Psychological disease and action-guiding impressions in early Stoicism

ABSTRACT

The early Stoics diagnose vicious agents with various psychological diseases, e.g. love of money and love of wine. Such diseases are characterized as false evaluative opinions that lead the agent to form emotional impulses for certain objects, e.g. money and wine. Scholars have therefore analyzed psychological diseases simply as dispositions for assent. This interpretation is incomplete, I argue, and should be augmented with the claim that psychological disease also affects what kind of action-guiding impressions are created prior to giving assent. This proposal respects the Stoic insistence that impression-formation, no less than assent, is an activity of reason. In so far as the wine-lover’s reason is corrupted in a different way from the money-lover’s, the two vicious agents will form different action-guiding impressions when faced with similar stimuli. Here I juxtapose the Stoic account of expertise, according to which experts form more precise action-guiding impressions compared to the amateur, in virtue of possessing a system of grasps (katalēpseis). So expertise enhances, whereas psychological disease degrades, the representational fidelity of the impressions that prefigure action. With these commitments, the Stoics can be seen to offer a nuanced and principled theory of cognitive penetration and to anticipate some recent proposals in epistemology and cognitive science.

KEYWORDS:

• Stoicism
• vice
• impression
• cognitive penetration
• emotion

According to the early Stoics, human agents fall into two sharply opposed categories. We are either virtuous and happy, or vicious and miserable. Underpinning this stark dichotomy are the Stoic doctrines that virtue is necessary and sufficient for happiness, and that virtue is a disposition that does not admit of degrees. One drowns just as easily one foot from the surface as on the bottom of the sea – and similarly, the Stoics hold, an agent who has made some progress toward virtue, but does not yet possess it, will be as unhappy as one who has made no progress at all (Plutarch, Com. Not. 1063a-b; Cicero, Fin. 3.48, 4.64).

These commitments may lead one to expect a monolithic treatment of vice, but in fact Stoic theory recognizes different ways of being miserable and sets out to distinguish the psychological profiles of the non-virtuous. The Stoic founders fully accept the ancient commonplace that ordinary human beings cluster around the pursuit of wealth, bodily pleasure, and fame (Cleanthes, Hymn 23-29), and accordingly they attribute to vicious agents different defective traits of character (Gellius, 7.2.8; Cicero, Fat. 8; Tusc. 4.27-9).¹ To describe such traits, the early Stoics use medical imagery. Vicious agents are thus diagnosed with various proclivities, infirmities, aversions, and diseases of the soul.²

Given their psychological monism, the Stoic founders reject a Platonic-Aristotelian position on which vicious character traits are partly constituted by conditions of non-rational spirit or appetite – habituated tendencies for desires and feelings centred on social standing and bodily pleasure that arise independently of reason. Orthodox Stoic psychology recognizes no non-rational parts of the adult human mind, and so the early Stoics must understand psychological disease and related conditions as non-ideal configurations of the vicious agent’s rationality.

Consistent with this approach, the Stoics define psychological disease as a false evaluative opinion that has “turned into a hardened condition”, on the basis of which its possessor “thinks that things not worth choosing are extremely worth choosing”.³ As scholars have noticed, in characterizing the objects for which one has a psychological disease as ‘not worth choosing’, the Stoics have in mind what they elsewhere call the ‘indifferents’ – types of objects whose possession, they claim, makes no difference to one’s happiness and so, in the Stoics’ technical terminology, ought to be ‘selected’ rather than ‘chosen’.⁴ Thus we find love of money, women-madness, love of fame, and love of wine (among others) as examples of psychological disease.⁵ So insofar as vicious agents suffer from different psychological diseases, they will vary in the items they mistakenly regard as choice-worthy and for which they have developed an ingrained attraction.

The Stoics see no difficulty, of course, in claiming that psychological disease is both an impairment of reason and also a condition that affects the agent’s decision-making and behaviour. This again reflects their commitment to psychological monism: the Stoics identify reason as the sole source of action and impulse in human agents, and so assume that a corrupted reason brings about errors not only in cognition but also in motivation. The wine-lover, for instance, holds the false opinion that wine is extremely worth choosing, and consequently is led to form an impulse (hormē) to drink wine in circumstances where a different vicious agent, without this particular disease, would not.⁶ Indeed, these impulses of the wine-lover will often be excessive and recalcitrant and therefore count as emotions as the Stoic theory understands them.⁷

But how exactly does the possession of a false evaluative opinion – e.g. that wine is extremely worth choosing – result in the wine-lover forming an emotional impulse to drink? To answer this question and integrate the Stoic treatment of psychological disease into their more general psychological framework, recent commentators have settled on a promising interpretation that I will call the Assent Proposal. This interpretation analyzes psychological diseases as dispositions for assent: they are characterized as conditions of the agent’s mind which determine when she gives or withholds an act of assent (sunkatathesis).⁸ After all, on Stoic theory, action is constituted by impulse, and in human agents an act of assent is necessary and sufficient for any impulse to be created.⁹ Moreover, it is clear that the opinions an agent holds will inform her acts of assent (SE M 7.253-6), and, as we have seen, the Stoics identify psychological diseases with false opinions concerning the choice-worthiness of indifferents.

To illustrate the Assent Proposal, consider two vicious agents, only one of whom has the psychological disease of love of wine. At breakfast, both agents are offered a glass of Merlot, but only the diseased agent reaches for the glass and feels elated at its presence. The Stoics assume that neither this emotional response nor the action of reaching for the glass can arise without the diseased agent’s act of assent (Plutarch, St. Rep. 1057a). And so to explain the differential response of the two agents when confronted by the same stimuli, the Stoics can appeal to the distinctive element in the wine-lover’s psychological profile – her hardened false opinion that wine is extremely worth choosing – as the cause of her act of assent.

The Assent Proposal is correct so far as it goes, I think, but to deliver on its promise to fully incorporate psychological disease into the Stoics’ broader psychological theory, I will argue here that an addition must be made. According to my proposed amplification to the Assent Proposal, psychological disease also controls what kind of action-guiding impressions (phantasiai) the agent forms prior to giving assent. As we have seen, no human action or impulse arises without an act of assent. But one never gives assent in a vacuum, on Stoic theory: strictly speaking, what one gives assent to and affirms as true is always the propositional content of an impression.¹⁰ Some impressions are ‘impulse-prompting’ (hormētikē), the Stoics say, because giving assent to their content creates an impulse (hormē): these are the impressions that guide our behaviour, insofar as they are the states responsible for bringing into our awareness potential courses of action and the value of external objects and objectives. My proposal, then, is that, in addition to controlling when the agent gives assent to such impressions, psychological disease also determines when they are generated in the first place.

This proposed amplification needs to be argued for explicitly, since it is not entailed by the Assent Proposal, and no existing study unambiguously attributes to psychological disease a role in creating the vicious agent’s action-guiding impressions, or attempts to work out in detail how the Stoics could have defended this claim. To be sure, the Assent Proposal is compatible with the addition I propose here, but it is also compatible with an alternative account, on which vicious agents always form the same action-guiding impressions, regardless of which psychological disease they have, and differ only in which of these impressions they go on to accept. On this alternative account, action-guiding impressions are passive responses to the world, created without any interference from the agent’s particular profile of vice, so that the psychological effect of love of money or love of wine will be limited to the activities of assent. Thus the money-lover and the wine-lover will both form the impression that the Merlot at breakfast is fitting to drink, but only the wine-lover, because of her psychological disease, will give assent to it. By contrast, according to my proposal, the money-lover will not even entertain the impression that it is fitting to drink wine at breakfast, nor the wine-lover, to refuse to tip the waiter when the bill comes.

The choice between these alternatives is philosophically and exegetically important. In what follows, I will show that my account, more so than the alternative sketched here, does justice to the Stoic insistence that impression-formation is an activity of reason (DL 7.51) and, as such, a process that involves the deployment of concepts in a way that is sensitive to the opinions and beliefs the agent holds. The alternative account obscures these Stoic commitments and fails to respect that for the Stoics all of our impressions, including those that prefigure action, are cognitively penetrated. Since the wine-lover and money-lover differ in their psychological make-up and evaluative opinions, they will generate different action-guiding impressions in response to the same stimuli.

Proponents of the Assent Proposal have not discussed the interaction between psychological disease and impressions, and they do not decide between the two alternatives set out here. Brad Inwood, for instance, describes psychological disease as a ‘tendency to assent inconsistently’, which makes the agent “liable to fall prey to the persuasiveness of external things” (Ethics and Human Action, 164–5). But Inwood does not consider whether, before assent is given, psychological disease might also be responsible for these external things being represented in an attractive, persuasive guise in the agent’s action-guiding impressions. Similarly, Margaret Graver allows that psychological disease will influence “the agent’s implicit recognition of logical relations between his or her long-term beliefs and new impressions that arise” (Stoicism and Emotion, 140), but does not specify whether the influence of psychological disease extends also to determining what these ‘new impressions’ are like or under what circumstances they are created. Moreover, according to Tad Brennan, “Stoic psychology conceives of the human mind as an assent machine”, and so “the difference between the greedy person and the one who cares little for money can be captured by providing a complete list of the kinds of impressions to which their assents would differ” (The Stoic Life, 105). However, Brennan does not clarify whether this list also includes the action-guiding impressions entertained by the two agents prior to giving assent. If it does not include them, then it fails to capture all the psychological manifestations of love of money (or so I will argue).

In highlighting these remarks, I do not wish to suggest that the proponents of the Assent Proposal would deny the possibility of it being further amplified to take account of the role of psychological disease in generating the agent’s action-guiding impressions. Rather, I mean to emphasize that this claim does not obviously fall out of the Assent Proposal and so must be argued for separately, since, in general, we are not entitled to infer that whatever is true of assent is also true of our impressions. The Stoics sharply distinguish the activities of assent and impression as two different movements of the soul (peri tēn psuchēn kinēmata: Plutarch, Ad. Col. 1122b-c).¹¹ So insofar as it posits a connection between psychological disease and assent, the Assent Proposal cannot be taken to imply any specific account of the relation between psychological disease and our impressions: absent a further story, we cannot simply assume that the analysis of psychological disease as a disposition for assent will have consequences for the kinds of impressions the agent forms prior to giving assent on a particular occasion.

To help motivate my amplification to the Assent Proposal, I juxtapose the Stoic account of expertise. Expertise, as a system of grasps (katalēpseis), determines how the expert mind applies concepts as it forms action-guiding impressions. Consequently, the expert receives more precise representations than the amateur and thereby discerns the appropriate technical response to the objects in her domain of expertise. I will argue that psychological diseases, as false opinions concerning what is choice-worthy, also bring about a change in the agent’s pattern of concept application, but with the inverse effect on representational accuracy: the wine-lover’s action-guiding impressions are persistently skewed when depicting the value of wine. In this way, the Stoics uphold their claim that virtue is the only good and vice the only bad, while recognizing that ordinary agents – consistent with the particular psychological diseases that afflict them – differ in the items besides virtue and vice that their action-guiding impressions falsely represent as worthy of choice and avoidance. Indeed, our ancient evidence makes clear that psychological disease prompts the mind to apply the concepts good and bad incorrectly (e.g. to wine and sobriety in the case of love of wine, or to amassing wealth and giving it away in the case of love of money).¹²

The Stoics therefore acknowledge both beneficial and dangerous consequences of cognitive penetration. Expertise enhances, whereas psychological disease degrades, the representational fidelity of our action-guiding impressions. In these Stoic commitments, we find a forerunner of some contemporary proposals in epistemology and cognitive science. To explain the findings of the ‘Shooter Task’ experiment, for instance, theorists today have suggested that racist beliefs penetrate into and affect the content of perceptual experience.¹³ Because of the false beliefs the subject holds, men of colour appear to him to be more threatening than whites and thus incline him toward a violent response. The Stoics, as I interpret them here, would be attracted to this kind of account of how the presence of false evaluative beliefs interferes with the character of the mental representations that prefigure action: for them, psychological disease is a corruption of reason that manifests not only in the vicious agent’s acts of assent, but also in which evaluative claims and potential courses of action are brought to mind in the first place.

Action-guiding impressions and reason

To begin, I will present some basic elements of Stoic psychology, emphasizing their claim that, for human agents, impression-formation is an activity of reason.¹⁴ Then, to set the scene for my augmentation to the Assent Proposal, I will examine the structure of our action-guiding impressions in particular.

In general, impressions function as non-doxastic representational vehicles. They are information-carrying modifications of the soul – psychological states responsible for bringing objects and their features into the agent’s awareness (Chrysippus, SVF 2.54) – which do not yet imply any doxastic commitment on the part of the agent; for that, an act of assent is required in addition. So, for instance, suppose that a perceiver encounters a cat. The Stoics characterize the impression as the state that enables the perceiver to notice the cat and to attend to its sensory characteristics (e.g. its black colour). This kind of impression the Stoics call ‘perceptual’ (aisthētikē, DL 7.51), because it engages the subject’s sense-organs. Having been affected by the cat, the sense-organs in turn affect the subject’s mind, and it is this change – not the one which occurs in the sense-organs – that the Stoics identify as the impression (SE M 7.232-3). The Stoics also recognize that an impression can be ‘non-perceptual’ (ouk aisthētikē, DL 7.51), in the sense of being created in the mind without any activity of the sense-organs, for instance the impression that there is no afterlife.

Whether perceptual or not, however, impressions are always formed in the part of the soul possessing reason – i.e. in the mind (dianoia) or ruling part (hēgemonikon) (DL 7.159; Aëtius, SVF 2.836) – and are states with content. With the aid of their logical theory, the Stoics claim that such content will take the form of a ‘sayable’ (lekton). For example, the perceptual impression above says that << This is a black cat >> and the non-perceptual impression, that << There is no afterlife >>.¹⁵ Indeed, the Stoics assume that the sayable p is accessible to the subject only if the subject creates the impression that p. Stoic theory therefore does not countenance the possibility of entertaining a sayable ‘directly’, i.e. in the absence of forming an impression with that sayable as its content (SE M 7.385-6). So whenever the agent gives or withholds assent on the proposition that p, we can assume that p has been brought to mind through the impression that p.

Our examples so far involve impressions which merely describe the world without suggesting any practical response, but the Stoics also think there are impressions that are action-guiding and impulse-prompting (hormētikai). This label reflects the Stoic claim that by affirming these impressions as true and giving assent to their content, the agent creates an ‘impulse’ (hormē), and an impulse, on the Stoic view, is a “motion in the soul toward something”, the psychological change which in human agents constitutes action.¹⁶ This relationship is distilled in the Stoic thesis that “what stimulates an impulse is nothing other than an impulse-prompting impression of what is immediately fitting”.¹⁷ Presumably, the impulse-prompting impression is ‘of what is immediately fitting’ in that it represents a particular course of action as fitting – as the thing to do here and now.

As scholars have noticed, these commitments entail a strong version of motivational cognitivism in Stoic psychology.¹⁸ For human agents, no impulse arises without the subject entertaining the impression that this particular action is fitting to do in the circumstances. Indeed, our sources indicate that the Stoics offer a rough formula for the propositional content of such impressions:

Now acts of assent are to one thing, and impulses toward another. Acts of assent are to certain propositions, while impulses are toward predicates, the ones somehow included in the propositions assented to.¹⁹

(Arius, Ecl., 2.88.2-6, tr. Sauvé Meyer slightly modified)

In assenting to the propositional content of an impulse-prompting impression, one thereby forms an impulse. But the act of assent and the impulse have different objects. The object of the impulse is a predicate (katēgorēma), one that is ‘somehow included’ in the assented-to proposition.²⁰ One way to interpret this remark, now widely endorsed in the secondary literature, is to construe the propositional content as including a marker along the lines of << it is fitting for me …  >> (kathēkei emoi …).²¹ The hole in the marker << … >> is completed with a predicate (expressed with a verb in the infinitive) describing the action to be done by the agent. So, for instance, the impulse to drink wine arises from the agent’s assent to the impulse-prompting impression with the propositional content << it is fitting for me to drink this glass of wine >>, where << … to drink this glass of wine >> is the predicate included in the proposition and serving as the object of the impulse (Cicero, Tusc., 4.21; Arius, Ecl., 2.97.22-98.1).²²

Psychological disease and action-guiding impressions

With this background material in place, I will now argue for my amplification to the Assent Proposal, on which psychological disease controls which impulse-prompting impressions are formed by the vicious agent and so which prospective courses of action she entertains prior to giving assent. Using evidence from the Stoic account of expertise, I will show that the Stoics operate with a porous boundary between the larger cognitive condition of the agent and her occurrent impressions, such that her opinions and beliefs influence these impressions’ representational character. Expertise upgrades their precision and detail, whereas psychological disease mars them, leading the agent to entertain thoughts that falsely represent the value of certain objects and that misjudge the appropriate practical response. In possessing a psychological disease, then, one takes on a propensity for evaluative error in one’s action-guiding impressions.

To start, I will present the Stoic account of the cognitive benefits of expertise, and then compare it to the cognitive defects brought about by psychological disease.

The comparison with expertise

The Stoics define expertise as a system of grasps (katalēpseis).²³ Since every grasp is an assent to a ‘grasp-prompting’ (katalēptikē) impression, to possess expertise is to possess certain beliefs (hupolēpseis) that the non-expert lacks.²⁴ But the cognitive significance of expertise is not exhausted by belief. On the Stoic view, expertise also influences the character of the impressions the expert forms, whenever an object from the relevant domain is encountered. The degree to which an agent has refined her rationality is reflected in the detail and scope of the impressions she entertains prior to giving assent. This is because to form an impression is itself an activity of reason, and more specifically one which involves the activation of concepts.

Consider the following example, recorded in several texts (Cicero, Acad. 2.20; Philodemus, Mus. 4.34.1-8, see also 4.115.26-116.5). A musical expert and amateur hear the same piece, in the same auditory conditions. Here the Stoics claim that, in response to the same auditory input, the impressions generated by these two subjects will differ. The expert receives a ‘scientific perception’ (epistēmonikē aisthēsis) of a genuine property of the music, while the non-expert does not (Philodemus, Mus. 4.34.1-8). The expert’s achievements in the domain of music are thus registered in the content of her impression of the piece. So for instance the expert gets the impression that << This melody is in the Dorian mode >>, whereas the amateur hears only that << This song sounds sad >>.

To explain this difference in content, the Stoics appeal to the expert’s use of concepts such as melody and dorian mode, which she has acquired through musical training and practice. In general the Stoics hold that, as alterations of the mind, impressions are generated under the influence of the concepts the agent possesses. Indeed, Chrysippus characterizes reason as a collection of concepts (Galen, PHP 5.3.1), and identifies concepts as the parts of reason (Galen, PHP 5.3.2, 5.4.9). Given the holistic functioning of the mind and the Stoic claim that impressions are formed there, it is no surprise that reason’s parts (its concepts) are held to influence reason’s activities (its impressions). This much, at least, seems clear from the sources (see Galen, PHP 2.5.12-13) and is largely uncontroversial among interpreters of Stoic epistemology.²⁵ Therefore, since musical training results in the musical expert’s acquisition of specialized concepts, above and beyond those which every rational agent possesses (Aëtius, SVF 2.83; Epictetus, Diss. 2.11.2-3, 3.6.8), the informational background she uses to process auditory stimulus is richer and more comprehensive than the amateur’s. The result is a difference in the auditory impressions of the musical expert and amateur (DL 7.51).

Furthermore, it is clear that some expert impressions will be impulse-prompting and action-guiding. This is because the Stoics recognize forms of expertise such as horsemanship and hunting (Arius, Ecl. 2.67.5-7), and no doubt assume that these cognitive accomplishments allow their possessor to behave in distinctive ways (e.g. when holding the reins or pursuing a deer). But since one cannot have an impulse to φ without first generating the impulse-prompting impression that << It is appropriate for me to φ >>, the Stoics must think that, in the various circumstances where their expertise is called for, the horseman and hunter will entertain more determinate action-guiding thoughts than the neophyte. Likewise, we should expect that the musician, too, will receive expert, action-guiding impressions, when he (for example) undertakes to sing in harmony or pluck the strings of his lyre. In these settings, there is a range of action-types that – to him but not to the amateur – seems fitting to carry out, thanks to advanced training in his craft.

For now, the important upshot is that the condition of the agent’s mind affects the content of the impressions she forms. To illustrate this point, the early Stoics compare the relationship between reason and impression to the intermingled flavours and scents of an apple (Iamblichus, SVF 2.826): just as one cannot get a taste of the apple without also smelling its scent, so too one cannot form an impression without some contribution of reason and its parts (concepts). Given this framework for understanding the relationship between reason and impression, what would the Stoics say about a subject who has taken on false opinions and has to this extent corrupted their reason? Does this kind of cognitive deficiency also penetrate into and affect the content of their impressions, as an inverse effect of expertise?

As I read the source texts, the Stoics fully embrace this implication. Recall that psychological diseases are identified with hardened, false evaluative opinions. The Stoics hold that just as a system of grasps (expertise) expands the information revealed by one’s impressions, so the presence of false opinions (psychological diseases) degrades their representational accuracy: the defective reason of the vicious no less than the advanced reason of the expert influences the impressions they form. In one sense this is an unsurprising result, since the Stoics recognize a physiological analogue to psychological disease – lack of tension (atonia) in the corporeal breath (pneuma) constituting the agent’s mind (Galen, PHP 4.6.1-12) – and characterize the impression as an imprint or alteration in the mind (DL 7.45; SE M 7.232-3). Moreover, the Stoics define knowledge as a condition governing the creation of impressions (DL 7.47) and therefore will assume that the wise, in virtue of their knowledge, form different impressions than the vicious.²⁶ The point I mean to emphasize now, however, is that psychological diseases control which concepts are called upon in the generation of one’s impressions, namely, in the impulse-prompting impressions prefiguring an emotional reaction to the object of the disease (e.g. wine for the wine-lover). It is therefore through the mechanism of concept application that both expertise and psychological disease govern which action-guiding thoughts the agent entertains prior to giving assent.²⁷

Concept mis-application in action-guiding impressions

Let’s take a closer look at the Stoic account of concepts (ennoiai).²⁸ These cognitive states allow the rational agent to articulate the significance of the objects she encounters in her surroundings and thereby to depict them in an impression. On Stoic theory some concepts are acquired naturally, as a result of the process by which a child acquires reason, while others, such as those deployed by the musical expert above, arise “through instruction and attention” (Aëtius, SVF 2.83). Included among the naturally-acquired concepts – also called ‘preconceptions’ (prolēpseis) – are our colour concepts and, we can assume, such notions as plant or human being.²⁹

Cicero reports that in possessing human being one has access to the proposition << if it is human, then it is a mortal, rational animal >> (Acad. 2.21). Propositions of this form – also called ‘indefinite conditionals’ (Cicero, Fat. 15) or ‘universal’ (katholika) propositions (SE M 1.86; Epictetus, Diss. 2.20.2-3; Plutarch Com. Not. 1080c) – are central to Stoic logic, where they are introduced to render the content of universal utterances. For the Stoics, the utterance ‘every human is a mortal rational animal’ expresses the proposition Cicero alludes to above, << if it is human, then it is a mortal, rational animal >> (SE M 11.8-9).³⁰ In the context of Stoic psychology, then, Cicero’s example should be understood as formulating the features the agent associates with human beings and so the general information which underlies her cognition of her surroundings and the impressions she forms of it.³¹

Crucially for our purposes, also included among the preconceptions are certain evaluative concepts, most prominently good and bad.³² The Stoics think that every adult human has at least a basic handle on good and bad, insofar as it is part of being rational to see that if something is good, it is necessarily beneficial, and that if something is bad, it is necessarily harmful.³³ Therefore generating the impression that, e.g. << Money is good >> requires activating the preconception good, with the result that one takes on at least a provisional commitment to money being necessarily beneficial. Of course, to believe that money is good (and so necessarily beneficial) requires the further activity of giving assent to this impression. But merely entertaining the claim that << Money is good >> also entails some rational activity, on the Stoic view, insofar as it depends on a kind of implicit calculation to the effect that money instantiates goodness and is therefore necessarily beneficial.

Despite sharing the same preconceptions of good and bad, however, agents obviously differ from one another in the particular objects they take to be necessarily beneficial and harmful. Chrysippus identifies childhood experience as one factor influencing the evaluative views one holds as an adult (Cicero, Fat. 7-8; Gellius, 7.2.8) – presumably because each agent acquires the preconceptions good and bad partly on the basis of such experience. We are unfortunately not provisioned with a surfeit of Sages, and so it cannot be by directly encountering genuinely good objects that we acquire the notion of good (as we might acquire the concept white by directly encountering white objects).³⁴ Instead, we come to form this preconception by a kind of ‘rational inference’ (collatione rationis) from objects which accord with human nature and are thus fitting for us to pursue, but which are not themselves good and whose possession makes no difference to our happiness (Cicero, Fin. 3.33; see also DL 7.85-89).³⁵ Clearly, the Stoics recognize that different agents will encounter and pursue different fitting objects in the course of their maturation – different kinds of food and sources of social esteem depending on one’s cultural environment.³⁶ So since childhood experience varies from agent to agent, there will also be differences in the objects which spark the development of our evaluative preconceptions. These differences could show up in the patterns of concept application we perform later as adults, if one mistakes the target of the inference (goodness) with its source (the fitting objects encountered in childhood).³⁷

This picture sketched in our early Stoic sources comes into sharper focus in the Discourses of Epictetus, where mis-predicating the preconceptions good and bad is characterized as the fundamental expression of vice. According to Epictetus, differences among the non-wise can be analyzed in terms of the objects to which they wrongly predicate good and bad and of the impressions generated thereby. Consider this revealing passage:

For this is the cause of all bad things for human beings: not to be able to attach correctly their shared preconceptions to particulars. And some of us think one thing, others of us something else.³⁸

(Diss. 4.1.42-3, tr. Oldfather modified)

Epictetus assumes throughout the Discourses that to ‘attach’ (epharmozein) a preconception to a particular involves the creation of an impression, in which the feature for which one has a preconception is predicated of a particular object (e.g. Diss. 1.22.1-9; 1.28.11-13).³⁹ He also assumes that preconceptions can be attached correctly, as in the thoughts of the Sage, or incorrectly, by the vicious (Diss. 2.11.6-10). Epictetus’ model therefore harmonizes with the early Stoics’, since both he and the Stoic founders agree that forming an impression requires the predication of concepts, and that the concepts activated in response to a given object determine the propositional content of the resulting impression.⁴⁰ Furthermore, in Diss. 2.18.8-12, Epictetus evinces the early Stoic view that psychological disease and infirmity consist in ingrained, false evaluative opinions. In Epictetus’ mouth, then, the claim that ‘some of us think one thing, others of us something else’ (4.1.43) would seem to reflect the assumption that different vicious agents suffer from different diseases and infirmities of the soul. Indeed, Epictetus also refers to a ‘corresponding impression’ (katallēlos phantasia) for each psychological infirmity (2.18.9-10) – presumably an action-guiding impression corresponding to the relevant object (e.g. wine in the case of love of wine).⁴¹

Immediately following his remarks at Diss. 4.1.42-3, Epictetus turns to the psychological effects of such vicious conditions:

One person [thinks] that he is sick. Not at all. Rather, he does not correctly attach his preconceptions. Another thinks that he is a beggar, another that he has a difficult father or mother, and yet another that Caesar is not gracious to him. But this is one thing and one thing only: not knowing how to attach one’s preconceptions (to tas prolēpseis epharmozein mē eidenai). Indeed, who does not have the preconception of bad, that it is harmful, to be avoided, that it is something to be rejected in every way? One preconception does not conflict with another, but [conflict does arise] whenever one proceeds to attach them (prolēpsis prolēpsei ou machetai, all’ hotan elthēi epi to epharmozein). What, then, is the bad that is harmful and to be avoided? One says that it is not to be friends with Caesar. But he is off-track and has missed the proper application (apepesen tēs epharmogēs).

(Diss. 4.1.43-5, tr. Oldfather modified, reading Schenkl’s text)

Whereas everyone who possesses the preconception agrees that if something is bad, then it is “harmful, to be avoided” (4.1.44), only the politically ambitious attribute these negative features to isolation at court. Epictetus then suggests that it is this psychological disease – political ambition, perhaps a relative of love of fame – that is responsible for the agent ‘attaching’ bad to political isolation, and so for forming the impression that << It is bad not to be friends with Caesar >>. This is a mis-application of bad, on the Stoic reckoning, insofar as it is implicitly asserted of something that is not necessarily harmful.

When read in light of this final example, the initial examples from the passage take on a new cast. Epictetus likely has in mind not merely descriptive appearances such as << I am sick >> but rather << It is bad to be sick >>, not << I am a beggar >> but << It is bad to be a beggar >>. In these impressions, the preconception bad is predicated incorrectly of objects that are not necessarily harmful according to Stoic axiology, for bodily disease and poverty are merely dispreferred. But here too Epictetus emphasizes that it is the differences in what one thinks about goods and bads (4.1.43), i.e. in one’s profile of psychological disease and infirmity, that explain why the agent erroneously attaches her evaluative preconceptions to the objects that she does. The claim that emerges from the Epictetus passage, then, is that differences in beliefs about value account for the diverse range of objects to which the non-wise incorrectly attach good and bad.

Could it be that psychological disease also determines how concepts other than good and bad are attached? For instance, could political ambition affect how the concepts power and freedom are applied, as well as good and bad? This possibility is intriguing and certainly Stoic in spirit. After all, the school maintains that only the Sage is free and that every vicious agent is a slave, regardless of the offices they may hold in their city (DL 7.121). The Stoics surely think, then, that the non-wise misunderstand the nature of power and freedom. But in our text above, Epictetus explains only how the preconception bad is mis-applied, and my reconstruction tracks his point of emphasis. Epictetus’s discussion therefore suggests a partial disanalogy between psychological disease and forms of expertise such as the musician’s: as we saw earlier, musical expertise manifests in the agent applying specialized concepts from a particular domain (e.g. dorian mode), rather than the evaluative preconceptions shared among all rational agents (e.g. good and bad).⁴²

Now, even if we are convinced that the pattern in which one attaches good and bad can be traced back to one’s profile of psychological disease and infirmity, how does this result support my proposal that such conditions influence the content of one’s impulse-prompting impressions? After all, such impressions are distinguished by their content including a certain marker – e.g. << it is fitting for me … >> – but this is not the case for any of Epictetus’s examples above.

The worry dissolves, however, once we realize that predications of good and bad are indeed included in the content of the impulse-prompting impressions summoned by psychological disease and related conditions. These are the impressions prefiguring the emotions (pathē), whose content is given by a conjunctive proposition combining both (i) a predication of value and (ii) an assertion that it is fitting to φ in the circumstances.⁴³

Recall that the Stoics appeal to the diseases an agent has to explain why she suffers emotional appetites in response to certain objects but not others (Arius, Ecl. 2.93.1-3; Galen, PHP 4.5.30). Moreover, as we have just seen, for the Stoics, every emotion requires an assent to a proposition with two components: (i) a general value claim about the goodness or badness of a given object-type (usually false) and (ii) a more specific claim about which action is fitting in the circumstances.⁴⁴ (Often the action in (ii) is phenomenological, such as a sense of uplift with delight or a downcast feeling with grief.⁴⁵) Here the Stoics assume that the value claim (i) is tightly connected with the fittingness claim (ii), as two pieces of the same action-guiding thought. Indeed, recent commentators have gone so far as to propose that (i) serves as something like the perceived grounds for (ii). For instance, it is because wine is taken to be good, and so necessarily beneficial, that the wine-lover thinks it fitting to respond with elation whenever it is present.⁴⁶

Epictetus’s discussion above is entirely consonant with this account, and should be understood as identifying the origin of the value claim (i) featuring in the content of the diseased agent’s emotion-prompting impression. Psychological diseases and related conditions determine the more general evaluative errors which underwrite particular cases of emotional response, insofar as they are responsible for the pattern in which the agent mistakenly applies good and bad.

Consider the psychological effect of ‘hostility to guests’ (inhospitalitas, Cicero, Tusc. 4.27). A vicious agent with this aversion is committed to the general evaluative view that hosting guests at home is something bad. So whenever a guest arrives, bad is activated, resulting in an impression like << Having a guest in my home is bad, and it is fitting for me to feel anxious and to lie about a bedbug infestation >>. The thought that hosting guests is bad grounds the judgement that both a phenomenological feeling (anxious constricting) and some particular behaviour (telling a lie) is fitting in this scenario. By contrast, an agent who is not hostile to guests would form a different impression in response to the same stimulus, perhaps that << Having a guest in my home is pleasant, and it is fitting for me to prepare the spare bedroom >>.

Conclusion

What emerges from our close look at the ancient evidence is a well-developed Stoic account of the benefits and dangers of cognitive penetration, grounded in the school’s insistence that impression-formation is an activity of reason. Insofar as the wine-lover’s reason is corrupted in a different way from the money-lover’s – though both are equally miserable – the two vicious agents will generate different action-guiding impressions when faced with similar stimuli. Stoic psychology identifies concept application as the mechanism through which the beliefs one holds affect the content of one’s action-guiding impressions – marred in representational accuracy by the false evaluative opinions of the vicious, and enhanced by the grasps of the expert. I have shown how this reconstruction is compatible with the early Stoic theory of the emotions, as well as with the founders’ account of concept acquisition and expertise. The Assent Proposal ought to be augmented, then, with the claim that, for the Stoics, psychological disease is a vicious condition that takes effect prior to the act of assent, by controlling which objects and objectives are first brought to mind in an action-guiding impression.

Acknowledgements

For helpful discussion and constructive comments on earlier drafts, I am grateful to Elena Cagnoli Fiecconi, Christopher Gill, Ursula Coope, Karen Margrethe Nielsen, Jake Quilty-Dunn, Jeremy Reid, Natalie Hannan, Anna Schriefl, Gail Fine, Jacob Klein, Luca Castagnoli, Katja Vogt, and Whitney Schwab. I would also like to thank several anonymous referees for their sharp critical eyes and for many suggestions that improved the paper. Last but not least, I would like to recognize the editors of this journal for their efficient practical assistance.

Notes

1 All vicious agents are said to possess the disposition (diathesis) of vice but to differ in their conditions (hexeis) or traits (ēthē) of character. For the Stoic distinction between dispositions and conditions, see Simplicius, Cat. 8.237-8; Arius, Ecl. 2.70.21-71.14; and helpful discussion in Graver, Stoicism and Emotion, 133-5. A more detailed account of character traits (as part of the celebrated personae theory) is found in Cicero, Off. 1.107-121, but this discussion reflects later trends in the Stoa initiated by Panaetius and so I set it aside here. My focus will be the views of the Stoic founders, i.e. Zeno, Cleanthes, and Chrysippus – though I will argue below that Epictetus aligns with their position in crucial respects.

2 DL 7.115; Arius, Ecl. 2.93.1-13; Cicero, Tusc 4.23-33; Galen, PHP 4.5.21-3, 5.2.3; Plutarch, St. Rep. 1050d.

3 Arius, Ecl. 2.93.6-8: Νόσημα δ’ ϵἶναι δόξαν ἐπιθυμίας ἐῤῥυηκυῖαν ϵἰς ἕξιν καὶ ἐνϵσκιῤῥωμένην, καθ’ ἣν ὑπολαμβάνουσι τὰ μὴ αἱρϵτὰ σφόδρα αἱρϵτὰ ϵἶναι. See also DL 7.115. Of the various types of psychological ailments, we have the best evidence on psychological diseases (nosēmata), and these will be my focus here. Infirmities (arrōstēmata) seem to form a subset of diseases: thus Graver, Stoicism and Emotion, 139–40, and Brennan, “The Old Stoic Theory of the Emotions”, 40–1. And aversions (proskopai) (e.g. hatred of wine) are simply negative analogues to diseases proper (e.g. love of wine): hardened opinions that an object should be avoided rather than chosen. For discussion of the ‘proclivities’ (euemptōsiai) and their possible origins in Posidonius’ heterodox psychology, see Kidd, “Proneness to Disease”; Graver, Stoicism and Emotion, 142–5; and Ranocchia, “The Stoic Concept of Proneness to Emotion and Vice”.

4 Graver, Stoicism and Emotion, 139; Brennan “The Old Stoic Theory of the Emotions”, 41-3. For the claim that goods are chosen, whereas (preferred) indifferents are selected, see e.g. DL 7.99; Arius, Ecl. 2.82.20-83.1, 97.15-22; and discussion in Klein, “Making Sense of Stoic Indifferents”, 239.

5 Love of money: Arius, Ecl. 2.93.9; Cicero, Tusc. 4.25; Plutarch, St. Rep. 1050d. Women-madness: Arius, Ecl. 2.93.9; Cicero, Tusc. 4.25. Love of fame: DL 7.115; Plutarch, St. Rep. 1050d. Love of wine: Arius, Ecl. 2.93.9; Athenaeus, Deipn. 11.464d-e.

6 Thus Arius says that all moral errors are equal, insofar as they are all decisions (kriseis) stemming from the disposition of vice, but differ in their external causes, since there are differences in the intermediate things (i.e. indifferent objects) which trigger the decision (Ecl. 2.106.21-26).

7 For emotions (pathē) as excessive impulses, see e.g. Arius, Ecl. 2.88.6-10 and DL 7.110. Note that Arius defines psychological disease as “an opinion related to appetite” (doxan epithumias: Ecl. 2.93.6-7), and that appetite is one of the four highest genera of emotion. See also Galen, PHP 4.5.30. DL 7.111 conflates occurrent emotions with psychological diseases, but this oversimplifies the Stoic view: see Inwood, Ethics and Human Action, 127–8; Brennan “The Old Stoic Theory of the Emotions”, 50, and Graver, Stoicism and Emotion, 39.

8 Inwood, Ethics and Human Action, 162–5; Brennan, The Stoic Life, 105; Graver, Stoicism and Emotion, 63–5, 138–45, 166–7. For scepticism that there is any precise account of the functioning of psychological disease, see Gill, The Structured Self, 261–4. However, in common with the proponents of the Assent Proposal, I regard this assessment as unnecessarily pessimistic.

9 For insightful discussion of the relation between action and impulse, see Sauvé Meyer, “Passion, Impulse, and Action in Stoicism”, 112–24. I follow her in supposing that action is constituted, rather than caused, by impulse, but nothing hangs on this question here: my interpretation of psychological disease is compatible with impulse either causing or constituting action. See also Inwood, Ethics and Human Action, 48–53; Brennan, The Stoic Life, 86–93; and Vogt, Law, Reason, and the Cosmic City, 171-3. That assent is necessary and sufficient for impulse in human agents is non-controversial.

10 Cicero, Fat. 42; Arius, Ecl. 2.88.1-6; SE M 7.154, 370, 385. See Brennan, The Stoic Life, 54–7.

11 The activities of assent and impression are also distinguished in Zeno’s hand simile (Cicero, Acad. 2.145), and Iamblichus reports that the Stoics attribute them to different powers of the mind (SVF 2.826). As the quotations above make clear, the Assent Proposal is a claim about the activity of assent (under what circumstances acts of assent are given). It is thus silent on the connection between psychological disease and the activity of impression (when impressions are formed and why), which is my focus here.

12 I use small caps to indicate concepts.

13 In the original ‘Shooter Task’ study, participants play a “simplified videogame, which roughly simulates the situation of a police officer who is confronted with an ambiguous but potentially hostile target, and who must decide whether or not to shoot. In the game, images of people who are either armed or unarmed, and either African American or White, appear unexpectedly in a variety of contexts … this game requires participants to make a behavioral shoot / don’t shoot decision similar to that of a police officer” (Correll et al., “The Police Officer’s Dilemma”, 1315). One result is that subjects decide to shoot more frequently when presented with an unarmed African American male than with an unarmed white male. For the possibility that this result can be explained by cognitive penetration, see Siegel, “Bias and Perception”.

14 I do not consider here the impressions or impulses of non-rational animals, since they are irrelevant to my topic. Psychological disease is a condition that afflicts only rational creatures: Cicero, Tusc. 4.31.

15 I indicate sayables with double angled brackets. Paradigmatically, as in the two examples here, the sayable expressed by an impression will be a proposition (axiōma). Though this claim has recently been questioned (Bronowski, The Stoics on Lekta, 99–112), there is no doubt that Chrysippus understands rational impressions to express propositions: see his Logical Investigations col. IV 20–35 (ed. Marrone) and SE M 7.416-21.

16 Arius, Ecl. 2.86.19: τὴν δὲ ὁρμὴν ϵἶναι φορὰν ψυχῆς ἐπί τι. Rational creatures (i.e. mature adult humans) form rational impulses, which are defined as “a motion in the mind toward something in the domain of action” (φορὰν διανοίας ἐπί τι τῶν ἐν τῷ πράττϵιν, Arius, Ecl. 2.87.3-5). See also Seneca, Ep. 113.23 and Origen, Princ. 3.1.2. For the relation between impulse and action, see note above.

17 Arius, Ecl. 2.86.17-18: Τὸ δὲ κινοῦν τὴν ὁρμὴν οὐδὲν ἕτϵρον ϵἶναι λέγουσιν ἀλλὰ φαντασίαν ὁρμητικὴν τοῦ καθήκοντος αὐτόθϵν. For the translation of αὐτόθϵν, see Brennan, “Stoic Moral Psychology”, 266n20; Long and Sedley, The Hellenistic Philosophers, v.2, 318; and Inwood, Ethics and Human Action, 224.

18 Klein, “Making Sense of Stoic Indifferents”, 233; Graver, Stoicism and Emotion, 26–8; Brennan, The Stoic Life, 174–5.

19 Reading Wachsmuth’s text and accepting his emendations: Ἤδη δὲ <ἄλλῳ> μὲν ϵἶναι συγκαταθέσϵις, ἐπ’ <ἄλλο> δὲ ὁρμάς· καὶ συγκαταθέσϵις μὲν ἀξιώμασί τισιν, ὁρμὰς δὲ ἐπὶ κατηγορήματα, τὰ πϵριϵχόμϵνά πως ἐν τοῖς ἀξιώμασιν, <οἷς συγκατατίθϵσθαι>.

20 Sauvé Meyer, “Passion, Impulse, and Action in Stoicism”, 112–6, 123–4; Vogt, Law, Reason, and the Cosmic City, 168–73; Brennan, The Stoic Life, 86–8; Long and Sedley, The Hellenistic Philosophers, v.2, 200; Inwood, Ethics and Human Action, 55–6.

21 Or some other normative guise, e.g. <<it is advantageous for me…>>. See Epictetus Diss. 1.18.2, 2.26.2; Seneca, Ep. 113.18; and discussion in Sauvé Meyer, “Passion, Impulse, and Action in Stoicism”, 115–6, 130, and Brennan, “Stoic Moral Psychology”, 268.

22 See Klein, “Desire and Impulse in Epictetus and the Older Stoics”, 12–16; Graver, Stoicism and Emotion, 27–8; Brennan, The Stoic Life, 87–8; Long and Sedley, The Hellenistic Philosophers, v.1, 202; Inwood, Ethics and Human Action, 56–66. For an alternative view, see Blackson, “Impulsive Impressions”. For detailed discussion of Stoic predicates, see Bobzien and Shogry, “Stoic Logic and Multiple Generality”, 4–7.

23 Alternative translations for katalēpsis include ‘cognition’ and ‘apprehension’. See SE M 11.182: τέχνη γάρ ἐστι σύστημα ἐκ καταλήψϵων, καὶ κατάληψίς ἐστι καταληπτικῆς φαντασίας συγκατάθϵσις. Note that grasp (katalēpsis) is a factive mental state open to Sage and non-Sage alike, unlike knowledge (epistēmē, scientia), which is found only in the Sage (Cicero, Acad. 2.144-5 et al.).

24 The Stoics tend to use belief (hupolēpsis) to cover all the cognitive states produced by an act of assent: opinion (doxa), grasp (katalēpsis), and knowledge (epistēmē). On this point, see Moss and Schwab, “The Birth of Belief”, 24–5; Vogt, Belief and Truth, 165–6; and Brennan, The Stoic Life, 64–6. The Stoic definition of technē (see SE M 11.182 and note above) does not mention knowledge and therefore allows that one can possess technē (e.g. musical expertise) without being a Sage.

25 Thus de Harven, “Rational Impressions and the Stoic Philosophy of Mind”, 224–6; Coope, “Rational Assent and Self–Reversion”, 246n27; Brittain, “The Compulsions of Stoic Assent”, 334n8; and Long and Sedley, The Hellenistic Philosophers, v.1, 240. For a recent defence of this interpretation against the claim that concept activation occurs subsequently to the formation of an impression, see Shogry, “What Do Our Impressions Say?”, 42–5, and de Harven, “Rational Impressions and the Stoic Philosophy of Mind”, 225–8.

26 DL 7.47: “and knowledge itself, they say … is a condition related to the reception of impressions that is unchangeable by argument” (αὐτήν τϵ τὴν ἐπιστήμην φασὶν … ἕξιν ἐν φαντασιῶν προσδέξϵι ἀμϵτάπτωτον ὑπὸ λόγου). See Cooper, “The Emotional Life of the Wise”, 183-4. On my account, psychological disease is also a condition (hexis) related to the reception of impressions, albeit not one that is unchangeable by argument.

27 On my view, since one’s past assents inform which concepts one has (whether they are specialized or not), and since concepts are deployed in the formation of impressions, one’s past assents indirectly inform the character of one’s impressions. Crucially, however, this claim is neither part of, nor entailed by, the Assent Proposal, since the Assent Proposal is silent on the question of impression-formation and a fortiori on the relation between one’s impressions and one’s past assents. Rather, the Assent Proposal concerns how one’s past acts of assent inform one’s occurrent acts of assent. As mentioned in the Introduction and note 11, the Assent Proposal is not a wide-ranging thesis about the importance of assent in Stoic moral psychology, but rather a narrower claim about the causal origins of one’s occurrent acts of assent.

28 Throughout the paper, I use ‘concepts’ to render ennoiai, not ennoēmata.

29 The interpretation of the Stoic theory of preconceptions is controversial. However, we can safely assume that notions such as plant or human being number among the naturally-acquired preconceptions, since, according to Aëtius, preconceptions can be formed on the basis of repeated perceptual encounters, and we typically encounter plants and human beings in this way – just like we repeatedly encounter white things and thus acquire the preconception white in the course of our natural development (SVF 2.83). On this point and others noted below, I follow Brittain, “Common Sense”, 169-79, but my claim here is also accepted by Dominic Scott, who agrees that concepts acquired on the basis of repeated perceptual encounters are “formed naturally” (Recollection and Experience, 202).

30 See Bobzien and Shogry, “Stoic Logic and Multiple Generality”, 15–20; Bobzien, “Logic: The Stoics”, 112–4; and Crivelli, “Indefinite Propositions and Anaphora in Stoic Logic”, 198–201.

31 See Crivelli, “The Stoics on Definitions”, 378–9. Cf. DL 7.54: “And a preconception is a natural concept pertaining to universals” (ἔστι δ᾽ ἡ πρόληψις ἔννοια φυσικὴ τῶν καθόλου).

32 Vogt, “The Good is Benefit”, 11–14, argues persuasively that both early and later Stoics posit a preconception of good that is shared among all rational agents. This claim is also accepted by Brittain, “Common Sense”, 171; Inwood, “Getting to Goodness”, 284–5; Frede, “On the Stoic Conception of the Good”, 75, 78; and Scott, Recollection and Experience, 201–10. Jackson-McCabe, “The Stoic Theory of Implanted Preconceptions”, 334–41, affirms that good numbers among the shared preconceptions, and makes the helpful observation that good must arise in a different manner than preconceptions of objects directly experienced (e.g. white). The Epictetus passage cited below (Diss. 4.1.42-45) clearly assumes that there is a shared preconception of bad, while the preconception good is explicit at Diss. 1.22.1. Cf. Diss. 2.11.3. Throughout I am focusing on the preconceptions of good and bad, the basic notions common to all rational agents, and not on the specialized refinements of these concepts possessed by the Sage (and arguably the subject of Cicero, Fin. 3.20-22).

33 SE M 11.22; Epictetus Diss. 4.1.42-45 quoted below; cf. DL 7.53. Here I rely on the interpretation of SE M 11.22 given by Vogt, “The Good is Benefit”, 4–6. Brittain, “Common Sense”, 178–9, plausibly identifies the preconceptions shared among all rational agents as those which the Stoics describe as ‘shared concepts’ (koinai ennoiai).

34 DL 7.53-54; cf. Seneca, Ep. 120.4-5, 8-11. This is a main theme of Inwood’s interpretation of Seneca Ep. 120 (“Getting to Goodness”, 277–8, 284–6). For the distinction between preconceptions acquired through direct encounters with their objects and those acquired by analogy or inference, see Jackson-McCabe, “The Stoic Theory of Implanted Preconceptions”, 334–7.

35 This claim forms part of the Stoics’ larger theory of oikeiōsis, whose details need not detain us here. See helpful discussion in Klein, “The Stoic Argument from oikeiōsis”, 181-91.

36 Moreover, the early Stoics recognize hereditary influences on both the constitution of the body and psychological character: children are said to resemble parents in both respects (Cleanthes in SVF 1.518; Chrysippus in Plutarch, St. Rep. 1053c-d). Thus differences in hereditary factors could affect which fitting objects are pursued in childhood in preference to others, or the extent to which the same object is pursued in different agents. See Boys-Stones, “Physiognomy and Ancient Psychological Theory”, 80–3, and Bobzien, Determinism and Freedom in Stoic Philosophy, 291–4.

37 For an elaboration and defence of this proposal, see Lorenz, “Posidonius on the Nature and Treatment of the Emotions”, 198–9; Graver, Stoicism and Emotion, 157–63; and Bobzien, Determinism and Freedom in Stoic Philosophy, 291–4. Note that, for Chrysippus, one possible result of pursuing fitting objects in childhood is that “there arises a kind of natural opinion that everything sweet and pleasurable is good, and that what brings pain is bad and to be avoided” (opinio quaedam naturalis exoritur omne suaue ac delectabile bonum, contraque quod dolorem afferat malum esse atque uitandum, Calcidius, Tim. 165). In the background here is the Stoic claim that obtaining fitting objects usually brings a feeling of pleasure, and lacking them, a feeling of pain (DL 7.86; see also Seneca, Ep. 121.8). Boys-Stones defends a similar conclusion on the basis of the Stoic claim that our pursuit of fitting objects is mediated by ‘co-perception’ (sunaisthēsis) of the agent herself and the object, and so what vicious agents “in their pre-rational lives co-perceived (rightly) as harmful and beneficial … will become in their mistaken view of things ‘evil’ and ‘good’ respectively” (‘Physiognomy and Ancient Psychological Theory’, 87).

38 τοῦτο γάρ ἐστι τὸ αἴτιον τοῖς ἀνθρώποις πάντων τῶν κακῶν, τὸ τὰς προλήψϵις τὰς κοινὰς μὴ δύνασθαι ἐφαρμόζϵιν τοῖς ἐπὶ μέρους. ἡμϵῖς δ’ ἄλλοι ἄλλο οἰόμϵθα.

39 Admittedly the metaphor of ‘attaching’ is never fully spelled out: Epictetus does not specify, for instance, whether attaching a preconception consists merely in forming an impression, or in forming an impression and then giving assent to it. However, the texts leave no doubt that attaching a preconception will always minimally involve the creation of an impression. Attaching is never only a matter of assent, separated from the formation of an impression. For discussion of the origins of this Stoic metaphor in Plato’s Theaetetus, see Shogry, “The Stoic Appeal to Expertise”, 22–8.

40 See Crivelli, “The Stoics on Definitions”, 387–90, for discussion of successful and unsuccessful concept application in the context of the early Stoic theory of definition.

41 For helpful discussion of Diss. 2.18.8-12, see Graver, Stoicism and Emotion, 164–7. Graver’s reconstruction emphasizes Epictetus’ claim that repeated cases of appetite (and so repeated acts of assent) cause evaluative beliefs to become ingrained and thus turn into psychological infirmities. According to Graver, for an evaluative belief to become ingrained brings about “a change in the structure of the belief-set which ensures that that belief is consulted especially often in one’s practical reasoning” (166), so that one more easily assents and suffers emotions in the future. She thus interprets the passage in line with the Assent Proposal. However, Graver does not venture a suggestion about the ‘corresponding impression’ Epictetus associates with each psychological infirmity at 2.18.9-10 or its possible content and causal origins (see also Marcus Aurelius, 7.2). This remark could signal Epictetus’ acceptance of the claim that psychological infirmity controls which impressions are created prior to giving assent, as well as the acts of assent themselves. But admittedly the text is indeterminate. In any event, I will show that in other passages Epictetus supports my proposed amplification to the Assent Proposal.

42 Though expertise may control how other preconceptions are applied: for instance, the expert horseman could correctly apply strong or fast to particular horses, in a way that the amateur could not. A second disanalogy between the effect of psychological disease and expertise on our impressions is that non-expert impressions (e.g. <<This song sounds sad>>) are typically true, although imprecise, whereas the impulse-prompting impressions summoned by psychological disease are always false. This difference, and the one noted in the main text above, confirm that the comparison between expertise and psychological disease is imperfect, although both are conditions of the agent’s reason which influence the character of the action-guiding impressions she forms.

43 This is certainly the case for the secondary genera of the emotions, i.e. pleasure and pain (see Arius, Ecl. 2.90; Cicero, Tusc. 4.14-15; Galen, PHP 4.2.4-6; Ps-Andronicus, SVF 3.391). But it is likely that the primary genera of emotions (i.e. fear and appetite) also require assent to an impulse-prompting impression whose content combines both a value predication and fittingness assertion. On this point I follow Inwood, “Stoic Ethics”, 700–1; Brennan, The Stoic Life, 93–6; Graver, Stoicism and Emotion, 42; and Klein, “Desire and Impulse in Epictetus and the Older Stoics”, 18–21.

44 I say (i) is usually false because the Stoics seem to have allowed that Alcibiades suffers grief at his present vice, i.e. assents to the proposition that his present vice is something bad, which is true on Stoic axiology (Cicero, Tusc. 3.76-77). Even so, Alcibiades is taken to respond to this genuine evil in the wrong manner. See Graver, Stoicism and Emotion, 44–46, 196–206, and Brennan, “Stoic Moral Psychology”, 290.

45 Thus the Stoics speak of ‘swelling’ of the soul associated with delight and of the ‘contraction’ of the soul associated with grief (Arius, Ecl. 2.90.14-18; see also Cicero Tusc. 3.26, 83; 4.14, 63).

46 The extant Stoic sources are not explicit about how the emotional agent sees (i) as related to (ii). Graver, Stoicism and Emotion, 44–5 compares (i) to the major and (ii) to the minor premise in a practical syllogism. She is followed on this point by Klein, “Desire and Impulse in Epictetus and the Older Stoics”, 16–21. See also Cooper, “Posidonius on the Emotions”, 463–4. The important point for present purposes is that, according to the consensus view I accept here, both (i) and (ii) are part of the content of the impression prefiguring an emotional response.