2

The Stoics on the Voluntariness of the Passions

Steven K. Strange

One of the most characteristic and, to the contemporary mind at least,

bizarre theses of orthodox Chrysippean Stoicism is its insistence on the
wholly rational character of the human mind or soul. The whole of the
soul, or at least the whole of the h¯e gemonikon or “leading part” of the soul,
is said to be reason (logos): there are no irrational parts or powers within
the soul, no independent sources or faculties of emotion or nonrational
desire, and all cases of apparent irrationality are supposed to be referred
to the operations of thought itself (dianoia, another name for the mind).
This seems to fly in the face of much of what we assume or feel sure
we know about human psychology: not only do modern psychological
theories tend to stress the role of irrational or subrational factors in hu-
man motivation, but we seem to be clearly aware in ourselves of major
nonrational or irrational aspects of our own behavior and thought pro-
cesses. Thus when confronted with Stoic claims such as that all emo-
tions and desires are really just judgments or beliefs, or that passions
or emotions are unnatural conditions that ought to be extirpated, we
find these not only hard to swallow but difficult and nearly impossible
even to grasp. However, there is nothing about this sort of difficulty with
Stoicism that is specifically modern. Every major theory of moral psy-
chology since the Hellenistic period has taken for granted the originally
Platonic-Aristotelian separation of reason from the passions against which
the strict rationalism or intellectualism of the Stoics was a reaction: I need
only mention such examples as Plotinus, Augustine, Aquinas, Hume, and
Kant. Indeed, this held true even during the brief but important revival
in popularity of Stoic moral thinking in the seventeenth century: both
Descartes and Spinoza, each in his own way quite sympathetic to Stoicism
32
 On the Voluntariness of the Passions 33

in ethics, accept in some form the standard division of passive affections

of the soul from active judgments. No doubt the dominance of this dis-
tinction between reason and the passions in the history of philosophy
contributes to our feeling that it is just philosophical common sense, as
well as to the perceived uncongeniality of the Stoic view and our failure
to comprehend it.
But even though later philosophy, as well as historical common sense,
seems to have wholly rejected this part of the Stoic view, at least one aspect
of it has continued to be deeply influential, the part of it that concerns
responsibility and the human capacity for moral choice. We have a con-
cept of the will as a faculty of choice or of deciding to perform or not to
perform certain actions. Such a concept in fact represents a rather late de-
velopment in the history of philosophy, and its origins remain obscure.1
It is well known that the standard Latin term for the will, voluntas, was
originally just the Latin translation of the Greek term boul¯e sis or (as I shall
be rendering it) “wish,” translated this way by Cicero in his discussion of
the Stoic theory of the emotions in the Tusculan Disputations (4.12), and
taken over from Cicero by Seneca and later writers, such as Lactantius.
The Stoics, of course, did not invent the term boul¯e sis: it is used by Plato,
among other places, in the fourth book of the Republic to designate the
kind of desire proper to the rational part of the soul, that is, desire for
the good, as contrasted to thumos or spirit, desire for respect, fame, and
more generally one’s personal interests, and epithumia or bodily appetites,
paradigmatically the desires for food, drink, and sex and their associated
pleasures. Aristotle uses the term boul¯e sis in the same way as Plato, as the
desire (orexis) for the good, in his Ethics and in the De Anima. In both
Plato and Aristotle, wish is a function of the rational part of the soul,
not of the nonrational or “lower” parts – which by the way indicates that
Platonic-Aristotelian moral psychology is itself perhaps not as familiar
or commonsensical (by our lights) as we might at first have supposed:
the separation between the rational and nonrational soul in Plato and
Aristotle is not the, for us, expected distinction between pure thought or
belief on one side and feelings and desires on the other, because wish,
which belongs essentially to the rational part of the soul, counts as a type
of desire.
The concept of wish or boul¯e sis in Stoic moral psychology has affinities
with that in Plato and Aristotle, and indeed they probably took over the
use of the term from the Old Academy, while adapting it for their own use.
Boul¯e sis for the Stoics is defined as reasonable desire or pursuit (eulogos
orexis),2 reasonable because it is desire that aims at what is in fact good
 34 Steven K. Strange

orexis eparsis ekklisis sustole

desire/pursuit elation (upon avoidance contraction
(toward future present apparent (away from (upon present
apparent good) good) future apparent evil) apparent evil)

1. epithumia A. boulesis 2. hedone B. khara 3. phobos C. eulabeia 4. lupe [No eupatheia

lust (for future wish (for delight (upon joy (upon fear (against caution distress (upon corresponding
merely apparent future real present merely present real future merely (against present to present real
good) good) apparent good) good) apparent evil) future real merely evil]
evil) apparent evil)

(1)–(4) are passions (pathe)

(A)–(C) are rational affections (eupatheiai). No eupatheia corresponds to the passion of distress or lupe.

figure 1. Psychological responses to perceived good or evil

and is not merely apparently so. As such, wish belongs to the class of
the eupatheiai or “rational affections” in Stoic theory – that is, it is one of
the emotions or analogues of emotion that are possessed solely by the
sage or wise person, which in his case have come to replace the normal
passions or path¯e , which according to the Stoics are diseased conditions
of the vicious soul, and which they insist on calling by separate names to
bring out the crucial difference between the two kinds (see Figure 1).
It is nevertheless a species of the same genus, pursuit or desire (orexis),
as epithumia or lust, bad or irrational desire (which aim at things that
seem to be good but are not). Wish, as opposed to will, is thus a kind of
desire, not a faculty, and has no obvious connection with the notion of
choice; moreover, it is by definition aimed at the genuine good, so that
the idea of a bad wish or will makes no sense. Indeed, it is only possessed
by the fully morally good person: a bad person cannot be said to have
boul¯e sis.3 As far as I am aware, it seems to be Saint Augustine, in his De
Libero Arbitrio, who first clearly introduces the notion of a bad voluntas –
but even Augustine’s bad voluntas is aimed at things, namely earthly or
corporeal goods, that are good considered in themselves, just not good
for the spiritual beings that he considers us to be. Augustine, even if he
does not – as some might want to suggest – create the modern notion
of the will, is a much less remote ancestor of it than are the Stoics. (Yet
his notion of voluntas seems to owe something to the Stoic conception of
boul¯e sis, as well as to Plato’s and Aristotle’s.)
The Stoic notion of boul¯e sis thus lies fairly far from its descendant,
the later and familiar notion of the will. The specific contribution of
Stoic moral psychology to the history of the notion of moral choice does
not seem, therefore, to lie in the term used for its faculty. Of greater
importance here is the Stoic conception of what makes someone’s action,
 On the Voluntariness of the Passions 35

or someone’s emotion or desire, voluntary or a locus of responsibility, so

that it may be evaluated in terms of praise or blame. The key notion here
is that of assent (sunkatathesis), which Cicero, usually eager to play down
Stoic originality and to portray the Stoics as mere deviationists from the
Old Academy, identified as one of Zeno’s most important innovations in
philosophy.4 Assent is the acceptance of a proposition (axiōma) presented
to us in experience or thought as the content of a phantasia or impression
and as being true: assenting to such a proposition is what we would call an
act of judgment.5 According to the Stoic view, it is primarily assent that is
“in our power” or “up to us” (eph’ h¯e min), and it is only because this lies in
our control that anything at all does. So if we are responsible for any of our
actions, for any of our desires, or for any of our passions or emotions, it is
because these all depend upon our capacity to give or withhold our assent
in particular cases. Similarly, if we are to bear any responsibility for our
own moral character, this too will be due to our capacity for assent. But
responsibility for desires and emotions is primary here, for our actions
and our moral character are functions of them, of what we want and
how we respond to the world. And our desires, in a fairly clear sense,
just are emotions, for the Stoics, for they are the impulses to pursue or
avoid, to accept or reject, certain objects (technically, certain prospective
or actual properties applying to ourselves) as being respectively good
or bad. Therefore, according to the Stoics, they are also immediately
dependent on our value judgments and views about the good. So, in the
end, everything will depend on what we take to be good and evil.
Of course, what we should take to be good or evil according to Stoicism –
as is well known to anyone even minimally acquainted with the tenets
of Stoic ethics – is only our own moral condition, our own virtue and
vice, respectively. This is the attitude to the world adopted by the Stoic
sage or sophos, which she has attained through the excellence of her
understanding of the cosmos and her place in it, due to the perfection of
her reason and proper oikeiōsis or accommodation to the world. Ordinary
people such as we are, however, as the Stoics were aware, do not look upon
the world and the good and evil in it in this way, and seemingly cannot
help taking things such as pleasure and pain, health and sickness, wealth
and poverty, the life and death of themselves and loved ones as good and
evil, respectively. They cannot help, it seems, being grieved at the death
of a friend, or elated at the success of whatever enterprise they may have
been engaged in. It seems to us not only natural but inevitable that human
beings should have such feelings. Yet the Stoics want to insist that we can,
in principle, avoid experiencing such feelings, and that because they are
36 Steven K. Strange

bad and harmful to ourselves (as well as to others), and are impediments
to our happiness and freedom, that we should and must strive to and
learn to avoid feeling them.
Now, even if we might be inclined to concede the first of these claims,
the great majority of us would surely, at least initially, strongly resist the
second, namely that we would be much better off with all of our passions
extirpated. I do not propose to defend this thesis here, although it is
well worth reflecting upon in the course of a close and careful study of
Epictetus’ Discourses or Seneca’s Letters. My concern is rather with the first
claim, that passions are voluntary and that we therefore can in principle
learn to avoid having them. This seems generally taken to be a rather
straightforward if paradoxical claim on the part of the Stoics. However,
it seems to me to be very closely bound up, in interesting ways, with the
rationalism and monism of their moral psychology and to merit closer
discussion for this reason. Moreover, I think consideration of their views
about this can throw some interesting light on the general Stoic concep-
tion of responsibility and the nature of their compatibilism (by which I
mean the compatibilism of Chrysippus). Perhaps the discussion might
even throw some light on the antecedents of the original notion of the
will.

the passions as hormai pleonazousai and stoic

intellectualism
We first need to go over some ground that has already been well covered
in recent studies, namely the basic outline of the Stoics’ conception of
the emotions within the general context of their philosophy of action.6
It is important to remind ourselves of the details in order not to be mis-
led by the oddity of Stoic moral psychology into drawing false analogies
with more familiar theories. I cannot deal in detail with all controversial
points, so I restrict myself to a few crucial observations. My main focus
is the passions or path¯e that are experienced by ordinary people, that is,
“fools” or nonsages; but we need to compare these with the eupatheiai
or rational affections of the good person or sage. In genus, both pas-
sions and rational affections in human beings (rational animals) are the
proper motions (kin¯e seis) of the mind or h¯e gemonikon, that is, impulses
(hormai),7 intentional motions of the soul, that is, of thought (dianoia),
toward or away from some object (again, technically, toward or away from
some predicate or kat¯e gor¯e ma taken as worthy of being sought or avoided –
crudely, practical evaluative predicates that may (right away) or do (now)
 On the Voluntariness of the Passions 37

affect the evaluating subject. Hormai, the intentional motions of the soul,
are not dispositional states of desire or aversion, but actual occurrent
motions of the soul toward or away from prospective or actual valued or
disvalued objects. The theory can countenance dispositional desiderative
states as tendencies to pursue or avoid certain kinds of objects,8 but hormai
themselves are always actual movements occasioned by the (also actual)
positive or negative evaluation of some practical property.
The summa genera of passions are four (lust or appetite, fear, delight,
and distress), depending on whether the associated object is present or
prospective, and conceived as good or bad, with three summa genera of
rational affections (wish, caution, and joy – there is no rational affection
having as its object a present evil, briefly because the only real evil is vice,
which can never be present for a sage) (see Figure 1). Passions differ from
these latter, fully rational hormai in being disturbances of the soul: they are
referred to by Stobaeus’ source (SVF 3:378) as “flutterings” (ptoiai), and a
similar thought may lie behind Cicero’s somewhat misleading translation
in the Tusculan Disputations of pathos by perturbatio and commotio.9 We may
thus think of them as analogous to the disorderly and disturbing motions
of soul that have been eliminated in the competing Epicurean ideal of
ataraxia. The usual Stoic term for the analogous state, attained by their
sage, is apatheia or absence of passion. (“Dispassionateness” is perhaps
the best available translation.) The sage experiences instead of passions
the calm or orderly motions of the soul that are the eupatheiai (which are
still a type of hormai).
Connected with the notion of passions as disturbances within the soul
is Zeno’s description of them as irrational and excessive impulses. Accord-
ing to Diogenes Laertius (7.110; cf. Cicero, Tusculan Disputations 4.11),
Zeno defined passion as an irrational or unnatural motion of the soul,
and as a horm¯e pleonazousa, an “excessive impulse,” one that goes beyond
the bounds or reason.10 It appears from the parallel with the Cicero
passage that “going beyond the bounds” here means turning away from
(or, elsewhere, disobeying) “right” or correct reason.11 This is captured
in Chrysippus’ famous comparison of a walking person, who is able to
stop moving immediately any time he wants, with a runner, who is unable
to stop himself right away. This disobedient violation of the appropriate
bounds set by reason for the relevant motion of the soul is presumably
the same as the “certain voluntary intemperance” that Cicero says Zeno
declared to be the mother of all emotions (i.e., passions, Academica 1.39).
Passions represent disturbances in the soul – more precisely, in the soul’s
h¯e gemonikon or controlling part – precisely because they exceed rational
38 Steven K. Strange

bounds. For the soul’s h¯e gemonikon is the mind or thought (dianoia), and
a disturbance in it is a disturbance of its rationality. We can come to see
this by reflecting on what is meant by the Stoic talk of a “motion of the
soul.” The Stoics, of course, do not conceive of the soul as an immaterial
entity: the soul, and in particular the h¯e gemonikon or controlling portion
of the soul, which is also the mind (dianoia), is a bit of pneuma, a mixture
of corporeal air and fire that is concentrated about the main organ of
the body, which for the Stoics is the heart.
So the motion of the soul, under one aspect, is a literal movement
within this pneuma, which is itself a part of the active element in the uni-
verse, which the Stoics identify with Zeus or God. Under another aspect,
the motion of the soul is intentional: in the case of appetite and fear,
it is a movement of thought toward what is judged to be an apparently
good object, or away from an apparently bad one. These movements are
also called desire or pursuit (orexis) and aversion (ekklisis). In the case
of delight or pleasure and distress, where the material movement corre-
sponding to the passion is an actual physical swelling (elation, eparsis)
or contraction (sustol¯e ) of the soul, presumably again to be considered
as the region of pneuma around the heart, the intentional aspect is an
opinion (doxa) about the presence of some good or evil object. So while
horm¯e quite literally is a motion in the pneuma constituting the h¯e gemonikon
either outward, toward something that is desired or welcomed, or away
from something unwanted or rejected, it is also at the same time a move-
ment of thought, as I said earlier. These two aspects of the movement
which is the horm¯e , the literal, spatial movement toward the object, and
the intentional movement of thought, are not in any way separate: they
are quite literally two ways of looking at the very same phenomenon –
from the outside and from the inside, as it were, from the third-person
and from the first-person perspective, respectively.
Note that passions have been defined not only as pursuits and avoid-
ances but as out-of-bounds impulses (hormai).12 This applies in the first
instance to lust and fear (corresponding to the eupatheiai wish and
caution), which are the primary forms of passion: delight and distress
(and joy) are said to supervene (epigignesthai) on these,13 perhaps as
the results of satisfaction of lust (or, respectively, wish) or its opposite.
Officially, orexeis and ekkliseis are species of hormai (or, respectively, aphor-
mai), impulses: strictly, they are impulses toward actual, occurrent pur-
suits and avoidances of immediately available objects, as opposed to ones
expected to be pursued or avoided in the future.14 Most discussions
naturally focus on the case of orexis or desire (or, sometimes, ekklisis or
 On the Voluntariness of the Passions 39

avoidance), though presumably one could also have an excessive (or, re-
spectively, nonexcessive and fully rational) occurrent impulse toward the
object of some future plan – for example, that one will be prepared to do
whatever is necessary when one reaches the age of thirty-five to become
president of the United States. Some notion of moral evaluation of fu-
ture plan or intention, if that is what orousis is, would seem to be required,
as Inwood notes, and orousis will clearly also be “up to us.” Nonrational
animals also have impulses, but in this discussion we are only concerned
with rational impulses, that is, the impulses of rational animals, all of
which are either actual or future pursuits (or avoidances) on the basis of
judgments.
I have focused so far on the standard Old Stoic definitions given in our
texts of passions as motions, that is, impulses of soul, and have brought
in the intentional element, and the element of judgment, as consequent
upon this – though, as I have stressed, these should not be seen as two
separable aspects of the phenomenon but two ways of looking at the same
thing. Chrysippus, however, notoriously seems to have given definitions
of the passions merely in terms of judgment (krisis) and opinion (doxa) –
as the judgment, which is also an opinion, that there is a good or evil in
the offing, or now present, which good or evil is of a certain particular
sort – without making a reference to motion or impulse a part of this
definition. For this he was notoriously criticized by Galen, who accused
him of contradicting the Zenonian definitions of passions in terms of the
soul’s motion or horm¯e . Galen probably is following Posidonius on this
point, because it appears from some of the evidence in Galen’s Doctrines
of Hippocrates and Plato that Posidonius was concerned not only to refute
what he saw as Chrysippus’ extreme monistic rationalism in psychology
but to claim Stoic orthodoxy for his new, more moderate (and Platonically
influenced) moral psychology by arguing that it was more in line with the
moral psychology of Zeno and Cleanthes,15 the canonical founders of
Stoicism, than was Chrysippus’ own. (To what extent Posidonius’ alterna-
tive, less rationalistic moral psychology might have also involved a rejec-
tion of psychological monism is a question I wish to leave to one side.)16
I agree with the those scholars who argue that this criticism of
Chrysippus’ definitions is not well grounded. For one thing, it is clear
from Galen’s evidence that Chrysippus attempted to defend Zeno’s def-
initions in terms of his own. This sort of phenomenon is not unknown
elsewhere in Stoicism. Every important Stoic tried to give his own indi-
vidual definition of the telos or happiness, for instance, but we should not
think that in so doing that they meant to imply that Zeno’s definitions
40 Steven K. Strange

in terms of consistency (homologia) or the well-flowingness of life were to

be rejected: rather, they should be seen as expanding and commenting
upon Zeno’s definitions, the acceptance of which was probably seen as
defining orthodoxy, or at least one’s adherence to the Stoic school.
It will be clear, I hope, why I do not think that Chrysippus’ definitions
of the passions in terms of evaluative judgments need conflict with the
standard definitions (which he apparently also defended) of passions as
excessive impulse and as irrational movements of the soul caused by such
judgments. For one thing, the two sorts of definitions are often found
together (e.g., at Pseudo-Andronicus, On Passions, init.), which is what
we should expect, if indeed they are supposed to express the same basic
Stoic viewpoint. The judgment, given the Stoics’ dual-aspect metaphysics,
should also have a physical correlate that is intentionally directed in the
same way that it is (one is reminded here of the direct connection made
by Aristotle between the intentional experience of fears, confidences,
and sexual stimulations and heatings and chillings of the pneuma in the
body in De Motu Animalium 8).17 The judgment and the motion are con-
ceptually inseparable; besides, if the h¯e gemonikon or controlling part of
the soul is just supposed to be thought (dianoia), then any motion of it
will presumably be a thought. Chrysippus fixes on the judgment as the
key element in the phenomenon of emotion because it is the cause of
the motion or impulse as effect – necessarily so, because the impulse in
the soul is continuous with the visible bodily action (actually reaching
for the food, say), and thus is the link that explains how the value judg-
ment manages to cause the action that it is held to motivate (again, we
are reminded of the chain of psychic and bodily events that result in the
action to be explained in Aristotle’s De Motu Animalium). And because
the judgment is the locus of responsibility, what is “up to us,” it is what
accounts for why the action itself is voluntary and responsible.
What of the nature of this judgment, which is supposed to constitute
the passion? The sources for Chrysippus’ moral psychology present a dis-
tinction between two moments in the judgment that is cause of a passion,
as I have already noted. The subject judges (assents to the impression
that) a certain object is good or evil (and that it is either present or in the
offing), but also that it is appropriate for one to react to this realization in
the way specified, by pursuing or avoiding the prospective object, or con-
tracting or expanding (i.e., feeling elated or depressed) at the thought
that one has or has not got it. The hormetic reaction of the passion then
follows automatically, once these requisite judgments have been made –
the motions are not under independent conscious control.
 On the Voluntariness of the Passions 41

The dual nature of the judgment is particularly clear in the case of

delight and distress, where the material movement corresponding to the
passion is an actual physical elation or contraction of the soul, which
one judges to be appropriate in the circumstances. The link between
the two opinions is sometimes expressed merely by specifying that the
original evaluative judgment is “recent” or “fresh” (prosphatos): this term
is apparently intended to account for the transitory and unstable nature
of passions, or at least for the fact that they tend to diminish in intensity
monotonically over time, even though the person in question has not
ceased to believe in the goodness or badness of the object.18 It seems
clear that there was a corresponding secondary moment of judgment in
the case of lust and fear as well. This is sometimes expressed by talking of
the prospective object being taken to be a very great good or evil, even
“the greatest good,” so great that one would be justified in taking it to
be one’s primary or only object of concern (more so than one’s virtue
or the overall good of the universe, for instance, which are really and
in truth what is good). It also appears that there was a similar duality in
the judgment that was held to be the cause of desire or orexis (i.e., wish
as well as lust or appetite), since Galen in On the Doctrines of Hippocrates
and Plato (4.2.3–4) reports that Chrysippus defined orexis as “rational
horm¯e [i.e., the horm¯e of a rational animal] to something appropriately
pleasant [hoson khr¯e h¯e don],” which must mean something judged to be
appropriately pleasant to pursue.19 Despite the reference to the pleasant
here, the context in Galen seems to make clear that this was Chrysippus’
definition of orexis as a genus.
The two opinions or judgments are, however, clearly distinguished in
direct quotations from Chrysippus, and it seems clear that he stressed
that both were in principle under the voluntary control of the agent. For
he advocated trying to influence the secondary judgment rather than
the primary one sometimes in therapy of passions. The two judgments
are logically related, however, in that the second constitutes a further
observation on the sort of good or evil that the object of the primary
judgment is taken to be.
Once we notice that the rational affections differ from the passions in
that their objects actually are good, rather than just seeming so, whereas
in the case of passions the objects actually are not good, although they may
seem so, we are entitled to conclude that the upshot of this complicated
classification (which is actually a good deal more complex than I have
made out) is that the class of Stoic emotions actually coincides with the class
of what we would call human desires and their respective satisfactions,
42 Steven K. Strange

and their opposites – that is, with every valued and disvalued experience.
That emotions are at base desires is not an idea that is new with the
Stoics: Aristotle defines emotions in terms of desires, for instance in his
well-known definitions of anger in the De Anima and the Rhetoric. What
is significant about the Stoic theory is that it allows us to see how all
desires – or at least all practical desires, ones capable of being put into
action (for I could desire that some purely theoretical claim be true,
for instance that there be life on Mars)20 – can be conceived along the
lines of emotions. For all such desires will be either rational or irrational
pursuits or avoidances. Moreover, all actions have such desires as their
motivations. So the scope and role of the emotions and passions in Stoic
theory turn out to be very broad.

the passions not opposed to reason

Now we come to the crucial question of how the Stoics suppose that all
emotions and desires (and thus actions) can be fully under the control
of reason (and that they are in fact functions of reason). This, I take
it, is because they are all the direct results of evaluative judgments, and
judgments for the Stoics are acts of assent, which are supposed to be
wholly voluntary, and in fact the primary case of what counts as voluntary.
It is up to me what appearances or impressions (phantasiai) I should
accept or reject, at least in doubtful cases. This is all that is up to me, but
it is enough.21
The point here is not merely that the emotions are “cognitive” – that
is, that they essentially involve judgments. This is true of Aristotle’s con-
ception of the emotions as well, as can be seen, for example, from the
definition of anger that he gives in Rhetoric 2.2.22 But one would not wish
to claim that for Aristotle that the emotions in any way are the judgments,
as the Stoics claim – despite the fact that for Aristotle, as well as for the
Stoics, they are all by definition types of desire or orexis. In Aristotle,
following Plato, emotions and judgments reside in distinct and separate
parts or faculties of the soul, even if emotions are caused or occasioned by
certain judgments – for example, that one has been unjustly or inappro-
priately harmed. On the Stoic view, however, the emotion is a movement
of the soul necessarily consequent upon an evaluative judgment and oc-
curs in one and the same h¯e gemonikon or controlling part of the soul as
the judgment, and there is full continuity between a particular occur-
rence of a type of evaluative judgment (that a certain future available
object is good or appropriate, say) and the particular resulting motion
 On the Voluntariness of the Passions 43

(that I pursue it on that occasion: see Cicero’s careful explanation of

the Zenonian position in Academica 1.39). The cause (the particular act
of judgment) and effect (the particular pursuit of the object) are always
present together, and there are no grounds for separating them.
Plato and Aristotle, on the other hand, thought that emotions (with
the exception, at least in Plato, of wish) and reasoning must be separated
in the soul. They held this on the basis of the argument that Plato gives in
Republic 4 (436–39), which is endorsed by Aristotle in Nicomachean Ethics
1.13. According to this argument, reason and appetite (alternatively, for
Plato at least, spirit and appetite) can oppose one another as faculties,
in the sense that one can spur us to pursuit and the other to avoidance
of the very same object at the very same time: for example, a particular
opportunity to drink that presents itself – appetite, in the form of thirst,
spurring us to drink, while reason impels us not to do so. (The pursuit
and avoidance in question seem to be conceived as actual movements in
the soul, as is suggested by Plato’s physical examples of the application of
his principle concerning opposites, for example, the case of the archer. In
fact, the argument seems to depend on this.) The argument is intended
to show that this is only possible if this particular judgment of reason and
this particular thirst occur in distinct parts or aspects of the soul.23 As
Plato sees and notes in the passage, the argument can only get off the
ground if we assume that thirst is for drink unqualified, not for good drink
(438a), but as we have already seen, the Stoics (presumably following
Socrates) reject such an assumption: lust for them is precisely for what
seems good, and not as in the Republic for unqualified sorts of objects,
or objects qua pleasant, even if the reason that a particular object seems
good to me on some occasion is the obviously inadequate one that I
believe that it would serve to quench my raging thirst, that is, merely
that it is a potential drink, irrespective of whether it may be poisoned
or otherwise unhealthy. The Stoics have other reasons for rejecting the
argument as well, besides this one that is invited by Plato’s text itself: the
oscillation model, which involves the denial that the opposite movements
in the soul appealed to in Plato’s argument are really simultaneous,24
and the insistence that the mental acts labeled by Plato as “pursuit” and
“avoidance” may not really be contrary to one another, as assumed by the
argument.
It appears to be this same form of argument that Aristotle alludes to
in Nicomachean Ethics 1.13, when he cites his reason for positing a non-
rational element in human moral psychology (1102b16–21). Aristotle
makes explicit what is implicit in Plato’s presentation of the argument,
44 Steven K. Strange

that the phenomenon appealed to is that of akrasia or weakness of will,

where notoriously appetite or emotion does appear to go against reason.
(Note Aristotle’s reference to the opposite hormai of the akratic and the
paralytic: here, as in Stoic terminology, hormai would seem to denote, if
not actual opposed motions, at least opposed forces operating within the
soul. As we have seen, some such opposition seems required by Plato’s
formulation of the argument, which Aristotle is summarizing here.)
Now, if we mean by akrasia or “weakness of will” acting or doing some-
thing knowingly or intentionally against one’s better judgment, it seems
clear, as Engberg-Pederson has argued in his book on oikeiōsis,25 I believe
correctly, that for the Stoics the passions are cases of akrasia or, more
precisely, the classes of passions and cases of akrasia are coextensive with
one another. The Stoics do not therefore wish to deny the existence of
akrasia, as has often been maintained – nor could they reasonably have
defended the Socratic Paradox, which declares that no one does wrong
willingly. On the contrary, they wish to maintain that wrong actions and
their motivations (i.e., passions) are all voluntary and open to moral eval-
uation. In this they resemble Aristotle, who likewise holds that akrasia
is voluntary and blameworthy, and who likewise seems to allow akrasia
(along with its opposite enkrateia) an important role in the formation or
modification of character. Passions are genuinely akratic because the lo-
gos that they are disobediences to is one’s own dominant or hegemonic
logos, not merely the “right reason” which states the objective prescription
that applies to the particular situation in which one finds oneself. This
is because all rational animals have bestowed on them and continue to
possess natural preconceptions about what is good, however confused or
inarticulate these may be in the case of an individual person. (Oikeiōsis,
or anyway proper oikeiōsis, involves the process of coming to articulate
these preconceptions properly, and to fit them correctly to particular sit-
uations.) Thus one will always possess, or at least have available to one,
a correct and relevant doxa about the good – unless, of course, one has
attained sagehood, in which case one will possess epist¯e m¯e or knowledge
about the good.
For the Stoics it is this doxa that is overcome by epithumia (which is in ef-
fect another doxa) in the case of akrasia, a position that Aristotle explicitly
tries to refute. Aristotle holds, at least in the De Anima, rather that akrasia
is orexis overcoming orexis (3.11), which fits with his general view that de-
sire is the primary cause of action, and that thought is subordinate to it
(3.10; De Motu Animalium). The focus on the role and function of practical
reason in Aristotle’s “physical” discussion of akrasia in Nicomachean Ethics
 On the Voluntariness of the Passions 45

7.3 is due to it being Socrates’ views on akrasia that are there being dis-
cussed. But Aristotle’s “physical” solution to the Socratic puzzle about
akrasia is remarkably similar to the Stoics’ own picture, as I have devel-
oped it, especially in attributing akrasia to beliefs that conflict, but not
per se (logically).
For the Stoics, in cases of (real or apparent) weakness of will, it is not
passion that is opposed to reason: rather, reason is opposing itself, reason
is as it were disobeying reason. If one decides to do something against
one’s better judgment, as we say, what is occurring is that there are two
judgments in the soul that are opposed to one another, the one declaring
that it is appropriate to perform a certain action A, and the other that
it is not right to do so.26 Logically, these two judgments are, of course,
contradictory, but this only entails that they cannot both be entertained,
or in Stoic terms assented to, by the same mind at the same time. One will
find oneself wavering between them. Copious evidence in the form of lit-
eral quotations from Chrysippus’ treatise on the passions cited by Galen
in the central books of his work On the Doctrines of Hippocrates and Plato
(evidence that Galen himself seems often willfully to misunderstand) re-
veals that Chrysippus tried to show in detail the sorts of psychological
mechanisms a person may engage in so as to accommodate the presence
of such a contradiction in her practical commitments. He compares this
kind of situation, for instance, with a person’s refusal to believe anything
bad about his lover, even though one may possess the incriminating re-
port on unimpeachable authority. That is, one may decide not even to
entertain a certain proposition, in the knowledge that if one did enter-
tain it, one would be compelled to assent to it. Chrysippus, as reported by
Galen, quotes a fragment of the comic poet Menander: “I got my mind
in hand and stuck it in a pot.”27
Chrysippus also seems to have undertaken an elaborate exegesis of
Medea’s famous monologue in Euripides’ play (Medea 1021–80), where
she addresses her own thumos or anger at her husband as the motive for
her prospective murder of her children, and utters the famous lines:

I know what kind of evils I am about to do

but anger [thumos] is strongest among my deliberations.28

Traces of this exegesis are no doubt to be found in Epictetus’ discussion

of the same scene in his Discourses (1.28.7). Contrary to what Galen insists,
Chrysippus need not have been compelled to see here a direct opposition
between Medea’s reasoning and her emotion of anger, as something
separate from reason. Rather, the source of her anger, according to
46 Steven K. Strange

Epictetus and Chrysippus, is her judgment that it would be an overwhelm-

ingly good or appropriate thing for her to take the most terrible revenge
upon her husband for his abandonment of her. Of course, she also thinks
that it would also be a good thing to spare her children’s lives. If she also
thinks that this would be an appropriate action, then we would find her in
the sort of akratic situation mentioned earlier: the overwhelming power
of her positive evaluation of revenge would cause her just to refuse to
think about the wrongness or moral inappropriateness of the murder
of the children, at least until after the savage act has been carried out.
But this does not seem to be Epictetus’ analysis of the situation. On his
interpretation, Medea thinks that these two goods are in conflict but that
revenge in the present situation is “more profitable” (sumphorōteron) for
her than saving her children (1.28.7). Of course, this belief is false: the
gratification of anger is never a real good, only an apparent one, and as
Epictetus repeatedly insists throughout his Discourses, real goods actually
cannot conflict: the world and human life are so constructed that true
goods form a single, coherent, mutually realizable structure. But Medea,
not being a Stoic and trapped in the confusion of ordinary human moral
beliefs, fails to realize these facts. Given her beliefs, she must do what
she thinks she must do. Epictetus even pretends to admire her for the
strength of her soul (2.17.21). This is odd, since it seems we should an-
alyze Medea’s akrasia as the species of it that Chrysippus called “from
weakness” rather than “from rashness.”29
To express admiration for Medea may seem a strange thing for a Stoic
moralist to say, but in fact it is quite consistent with the general Stoic atti-
tude about the moral evaluation of wrongdoing. I return to this point
in closing. For now, I want to point to another important feature of
Medea’s situation, namely the relative intensity of the evaluative judg-
ment involved in the passion. Chrysippus emphasizes that in the case of
a strong passion at least, the object is taken by the agent as being over-
whelmingly good or bad (as being the greatest good or evil). I suggested
earlier that this was the gist of the second in Chrysippus’ analysis of the
two judgments that found emotion. We know that in his psychotherapy
he emphasized against Cleanthes the importance not only of trying to
persuade people of the Stoic doctrine that nothing outside the mind
is actually really good or evil and hence that no pathos can be justified,
since all passions involve positive or negative evaluative judgments about
external objects but, first, that external things are less important than
we ordinarily think, and less than virtue and vice (a view, of course, that
Peripatetics and Platonists agreed upon with the Stoics).
 On the Voluntariness of the Passions 47

assent and the voluntariness of passions

Passions, and the emotions more generally, thus depend on the Stoic
view entirely on the judgments of reason concerning the goodness or
badness of particular objects, or the appropriateness or inappropriate-
ness of particular actions, that is, pursuits or avoidances of objects. There
is an additional complication here that deserves to be mentioned. In
the case of the truly morally good acts of the sage, the object pursued
seems actually to be what the Stoics sometimes call the proper “selec-
tion” (eklog¯e ) of external objects, and this forms the practical content of
virtuous action (this is the notion of kath¯e kon or obligation). This is what
should be meant by the idea that the sage pursues only his own moral
virtue. The sage is nevertheless said to have horm¯e toward the external
objects, as when from obligation he acts to preserve his health (rather
than because it is good in itself to do so). In this case the objects are said
and seen by the sage to possess positive value (axia) rather than goodness.
But the object of his orexis or pursuit will be the goodness of the action
of proper or appropriate selection.
Nevertheless it is always the goodness or perceived apparent good-
ness of an object (or its badness, in the case of aversion) that motivates
an action. This in turn will form the content of an evaluative judgment
or act of assent. Actions, emotions, and desires are all subject to praise
and blame or moral evaluation, the Stoics think, for precisely the same
reason: because this sort of act of assent lies in our power to perform or
not to perform. This is true not only in the case of evaluative judgments:
whether we believe that any given proposition is true or false will be an
act of assent, which will be up to us in the same way.30 For instance, even
perceptual beliefs or aisth¯e seis involve acts of assent, as is evident when
we reflect on the fact that we sometimes do not accept what our senses
seem to be telling us – for example, that the pencil in the glass of water
is actually bent, or that the sun is a relatively small object up in the sky.
The example of perception also helps make clear the important fact that
it is not always within our power to withhold a particular act of assent:
the Stoics hold that we cannot refuse to believe the clear evidence of
our senses – for instance, I cannot fail to believe that I am now speak-
ing to you or that you are now listening to me. (In fact, our very word
“evidence” seems to derive from this Stoic view about what they call
enargeia, obviousness, although the proximate source is Cicero’s use of
the term in his discussion of skeptical epistemology in the Academica.)31
This is their famous doctrine of the katal¯e ptik¯e phantasia or compellingly
48 Steven K. Strange

clear-and-distinct (cataleptic) impression, which we are unable to resist

assenting to.
The same holds in the practical or moral case. Here the phantasia or
impression is of the appropriateness or inappropriateness of an act, of the
pursuit or avoidance of a particular object as being good or bad for us (cf.
Epictetus, Discourses 1.28.5–6). We cannot fail to pursue what we grasp as
obviously good or to avoid what is obviously bad, which is to say that there
are practical cataleptic impressions as well as theoretical ones. But in nor-
mal circumstances (which is to say, as long as we have not yet attained to
sagehood) these are not the sort of practical (or hormetic) impressions
we are characteristically presented with. Fortunately, the Stoics think, the
world is so providentially arranged that all irresistible (cataleptic) impres-
sions are, in fact, true. (The sage assents only to cataleptic impressions: if
not the cataleptic impression that a certain action is either good or bad,
at least the impression that it is reasonable that it is good or bad.) The pur-
pose and, indeed, the very nature of our reason is to distinguish (krinein)
between true and false impressions (i.e., between those that present us
with true as opposed to false propositions) – in the practical case, be-
tween true and false impressions that it would be good or bad to pursue
or avoid certain things. Because all cataleptic or irresistible impressions
are in fact true, nature never leads us completely astray. In the case of
noncataleptic impressions, it is wholly up to us to assent or not.
Besides, the Stoics argue, there is naturally available to us a way to train
ourselves so that we never assent to false propositions or impressions –
this is apparently what the method of Stoic moral education, of making
progress toward wisdom, principally consists in. Hence it is always possible
that we could assent only to true propositions, that is, if we had properly
(and voluntarily) trained ourselves to do so. Hence it follows that virtue
and vice, and the complete avoidance of passions, which all involve false
evaluative judgments, lies completely within our power.32
If our assent, emotion, desire, and action are all up to us in the way
specified, then clearly our characters will be up to us too. Despite the
persuasiveness of the hormetic impressions signifying pleasures, and the
seductive power of our bad companions (the two sources of moral per-
version, according to the Stoics), we have in principle the possibility of
resisting them – in the precise Chrysippean sense of possibility, that this
ability could have naturally occurred in us in this situation and that noth-
ing external prevents its having done so.
Aristotle in the second part of Nicomachean Ethics 3.5 strives to answer
an objection, based on the Socratic Paradox,33 that we cannot be held
responsible for our own bad characters, because it is not up to us how
 On the Voluntariness of the Passions 49

the good appears to us. But Aristotle’s position in that chapter – that
we can be held responsible for our bad moral characters because we are
responsible for the voluntary actions that went into our habituation that
produced those characters – won’t do, because by Aristotle’s own lights
how the good appears to us, that is, whether we possess a concept of the
good that is at all adequate, already depends on our character. Hence,
Aristotle is committed to saying that a person can be held responsible for
her conception of the good if she is already assumed to be responsible
for her character, so that Aristotle’s argument is circular. It is clear what
the Stoics’ way of dealing with this sort of problem will be: to insist on
the basic principle that we are responsible for how the good appears to
us. This also means that they reject Aristotle’s further claim that a vicious
person has irretrievably lost the proper conception of the good and hence
is a morally hopeless case. This is not true for the Stoics – fortunately so,
since all of us, at least nowadays, are vicious fools.
The notion of power or possibility appealed to here must be un-
derstood within the context of Stoic (i.e., Chrysippean) determinism.
Chrysippus holds that everything that happens, happens in accordance
with fate (heimarmen¯e ), and thus must occur in exactly the way that it
does both because it is causally determined to happen in just that way,
and because in so doing it makes a contribution to the overall perfec-
tion of the universe (universal teleological determinism). Included in
everything that happens are, of course, all human assents, pursuits and
avoidances, and emotions, which are just as fated as is anything else.
Nevertheless, these are held by the Stoics to be open to moral evaluation
and praise and blame. Important passages from Cicero (De Fato 40–44),
Aulus Gellius (Noctes Atticae 7.2.1–14), and Plutarch (De Stoicorum Repug-
nantiae), supplemented by some others (perhaps, for instance, Origen De
Principiis 3.1–6) show us Chrysippus’ way of constructing a compatibilist
defense against objectors who claim that Stoic determinism undermines
the legitimacy of any attribution of responsibility. I think we are now in
a position to see how this defense works. It crucially depends on the vol-
untariness of the passions, our control over our evaluative attitudes and
judgments.

Notes
1. Its origins have often but usually unconvincingly been discussed. They have
in recent years been the subject of a set of Sather Lectures by Michael Frede,
which will hopefully soon be published.
2. Cf. Cicero, Tusculan Disputations 4.12, Diogenes Laertius 7.115, and SVF 3:173,
also Inwood 1985: appendix 2, p. 237.
50 Steven K. Strange

3. Compare the conditions of the vicious persons described in Plato’s Repub-

lic, books 7 and 9, where boul¯e sis seems to be, even if present, at least not
operative in the preferred way, as a desire for the person’s true good.
4. In Varro’s speech in Academica 1.40–41.
5. This is somewhat of an oversimplification, since we do not know the precise
details of the Stoic view of practical judgment (cf. Bobzien 1998: 240–41);
however, it seems safe to assume that it involved assent to or approval of a
practical impression, analogous to assent to the truth of an impression in
the theoretical case.
6. One could mention here, among others, Frede 1986, Engberg-Pedersen
1990, Nussbaum 1994 (and Nussbaum 2001), Brennan 1998, and the rele-
vant section of Long and Sedley 1987.
7. See Inwood 1985: 228–29. Epictetus makes orexis or desire or pursuit and
horm¯e distinct genera, the former seemingly directed at the good (and away
from evil), and the latter directed at appropriate or preferred objects. I
assume that this represents a later development in the Stoic theory.
8. Perhaps this what is meant by hexis horm¯e tik¯e at Stobaeus 2.86.17 = SVF 3:169.
9. But perhaps commotio is just a translation of kin¯e sis.
10. Cf. huperteinousa ta kata ton logon metra, Clement of Alexandria in book 2 of
the Stromateis (SVF 3:377).
11. Cf. Chrysippus in LS 65J4 for both expressions.
12. Recall again that the eupatheiai, by way of contrast, are properly restrained
impulses.
13. LS 65A4.
14. Here the contrast term to orexis is orousis; on all this, see Inwood 1985,
appendix 2.
15. Galen 1981: 4.3.2, 4.4.38, and 5.6.33–38, respectively.
16. See Cooper 1999, following the lead of Fillion-Lahille 1984.
17. Cf. also the “boiling of the blood around the heart” in his “physicist”’s defini-
tion of anger in Aristotle, De Anima 1.1 fin., corresponding to the intentional
“dialectician”’s definition as a desire for revenge.
18. The details are complex and controversial. It is, however, clear that the term
prosphatos applied to the value judgment causing an emotion was originally
due to Zeno. The secondary judgment seems to have been brought in by
Chrysippus to explicate this term of Zeno’s.
19. I assume the accusatives ti and h¯e don are to be retained; cf. also 250.9 and
342.30 Müller.
20. Cf. LS 53Q4.
21. Epictetus (Encheiridion 1 init.) lists as what is up to us is hupol¯e psis, suppo-
sition (which includes belief), impulse, desire, and aversion, “and what are
our doings [erga].” All the items on the list are, in terms of the Old Stoic
classification, either assents or hormai, and thus depend on assent.
22. Compare Seneca, De Ira 3.3, and Posidonius as reported by Lactantius, On the
Anger of God 17.13 (reporting the contents of the lacuna at Seneca, De Ira 1.2).
23. Galen 1981: 5.7, claims that the argument establishes indifferently either that
there are different parts or different faculties (of reason and appetite, or
appetite and spirit) within the soul (cf. 5.7.49–50). He does not seem to
 On the Voluntariness of the Passions 51

think it matters which. His interpretation certainly seems to gain support

from the fact that Aristotle, who uses the same argument for dividing the soul,
seems to think that reason and the desiderative part may only be different
faculties, not different parts, of the soul. But this is so, it seems to me, only if
a difference of faculty is enough to allow simultaneous opposite intentional
movements having the same object in a unitary soul. I am unsure as to whether
this is, in fact, the case.
24. Plutarch, De Virtute Morali 446f–447a = LS 65G.
25. Engberg-Pedersen 1990.
26. Again, cf. Engberg-Pedersen 1990 for a clear exposition of the correct idea
here.
27. Ton noun ekhōn hupokheirion / eis ton pithon dedōka, Galen 1981: 4.6.34.
28. Alternatively, “Anger is stronger than my deliberations”: kai manthanth’ men
hoia dran mellô kaka / thumos de kreissōn tōn emōn bouleumatōn (1078–79).
29. See also Chrysippus’ analysis of Menelaus’ failure to carry out his resolve to
kill his faithless spouse Helen in the sack of Troy, Galen 1981: 4.6.
30. See note 6 on the analogy between theoretical and practical judgments
(assents) for the Stoics.
31. For discussion, see Zupko 2003: 360, n43.
32. See Origen, De principiis 3.1.4–5, a very Chrysippean-sounding passage, for-
tunately preserved in Greek, in which Origen is responding precisely to the
objection that in at least some cases of immoral choices (the example is of
a monk being seduced to violate his vow of chastity) “the impression from
without is of such a sort that it is impossible to resist it.”
33. Probably, however, as this was maintained by Plato in the Laws, and not as
held by the historical Socrates.