Posidonius’ Two Systems: Animals and Emotions in Middle Stoicism

: This paper attempts to reconstruct the views of the Stoic Posidonius on the emotions, especially as presented by Galen’s On the Doctrines of Hippocrates and Plato . This is a well-studied area, and many views have been developed over the last few decades. It is also significant that the reliability of Galen’s account is openly at issue. Yet it is not clear that the interpretative possibilities have been fully demarcated. Here I develop Galen’s claim that Posidonius accepted a persistent, non-rational aspect of the soul that he connects with the merely animal part of humans. The aim is to begin from this testimony in answering two questions: (1) How might the possession of a non-rational element of the soul operate alongside the hêgemonikon (leading-part of the soul) as a source of impulse for Posidonius. (2) How does this persistent animal aspect conform to the Stoic ontological classification found in their scala naturae ? I shall argue in response to these that (a) Posidonius distinguished the merely cognitive aspects of the soul from those that are rational, and (b) that the hêgemonikon itself is not to be identified with what is rational. Accepting a persistent non-rational source of emotional impulses allows Posidonius a richer framework for explaining human affective responses and behaviours. I also briefly address Galen’s motivation for the account he offers. It is in view of Posidonius’ approach to Plato’s Timaeus that Galen’s discussion finds its most plausible interpretation. phenomena broad-chested by


Introduction
The Stoic Posidonius (c. 135-51 BCE) is a very attractive figure in the history of philosophy. Cicero calls him "the most noble of philosophers"; Strabo "the most learned philosopher in my time"; Seneca "among those who contributed most to philosophy".1 His basic commitment to the philosophical approach of the Stoa, as most extensively articulated by , is clear, but there is also ample evidence that Posidonius pushed Stoicism into new domains largely avoided by his predecessors, including mathematics, the exact sciences, history, and geography.2 We even have evidence that he wrote an influential book on military tactics.3 Exactly how Posidonius' development of the earlier Stoic tradition should be characterised, and whether there is compelling evidence for the existence of heterodox positions on his part are questions that are as long-standing as the modern scholarship is on the philosopher.4 My aim in the following is to offer a contribution to the reconstruction of his position on the emotions (pathê), what are termed "emotional movements",5 and how these relate to the Stoic theory of impulse. Specifically, this theory holds that for an adult human-being an impulse is an act of assent to or endorsement of a value-judgement brought about by a particular type of impression (phantasia) within the commanding-part (hêge monikon) of the soul.6 Such an impulse is a cause of action, and because we are responsible for our judgements, we are responsible for our impulses.7 The emotions for the Stoics, Stobaeus tells us, are impulses "excessive and disobedient to reason".8 Posidonius' "emotional pull" (παθητικῆς ὁλκής: PHP 5.5.21=F169.101-3), or his "emotional movements", these may be attributed to this merely animal aspect of humans. (3) As such, emotional impulses can come about in adult humans that are independent of rational judgement, or assent, as the result of the activity of the non-rational, affective, animal part of the soul.
Of these three claims, I take it that by far the most controversial is (3), and it is worth being explicit about what I am arguing and what I am not. This is not an argument that the emotions are always generated independently of judgement. Nor do I maintain that the emotions are thought to arise independently of any cognitive endorsement of their emotional content.14 Rather the point is that, in adult humans, the emotions are sometimes to be attributed only to an ineliminable non-rational animal aspect within their souls. This animal part is non-rational, but the Stoics are also clear that non-rational animals have the cognitive resources to experience both impression and impulse (LS 53P and T). Notably, Posidonius goes further than other Stoics by attributing full-blown emotions to animals. There is good evidence too that assent in some attenuated form, understood as a sort of non-rational endorsement (quasi-assent), was attributed by the Stoics to non-rational animals (LS53, 62G6).
The idea might be made clearer if we look at the following passage from Engberg-Pedersen: He (Posidonius) kept complaining that Chrysippus, who explicitly denied that there is more than one root, viz. that of understanding and belief, could not explain passion. In order to explain passion, so Posidonius claimed, one needs something more than understanding or belief, viz. a non-rational function of the mind.15 This is partly correct and partly misleading. Yes, Posidonius demanded a more complete causal account of the rise of the emotions,16 but this should not be taken to imply that the non-rational aspect of the soul he posited operated independently of its cognitive features. 'Something more' does not amount to 'something separable'. Nor should we presume that such an aspect amounted to a function for Posidonius in the manner Galen suggests. This should not be very surprising for a Stoic account of such cognitive or intentional states as belief or holding an opinion (doxa). Such weak epistemic states are explicitly denied to the perfectly rational Stoic sage and held to be befitting only for the insecure and changeable condition of the non-rational, non-sage.17 In fact, it is because the morally imperfect have mere beliefs that they are subject to emotions in the first place. 18 The principal claim, then, is that Posidonius thinks adult humans have rational animal and non-rational animal aspects of their souls that are responsible for emotional impulses. Indeed, I will be canvassing the view that Posidonius accepts the simultaneous possession of both rational and non-rational cognitive faculties.19 This seems a straightforward violation of the Stoic commitment to a unitary soul where any emotion is to be interpreted as the product of a mistaken judgment produced by the rational hêgemonikon.20 However, the story for Posidonius is more complex than straightforward heterodoxy. I will suggest below, building partly on the reading of Hendrik Lorenz, that Posidonius develops a familiar earlier Stoic distinction between distinct types of cognitive endorsement to impressions -a 'yielding' (εἴξις) and a more reflective manner of assent connected with the possession of reason.21 That 'yielding' is how the Stoics understood endorsement to impressions in non-rational animals is a familiar position in the literature. 22 The view I develop here is that Posidonius advances on this earlier Stoic position by identifying human 'yielding' with its non-rational animal counterpart, allowing for a type of non-rational impulse in otherwise rational adult humans. This is the upshot of identifying an animal aspect of adult human souls. If this is right, humans have the capacity to experience impulse that has not been processed by their rational faculty.23 This, however, is a very different claim than that no processing has occurred in 17 LS41C. 18 By 'emotions', I mean what the Stoics termed pathê. These are sometimes termed 'affections' or 'passions'. All are acceptable translations so long as we keep in mind that these states are to be understood as corporeal occurrences within the soul (SVF 3.463) that are disobedient to reason (SVF 3.378=LS65A). Such undesirable states must also be contrasted with what the Stoics called eupatheiai, or feelings (i. e. joy, caution, rational wishing) which do conform to reason (DL 7.116). These are restricted, at least in the early Stoa, to sages. 19 I borrow this formulation from Brittain 2002, 255. I stop short, however, from ultimately endorsing Posidonius' non-rational aspect as an independent 'faculty'. 20 See, for example, SVF 3.459=LS65G, and below. 21 LS53S. 22 LS 1, 322, following Inwood 1985. 23 We might, at this stage, want to keep separate the human rational faculty from 'reason' itself. See further below.

such instances. For Posidonius, what is cognitive is not be identified with what is rational.
It should be obvious that in the above scene-setting I have been imprecise in describing the non-rational 'aspect' or 'element' of the soul I wish to attribute to Posidonius' account, following Galen. If we were to take the latter at his word, this aspect simply would amount to a capacity on the Aristotelian model of a dunamis. Yet, as indicated above, there is reason to be cautious; notably, Galen's verbatim quotations of Posidonius do not demand that this is how we must understand such an aspect. Of course, we cannot dismiss outright the possibility that Galen is entitled to interpret Posidonius' soul in this way. However, there is another option I hope to sketch out for consideration. This turns on two characteristic features of Posidonius' contribution to the Stoic tradition: (1) his intense focus on the tools of causal explanation, and (2) an evident interest in how the corporeal soul of the Stoics is itself to be conceptualised, which we learn was developed through an interest in the Plato's thinking, especially as found in the Timaeus.
The idea then is that the non-rational aspect Galen isolates, and which I attempt to preserve, is best construed in causal terms as a locus of responsibility and is less plausibly presented in functionalist terms as an Aristotelian capacity, following Galen. I conclude our discussion below with this suggestion.
Taking stock: what distinguishes Posidonius from Chrysippus is how logos is thought to relate to the pre-rational soul after the actualisation of the potential for reason at age 14, as the Stoics held.24 Put another way: Is the advent of a rational capacity an addition to the souls of pre-rational children (as I maintain Posidonius thought), or does reason transform the soul, rendering all impulses of the soul subject to its judgement?

The Animal Aspect
In this first section, my larger goal is to outline Galen's evidence as plainly as possible, foregoing much of the scepticism of it that has become increasingly common in the literature. In particular, I hope to demonstrate that this evidence 24 Reason is 'perfected' at this age: see SVF 2.764. This does not suggest that the fourteen-yearold is rational in a complete sense. Rather a human's potential for virtue is actualised at this age; this allows for the acquisition of the capacity that might one day result in reason's perfection, i. e. achieving virtue and a state of knowledge. This suggests that possessing some concepts as one develops, prior to full rationality, is consistent with the Stoic view of reason. See Caston 2021, 26 f., for discussion. reports that Posidonius held that there is a non-rational animal aspect of the human soul that is both ineliminable and identified as a source of the emotions. I take it that the evidence for this somewhat circumscribed claim is quite strong.25 Certainly, Posidonius links innate features of the human species with the genesis of the emotions: τὸ δὴ τῶν παθῶν αἴτιον, τουτέστι τῆς τε ἀνομολογίας καὶ τοῦ κακοδαίμονος βίου, τὸ μὴ κατὰ πᾶν ἕπεσθαι τῷ ἐν αὑτῷ δαίμονι συγγενεῖ τε ὄντι καὶ τὴν ὁμοίαν φύσιν ἔχοντι τῷ τὸν ὅλον κόσμον διοικοῦντι, τῷ δὲ χείρονι καὶ ζῳώδει ποτὲ συνεκκλίνοντας φέρεσθαι. οἱ δὲ τοῦτο παριδόντες οὔτε ἐν τούτοις βελτιοῦσι τὴν αἰτίαν τῶν παθῶν οὔτ' ἐν τοῖς περὶ τῆς εὐδαιμονίας καὶ ὁμολογίας ὀρθοδοξοῦσιν· οὐ γὰρ βλέπουσινὅτι πρῶτόν ἐστιν ἐν αὐτῇ τὸ κατὰ μηδὲν ἄγεσθαι ὑπὸ τοῦ ἀλόγου τε καὶ κακοδαίμονος καὶ ἀθέου τῆς ψυχῆς. PHP 5.6.4-6=F187EK (part).26 The cause of the affections, that is, of disharmony and the unhappy life, is not to follow in all things the daimon in oneself, which is both inborn and similar in nature to the one that rules the whole cosmos, but to deviate and be carried away by the worse and the beastlike. Those that have failed to see this neither give the better cause for the affections in these matters, nor do they have correct beliefs about happiness and consistency. For they do not see that that primary thing in happiness is to be led in no way by the irrational and the unhappy, i. e. the godless aspect of the soul.27 While this passage suggests that human beings acquire a divine capacity to achieve virtue at birth, Posidonius is also clear that such a capacity is balanced by the possibility of beast-like (ζῳώδει) behaviour that is also implicitly innate within us.28 Vicious behaviour is framed as essentially animalistic. It is notable 25 Boys- Stones 2007, 89-91, andLorenz 2011, 209-11, both accept, in  that this beast-like behaviour is not explained by any reference to external influences on the soul.29 Human nature as such is the relevant explanans.
Notable too is that Posidonius confirms in this verbatim quotation that his position on the soul was developed in an explicit dialectic exchange where psychology grounds the determination of the human telos. His opponents are most likely Chrysippus and the Chrysippeans, and Galen very much encourages us to come to this conclusion. Yet the fragment itself is compatible with a much broader dialectical context that includes those in the Stoic tradition but also others found within Posidonius' creative use of the traditions he inherits, including Platonism. I return to this point towards the end of our discussion.
The theme of an internal non-rational psychic component is continued in Galenic evidence outside of PHP. Vice is cast as internal to the agent and rooted in their embodied soul: Posidonius, on the other hand, was convinced that he should betray the Stoic sect rather than the truth. In his work On Affections […] his view is totally opposed to that of Chrysippus […] It is not, then, Posidonius' belief, either, that vice enters from the outside, without there being any specific root for it in our souls, from which it takes its first impulse, and then sprouts and grows; his view, rather is quite the opposite. There is, indeed, a seed of vice within our selves. (QAM 78.8 Müller=820 K., trans. after P. N. Singer).
The connection presupposed in these two passages between identifying the cause of the emotions and the virtuous life is a persistent Posidonian concern. Galen quotes verbatim: ταύτην τε δὴ τὴν ἀποπίαν διέλυσεν ἡ αἰτία τῶν παθῶν ὁραθεῖσα καὶ τὰς ἀρχὰς ἔδειξε τῆς ἐν τοῖς ὀρεκτοῖς καὶ φευκτοῖς διαστροφῆς καὶ τοὺς τρόπους τῆς ἀσκήσεως διεῖλε καὶ τὰ διαπορούμενα περὶ τῆς ἐκ πάθους ὁρμῆς ἐξέφηνεν.30 When the cause of the affections was seen, it removed this absurdity [i. e. what Chrysippus said about the end (telos)]; it showed the sources of distortion in what is to be sought and avoided; it distinguished the methods of training; and it cleared up the difficulties about the impulse that arises from affection. (PHP 5.6.14-16, trans. De Lacy. modified).
29 Boys-Stones 2007, 91 f., emphasises the importance of Posidonius' appeal to 'internal pressures' on judgement and impulse, distinguishing this from their absence in earlier Stoic theorizing. Contrast Chrysippus' appeal to oral communication and the nature of (external) things as such as corrupting, as reported by Galen (F169EK, sec. d). 30 See Lorenz 2011, 195-97, for the somewhat curious formulation 'impulses that arise from affection'. On the standard Stoic account, affections/emotions simply are impulses. Lorenz plausibly suggests that 'affection' here denotes what are termed 'affective motions' and 'affective pulls' elsewhere in Galen's evidence. This largely follows Cooper 1998, 71-111, who argues that 'affective movements' are not the emotions themselves but something that underlies them.
The relevance of the emotions to both the definition of the Stoic telos and to the education of pre-rational children suggests the paramount importance of an innate and persistent affective aspect of the soul in Posidonius' theory. How the end (telos) of life is to be understood received attention from all significant Stoics, and Posidonius is no exception. His suggestion, as we have seen above, is that the happy life is achieved by following in everything one's internal daimon. Galen reports (PHP 5.469-76=F187EK) that Posidonius took issue with Chrysippus, and earlier Stoics generally, on the formulation of the end. It appears that Posidonius' criticism figures within a complex, inter-school, dialectical exchange about the integration of elements, indifferent in themselves, but in accordance with nature, within the end itself of living in agreement with nature.31 The primary Academic, Carneadean criticism of the Stoic view of the end was that they commit themselves to circularity or to two different ends in attributing some value to the selection of things according to nature as well as value to nature itself.32 What is relevant for us in this debate is that we have strong evidence that an innate, persistent aspect of soul figures within Posidonius' attempt at such a definition of the telos. And the evidence notably comes outside of Galen's account. Clement reports (Strom. 2.21.129.4-5=F186EK) that Posidonius proposed the end to be: To live contemplating the truth and order of all things (θεωροῦντα τὴν τῶν ὅλων ἀλήθειαν καὶ τάξιν) together and helping in promoting it as far as possible, in no way being led by the irrational part of the soul (κατὰ μηδὲν ἀγόμενον ὑπὸ τοῦ ἀλόγου μέρους τῆς ψυχῆς). Trans. Kidd; emphasis mine.
We can go even further on this point. The acceptance of something both persistent and non-rational is also motivated by what Galen reports is Posidonius' revision of the Stoic theory of oikeiosis, or appropriation. This concept describes how animals are born with an innate drive that impels them towards their self-preservation. In the case of humans, as children develop into rational adults, this striving for self-preservation makes one's rational nature central.33 31 For interpretative possibilities, see LS 1, 408-10, and EK, 674-83. The historical reconstruction of the debate is complex, but the main issue for Posidonius is how his formulation of the end is an improvement on Antipater's, which seems to be a deviation from the Chrysippean norm designed to answer a trenchant Academic (Carneadean) criticism. I leave discussion of most of this reconstruction aside, as peripheral to my main objective. 32 See Plutarch, Comm. Not. 1070F-1071E and SVF 3.15. 33 See the texts collected by LS57.
Our evidence from Galen suggests that Posidonius modified this picture by incorporating additional objects of appropriation into his understanding of rational development.34 These additional objects are divided into two groups: pleasure; and power and victory (F158EK). They are reported to be certain capacities (δυνάμεις τινάς) in our soul with these objects as their natural goals. Galen then notes that Posidonius observed that these powers operate obviously in other animals (i. e. those that are non-rational) as well, marking out his position from the early Stoics.35 On Galen's account, the upshot of this acceptance of a persistent, non-rational aspect within the soul is that the proper mode of childhood education is clarified (F168EK) and the evidence of incorrigible viciousness in rational adults explained (F169EK). As such, Galen gives us every reason to believe that the advent of reason in one's development from childhood to adulthood does not erase or transform such non-rational capacities; rather it works to add another object of appropriation, i. e. the good conceived of in the holistic manner Clement (above) suggests. The merely animal goals of pleasure and the desire for victory remain as persistent features of the otherwise rational adult soul.
However, these additional objects of appropriation are not just claimed to be ineliminable, they are also, we are told, meant to be explanatory of the origins and eradication generally of the emotions. This is how Galen understands these additional objects (F161EK), and he does so by claiming that the natural objects of pleasure and victory are functions of the merely animal aspect of the soul: Some, you see, are deceived into thinking that what belong to the irrational powers of the soul as natural goals, are natural goals without qualification; what they don't know is that pleasure and power over one's neighbour are goals of the animal aspect of our soul (τοῦ ζῳώδους τῆς ψυχῆς), while wisdom and all that is good and moral together are the goals of the rational and divine aspect. Trans. Kidd. Given the connection seen above between the cause of the emotions and the telos, it should not be surprising that Galen frames this interpretation of Posidonius' emotional, animal aspect of the soul by reporting that it is this that must be supressed in order to 'live in agreement with nature'.36 We find no hint, however, that the animal aspect is discarded at the point of the acquisition of reason or that 34 Here I follow the account offered by Boys-Stones 2007, 88-92. 35 Cf. Panaetius' claim that courage ought to be attributed to non-human animals (Cicero, De Off. 1.50). Galen is clear that Chrysippus denied the emotions occur in non-rational animals, see PHP 4.5.4. 36 PHP 5.6.15 f. it may be eliminated by moral education; it is rather a quality of the human soul that must be addressed and mastered but may not be excised. Now, we must be cautious here because the above passage is clearly Galen's reading of the Posidonian evidence and is not presented as a verbatim quotation. Yet it is significant that he insists that Posidonius broke from the orthodox Stoic line by accepting that both animals and children, despite being non-rational, are capable of emotion (F159EK, cf. F33EK).37 If this is correct, Posidonius did not accept that reason per se was a requirement for impulse, at least in the case of animals and children. He had, then, a model for the explanation of the emotions that did not assume the rational processing of their content. However, the possibility is very much alive, as we shall see, that some cognitive processing within the non-rational soul is assumed by Posidonius and that this is fully in keeping with the Stoic view of the non-rational souls of children and animals.
In this respect, F154EK is an important piece of evidence. This striking, if controversial, report from ?Ps.? Plutarch (De Libidine et Aegritudine, 5 f.) discusses Posidonius' account of the effects of the connection between body and soul on the affections.38 Certainly Posidonius at least says in his classification that (1) some are of the soul, (2) some are of the body, and (3) some do not belong to the soul but are physical with mental effects, and (4) others do not belong to the body but are mental with physical effects.
(1) Instances of what belongs to the soul without qualification are those having something to do with judgements and suppositions (κρίσεσι καὶ ὑπολήψεσιν), like desires, fears, fits of anger. (2) Those which belong to the body without qualification are fevers, chills, contractions, opening up of the pores. (3) Those which are physical with mental effects are lethargies, madness arising from black bile, mental pangs from physical gnawing pains, sense presentations, feelings of relaxation. (4) And the other way round, those which are mental with physical effects are tremors and pallors, that is, changes of appearance in fear and grief. (trans. Kidd, modified).39 37 See Fillion-Lahille 1984, 156-59, for some doubts. Posidonius' reported insistence on animal emotions did not become a mainstream Stoic view; cf. Seneca, De ira 1.3.7 on the 'quasi'-emotions of animals, with SVF 2.143. On the uniqueness of emotions to humans in Greek thought, see Newmyer 2016, esp. chapter 7, with Stephens 2014 on Epictetus. 38 This passage comes from the so-called 'Tyrwhitt's Fragments' and is usually, but not always, taken to be spuriously attributed to Plutarch; see Kidd 1988, 560-62 for discussion. See Sandbach 1969, 211-16, for an argument for authenticity. 39 Kidd translates κρίσεσι καὶ ὑπολήψεσιν as 'rational decisions and suppositions', but this is misleading as it implies that both kriseis and hupolepseis are rational cognitive processes associated by Posidonius with the activity of reason. Yet, as we will see below, there is evidence to support discriminating between these two terms as the former is connected with rational, reflective assent while the latter is associated with a cognitive process more or less automatic, This testimony makes it plain that Posidonius emphasised and categorised the reciprocal influences between body and soul at the level of the emotions and connected these with cognitive presentation. That is to say that a dynamic account of the causal interactions between body and soul, including sense presentations, was a desideratum for his account of the soul.
Most importantly, however, this passage confirms in (1) that products of the soul, including the emotions of fear and anger, are explicitly cognitive for Posidonius. There may be room for attributing non-cognitive factors to his account of emotions which are not described here, but some cognitive processing seems a prima facie minimal requirement for their generation. If so, any interpretation of Posidonius' non-rational aspect of the soul, if it is to be held responsible for generating the emotions, cannot be identified, as Galen would have us believe, with the Platonic account of the Timaeus (77b-d), which denies cognitive status to the appetitive part of the soul. We will see below that Posidonius' interest in Platonic psychology, while genuine, is far subtler than simple appropriation.
Taking stock of this initial survey of the evidence: I have tried to suggest that Posidonius accepted a non-rational affective aspect of the soul as a persistent feature of humans and that he identified it with the merely animal part of the human rational animal. It is difficult to deny, at the very least, that this is the picture Galen wishes us to accept and that it is worth attempting an interpretation of Posidonius that can make sense of it. Clement's evidence on the telos and the persistence of a non-rational part of the soul is also hard to dismiss.
However, one could acknowledge this broad outline and still insist that the existence of a merely animal quality within rational adults tells us little about how such an aspect functions and nothing about whether it could be responsible for impulse absent reason in the case of adults.40 Such an aspect may simply amount to a persistent inclination, as Boys-Stones and Lorenz have it, that somehow affects or 'puts pressure' on judgement but does not function itself independently of reason. What needs to be shown is that Posidonius had a framework for two distinct cognitive systems within the human hêgemonikon and that reason is not required for the operation (i. e. the production of impulse) of the first, persistently animal, psychological mechanism. less reflective, and more prone to error. It is this distinction between types of cognitive processes that Posidonius wishes to emphasize and to incorporate within his account of pathê. On (4), cf. Seneca, De ira 2.2-4, with Sorabji 2000, 66-75. 40 The latter worry is what Boys-Stones 2007, 90 fn156, seems to have in mind, contra Sorabji 2000, 127-29, by arguing that the fact that children and animals have emotions without judgement does not tell us that the emotions are generated in this way in rational adults.
Indeed Posidonius criticises [Chrysippus] also on these matters, and he tries to show that the causes of all false suppositions in the sphere of rational consideration <arise through ignorance, but when they occur in the emotional sphere> they arise because of the affective pull, but false opinions are antecedent to this pull, as the rational part is weakened in respect of judgement. For he says that impulse sometimes comes about in the animal because of the judgement of the reasoning part, but often as the result of the movement of the affective part. Trans. Kidd (modified).
Commentators have expressed significant worries about this passage. As indicated by the diamond bracketed addition to the text above, many have thought that there is a lacuna where part of an antithesis opposing ἐν μὲν τῷ θεωρητικῷ has disappeared, suggesting the theoretical sphere was contrasted with the practical, or emotional.41 Fillion-Lahille, Tieleman, and others have tried to make sense of the text without supplement.42 This textual conservatism is tempting but, in the final analysis, difficult to accept. The Greek construction itself seems to demand an antithesis of some sort, and this is very strongly implied by Galen's report on impulse in the final sentence. So, even if we ignore the textual crux, it is clear that Galen is reporting that Posidonius accepted two distinct sources of impulse. As such, I tentatively follow Kidd's text.
The first thing to note here is that Galen explicitly connects this discussion of impulse with the animal aspect of the soul. We find this in the above passage with the mention of the rise of impulse in animals (τῷ ζῴῳ), and Galen continues immediately after the above passage: 41 See Kidd 1999, 233. De Lacy 1978, adds γίγνεσθαι δι' ἀμαθίας, ἐν δὲ τῷ πρακτικῷ. 42 Tieleman 2003, 231-42. Posidonius reasonably attaches to this discussion (i. e. on the sources of impulse) the phenomena from physiognomy: all broad-chested and warmer animals and humans are more spirited by nature, the broad-hipped and colder, more cowardly.
The second point is that the rise of impulse in either source depends on cognitive features of the soul. In the first case, impulse is the product of judgement (κρίσει); in the second, false beliefs (ψευδεῖς δόξας) weaken the reasoning part, allowing for the action of the emotional, or affective, pull (παθητικῆς ὁλκῆς). While the pull is not explicitly confirmed to operate using typical cognitive functions (perception, impression, belief, and assent), this is not ruled out either. It is simply that reason itself is not involved. In fact, as we have seen, there is good evidence elsewhere that Posidonius understood the emotions to involve necessarily judgement or supposition (F154EK).43 This is further emphasised if we explore what the missing antithesis likely amounted to for Galen himself. One passage that has not, as far as I am aware, been adequately considered in this discussion is PHP 4.7.39-44. In this text, Galen is drawing conclusions about emotional movements (παθητικὰς κινήσεις) to illuminate the errors in Chrysippus' view that the emotions are judgements. The difficulties Posidonius is said to have raised about the weakening of emotions over time, despite the persistence of the correlated judgement, and about those who unwillingly weep, are discussed in the previous pages.44 In this summarising passage, Galen supports Posidonius' view on how the passage of time affects the emotions by distinguishing between two different types of opinion. The first is connected by Galen with rational views, judgements, sciences and arts (αἱ δὲ λογ ικαὶ γνώσεις τε καὶ κρίσεις καὶ ὅλως ἐπιστῆμαι πᾶσαι καὶ τέχναι). Two examples of this type of opinion are given: 'two times two are four' and 'all radii of a circle are equal'. These are theorems (θεωρημάτα) that nobody would ever 'have their fill of' (οὐδείς ἐστιν ὅστις ἐμπλησθεὶς). On the other hand, there are the emotional activities of weeping, grieving, groaning, wailing, and mourning that are subject to satiation despite the persistence of the original supposition (hupolêpsis). Both examples imply the cognitive workings (impression and belief, at the very least) of the soul; the language of judgement (krisis), however, is reserved for the areas of the most stable intellection.
This, I submit, is how Galen himself understood the antithesis. Suppositions are to be divided into those that reliably result in assent (on the example of the 'theorems' above) and a second type that are less securely tied to impulse and action. In the latter case, the weakness of the connection between supposition and impulse is supported by two groups of phenomena that Posidonius -widely reported for his interest in causal explanation -wishes to explain.45 First there are the examples noted above (the passage of time on the emotions, unwilling weeping) as well as the Homeric case of Agamemnon both rejecting reason by succumbing to emotion and yet embracing it by seeking counsel on the basis of the same supposition about the same state of affairs. Secondly, there is the impact of habituation on the emotions, something of particular concern to Posidonius (F165EK, F168EK).
For Posidonius, habituation is what severs the link between weak supposition and impulse; thus it may serve as a therapy for the emotions. Galen quotes verbatim: Two persons may have the same weakness and receive a similar presentation of good or evil, yet one may incur an emotion, the other not […]. At least the more unused a man is to a situation, the more affected he is in circumstances of fear, distress, desire, and pleasure, and the more vicious, the more speedily seized by emotions. (F164EK, Trans. Kidd).
Each of the persons here are clearly 'weakened in respect of judgement', i. e. in the position of holding mere beliefs, yet the one more familiar with the experience of the emotion is less likely to succumb.46 If this is right, there seem to be two possible ways of eradicating the emotions for Posidonius. In the first, the rational aspect of the soul is improved through the acquisition of knowledge, cutting out the condition of weakness that allows for the 'emotional pull'. However, progress on avoiding emotion does not require such rational gains. Habituation can improve even those weak in reason.
How does this emphasis on habituation relate to the 'emotional pull'? One promising approach, outlined by Lorenz, is to think that this 'pull' is simply an attraction of some sort (whether to an emotion, action, etc.) that occurs when reason, as a capacity, is weakened.47 On this view, what happens is that a supposition is endorsed that reason, in its poor state, fails to oppose. This approach generates a distinction between different types of assent and between the cognitive vocabulary correlated to each. On the one hand, there is an active, rational, reflective type associated with judgement (krisis); on the other, a more passive 45 See F18EK and T85EK. 46 It is notable that the Stoics denied doxai to the sage on the basis that opinions are 'weak' and 'changeable', cf. SVF 3.548 and SE M. 7.151-57, with LS 41. 47 Lorenz 2011, especially 191-94. type connected with a less stable supposition (hupolêpsis) and belief (doxa).48 Habituation, then, is most plausibly concerned with the improvement of this second, more passive type of endorsement and works by weakening the power of the affective pull's attraction. This occurs through pre-rational education, which is 'defined by the recognition of the cause of the emotions' (F168EK), as well as the passage of time (passim), and 'dwelling on things in advance' (F165aEK).
This division between types of assent takes us in the right direction. It neatly captures what seems to be behind the difference between yielding (εἴξις) and assent (συγκαταθέσις) attributed to Chrysippus and the Stoics generally.49 Notably, for our purposes, this is the very distinction that has been used to reconstruct how the Stoics explain how non-rational animals can endorse their impressions absent rational assent.50 It is true that assent is sometimes attributed to animals generally (SVF 2.991=LS53O), but it is also strongly associated with reason (SVF 2.826=LS53K). Accepting animal εἴξις as a weak, non-rational sort of assent, or endorsement, provides a very attractive solution to the conflicted evidence.
Unfortunately, we lack direct support for attributing εἴξις to animals, but we might proceed on the firmer basis that Lorenz is right to think that Posidonius accepted that acts of endorsement are not uniform. Yet it is not clear to me that understanding one as active and the other as a more passive sort of endorsement gets the difference right. Galen's report indicates that the false suppositions involved in the emotions come about because the reasoning part fails to oppose this process. Certainly, this suggests that reason has yielded or become passive. Yet this is also obviously not what is interesting for Posidonius in his debate with Chrysippus. What is at issue is how emotional impulse can come about when the reasoning part has become weakened. That the soul has been so weakened and that this is a condition of moral error is agreed by all sides (PHP 4.6.2). So, the explanandum is not how or why reason yields, but what operates in its place. This is most plausibly, on the evidence we have assembled, a non-rational cognitive element that Posidonius identified with the persistent animal aspect of the soul. The affective pull is the non-rational, merely animal, part of the human soul in action.
48 Lorenz 2011, 197-202. He tries to show that Chrysippus, evidenced by his interest in the persuasiveness of impressions (PHP 4.7.16 and 5.5.19), also accepted these two types of assent. This is plausible enough. What distinguished Posidonius is that he connects the weaker sort with an animal aspect. 49 SVF 3.177 and SVF 3.459=LS53S and LS65G. See, too, DL 7.51. 50 See Long 1982 andLS 1, 322, with Inwood 1985. In fact, the distinction Posidonius has in mind between two types of assent or endorsement is well attested in the later Stoic tradition. Consider how Epictetus uses 'supposition' (hupolêpsis) in his Encheiridion (20): Μέμνησο, ὅτι οὐχ ὁ λοιδορῶν ἢ ὁ τύπτων ὑβρίζει, ἀλλὰ τὸ δόγμα τὸ περὶ τούτων ὡς ὑβριζόντων. ὅταν οὖν ἐρεθίσῃ σέ τις, ἴσθι, ὅτι ἡ σή σε ὑπόληψις ἠρέθικε. τοιγαροῦν ἐν πρώτοις πειρῶ ὑπὸ τῆς φαντασίας μὴ συναρπασθῆναι· Bear in mind that it is not the man who reviles or strikes you that insults you, but it is your judgement that these men are insulting you. Therefore, when someone irritates you, be assured that it is your own supposition which has irritated you. And so make it your first endeavour not to be carried away by the external impression. (Trans. Oldfather, modified).
'Supposition' here is one's automatic response to an external impression -something that has the power to instigate an immediate reaction.51 It has not yet been subject to rational reflection; this process, it seems, will reveal that one's initial view of the externally caused impression was unnecessary and of one's own making.
This distinction between what seems to be true according to an automatic response and what is really the case on reflection underpins Epictetus' principal theme that reflection on one's impressions is what is truly in one's power. Such an approach also seems to lie behind the frequent self-reproach we find in Marcus Aurelius' Meditations 'to erase your impressions'.52 Notably, for both Epictetus and Marcus, the ability to subject one's impressions to scrutiny and to act as their interpreter is precisely what divides rational from non-rational animals. The former are capable of withholding assent and engaging in reflective contemplation; the latter are not.53 Connecting Posidonius with these later discussions suggests that a limited sort of non-reflective cognitive ability in rational adults is associated for the Stoics with mere animal psychology. Marcus' 'to transpire like plants or to breath like cattle or wild beasts is not a thing to value, nor to be stamped by sense impres-51 Just how automatic the judgement is that occurs is not specified by Epictetus. Perhaps the reaction he has in mind takes some time to develop. However, Epictetus was obviously concerned with the difference between reflective and unreflective assent (see his Fr. 9), i.  Berryman 2010, 198-202. See Brittain 2002 fn50, for an emphasis on the automatic character of animal cognition. sions or drawn by the strings of impulse, nor to live in herds or take in nourishment -this last is on a level with relieving the body of the dregs of that nourishment' suggests this connection.54 However, where I take Marcus to be making a metaphorical point in making such a comparison, Posidonius pointedly accepts a non-rational cognitive aspect within the human soul identified as its animal part. This is consistent with the evidence we have elsewhere, including Seneca in his Ep. 121 and in Hierocles' Elements of Ethics.55 Thought is denied to mere animals,56 as are senses of the future and the past.57 Yet it is also clear that some conscious self-perception is to be attributed to non-rational animals (Seneca, Ep. 121.10),58 as well as goal-directed activity more generally.59 There also seems to be a central co-ordinating faculty able to bring together sensory inputs (Calcidius, 220).
Such non-rational abilities and the complexity of non-rational mental representations have been subject to wide recent study,60 and it is increasingly recognised that the contents of animal minds have a great deal more richness than previously understood. Much of this debate has turned on whether the contents of non-rational representations are propositional in nature, i. e. mental or intentional states that something is the case despite the Stoic restriction of concepts to rational humans. In short: Is conceptualisation or a correspondence between a lekton (a sayable) and a rational impression needed to give content to mental representations? Many interpreters answer affirmatively to this,61 others have insisted on representational complexity without propositionality,62 and some argue that non-rational animals may indeed entertain a proposition, or something that would amount to a proposition in the case of rational animals.63 This is no idle worry for the interpretation we have been canvassing. If the sort of mental representation I attribute to the animal aspect of the soul depends on the possession of concepts (ennoiai), there is the immediate tension that for the Stoics such concepts seem the hallmarks of reason and exclusive to the souls 54 III.16, trans. Farquharson. 55 See Long 1996, 256-61, andBrittain 2002, 266- Labarrière 1993 and1997;Sorabji 1993, 20-28, 40-44, 58-60;Brittain 2002, 256 f.;Klein 2016;and Caston 2021. 61 E.g. Frede 1983, particularly 153-57, andSedley 1987, 240. See Caston 2021, 23 fn43, for a comprehensive overview of the debate. 62 Brittain 2002, 258 f. See also Klein 2016, 174. 63 Sorabji 1993 of rational animals. Reason, Galen reports, is a 'collection of certain concepts and preconceptions', suggesting that it is only rational animals that possess these.64 If this tension cannot be dissolved, accepting the primary claims made here also entails accepting a great deal of innovation on Posidonius' part of the Stoic understanding of the mental functioning of the soul and what rationality consists in.
However, we need not assume any real heterodoxy on Posidonius' part in how the mental representations of non-rational animals are thought by the Stoics to gain the complex content needed to explain their behaviour. Without going into excessive detail, I take it that there is good reason to follow Victor Caston's contributions on this issue.65 He rightly notes the difficulty of explaining the Stoic view of goal-directed animal behaviour and of the pre-rational development of children if only adult, rational representation has content as a product of conceptualisation. Rather, he argues that the content of an impression or mental representation is a causal property of an effect on an observer that operates inde pendently of whatever concepts they happen to possess, or indeed whether they possess any at all. Non-rational representations then have their content in the same way as rational ones do as causal effects of the interactions between bodies that make predicates true of the bodies in question.66 This fits very well with the report in Diogenes Laertius (7.49) that mental representation takes priority over language or conceptualization.67 Such content might not rise to the level of 'thoughts' in non-rational animals,68 but this in no way entails that it is hopelessly impoverished.
In any case, there is some evidence in the testimonia of Posidonius that mental representation is prior to either language or conceptualisation. In a report 64 PHP 5.3.1=SVF 2.851 (part). 65 Usefully surveyed in Caston 2021, especially 20-35. See, too, Sorabji 1993. Gourinat 2018, 151-63, particularly 157-60, offers a parallel analysis. He very attractively explains how sayables may be said to subsist in thought for the Stoics without making these either mind-dependent or linguistic items. 66 LS55B-D. 67 Now, accepting this interpretation entails that we also accept a particular view of impressions and their connection with lekta that is far broader than the more orthodox position limiting the connection of lekta to rational representations. This depends on a 'generous' view of DL 7.63 ('the Stoics state that a lekton is what subsists in accordance with a rational representation') that allows also for lekta that correspond to non-rational impressions. See Sorabji 1993, 24 f., for a convincing argument for this more expansive view, prefiguring Caston's discussion. Gourinat 2018 is useful here. 68 Cf. DL 7.51=LS 39A6. on the rhetorical theory of status (στάσις, the boundaries of what is at issue in a given case), Quintilian preserves the following classification:69 Posidonius too divides status into two: language and facts. Under language he thinks come the questions: 'Does this have any meaning?'; 'What is its meaning?'; 'How many meanings has it?'; 'And how?' Under fact he classifies conjecture, which he says depends on sense perception (κατ' αἲσθησιν); 'quality', that is definition of the fact, which Posidonius names as conceptual (κατ' ἔννoιαν); and 'relation'. (Trans. Kidd) Clearly, this is a specific discussion of rhetorical theory and classification. As such, we might hesitate to draw too firm a conclusion. However, Stoic logic and their theory of meaning places lekta and their subsistent relationship with corporeal states of the soul at its centre.70 This ties their theory of meaning, elaborated within the domain of logic, directly to their views on the soul. Significant, too, as Kidd notes, is the fact that rhetoric for the Stoics was one half of logic alongside dialectic, and here Posidonius seems to be operating using the terminology of the latter.71 Stoic logic -a part of their philosophy more broadly construed than its modern counterpart -then bears directly on the elements of epistemology and psychology with which we are concerned. With these points in mind, Quintilian's report -especially his striking preservation of Posidonius' Greek terms (κατ' αἲσθησιν, κατ' ἔννoιαν) -seems germane to our discussion.
'Conjecture', a term Strabo's reporting connects with the carefully qualified statements Posidonius makes on Homer, for example, and the Thracian Mysians (T88EK), and the cycle of the tides (F217EK), seems independent of, and prior to, conceptualisation as simply the result of sense-perception. It is only in the next step of forming a definition, something tied closely to conceptualisation for the Stoics,72 that concepts then become central.
Does this suggest that concepts are irrelevant to such statements of belief as 'conjecture'? It is difficult to see how concepts tout court could be excluded here. Articulating beliefs on the basis of available evidence seems a prima facie strong candidate for something that involves concepts. Yet we need not think that simply involving concepts is what Posidonius has in mind by κατ' ἔννoιαν. The emphasis on the formation of definitions suggests it is something more robustly rational than mere concept possession. After all, the account of oikeiosis the Stoics adopt seems to imply, as discussed above, that children may acquire concepts prior to 69 Institutio oratorica III.6.37=F189EK. 70 LS33. 71 Kidd 1988, 687. 72 LS32D and F. The latter report from Augustine confirms that the senses lead the mind to conceptions, which in turn can be articulated by definitions. full rationality. Yet this does not entail that they are capable of the abstraction needed to form general definitions. We might, then, think that κατ' ἔννoιαν suggests the stronger claim that the stable and integrated 'collection' of concepts, which the Stoics identify with reason, is what Posidonius means.
On this view, what demarcates what is κατ' αἲσθησιν from κατ' ἔννoιαν mirrors the divide between weak and strong endorsement discussed above. What is 'according to sense perception' falls short of stable understanding and the complete possession of the concepts that the Stoics associate with reason and rational reflection. It is something akin to the sorts of cognition children and non-rational animals experience in being preliminary (i. e. not the product of a fully developed rational capacity) and fallible. As Seneca discusses animal perception in Ep. 121, it is uncertain and not fully clear. What is κατ' ἔννoιαν is concerned with definition, and this is a product of abstraction and thus of rational understanding and cognitive stability. The idea, once again, is not whether concepts are somehow entailed but whether they are stably and completely held by, or available to, the agent.
If this is along the right lines, attributing enough mental content to non-rational souls to allow for some (automatic) cognitive processes seems both plausible for the case of Posidonius and the most convincing approach for the Stoa as a whole. This, I take it, is the very reason why Posidonius advanced his 'emotional pull', and why he connects it with non-rational animals: it functions as a cognitive aspect characterised by its reflexive responsiveness, weak relation to its correlated supposition, and its lack of rational reflection. Such characteristics, Posidonius suggests, are proper both to non-rational animals and imperfect humans. 'Reason' then is less a matter of possessing concepts than of possessing them as a collection, as Galen reports it. Reason allows for reflection, discernment, testing for consistency, and judgement. For Posidonius the emotions are generated, at least in some cases, through weak supposition and untutored reflexive response, i. e. they are contrary to reason because they lack its distinctive reflective aspect.
On this account, habituation then may serve to sever the link between sense impression and the 'emotional pull' and thus it allows for the operation of the activity of reason. But it is not habituation as such that achieves this. Of course, habits may reinforce ignorant beliefs and prompt foolish assents. However, what Galen's evidence preserves is that Posidonius realised the power of habituation to achieve, in some, empirically supported cases, the opposite effect, i. e. the weakening of the connection between phantasia and impulse. It is in demonstrating this phenomenon that the examples of the greater affection that comes to the inexperienced (F164EK) and the trainer who allows young animals to wear out their desires (F166EK) become relevant. Presumably, this is also why Posidonius accepts the value of 'living with things in advance' as a means of reducing 'emo-tionally disturbed movements'. These are caused, according to Galen's reading of Posidonius, not only because of the 'opinion of present evil' but also because they are 'fresh'.73 Cognitive states, then, including opinion are relevant in the explanation of pathê but are, by themselves, insufficient from the perspective of Posidonius' aim to offer a complete account of the emotions.
Looking ahead: one central question needs to be answered to sustain such an interpretation: How does such a persistent aspect bear on the Stoic account of the rational soul as the differentia of the human species? I shall argue that our evidence for Posidonius on this question further supports the view we have canvassed above.

Human and Animal Souls
It is worth determining what the persistence of an animal aspect would mean for the Stoic view of the human, rational soul. To distinguish this rational nature from that of everything else in the cosmos, the Stoics developed a hierarchy of capacities (hexis, phusis, psuche, and logos) that served to provide the conceptual framework for a scala naturae.74 This hierarchy accounted for things in the world by reference to ascending states of tensional pneuma and the corresponding qualities and capacities that make each thing (and type of thing) what it is.
Referring to a non-living object's hexis analyzes the pneumatic state that gives it structure and unity; plant life is accounted for by phusis; animal capacities by psuche; and finally, rational humans by logos.75 It is clear that when moving up to the analysis of higher beings the capacities of the lower orders that are held in common are modified by their new reference but are also presumed to persist in some form. As Philo puts it, 'phusis is hexis in actual motion, psuche is phusis with impression and impulse added'.76 However, there has been a well-founded worry that a picture of total subsumption of the lower functions by the umbrella of the higher fails to capture the important difference between automatic vegetative capacities (e. g. growth) and the rational, self-directed powers of the soul.77 The interpretative crux noted in the literature is to be found in the transition from phusis to psuche.  Inwood 2014, 64-67. question: do some aspects of phusis simply remain present after the plant-like embryo is transformed at birth into the animal soul?78 This discussion closely mirrors the distinction in the study of Aristotle's anthropology and elsewhere between 'additive' and 'transformational' theories of human/animal differentiation. The latter holds that the lower capacities (vegetative, locomotive, etc.) are fundamentally altered at the level of the human and not simply enhanced by the addition of reason.79 On the 'additive' interpretation, certain capacities shared between animal and non-animal (growth, nutrition, etc.) can be construed as fundamentally alike across hierarchical levels, e. g. that the nutrition capacity of non-rational animals is meaningfully akin to that of humans.
In the case of the Stoics, the evidence is somewhat heterogeneous, but the 'additive' view receives the strongest support from the evidence extant. Ju, for example, has persuasively argued that Calcidius' (SVF 2.879=LS53G) inclusion of growth and nutrition, alongside locomotion, sensation and impulse, within the soul mistakes these first two functions, properly accounted for by phusis, for genuine soul parts.80 Such evidence, taken at face value, would seem to suggest that the lower order capacities of humans become properly subject first to psuche at birth and then to logos at the point of the acquisition of the capacity of reason. However, as Ju has noted, there is strong textual evidence, including DL 7.86, that confirms the Stoic view that phusis remains active in animals and is responsible for vegetative processes.81 So then, ascending the scala does not imply that lower-order capacities are jettisoned or wholly modified at each move upwards.
For our purposes, what is important is that there is good reason to think that the details of the responsibility of logos for all human qualities, rational and non-rational, was at the very least unsettled for the 'Middle' Stoics. Posidonius' teacher, Panaetius, is a case in point. Nemesius reports (LS53I) that Panaetius took the vocal faculty to be governed by impulse (hormê), while the reproductive faculty was to be attributed not to soul but to phusis. This represents a modification to what seems to have been the traditional picture, as presented by Aetius.82 78 See Hierocles 1.5-33, 4.38-53=LS53B. 79 See Glock 2019, 155-57. 80 Ju 2007. At 105 f. Ju notes, with Long 1982, 41, that 'soul' for the Stoics is used in several ways in the evidence, and that we should accept that it is only in the specific sense of 'soul' as the hêgemonikon in their eight-part theory (below) that 'soul' is marked out from 'nature'. But as it is just this that is operative in their scala naturae, the wider Stoic use of the term is only peri- In this testimonium, the rational hêgemonikon is said to produce impressions, assents, perceptions, and impulses, with seven soul parts reaching out like tentacles from the commanding part, accounting for the five senses, reproduction, and voice. The latter two, on this account, are attributable to psuche. The evidence for Panaetius, in contrast, suggests that the vocal faculty is elevated to what is proper to logos, i. e. to humans, while reproduction lowered from psuche to phusis.83 Direct evidence for Posidonius on the hêgemonikon is harder to pin down but not entirely absent. Tertullian (F147EK) preserves a textually problematic but striking account of Posidonius' division of the soul into parts put into comparison with the numbers accepted by other Stoics.84 […] but the soul is also divided into twelve parts by certain Stoics, and into two more by Posidonius, who, starting from two labels, 'governing' (which the Stoics call hêgemonikon) and 'rational' (called by them logikon), proceeded to cut up the soul into seventeen parts. (Trans. Kidd) The relevance for us of the testimonium is not the number of parts but its report that Posidonius classified aspects of the soul by making a distinction between the governing (hêgemonikon) and the rational (logikon), then proceeding from this heading classification. What is governing, then, is not to be identified, at least exhaustively, with what is proper to reason.
The significance of this distinction might be clearer if we compare what seems to have been the more standard Stoic account. Plutarch reports: "For appetite and anger and fear and all such things are corrupt opinions and judgements, which do not arise about just one part of the soul but are the whole commanding faculty 's inclinations, yieldings, assents, and impulses."85 In this account of the operation of the hêgemonikon, rational capacities are undertaken by a unified governing part.86 Tertullian's report on Posidonius suggests, in contrast, that the 83 Possibly also relevant is Diogenes of Babylon, a Stoic sometimes supposed to have accepted a dualist account of the soul (Obbink and Vander Waerdt 1991), who wrote a book on the hêgemonikon. For discussion, see Alesse 1994, 163-217. See also PHP 2.8.40-44, with Tieleman 1996 The number of soul-parts he accepted is reported to be seventeen. This seems to contradict an earlier sentence where Posidonius is said to have added two more to the twelve parts other Stoics accepted. How to make the arithmetic work has yielded a number of suggestions. See Kidd 1988, 546-48, for a discussion of the proposals on offer. 85 SVF 3.459=LS65G. Trans. LS. Emphasis mine. 86 Cf. the Stoic evidence that the soul is identified, as a whole, as a daimon: Sextus Empiricus, M 9.74 with DL 7.87-89, and the commentary of LS, vol. 2, 391. However, as Galen's report (PHP 5.6.4=F187EK (part)) on the cause of the emotions for Posidonius, discussed above, makes clear, activity of the governing part of the soul is (a) not exhausted by such activities and the 'rational' is not the same as the 'governing', and it leaves open the possibility (b) that the capacities undertaken by the unified, rational governing part for Plutarch's Stoic may be attributed to the merely 'governing', but non-rational, part for Posidonius. For our purposes, the significance of this report is that it explicitly leaves room open for the governing part to be responsible for impulse that is not to be connected with reason. This is consistent with holding that the human hêgemonikon may be simultaneously responsible for activities proper to either psuche or to logos.
We should be cautious not to take the evidence of Tertullian too far. The motivation for the division he reports is not specified and other criteria for partition are possible.87 For example, Nemesius, a reporter of similar divisions of the soul in the Stoa and elsewhere, attests to the variety of ways psychic aspects may be classified, including those functional, physical, and moral.88 Yet, even with this note of caution in mind, there is evidence that Plutarch's picture of a unified soul undertaking 'psychic' activities as a whole needs qualification for the Stoics generally. Certainly, such activities of the soul (e. g. impression and impulse) are the proper function of the hêgemonikon.89 Yet it is not the case that this commanding part is to be simply identified with reason or the rational, even in adult humans. Indeed, even if the hêgemonikon as a whole may be said to produce the activities of reason, this does not amount to the view that each of its parts or aspects is so responsible. Diogenes Laertius, for example, reports the Stoic definition of hêgemonikon in a way that makes reason a product of its capacities: The hêgemonikon (ruling part) is the dominant part of the soul; it is where presentations and drives arise, and the source from which reason (logos) issues; it is located in the heart. (7.159).90 our in-born daimon is capable of being repressed and ignored. Surely, it is the soul in some form that allows for such transgressive action and a failure to obey. If we accept that the ruling-part must have a role in this disobedience, as I think we must, there is good reason to doubt any simple identification of the hêgemonikon with one's daimon. 87 A reviewer helpfully notes that those following the Placita tradition (including Tertullian) draw from distinctions made for various reasons; sometimes these are unknown to the reporter. Perhaps we have a distinction between moral and physical categories here; other motivations are also possible. However, the report is worth considering at face value. 88 Cf., Nemesius,Nat. Hom.,15 f.,with 26. 89 LS53G and H. 90 This translation owes something to White 2021, but I doubt 'logos' means 'speech' here. Rather Diogenes is reproducing the standard Stoic view that the ruling part issues the products What Plutarch's testimony claims is that the Stoics refuse to acknowledge psychic conflict, which he takes to be evident. This is the context in which his passage quoted above appears.91 But the Stoic point he notes does not so much insist on a monistic conception of the soul as much as it seeks to show that there is a particular form of the relationship between the commanding-part and its psychic activities. This relationship makes such activities or faculties qualities of the command ingpart rather than its parts.92 While Plutarch seeks the identification of a part, the Stoics take such a notion to be absurd because qualities are not manifested in identifiable sub-regions of their substrate but in the tensional state of the commanding-faculty as a whole. As Iamblichus points it, an apple may possess sweetness and fragrance in the same body, just as the commanding-part may combine the activities of impression, assent, impulse, and reason.93 Galen's criticism of Chrysippus is apt here. At PHP 5.2.49-5.3.11, Chrysippus is attacked, fairly or not, for allegedly taking the soul's activities (energeiai) to be both its parts and also what constitutes reason. Galen connects these activities with Chrysippus' definition of reason as 'a collection of concepts and preconceptions'. He is 'twice wrong', Galen claims, because (1) such parts must be of reason and not the soul because 'the soul and reason are not the same'.
(2) Reason may be constituted in the soul, but the soul and what is constituted in it are not the same.
We may doubt the fairness of Galen's reading here, but his verbatim quotation of Chrysippus is telling: They are the parts of the soul through which its reason and the character of its reason are constituted. A soul is noble or base according to the state of its commanding-part with respect to its own divisions. (5.5.49 f., Trans. LS).
The 'governing' and the 'rational' are not identical but related as a source is to an activity, or a body to a quality. This is because reason is a capacity which is realized as a quality of the governing part of the soul. If this is right, there is nothing of reason: cf. the Aetius report at SVF 2.836=LS53H. As LS (pg. 315) translate: 'The Stoics say that the commanding-faculty is the soul's highest part, which produces (ποιοῦν) impressions, assents, perceptions and impulses. They also call it the reasoning faculty.' 91 Plutarch, On moral virtue 446E: Ὅτι μὲν οὖν γίνεταί τις ἐν τῇ ψυχῇ τοιαύτης ἑτερότητος αἴσθησις καὶ διαφορᾶς περὶ τὰς ἐπιθυμίας, ὥς τινος μαχομένου καὶ τἀναντία λέγοντος Fαὐταῖς, οὐκ ἄδηλόν ἐστιν. 92 SVF 2.826=LS28F. For qualities of bodies as themselves corporeal, see Simplicius, On Aris totle's Categories 217.32=218.1=LS28L. 93 SVF 2.826=LS53K. See Inwood 2014 for an important discussion of this passage and Stoic soul-divisions, more generally. essential to the hêgemonikon of the adult human soul that prima facie prevents it from also being responsible for a non-rational aspect that gives rise to the emotions. This is true whether or not Stoics prior to Posidonius recognised the possibility. All that Chrysippus' view requires is that the relation of the hêgemonikon's aspects (whether or not they are called 'parts') to each other is such that reason may be successfully realised in the wise. That each part, as Plutarch implies, contributes similarly to impressions etc. is not entailed. Of course, in the case of the wise, where reason is strong and fully in control, such an aspect would be subject to the infallible rationality characteristic of the sage. But Posidonius' concern is with the mere proficiens, vulnerable to weak beliefs, and under the control of the 'emotional pull'.
On this account, the evidence for Posidonius suggests that the difference between the hêgemonikon in non-rational animals and humans is not one of kind. Rather, the relevant point is that reason is added to the commanding part of the soul in the case of adult humans as something they are capable of engaging. But this addition does not, on this account, make all functions of the adult human hêgemonikon subject to such an addition. The fragments suggest that Posidonius insisted on a persistent non-rational animal aspect of the soul, even in the rational. As we have seen, however, this does not demand that such an aspect is sundered from the cognitive faculties of supposition and belief.
Two final pieces of evidence may be adduced for the 'additive view'. The first comes from Galen's presentation of Posidonius' own approach to the Stoic scala naturae, the second from Plutarch's On Moral Virtue: (1) He (Posidonius) says that all those that are least mobile and are naturally attached like plants to rocks and the like, are governed by desire alone; all other irrational animals use both powers, the desiring and the spirited, whereas man alone uses all three, because he has acquired the rational ruling principle. (F33EK, Trans. Kidd).
(2) And in general, both as my opponents (i. e. the Stoics) themselves admit and as is quite obvious, in this world some things are governed by an acquired disposition, others by a natural one, some by an irrational soul, others by a rational and intellectual one; and in practically all these things man participates and he is subject to all the differences I have mentioned. For he is controlled by his acquired disposition, nurtured by his natural disposition, and makes use of reason and intellect. He has, therefore, some portion of the irrational also and has innate within him the mainspring of emotion, not as an adventitious accessory, but as a necessary part of his being, which should never be done away with entirely, but must needs have careful tending and education. (451b-c=SVF 2.460, part, trans. Helmbold).
The first passage is quite clearly filtered through a Platonising tripartition, where 'desire' and 'spirit' have been grafted onto the standard Stoic hierarchy of capa-cities.94 Having said that, we should not rule out that Galen is preserving Posidonius' view of reason as something that is operated in addition to other elements in the soul. As we have seen, this is consistent with the general (Chrysippean) Stoic view, if properly qualified with the caveat that only judgement can give rise to impulse. Everything in our evidence so far leads us to accept that there is room for activities of the adult human soul that are not properly subject to reason. Posidonius is able to capitalise on this to further his explanatory account of the emotions as the product of the non-rational animal 'part' of our souls.
The second passage does not mention Posidonius or any Stoic in particular; however, the parallels with Galen's account of Posidonius should be obvious. We might, on that basis, dismiss such evidence as simply a common polemical attack on Stoicism. However, even on a sceptical approach it should be admitted that such an attack assumes that an additive view of rationality is reasonably attributed to a Stoic that looks undeniably much like Galen's portrait of Posidonius. It cannot then be denied that attacking the Stoics for accepting an innate, irrational aspect of the soul responsible for the emotions had some polemical plausibility. Our collection of evidence suggests that this plausibility is grounded by Posidonius' own account.

What is the Non-Rational Aspect?
What I have tried to suggest above is that there is a strong case for seeing Posidonius' approach to the emotions within the context of an ongoing Stoic debate about the classification of human and animal aspects under the rubric of their scala naturae. Preserving a permanent animal element of human nature responsible for instigating impulse ought then to be understood as Posidonius' contribution to this larger discussion.
Above, I have used the non-specific place-holders 'aspect', 'element', and 'part' to refer to this non-rational component. Can we be more precise? As we have seen, Galen insists that it is a capacity or power (dunamis), following the Aristotelian model, capable of generating an emotion. 95 We have also noted that 94 See Tieleman 2003, 214-20. 95 See, for example, F146EK. Here Galen suggests that Posidonius and Aristotle opt for the language of capacities of the soul, while Plato prefers 'forms, or species and parts'. The distinction, however, is not made on the basis that an Aristotelian or Posidonian capacity is more of a psychic faculty than a Platonic 'part'. Rather, it is because Plato's parts of the soul are spatially distinct in the body with peculiar locations.
the absence of such a description from Galen's verbatim quotations of Posidonius has rightly encouraged the criticism that such a connection may be Galen's own and a misleading interpretation of his material.
Here I attempt to sketch out, in outline, a reading of the evidence that incorporates such an aspect of the soul within Posidonius' philosophical project while also attempting to isolate what Galen found in the text that encouraged the strong connection he makes with Platonic tripartition. Put another way: There might be a way of avoiding the attribution of Galen's discussion of capacities to Posidonius which also helps to explain why Galen was encouraged to develop the interpretation he offers.
The first point to note is that it is undeniable that Posidonius sought to further the understanding of causes within philosophical explanation.96 This aim he applied to the emotions in particular: Advancing a non-rational capacity is one approach to specifying such a cause, but hardly the only one. What Posidonius really seeks is a cause to serve as a locus of responsibility for explaining the emotions which accounts for the empirical phenomena Galen describes as motivating the criticism of the Chrysippean account. For Galen this source of responsibility may be most easily explained in terms of a capacity, but this is not Posidonius' approach. Rather it is in the char acter of the relationship between the objects of cognition and the aspects of the soul themselves that such causes are to be found. Such a suggestion has particularly strong explanatory power for us because there is suggestive evidence that it is on the basis of Plato's Timaeus that he developed such a view.
Consider Posidonius' discussion of logos when interpreting this dialogue. Minimally, the analogy drawn here, using the principle that like is cognised by like, establishes that reason has its own peculiar object and must be distinguished from other cognitive processes. Whether Posidonius actually accepted such a view of vision and hearing or is merely developing an analogical point using Platonic materials to describe reason is beyond our scope. However, the connection made here between the 'nature of the wholes' and our logos, which is kin to the former, we have already seen (above, section 1a) Posidonius establish in F187EK to explain the cause of the emotions. In that verbatim fragment, we are told that following our daimon allows for the realisation of our kinship with the nature that governs the universe. Embracing this kinship prevents us from being swept along by our animal aspect. Thus, as Kidd suggests, τῶν ὅλων φύσις must be the Stoic active principle.98 Human reason finds its object in this through a kinship just as sight does in what is luminous and hearing in what is airy. But Posidonius' point is even stronger here in Sextus' passage. In Galen's report at F187EK our daimon has a similar φύσις to the one "which governs the whole universe" (τῷ τὸν ὅλον κόσμον διοικοῦντι). In Sextus, however, we find τῶν ὅλων rather than τὸν ὅλον. The difference might appear slight, but it is significant. In its singular form, τὸν ὅλον (the whole) refers for the Stoics to the created cosmos and was distinguished from τὸ πᾶν (the all), which also incorporates the external void.99 Although it is not specified, the latter is said to be infinite, while the former finite, suggesting that 'the all' might also refer to the infinite number of Stoic world-orders implied by their theory of everlasting recurrence.
What does this suggest about our non-standard plural τῶν ὅλων? If nothing else, it should indicate that we are to be careful not to accept that Posidonius simply means the plural to stand in for the singular.100 In fact, we find two other notable uses of this plural in our fragments; in both it is the power of human rationality that is paramount. In F186EK (above, section 2), we find τῶν ὅλων figure within Posidonius' novel incorporation of contemplation in his description of the human telos.101 In a second parallel, at F92EK, the plural is used in a report on Posidonius' discussion of how the substance of the whole (τὴν τῶν ὅλων οὐσιαν) relates to matter (ὕλν). This is a difficult passage to construe confidently, but the distinction Posidonius is said to make between these two in 98 Kidd 1988, 343. Contra Ju 2012. 99 LS44A. 100 Kidd 1988, 672, makes this point for F186EK but seems to ignore its implications for F85EK. He does, however, note the parallel. 101 To live contemplating the truth and order of all things (θεωροῦντα τὴν τῶν ὅλων ἀλήθειαν καὶ τάξιν) thought (ἐπινοίᾳ), as opposed to reality (ὑπόστασις), clearly turns on the operation of reason at the greatest level of generality.102 If τὸν ὅλον is the ordered cosmos of bodies, τῶν ὅλων seems to extend this to refer to everything that is, perhaps including the non-corporeal items appealed to in this passage.103 The possibility is raised, then, that it is not just the corporeal bodies that constitute the ordered 'universe', but the nature of everything that is, including incorporeals and mere thought-constructs, that Posidonius aims to emphasise in using the plural.104 At the very least, the plural form occurs in contexts where absolute generality (i. e. what there is in general) is important to the content of the argument.
With this in mind, Sextus' report encourages us to not only identify the proper object of logos with the Stoic active principle but also to incorporate this point within Posidonius' broader discussion of the application of reason by the philosopher. It is this person, above all, who is capable of the contemplation and abstraction Posidonius describes. Sextus' report on the distinction between rational cognition and sense perception, made in light of the Timaeus, suggests a further distinction from that dialogue.105 At 27d5-28b1, we find the famous distinction between Being and Becoming put in epistemic terms: As I see it, then, we must begin by making the following distinction: What is that which always is and has no becoming, and what is that which becomes but never is? The former is grasped by understanding (νοήσει), which involves reason (μετὰ λόγου), being unchanging. The latter is grasped by opinion, which involves unreasoning (ἀλόγου) sense perception. It comes to be and passes away, but never really is. Now everything that comes to be must of necessity come to be owing to some cause, for it is impossible for anything to come to be without a cause. So whenever the craftsman of any object looks to (βλέπων) what is always changeless and, using a thing of that kind as his model, reproduces its form and character, then, of necessity, all that he so completes is beautiful. (Trans. Zeyl, modified. Emphasis mine).
102 For the interpretative possibilities of this passage, see Kidd 1988, 368-74. See F16EK=DL 7.135 for a further instance of Posidonius' use of the distinction between 'in thought' and 'in reality'. 103 It seems significant that the singular is used of the cosmos at F187EK and of earth at F49.6EK, while another plural form is found at F96EK in highly generalised discussion of the species of destruction and generation. 104 The ontological status of concepts or 'thought-constructs', e. g. fictional entities or mathematical limits, for the Stoics, is not fully clear; they are neither obviously bodies nor are they associated with their canonical incorporeals. It is clear, however, that Posdionius' use of ἐπινοίᾳ refers to such a thought-construct; cf. LS50D and E (F16EK). Generally, see LS, 165, for discussion. 105 Ju 2012, 107 f., followed by Opsomer 2012, 191 n70, note this connection.
The Posidonian reception of this passage has not been much explored, but it is beyond much doubt that it figured prominently for him. Notably, in his discussion of philosophical practice in F18EK, Posidonius distinguishes between the methodology of the philosopher and that of the astronomer in terms strikingly similar to what is found here: καὶ ὁ μὲν φυσικὸς τῆς αἰτίας πολλαχοῦ ἅψεται εἰς τὴν ποιητικὴν δύναμιν ἀποβλέπων, ὁ δὲ ἀστ ρολόγος ὅταν ἀπὸ τῶν ἔξωθεν συμβεβηκότων ἀποδεικνύῃ, οὐχ ἱκανὸς θεατὴς γίνεται τῆς αἰτίας, οἷον ὅτε σφαιροειδῆ τὴν γῆν ἢ τὰ ἄστρα ἀποδίδωσιν […] The philosopher will often fasten on to the cause, looking to its creative force. The astronomer, however, whenever he proves facts from external conditions, is not an adequate observer of cause, e. g. when he states that the earth or the stars are spherical […] (trans. Kidd).
Posidonius latches onto Plato's emphasis on cause by transforming the Timaean demiurge into the philosopher and directly borrowing the characteristic Platonic vocabulary of ἀποβλέπων/ βλέπων.106 In Plato, such language is associated with the transcendent Forms, and here with the παράδειγμα, but in Posidonius' hands it becomes connected to the Stoic active principle itself, here put into terms (τὴν ποιητικὴν δύναμιν) that further the connection between demiurge and philosopher. We have then a clear example of Posidonius Stoicizing a Platonic precedent for the sake of explaining the object of reason and giving us a model for its operation in the philosopher. We also find here the connection between reason and the active principle as cause further confirmed.
What is central for us is that Plato's invocation of the demiurge proceeds from the very epistemological division that Posidonius interprets using the Timaeus in F85EK. What is even more clearly put here is that logos is to be distinguished from unreasoning sense perception using the difference between their respective objects. Sense-perception has its object in the unstable Becoming; rational understanding has its in what is fixed and eternal. This fits nicely with our discussion above: the non-rational aspects of the soul are to be characterised as unstable, impermanent, and preliminary. Reason, however, is reflective and stable.
Posidonius, of course, had no desire to adopt the Platonic metaphysics of Being and Becoming, but one can see what attracted him to this discussion. This ontological distinction could be converted into a means of dividing rational from non-rational impulse in terms of the Stoic active principle. This is what is shared between human and cosmos; reason then finds its proper object in this cause, which is to be isolated. This action is performed by the philosopher in F18EK, just as the demiurge focuses on the eternal paradigm in the Timaeus. In both cases, losing the characteristic objects of logos or eternal paradigm as one's exclusive focus has disastrous results.
Thus contemplation of this active cause, understood as the nature of everything that there is, integrates reason within the telos Posidonius specifies. Further, if reason is to be so closely associated with the active principle, we should consider the converse: the passive principle of matter may connect with what is non-rational. This makes some sense. As we have seen in F169EK, for example, Posidonius' closely connects physiognomy with the emotions.107 Crucially, sense perception here in the Timaeus, though unstable and lacking reason, is explicitly cognitive on the model provided. It is the preserve of opinion (δόξῃ), while reason has its place in understanding (νοήσει). This very much reflects the cognitive model of F154EK, where Posidonius involves judgements and assumptions in the emotions generally. However, compare the very different model of sensation within the appetitive part of the soul at Timaeus 77b3-6: And, in fact, what we are talking about now partakes of the third type of soul, the type that our account has situated between the midriff and the navel. This type is totally devoid of opinion, reasoning or understanding (ᾧ δόξης μὲν λογισμοῦ τε καὶ νοῦ μέτεστι τὸ μηδέν), though it does share in sensation, pleasant and painful, and desires. (trans. Zeyl).
Here the appetitive part has no share in opinion. If this is right, the Timaeus provides more than one model of how sense-perception operates, with the earlier, cognitive model strongly associated with Posidonius. This also suggests an answer to our question about Galen's interpretative framework and his imputation of non-rational capacities to Posidonius' theory. If the Timaeus did provide fertile ground for Posidonius' treatment of the emotions, as I have suggested, and this account is a development of what he inherits from the Stoic tradition, we find a plausible explanation of what Galen provides. After all, it is no great leap from the claim that Posidonius pressed the Timaeus into his service to one that 107 Posidonius need not be thought to be adopting any Platonising suspicion of matter as such here. Rather the point is that a focus on the material aspect of any body (i. e. matter qualified by logos) comes at the expense of isolating its active cause. As it is only when the divine active principle qualifies matter into the form of human bodies that vice may emerge, such an association between corporeality and emotional wickedness is naturally made. This is supported by what is alleged to happen during the conflagration; evil is eliminated (LS46N), the gods are blended into one, and Zeus gives himself over to his thoughts (LS46O). interprets this to mean that Posidonius accepted Timaean psychic tripartition and non-rational capacities. As Galen was keen to find a Platonising critic within the Stoa, such a jump required little encouragement.

Conclusion
In the above, I have tried to ask what Galen's testimony on Posidonius amounts to if we put some of the methodological worries about his reliability to one side. I have not suggested that these concerns are unfounded; rather the point is that it is only sensible and fair to our evidence to find the point of departure for the Platonising tripartition Galen imposes. Galen does, after all, frequently claim to quote verbatim from the Posidonius' text. We have seen that it is by Posidonius' reading of the Timaeus that Galen's approach is best explained.
More broadly, I have suggested that the gap between Posidonius and Chrysippus on the emotions is not one that turns on different approaches to cognitive features of the soul, i. e. on whether or not the emotions depend on suppositions, beliefs, judgements, etc. The evidence is clear that such features are necessary for both in explaining how the emotions come about. Both also go to some lengths to distinguish stable judgement from weak supposition and apply the vocabulary of such cognitive states carefully.
Rather the difference is best taken as one concerned with reason and its relation to impulse. As Lorenz has argued, there is evidence that both Chrysippus and Posidonius worried about the conditions necessary for the endorsement of impressions that lead to emotional impulses. This seems to graft neatly onto the distinction between yielding (εἴξις) and assent (συγκατάθεσις) attributed widely to the Stoics. Posidonius' contribution to the debate is to posit a non-rational animal aspect of the soul that can explain how weakly formed beliefs and suppositions can give rise to impulses in the absence of reason. Posidonius, then, can give a rich account of the very real, and empirically supported, psychological phenomena he is concerned to describe. Galen may wish to construct an antithesis between the two, but this is ultimately illusory. Posidonius develops gaps in his Stoic inheritance; he seeks to perfect the system, not to destabilise it.108 It is notable that accounting for such things, including precipitate impulses and mental conflicts, remains both a central concern of the Stoic tradition and one where appealing to non-rational animal qualities and behaviours persists in their attempts at explanation. The solution canvassed here gives Posidonius a very neat response to this apparently long-standing explanatory demand.109