I.2 From σύνδεσμοι to particulae

2.1 Introduction

§1. The group of lexical items generally called particles has never been clearly defined. Despite many attempts, no one has yet succeeded in isolating them based either on their form or their function. [1] The ancient grammarians had the same problem, barring the fact that they did not have to wrestle with the term particles, as it was not defined as a single word group yet. From the fourth century BCE onward, the lexical items under consideration were gathered among the conjunctions and in some cases the adverbs. Establishing the history of the scholarship on particles is complicated by the fact that our knowledge of the early study of grammar is sketchy at best. Therefore, the genesis of grammar as a field of study needs to be outlined, before we will attempt to reconstruct the ancient views on particles.
§2. In the following, I briefly outline the history of ancient Greek grammar in antiquity, first presenting in broad terms the vexed nature of our evidence, then focusing on the discussion surrounding the Téchnē Grammatikê attributed to Dionysius Thrax. I then turn to descriptions of particles themselves, starting from Aristotle’s and Diogenes of Babylon’s reported descriptions of the σύνδεσμος, the ancient word group closest to the modern notion of particle. A considerable section is devoted to σύνδεσμοι in the scholia, followed by the few extant discussions of the topic in the ancient grammarians up to Apollonius Dyscolus. Apollonius Dyscolus’ Περὶ συνδέσμων, the only extant monograph on the word group from antiquity, then receives ample attention, before turning to the ancient grammarians after Apollonius Dyscolus through late antiquity and the medieval period, and finally early modern works on particles.
§3. Since it is my purpose to cover discussions of particles spanning almost two millennia, the chapter is necessarily selective. I have, however, tried to represent every kind of relevant source, offering closer analysis wherever the content warranted it. Finally, this chapter is meant as a background for the “Online Repository of Particle Studies” (volume V) and the research presented in the individual volumes. As such, it aims to offer both an idea of the starting point for early modern researchers, as well as a broader perspective on the degree of innovation – or lack of it – portrayed by later authors.

2.2 Early study of grammar

§4. It is now broadly agreed that the Stoic philosophers played a crucial role in developing thought on language and in defining the parts of speech. [2] In the third and second centuries BCE, the Alexandrian librarians Aristophanes of Byzantium and Aristarchus adopted the terms coined by the Stoics in their commentaries on ancient texts. Since direct sources for the Stoic philosophers are lacking, the Alexandrian commentaries form an important source for this early period. However, these commentaries did not survive directly, but eventually ended up as separate notes in the margin of manuscripts of literary texts. As a result, there is little certainty if the terminology found in these scholia can actually be traced back to the third- and second-century BCE. [3]
§5. In an important work on the grammatical thought of Aristarchus, Matthaios argues that the Alexandrian librarian already distinguished the eight parts of speech that would become standard later. [4] This claim notwithstanding, the several layers through which the notes of the Alexandrian commentators are filtered keep us from establishing firmly how far they had progressed to an actual theory of grammar and/or syntax. In fact, the evidence from contemporary and later sources suggest that grammar was not yet established as an independent field of study in the third and second century BCE, neither as a theoretical exercise in philosophy nor as an ad hoc terminology in service of philology. [5]
§6. The scarcity of material between the Alexandrian commentators and the fully developed study of grammar by Apollonius Dyscolus in the second century CE (roughly four hundred years later), makes it practically impossible to know when exactly the decisive step to grammar as an autonomous field was taken. The evidence we do have has led scholars to posit an important shift in attitude in the first century BCE. [6] Taylor claims that Varro’s Lingua Latina, composed in this period, is the first work to clearly show an application of the Stoic ideas in an independent study of grammar. But as we have lost all of the books on syntax, the part of his work that might best have substantiated or contradicted this claim, it is hard to agree or disagree with Taylor, even if his argument is persuasive. [7]
§7. The material from the second century CE shows that the study of grammar had been established by this point at the latest. The two crucial bodies of evidence from the first two centuries of our era are the grammatical papyri and the works of Apollonius Dyscolus. [8] The papyri contain basic short textbooks on grammar, often referred to as “school grammars.” Although they differ from each other in form and content, they are all clearly part of a certain genre. The work of Apollonius Dyscolus stands at the other end of a methodological spectrum. Rather than little handbooks, his treatises are scholarly discussions of different aspects of grammar into the very details. [9]
§8. Current scholarship generally posits a roughly linear development from the pragmatic textual notes by third century BCE commentators to the full-fledged grammatical analysis in Apollonius Dyscolus. However, this hypothesis has two significant problems: the first is that it based upon scanty evidence, and the second is that the evidence we do have is hard to date. Central to both issues is the Téchnē Grammatikê attributed to Dionysius Thrax. If this treatise is indeed accepted as the work of the second century BCE grammarian, it is proof of the early systematic study of language, but recently numerous scholars have challenged its authenticity, dating at least part of the Téchnē to after Apollonius Dyscolus. The question of the authenticity of this text is therefore a crucial point in any discussion of the development of the study of grammar. In the following section we discuss the most important primary and secondary material pertaining to the issue.

2.3 The Téchnē attributed to Dionysius Thrax

§9. Dionysius Thrax was a student of Aristarchus and worked from the second to first century BCE. Several works are attributed to him, but not a single treatise has survived in its entirety, except for the Téchnē Grammatikê, transmitted in multiple manuscripts and some papyri. [10] The Téchnē is a grammatical treatise that consists of twenty chapters: the first provides a definition of grammar, the second to fourth discuss prosody, the fifth traces the history and etymology of rhapsody, and the remainder of the work provides an overview of word classes and their forms and functions. In the second century CE, Sextus Empiricus quotes from the first four chapters of the treatise. The fact that this part of the Téchnē was attributed to Dionysius Thrax suggests that it does actually go back to the second century BCE author. There is no such direct or indirect evidence for the rest of the treatise. [11]
§10. The Téchnē’s authenticity was doubted in antiquity, but these doubts were laid to rest in the nineteenth century. [12] The problem was not picked up again until the second half of the twentieth century. The current discussion is roughly divided into two camps, with several scholars attempting – with differing degrees of success – to reconcile the two. On one side are the scholars who rekindled the discussion of the Téchnē’s authenticity, led by Di Benedetto and including Pinborg, Siebenborn, and Fehling. [13] They argue that the Téchnē as we have it cannot have been a second century BCE composition. Instead, they believe that it obtained its current form only in the third or fourth century CE. Most strongly opposed to the idea that the Téchnē is a compilation, with different parts authored at different times, are Pfeiffer, Erbse, and more recently, Wouters. [14] Wouters bases his conclusions on grammatical papyri dating back to as early as the first century CE. [15] Schenkeveld, with help from Wouters’ grammatical papyri, follows a middle road that has come to be the communis opinio. [16]
§11. Schenkeveld pays particular attention to the problems of the Téchnē’s internal coherence and concludes that a large part of the treatise (chapters 6-20) is more likely to have been the product of third century CE scholarship than the work of Dionysius Thrax. [17] On the other hand, he regards it as beyond doubt that Thrax wrote a work on grammar that “was of a systematic character.” [18] In his view, the first four chapters of the Téchnē formed part of this work, but the rest of the Téchnē in its current form cannot be regarded as authentic. Robins has proposed that exactly because it was an actual grammar textbook the Téchnē developed and changed continuously. [19] He points out that modern-day textbooks are likewise constantly updated, while generally retaining the name of the original author; he uses this analogy to explain the discrepancies between the ideas expressed in the Téchnē and the time in which Thrax lived. Against Robins’ hypothesis, one might argue that one would expect continual development to grant more internal coherence to the work, especially between the first parts and the rest.
§12. In spite of remaining doubts, the most prudent conclusion appears to be that only the first four chapters of the Téchnē can be attributed to Dionysius Thrax with any certainty, [20] and the other parts are later additions or redactions. As Pagani notes, this is not a completely negative conclusion. If we accept part of the work as authentic, this means that Thrax did write a work on grammar in the second century BCE, albeit one that has largely been lost to us. [21] The existence of this grammatical work by Thrax reflects a growing interest in the theoretical aspects of language and linguistics, an interest that would continue to develop in the following centuries, and culminate in Apollonius Dyscolus’ works. [22]
§13. To conclude, language and its components steadily became more of an object of study from the third century BCE onwards. However, the establishment of grammar as an object of study per se appears to have been a development of the first century BCE, leading eventually to Apollonius Dyscolus’ seminal work. [23] The following material on the study of particles will reinforce the argument that Thrax’s Téchnē, whatever its original form, was an important stage in the development of grammatical studies but not its culmination.

2.4 Early Definitions of σύνδεσμοι

§14. The very first authors who wrote about the parts of speech referred to most of the words that we consider particles as σύνδεσμοι. Grammarians writing in Latin later rendered the term as coniunctio. However, despite the formal likeness this is not the equivalent of the English “conjunction.” After all, in English grammar the word conjunction has very specific connotations that in many cases do not apply to the words that the Greeks and Romans gathered respectively under the terms σύνδεσμος and coniunctio. Instead, it is more productive to use Swiggers and Wouters’ translation “combiner,” since this term reflects the neutral nature of σύνδεσμος, something that “binds together.” [24]
§15. Initially σύνδεσμος was used to cover anything from conjunctions to prepositions to interjections to noun phrases, in addition to the words we call particles. The term had such a wide application because it appears to have been coined to cover the words that were neither noun/adjective (ὄνομα), verb (ῥῆμα), nor adverb (ἐπιρρῆμα/μεσοτῆς). The σύνδεσμος was probably the fourth lexical category to be introduced, just before or at the same time as the ἄρθρον (“joint”), the category that would later become the article. Aristotle is the first we know of to distinguish the σύνδεσμος and ἄρθρον in addition to the noun/adjective, verb, and adverb. [25] As the following definition shows, σύνδεσμος was by no means clearly defined – at least grammatically – at this point: [26]
(t1)
σύνδεσμος δὲ ἐστιν φωνὴ ἄσημος ἣ οὔτε κωλύει οὔτε ποιεῖ φωνὴν μίαν σημαντικὴν ἐκ πλειόνων φωνῶν πεφυκυῖαν συντίτεσθαι καὶ ἐπὶ τῶν ἄκρων καὶ ἐπὶ τοῦ μέσου ἣν μὴ ἁρμόττει ἐν ἀρχῇ λόγου τιθέναι καθ᾽ ἁυτόν, οἷον μέν ἤτοι δέ. ἢ φωνὴ ἄσημος ἣ ἐκ πλειόνων μὲν φωνῶν μιᾶς σημαντικῶν δὲ ποιεῖν πέφθκεν μίαν σημαντικὴν φώνην.
Poetics XX, 6, 1456b38-1457a6
A combiner is a non-significant sound which neither precludes, nor brings about, the production of a single significant sound that by nature is composed of several sounds [i.e. an uttered sequence], and which can be used at either end and in the middle, but which it is not appropriate to place at the beginning of an utterance on its own, [27] e.g. μέν ἤτοι, δέ. Or a non-significant sound, which by nature produces, as a result of [uniting together] several sounds that are significant, a single significant sound.
Swiggers and Wouters 2002:108
The passage is broadly regarded as corrupt and highly opaque, [28] but what we may roughly deduce is that Aristotle here presents the σύνδεσμος as a word that can combine other significant (signifying) sounds into a significant whole. As is clear from the context, he is not discussing parts of speech (μέρη τοῦ λόγου) here, but parts of the lexis (μέρη τῆς λέξεως) – “diction” in Swiggers and Wouters’ translation – one of the six components of (good) tragedy. As such his focus is not to offer “a systematic treatment of word-classes, but [to offer] us a list of definitions of elements constitutive of the λέξις, (oral) poetic expression.” [29] A definition of ἄρθρον follows this passage; like the definition of σύνδεσμoς it too shows that ἄρθρον is not yet conceived as denoting only the article. In view of the opaque nature of the passage, as well as Aristotle’s non-grammatical concerns here, it is unclear which words fall in the category σύνδεσμος. [30] Swiggers and Wouters conclude tentatively that the connective and disjunctive particles would be covered by the term σύνδεσμος, whereas expletive particles would fall under ἄρθρον.
§16. The next extant definition of σύνδεσμος appears to go back to Diogenes of Babylon, a Stoic philosopher who lived in the third and second centuries BCE. In his Lives of the Philosophers the third century CE biographer Diogenes Laertius reports that the Stoics Chrysippus (third century BCE) [31] and Diogenes of Babylon distinguished five parts of speech: ὄνομα (proper name), ῥῆμα (verb), προσηγορία (apellative), ἄρθρον (joint), and σύνδεσμος (combiner); the μεσότης (adverb) was added by Diogenes of Babylon’s student Antipater of Tarsus. [32] Diogenes Laertius then goes on to give definitions, introducing the first one as follows: Ἔστι δὲ προσηγορία μὲν κατὰ τὸν Διογένην (“the προσηγορία, according to Diogenes [of Babylon], is:”). The explicit attribution of this definition suggests that the subsequent definitions have their origin with Diogenes of Babylon as well. The definition of σύνδεσμος is as follows:
(t2)
σύνδεσμος δέ ἐστι μέρος λόγου ἄπτωτον, συνδοῦν τὰ μέρη τοῦ λόγου
Diogenes Laertius Lives of the Philosophers 7.58 [33]
a combiner is an indeclinable part of speech, to bind together the parts of speech
Clearly, the category has been redefined, and a crucial step has been taken from Aristotle’s philosophical comments on what the lexical category is to Diogenes’ attempt to describe the morphology (ἄπτωτον) and function of its members. The description provided in the text, “indeclinable” (ἄπτωτον), is a characteristic of particles that seems natural to the modern reader, but barring the Stoic tradition the feature was, in fact, not a requisite part of the category’s definition in antiquity until Priscian (sixth century CE). [34]
§17. Apart from the definition transmitted by Diogenes Laertius, the sources from the period between Aristotle and Apollonius Dyscolus (second century CE) cannot be securely assigned to an author or even a certain period. There are two bodies of evidence to draw from: (1) the Homeric scholia, and (2) grammatical handbooks, such as the Téchnē and some grammatical papyri, authored before Apollonius Dyscolus’ work. In the following sections we first discuss the scholia, then the grammatical handbooks; because of the problems of dating the material, it cannot be presented in a strictly chronological way.

2.5 The scholia

2.5.1 Terminology

§18. The term σύνδεσμος is the closest thing to an equivalent of the modern “particle.” However, particle comes from the Latin particula, whereas σύνδεσμος became coniunctio in Latin. There is, then, a mismatch of terminology, and this led Schenkeveld to investigate the use of the ancient Greek equivalent of particula: μόριον. After examining the use of the term in early Greek works about language, he concludes that μόριον was in fact never used as particula would be. [35]
§19. The scholia to the Iliad may be adduced to illustrate this. The word μόριον, when applied to language, almost without exception denotes an affix or suffix, such as the alpha privans, [36] νη privans, [37] the suffix –θεν, and several other syllables at the beginning and end of words. Single words called μόρια include ὡς, [38] ἔτι, [39] εὖ, [40] articles or pronouns, [41] τῶς, [42] and μα. [43] The only case where μόριον is used in a description of a word that we consider a particle is in the scholion to A 210:
(t3)
A 210 ἀλλ’ ἄγε.
ἀλλὰ φέρε. ἔστι δὲ ἐπίῤῥημα παρακελεύσεως, ἢ ἐπιρρηματικὸν μόριον.
[this means] “but come”, and it is an exhortative adverb, or an adverbial μόριον.
This scholion hints at valuable insights, but the scholiast is not very helpful. First of all, it is not clear if his comment explains only ἀλλά or the whole construction ἀλλ᾽ ἄγε. The paraphrase ἀλλὰ φέρε suggests the latter, but it seems completely superfluous: surely ἄγε would be as natural as φέρε to the audience in later antiquity, if not more so. More problematic is the following description of “exhortative adverb.” Its sense is clear enough, but it is not entirely clear how ἐπίρρημα can be used to describe the combination ἀλλ᾽ ἄγε. ἀλλά on its own might be regarded as an adverb or “adverbial” in the sense that it co-occurs with a verb form (the imperative ἄγε). By extension the comment “or it is an adverbial μόριον” would refer just to ἀλλά, making this instance the only one in the scholia where μόριον describes a particle. [44] The weight of the evidence in the scholia, then, suggests that in early linguistic discussions σύνδεσμος was the normal term to refer to the words we would call particles. Note however, that although most particles were called σύνδεσμοι, not all σύνδεσμοι were what we would call particles.

2.5.2 σύνδεσμοι in the scholia

§20. The Iliad scholia, marginal notes found in several manuscripts, display a wealth of insights on σύνδεσμοι. [45] These marginalia contain textual commentaries by Aristarchus, Aristophanes of Byzantium, and Zenodotus – among many named or anonymous others. These librarians of the great Mouseion in Alexandria edited the most important archaic and classical Greek texts in the third and second centuries BCE. They worked in the same period as the stoic philosophers, by whom they seem to have been influenced. [46] As a source, the scholia are problematic for two reasons: first, only a small number of scholia can be traced back confidently to a specific author, and second, even if the author is established we have no way of knowing beyond doubt that the wording of the note is original. These issues make it hard to determine the date of the terminology used, which is an especially pertinent question in the current attempt to sketch the development of the study of σύνδεσμοι.
§21. It will be useful at the outset to analyze a single scholion in detail to give a general sense of the kind of discussions we find in the scholia. Consider the scholion to 1.219a.
(t4)
1        [ἦ καὶ ἐπ’ ἀργυρέῃ] τοῦτο τὸ ἦ ψιλῶς λεγόμενον καὶ περισπωμένως
          δηλοῖ σύνδεσμον παραπληρωματικὸν τὸν ἤτοι, ἴσον τῷ
          δή, οἷον „ἐπειῆ πολὺ φέρτερόν ἐστι“ (1.169), καὶ παραλλήλως „ἦ δὴ
          λοίγια ἔργα“ (1.518). δηλοῖ δὲ καὶ ἀπορρηματικὸν σύνδεσμον· „ἦ οὐχ
5        Ἑλένης ἕνεκ’ ἠϋκόμοιο; / ἦ μοῦνοι φιλέουσ’ ἀλόχους μερόπων ἀν-
          θρώπων / Ἀτρεῖδαι;“ (9.339—41). δηλοῖ δὲ καὶ τὸ ἔφη· „ἦ καὶ κυανέῃσιν“ (1.528). |
          καὶ σεσημείωται Ἀρίσταρχος ὅτι ὁ μὲν Ὅμηρος ἀεὶ
          ἐπὶ προειρημένοις λόγοις ἐπιφέρει τὸ δηλοῦν τὸ ἔφη, ὡς ἐπὶ τοῦ προ-
          κειμένου, ὁ δὲ Πλάτων μετ’ αὐτὸ ἐπιφέρει τὸν λόγον. |
10      ψιλούμενον δὲ καὶ βαρυτονούμενον δηλοῖ σύνδεσμον διαζευκτικόν· „ἦ εὖ ἠὲ κακῶς“
          (2.253). ἔστι δὲ ὅτε καὶ ἀντὶ συναπτικοῦ τοῦ εἴ τίθεται, οἷον „οὐδ’ ἀφα-
          μαρτοεπής, ἢ καὶ γένει ὕστερος ἦεν“ (3.215). ποτὲ δὲ παρέλκει· „ἀλλὰ
          τίη με ταῦτα παρεξερέεσθαι ἕκαστα;“ (10.432). δασυνόμενον δὲ καὶ
          ὀξυτονούμενον ἄρθρον προτακτικὸν δηλοῖ· „ἣ δ’ ἑτέρη θέρεϊ προρρέει
15      εἰκυῖα χαλάζῃ“ (22.151). δηλοῖ δὲ καὶ ἄρθρον ὑποτακτικόν, οἷον „ἣ
          μυρί’ Ἀχαιοῖς“ (1.2). δηλοῖ δὲ καὶ ἀναφορικὴν ἀντωνυμίαν· „ὣς ἥ γ’
          ἀμφιπόλοισι μετέπρεπεν“ (6.109).
          δηλοῖ δὲ καὶ τὴν σύναρθρον
          ἀντωνυμίαν τρίτου προσώπου, συζυγοῦσαν τῇ ἐμή, σή. ὑποδείγματα
20      δὲ ταύτης παρ’ Ὁμήρῳ οὐχ εὑρίσκεται, ἐκ δὲ τοῦ ἀναλόγου νοεῖται·
          αἱ γὰρ πλάγιοι πᾶσαι δι’ αὐτῆς παρ’ Ὁμήρῳ σώζονται.

1        [ἦ καὶ ἐπ’ ἀργυρέῃ] this unaspirated ἦ and with the circumflex
          is the filling combiner ἤτοι, similar to
          δή, as in “ἐπειῆ πολὺ φέρτερόν ἐστι” (1.169), and elsewhere “ἦ δὴ
          λοίγια ἔργα” (1.518). It is also the interrogative combiner: “ἦ οὐχ
5        Ἑλένης ἕνεκ’ ἠϋκόμοιο; / ἦ μοῦνοι φιλέουσ’ ἀλόχους μερόπων ἀν-
          θρώπων / Ἀτρεῖδαι;” (9.339—41). And it also means ἔφη: “ἦ καὶ κυανέῃσιν” (1.528). |
          And Aristarchus noted that Homer always
          uses it to mean ἔφη after the words have been spoken, as in the
          current example, while Plato starts the speech after it [sc. ἦ]. |
10      Unaspirated and with the grave accent, it is the disjunctive combiner: “ἦ εὖ ἠὲ
          κακῶς.” (2.253) And it happens that it is even placed instead of hypothetical εἰ, as in:
          “οὐδ’ ἀφαμαρτοεπής, ἢ καὶ γένει ὕστερος ἦεν” (3.215). And sometimes it is redundant:
          “ἀλλὰ τίη με ταῦτα παρεξερέεσθαι ἕκαστα;” (10.432). Aspirated 15 and
          with the acute accent, it is the prepositive article: “ἣ δ’ ἑτέρη θέρεϊ προρρέει
15      εἰκυῖα χαλάζῃ” (22.151). It is also the postpositive article, as in: “ἣ
          μυρί’ Ἀχαιοῖς” (1.2). And it is also the anaphoric pronoun: “ὣς ἥ γ’
          ἀμφιπόλοισι μετέπρεπεν” (6.109).
          It is also the possessive
          pronoun of the third person, to be added to ἐμή, σή. Signs
20      of this are not found in Homer, but from analogy it may be reasoned,
          since all the oblique cases maintain that paradigm in Homer.
This long discussion of η (ἤ, ἥ, ἦ) is found as a whole in the Venetus A manuscript, but that does not mean it was conceived in its entirety by one person. It is typical of the scholia in containing an explicit references to a specific scholar, in this case Aristarchus (second century BCE). The embedded reference to a named scholar indicates that this scholion is probably a composite, that is, authored by someone who includes information from Aristarchus while adding other information drawn either from his own experience or, more likely, from other sources. [47]
§22. If the form of the scholion is typical, its content is not. As an exception among many literary and short linguistic notes, this scholion devotes a long discussion to a word that we would call a particle. [48] The reason for this inclusion illustrates a problem that is relevant to our work. As opposed to most other particles, η is inherently ambiguous. Although small words were probably always vulnerable in the process of transmission, η is a particularly unstable lexical item, because its possible force and function depend entirely on accentuation and breathings. In performance the audience must have had no problem distinguishing between ἤ, ἦ, and ἥ, but in the transition to written versions, this disambiguation was lost. Moreover, there seems to have been a significant period where nothing was done to resolve this problem. [49] It was not until the fourth or third century BCE that we find some accentuation added in papyri. Understandably, disambiguation for performance seems to have been the main reason for the first diacritical signs and accents. [50] It was not until the third or second century BCE that Zenodotus or Aristophanes of Byzantium introduced a comprehensive system to provide literary texts with accentuation throughout. In any case, there must have been a period of unaccented written versions of the Homer epics.
§23. Eventually, fully accented editions of most canonical texts – naturally including the Homeric corpus – circulated in antiquity, but that is not the end of the story. In the transition from papyrus roll to codex, a certain font, the so-called Biblical Uncial, became generally used, from around the fourth or fifth century CE. This type of writing, all in capitals, was not well-suited to accents, so accents fell out in many cases, until writing in minuscule around the ninth century CE brought back accentuation. Around this period it appears that accentuation was added to entire texts, following the Byzantine system. [51]
§24. At two separate moments then, then, a decision had to be made about the accentuation of the texts: first when the Alexandrian scholars produced the first fully accented editions, and later when the Mediaeval copyists made the transition from capital to minuscule. This means that we read the instances of η at a double remove, through the interpretations of at least two post-Homeric scholars. The problems with η are of course not unique, but they serve as a concrete illustration of the challenges that the process of transmission posed, not only in the mediaeval and modern era, but even for scholars in antiquity.
§25. The scholion to A 219 shows that scholars in antiquity were aware of these ambiguities, and consciously made a decision to accentuate in a certain way, based on an analysis of the passage. For other particles these decisions are less problematic; exceptions are ταρ, δαί, and elided particles that are ambiguous, like τ᾽ for τε, τοι, or the article τά. A large part of the scholia that concern σύνδεσμοι discuss exactly these questions of form.
§26. In two other respects, this scholion on ἤ is representative of the scholiasts’ approach to particles. First, note that the word παραπληρωματικός occurs in line 2: the LSJ renders it as “expletive,” working as a “filler.” A typical way of describing σύνδεσμοι in antiquity, especially in the scholia, the term was used where the σύνδεσμος in question seemed redundant; we discuss the use of this term in greater detail below. Second, the scholiast of A 219 explains ἤ through analogy, comparing it to ἤτοι (l. 2) and δή (l. 2-3). Lacking a shared terminology, and probably also lacking any reason to offer in-depth analysis, giving a paraphrase or analogy is the scholiasts’ preferred method to explain σύνδεσμοι.
§27. Before outlining the general tenor of discussions of particles in the scholia, we turn to the oldest traceable author in the scholia: Aristarchus. To keep up the attempt to present the material roughly chronologically, we first discuss the notes on σύνδεσμοι that Matthaios ascribes to him. [52]

2.5.3 Aristarchus on σύνδεσμοι

§28. From the limited number of scholia attributed to Aristarchus, Matthaios tries to establish the Homeric scholar’s methods and terminology. His analysis of the complete corpus of Aristarchus’ fragments allows him to establish better whether Aristarchus, and by extension the other librarians, study grammar for grammar’s sake, or if he only uses it as a philological and exegetical tool. For the purpose of tracing the history of scholarship on particles, we focus only on his discussions of σύνδεσμοι.
§29. First there is the question of terminology. Praxiphanes, a fourth century BCE Peripatetic philosopher, reportedly already discussed redundant conjunctions, [53] but the term παραπληρωματικός probably does not go quite so far back. Similarly, when we find the term in Aristarchus’ scholia, we cannot know if the scholia represent only his thoughts or actually his ipsissima verba. The same problem applies to Aristarchus’ use of the term συμπλέκειν (Π 636a1) and συμπλεκτικός (copulative), which would later become another fixed category of σύνδεσμοι. [54]
§30. As regards the content of Aristarchus’ scholia, his treatment of σύνδεσμοι is typical for the kind of comments we find in the scholia at large: they discuss σύνδεσμοι (1) as redundant or (2) in terms of interchangeability. [55] Although scholia most commonly mark particles as redundant, [56] Matthaios argues that it cannot be established that Aristarchus regards the παραπληρωματικοὶ σύνδεσμοι as a word category. [57] He solves other troublesome instances of particles by positing that one particle is used for another, like περ used instead of δή or γε, in the scholion to A 131. Elsewhere, Aristarchus is reported as regarding δαί as an equivalent of δέ and as a connective (συμπλεκτικός). [58] More remarkable is his note that in Homer a γάρ clause often comes first in causal constructions, commonly called anticipatory γάρ. [59] This argument may go back to his teacher and predecessor as head of the Mouseion, Aristophanes of Byzantium. [60] In a similar vein, Apollonius Dyscolus reports that Aristarchus chooses the reading δαί over δ᾽αἱ (elided δέ + feminine nominative plural article αἱ), because Homer was wont to use δαί after question words. [61]
§31. Matthaios concludes that Aristarchus and the other Alexandrian scholars did contribute to the development of a study of grammar, but that they never practiced it as a goal in and of itself; [62] the selection of scholia discussed below corroborates this view, and we discuss a range of examples concerning redundancy (2.5.4) and interchangeability (2.5.5). Then we give a brief overview what appears to have been a hot topic among scholiasts, namely the difference between ἄν and κεν (2.5.6). Finally, we turn to glimpses of deeper insight hidden among the many paraphrases and dismissals (2.5.7).

2.5.4 Redundancy

§32. By far the most frequent kind of comment on σύνδεσμοι in the scholia takes the form “X is redundant.” [63] Several words may be used to describe the function of many combiners as superfluous, or as a filler. In many scholia we find the term σύνδεσμος παραπληρωματικός, a “filling combiner.” The adjective παραπληρωματικός is also the term that would become the standard in grammar treatises to describe a certain group of σύνδεσμοι that appears (to the ancient grammarians before Apollonius Dyscolus) to have no clear function. In the scholia the words commonly used to describe redundant particles are περισσός (or –ττ-) and the verb περισσεύω (or –ττ-). The other options used regularly are forms of παρέλκω, πλεονάζω, παραπληρόω, and the phrase ἐκ πλήρους. Although the words clearly do not mean exactly the same thing, they appear to be used rather interchangeably. Particles that get the predicate ‘redundant’ or ‘filler’ most often are: κε, [64] δέ, [65] περ, [66] πω/που, [67] and τε. [68]

2.5.5 Interchangeability

§33. In his work on the scholia by Aristonicus, Friedländer remarks about a note on περ used for δή: Praeter illam particulae περ cum δή commutationem paucissimas conjunctionum enallagas notatas invenimus (“Apart from this exchange of περ for δή, we have found very few exchanges of combiners noted”). [69] It is unclear if he is restricting himself only to scholia attributed to Aristonicus in this statement, but in any case offering a paraphrase in the form of another particle is a reasonably standard method in the scholia.
§34. The most frequent example of this type of comment is δὲ ἀντὶ τοῦ γάρ, δέ used instead of γάρ, as in the scholion to β 6: ἔστι δ᾽ ὅτε καὶ ἀντὶ τοῦ γάρ αἰτιολογικῶς λαμβάνεται [sc. δέ] (“It also happens that [δέ] is used instead of γάρ, with a causal sense”). [70] The terminology in this scholion is probably late, not Aristarchus or Aristonicus. [71] Similar is ἠδέ for καί (Ζ 149a, I 134a), and the other way around in Ο 670.
§35. Ιn line with the discussions of δ᾽, which could stand for δέ or δή, δέ is equated with δή. [72] In general, δή seems to have been a stable reference point in the centuries just before and after the beginning of the Common Era. In the scholion to E 258, γε is paraphrased as δή. In 1.131a, περ is said to be used instead of γε or δή. ἄρ᾽ (or τ᾽ἄρ) is also equated with δή, in 18.6b. Likewise, ἄρα is simply rendered δή in the D scholion to 1.308, [73] which may be compared to the many glosses of the form: ἦ μάν (or ἦ μέν, or ἦ ῥα, or ἦ που, or ἦ μάλα, or ἦ θήν): ὄντως δή. [74] δή clearly had some kind of emphatic force at this time, and was thought to render the different nuances contained in this list of words. ἦτοι, ἄρα, ἔπειτα, δέ, μάν, μέν, γε, περ, and τ᾽ἄρ are all at one point or another paraphrased as δή.
§36. In the same way that the scholiasts substitute δή for a host of difficult particles, ἄρα is often noted as having an alternative reading in other manuscripts. In the scholion to 18.151 we find [οὐδέ κε] ἐν ἄλλῳ “οὐδ’ ἄρα”; and likewise: 19.96a: [ἀλλ᾽ ἄρα καὶ τόν] παρὰ Ἀριστοφάνει “ἀλλά νυ καὶ τόν” and 23.362 [ἄρα] γράφεται, {οἱ δ’} “ἅμα.” Finally, a scholiast observes that εἰ may be used as an equivalent of interrogative ἆρα (21.567a).
§37. The scholion to 7.89d (and 13.622b, 23.311c, 24.488a, 24.732a) explains an utterance-initial μέν thus: τὸ μέν ἀντὶ τοῦ μήν. A scholion to the same verse (H 89c) makes it a bit more complicated and argues that μέν is used instead of δέ, and that this hypothetical δέ would have the force of δή here, i.e. μέν = δή here.
§38. Another common interchange is found in the scholia to Pindar, which argue that ἀλλά is used instead of δέ: Ol. 3.40[23], P. 4.270a[152], N. 1.59[39], Ν. 2.32b[20], I. 6.47e[35]. One scholiast (to P. 8.20[15]) reverses the interchange: δέ for ἀλλά.

2.5.6 ἄν and κε(ν)

§39. Two other particles that were often treated as interchangeable are κε and ἄν. In discussing these words, the scholia are far from consistent. There is little doubt in our minds that the different scholiasts regarded κε and ἄν as words with different functions and possibilities, but it is unclear what these differences might have been, or if they would even have agreed on what distinguished the two. What the scholiasts appear to agree on is that κε is often redundant; by contrast, they say this much less often of ἄν. This tendency may well be a result of the obscure nature of κε, a word extremely rare in Greek literature outside of Homer and Hesiod, except in instances of allusive imitation. The confusion over the particle κε emerges from the scholia to 1.175a and 5.212:
(t5)
1.175 οἵ κε με τιμήσουσι
ὅτι [75] περισσὸς ὁ κέ σύνδεσμος, ἢ τὸ τιμήσουσιν ἀντὶ τοῦ τιμήσειαν.
[the sign] because the combiner κε is redundant, or τιμήσουσιν instead of τιμήσειαν.
5.212 εἰ δέ κε νοστήσω
ἀντὶ τοῦ νοστήσαιμι, ὡς “πληθὺν δ᾽ οὐκ ἂν ἐγὼ ὀνομήνω” (2.488). περιττεύει δὲ καὶ ὁ κέ σύνδεσμος.
[νοστήσω] instead of νοστήσαιμι, like “I could not name the multitude” (Iliad 2.488). Also, the combiner κε is redundant.
The authors of these scholia are rightly confused about the forms in these two constructions, arguing in both cases that the verb form should be replaced by an optative. In the first case the argument is understandable: “either κε is redundant, or the verb form should be optative.” The second, however, is confused: “[νοστήσω] instead of νοστήσαιμι, and κε is redundant.” Possibly the scholia do not regard κε as an equivalent of ἄν. The former is often discussed as being redundant (more than any other particle), whereas for ἄν redundancy is suggested only four times in the Iliad scholia. [76]
§40. In the scholion to I (9) 262a, the interchangeability is discussed in a typical way:
(t6)
9.262 ἐγὼ δέ κέ τοι καταλέξω
ἡ διπλῆ δὲ πρὸς τὸ σχῆμα, ὅτι ἀντὶ τοῦ ἐγὼ δ᾽ἄν σοι καταλέξαιμι, ἢ περισσὸς ὁ κέ.
the diplê placed at this construction, because [this construction] is instead of ἐγὼ δ᾽ἄν σοι καταλέξαιμι, or κε is redundant.
Either the construction is a variant for ἄν plus the optative, or it is a future, and κε is simply superfluous. For the scholiasts, clearly, ἄν was the standard particle that goes with the optative, and the standard means to explain the usage of κε. However, they seem to have believed that κε could be used simply as a filler too, without any bearing on the reading of the verb. The few notes on redundant ἄν (see note 27 below) fall in the same group, but the longer discussions suggest that the scholiasts had more trouble accepting a redundant instance of ἄν than of κε. [77]

2.5.7 Noteworthy readings of σύνδεσμοι

§41. Now that we have listed some of the most common discussions of σύνδεσμοι in the scholia, it is worthwhile to highlight some less common, and especially insightful comments. Among the repetitive kinds of comments discussed above, some others reveal a nascent awareness of the possible polyfunctionality of some σύνδεσμοι. These very few notes illustrate a broader interest in particles for which we have otherwise very little evidence outside of Apollonius Dyscolus’ treatise.
§42. The quintessentially Homeric particle ἄρα clearly caused problems for the scholiasts, who knew the particle only in its classical “conclusive” sense; in most scholia the particle is simply ignored. [78] In the scholion to 16.147a, a scholiast paraphrases ἄρα as ὡς ἔοικεν, given it the force of rendering a realization by the speaker. For comparison the scholiast adduces Hesiod Works and Days 11: “οὐκ ἄρα μοῦνον,” adding that the realization marked by ἄρα may be contradictory to something previously stated or thought. Likewise, the scholion to 17.33 paraphrases οὐκ ἄρα σοί γε πατὴρ ἦν with οὐκ ἦν ὡς ἔοικέ σοι πατήρ. These readings of ἄρα may well be much more productive than the explanations found in modern standard works, which regard marking “surprise” or “conclusion” as the main function of the particle. [79]
§43. The challenge of understanding ἄρα extends to the problematic particle or particle combination τ᾽ἄρ/ταρ. [80] Modern editors vacillate between spelling τ᾽ἄρ or τάρ, and this discussion goes back to the scholia. In the scholion to 1.93, τάρ is reported as one word (τέλειος), not from τε and ἄρ, [81] and the scholiast notes: οὐ δή, “it is not δή.” This is a puzzling note, as δή in this construction (οὐ τ᾽ἄρ, which would make οὐ δή) never occurs in the Iliad, and only once in the Odyssey (7.239). [82] It does make sense if the scholiast normally takes (τ᾽)ἄρ as an equivalent of δή, to which this instance is an exception. This is supported by the scholion 18.6b, where τί τ᾽ἄρ᾽ αὖτε is paraphrased τί δὴ πάλιν. Another problem with τ᾽ἄρ/ταρ is the palaeographic similarity to γάρ—majuscule τ and γ are easily confused. In the scholion to 18.182, Didymus discusses the variants γάρ and τάρ, and decides in favour of the former (contra Aristarchus) on the basis that Homer was wont to start with γάρ. [83]
§44. Another recurring discussion in the scholia is the concept of μέν solitarium, μέν used on its own. The scholiast on 4.301a finds a solitary μέν and asks simply: ποῦ ὁ δέ; “Where is the δέ?” This question is later integrated into the larger problem of the forms μέν (Ionic) – μήν (Attic, koinê) – μάν (Doric). [84] The scholion to 7.89d, mentioned above, explains an utterance-initial μέν as τὸ μέν ἀντὶ τοῦ μήν, but this comment reveals no awareness of the dialectal connection between the words. However, scholion 15.16a does remark that μάν is Doric. [85]
§45. Finally, the scholia contain some valuable notes on the position of σύνδεσμοι. Scholiasts, when confused about the sense of a passage, have recourse to rearranging the particles in the sequence, so that logical links between clauses are attained. 1.211-212a, for example, explains the sequence: ὡς ἔσεταί περ, ὧδε γὰρ ἐξερέω with the paraphrase: ὥσπερ γὰρ ἔσται, οὕτω καὶ ἐρῶ. The scholiast speaks of a hyperbaton of γάρ, apparently to mean that γάρ is set far apart from its host clause.
§46. The comments selected above represent what we regard as the most remarkable discussions of σύνδεσμοι in the scholia. We have chosen them specifically because they anticipate a number of discussions about particles that still persist. The use of ἄρα in Homer still perplexes scholars, just as it confused the scholiasts, but their reading as “it appears” (ὡς ἔοικεν) may well be more helpful than the expression of a “lively feeling of interest” posited by Denniston. [86] The discussion surrounding τ᾽ἄρ/ταρ has recently revived because a particle “tar” has been found in Luvian. The word in Homer may thus be either a loanword or an inherited Indo-European word that has disappeared from later Greek. The scholia show that the possibility of a word ταρ was entertained even back then. The relation between μέν-μήν-μάν was discussed mostly in the nineteenth century, but our understanding of μέν in Homer still strongly depends on whether we assume a link with μήν or not. [87] Finally, the phrasing in the scholion to 1.211-212, “a hyperbaton of γάρ,” reveals an awareness of the link between a particle and its host act, or more precisely of the scope of particles. [88] The accumulated intuitions of several generations of scholiasts have thus laid the foundations for centuries of particle research.

2.6 The Téchnē and other early scholarship

§47. The particle καί may serve as an illustration of the development of the study of particles from the scholia to the early grammarians. The scholia appear to have given little attention to καί. Whenever the scholiasts find καί in an unexpected position (particularly when it is in second position and not copulative) they merely comment: “καί is redundant.” [89] One possible exception is a comment in Eustathius (to Iliad 2.827), which Erbse believes might go back to a scholion: ἢ περιττὸς ὁ καί σύνδεσμος ἢ συμπλέκει καὶ ἕτερα θεόσδοτα τῷ τόξῳ ἀγαθά (“either the combiner καί is redundant, or it conjoins also other god-given goods to the bow”). [90] The comment indicates that καί is read as “also”. It is probable that in more cases “also” was a natural reading of καί, but we find this explanation only in one scholion.
§48. Even more exceptional is the scholion to Iliad 12.301. Its approach gives a glimpse into ways of describing more complicated uses of particles before Apollonius Dyscolus. The passage from the Iliad is as follows:
(t7)
βῆ ῥ’ ἴμεν ὥς τε λέων ὀρεσίτροφος, ὅς τ’ ἐπιδευὴς
δηρὸν ἔῃ κρειῶν, κέλεται δέ ἑ θυμὸς ἀγήνωρ
μήλων πειρήσοντα καὶ ἐς πυκινὸν δόμον ἐλθεῖν·
Iliad 12.299-301
And so he went, like a mountain-born lion that was without
meat for a long time, and whose proud spirit urges him
to go and attack καί the closely built sheep-fold.
About καί in line 301 the scholion adduces the analysis and paraphrase of Dionysius Thrax:
(t8)
...καὶ ὅτι ὁ καί περισσός ἐστιν. ὁ δε Διονύσιος [Dionysius Thrax] ὅτι δύναται σημαίνειν τι πλέον, οὕτως ἐνδεὴς τροφῆς ὥστε καὶ ἐπι πεπυκνωμένον καὶ ἠσφαλισμένον δόμον ἐλθεῖν.
Scholion to 12.301 [91]
… and because καί is redundant. Dionysius [Thrax], however, [argues] that it can signify something more: “so hungry for food that even to a fenced and secured fold he went.”
The scholiast proposes that καί is redundant here, meaning it should not be translated: “to go and attack the closely built sheep-fold.” Dionysius, conversely, proposes a paraphrase containing the construction οὕτως … ὥστε (so hungry … that) but without finding an adequate synonym for καί. We propose that this is an attempt by Dionysius to render what we now call the scalar function of καί with the addition of οὕτως … ὥστε. With this paraphrase, then, Dionysius gives us a first attempt to express the force of καί as a scalar particle. [92]
§49. The scholion suggests that Dionysius’ analysis was a departure from the ideas of his predecessors, and illustrates that he was devoting time and energy to σύνδεσμοι. In his readings, Dionysius was both innovative and traditional, but his work is still a few steps removed from the analyses of Apollonius Dyscolus. Unfortunately, nothing more of Dionysius Thrax’ work on σύνδεσμοι is extant, so we will never know if he was an important precursor to Apollonius.

2.6.1 Trypho

§50. The first century BCE grammarian Trypho seems to have been a recognized authority in antiquity, as Apollonius Dyscolus refers to him most out of all his predecessors, [93] twenty times to discuss issues regarding σύνδεσμοι. [94] Trypho’s notes concern particles (43-46, 59-60) as well as words we now call conjunctions (ὅτι, τηνίκα, τοὔνεκα, διότι) or adverbs (ἕκητι, χάριν). [95]
§51. The observations on σύνδεσμοι that we find in these notes are very much in line with those found in the scholia. Trypho comments on the interchangeability of σύνδεσμοι, as in the following note on Odyssey 10.501-2: τὸν γάρ ἀντὶ τοῦ δέ καὶ τὸν δέ ἀντὶ τοῦ γάρ. [96] Likewise, in another fragment he claims that μὲν γάρ is to be regarded as one particle, with the value of δέ. [97] Trypho also uses analogy to show that ἦ and δή are not different forms of the same word. [98] Most valuable, however, is the (partial) definition of redundant σύνδεσμοι that Apollonius attributes to Trypho: [99]
(t9)
Ὁ γοῦν Τρύφων ἐν τῷ ὅρῳ βουλόμενος καὶ αὐτοὺς (sc. τοὺς παραπληρωματικούς) ἐμπεριλαβεῖν φησί: “...καὶ τὸ κεχηνὸς τῆς ἑρμηνείας ἔστιν ὅπου παραπληρῶν”, ἀπείκασε δὲ καί τινας αὐτῶν ταῖς καλουμέναις στοιβαῖς: “ὃν γάρ,” φησι, “τρόπον εἰς τὰς συνθέσεις τῶν ἀμφορέων εὐχρηστεῖ ἡ τῶν στοιβῶν παρένθεσις ὑπὲρ τοῦ μὴ καταθραύεσθαι τοὺς ἀμφορεῖς, τὸν αὐτὸν δὴ τρόπον ὑπὲρ τοῦ τὰ τῆς φράσεως μὴ τραχύνεσθαι ἥδε ἡ σύνταξις τῶν μορίων παραλαμβάνεται.
Apollonius Dyscolus, Combiners 247.23-29
Trypho, for example, wanting to incorporate them [sc. the σύνδεσμοι παραπληρωματικοί] in the definition too, says: “it also happens that they fill the gap(s) of the utterance,” and moreover he compares some of them with what we call stuffings: “For,” he says, “in the way it is helpful, when putting amphoras together, to put stuffings in between so that the amphoras are not damaged, in just the same way, so that the constituents of the phrase do not become harsh, this combination of the μόρια [100] is used.”
Here we find a definition of σύνδεσμοι παραπληρωματικοί that resonates strongly with the definition of σύνδεσμοι in the Téchnē, but with a focus on filling the gaps (παραπλῆρων, compare the variant reading πληροῦσα in the Téchnē). Trypho explains this with a metaphor of stuffing between amphoras: σύνδεσμοι παραπληρωματικοί serve to keep the other words from becoming “harsh.” It may remind us of the Téchnē’s definition of σύνδεσμοι παραπληρωματικοί, with its focus on metre and arrangement or euphony. Trypho further argues that σύνδεσμοι should be regarded as words (as opposed to syllables). Elsewhere, Trypho has posited: “[if] they are [words] they must mean something.” [101] If Trypho pursued this line of thought, his discussion is no longer extant. However, the combination of these two thoughts will form the basis of Apollonius Dyscolus’ discussion of σύνδεσμοι παραπληρωματικοί.

2.6.2 Apollonius the Sophist

§52. The author of a first century CE Homer Lexicon work appears to be of little significance in the development of the study of particles. However, his work has in places an overlap with the scholia, most notably with the scholion to 1.219 on η (quoted above on page 14-15). [102] Apollonius the Sophist gives a shorter version of the same note – omitting the references to Aristarchus, for example – so it is more likely that he incorporated this note from an earlier commentary on the Iliad than that the entry in his Lexicon was later incorporated into the scholia.
§53. Apollonius the Sophist’s work deserves some attention, however, because he notes diachronic development in the use of certain particles. In two instances he notes a difference in the use of a particle between Homer and “others,” as well as his own time. ἄρα, for example, is used for δή throughout Homer, but is in fact a syllogistic, conclusive particle in other authors. [103] Likewise, ἤτοι stands for μέν in Homer, but is a disjunctive particle in other poets and in Sophistes’ time. [104] It is of paramount importance to understand that particle use changes over time, and Apollonius the Sophist is one of the first sources to bring up the topic.

2.6.3 σύνδεσμοι in the Téchnē

§54. All of chapter 20 of the Téchnē is devoted to σύνδεσμοι. As we have discussed above, this part of the Téchnē probably does not go back to Dionysius Thrax, but looks like a later addition. Regardless of its exact date, perhaps somewhere between the first and third centuries CE, it is one of the few early discussions of σύνδεσμοι as a category. This systematic account of combiners begins with the following definition:
(t10)
Σύνδεσμός ἐστι λέξις συνδέουσα διάνοιαν μετὰ τάξεως καὶ τὸ τῆς ἑρμηνείας κεχηνὸς δηλοῦσα
Téchnē 20.1-2
A combiner is a word that conjoins the thought through order and clarifies the gap{s} of the expression.
This definition assigns two functions to a σύνδεσμος: on the one hand a σύνδεσμος knits together units of thought (διάνοια) in an utterance by imposing order, and on the other it has an effect on gap(s) in the expression. [105]
§55. The first part is clear enough, but the second requires more discussion; the difficulty lies in κεχηνὸς δηλοῦσα: how can anything “clarify a gap”? In antiquity Heliodorus, in his commentary on the Téchnē, explains it as referring specifically to the disjunctive (διαζευκτικοί) combiners. [106] Modern editors do not lean that way. Swiggers and Wouters argue that it is because σύνδεσμοι have no propositional content that they can be said to “show the void in linguistic symbolization” (ἑρμηνεία). [107] Lallot, likewise, translates: “… et qui révèle l’implicite de l’expression.” [108] Kemp prefers the alternative reading πληροῦσα found in several manuscripts, a lectio facilior that allows him to translate: “… and fills up gaps in the expression.” In our translation we have followed the Greek as closely as possible, and aligned with Lallot as to the interpretation of δηλοῦσα. However opaquely, it appears that the definition in the Téchnē states that σύνδεσμοι serves to clarify what is unexpressed in language, such as those things implied, or perhaps the mode rather than the content of the utterance. Other definitions generally highlight the two functions of σύνδεσμοι to impose order, and show some force (δύναμις). [109] The Téchnē does not express it in those terms, but its author may have been thinking of the same thing.
§56. The Téchnē continues with enumerating eight or nine subcategories groups of σύνδεσμοι: (1) συμπλεκτικοί (copulative), [110] (2) διαζευκτικοί (disjunctive), [111] (3) συναπτικοί (hypothetical), [112] (4) παρασυναπτικοί (continuative), [113] (5) αἰτιολογικοί (causal), [114] (6) ἀπορηματικοί (dubitative), [115] (7) συλλογιστικοί (conclusive), [116] (8) παραπληρωματικοί (filling/redundant), [117] the ninth subcategory, ἐναντιωματικοί (concessive), is reported as accepted only by “some.” [118]
§57. These subcategories, unfortunately, receive only cursory definitions, such as the following explanation of σύνδεσμοι παραπληρωματικοί:
(t11)
παραπληρωματικοὶ δέ εἰσιν ὅσοι μέτρου ἢ κόσμου ἕνεκεν παραλαμβάνονται
Téchnē 20.24-25
… and fillers are those that are employed for the sake of metre or beauty [119]
Beyond such brief definitions, the Téchnē only gives a few examples for each subcategory; i.e. καί is συμπλεκτικός, ἤ is διαζευκτικός, ἵνα is αἰτιολογικός et cetera. This discussion of σύνδεσμοι reveals none of the insight or interest found in Thrax’ comment quoted in the scholion to Iliad 12.301, and may serve as another argument that it is unlikely that chapter 20 of the Téchnē can be attributed to Thrax.

2.6.4 Demetrius’ Style

§58. Before moving on to Apollonius Dyscolus Ι mention one more hard-to-date work, Περὶ Ἑρμενείας, a treatise on Style traditionally attributed to the fourth to third century BCE philosopher Demetrius of Phaleron but whose authenticity has been questioned. [120] This work – now believed to have been written anytime in a five-hundred year time span (from third century BCE to second century CE) – is innovative because it discusses σύνδεσμοι with style, rather than grammar, in mind. The author’s interest in style is apparent in the following passage on σύνδεσμοι παραπληρωματικοί:
(t12)
λαμβάνεται δὲ κἀν παθητικοῖς πολλάκις ὁ σύνδεσμος οὗτος [sc. δή], ὥσπερ ἐπὶ τῆς Καλυψοῦς πρὸς τὸν Ὀδυσσέα
Διογενὲς Λαερτιάδη πολυμήχαν’ Ὀδυσσεῦ,
οὕτω δὴ οἶκόνδε φίλην ἐς πατρίδα γαῖαν... (Odyssey 5.204)
εἰ γοῦν τὸν σύνδεσμον ἐξέλοις, συνεξαιρήσεις καὶ τὸ πάθος.
Pseudo-Demetrius, Style 57
Also in emotional passages this combiner [sc. δή] is often used, as in the scene with Calypso in the Odyssey:
Zeus-born Laertes’ son, creative Odysseus,
just like that, homeward to your beloved fatherland...
so if you were to take out the combiner, you would also take out the πάθος.
This explicit discussion of the πάθος that a σύνδεσμος contributes is unparalleled in early sources, other than Apollonius Dyscolus’ discussion of γε. [121]
§59. Demetrius discusses these combiners only in the context of how they contribute to the “grand style,” he does not attempt to define them or categorize them. Perhaps exactly because of that reason, his notes contain a philological angle that is suprisingly rare in the scholia and early grammars. Because it falls outside the scope of his work, the author of Style does not make generalizations about σύνδεσμοι – for these, we have to go on to Apollonius Dyscolus. [122]

2.7 Apollonius Dyscolus

§60. The grammatical work done by Apollonius Dyscolus forms a watershed in the study of language, as far as can be established from the extant literature. Among his extant works we discuss in the present section first his Syntax, since in this general work he presents some general ideas about σύνδεσμοι. Then we turn to Περὶ συνδέσμων, the oldest extant study devoted entirely to combiners.
§61. Early on in his seminal work on Syntax, Apollonius attempts to define the category of σύνδεσμοι. He finds that these words tend to work in two important ways: first, they conjoin two or more phrases such that an essential connection is lost without their presence; and second, like ἄρθρα (articles) and προθέσεις (prepositions), σύνδεσμοι can only co-signify (συσσημαίνειν): τὰ γὰρ τοιαῦτα τῶν μορίων ἀεὶ συσσημαίνει. [123] That is, prepositions, articles, and combiners obtain meaning only when used in combination with nouns, verbs, and/or adverbs. Combiners co-signify in the following way:
(t13)
οἵ τε σύνδεσμοι πρὸς τὰς τῶν λόγων τάξεις ἢ ἀκολουθίας τὰς ἰδίας δυνάμεις παρεμφαίνουσιν
Apollonius Dyscolus, Syntax I.14.4-6
and the σύνδεσμοι, with respect to the positions or constructions of the phrases, [124] show [125] their individual forces [126]
Several different elements in this cryptic passage require explication. What is Apollonius saying here? Lallot interprets Apollonius to mean that σύνδεσμοι have their own force, but that context determines which specific σύνδεσμος should be used, as well as the particular force that the chosen σύνδεσμος acquires in the sentence. [127]
§62. The two aspects of context that determine a combiner’s force are the τάξις and the ἀκολουθία of λόγοι. The meaning of τάξις here is clear: it refers to the order of phrases, as elsewhere in Apollonius. [128] Ἀκολουθία is more ambiguous. The term can be used to describe the succession of an argument, but this sense is unlikely given Apollonius’ usage of the word: in general, ἀκολουθία in Apollonius refers either to agreement between words or to a pattern of regularity. [129] As an extension of this latter meaning we have translated ἀκολουθίας as “constructions”: the relation between sequences of words and their specific meanings. Thus combiners co-signify not only with phrases, but also with the constructions that those phrases make up. [130]
§63. At first sight, Apollonius’ definition of σύνδεσμοι may not appear to cover those combiners commonly gathered under the name σύνδεσμοι παραπληρωματικοί. Unlike his predecessors, however, Apollonius has concrete ideas about these “filling” combiners, and later in his Syntax he refers to the Combiners, the work in which he had developed these ideas more fully:
(t14)
οἵ γε μὴν καλούμενοι παραπληρωματικοὶ [sc. σύνδεσμοι] οὐκ ἀπὸ τοῦ δηλουμένου τὴν θέσιν ἔσχον. οὐ γὰρ ἀληθές ἐστιν, ὥς τινες ὑπέλαβον, μόνον αὐτοὺς ἀναπληροῦν τὸ κεχηνὸς τῆς ἑρμηνείας καὶ διὰ τοῦτο εἰρῆσθαι παραπληρωματικούς· ὅτι γὰρ ἕκαστος αὐτῶν [sc. συνδέσμων παραπληρωματικῶν] ἔχει τινὰ δύναμιν, παρεστήσαμεν ἐν τῷ Περὶ Συνδέσμων.
Apollonius Dyscolus, Syntax III.378.4-379.1
However, those called filling [sc. combiners] do not get their name from their meaning. For it is not true, as some have understood it, that those [sc. filling combiners] only fill in the gaps of the expression, and it is because of this that they are called filling. That every one of them has some force, we showed in the work Combiners.
And further down:
(t15)
...σχεδὸν γὰρ ἕκαστος αὐτῶν ἴδιόν τι ἐπηγγέλλετο
Apollonius Dyscolus, Syntax III.380.13-14
...for nearly each one of them signals something specific
These passages outline Apollonius’ approach to those σύνδεσμοι he calls redundant, on which he expands in Combiners, the first and only work from antiquity known to be dedicated solely to σύνδεσμοι. The text of the treatise is damaged, apparently corrupt in places, and incomplete. Even so, it gives us more material on the subject than any other work. As Apollonius’ definitions in the Syntax already suggest, his Combiners contains innovative insights about the workings of σύνδεσμοι in general and of specific instances.
§64. In Combiners 222.12, where he argues that negations are adverbs, not combiners, [131] Apollonius comes closest to defining combiners. [132] Later he argues that a combiner can be taken out and replaced with another, which is an important characteristic of words for Apollonius. [133]
§65. It is clear that Apollonius went much further than any of his predecessors in trying to grasp and define σύνδεσμοι. He gives many insightful discussions of individual σύνδεσμοι as a result. Apollonius discusses a number of words that we would consider conjunctions (ὅτι, ἕνεκα and cognates, διότι, ἵνα, ὅπως, ὄφρα) and the adverbially used χάριν. The individual particles discussed are (in alphabetic order): ἄρα, ἆρα, αὐτάρ, γάρ, γε, δή, ἦρα, ἤ, ἦ, ἤτοι, θην, καίπερ, καίτοι, μέν, μέντοι, μών, νυ, οὐκοῦν, οὔκουν, οὖν, περ, που, τοίνυν, τοιγάρτοι, τοιγαροῦν, ὦν, as well as the combinations ἀλλὰ μήν, ἀλλὰ γάρ, ἆρα γε, δέ γε, καὶ μήν, μὲν γάρ. In the following subsections we first discuss Apollonius’ notes on the subcategories of σύνδεσμοι that he identifies, and then we turn to his insights on individual σύνδεσμοι.

2.7.1 Subcategories

§66. As far as can be reconstructed from his damaged treatise, Apollonius uses roughly the same set of subcategories as can be found in the Téchnē although he adds a few and does not always use the same term. Unlike the Téchnē, he does not appear to discuss the subcategories one-by-one in a systematic order, but that lack of order may be a result of the state of the work. Probably for the same reason, not all of his definitions of the different subcategories are extant. For disjunctive, subdisjunctive, affirmative (Dalimier 2001:87 translates “manifestantes”), dubitative, syllogistic, and expletive conjunctions there are longer discussions; the other categories are mentioned only in passing. [134] In this section, we focus on Apollonius Dyscolus’ definitions of the disjunctive, conclusive, and redundant combiners.
§67. Regarding disjunctive combiners, Apollonius makes a very pertinent observation. He asks the question: How can a word that combines, or conjoins (σύνδεσμος), be disjunctive (διαζευκτικός)? Is this not contradictory? His answer is that a disjunctive combiner conjoins words by presenting a disjunction in the words referred to:
(t16)
εἴρηνται μὲν σύνδεσμοι ἕνεκα τοῦ συνδεῖν τὰς φράσεις (...) ἕνεκα δὲ τοῦ ἀπ᾽ αὐτῶν δηλουμένου διαζευκτικοὶ ὠνομάσθησαν, ὅλης γὰρ τῆς φράσεως [135] πράγματα διαζευγνύουσιν.
Apollonius Dyscolus, Combiners 216.2-6
The are called σύνδεσμοι because they join together the expression (...) and because of what they clarify they are called disjunctive, for of the entire expression they disjoin the facts. [136]
Apollonius’ explanation bears a striking resemblance to the definition given in the Téchnē:
(t17)
διαζευκτικοὶ δέ εἰσιν ὅσοι τὴν μὲν φράσιν ἐπισυνδέουσιν, ἀπὸ δὲ πράγματος εἰς πρᾶγμα διιστᾶσιν.
Téchnē 20.10-11
Disjunctive are those [sc. combiners] that conjoin the expression, but set one fact apart from another.
The resonance of the terms φράσις, συνδεῖν and most of all πράγματα must be significant. [137] However the lacuna is resolved, the explanation in Apollonius is more extensive than the definition in the Téchnē, but its meaning is not as clear. Moreover, the term διιστᾶσιν in the Téchnē’s definition seems more advanced than διαζευγνύουσιν, as it explains the category (διαζευκτικοί) without resorting to the same root. The definition in the Téchnē appears to be further developed than that given by Apollonius; yet another argument to regard this part of the Téchnē as composed later than Apollonius Dyscolus.
§68. Apollonius Dyscolus discusses conclusive (συλλογιστικοί) σύνδεσμοι several times, and mentions they used to be called ἐπιφορικοί by the Stoics. [138] If the Téchnē was indeed written by Dionysius Thrax, we may have perhaps expected to find this term, but this is not the case. Rather, like Apollonius, the Téchnē uses the word ἐπιφορά, “conclusion,” in his definition, but never the adjective ἐπιφορικός. [139]
§69. Doubtlessly, Apollonius’ most important contribution to the study of σύνδεσμοι is the contention that redundant combiners (σύνδεσμοι παραπληρωματικοί) are not only fillers. Recall that Trypho had compared σύνδεσμοι to pieces of cloth placed as buffers between amphorae. [140] Apollonius discusses the fact that some scholars say that σύνδεσμοι παραπληρωματικοί should not be called σύνδεσμοι, since they do not conjoin parts of the utterance as such. [141] While Apollonius concedes that not all σύνδεσμοι παραπληρωματικοί actually combine, [142] he leans toward the views of one author (conjectured to be the Stoic Chairemon) [143] who argues that since these words look like combiners formally (τύπῳ), they should be designated as such. [144]
§70. After reporting Trypho’s discussion of σύνδεσμοι παραπληρωματικοί (see §51 above), Apollonius adds the following “for this we will plead too, having added something extra.” [145] The “something extra” Apollonius actually deduces from a premise by Trypho quoted earlier in the work: “[if] they are [words] they must mean something.” [146] What Apollonius is hinting at, and what he develops in his treatise Combiners, is that even expletive particles (can) contribute meaning to a sentence. [147] He argues that one characteristic of words is their interchangeability, the fact that they can be replaced by synonyms, such as αὐτάρ for δέ. [148] Moreover, even enclitics are actual words, as shown by the fact that they can bear the accent when placed next to each other. Thus, Apollonius concludes, if on formal grounds σύνδεσμοι may be regarded as words, they should mean something. Finally, he argues that just because a word is redundant in one utterance does not mean it ceases to be a word. He points out that an adjective λευκώλενος that is redundant in one passage is not redundant at all in another context. [149] Likewise, he argues, so-called fillers are not always redundant.
§71. Apollonius then moves on to another problem with the category. Unlike the copulative, disjunctive, or causal σύνδεσμοι, the παραπληρωματικοί cannot be said to all do roughly the same thing: [150]
(t18)
οἱ μέντοι παραπληρωματικοὶ οὐχ ἓν ἐπηγγέλλοντο κατὰ τὸ δηλούμενον. εἴγε ὁ μὲν δή περιγραφήν τινα ἐδήλου, ὁ δὲ πέρ ἐναντιότητά τινα μετ’ αὐξήσεως, καὶ ἔτι ὁ γέ μειότητα ἢ ἐπίτασιν θαυμασμοῦ, καὶ εἰ διάφοροι κατὰ τὸ δηλούμενον, πῶς ἦν δυνατὸν μίαν ὀνομασίαν ἀπὸ τοῦ δηλουμένου χωρίσαι;
Apollonius Dyscolus, Combiners 253.15-18
Still, the filling [combiners] do not say one thing as to their meaning. If δή signals a kind of conclusion, [151] περ indicates an opposition and amplification, and further γε signals limitation or underlines the amazement, and if they differ as to their meaning, how would it be possible to set them apart under one name [i.e. category] based on their meaning?
Apollonius answers this question by arguing that the majority of instances must have priority over the minority: copulative (συμπλεκτικοί) combiners often connect even if they are sometimes redundant – so they are rightly called copulative combiners. Similarly, σύνδεσμοι παραπληρωματικοί are filling in the majority of instances, hence their name. [152]

2.7.2 Important topics raised by Apollonius

§72. In the present section we focus on three discussions in the Combiners that coincide with three lines of research important to our monograph: (1) particles and prosody, (2) particles working above the sentence level, and (3) the polyfunctionality of particles. [153]
§73. First, on the issue of prosody, consider the following comment on καίπερ:
(t19)
ὑπεναντίωσιν γὰρ ἐδήλωσεν ὁ καί<περ>, καὶ δῆλον ὅτι διὰ τὸν πέρ, ὅπου γε καὶ κατ᾽ ἰδίαν ὁ πέρ ἐναντιωματικός ἐστι μετ᾽ αὐξήσεως.
Apollonius Dyscolus, Combiners 251.3
καίπερ shows an opposition, [154] and it is clear that this is because of περ, as also on its own περ is concessive with amplification.
The combination καίπερ is described as getting its force from περ, as both combiners have the same function. The concessive force of περ is clear enough, and repeated in all later descriptions of the particle, but more notable is the comment on amplification (αὐξήσις). When he talks of “amplification,” Apollonius appears to be thinking of some form of emphasis: compare a similar comment in the Syntax, where Apollonius is more explicit: [σημαίνει] ἐναντιότητα ὁ πέρ μετ’ αὐξήσεως ἐμφατικῆς (“περ [conveys] a concessive force, along with emphatic amplification”). [155] What Apollonius has in mind by “amplification” is not self-evident, but we envision some kind of prosodic emphasis resulting from the addition of the enclitic περ. [156]
§74. Apollonius touches upon another important dimension of the function of σύνδεσμοι παραπληρωματικοί in his attempt to describe the function of δή, often called redundant. He argues that it can effect a transition in discourse, having an effect similar to that of a περιγραφή, a conclusion or summary, before one moves on to the next topic:
(t20)
Ἔτι ὁ δή ὡς μὲν παρέλκει, παντὶ προῦπτον· ὡς δὲ καὶ πολλάκις μετάβασιν λόγου ποιεῖται, σαφὲς ἐκ τῶν τοιούτων
οἱ μὲν δὴ παρ’ ὄχεσφιν ἐρητύοντο μένοντες (Iliad 15.3)
καὶ τῶν παραπλησίων. νοοῦμεν γὰρ λόγου ἔκλειψιν καὶ ἀρχὴν ἑτέρου, ὡς εἰ καὶ ἐν περιγραφῇ κατελιμπάνετο ὁμοίως τῷ
ὣς οἱ μὲν Τρῶες φυλακὰς ἔχον (Iliad 9.1)
ὣς ὁ μὲν ἔνθ’ ἠρᾶτο (Odyssey 7.1).
Apollonius Dyscolus, Combiners 251.19-26
Also for δή, that it is redundant is obvious, but that it also often effects a transition of discourse is clear from the following:
And they then were stopped around their chariots, waiting, (Iliad 15.3)
and similar [verses]. For we see a leaving off from [one part of] the discourse and the beginning of another, as if also leaving it with a conclusion, similar to:
Thus, the Trojans set up guards (Iliad 9.1)
Thus, he prayed there (Odyssey 7.1).
 Apollonius compares the constructions οἱ μὲν δή and ὣς X μέν, both occurring often in transitional passages in Homer. Both constructions signal to the reader that one episode is left behind, and another begins. [157] Apollonius’ analysis of δή in this context shows a crucial new step in the study of σύνδεσμοι. Rather than discussing the function of a σύνδεσμος only with respect to the surrounding words or the host clause, he shows how important it is to consider the place of a σύνδεσμος, or string of σύνδεσμοι, in the larger discourse. Here he has noted the common occurrence (καὶ τῶν παραπλησίων) of δή in such concluding verses just before a transition to a new episode.
§75. Finally, Apollonius shows a nascent awareness of the polyfunctionality of particles, that is, they can do different things depending on the context (remember his claim that combiners co-signify). When discussing σύνδεσμοι and the subcategories to which they are assigned, Apollonius often points out that categorization is not sacred: ἄρα, for example, though it generally has a conclusive force, can also be redundant, in clauses like ὣς ἄρα μιν εἰπόντα τέλος θανάτοιο κάλυψε (“so then, when he had spoken, the end of death enveloped him,” Iliad 16.502). [158] Likewise, γε is often used as a filler, but that is not the case in constructions like καλῶς γε, where it intensifies the emotion (ἔκπληξις), nor in τοῦτό γέ μοι χάρισαι, “grant me just this one thing.” He adds: ἔμφασις ἱκανὴ μειότητος καὶ τοῦ μηδὲν ἀποφαινομένου, “the suggestion [159] of limitation is sufficient, even if nothing expresses it explicitly.” [160] That is to say, τοῦτό γέ μοι χάρισαι on its own is enough to convey “grant me just this one thing” rather than “grant me this one thing.” Here, Apollonius is in fact contemplating the polyfunctionality of certain combiners. This reminds us of the brief definition he gives in the Syntax, [161] where he argues that combiners can have different forces depending on context. Even though the examples in the Combiners describe combiners that can either have a function or be redundant, he is the first to attempt explanation of this phenomenon. [162]
§76. Apollonius’ ad hoc discussions of σύνδεσμοι contain several observations essential even for the modern student of particles. While his reading of περ may seem vague, it hints at an insight that is missing in all early and much of the modern literature. His observation that σύνδεσμοι may be connected to emphasis may be linked to prosody, an issue that we will pick up in several of the research chapters. His analysis of δή as a word that can effect transitions between larger stretches of narrative suggests that he thinks that σύνδεσμοι are relevant not only to the current and preceding clause, but also to larger subdivisions of discourse. [163] Finally, Apollonius reminds us throughout that σύνδεσμοι need not be limited to doing or meaning one thing, but can have multiple functions, depending on the context. Although he never comes to an all-encompassing description of redundant σύνδεσμοι that does justice to the combiners he collects under that term, the multiple angles from which he approaches these words must have made the ancient reader nod to the rhetorical question he asks:
(t21)
Πῶς οὖν ἔτι οὐδὲν πληροῦσιν οἱ παραπληρωματικοί;
Apollonius Dyscolus, Conjunctions 250.11-12.
How can we then still [say] that fillers add nothing?

2.8 After Apollonius Dyscolus

§77. Apollonius’ work was not preserved by chance; his work on grammar remained the standard work on Greek grammar all through the middle ages. Still, some later scholars deserve mention: his son and pupil Aelius Herodianus, Nicanor “the punctuator,” and Priscian, who wrote a Latin grammar in the fifth century CE based strongly on Apollonius’ work. In between these figures come the majority of the extant grammatical papyri. To these scholars and works we now turn.
§78. Herodian, like his father, produced works on grammar himself, and these works include some notes on σύνδεσμοι. [164] The work’s manner of organization enables us to establish what categories of σύνδεσμοι Herodian recognized. [165] Since most of his work is devoted to accents, however, his work does not yield any analyses of the function of σύνδεσμοι. [166] Nicanor’s work cannot rightly be called grammatical, since he was first and foremost a scholar of Homer. He gained his nickname “the punctuator” (ὁ Στιγματίας) because he undertook to help the reader of the Homeric text by adding punctuation. In his system, particles play an important part in deciding what kind of punctuation he introduced: he linked different particles to pauses of different lengths (χρόνοι); his work is discussed at length in IV.3 §§31-36.

2.8.1 Early grammars

§79. The oldest substantial extant grammatical papyrus (P.Yale I. 25, first century CE) discusses fewer subcategories of σύνδεσμοι than Apollonius and Herodian. The seven categories in the papyrus are, conversely, roughly the same as those listed in the Téchnē attributed to Thrax, though worded differently. [167] More importantly, the papyrus attempts a definition of σύνδεσμος:
(t22)
σύνδεσμος δ’ ἐστὶν λέξις συνάπτουσ̣α
τὰ μέρη τῆς ἑρμεν<ε>ίας.
P.Yale I 25, 54-55
and a combiner is a word that joins together
the parts of the expression.
The simplicity of the definition is striking, especially compared to all the other extant definitions. [168] Scholars have therefore concluded that it is a simple school grammar, designed with the practical aim of listing and describing lexical categories rather than providing a philosophical foundation of the workings of language. This conclusion may be extended to most of the papyri from this genre, dating from the second to the sixth century CE. [169]
§80. This brings us to the end of antiquity, and to a grammarian who flourished around 500 CE, Priscian. His Institutions of Grammar describes the Latin language and is strongly influenced by Appollonius. In the book devoted to combiners (book XVI), Priscian gives one of the most comprehensive definitions of combiners in antiquity: [170]
(t23)
Coniunctio est pars orationis indeclinabilis, coniunctiva aliarum partium orationis, quibus consignificat, vim vel ordinationem demonstrans
Priscian Institutions of Grammar XVI I.1-2 (Keil III.93.1-2)
A combiner is an indeclinable part of the expression, which connects the other parts of the expression, with which it co-signifies, signalling either a force [of its own] or the arrangement [of the parts of the utterance]
All separate aspects of Priscian’s definition can be traced back to his predecessors, but this is the first time we find the elements combined. In book XVI of the Institutions of Grammar, Priscian discusses only a small selection of combiners, and often has recourse to giving Greek translations. The list of subcategories that Priscian identifies comes closest to that found in Herodian, but he has added a few categories not mentioned before. After listing all possible kinds of combiners, [171] Priscian devotes only a few lines to most kinds and none to some. Two apparent additions to the list, the coniunctiones ablativae and praesumptivae, remain undiscussed, which makes it hard to guess what he meant with them (hence the lack of translation) and which combiners he had in mind.
§81. In the centuries following Apollonius Dyscolus, grammars leaned heavily on his work, but there is little evidence for actual progress in the thinking about combiners. The genre of the school grammar emerged, and several examples are extant. For all these papyri their function is directly reflected in a focus on categorizing and exemplifying rather than analyzing combiners. Priscian wrote an entirely different kind of work, but as far as the study of combiners is concerned, he presents a synthesis of earlier ideas rather than innovative analyses.

2.8.2 Late Antique Scholia to the Téchnē

§82. By the fourth century CE, the Téchnē attributed to Dionysius Thrax appears to have become a common and often-copied work, and its manuscripts gathered a mass of scholia and commentaries over the centuries. [172] These scholia and commentaries are some of the few grammatical sources we can date with certainty, falling between the fifth and tenth centuries. Overall, the additions they propose as regards the analysis σύνδεσμοι appear to be taken directly from Apollonius Dyscolus or other sources.
§83. Heliodorus, [173] in his commentary on the Téchnē, offers a revised definition of σύνδεσμος, which according to Pecorella is to be attributed to Apollonius. [174] The definition adds several characteristics in comparison with the definition in the Téchnē: [175]
(t24)
σύνδεσμος ἐστι μέρος λόγου ἄκλιτον, συνδετικὸν τῶν τοῦ λόγου μερῶν, οἷς καὶ συσσημαίνει, ἢ τάξιν ἢ δύναμιν <ἢ καὶ τάξιν καὶ δύναμιν> παριστῶν.
Heliodorus Commentary to Dionysius Thrax [176]
a combiner is an indeclinable part of speech, to bind together the parts of speech, with which it co-signifies, showing either order or force .
Note the striking similarity with Priscian’s definition, which we discuss above: “a combiner is an indeclinable part of the expression, which connects the other parts of the expression, with which it co-signifies, signalling either a force [of its own] or the arrangement [of the parts of the utterance].” [177] It is generally assumed, and may be observed in comparison, that Priscian based much of his Institutions on Apollonius Dyscolus’ work, which supports Pecorella’s hypothesis that this definition is in fact to be attributed to Apollonius. The definition combines different ideas in a way that fits well with our sense of Apollonius’ approach to σύνδεσμοι. It is, in addition, more intricate than the definitions we find in the different Téchnai. Since a definition of σύνδεσμοι is extant in Apollonius’ Syntax, [178] we find it more likely that the source both for Priscian and for Heliodorus was the definition of σύνδεσμοι in a lost part of Apollonius’ Combiners.
§84. Apart from this definition, which is thus probably to be attributed to Apollonius, the scholia to the Téchnē cannot be said to offer new insights. Although the scholia expand upon their source text, the Téchnē attributed to Thrax, they do not form a body of innovative research when regarded in the broader scholarly context.

2.8.3 The Medieval lexicographers

§85. Other more or less datable sources are Hesychius’ Lexicon, the Suidas, and Eustathius’ commentary to Homer. Hesychius’ lexicon, as we would expect from the genre to which it belongs, rarely offers more than synonyms for σύνδεσμοι, which in turn usually have an analogue in the scholia. However, the lexicon’s value lies in the fact that it was readily available to early modern researchers, as opposed to the scholia that were by that period attested in only a few manuscripts and thus were hardly accessible. The lexicon’s popularity is probably why we find regular references to Hesychius in works from the fifteenth, sixteenth, and seventeenth century.
§86. The Suidas lexicon from the tenth century generally gives some more information per lemma than Hesychius, but like Hesychius the Suidas does not so much add new knowledge as compile older knowledge.
§87. Moderately more significant are Eustathius’ commentaries on the Iliad and Odyssey, based in large part on the scholia. Indeed, Eustathius’ explanations of σύνδεσμοι do not add much to the scholia. His classifications of individual σύνδεσμοι have a different focus than the scholia – he regards καί as redundant more often than δέ, for example – few of his comments on σύνδεσμοι are noteworthy, except for the following two on γάρ and μήν.
§88. On γάρ at Iliad 17.221, Eustathius remarks that it is either redundant, or gives the cause for something left out (“ἐλλειψεώς τινος”). The latter part of his comment adumbrates a whole tradition of positing ellipsis to explain the use of γάρ in problematic instances. [179] The topic of ellipsis returns in a discussion of γάρ in dialogue. [180] In this context implications and presuppositions are extremely relevant to particle use; as will become apparent, our approach acknowledges the value of these early views.
§89. On μήν, Eustathius is the first attested author to note its relation to μέν and μάν in other dialects, where μέν is Ionic, μάν Doric (which had already been noted in the scholia), and μήν is “most common,” presumably in the Greek that he knew. [181] This realization would later lead to a reassessment of μέν as more than just one part of a μέν - δέ construction. [182]

2.9 A renaissance of the particle

§90. The incomplete material left to us strongly suggests that after Apollonius Dyscolus not much innovative work on σύνδεσμοι was done until the sixteenth century. This is when Devarius wrote his Liber de Graecae Linguae Particulis (1588), a fully extant work on particles that would form the basis for all discussions of particles until well into the nineteenth century. He set a new standard on form, quantity, and quality, yet his definition illustrates his strong dependence on the ancient scholarship:
(t25)
Constitui de Graecae linguae vocibus quibusdam agere, quae tametsi rem ipsae per se nullam fere significant; tamen in aliarum vocum constructione positae vim aliquam habent, efficacitatemque: seu emphasim aut certe qualitatem aliam sermoni tribuunt. (…) modique alicuius aut ordinis veluti signa sunt; continuandique aut transeundi , aut interrogandi, aut dubitandi, aut ratiocinandi, aut qualitatem aliam referendi vim habent.
Devarius 1588:1
I have undertaken to discuss a number of Greek words, which, although they themselves signify almost nothing, still when put into constructions with other words have some force (vis) and power (efficacitas): they add either emphasis or some other quality to the utterance. (…) They are like signs of some mode or order, and have a continuative, transitional, interrogative, dubitative, or ratiocinative force, or a force that conveys some other quality.
Unlike Apollonius and his other predecessors, Devarius discussed the particles not in groups of functions, but rather per lexical item, roughly in alphabetical order. [183] This lexicographical organization made for a much more systematic work, and required Devarius (the first to do this as far as we know) to decide which specific lexical items to include and exclude from the category he defined as particulae. [184] As the list shows Devarius’ selection is extremely diverse, ranging from single words to full phrases and including particles, conjunctions, interjections, pronominals, and verb forms.
§91. In later works on Greek particles the process of selection became central: from the long list of Devarius, Denniston only discusses thirty-two particles and their combinations. [185] There is hardly any agreement between the authors who worked from the sixteenth to the twenty-first century about the exact definition of the word “particle,” or about the delineation of the group of words to be gathered under the term. Note, however, that it was not until the sixteenth century that delineation of the category was regarded as an actual problem. [186] Devarius’ landmark work led to a new appreciation for particles and laid the foundations for the important scholarly works by Hoogeveen 1769, Hartung 1832-1833, Klotz 1842, Bäumlein 1861, and Denniston 1934.

Footnotes

[ back ] 1. As implicitly acknowledged, but summarily put to one side, by Sicking 1986:25-26, Schenkeveld 1988:81, and Duhoux 2006:520-523. Denniston 1934:xxxvii claims that he will attempt a definition, but what follows is a forty-five page description of those words that he regards as particles, not anything that qualifies as a true definition. See also Pinkster 1972:135n2 for problems with the term particle in general and I.3 for an overview of the discussion on particles and discourse markers in contemporary linguistics.
[ back ] 2. Their activities are generally not regarded as linguistics, however, since if their starting point was always language as a reflection of the thinkable, i.e. the sayable, instead of instantiations of language, that is, actual texts or speech. Pagani 2011:23n27 gives the most relevant references regarding the Stoics and their influence on the study of linguistics.
[ back ] 3. See Matthaios 1999:193-198 with notes for an exposition of the problems.
[ back ] 4. See Matthaios 2002:163-169, the eight categories are: ὄνομα (noun), ῥῆμα (verb), μετοχή (participle), ἄρθρον (article), ἀντωνυμία (pronoun), πρόθεσις (preposition), ἐπίρρημα (adverb), σύνδεσμος (combiner). From the first grammars onward there appears to have been a constant discussion of the number of grammatical word categories, most grammarians arguing for eight or nine categories. See for instance Matthaios 2002:171 for a schema of the different grammarians and their particular system, and 171-213 for a discussion of the possible development. As far as Aristarchus is concerned, De Jonge and Van Ophuijsen 2010:496 take it from Matthaios and regard it as the communis opinio that Aristarchus already used a system of eight word categories.
[ back ] 5. Ax 1991:288 describes Aristarchus’ approach to grammar as a “Grammatik im Kopf,” a grammar in the mind.
[ back ] 6. Di Benedetto 1958:196-206; 1959:118; he reiterates his argument in Di Benedetto 1973, Pinborg 1975:110-114, Siebenborn 1976:13, Fehling 1979:489, Taylor 1987:11; discussed in Schmidhauser 2010:508. Schenkeveld 1994:287 adds the rise of hellenismos as a factor in the development of grammar in the first century BCE.
[ back ] 7. Varro, Lingua Latina, edited by Traglia 1974; of the originally twenty-five books, only Books 5-10 are partially extant. Taylor makes his argument in 1987:14-16.
[ back ] 8. See Di Benedetto 1958:185-196 for discussions, Wouters 1979 for additional editions, and also Wouters 1995 for discussions of the grammatical papyri.
[ back ] 9. Apart from his Syntax, treatises on adverbs, σύνδεσμοι (on which more below), and pronouns are (partially) extant.
[ back ] 10. The earliest papyri date from the fifth or sixth centuries CE: Pap. Hal. 55 a and PSI I 18, see Di Benedetto 1973:801-802.
[ back ] 11. Sextus Empiricus quotes Thrax’s definition of grammar in Against the Mathematicians I 57, with only a slight variation compared to the manuscripts of the Téchnē. A discussion of the variation can be found in Pagani 2011:18, with further bibliography in notes 6 and 7.
[ back ] 12. Taylor 1987:8. The most recent editions are Kemp 1987, Lallot 1998, and Callipo 2011.
[ back ] 13. Di Benedetto 1958-1959, 1973, and 1990, Pinborg 1975:103-106, Siebenborn 1976:12-13, and Fehling 1979:488-489. For a discussion of the most important discrepancies in the tradition, see Di Benedetto 1958:171-178, Pagani 2011:33, and the scholia to the Téchnē, Hilgard 1901:160.24-161.18.
[ back ] 14. Pfeiffer 1968:267-272, Erbse 1980:244-258, and Wouters 1995:95-97. Frede 1977:52-54 believes in the Téchnē’s authenticity, though he does not argue for it, but he says (52) that it is generally accepted as genuine.
[ back ] 15. See Wouters 1995:99, “the kernel of the Téchnē must have undergone clear changes between the moments of its composition and the copying in the first medieval manuscripts.”
[ back ] 16. Taylor 1987:8-9 and Schenkeveld 1994:266-269, who refers to Wouters 1979.
[ back ] 17. Schenkeveld 1994:269 and 1995:50-52; see Di Benedetto 1973:802. In a recent handbook, Schmidhauser 2010:508 cites this view as the communis opinio.
[ back ] 18. Schenkeveld 1995:42.
[ back ] 19. Robins 1995:18-24.
[ back ] 20. Although it is likely that chapter 5 was moved to its current position in the Téchnē from elsewhere.
[ back ] 21. Pagani 2011:36.
[ back ] 22. Pagani 2011 gives a neat summary of the development on 60-62.
[ back ] 23. See Taylor 1987:11 and Schenkeveld 1994:267-269 and 1995:42-44. A similarly problematic work is the Téchnē that is signed Tryphonos technê grammatikê on a third century CE papyrus (P.Lond. 126, see §50n95 below). The extant part of the text gives the last four of the eight parts of speech as listed in the Téchnē attributed to Thrax, which suggests that the two works are part of a similar tradition.
[ back ] 24. Swiggers and Wouters 2002:102n7.
[ back ] 25. Poetics 1456b20-21.
[ back ] 26. Van Bennekom 1975:408 takes it one step further, referring to Simplicius (Commentary in Aristotle’s Categories 10), who mentions that Theophrastus (fourth century BCE) “and his associates” had already dealt with the question of whether to include ἄρθρα and σύνδεσμοι among the parts of speech.
[ back ] 27. Swiggers and Wouters take καθ᾽ αὑτόν to refer to λόγου, but I believe it must here be taken to refer back to σύνδεσμος (even though one might have expected a female form here, as φωνή is the actual antecedent. In this case the first Latin translation (Moerbeke, 13th century) might be adduced, as it translates ipsam, to refer back to vox. This was followed by Kassel who emends αὑτήν 1965:32, which was then, wrongly I believe, thought by Swiggers and Wouters 2002:107 to refer back to ἀρχή, when it must also in Kassel be meant to refer back to φωνή), to mean: ‘on its own’ i.e. alone. After all, Aristotle has already said it can be found at the beginning, but not on its own. This would explain why he includes ἤτοι, which occurs only at the beginning of clauses, but of course always followed by another word.
[ back ] 28. For full discussions, see Dupont-Roc 1980:321-328, Laspia 1997:79-116, and a summary of the main points in Swiggers and Wouters 2002:107-112.
[ back ] 29. Swiggers and Wouters 2002:110.
[ back ] 30. Swiggers and Wouters 2002:112. Van Bennekom 1975:406 lists μέν, δέ, ἤτοι (possibly corrupt), τε, καί, ἐπεί and γάρ as words expressly regarded by Aristotle as σύνδεσμοι.
[ back ] 31. Schmidhauser discusses Chrysippus, the “first philosopher of language” in 2010:502-507.
[ back ] 32. Diogenes Laertius Lives of the Philosophers 7.57: τοῦ δὲ λόγου ἐστὶ μέρη πέντε, ὥς φησι Διογένης τ’ ἐν τῷ Περὶ φωνῆς καὶ Χρύσιππος, ὄνομα, προσηγορία, ῥῆμα, σύνδεσμος, ἄρθρον· ὁ δ’ Ἀντίπατρος καὶ τὴν μεσότητα τίθησιν ἐν τοῖς Περὶ λέξεως καὶ τῶν λεγομένων.
[ back ] 33. This is part of Diogenes of Babylon, fr. 22 in Von Arnim 1903:210-243.
[ back ] 34. The one exception is the fourth to fifth century grammarian Theodosius of Alexandria, who gives a definition of σύνδεσμοι that is based almost fully on the one by Diogenes: σύνδεσμός ἐστι μέρος λόγου ἄπτωτον συνδοῦν τὰ μέρη τὰ λοιπὰ τοῦ λόγου (Grammar 17.21-23). See also §83 below, on a definition (with ἄκλιτον) found in Heliodorus, but attributed to Apollonius Dyscolus by Pecorella.
[ back ] 35. Schenkeveld 1988:83.
[ back ] 36. 5.341; in the hypotheseis to the following verses: 1.214, 3.267, 3.444, 4.489, 4.824, 5.402, 6.83, 10.570, and 16.57.
[ back ] 37. In the hypotheseis to the following verses: 4.439, 17.317, 19.498.
[ back ] 38. 1.512b, 2.139b, 2.463b, 3.31b.
[ back ] 39. 4.539a.
[ back ] 40. Hypothesis to 17.28.
[ back ] 41. 19.63b.
[ back ] 42. 2.330.
[ back ] 43. 1.234 (where it is read as the negation μή).
[ back ] 44. However, note the problematic use of μόριον in Trypho’s definition of the redundant combiners, §§50-51 below.
[ back ] 45. The Homeric scholia to the Iliad form the source for most of the material in this paragraph. The scholia vetera to Pindar were also studied, and follow the same pattern. They will be referred to mainly in the footnotes, referenced in the main text only where they offer insights absent from the Homeric scholia.
[ back ] 46. See Matthaios 2002:163-169 for a well-annotated argument that the Alexandrians already distinguished and used the eight word groups established by the stoics.
[ back ] 47. Erbse attributes the scholion in its current form to Didymus or Aristonicus, who worked in the first century BCE and CE, respectively.
[ back ] 48. See V.ἦ for modern literature on ἦ.
[ back ] 49. This is a clear reflection of how written versions and oral performance must still have gone hand in hand.
[ back ] 50. For early accentuation, see Probert 2006:48-50.
[ back ] 51. See Probert 2006:49-50. In this second development, the possibility that the ninth-century scribes were in possession of an earlier, fully accented papyrus must not be discounted. However, this will certainly not have been the case for every text.
[ back ] 52. The fragments are gathered in Matthaios 1999:162-168, with an analysis in 566-585.
[ back ] 53. Demetrius, Style 55 = Praxiphanes fr. 13 in Wehrli.
[ back ] 54. Matthaios 1999:573 does not assume that Aristarchus already used the term συμπλεκτικός for copulative conjunctions.
[ back ] 55. It is open to discussion whether this means that (1) many of the anonymous comments in the scholia should also be ascribed to Aristarchus, or that (2) Aristarchus’ notes functioned as a model for later scholiasts.
[ back ] 56. See Friedländer 1853:34; Aristarchus already discusses so-called apodotic δέ, see Matthaios 1999:571.
[ back ] 57. For a discussion of the history of the category of σύνδεσμοι παραπληρωματικοί, see Sluiter 1997:234-245.
[ back ] 58. In the scholion to Iliad 10.408, Matthaios 1999:123, fragment 99:42-43, with commentary on pages 581-582.
[ back ] 59. See Matthaios 1999:165 fr. 173 and the scholia in the notes for the relevant places, and page 574 for Aristarchus on γάρ. Also noted in Ax 1982:102-104 and Pagani 2011:43.
[ back ] 60. Matthaios 1999:575, see also the scholion to Euripides Phoenician Women 88657.
[ back ] 61. Apollonius Dyscolus, Syntax I.6, I.127; see Matthaios 1999:581. A similar note may be found in the scholion to K 408. The scholia to the Téchnē repeat the observation: scholia to [Dionysius Thrax], in Uhlig’s Grammatici Graeci I 3, 106.3-5 and 441.37.
[ back ] 62. Matthaios 1999:625 and 2001:90.
[ back ] 63. As Friedländer noted too in his work on Aristonicus. Almost all his notes on coniunctiones found in the scholia concern redundancy: Friedländer 1853:33-35.
[ back ] 64. For a list, see Erbse VI 1983:181. The listings given here and below are as complete as possible. The indices offered by Erbse are invaluable and in combination with the TLG give unheard-of access to the scholia, but even so some relevant notes will be missed. There is therefore no claim that the material is exhaustive, but we are confident that the themes and instances presented here at least form a good representation of the discussions of σύνδεσμοι in the scholia.
[ back ] 65. For a list, see Erbse VI 1983:154.
[ back ] 66. Erbse does not give a separate list, but see at least the following: A 352b, N 317, Ξ 1e, O 372-4, Y 21c, Φ 189 (where the scholiast compares ρ 46), X 416c, and Ω 750.
[ back ] 67. For a list, see Erbse VI 1983:213 for πω and Ρ 643a for που (παραπληρωματικός).
[ back ] 68. For a list, see Erbse VI 1983:219.
[ back ] 69. Friedländer 1853:35.
[ back ] 70. See Erbse VI 1983:154 for an apparently exhaustive list of instances. To this may be added the list of Odyssey scholia in Matthaios 1999:164n1, and the scholia to Pindar: Ol. 2.106a[64] (quoted on the same page), Ol. 4.34b+c[22], Ol. 6.4b+c[3], Οl. 10.36-46[30], Ol. 10.47-50[39], Ol. 12.6-18[5-12], Ol. 13.83[60], P. 5.132[98], P. 6.38[38], P. 10.2-3[2], Ν. 4.95b[59].
[ back ] 71. Matthaios 1999:573n43.
[ back ] 72. For a list, see Erbse VI 1983:154, see also the scholia to Pindar, Ol. 2.102a[62] (δέ = δή), Ol. 9.33[21] (δέ τοι = δὴ οὖν), Ol. 10.63b[51] (δέ = δή), Ol. 13.69a[49] (δέ = δὴ οὖν).
[ back ] 73. Note also the following scholia to Pindar Ol. 8.61-70[46], Ol. 10.51b[43] (ἄρα = δή), P. 4.337[189] (ῥα = δή),
[ back ] 74. 1.77, 10.322 (ἦ μέν), 3.43 (ἦ που), 3.204, 8.102 (ἦ μάλα), 13.354 (ἦ μάν), 13.813 (ἦ θήν), 20.347 (ἦ ῥα).
[ back ] 75. In scholia ὅτι goes back to a construction like [ἡ διπλῆ] ὅτι (as preserved in example t8), explaining the critical sign in the edition; see Dickey 2007:122.
[ back ] 76. Erbse VI 1969:141, s.v. περισσὸς ὁ ἄν.
[ back ] 77. See for Aristarchus on these particles Matthaios 1999:107-109, frr. 73-78 and pp. 363-373.
[ back ] 78. In the scholion to 13.521, for example, οὐδ᾽ ἄρα πω is rendered οὐδέπω, eliminating ἄρα in the paraphrase.
[ back ] 79. See V.ἄρα, II.4.3.2, and II.4.4.
[ back ] 80. Katz 2007 argues that ταρ is one word, and that it might be a loanword from Luvian; see there for additional literature.
[ back ] 81. In 1.65a this is echoed: οὐ γάρ ἐστιν ὁ τέ σύνδεσμος; moreover, ταρ is described as a conclusive (ἐπιφερόμενος) enclitic combiner.
[ back ] 82. This does leave instances of οὐδ᾽ where the scholiasts may have wanted to read οὐ δ(ή).
[ back ] 83. Elsewhere, in scholion 2.284a, Aristarchus himself argues for reading Ἀτρεΐδη, νῦν γάρ σε instead of Ἀτρεΐδη, νῦν δή σε (as most manuscripts), on the basis that in causal constructions Homer often starts with γάρ: ἔθος δὲ αὐτῷ (i.e. Ὁμήρῳ) ἀπὸ τοῦ γάρ ἄρχεσθαι; see Matthaios 1999:574-5 for a discussion. See II.3.2.2 and II.4.2 for more on γάρ beginning new parts of discourse in Homer and Pindar.
[ back ] 84. Apparently first in Eustathius, see §§87-89 below.
[ back ] 85. See also Haslam 2013:2-3 on a new scholion that reports Dionysius of Sidon preferring μήν over Aristarchus’ μάν at Iliad 4.512.
[ back ] 86. Dennniston 19502:33.
[ back ] 87. See II.2.4 for μέν in Homer and Pindar.
[ back ] 88. See I.1 for the concept of scope and its relevance to particle analysis; for an application, see II.3.3 on Homeric δή.
[ back ] 89. Erbse gives a list in part VI on page 179; Linke 1977:61 offers a few examples for the scholia to the Odyssey: 1.33, 8.154, 10.471, 11.453, 16.216.
[ back ] 90. The line is Πάνδαρος, ᾧ καὶ τόξον Ἀπόλλων αὐτὸς ἔδωκεν, the comment in Eustathius 354, 32.
[ back ] 91. The scholion is attributed to Aristonicus.
[ back ] 92. See IV.2.4.8 for a discussion of this function of καί.
[ back ] 93. Fifty-two instances, for the numbers see Lallot 1997:I.16-17.
[ back ] 94. Frr. 40-60, discussed in De Velsen [1853] 1965:34-45. De Velsen adds to Trypho’s fragments on σύνδεσμοι one note (fr. 61) by an anonymous grammarian on Trypho’s discussion of ὡς.
[ back ] 95. Quite recently a Téchnē bearing the name of Trypho was found on a third century CE papyrus (P.Lond. 126). This text, like the one attributed to Dionysius Thrax, takes the form of a short (school) grammar. Towards the end of the extant part the section on conjunctions begins. The crucial first lines, containing a rough definition, are quite opaque, unfortunately. Whoever may have been the author of this papyrus, it is unlikely to have been the first century BCE grammarian.
[ back ] 96. De Velsen [1853] 1965:40, fr. 52; the source is Apollonius Dyscolus, Combiners 240. In the same fragment, Trypho argues that γάρ can be redundant.
[ back ] 97. De Velsen [1853] 1965:41, fr. 54 = Apollonius Dyscolus, Combiners 241.
[ back ] 98. De Velsen [1853] 1965:44, fr. 59 = Apollonius Dyscolus, Combiners 257. Trypho notes that although the form of two words may be very similar, this does not mean anything for their meaning, adducing γαῖα - αἶα, μία - ἴα, and σῦς - ὗς.
[ back ] 99. De Velsen [1853] 1965:35, fr. 41 = Apollonius Dyscolus, Combiners 247.23-29.
[ back ] 100. See above, §§21-22 for a short note on μόρια. Its use here is not entirely clear, and thus difficult to translate.
[ back ] 101. Combiners 249.9-10: [εἰ λέ]ξεις, ὀφείλουσί τι δηλοῦν.
[ back ] 102. Apollonius Sophistes, Lexicon Homericum 81.27-82.8. The edition referred to is that by Bekker, published in 1833 and reproduced in 1967.
[ back ] 103. Apollonius Sophistes, Lexicon Homericum 41.6, 43.13: ἄρα ἀντὶ τοῦ δὴ παρ’ Ὁμήρῳ διὰ παντός, παρὰ δὲ τοῖς ἄλλοις ἐν τῷ βίῳ συλλογιστικὸς σύνδεσμος.
[ back ] 104. Apollonius Sophistes, Lexicon Homericum 85.5-7 καθ’ Ὅμηρον μὲν ἰσοδυναμεῖ τῷ μέν συνδέσμῳ, (…) παρὰ δὲ ἡμῖν καὶ ἄλλοις ποιηταῖς διαζευκτικὸς σύνδεσμος.
[ back ] 105. See Dickey 2007:239 for this translation of ἑρμενεία.
[ back ] 106. Commentary to Thrax’ Téchnē, attributed to Heliodorus, in: Hilgard, Grammatici Graeci 1.3:103 “Καὶ τὸ τῆς ἑρμηνείας κεχηνὸς δηλοῦσα.” Τοῦτό φησι διὰ τοὺς διαζευκτικοὺς συνδέσμους· ἐκεῖνοι γὰρ τὸ τῆς ἑρμηνείας, ὅ ἐστι τὸ τῆς διανοίας, διεζευγμένον καὶ διεστηκὸς δηλοῦσι·
[ back ] 107. Swiggers and Wouters 1998:3.
[ back ] 108. See Lallot 1998:231-241 for an extensive discussion of the definition. His definition is attractive because it highlights the importance of combiners with respect to the implicit or the “unsaid” in interaction. On the other hand, I am not sure if we can read quite as much into the κεχηνὸς of the definition.
[ back ] 109. The term δύναμις is already applied to particles in the Homeric scholia, especially in forms of the verb ἰσοδυναμέω, e.g. in the scholion to Iliad 9.134 τὸ δὲ ἠδέ ψιλωτέον· σύνδεσμος γάρ ἐστιν ἰσοδυναμῶν τῷ καί (“ἠδέ has a smooth breathing. For it is a combiner with the same force as καί”).
[ back ] 110. The Téchnē lists: μέν, δέ, τε, καί, ἀλλά, ἠμέν, ἠδέ, ἰδέ, ἀτάρ, αὐτάρ, ἤτοι, κεν, ἄν.
[ back ] 111. ἤ, ἤτοι, ἠέ.
[ back ] 112. εἴ, εἴπερ, εἰδή, εἰδήπερ.
[ back ] 113. ἐπεί, ἐπείπερ, ἐπειδή, ἐπειδήπερ.
[ back ] 114. ἵνα, ὄφρα, ὅπως, ἕνεκα, οὕνεκα, διὅ, διὅτι, καθ’ ὅ, καθ’ ὅτι, καθ’ ὅσον.
[ back ] 115. ἆρα, κᾶτα, μῶν.
[ back ] 116. ἄρα, ἀλλά, ἀλλαμήν, τοίνυν, τοιγάρτοι, τοιγαροῦν.
[ back ] 117. δή, ῥα, νυ, που, τοι, θήν, ἄρ, δῆτα, περ, πω, μήν, ἄν, αὖ, νῦν.
[ back ] 118. ἔμπης ὅμως.
[ back ] 119. Kemp 1987 and Lallot 1985 translate κόσμου with “embellishment” and “ornament”, respectively; I have chosen a more neutral translation.
[ back ] 120. Schenkeveld 1964:135-148 presents the discussion with relevant literature, and argues that the work must have been written in the first century CE or later. Morpurgo-Davies 1980:145 calls this a thesis born out of desperation. This is not the place to present the discussion, let alone join in, so I will focus only on the notes on σύνδεσμοι in the work.
[ back ] 121. See below, §75. Elsewhere (Demetrius, Style 56) Demetrius notes that δή is used to mark a new beginning (Iliad 21.1), and that if the combiner had not been used, you might have thought Homer was still talking about the same thing.
[ back ] 122. See IV.3.5 for Demetrius’ discussion of kôla and kómmata in Greek prose.
[ back ] 123. Apollonius Dyscolus, Syntax I.11.3-7 and I.14.2. The edition used is Lallot 1997, vol. I.
[ back ] 124. On the difficult term λόγος in Apollonius Dyscolus, see Lallot 1997:II.10 and Dickey 2007:245.
[ back ] 125. Or “add,” as Lallot 1997:II.18 argues. The difference in sense is minimal, so I chose the more literal translation.
[ back ] 126. Householder 1981:23 translates: “Conjunctions, too, may vary in force according to their position in the sentence or the context,” which is a free translation amounting to the same thing.
[ back ] 127. Lallot 1997:II.18, chapter 227.
[ back ] 128. Consider, for example, Syntax I.81.8-9 ἐν δευτέρᾳ τάξει “in second position.”
[ back ] 129. See Dickey 2007:220; compare also the idiom ἐν ἀκολουθίᾳ in Apollonius Dyscolus, which means “regular” in the sense of regular versus irregular verbs or nouns.
[ back ] 130. This translation of ἀκολουθία would also work very well for Apollonius’ discussion of the σύνδεσμοι παρασυναπτικοί in Combiners 82.13-16: ὁ καλούμενος παρασυνα<πτικός,> ἔχων καὶ ἐπαγγελίαν τὴν τοῦ συμπλεκτικοῦ, ἐν οἷς <συμπλέκει λόγους>, ἔχων δὲ καὶ τὴν τοῦ συναπτικοῦ, ἐν οἷς ἀκολουθίας ἐστὶ παραστατικός (“the [kind of combiner] called παρασυναπτικός, which has both the meaning of the connective [sc. combiner], in that it connects words, and that of the hypothetical [sc. combiner], in that it is indicative of a construction”).
[ back ] 131. The five differences (between οὐ and ἤ, in this case) are the following: first, the two cannot be interchanged indiscriminately: οὐ is not the same as ἤ, even if the one can sometimes be placed instead of the other. Second, combiners co-signify, whereas negations have a clear inherent meaning. Third, combiners cannot form an utterance on their own, whereas negations can. Fourth, negations can have derived forms (οὐ and οὐχί), while ἤ cannot. Fifth, a negation with a verb forms a complete predicate, whereas ἤ with a verb needs another form, a second verb in this case.
[ back ] 132. Although the very first part of the treatise (Combiners 214.4-215.26) discusses the form of the words briefly, an actual definition of the category is missing. This is striking, but this first part is so lacunose that we may posit that a concise definition was lost in transmission.
[ back ] 133. See, for instance, Combiners 249.12-16. See also §70 on how Apollonnius connects this with Trypho’s thesis that “if they are words, they must mean something.”
[ back ] 134. συμπλεκτικοί (copulative), διαζευκτικοί (disjunctive), παραδιαζευκτικοί (subdisjunctive), συναπτικοί (hypothetical), παρασυναπτικοί (continuative), διασαφητικοί (comparative), αἰτιολογικοί (causal), ἀποτελεστικός (resultative), διαπορητικοί (dubitative), συλλογιστικοί (conclusive), παραπληρωματικοί (redundant).
[ back ] 135. There is a lacuna in the text after φράσεως, Dalimier 2001:71n1 gives Schneider’s conjecture: ὅλης γὰρ τῆς φράσεως < ὄντες συνδετικοί, τὰ ἐν αὐτῇ > πράγματα διαζευγνύουσιν: “for, being conjoiners of the entire phrase, they disjoin the realities in them”; also giving a translation of Bekker’s proposal. Neither seems convincing to me, as what is expected in this sentence is not a revisiting of both sides of the σύνδεσμοι διαζευκτικοί, but (ἕνεκα δὲ) an explanation of why they are called διαζευκτικοί. In that sense, the lacunose clause as it stands is clear enough: “they disjoin the realities of the entire utterance.” Whatever is missing in the lacuna must be expected to only add clarity to the construction, no more.
[ back ] 136. See De Kreij 2013 for a discussion of πρᾶγμα in the definitions given in the Téchnē and in Apollonius’ Combiners; I follow Dalimier 2001:467, who reads πρᾶγμα as “Réalité pensée correspondant à un énoncé,” which I paraphrase as “fact.”
[ back ] 137. Dalimier 2001:249n1+2notes the similarities with definitions in the scholia to Dionysius Thrax, but fails to discuss the similarities with the definition in the Téchnē.
[ back ] 138. Apollonius Dyscolus, Combiners 251.28.
[ back ] 139. See Dalimier 2001:411-412 for a discussion of the two definitions.
[ back ] 140. See §51 for the text and the reference.
[ back ] 141. Combiners 247.30-248.1 Ἔτι δὲ καί τινές φασιν οὐ δεόντως αὐτοὺς συνδέσμους εἰρῆσθαι, εἴγε συνδέσεως λόγων οὐκ εἰσὶν αἴτιοι, see §§53-54 above.
[ back ] 142. Combiners 247.30-248.1: ἔτι δὲ καί τινές φασιν οὐ δεόντως αὐτοὺς συνδέσμους εἰρῆσθαι, εἴγε συνδέσεως λόγων οὐκ εἰσὶν αἴτιοι.
[ back ] 143. On the conjecture <Χαιρήμων> (248.1) by Bekker, see Dalimier 2001:385-386. For the argument that Apollonius Dyscolus’ work has much in common with the ideas of the Stoics, see Blank 1982.
[ back ] 144. After adducing the examples of patronymics that may not give a father’s name, but are still called patronymics, and of masculine words that do not actually denote something masculine, but are still called masculine, he says οὕτω καὶ ἂν τύπῳ ᾖ ὁ παραπληρωματικὸς κεχορηγημένος συνδεσμικῷ, μὴ μὴν δηλουμένῳ, εἰρήσεται σύνδεσμος, in Apollonius Dyscolus, Combiners 248.8-9.
[ back ] 145. ᾧ καὶ συνηγορήσωμεν, ἔτι τινὰ προσθέντες. De Velsen [1853] 1965:44 also notes this addition by Apollonius.
[ back ] 146. Combiners 249.9-10: [εἰ λέ]ξεις, ὀφείλουσί τι δηλοῦν.
[ back ] 147. Regarding the argument that Apollonius Dyscolus is the first to claim this, see Wouters 1979:85n55 and Pecorella 1962:187-188.
[ back ] 148. Apollonius Dyscolus, Combiners 249.13.
[ back ] 149. Apollonius Dyscolus, Combiners 249.21-30; he compares λευκώλενος in Iliad 6.377 (πῇ ἔβη Ἀνδρομάχη λευκώλενος;) and Iliad 1.55 (λευκώλενος Ἥρη).
[ back ] 150. At the same time, some σύνδεσμοι παραπληρωματικοί have the same function and therefore need not be discussed separately: “It would be redundant to discuss ῥα after δή ... as we use them for the same thing [i.e. marking transitions]” (252.11-13).
[ back ] 151. See the discussion in §74 below.
[ back ] 152. Apollonius Dyscolus, Combiners 252.22-28.
[ back ] 153. Notes on particles by Apollonius will be referenced in the relevant places in the monograph.
[ back ] 154. Dalimier translates “contrariété supposée,” with a discussion on page 407, but a quick look at the other instances of this word in the TLG shows there is no reason to translate it as anything other than simply: opposition or contradiction. The question remains why Apollonius chooses ὑπεναντίωσιν over ἐναντίωσιν. It is tempting to regard the υπ- part as an indication that καίπερ shows an implied opposition. καίπερ signals that the clause it is in will in some way contradict the clause that follows after or comes before (such as: Even though he was a good man, he died a horrible death).
[ back ] 155. Apollonius Dyscolus, Syntax III.380.16.
[ back ] 156. Out of the five instances of αὔξησις and cognates in Apollonius three are in the description of περ, the other two in descriptions of the inflections of cases. As a result, they are not much help in explaining what Apollonius means here.
[ back ] 157. For more literature on δή, see V.δή.
[ back ] 158. Quoted in Apollonius Dyscolus, Combiners 254.22-25.
[ back ] 159. ἔμφασις can mean “suggestion (as opposed to expression),” Dickey 2007:237.
[ back ] 160. Apollonius Dyscolus, Combiners 250.6-10.
[ back ] 161. See §61 above.
[ back ] 162. Compare for example the Téchnē, which simply list some particles under more than one category, without further explanation.
[ back ] 163. See for this point II.3, III.4, IV.3, and IV.5 especially.
[ back ] 164. Aelius Herodianus, General Prosody 515-520.
[ back ] 165. συμπλεκτικοί, διαζευκτικοί, παραδιαζευκτικοί, συναπτικοί, παρασυναπτικοί, διασαφητικοί, αἰτιολογικοί, διαπορητικοί, συλλογιστικοί, παραπληρωματικοί.
[ back ] 166. Many of his notes are also preserved in the scholia to the Iliad.
[ back ] 167. παραλαμβάνεται (…) χάρις συνπλοκῆς, [χάρις] διαζεύξεως, [χάρις] ἀκολουθίας, [χάρις] αἰτίας, [χάρις] ἀπορίας, [χάρις] συλλογισμοῦ, [χάρις] τοῦ μὴ κεχηνέναι τὴν σύνθεσιν.
[ back ] 168. See I.3.2 for the similarity between this simple definition and modern attempts at defining particles and discourse markers.
[ back ] 169. See Di Benedetto 1958-1959 and Wouters 1979 and 1995; Wouters 1995:97 promises more papyri to be published in the Oxyrhynchus series, but as far as I have been able to establish, this has not happened yet.
[ back ] 170. The edition used is Keil 1859 Grammatici Latini, volumes II and III.
[ back ] 171. Copulativa, continuativa, subcontinuativa, adiunctiva, causalis, effectiva, approbativa, disiunctiva, subdisiunctiva, disertiva, ablativa, praesumptiva, adversativa, collectiva vel rationalis, dubitativa, confirmativa, completiva, see Keil III.93.13-16.
[ back ] 172. Pecorella 1962:186-196.
[ back ] 173. These commentators are hard to date, but Heliodorus is certainly later than the sixth century, as he excerpts the sixth century grammarian Choeroboscus. Schmidhauser 2010:509 dates him to the ninth century.
[ back ] 174. Pecorella 1962:188 posits that it would have figured in one of the lost parts of the Syntax (strikingly, not in the Combiners).
[ back ] 175. See §57 above.
[ back ] 176. The edition used is Hilgard 1901, with this passage on page 102. The words between are an emendation by Hilgard.
[ back ] 177. Pecorella fails to mention this similarity in his commentary.
[ back ] 178. Apollonius Dyscolus, Syntax I.14.4-6, quoted and discussed in §61 above.
[ back ] 179. See V.γάρ for literature.
[ back ] 180. See III.3.3.1.5 for discussion.
[ back ] 181. Eustathius, Commentary chapter 500.8-9.
[ back ] 182. See V.μέν for more literature, and II.2.4 for an analysis of μέν in Homer and Pindar.
[ back ] 183. The extremely inclusive list of particles and combinations that Devarius discusses are, in order: ἀλλά, ἀλλὰ γάρ, ἀλλ᾽ εἴπερ, ἀλλ᾽ εἰ ἄρα, ἀλλ᾽ ἤ, ἀλλ᾽ ἦ, ἀλλὰ μή, ἀλλὰ μήν, ἀλλά νὴ Δία, ἄλλοτι, ἀλλ᾽ ὅτι, ἄλλοτι ἤ, ἄλλοτι οὖν, ἄλλως τε καί, ἀμέλει, ἄν, ἄνθ᾽ ὧν, ἄνθ᾽ ὁτοῦ, ἄρα, ἀτάρ, αὐτάρ, ἄτινα, ἀτεχνῶς, αὖ, αὐτίκα, αὐτός, ἄχρι, μέχρι, γάρ, γε, γε δή, γέ τοι, γοῦν, δέ, δαί, δεῖνα, δή, δήπου, δῆθεν, δῆτα, Διά, δ᾽ οὖν, ἐάν, εἰ, εἰ βούλει δέ, εἰ δὲ μή, εἰ καί, εἰ μή, εἴπερ, εἰς, εἶτα, ἐμοί, ἕνεκα ἐμοῦ, ἐπεί, ἐπειδή, ἔπειτα, ἐξ ὧν, ἀφ᾽ ὧν, ἔξω, ἐς, ἔστε, ἔτι τοίνυν, ἐφ᾽ ὅσον, ἐφ᾽ ᾧ, ἐφ᾽ οἷς, ἐφ᾽ ὁ, ἕως, τέως, ἦ, ἤ, ἦ που, ἦ γάρ, ἦ γὰρ ἄν, ἤδη, ἦ μάλα δή, ἦ μήν, ἤπου, ἢ οὔ [sic], ἤτοι, ἵνα, καθό, καθ᾽ ὅτι, καθ᾽ ὅσον, καί, καὶ μήν, καὶ μέντοι, καὶ μὲν δή, καὶ τοίνυν, καίτοι, καὶ δή, κἄν, κοὐδ᾽ οὖν, μά, μέν, μέντοι, μὲν οὖν, μέχρι, μέχρις, μή, μήν, μὴ γάρ, μὴ δῆτα, μὴ ὅτι, μὴ οὐ, μὴ οὐκ, μήποτε, μήτι, μήτις, μή τοι γε, μόνον οὐ, μόνον οὐκ, μῶν, ναί, νὴ δία, νῦν δέ, ὁ, ὅδε, ἥδε, τόδε, ὃ δήποτε, οἷος, οἷον, οἴμοι, ὅμως, ὅπη, ὅπου, ὁπόταν, ὅπως, ὅς, ὅσος, ὅση, ὅσον, τόσος, ὅτε, ὅτι, ὅτι μή, ὅτι τοίνυν, οὐ, οὐκ, οὐ γάρ, οὐ γὰρ δή, οὐ γὰρ ἄν που, οὐ γάρ που, οὐ γὰρ ἀλλά, οὐδ᾽ εἰ, οὐδέν, οὐδέν τι μᾶλλον, οὐδὲ γὰρ οὐδέ, οὐδὲ μήν, οὐ δῆτα, οὐκοῦν, οὔκουν, οὐ μέν, οὐ μέντοι ἀλλά, οὐ μή, οὐ μήν, οὐ γὰρ ὅπως, οὕτως, οὐ μὴν οὐδέ, οὐ μὴν ἀλλά, οὐδέ, οὐ μήν πω, οὖν, οὔτι, οὔ τοι, οὐχ ὅτι, οὐχ ὅπως, οὐχ οἷον, οὐχ ὅσον, οὐ τοίνυν, πάνυ γε, πάνυ τοι, πάλιν, παρ᾽ ὅ, παρ᾽ ὅσον, περ, πῆ, πλήν, ποῦ, πρίν, προς, πῶς, σχολῇ, τὸ μέν, τὸ δέ, τὰ μέν, τὰ δέ, ταῦτα, προ τοῦ, ταύτῃ, ταχύ, τί, τί δέ...εἰ μή, τί δή, τί μήν, ποιός, τὸ καὶ τό, τὸ μέν τι, τὸ δέ τι, τοιοῦτος, τοῦτο μέν, τοῦτο δέ, τοι, τοίνυν, τοῦ, τῳ, ὑπέρ, ὑπὲρ ὧν, φεῦ, ῶ, ῶν, ὡς, ὥστε, ἄγε, φέρε, ἴθι, δεῦρο, ἴτε, ἔστιν ὅτε, ἔστιν οὗ, ἔστιν ὡς, ἐστί, τί δέ, τί οὖν, ἄπαγε, οὐαί, ὤ.
[ back ] 184. Both Budé 1529 and Estienne 1572 already used the term, like the ancient Latin grammarians, but Devarius first treated it as an independent – if very diverse – group.
[ back ] 185. ἀλλά, ἄρα, ἆρα, ἀτάρ, αὐτάρ, γάρ, γε, δέ, οὐδέ, μηδέ, δή, δαί, δῆθεν, δήπου, δήπουθεν, δῆτα, ἦ, θήν, καί, μάν (μέν/μήν), μέν, μέντοι, οὖν (ὦν), περ, που, τε, τοι, καίτοι, τοιγάρ, τοιγαροῦν, τοιγάρτοι, τοίνυν.
[ back ] 186. This is not to say that there were no discussions about which grammatical category several lexical items belonged to, as we find these as early as in the Homeric scholia. However, there was no attention given to a normative definition of the category. As n. 1 shows, however, even in modern times there are more modern scholars who note the problem than those who attempt to tackle it.