Barnes Jonathan. Truth, etc

Подождите немного. Документ загружается.

230 What is a Connector?

Thecompoundphrase‘tu et ego’ is a unity—a single pronominal for-

mula—inasmuch as it functions syntactically in the same way as a pronoun

does. The compound phrase ‘silently and very fast’ is an adverbial phrase—a

single or uniﬁed adverbial phrase—inasmuch as its syntax is the syntax of the

adverb. More precisely, ‘tu et ego’ is a pronominal phrase because it can be

embedded in a pronominal context, and ‘silently and very fast’ is an adverbial

phrase because it can be embedded in an adverbial context. Take any sentence

which contains an adverb—say:

He driveth furiously.

Remove the adverb and contemplate the shell which remains:

He driveth—.

That, I shall say, is an adverbial context. An item can be embedded in

that context if and only if the result of putting it in place of the dash is

grammatically well formed. So ‘silently and very fast’ can be embedded in

He driveth—;

for

He driveth silently and very fast

is grammatically impeccable.

An item is an adverbial formula if and only if it can be embedded in at

least one adverbial context. And in general, an item is a formula of a given

syntactical type if and only if it can be embedded in at least one context of

that type. That is what I shall call the embedding test.

Is there a similar embedding test for sayings? Frege argued that expressions

I came, and I saw, and I overcame

constitute a single sentence by observing that they can be negated. You can

say

It’s not the case that I came, and I saw, and I overcame.

That is syntactically impeccable and semantically pellucid. You cannot do

the same for

I came, I saw, I overcame.

So that doesn’t constitute a single sentence. Frege’s negation test is a special

case of the embedding test. He supposes that

It is not the case that—

is a sentential context: an item is a sentence if and only if it can be embedded

in that context or in some other such sentential context. Other sentential

contexts will include items like ‘If—, then I’m a Dutchman’, ‘God knows

whether or not—’, ‘Firmly I believe and truly that—’, and so on.

Glue 231

Or is that a test not for sentences in general but for assertoric sentences in

particular? Well, is the following item a sentential context?

—or let nature deal with the problem.

If it is, then it will ensure that, say,

Cut off the affected parts and spray with Bordeaux mixture

is a sentence. If not, not. In other words, and unremarkably, an embed-

ding test will test for sentences in general or for assertoric sentences in

particular according to whether the sentential contexts to which it appeals

are determined by assertoric sentences in particular or by sentences in

general.

The sentential test is a test for sayings only if sayings are sentences. There

are also tests for subsentential sayings. But it is hard to think of any test

for suprasentential sayings. Why—to repeat an earlier question—should the

Iliad count as a single saying whereas the Iliad and the Odyssey together must

count as two?

GLUE

So much for the grammar of connectors. What, next, of their semantics?

What do connectors mean? Indeed, do they mean anything at all?

Apollonius says, negatively, that ‘connectors never signify anything in their

own right’ (adv 133.25);⁹⁶ and he says, positively, that ‘connectors co-signify’

(conj 222.12–13).⁹⁷ Indeed, like prepositions and articles, connectors always

co-signify (synt i 12 [14.1–2]). The co-signiﬁcation of connectors is part of

what I called the improved deﬁnition; and it is a commonplace in the later

tradition. For example,

Connectors are among the expressions which co-signify. They do not signify anything

in their own right but they connect a gaping thought—that is why they are called

connectors. For in themselves connectors signify nothing, but when they are put in

the construction they bind together what was missing or gaping or dissolved.

(scholiast to Dionysius Thrax, 284.6–10)⁹⁸

⁹⁶ οἱ ... σύνδεσμοι οὔποτε κατ᾿ ἰδίαν σημαίνουσί τι.

⁹⁷ οἱ σύνδεσμοι συσσημαίνουσιν.

⁹⁸ ὁ σύνδεσμος τῶν συσσημαινουσῶν ἐστι λέξεων· οὐ γὰρ καθ᾿ ἑαυτόν τι σημαίνει, συνδεῖ

δὲ διάνοιαν κεχηνυῖαν· διὰ τοῦτο γὰρ καὶ σύνδεσμος ὠνομάσθη· καθ᾿ ἑαυτὸν γὰρ ὁ σύνδεσμος

οὐδὲν σημαίνει, συντασσόμενος δὲ τὰ ἐλλείποντα ἢ κεχηνότα ἢ διαλελυμένα σφίγγει.

232 What is a Connector?

So too in the Latin tradition, where Priscian says of connectors that

they always co-signify, i.e. signify when connected to other items, but not by

themselves.

(inst xvii i10[iii 114.18–20])⁹⁹

And the idea was familiar outside the grammatical tradition.

One of the questions with which Dexippus deals in his set of Questions

and Answers on Aristotle’s Categories is this: ‘Why, they worry, did he omit

the connectors?’—that is, why is the account of homonymy and synonymy

at the beginning of the Categories apt for some sorts of words but not for

others, and in particular not for connectors? After all, some connectors are

ambiguous, and an account of homonymy which can say nothing about them

is, pro tanto, inadequate. Here is Dexippus’ answer:

It is because, or so we claim, the utility of connectors for sayings is not primary but

secondary, not complete but incomplete, not expressive but rather symbolic—indeed,

connectors do not even primarily signify but rather co-signify, like the double lines

which we write in the margin and which along with what has been written co-signify

the ﬁnished nature of the thought while themselves in themselves signifying nothing.

So connectors, too, co-signify along with the other parts of sayings but themselves

in themselves are not signiﬁcant—rather, they are like glue. That is why we do not

count them among the elements of sayings but, if anything, as parts of expressions.

(in Cat 32.17–26)¹⁰⁰

Dexippus doubtless took the question, and his answer to it, from Porphyry.

Simplicius also took the matter over; and he added the information that the

objection which the answer addresses had been advanced by Lucius (see in

Cat 64.18–28). Lucius thinks that any account of signiﬁcant expressions

must surely make room for connectors. Simplicius disagrees: connectors are

not—or not really—signiﬁcant expressions. After all, Aristotle had said that

‘a connector is a non-signiﬁcant expression’ (Poet 1456b38).¹⁰¹

⁹⁹ eae enim semper consigniﬁcant, id est coniunctae aliis signiﬁcant per se autem non.

¹⁰⁰ ἀλλὰ διὰ τί τοὺς συνδέσμους παρέλιπεν ἀποροῦσιν.—ἐπειδή, φαμὲν ἡμεῖς, οὔτε προ-

ηγουμένη ἐστὶν αὐτῶν ἡ χρεία τοῦ λόγου ἀλλὰ δευτέρα, οὔτε τελεία ἀλλ᾿ ἀτελής, οὔτε λεκτικὴ

συμβολικὴ δὲ μᾶλλον· ἀλλ᾿ οὐδὲ σημαίνει προηγουμένως συσσημαίνει δὲ μᾶλλον, ὥσπερ

τὰς διπλᾶς εἰώθαμεν παραγράφειν, αἵτινες μετὰ τῶν γεγραμμένων μὲν συσσημαίνουσι τὸ

ἀπαρτίζον τῆς διανοίας αὐταὶ δὲ καθ᾿ ἑαυτὰς οὐδὲν δηλοῦσι. καὶ οἱ σύνδεσμοι τοίνυν συσση-

μαίνουσι μετὰ τῶν ἄλλων μερῶν τοῦ λόγου, αὐτοὶ δὲ καθ᾿ ἑαυτοὺς οὐκ εἰσὶ σημαντικοί, ἀλλ᾿

ἐοίκασι τῇ κόλλῃ· διόπερ οὐδὲ λόγου στοιχεῖα αὐτοὺς τιθέμεθα, ἀλλ᾿ εἴπερ ἄρα μέρη λέξεως.

¹⁰¹ σύνδεσμος δέ ἐστιν φωνὴ ἄσημος.

Glue 233

The commentators were appealing to an ancient theory, ascribed to Plato

and associated with Aristotle and with Theophrastus:

Insofar as they [i.e. predicates] are expressions, they involve other inquiries, which

Theophrastus stirred up in his On the Elements of Sayings and which his followers

wrote about—for example, whether names and verbs are the elements of sayings or

whether articles and connectors and certain other items are so too (these are indeed

parts of expressions, whereas names and verbs are parts of sayings).

(Simplicius, in Cat 10.23–27)¹⁰²

There are only two genuine parts of sayings, names and verbs. All the other

sorts of item which turn up in complex expressions are parts of those expres-

sions but not parts of the sayings; they do not signify anything in themselves

but co-signify; they are compared to glue and nails and dowels and the like.

The comparison with glue and nails was popular: it was exploited by

Apuleius (or whoever wrote the Latin On Interpretation); it is one of several

similar metaphors which Plutarch uses in his essay on the parts of speech

(quaest Plat 1009f –1010d);anditiselaboratedbyAmmonius:

Those items which signify a nature or a person or an activity or a passivity or some

combination of person with activity or passivity—all those Aristotle divides into

names and verbs. He calls verbs those which are timed or which are predicated in

propositions, and names those which are without time or which supply the need for

subjects. Items which are used in neither of those two r

oles,evenifinsomeother

fashion they are appended to the propositions (signifying that the predicate holds

of the subject, or that it does not hold, or that it holds at a certain moment or

in a certain manner or a certain number of times, or indicating any other relation

between the two), these he claims are not strictly called parts of sayings. For just as

the planks are strictly parts of a ship while the dowels and the oakum and the pitch

are used for the sake of connecting the planks and unifying the whole ship, in the

same way in sayings connectors and articles and prepositions and even adverbs satisfy

the need for a sort of dowel and are not justly called parts—after all, when combined

on their own they cannot produce a complete saying. These items, then, are not

parts of sayings but parts of expressions, of which sayings are themselves parts, as he

says in the Poetics; and they are useful for combining and constructing the parts of

sayings one with another, just as a cord is useful for the artiﬁcial uniﬁcation of the

items corded, and glue for the items which hold together by its means—but they are

¹⁰² καθὸ μὲν γὰρ λέξεις, ἄλλας ἔχουσι πραγματείας, ἃς ἐν τῷ Περὶ τῶν τοῦ λόγου

στοιχείων ὅ τε Θεόφραστος ἀνακινεῖ καὶ οἱ περὶ αὐτὸν γεγραφότες, οἷον πότερον ὄνομα καὶ

ῥῆμα τοῦ λόγου στοιχεῖα ἢ καὶ ἄρθρα καὶ σύνδεσμοι καὶ ἄλλα τινά (λέξεως δὲ καὶ ταῦτα

μέρη, λόγου δὲ ὄνομα καὶ ῥῆμα) ...

234 What is a Connector?

not themselves parts of the items corded or glued, nor are connectors or articles or

prepositions or adverbs parts of sayings.

(in Int 12.16–13.6)¹⁰³

The comparisons or metaphors are elaborated in the philosophical texts.

But they are unsurprising—indeed, the metaphor of binding is present in the

very nomenclature of connection—and we might expect the grammarians to

exploit them. Apollonius does so, lightly and in one passage:

After the parts which we have listed we mentioned the connector which is connective

of them: on its own and apart from the matter of expressions it cannot establish

anything—just as the bindings of bodies are of no use if the bodies do not exist.

(synt i 28 [27.10–13])¹⁰⁴

But the later grammarians did not adopt, let alone elaborate, the topic. Rather,

they ascribed the metaphors to ‘the Peripatetics’ (scholiast to Dionysius

Thrax, 515.19–29), or to ‘certain philosophers’ (Priscian, inst xi ii 6–7 [ii

551.18–552.17]). Moreover, they sometimes argued against the aptness of

the ﬁgures. Priscian, for example, has a long discussion which concludes that

it is far better to side with those who call the name and the verb principal and

pre-eminent parts, and the rest their appendages.

(inst xi ii 7 [552.12–14])¹⁰⁵

¹⁰³ τὰ μὲν οὖν φύσεων ἢ προσώπων ἢ ἐνεργειῶν ἢ παθῶν ἢ ποιᾶς συμπλοκῆς προσώπου

πρὸς ἐνέργειαν ἢ πάθος σημαντικὰ πάντα ὁ ᾿Αριστοτέλης εἰς ὀνόματα διαιρεῖ καὶ ῥήματα,

τὰ μὲν κατὰ χρόνον λεγόμενα ἢ κατηγορούμενα ἐν ταῖς προτάσεσι ῥήματα καλῶν, τὰ δὲ

ἄνευ χρόνου λεγόμενα ἢ τὴν χρείαν συμπληροῦντα τῶν ὑποκειμένων ὀνόματα· τὰ δέ γε

ἐν μηδετέρᾳ τούτων χώρᾳ παραλαμβανόμενα, κἂν ἄλλως προσκέωνται ταῖς προτάσεσι, τὸ

ὑπάρχειν ἢ μὴ ὑπάρχειν ἢ πότε ἢ πῶς ἢ ποσάκις ὑπάρχει τὸ κατηγορούμενον τῷ ὑποκειμένῳ

σημαίνοντα ἤ τινα ἄλλην αὐτῶν πρὸς ἄλληλα σχέσιν, οὐδὲ κυρίως ἀξιοῖ μέρη τοῦ λόγου

καλεῖν· ὥσπερ γὰρ τῆς νεώς αἱ μέν σανίδες εἰσὶ τὰ κυρίως μέρη, γόμφοι δὲ καὶ λίνον καὶ

πίττα συνδέσεως αὐτῶν καὶ τῆς τοῦ ὅλου ἑνώσεως ἕνεκα παραλαμβάνονται, τὸν αὐτὸν

τρόπον κἀν τῷ λόγῳ σύνδεσμοι καὶ ἄρθρα καὶ προθέσεις καὶ αὐτὰ τὰ ἐπιρρήματα γόμφων

τινῶν χρείαν ἀποπληροῦσι, μέρη δὲ οὐκ ἂν λέγοιντο δικαίως, ἅ γε μὴ δύνανται συντεθέντα

καθ᾿ ἑαυτὰ τέλειον ἐργάσασθαι λόγον. λόγου μὲν οὖν ταῦτα οὐ μέρη, λέξεως δὲ μέρη, ἧς καὶ ὁ

λόγος αὐτὸς μέρος, καθάπερ ἐν τοῖς Περὶ ποιητικῆς εἴρηται· καὶ εἰσὶ χρήσιμα πρὸς τὴν παρ᾿

ἄλληλα ποιὰν σύνθεσίν τε καὶ σύνταξιν τῶν τοῦ λόγου μερῶν, ὥσπερ καὶ ὁ δεσμὸς πρὸς τὴν

ἐπίκτητον ἕνωσιν τῶν δεδεμένων καὶ ἡ κόλλα τῶν δι᾿ αὐτῆς συνεχομένων, ἀλλ᾿ οὔτε ἐκεῖνα

μέρη τῶν δεδεμένων ἡ κεκολλημένων οὔτε σύνδεσμοι ἢ ἄρθρα ἢ προθέσεις ἢ ἐπιρρήματα τοῦ

λόγου μόρια.

¹⁰⁴ ἐπὶ πᾶσι δὲ τοῖς κατειλεγμένοις ὁ τούτων συνδετικὸς σύνδεσμος παρελαμβάνετο, οὐδὲν

δυνάμενος ἰδίᾳ παραστῆσαι χωρὶς τῆς τῶν λέξεων ὕλης, καθάπερ οἱ τῶν σωμάτων δεσμοὶ

οὐκ εἰσὶ χρειώδεις ἀνυποστάτων ὄντων τῶν σωμάτων.

¹⁰⁵ multo melius igitur qui principales et egregias partes nomen dicunt et verbum, alias autem his

appendices.

Connectorless Connections 235

Connectors—along with prepositions and adverbs—may be called append-

ages or adjuncts to verbs and names. But they should not be likened to glue or

pitch or cord—the comparisons are diverting but they are not illuminating.

CONNECTORLESS CONNECTIONS

Are the grammarians right? The comparison was intended to explain, or to

help to explain, how connectors are ‘non-signiﬁcant’. Apollonius says that

Posidonius, in his On Connectors argues against those who afﬁrm that connectors do

not show anything but simply connect the phrase.

(conj 214.4–6)¹⁰⁶

Posidonius’adversariesapparentlyheldthat connectors have no semantic force.

Such an idea is perhaps suggested by Dexippus’ comparison between

connectors and ‘the double lines which we write in the margin’: connectors

are, so to speak, punctuation marks; they indicate the articulation of a saying

but they do not contribute to what it says. The Dionysian Art has this to say

on punctuation marks or points:

There are three points: complete, intermediate, subpoint. A complete point is a sign

of a ﬁnished thought; an intermediate point is a sign introduced for the sake of taking

breath, and a subpoint is a sign of a thought which is not yet ﬁnished but still lacks

something.

(6 [7.4–7])¹⁰⁷

The remarks on complete points and subpoints—on full stops and com-

mas—might be developed, and the comparison with connectors elaborated.

Later texts used a large number of points and signs—Isidore enumerates

no fewer than twenty-six (etym i xx), one of them being the double line or

διπλῆ to which Dexippus refers. But the signs and points are not all of them

punctuation marks. The double line itself had several functions—in editions

of Plato, for example, it was used to mark ‘the doctrines and opinions of

Plato’ (Diogenes Laertius, iii 65).¹⁰⁸ But there is no evidence that it was ever

used as a sign of punctuation.

¹⁰⁶ Ποσειδώνιος ἐν τῷ Περὶ συνδέσμων ἀντιλέγων πρὸς τοὺς φάσκοντας ὡς οἱ σύνδεσμοι

οὐ δηλοῦσι μέν τι, αὐτὸ δὲ μόνον τὴν φράσιν συνδέουσι ...

¹⁰⁷ στιγμαί εἰσι τρεῖς· τελεία, μέση, ὑποστιγμή. καὶ ἡ μέν τελεία στιγμή ἐστι διανοίας

ἀπηρτισμένης σημεῖον, μέση δὲ σημεῖον πνεύματος ἕνεκεν παραλαμβανόμενον, ὑποστιγμὴ δὲ

διανοίας μηδέπω ἀπηρτισμένης ἀλλ᾿ ἔτι ἐνδεούσης σημεῖον.

¹⁰⁸ διπλῆ πρὸς τὰ δόγματα καὶ τὰ ἀρέσκοντα Πλάτωνι.

236 What is a Connector?

So Dexippus cannot in fact have meant to compare connectors with punc-

tuation marks. What did he have in mind? None of the known functions of

the double line seems peculiarly appropriate to a comparison with connectors;

and I suppose that Dexippus took the sign as one example from many. The

signs which we write in the margins do not constitute a part of the text, and

yet they have no sense when divorced from the text: they are, so to speak, com-

ments upon the text. I cannot think that such considerations help us greatly to

understand the semantics of connectors—but then Dexippus only mentions

the double lines en passant and no other ancient text adduces them in the same

context.

In the ﬁrst version of his artiﬁcial language, Frege introduced a special sign

which he called the content-stroke and printed as a longish horizontal line.

The symbol served to bind together the expressions which followed it into a

judgeable whole. If I write

2 + 2 = 4

then those ﬁve symbols do not stick together. But write

—2+ 2 = 4

and everything is dandy. The initial horizontal glues together the items which

follow it. Its function is purely connective, purely syntactical.

Frege’s view has often seemed absurd, and for several reasons. One of

the reasons is this: if the ﬁve symbols are likely to fall apart, how could

the addition of a further symbol help them? The addition will simply mean

that there are six items to fall apart rather than ﬁve. That is clever; but it is

unpersuasive. And here at least the ancient invocation of glue is illuminating.

There are two sorts of parquet ﬂooring, ﬂoating and ﬁxed. I ordered ﬁxed.

The parquetier told me that I needed glue to ﬁx it to the ﬂoor. ‘What’s the

use of that?’, I said. ‘If I buy the glue, then I’ll simply have three items to

ﬁx together rather than two.’ The parquetier was unmoved. Some ancient

theorists took their cue from one of his forebears: connectors are syntactical

glue, not semantic planks.

That idea might be supported by reﬂecting on the fact that the connecting

which connectors do may also be done without them. For example, you may

connect

The tenor was new to the part

and

The tenor sang like a strangled hyena

Connectorless Connections 237

by using a relative pronoun:

The tenor, who was new to the part, sang like a strangled hyena.

Or by way of a participial clause:

Being new to the part, the tenor sang &c

Or by a prepositional construction (here, but not always, rather cumber-

some):

On account of his being new to the part, the tenor sang &c

And there are other familiar devices—genitive absolutes in Greek, ablative

absolutes in Latin. In such cases, the role of the connector is played by

something else: the connecting is done by a purely syntactical device.

Aristotle’s Rhetoric has a certain amount to say about the proper use of

connectors. In one passage he refers to items which ‘are without a connector

but not unconnected’ (Rhet 1407b38–39).¹⁰⁹ His example is

Having travelled, I spoke.

The saying is connected, but the connection is done by a syntactical turn

rather than by the interpolation of a connector. Ammonius has something a

little more elaborate:

You might wonder what we shall say about a saying which says

The sun being above the earth, it is day.

For that saying is neither simple nor yet does it seem to have needed a connector

to unify it. In answer, we say that it is impossible for two complete assertions to be

amalgamated to produce a single saying unless a connector is actually present. Then

how could

The sun is above the earth,

which is complete, be mingled with

It is day,

which itself too is complete, without a conditional connector? But often we alter the

antecedent proposition together with the connector in such a way that, although it

is no longer complete with regard to assertion, nonetheless inasmuch as it contains

potentially a connector (or an adverb which is in this respect equivalent to a connector)

it fuses together with the consequent proposition which has remained complete. And

that is so in the present case. For

The sun being above the earth

is incomplete with regard to assertion, but it contains potentially either the adverb

‘when’ or the connector ‘if ’, and it expresses the same as

¹⁰⁹ ἐὰν δὲ συντόμως, ἄνευ μὲν συνδέσμου, μὴ ἀσύνδετα δέ.

238 What is a Connector?

When the sun is above the earth

If the sun is above the earth.

(in Int 68.12–26)¹¹⁰

Two complete sentences cannot become one by mere juxtaposition. But they

may fuse without the aid of a connector—so long as one or other of them is

transposed in a suitable way.

The point was not lost on the grammarians. Apollonius will sometimes

speak of a connective power or construction, which may be possessed by

items which are not connectors.

When the preposition ‘διά’ has a connective construction, it is used with the

accusative.

(synt iv 30 [461.1–2])¹¹¹

Or again:

A connective power accrues even to a prepositional juxtaposition:

Because of its being day, it is light

Because of Dionysius Apollonius was present

From idleness vices are born

—as though ‘On account of idleness’. So why is it strained to say that a connector

may also effect a prepositional force?

(adv 181.32–182.3)¹¹²

There are connections without connectors: a preposition, or a syntactical

turn, may do the same job as a connecting particle.

But something, according to Ammonius, must do the job: the mere

juxtaposition of two complete sayings cannot produce a compound saying. If

¹¹⁰ ἐπιζητήσειεν ἄν τις τί ἐροῦμεν περὶ τοῦ λόγου τοῦ λέγοντος ἡλίου ὑπὲρ γῆν ὄντος ἡμέρα

ἐστίν· οὔτε γὰρ ἁπλοῦς οὗτος ὁ λόγος οὔτε συνδέσμου πρὸς τὴν ἕνωσιν φαίνεται δεηθείς. πρὸς

ὃν ῥητέον ὅτι δύο μὲν αὐτοτελεῖς ἀποφάνσεις συμπλακῆναι ἀλλήλαις πρὸς ἑνὸς λόγου γένεσιν

συνδέσμου χωρὶς ἐνεργείᾳ ληφθέντος ἀδύνατον· τὸ γοῦν ἥλιος ὑπὲρ γῆν ἐστιν αὐτοτελές ὂν

τῷ ἡμέρα ἐστίν αὐτοτελεῖ καὶ αὐτῷ ὄντι πῶς ἂν συγκραθείη δίχα τοῦ συναπτικοῦ συνδέσμου;

πολλάκις δέ γε τὴν ἡγουμένην τῶν προτάσεων ἅμα τῷ συνδέσμῳ μεταρρυθμίζομεν οὕτως

ὥστε μηκέτι μὲν αὐτοτελῆ εἶναι πρὸς ἀπόφανσιν, τῷ δὲ δυνάμει περιέχειν τὸν σύνδεσμον ἢ

τὸ τῷ συνδέσμῳ ἰσοδυναμοῦν κατὰ τοῦτο ἐπίρρημα πρὸς τὴν ἑπομένην τῶν προτάσεων καὶ

αὐτοτελῆ μεμενηκυῖαν συμφύεσθαι, καθάπερ ἔχει καὶ ἐπὶ τοῦ προκειμένου· τὸ γὰρ ἡλίου ὑπὲρ

γῆν ὄντος ἀτελὲς πρὸς ἀπόφανσιν δυνάμει δὲ περιέχον ἢ τὸ ὅτε ἐπίρρημα ἢ τὸν εἰ συναπτικὸν

σύνδεσμον καὶ ταὐτὸν φθεγγόμενον τῷ ὅτε ἥλιος ὑπὲρ γῆν ἐστι καὶ τῷ εἰ ὁ ἥλιος ὑπὲρ γῆν

ἐστιν.

¹¹¹ προθέσεως τῆς διά κατὰ συνδεσμικὴν σύνταξιν φερομένης ἐπ᾿ αἰτιατικὴν ...

¹¹² καὶ προθετικῇ παραθέσει συνδεσμικὴ δύναμις ἐγγίνεται· διὰ τὸ ἡμέραν εἶναι φῶς ἐστι,

διὰ ∆ιονύσιον παρεγένετο ᾿Απολλώνιος, ἐκ τῆς ῥᾳθυμίας αἱ κακίαι γίνονται (ὡς εἰ ἕνεκα τῆς

ῥᾳθυμίας). τί οὖν βίαιον τὸ καὶ σύνδεσμον προθετικὴν δύναμιν ἀποτελέσαι;

Connectorless Connections 239

two complete sentences are juxtaposed, then there is an unconnectedness or

asyndeton (ἀσύνδετος´ being the Greek for ‘unconnected.’) Such items were

of concern to rhetoric—thus Aristotle, for example, declares that

an unconnected item has this peculiarity—many things seem to have been said at

the same time. For the connector makes the plurality one, so that if it is removed it

is plain that, conversely, the one will be many. That produces grandeur:

I came, I spoke, I implored.

(Rhet 1413b31–1414a1)¹¹³

I came, I spoke, I implored

there is not only an absence of connectors: there is no connection at all—mere

juxtaposition.

But does juxtaposition never connect? A curious passage in Apollonius’

Connectors to which I have already alluded is pertinent. Of disjuncts, he

claims,

some have received a true disjunction and others have taken a non-true disjunction.

The expression

Either it is day or it is night

is in a true disjunctive—for those circumstances will never come about at the same

time. The expression

Either Apollonius will be there or Trypho will be

announces a temporary disjunction. The former example will be in disjunction even

if it does not take a disjunctive connector

It is day, it is night.

(The one of them is true—if we speak so while it is day, then it is day.) The other is

not at all so:

Trypho will be here, Apollonius will be here.

For those items are not disjoined unless they take a disjunctive connector.

(conj 216.16–217.10)¹¹⁴

¹¹³ ἔτι ἔχει ἴδιόν τι τὰ ἀσύνδετα· ἐν ἴσῳ γὰρ χρόνῳ πολλὰ δοκεῖ εἰρῆσθαι· ὁ γὰρ σύνδεσμος

ἓν ποιεῖ τὰ πολλά, ὥστε ἐᾤν ἐξαιρεθῇ, δῆλον ὅτι τοὐναντίον ἔσται τὸ ἓν πολλά. ἔχει οὖν

αὔξησιν· ἦλθον, διελέχθην, ἱκέτευσα.

¹¹⁴ καὶ ἔστι πάλιν αὐτῶν ἃ μὲν <ἀληθῆ> τὴν διάζευξιν ἀναδεδεγμένα, ἃ δὲ οὐ<κἀληθῆ>

τὴν διάζευξιν παρειληφότα. τὸ γὰρ λεγόμε<νον ἢ> ἡμέρα ἐστὶν ἢ νύξ ἐστιν ἐν ἀληθεῖ

καθέστηκε διεζευγμένῳ· ταῦτα γὰρ τὰ καταστήματα ο<ὐδέποτε> κατὰ ταὐτὸ γενήσεται. τὸ

δὲ λεγόμενον ἢ ’Απολλώνιος παρέσται ἢ Τρύφων ὡς πρὸς καιρὸν τὴν διάζευξιν ἐπαγγέλλεται.

τὸ γοῦν πρότερον ὑπόδειγμα, κἂν μὴ <λάβῃ> τὸν διαζευκτικὸν σύνδεσμον, πάλιν ἐν διαζεύξει

<ἔσται>· ἡμέρα ἐστί, νύξ ἐστι. (τὸ ἕτερον ἀληθές· εἰ φαί< ημεν οὕτως> ἡμέρας οὔσης ἡμέρα

ἐστί.) τὸ δὲ ἕτερον οὐ πάντως· Τρύφων παρέσται, ᾿Απολλώνιος οὐ παρέσται· οὐ διαζεύγνυται

γὰρ τὰ τοιαῦτα ἐὰν μὴ λάβῃ τὸν διαζευκτικὸν σύνδεσμον.