1. Introduction
How does one become a better person? Kant answers that one can always freely choose to subordinate one’s self-interest to the demands of morality, thereby acquiring a good intelligible character. However, Kant also holds that one is part of a world of experience, where one has an empirical character, and where one’s actions are all causally determined in accordance with necessary laws of nature. Kant’s account of the compatibility of these characters invokes what I call Empirical Character Grounding (ECG): The empirical character is grounded in one’s freely chosen intelligible character. This paper will propose a novel account of what empirical character is and how it is grounded by intelligible choice.
According to an increasingly popular interpretation of empirical character, which I will call the Causal Laws reading, the empirical character is the set of causal laws governing one’s empirical psychology and behavior.Footnote 1 The Causal Laws reading explains ECG and the compatibility of the intelligible and the empirical character as follows: All one’s mental actions are determined in accordance with laws of nature, and thus in accordance with the empirical character (since the empirical character simply is the particular laws of nature governing one’s psychology). But these laws of nature themselves are in some sense freely chosen, grounded in one’s choice of intelligible character. One freely chooses (as a thing in itself) the psychological laws by which one is (as an object of experience) bound.
This article will present a challenge to the Causal Laws reading based on Kant’s account of moral change. At the intelligible level, becoming a better person can, according to Kant, only be a momentous change of intelligible character from evil to good, a ‘revolution (…) in the mode of thought [Denkungsart]’ (Rel, 6:47).Footnote 2 To this revolution corresponds, on the empirical level, a ‘gradual reformation in the mode of sense [Sinnesart]’ (Rel, 6:47). This implies a gradually changing and improving empirical character over time, and hence what I call Empirical Character Changeability (ECC). However, a strict Causal Laws reading is incompatible with ECC: Laws of nature cannot change over time.
I propose instead an alternative Causal Powers reading of empirical character: One’s empirical character is not just a set of psychological and behavioral causal laws, but also the powers [Kräfte] of the mind over which the laws hold. Unlike laws, these powers can change over time, for instance by being weakened or strengthened. Kant famously uses this point in the KrV to argue against the substantiality of the soul (B413-18). Interpreters have hitherto overlooked that this setback for traditional rational psychology has an important positive upshot within Kant’s system: Kant’s theoretical account of changeable empirical mental powers accommodates ECC, and hence the empirical expression of moral improvement. Moral cultivation, the ‘gradual reformation’ towards virtue, takes place through the intelligible character’s influence on one’s mental powers. I do not claim to show that Kant himself explicitly advocated the Causal Powers view. However, I take it to be a promising candidate for interpreting Kant’s sparse remarks concerning empirical character: It is consistent with and supported by his texts, avoids the objections raised against the Causal Laws reading, and can be defended against objections based on Kant’s empirical determinism and his account of the empirical subject.
2. Empirical character as causal laws
The Causal Laws reading is partly motivated by some serious, well-known objections (which I will not rehearse here) raised against the common interpretation on which the intelligible character directly grounds our empirical choices as causal events (we may call this the Causal Events reading).Footnote 3 The Causal Laws conception of empirical character arguably avoids these objections, while also boasting significant textual support. Kant explicitly refers to a thing’s character as the ‘law of its causality’:
[E]very effective cause must have a character, i.e. a law of its causality, without which it would not be a cause at all. And then for a subject of the world of sense we would have first an empirical character, through which its actions, as appearances, would stand through and through in connection with other appearances in accordance with constant natural laws, from which, as their conditions, they could be derived. (A539/B567, italics mine)
Later, Kant again speaks of the ‘empirical character, i.e. the law of its causality’ (A540/B568). When introducing the ECG hypothesis, i.e. the grounding of the empirical character in a free, intelligible choice, Kant similarly refers to the laws of empirical causality:
Is it not (…) possible that although for every effect in appearance there is required a connection with its cause in accordance with laws of empirical causality, this empirical causality itself (…) could nevertheless be an effect of a causality that is not empirical, but rather intelligible[?] (A544/B572; cf. Prol, 4:346; GMS, 4:453).
The Causal Law conception understands this as follows: The empirical character of a subject is constituted by the causal laws determining its mental states and actions. These ‘laws of empirical causality’ themselves – the psychological laws of nature governing my mental life – are products of my free choice, grounded in my intelligible character.
ECG may be interpreted metaphysically, or in a more deflationary manner.Footnote 4 In any case, ECG understood in terms of the Causal Laws conception chimes well with the overall tenor of Kant’s transcendental idealism, where the subject’s spontaneity is in some sense the author of laws of nature: ‘Categories are concepts that prescribe laws a priori to appearances, thus to nature as the sum total of all appearances’ (B163; cf. Prol, 4: 319–320). Admittedly, categories prescribe only the laws ‘on which rests a nature in general, as lawfulness of appearances in space and time’ (B165), and hence no particular laws of nature. However, once objects are understood as appearances grounded in the subject’s forms of sensibility and thinking, the possibility that some particular laws (the laws that constitute one’s empirical character) might further be grounded in the subject’s free choice no longer seems out of the question.Footnote 5
3. The causal laws reading and the problem of moral change
Most theories of moral psychology allow for the possibility of moral change: the change of a person’s moral character over time. People are not irredeemably evil, nor incorruptibly good, and circumstances, experiences, as well as personal realizations and conversions can make a difference to our moral personality over time. Assuming that such moral change, for Kant, involves a change in the empirical character, we get:
ECC (Empirical Character Changeability): Empirical character can change over time.
At least some proponents of the Causal Laws conception of empirical character appear committed to a strict version, where the empirical character is nothing but a set of laws of nature governing one’s psychology.Footnote 6 This gives rise to the following inference:
(1) Empirical character is a set of causal laws of nature.
(2) A set of causal laws of nature cannot change over time.
(3) (From (1), (2)) Empirical character cannot change over time.
The inference is valid, and the conclusion 3 evidently contradicts ECC. Premise 1 is just the strict Causal Laws conception of empirical character. Hence the only contentious premise is 2: However, that laws cannot change seems implied by Kant’s references to the ‘invariable [unwandelbaren] natural laws’ (A536/B564, cf. A539/B567, A798/B826; Prol, 4:295). Could it instead be that the set of laws changes, i.e. that what changes over time is which laws are ‘picked out’ by the intelligible character to be the laws of our psychology? This seems highly implausible, given that the laws of nature, as rules, are both necessary and ‘constant [beständige]’ (A539/B567; cf. A113, A766/B794; Prol, 4:295, 343, 345–346). Watkins’ prominent treatment of Kant’s conception of causality accordingly states: ‘Whatever grounds and causal laws have held in the past will not change in the future’ (Watkins Reference Watkins2005, 290).
Proponents of the strict Causal Laws reading may embrace the conclusion 3 and reject ECC. This option has two different inflections: The more radical view rejects the reality of moral change of character tout court. Schopenhauer, inspired by Kant’s theory of freedom and character, takes this path and argues that ‘character is unalterable’ (Schopenhauer Reference Schopenhauer, Zöller and Payne[1839] 1999, 45). Jacquette summarizes:
The unalterability of the character of the willing subject deprives the moral agent of a certain type of freedom: the freedom to change. If Schopenhauer is right, then we can never become something different, change from being virtuous to vicious, or the reverse, or reform ourselves through force of will and dedication of effort to alter the kinds of people we are. (Jacquette Reference Jacquette2005, 187)
Given Kant’s commitment to the possibility of a ‘revolution (…) in the mode of thought [Denkungsart]’ (Rel, 6:47) and progress in the mode of sense, this is unlikely to be Kant’s view.
More exegetically plausible is a second, less radical rejection of ECC: moral change is possible, but does not require change in one’s empirical character. At least two different approaches are possible: The fundamental, invariable laws constituting one’s empirical character may be (1) sufficiently fecund, or (2) sufficiently numerous and variegated, to allow for satisfactory expressions of moral change. McCarty opts for (1), drawing an analogy to water and its phase changes (cf. McCarty Reference McCarty2008, 445–446): The invariable laws governing the behavior of water allows for ‘change of character’ in a less fundamental sense – like change from liquid water to ice or water vapor. Hanna instead suggests a version of (2), where ‘transcendentally free rational animal choices produce natural causal singularities, and one-time laws’ (in Hanna and Moore Reference Hanna and Moore2007, 121; cf. Vilhauer Reference Vilhauer2010).
I believe that each of these approaches has problematic consequences, but cannot pursue detailed arguments here.Footnote 7 The main concern I want to highlight is textual: Maintaining a strict Causal Laws conception, where empirical character is nothing but a set of causal laws of nature, means locating change outside the empirical character in sensu strictu.Footnote 8 Textual evidence indicates, however, that change in empirical character is possible: The ‘gradual reformation in the mode of sense [Sinnesart]’ (Rel, 6:47) concerns virtue ‘in its empirical character (virtus phaenomenon), ’ something that can be ‘acquired little by little (…), in virtue of which a human being, through gradual reformation of conduct and consolidation of his maxims, passes from a propensity to vice to its opposite’ (Rel, 6:47). KrV indicates that ‘mode of sense’ and ‘empirical character’ are equivalent terms, referring to ‘the mode of sense (the empirical character)’ (A551/B579), and, conversely, reason’s ‘empirical character (in the mode of sense [der Sinnesart])’ (A551/B579).
Many commentators emphasize the possibility of gradually cultivating a virtuous character.Footnote 9 One’s intelligible character enacts a gradual influence on the empirical one: ‘Reason gradually draws sensibility into habitus, arouses incentives, and hence forms [bildet] a character, which however is itself to be attributed to freedom’ (R5611, 18:252); the Religion similarly speaks of ‘gradual reformation’ and ‘gradual influence (…) on the mind’ (Rel, 6:83). Central to this process is the goal of attaining two things: ‘[B]eing one’s own master in a given case (animus sui compos), and ruling oneself (imperium in semetipsum), that is, subduing one’s affects and governing one’s passions’ (MS, 6:407). Kant continues: ‘In these two states one’s character [Gemüthsart] (indoles) is noble (erecta); in the opposite case it is mean (indoles abjecta, serva)’ (MS, 6:407). One has a duty to gradually change one’s Gemüthsart towards virtue.
These points suggest that a non-strict Causal Laws reading fits better textually: the gradually changing aspect of one’s empirical character cannot be the causal laws. Coupling it with the Causal Events reading is possible.Footnote 10 However, the next section proposes instead looking to Kant’s account of causal powers, giving rise to what I call a Causal Powers reading of empirical character.
4. The causal powers reading of empirical character
The core of my proposal is to conceive of our empirical character not just as the psychological laws of causality, but also as the set of causal powers of the mind. While some recent literature implicitly suggests this Causal Powers reading, it has not been explicitly put forward or defended as a distinct alternative.Footnote 11 The reading has textual support and significant philosophical benefits – including, crucially, its compatibility with ECC.
The Causal Powers reading can helpfully proceed from Watkins’ (Reference Watkins2005) re-evaluation of Kant’s metaphysics of causality. According to Watkins causal powers, understood as the ‘causality of the cause’ (Watkins Reference Watkins2005, 249; cf. e.g. A203/B248; VMe, 28:573), are essential to Kant’s account of causality. Objects causally influence other objects, and are influenced by them, in virtue of their powers.Footnote 12 The fundamental causal powers of outer, physical objects are moving powers of attraction and repulsion (cf. MAN, 4:496f.). Powers of the mind, according to Kant, are the proper subject matter of psychology: ‘Psychology, which explains what happens, and does not prescribe what ought to happen, concerns itself with mental powers [Gemüthskräften]’ (R5864, 18:371–372). KrV provides an extensive list of such powers, while also indicating that the list can be shortened by reducing some powers to other, more fundamental ones:
In the human mind there are sensation, consciousness, imagination, memory, wit, the power to distinguish, pleasure, desire, etc. (…) A logical maxim bids us to reduce this apparent variety as far as possible by discovering hidden identity through comparison, and seeing if imagination combined with consciousness may not be memory, wit, the power to distinguish, or perhaps even understanding and reason. (A649/B677)
Kant reportedly begins the empirical psychology part of one of his lectures by stating: ‘The powers of the human soul can be reduced to three, namely: (1) The faculty of cognition, (2) the feeling of pleasure and displeasure, and (3) the faculty of desire’ (VMe, 28:584; cf. EE, 20:206). Importantly, Kant holds that there is an essential and irreducible distinction to be made between the lower powers and the higher powers of the mind; the lower often subsumed under the umbrella of ‘sensibility’, the higher under the umbrella of ‘understanding’ or ‘reason’ (cf. Anth, 7:140–141; Päd, 9:472; see Frierson Reference Frierson2013, 50–51).
Further examination of Kant’s system of empirical powers of the mind can be found elsewhere (cf. Frierson Reference Frierson2014); I focus here on whether these empirical powers constitute one’s empirical character. If causal powers are understood as ‘the causality of the cause’ – an object is a cause qua exercising a causal power – the mind causes its representations and actions by exercising mental powers. Psychological laws, meanwhile, are necessary rules in accordance with which these powers operate. Returning to the passages discussing empirical character, we see Kant describing character not just as laws of causality, but also as the empirical causality itself: ‘[T]he human being himself is an appearance. His power of choice has an empirical character, which is the (empirical) cause of all his actions’ (A552/B580). Similarly, ‘[e]very human being has an empirical character for his power of choice, which is nothing other than a certain causality of his reason, insofar as in its effects in appearance this reason exhibits a rule’ (A549/B577). When introducing ECG at A544/B572, Kant’s hypothesis is that the ‘empirical causality itself’ might be ‘an effect of a causality that is not empirical, but rather intelligible’ (cf. also KpV, 5:114, 115; KU, 5:195n.). And after stating that ‘[the human being] must (…) have an empirical character, just like all other natural things, ’ Kant remarks that ‘[w]e notice it through powers and faculties which it expresses in its effects’ (A546/B574; cf. KpV, 5:98).
Importantly, Kant holds that the empirical powers of the mind can change over time.Footnote 13 The most conspicuous evidence for this is found in the B Paralogisms of KrV. Kant considers Mendelssohn’s via negativa proof of the soul’s substantiality, arguing that the soul, as simple, cannot coherently (within the bounds of experience) cease to exist. Kant criticizes Mendelssohn for overlooking that the soul could disappear through gradual remission of its powers:
[O]ne nevertheless cannot deny to [the soul], any more than to any other existence, an intensive magnitude, i.e. a degree of reality in regard to all its faculties, indeed to everything in general that constitutes its existence, which might diminish through all the infinitely many smaller degrees (…) [T]he supposed substance (the thing whose persistence has not otherwise been established already) could be transformed into nothing, although not by disintegration, but by a gradual remission (remissio) of all its powers (…). Thus the persistence of the soul, merely as an object of inner sense, remains unproved and even unprovable. (B414-5; cf. MAN, 4:542; VMe, 28:761, 763–764, 29:905–906, 912–913, 1037; R5650, 18:299)Footnote 14
Kant’s deeper reasons for holding that the object of inner sense must have powers with degrees that ‘might diminish’ (or, presumably, increase) are found in the earlier Anticipations of Perception part of KrV.Footnote 15 While other kinds of change may be possible, for the purposes of this article change of degree stand in for possibly more complex or different manners of change through cultivation of mental powers that, according to the Causal Powers reading, constitute the ‘gradual reformation in the mode of sense’ (Rel, 6:47).Footnote 16
Is change in powers possible without also changing the laws of nature? A negative answer is fatal to my proposal, since it resurrects the problems concerning ECC from the previous section. Obviously, the answer depends on the relation between powers and laws in Kant. A comprehensive investigation cannot be undertaken here, but let me offer some reasons why a positive answer – that causal powers can change without also changing the laws – is plausible:
We can distinguish, roughly, between ‘top-down’ and ‘bottom-up’ interpretations of Kantian laws; either laws are imposed on objects ‘from above, ’ or they arise from properties of the objects.Footnote 17 Top-down approaches where, for instance, the understanding prescribes laws that powers obey, seem unproblematic: change in powers may at most change which laws these powers are governed by, but not the laws themselves. However, bottom-up interpretations where laws of nature are grounded in causal powers are growing increasingly prominent (see e.g. Kreines Reference Kreines2017; Massimi Reference Massimi, Breitenbach and Massimi2017; Messina Reference Messina, Breitenbach and Massimi2017; Patton Reference Patton2017; Watkins Reference Watkins2005). Moreover, due to their greater emphasis on causal powers, these accounts appear generally more hospitable to the Causal Powers reading – hence the importance of investigating their compatibility with ECC.
Everything depends on a more precise specification of the grounding in question; how is the grounding relation to be understood, and which aspect(s) of the causal powers ground the causal laws? In the briefest of overviews, the following points seem crucial and relatively uncontroversial: First, the grounding in question is commonly understood as, at the least, entailing supervenience: laws of nature supervene on the causal powers of objects in nature (Ellis Reference Ellis2001, 1; Massimi Reference Massimi, Breitenbach and Massimi2017, 169; Messina Reference Messina, Breitenbach and Massimi2017, 137; Molnar Reference Molnar and Mumford2003, 199). Second, Kant interpretations hitherto proposed suggest that the natures, the essential or necessary aspects of causal powers, ground laws (Watkins Reference Watkins2005, 335; Messina Reference Messina, Breitenbach and Massimi2017, 137; Massimi Reference Massimi, Breitenbach and Massimi2017, 169; Kreines Reference Kreines2017, 326; Patton Reference Patton2017, 349f.).
These specifications are compatible with my proposal: First, supervenience relations only require that different supervenient properties (the laws) entails a different supervenience base (the powers) – not vice versa. Supervenience thus allows the supervenience base, i.e. the causal powers, to change without change in the supervenient properties, i.e. the laws of nature. Second, and more significantly, the particular quantity of the powers, i.e. their degree, need not be part of the essence or nature of objects of experience. This is especially plausible when it comes to mental powers: whereas Kant’s works in transcendental philosophy explicate essential aspects of our mental powers of sensibility, understanding, reason, and so on, the amount of different powers in different persons (and also in different genders, races, and temperaments) is among the features discussed in anthropological, psychological and pedagogical writings that focus on our contingent properties. The pragmatic aim of Kant’s anthropology (cf. Anth, 7:119) and pedagogy targets contingent aspects susceptible to development and improvement – including the mental powers of oneself or tutored children (see further below; for children, see Päd, 9:472f.).Footnote 18
The variable, contingent degrees of power may still figure in laws grounded in essential aspects of the causal powers. To see this, consider laws that are necessary functions from variables (causes) onto other variables (effects).Footnote 19 This is a conception of law that fits well the inverse-square and -cube laws discussed in MAN (4:519f.; see Friedman Reference Friedman2013, 221f.), for instance concerning the diffusion of light through space: The law gives the degree of illumination of a surface as a function of the total ‘light quantum’ (MAN, 4:519) (the degree of power of the light source) and the surface’s distance from this light source (where the resulting degree of illumination equals the total light quantum divided by the distance squared). Here, light quantum and distance are cause-variables, and the inverse-square law a necessary function from these variables to the degree of illumination as effect-variable.Footnote 20
Kant sees this as a typical law of outer nature: ‘[S]o, too, with all other powers, and the laws whereby they must diffuse, either on surfaces or on volumes, so as to act on distant objects in accordance with their nature’ (MAN, 4:519). The more general framework, of laws as functions relating variables, easily transposes even to the psychological realm: One could, for instance, envisage a law expressing a function from the degree of one’s power of attention (as cause-variable) to the degree of consciousness of one’s representations (as effect-variable).Footnote 21 On a bottom-up interpretation, these laws are grounded in essential properties (the essential properties of extension and light for the inverse-square law of illumination; the essential properties of the power of attention for the envisaged law of conscious representation). Since they relate variables, they nonetheless incorporate reference to contingently varying degrees of power.
Returning now to the gradual reformation of empirical character, there is textual evidence for understanding it as gradual change in one’s powers. Consider again the aspects of a virtuous empirical character mentioned above, of ‘[B]eing one’s own master (…) (animus sui compos), and ruling oneself (imperium in semetipsum)’ (MS, 6:407). Ethics lecture notes from 1793/94 explicate how to attain this more in detail, as aspects of ‘the duty of man to develop his powers quoad maxime’ (VMo, 27:625–626):
a. To possess oneself, i.e. to determine all actions by way of a free choice. This is what is called animi sui compos, or having a settled disposition. (…) [One] attains this only by subjecting all his powers and capacities solely to his free choice, and employing them accordingly. (VMo, 27:626)
b. The duty to govern oneself. This involves cultivation of the mental powers to those ends with which they are collectively compatible, and constitutes, therefore, the essential in the soul’s capacity or readiness to enlarge the facultates animi for all moral ends, and to direct them thereunto. (VMo, 27:627)
Other passages in Kant’s works confirm that attaining a virtuous empirical character involves changing and cultivating one’s mental powers (cf. e.g. MS, 6:386–387, 6:391f.; VMo, 27:360f.). The same lecture notes also support a Causal Powers reading of ECG: ‘In respect of its power, the sensory being is so far dependent on the noumenon, as intellectual being, that it is subordinated thereto’ (VMo, 27:603, my emphasis). Since this account also allows ECC, the intelligible character’s influence can effect a ‘gradual reformation in the mode of sense’ (Rel, 6:47): The gradual cultivation of one’s mental powers towards empirical virtue.
Plausibly, this includes both the lower and the higher powers: The higher powers by e.g. increasing the ‘strength of one’s resolution’ (MS, 6:390, cf. 6:384) acquiring the ‘moral strength of a human being’s will in fulfilling his duty’ (MS, 6:405), and combating the ‘weakness in the use of one’s understanding’ (MS, 6:408; cf. VMo, 27:293); the lower powers by e.g. strengthening the susceptibility to ‘moral feeling’ (KpV, 5:38, 5:76; cf. MS, 6:399–400) and the capacity for ‘subduing one’s affects’ (MS, 6:407).Footnote 22 More can clearly be said to flesh out Kant’s account of moral improvement along these lines.Footnote 23 My aim is establishing that cultivation of a virtuous empirical character can be understood as cultivation of mental powers.
Virtue as such cannot be identified with suitably strengthened empirical powers. In isolation, this is ‘mere mechanism of applying power [Kraftanwendung]’ (Anth, 7:147) and merely empirical virtue, ‘virtus phaenomenon’ (Rel, 6:47). This may be just habit and a ‘happy constitution (merito fortunae)’ (A551/B579n.; cf. Rel, 6:37–38) without moral worth. Full virtue requires an empirical character constituted from, not just in accordance with, morality. Hence, it demands a good will actively and continually striving to form a moral empirical character (cf. MS, 6:383). The locus of human moral worth is not the empirical result, but the intelligible activity that influences the empirical character in the right direction, for the right reason.
Phrasing Kant’s position like this suggests how the Causal Powers reading can make sense of ECG: our empirical character comprises the changeable empirical powers of the mind, which can be influenced by the corresponding second-order power of our intelligible character.Footnote 24 ECG should be understood not as a completed result, but as an ongoing activity. Conceiving the intelligible character as ‘transcendental cause’ (A546/B574) of the empirical character thereby also involves thinking (though not cognizing) its ‘causality of the cause, ’ as power: a free, second-order noumenal power conceived as capable of ‘gradual influence (…) on the mind’ (Rel, 6:83), i.e. gradual reformation of the empirical powers of the mind.Footnote 25 This conception also gives, I will now argue, a plausible Kantian account of weakness of will.
5. Causal powers and weakness of will
Many recent interpreters have investigated how Kant can accommodate weakness of will.Footnote 26 Several conclude that weakness of will requires an evil will at the intelligible level. However, Kant sometimes seems to think weakness of will in a stronger sense is possible, namely failing to act morally even if one has a good will:
This weakness in the use of one’s understanding coupled with the strength of one’s emotions is only a lack of virtue and, as it were, something childish and weak, which can indeed coexist with the best will. (MS, 6:408, my emphasis)
I incorporate the good (the law) into the maxim of my power of choice; but this good, which is an irresistible incentive objectively or ideally (in thesi), is subjectively (in hypothesi) the weaker (in comparison with inclination) whenever the maxim is to be followed. (Rel, 6:29)
These seem to be actions contrary to duty and hence ‘transgressions [Übertretung], ’ but distinct from the kind of ‘intentional transgression [that] has become a principle’ which ‘is properly called a vice [Laster]’ (MS, 6:390). But is it possible to transgress not because one’s will is evil but because it is weak? Kant interpreters typically balk at the idea that a good-willed person can fail to act on what she cognizes morality as demanding: Given Kant’s account of freedom – and specifically the ‘ought implies can’ principle – the commitment to morality inherent in a good will presupposes the capacity to do, in some sense, what morality demands. A good-willed person’s failure to choose morally could not properly be characterized as a ‘doing, ’ but must rather be something like a physical compulsion, a mere reflex outside one’s control. These compulsions may be ‘morally unfortunate, ’ when what they effect is not in accordance with morality, but they are not something for which we are directly morally responsible.Footnote 27The power framework, where these powers have a greater or lesser degree, is better poised to explain Kant’s invocations of ‘weakness.’ Indeed, Kant explicitly distinguishes the ‘can’ of capacity, which must be presupposed, from the ‘can’ of strength, which one must gradually acquire: ‘[W]hile the capacity (facultas) to overcome all opposing sensible impulses can and must be simply presupposed in man on account of his freedom, yet this capacity as strength (robur) is something he must acquire’ (MS, 6:397). ‘Ought implies can’ entails the capacity, but not necessarily the strength, to overcome sensible impulses; rather, it means only that this strength can be acquired. The Causal Powers reading allows for weakness on two distinct levels, both of which help explain Kantian weakness of will:
First, weakness of will may manifest itself in the first-order, empirical powers. Whether or not one has a good will, these powers will have a finite degree and may therefore be ‘weak’ in relation to the specific obstacles or temptations they face. How can one be responsible for such weakness? Given that we have a duty to gradually reform, we may be culpable due to prior failings to sufficiently cultivate the mental powers. Ultimately, this may trace back to Kant’s doctrine of radical evil: all natural human beings ‘begin’ from an evil will, which is innate but must nonetheless be understood as freely chosen by each individual. If the insufficiency of one’s empirical powers to execute the moral action stems from a prior, freely chosen evil will, the weak-willed action is as imputable as the evil will it ultimately originates or is grounded in. There may even be a sense in which the finitude of our empirical powers, and hence the inextirpable possibility of weakness through being overpowered by sensible inclinations, is conclusive evidence for the radical evil of all humans within possible experience. Trying, yet failing, to do the right thing now, is not sufficient to absolve one of guilt: One remains culpable for lack of prior training and reformation. Virtue must be ‘exercised [geübt] and cultivated by efforts to combat the inner enemy within the human being (asceticism); for one cannot straightaway do all that one wants to do, without having first tried out and exercised one’s powers’ (MS, 6:477). Of course, the silver lining is that by continually striving for virtue, one can justifiably hope to do better next time around – ECC ensures that one is not stuck with permanent deficiencies of empirical character, that strength can be gradually acquired.
Second, weakness could conceivably arise on the second-order, noumenal level: A good will may try to do the right thing, but not try all that hard. Admittedly, this idea has a speculative air, as it introduces degrees of strength into the intelligible character’s noumenal power (although see Rel, 6:71 for apparent textual support). However, while we lack any ability to cognize the degree of intelligible strength, the thought of such a strength might be required: How else do we explain the fact that a good will must gradually reform the empirical character, rather than instantly (or almost instantly) change our empirical character into something approximating perfect virtue? If even a good will ‘cannot straightaway do all that [it] wants to do’ (MS, 6:477), a reasonable explanation is that its ability to do so is somehow limited in degree.
By bringing strength into the equation, on the side of both the intelligible and the empirical character, the Causal Powers reading provides a framework for explaining weakness of will. Moreover, it does so in a way that maintains responsibility, insofar as the strength is not something permanently fixed, but something that can – and ought – to be gradually acquired.
6. Substantiality and the powers of the mind
I now consider objections to the Causal Powers reading. One worry concerns its consistency with Kant’s theoretical account of the mind: If the empirical character of the mind consists in its mental powers, does this not imply that the empirical mind is a substance possessing these powers? This seemingly contradicts Kant’s criticisms of rational psychology in the Paralogisms of the KrV, centering precisely on the impossibility of showing the soul’s substantiality.
This is a complex issue. Kant doubtlessly sees an intimate relation between substance and power. Arguably, an analytical entailment holds between these two concepts: Something conceived as having power must also be conceived as being or being composed of substance(s), and vice versa – powers, by definition, are had by substances. However, an analytical entailment of this kind means only that cognizing the empirical powers of the mind entails thinking (but not cognizing) the mind as a substance, which the Paralogisms allow.Footnote 28
More worrisome is the apparent synthetic connection Kant draws between substantiality and power, most prominently in the Analogies of Experience of the KrV. Kant seems to argue that empirical substances must underlie all empirical powers and all alteration (cf. e.g. A188/B231, A204/B250). One option is to concede this point and argue that despite the Paralogisms, Kant can allow for a substantial soul within empirical psychology (cf. Frierson Reference Frierson2014, ch. 1). A different approach appeals to suggestive evidence that the synthetic connection between substance and power holds only for outer appearances. Kant repeatedly indicates that the First Analogy of substantiality is valid only for outer appearances,Footnote 29 and states that when it comes to empirical cognition of ourselves: ‘Instead of the word ‘soul, ’ we have taken to using that of living power (and rightly so, since from an effect we can certainly infer to the power that produces it, but not forthwith to a substance specially adapted to this kind of effect)’ (FP, 8:413). Kant continues, post-Paralogisms, to refer to powers of the mind in his empirical psychology and anthropology, and insists that we ‘cannot become aware of (…) inner powers in any thing except our soul. For we cannot perceive them through outer, but rather only through inner sense’ (VMe, 29:929). Hence a viable case might be made for the possibility of cognizing powers of the mind even without being able to cognize its substantiality.Footnote 30
Indeed, Kant’s argument against Mendelssohn suggests an interesting twist: The alterability of mental powers precludes Mendelssohn’s theoretical proof of substantiality, hence aiding the downfall of rational psychology’s theoretical aspirations. But this alterability simultaneously proves essential to Kant’s practical account of the human being, allowing for the possibility of cultivation on the basis of free choice, hence of moral progress (or regress). Kant’s theoretical account of mental powers denies substantiality in order to make room for Bildung.
7. Powers and determinism
A second worry concerns the compatibility of the Causal Powers reading with a different aspect of Kant’s theoretical framework: Is the intelligible character’s influence on the empirical character compatible with Kant’s empirical determinism? I have emphasized the changeability of the empirical powers of the mind, but do these changes have sufficient empirical causes? If so, the Causal Powers reading encounters its own ‘scope problem, ’ since to ground a change in power, the intelligible character must ground the empirical causes of this change, the causes of these causes, and so on back indefinitely through the temporal causal series.
One could bite the bullet, accept that changes in powers must have sufficient empirical causes, and tackle the scope problem – perhaps along lines already proposed in the literature. However, a more interesting and promising, albeit radical, option is to deny the premise: Changes in our empirical powers need not have sufficient empirical causes. How does this not fly in the face of Kant’s empirical determinism? The answer parallels one given by the Causal Laws reading: Kant’s empirical determinism holds between events (Begebenheiten) (cf. e.g. Prol, 4:345f.; KpV, 5:114) or accidents. Causal laws are not events – and neither are causal powers. Kantian powers are neither accidents nor substances (cf. VMe, 28:431, 29:771; for discussion, see Watkins Reference Watkins2005, 259f.). Hence, increase or decrease in their strength are not events in Kant’s sense, and need not be deterministically caused.
On the Causal Laws reading, all events are causally determined by preceding empirical events given the laws of nature; however, the laws of nature themselves are not causally determined by anything empirical. The Causal Powers reading extends this point: All events are causally determined by preceding empirical events given laws of nature and relevant causal powers; but the laws of nature themselves and the nature and strength of the causal powers are not (necessarily) causally determined by anything empirical.Footnote 31 To exemplify: Take a certain event, say, object A striking object B. The effects of this events – change of motion in A and B, etc. – are determined by the causes, i.e. the prior motions and locations of A and B. The effects necessarily follow from the causes given the causal laws that govern the interactions between physical bodies. But on the causal powers framework, the effects only follow with necessity given the causal laws and the causal powers of A and B. The same prior motions and locations, and the same laws of nature, give different effects if A and B have different causal powers, e.g. stronger or weaker repulsive powers. And the existence and strength of these powers themselves need not have sufficient empirical causes.
This allows Kant’s determinism to be compatible with the claim that ‘this empirical causality itself (…) could nevertheless be an effect of a causality that is not empirical, but rather intelligible’ (A544/B572; cf. KpV, 5:144): Causal powers as the ‘empirical causality itself’ can be effects of intelligible causality, and their gradual reformation the effect of intelligible influence.Footnote 32 While several interpreters argue that Kant’s empirical determinism does not, or at least should not, hold when it comes to the mental realm,Footnote 33 my account respects the plentiful textual evidence that Kant’s determinism applies to both inner and outer experience.Footnote 34 Instead, the ‘elbow room’ for freedom resides on a different metaphysical level in nature – changing states (both inner and outer) are deterministically governed, laws of nature are unchanging, but the causally efficacious natural powers may change in ways that are not themselves empirically necessitated (through the influence of our freedom).
Now, there are passages that appear problematic for this denial of empirical necessitation with respect to powers. Kant suggest that to explain an immoral action (a malicious lie) you can go
into the sources of the person’s empirical character, seeking them in bad upbringing, bad company, and also finding them in the wickedness of a natural temper insensitive to shame, partly in carelessness and thoughtlessness; in so doing one does not leave out of account the occasional causes. Now even if one believes the action to be determined by these causes, one still blames the agent. (A554-5/B582-3)
Kant proceeds to point out that we still, despite the empirical determination, blame the person when ‘the action is ascribed to the agent’s intelligible character’ (A555/B583). Kant may be read here as saying that the empirical character is necessitated by empirical causes such as ‘bad upbringing.’ However, this reading is in fact compatible with the Causal Powers reading: Even if the empirical character, in this particular case, is fully determined by empirical causes, Kant’s point may be that it ought not to have been. The intelligible character ought to have influenced the empirical character so as to counteract the unfortunate empirical influences. In other words, the empirical character has sufficient empirical causes because the intelligible character has not freely exercised a (moral) influence.Footnote 35 Just after this passage, Kant reiterates the potential influence of the intelligible character: ‘Another intelligible character would have given another empirical one’ (A556/B584).Footnote 36
Another passage poses a different variation of the same challenge:
If it were possible for us to have such deep insight into a human being’s cast of mind [Denkungsart], as shown by inner as well as outer actions, that we would know every incentive to action, even the smallest, as well as all the external occasions affecting them, we could calculate a human being’s conduct for the future with as much certainty as a lunar or solar eclipse and could nevertheless maintain that the human being’s conduct is free. (KpV, 5:99)
One could read Kant here as saying that complete empirical knowledge of a human being at a specific point in time would allow us to ‘calculate [that] human being’s conduct for the future’.Footnote 37 The Causal Powers reading cannot accept this: Complete empirical knowledge of a human being at a certain point in time would not afford complete knowledge of whether and how the mental powers might change through intelligible influence, and hence not complete knowledge of future conduct (as determined by the empirical powers).
However, Kant does not say expressly that future conduct can be calculated by knowing a human being’s cast of mind at a certain point in time. Rather, he simply states that this would be possible given sufficiently ‘deep insight’.Footnote 38 Of course, if we knew the strength of the powers of the mind now as well as in the future, we could calculate the conduct and ‘nevertheless maintain that the human being’s conduct is free’.Footnote 39 The Causal Powers reading can admit that if (perhaps per impossibile) we had that deep insight, future conduct could be predicted.
Is it not at least implicitly obvious that Kant means insight solely at a specific point in time? I do not think so: Even with complete empirical knowledge of a human being at a certain point in time, we cannot calculate future conduct if we only know the ‘external occasions’ at that point in time (unless, implausibly, complete knowledge of the external occasions affecting me at a specific point in time enables calculation of the future progression of the entire external world). My psychological state two hours from now will vary depending on e.g. whether I then look up at a clear sky, or a sky full of descending hail. Kant’s main point seems to be that even with complete knowledge of empirical conditions and thus complete power of prediction, we could still justifiably maintain freedom (because it could still be grounded in freedom). But the passage need neither be read as implying that complete knowledge of present empirical conditions gives complete power of prediction, nor as saying that it really is possible, even in principle, to know at a given point in time the future development of one’s empirical powers.Footnote 40
Evidently, the Causal Powers interpretation I propose requires a significant re-evaluation of the nature and limits of Kant’s empirical determinism. I do not think that it can be decisively shown that Kant exempted change in powers from needing sufficient empirical causes, though I hope to have made a preliminary case that it is compatible with his explicit pronouncements concerning determinism and causality. I submit, further, that it offers an attractive way of accommodating the letter of Kant’s determinism concerning empirical events, while leaving room for freedom to have a tangible and gradual influence on our empirical conduct, through its influence on the mental powers constituting our empirical character.
8. Conclusion
I have argued that our moral agency is exercised in the empirical world through influencing our empirical powers. Kant holds that our empirical character is grounded in the intelligible character, and I interpret our empirical character as constituted by our empirical mental powers. I have defended this Causal Powers reading of empirical character from several objections, and pointed to significant benefits gained by adopting it. Chief among these benefits is providing a robust and plausibly Kantian account of moral change, i.e. of the gradual progress involved in becoming a better person in the empirical world.
The Causal Powers reading could be developed into a general account of rational agency in Kant. Epistemic agency may also be understood in terms of influencing one’s powers, e.g. increasing and directing one’s power of attention and reflection to notice the grounds of one’s judgments. This interpretation would be well-placed to understand Kant’s account of error as the ‘unnoticed influence of sensibility on the understanding’ (A294/B351), and the normative demand to counteract such error. The Causal Powers reading thus provides a promising general framework for understanding Kant’s account of the relation between normativity and psychology: Psychology concerns itself with mental powers, and normativity is possible because freedom can influence these powers. The present article, however, has focused on showing how the Causal Powers reading accounts for empirical character and the possibility of moral progress.
References to Kant’s works give Kant (Reference Kant1900–) volume and page, except the Kritik der reinen Vernunft where I use the standard A/B edition pagination. Translations are from Kant (Reference Kant, Guyer and Wood1992–), or, where unavailable, my own.
Abbreviations
- Anth
-
Anthropologie in pragmatischer Hinsicht
- EE
-
Erste Einleitung in die ‚Kritik der Urteilskraft’
- FP
-
Verkündigung des nahen Abschlusses eines Tractats zum ewigen Frieden in der Philosophie
- GMS
-
Grundlegung zur Metaphysik der Sitten
- KpV
-
Kritik der praktischen Vernunft
- KrV
-
Kritik der reinen Vernunft
- KU
-
Kritik der Urteilskraft
- MAN
-
Metaphysische Anfangsgründe der Naturwissenschaft
- VMe
-
Vorlesungen über Metaphysik
- MS
-
Metaphysik der Sitten
- VMo
-
Vorlesungen über Moralphilosophie
- N
-
Nachträge zur ‚Kritik der reinen Vernunft’ (1. Auflage)
- Päd
-
Über Pädagogik
- Prol
-
Prolegomena zu einer jeden künftigen Metaphysik, die als Wissenschaft wird auftreten können
- R
-
Reflexionen
- Rel
-
Religion innerhalb der Grenzen der bloßen Vernunft
- SF
-
Der Streit der Fakultäten
Funding
This work was supported by Sun Yat-sen University.
Disclosure statement
No potential conflict of interest was reported by the author.
Acknowledgements
For invaluable discussion and/or comments on previous drafts, I am grateful to Toni Kannisto, Jacob Lautrup Kristensen, Michael Oberst, Camilla Serck-Hanssen, Feroz Mehmood Shah, and Eric Watkins. I also thank audiences in Oslo, Bergen, and Leuven and previous anonymous referees for feedback on the ideas presented in this article in various stages of development, and two anonymous referees for this journal, whose incisive and constructive comments greatly improved the end result.
