Navigation – Plan du site
Dire et vouloir dire dans les arts du langage anciens et tardo-antiques

What Does “This” Mean? Deixis and the Semantics of Demonstratives in Stoic Propositions

Marion Durand

Résumés

Cet article vise à comprendre la théorie stoïcienne de la deixis afin d’expliquer l’importance accordée par les stoïciens aux pronoms démonstratifs et aux énoncés qu’ils composent, c’est-à-dire les propositions dites définitives. Nous montrons que ces propositions sont privilégiées pour des raisons à la fois ontologiques et épistémologiques en raison des propriétés sémantiques de leur sujet. Elles sont privilégiées d’un point de vue ontologique parce que la deixis grâce à laquelle leur sujet fait référence au réel crée une relation privilégiée à la matière, la catégorie ontologique stoïcienne fondamentale. De plus, ces sujets requièrent que leur référent existe et soit présent et saisissable. Ces énoncés sont d’autre part privilégiés d’un point de vue épistémologique parce qu’ils sont fondamentaux dans l’acquisition de la connaissance. Nous suggérons enfin que tous les énoncés apparemment définis ne le sont pas en effet, et considérons quand et sous quelles conditions une référence déictique est réussie.

Haut de page

Texte intégral

I am indebted to audiences at the New Perspectives on Stoicism Workshop at Cornell, where I also received a very thoughtful response from Simon Shogry, the 3rd Ancient Philosophy Workshop for Female Graduate Students and Early Career Researchers at Humboldt Universität Berlin, and the Colloque Sēmainō in Lille, for their feedback on this paper. For comments on earlier versions of the work, I am grateful to Susanne Bobzien, Charles Brittain, and Stephen Menn, and of course to Brad Inwood, James Allen, Rachel Barney, and Gurpreet Rattan, for guiding me every step of the way as I formulated my thoughts on this topic. Many thanks, lastly, to the anonymous reviewers for their valuable suggestions for improvements.

  • 1 See esp. Charles H. Kahn (1969), “Stoic Logic and Stoic Logos”, Archiv fur Geschichte der Philosoph (...)
  • 2 The most noteworthy exception is Frédérique Ildefonse (1997), La naissance de la grammaire dans l’A (...)

1The demonstrative houtos (“this”) is given pride of place in Stoic semantic theory. Yet, although the related concept of deixis has often been highlighted as an interesting issue in Stoic semantics1, little attention has been paid in scholarship to demonstratives and their semantics2. Demonstratives are often discussed in passing. These discussions tend to be based on uncritical assumptions or take a certain conception of deictic reference for granted. In what follows, I question these assumptions and take a fresh look at the evidence, with a view to offering a new and comprehensive interpretation of the Stoic theory of demonstrative reference. I argue for a consistent and in-depth Stoic theory according to which deixis is a particular way to refer which confers a privileged status on the propositionsusing it, on both ontological and epistemological grounds. Demonstratives that refer by deixis have a privileged relationship to what is most ontologically fundamental and form the most epistemologically fundamental propositions.

  • 3 I translate axiōma by “proposition”, which I take to be uncontroversial. While “proposition” is a (...)
  • 4 This account is echoed by Diogenes Laertius, Lives of Eminent Philosophers (DL) VII.70. The differe (...)

2The Stoics distinguished between the following types of propositions3: definite (hōrismena), indefinite (aorista) and middle or intermediate (mesa), as clearly set out by Sextus Empiricus4:

  • 5 Sextus Empiricus, Against the Mathematicians (SE M) VIII.96: τῶν δὲ ἁπλῶν τινὰ μὲν ὡρισμένα ἐστὶν τ (...)

“Of the simple propositions, some are definite, some indefinite, and some middle. Definite are the ones expressed deictically, for example ‘this one is walking’, ‘this one is sitting’ (for I am indicating a particular man). Indefinite propositions, according to them, are ones in which something indefinite governs, for example ‘someone is sitting’, middle propositions are ones such as ‘man is sitting’ or ‘Socrates is walking’”5.

3What sets the different types of propositions apart is the grammatical subject of the sentence that expresses them: the subject of what expresses a definite proposition is a demonstrative pronoun (paradigmatically houtos, “this one”), that of an indefinite proposition is an indefinite pronoun (tis, “someone”) and that of a middle proposition is a proper or common noun (“Socrates” or “man”). More precisely, what distinguishes them is the way in which these subjects pick out things in the world:

  • 6 Sextus Empiricus, M VIII.97: τὸ μὲν οὖν “τὶς περιπατεῖ” ἀόριστόν ἐστιν, ἐπεὶ οὐκ ἀφώρικέ τινα τῶν ἐ (...)

“‘Someone is walking’ is indefinite because it has not delineated a particular walking individual; for it can be expressed with reference to each of them in common. On the other hand, ‘this one is sitting’ is definite, since it designates the person indicated. And ‘Socrates is sitting’ is middle, according to them, since it is neither indefinite (for it has delineated a species), nor definite (for it is not expressed with deixis), but it seems to fall between the indefinite and the definite.”6

  • 7 On the Stoic notion of eidos, see especially Jacques Brunschwig (1988), « La théorie stoïcienne du  (...)

4The subjects of definite propositions are expressed with deixis, or demonstratively. Indefinite pronouns do not designate any particular or individual, making their proposition indefinite. Nouns, the subjects of middle propositions, fall somewhere in between: they do not refer deictically, but do designate a “species”, eidos, by which the Stoics mean a concept which covers both species as we would normally understand it and species with just one member, that is, individuals7.

5The different propositions have different relationships with one another. We are told, for example, that the truth-value of indefinite propositions is in some sense dependent on that of a corresponding definite proposition:

  • 8 Sextus Empiricus, M VIII.98: γίνεσθαι δέ φασι τὸ ἀόριστον ἀληθές, τὸ “τὶς περιπατεῖ” ἢ “τὶς κάθηται (...)

“They say that the indefinite, namely ‘someone is walking’ or ‘someone is sitting’ is true whenever the definite, ‘this one is sitting’ or ‘this one is walking’ is found to be true. For if no individual is sitting, the indefinite ‘someone is sitting’ cannot be true”8.

  • 9 For more on this passage, see Paolo Crivelli (1994), “Indefinite Propositions and anaphora in Stoic (...)
  • 10 Some (e.g. Andreas Graeser (1978), “The Stoic theory of Meaning”, in John Rist (ed.) The Stoics, Be (...)
  • 11 This would have avoided an issue: “someone is dead” can apparently never be true, since, as we will (...)
  • 12 Diogenes Laertius VII.45.

6This passage has usually been taken to suggest that “someone is walking” is true only if an instance of “this one is walking” is true.9 It is particularly interesting that indefinite propositions rely on a corresponding definite proposition but not a corresponding middle proposition10. It would have been easy for the Stoics to claim that an indefinite is true if a corresponding definite or middle proposition is true. “Someone walks” would then be true if either “this one walks” is true or “[insert proper or common noun] walks” is true11. Something about the way in which demonstratives refer – the distinctive feature of definite propositions – confers a privileged status on definite propositions when it comes to assessing truth-value. This is especially important because one of the primary goals of logic, or dialectic, according to the Stoics, is to distinguish the true from the false, so that the sage can identify and use infallible arguments, and assent only to true impressions12.

I. Demonstratives and Ontology

  • 13 The details of the correspondence are still subject to debate. Anthony Lloyd (1971), Problems in St (...)
  • 14 “Matter” and “qualities” are thus technical notions for the Stoics, and distinct from Aristotle’s u (...)

7Part of the privileged status of demonstratives stems from the well-known correspondence between language and ontological categories in Stoic theory13. The Stoics distinguished four ontological categories: matter (ousia, which is a hupokeimenon), quality (poiotēs, which is a poion, and subdivided into common and proper), disposition (pōs echon) and relation (pros ti pōs echon)14. The first two of these (matter and quality) correspond to demonstrative pronouns and nouns respectively. Diogenes Laertius tells us:

  • 15 Diogenes Laertius VII.58: Ἔστι δὲ προσηγορία μὲν κατὰ τὸν Διογένην μέρος λόγου σημαῖνον κοινὴν ποιό (...)

“According to Diogenes [of Babylon], a common noun is a part of speech which signifies a common quality, for example ‘man’, ‘horse’. A proper noun is a part of speech which designates a proper quality, for example ‘Diogenes’, ‘Socrates’.”15

  • 16 The grammarian Apollonius Dyscolus was heavily influenced by the Stoics, as shown by, e.g., David L (...)

8This matches what we find in Apollonius Dyscolus, who adds the corresponding information regarding demonstrative pronouns16:

  • 17 Ap. Dysc. S II.22.3-6, 24.1-4= GG 2.2.142.1-4, 143.9-144.4: Ἡ τῶν ὀνομάτων θέσις ἐπενοήθη εἰς ποιότ (...)

“The imposition of nouns was thought of with an eye to common and proper qualities, such as “man”, “Plato”, and since these are not accompanied by deixis or anaphora, a great number of them was established so the characteristics of each can render the quality of each referent. […] But pronouns, although they focus on nothing other than the underlying matter (ousia) through the deixis inherent in them, also capture easily grasped accidents which accompany quality (poiotēs), namely those apprehensible by sight such as white or black, long or short.”17

  • 18 See also I.120, and Priscian, Institutiones Grammaticae (IG), 15.63-65 ; 17.105. Cf. Jean Lallot (1 (...)
  • 19 I take ousia and poiotēs to be used in the technical Stoic sense here, and accordingly translate t (...)
  • 20 A note on the translation of those two key terms: ἀποβλέπειν appears only one other time in Apollon (...)
  • 21 Similarly, both nouns and pronouns capture their referent’s matter, but only pronouns focus on it. (...)

9This passage is difficult18. The central idea is clear enough: nouns are, in some way, linked to quality while pronouns are first and foremost linked to matter (they “focus on” matter), but nonetheless also include, capture, or render qualities19. The exact nature of the relationships between words and categories, in particular what Apollonius means by “focusing” (apoblepein) and “capturing” (echein) is unclear20. The crucial distinction appears to be precisely between those two: what each part of speech focuses on and what it captures. Both nouns and pronouns capture qualities of the referent, but what they focus on is different: nouns both capture and focus on qualities while pronouns capture but do not focus on qualities21.

  • 22 This is in line with the argument in David N. Sedley (1982), “The Criterion of Stoic Identity”, Phr (...)
  • 23 I do not agree with Jean Lallot (1997) Apollonios Dyscole, ad loc., who interprets the passage as s (...)
  • 24 I take the somewhat puzzling phrase τὰ ἐπισυμβεβηκότα τῇ ποιότητι (“the accidents which accompany q (...)
  • 25 I take it that I could, by contrast, see it without grasping that it is an elm. The difference, as (...)
  • 26 This is why I take the participle ἀποβλέπουσαι to be concessive (contra Lallot and Householder who (...)
  • 27 Casper De Jonge (2008), Between Grammar and Rhetoric. Dionysius of Halicarnassus on Language, Lingu (...)

10What Apollonius has in mind must be something like this: nouns designate the object to which they refer by its quality. They secure reference to it insofar as it is a man (a common quality), or insofar as it is Plato (a proper quality). Deictics, on the other hand, refer by deixis, which focuses on the matter (ousia) of the object they are referring to.22 They do pick out qualities of their referent, but, unlike nouns, do not focus on them. Their picking out qualities is, in this sense, incidental to their picking out the matter of the referent.23 Since demonstratives pick out an object’s underlying matter, they also incidentally pick out some of the qualities which organise the object’s matter. Note that Apollonius tells us that the properties picked out in this way are easily grasped (eulēpta), or perceptible qualities24. I cannot see a tree’s matter without also seeing its height, or the colour of its bark and leaves25. However, these qualities do not play a part in securing the reference: it is precisely because the deictic reference has been successful that I am able to grasp the visible qualities that organise the matter26. Demonstratives secure reference to objects by appeal to matter alone. By echein, then, Apollonius must mean something along the lines of generally picking out in some (possibly indirect or incidental) way. Apoblepein points to a more fundamental relationship between words and categories: that by means of which reference is secured. Nouns are in this privileged relationship to qualities while pronouns are in this privileged relationship to matter. Matter, for the Stoics, is the most ontologically fundamental category. One reason why demonstratives are privileged, then, is because they refer by appealing to matter, which is ontologically fundamental27.

II. Demonstratives and Deixis

  • 28 cf., e.g., Simplicius in Ar. Cat. 222.30-33. I am grateful to Simon Shogry for pointing this out to (...)

11This relationship of demonstratives to matter raises an issue. According to Stoic ontology, all matter is the same. Two objects differ only by virtue of the qualities arranging their matter. Talking about a particular object’s matter as opposed to another bit of matter is itself somewhat misleading. Insofar as they are purely unqualified matter, Socrates, an aircraft, and a banana are the same and entirely indistinguishable28. This is why there are, as Apollonius points out, very many nouns (to designate the various and numerous qualities), while the same demonstrative can successfully refer to a large number of objects: their matter is one and the same. However, if this is so, how can demonstratives successfully identify anything in particular? In order to successfully refer to something, it seems that I would have to refer to its matter qualified to some degree - at least in terms of place or size. If I refer to it entirely unqualifiedly, I cannot possibly successfully distinguish it from anything else. How, then, can one successfully designate an object by means of a demonstrative pronoun if it focuses solely on its matter?

  • 29 It is worth noting that the exact way in which deixis features alongside deictic pronouns is unclea (...)
  • 30 As I argue in section V below, this ostension need not be physical, nor is any physical ostension s (...)

12The answer, I suggest, lies in the deixis itself. The Stoics described demonstrative phrases as expressed deictically (μετὰ δείξεως ἐκφέρεται) and made deixis their defining feature29. That is, they recognised that demonstratives refer by deixis (διὰ τῆς ἐν αὐταῖς ἐγκειμένης δείξεως, as Apollonius puts it). The deixis is what fixes the reference. Now, deixis is a demonstration or ostension of some kind performed by the speaker as she utters the expression which the deixis accompanies30. Deixis is necessarily indexical: the deixis cannot be performed by anyone but the speaker. This indexical ostension is what allows the relevant particular lump of matter to be delineated and identified. It indicates its location relative to the speaker and enables the interlocutor to grasp what it is that is being talked about. The ostension can indicate other aspects of the object as relevant and necessary: for example, a speaker referring to a flower bed on the ground to her right can make a low sweeping gesture with her right arm to indicate that what she is showing is not only located low and to her right but also stretched out over a certain distance or of a certain size. In this sense, the deixis will pick out some qualities. That is, it will make them graspable. However, it does not require any previous knowledge from the interlocutor about the object thus designated or its qualities. The interlocutor will be able to locate the flowerbed to the right of the speaker even without knowing that it is to her right, or what “right” and “left” mean.

13Indeed, the relevant difference between nouns and demonstrative pronouns and their relationship to ontological categories is not so much what they pick out, but what they focus on, that is, what plays a reference-fixing role. Demonstratives have the capacity to refer to an object through deixis alone and render its qualities graspable to interlocutors. This is the meaning of Apollonius’ condensed sentence in the passage above: “But pronouns, although they focus on nothing other than the underlying matter, through the deixis inherent in them, capture easily grasped secondary accidents which affect quality”. Demonstratives focus on nothing other than matter, that is, they appeal to nothing else in the object to fix the reference. The rest of the work is done by the deixis, which locates the object relative to the speaker. This act of reference in turn captures the object’s qualities so as to make them graspable.

  • 31 Chrysippus argues that egō is deictic: we indicate ourselves with our chin as we say it (Galen, de (...)

14This interpretation of the passage is strengthened by a passage of Apollonius’ work On Pronouns in which he uses similar language to describe the first- and second-person pronouns, which he, like Chrysippus, considered to be deictic31:

  • 32 Ap. Dysc. P 11.27-30: αἱ μέντοι πρωτότυποι διὰ μὲν τῆς φωνῆς γένους ἀδιάστολοί εἰσι πάντοτε κατὰ πρ (...)

“However, the archetypes in the first and second person do not in any case distinguish gender by means of their form, rather the distinction of the genders is made manifest by the deixis in them. For the deixis captures the gender which is made clear at the same time.”32

15Greek first- and second-person pronouns (egō and su) do not have gendered endings: the same form is used for both the feminine and the masculine. The forms, then, do not distinguish the gender of the referent. Rather, the deixis in them makes the gender of the referent manifest. Gender, one of the referent’s qualities, is here being made perceptible by the deixis. There can be no doubt that gender does not play a part in fixing the reference, for the reference must have been fixed in order for the gender to be perceptible. The explanation given for this is that the deixis captures and makes clear the gender (συνεξηγούμενον γὰρ ἔχει τὸ γένος ἡ δεῖξις). The use of ἔχει here is parallel to the one in the previous passage. Just as the deixis of personal pronouns in the first and second person makes a quality, i.e. gender, manifest – that is, perceptible and graspable – in the same way, the deixis of demonstrative pronouns, while focusing on matter, captures easily grasped and visible qualities of the referent and renders them graspable. In that sense, these qualities are referred to incidentally. They do not, however, play any part in fixing the reference.

  • 33 This is to some extent suggested by the interpretation of Frédérique Ildefonse (1997), La naissance (...)
  • 34 This is only a brief and compressed sketch of the apparatus. For more on the semantics of nouns see (...)
  • 35 DL VII.60. cf. n. .
  • 36 I use angled brackets to signal reference to concepts throughout.

16In this respect, the privileged status of demonstrative pronouns is epistemological33. While demonstrative pronouns refer by means of deixis alone, nouns do so by signifying a quality. Nouns in fact rely on concepts to refer34. They delineate an eidos, which, according to the Stoics is a concept (ennoēma)35. To this concept corresponds a physical quality (poiotēs) of the noun’s referent. When I utter the noun “Socrates”, I delineate the concept <Socrates>36 which corresponds to the quality of being Socrates. The noun calls to my interlocutor’s mind the same concept, and therefore allows her to pick out the corresponding quality in the world, that is, to understand that I am saying something about Socrates.

17The concept therefore plays an essential role in fixing the reference of nouns. In particular, it requires that this concept be previously known to the interlocutor. When I talk about Socrates, or I say “man”, I rely on my interlocutor’s knowledge of who Socrates is, or of what a man is. If I were speaking to someone who I thought might have never heard of Socrates, I would refer to him differently. This is not the case for demonstrative pronouns. Apollonius Dyscolus tells us that a demonstrative, by contrast, refers to something introduced for the first time to the conversation:

  • 37 Ap. Dysc. S I.96 = GG 2.2.80.13-81.3: προσώπων ἀοριστουμένων διακριτικά ἐστι τὰ μόρια, διὸ καὶ τὰ ἐ (...)

“These parts of speech distinguish previously indeterminate persons; this is why the persons intended by them are grasped with determinacy. And it is clear that the deixeis originating in them are the first introductions to the persons under discussion and do not require a construction with an article. For these persons are not referred back to but shown to the eye”37.

  • 38 The relevant knowledge here is that of concepts corresponding to qualities. For successful communic (...)

18In the case of demonstrative pronouns, no knowledge either of the referent or their qualities, nor of the corresponding concepts is presumed or necessary for successful communication38. Thanks to the deixis which accompanies them, demonstrative pronouns do not need to appeal to a concept in order to refer.

  • 39 Much like Apollonius Dyscolus, whom he follows extremely closely, Priscian shows signs of Stoic inf (...)

19In addition, because demonstratives do not rely on a concept to secure reference, they are less prone to misunderstandings. Nouns leave more room for ambiguity or misinterpretation, as noted by Priscian39:

  • 40 Priscian, IG XVII.63.7-15 = GL 3.145.22-146.4: Quare non mediocriter disturbant qualitatis signific (...)

“For which reason, there is a not insignificant confusion in the expression of quality, when several impositions of names coincide in one form, in the case of proper nouns as much as in the case of common nouns. Therefore the person whom we understand in the name itself lacks determinacy. And this is not surprising, since even proper nouns, although they were established in order to distinguish each individual from all others, are nonetheless indeterminate since they cannot indicate all the qualities which separate this individual from all others without the help of deixis, which is produced through the pronoun.”40

  • 41 It is worth noting that the surviving evidence does not include much explicit consideration of prag (...)

20Nouns, including proper nouns, are not as reliable a way of identifying a referent as demonstratives. If I say “Diogenes wrote it”, I assume that my interlocutor knows which Diogenes I mean. But I could well mean Diogenes of Oenoanda and they could think that I mean Diogenes Laertius. Common sense and norms of conversation dictate that we speak accurately (for example, by using an author’s full name the first time we mention him) to avoid such confusion, but slips and misunderstandings happen41.

21Third-person demonstrative pronouns are not entirely free from such concerns. Indeed, Alexander of Aphrodisias suggests that using the wrong gender of the demonstrative pronoun houtos, “this”, can cause the proposition to be false:

  • 42 Alex. Aphr. in Ar. An. Pr. 402.20-4: ἔτι τοῦ μὴ δεῖν οὕτως τὴν ἀπόφασιν ποιεῖσθαι πίστιν καὶ τοιαύτ (...)

“In addition, they introduce this kind of evidence that one should not form a negation in this way: ‘this [masculine] one is walking’, ‘this [masculine] one is not walking’, when indicating a woman. For they say that again both of these are false when taken in this way, whether or not what is indicated is walking.”42

  • 43 This goes hand in hand with the rejection of the idea that a mistake in gender is a solecism. cf. A (...)

22Third-person pronouns, unlike first- and second-person pronouns, are gendered. The gender of a demonstrative is considered to be part of the description of the referent and must be taken into account when assessing the truth-value of the proposition. If the wrong gender is used, the proposition is false. However, just as we have seen in the case of first- and second-person pronouns, the gender is independent of the deixis and plays no part in fixing reference. The deixis is what makes the gender manifest. I must know what has been referred to in order to judge that the wrong gender has been used. In fact, whatever qualities are picked out by deixis are incidental and do not play a part in fixing reference. The reference will be successful regardless of the accompanying description43.

23There is one passage in which Apollonius suggests that gender can help where the deixis is insufficient to fix the reference because, for some reason (usually distance), it is not precise enough:

  • 44 Ap. Dysc. S II.26.14-16 = GG 2.2.145.14-146.1: ἡ γὰρ ἐξ αὐτῶν νοουμένη ἀπόστασις ἀμαυροτέραν τὴν δε (...)

“For the distance which they express obscures the deixis and for this reason the gender is introduced, not to indicate the matter but to make clear the gender.”44

  • 45 It is not entirely clear why Apollonius thinks the gendered endings of houtos are not akin to that (...)
  • 46 Another reason not to give too much weight to this passage is emphasised by Jean Lallot (1997) Apol (...)

24This is the only passage which may be taken to suggest that something external to the ostension plays a reference-fixing role, as gender could be introduced in order to help fix the reference in cases where the deixis itself cannot be sufficiently precise due to distance. It is not clear from the passage that gender is in fact thought to play a part in reference-fixing in such cases. If it does, it may seem to pose a problem for the argument laid out so far. However, this passage is a discussion of ekeinos specifically, presenting it as an exception to the rules previously laid out for pronouns, and in particular as a contrast to houtos45. If anything, it seems rather to further emphasise the fact that gender does not play a part in securing the reference in the case of houtos46. Nonetheless if we were to take this passage as suggesting that, at times, the gender of a pronoun – including houtos – could work in conjunction with the deixis to fix reference, I do not think that it would be devastating: gender is a very basic concept, one which all competent speakers of Greek would have acquired. As such, appeal to it would hardly be comparable to appeal to concepts such as <horse> or even <white>, and demonstratives would remain simpler, more basic expressions.

III. Definite Propositions as Epistemologically Fundamental

25The epistemic simplicity of pronouns – the fact that they do not appeal to concepts to secure reference – in turn makes sense of the Stoic theory of concept-formation as reported by Cicero:

  • 47 Cicero, Academica II.21:Atqui qualia sunt haec quae sensibus percipi dicimus talia secuntur ea quae (...)

And such are the things which we claim are apprehended by the senses. The next set are just like them, though we don’t claim that these are apprehended by the senses themselves, but by the senses in a certain respect – e.g., ‘this is white’, ‘this is sweet’, ‘this is melodious’, ‘this right here is fine-scented’, ‘this right here is rough’. For we now grasp this set with the mind, not with the senses. Next comes: ‘this is a horse’, ‘this is a dog’. Then comes the rest of the series, which connects more significant things and encapsulates what we might call a filled-out apprehension of things – e.g., ‘if something is a man, it is a mortal animal partaking in reason’.”47

  • 48 This epistemological grounding is consistent with the alethic independence discussed in n.10 above.
  • 49 By grasping a proposition of the form “this is a horse”, I do not mean that I need to have uttered (...)
  • 50 It is not clear how the Stoics thought one could acquire concepts for things they never directly pe (...)

26The first two stages of concept-formation are presented in the form of demonstrative expressions of the form “this is white”, then, moving to a more specific concept, “this is a horse”. In other words, grasp of definite propositions comes first, and grounds our concepts before we can grasp middle propositions and complex propositions.48 Definite propositions are therefore fundamental in an epistemological sense. In fact, middle propositions rely epistemologically on definite propositions, in that the grasp of a middle proposition relies on the previous grasp of a definite proposition. I must have grasped “this is Socrates” before I can grasp “Socrates is walking”49. Competent use of the nouns “horse” and “Socrates” to express a proposition, and successful reference with them requires a prior acquisition of the concepts <horse> and <Socrates>, respectively. This would be first grasped in the most basic way through a thought of the form “this is a horse” and “this is Socrates”50.

  • 51 We may raise an objection, debating the relevant sense of “grasp” or “understand” at play here, and (...)

27This is a result of the semantics of definite propositions, which are such as to make a qualities corresponding to previously unknown concept graspable. We have seen Apollonius claiming that deixis was such as to make qualities of the referent manifest, and in particular easily grasped qualities. When deictic reference of a tree is secured, its height, the colour of its bark, and the shape of its leaves are visible to the interlocutors. I cannot perceive the tree without perceiving these properties of the tree. Perception of the qualities is, it seems, the first step to forming a concept for them. Note that I need none of the concepts associated with the qualities to perceive them. I do not need the concept <green> to perceive a tree’s greenness. Nor do I need the concept <green> to understand, when someone points to a tree and (successfully referring) says “this is green”, that this thing is green51. It is by means of a process of this sort that I will come to acquire the concept <green>, precisely because the deictic demonstrative has the capacity to refer despite my lacking the concept <green> and to make the quality green graspable.

28When deictic reference is secured to an object which has a number of qualities, some of which I have a concept for and some of which I do not, I suggest the following occurs: I perceive that the object has whatever qualities I have concepts for and I perceive its other qualities without perceiving that it has them. That is, if I have the concepts <tree>, <bark>, <leaves>, <green>, <brown> but not <elm> and someone points to an elm and, successfully referring, says “this is mine”, I will understand that this brown-barked green-leaved tree is the speaker’s. Someone else with the concept <elm>, who would recognise the tree as an elm, would understand that this elm is the speaker’s. I will nonetheless perceive its elm-ness, without being able to articulate it as such, until I have formed the concept <elm> by means of a proposition of the form “this is an elm”. I perceive the elm-ness in the same way that I may perceive its height without knowing that it is exactly 18.2m tall. Deixis, then, has the ability to make sensible qualities of the referent perceptible to the interlocutor without requiring any previously acquired concepts. It thereby enables the acquisition of new concepts. Furthermore, it has the ability to make perceptible as such any quality for which a concept has previously been acquired. It is for these reasons that deictic propositions are privileged as a powerful and fundamental epistemological tool.

  • 52 I mean by this that they do not make an existential claim, as middle propositions do, according to (...)

29The epistemological fundamentality of definite propositions is also explained by their simplicity. Not only are they basic in that they do not appeal to concepts to secure reference, they are also simpler in that, unlike propositions with nouns as their subjects, they do not have existential import52. This is highlighted by Chrysippus’ answer to the Master Argument. As we noted above, he argues that “this one has died” when pointing at Dion is impossible:

  • 53 Alex. Aphr. in Ar. An. Pr. 177.27-178.1: φησὶ γὰρ ἐν τῷ συνημμένῳ τῷ ‘εἰ τέθνηκε Δίων, τέθνηκεν οὗτ (...)

“For he says that in the true conditional ‘if Dion has died, this one (indicating Dion) has died’, the antecedent ‘Dion has died’ is possible since it can at some time come to be true that Dion has died, but ‘this one has died’ is impossible; for, after Dion has died, the statement ‘this one has died’ is destroyed since the recipient of the deixis no longer exists; for the deixis was to a living man and insofar as he was living. So if, since Dion has died, ‘this one’ is no longer possible, and if Dion does not exist again so that it is possible to say of him ‘this one has died’, then ‘this one has died’ is impossible.”53

  • 54 Nicholas Denyer (1988), “Stoicism and Token Reflexivity”, in Jonathan Barnes & Mario Mignucci (eds. (...)
  • 55 SE PH II.234; Epictetus Diss.1.7.1.10-21; Simplicius in Ar. Phys.1299.36-1300.10.

30“This one has died” is false before Dion dies and destroyed after Dion has died. Chrysippus relies on two important assumptions here. Firstly, that the proposition under consideration (“if Dion has died, this one has died”) is uttered before Dion’s death54. The truth-value of Stoic propositions can famously change over time55. Here, we are asked to consider the truth-value before and after Dion’s death of a complex proposition (“if Dion has died, this one has died”), composed of two simple propositions (“Dion has died” and “this one has died”). The proposition must have been uttered before his death because the sentence “this one has died” (pointing at Dion) uttered after Dion had already died would fail to express a proposition in the first place for just the same reason that it is destroyed after his death. “This one has died” must have been a proposition in order to later be destroyed, and for its truth-value to ever have been evaluable. If it had not been uttered before Dion’s death, it would never have had a truth-value.

  • 56 This is similar to the reading of Susanne Bobzien (1999), “Logic: The Stoics”, CHHP, p. 100 and 117 (...)

31Secondly, Chrysippus assumes that what the deixis refers to is Dion insofar as he is alive. Immediately after his death, his corpse might be present to be pointed at and referred to, but this corpse would not be Dion so that you could not truly say of it “this one has died”, where the deictic (and not anaphoric) “this one” would refer to the same thing as “this one” did when it was uttered by someone pointing at the living Dion. Once Dion is dead, he can no longer be pointed out. The demonstrative then fails to refer and the sentence fails to make an assertion, it no longer expresses a proposition, so that there is no proposition to bear a truth-value56.

  • 57 This would mirror the symbolization for middle propositions suggested by Anthony Lloyd (1978), Les (...)
  • 58 I assume here that the Stoics would analyse x (Gx x=δ) as well-formed even when there is no δ, w (...)
  • 59 cf. Jonathan Barnes (1999), “Aristotle and Stoic Logic”, in Katerina Ieradiokanou (ed.), Topics in (...)
  • 60 DL VII.63
  • 61 This highlights the need for a better account of completeness and incompleteness of lekta, since it (...)

32This suggests that “this one is G’ing” (for any property or activity G) should not be symbolised as either ∃xGx (which would surely make “someone is G’ing” redundant), or as ∃x (Gx ∧ x=δ)57 where δ is the thing pointed out by ostension. If we understand “this one is G’ing” as ∃x (Gx ∧ x=δ), then Chrysippus’ claim that “this one is dead” is not false but destroyed after his death cannot be correct: it would be false (because there is no x such that x = δ)58. Rather, we must understand “this one is G’ing” as the very simple expression Gδ59. This allows us to make sense of the claim about destruction: when there is no object of deixis, the expression is incomplete. Stoic propositions are said to be distinct from other sayables (lekta) in that they are complete in themselves (autoteles), and the bearers of truth-value60. When reference fails (when there is no δ), the expression is no longer complete in itself.It no longer satisfies the requirement for being a proposition, thereby being destroyed (φθείρεσθαι)61. In other words, when reference fails because there is nothing to refer to, the proposition becomes idle and loses its status as a proposition. It can in fact never be the case that a demonstrative falsely attributes existence to its referent: it can only refer successfully if its referent exists.

  • 62 Charles Kahn (1969), AGPh 51, p. 160. Cf. SVF II. 509, 517-519. See also Victor Goldschmidt (1969),(...)

33This is important for epistemological purposes. As Kahn points out: “the existence of the subject and the truth of the assertion are thus guaranteed by the familiar Stoic criterion of evident perception or ‘irresistible impression’ (phantasia katalêptikê). Stoic ontology concurs with their epistemology here in preferring the definite proposition with deixis since for the Stoics only the present is fully real”62. Insofar as a definite proposition is a proposition (a self-complete bearer of truth-value), that is, insofar as reference has been successful, the referent of its subject must exist, and be present. The referent can therefore be grasped and, as such, can be the source of a kataleptic impression. It is also because of this guarantee of immediate graspability and full reality of the referent of their subject that definite propositions provide a solid basis for forming a concept.

34Demonstratives, then, refer by deixis and deixis alone. They pick out first and foremost matter, the most ontologically fundamental category, and do not appeal to quality or concepts to secure reference. They are simple expressions, ensure the immediate graspability of their referent, and form the most epistemologically fundamental propositions. It is for these reasons that they have a privileged status.

IV. Failures of Deixis

35As we have seen, what confers this privileged status on demonstratives is the way in which they refer, namely by deixis. Chrysippus’ answer to the Master Argument tells us something important: deixis can fail. We have seen that demonstratives can successfully refer only if their referent is immediately graspable. We might infer from this that insofar as I use a demonstrative, I will successfully refer deictically. That is clearly not the case. So let us consider what makes deixis successful.

36Deixis firstly requires the presence of both an indicator and an indicated as Apollonius highlights:

  • 63 Ap. Dysc. S II.42.1-9 = GG 2.2.155.5-8: Ἕνεκεν τούτου πρὸς οὐδὲν χρειώδεις εἰσὶν αἱ ἀντωνυμίαι στερ (...)

“Because of this, pronouns are not at all useful when they are removed from the person indicating and the indicated, indeed when they are written, they are entirely indeterminate, because they have been thrust away from the particular matter.”63

  • 64 Ap. Dysc. P. 2.1.1.37.8-12: αἱ γοῦν διὰ τοῦ ι ἐπεκτεινόμεναι καὶ ἔτι τὴν δεῖξιν ἐπιτείνουσαι καὶ τὸ (...)

37If a demonstrative is disconnected from either the person indicating or the thing indicated, as may be the case in written text, it will be idle. In addition, Apollonius seems to suggest elsewhere that there are degrees of deixis, with the deixis being strengthened, for example, by the addition of an ι suffix64.

  • 65 I follow Philipp Brandenburg (2005), Apollonios Dyskolos. Über das Pronomen, Leipzig, Teubner, in t (...)

“So those intensified by the <ι> both further intensify the deixis and sharpen the aim, for example ekeinosi (that very one), houtosi (this very one). It is clear that this kind of intensification is for the sake of the deictic expression, since autos is not intensified, because it is not at all deictic.”65

38Deixis, then, can be intensified, with the intensification being reflected at the level of words by the addition of the ι suffix. It is worth noting that this can happen to any deictic expression, including houtos, the paradigmatic case of a deictic demonstrative pronoun. This should not surprise us: there are certainly contexts in which “this” may fail to be sufficiently precise. I might say “look at this” pointing to a small bug crawling across the floor. Say the bug is not only small but also happens to be a similar colour to the carpet. My interlocutor may not see it immediately and ask what it is that she is supposed to look at. I would likely respond “this right there” (houtosi, in Greek) and point again, perhaps getting a little closer. Even houtos, then, may turn out to fail to uniquely identify a referent, and need its deixis intensified to succeed. This intensification in deixis seems to bring an increase in its precision.

  • 66 SE M VIII.99-102
  • 67 While it may seem plausible that pragmatics were (or should have been) appealed to here – and elsew (...)

39This highlights a problem with regards to the indeterminacy of demonstrative reference. This is something which the Stoics must have been aware of, and something which Sextus Empiricus seems to have wielded against them arguing that “this” does not indicate Socrates as a whole, nor his soul nor his body, nor a part of him.66 One would expect the Stoics to have had an account of how demonstratives can always have the accuracy necessary to successfully refer. Intensification of deixis is presumably one answer, but it will not always be successful. For example, it is not clear how helpful it would be if, instead of the bug on the floor, I wanted to point to the bug’s exoskeleton. The Stoics likely recognised that we can and do, in fact, refer successfully with demonstratives to a wide range of things, and used this pragmatic observation to brush such concerns aside. We may nonetheless like them to have had some account for the indeterminacy of demonstrative reference in problematic cases, in order to ward off criticism of unproblematic cases. What that account may have been is difficult to say, and no clear indication of a possible answer exists in the surviving evidence67.

  • 68 DL VII.70.9: τις περιπατεῖ, ἐκεῖνος κινεῖται.
  • 69 See esp. Paolo Crivelli (1994), Phronesis 39, p. 189-190.
  • 70 Cf. SE M VIII.93-8, 100, where every subclass of proposition is illustrated with two separate examp (...)

40Nonetheless, the idea that demonstratives can be more or less deictic can help us make sense of a contentious passage in Diogenes Laertius. When listing the different types of simple propositions, Diogenes illustrates indefinite propositions with the examples “someone walks, that one moves”68. Some have tried to read the latter as anaphoric to explain the fact that a demonstrative pronoun, ekeinos, might be considered indefinite69. This is implausible: giving pairs of (independent) examples is the norm in our surviving evidence70; why would this particular pair be connected in this way?

  • 71 P 2.2.18.10, 2.1.1.57.10-12, 2.2.136.10. See also Galen de Hipp. et Plat. plac. II.2 (below). Teren (...)

41The solution lies rather in the idea that deixis comes in degrees and only fully deictic pronouns make propositions definite. As Apollonius frequently recalls, ekeinos (“that”) is used in Greek to refer to objects further removed or more distant than those referred to by houtos (“this”)71. Since the referent of ekeinos tends to be at a distance, it is more likely that what is being pointed out is not clear. I might use ekeinos to say “that one is my brother” while pointing to a group of men standing together at a distance. It might be difficult for my interlocutor to identify exactly which man I am pointing at. She would fail, or risk failing to identify the referent of the demonstrative and therefore not be in a good position to assess the truth-value of the proposition. In such a situation, in fact, the content of the sentence “that one is my brother” would not be unlike the content of “someone is my brother”: I have asserted that there is a person in the world (in the distance, towards which I point) who is my brother, but I have failed to identify specifically which individual that is. To do so successfully, I would need an additional, more deictic proposition, which may require me to get closer and would likely take the form “this one right there (houtos, or hode) is my brother”. This new proposition would be definite, because it would successfully identify the person I am talking about.

  • 72 This is, at least to some extent, in line with the use of houtos and ekeinos in Greek more generall (...)
  • 73 Alternatively, we could infer from Diogenes’ description that we should limit our understanding of (...)

42To make its proposition definite, the demonstrative and its accompanying ostension need to make the referent evident – again, an epistemological criterion. There can be no doubt as to whom I am referring to. The subject must be unambiguous and unequivocal. Ekeinos most often does not meet this requirement, which is why it can be given as an example of an indefinite subject72. Even houtos can fall short: as we have seen, its deixis can sometimes require strengthening. Only full deixis – that is, successful deixis, which unequivocally identifies its referent – can impart definiteness. Therefore, when Diogenes Laertius tells us that a definite proposition is composed of a deictic nominative and a predicate, he is giving a necessary but not sufficient condition for definiteness: not all forms of deixis ensure the definiteness of the proposition73.

V. Deixis and Physical Ostension

  • 74 e.g. S II.13, 83= GG 2.2.136.5-10, 187.14-17.
  • 75 As has often been assumed in scholarship. See esp. Jacques Brunschwig (1984), Histoire Épistémologi (...)

43The status of ekeinos highlights another salient feature of deixis, namely that pointing does not guarantee successful reference. There is no denying that ekeinos is a demonstrative. We know that it was thought by Apollonius to be naturally deictic74. It could be and surely regularly was accompanied by pointing. If pointing were enough to refer deictically, propositions using ekeinos should be definite. Yet they are not. It is not the case, then, that pointing is sufficient to refer deictically75.

44In addition to not being sufficient, pointing is also not necessary. “Pointing at” is too restrictive for the kind of ostension required to secure deictic reference. Ostension can be made in different ways. Chrysippus highlights this in his On the Soul:

  • 76 Galen de Hipp. et Plat. plac. II.2: Οὕτως δὲ καὶ τὸ ἐγὼ λέγομεν, κατὰ τοῦτο δεικνύντες ἑαυτοὺς ἐν ᾧ (...)

“We also say ‘egō’ in this way, indicating ourselves in the place in which thought appears to be, with the deixis carried there naturally and appropriately, and without such a deixis with our hand, we say ‘egō’ nodding towards ourselves; the word ‘egō’ is such and is expressed together the deixis described next. For as we say ‘egō’, with the first syllable, we drop our bottom lip deictically towards ourselves. Following the movement of the chin, the nod to the chest and such a deixis, the next syllable is juxtaposed suggesting no distance, as is the case with ekeinos.”76

45Egō is deictic because, as we say it, our chin nods to our chests. Chrysippus’ expression “without such a deixis with our hand” suggests that deixis can happen in a variety of ways: “pointing” can be done by other means than our hands. Perhaps this is a tenuous point, and we should simply understand pointing as encapsulating other kinds of physical ostensions.

46Nonetheless, we should be open to the possibility of non-physical types of ostension. Our sources provide us with examples of definite propositions whose subjects do not seem to be easily pointed out physically. Sextus Empiricus lists the following argument as an example of a Stoic argument:

  • 77 SE Outlines of Pyrrhonism (PH) II.141: εἴ τίς σοι θεῶν εἶπεν ὅτι πλουτήσει οὗτος, πλουτήσει οὗτος· (...)

“If some one of the gods said to you that this one will be rich, then this one will be rich. This god (let’s say I am pointing to Zeus) said that this one will be rich. Therefore this one will be rich.”77

  • 78 Susanne Bobzien (1999), CHHP p.99 also rejects deixis by proxy, though on somewhat different ground (...)
  • 79 Michael Frede (1974), Die Stoische Logik, Göttingen, Vandenhoeck und Ruprecht, p. 56, claims that t (...)
  • 80 Anthony Lloyd (1978), Les Stoïciens, p. 286 brushes this passage aside on the grounds that the deix (...)

47It seems plausible that one might want to say “this one” to refer to a god who is not currently present around us in an immediately graspable way. Perhaps one might point at their statue or their temple as a proxy78. For this, and occurrences like it, to count as definite it is important that the god be picked out for the first time and the reference not be established on the basis of assumed shared knowledge among the interlocutors. Reference by proxy in this case would surely be relying on previous knowledge of what the proxy is a proxy for. In addition, it would not be pointing at the object’s matter. How does one physically point at a god’s matter? The Stoics believed Zeus to be co-extensive with the universe. In order to point to Zeus, I would have to point to the whole of the world79. It is not clear how this could be achieved through physical ostension: waving one’s hands in the air would hardly capture the whole cosmos. “I point to Zeus”, must mean something other than physical ostension80.

  • 81 Alex. Aphr. in Ar. An. Pr. 180.9: ὁ δ' αὐτὸς λόγος καὶ ἐπὶ τοῦ εἰ νύξ ἐστιν, οὐκ ἔστιν αὕτη ἡμέρα. (...)

48Similarly, Alexander presents the true conditional “if it is night, this is not day” as an alternative or parallel to “if Dion has died, this one has died”.81 Just as in the case of “this one has died”, the consequent “this is not day” is presented as an impossible definite proposition. “This” indicating the day, then, makes its proposition definite. Just as “this one has died” is false while the referent is still alive, “this is not day” is false while it is still day and destroyed when the day is over because the day can no longer be indicated. Only expressions whose subjects have successfully referred can express propositions and only propositions can be determined to be true or false, so it is assumed that “this” (indicating day) can successfully refer while it is day. Here again, as in the case of the bug’s exoskeleton, it is not entirely clear how the Stoics would have explained the details of the successful reference. Nonetheless it is clear from the use of the example that Stoics assumed it possible and perhaps this is sensible: a competent speaker should be able to identify that the referent is the day of utterance.

49Yet it is unclear how one would physically point to “day”, even though it is surely graspable. Even if one were to point to the ground, or to wave her arms around to indicate what she means, it is not obvious that those gestures would make it any clearer. What makes it clear to competent speakers does not seem to have anything to do with physical ostension. One could argue that once again we rely on assumed shared knowledge: reference is established because of the shared knowledge of the time of utterance. But if this were so, the proposition would not be definite. As we have seen a child, or someone who has recently recovered sight, who does not have the concept <day>, should be able to acquire it from the definite proposition “this is day”, so that we cannot assume this previous knowledge. If the Stoics considered such a proposition definite, they must have thought that “this” indicated the day, which was immediately present, graspable and could be referred to deictically. This once again suggests that, for the Stoics, deixis includes a non-physical kind of ostension which nonetheless fixes the reference.

  • 82 Similarly, given the relationship between definite and indefinite propositions outlined in SE M VII (...)
  • 83 As e.g. Long (1971), “Language and Thought in Stoicism” argues.
  • 84 It is worth noting again (cf. n.79) that what I have in mind here is not what Apollonius Dyscolus c (...)

50One final reason to think that non-physical ostension is possible has to do with the metaphysical status of thoughts and lekta for the Stoics. Indeed, it seems likely that we have deictic thoughts about thoughts, or incorporeal lekta, or about concepts (which are also non-existents). I can think, or indeed say, about a proposition, “this is stupid”, or “that’s what I meant”82. Surely, these non-corporeal things cannot be pointed out by means of physical ostension. In fact, the very fact that we have deictic thoughts suggests that non-physical ostension is both possible and common. We can think or grasp the thought “this is beautiful”, without saying it, or physically pointing. If lekta are the content of thoughts83, it seems that we ought to account for instances such as this too, where the deixis appears to operate solely in our mind84.

VI. Anaphora and Deixis

  • 85 The status of anaphora has long been debated. Some, such as Michael Frede (1974), Die Stoische, p.  (...)

51We may be tempted to think that such instances of non-physical deixis to things said or thought would be instances of anaphora, which would suggest that anaphoric propositions are definite, as some have claimed85. Yet the epistemological underpinnings of the semantics of demonstratives and definite propositions help us rule out anaphora as making its proposition definite. In fact, both Apollonius and Priscian clearly distinguish deictic and anaphoric pronouns. Apollonius emphasises that while anaphora can be built on a deictic, a deictic cannot pick up on an anaphora:

  • 86 Ap. Dysc. P. 61.3-8: τοῦτο δὲ συμβέβηκεν, ἐπεὶ ἡ μὲν ἐκεῖνος καὶ ἡ οὗτος, δεῖξιν σημαίνουσαι, τὴν ὑ (...)

This happens since ekeinos and houtos, because they signify deixis, provide immediate perception of the person while autos refers back to a person brought up again. Rightly then does autos refer back to the deixis shown by ekeinos but ekeinos and houtos cannot refer back again to autos. For the acquaintance from the deictic pronouns comes first.”86

52The use of “shown” (δηλουμένην, instead of, for example, “signified”, σημαινομένην) in the second sentence suggests that, despite the odd turn of phrase, what autos refers back to is the person referred to by the deixis (and not the deixis itself). Priscian makes the point more clearly:

  • 87 Priscian, IG XVII.56.9-11 = 3.142.1-4: Et sciendum quod demonstrativa pronomina non aliquorum praed (...)

It should also be known that deictic pronouns do not replace such or such nouns previously used, as anaphorics do, but those nouns which cannot be expressed deictically.”87

  • 88 cf. n.29 above.

53Anaphoric pronouns are used to replace or refer back to the referents of nouns previously mentioned. When we want to refer to something new with deixis, only a demonstrative will do, a noun or an anaphoric cannot properly be expressed deictically88.

54Deictic and anaphoric pronouns are relevantly different. They differ firstly in that anaphoric pronouns do not refer by means of deixis, and secondly in that anaphoric pronouns refer to someone or something previously mentioned while deictic pronouns do not. This makes sense: as we have seen, an essential characteristic of deixis is that it does not require previous knowledge to secure reference. In particular, it does not assume any previous knowledge of the referent. Anaphoric pronouns thus lack deixis as well as the epistemic simplicity required to form definite propositions, so that instances of anaphora cannot be definite.

  • 89 To be sure, in cases where the noun plays no part in fixing the reference and leaves all the work t (...)

55We can similarly rule out complex expressions, such as “this man”, houtos ho anthrōpos. The expression “this man” features the noun “man” which appeals to the concept <man>, so that it does not refer purely deictically, focusing on matter alone, but also focuses on the quality of being a man by means of a concept89. As such, it is not simple or fundamental in the way which we have seen is characteristic of definite propositions. Not all demonstrative expressions, then – very few, in fact: only pronouns and not all pronouns – confer their privileged status on the propositions which use them as subjects.

Conclusion

56We set out to give an account of the semantics of demonstrative subjects of definite propositions in order to explain their privileged status. We have found that this privilege stems from ontological and epistemological implications of the semantic properties of demonstratives. Demonstrative subjects of definite propositions refer by deixis, so that they focus on matter – the most fundamental ontological category – and by means of a simple expression that does not appeal to any quality or previously known concept. In addition, their referent is necessarily immediately graspable. Definite propositions are in turn epistemologically fundamental and constitute the first step of concept formation. However, not all demonstratives will fulfil the requirements for making their propositions definite. Only fully deictic demonstratives will, namely ones which successfully and unequivocally refer. This full deixis is not guaranteed by pointing, nor does it require pointing, or any other kind of physical ostension. As we have seen, this rules out ekeinos in favour of houtos which guarantees closer proximity and therefore fuller deixis. Successful deictic reference, then, confers ontological and epistemological privilege to definite propositions which employ it, and which are, as a result, central to Stoic semantics.

Haut de page

Bibliographie

Editions of Primary Sources

Dorandi, T. (2013), Diogenes Laertius: Lives of Eminent Philosophers, Cambridge, New York, CUP.

Keil, H. (1822-94), Grammatici latini, Leipzig, Teubner.

Mau, J. & H. Mutschmann (1914-1961), Sexti Empirici opera, vols. 2 & 3, 2nd ed., Leipzig, Teubner, 2, 1914; 3, 1961: 2: p. 3-429; 3: p. 1-177.

Mutschmann, M. (1912), Sexti Empirici opera, vol. 1, Leipzig, Teubner, p. : 3-131, 133-209.

Plasberg, O. (1922), M. Tulli Ciceronis Scripta Quae Manserunt Omnia, fasc. 42, Leipzig, Teubner.

Uhlig, G. (1910), Grammatici Graeci, vol. 2.2, Leipzig, Teubner (repr. Hildesheim, Olms, 1965)

Wallies, M. (1883), Alexandri in Aristotelis analyticorum priorum librum i commentarium [Commentaria in Aristotelem Graeca 2.1, Berlin, Reimer].

Secondary Literature

Alessandrelli, Michele (2013), Il Problema del Lekton nello Stoicismo Antico, Firenze, Olschki.

Bailey, Dominic T. J. (2014), “The Structure of Stoic Metaphysics”, Oxford Studies in Ancient Philosophy 46, p. 253–309.

Bakker, Egbert J. (2010), “Pragmatics: Speech and Text”, in Egbert J. Bakker (ed.), A Companion to the Ancient Greek Language, Oxford, Blackwell, p. 151-168.

Baratin, Marc (1989), La naissance de la syntaxe à Rome, Paris, Les Éditions de Minuit.

Baratin, Marc, Bernard Colombat & Louis Holtz (éds.) (2009), Priscien : transmission et refondation de la grammaire, de l’Antiquité aux Modernes, Turnhout, Brepols (Studia Artistarum, 21).

Barnes, Jonathan (1999), “Aristotle and Stoic Logic”, in Katerina Ieradiokanou (ed.), Topics in Stoic Philosophy, Oxford, OUP, p. 23-52.

Blank, David L (1982), Ancient philosophy and Grammar: The syntax of Apollonius Dyscolus, Chico, CA, Scholars Press.

Blank, David L. & Catherine Atherton (2003), “The Stoic Contribution to Traditional Grammar”, in Brad Inwood (ed.), The Cambridge Companion to the Stoics, Cambridge, CUP, p. 310-327.

Bobzien, Susanne (1999), “Logic:  The Stoics”, in Kempe Algra, Jonathan Barnes, Jaap Mansfeld & Malcolm Schofield (eds.), The Cambridge History of Hellenistic Philosophy, Cambridge, CUP, p. 77-176.

Brandenburg, Philipp (2005), Apollonios Dyskolos. Über das Pronomen, Leipzig, Teubner.

Brittain, Charles (2012), “Antiochus’ epistemology”, in David Sedley (ed.), The Philosophy of Antiochus, Cambridge, CUP, p. 104-130.

— (2006), On Academic Scepticism, Indianapolis, Hackett.

Bronowski, Ada (2014), « La structure logique du langage ordinaire chez les stoïciens », in Jean-Michel Counet (éd.), Philosophie et langage ordinaire de l'Antiquité à la Renaissance, Louvain, Peeters, p. 83-96.

Brunschwig, Jacques (1984), « Quelques remarques sur la conception stoïcienne du nom propre », Histoire Épistémologie Langage 6, p. 3-19.

— (1986), « Remarques sur la classification des propositions simples dans les logiques hellénistiques », Philosophie du langage et grammaire dans l’Antiquité, Brussels/Grenoble, Ousia, p. 287-310.

— (1988), « La théorie stoïcienne du genre suprême et l'ontologie platonicienne », in Jonathan Barnes & Mario Mignucci (eds.), Matter and Metaphysics, Naples, Bibliopolis, p. 19-27.

Caston, Victor (1999), “Something or Nothing: The Stoics on Concepts and Universals”, Oxford Studies in Ancient Philosophy 17, p. 145-213.

Crivelli, Paolo (1994), “Indefinite Propositions and anaphora in Stoic logic”, Phronesis 39, p. 187-206.

de Jonge, Casper (2008), Between Grammar and Rhetoric. Dionysius of Halicarnassus on Language, Linguistics and Literature, Leiden, Brill.

Denyer, Nicholas (1988), “Stoicism and Token Reflexivity”, in Jonathan Barnes & Mario Mignucci (eds.), Matter and Metaphysics, Naples, Bibliopolis, p. 375-396.

Dyson, Henry (2009), Prolepsis and Ennoia in the Early Stoa, Berlin, De Gruyter.

Ebert, Theodor (1993), “Dialecticians and Stoics on the classification of Propositions”, in Karl Döring and Theodor Ebert (eds.), Dialektiker und Stoiker zur Logik der Stoa und ihrer Vorläufer, Stuttgart, Steiner, p. 111-127.

Frede, Michael (1974), Die Stoische Logik, Göttingen, Vandenhoeck und Ruprecht.

Goldschmidt, Victor (1969), Le système stoïcien et l’idée de temps, Paris, Vrin.

Goulet, Richard (1978), « La classification stoïcienne des propositions simples selon Diogene Laërce VII.69-70 », in Jacques Brunschwig (éd.), Les Stoïciens et leur logique, Paris, Vrin, p. 171-198.

Graeser, Andreas (1978), “The Stoic theory of Meaning”, in John Rist (ed.), The Stoics, Berkeley, University of California Press, p. 77-100.

Hadot, Ilsetraut (2014), “Getting to Goodness: Reflections on Chapter 10 of Brad Inwood, Reading Seneca”, in Marcia L. Colish & July Wildberger (eds.), Seneca Philosophus, Berlin, De Gruyter, p. 9-41.

Householder, Fred W. (1981), The Syntax of Apollonius Dyscolus, Amsterdam, Benjamins.

Hülser, Karlheinz (1987), Die Fragmente Zur Dialektik der Stoiker, Stuttgart, Verlag.

Ildefonse, Frédérique (1997), La naissance de la grammaire dans l’Antiquité grecque, Paris, Vrin.

Inwood, Brad (2005), “Getting to Goodness”, in Brad Inwood, Reading Seneca: Stoic Philosophy at Rome, Oxford, Clarendon, p. 271-301.

Irwin, Terence H. (1997), “Aristotelian Substances and Stoic Subjects”, Revue Internationale de Philosophie, 51, N° 201,3, p. 397-415.

Kahn, Charles H. (1969), “Stoic Logic and Stoic Logos”, Archiv fur Geschichte der Philosophie 51, p. 158-172.

Kovacs, David (1995), Euripides : Hippolytus, Cambridge, Loeb.

Lallot, Jean (1997), Apollonios Dyscole, De la construction, Paris, Vrin.

Lloyd, Anthony C. (1970), “Activity and description in Aristotle and the Stoa”, Proceedings of the British Academy 56, p. 227-240.

— (1971), “Grammar and Metaphysics in the Stoa”, in Anthony A. Long (ed.), Problems in Stoicisim, London, Athlone, p. 58-74.

— (1977), Review of M. Frede, Die Stoische Logik, Mind 86, p. 286.

— (1978), “Definite propositions and the concept of reference”, in Jacques Brunschwig (éd.), Les Stoïciens et leur logique, Paris, Vrin, p. 285-296.

Long, A.A. (1971), “Language and thought in Stoicism”, in A.A. Long (ed.), Problems in Stoicisim, London, Athlone, p. 75-113.

Luhtala, Anneli (2000), On the origin of syntactical description in Stoic logic, Münster, Nodus.

— (2005), Grammar and philosophy in late antiquity: A study of Priscian's sources, Amsterdam, John Benjamins.

Menn, Stephen (1999), “The Stoic Theory of Categories”, Oxford Studies in Ancient Philosophy 17, p. 215-247.

Mueller, Ian (2006), Alexander of Aphrodisias: On Aristotle Prior Analytics 1.32-46, Ithaca, N.Y, Cornell University Press.

— (2013), Alexander of Aphrodisias: On Aristotle Prior Analytics 1.14-22, Ithaca, N.Y, Cornell University Press.

Pachet, Pierre (1975), « La déixis selon Zénon et Chrysippe », Phronesis 20, p. 241-246.

Schmidt, Rudolf Traugott (1839), Stoicorum Grammatica, Hakkert.

Sedley, David N. (1982), “The Criterion of Stoic Identity”, Phronesis 27, p. 255-275.

— (1984), “The Stoic Theory of Universals”, The Southern Journal of Philosophy 23, p. 87-92.

— (1999), “Hellenistic Physics and Metaphysics”, in Kempe Algra, Jonathan Barnes, Jaap Mansfeld, & Malcolm Schofield (eds.), The Cambridge History of Hellenistic Philosophy, Cambridge, CUP, p. 355-411.

Sellars, John. (2011), “Stoic Ontology and Plato's Sophist”, Bulletin of the Institute of Classical Studies 54, p. 185-203.

Sluiter, Ineke (1990), Ancient Grammar in Context. Contributions to the Study of Ancient Linguistic Thought, Amsterdam, VU University Press.

Haut de page

Notes

1 See esp. Charles H. Kahn (1969), “Stoic Logic and Stoic Logos”, Archiv fur Geschichte der Philosophie 51, p. 158-172; Pierre Pachet (1975), “La deixis selon Zénon et Chrysippe”, Phronesis 20, p. 241-246; Anthony C. Lloyd (1978), “Definite propositions and the concept of reference”, in Jacques Brunschwig (éd.), Les Stoïciens et leur logique, Paris, Vrin, p. 285-296; Jacques Brunschwig (1984), « Quelques remarques sur la conception stoïcienne du nom proper », Histoire Épistemologie Langage 6, p. 3-19; Jean Lallot (1997), Apollonios Dyscole, De la construction, Paris, Vrin; Susanne Bobzien (1999), “Logic: The Stoics”, in Kempe Algra, Jonathan Barnes, Jaap Mansfeld & Malcolm Schofield (eds.), The Cambridge History of Hellenistic Philosophy, Cambridge, CUP, p. 77-176, here p. 98-9.

2 The most noteworthy exception is Frédérique Ildefonse (1997), La naissance de la grammaire dans l’Antiquité grecque, Paris, Vrin.

3 I translate axiōma by “proposition”, which I take to be uncontroversial. While “proposition” is a loaded term in contemporary philosophy of language, it captures well what the Stoics mean by axioma: the incorporeal meaning of an assertive sentence which is the bearer of truth-value.

4 This account is echoed by Diogenes Laertius, Lives of Eminent Philosophers (DL) VII.70. The differences between the two accounts matter little for my purposes here since they agree on the aspects most salient to my investigation. I adopt Sextus’ terminology because it is the simplest and clearest. For comparative discussions of the two accounts see Brunschwig, Jacques (1986), « Remarques sur la classification des propositions simples dans les logiques hellénistiques », Philosophie du langage et grammaire dans l’Antiquité, Brussels/Grenoble, Ousia, p. 287-310; Theodor Ebert (1993), “Dialecticians and Stoics on the classification of Propositions”, in Karl Döring & Theodor Ebert (eds.), Dialektiker und Stoiker zur Logik der Stoa und ihrer Vorläufer, Stuttgart, Steiner, p. 111-127; and Susanne Bobzien (1999), “Logic: the Stoics”, CHHP, p. 97, esp. n.56.

5 Sextus Empiricus, Against the Mathematicians (SE M) VIII.96: τῶν δὲ ἁπλῶν τινὰ μὲν ὡρισμένα ἐστὶν τινὰ δὲ ἀόριστα τινὰ δὲ μέσα, ὡρισμένα μὲν τὰ κατὰ δεῖξιν ἐκφερόμενα, οἷον ‘οὗτος περιπατεῖ, οὗτος κάθηται’ (δείκνυμι γάρ τινα τῶν ἐπὶ μέρους ἀνθρώπων). ἀόριστα δέ ἐστι κατ' αὐτοὺς ἐν οἷς ἀόριστόν τι κυριεύει μόριον, οἷον ‘τὶς κάθηται’, μέσα δὲ τὰ οὕτως ἔχοντα ‘ἄνθρωπος κάθηται’ ἢ ‘Σωκράτης περιπατεῖ’. Unless otherwise noted, translations are my own.

6 Sextus Empiricus, M VIII.97: τὸ μὲν οὖν “τὶς περιπατεῖ” ἀόριστόν ἐστιν, ἐπεὶ οὐκ ἀφώρικέ τινα τῶν ἐπὶ μέρους περιπατούντων· κοινῶς γὰρ ἐφ' ἑκάστου αὐτῶν ἐκφέρεσθαι δύναται· τὸ δὲ “οὗτος κάθηται” ὡρισμένον ἐστίν, ἐπείπερ ἀφώρικε τὸ δεικνύμενον πρόσωπον. τὸ δὲ “Σωκράτης κάθηται” μέσον ὑπῆρχεν, ἐπείπερ οὔτε ἀόριστόν ἐστιν (ἀφώρικε γὰρ τὸ εἶδος), οὔτε ὡρισμένον (οὐ γὰρ μετὰ δείξεως ἐκφέρεται), ἀλλ' ἔοικε μέσον ἀμφοτέρων ὑπάρχειν, τοῦ τε ἀορίστου καὶ τοῦ ὡρισμένου.

7 On the Stoic notion of eidos, see especially Jacques Brunschwig (1988), « La théorie stoïcienne du genre suprême et l'ontologie platonicienne », in Jonathan Barnes & Mario Mignucci (eds.) Matter and Metaphysics, Naples, Bibliopolis, p. 19-27; David N. Sedley (1984), “The Stoic Theory of Universals”, The Southern Journal of Philosophy 23, p. 87-92; (1999), “Hellenistic Physics and Metaphysics”, in Kempe Algra, Jonathan Barnes, Jaap Mansfeld & Malcolm Schofield (eds.), The Cambridge History of Hellenistic Philosophy, Cambridge, CUP, p. 355-411; Victor Caston (1999), “Something or Nothing: The Stoics on Concepts and Universals”, Oxford Studies in Ancient Philosophy 17, p. 145–213; John Sellars (2011), “Stoic Ontology and Plato's Sophist”, Bulletin of the Institute of Classical Studies 54, p. 185-203.

8 Sextus Empiricus, M VIII.98: γίνεσθαι δέ φασι τὸ ἀόριστον ἀληθές, τὸ “τὶς περιπατεῖ” ἢ “τὶς κάθηται”, ὅταν τὸ ὡρισμένον ἀληθὲς εὑρίσκηται, τὸ “οὗτος κάθηται” ἢ “οὗτος περιπατεῖ”· μηδενὸς γὰρ τῶν ἐπὶ μέρους καθημένου οὐ δύναται ἀληθὲς εἶναι τὸ “τὶς κάθηται” ἀόριστον.

9 For more on this passage, see Paolo Crivelli (1994), “Indefinite Propositions and anaphora in Stoic logic”, Phronesis 39, p. 187-206, and Anthony C. Lloyd (1971), “Grammar and Metaphysics in the Stoa”, in Anthony A. Long (ed.), Problems in Stoicisim, London, Athlone, p. 58-74.

10 Some (e.g. Andreas Graeser (1978), “The Stoic theory of Meaning”, in John Rist (ed.) The Stoics, Berkeley: University of California Press, p. 77-100, here p.78) have assumed that middle propositions also take their truth-value from corresponding definite propositions. “Socrates is walking” would be said to be true only if “this one is walking” is also found to be true. I see no evidence for this. In fact, it seems clear that it is not the case. In his response to the Master Argument, Chrysippus uses the example “If Dion is dead, this one is dead”, which he calls a true conditional, to show that the impossible can follow from the possible. Alexander of Aphrodisias, Commentaria in Analytica priora Aristoteli (Alex. Aphr.), In An. Pr. 177.25-178.1: τῷ συνημμένῳ τῷ εἰ τέθνηκε Δίων, τέθνηκεν οὗτος δεικνυμένου τοῦ Δίωνος ἀληθεῖ ὄντι). Chrysippus claims that “Dion is dead” can be true (and is therefore possible): it is true once Dion has died; but that “this one is dead”, can never be true (and hence is impossible). I return to this passage in more detail below. For now, let us just note that it cannot be the case that the middle proposition “Dion is dead” relies on the corresponding definite proposition “this one is dead” for its truth-value, since the former can be true but the latter cannot. In other words, there is a distinction to be made between the logical implication (captured by the true conditional “if Dion is dead, then this one is dead”) and a metalogical claim about the truth conditions of the two propositions which form the antecedent and the consequent of the conditional. The propositions and their truth-value relate to each other on two different levels. Since the conditional “if Dion is dead, this one is dead” is true, the antecedent, “Dion is dead”, logically implies the consequent, “this one is dead”. However, it is not the case that the truth-value of the antecedent depends on the truth-value of the consequent: “Dion is dead” can have a truth-value even when “this one is dead” does not, since “Dion is dead” is true after Dion’s death and “this one is dead” is neither true nor false. Sextus Empiricus, in the text above, is usually understood to be stating metalogical commitments. On the other hand, when calling the conditional “true”, Chrysippus is committing to the logical implication but he is not making any metalogical claim. Given these commitments, there are no grounds to think that Chrysippus had a theory according to which middle propositions rely on definite propositions for their truth-value.

11 This would have avoided an issue: “someone is dead” can apparently never be true, since, as we will see in section IV below, “this one is dead” can never be true. The fact that they chose the definite proposition rather than the middle, or both, is significant.

12 Diogenes Laertius VII.45.

13 The details of the correspondence are still subject to debate. Anthony Lloyd (1971), Problems in Stoicism, p. 62ff (following Rudolf Traugott Schmidt (1839), Stoicorum Grammatica, Hakkert) provides a thorough account of it. For a different interpretation see Frédérique Ildefonse (1997), La naissance de la grammaire, p. 37 and p.224-228. An important difference between the two concerns the linguistic equivalents to disposition and relation, which do not bear on my argument.

14 “Matter” and “qualities” are thus technical notions for the Stoics, and distinct from Aristotle’s use of the terms. For a detailed account of the Stoic theory of categories, see Stephen Menn (1999), “The Stoic Theory of Categories”, Oxford Studies in Ancient Philosophy 17, p. 215-247, and David Sedley (1999), “Hellenistic Physics and Metaphysics”, CCHP.

15 Diogenes Laertius VII.58: Ἔστι δὲ προσηγορία μὲν κατὰ τὸν Διογένην μέρος λόγου σημαῖνον κοινὴν ποιότητα, οἷον Ἄνθρωπος, Ἵππος· ὄνομα δέ ἐστι μέρος λόγου δηλοῦν ἰδίαν ποιότητα, οἷον Διογένης, Σωκράτης·

16 The grammarian Apollonius Dyscolus was heavily influenced by the Stoics, as shown by, e.g., David L. Blank (1982), Ancient philosophy and Grammar: The syntax of Apollonius Dyscolus, Chico, CA, Scholars Press; Ineke Sluiter (1990), Ancient Grammar in Context. Contributions to the Study of Ancient Linguistic Thought, Amsterdam, VU University Press; Frédérique Ildefonse (1997), La naissance de la grammaire; Jean Lallot (1997), Apollonios Dyscole; Anneli Luhtala (2000), On the origin of syntactical description in Stoic logic, Münster, Nodus; David L. Blank & Catherine Atherton (2003), “The Stoic Contribution to Traditional Grammar”, in Brad Inwood (ed.), The Cambridge Companion to the Stoics, Cambridge, CUP, p. 310-327. Marion Durand (2018), Language and Reality: Stoic Semantics Reconstructed, University of Toronto (dissertation available from the author upon request) argues at greater length that he in fact worked within a Stoic framework for linguistic theory, so that his work, albeit not strictly speaking Stoic, can serve as a source for Stoic theory. This particular passage, though not included in Karlheinz Hülser (1987), Die Fragmente Zur Dialektik der Stoiker, Stuttgart, Verlag, is particularly likely to be evidence of a Stoic view, since the claim it makes about nouns (that they signify quality) is precisely the one reported by Diogenes Laertius. To be sure, this needn’t mean that Apollonius is committed to every detail of the Stoic view, only that this passage can be used as evidence for it.

17 Ap. Dysc. S II.22.3-6, 24.1-4 = GG 2.2.142.1-4, 143.9-144.4: Ἡ τῶν ὀνομάτων θέσις ἐπενοήθη εἰς ποιότητας κοινὰς ἢ ἰδίας, ὡς <ἄνθρωπος, Πλάτων>, καὶ ἐπεὶ οὔτε σὺν δείξει τὰ τοιαῦτα οὔτε ἀναφορᾷ, πάμπολλος ἡ ἐπὶ τούτων θέσις ἐγίνετο, ἵν' ἑκάστου τὸ χαρακτηριστικὸν ἀπονείμῃ τὴν ἑκάστου ποιότητα. [...] Αἱ μέντοι ἀντωνυμίαι εἰς οὐδὲν ἄλλο ἀποβλέπουσαι διὰ τῆς ἐν αὐταῖς ἐγκειμένης δείξεως ἢ εἰς τὴν ὑποκειμένην οὐσίαν, ἔχουσιν εὔληπτα καὶ τὰ ἐπισυμβεβηκότα τῇ ποιότητι, λέγω τὰ δυνάμενα δι' ὄψεως ἐπινοηθῆναι, <λευκὸν ἢ μέλαν, μακρὸν ἢ βραχύ>. I do not follow Lallot in accepting Dudith’s addition (τὴν ἑκάστου <τῶν ὑποκειμένων> ποιότητα).

18 See also I.120, and Priscian, Institutiones Grammaticae (IG), 15.63-65 ; 17.105. Cf. Jean Lallot (1997) Apollonios Dyscole, n.II.263 ; Richard Goulet (1978), « La classification stoïcienne des propositions simples selon Diogene Laërce VII.69-70 », in Jacques Brunschwig (éd.), Les Stoïciens et leur logique, Paris, Vrin, p. 171-198, here p. 175.

19 I take ousia and poiotēs to be used in the technical Stoic sense here, and accordingly translate them as “matter” and “quality”, respectively. While Apollonius may not have himself been committed to Stoic ontological theory, there can be little doubt that the use of these terms in this passage is due to their presence in a Stoic source on which Apollonius is building.

20 A note on the translation of those two key terms: ἀποβλέπειν appears only one other time in Apollonius, at P 2.1.1.31.4, where its meaning is equally debatable. In general, the term can mean “looking to”, or “looking at”, with an intensive force (the LSJ adds “with love, wonder or admiration”), but when constructed with a prepositional phrase governed by εἰς, as it is frequently, most notably in Plato, it tends to mean “to consider”, “to take into consideration”, or “to focus on”, sometimes for the purpose of an investigation, or the forming of an opinion (it occurs multiple times in this sense in the Republic, e.g. 466a5, 472c7, 530a4, 578b5, but also in the Philebus, e.g. 44d9-45a2, where the verb is used twice in just this sense, or again in Symposium 82a3, Phaedrus 237d1,…). It often seems to have the force of considering something rather than something else (that is, focusing on it), whether in the context of an investigation (as is the case in the passages cited above) or not (e.g. Phaedrus 239b6-7). Apollonius seems to be using it in just this way here, hence my translation, “focusing”. For ἔχειν, I prefer “capture” to “include” in Fred W. Householder (1981), The Syntax of Apollonius Dyscolus, Amsterdam, Benjamins, ad loc., and “inclure” in Jean Lallot (1997) Apollonios Dyscole, ad loc., as “include” seems to invite a variety of questions regarding the capacity in which those qualities are included in or by the pronouns. I follow them, however, in taking εὔληπτα not as a predicate accusative but as attributive, modifying τὰ ἐπισυμβεβηκότα. One could take εὔληπτα as a predicate accusative, in which case the phrase would mean “make visible qualities easy to grasp”. Although the word order may seem to make this a probable construal, there are very few parallels. The closest one is found at Eur. Hipp. 953, where Ὀρφέα τ' ἄνακτ' ἔχων is sometimes rendered “make Orpheus your lord” (so David Kovacs (1995) Euripides: Hippolytus, Cambridge, Loeb). This is an attractive interpretation, and in fact, very close to what I think the view entails, as I argue below. Nonetheless the construction with a predicate accusative is rare and more often means “to hold x as y”, which cannot be the meaning here.

21 Similarly, both nouns and pronouns capture their referent’s matter, but only pronouns focus on it. Note that Apollonius does not, strictly speaking, say that nouns focus on quality, but rather that they were established with a view to qualities. My interpretation here conflates the aim of the imposition of nouns with what the nouns themselves aim at or focus on. Given the way the passage is laid out, with the use of the preposition εἰς in both cases, and the distinction made between nouns and pronouns, I think this is not an implausible slip to impute on Apollonius.

22 This is in line with the argument in David N. Sedley (1982), “The Criterion of Stoic Identity”, Phronesis 27, p. 255-275, which convincingly shows that a metaphysical argument for the privileged status of demonstratives based on their ability to track identity over time in a constantly changing world would be misguided. As Sedley argues, names are better able to track identity through time than demonstratives, so that, metaphysically speaking, demonstratives signify something less stable than nouns. This metaphysical feature of their referent can thus not be the source of their privilege.

23 I do not agree with Jean Lallot (1997) Apollonios Dyscole, ad loc., who interprets the passage as suggesting that pronouns are able to say just as much as nouns and in fact replace them even in their descriptive function. The way in which the deixis “includes” or captures quality is precisely not by describing and it is in this respect that demonstratives’ way of securing reference differs from that of nouns.

24 I take the somewhat puzzling phrase τὰ ἐπισυμβεβηκότα τῇ ποιότητι (“the accidents which accompany quality”) to mean little more than the quality in question. The verb ἐπισυμβαίνω is not very common, though it appears in Aristotle and Sextus Empiricus, where it means “supervene”, or “happen besides/after”. The participle ἐπισυμβεβηκός seems to be first used by Apollonius and its subsequent usage is sparse and confined to rhetorical texts. In Apollonius, it appears just three times, once just a paragraph above our passage, where it seems to mean “accidents” in the Aristotelian sense. Apollonius seems to be muddling terminologies here and have in mind something a little confused, perhaps best described as the qualities as they appear to us as a result of the way the Stoic-conceptualised qualities arrange matter.

25 I take it that I could, by contrast, see it without grasping that it is an elm. The difference, as will transpire, may be the complexity of the concepts associated with those qualities. Note that, according to Stoic epistemology, all adult human impressions are affections of the soul – which is the commanding faculty – and rational, that is, with propositional content. I use sight here as the means of perceiving qualities, though Apollonius’ phrasing leaves open the option of other modes of perception. Priscian, is in this respect more restrictive, since he translates eulēpta by qui possunt oculis conspici (IG 17.64 = GL 2.2.146.18).

26 This is why I take the participle ἀποβλέπουσαι to be concessive (contra Lallot and Householder who both translate it with a relative clause): the point is that although, unlike nouns, the demonstrative does not focus on the qualities (but solely on the matter), it nonetheless captures them.

27 Casper De Jonge (2008), Between Grammar and Rhetoric. Dionysius of Halicarnassus on Language, Linguistics and Literature, Leiden, Brill, p. 278ff. shows that this same idea motivates Dionysius of Halicarnassus’ ranking of parts of speech.

28 cf., e.g., Simplicius in Ar. Cat. 222.30-33. I am grateful to Simon Shogry for pointing this out to me.

29 It is worth noting that the exact way in which deixis features alongside deictic pronouns is unclear. As is clear from the texts cited above and below, deixis is considered an inherent feature of demonstrative pronouns (ἡ ἐν αὐταῖς ἐγκειμένη δεῖξις), but its relationship to pronouns is described in many ways. It is not always said to be in a pronoun, but sometimes to accompany the pronoun (μετὰ δείξεως) and to come from the pronoun (ἐξ αὐτῶν, and even ὑπ’ αὐτῶν). Given our evidence, I see no way of decisively judging which of these we ought to adopt as the mainstream view, and I will, in my discussion, try to mirror the text at hand, therefore alternating between the two. My inclination is to think that the deixis should be thought of as something which, strictly speaking, accompanies a pronoun. It is performed by the speaker as she utters the pronoun. It is described as being inherent to pronouns, and as being in them or carried by them, because a competent speaker would not utter a demonstrative pronoun without also performing an ostension. There would be something strange about a demonstrative that is not accompanied by ostension; its reference would fail, and the utterance would not express a proposition or any sort of complete lekton (more on this below). Therefore, the pronoun compels, as it were, the person uttering it to perform the accompanying deixis, which is the reference-fixing mechanism. It is in this sense that the deixis is said to be in it or carried by pronouns. In addition, deixis is a unique and defining feature of demonstratives. Priscian tells us, for example, that nouns do not refer by deixis (IG XVII.63.15, quoted below), nor do anaphoric pronouns (IG XIV.2.1-4). Now, it seems that we could well say “Socrates” and at the same time point to Socrates. So it cannot be the case that nouns cannot be accompanied by deixis. Rather, the point seems to be that, if I need to refer by deixis, then I ought to use a pronoun, not a noun. Conversely, if I utter a noun and successfully refer by means of deixis, that is, if what fixes the reference is deixis, rather than appeal to a concept, then I did not, strictly speaking, refer by means of a noun. Put another way, if I point at Socrates and say “Socrates is sitting” and my interlocutor does not know who Socrates is but successfully identifies Socrates as the referent, I have in effect said “this is Socrates and he is sitting”, that is, a definite proposition and an anaphoric proposition, and not a middle proposition. This is perhaps even clearer in cases where I point at Plato and say “Socrates is sitting”. In such cases it seems that there are two interpretations available: (i) I did express a middle proposition, “Socrates is sitting”, and referred to Socrates, in which case the proposition will be true if sitting obtains of Socrates (wherever he is); or (ii) I in fact expressed a definite proposition, “this is Socrates and he is sitting”, and referred to Plato by means of deixis, in which case the proposition will be false by virtue of the referent not being Socrates. (Such instances are not dissimilar to uses of houtos (instead of houtē) to refer to a woman, as discussed below, or to the case of “Kallias the teacher is walking” being false if Kallias is not a teacher (Alex. Aphr. in Ar. An. Pr. 402.12-19).) This highlights the importance of semantic properties in the definition of parts of speech, as well as the epistemological concerns at play in the semantics.

30 As I argue in section V below, this ostension need not be physical, nor is any physical ostension sufficient for successful deixis.

31 Chrysippus argues that egō is deictic: we indicate ourselves with our chin as we say it (Galen, de Hipp. et Plat. II.2, on which see section V below). This suggests that he saw “I” and “you” as referring in much the same way as houtos, by deixis alone. If the relevant feature of a part of speech is the way in which it secures reference, egō and houtos would seem to belong to the same category and “I walk” would have been classified as a definite proposition.

32 Ap. Dysc. P 11.27-30: αἱ μέντοι πρωτότυποι διὰ μὲν τῆς φωνῆς γένους ἀδιάστολοί εἰσι πάντοτε κατὰ πρῶτον καὶ δεύτερον πρόσωπον, διὰ δὲ τῆς ὑπ' αὐτῶν δείξεως ἡ τῶν γενῶν διαστολὴ παρεμφαίνεται· συνεξηγούμενον γὰρ ἔχει τὸ γένος ἡ δεῖξις. cf. S II.24.19-21 = GG 2.2.144.8, where Apollonius claims that first- and second-person pronouns do not have gendered endings because the deixis in them focuses solely on the matter. It is worth noting that while this passage is strictly speaking about personal pronouns, he goes on to list exceptions to this rule, of which houtos is explicitly not one (S II.26 = GG 2.2.145.8-146.7).

33 This is to some extent suggested by the interpretation of Frédérique Ildefonse (1997), La naissance de la grammaire, p. 172ff and passim, whose account has much in common with mine, though I would reject her suggestion that determinacy of the referent is the central notion here – and that demonstratives are the first and logically least determinate stage. Nor do I think that the different propositions differ first and foremost in the role that they play in inquiry. Instead, I am inclined to place emphasis on the role of the different propositions in the acquisition of concepts.

34 This is only a brief and compressed sketch of the apparatus. For more on the semantics of nouns see especially Anthony Lloyd (1978), Les Stoïciens, and Jacques Brunschwig (1984), Histoire Épistemologie Langage 6.

35 DL VII.60. cf. n. .

36 I use angled brackets to signal reference to concepts throughout.

37 Ap. Dysc. S I.96 = GG 2.2.80.13-81.3: προσώπων ἀοριστουμένων διακριτικά ἐστι τὰ μόρια, διὸ καὶ τὰ ἐξ αὐτῶν πρόσωπα νοούμενα ὁρισμῷ καταλαμβάνεται. καὶ σαφὲς ὅτι αἱ ἐξ αὐτῶν δείξεις πρῶται ἔφοδοί εἰσι τῶν ὑποκειμένων προσώπων, οὐ δεόμεναι τῆς τῶν ἄρθρων συντάξεως· οὐ γὰρ ἀναφέρεται τὰ πρόσωπα, ὑπ' ὄψιν δὲ δείκνυται.

38 The relevant knowledge here is that of concepts corresponding to qualities. For successful communication, interlocutors will need some knowledge, including a grasp of conventions (for example understanding how pointing works), as well as some basic understanding of how the world is organised in order to be able to pick out discreet parts of it as they are being pointed out. This kind of knowledge is presumably required for any successful communication – not just for successful use of deictic pronouns – and any competent speaker (one who can partake in conversations) will have this basic knowledge.

39 Much like Apollonius Dyscolus, whom he follows extremely closely, Priscian shows signs of Stoic influence and can, with caution, serve as evidence of views which stem from Stoic grammar. cf. n., and Marc Baratin (1989), La naissance de la syntaxe à Rome, Paris, Les Éditions de Minuit; Anneli Luhtala (2005), Grammar and philosophy in late antiquity: A study of Priscian's sources, Amsterdam, John Benjamins; Marc Baratin, Bernard Colombat & Louis Holtz (eds.) (2009), Priscien : transmission et refondation de la grammaire, de l’Antiquité aux Modernes, Turnhout, Brepols.

40 Priscian, IG XVII.63.7-15 = GL 3.145.22-146.4: Quare non mediocriter disturbant qualitatis significationem, cum in unam concidant vocem nominum positiones tam in propriis quam in appellativis. Inde caruit definitione persona ea quae in nomine ipso intellegitur; nec mirum, cum propria quoque nomina, quamvis ideo ponantur ut unumquemque ab aliis omnibus discernant, incerta sint tamen, cum non possint omnes eius qualitates quae illum separant ab aliis omnibus ostendere absque demonstrationis auxilio, quae fit per pronomen.

41 It is worth noting that the surviving evidence does not include much explicit consideration of pragmatics, though they may have helped here and elsewhere. In particular, this leads to difficulties with regards to the preciseness and accuracy of deictic reference. It is not clear exactly how the Stoics account for successful deictic reference in cases where pragmatic calculations seem to be required for it. For more discussion of such difficulties, see sections IV below.

42 Alex. Aphr. in Ar. An. Pr. 402.20-4: ἔτι τοῦ μὴ δεῖν οὕτως τὴν ἀπόφασιν ποιεῖσθαι πίστιν καὶ τοιαύτην προφέρουσιν· οὗτος περιπατεῖ, οὗτος οὐ περιπατεῖ δεικνυμένου θήλεος· ἄμφω γὰρ πάλιν τὰ οὕτως λαμβανόμενα ψευδῆ γίνεσθαί φασιν, εἴτε περιπατοίη τὸ δεικνύμενον εἴτε μή. trans. modified from Ian Mueller (2006), Alexander of Aphrodisias: On Aristotle Prior Analytics 1.32-46, Ithaca, N.Y, Cornell University Press.

43 This goes hand in hand with the rejection of the idea that a mistake in gender is a solecism. cf. Ap. Dysc. S III.8 = GG 2.2.273.9-274.7 (cf. Ap. Dysc S III.9 = GG 2.2.274.8).

44 Ap. Dysc. S II.26.14-16 = GG 2.2.145.14-146.1: ἡ γὰρ ἐξ αὐτῶν νοουμένη ἀπόστασις ἀμαυροτέραν τὴν δεῖξιν καθιστάνει, καὶ ἔνθεν παρεισεδύετο τὸ γένος, οὐχ ἵνα τὴν οὐσίαν παραστήσῃ, ἀλλ' ἵνα τὸ γένος διαστείλῃ.

45 It is not entirely clear why Apollonius thinks the gendered endings of houtos are not akin to that of ekeinos. He gives only a very brief justification, based on its similarity with the adverb τημοῦτος, itself a rare alternate form of τῆμος (S 145.10-11).

46 Another reason not to give too much weight to this passage is emphasised by Jean Lallot (1997) Apollonios Dyscole, ad loc., who notes that this particular passage, and especially the argument laid out with respect to ekeinos is very confused.

47 Cicero, Academica II.21:Atqui qualia sunt haec quae sensibus percipi dicimus talia secuntur ea quae non sensibus ipsis percipi dicuntur sed quodam modo sensibus, ut haec : “illud est album, hoc dulce, canorum illud, hoc bene olens. Hoc asperum” : animo iam haec tenemus conprehensa non sensibus. “ille” deinceps “equus est, ille canis. Cetera series deinde sequitur maiora nectens, ut haec quae quasi expletam rerum conprehensionem amplectuntur : “si homo est, animal est mortale rationis particeps”. trans. modified from Charles Brittain (2006), On Academic Scepticism, Indianapolis, Hackett.

48 This epistemological grounding is consistent with the alethic independence discussed in n.10 above.

49 By grasping a proposition of the form “this is a horse”, I do not mean that I need to have uttered it myself. I can grasp a proposition uttered by someone else. In fact, it seems most plausible that I will acquire the concepts <white> or <horse> with the help of a teacher, who will point to white things and horses and say “this is white” and “this is a horse”, respectively. This leaves the problem of the prōtos heuretēs, or first inventor. We know too little about the Stoic theory for the origins of language to know how they might have dealt with this. Note also that the picture presented here assumes that learning a concept and learning the word for it are, as it were, one and the same process. Some believe that Cicero has a pre-rational stage of perception (see Charles Britain (2012), “Antiochus’ epistemology”, in David Sedley (ed.), The Philosophy of Antiochus, Cambridge, CUP, p. 104-130), which might suggest that one first forms a concept (through repeated perception) at a prior level of pre-rational discrimination, and then learns the name of such concepts through an identification of the form “this is F”. The details of the Stoic theory of concept-formation are very much unclear and the view presented here may need to be slightly altered to fit different reconstructions of it. I take it that whatever the details are, however, definite propositions will remain fundamental and play a similarly fundamental role as the one I ascribe to them here.

50 It is not clear how the Stoics thought one could acquire concepts for things they never directly perceiveid, such as historical figures. On this view, in order to say “Socrates was a philosopher”, I should have previously acquired the concept <Socrates>, but surely neither I nor Chrysippus were ever in a position to acquire it by means of a proposition of the form “this is Socrates”. Aetius IV.11.1–6 suggests that some concepts can be acquired by teaching, others by analogy. How this would help in the case of <Socrates> is not obvious (nor is the place of definite propositions in the process in those cases). A similar worry is raised by Seneca Ep. 120, on which see Brad Inwood (2005), “Getting to Goodness”, in Brad Inwood, Reading Seneca: Stoic Philosophy at Rome, Oxford, Clarendon, p. 271-301. Cf. Henry Dyson (2009), Prolepsis and Ennoia in the Early Stoa, Berlin, De Gruyter, and Ilsetraut Hadot (2014), “Getting to Goodness: Reflections on Chapter 10 of Brad Inwood, Reading Seneca”, in Marcia L. Colish & July Wildberger (eds.), Seneca Philosophus, Berlin, De Gruyter, p. 9-41.

51 We may raise an objection, debating the relevant sense of “grasp” or “understand” at play here, and arguing that I cannot fully grasp the proposition “this is a horse” until I have in fact fully formed the concept <horse>. It is worth noting, firstly, that the concept-acquisition need not be immediate or instantaneous, and it need not occur on the first instance of someone saying to me “this is green”. Secondly, the acquisition of the concept and the grasping of the proposition could well coincide, and be simultaneous. That is, the point in time at which I grasp the proposition “this is a tree” could well be exactly the point in time at which I acquire the concept <tree>. We need not posit that the concept-acquisition must be temporally prior to and distinct from grasping the proposition. We need only accept that both grasping the proposition “this is a tree” and acquiring the concept <tree> are prior to grasping a proposition such as “the tree is green”.

52 I mean by this that they do not make an existential claim, as middle propositions do, according to the widely accepted view of nouns developed by Anthony Lloyd (1978), Les Stoïciens and Jacques Brunschwig (1984), Histoire Épistémologie Langage 6, on the basis of Alex. Aphr. In Arist. An. pr. 402-3.

53 Alex. Aphr. in Ar. An. Pr. 177.27-178.1: φησὶ γὰρ ἐν τῷ συνημμένῳ τῷ ‘εἰ τέθνηκε Δίων, τέθνηκεν οὗτος’ δεικνυμένου τοῦ Δίωνος ἀληθεῖ ὄντι τὸ μὲν ἡγούμενον <τὸ> ‘τέθνηκε Δίων’ δυνατὸν εἶναι τῷ δύνασθαί ποτε ἀληθὲς γενέσθαι τὸ τεθνηκέναι Δίωνα, τὸ δὲ τέθνηκεν οὗτος’ ἀδύνατον· ἀποθανόντος γὰρ Δίωνος φθείρεσθαι τὸ ἀξίωμα τὸ ‘οὗτος τέθνηκε’ μηκέτ' ὄντος τοῦ τὴν δεῖξιν ἀναδεχομένου· ἐπὶ γὰρ ζῶντος καὶ κατὰ ζῶντος ἡ δεῖξις. εἰ οὖν μή<τε> τεθνεῶτος αὐτοῦ ἔτι τὸ ’οὗτος’ οἷόν τε, μήτε πάλιν [ἢ] ὑφίσταται ὁ Δίων ὡς δύνασθαι ἐπ' αὐτοῦ  ῥηθῆναι τὸ ‘τέθνηκεν οὗτος’, ἀδύνατον τὸ ‘τέθνηκεν οὗτος’. trans. modified from Ian Mueller (2013), Alexander of Aphrodisias: On Aristotle Prior Analytics 1.14-22, Ithaca, N.Y, Cornell University Press.

54 Nicholas Denyer (1988), “Stoicism and Token Reflexivity”, in Jonathan Barnes & Mario Mignucci (eds.) Matter and Metaphysics, Naples, Bibliopolis, p. 375-396.

55 SE PH II.234; Epictetus Diss.1.7.1.10-21; Simplicius in Ar. Phys.1299.36-1300.10.

56 This is similar to the reading of Susanne Bobzien (1999), “Logic: The Stoics”, CHHP, p. 100 and 117. A significant difference between our views is that unlike Bobzien, I do not believe that the lekton ceases to subsist altogether, but rather that it stops being a proposition. On this, see also Dominic T.J. Bailey (2014), “The Structure of Stoic Metaphysics”, Oxford Studies in Ancient Philosophy 46, p. 253–309, here p. 281, who follows Bobzien, and Nicholas Denyer (1988), “Stoicism and Token Reflexivity”, p. 380, who comes to the conclusion that the proposition “becomes idle” and ceases to be a proposition when there is nothing to be pointed at.

57 This would mirror the symbolization for middle propositions suggested by Anthony Lloyd (1978), Les Stoïciens, and Jacques Brunschwig (1984), Histoire Épistémologie Langage 6.

58 I assume here that the Stoics would analyse x (Gx x=δ) as well-formed even when there is no δ, which would be coherent with the analysis of nouns found at Alex. Aphr. in Arist. An. Pr. 402-3.

59 cf. Jonathan Barnes (1999), “Aristotle and Stoic Logic”, in Katerina Ieradiokanou (ed.), Topics in Stoic Philosophy, Oxford, OUP, p. 23-52, here p. 44, who arrives at very much the same conclusion.

60 DL VII.63

61 This highlights the need for a better account of completeness and incompleteness of lekta, since it suggests that a lekton of the form subject-predicate is incomplete if the subject is an empty term. The description of a complete lekton as one being expressed by a grammatically complete sentence (with an explicit nominative) is therefore inaccurate. At best, that is a necessary but not sufficient condition. Cf. Anthony Long (1971), Problems in Stoicism; Ada Bronowski (2014), « La structure logique du langage ordinaire chez les stoïciens », in Jean-Michel Counet (éd.), Philosophie et langage ordinaire de l'Antiquité à la Renaissance, Louvain, Peeters, p. 83-96; Michele Alessandrelli (2013), Il Problema del Lekton nello Stoicismo Antico, Firenze, Olschki.

62 Charles Kahn (1969), AGPh 51, p. 160. Cf. SVF II. 509, 517-519. See also Victor Goldschmidt (1969), Le système stoïcien et l’idée de temps, Paris, Vrin, here p. 83-87, and Anthony C. Lloyd (1970), “Activity and description in Aristotle and the Stoa”, Proceedings of the British Academy 56, p. 227-240, here p. 232-240.

63 Ap. Dysc. S II.42.1-9 = GG 2.2.155.5-8: Ἕνεκεν τούτου πρὸς οὐδὲν χρειώδεις εἰσὶν αἱ ἀντωνυμίαι στερόμεναι τοῦ τε δεικνύντος προσώπου καὶ τοῦ δεικνυμένου, εἴγε αἱ ἐγγραφόμεναι πάνυ ἀοριστόταταί εἰσιν, ὅτι καὶ τῆς ἰδίας ὕλης ἀπεώσθησαν.

64 Ap. Dysc. P. 2.1.1.37.8-12: αἱ γοῦν διὰ τοῦ <ι> ἐπεκτεινόμεναι καὶ ἔτι τὴν δεῖξιν ἐπιτείνουσαι καὶ τὸ τέλος ὀξύνουσιν, <ἐκεινοσί, οὑτοσί>. καὶ δῆλόν ἐστιν, ὡς ἕνεκα τῆς δεικτικῆς ἐκφορᾶς ἡ τοιαύτη ἐπέκτασις, τῷ μὴ τὴν αὐτός ἐπεκτείνεσθαι, καθὸ οὐδὲ ὅλως δεικτική. cf. S II.6 = GG 2.2.133.

65 I follow Philipp Brandenburg (2005), Apollonios Dyskolos. Über das Pronomen, Leipzig, Teubner, in the translation of the condensed and difficult last sentence. An alternative translation would take οὐδὲ ὅλως to mean “not entirely” rather than “not at all”. However, the distinction seems to be precisely that those words which are deictic can see their deixis intensified, whereas those which are not deictic, like autos, cannot.

66 SE M VIII.99-102

67 While it may seem plausible that pragmatics were (or should have been) appealed to here – and elsewhere in their discussion of deictic reference – there is no evidence of such considerations. cf. n.41.

68 DL VII.70.9: τις περιπατεῖ, ἐκεῖνος κινεῖται.

69 See esp. Paolo Crivelli (1994), Phronesis 39, p. 189-190.

70 Cf. SE M VIII.93-8, 100, where every subclass of proposition is illustrated with two separate examples.

71 P 2.2.18.10, 2.1.1.57.10-12, 2.2.136.10. See also Galen de Hipp. et Plat. plac. II.2 (below). Terence H. Irwin (1997), “Aristotelian Substances and Stoic Subjects”, Revue Internationale de Philosophie, Vol. 51, No. 201,3, p. 397-415, here p. 402, suggests this may be the grounds on which ekeinos is classified as indefinite.

72 This is, at least to some extent, in line with the use of houtos and ekeinos in Greek more generally, as argued by Egbert J. Bakker (2010), “Pragmatics: Speech and Text”, in Egbert J. Bakker (ed.) A Companion to the Ancient Greek Language, Oxford, Blackwell, p. 151-168. He shows that, for example, in dramatic texts, houtos is traditionally used for objects on stage and immediately graspable while ekeinos is used for what is only conceptually grasped and off stage or having just entered the stage.

73 Alternatively, we could infer from Diogenes’ description that we should limit our understanding of deixis to only successful deixis. Given Apollonius’ description of degrees of deixis, however, the former interpretation seems preferable.

74 e.g. S II.13, 83= GG 2.2.136.5-10, 187.14-17.

75 As has often been assumed in scholarship. See esp. Jacques Brunschwig (1984), Histoire Épistémologie Langage, p. 47 and Susanne Bobzien (1999) CHHP p. 100, who talks about deixis in terms of “point-at-ability”.

76 Galen de Hipp. et Plat. plac. II.2: Οὕτως δὲ καὶ τὸ ἐγὼ λέγομεν, κατὰ τοῦτο δεικνύντες ἑαυτοὺς ἐν ᾧ φαίνεσθαι διάνοιαν εἶναι, τῆς δείξεως φυσικῶς καὶ οἰκείως ἐνταῦθα φερομένης· καὶ ἄνευ δὲ τῆς κατὰ τὴν χεῖρα τοιαύτης δείξεως νεύοντες εἰς αὑτοὺς τὸ ἐγὼ λέγομεν, εὐθὺς καὶ τῆς ἐγὼ φωνῆς τοιαύτης οὔσης καὶ κατὰ τὴν ἑξῆς ὑπογεγραμμένην δεῖξιν συνεκφερομένης. τὸ γὰρ ἐγὼ προφερόμεθα κατὰ τὴν πρώτην συλλαβὴν κατασπῶντες τὸ κάτω χεῖλος εἰς αὑτοὺς δεικτικῶς· ἀκολούθως δὲ τῇ τοῦ γενείου κινήσει καὶ ἐπὶ τὸ στῆθος νεύσει καὶ τῇ τοιαύτῃ δείξει ἡ ἑξῆς συλλαβὴ παράκειται οὐδὲν ἀποστηματικὸν παρενσημαίνουσα, ὅπερ ἐπὶ τοῦ ἐκεῖνος συντέτευχεν. Note that this passage confirms that the Stoics thought of ekeinos as a deictic but indicating some distance.

77 SE Outlines of Pyrrhonism (PH) II.141: εἴ τίς σοι <θεῶν> εἶπεν ὅτι πλουτήσει οὗτος, πλουτήσει οὗτος· οὑτοσὶ δὲ ὁ θεός (δείκνυμι δὲ καθ' ὑπόθεσιν τὸν Δία) εἶπέ σοι ὅτι πλουτήσει οὗτος· πλουτήσει ἄρα οὗτος.

78 Susanne Bobzien (1999), CHHP p.99 also rejects deixis by proxy, though on somewhat different grounds.

79 Michael Frede (1974), Die Stoische Logik, Göttingen, Vandenhoeck und Ruprecht, p. 56, claims that the Stoics could not have believed that it was possible to point to Zeus and uses this example to support the view that there is such a thing as a “deixis of the mind”. The main evidence for this is a passage of Apollonius Dyscolus (S II. 12 = GG 2.2.135.13-136.4) where he talks of anaphoric pronouns as operating a “deixis of the mind.” This passage is hardly reliable: Apollonius is scrambling to make his theory consistent by making sense of cases where pronouns, which he repeatedly describes as inherently deictic, are not used in an obviously deictic way. The deixis he appeals to here is unlikely to be the sort of deixis the Stoics believed to make propositions definite. For a brief discussion of anaphora, see section VI below.

80 Anthony Lloyd (1978), Les Stoïciens, p. 286 brushes this passage aside on the grounds that the deixis is reported as conditional (καθ' ὑπόθεσιν). While we may perhaps take this as highlighting the surprising or unusual nature of the ostension, I do not think it is sufficient to dismiss the passage entirely.

81 Alex. Aphr. in Ar. An. Pr. 180.9: ὁ δ' αὐτὸς λόγος καὶ ἐπὶ τοῦ εἰ νύξ ἐστιν, οὐκ ἔστιν αὕτη ἡμέρα. “the same reasoning applies to the proposition ‘if it is night, this is not day’.”

82 Similarly, given the relationship between definite and indefinite propositions outlined in SE M VIII.98, cited above, if the Stoics believed propositions such as “something is incorporeal” or “something is a proposition” to be true, then they must have held true propositions such as “this is an incorporeal” and “this is a proposition”, which require successful deictic reference to incorporeals.

83 As e.g. Long (1971), “Language and Thought in Stoicism” argues.

84 It is worth noting again (cf. n.79) that what I have in mind here is not what Apollonius Dyscolus calls deixis of the mind (S II.12 = GG 2.2.135.13-136.4).

85 The status of anaphora has long been debated. Some, such as Michael Frede (1974), Die Stoische, p. 52-53 have argued that anaphora is always definite, an argument convincingly refuted by Anthony C. Lloyd (1977), Review of M. Frede Die Stoische Logik, Mind 86, p. 286; others, such as Richard Goulet (1978), Les Stoïciens, p. 178, have argued that anaphora will inherit the definiteness of the proposition on which it is anaphoric. A fuller discussion of anaphora is beyond the scope of the present paper and I leave it to another occasion, focusing here only on the definiteness of anaphora.

86 Ap. Dysc. P. 61.3-8: τοῦτο δὲ συμβέβηκεν, ἐπεὶ ἡ μὲν ἐκεῖνος καὶ ἡ οὗτος, δεῖξιν σημαίνουσαι, τὴν ὑπόγυιον γνῶσιν τοῦ προσώπου παριστᾶσιν, ἡ δὲ αὐτός ἐπ' ἀναπολούμενον πρόσωπον φέρεται. ὀρθῶς οὖν ἐπὶ τὴν διὰ τοῦ ἐκεῖνος δηλουμένην δεῖξιν ἐπαναπολεῖται ἡ αὐτός, οὐκέτι μέντοι <ἡ> ἐκεῖνος ἢ οὗτος ἐπὶ τὴν αὐτός δύναται ἀναπέμπεσθαι· πρώτη γὰρ ἡ διὰ τῶν δεικτικῶν ἀντωνυμιῶν γνῶσις.

87 Priscian, IG XVII.56.9-11 = 3.142.1-4: Et sciendum quod demonstrativa pronomina non aliquorum praedictorum loco nominum, quomodo relativa sed eorum quae proferri demonstrative non possunt accipiuntur.

88 cf. n.29 above.

89 To be sure, in cases where the noun plays no part in fixing the reference and leaves all the work to the deixis, such expressions would form definite propositions (cf. n29, above). I take it such cases are rare. I should note that I do not think οὑτοσὶ δὲ ὁ θεός, at SE PH II.141 (above), is a counter-example to this. The noun is added to the demonstrative because this is a hypothetical situation. Because the situation described is hypothetical, there is nothing to be pointed at. The noun θεός is added in this context (the description of a hypothetical situation) to clarify what would be pointed at (namely a god, as opposed to, for example, the man who, the speaker says, will be rich). If this had been a real-life situation, the demonstrative would have been sufficient to secure reference and the noun would have been omitted.

Haut de page

Pour citer cet article

Référence électronique

Marion Durand, « What Does “This” Mean? Deixis and the Semantics of Demonstratives in Stoic Propositions », Methodos [En ligne], 19 | 2019, mis en ligne le 04 février 2019, consulté le 20 mars 2019. URL : http://journals.openedition.org/methodos/6023 ; DOI : 10.4000/methodos.6023

Haut de page

Auteur

Marion Durand

University of Toronto

Haut de page

Droits d’auteur

Licence Creative Commons
Les contenus de la revue Methodos sont mis à disposition selon les termes de la Licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Haut de page
  • Logo DOAJ - Directory of Open Access Journals
  • Logo UMR Savoirs, Textes, Langage
  • Logo CNRS - INSHS
  • Logo Université de Lille
  • OpenEdition Journals