Iliad XVII 75a (Nicanor)

Ἕκτορ, νῦν σὺ μὲν ὧδε <θέεις ἀκίχητα διώκων>: μετὰ τὸ ὄνομα στικτέον· προσαγορευτικὴ γάρ ἐστι. τὸ δὲ ἀκίχητα φασὶ Ποσειδώνιον τὸν Ἀριστάρχου ἀναγνώστην τοῖς ἑξῆς προσνέμειν καὶ τὸν Ἀρίσταρχον ἀποδέχεσθαι. {59|60}

Ἕκτορ, νῦν σὺ μὲν ὧδε <θέεις ἀκίχητα διώκων>: It should be punctuated after the noun [= ἀκίχητα], for it is vocative. But they say that the word ἀκίχητα was read by Posidonius the reader [anagnōstēs] of Aristarchus as belonging to the words that follow, and that Aristarchus accepted [apodekhesthai] this as traditional.

P§160 Evidently the phonology of the reader’s performed reading follows one of the two possible ways of performing the verse, and it is said here that Aristarchus accepted (apodekhesthai) as traditional the version performed by his reader. In terms of editing, Aristarchus considered the reading of Posidonius to be one of two syntactical variants, and he preferred the variant performed by his reader. The wording of Eustathius confirms the essentials of the approach taken by Aristarchus:

Pⓣ30 Eustathius Commentary 4.17|18.25-26|1 on Iliad XVII 75-76

Τὸ δὲ πλῆρες τῆς ῥηθείσης παροιμιώδους συντάξεως τοιοῦτον· σὺ μὲν ὧδε θέεις, εἶτα ὡς ἄλλης ἀρχῆς, ἀκίχητα διώκων ἵππους, καὶ τὸ ἑξῆς. Ποσειδώνιος δέ, ὁ Ἀριστάρχου ἀναγνώστης, οὕτω λέγει· σὺ μὲν ὧδε θέεις ἀκίχητα, εἶτα· διώκων ἵππους Αἰακίδαο, καὶ τὸν ᾿Αρίσταρχον ἀποδέχεσθαι τοῦτο.

The filled-out meaning of this sententious utterance goes something like this: σὺ μὲν ὧδε θέεις, and then, as if starting the syntax anew, ἀκίχητα διώκων ἵππους, and then the rest of the syntax in sequence. But Posidonius, the reader [anagnōstēs] of Aristarchus, says it this way: σὺ μὲν ὧδε θέεις ἀκίχητα, and then διώκων ἵππους Αἰακίδαο. And it is said that Aristarchus accepted [apodekhesthai] this version.

P§161 In contradiction of what is said in scholia A, Eustathius says that Posidonius read ἀκίχητα as going with what precedes syntactically, not with what follows. Either Eustathius or the transmitter of the scholia A has evidently reversed the direction of syntactical linkage. Still, the point remains that Aristarchus considered the reading of Posidonius to be one of two syntactical variants, and that he chose the variant he heard performed by his reader.

P§162 In the other of the two references in the Homeric scholia, it is reported that Posidonius the anagnōstēs was reading out loud the wording of {60|61} the verse that we know as Iliad VI 511, pronouncing rhimpha-e where we now read ῥίμφα ἑ in the expression ῥίμφα ἑ γοῦνα φέρει: [95]

Pⓣ31 Scholia A for Iliad VI 511a (Aristonicus)

ῥίμφα <ἑ γοῦνα φέρει>: Ζηνόδοτος ῥίμφ’ ἑὰ γοῦνα φέρει. Ποσειδώνιος δὲ ὁ ἀναγνώστης Ἀριστάρχου <ἄνευ> διαιρέσεως τὸ ε ψιλῶς προφέρεται, παρέλκειν αὐτὸ λέγων ὡς ἐν τῷ ἠὲ σὺ τόνδε δέδεξο, καὶ λύεται τὸ σολοικοφανές. ὁ δὲ Ὅμηρος ὑπὸ τῶν γονάτων καὶ ποδῶν φέρεσθαι λέγει· τὸν μὲν ἄρ’ ὣς εἰπόντα πόδες φέρον.

ῥίμφα <ἑ γοῦνα φέρει [‘lightly his knees carry him along’]>: Zenodotus has ῥίμφ’ ἑὰ γοῦνα φέρει [‘lightly he carries his own knees’]. But Posidonius the reader [anagnōstēs] of Aristarchus produces the ε without word-division and with smooth breathing, saying that it is suffixed as in the expression ἠὲ σὺ τόνδε δέδεξο [Iliad V 228] [where the -ὲ of ἠὲ is ‘suffixed’]. This way, the problem of the apparent solecism [= as in the reading of Zenodotus] is solved. In Homeric idiom, one can be ‘carried’ [φέρεσθαι] by one’s knees and feet: τὸν μὲν ἄρ’ ὣς εἰπόντα πόδες φέρον [Iliad XV 405].

Pⓣ32 Scholia A^im for Iliad VI 511b (exegetical scholia) [96]

<ῥίμφα ἑ γοῦνα φέρει:> μετέβη τὴν πτῶσιν· τὸ γὰρ ἕ αὐτόν σημαίνει.

<ῥίμφα ἑ γοῦνα φέρει:> He [= Homer] omitted the declined form. For the ἕ means ‘him’. {61|62}

Pⓣ33 Scholia T for Iliad VI 510-511a1 (Aristonicus | exegetical scholia)

ὁ δ’ ἀγλαΐηφι <πεποιθώς ῥίμφα ἑ γοῦνα φέρει>: ἀντὶ τοῦ τοῦτον. {|} καὶ τὸ ἕ ἀντὶ τοῦ αὐτόν, φημὶ τὸ ῥίμφα †ι. | Ποσειδώνιος δὲ ψιλῶς τὸ ἕ προφέρεται καί φησιν αὐτὸ πλεονάζειν ὡς ἐν τῷ ἠὲ σύ.

ὁ δ’ ἀγλαΐηφι <πεποιθώς | ῥίμφα ἑ γοῦνα φέρει>: The ἕ is in place of ‘him’. [In the corresponding Greek text, I mark this point with the sign “{|}”: it is where the testimony of one source stops and the testimony of another source begins.] The ἕ is in place of ‘him’. I say ‘ῥίμφα †ι. | Posidonius produces the ἕ with smooth breathing and says that it [= the ἕ] is pleonastic, as [the -ὲ] in the expression ἠὲ σύ [Iliad V 228].

Pⓣ34 Eustathius Commentary 2.377|378.19-21|1-3 on Iliad VI 510-511

Τινὲς δέ φασι τὸν τοῦ ᾿Αριστάρχου ἀναγνώστην Ποσειδώνιον ῥίμφαε λέγειν ἐν ἑνὶ τρισυλλάβῳ μέρει λόγου, πλεονάσαντος, φησίν, ἐν παρολκῇ τοῦ ε, ὡς ἐν τῷ ἠὲ σὺ τόνδε δέδεξο. καὶ αὐτοὶ μὲν τοιαῦτα. Ὁ δὲ Ὁμηρικὸς ἀνὴρ ἥδιον ἀκούοι ἂν καὶ νῦν καινοφωνοῦντος ἐλλόγως τοῦ ποιητοῦ τὰ ἀστεῖα ταῦτα σολοικοφανῆ ἤπερ ἀνέχοιτο Ἀττικῶς φθεγγομένων ἀκροᾶσθαι τῶν ἐπιδιορθουμένων αὐτόν

And some people say that the reader [anagnōstēs] for Aristarchus, Posidonius, said ῥίμφαε, as if it were a trisyllabic word, with a pleonastic suffixation of ε, he says, as in the expression ἠὲ σὺ τόνδε δέδεξο. Well, that is the kind of thing that these people say. But a man who knows his Homer would more gladly hear, even today, the Poet’s neologizing of these elegant words, though they appear to be solecisms, instead of putting up with listening to Attic-sounding speakers who try to correct him [= the Poet].

P§163 In the last of these passages that I just quoted, Eustathius is evidently distancing himself from the methodology of Aristarchus, who tests the usage of ‘Homer’ by comparing ‘Attic’ and other supposedly extraneous evidence.

P§164 The rationalizing explanation given by Posidonius for the pronunciation, that the -e is ‘pleonastic’ like the -e of the conjunction ē-e (ἠὲ), may be diachronically invalid, but the actual need for a rationalization indicates the reality of the traditional way in which this Homeric verse was performed for Aristarchus by ‘his’ reader. The variant reading adduced by Zenodotus, {62|63} by contrast, is ῥίμφ’ ἑὰ γοῦνα φέρει. From the short-term perspective of the immediate context in Iliad VI 511, this version as adduced by Zenodotus is the lectio facilior. [97] From the long-term perspective of Homeric diction involving analogous phraseology, however, Aristarchus preferred the lectio difficilior. [98] The pronunciation of his anagnōstēs, who read rhimphae, helped Aristarchus reconstruct the syntactically and morphologically valid wording ῥίμφα ἑ. [99] Such a pronunciation, even if Posidonius himself interpreted it wrongly, is I think a plausible reflex of the diachronically valid wording ῥίμφα ἑ, which was evidently the interpretation of Aristarchus. Moreover, the reading ῥίμφα ἑ is evidently the Koine version, reinforced here by the Koine pronunciation rhimphae of the official reader or anagnōstēs. [100] It is the Koine reading, the koinē anagnōsis. [101]

P§165 The explanation I have offered here is at odds with another explanation, formulated as follows: “Evidently Aristarchus had Zenodotus’ text read aloud to him by Posidonius, who had to decide between alternative articulations. Aristarchus occasionally thought it worth noting his choices and giving him credit for the m.” [102] As I have just argued, there is no reason to infer that it was specifically the text of Zenodotus that had been read out loud for Aristarchus by Posidonius. [103] Nor is it justified to assume that Posidonius was arbitrary in following one of two alternative articulations. The point of the Aristarchean testimony about the reading of Posidonius to Aristarchus is that the reading of this anagnōstēs was a traditional reading. Aristarchus accepted the readings of Posidonius as traditional: such is the force of the expression apodekhesthai ‘accept as traditional’.

P§166 This editorial practice, where an anagnōstēs ‘reader’ reads out loud to an editor a text to be approved by the editor, is reflected in a traditional formula attested in texts dated to later periods. The formula consists of a genitive {63|64} absolute construction, ἐκδόσεως παραναγνωσθείσης in combination with the use of the dative case in referring to the editor:

Pⓣ35 Explicit for the commentary of Eutocius of Ascalon (sixth century CE) on Book I of Archimedes On Sphere and Cylinder (ms. Laurentianus 28.4) [104]

Εὐτοκίου Ἀσκαλωνίτου ὑπόμνημα εἰς τὸ πρῶτον τῶν Ἀρχιμήδους περὶ σφαίρας καὶ κυλίνδρου, ἐκδόσεως παραναγνωσθείσης τῷ Μιλησίῳ μηχανικῷ Ἰσιδώρῳ ἡμετέρῳ διδασκάλῳ

The commentary [hupomnēma] of Eutocius of Ascalon to the first volume of the works of Archimedes On Sphere and Cylinder, and the ekdosis had been read out loud to the Milesian engineer Isidore, my teacher, for cross-checking [paranagignōskein] [105]

P§167 My interpretation of paranagignōskein here, ‘read out loud for cross-checking’, needs further comment. The contexts of some earlier attestations of paranagignōskein indicate that the text that is being read out loud is a model text in the making, that is, it is a text that becomes a model through the actual process of cross-checking and correcting. I have discussed such contexts in my previous work. [106] Here I concentrate on specialized contexts where the model text is being read out loud to its editor, whose task is to cross-check this text with other texts in order to correct any mistakes. In the wording I just cited, the editor of the text of Archimedes is Isidore of Miletus. [107] The model text that Isidore edited is in this case the actual text of Archimedes that is being read back to Isidore for cross-checking and correction. [108] The process of cross-checking may involve reading from – or remembering from – parallel texts, leading to corrections and, ultimately, approval by the editor.

P§168 I proceed to examine some further examples of this formula: {64|65}

Pⓣ36 From the heading to Book 3 of the commentary by Theon of Alexandria on Ptolemy’s Almagest (ms. Laurentianus 28.18) [109]

Θέωνος Ἀλεξανδρέως εἰς τὸ τρίτον τῆς μαθηματικῆς Πτολεμαίου Συντάξεως ὑπόμνημα ἐκδόσεως παραναγνωσθείσης τῇ φιλοσόφῳ θυγατρί μου Ὑπατίᾳ

The commentary [hupomnēma] of Theon of Alexandria to the third volume of the mathēmatikē Suntaxis of Ptolemy, and the ekdosis had been read out loud to my philosopher daughter Hypatia for cross-checking [paranagignōskein].

P§169 The role of Theon’s learned daughter Hypatia can be better understood in the light of the formula describing the role of Theon himself:

Pⓣ37 From the headings to Books 1 and 2 of Theon’s commentary on Ptolemy’s Almagest [110]

Θέωνος Ἀλεξανδρέως τῆς παρ᾿ αὐτοῦ γεγενημένης ἐκδόσεως εἰς τὸ πρῶτον τῆς Συντάξεως Πτολεμαίου ὑπόμνημα

Θέωνος Ἀλεξανδρέως τῆς παρ᾿ αὐτοῦ γεγενημένης ἐκδόσεως εἰς τὸ δεύτερον τῆς Συντάξεως Πτολεμαίου ὑπόμνημα

The commentary [hupomnēma] of Theon of Alexandria of the edition [ekdosis] made by him of the first / second volume of the Suntaxis of Ptolemy

P§170 Just as the father edited Books 1 and 2 of Ptolemy’s Almagest, so also the daughter edited Book 3: for the first two books, Theon produced both the edition and the commentary, but for the third he produced only the commentary while Hypatia produced the edition. [111] As we see in the more expanded wording that describes the contribution of the daughter, the actual process of editing involved the reading out loud of the text to the editor. This process is mentioned explicitly only when the person who produced the commentary or hupomnēma is different from the person who produced the edition or ekdosis.

P§171 Unlike the roles of Isidore and Hypatia, who produced only the ekdosis ‘edition’ but not the hupomnēma ‘commentary’ for their respective {65|66} texts, the role of Theon is directly comparable to that of Aristarchus, in that he produced both ekdosis and hupomnēma for the Homeric texts. And, like Theon, Aristarchus must have had the text read out loud to him for the purpose of cross-checking. The process of cross-checking leads to diorthōsis ‘correction’ and, ultimately, approval, so that the edited text becomes the model text. [112]

Pⓢ15. Conclusions: the Homeric Koine as a classical text

P§172 I conclude that the base text of the edition of Homer by Aristarchus was at least notionally the Koine version. I conclude also that the criteria of Aristarchus in editing Homer were relatively more rigid than those established by one of his predecessors as director of the Library of Alexandria, Zenodotus of Ephesus, who was active over a hundred years earlier, in the age of Callimachus.

P§173 It is all-important for me to stress, however, that the criteria of Zenodotus as editor of Homer were rigid in their own right. They can be considered relatively less rigid only if we compare them to the later criteria of Aristophanes and to the still later criteria of Aristarchus himself. The earlier criteria of Zenodotus, by comparison with those of his own contemporaries in the age of Callimachus, turn out to be almost as rigid as those of Aristophanes, Aristarchus, and later Aristarcheans such as Didymus in the age of Virgil.

P§174 The Homeric koinē, equated by Aristarchus with the base text he used for editing Homer, was rigid not only because it was the ‘definitive’ or ‘standard’ Homer: it was rigid also because it was the ‘common’ Homer, the ‘vulgate’ Homer. This last term ‘vulgate’ tends to be misunderstood and misused nowadays. The casual reader of Homeric scholarship is often led to believe that the ‘vulgate’ Homer was unstable and even chaotic – or let us say fluid. As I will argue in Chapter 3, however, the Homeric koinē or ‘vulgate’ was the opposite: it was relatively stable, orderly, and even rigid. To put it more precisely, the Homeric koinē was notionally rigid, that is, it was thought to be unchangeable in terms of its reception, from the standpoint of those who {66|67} considered Homer to be their common possession, their own standard Homer. The way I have just spoken about the vulgate Homer is meant to evoke the way Saint Jerome spoke about the vulgate Bible. [113]

P§175 In contemplating the reality of multiple koinai and the ideal of a singular koinē of Homer, I see an analogy with the vulgata or ‘vulgate’ of the Bible. [114] In Jerome’s Epistle to Sunnia and Fretela, the word koinē, which he glosses in Latin as the vulgata or ‘vulgate’, is applied to two ‘common’ Greek versions of the Hebrew Bible, one of which is the editio or ‘edition’ of one Lucian while the other is the editio of Origen of Alexandria (late second to mid-third century CE) – that is, the Septuagint as edited in the Hexapla of Origen. [115] As in the usage of the Homeric scholia, there is an element of negative comparison here as well: conceding that the Greek term koinē is applicable to both of the Greek-language biblical ‘editions’ in question, Jerome goes on to contrast the ‘old corrupt edition’ of Lucian with the ‘uncorrupted and immaculate’ version resulting from the edition of Origen, which serves as the source for Jerome’s Latin vulgate translation:

Pⓣ38 Jerome Epistles 106.2

quaeritis a me rem magni operis, et majoris invidiae; in qua scribentis non ingenium, sed eruditio comprobetur; ut dum ipse cupio de caeteris judicare, judicandum me omnibus praebeam: et in opere Psalterii juxta digestionem schedulae vestrae, ubicumque inter Latinos Graecosque contentio est, quid magis Hebraeis conveniat, significem. In quo illud breviter admoneo, ut sciatisstrong aliam esse editionem, quam Origenes et Caesariensis Eusebius, omnesque Graeciae tractatores κοινήν, id est communem appellant, atque vulgatam, et a plerisque nunc Λουκιανός dicitur; aliam Septuaginta interpretum, quae in ἑξαπλοῖς codicibus reperitur, et a nobis in Latinum sermonem fideliter versa est, et Jerosolymae atque in Orientis ecclesiis decantatur. … κοινή autem ista, hoc est communis, editio ipsa est quae et septuaginta, sed hoc interest inter utramque {67|68} quod κοινή pro locis et temporibus et pro voluntate scriptorum vetus corrupta editio est, ea autem quae habetur in ἑξαπλοῖς et quam nos vertimus ipsa est quae in eruditorum libris incorrupta et immaculata septuaginta interpretum translatio reservatur.

You ask me about a matter of great importance, which is a source of even greater invidiousness. In this matter it is not the inborn gifts of the one who is writing but his erudition [eruditio] [116] that is being tested. The result is that, in wishing to make judgments about others, I am setting myself up to be judged by all. Juxtaposing what is in my work on the Psalter with a listing of your queries, wherever there is a difference between the Latin and Greek texts I will indicate what agrees more with the Hebrew texts. In this matter, I caution you about something, just briefly. Here is what you have to understand. Thstrongere is an edition that Origen, Eusebius of Caesarea, and all those who are experts in the Greek tradition call the Koine edition, that is, the common edition, the vulgate edition, and most people nowadays call it the Lucianus edition. Then there is another edition [= the Septuagint], which comes from the Seventy Interpreters, which is found in the hexapla codices and which I faithfully translated into the Latin language. This is the edition that is liturgically sung in Jerusalem and in other congregations of the East. … That Koine of yours [= the Lucianus edition that you are using], that is, the common edition, is the same thing as the edition of the Seventy [= the Septuagint = the edition that I am using], but there is this difference between the two: that the Koine – in line with different times and different places and different whims of people who wrote it down – is an old corrupt edition, whereas by contrast the one that is contained in the hexapla and which I have translated is the same thing as the actual translation [into Greek from Hebrew] that has been conserved in the books of the erudite [eruditi] [117] without corruption and without blemish – that is, the translation by the Seventy interpreters.

P§176 In other words, the ‘edition’ of the Septuagint that Jerome uses as his own textual source – that is, the edition of Origen as found in the fifth {68|69} column of his six-column display or hexapla – is the same thing as the koinē to the extent that it is ‘common’ in the sense of ‘general’ or even ‘universal’, but it transcends the designation of koinē to the extent that it is a ‘corrected’ text, freed from ‘corruptions’ associated with the other koinē text that is ‘common’ in the sense of ‘vulgar’.

P§177 For Jerome, then, the word koinē has the aura of an authoritative but relatively ‘uncorrected’ text. Similarly in the case of Aristarchus, I am saying that his category of koinai or ‘common’ texts of Homer is notionally derived from an authoritative but relatively ‘uncorrected’ textual source – a Koine in the sense of a ‘standard’ version. As I have argued in earlier work and will argue further in the chapters that follow, especially in Chapter 3, such an authoritative but relatively ‘uncorrected’ Koine is the Homeric tradition as it existed in Athens in the fourth century. [118]

P§178 A piece of evidence that supports my reconstructing a fourth-century Athenian Koine is the fourth-century Athenian usage of the adjective koinos / koinē as ‘common’ in the ideological sense of ‘general, standardized, universalized’. [119] As I have argued in earlier work, such a meaning of Koine would fit the Iliad and Odyssey as “owned” by the Athenian State, on the occasion of seasonally recurring performances at the Festival of the Panathenaia. [120]

P§179 In my earlier work, I described such a notionally single text, such a single Koine, in terms of an Athenian “City Edition” of Homer. [121] But the very idea of an Athenian “City Edition” can be called into question on the grounds that Aristarchus never refers to such a thing. According to this line of reasoning, there never was an Athenian “City Edition,” since Aristarchus would most likely have preferred it over other editions – if such a thing had ever existed. [122] It all depends, however, on what we mean when we speak of Aristarchus’ preferences. In one sense, Aristarchus preferred readings judged {69|70} to be khariesterai over readings that were koinai. In another sense, however, the overall practice of Aristarchus can be described as just the opposite: we have seen that he chose as a rule to retain in his base text the readings he judged to be koinai. In that sense, he did indeed prefer what I have just described as an Athenian “City Edition.” Still, I now think it would be better not to use the word edition in this context. After all, the Homeric Koine was for Aristarchus not an edition but rather a preedition. For Aristarchus, the Koine was a tradition that still needed to be edited. The Homeric scholia make it clear that Aristarchus’ criterion for distinguishing a superior from an inferior ekdosis or ‘edition’ was the variable scholarly quality of the editing process, that is, of diorthōsis or ‘correction’ – in the sense of restoring ‘genuine’ or ‘original’ readings to a ‘corrupted’ text.

P§180 To the extent that the koinē Homer is ‘common’ in the uneroded and privileged sense of a ‘general, standardized, universalized’ text stemming from an earlier past, we can expect Aristarchus to value it; to the extent that this same koinē is ‘common’ in the eroded and non-privileged sense of ‘vulgar’, we can expect him to prefer the more ‘corrected’ editions from the more recent past, including those of Aristophanes and Zenodotus. This pattern of preference could only be expected to intensify in the post-Aristarchean era, by which time the privileged sense of koinē would have eroded further.

P§181 In Jerome’s Letters (106.2), as we saw, he speaks about achieving a perfect biblical text by correcting any mistakes that may have corrupted or stained the vulgata or ‘vulgate’. Correction, he says, recovers a vulgata that is incorrupta et immaculata, freed of all corruptions and stains. Implicitly, the vulgata had been authoritative from the start, and that is why it was worthy of serving as a base text in the first place. The base text simply needed to be corrected in order to become the master text. In the passage from the Letters of Jerome that I am citing, the Greek word that Jerome says is equivalent to this authoritative base text, this vulgata, is koinē / κοινή.

P§182 I propose to illustrate this essential concept of the Koine as an authoritative base text by showing a snapshot image that I invoke as a metaphor for the fixity sought by Jerome for the vulgate Bible. The image I have in mind is in the neighborhood where I live. It is a bronze statue that sits in front of the Croatian Embassy on Massachusetts Avenue in Washington, District of Columbia (Figure 1).

Figure 1. Ivan Meštrović (1883-1962), St. Jerome the Priest, 1954. Bronze, over life size. Located in front of the Embassy of the Republic of Croatia, Washington, DC.

P§183 The statue represents Saint Jerome, or “Jerome the Priest,” as the lettering inscribed at the base of the statue calls him. We see him seated on top of a block of stone, wearing nothing, hunched over a codex of the Bible, the Latin vulgate version of which is primarily his own achievement – the result of {70|71} his own agonizing efforts at perfection. I walk past Jerome every time I head for Sheridan Circle. There he sits, day after day, stooped over his beloved text, lost in his thoughts as he seeks ever to find the exact meaning of the Book. The rigidity of the bronze that freezes Jerome in this moment of deep concentration extends from the fixity of the text he is forever reading. Oblivious to his external surroundings, he sits there in all his nakedness, all exposed to the elements, rain or shine, and unwittingly inviting the gaze of gawking passers-by. Day in and day out, he just sits there, steadfastly indifferent to all that comes and goes outside his inner world of words, those perfect words of perfection in the making.

P§184 With this image in mind, I conclude my argument about the Homeric Koine as ‘corrected’ by Aristarchus, which had become the definitive Homeric text of later ages, including the age of Virgil. {71|72}

P§185 Aristarchus accepted the already rigid Athenian Koine version of Homer as his base text not because he thought it went back directly to Homer. In fact, we have seen that Aristarchus distanced himself from the Koine by imposing even more rigid criteria for determining what is or is not Homeric. Still, Aristarchus adhered to the Koine version as his point of departure, as a base text that required massive editorial correction, diorthōsis, which he displayed fully in his commentaries. Although Aristarchus’ narrower system of preferred readings – his diorthōsis – has not survived as a system, the broader system of the Koine textual tradition that he used as his base text, which was still relatively narrower than the augmented base text used by Zenodotus and Crates, has indeed survived, by and large, in the so-called medieval vulgate, that is, in the overall medieval manuscript tradition of Homer. [123] It has also survived, by and large, in papyri dating from the post-Aristarchean period. It has even survived in Homeric quotations dating from that later period. These survivals represent the classical text of Homer. {72|73}

Footnotes

[ back ] 1. My approach to such debates is outlined in HR 1-3.

[ back ] 2. For alternatives to such ideas, see Graziosi 2002:90-93. I am in general agreement with her alternative ideas.

[ back ] 3. HTL xi.

[ back ] 4. PP and HQ .

[ back ] 5. On the term Panhellenic, see HQ 39-42.

[ back ] 6. See PP 110 and following, with working definitions of the descriptive terms transcript, script, and scripture. Chapters 5 / 6 / 7 of PP focus respectively on transcript / script / scripture. See also HQ 41, where the same descriptive scheme of five consecutive periods is more explicitly situated in an overall evolutionary model.

[ back ] 7. HQ 34-36, 65-69.

[ back ] 8. HQ 34-36, with further citations.

[ back ] 9. PP 153-186, HQ 32-34.

[ back ] 10. PP 187-206.

[ back ] 11. A synchronic standpoint is needed for viewing the current state of an existing structure, while a diachronic standpoint is needed for viewing different phases in the evolution of that structure. In HR 1, I offer background on the application of these terms synchronic and diachronic to Homeric studies.

[ back ] 12. In the Suda, the life span of Didymus is described thus: γεγονὼς ἐπὶ Ἀντωνίου καὶ Κικέρωνος καὶ ἕως Αὐγούστου ‘flourished in the era of Antonius and Cicero, up to the era of Augustus’. Pfeiffer 1968:256 comments on γεγονώς as ‘floruit’. More on Didymus in West 2001:48. By way of comparing dates traditionally ascribed to various scholars in this era, Cameron 1995:191n33 calculates that Didymus was born in 80 BCE and died in 1 BCE.

[ back ] 13. For example, West 2001:46-85 thinks that Aristarchus neither collected nor collated Homeric texts. For counterarguments, see HTL 87-109; also Rengakos 2002, Montanari 2002a, and 2004a.

[ back ] 14. That Aristarchus actively assembled and collated khariesterai texts is argued in HR 4-7; see also Rengakos 2002, Montanari 2002a and 2004a.

[ back ] 15. I offer an overview in HTL 87-109.

[ back ] 16. Again, I offer an overview in HTL 87-109. See Rengakos 2002 on scholia A for Iliad XIX 386b1, where the testimony of Didymus indicates that Aristarchus himself must have consulted the politikai ‘city editions’. For the relative dating of the politikai, see Citti 1966. For a defense of the authenticity of variant readings found in the politikai, see PP 147-148, following Citti 1966. For further comments, see Haslam 1997:69-71. West 2001:67 notes that the Homeric scholia nowhere refer to the politikai in terms of a diorthōsis. So perhaps they are ‘uncorrected’ texts. In other words, perhaps they are texts that have not undergone the editorial procedures of diorthōsis ‘correction’. As I note in PP 116n46, however, there is one place where the politikai are called ekdoseis (scholia A for Iliad III 10, with reference to the Homer texts from Chios and Massalia).

[ back ] 17. I postpone till P§46 my discussion of the terms athetesis and athetize.

[ back ] 18. See the previous note.

[ back ] 19. Montanari 2002a:124-127.

[ back ] 20. Haslam 1997:71, following S. West 1967:26.

[ back ] 21. I introduced these concepts in Nagy 2000a; see also HTL 54, 56, 59-64, 66, 73.

[ back ] 22. See also in general Apthorp 1980.

[ back ] 23. See the previous note.

[ back ] 24. Apthorp 1980:xv. See also HTL 36. Later on, in ch. 2, I focus on one particular plus verse, Iliad XIV 246a. That verse, concerning the Ōkeanos, was excluded from the base text of Aristarchus.

[ back ] 25. Montanari 2002a:124 says: “The plus-verses present in the Zenodotean text were not his own interpolations but were instead typical of exemplars that were current in his day.” West 2001:40n33 offers a list of such plus-verses: Iliad I 69ab, III 338a, V 808, VIII 52a-d, X 349, XIII 808a, XIV 136a, XVII 456a, XVIII 156a.

[ back ] 26. Montanari 2002a:120-125.

[ back ] 27. Montanari 2002a:124.

[ back ] 28. Montanari 2002a:124-125.

[ back ] 29. Montanari 2002a:124.

[ back ] 30. Montanari 2002a:125.

[ back ] 31. Montanari 2002a:125.

[ back ] 32. Montanari 2002a:124.

[ back ] 33. Montanari 2002a:124.

[ back ] 34. For an overview of this concept of numerus versuum, see PP 138-140, 143-144, 155, following especially Apthorp 1980.

[ back ] 35. HTL 54, 56, 59-64.

[ back ] 36. Montanari 2002a:124-127.

[ back ] 37. LP (Nagy 1998) 213-228.

[ back ] 38. LP (Nagy 1998) 213-228.

[ back ] 39. For a general survey, see Fraser 1972 II 674, with reference to the discussion of Pfeiffer 1968:239-242. Pfeiffer sums up this way (p. 242): “Even this short survey of scanty evidence gives the impression that Crates was a serious scholar capable of displaying solid learning who did not disregard the results of previous research, even though it was the work of scholars who were his opponents in principle.” See also in general Wachsmuth 1860 and Broggiato 2001.

[ back ] 40. The editorial disagreements of Crates and Aristarchus extend to the textual tradition of Hesiod and other classics as well. For an illuminating example, see the discussion of the variant readings mentioned by Crates (and also by Zenodotus) at Hesiod Theogony 5, as summarized by West 1966:153-154. I do not share, however, West’s assumptions about the methodology of Crates: for example, West p. 208 accuses Crates, without justification, of making up a verse (a variant of the medieval manuscript reading adopted by West for Theogony 142).

[ back ] 41. See in general McNamee 1992; also McNamee 1981. Diogenes Laertius 3.66 says that owners of texts featuring margins marked up with signs could get away with charging fees from readers who wished to have access to such texts. I will discuss these signs in ch. 2.

[ back ] 42. Nagy 2000a.

[ back ] 43. Nagy 2000a. HTL ch. 3.

[ back ] 44. Relevant are the scholia A for Iliad IX 222b; also the scholia A for Iliad II 665a and XVI 467c1. {On the scholia for Iliad IX 222, see C11 p. 86B and in PP. On the scholia for Iliad II 665 and XVI 467, see C11 p. 86B.}

[ back ] 45. West 2001:61n44 gives a list of the references by Didymus to two distinct readings in these two ekdoseis of Aristarchus.

[ back ] 46. This interpretation, as entertained by West 2001:62-63, is criticized by Montanari 2002a:126n24.

[ back ] 47. Montanari 2002a:126; see also Nagy 2003b:488, with reference to Montanari 1998:11-20. [[pleiones_ekdoseis]]

[ back ] 48. Montanari 2002a:126.

[ back ] 49. Montanari 2002a:125.

[ back ] 50. Montanari 2002a:125. The reading given in the edition of Aristophanes was evidently Ἰλίου, which is likewise given in the lēmma by scholia A for Iliad II 133a; it is also the reading of the vast majority of the medieval manuscripts. Evidently it is the Koine reading, as opposed to the variant Ἴλιον found by Aristarchus and recorded in his earliest hupomnēmata. [[variants<hupomnēmata]]

[ back ] 51. Montanari 1998:11-20. In Nagy 2000a, reviewing West 1998b, I summarized the relevance of the observations made by Montanari.

[ back ] 52. Montanari 1998:19. I added in Nagy 2000a the following remark. “But I am not sure that we need to infer, as Montanari does (p. 19), that the contents of D2 were written into the same ‘copy’ that contained the contents of D1.”

[ back ] 53. The expression pleiones ekdoseis ‘a number of editions’ with reference to the editing of the Conica by its author, Apollonius of Perga, is comparable to the term pleiones in the Homeric scholia, with reference to ekdoseis available to Aristarchus (on which see PP 116n48). [[pleiones_ekdoseis]]

[ back ] 54. My translation of this passage differs from that of Cameron 1990:117.

[ back ] 55. Villoison 1788:xxvii, followed by Wolf 1795 / 1985:195 (ch. 47). See also Apthorp 1980:132 and West 2001:62n50.

[ back ] 56. For a comparable though slightly different formulation, see Montanari 2002a:127, citing Pfeiffer 1968:217: “Whether Didymus was able to work on copies of these original διορθώσεις of Aristarchus and of his monographs, the συγγράμματα, is an insoluble problem.” See also West 2001:66, who says that the two ἐκδόσεις of Aristarchus “were not claimed [by Didymus] to be the master’s autographs and accordingly could not be relied on absolutely.”

[ back ] 57. Strabo, born around 64 BCE, refers to Aristonicus as a contemporary of his (1.2.31C38). See West 2001:47. For an edition of the fragments of Aristonicus’ work on the Iliad, see Friedländer 1853.

[ back ] 58. West 2001:65.

[ back ] 59. West 2001:65 says about the marginal signs of Aristarchus: “they had probably been transcribed into many copies, any of which he [= Aristonicus] could use as the basis for his work.”

[ back ] 60. West 2001:65.

[ back ] 61. I need to emphasize that the Aristarchean standard does not always favor the variants that Aristarcheans found in the master’s hupomnēmata. Whenever Aristarchus expresses a distinct preference for the Koine reading over a reading he adduces in his hupomnēmata, then the base texts of the post-Aristarchean ekdoseis can also be expected to show the Koine reading. [[variants<hupomnēmata]]

[ back ] 62. So West 2001:47-49.

[ back ] 63. West 2001:48n8 remarks: “It is curious that these three Alexandrians who went to Rome, Aristonicus, Philoxenus, and Selecucus, all wrote works about Aristachus’ critical signs.”

[ back ] 64. I have already cited two examples at earlier stages of my argumentation. The first was a citation from hupomnēmata of Aristarchus for the Homer edition of Aristophanes (Didymus via scholia A for Iliad II 133a ἐν τοῖς κατ’ Ἀριστοφάνην ὑπομνήμασιν Ἀριστάρχου) and the second was a citation from hupomnēmata of Aristarchus that had been ‘corrected for greater accuracy’ (Didymus via scholia A for Iliad II 111b ἔν τινι τῶν ἠκριβωμένων ὑπομνημάτων γράφει ταῦτα κατὰ λέξιν). [[variants<hupomnēmata]]

[ back ] 65. What I say here differs slightly from what I said in Nagy 2003b:497-498, which I revised already in HTL 102-104.

[ back ] 66. On the post-Aristarchean practice of transferring information from hupomnēmata into the margins of the edited text, see Wilson 1967 and 1984. See also Cameron 1990:117-118, who stresses that abridgement is generally necessitated by such transfers. [[variants<hupomnēmata]]

[ back ] 67. PP 122-124.

[ back ] 68. A classic example of such an absolutizing sense is the Homeric Greek skaioteros as opposed to dexios, meaning ‘left’ as opposed to ‘right’, not ‘more left’ as opposed to ‘right’; conversely, dexiōteros as opposed to skaios means ‘right’ as opposed to ‘left’, not ‘more right’ as opposed to ‘left’. See Benveniste 1948.

[ back ] 69. I owe this reference to Paul Psoinos.

[ back ] 70. In the Iliad, we see an attestation of the word kharis with reference to myrtle blossoms. Analysis in HPC II§425.

[ back ] 71. Translation after Usher 1974:37-39.

[ back ] 72. Translation after Usher 1974:41.

[ back ] 73. Haslam 1997:71 and n35.

[ back ] 74. Haslam 1997:71.

[ back ] 75. This paragraph derives from HTL 22n92.

[ back ] 76. Haslam 1997:71, following S. West 1967:26. It may be objected that many of the readings of early Ptolemaic papyri reflect readings that Aristarchus would have considered khariestera, not koina.

[ back ] 77. Janko 1992:26.

[ back ] 78. PP 148-149; see also Muellner 1976:58-62 and Bird 1994.

[ back ] 79. PP 118. When I say learned editions here, I include the politikai.

[ back ] 80. PP 129-132.

[ back ] 81. I think it is unnecessary to emend the manuscript reading from κατέκτα to κατέκτα<ν>, as we find it in Erbse’s edition.

[ back ] 82. Chantraine 1958:381.

[ back ] 83. Chantraine 1958:380.

[ back ] 84. I am translating ἀνέγνω here as a gnomic aorist; the readership that ‘reads’ (ἀνέγνω) the text this way is exemplified by an authority like Aristarchus.

[ back ] 85. I am also translating εὑρέθη as a gnomic aorist; this kind of reading ‘is found’ (εὑρέθη) by an authority like Aristarchus.

[ back ] 86. Again I am also translating εὑρέθη as a gnomic aorist.

[ back ] 87. Again I am translating ἀνέγνω as a gnomic aorist.

[ back ] 88. The form λέλυνται, which comes after σέσηπε in the Homeric verse, is here placed before σέσηπε; I infer that the focus of the commentator’s interest is on the plural λέλυνται as an example of both ‘Homeric’ usage and ‘our’ usage.

[ back ] 89. The variant reading of Zenodotus here (σῆς ἕνεκ’ ἀγγελίης) does not affect the point made by Aristarchus about the syntax of ἀγγελίης – except to the extent that the reading of Zenodotus requires the ‘customary’ interpretation of a genitive singular of ἀγγελίη whereas the reading favored by Aristarchus (σεῦ ἕνεκ’ ἀγγελίης) leaves it open whether ἀγγελίης is the genitive singular of ἀγγελίη or a nominative singular ἀγγελίης. The reading as produced in the lēmma can still be a Koine reading even if the interpretation of ἀγγελίης reflects a non-Koine usage.

[ back ] 90. On the semantics of apodekhesthai as ‘accept as a genuine tradition’, see PH 8§§4, 7 (= pp. 217-218, 220).

[ back ] 91. For Aristarchus, I should note in passing, another example of a contrast between ‘Homeric’ and ‘Attic’ usage was the use of the plural instead of the dual, as Aristonicus reports by way of the scholia A for Iliad XIII 197.

[ back ] 92. For the idiom here, compare the wording of the scholia A for Iliad VIII 276a1: πιθανώτερον γάρ ἐστι κύριον αὐτὸ παραλαβεῖν ‘it is more reliable to take it as proper [= not common]’.

[ back ] 93. The οὕτως, indicating ‘this is the way it is read’, evidently refers to the pronunciation of smooth breathing in ἔασον, as opposed to rough breathing as in ἕασον, a way of reading that yields a different word and a different meaning: see the scholia bT for Iliad XXI 221 b1 c1 b2 c2, as opposed to the scholia A for the same verse (where it is said explicitly: ψιλῶς ‘with smooth breathing’).

[ back ] 94. For background on this word anagnōstēs ‘reader’, see PP 149-150, 168n49, 176-177, 201.

[ back ] 95. West 2001:55 thinks that Posidonius pronounced rhimphea, that is, ῥίμφεα. Such an interpretation cannot be justified, I think, on the basis of what we read in the original Greek of the scholia for Iliad VI 511. In the same discussion, West claims that Posidonius was “the author of opinions on the articulation of Zenodotus’ text.” But I see no specific association between Posidonius and the text of Zenodotus per se. In fact, the reading of Posidonius represents an alternative to the reading of Zenodotus. What the Greek says in the scholia for Iliad VI 511a is that Posidonius reads ε where Zenodotus had read ἑὰ. Further, Posidonius pronounces the …a-e of ῥίμφα ἑ in a way that parallels the way he pronounces the ē-e of ἠὲ at Iliad V 228.

[ back ] 96. I follow here the attribution given by Erbse.

[ back ] 97. This way, ῥίμφ’ ἑὰ γοῦνα φέρει allows the subject of φέρει to agree with the subject marked by the following participle πεποιθώς.

[ back ] 98. Scholia for Iliad VI 511a: ὁ δὲ Ὅμηρος ὑπὸ τῶν γονάτων καὶ ποδῶν φέρεσθαι λέγει ‘in Homeric idiom, one can be “carried” (φέρεσθαι) by one’s knees and feet’.

[ back ] 99. The sequence φημὶ τὸ ῥίμφα †<ι> in the scholia for Iliad VI 510-511a1 may perhaps be restored as a direct quotation from Posidonius: φημὶ τὸ ῥίμφαε ‘I say rhimphae’.

[ back ] 100. As we saw earlier, Eustathius (Iliad commentary vol. 2 pp. 377 / 378 lines 19-21 / 1-3) expresses his own preference for this version, ῥίμφα ἑ γοῦνα φέρει, which he contrasts negatively with what he describes as the strained ‘Attic’ version represented by ῥίμφ’ ἑὰ γοῦνα φέρει, which results from what he calls epidiorthōsis.

[ back ] 101. See my note above on the concept of ἡ … κοινὴ ἀνάγνωσις ‘the Koine reading’.

[ back ] 102. West 2001:55.

[ back ] 103. See the notes above.

[ back ] 104. See Cameron 1990:103-107.

[ back ] 105. On Isidore of Miletus, the ‘Milesian Engineer’, who along with Anthemius of Tralles was commissioned by Justinian in 532 CE to design and build the new Hagia Sophia, see Cameron 1990:103-104. On the editing procedures of Eutocius, see the data assembled by Cameron pp. 116-118 with reference to Eutocius’ edition of the Conica of Apollonius of Perga: scholars who prepared a ninth-century Arabic translation collated Eutocius’ edition of Books 1-4 with earlier versions of the text and found that Eutocius made important ‘corrections’. In his commentary, Eutocius often refers to variants found ‘in some copies [antigrapha]’ (ἐν τισιν ἀντιγράφοις).

[ back ] 106. PP 174-176. See also the contexts of paranagignōskein assembled by Cameron 1990:124-125.

[ back ] 107. Cameron 1990:118-120. Commenting on the Doric dialectal features and the technical terminology inherent in the textual tradition of Archimedes, Cameron p. 120 offers the opinion that “both dialect and terminology had already been largely modernized before Isidore’s day.”

[ back ] 108. Cameron 1990:118.

[ back ] 109. Cameron 1990:106.

[ back ] 110. Cameron 1990:111.

[ back ] 111. Cameron 1990:115, who adds that he has no way of determining whether Hypatia also edited the text of the remaining ten books of the Almagest.

[ back ] 112. I revise the interpretation I gave in the original version of PP 175-176n83 for the formula ἐκδόσεως παραναγνωσθείσης plus dative, and I offer the following rewording (see also Davidson 2001b:410n19): “I interpret the formula ἐκδόσεως παραναγνωσθείσης plus dative to mean ‘and the edition [ekdosis] was read out loud (by reader X) for (editor) Y’.” I withdraw my translation ‘with the edition [ekdosis] read out loud, as a model, by X’. I rewrite as follows the rest of PP 175-176n83: “The ekdosis ‘edition’ in question is the text of the work about which the commentary is written, not the commentary itself, and this text is ‘corrected’ by the one who ‘has it read out loud’ [corrected from ‘reads it out loud as a model’] (this editor is sometimes but not always the same person as the commentator), with variant readings placed at the margins of the ‘edited’ text (Cameron 1990:116-117). I suggest that the idea of ‘reading out loud’ is a way of expressing the process of establishing a definitive text as if it were a speech-act.”

[ back ] 113. The next three paragraphs are a revised version of what I offer in HTL 20-21.

[ back ] 114. So I agree basically with Allen 1924:317.

[ back ] 115. Jerome Epistles 106.2, as discussed by Allen 1924:317, 319. See in general Neuschäfer 1987. See also Lührs 1992:8 on Origen’s editorial policy of avoiding personal emendations or conjectures in editing the text of the Septuagint.

[ back ] 116. On Jerome’s concept of eruditio, see the next note.

[ back ] 117. Jerome’s concept of ‘the books of the erudite [eruditi]’ corresponds to the Aristarchean concept of the khariesterai.

[ back ] 118. PP 187-200, following (in part) Jensen 1980:109. This possibility is entertained but rejected by Haslam 1997:71.

[ back ] 119. See Lycurgus Against Leokrates 102 and Demosthenes 18.170; also Isaeus 7.16, on the care taken in the legitimizing of texts recorded by the state of Athens: only after full verification ‘are they to be written down into the koinon grammateion’ (εἰς τὸ κοινὸν γραμματεῖον ἐγγράφειν). This usage confirms that the expression ἐν κοινῷ goes with both γραψαμένους and φυλάττειν in “Plutarch” Lives of the Ten Orators 841f (as discussed in PP 175n77): in this context, what is recorded and preserved by the Athenian state in standardized form is the corpus of tragedies attributed to Aeschylus, Sophocles, and Euripides. See also Bollack 1994.

[ back ] 120. PP 189. For the moment, this formulation is unaffected by the existing distinction between the feast of the Great Panathenaia, celebrated every fourth year, and the feast of the Lesser Panathenaia, celebrated in the other years.

[ back ] 121. PP 187-190.

[ back ] 122. I am paraphrasing here (and disagreeing with) the formulation of Haslam 1997:71.

[ back ] 123. The point that I am making here can easily be misunderstood if we translate koinē simply as ‘vulgate’, without further clarification.

Prolegomena: A classical text of Homer in the making

Pⓢ1. The Homeric Koine

Pⓢ2. Twin books about six ages of Homeric reception

Pⓢ3. An evolutionary model

Pⓢ4. Metaphors of rigidity and definitiveness

Pⓢ5. Aristarchus and the Homeric Koine

Pⓢ6. Homeric koinai and khariesterai

Pⓢ7. “Vertical” variation and the numerus versuum

Pⓢ8. “Horizontal” variation

Pⓢ9. Homeric editions attributed to Aristarchus

Pⓢ10. The base text of Aristarchus

Pⓢ11. The kharis of the “real ” Homer

Pⓢ12. The Koine standard reaffirmed

Pⓢ13. Reading out loud the Homeric Koine

Pⓢ14. Reading Homer out loud to Aristarchus

Pⓢ15. Conclusions: the Homeric Koine as a classical text

Footnotes

About CHS

Programs

Publications

CHS Greece

Who We Are:

Prolegomena: A classical text of Homer in the making

Pⓢ1. The Homeric Koine

Pⓢ2. Twin books about six ages of Homeric reception

Pⓢ3. An evolutionary model

Pⓢ4. Metaphors of rigidity and definitiveness

Pⓢ5. Aristarchus and the Homeric Koine

Pⓢ6. Homeric koinai and khariesterai

Pⓢ7. “Vertical” variation and the numerus versuum

Pⓢ8. “Horizontal” variation

Pⓢ9. Homeric editions attributed to Aristarchus

Pⓢ10. The base text of Aristarchus

Pⓢ11. The kharis of the “real ” Homer

Pⓢ12. The Koine standard reaffirmed

Pⓢ13. Reading out loud the Homeric Koine

Pⓢ14. Reading Homer out loud to Aristarchus

Pⓢ15. Conclusions: the Homeric Koine as a classical text

Footnotes

About CHS

Programs

Publications

CHS Greece

Who We Are:

Sign In

Forgot Password?