English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
Aristotle says that animals are agents, and also that only humans are.1 To avoid contradiction there must be two senses of agency at play: ‘broad’ agency, common to animals and humans, and ‘narrow’ agency, proper to human adults. So the question arises: what makes human adult agency distinct from that of animals and children? Perhaps the latter lack, but the former have, logos (e.g. Pol.I.2.1253a9-18) and can use reasoning (logismos) in action production (EE.II.8.1224a28-30). Thus, rationality seems a clear distinguishing feature of human agency. This intellectualist interpretation is prevalent in contemporary scholarship, but an anti-intellectualist challenge has recently emerged: although only humans have reason, non-rational cognition (of the kind we share with animals, i.e. perception and phantasia) still governs the startingpoint of human action, i.e. the setting of goals.2 While the recent charge against the intellectualist orthodoxy has received mixed responses,3 it has renovated the debate about Aristotle’s agency theory.
This essay argues against both intellectualist and anti-intellectualist views: neither can do justice to all of Aristotle’s key claims about agency. The essay (1) characterizes the principles of animal agency, and tests whether intellectualists or anti-intellectualists can successfully explain human agency. First, (2) after assessing prominent intellectualist accounts, it becomes apparent that they cannot accommodate solid textual evidence according to which the human soul’s non-rational part sets the goals, and the rational part sets the things relative to the goal. Further, (3) recent anti-intellectualism is unable to explain how the rational soul can prescribe over its non-rational counterpart. This calls for (4) a new interpretation that preserves the merits and avoids the pitfalls of both traditional views.
1. Principles of animal agency
In the Aristotelian texts, actions are primarily locomotions. The discussions of animal movement (DA.III.7-11; MA.6-11) treat ‘acting’ (prattein) and ‘moving’ (kineisthai) largely interchangeably, and ‘action’ (praxis) refers to an animal’s locomotive pursuit or avoidance of some object. Aristotle calls locomotion ‘voluntary’ and labels it ‘action’ whenever the principle of motion is internal to the mover.4 The crucial notion of a ‘principle of motion’ gets expressed in several different ways. For instance, agents are called kurioi of their actions, and our actions are said to be eph’ hēmīn. Aristotle gives the word ‘kurios’ (originally meaning ‘master’ or ‘lord’) an abstract turn as ‘commander’ or ‘controller’. Thus, we are kurioi of our actions in the sense that we are in control of them. On the other hand, Aristotle claims that X is eph’ hēmīn, or up to us, meaning that it is up to us to do X or not to do X, i.e. that whether X happens or not depends on our causal influence in the world.5
Thus agents are entities capable of producing bodily movements of which they themselves are principles. Such movements can hence be called ‘actions’, and insofar as they are generated by inner principles, they are voluntary and up to us. Thus, the difference between human and animal agency should depend on the principles of motion proper to each.
1.1. The ‘faculty of desire’ as a principle of animal motion
In a crucial passage, Aristotle claims there are two key principles of animal motion: cognition and desire.
We see that the things that move the animal are cognition, phantasia, decision, wish, and appetite. But all of these come down to cognition [noūs] and desire [orexis]. For phantasia and perception take the same place as cognition, since they are all discerning faculties [kritika], although they differ from one another in the ways we have mentioned elsewhere. On the other hand, wish, spiritedness, and appetite are all desire, and decision is common to both cognition and desire; hence the things that move primarily are the object of desire and the object of cognition. (MA.5.700b17-24)
The faculties of perception, phantasia, and intellect are all grouped under the label ‘discerning faculties’ (kritika) because they are all cognitive in a broad sense of the term: they allow animals to discern (krinein) the world’s entities. So for all animals a cognitive, discerning faculty is one of the principles of action.6 The same goes for desire: animals can have desires for simple, immediate objects (i.e. ‘appetites’ or epithumiai), or more complex desires (i.e. ‘wishes’ or boulēseis, desires implying reason), but in any case some kind of motivational process must play a role in explaining why they move. Hence cognition and desire are principles of action, inner sources of animal motion.7
Now, these capacities are not independent; cognition and desire are intrinsically linked by pleasure: all animals who have perception have desire, Aristotle holds, because perception implies the ability to discern pleasure from pain, pleasant from painful objects; and if an animal perceives these, then it has the most basic form of desire: appetite (DA. II.3.414b1-6). So pleasure and pain are perceptual experiences, and viceversa, perception can reveal objects as pleasant or painful. Therefore, perceptions (and the phantasiai that follow them) are motivational: they can direct the animal towards certain objects or away from others. Cognition moves the animal according to its nature:
Taste [belongs to all animals] because of nutrition. For by means of it they discern [diakrinei] the pleasant and the painful concerning nutrition, so that they flee from the latter and pursue the former […]. (Sensu.1.436b15-17)
Perceptual objects appear as pleasant or painful, i.e. as worth pursuing or avoiding, and thereby also as good or bad. In other words, goodness and badness, insofar as they coincide with pleasure and pain, are also perceptible properties of objects. ‘Perceiving something as good’ does not require intellectual cognition; it just means that in perception it is revealed as worth being pursued; as to-be-sought, or to-be-done-that is, as pleasurable. The argument for taste applies to the other sensory modalities: if an animal is to find nutrition and preservation by moving, then it needs senses that guide it toward naturally pleasant things, and away from naturally painful objects (Sensu.1.436b18-437a2).
Thus, pleasure links animal cognition and desire toward the pursuit or avoidance of pleasant or painful objects. So the sensory, volitional, and motor capacities work together. This seems to be what Aristotle means by claiming that perception, desire, and motion are identical in activity, but different in being (DA.III.7.431a8-14). These capacities, though definitionally different, constitute a functional unity whose different aspects are always co-activated, and which Aristotle sometimes calls “the faculty of desire [to orektikon]”.8 Whenever this complex faculty causes the animal’s motions, we can say that the principle of motion is internal to the animal.
1.2. Natural dispositions as principles of animal motion
But one more element is missing. For although Aristotle claims that there is only one mover, namely the faculty of desire (to orekton), he goes on to say that the first mover is not said faculty but “the good itself”, the external object of desire. Thus, “the unmoved mover is the good attainable in action, and the mover that is moved is the faculty of desire” (DA.III.10.433b15-17). The object of desire seems therefore to be the first principle of action, its presence (sensed through perception or imagined through phantasia) being what triggers the faculty of desire’s activation. If so, the first principle seems to be external to the agent (cf. Labarrière 1997). No movement would be initially originated by the agent’s internal principles, and animals would therefore not be agents: the faculty of desire merely reacts to the object’s presence.
However, what moves the animal is not the object as such, but the object as pleasant or as painful. And the object is not pleasant or painful in itself, but only in an animal’s discerning experience of it. In fact, what makes a certain animal cognize a given object as pleasant or painful is largely established by nature, since “what is in accordance with nature is pleasant; and each animal pursues that which is pleasant in accordance with nature” (HA.VIII.1.589a2-9). So rather than the external object, the principle of motion should be identified with the animal’s natural dispositions of practical perception. They determine which objects become objects of desire by appearing pleasant or painful for each animal in accordance with its organic functions.9
Thus, the principles of animal agency are its “faculty of desire” and its natural cognitive dispositions. Whenever these two principles explain the animal’s motion, such motion is an action, and the animal an agent. Now, what distinguishes the principles of animal agency from those of human adult agency? The next two sections examine the two dominant ways of answering that question.
2. Assessing intellectualism about human agency
Aristotle mentions many features that make humans unique: we are the only animal that has not just memory, but also recollection (HA.I.1); humans are the most imitative animals, and learn their first lessons through imitation (Poet.I.4); human sight, hearing, and smell are comparatively underdeveloped, but our senses of taste and touch are the most accurate of all; and this makes us the most intelligent [phronimotaton] animal (DA II.9). Only humans take pleasure in smells that are not related to food, like those of flowers (Sensu.5). Man has the smoothest skin because, in proportion to size, the human male emits more semen than any other animal (HA.VII.2). The list goes on.10 But what among these features makes us more complex agents?
Here is a common answer: if we add many of these specific traits (like our capacity to recall past events, great imitative skills, and superior senses of taste and touch), they lead to the claim that we are the animal that has reason (logos) and calculation (logismos).11 We may safely say that a key distinctive feature of humans is the possession and use of reason in action production.
But what exactly is reason’s role in human agency? This is where the debate begins. Intellectualists claim that reason, among other things, determines the goals of human action. Anti-intellectualists hold that goal-setting does not belong to reason, but to non-rational faculties.
This section assesses the merits of different versions of intellectualism, by first (§2.1) distinguishing different functions that can be assigned to practical reason, and then (§2.2) discussing the merits and limitations of diverse intellectualist positions.
2.1. Varieties of intellectualism
Aristotle claims multiple times that the non-rational part of the soul determines the goals of action, whereas reason is in charge of establishing only “the things toward the goal” (more on this below). Thus, reason’s realm of operation differs according to the interpretation of the phrase “things toward the goal”. These are three of the most influential interpretations of practical reason’s functions:
Instrumental function - The most restricted interpretation of “things toward the goal” takes it to refer simply to the means required for the goal. This gives reason the merely instrumental function of identifying steps necessary to achieve an end. Such interpretation can be called ‘Humeanism’, since reason turns out to be little more than an assistant for non-rational goal-setting processes.12
Specification function - More possibilities open up if “things towards the goal” includes more than just means. An influential line of interpretation holds that the goals that character specifies are rather abstract (e.g. ‘happiness’, ‘doing well’, ‘the fine’ or ‘the noble’), and therefore reason must not only establish the means, but also determine what counts as attaining the goal in each practical situation-what counts as doing well, or as a fine and noble action, here and now. Thus, practical reason’s role is not solely instrumental, but also specificatory: the goal’s specifics are determined by deliberation. Interpreters have pointed to this function of practical reason as substantiating intellectualism about goal-setting.13
Coordination function - Reason can also play the role of assessing a given goal by contrasting it to the agent’s other goals: whether it fits with her overall conception of the good, whether it is preferable to competing goals, or whether seeking it would generate negative longterm consequences. If reason determines that a given goal is not in accordance with one’s broader practical concerns, then the agent can override it. This interpretation gives reason a coordination role, which allows the agent to lead a consistent and unified life, responsive to the general concerns and values purportedly established by character: Since “the eligibility of a target […] is tested by deliberation” (Price 152-153), reason enables humans to exert more control over their actions than other animals, and may even be said to determine which goals are pursued: thus intellectualism can also be defended via reason’s coordinating function.14
These descriptions of Aristotelian practical reason are not incompatible. In fact, we may stack all these functions together and argue that human agency differs from animal agency due to the greater control afforded by practical reason’s three functions. Reason receives a somewhat indefinite motivational input from the agent’s character, and structures it by making it specific, rendering it consistent with broader goals and plans (or override it if inconsistent), and providing a specific set of steps to take toward its achievement. All of these features are unique to humans, since they are different uses of calculation (logismos) and deliberation (boulē), capacities absent from all other animals. For them, the appearance of an object as good or bad immediately generates a desire, which leads immediately and invariably to action; but human capacities for rational deliberation and calculation stand in between desire and action, allowing us to not act even in the presence of an apparent good. It is truly up to us (and not up to some innate natural disposition) whether an action occurs or not. All of this supports intellectualism:
[H]uman beings are special among animals in having a capacity for articulable thought. Purposive behaviour in brutes is an immediate response to an opportunity for gratification of non-rational motivational impulses […]. The peculiarly human capacity for thought allows for purposiveness without that immediacy; thought can mediate gaps between project and execution. (McDowell 1998, 23)
2.2. The problem with intellectualism: the division of labour
Two of the three main functions of practical rationality (specification and coordination) seem to support the intellectualist claim that reason can determine the goals of action. But there is an obstacle, a well-documented claim in Aristotle’s extant texts:
(DL) The Division of Labour
Character, and not reason, sets the goal of human action. Reason must take the goals set by character as its starting points.
In other words, reason cannot directly control the appearance of the good (i.e. the appearance of objects as pursuable or avoidable): reason cannot make something appear good if character did not already present it as good. Reason can, of course, present something as a good means to something else, or as a specification of an already given end, but those rational appearances, however important, are strictly dependent on prior character-based, non-rationally-determined appearances of the good. If so, reason cannot use its aforementioned functions to control the first principle of our actions.
The remainder of this section (§2.2.1) summarizes the textual evidence in favour of DL, and (§2.2.2) shows that current versions of intellectualism cannot accommodate DL.
2.2.1. The evidence for DL: Goal setting is out of reason’s reach
Reason’s three functions are applications of deliberative and calculative capacities, i.e. of logismos. But [i] practical reasoning cannot provide its own starting points. And since those starting points are goals, [ii] deliberation cannot be about goals, but only about things toward the goals. Moreover, [iii] goals are determined not by reasoning, but by habituation and character.
[i] Aristotle claims that there is no logismos of theoretical or practical starting points. He asks whether virtue produces the goal (τὸν σκοπὸν) or the things toward the goal (τὰ πρὸς τὸν σκοπόν), and replies: “We posit the goal, because there is no argument [sullogismos] or reasoning [logos] of it. Rather, it must be presupposed like a starting point.” And after an analogy between hypotheses in theoretical reasoning and goals in practical reasoning, he concludes: “Therefore, if either virtue or reason is the cause of all correctness, and reason is not the cause, then virtue will be the cause of the correctness of the goal, but not of the correctness of the things toward the goal”.15 These remarks fit nicely with the view of reason as a mediator between goal and action, but also imply that reason cannot establish its own starting point: the goal.
[ii] Moreover, as Aristotle goes on to argue in the next few lines, decision (prohairesis) is concerned only with things toward the goal, but is never concerned with the goal itself (EE.1227b38-39; cf. EN.III.2.1111b26-29). Decision, itself a product of deliberation,16 can have only the things toward the goal as its object, because, as Aristotle repeatedly claims, “we deliberate not about the goals but about the things toward the goals”.17 So practical reasoning is about the things toward the goals rather than about goals themselves. But if not through reason, then how are goals determined?
[iii] The non-rational part of the soul determines goals because it is in charge of pleasure and pain,18 on the one hand, and of habits and character, on the other. As seen above (§1.1), an animal’s natural dispositions quite rigidly determine its tendency to experience things as pleasant or painful via perception and phantasia. But human innate dispositions are much less definitive. Our initial natural dispositions are similar to those of other animals, but our practical perception can be largely re-shaped through habituation, so that we end up being pleased or pained by the things that we get used to being pleased or pained by, and enjoy the activities that become familiar to us through repetition.19 We are also born with natural dispositions, and to that extent the souls of children “do not differ at all, so to speak”, from the souls of animals (HA.588a32-b2).20 But while their natural dispositions provide animals with sufficient cognitive adaptation to their environments, human innate dispositions are quite deficient: we are born less hard-working and more pleasure-loving than is good for us.21 Thus we need habituation processes that build our character and shape our phantasia of the good. Therefore, instead of natural dispositions, it is our character which determines our cognitive reactions to the practical world.22
Moreover, character determines not only which things one loves and hates, but also whether one listens to the voice of reason or simply lives in accordance with one’s passions-and most people could not care less for the voice of reason (EN.X.9.1179b23-31). Arguments do not affect most people’s actions, because they were not habituated to the use of practical reason.
In sum, DL is supported by at least three strong lines of textual evidence: the starting points of practical reasoning cannot be obtained via reasoning; practical reasoning (i.e. deliberation) is not about goals but only about things toward the goals; our phantasiai of the good are set by the characterbuilding processes of habituation, and only those habituated into rational activity can understand reasons and act accordingly.
2.2.2. Intellectualism is incompatible with DL
Anyone with strong intellectualist intuitions (and that includes most Aristotle scholars) would be unhappy with the way this is going. If reason does not play a role in it, then goal-setting seems an entirely unintelligent matter, determined by the blind mechanism of habituation. This kind of anti-intellectualism seems to turn Aristotle’s ethical texts themselves into superfluous exercises. Aren’t those texts largely dedicated to rationally establishing the goals of human life? Why would Aristotle have dedicated such theoretical effort to a task his own theory renders impossible? These are serious concerns that anti-intellectualism must face (§3). But first, several forms of intellectualism must be put to the test. Despite its popularity, intellectualism is indeed a difficult position to hold. After all, in its strictest form, “the intellectualist case requires that an agent can form a conception of goodness through reasoning without appeal to values already desired by the agent. The idea is that the conception so formed produces desires in the agent rather than reflecting desires she already has” (Grönroos 2015, 66).
Such radical intellectualism is very much at odds with the textual evidence for DL (§2.2.1). But many scholars have argued for it nonetheless. I now examine three differently extreme versions of intellectualism: (A) that through its goal-specification function reason constitutes those goals; (B) that wish, being a specifically rational kind of desire, allows us to set fully rational goals; and (C) that the intellectual capacity called noûs can grasp practical starting-points.
(A) Specification as constitution
Given reason’s specification function, intellectualists can argue that, beyond determining goal-extrinsic details (like means or consistency relations among them), logos also determines goal-intrinsic features, i.e. what exactly the goal is in particular situations. They know that this implies rejecting Aristotle’s explicit division of labour between character and reason. And they are prepared to do just that, by considering DL a consequence of unfortunate phrasing, or an outright mistake on Aristotle’s part. Thus, McDowell (1998, 30) claims Aristotle risks obscuring his own view with DL; and Irwin considers DL “at least misleading” (1975, 576) because “a wise man’s grasp of the end is the result of his deliberation”, and so “Aristotle is wrong to claim that there is no reasoning about ethical first principles” (578).23
To justify DL’s dismissal, specification-intellectualists often argue that character-set goals are too vague, and therefore rational calculation must turn them into something concrete. Thus, for Wiggins “the problem is not to see what will be causally efficacious in bringing [the goal] about, but to see what really qualifies as an adequate and practically realizable specification of what would satisfy this want” (1975, 38). Given the extreme vagueness of character’s goals, practical reasoning must step in to actually “determine the content of the correct conception of living well” (McDowell 1998, 27; cf. Irwin 1975, 572).
But one must ask: are character-determined goals so vague as to require deliberation to acquire some minimally specific content? This cannot be right.24 First, Aristotle compares the practical starting points with those of the theoretical sciences. The latter must be sufficiently defined to make subsequent deductive processes possible: the content of geometric axioms, definitions, and postulates must be specific enough to make all of Euclid’s theorems deducible. Likewise, habituation into a certain way of acting should give agents starting points powerful enough to produce corresponding actions.
Further, Aristotle’s examples of practical starting points are often much less vague than ‘happiness’, ‘doing well’, or ‘the fine and the noble’. In discussing DL, Aristotle claims that the doctor does not deliberate about whether to heal, or the gymnast about whether to be in shape-healing and being in shape playing the role of deliberation’s starting points (EE.II.11.1227b22-30; cf. EN.III.3.1112b12-16). Earlier on (1227a13-15) Aristotle mentioned wealth and pleasure as potential starting points. If wealth, pleasure, healing, etc., count as character-determined goals, the view that reason must specify character’s overly-vague goals loses steam. Of course, there is still a lot of specification to do to take the broad goal of pleasure and turn it into a particular action, but the goal’s main content is already there when reason steps in. This is actually good news: it means we do not have to engage in outright rejection of solid textual evidence in order to deal with DL. For the evidence suggests that reason, even in its specificatory role, does not constitute the goal’s content, but rather “merely makes explicit what is already contained in the appearance, and thus in no way affects what goal we pursue (although it makes all the difference to how we pursue it)” (Moss 2012, 155).
Intellectualists may reply that some passages explicitly endorse intellectualism, e.g.: “it seems it is proper of the phronimos to deliberate finely about the things that are good and convenient for him, not piecemeal […], but with respect to the good life in general” (EN. VI.5.1140a25-28). But this is a report of someone else’s view in a dialectical passage (hence the “it seems”), and so should not be taken at face value. Even if it is, one may argue that “the things that are good and convenient” refer to the things toward the goal.
Consider then this: “the good deliberator without qualification is he who is skilful in aiming at [stochastikos] what is best for the human being in accordance with reasoning” (EN.VI.7.1141b12-14), which seems to connect deliberating well with properly determining the goal. But ‘being good at aiming at’ a target is different from determining the target itself. In fact, the target must be already sufficiently determined before someone can successfully aim at it.
There is also this:
So if having deliberated well is proper to the phronimos, good deliberation [euboulia] seems to be correctness about what is convenient with respect to the goal, of which phronēsis is a correct supposition [hupolēpsis]. (EN.VI.9.1142b31-33)
But the passage is ambiguous: “of which [οὗ]” can be read as referring back to “the goal” (thereby supporting intellectualism) or to “what is convenient with respect to the goal”. So an anti-intellectualist reading is available. Furthermore, even if we accept that phronēsis is the correct supposition of the goal, the meaning of ‘supposition’ [hupolēpsis] must still be specified. Another passage (EE.II.10.1226b20-30) leads Moss to interpret hupolēpsis as “the recognition that one is working towards a given end” (2012, 180-182). If so, then phronēsis entails having a correct recognition of one’s goal, but this recognition needn’t imply establishing or specifying the goal.
(B) Wish, the rational desire
Any view that takes goals to be a product of deliberation, or of any kind of logismos, faces the dead-end of DL: logismos simply cannot set goals. But intellectualists may take a different route: reason sets goals by producing wish, “a basic, and unreasoned, desire of the reasonpossessing part of the soul”.25
This intellectualism is more radical than the previous ones, because the specification process depends on prior habituated desires, but Grönroos’ wish is rational from the start.
Aristotle’s texts make the nature of wish far from clear. For one, it is unclear whether it belongs in the rational or the non-rational part of the soul.26 However, wish-intellectualism relies more strongly on clearer evidence: Aristotle holds that wishes are for the end as opposed to the things toward the end, and that they are for the good as opposed to the apparent good.27 Grönroos constructs an illuminating account of wish out of these two elements, but despite its strengths it does not seem sufficiently to justify intellectualism.
Wish is a desire for goals rather than for things toward the goals. But which goals is wish for? Aristotle says that we wish for many things, e.g. health or victory, and even things that are impossible, like immortality, or not achievable by our own action, like success for some actor or athlete (EN.III.2.1111b19-26). Despite this variability in its objects, Grönroos convincingly argues that wish has one proper object: being a desire for the good, if we take ‘the good’ to mean ‘happiness’ (as it often does), the specific object of wish turns out to be human happiness.
To justify this, Grönroos stresses the distinction between natural and non-natural objects of wish, which Aristotle draws through a medicine analogy: the natural goal of medicine is health, even though because of a perversion [strophē] it can pursue a goal contrary to nature. “And similarly, also wish is by nature for the good, but contrary to nature it is also for the bad, and the good is wished by nature, but contrary to nature, due to perversion [diastrophēn] the bad is also wished” (EE. II.10.1227a28-31).
Thus, wish has a ‘natural teleology’: it naturally pursues one object, even if perversions can make it pursue others. Further clarification comes from Aristotle’s treatment of a famous puzzle concerning the object of wish (EN.III.4): if the object of wish is the good, then people who choose things that are not good (e.g. intemperates choosing excessive pleasures) are not really wishing at all; and if wish is for the apparent good, then wish has no natural object. To solve this aporia, Aristotle argues that the natural object of wish is the good because, although the object of each person’s wish is what appears good to each one, in the virtuous person the good and the apparent good coincide (EN.III.4.1113a23-33).
But how can wish be at the same time only for the good, and also for whatever appears good to anyone? Grönroos argues that to clarify this we must distinguish two constitutive elements of wish: as a source of motivation, wish naturally aims at the human good; and as containing a representation of the good, wish directs the agent toward whatever appears good to her. As a source of motivation, wish is an originary rational desire for the human good, independent from prior non-rational motivations or prior reasoning processes. (This is good news for intellectualism.) However, wish often aims not at the real human good, but at other objects because the representation of the good it contains may be perverted. Crucially, even when perverted wish leads the agent toward the wrong object, wish still motivates the agent toward the real good (due to its intrinsic motivational nature). The error is representational, not motivational.
What causes this representational error? Since wish-intellectualists construe wish as a primordially rational desire, they must argue-as Grönroos indeed does-that reasoning errors produce representational errors: if someone has not reached the correct conclusions about the true nature of the human good, then she is led to wishing for non-natural objects. But this cannot be correct: mistakes about the apparent good are caused not by bad reasoning, but, as DL attests (§2.2.1), by bad habituation. Aristotle’s explicit view is that each agent’s representation (i.e. phantasia) of the good originates in habituation processes that determine her experience of objects as worthy of pursuit or avoidance. The EN.III.4 passage is explicit about this: “the fine and pleasant things are proper to each character disposition (hexis)”. So, the perversion that leads wish to aim at the wrong objects is not due to reasoning; the representational content of wish varies according to “each character disposition”, not argument.28 Saying otherwise would imply denying DL.29
Thus, although Grönroos provides powerful evidence that wish is a non-derivative source of rational motivation, this remains insufficient to ground the intellectualist view that reason can set an agent’s goals, because wish, although motivationally independent, is not representationally independent from the non-rational soul. Wish is therefore not a source of rational goal-setting: it is a source of rational motivation for whatever the agent was habituated into representing as a goal. Character determines the content of all desire-even wish.
Intellectualists may counter that people who were badly habituated, and now misrepresent the good, still wish for the real human good and feel inner conflicts due to their apparent good not matching the real good.30
This may be so, but it still falls short of demonstrating wish-intellectualism, which would require reason to provide remorseful vicious people with a character-independent representation of the goal. If reason could do this, we would expect to find in Aristotle’s texts cases of reflective conversion, in which vicious people correct their representation of the good through philosophy or some other rational activity. But such cases are nowhere to be found. What is easily found is Aristotle’s view that only properly habituated people can profit from practical reasoning, and that most people listen to their passions (like fear and pleasure), but not to reason (EN.X.9.1179b23-31; VII.7.1150a9-16).
Thus, “the decent person, living with a view to the fine, will listen to reason, whereas the base person, desiring pleasure, is to be punished just like a beast of burden” (EN.X.9.1179b31-1180a14).
So, although there is evidence in favour of the claim that wish is a basic, rational, character-independent source of motivation, the representational content of wish is provided by character. No solid foundation for intellectualism is to be found here.
(C) Practical noûs
Those who admit there is no deliberation about ends as such, but still want to say grasp of the ends is rational, can say that the grasp of the end is an intellectual intuition of the kind Aristotle calls noûs. After admitting that there is no deliberation about the ultimate end, and that the grasp of the good comes through habituation, Cooper concludes that the virtuous person’s knowledge of the ultimate goal “must be a kind of intuitive knowledge, not based on reasons of any kind” (1975, 62). However, he worries that a phronimos, if questioned about his conception of the good, may be unable to provide an articulable justification for it. The virtuous person should be able to argue in favour of his view of the good; but if not through practical reasoning, then what kind of reasoning could he employ?
Drawing a parallel between the theoretical and practical realms, Cooper argues that dialectic may do the job. In the theoretical sciences, dialectic allows for the grasp of a science’s non-demonstrable first principles. Dialectic argument is more informal than deduction: its raw materials are common observation and reputable opinion. Since there is no demonstration of a science’s first principles, they are said to be known by noûs, i.e. intellectual intuition.31 Arguably, EN.I is a dialectical argument seeking to establish the first principles concerning the human good. If so, the virtuous person can justify his views about the good, if not through practical reasoning (i.e. deliberation), then at least through dialectic. Thus, practical principles are grasped by noûs, and can be established by dialectical argument (60-65).
Cooper’s interpretation, though compelling, over-stretches intellectualism in one crucial respect. He claims that we grasp the good by intelectual intuition (or noûs), but while the intellect must no doubt be involved in dialectical argument, the dialectic concerning the human good does not have rational starting points. Recall that, although the EN.I dialectic provides the why concerning the human good, only those who already have acquired the what through proper habituation can properly grasp it.32 Because the habituated what is a prerequisite for fruitfully engaging in ethical dialectics, Cooper’s dialectic justification of intellectualism fails, since it cannot provide evidence that the starting points of dialectic itself are grasped intellectually. The only way to argue that this habituation-based cognition of the good is intellectual would be to say that habituation itself is an intellectual process of some kind. Cooper- and many others-think that this is the case, but that is a separate topic that must be considered independently. See Bermúdez (2016) for discussion of whether habituation has a rational component. I argue it does not.33 However, he should agree that the originary grasp of the good is not intellectual given his own theoretical commitments. Cooper distinguishes between the question “whether for any agent […] the commitment to an ultimate end is based on reasons”, and the question “whether or not considerations can be advanced which tend to show that a given end […] is the best end to treat as ultimate” (1975, 60). Although he gives a positive answer to the second question (dialectical argumentation can produce considerations in favour of treating a certain end as ultimate), he thinks Aristotle must answer “No” to the first question, precisely because grasp of the ultimate end originates not in reasoning, but in habituation. Thus,
It is prudent and reasonable, I think, of Aristotle not to insist that ordinary moral agents know, or need to know, how to defend their own view of what kind of life is best: their intuitive conviction is a perfectly adequate foundation for them. But those dialectical and critically alert persons who are moved to deepen their understanding of this foundation can do so by examining, in a dialectical spirit, the deliverances of moral intuition. (Cooper 1975, 71)
This is perfectly sensible: human agents (even those with a correct grasp of the good) need not have a sophisticated theoretical articulation of the good. Having it implies dialectic engagement with the issue-but that is an entirely different matter from having a habituated grasp of the what. So, although the why provides an intellectual grasp of the good, its foundation, the what, is a non-intellectual grasp of the good.
3. Assessing anti-intellectualism about human agency
In the last section I have argued that available intellectualist interpretations cannot account for the textual evidence. If so, does antiintellectualism fare any better? This part presents (3.1) the arguments for anti-intellectualism raised by its most prominent advocate, Jessica Moss, who reads Aristotle as a “practical empiricist”. Despite its strengths, I argue that (3.2) anti-intellectualism fails to account for the priority of reason over desire that Aristotle states explicitly.
3.1. Aristotle as a practical empiricist: Moss’s anti-intellectualism
Through a careful analysis of the division of labour passages, and of the role of phantasia in practical cognition, Moss argues that Aristotle endorses a form of “practical empiricism”: all the contents of practical thought come originally from sense-perception processes. Every object of wish, deliberation, and other rational processes originates in prior perceptual experience. Specifically-and following the analogy between theoretical and practical reasoning proposed in some of the DL passages-, just like theoretical principles are presupposed by reasoning, similarly the practical principles are established perceptually (or quasiperceptually via phantasia).
Practical empiricism entails that practical reasoning is constrained by earlier non-rational processes of perception and habituation: “On an empiricist epistemology like this, one will not come to […] make intellectual judgments about goodness which apply to things radically different from things one has actually pleasurably perceived, because it is those things which inform the concept” (Moss 2012, 234). This view can easily accommodate DL and its consequences. But Moss goes several steps further: since “reason does not rule in us as an independent force inserted as it were from above”, but instead “is dependent on non-rational cognition, both in genesis and in operation” (235), we must demote its relevance as a psychological force, acknowledging the primacy of non-rational forms of cognition.
This apparently turns Aristotle’s moral psychology into a kind of Humeanism: reasoning is not in control of volitions, but rather volitions control the cognitive faculty. Moss recognizes this as problematic, so she resists the inference: “Aristotle can hold a Humean view of practical reasoning as restricted to working out “things towards ends,” while still holding the very un-Humean view that we want our ends because we find them good, for his claim is that we find them good through phantasia” (Moss 159-160). Her view is thus that practical cognition still rules over desire and the passions, because the latter are determined by phantasia, which, though non-rational, is still a form of cognition (§1.1). So, goals are not blindly determined by desires.
There are two key claims of practical empiricism:
(PE1) Humeanism about reason Practical reason is restricted to working out the things relative to the ends that are established via perception, quasi-perceptual phantasia, and habituation.
(PE2) Non-Humeanism about cognition Goal-setting depends on phantasia, a non-rational kind of cognition.
These claims are meant to strike the proper balance between character’s goal-setting function and the dominance of practical cognition over desire. But do they?
3.2. The problem with anti-intellectualism: the priority of reason
Any intellectualist would certainly be upset about the minimal role Moss gives reason in the constitution of goals. And rightly so, because key aspects of Aristotle’s doctrine remain unaccounted for. Consider Aristotle’s claim that the non-rational soul “participates somehow” in reason, “insofar as it listens to it and obeys it”.
Also, correction and every punishment and exhortation reveal that the non-rational [part] is somehow persuaded by reason. And if it is necessary to say that this [part] has reason too, two [parts] will have reason: one will have it properly and in itself, and the other will have it in a way, by listening to it like a father. (EN.I.13.1102b30-1103a3; cf. EE.II.1.1219b29-1220a2)
So, the non-rational part must listen to reason’s ‘parental’ commands. This is a widespread hierarchical view, according to which the soul’s rational part should govern the appetitive part. Reason also sometimes plays other caregiving roles: “as the child should live according to the direction of his tutor, so the appetitive element should live according to reason”.34 Aristotle also clearly considers reason the best part of our nature, and thinks that it should therefore be the commanding force.35 This seems to counter the view of reason as nothing but an intermediary between character-established desire and action. For if reason cannot determine goals, but is rather constrained by them, reason seems to be obeying desire, not the other way around.
Moreover, the worry about what the Ethics and political science in general are for remains. Aristotle’s ethical project is undoubtedly a rational study of the human telos, and he is quite explicit that we study ethics “not so that we may know what virtue is, but so that we may become good” (EN.II.2.1103b27-8; cf. EN.X.9; EE.I.5.1216b16-25). How can we become good by studying ethics, if reasoning cannot contribute to shaping our goals?
These are both important concerns.36 Given the textual evidence mentioned in this subsection, Aristotle clearly holds something like the following:
(PR) The Priority of Reason
The human soul’s rational part must guide the non-rational part, and the non-rational part must in turn listen to reason. Reason should prescribe, character should follow.
Moss’ practical empiricism is unable to accommodate PR. The reason is not that PE1 is too Humean, taking too much away from reason (PE1 merely restates DL). The problem is rather that PE2 is not sufficiently non-Humean, not intellectual enough.
To avoid conflating Aristotle and Hume, Moss claims that Aristotle would have phantasia rule over the passions, instead of the passions rule over cognition. That much is certainly correct: the cognition of an object as good or bad (via perception or phantasia) sets the faculty of desire in motion (§1.1). However, this non-rational cognitivism is inconsistent with PR, which states that reason (not non-rational cognition) prescribes over desire.
Against PR, Moss may contend that practical reason’s activities are effectively constrained by the contents of the agent’s phantasiai, which are in turn determined by her prior habituation and perceptual experiences. This has to be accepted as well, because its rejection over-intellectualizes the starting points of human action (§2.2.2). But we must also be careful not to under-intellectualize Aristotle’s view: for if, as Moss says, reason is fully “dependent on non-rational cognition, both in genesis and operation”, it follows that character prescribes, and reason must follow character’s prescriptions. Thus, reason becomes a slave, not to the passions, but to non-rational cognition. This practical-empiricist position, while more cognitivist than radical Humeanism, is still not intellectual enough for PR. For, according to Moss (2012, 140),
practical intellect does quite a bit less […] than is widely thought. We are indeed able, in virtue of our rationality, to have cognitions about the good, and corresponding motivations, that are unavailable to lower creatures. But this is not so much a matter of grasping distinctive content as a matter of manipulating and conceptualizing content in sophisticated ways. The content of our thoughts about the good […] is determined by non-rational cognition: by evaluative perception and phantasia, the cognitive capacities of the non-rational soul.
There is much truth in this. But it cannot be the whole truth, since Aristotle still considers reason the most important, and therefore ruling, element. How can it be the ruling element, if it works only within the confines that character has set for it?
This highlights the ultimate problem for interpreting Aristotle’s agency theory: we must make the view that only character establishes the goals (DL) compatible with the claim that reason must prescribe and character must follow (PR). Intellectualism fails by disregarding DL, and antiintellectualism fails by disregarding PR, despite solid textual evidence for both claims. How are we to make progress in this situation?
4. Conclusion: Toward indirect intellectualism
If the two traditional interpretive approaches are flawed, it is time to try something different. The key intuition for a new approach is that reason cannot set the goals of action directly (that is character’s role), but it may be able to set goals indirectly, by shaping character. This is indirect intellectualism, a view that seeks to preserve the virtues of both intellectualism and anti-intellectualism while avoiding their pitfalls. Indirect intellectualism has two main claims:
[I.I.1] The care of one’s self: Action as practice
Reason can set the goals of action indirectly, by carefully shaping character through the practice of appropriate actions.
[I.I.2] The care of others as a precondition for individual self-care
An individual’s reason can set the goals of action indirectly only in a political context in which individuals receive proper care for the development of their character from early childhood.
I.I.1 makes DL and PR consistent: since-as Aristotle often claims-a certain kind of action generates the corresponding kind of character disposition, agents can rationally establish their goals indirectly, by treating each action as practice toward their desired character. If reason succeeds in sculpting character, then it will succeed in (indirectly) determining the agent’s goals.
But I.I.1 cannot be the whole answer, for it presupposes two things: first, that the given individual already has the desire to attain a certain character. And second, that this individual already is the kind of person who can design and execute a rational plan of long-term, effortful character construction; but this is possible only for persons who are already rational, i.e. who do not allow their immediate affections to deviate them from their rational plan. In short, I.I.1 presupposes that people are already able to care for themselves. But, as mentioned above, Aristotle claims most people fail to satisfy these presuppositions, unable to properly care for their own characters, simply living in accordance with their immediate passions (EN.X.9.1179b23-31). So how do people get to that point (i.e. the point of having the desire to care for themselves, and of being able to devise and execute a long-term rational plan)?
This is where I.I.2 comes in. Reason can take control of an agent’s goals as long as the agent herself has been properly habituated “straight from birth”.37 Such early habituation initiates the agent into a way of life that allows her later to take habituation into her own hands, by rationally developing the character that will make her pursue certain goals and avoid others. This is how reason can set goals: not directly, by means of reasoning (this is impossible given DL), but indirectly, via habituation and careful character formation (which secures PR without denying DL).
For reasons of space, the details of indirect intellectualism must be developed elsewhere.38 My goal here has been to show that the traditional interpretative strategies fail to consistently accommodate both DL and PR. This discussion opens the path toward a novel interpretation of the interplay between character and reason in Aristotle’s theory of human agency.39
