Barnes Jonathan. Truth, etc

Подождите немного. Документ загружается.

200 What is a Connector?

Are not non-simple expressions sentences, and does not the text imply that

connectors are essentially sentential? Alas, no. All that Plutarch means—as the

context makes quite plain—is this: whereas rhetoricians customarily praise

unconnected ﬁgures or asyndetons, logicians sometimes ﬁnd connectors

necessary for binding their assertibles together. Plutarch suggests, perhaps, that

logicians use connectors primarily in order to connect indicative sentences.

He does not suggest that that was, in general, the primary function of

connectors.

Or again, the deﬁnition of connectors offered by the Dionysian Art has

been taken to hide a sentential heart up its sleeve:

A connector is an expression connecting thoughts together with order.

Does not that mean that a connector connects one thought to another? And

is not a thought the sense of a sentence? So are not Dionysian connectors

deﬁned—at least so far as the ﬁrst clause in the deﬁnition is concerned—as

sentential links? Alas, no. The ancient commentators do not remark upon

that supposed difference between Dionysius and all the other grammarians,

who make connectors connect parts of sayings. Rather, they suppose that

‘connect thoughts’ and ‘connect parts of sayings’ are two different expressions

for the same thing: a connector may be said, indifferently, to connect a

thought (to bring it about that a string of expressions expresses a continuous

thought) or to connect the parts of sayings (to bring it about that the parts

stick together to form the expression of a continuous thought).

There are a few other authors who might be adduced; but they are all, I

think, equally irrelevant—all except for one man. For many modern scholars

agree with James Harris and claim that Apollonian connectors are essentially

sentential. Various arguments are advanced in favour of that agreeable thesis.

First, the vast majority of the illustrative examples in Connectors show

connectors linking indicative or assertoric sentences. Secondly, Apollonius

often explains the sense of a connector in a way which presupposes that it

makes a sentential connection. For example,

the announcement of the disjunctive connectors announces the holding of one item

and the rejection of the other or others.

(conj 216.14–16)⁵⁴

⁵⁴ ἡ ἐπαγγελία τῶν διαζευκτικῶν <ἑνὸς ὕπαρ>ξιν ἐπαγγέλλεται, τοῦ δ᾿ ὑπολειπομένου ἢ

<τῶν ὑπολειπομένων> ἀναίρεσιν.—The Greek text can be restored with certainty thanks to the

scholiast to Dionysius Thrax, 104.28–30.

What do Apollonian Connectors Connect? 201

The items in question are evidently sentential in their nature; and the

disjunctive connectors are thus deﬁned as sentential links.

Thirdly, Apollonius sometimes implies—or at least seems to imply—that

by their very nature connectors bind sayings or sentences. Here are two

passages from Connectors and three from the Syntax whichtellinthatsense.

The negation when placed before a verb makes a complete saying: I’m not writing,

I’m not walking. That is characteristic of adverbs. But ‘or’ never does that … rather,

another saying must be taken … and that is characteristic of connectors: I’m writing

or I’m talking, I’m conversing or I’m reading.

(conj 222.18–23)⁵⁵

The text is lacunose, the general sense certain: the expression ‘or’ cannot be

added to a verb to make a saying—it calls for a second saying. The second

passage from Connectors is this:

Again, ‘It is day and’ is not complete: it demands another phrase—‘Both it is day

and it is light’. It is the same with disjunctive connectors: ‘Either it is day or it is

night’. So we have proved that they are connectors inasmuch as they connect phrases.

(conj 216.6–10)⁵⁶

This passage uses ‘φράσις’ rather than ‘λόγος’; but the words are synonymous.

According to the Syntax,

in sayings, too, the accompanying connectors sometimes unify two sayings or even

more—for example, sayings constructed from conditionals or quasi-conditionals or

again conjunctives. And conversely their absence dissolves the sayings, as in

We went up into the forest as you ordered, noble Odysseus.

We found in the woods ﬁne houses built.

It should have been bound together with ‘and’:

And we found in the woods …

(synt i 10 [11.3–10])⁵⁷

⁵⁵ ἡ ἀπόφασις πρὸ ῥήματος ἐπιτασσομένη ποιεῖ αὐτοτέλειαν—οὐ γράφω, οὐ περιπατῶ.

ἦν δὲ τὸ τοιοῦτον ἴδιον ἐπιρρημάτων. ὁ δὲ ἤ οὐδέποτε τὸ τοιοῦτον ἀπετέλεσε ... τοῦ δ᾿ ἑτέρου

λόγου ἐξ ἀνάγκης παραλαμβανομένου ... ὅπερ ἦν ἴδιον συνδέσμου—γράφω ἤπερ λέγω,

διαλέγομαι ἤπερ ἀναγινώσκω.

⁵⁶ ἔτι δὲ οὐκ αὐτοτελὲς τὸ ἡμέρα ἐστὶ καί, ἀλλ᾿ ἐζήτει ἑτέραν φράσιν· καὶ ἡμέρα ἐστὶ

καὶ φῶς ἐστί. τὸ αὐτὸ δὲ καὶ ἐπὶ τῶν διαζευκτικῶν· ἢ ἡμέρα ἐστὶν ἢ νύξ ἐστι. καὶ ὡς μὲν

σύνδεσμοι κατὰ τὸ συνδεῖν τὰς φράσεις εἰσίν, ἀπεδείχθη.

⁵⁷ ἀλλὰ κἂν τοῖς λόγοις οἱ παρεπόμενοι σύνδεσμοι ἔσθ᾿ ὅτε ἑνοῦσι δύο λόγους ἢ καὶ πλείους,

καθάπερ οἱ συνδεόμενοι λόγοι ἐκ συνημμένων ἢ παρασυνημμένων ἢ καὶ ἔτι συμπεπλεγμένων·

ἢ πάλιν ἀποστάντες διάλυσιν τῶν λόγων ποιοῦνται, ὡς ἔχει τὸ ᾔομεν, ὡς ἐκέλευες, ἀνὰ

δρυμά, φαίδιμ᾿ ᾿Οδυσσεῦ· εὕρομεν ἐν βήσσῃσι τετυγμένα δώματα καλά. ἔδει γὰρ συμπλέξαι

τῷ καί· καὶ εὕρομεν ἐν βήσσῃσι.

202 What is a Connector?

A little later, Apollonius remarks that a simple sentence may contain all

but one of the seven parts of sayings which he recognizes, for example:

The same man, slipping, fell today.

The parts of sayings are there, apart from the connector, since when that is added it

will demand another saying.

(synt i 14 [17.4–6])⁵⁸

The saying contains, in order of occurrence, an article, a pronoun, a name,

a participle, a verb, an adverb. It does not contain a connector, and in fact

you cannot provide an example of a simple saying which uses all seven parts

of sayings; for one of them, the connector, necessarily introduces a second

saying.

Again, to show that certain items are not connectors Apollonius urges that

no one will think this—that they are connectors—inasmuch as they do not connect

the adjunction of another saying, which is characteristic of connectors.

(synt iii 69 [334.5–6])⁵⁹

Other passages might be adduced. Collectively, they provide strong evid-

ence—conclusive evidence, many scholars have thought—that Apollonian

connectors essentially connect sentences.

And that conclusion is strengthened by a passage which seems at ﬁrst

glance to refute it. In Adverbs Apollonius observes that

some connectors connect name and verb universally and others are particular.

(adv 121.12–13)⁶⁰

That is offered as an exhaustive disjunction: connectors, in other words,

connect names and verbs, some of them connecting any old names and verbs

and others being more particular in their tastes. So connectors connect names

and verbs: that is to say, surely, that they connect names with names and verbs

with verbs, or else names with verbs—and in any event that they connect

parts of sayings rather than sentences.

The text readily lends itself to that reading, and Apollonius does nothing

to warn his readers against it. But it is quite certainly a false reading.

⁵⁸ ὁ αὐτὸς ἄνθρωπος ὀλισθήσας σήμερον κατέπεσεν· ἔγκειται τὰ μέρη τοῦ λόγου παρὰ τὸν

σύνδεσμον, ἐπεὶ προστεθεὶς ἕτερον λόγον ἀπαιτήσει.

⁵⁹ οὐδὲ γὰρ ἐκεῖνό τις οἰήσεται, ὡς σύνδεσμοί εἰσιν, καθὸ οὐ συνδέουσιν ἐπιφορὰν ἑτέρου

λόγου, ὅπερ ἴδιον συνδέσμων.

⁶⁰ τίνες τε ἐν τῷ καθόλου σύνδεσμοι συνδέουσιν ὄνομα καὶ ῥῆμα, καὶ τίνες εἰσὶ μερικοί.

What do Apollonian Connectors Connect? 203

The sentence occurs in a paragraph in which Apollonius is remarking on

the primacy of names and verbs among the parts of sayings: they are the

‘thematic’ or fundamental parts of sayings—the other parts merely serve

their needs. Now a connector may serve the needs of names and verbs by,

say, making a complex sentence from a couple of simple sentences. In doing

so, it ‘connects name and verb’—that is to say, it connects one name-verb

couple to another. That is indubitably what Apollonius meant to say; and

the connectors of that passage in Adverbs are therefore sentential.

For almost all the grammarians, connectors ofﬁcially connect parts of

sayings. For Apollonius, it now appears, they connect sayings or sentences.

If that is so, then the fact that Apollonius thus distinguishes himself from

his colleagues must have escaped the attention of the ancient commentators.

There are, admittedly, traces of the apparently Apollonian view here and

there. Thus one of the commentators on the Dionysian Art states that

the items in question are called connectors in virtue of the fact that they connect

expressions and phrases, and they are called disjunctive connectors in virtue of their

meaning. For while they are connective of the whole phrase, they disjoin the objects.

(scholiast to Dionysius Thrax, 104.25–28)⁶¹

He is tacitly paraphrasing a passage from Apollonius’ Connectors;buthedoes

notnoticethatitis—orappearstobe—quiteatoddswithhisofﬁcialview

about the nature of connectors.

Apollonian connectors are sentential—or at any rate, they are essentially

or fundamentally or paradigmatically sentential. That conclusion is not easily

resisted—and it is, of course, logically seductive. It has certain consequences.

For example, if it is accepted, then we must accuse Apollonius of carelessness

when he declares at the beginning of the Syntax that connectors are connective

of the parts of sayings. We shall also be inclined to judge that the improved

deﬁnition of connectors is not, after all, Apollonian. And we may like

to conjecture that the Stoics originally explained connectors as sentential

connectives; that Apollonius adopted, and also elaborated, their deﬁnition;

and that the other Greek grammarians, together with all the Latins, either

misunderstood the deﬁnition or else deliberately modiﬁed it, and thereby

introduced the notion that connectors essentially connect parts of speech, a

notion which came to dominate the whole of the grammatical tradition.

⁶¹ καὶ οἱ προκείμενοι οὖν σύνδεσμοι μὲν εἴρηνται ἕνεκα τοῦ συνδεῖν τὰς λέξεις καὶ τὰς

φράσεις, ἕνεκα δὲ τοῦ ἀπ᾿ αὐτῶν δηλουμένου σύνδεσμοι διαζευκτικοὶ ὠνομάσθησαν· ὅλης γὰρ

τῆς φράσεως ὄντες συνδετικοὶ τὰ ἐν αὐτῇ πράγματα διαζευγνύουσιν.

204 What is a Connector?

APOLLONIAN SAYINGS

The argument which makes Apollonian connectors sentential depends partly

on the claim that by ‘λόγος’—and also by ‘φράσις’—he means ‘sentence’.

So I must say something about Apollonian sayings.

I have already noticed that although Aristotle frequently uses the word

‘λόγος’ to denote sentences—and in particular, indicative sentences—his

formal deﬁnition determines as a λόγος any expression whatever which is

more complex than a single name or verb. The same is true of Plato, who has

this to say when, in the Cratylus, he sets out to describe the nature of language:

So we shall apply the letters to the objects, either one by one, wherever they seem

to be required, or several together, thereby making what they call syllables, and then

putting syllables together, from which names and verbs are put together; and then,

again, from the names and verbs we shall assemble something grand and noble and

whole—just as painters compose a picture by the art of painting, so we shall compose

a saying by the art of naming or speaking or whatever it is.

(Crat 424e –425a)⁶²

We start with letters, which are the elements or atoms of language. From

letters we make syllables, and from syllables words. Finally, from words we

make sayings.

Letters, syllables, words, sayings—the four-tiered hierarchy described by

Plato became a standard part of ancient linguistic science. It is set out, for

example, on the ﬁrst pages of Apollonius’ Syntax (i 2 [2.3–3.3]), and it serves

to introduce his general conception of linguistic construction. The hierarchy

recognizes no level above the level of the saying, and no level between the

level of the word and the level of the saying. In general, ancient grammarians

recognized the four Platonic levels, and no others.

It is true that a number of ancient theorists talk about linguistic units

which they call ‘commas’ and ‘colons’; and commas and colons are more than

words and less than complete sayings (and commas are shorter than colons).

For example, in his brief discussion of the deﬁnition of connectors given by

the Dionysian Art, one of the ancient commentators asks:

⁶² οὕτω δὴ καὶ ἡμεῖς τὰ στοιχεῖα ἐπὶ τὰ πράγματα ἐποίσομεν, καὶ ἓν ἐπὶ ἕν, οὗ ἂν δοκῇ

δεῖν, καὶ σύμπολλα, ποιοῦντες ὃ δὴ συλλαβὰς καλοῦσιν, καὶ συλλαβὰς αὖ συντιθέντες, ἐξ ὧν

τά τε ὀνόματα καὶ τὰ ῥήματα συντίθενται· καὶ πάλιν ἐκ τῶν ὀνομάτων καὶ ῥημάτων μέγα

ἤδη τι καὶ καλὸν καὶ ὅλον συστήσομεν, ὥσπερ ἐκεῖ τὸ ζῷον τῇ γραφικῇ, ἐνταῦθα τὸν λόγον

τῇ ὀνομαστικῇ ἢ ῥητορικῇ ἢ ἥτις ἐστὶν ἡ τέχνη.

Apollonian Sayings 205

What does interpretation mean?—The sentence. For a connector harmonizes to

itself—I mean, by means of itself—the colons and commas.

(scholiast to Dionysius Thrax, 436.34–35)⁶³

The sense of that remark is less than pellucid; but one thing is plain: an

ancient grammarian here refers to commas and colons, and treats them as

syntactical units (which may be united by connectors).

But the commentator says nothing more on the matter; and so far as

I am aware, only one other grammatical text ever mentions commas and

colons. That text is found in another of the commentaries on the Dionysian

Art —and there the commentator associates the items with the rhetoricians,

who are explicitly contrasted with the grammarians (scholiast to Dionysius

Thrax, 295.41–296.1). In other words, one errant passage apart, commas

and colons are not treated as grammatical or syntactical units. And that is

because in fact they are not grammatical or syntactical units: they are metrical

units, comparable to lines of verse or to strophes or to stanzas. Hephaestion

discusses them brieﬂy in his Handbook on Metre (ench xvi 114); Dionysius

of Halicarnassus mentions them in his remarks on prose rhythm (comp verb

xxvi 213).

In short, it is true that commas and colons are linguistic units, and it is

true that they are (normally, at least) more than words and less than complete

sayings. But they do not, for that reason, stand between words and sayings:

they do not belong to the same sort of classiﬁcation as words and sayings,

they have nothing to do with syntax or with grammatical theory.

So as far as ancient grammar is concerned, there is no class of items between

the class of words and the class of sayings, and there is no class of items

more complex than sayings. In that case, sentences will surely count among

sayings—and, of course, no one has ever doubted that they do so count.

But equally surely, certain types of non-sentence ought to count among

as sayings—and in fact Aristotle, as I have already recorded, recognizes

subsentential units (such as deﬁnitional formulas) and also supersentential

units (such as the Iliad ) as sayings. Indeed, anything from a two-word phrase

to a twenty-four book poem ought to count as a saying—provided, of course,

that it has some syntactical unity. Perhaps it seems curious to place The

Tempest and ‘the tempest’ in the same syntactical class; but what else can you

do once you have kowtowed to the four-level hierarchy?

⁶³ τί σημαίνει ἑρμηνεία; τὴν φράσιν· ὁ γὰρ σύνδεσμος τὰ κῶλα καὶ κόμματα συναρμόττει

αὑτῷ, τουτέστι δι᾿ ἑαυτου.

206 What is a Connector?

Even if they had not read their Aristotle, the grammarians can hardly have

failed to notice the existence of items larger than sentences—of sequences of

sentences. So what sort of unit is a sequence of sentences? Well, if you take

two or more sayings and make some one item out of them, the one item must

be a saying—there is nothing else for it to be. If you regard The Tempest

as a single unit—I mean, as a single syntactical unit—then The Tempest is

a saying. Of course, an ancient grammarian was not obliged to classify The

Tempest as a saying; for he was not obliged to think of it as a syntactical unit

of any sort. Indeed, the play surely isn’t such a unit. Nonetheless, there are

certain sequences of sentences which the ancient grammarians did take to

be syntactical units—and those items can only be classiﬁed as sayings. (The

items in question will come in for a later discussion.)

If the grammarians allow that sequences of sentences are sayings, then their

class of sayings cannot be identiﬁed with the class of sentences.

Again, consider a subsentential item. In the sentence

Jehu driveth furiously

an ancient grammarian would have discovered three words (and eight syllables,

and so many letters). But it is surely plain that the sentence also consists of the

name ‘Jehu’ together with the verbal formula ‘driveth furiously’, which is itself

composed from the words ‘driveth’ and ‘furiously’. True, someone might try

to deny that ‘driveth furiously’ is a proper part of the sentence; someone

might try to claim that the sentence simply consists of a string of three words

and that there is no structure to it apart from its string-like structure. But

such an attitude is not only hopelessly inadequate from a theoretical point of

view: it is also an attitude which an ancient grammarian could hardly have

contemplated with approval. For the ancient grammarians insist that adverbs

are ‘predicated’ of verbs, that they attach to verbs, that they modify verbs. In

other words, they implicitly recognize that within the sentence

Jehu driveth furiously

‘driveth furiously’ forms a unit.

What sort of unit? Not a word; for it has parts which signify separ-

ately—and by any criterion for counting it will be deemed to consist of at

least two words. So it must be a saying. There is nothing else for it to be. But if

the grammarians allow that items such as ‘driveth furiously’ are sayings, then

their class of sayings cannot possibly be identiﬁed with the class of sentences.

And nevertheless they do seem to identify sayings and sentences. No

deﬁnition of the word ‘saying’ is explicitly ascribed to Apollonius; but

there are deﬁnitions in the other grammarians, Greek and Latin. The

Apollonian Sayings 207

Greek word ‘λόγος’, as every schoolboy knows, is multiply ambiguous.

The commentators on the Dionysian Art offer lengthy accounts of its

several senses (e.g. 213.6–214.2; 353.27–355.15); and they are, of course,

particularly interested in the sense which the Art deﬁnes. This is what is

found in the Art:

An expression is a smallest part of a constructed saying. A saying is a combination of

prose expressions which indicates a complete thought.

(11 [22.4–23.1])⁶⁴

The phrase ‘complete thought’ is hardly precise. But it is precise enough

for me here. For it is plain that a paragraph does not, as a rule, express a

complete thought—rather, it expresses a string of complete thoughts; and it

is quite plain that an expression such as ‘driveth furiously’ does not express a

complete thought. So a saying, according to the Art, is a sentence of prose.

The commentators were not prepared to limit sayings to prose. One of

them objects briskly:

This excludes verse sayings. So a saying should be deﬁned as follows: A combination

of expressions which is well-formed and which rounds off a thought.

(scholiast to Dionysius Thrax, 214.5–6)⁶⁵

This revised deﬁnition is found in one variant of another in several Greek

and Latin texts. Diomedes offers two deﬁnitions—ﬁrst there is something

which I do not understand and then comes a version of the commentator’s

Greek, thus:

A saying is a construction of words in which an outcome is composed and which

ends in a closure. Some deﬁne it thus: A saying is a combination of expressions which

rounds off a thought and signiﬁes a complete object.

(ars gramm i 300.17–19)⁶⁶

More signiﬁcantly, Priscian comes even closer to the Greek:

A saying is a sequence of expressions which is well formed and which indicates a

complete thought.

(inst ii iv 15 [ii 53.28–29])⁶⁷

⁶⁴ λέξις ἐστὶ μέρος ἐλάχιστον τοῦ κατὰ σύνταξιν λόγου. λόγος δέ ἐστι πεζῆς λέξεως

σύνθεσις διάνοιαν αὐτοτελῆ δηλοῦσα.

⁶⁵ τοῦτο ἐκβάλλει τοὺς ἐμμέτρους. ὁριστέον οὖν οὕτως· σύνθεσις λέξεων κατάλληλος

διάνοιαν ἀπαρτίζουσα.

⁶⁶ oratio est structura verborum composito exitu ad clausulam terminata. quidam sic eam deﬁniunt:

oratio est compositio dictionum consummans sententiam remque perfectam signiﬁcans.

⁶⁷ oratio est ordinatio dictionum congrua sententiam perfectam demonstrans.

208 What is a Connector?

Priscian habitually follows Apollonius, and here he paraphrases or trans-

lates the revised deﬁnition of what sayings are: so is it not likely that

the revised deﬁnition is Apollonian? If so, then Apollonian sayings are,

ofﬁcially, sentences.

Now the subject of Apollonius’ Syntax is the construction—the syntax—of

what he calls complete sayings. That is what the opening words of the work

announce:

In the lectures which we have already made public an account of utterances was

laid out as the matter demanded. The publication which is now to be presented

will contain the construction of utterances into well-formed and complete sayings,

a matter which I have decided to set out with all exactitude inasmuch as it is

indispensable for the interpretation of poetry.

(synt i 1 [1.1–2.2])⁶⁸

If anything is a complete saying, then surely a sentence is. So, for example, in

Connectors Apollonius notes that

‘It is day’ is complete, but ‘Either it is day’ is not complete.

(conj 225.9–10)⁶⁹

‘It is day’ is a sentence, ‘Either it is day’ is a sentence fragment. The

sentence is a complete saying, the fragment is not. That example takes an

indicative or assertoric sentence; but Apollonius sometimes illustrates his

remarks with other sorts of sentences—with imperatives, for example, or

with interrogatives—and there is no reason in the world to doubt that they

too are to be reckoned among complete sayings.

Should complete sayings be identiﬁed with sentences? The identiﬁcation is

certainly tempting, and temptation is not a thing I like to resist. But consider

the following case. ‘Mimi, Mimi’, sobs Rodolfo—and the curtain falls. Does

La Boh`eme end with a sentence? According to Apollonius, it ends with a

complete saying; for

the vocative, being complete, demands a mark of punctuation.

( pron 53.16)⁷⁰

⁶⁸ ἐν ταῖς προεκδοθείσαις ἡμῖν σχολαῖς ἡ περὶ τὰς φωνὰς παράδοσις, καθώς ἀπῄτει ὁ περὶ

αὐτῶν λόγος, κατείλεκται· ἡ δὲ νῦν ῥηθησομένη ἔκδοσις περιέξει τὴν ἐκ τούτων γινομένην

σύνταξιν εἰς καταλληλότητα τοῦ αὐτοτελοῦς λόγου, ἣν πάνυ προῄρημαι, ἀναγκαιοτάτην

οὖσαν πρὸς ἐξήγησιν τῶν ποιημάτων, μετὰ πάσης ἀκριβείας ἐκθέσθαι.

⁶⁹ τὸ ἡμέρα ἐστίν αὐτοτελές, ἀλλὰ τὸ ἤτοι ἡμέρα ἐστίν οὐκ αὐτοτελές.

⁷⁰ ἡ κλητικὴ αὐτοτελὴς οὖσα στιγμὴν ἀπαιτεῖ.

Apollonian Sayings 209

And again, more expansively:

I have not forgotten that completeness is evidence for the vocative. Take ‘Helicon’—if

it requires a verb, that testiﬁes to the nominative; if not, it is in the vocative case, e.g.

OHelicon.

(synt iii 119 [372.7–10])⁷¹

A saying is complete if it can be used to say something complete—if, in the

Stoic jargon, it can be used to express a complete sayable. Having sobbed his

‘Mimi, Mimi’, Rodolfo has nothing more to add.

If you feel qualms about classing isolated vocatives as sentences, then you

will deny that all Apollonius’ complete sayings are sentences: all sentences are

complete sayings, some complete sayings are not sentences.

What about suprasentential units? What about the Iliad? Is Homer’s poem,

according to Apollonius, a complete saying? There is, so far as I have noticed,

nothing in the surviving texts which determines an answer to that question.

But there are a few passages—which will be cited in a later context—which

explicitly classify as sayings, and implicitly as complete sayings, items which

we should be inclined to count as sequences of sentences rather than as

sentences.

However that may be, in practice Apollonius’ complete sayings virtually

coincide with sentences. What, then, is the relation between a saying and a

complete saying? Scholars seem generally to suppose that the phrase ‘complete

saying’ is pleonastic, so that something is a complete saying if and only if it is

a saying. In that case, Apollonian sayings will, in practice, virtually coincide

with sentences.

But why suppose that the phrase is pleonastic? After all, the notion of

an incomplete saying is far from absurd—and in many ways it seems to be

just the sort of notion which Apollonius would and should have embraced.

For the Stoics distinguish between complete sayables and incomplete or

deﬁcient sayables, among which they number predicates; and just as a

complete saying—a sentence—is an expression of a complete sayable,

so an incomplete saying will be an expression of an incomplete sayable.

Since verbs and verbal phrases express predicates, such items as ‘runs’, and

‘runs silently’, and ‘runs silently and very fast’ will count as incomplete

sayings.

⁷¹ οὐ λέλησμαι ὅτι καὶ ἡ αὐτοτέλεια τεκμήριόν ἐστιν κλητικῆς· ἰδοὺ γὰρ καὶ αὐτὸ τὸ

῾Ελικών ἐλλεῖπον μὲν ῥήματι εὐθεῖαν ὁμολογεῖ, οὐ τῇδε δὲ ἔχον κλητικῆς ἐστιν πτώσεως τὸ

τοιοῦτον, οἷον ὦ ῾Ελικών.