Cet article est disponible en ligne à l’adresse : http:/ / www.cairn.info/ article.php?ID_ REVUE=RMM&ID_ NUMPUBLIE=RMM_ 0 54&ID_ ARTICLE=RMM_ 0 54_ 0 489 Early St oic Det erminism par Susanne BOBZIEN | Presses Universit aires de France | Revue de Mét aphysique et de Mor ale 2005/ 4 - n° 48 ISSN 0035-1571 | ISBN 2-1305-5372-9 | pages 489 à 516 Pour cit er cet art icle : — Bobzien S., Early St oic Det erminism, Revue de Mét aphysique et de Morale 2005/ 4, n° 48, p. 489-516. Distribution électronique Cairn pour les Presses Universitaires de France. © Presses Universitaires de France. Tous droits réservés pour tous pays. La reproduction ou représentation de cet article, notamment par photocopie, n'est autorisée que dans les limites des conditions générales d'utilisation du site ou, le cas échéant, des conditions générales de la licence souscrite par votre établissem ent. Toute autre reproduction ou représentation, en tout ou partie, sous quelque forme et de quelque manière que ce soit, est interdite sauf accord préalable et écrit de l'éditeur, en dehors des cas prévus par la législation en vigueur en France. Il est précisé que son stockage dans une base de données est également interdit. Early Stoic Determinism RÉSUMÉ. — Bien que, du second siècle avant Jésus-Christ au troisième après, les problèmes du déterminisme aient été quasi exclusivement traités dans le cadre de la question du Destin, le déterminisme avancé par Zénon et développé par Chrysippe fut largement élaboré dans des traités physiques, indépendamment de toute « doctrine du Destin ». Le déterminisme stoïcien était fondé sur la cosmologie stoïcienne et, pour le comprendre, il est indispensable de tenir compte de la conception stoïcienne des causes et des effets, les premières étant corporelles et responsables à la fois de la subsistance des choses et de leurs changements, les seconds étant incorporels et consistant en prédicats. À l’origine, le déterminisme stoïcien n’était pas présenté en termes de causes et possédait une dimension téléologique importante, liée à une doctrine des mouvements et des états naturels. Chrysippe mobilisa toutefois aussi sa conception de la causalité pour expliquer le déterminisme qu’il défendait, dont on peut montrer qu’il s’agissait d’un déterminisme au sens moderne du terme, causal et universel. Les dimensions téléologiques et mécaniques du déterminisme stoïcien étaient combinées dans la conception chrysippéenne du destin, qui introduit des éléments rationnels dans chaque cause particulière. ABSTRACT. — Although from the 2nd century BC to the 3rd AD the problems of determinism were discussed almost exclusively under the heading of fate, early Stoic determinism, as introduced by Zeno and elaborated by Chrysippus, was developed largely in Stoic writings on physics, independently of any specific “theory of fate”. Stoic determinism was firmly grounded in Stoic cosmology, and the Stoic notions of causes, as corporeal and responsible for both sustenance and change, and of effects as incorporeal and as predicates, are indispensable for a full understanding of the theory. Stoic determinism was originally not presented as causal determinism, but with a strong teleological element, in the context of a theory of natural motions, which makes use of a distinction between a global and an inner-worldly perspective on events. However, Chrysippus also employed his conception of causality in order to explicate his determinism, and can be shown to have maintained a universal causal determinism in the modern sense of the term. The teleological and mechanical elements of early Stoic determinism were brought together in Chrysippus’ conception of fate, which places elements of rationality in every cause. Revue de Métaphysique et de Morale, No 4/2005 490 Susanne Bobzien P H Y S I C A L A N D O N TO L O G I C A L F O U N DAT I O N S I start with some basic elements of Stoic physics and ontology that are required for an understanding of Stoic determinism 1. These are, in order, the active principle, causation, and changes, qualitative states and events. For the Stoics, the world is a unitary and continuous body, without any gaps ; it is located in the void, and it contains no smallest parts 2. Stoic physics is thus a continuum theory 3. The world is constituted of two principles, the active and the passive 4. The passive principle is called “matter” and “substance” ; it is amorphous and unqualified ; it possesses neither power of cohesion nor power of movement 5. The active principle is a power that is eternal and self-moved ; it is responsible for all form, quality, individuation, differentiation, cohesion and change in the world 6. Both principles are material 7. In physical terms, for Chrysippus, the active principle is pneuma or breath, which is a special combination of air and fire ; the passive principle is a combination of earth and water 8. (For Zeno, physically the active principle is creative fire.) The two principles form a complete blending (kra'si") both in the world as a whole, and in any object in the world 9 ; that is, they are completely co-extended, but they and their respective qualities are fully preserved in that mixture, and they are in principle separable again. The function of the active principle can be contemplated from two viewpoints : from a global perspective, which considers the whole cosmos as one unified entity ; and from the innerworldly perspective, which looks at particular objects and their interrelations. The active principle is responsible both for the cohesion, form and change of the cosmos as a whole, and for the individuation, cohesion, form, change and duration of the objects in the world. The individual objects are each held together (as the objects they are) by the active principle, which gives them a certain tension or tenor (e{xi"). Different objects have different complexity, owing to the complexity of their tenor. With increasing complexity, inanimate objects have tenor (in the specific sense) ; plants have nature ; non-rational animals have soul ; and rational beings have 1. A more detailed discussion of early Stoic determinism, its relation to freedom and moral responsibility, and later Stoic determinism can be found in Bobzien 1998a. For an in-depth study of Chrysippus’ theory of causes see Bobzien 1998b. 2. DL 7.140, 143, 150 ; Sextus AM 9.332. 3. For details on this point and for Stoic physics in general cf. Sambursky 1959, Bloos 1973 and Lapidge 1978. 4. DL 7.139. 5. DL 7.134, 139, 150 ; AM 9.75 ; Seneca Ep. 65.2. 6. DL 7.139 ; AM 9.75-6 ; Ep. 65.2 ; Stoic. rep. 1054a. 7. E.g. Aristocles in Eusebius Praep. Ev. XV 14.1. 8. Nemesius Nat. hom. 52.18-9 ; Comm. not. 1085c-d. 9. Alexander Mixt. 224-5. Early Stoic Determinism 491 reason as their highest organising principle. They each also have all the lower kinds of tenor. Physically, these kinds of tenor are pneuma of increasing purity or fineness. The finest pneuma, reason, is situated in beings of the highest order, i.e. rational beings, in the ruling part of their soul. The world as a whole is a being of the highest order, i.e. a rational being, too 10. Like human beings, in addition to tenor, nature and soul, it has reason. This rational organising principle, the reason of the cosmos, is called “god” or “god’s soul”, “nature”, or “ruling principle” 11. Sometimes it is placed in the aether, as the accumulation of finest pneuma in the cosmos. Sometimes all of the active principle, sometimes only its finest part is called the “reason of god”. Stoic determinism is causal determinism in one of its main aspects. However, we must beware of rash comparisons with modern theories of causal determinism. Modern theories may consider causes as events, facts, things, or properties, but mostly there is the assumption that cause and effect belong to the same ontological category, and that what is effect in one instance of causation can be cause in a subsequent instance. On this point, the Stoics differ : for them causes and effects belong to two different ontological categories ; this has various consequences for their account of determinism 12. Here are two Stoic accounts of “cause” : Chrysippus states that a cause is that because of which ; and that the cause is an existent thing and a body, <and that of which it is a cause is not existent and is a predicate ;> 13 and that the cause is “because”, and that of which it is a cause is “why ?” (Ecl. I 138.23-139.2). ... the Stoics say that every cause is a body which becomes a cause, to a body, of something incorporeal ; as for instance the scalpel, which is a body, becomes a cause, to the flesh, which is a body, of the incorporeal predicate “being cut” (AM 9.211). Every instance of causation involves at least three main factors, two corporeal, one incorporeal. (For reasons of convenience, I individuate instances of causation by assuming one such instance per effect.) 14 One corporeal is the cause, the other the object to which it is the cause, and at which the effect obtains. The effect, or that which is caused, is incorporeal and obtains at the second corporeal ; and it is a predicate 15. In cases in which different causes work 10. DL 7.142-3. 11. Plutarch Stoic. rep. 1052c ; Mixt. 225.1-2 ; Stoic. rep. 1053b. 12. For Stoic theory of causation in general see Frede 1980 and Bobzien 1998b. 13. Some emendation is required. My suggestion is based on the close parallels for Zeno (Ecl. I 138.5-16) and Posidonius (Ecl. I 139.7-8). 14. The individuation of effects is discussed below. 15. Ecl. I 138.15-16 ; 139.7-8 ; AM 9.211 ; Clement Stromata 8.9 96.23-97.1. 492 Susanne Bobzien together in one instance of causation the factors are multiplied. Moreover, cause and effects are relative ; that is, they are relative to each other, and inseparable : a cause is not a particular thing, but that thing insofar as it produces its effect 16. For the Stoics, thus, all interaction is interaction between bodies, with one body acting upon another. Incorporeals can neither act nor be acted upon 17. Hence effects, being incorporeal can neither bring something about, nor can they be acted upon. When the interaction between bodies is described as causation, it is described as the relation between one or more bodies as causes and the incorporeal effect as obtaining at another body. The most basic Stoic distinction between interactions of bodies is that between (I) sustenance (sustaining-something / being-sustained-by-something) and (II) change (changing-something / being-changed-by-something) 18. Correspondingly, the most basic distinction of causes is that between causes of qualitative states (scevsei") and causes of changes (kinhvsei") : On the level of causal explanation, the Stoics not only ask for the cause why something has changed, but also for the cause why something keeps on being what it is and in what state it is. The requirement of causes of (the sustenance) of qualitative states 19 is a consequence of the Stoic theory of the active and passive principle. The active principle is not only the sole source of change, it is also required for any qualitative state of an object to continue, and for any object to continue to be that object. The active principle is regarded as the cause of both changes (it brings them about) and qualitative states (it keeps them up) 20. The active principle is also the main reason why for the Stoics all causes are active causes, and Aristotelian formal, material, and final causes, as well as mere necessary conditions, do not qualify, but would merely count as circumstances 21. However, the active causes combine some of the functions of Aristotelian formal, final and effective causes, and are involved equally in causation of change and of qualitative states. A main characteristic of causes of qualitative states is that they are simultaneous with their effect. The most prominent causation of qualitative states is that involving a cohesive cause (sunektiko;n ai[tion) : the cohesive cause is that portion of pneuma in an object that is – and insofar as it is – responsible for 16. Cf. e.g. Sextus PH 3.25, AM 9.207. 17. Nemesius Nat. hom. 21.6-7 (Morani) ; AM 8.263 ; Comm. not. 1085b. 18. All sustenance and change is two-faced in this way : it depends on one’s chosen perspective whether one describes one body as affecting another, or that other body as being affected by the first ; cf. Simplicius Cat. 306.14-15. 19. I render scevsi" by “qualitative state”, in order to remind the reader that this is a kind of effect, i.e. something that requires a cause for its sustenance. 20. Cf. e.g. Stoic. rep. 1054a. 21. Cf. Ep. 65.4, PH 3.14. Early Stoic Determinism 493 the object being the object it is. This seems to be the same pneuma as the tenor in each object. Thus for the ongoing existence of an object, the presence of a cohesive cause is a necessary condition. The main characteristic of causation of change is that a co-operation of two causal factors is required, of which one has to be antecedent to its effect. Since for the Stoics all causes are corporeal, some elucidation is needed for what it means for a body to be such an antecedent cause. For in most cases of causation this body will exist before, during, and after the effect obtains. “Antecedent” can thus not refer simply to the time at which the body exists. One needs to take into account that qua cause the body actively contributes to its effect, and that causes are relative. Thus the body is the antecedent cause only insofar as it actively contributes to the effect. I thus suggest that, as a minimal condition, c is an antecedent cause of an effect e, if the period of time at which c is active in contributing to e precedes, at least in part, the period of time at which the effect obtains. The role of causation of qualitative states in Stoic theory of determinism is often neglected. This is so partly because it has no equivalent in later physical theories, and partly because it is the causal predetermination of changes that was the primary matter of contention, not the causal simultaneous determination of states. But the pair of nouns “change / qualitative state” together with the corresponding pair of verbs “to change / to be in a qualitative state” appear frequently in Stoic philosophy, and they denote a standard Stoic distinction 22. The individual changes and qualitative states are of special interest, since they are precisely those entities that are effects of instances of causation. What is their ontological status ? We can be confident that they are incorporeals, since they are effects, and effects are incorporeal. If they are incorporeals, they must be void, time, place or “sayables” (lektav) 23. Of these, they can reasonably only be sayables. This tallies with the account of effects as belonging to the “things that can be predicated” or “predicates”, which are a subclass of the sayables. But matters are more complex. First, the definitions of changes and qualitative states : Chrysippus had, it seems, a narrower and a wider concept of change. He defined change as “alteration concerning space, either to the whole <body> or to a part of it” 24. Here change is locomotion. Another definition is “alteration concerning space or 22. Cf. Stoic. rep. 1050c-d, 1056c ; Comm. not. 1076e ; Ecl. II 73.1, 82.11-17 and 95.6-8, DL 7.104, Cicero Fin. III 33 ; Origen Orat. 2.368, Plotinus Enneads III 1.7 ; Ecl. I 166.24-167.14 ; Simplicius Cat. 212-13. 23. AM 10.218. 24. Ecl. I 165.15-17, cf. AM 10.52. 494 Susanne Bobzien shape <either to the whole body or to a part of it>” 25. In this account, what is at issue includes any change or process. For “shape” (sch'ma) stands for all qualitative features of an object 26. This wider definition is the one of primary relevance for the theory of determinism. Chrysippus also produced two definitions of “continuity” (monhv) : “absence of change of a body”, and “qualitative state of a body, concerning the same features and in the same way, now and before” 27. The qualitative states, here, being a type of continuity, must be incorporeal. This is confirmed by Apollodorus’ 28 definitions of “qualitative state” as (I) “continuity/coherence (sunochv) <of a body> concerning space or form” and (II) “the being in such a state” 29. The second definiens defines an incorporeal. And it is qualitative state as equivalent to being-in-a-qualitativestate, and change as equivalent to changing, which are incorporeals, predicates, and the effects of causation. Qualitative states, as is clear from the definition, have duration. So, too, do changes : In line with Stoic continuum theory and the indefinite divisibility of time, Apollodorus states that for every change there is a change that is part of it ; and for every qualitative state there is such a state that is part of it 30. Recalling the definitions of qualitative states, we can conclude that qualitative states are homogeneous : every state that is a temporal part of a state is of exactly the same kind as the state of which it is a part. Changes, though, are not all homogeneous : the fact that for every change there is a change that is part of it does not entail that the change that is part of it is of the same kind as the change of which it is a part. Rather, some Stoic changes are homogeneous, some are not. For further understanding of the Stoic concepts of change and qualitative state we have to follow the trail that they are predicates (since they are effects and effects are predicates). The Stoics define a predicate as “what is asserted of something, or a thing (pra'gma here, as often, meaning ‘sayable’) attachable to some thing or things” 31. Examples for predicates are “to walk” and “to be alive”. Predicates can be actualised, i.e. subsist at something or hold of some25. Ecl. I 165.17-18. The supplement is taken from Chrysippus’ previously quoted narrower definitions of change ; it finds justification in its occurrence in Apollodorus’ parallel definition, Ecl. I 166.24-6. 26. Cf. also AM 9.75, Comm. not. 1054a. 27. Ecl. I 165.19-21. 28. The Stoic Apollodorus was presumably a pupil of Antipater (Index Stoicorum col 53.7). The excerpt on qualitative states and changes from his Handbook of physics in Ecl. I 166-7 overlaps in large parts with Chrysippus’ accounts (Ecl. I 165-6). As there are no inconsistencies with earlier Stoic theory, I assume that the excerpt generally squares with Chrysippus’ view. 29. Ecl. I 166.26-7. 30. Ecl. I 167.9-14. Cf. also, for Chrysippus, Ecl. I 106.8-9. 31. DL 7.64. Early Stoic Determinism 495 thing, or can be not actualised. When actualised, they are called “actualised predicates” or “attributes” (sumbebhkovta) 32. The being actualised of a predicate is best explained by example : the predicate “to walk” holds of me, and is thus actualised at me, precisely when I am walking 33. Whenever a predicate is actualised, it is so for a period of time, tm – tn. The same predicate, “to walk”, can be actualised repeatedly, e.g. (at Thea) from 9am-10am and again (at Thea) from 3pm-5pm today ; it would then presumably also be actualised (at Thea) from 9.30am-9.45am, etc. 34. We can now draw the connection between predicates and changes and qualitative states as follows : whenever a predicate is actualised, one can discern an actualisation of that predicate. So, in place of talking about a predicate’s being actualised many times, one can talk about as many actualisations of the one predicate. Thus the various cases of walking are individuated, and what one arrives at are individual changes and qualitative states. For instance in our example we have an actualisation of the predicate “to walk” from 9am-10am, from 3pm-5pm, and from 9.30am-9.45am, i.e. three different actualisations. Each has duration. The same holds for predicates that are concerned with qualitative states, e.g. “to be alive”, actualised at Dio ; there were in fact indefinitely many actualisations of that predicate at Dio, during Dio’s lifetime. Changes and qualitative states can thus be understood as a subclass of all actualisations of predicates : namely the subclass of those things that are effects. For instance, in the above example of “to walk”, actualised at Thea, we specified three motions. And in the case of “to be alive”, actualised at Dio, we had indefinitely many qualitative states. But is the actualisation of a predicate a predicate ? Asked in this way, the answer seems “no”. However, the phrase “actualisation of a predicate” was something I introduced for reasons of explication only. In fact, we can just as well say : the motion of Thea’s walking from 4pm-5pm is the predicate “to walk” actualised (at Thea) from 4pm-5pm. Thus individual changes and qualitative states can be understood as something like phases of predicates, more precisely, phases of predicates-while-they-are-actualized. If we look at changes and qualitative states qua effects, we obtain similar results. Effects are predicates. And effects, as dealt with by the Stoics in the context of determinism, are particulars. Thus, effects also must be predicates actualised at some thing from tm – tn. And indeed it is testified, at least for Zeno, that effects are 32. Ecl. I 106.20-3. 33. Ecl. I 106.20-3. 34. If the time span becomes very small, perhaps no longer the predicate “to walk” but some other predicate of change would be actualised at those times (at Thea), e.g. “is moving her right leg forward”. 496 Susanne Bobzien sumbebhkovta, that is actualized predicates 35. Thus changes and qualitative states, qua effects, are tokens, not types ; they obtain in the world during some time tm – tn 36. Besides “change” and “qualitative state”, there is another expression prevalent in the debate over Stoic determinism : this is “that which happens” or “events” (gignovmena), derived from the verb “to happen/to be” (givgnesqai). In Stoic sources “event” and “change” are sometimes used interchangeably 37. However, “event” is equally used to cover both changes and qualitative states 38. In this second sense, events make up that important subclass of all actualizations of predicates that are effects. The Stoic concepts of events, changes, and qualitative states share elements of both modern concepts of events and facts. The above exposition is however not meant to represent a Stoic theory of events or facts, as they consciously developed it. Rather, I have tried to make apparent several features of their concepts of change, qualitative state, and event which follow from theories we know they had about other things : causation, effects, sayables, predicates and actualised predicates. As many modern philosophers build their theories without a fully worked out theory of meanings, facts, and events, and indeed sometimes mix up their areas, so did the Stoics. TELEOLOGICAL DETERMINISM DEFINED The only surviving passages in which Chrysippus systematically sets out and justifies his determinism are a number of quotes from his first book On Nature, preserved by Plutarch. This sequence of short passages shows how Chrysippus’ determinism grows out of the basic assumptions of Stoic cosmology and is thus anchored in Stoic physics : (1)... in his first book On Nature he said this : Since the organisation of the universe proceeds in this way, it is necessarily in accordance with this organisation that we are in whatever qualitative state we may be, whether contrary to our individual nature we are ill or maimed or have become grammarians or musicians.... 35. Ecl. I 138.15. 36. This is not to deny that the Stoics ever talked about motions, qualitative states, and effects on a general level as well. On the contrary – it is likely that when the Stoics talked about general motions, qualitative states, or effects, they precisely took them to be predicates (kathgorhvmata). 37. e.g. Cicero De fato. 20 ; Alexander Fat. ch.13. 38. e.g. Stoic. rep. 1049f-1050d. Early Stoic Determinism 497 (2) We shall on this principle state similar things both about our virtue and about our vice and generally about skills and the lack of skills as I have said.... (3a)... that nothing at all, not even the smallest, is in a qualitative state or changes otherwise than in accordance with the reason of Zeus, which is the same as fate. (4) For since the common nature stretches into all things, it must be the case that everything that happens in any way whatsoever in the universe and in any of its parts will have happened in accordance with that nature and its reason in unimpeded sequence ; for neither is there anything to obstruct the organisation from outside nor can any of its parts change or be in any qualitative state except in accordance with the common nature. (5)... everywhere, but mainly in his physical books he has written : that there happen many obstacles and impediments for particular natures and (their) changes but none for the nature of the universe 39. Passages (1) to (5) appear all to be part of one argumentative context : the thesis Chrysippus wants to back up is that every change and every qualitative state of any object is in accordance with the common nature of the world 40. The above-mentioned distinction between global and innerworldly perspective is essential for Chrysippus’ reasoning, and is manifested in his distinction between common and individual nature. The common nature is the same as the active principle or god (3a, 3b) 41. This universal nature extends into all things (4) 42. It is responsible for the organisation of the universe as a whole (4, 1, 3b), and it is unhindered (4). On the intra-cosmic level naturally things are more complex. All objects in the world have their individual natures (1), and these go together with qualitative states and changes, which are in accordance with the object’s individual nature ; e.g. for human beings, to be in a state of health and to move in accordance with reason are qualitative states and changes in accordance with nature 43. But there are also counter-natural changes and qualitative states, i.e. those contrary to an object’s individual nature : being ill or maimed are examples for states (1), bodily movements that go beyond a person’s impulse for movement 44 ; and we can deduce from (1) that there are changes of an object that are externally forced, like being maimed. Counter-natural states can be the result of counternatural changes. But they can also be frustrated natural changes : a tree’s growing 39. Stoic. rep. : (1) 1049f-1050a ; (2) 1050a ; (3) 1056c (cf. Stoic. rep. 1050a and Comm. not. 1076e) ; (4) 1050c ; (5) 1056d. 40. Cf. Philodemus De pietate col. 11.32 col. 12.1 : ND I 39. 41. Cf. Philodemus, loc. cit., Cicero, loc. cit. 42. The same verb, “to extend” (diateivnein) is used to describe how the pneuma of the soul stretches through an animal’s body. E.g. Ecl. I 368. “Aetius” IV 21, Doxographi Graeci 410-11 ; cf. also DL 7.138. 43. PHP 4.2.12, quoting from Chrysippus. 44. PHP 4.2.16-18, quoting from Chrysippus. 498 Susanne Bobzien can be prevented by a roof that is “in the way”. That is, on the innerworldly level there can be impediments that interfere with the natural (and also with counternatural) changes of individual objects. An impediment is external to the object (or part of object) whose changing is prevented. Changes and qualitative states that are not natural have the reason of their occurrence in the fact that there is a multitude of objects in the world and that they can get in each other’s way, when performing their natural changes. Both counter-natural changes (i.e. changes by force) and counter-natural states (prevention of an object’s natural changes or results of counter-natural changes of an object) are the result of such clashes. The interplay of the objects leads to the result that one object – in moving naturally – hinders another from moving or from being in a certain state ; and one object can by force move another object, or change the state it is in, or even destroy it. Chrysippus’ reasoning in his first book On Nature is then based on the relation between cosmic and intra-cosmic perspective. His main thesis is that “every qualitative state and every change in the world is in accordance with the universal nature”. Our passages suggest that Chrysippus put forward the following two reasons to back up this thesis : First, nothing can obstruct or destroy the organisation of the universe from outside, for the reason that there is nothing (no thing) external to of the universe – only the void. The natural movement of the universe as a whole can thus not be thwarted 45. (Why can the impediments on the level of individual objects not be regarded as obstacles to the nature of the universe ? This is Plutarch’s criticism 46. We should expect Chrysippus’ reply to be (a) that the common nature extends into all these obstacles, too ; (b) that all obstruction can in the end be reduced to natural changes of some things (that are part of the world) ; and (c) a teleological reason, viz. that these obstructions lead to the realisation of the best possible world.) Second, nothing in the world can be in a qualitative state or change otherwise than in accordance with the universal nature, since the universal nature is “allembracing” : it includes the individual natures, in that it extends into the individual objects, making up their individual nature (4). Although it would make sense, say, that my individual nature goes against yours, it makes no sense to say that my individual nature goes against the universal nature, for my individual nature is part of the one universal nature 47. 45. Cf. ND 2.37-8. 46. Stoic. rep. 1056d. 47. What about the passive principle, matter ? In the same first book On Nature Chrysippus applies the same distinction of global and innerworldly perspective to it, too. Whereas the matter of the individual things both increases and decreases, as result of the workings of the active principle, the matter of the whole world remains constant (DL 7.150 ; Ecl. I 133.6-11). Early Stoic Determinism 499 Combining the two reasons, we can see that the thesis that every state and every change is in accordance with the one universal nature (and its organisation) is a consequence of the Stoic conception of the world as a unity which is held together by the one active principle. The teleological aspect of the active principle is taken into account in our passage by the references to the organization of the world (1), (4) and to the reason of the universal nature (3), (4). There can be little doubt that Chrysippus proclaims some sort of universal determinism in these passages from On Nature. His formulations in (1), (3), and (4) suggest that the universal nature leaves no room for alternative developments of the world. There is exactly one course of events that is in accordance with the rational universal nature, and that is the course of events which is the actual one. But what sort of determinism is this ? (I) As to its scope, it is universal : first, changes and qualitative states are equally determined (by the rational universal nature). Second, every qualitative state and every change is included. Given Stoic physics is a continuum theory, changes and qualitative states can be indefinitely small ; but however small, they are included in the universal nature. There is no room for quantum leaps ; nor is Nature or god concerned with the “weighty” events only. Later Stoics illustrated this point with the raising of one’s eyebrow, stretching out of one’s finger or turning one’s head as examples for events that are trifling but predetermined nonetheless 48. (II) The world, as it is determined, is not chaotic, or like something produced by a random generator. Rather, it is an ordered and organised whole (1), (4). (III) This organised whole has not evolved randomly (“order from chaos”), but its development follows a rational principle of organisation : the reason of nature or Zeus (3), (4). (IV) This rational organising principle works completely immanent to the world, and also from the inside of each individual object. (V) There is no universal innerworldly harmony of the kind that every object can realise its individual nature unhindered, and unforced. Rather, the rational organising principle works like this : the individual nature in each object provides the object with its characteristic qualities and activities, which in favourable circumstances it will display and perform. However, not all objects realise their individual nature in all respects. Rather, some objects prevent others from performing their natural changes, and some objects force others to go through certain counter-natural changes. The world is such that the objects are, as it were, left to battle the conflicts out between themselves. Yet – from the cosmic perspective – the way this happens does not include any element of chance ; 48. Fat. 175.6-13. 500 Susanne Bobzien for is in accordance with the reason of the world, which works from the inside of these objects. (VI) Chrysippus expressly mentions our virtues and vices as among the qualitative states that cannot be in any other way than they are (1), (2) ; and there is no hint that he considered this problematic. Virtues and vices, being dispositions (diavqesei"), are part of the portion of active principle in us. This picture of universal determinism is far removed from modern theories of mechanical, causal determinism, in particular on two counts ; first, the overall picture is strictly teleological : there is a rational organisation of the world, and every state and change conform to the rational organising principle. Second, causation is not mentioned in any of the passages, nor is there any idea of empirically detectable regularity or individual laws of nature that governs all events. The changes in the world – like the qualitative states – are defined in terms of the natural changes and states of the objects in the world, and of the counter-natural changes and obstructions that are the consequence of many different objects trying to realise their natural changes. What makes an object change “naturally” is primarily the active principle in that object, although such changes always presuppose some external stimulus. One could express the Stoic theory in terms of “laws of nature” : but these would be laws concerning the individual natures of individual objects. They would have the form : an object with the individual nature n will perform changes of type m as long as the circumstances are favourable, i.e. as long as it is (appropriately prompted and) not prevented by external circumstances from doing so. Such “laws of individual natures” connect objects that have certain qualities with certain changes, in line with Stoic theory of causation 49. They are however not suitable to define determinism – not so much because they do not connect events with events, but rather (a), since they do not take into account the circumstances in which the objects find themselves, and (b), since they do account for natural changes only. Precise prediction of the future on the basis of such laws, even if one knew them all, is not possible. But it would be mistaken to infer from this that Chrysippus’ determinism was not a causal determinism in which efficient causes play a central role. In order to see how Stoic determinism and Stoic theory of causation fit together, we have to turn to another passage in Plutarch. 49. But it is unlikely that the Stoics themselves thought of laws of such a kind as governing the natural movements, or that they would take recourse to such laws for explaining such movements. Rather, the rational pneuma in every natural object will have been thought to contain the relevant “information”, embodied in the specific state of tension of their pneuma, and this will make the object perform its natural movements, if prompted. The regularity results from the fact that natural objects of a certain kind all have the same kind of “information” stored in their pneuma. Early Stoic Determinism 501 C AU S A L D E T E R M I N I S M D E F I N E D In antiquity, Stoic physics stands out not so much because it is a deterministic system, but because it contains a worked out theory of universal causal determinism. No early Stoic exposition of this theory has survived, but a passage in Plutarch gives us the basic information we need. The passage is a summary from a Chrysippean work ; it consists of an argument by some opponents of the Stoics (the “arguments from indistinguishables”) and Chrysippus’ reply. I start with these. Here is Plutarch’s presentation of the opponents’ argument from indistinguishables : (1) Some philosophers, believing that they bring about release for the impulses from being forced by external causes, construct in the ruling faculty of the soul some adventitious motion (2) which becomes evident best in the case of indistinguishables. (3) For when it is necessary to take one of two things, (4) when the two are of equal power and are in the same state, (5) <and> when no cause leads to either of them, (6) since it in no way differs from the other, (7) <then> the power of spontaneity itself in the soul, by taking from itself an inclination, (8) cuts through the puzzle. (Stoic. rep. 1045b-c) The philosophers in this passage try to justify the existence of spontaneous motions by singling out a subclass of the situations in which such changes are assumed to occur and in which their introduction appears most plausible. These are situations in which an agent encounters some indistinguishable alternatives between which to choose (2). The philosophers’ argument for their point is crammed into (3) – (8) 50 : They make the assumption that if two alternatives are indistinguishable, then no (external) cause will lead the faculty of impulse to either (6, 5). They also presuppose that at least sometimes an action ensues. From these two assumptions they seem to infer that in the cases in question impulse and action occur without an external cause. The power of spontaneous motion in the mind is thus introduced to explain how the agent’s predicament was solved (8). This power takes from itself (i.e. not from outside) an inclination for one of the two alternatives. What sort of power it is we are not told. The argument neither states nor entails motion without a cause (with the Hellenistic concept of corporeal causes). Rather, the passage propagates some kind of “agent-causality”, based on a self-moving force in the agent’s soul (7). The 50. I pass over various ambiguities. 502 Susanne Bobzien spontaneous motion is introduced not as causally undetermined, but as “unpredetermined”. Plutarch summarises Chrysippus’ reply thus : (9) Chrysippus, speaking against these, as people who violate nature through the uncaused, (10) adduces in many places the die and the scales and many of the things that cannot take now one fall or inclination, now another (11) without some cause and <without> there being a difference either concerning just the things themselves or concerning the external circumstances ; (12) for, he thinks, the uncaused is completely non-existent, and so is self-motion, (13) and in the case of these things which some people invent and name “spontaneous”, (14) concealed causes sneak in and, (15) without our noticing it, they lead the impulse to one of the two alternatives. (Stoic. rep. 1045c.) Chrysippus proceeds as follows : First he states that the introduction of the power of spontaneous movement amounts to the introduction of uncaused motion (9). Then he gives an alternative explanation of what happens in the situations of indistinguishables (10) – (15). Since the intra-psychical events are not open to direct investigation, instead of a direct account, he presents an analogy from “everyday physics” to the situations of indistinguishables 51. The dice and scales analogy is given in a very compressed form. But we can secure from the text a few general points about the situations at issue in the analogy : • The analogon looks at the same object (a die, a pair of scales), or at two identical looking objects, in two spatio-temporally distinct situations. • In these two situations the object, or the two objects, react in noticeably different ways. • The reactions are characteristic movements of the objects at issue : falling on one of its sides (die) ; inclining to one side (scales). • From the fact that the situations are meant to be in some way analogous to those of the indistinguishables, we can infer that in each case the two starting situations must at least prima facie look indistinguishable in all relevant respects. (For example, a die, thrown twice, under seemingly identical circumstances, comes once up with one side, then with another.) Chrysippus then explains what is really going on in these cases : the seemingly indistinguishable starting situations are in fact not completely the same : there is a difference either in the object (die, scales) or in the surroundings, which is responsible for the difference in the outcome. And – given the assump51. Chrysippus and Zeno used this method repeatedly, cf. Gellius NA 7.2.11, Cicero De Fato 42-3, Acad. 2.145, PHP 4.2.14-18. Early Stoic Determinism 503 tion that the situations appear alike in all relevant respects – this difference must be hidden to the observer. Thus Chrysippus clearly acknowledged the fact that tiny variations in the antecedent situation can lead to – comparatively – large changes in the outcome. Since Chrysippus intends to argue by analogy, we can assume that he thought that his audience accepts that there are such unnoticeable differences present in the case of dice and scales. For only then can the analogon be used to elucidate Chrysippus’ explanation of what happens in the human mind in the situations of indistinguishables. Chrysippus draws the analogy in (13) to (15) : as in the case of dice and scales, in the case of indistinguishables, the reactions are neither uncaused nor spontaneous ; rather there are factors concealed to us which are responsible for the agent’s impulse to go for one and not the other alternative. In parallel with the case of the dice and scales these hidden factors should be either in the person’s mind, or in the surroundings. We can assume Chrysippus to have concluded that hence the so-called indistinguishable alternatives do not prove the existence of spontaneous – and thus uncaused – self-motion. So far Chrysippus’ reply 52. How is Chrysippus’ argument related to causation ? We know from elsewhere that the causal factors responsible for the movements in the case of the die would be the person throwing the die and the nature (i.e. shape, etc.) of the die 53. But Chrysippus, in his explanation of the analogon, says there is a difference either in the object (the die) or in the surroundings (which should include more than just the person who throws the die). That means that there is not necessarily a difference in the causes in the Stoic strict sense. There is a difference somewhere in the overall situation – perhaps in the causal factors, perhaps in the circumstances. And this difference explains the difference in outcome. This is why we have to read (11) as saying that dice cannot fall now in one way, now in another “without some cause and (without) there being a difference either concerning just the things themselves or the external circum52. A comparison between the two arguments shows that the controversy between Chrysippus and his opponents is at base metaphysical : the opponents, starting from the premises that there is no difference in the antecedent data and that one option is actually chosen, conclude that there is spontaneous “self-movement”. Chrysippus, starting from the premises that there is no uncaused motion (and that spontaneous self-motion implies uncaused motion) and that one option is chosen, concludes that the opponents’ assumption of ontological indistinguishability in the antecedent data is false ; there are differences, but they are hidden ; the indistinguishability is merely epistemic. Chrysippus thus rightly rejects the validity of his opponents’ step from epistemic to ontological indistinguishability. But he in turn can conclude that there always is a hidden difference only by virtue of his assumption that there are no uncaused motions. And this assumption is as metaphysical as the one that there are ontologically indistinguishable situations. 53. Cf. NA 7.2.11, Cicero De Fato 42-3. 504 Susanne Bobzien stances”. That is : there are causes for every fall of a die ; and if there are different outcomes in comparable situations, this is due to a difference either in those causes or in the surroundings. When Chrysippus draws the analogy, he emphasizes the presence of causes : “and in the case of these things which some people invent and name ‘spontaneous’, concealed causes sneak in.” ((13), (14) in response to (5)). This is so, because the situation of indistinguishables is assumed to leave the agent in a stalemate, i.e. inactive, if there is neither a difference in overall situation, nor a spontaneous motion. Thus, in order for something to happen in such situations, there must be an active causal factor involved – whichever way the agent decides. What does Chrysippus’ argument tell us about Stoic determinism ? The Plutarch passage allows us to extract two types of principles of universal causality which were part of Chrysippus’ theory. First, Chrysippus assumes a principle “the uncaused and the self-moved are non-existent”. Variations of this principle are recorded for him elsewhere. Thus Cicero reports that Chrysippus maintained that there is “no change without cause” and that “nothing can happen without a cause” 54. This principle is, in the first instance, about changes or events ; it links corporeal objects (the causes) with motions. (The formulation in Plutarch does not mention motion, but the whole argument is about the motions of the soul.) I call it the “General Causal Principle” : (GCP) Nothing happens without a cause. Elsewhere, Chrysippus backed up this principle with the Principle of Bivalence. Here, he seems to justify it by the idea that uncaused motion would violate nature (9). As long as the General Causal Principle is not further specified, it does not, within Stoic philosophy, entail determinism, since causes are not events, but bodies, considered in relation to the event they bring about. Thus the principle does not preclude that a change is not fully determined by its causes. Nor does it rule out that in relevantly similar situations an object is once the cause of one kind of change, once of another. That is, it is not ruled out that there is an element of spontaneity in the (thing that is the) cause. The General Causal Principle does not guarantee universal regularity or uniformity between cause and effect. But such uniformity is the hallmark of most modern theories of causal determinism. 54. De Fato 20 and Div. II 61. And Galen writes : “For in this way they <i.e. Chrysippus and his adherents> would concede some uncaused change, something they commend us to beware of.” (PHP 4.4.35-6). Early Stoic Determinism 505 This lack of specification cannot be made up for by the fact that we have determinism of some (teleological) kind secured for Chrysippus elsewhere (cf. p. 496-500). For it does not follow that, if we combine this determinism with the General Causal Principle, we obtain universal causal determinism. In order to show that Chrysippus’ determinism was universal causal determinism, we need to show that he holds a different kind of physical principle. We find such a principle underlying our Plutarch passage. The kind of principle I have in mind belongs to the family described by catch-phrases as “same causes – same effects” or “like causes – like effects”. I dub such principles “specified causal principles”. We are nowadays all familiar with such principles, and indeed many consider them trivial. However, this “triviality” is based on the fact that in everyday life we have absorbed a certain physical world view, thinking of causes and effects as events, and as governed by “laws of nature” which somehow connect types of causes and types of effects. And we seem to hang on to this kind of idea of the world fairly undisturbed by the fact that it has little to do with modern physics – or philosophy of causation for that matter. However, this world view has not always been the prevalent one. Someone had to come up with the idea of such a specified causal principle first ; and evidently, with a concept of cause as corporeal, any version of such a principle will have to look different than the modern ones. Let us return to Chrysippus’ reply to the proponents of spontaneous motions. It is plain that he maintains the General Causal Principle in his reply (12). But how is a specified causal principle involved in his reasoning ? One way of understanding specified causal principles of the kind “like causes, like effects” is that they imply the existence of a plurality of particular empirical causal laws which state some universal regularities, and which allow us to say something like “effect e happened because cause c preceded”, based on reasoning of the kind “ceteris paribus, whenever a cause of type C, then an effect of type E ; hence in this case e (which is of type E) followed, since c (which is of type C) preceded, and cetera pariba.” But this is not what we find for Chrysippus. In our passage, what Chrysippus intends to show is that in the case of any one motion (including those in situations of indistinguishables) causes and circumstances together determine the motion in every detail. To show this, he argues by analogy from two (or more) similar motions. The analogy makes use of the following type of situation : We have two sections of the spatio-temporal universe, each consisting of a starting situation s which includes an object o, and a subsequent change of o, which ends in a resultant state r. The two starting situations, as far as they are considered pertinent to the changes, are epistemically indistinguishable, but temporally and/or spatially distinct. The subsequent changes however differ from each other in that they lead to noticeably dissimilar 506 Susanne Bobzien resultant states. Chrysippus inferred from this – and takes his audience to agree with him – that the two epistemically indistinguishable starting situations are in fact ontologically distinct. So it looks as if Chrysippus has some kind of specified causal principle on the side of the analogon, roughly “difference in effects, difference in overall starting situations” ; or more precisely (SCP-analogue) When from two seemingly indistinguishable starting situations two different effects ensue, then these situations were ontologically distinct : either in the object or in the surroundings there was a factor that differed. This looks like the converse of the principle “like causes – like effects”. More accurately, it is contraposed to “no difference in starting situations – no difference in effects”. Why has Chrysippus the converse of what we usually find ? The answer could be that when formulating the principle in contraposition, starting with the effects in the antecedent, as is implied by our text, Chrysippus conveniently bypasses the difficulty of having to determine what is relevant to the starting position. For in the form beginning with the starting situation, a notorious difficulty is to determine what counts as part of the starting situation : First, the factors can be indefinitely many, owing to the Stoic continuum theory. Then, given the Stoic theory of sympathy, i.e. of the physical influence of everything on everything, the relevant starting situation may well include the whole universe ! Does Chrysippus base this principle on the assumption of the existence of particular empirical causal laws ? First, here we have to distinguish between what we would consider empirical causal laws, and what Chrysippus would. For the modern idea of uniformity between certain kinds of causes and certain kinds of effects, where both causes and effects are events, is very different from Chrysippus’ ideas of active causation, and corporeal causes. And our passage certainly does not entitle us to assume that Chrysippus had empirical laws built on his own concept of cause in mind, since in SCP-analogue the relevant differentiating factor in the comparable situations is either cause or surrounding. And if he had had something of the modern idea in mind, it would not have been qua causal laws. Second, neither from the SCP-analogue nor from the Plutarch passage itself can we infer that Chrysippus thought of the world as exhaustively determined by a plurality of particular empirical laws, short of laws that govern the entire situation of the world at a time as starting situation (and which I would not call “particular”). However, these points are in any case irrelevant, since principle SCP-analogue is not the one Chrysippus is after. For the empirical situation of scales and dice is only the analogon in his analogy. Chrysippus argues from this analogon, i.e. Early Stoic Determinism 507 from two empirical spatio-temporally different situations to one situation together with counterfactual reasoning about this one situation. First, the text implies that he elaborates from the empirical analogue of his Specified Causal Principle SCP-analogue as follows : in starting situation s1 there is some factor f1 that is responsible for the fact that e1 and not e2 (e3,...) happens, and in starting situation s2 there is some factor f2 that is responsible for the fact that e2 and not e1 (e3,...) happens. f1, f2,... need not be causal factors ; for instance, f1 can be the presence of some causal or some hindering element, in the object or the surroundings, and f2 the absence of that element. Chrysippus then draws the analogy to one starting situation in which there appear to be two equally likely outcomes, i.e. to the situations of indistinguishables. These are situations of human choice, but for the present this fact is immaterial. Chrysippus’ point, as relevant here, is this : (SCP-indist.) If in one starting situation s it looks as if there are two different but equally likely outcomes e, e*, and e occurs, then there is a factor f in s such that because of f e (and not e*) occurred ; that is, had instead of f f* (≠ f) been present, then e* (and not e) would have occurred. The empirical case of two numerically different starting situations was only an analogon. Here, now, we can see that Chrysippus’ Specified Causal Principle is based on counterfactual reasoning about one and the same situation. Particular empirical causal laws play no role in it. The point is made, as before, negatively, by introducing one factor that differs from the overall starting situation (this time this factor is counterfactual : “If the effect were different, there would have been a difference in the starting situation”). Accordingly, again there is no need to describe a “causally relevant section” of the world 55. By generalising from the situations of indistinguishables to all movements, we can formulate Chrysippus’ (unrestricted) Specified Causal Principle. If we include Chrysippus’ concept of cause and the General Causal Principle (as (i)), we obtain : (SCP) For every change e (i) there are causal factors c1... cn which are actively responsible for e, and (ii), for any possible alternative change e* to e there is a factor f in the starting situation s of e, such that because of f e (and not e*) occurred ; i.e. for e* (and not e) to occur there would have to have been some f* (≠ f) instead of f in s. 55. We have no evidence of converse formulations, starting with similarity of antecedent situations, rather then difference in resulting states, before the second century AD ; cf. Fat. 192.22-4 and Nat. hom. 105.18-21. 508 Susanne Bobzien Add to this the fact that for the Stoics motions are indefinitely divisible, and you can see that here we have a specified causal principle that defines a fully deterministic system. This principle does not involve particular empirical causal laws, nor does it rely on comparable empirical situations with comparable effects – except as analogy. Rather, it is concerned with the total actual state of the world or an unspecified part thereof. Backing of the principle would not be expected from empirical quarters. (Accordingly, no element of predictability comes in.) Instead, we would expect the justification of the principle to come from cosmological, theological, or teleological theory. Nonetheless, with this principle SCP, we have (in (ii)) a full formulation of causal determinism in the modern sense. For the principle entails that every change is fully determined by the antecedent situation. If the outcome were any different, in however minute a detail, then the antecedent situation would also have been different. However, remember two points : first, Chrysippus’ determinism is “stricter” : in addition to uniformity (of whole world-state starting situations and effects), there is active causation – in the Stoic sense – involved in (i), which moreover is regarded as the main determining factor. Second, what we may consider as “causality” in this account of determinism, was not regarded as causality by Chrysippus, and if he had considered his determinism as causal, then not because of (ii) in SCP. The question that remains is : how does Chrysippus’ causal determinism as tentatively captured by SCP combine with the teleological determinism proclaimed in the passages discussed p. 496-500 ? For an answer, we turn to the early Stoic concept of fate. TELEOLOGICAL AND MECHANICAL A S P E C T S C O M B I N E D : FAT E For Chrysippus, the causal and the teleological aspect of determinism are neither alternative nor mutually exclusive explanations of the world. Rather they complement each other in one comprehensive theory. The teleological part of the theory leaves it undetermined in which way exactly the rational principle makes the world develop in accordance with it. This is where causation comes in. The early Stoic theory of fate gives us some indication how this works. Chrysippus stated that fate was the same as the active principle, as god or Zeus, as providence, as the nature of the universe ; as the reason and the will of Zeus, and as the reason and the cause of the world 56. Most of these identi56. Stoic. rep. 1049f, Philodemus De Pietate col. 11.31-col. 12.2 ; Stoic. rep. 1056c, Comm. not. 1076e, cf. Diogenianus in Eusebius Praep. ev. 323.10-11 ; Ecl. I 79.5-7, Stoic. rep. 1055e. Early Stoic Determinism 509 fications are known already for Zeno, who adorned the creative Reason or fire with the names of “fate”, “god”, “soul of Zeus” and the “necessity of all things”, and maintained that fate is the same as providence and Nature 57. Chrysippus is thus taking over traditional Stoic doctrine. All these identity statements suggest that at least the extensions of the terms involved are the same. This is confirmed by the fact that physically all these entities are considered to be pneuma by Chrysippus 58. But the meanings of the terms were not all the same. For instance, although Chrysippus and Zeno identified fate with god or god’s reason, some properties that belong to god are never associated with fate : e.g. god is said to be perfect in happiness, blessed, benevolent, caring, beneficent, and not admitting any evil 59. On the other hand, all properties of fate seem to have parallels in those of god, except that for god the detailed description of the network of causes (below) is not recorded. Thus although co-extensive with god, fate seems to denote a particular set of aspects of god, i.e. god qua being the active principle that structures and moves the world. From Chrysippus’ accounts of fate we can extrapolate four aspects of the active principle that were particularly connected with the term “fate”. Some had been traditionally associated with fate, others are specifically Stoic. Gellius has preserved a Chrysippan definition of fate, in which all the aspects are assembled. Fate is a certain natural arrangement of the universe, with things following upon other things and being involved with other things from eternity, such a weaving-together being inexorable 60. Stobaeus presents a collection of accounts of fate, taken from Chrysippus’ works : Chrysippus <maintains that> the substance of fate is a power of breath, administering order of the all. This he does in his second book On the World. But in his second book On..., in his books On Fate, and occasionally in others he puts forward various views, stating that “Fate is the Reason of the universe” or “the Reason of the things in the universe administered by providence”, or “the Reason in accordance with which past events have happened, present events happen and future events will happen” ; 57. Tertullian Apologia 21, Lactantius SVF I 160, DL 7.135, cf. Ecl. I 133.3-5 and I 78.18-20. 58. Ecl. I 79.1-2 ; cf. PH III 218 ; [Plutarch] Epit. I 7 (Doxographi Graeci 306.1-6) ; Mixt. 224-5 ; Clement Stromata V 14.89.2. 59. DL 7.147 ; Comm. not. 1075e. 60. NA 7.2.3. The text is corrupt ; I follow Sharples’ suggestion, Sharples 1991, 96 and 197. 510 Susanne Bobzien and instead of “Reason” he uses “Truth”, “Cause”, “Nature”, and “Necessity” and adds other terms which apply to the same substance from different perspectives 61. Both passages confirm that “fate” was used to refer to a whole set of aspects of the active principle. First, there is an element of teleology, rationality, organisation and order. This is an aspect that was prevalent in Chrysippus’ exposition of the nature of the universe (cf. p. 496-500), which – as the passages just quoted show – is essential also to the meaning of “fate” : Fate is a natural arrangement or organisation, the reason of the world and the order which administers everything 62. What has been said about the common nature of the world above p. 496-500 holds generally for this aspect of fate. Second, like the active principle and god, fate is described as eternal 63. That is, (a), fate itself does not have a beginning, but has always been there as the organising principle of the world ; (b), the organisation is such that – in some way – whatever occurs had been organised, hence settled, before it occurred. And this advance organisation or fixing is also eternal : what occurs was always organised and settled to occur. Thus fate does not ever determine future events in any finite time before they happen 64. In particular, a person’s destiny is not determined at the time of their birth or conception, as some popular views had it 65. The third aspect, necessity, inevitability, and immutability, was traditionally connected with fate. It is also recorded for Zeno 66 and is repeatedly emphasised in Chrysippus’ accounts. Fate is called “greatest Necessity” 67, and there is a colourful selection of adjectives used to express this aspect of fate : e.g. “invincible”, “unpreventable”, “immutable”, “unchangeable”, and “inexorable” (ajparavbato"), which becomes the standard attribute of fate in later texts. We have seen above the main reason for the inevitability of fate : nothing external to the universal Nature (i.e. fate) can interfere with what occurs, since there is nothing external to it. In particular, human beings cannot interfere, since their natures are themselves part of the common Nature or fate. Connected with the eternal predetermination of all events is the point that it is futile to attempt to influence one’s fate, since at any time all future events have been settled already, 61. Ecl. I 79.1-12. 62. Cf. DL 7.149, Ecl. I 79.17-18. 63. NA 7.2.3 ; Cicero De Fato 20, 27, 28 ; Diogenianus in Praep. ev. 324.3-5 ; cf. also Ecl. I 79.15-16. 64. This is so despite the fact that for the Stoics the course of the world is cyclical. 65. e.g. Tacitus Ann. 6.22. 66. Tertullian Apologia 21. 67. Stoic. rep. 1055e and Stobaeus, quoted above. Early Stoic Determinism 511 and are thus unchangeable. This contrasts with theories which allow for the – occasional – later change of what has been predetermined as a result of prayers 68. The fourth factor concerns fate insofar as it links individual objects. Several Chrysippean accounts speak of fate as linking the things in the world in some orderly manner. Chrysippus named the way the things are connected an “interweaving” (ejpiplokhv) 69. This interweaving is explained by him as “things following upon other things and being involved with other things (from eternity)” 70. The idea of interconnection is found also in his etymological exegesis of “fate” as “connecting cause of the things” 71. What are the things that are linked ? Plotinus, reporting Stoic doctrine, mentions the “interconnection of causes with each other” 72. Then there is a later Stoic etymological explanation of “fate”, as “chain (or series) of causes” (eiJrmo;" aijtivwn). Thus Chrysippus probably was concerned with the concatenation of all causes or all things. There is not much difference here, since the causes are all things – considered in a certain respect, namely insofar as they have effects (see p. 492) : Fate, qua pneuma in all things, links these things through space and time ; through time by way of antecedent causes, through space by way of sustaining causes and sympathy 73. However, we need to exclude an interpretation which is sometimes read into the accounts from a modern perspective, viz. that we have a chain of causes and effects, in which the effect of one instance of causation is the cause of the next. This cannot be the early Stoic position, since cause and effect are ontologically distinct, one being a material thing, the other an actualised predicate. How are we to interpret the metaphors of interweaving and chain ? Stoic causes are corporeal and relative ; that is, they are bodies while and insofar they actively produce (or contribute to producing) an effect in a body. Both causes and effects thus have duration (see p. 491-496). We can then imagine the simplified case of a chain of causes thus : A body b1 is the cause c1 from t1 to t3, producing an effect e1 at a body b2, from t2 to t3. As a result, at t3 (perhaps a little earlier), body b2 starts being the cause c2 of another effect e2 at a body b3 ; c2 may last from t3 to t5, e2 from t4 to t5. At t5 (or a bit earlier), b3 starts being a cause of a further effect e3 at a body b4, etc. A case of a section of a 68. Nat. hom. 106.15-20. Mostly, however, the ancients assumed that if something is fated, it is immutable. 69. NA 7.2.3. Ecl. I 78.4-6 ; cf. also Aristocles in Praep. ev. XV 14.2. 70. NA 7.2.3. 71. Cf. Diogenianus in Praep. ev. 323.11-12 ; DL 7.149. 72. Plotinus Enneads III. 1.2. p. 236.30-1 ; [Plutarch] Epit. I 27.4 (Doxographi Graeci 322.1112). 73. These two dimensions of the connection of fate, time and space, are suggested by [Plutarch] De Fato 574e. 512 Susanne Bobzien network, which includes a combination of causes for one effect, would be this : A body b1 is the cause c1 from t1 to t3, of an effect e1 from t2 to t3 at a body b2. b2 as cause c2 from t3 to t5 and b3 as cause c3 from t4 to t6 are co-causes of effect e2 at a body b4 from t4 to t6. As a result, at t6 (or a little earlier) b4 may start being a cause of an effect e3 at a body b5 from t7 to t8, and perhaps also being a cause of an effect e4, perhaps at b1, etc. That is, we have a temporal concatenation of bodies insofar as they are causes. This is in line with the formulations of interweaving of things or chains of causes, i.e. of corporeal entities, as opposed to of the incorporeal effects, motions, states, occurrents, etc. In one respect, Chrysippus’ statement of the “interweaving” of things seems superior and better in keeping with his theory than the metaphor of a chain of causes. A chain leading through time has one link at a time, and conjures up the picture of isolated parallel “strings” of causation 74. The picture of an interweaving or concatenation of things on the other hand allows for indefinite complexity : a network rather than isolated strings, in which many instances of causation can occur at the same time, or be temporally staggered and overlap, etc. The account of fate as “concatenation of causes (in the plural)”, which arises from the inner-cosmic perspective, is not to be confused with the identification of fate with Cause (in the singular) from the global perspective, where both “fate” and “Cause” refer to the one active principle. The difference is marked by a terminological distinction. Chrysippus always referred to the one cause which is identical with fate and the active principle by the feminine noun aijtiva 75. Aijtiva is identified with fate, god, and the active principle (pneuma) of the universe, which are all one 76. This one Cause is the same as Reason. So Stobaeus : Fate is the Reason of the universe.... and instead of Reason he uses “Truth”, “aijtiva”, “Nature”, and “Necessity”.... (Ecl. I 79.5-10) And Seneca : As you know, our Stoics state that there are two <principles> in the nature of things from which everything occurs : Cause and matter.... Cause, i.e. Reason, moulds matter and turns it wherever it wants 77 (Ep. 65.2). 74. except if one assumes that at any time the whole world state counts as a link, which may be suggested by the singular noun “chain”, but for which there is otherwise no evidential support. 75. Cf. Stoic. rep. 1055e ; 1056b ; 1056c ; Ecl. I 79.5-10. Cf. also Plotinus, Enneads III 1.2 17-22. 76. Cf. Marcus Aurelius 5.8 ; 8.27 ; 9.29 ; Seneca De beneficiis IV 7 ; AM 9.75 ; Syrianus Met. (SVF II 308) ; Ecl. I 31.13-14. 77. Dicunt, ut scis, Stoici nostri duo esse in rerum natura ex quibus omnia fiant, causam et materiam. Materia... Causa autem, id est ratio, materiam format et quocumque vult versat. See also Ecl. I 132.27-133.5 ; DL 7.135. Early Stoic Determinism 513 “Reason” here is the pneumatic world-reason, which pervades the universe as a whole, including all things 78. What is the relation between the one Cause (aijtiva) and the many causes (aijtivai or ai[tia), in the context of fate ? Seneca uses the phrase “the cause of the causes” 79. The above evidence about the identification of Cause, Reason, and the active principle, suggests that the relationship is as follows : the one Cause is the pneumatic world-reason, which penetrates all material objects, and is responsible for their shape and changes. The Cause of any individual cause (ai[tion) is the portion of (rational) pneuma which permeates that cause 80. For example, in an instance of causation of bread-cutting, the pneuma in the cause knife which cuts the bread is the Cause of that knife, i.e. that portion of the world-reason in the knife that makes it cut the bread 81. Individual material objects are thus causes insofar as they are pervaded by a portion of the world-reason. This causal function of the worldreason is the ground why it is also called “Cause” and “Fate”. This relation between the active principle, qua Reason, or Cause, and the individual causes is crucial for understanding early Stoic determinism. It brings us closer to an answer to the question how the teleological and causal (“mechanistic”) elements in the theory combine ; how the predetermination of every movement by the situation of the world prior to it, and including antecedent causes of it, cohere with the eternal, rational world order. We know that the picture cannot be that of a transcendental deity who devises a plan and then realises it in the world. There is no space for either god or god’s plan outside the world. Both god and god’s will or reason are part of the one material world. God, qua Fate or Cause, is what makes everything a cause. It is the pneuma in every thing, by which all things are linked and through which the world progressively develops. God, or god’s will, thus works – in part – from the inside of all things. Hence we should imagine every individual cause as containing a piece of information about where it is heading. Every cause carries with it, and in itself, the relevant bit of god’s will or plan. The Stoics offer several analogies to explain this teleological element in their determinism. One compares god (the active principle, fate) to a human being : 78. Cf. Origen, Contra Celsum VI 7 (71.5-7 Borret), DL 7.134. 79. NQ II 45. 80. This explains the passage [Plutarch] Epit. I 12 ; (Doxographi Graeci p. 310.6-7) : “the Stoics hold that all causes are corporeal ; for they are portions of pneuma”. 81. These considerations suggest a plausible way of understanding the difficult closing sentence of Stobaeus’ report from Chrysippus’ theory of causes : “Ai[tia is the Reason in the cause (ai[tion), or the Reason in respect of the cause qua cause.” (Ecl. I 139.3-4) Making use of the above example, we can say that that aspect or part of the pneuma in the knife that makes it cut the bread is the reason in respect of the knife qua cause of the cutting. For although the whole knife is the cause (ai[tion) of the cutting, strictly speaking, it is the pneuma in the knife that is responsible for the effect. 514 Susanne Bobzien In god’s ruling part of the soul, placed by some Stoics in the aether, we find pure, condensed rationality. From this ruling part god or god’s reason literally stretches into everything, and connects all things, as the human soul stretches from the human ruling part into the whole body 82. Presumably, just as I can bend my finger, by sending impulses from my mind, via my nervous system, to my finger, so god can make something happen by sending impulses to the place where it is meant to happen. And just as my finger, when moving on the keyboard, shares in my rationality insofar as it is an extension of my mind, and when I have decided to type a “z” with my finger, then my finger will type a “z”, so everything shares in god’s rational pneuma and is directed by it. But this comparison is lacking – among other things – in the element of predetermination from eternity. Here a second analogy, taken from biology, is more successful : It is Zeno’s analogy with the development of things from seeds. Again, the world is analogous to one living being. As in a seed all the information is assembled to generate a plant or animal, and will do so, provided there is nourishment around, so in the world-pneuma is assembled all the information that is needed to make the world develop, provided matter (the passive principle) is around. And as the seed grows and develops into the thing and determines its life cycle (the form of which was already fully determined in the seed), so the world-pneuma (or creative fire) nourishes on the matter and develops into the world and determines its course 83. Another metaphor which belongs here is that of the “unrolling” or “unfolding” of fate through time 84. It is meant to illustrate the fact that everything has been predetermined from eternity, and nothing is newly created when it happens. All the analogies leave open some vexing theological questions : how is rationality translated into the perfection and orderliness of the world in accordance with a rational design ? For an answer we have to add to the second analogy from the first the deity’s awareness of the world, and/or its self-awareness. For this, again, will have to be thought of analogously to that of animals 85. But the awareness, being that of a rational being, will include not only consciousness of everything that happens at the present moment, but also of everything that has happened and that will happen. We are left to speculate on the question whether the awareness of what is still unrealised in the world is direct awareness 82. DL 7.138, 157. 83. The description of development of animals from sperm by the Stoic Hierocles (2nd cent. AD) is instructive here. The sperm absorbs matter from the pregnant body and develops into the embryo “in accordance with certain inexorable patterns” (Hierocles, Elementa Moralia 1.5ff). This terminology (tavxi", ajparavbato") is the same as is used in the description of fate. 84. Diogenianus in Praep. ev. VI.8.9 ; Div. I 127. 85. We have some brief passages on self-awareness from later Stoics : Seneca Ep. 121 6-15 ; Hierocles Elementa Moralia 1.34-9, 51-7, 2.1-9, 4.38-53. Early Stoic Determinism 515 of the “design” in the “world-sperm”, or god’s memory of the last world cycle 86, or still something else. Finally, it remains to determine how fate was thought to relate to the – incorporeal – effects of causation. Here one central principle of Stoic physics comes in, the “Fate Principle”, that (FP) everything happens in accordance with fate. This principle is documented for Chrysippus, Zeno, Boethus and Posidonius 87 ; no early Stoic is reported to have deviated from it. An unravelled version of the principle can be gained from one of Chrysippus’ accounts of fate : Fate... “the Reason in accordance with which the past events have happened, the present events happen and the future events will happen” (Ecl I 79.5-8) 88. Thus the word “happens” in the Fate Principle is to be taken as in the “atemporal” present, which covers all times from a global perspective. From the human perspective, events are always either past, present or future ; but all three classes (whose membership changes steadily with the course of time) are equally subordinate to fate. The ontological category to which the things that happen in accordance with fate belong are the – incorporeal – events, i.e. changes and qualitative states, past present and future ones 89. This ontological status must be the reason for the standard formulation “in accordance with (kata;) fate”. The corporeal causes each contain part of fate, but the effects, being incorporeal, can only be in accordance with fate. Thus we can see that fate and its workings were conceived of and described by the Stoics in two ontologically distinct ways : on the side of material entities, fate is responsible for the individuation of things, their qualities and their causal connection in the continuum through space and time. On the side of incorporeals, all qualitative states and all changes are the result or effect of fate. But these are only two ways of looking at the same world. 86. Cf. Nat. hom. 111.25-112.3. 87. DL 7.149 ; Diogenianus, in Praep. ev. 322 ; and numerous times in Cicero’s De Fato. 88. Cf. also Div. I 126 ; Aristocles in Praep. ev. XV 14.2 and Diogenianus in Praep. ev. VI.8.9. 89. The verb “to happen” in the formulations of the Fate Principle originally covered all events, both changes and qualitative states. Over time, the focus of criticism of Stoic fate theory turned to the claim that all changes are fated. Chrysippus’ claim that there are no changes without antecedent causes seems to have been the real bone of contention for his critics, and thus the Fate Principle came to be understood as being about changes, not states. In the longer run this isolation of one principle from Stoic physics sparked a debate which lost sight of the principle’s origin and roots ; the principle was transformed into a quasi-definition of fate, and came to be used as a catch-phrase for Stoic type determinism. 516 Susanne Bobzien The Fate Principle is a consequence of the early Stoic theory of fate as allpervading pneuma, and of Chrysippus’ theory of causation and of change. For the early Stoics, giving up the Fate Principle would be giving up the very basis of their physics and cosmology. Susanne BOBZIEN Yale University Bibliography K. ALGRA, J. BARNES, J. MANSFELD, M. SCHOFIELD (eds.), The Cambridge History of Hellenistic Philosophy, Cambridge, 1998. J. BARNES, M. BURNYEAT, M. SCHOFIELD (eds.), Doubt and Dogmatism. Studies in Hellenistic Epistemology, Oxford, 1980. L. BLOOS, Probleme der stoischen Physik, Hamburg, 1973. S. BOBZIEN, Determinism and Freedom in Stoic Philosophy, Oxford, 1998. S. BOBZIEN, “Chrysippus’ theory of causes”, in K. IERODIAKONOU (ed.), Topics in Stoic philosophy, Oxford, 1999, p. 196-242. S. BOBZIEN, “Stoic Logic”, in K. ALGRA, J. BARNES, J. MANSFELD, M. SCHOFIELD (eds.), The Cambridge History of Hellenistic Philosophy, Cambridge, 1999, p. 92-157. J.-J. DUHOT, La Conception stoïcienne de la causalité, Paris, 1989. M. FREDE, “The Original Notion of Cause”, in J. BARNES, M. BURNYEAT, M. SCHOFIELD (eds.), Doubt and Dogmatism, p. 217-249. W. GOERLER, “Hauptursachen bei Chrysipp und Cicero ? Philologische Marginalien zu einem vieldiskutierten Gleichnis (De fato 41-44)”, Rheinisches Museum für Philologie 130 (1987), p. 254-274. K. IERODIAKONOU (ed.), Topics in Stoic Philosophy, Oxford, 1998. A. M. IOPPOLO, “Il concetto di causa nella filosofia ellenistica e romana”, in W. HAASE, H. TEMPORINI (eds.), ANRW, part II 36.7 (1994). M. LAPIDGE, “Stoic Cosmology”, in RIST (ed.) The Stoics, Berkeley-Los Angeles, 1978, p. 161-185. J. MANSFELD, “The Sources”, in K. ALGRA, J. BARNES, J. MANSFELD, M. SCHOFIELD (eds.) The Cambridge History of Hellenistic Philosophy, p. 3-30. J. M. RIST (ed.), The Stoics, Berkeley/Los Angeles, 1978. S. SAMBURSKY, Physics of the Stoics, London, 1959. R. W. SHARPLES, Cicero : On Fate & Boethius : The Consolations of Philosophy IV.5-7, V, Warminster, 1991.