The 'Ought' and the 'Can'

CON-TEXTOS KANTIANOS.

International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347

ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006


The 'Ought' and the 'Can'1


KATERINA DELIGIORGI


University of Sussex, UK



Abstract


Kant’s ethics is objectivist. Like other objectivist ethics, it faces the problem of showing how what is objectively morally demanded hooks onto the moral deliberations of particular individuals. The issue is particularly acute for Kantian ethics given the centrality of the concept of autonomy, which expresses a demand for rational self-legislation. The paper focuses on the ‘ought implies can principle’ (OIC) and its role in Kant’s ethics. The argument shows how understanding the Kantian use of OIC helps also with the problem of establishing a link between individual deliberation and objective moral demands.


Keywords


Kant, autonomy, freedom, morality


Kant's conception of autonomy presents the following problem. If, following Kant's explicit lead, we consider autonomy as the universal principle of morality and ground of the actions of rational beings (e.g. G 4:452), then self-legislation is best understood as a prescription by reason to itself. Applied to individual cases of willing, the term 'autonomy' describes the bringing of a set of practical attitudes under rational legislation. Agents may count as autonomous then, insofar as and only to the extent that they are able to implement reason's prescription. This is the bare Kantian picture. The problem, as Schiller originally put it, is that this is also a picture of self-alienation, since parts of one's identity, feelings, emotions, and attachments, are kept at arm's length and treated with suspicion (e.g. AW XXb: 280).2 Schiller's point is that there must be something that makes autonomy different


1 Versions of the paper were given in seminars at Bonn and at York. I am thankful to the participants in these seminars and to the editor of this Dossier, Joe Saunders, for their comments.

2 Schiller's own ideal is of reason and nature being ‘in harmony with one another and man is at one with himself’ (AW XXb: 280). Schiller is of course raising here a normative ethics problem, and this is how his


[Recibido: 10 de septiembre 2018

Aceptado: 28 de septiembre 2018]

323


Katerina Deligiorgi


from mere rationomy. For Schiller this matters because he thinks that a rationalist prescriptive ethic is deeply unattractive and because, anticipating contemporary theories of personal autonomy, he wants to defend an integrative conception of autonomous agency. No such further commitments are needed, however, to see that the bare picture needs adding to it, to show how the principle of reason's self-legislation not only has a grip on individual agents, but also can express their autonomy.


Kant interpreters who are sensitive to this problem seek to show that an internal link can be established between the individuating aspects of agency and reason's own law. Understood in this way, the challenge is to show that, provided individuals reflect deeply and seriously enough on things they value in their daily lives, on their practical commitments and identities, they will be led to the universal principle of morality and recognize it as their own.3 My aim in this essay is to take a different approach to this issue and show how autonomy relates to the personal agential point of view by examining the meaning and scope of Kant's references to the 'ought implies can' (OIC) principle.


Although OIC appears unconnected to the topic of autonomy, I will argue that it is central to understanding how a standard that holds for all rational beings can also be expressive of the autonomy of individual agents. Here is a key reference for the discussion that follows:


Now this "ought" expresses a possible action, the ground of which is nothing other than a mere concept, whereas the ground of a merely natural action must always be an appearance. Now of course the action must be possible under natural conditions if the ought is directed to it; but these natural conditions do not concern the determination of the power of choice itself, but only its effect and result in appearance. (A 548/B 576).


The claim I want to examine is the conditional Kant presents here and which I have underlined in the quoted passage, namely that if the ought is directed to it, the action must be possible. I will argue that Kant endorses a strong version of OIC that sets out a relation between freedom and normativity on which rests the moral conception of autonomy. Section 1 is on the 'ought', section 2 on 'can', section 3 on modality and the relation between a possible action and an action that must be possible. In the final section, I show how the earlier discussion of OIC is useful in addressing Schiller's point about autonomy.


criticism has been received. However, he is also raising an important structural problem and it is the latter that interests me here. The problem is not with reason's formulating its own law; this is just an extension of the reflexive character of reason, it is continuous with the idea that reason in its theoretical application examines its claims to knowledge and sets its limits. The problem arises when it comes to individual agents and it is a demand for explanation of how a law that is for all rational beings has a grip on individual agents and expresses their autonomy. Not everyone thinks there is a problem here, see O'Neill 2003, for example, who argues that the 'ought' has a grip on individuals qua rational and that it has nothing to do with individual agents claims to their autonomy.

3 Perhaps the most sustained such project is undertaken by Christine Korsgaard 1996, 1998, 2002.

CON-TEXTOS KANTIANOS

324 International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347

ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

The 'Ought' and the 'Can'


  1. Ought: authority and normativity.


    The interpretative question about the nature of Kant's commitment to OIC arises because the principle appears to advise that we fit our moral standards to what we can do. While Kant often states that 'duty commands nothing but what we can do' (R 6:47, see too KprV 5:143, MS 6:380, R 6:47), it is not his intention to invite us to adjust the 'ought' to the 'can' of human capacities as in standard invocations of OIC in moral theory.4 When Kant writes that the human being 'must judge that he can do what the law tells him unconditionally that he ought to do' (MS 6:380), he seeks to assert that what is a priori valid is a posteriori commanding and action-guiding.5 As Bob Stern explains, the 'ought' is 'something that commands us', it 'tells us what to do unconditionally', it has 'authority over us' (Stern 2004: 56).


    Acknowledging the different direction in which Kant wants to take OIC, however, creates its own questions. I discuss these below under the headings of authority and normativity.


    1.2. Authority.


    The first question arises directly from the claim Stern makes about the 'ought' and the 'can', namely that Kant aims to say something about the authority of the 'ought' for individual agents: it is a demand that requires that individuals act on it. Henry Allison underlines this point: 'the moral law confronts us not merely as a lofty and admirable ideal but also as a source of an unconditional, inescapable demand upon the self' (Allison 1990:68). To put it more succinctly: 'What I ought to do, that I can do!' (Brown 1950:281). Still, enthusiastic commitment to one side, the prospective agent may still ask herself: Is what I am being commanded possible?


    In the quote from the first Critique above, the only constraint on the content of what is commanded is that it be a 'possible action'. 'Possible' in this context does not yet include any reference to natural conditions (this is coming later, in the description of an action that is possible). So the possibility invoked here must be logical possibility or absence of contradiction. This puts some constraints on the commanded action, that is, the agent can rest assured that there will be no command of the following form:

    O (A&~A)


    4 See Griffin 1992; discussed in detail by Stern 2004. For a contrasting ethical vision to that proposed in Griffin 1992, see Martin 2007.

    5 For detailed treatments of OIC in Kant see Timmerman 2003 and Stern 2004; for a discussion that distinguishes the Kantian and non-Kantian versions see Brown 1950. For probing of the precise relation of implication spelled out in OIC see Sinnott-Armstrong 1984.

    CON-TEXTOS KANTIANOS

    International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    325


    Katerina Deligiorgi


    Aside from the fact that the 'ought' would immediately disqualify as a rational command were it to command contradictory actions, there is evidence that Kant holds that non- contradiction is a condition for thinkability and thinkability a guide to possibility.6


    But from the perspective of the agent, assurance that the 'ought' does not command contradictory actions is not hugely reassuring. What she is asking, when she asks about possibility, is whether, in the real world, when she sets out to do what the 'ought' commands -precisely as Stern and Allison enjoin- that such commands do not generate conflicts. Conflict can be represented as:

    O (A) & O (~A)

    While O (A&~A) is false, O (A) & O (~A) could be true, if there are two conflicting imperatives in place. The problem with conflicts that do not admit of a moral solution is that their very prospect threatens to detach the content of the command from its authority, leaving the agent with sheer authority. A morality of sheer authority is a morality of 'Do not argue! Obey'.7


    Kant is alert to this problem and seeks to dissolve it by denying that conflict between oughts can arise:


    A conflict of duties (collisio officiorum s. obligationum) would be a relation between them in which one of them would cancel the other (wholly or in part). - But since duty and obligation are concepts that express the objective practical necessity of certain actions and two rules opposed to each other cannot be necessary at the same time, if it is a duty to act in accordance with one rule, to act in accordance with the opposite rule is not a duty but even contrary to duty; so a collision of duties and obligations is inconceivable (obligationes non colliduntur). (MS 6:224).


    Necessity is connected with the project of a priori ethics. Only an a priori law applies unexceptionally and only on the basis of an unexceptional law can the binding nature of moral commands be explained; as Jens Timmerman points out, in his discussion of this



    6 See A 220-1, B 267-8 where Kant makes the contrast between possible experience and possible thought, the latter has as its condition a logical condition of non-contradiction, whereas the former has as its condition agreement with formal conditions of experience. A simplified version of this discussion is included in the essay, "What real progress has metaphysics made in Germany?" 20:325-6. I discuss possibility in section 2, below. For a discussion of the universalizability test as a thinkability test for principles of action see Deligiorgi 2012:136-7.

    7 A morality that issues conflicting imperatives can be seen to embrace one horn of the Euthyphro dilemma: the divine (or here: moral) command is what takes precedence over the content of what is commanded. This seems to go against Kant's rationalist intentions. In addition, if the source of the command is, as Kant argues, reason itself, then rationally opaque commands, such as those that lead to conflicts, could be very corrosive (I call them opaque because they offer no rationally obvious way forward). This is one way of understanding what G. E. M. Anscombe found so troubling about non-theistic modern moral philosophies: the content that is commanded and we can see rationally why this is so is different from the authority of the command. Now the two can go together provided we do not get absurd results. When content and form separate, then the very authoritativeness of morality is threatened.

    CON-TEXTOS KANTIANOS

    326 International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347

    ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    The 'Ought' and the 'Can'


    passage, necessity accounts for 'the unrelenting normative force of moral principles' (Timmerman 2013: 41), and since necessity cannot be obliterated, there cannot be conflict.


    Note, however, how the non-conflict argument is secured: it is secured precisely by invocation of the necessity of the 'ought' that is its modal status for agents: the very fact that it appears authoritative, that it commands necessarily. No mention is made of the content of the 'ought'. So rational content and commanding form pull apart once again.8 The risk can be illustrated, if following Bernard Williams's (1973), we apply principle of agglomeration to commands. The principle states that if one has a duty to do something and also a duty to do something else, then one has a duty to do both:

    O (A) & O (B) → O (A & B).

    If both duties cannot be performed, a practical inconsistency arises. One way of getting round it is by rejecting the agglomeration principle, and so the implication of practical inconsistency, while insisting on the authority of the 'ought'. If we do this, as Alan Donagan comments, we would have a morality that is not inconsistent but at the cost of a morality that is 'absurd and nasty' (Donagan 1984:299).9


    Here is then the question that arises when we consider the Kantian version of OIC, which consists in the recognition of the authority of the 'ought' as action-guiding for individual agents: how can authority and a more substantive conception of rationality come together? Authority and rationality must come together in some way to avoid the sheer authority problem. On the other hand the authority of the 'ought' cannot be made conditional on content. So a better way of posing the questions is: how can we acknowledge the two sides of the 'ought', morally commanding and rationally obligating, without allowing them to detach from one another?10 I will make a first attempt at answering this question, when I discuss capacities in section 2. First, I want to examine another question that arises about the 'ought' and which concerns its normative character.


    1.2. Normativity.


    Timmerman (2003) argues that besides OIC, there is a distinctively Kantian principle at work in the form of 'Du kannst, denn du sollst' [you can, therefore you should], which


    8 I am just pointing at the split between rational content and form, I am not claiming that Kant commits himself to the axiological principle that there is always something that is best to do. The danger of espousing such a principle as he would see it is that it makes the authority of morality is not contingent on some estimation of goodness.

    9 Some philosophers have denied agglomeration for deontic logic, see Marcus 1980, on the grounds that there are moral dilemmas and ought implies can. See too van Fraassen 1973. But such an argument starts with moral dilemmas and so cannot be used in the Kantian context, hence here the rejection of agglomeration is potentially more damaging.

    10 By rationally commanding I mean that the explication of the moral ought in terms of reasons (that there is something one has most reason to do morally) naturally connects with content, 'reasons' are not generic favoring items they are rationally favoring and indeed morally rationally favoring. But commandingness is yet another thing. It matters for Kant that we don't assimilate the two but it also matters I hope to have shown that the two do not pull apart in the manner just described.

    CON-TEXTOS KANTIANOS

    International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    327


    Katerina Deligiorgi


    presupposes the objective validity of the 'ought', presented to the agent independently of any natural incentives, and is therefore a statement of the agent's freedom ('you can').11 This distinctively Kantian version of OIC depends on the moral law being the ratio cognoscendi of human freedom. The problem here, as Timmerman presents it, is that those who are not persuaded by Kant's theory of freedom will simply remain unmoved by this version of OIC. But there is a more general question that needs addressing, which concerns the normativity of the 'ought'. The question is this: what is it for an objective 'ought' to be ' inescapable' (Alison 1990:68), 'unrelenting' (Timmerman 2013:41) or as we said earlier 'necessary'? This is relevant to the current discussion because how we understand the necessity of the 'ought' has implications for how we understand 'can'.


    Kant characterizes the 'ought' as follows:


    The ought expresses a species of necessity and a connection with ground which does not occur anywhere else in the whole of nature. In nature the understanding can cognize only what exists or has been or will be, it is impossible that something in it ought to be other than what in all these time-relations it in fact is of no significance whatever. We cannot ask what ought to happen in nature...'(A 547/B 575; underline added)


    Exploring the contrast with nature Kant uses here, we may identify prescription and description as the relevant contrast-pair. To say that in nature the understanding cognizes what exists is to say that the form of the cognition is descriptive. This view concerning the cognition of nature is deeply entrenched and widespread among non-Kantian philosophers, who tend to dismiss the idea that natural laws are governing laws, in parallel with the laws of a state:


    There is something deeply unsettling about any theory of laws which allows them a kind of prescriptive, normative, obligating, or regulative force. This is because a plausible account would be required of the claimed governing relation that was supposed to exist between such laws and the events which they somehow controlled. (Mumford 2000: 281)12


    Having made the basic contrast between descriptive and prescriptive, how can we map necessity onto it? Kant is not a Humean regularist about natural laws. What the understanding cognizes also has necessity. So the point is to distinguish between different necessities.


    11 The thought is that because the ought is given independently of natural inclinations and predispositions, we know we must possess negative freedom. The point is made also in Brown 1950: 280-1; and Sebastian Rödl 'Why Ought Implies Can', presentation at the 2007 Manchester meeting of the UK Kant Society. Rödl further argues that an action represented through the representation of the law is doable by virtue of reason's practicality. This is not my concern here, though I do touch on this when I examine in detail the type of modality that Kant defends in his theoretical writings (see section 3 below).

    12 Mumford is discussing Lowe here and takes seriously the possibility of normative laws to see how a sophisticated descriptivist account can work.

    CON-TEXTOS KANTIANOS

    328 International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347

    ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    The 'Ought' and the 'Can'


    If one holds a univocal conception of law-hood, such as the one Eric Watkins defends, attempts at differentiation seem hopeless. For Watkins, there are two general features of law-hood for Kant: necessity and legislation. Laws of nature show necessity through 'determination' by causally determining the behavior of the bodies to which they apply (Watkins 2014: 482); they show legislation, i.e. that some authority 'enacts, brings about, or serves in some way as the source of the law and of its necessity' through the understanding which prescribes to nature a priori (Watkins 2014: 483). Watkins ascribes to Kant a governing conception of natural laws. So it would seem that, if we accept Watkins, the strong ought / is distinction Kant draws in the passage quoted above cannot stand.


    I do not think there is need to revise the line we have been pursuing so far: even accepting Watkins, the 'legislating' role of the understanding is exhausted in placing certain demands on what is to count as a law of nature, chiefly the demand that there be some natural necessities that determine the behavior of things in nature. On the basis of such legislation, we can cognize what is and also make predictions about what will be, as Kant says. This last permits a use of 'ought' as in 'the seedlings ought to appear in two weeks time', or 'given current conditions, there ought to be good visibility tonight'. 13 The natural necessities that sustain this kind of talk -be they metaphysical, transcendental or epistemic- describe how the world is; they obtain without anyone having to do anything. The necessity of natural laws is that of background conditions.


    By contrast, the necessity of the 'ought' that 'does not occur anywhere else in the whole of nature' legislates by directly demanding uptake by those who are under its command. The necessity of the 'ought' is the necessity of a norm that is to be followed. In what sense is it then necessary? If we understand 'necessary' as equivalent to 'irresistible' then it is unclear what space is left for the 'ought' to be follow-able and so to be guiding. If it is neither a necessity of an irresistible force nor that of a background condition, then what sort of necessity is it? It has a necessity that allows for it not to obtain necessarily. This may sound paradoxical but it is not; in contemporary semantics, for example, this feature of the 'ought' is recognized by analyzing its truth conditions in terms of best or optimal worlds (rather than all possible worlds).14


    The other thing that distinguishes the 'ought', in the passage cited, is what Kant calls its connection to its 'ground'. I take it that this ground is reason itself, which makes the 'ought' a rational command addressing rational agents. But this relation needs to be spelled out


    13 These are examples of the so-called epistemic flavor of 'ought'.

    14 The impetus for some of this work comes from attempts to offer a unified account of the logic and semantics of ought across the different uses; for an interesting application to OIC, see Wedgewood 2013. I am not claiming that this semantic analysis directly applies to the Kantian ought, simply that deontic necessity has a peculiar character that is not captured in standard analyses of necessity.


    CON-TEXTOS KANTIANOS

    International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006


    329


    Katerina Deligiorgi


    with a bit of care else we stand to lose again the sense in which the 'ought' is a follow-able rule.


    Mumford describes a puzzle in the context of normative or governing conceptions of laws of nature, that is, how the nomic relation governs particular observed phenomena. Note, however, that the situation is not crystal clear with respect to the 'ought' directing human agents either. This is because we need a relation between the 'ought' and the practical attitudes or actions it determines, such that the relation is not any of the following things: not a conceptual relation, e.g. the 'ought' describes what rational beings are qua rational, and so just part of what agents are anyway (a conceptual necessity); not a relation of irresistibility, for reasons discussed earlier; nor yet one of caprice, because it would leave the determination by the ought inexplicable, a matter of luck and so not real governance. Certainly, in contrast to natural events, we can say that the 'ought' connects with the practical attitudes it determines because there is a rational agent who 'gets' what the rule demands and adjusts her practical attitudes accordingly. But the question then is what is it for the agent to 'get' the 'ought'? And relatedly, how does the normative strength of the 'ought', its necessity indeed, fit with the possibility that it be not followed? I shall make a first attempt at answering these questions below by considering the abilities of the agent and her options, given the 'ought'.


  2. Can: abilities and options.


    Kant has a detailed and rich account of the mind, which comprises elements that must be considered essential to human mental life and those that form part of what we may call today a descriptive phenomenology of the mental; human beings receive sensations, respond to stimuli, have predispositions, inclinations and tendencies, they have functions and states of mind, they employ faculties, powers and capacities to cognize, judge, act, think, enjoy. Kant's own attempts to organize this complex material are guided in each case by the main topic he tackles, so it is not easy to offer a neutral overview, except for noting that his account of the human mind is dynamic: in interacting with its environment, the mind exercises a range of active and passive powers. Our topic here is the agential powers or capacities needed given the 'ought'.


    The previous examination of the 'ought' shows that it has the following characteristics: it is commanding and rationally obligating, and it is necessary yet capable of being not followed (and correlatively, capable of being followed freely).15 So minimally, given this specific 'ought' the agent must possess some receptive abilities in order to be alert to its command, some rational abilities to understand the 'ought' as giving her reasons, and


    15 Authority and normativity have 'force' of the same kind, the 'ought' is a command you can disobey and has a necessity that is not irresistible, but are not the same, they are at right angles to one another; in fact the 'ought' can be identified in a diagram with different values for 'imperative' and 'categorical'.

    CON-TEXTOS KANTIANOS

    330 International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347

    ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    The 'Ought' and the 'Can'


    finally discretion about how and indeed whether to follow the command; only then does the more traditional understanding of 'can', concerning practical application or implementation of the 'ought', arise. I will therefore leave this last to the final section, and focus on the elements of 'can' that directly flow from the analysis of the 'ought'.


    Here is how Kant presents the matter:


    The very concept of duty is already the concept of a necessitation (constraint) of free choice through the law. This constraint may be an external constraint or a self-constraint. The moral imperative makes this constraint known through the categorical nature of its pronouncement (the unconditional ought). Such constraint therefore does not apply to rational beings as such... but rather to human beings, rational natural beings, who are unholy enough that pleasure can induce them to break the moral law even though they recognize its authority... But since the human being is still a free (moral) being, when the concept of duty concern the internal determination of his will (the incentive) the constraint that the concept of duty contains can be only self-constraint (through the representation of the law alone) (MM 6:29-380)


    A number of things are noteworthy here:


    1. First, that the 'ought' is 'known' and 'recognized' fits with what we said about rational receptive abilities. So one thing to examine is how these abilities work so that the agent is not merely a passive recipient. The agent's being merely a passive recipient is problem given that we want to say that this is a morality of autonomy for individuals, not just for reason itself, broadcasting its commands to rational receptors, who somehow absorb and replicate the 'ought'.

    2. Secondly, what is 'known' and 'recognized' is not inert, it strikes the agent with a certain force as something to be done.

    3. Third, the doing is not automatic; while the agent recognizes the authority of the 'ought', she may or may not put it to practice.


      1. Abilities


        Starting with (a), we may ask what exactly is known and recognized by the agent? The 'ought' certainly. But what 'ought'? One answer is: the 'ought' as authoritative command, as something that is morally required. Another answer is: the 'ought' as the answer about why something is to be done; conclusive moral reasons.


        Let's take the former first. The receptive powers fit well the recognition of the call of morality: the 'ought' commands. Now it is possible for a command to strike one forcefully and yet for one to disobey. An officer can give an order, which is received and yet disobeyed. Similarly with the moral agent: she can recognize the 'ought' and fail to do it. So here is a simple solution to the problem of autonomy: the agent exercises discretion

        CON-TEXTOS KANTIANOS

        International Journal of Philosophy

        N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

        Doi: 10.5281/zenodo.2384006


        331


        Katerina Deligiorgi


        about what to do, even while she knows what ought to be done. She is passive in getting the 'ought' and active in breaking it. But, the charm of rebelliousness notwithstanding, there is something odd about valuing this kind of discretion, given that ex hypothesis there is a clear moral ought in view. Autonomy that consists in breaking, rather than following, the law is not a morality of autonomy, and certainly not what Kant is after.


        Here is an alternative: the agent expresses her autonomy, when she obeys the 'ought', because she could break the law, but she does not. Though this option is much more plausibly Kantian, it is not complete in itself. This is because we need to find out why she does not break the law. It would not count as freely obeying, if she were fearful of sanctions, for example (the fear would be a force pressing on her).16


        At this juncture, we may turn to the second answer we gave above: the agent does not disobey the 'ought', even though she could, because, besides its commandingness, she recognizes the moral reasons that speak for the action (the 'conclusive' reasons in the modern idiom).17 On this interpretation, the agent is using her discretion for something that is worth doing and in doing whatever it is she is doing she is expressing who she is, she is a good person.


        In the contemporary discussion, there is an important line of argument, which originates with Susan Wolf and identifies freedom with the recognition of moral reasons and action based on such recognition. This view is attractive from a Kantian perspective, because the agent does the right thing because she recognizes it as such. While the view is not a view of autonomy, it could fit Kantian autonomy because the idea that freedom is the freedom to do the right thing captures the positive conception of freedom as autonomy that is central to Kant's ethics.


        Wolf's basic thought is that we value freedom because of its connection to the attribution of moral responsibility and moral practices of praise and blame. So what we want is a notion of freedom that secures these practices. 'The freedom required for moral responsibility', Wolf argues, 'is the freedom to be good' (Wolf 1980:166). Dana Nelkin (2011) develops a full account of this position, which she calls, the 'rational abilities' view. The rational abilities view is that 'one is responsible for an action if and only if one acts with the ability to recognize and act for good reasons' (Nelkin 2011: 3) and 'what bestows responsibility is the possession of certain rational abilities' (Nelkin 2011: 7); specifically


        16 Equally she could be indoctrinated to think she cannot break the law, or if it just happens that she experiences no countervailing forces, such as pleasure, as Kant mentions in the paragraph cited above, she could be just morally lucky or depressed. There is a difference between contingent absence of countervailing forces and conscious effort to shape one's character so that it is good and so have a hand in setting the subjective conditions that 'help or hinder fulfilling the laws of a metaphysics of morals' (MM 6:217).

        17 Note that reasons connect to obeying the law not its authority. Reasons do not commit Kant to intellectualism, but to show this requires analysis of action per motiva.

        CON-TEXTOS KANTIANOS

        332 International Journal of Philosophy

        N.o 8, Diciembre 2018, pp. 323-347

        ISSN: 2386-7655

        Doi: 10.5281/zenodo.2384006

        The 'Ought' and the 'Can'


        the 'ability to recognize good reason for acting and the ability to translate those reasons into decisions and actions' (ibid.).


        While the existence of such abilities and their exercise certainly matches the requirements set by our previous analysis of the 'ought' -and additionally gives us a conception of freedom that fits with doing the right thing and with the attribution of moral responsibility, all of which are desiderata from the Kantian perspective- the view comes at a cost. The cost is the asymmetry between responsibility for good behavior and responsibility for bad behavior. The moral agent, who recognizes and does the right thing, is morally responsible and praiseworthy. The immoral agent, who fails to recognize and do the right thing confronts her judges with a puzzle: did she have the relevant ability properly to respond at the time of her wrongdoing or not? Only if this question is answered in the positive, is she to be held responsible and so be blameworthy. The asymmetry consists in this: 'the ability to do otherwise is required for moral responsibility when one does a bad action or an action for bad reasons, but not when one does a good action for good reasons' (Nelkin 2011: 16). So a very good person who cannot but do the right thing is praiseworthy and free even if she had no option but do what she does.18 She is free ex hypothesis: this is what the rational abilities view states. In other words, once she acts for the good reasons, which she recognizes as such, that's all that needs saying. The existence of alternatives plays no role. Hence someone who could not fail to obey the law is free. A very bad person who cannot but do the wrong thing is not blameworthy, because she is not free in the relevant sense, she simply lacks the relevant abilities. To be held responsible she must have the abilities and so the power to have done otherwise.


        To show why from a Kantian perspective the asymmetry is a cost, we need to take a closer look at the good agent. One of Nelkin's examples is Rosa, who helps a drowning child and is open to praise even though she could not have responded differently. Let us put to one side the importance Kant places on moral struggle to overcome contrary forces. Let us consider the matter in terms of Rosa's agency, not in terms of the praise due to her. The idea is that nothing interferes with her exercise of her rational abilities, so she is free. Taken out of its modern presentation this is a classical rationalist ethical picture: qua rational, agents are primed towards the good, doing the good is just a matter of the powers or abilities to manifest themselves. Rosa, in Nelkin's example, is primed towards the good and, as she suffers no interference to her powers, these powers effortlessly follow their course and manifest themselves.


        The problem with rational abilities is this: whether we consider rationality as a natural power or as a supernatural endowment, on the rationalist account, the good and freedom



        18 I have a minor concern here about the praiseworthiness of the agent; there is an elision between agent and act, that is, we may consider the act praiseworthy because it is in accordance with some good standard, but we may think the agent good because she is a living exemplification of a good standard, but not praiseworthy precisely because she could not have done otherwise.

        CON-TEXTOS KANTIANOS

        International Journal of Philosophy

        N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

        Doi: 10.5281/zenodo.2384006

        333


        Katerina Deligiorgi


        are just what it is for reason to realize itself. Such agency is clearly distinct from the passive recipience agency since practical reasons play a key role here. Still it is uncomfortably close to what Kant, uncharitably perhaps, calls the 'spiritual' and 'thinking' automaton 'driven by representations' (CPrR 5:97, 101) of rationalist compatibilism.


        The reason for Kant's dissatisfaction can be understood if we appreciate that for Kant freedom to do the right thing is not identical to the freedom of doing the right thing as Wolf and Nelkin would have it. The freedom of doing the right thing, the positive conception of freedom as autonomy and rational willing, presupposes a neutral freedom of choice of ends. Choosing the ends of reason, i.e. the agent's making use of her rational abilities as in Nelkin's account, presupposes for Kant that the agent is free with respect to ends, free to do the right thing. Unlike contemporary compatibilists, who consider actual interferences to the ability to do otherwise as the only threat to agential freedom, Kant is equally concerned with inevitability, this is why he insists that Willkür or the 'capacity for choice' is free (MM 6:213). Of course, what matters for Kant is freedom from stimuli, and just like the rational abilities view states, Kant thinks that choice is determinable by the intellect and therefore amenable to rational rules. However, he also has a notion of free Willkür that is not that of an already rationally determined Wille.


        To put it differently: to say that the necessity of the 'ought' is such that it allows for the 'ought' not to obtain necessarily (i.e. it is not the antecedent to a necessary consequent) is not to say that the 'ought' obtains just in case any obstacles are cleared from its path. The agent is not sliding, rationally, towards some end or other. Kant may be a Newtonian in science but he is not one in ethics. I think this is the deeper and non-reconcilable difference between the two accounts.19


        It is no surprise then that Nelkin considers OIC to be axiomatic (Nelkin 2011: 108) and basic (Nelkin 2011: 111). This is because the 'ought' and the 'can' are different ways of talking about the same rational ability, the power ('can') of seeing and doing the right thing ('ought'). While Kant must assume that we have abilities of response to good reasons and of resistance to contrary forces, this is not all the 'can' there is to the 'ought'.


      2. Options


    One way in which we obtain a more differentiated account, without going all the way to the notion of spontaneity, is by looking more closely at (b) and (c), starting with the agent's weighing of alternatives when confronted with a practical question about what is to be done in a given situation. We can secure the freedom of the agent's choice -the freedom to


    19 Nelkin links her position to Kant's but in order to emphasize the importance of freedom from the first person perspective of rational deliberation (118-9); I return to this in section 4. She also, like Hilary Bok, whose views I discuss below, reads Kant as a compatibilist (Bok 1998: 88-9); again more on this in 4.

    CON-TEXTOS KANTIANOS

    334 International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347

    ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    The 'Ought' and the 'Can'


    do- by considering the options she has, her power to do or to refrain from doing, which Kant associates with Willkür (see Vigilantius 29:1014 and MM 6:213). The possession of such a two-way power fits the view of the agent as deliberating about her options and can help as a complement and stepping stone to the full rational abilities given above. 20 Although not distinctively Kantian, the idea that agents possess a two-way power fits what Kant says in the passage about the agent being subject to external and internal forces and exercising some control over some of these forces.21


    The question is how to describe the availability of options to the agent. Hilary Bok discusses this extensively in her Kantian inspired compatibilism, the attempt that is to show 'freedom of the will and moral responsibility are compatible with mechanism' (Bok 1998: 56). Central to her argument is that 'theoretical and practical reason do not establish genuinely conflicting claims' (Bok 1998: 74). The former is concerned with causal history, the latter is characterized by deliberation about what we have most reason to do and presupposes claims about alternatives (Bok 1998: 121). So the two-way power I described above will feature in what Bok calls the practical point-of-view. I am interested here in the conditional account of the availability of options Bok develops to answer the question of what makes true counterfactuals about agent choices.


    Bok's answer is compatibilist in the following sense: possibilities are to be described in the 'general' sense that is relative to an incomplete description of circumstances 'one that abstracts from certain kinds of information' (Bok 1998: 94). This allows for a conditional analysis of agential options, which defines 'our possibilities relative to a description of our circumstances that treats the course of our deliberation and the outcome as variables' (ibid.). Theoretical information about causal history is abstracted. With this background, Bok argues that if we use 'can' in this broader compatibilist sense we have a satisfactory account of agential options, since the agent would have performed such and such action had she chosen so to do. More specifically:


    [I]t is possible for A to do x at t if the propositions ‘A does x at t’ is compatible with those propositions about the state of the world at t that can be inferred from some proposition expressing the state of the world at some instant prior or at A’s choice, together with those that express laws of physics if we prescind from any information about what A actually chooses or about events that constitute her choice (Bok 1998: 97, n.4).

    What is new in Bok's conditional account is the way she indexes 'can' to the practical point



    20 I say 'appears' because just a psychological account won't secure the reality of the two way power.

    21 For example the account fits traditional compatibilism, which states (1) determinism is true (2) free will matters (3) free will consists in the exercise of a two-way power or choice between alternatives (4) free will is compatible with determinism (see Campbell 2005). This is relevant to Bok's argument to which I turn below. A key reference is Moore 1912: 84-95, for articulating clearly the conditional analysis of free will that supports (3) and states that doing A rather than B counts as an exercise of free will just in case we would have done B had we chosen so to do.


    CON-TEXTOS KANTIANOS

    International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006


    335


    Katerina Deligiorgi


    of view. Possibilities exist prior to the choice, as alternatives in the agent's deliberations, because the practical perspective shields the agent from the events that constitute her choice and make it actual. It is from that point of view that the conditional analysis of 'can' makes sense.

    While I think Bok is right in trying to show the practical standpoint is not reducible to the theoretical one, I find her discussion of options and specifically of 'can' unsatisfactory. 22 Absent some information, we may make 'can' statements. For Bok this is a consequence of her acceptance of the truth of determinism, or 'mechanism'. It is this that motivates her 'general' sense of possibilities. But the basic point she is making is that there is no absolute sense of 'can'. There is support for this view if we think of the semantics of 'can', indeed there is a long-standing view that 'can' is best understood as an indexical, the truth conditions of which vary systematically according to context. Be this as it may, the point about the non-absoluteness of 'can' suggests, strongly, that no conditional analysis, such as Bok attempts, can capture possibilities for an agent, because any such attempt will involve some general antecedent, 'if she chose to...' in this case - so the threat is that the conditional analysis will be either vacuous or false.


    An obvious point to make is that the account of the freedom of the practical standpoint from which the agent endorses one of the alternatives she considers falls foul of Kant's criticisms of 'comparatist' freedom, the agent pursues a course of action caused by her ideas and her desires, but these 'determinations' have psychological causality, which may be classed as 'psychological freedom' but are still the product of natural necessity (CPrR 5:96-7).


    Clearly, those who do not want to go down the path of transcendental freedom may be happy enough with what Kant here calls 'psychological freedom'. If so, however, we need to know how the practical point of view deals with internal interference, such as manipulation or indoctrination, that is, barriers to choice itself. If we were to make the deliberation more demanding -which Bok wants to avoid (see Bok 1998: 75)- in order to show that there are ways in which the agent can exert control over her psychological states, then the practical standpoint would be over-intellectualized; it is implausible to think that prior to all action, we undertake the sort of demanding rational self-inspection that would allow us to detect such internal interferences. These problems are not insuperable and can be addressed by holding a gradualist conception of freedom, relative to how we cope with internal and external barriers. Whereas Kant does allow for such a gradualist conception, he considers it of secondary importance and supervening on an absolute conception.

    The upshot of this discussion is the following: either we go for a context sensitive 'can', which only gives us a relative 'can' and a gradualist conception of freedom, or we seek to


    22 I have my doubts about whether her way of going about it is the best, in particular whether the combination of epistemic limitations and evaluative practices and attitudes make for a coherent standpoint. If 'for all we know such and such is possible' is an essential component of the practical standpoint, as it seems it is, then it is not clear to me why this is not just an instance of the broader category of epistemic possibility.

    CON-TEXTOS KANTIANOS

    336 International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347

    ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    The 'Ought' and the 'Can'


    root the 'can' in spontaneity, in other words, we turn to the task of vindicating some powers of origination. From a Kantian perspective the former is unsatisfactory the latter impossible (strictly speaking: insofar as transcendental freedom is not subject to proof; there are arguments that are short of proof that Kant employs). In the next section I show what can be done with the material Kant provides us with to defend an unrestricted sense of 'can'.


  3. 'Must be possible': modality and inference.


    We have been led from two different interpretations of 'can' to the need to say something about spontaneity. I propose to simply acknowledge this for the moment and continue with a more literal task, which is to make sense of the 'can' of possibility which features prominently in the passage for the first Critique cited at the start, 'this 'ought' expresses a possible action' and 'of course the action must be possible under natural conditions if the ought is directed to it' (A 548/ B 576). I covered briefly the notion of possibility in relation to thinkability. I want here to discuss in a bit more detail the notions of possibility employed here and then open up the discussion to the Kant's conception of modality and the sort of inferences it licenses.


    A note of caution here: my aim in what follows is not to give an account and defense of Kant's theory of modality. Rather it is to focus on a specific innovation, roughly that there is more than logical and real possibility, and show how this innovation helps understand the 'can' in the context of OIC.23


    1. The 'ought' expresses a possible action.


I take the first reference to possibility to be a reference to a thin conception of possibility, understood as thinkability. In an earlier passage when Kant tries to explain the possibility of objects, he concedes that the 'character' of such possibility conforms to the necessary logical condition of non-contradiction (A 220/ B267-8).24 He then goes on to show what more is needed for the possibility of things. To return to the ought, the idea that it expresses a possible action is then to be understood in a minimal way as an action that is not contradictory (which, on some interpretations of the GW universalizability principle eliminates some actions at least in the so-called contradiction in conception test). Possible here has as its contradictory not-possible (impossible). But just like the point he makes in the earlier discussion with respect to the possibility of things, a similar demand for a more defined conception of possibility arises also with respect to the possibility of actions, which is captured by the phrase that the action must be possible. Kant completes the phrase with reference to natural conditions. I examined previously various substantive versions of


23 A fuller discussion will of course need to engage with Stang 2011, which focuses on real possibility.

24 See too Lectures on Metaphysics, L2 28:554-5 and esp. 28:557.


CON-TEXTOS KANTIANOS

International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006


337


Katerina Deligiorgi


these 'natural conditions' (capabilities etc.) in order to put some flesh on what is possible and not possible for agents as such, that is, with respect to what they must be able to do. Because both accounts examined were compatibilist, the natural conditions understood as the environment in which the agents exercised their abilities and considered their options were assumed to be deterministic. What I want to do now is to examine the augmented conception of possibility ('augmented' in comparison the 'mere' possibility of thinkability) that features in the consequent: 'the action must be possible under natural conditions'.


3.2. The action must be possible.


The question is: how does the agent consider her options? What makes the question tricky is that she takes herself to confront an open-ended future; this is the point -or one of the points- about spontaneity. Just like contemporary libertarians, Kant is committed to a notion of possibility that is sustained by a deep picture (indeterminism on the contemporary libertarian side, spontaneity on the Kantian side) that is irreducible to physical determinism. One direction in which this discussion can go is to examine how Kant defends the deep picture. But this is not quite our concern here, because, however Kant goes about it, this is not how an agent confronting options and having to make moral decisions is likely to be going about it. The question, 'how does the agent consider her options?' leads us to the following one: is there an account of modality that makes sense of the agent's consideration of her options? 25 Such an account should help explain the augmented sense of possibility employed by Kant but also secure a nexus between the agential standpoint and a conception of possibility whose contrary is necessity, rather than impossibility.


I think that Kant's account of formal modality fits exactly the bill. However it is quite distinctive and can look a bit of an outlier. So I want to start with a very brief overview of major rivals to show its attraction for anyone who wants to defend unrestricted agential optionality.26 The standard alternatives are Lewisian possibilism (or extreme realism, or reductivism), Aristotelianism and Platonism.


Lewis's position is reductively realist about modal facts, it aims to analyze and thus reduce modal facts to non-modal facts. The theory uses the notion of 'possible worlds'. Possible worlds are the same sort of thing as our world, they differ 'only in what goes on at them' (Lewis 1973:85). Our world, here and now, is to be designated as 'actual'. Possibility is


25 This availability also allows A to say, "I could have ψ-ed rather than φ-ed, but I chose to φ because of reasons R". Once she acts and 'A φ-ed' is true, then on the 'spilt milk' necessity conception, the proposition is necessary in the sense that nothing can make it untrue (see Lemmon 1956:389). The argument I present owes a lot to Baldwin (2002).

26 This already rules out forms of fictionalism about modal facts, since fictionalism is the product of scepticism about modality together with recognition of its importance. The desideratum is a position that captures the genuineness of the option confronting the agent, in the sense that it does not pick on a state of affairs that justifies the agent who considers 'either to φ or to ψ'.

CON-TEXTOS KANTIANOS

338 International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347

ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

The 'Ought' and the 'Can'


analyzed to states of affairs in some possible world. So the reduction of modal facts to non- modal facts takes the form:

It is possible that p, if p is true in some possible world.

A problem arises in the context of how agents consider possibilities. Take someone who considers whether to φ or to ψ. She takes both as options and so as possible. On the reductive model both are true in some possible world, one in one possible world, the other in another. But this does not capture the agential view that the two options are now possible for her, that is, it does not capture the proposition she encounters and considers, which is 'either to φ or to ψ'.27 Here the difficulty comes from the basic shape of the theory its ambition to parse out into possible worlds discrete possibilities. There is nothing that addresses the disjunctive way in which options appear to the agent.


Aristotelianism is realist about modal facts and, just like Platonism, which I consider below, takes modality to be a primitive feature of the world. The acorn is possibly an oak tree, in the sense of having the potentiality of becoming an oak tree. Possibilities are grounded in powers that entities possess. Aristotelians attribute powers to entities by tracking causal relations - the slogan is: actuality grounds possibility. So the modal profile of the entity is based ultimately on its causal profile as it interacts with other entities. Possibilities are then dependent on powers entities have and so on nested causal stories, which go all the way down to some basic substances. In short, possibilities qua powers, dispositions, capacities -but not qua occasions or opportunities- are indexed to kinds or, more accurately, essences, which are identified as sets of powers entities have given their behaviour. The position is attractive for those who want to assert that there are distinctive powers of the rational or the human kind. This is fine as far as the empirical picture of agency goes (i.e. it fits well, for instance, with the rational abilities view, and also with Kant's discussion of empirical psychology -see below). The attraction of the position also sets its limit however, because it can only explain local possibilities: given actual interactions with other agents and her environment, this agent now has this set of options rather than this other set of options. This is reasonable enough as a story of how such and such came to be the options confronting the agent, but it fails to capture the open- endedness that is characteristic of the unrestricted sense of 'can' we are after.


Platonism is realist about modal facts and treats possibility as a non-analyzable property of abstract things e.g. thoughts, propositions (or sets of propositions). This is useful for open- endedness because the options set out in the disjunctive proposition that confronts the



27 Plantinga makes the point that there is something morally odd to say the least to think that the option you discard as not right exists in some possible world. The point I try to make here is that it also voids of sense the disjunctive form of the proposition turning it into a conjunct. One might try to address this by saying 'necessarily either to φ or to ψ' but that's not how things stand ordinarily with agents and that's not how libertarians consider options. Furthermore it is important to keep the disjunctive form of the possibilities, because otherwise all sorts of possibilities can crowd the account that are irrelevant because never considered by the agent or not pertinent to the choice at hand or are simply outlandish (this last point is standard in the literature).

CON-TEXTOS KANTIANOS

International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

339


Katerina Deligiorgi


agent are not constrained by the agent's history or the history of her kind and so on. It does justice then to how the agent considers her options, namely as possibilities tout court. The only problem is that these possibilities are features of abstract things and so it is not clear what succor their existence can give the agent who is dealing with spatiotemporal particulars.


  1. 3. Modality


    Kant's formal conception of modality is a feature of judgments, yet it does not suffer from the Platonic problem identified above. What follows is an attempt to make good this claim.


    A judgment is a 'cognition' (A 68/ B 93). Specifically, it is a 'mediate' cognition of an 'object' (ibid.), which is to say that it is a cognition of objects through concepts, rather than directly through immediate apprehension of particulars. Concepts pull together different representations under a single common one and allow for relations between representations. Judgments are ordered applications of concepts, or in Kant's words, 'functions of unity among our representations' (A 69/B 93). Kant's example, 'bodies are heavy' is a judgment that brings together different representations, 'bodies' and 'heavy', through the concepts of 'body' and 'heavy', in an affirmative statement.


    The basic forms of judgments are quantity quality relation and modality. Modality differs the rest, because it is a feature of all judgments. What enables Kant to say this is that he is interested in formal modality. He is not denying that modal terms may apply to propositional content. It is just not what interests him here as transcendental philosopher. He is interested in the functions of thinking.28 Functions of thought are the ways in which the judgments are thought by thinkers, specifically how they feature in inferences. Modality 'is distinctive because it contributes nothing to the content of the judgment ... but rather concerns only the value of the copula in relation to thinking in general' (A 74/ B 100).29


    Key to understanding formal modality is the 'relation to thinking in general'. To take an example, let us consider a modal judgment of possibility, which Kant calls 'problematic'. We know already that the judgment is not one in which a predicate is applied qualifiedly to some thing. The reference to 'relation' suggests that it is not a detached qualified judgment


    28 The distinction just drawn is not quite the de dicto / de re distinction. De dicto modality describes the application of the predicate 'is necessary' to a dictum; de re modality describes the modal application of a predicate to a res. So while it is clear that Kant is ruling out here de re modality, it is not the case that he is opting for the de dicto variety, or so I want to argue; he is proposing, as he often does, a different take on the inherited distinctions. Whether it is successful or not is another issue.

    29 There are broadly two lines of interpretation one that takes Kant to be discussing epistemic modality - 'the modality of judgment reflects the attitude the judger has toward the acts of affirming or negating' (Mattey 1986: 426)- and the role of judgment in syllogism -'the mode of a judgement ... is where it appears in the course of reasoning of the judger' (Leech 2010:273); see Osborne 2016 for detailed discussion.

    CON-TEXTOS KANTIANOS

    340 International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347

    ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    The 'Ought' and the 'Can'


    either. This leaves the option that the relation that gives the judgment its modal value is to other judgments. The clearest example Kant gives is of the antecedent in a conditional statement, "There is perfect justice", which is not said assertorically, it is only a possibility. It is therefore a problematic judgment entertained by a thinker in the process of trying to figure something out - 'thinking in general'. The same goes with the judgments that compose a disjunctive judgment, they too are of 'problematic significance' (A 75/B 100). Someone, Kant says, 'might momentarily assume this proposition' (ibid.) in the process of thinking, for the purpose of finding truth.


    Two things are noteworthy here. First that Kant assigns importance to the role of the thinker and to the position of the judgment in thought. Second even though the focus is on form, the content is not simply irrelevant -otherwise there would be no relation to truth. My proposal of how these claims hang together is that modal terms specify how a thinker thinks about things. The relation to thinkers is not a matter of psychology. 30 It is a matter of the unifying function of the understanding, which is, in turn, a matter of judging, a matter of how judgments are located in inferential relations. We can of course understand Kant's references to the 'consciousness' of mere possibility as references to epistemic attitudes, but the modal value is not generated by the thinker, it is generated by the location of the judgment in the thinking. This is what allows Kant to affirm that all judgments have a mode, they do this as putative placeholders in webs of inferential relations.31 But this does not render the thinker redundant; the placement of judgments in inferential relations is done by a thinker. As a function of thought then, modality describes a three-place relation composed of a thinker who thinks in such and such way, forms of thought (the 'how') which judgments have by virtue of their relations in structured reasonings, and thought about things.32


    3.4. Inference


    If the previous sketch of Kantian formal modality is roughly on the right track, then inference plays a role in the very use of modal terms, in the assignation of value to the copula. What I want to look at now is at the implication relation in OIC (the implied 'then' in the conditional: if the ought is directed to it, the action must be possible). For this we need to look specifically at the thinker as agent and so reasoning with a view to acting.


    30 See Blecher 2013:55-6 for psychologism in the early reception of Kant's theory of modality. Blecher counters this by defending a progressive ascent of thought from problematic to apodeictic, which ties modality with a very ambitious conception of knowledge. While this is an interesting proposal, it is not necessary to rebut the accusation of psychologism.

    31 This third distinctive characteristic, that all judgments have a mode shows that the theory is not intended to be about de dicto modality.

    32 Baldwin sums it up perfectly, when he says that Kantian modality is about ‘critical reflection on reasonings, as when one asks whether there are possibilities not properly taken into account in the passage from premises to conclusion’ (Baldwin 2002:14).

    CON-TEXTOS KANTIANOS

    International Journal of Philosophy

    N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

    Doi: 10.5281/zenodo.2384006

    341


    Katerina Deligiorgi


    The modal category for the kind of thinking relevant to the agent considering options is possibility and the form of judgment disjunctive. So we are interested in what makes problematic disjunctive judgments about options true. In making modal claims, the thinker as agent considers how things are and states of affairs (this is the copula of actuality and empirical, spatiotemporal judgments), which that can justify her assigning a modal value to a judgment about her options. So what makes the problematic disjunctive judgment true is the thinker's appraisal of its fit with the antecedent judgments about actuality (which may in complex deliberations also include modal theoretical judgments). The problem is that this looks like Platonism with a thinker's perspective added to it, which takes away all the advantages of Platonism (i.e. that it sets out global rather than local possibilities) and a questionable conception of truth.


    The more detailed picture should alleviate these concerns: first an appraisal is needed of what is. From such appraisal, in Aristotelian fashion, conclusions about powers and possibilities are drawn (this can be settled knowledge on which the agent draws almost unthinkingly). More simply, the problematic disjunctive judgment does not spring out of nowhere, we already need the awareness that something is demanded morally and a reading of the situation in terms of reasons for action and the consequent appraisal of what is in the agent's power to do and how other things and beings around her are, what they do and their powers. But the assignation of modal value to the judgment describing options for prospective action is the done by the thinker who has to draw relations between judgments in setting out her problematic disjunct (i.e. her options). The thinker takes of course into account what is, but she is unrestricted in her assignation of modal values to specific judgments describing the possibilities she is entertaining, because this is something she is just then doing.


    This conception of the thinker as agent is close to Bok's description of the practical standpoint. It differs though because the 'can' -or options- is a function of thought, which is not further analysable into a conditional; it is rather the other way round: antecedents of conditionals are problematic judgments in some inferential chain. Pace Bok then it is possible to have a notion of 'can' tout court, but only in the context of a formal theory of modality.


    As regards the Platonic concerns, it is precisely because the possibilities are thought, and indeed first person intentional thought, that the theory avoids the problems of the relation to actuality. And because to entertain a problematic disjunct forms part of a thinker's forming an intention to do something, what is judged possible and incorporated in the agent's intention as the thing to do has also a straightforward relation to actuality; it's the agent's business to make actual the option she goes for.


  2. OIC and Autonomy


CON-TEXTOS KANTIANOS

342 International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347

ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

The 'Ought' and the 'Can'


Let us now gather the elements of the previous discussion to see how it helps with the original question concerning autonomy, how, that is, the principle of reason's self- legislation not only has a grip on individual agents, but can also express their autonomy.


The examination of the 'ought' gives a complex result: the 'ought' stands for something that is morally obligatory and the obligation derives from its authority, its commandingness. Commandingness is distinct from content, the moral reasons telling for the commanded action. Still, it is important that the authoritativeness of the command be not unconnected to its tracking of reasons that count in favor of some action. Establishing such connection is a matter of doing justice to the unconditionality of the imperatival character of the moral 'ought' (see e.g. G 4:462), and so its basicness (CPrR 5:31), and its rationality.33 As regards its normativity, the ought is categorical, yet it may not be taken up by agents, so again a connection is to be established between deontic necessity and the fact of its obtaining such that it is neither a random nor an inevitable occurrence.


The first step in the analysis of 'can', in terms of rational abilities, helps bring the two sides of the 'ought' together through the agent's ability to respond to the command on the basis of reasons. Again the situation is delicate here: the command remains commanding even when the agent does not heed it, the conditionalisation implied by ability has to do with uptake alone (this is a general feature of laws, they do not stop being commanding when citizens are incapacitated). In exercising this ability, the agent brings reasons to bear onto why she obeys a law she can disobey, and this shows that there is a link between the 'ought' and substantive reasons. The second step of the analysis of 'can' made the case for the existence of options for the agent, for the purpose of disambiguating free choice from moral choice. The final step of the analysis of 'can' came with the account of modality that fits the open-ended structure of deliberation.


On the preceding analysis of OIC, the 'ought' implies an unrestricted conception of 'can' that is only available from the agential standpoint. This unrestricted conception is tied to an understanding of the agent as thinker. The latter is not a gratuitous addition to the account, it emerges by asking by virtue of what the 'can' of abilities and then of options are possible. The question now is how this analysis of OIC can help with our original question about autonomy, that is, how the principle of reason's self-legislation not only has a grip on individual agents, but can also express their autonomy.


4.1. Involvement and control


33 The 'ought' is internally connected to reason: rationality is a property the moral ought has by virtue of being the moral ought. Rather controversially I suggest it is externally related to the first person perspective. This is not to deny that the moral ought addresses us first personally, which is a point of agreement between Kantians and non-Kantians; see e.g. Hill 2002 and Street 2012/13.

CON-TEXTOS KANTIANOS

International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

343


Katerina Deligiorgi


The moral 'ought' can be discussed without any reference to human abilities. Its relation to 'can', and specifically its relation to abilities, not just opportunities, becomes an issue when we ask questions about the nature of the morality it commands: how morality fits with reason and how it fits with freedom. The rationality of the moral 'ought' and its relation to freedom are human concerns. It is to address these concerns -fundamentally: what does the 'ought' require of us?- that a discussion of abilities is needed. What I am talking about here is not the inference of 'implies', but a shift in perspective: from an agent-neutral discussion of the 'ought', to a perspective in which the agent is centrally implicated. In what follows, I want to use this to show how the individual is a plausible candidate for the 'autos' of autonomy. Before that I want briefly to address the issue of implementation, the 'can' as 'can do'.


The moral agent must be able to follow through with her intention. This is the step that perhaps has attracted most attention in the reception of Kant's practical philosophy: the agent must exhibit control over her practical attitudes and not become distracted in the implementation of her choice. Obviously Kant is concerned with moral choice -of which more shortly- but he draws on a long tradition in which the encratic agent is one who manages impulses. The dynamic picture mentioned earlier is relevant also here; agents are subject to a range of forces, with respect to which they are patients; Kant calls these 'stimuli' or 'sensitive causes' (see e.g. MS 6:213, Mrongovius 29:896). He argues that these causes are resistible (A 534/B562, G 4:398-9, MS 6:213), and subject to appraisal by the understanding (A 802/B830), which provides us with 'motives' or 'intellectual' causes (Dohna 28:677, Vigilantius 29:1014-5).34 So 'can do' has to do in a large part, for Kant, with control over internal interferences.


The previous rudimentary account of self-governance, understood as exercise of self- control for the purpose of realizing a thought-through intention, can be untethered from the moral context in which Kant wants to place it. Indeed, one might say that the successfully self-governing agent who deliberates about what to do and controls herself by organizing her practical attitudes has all the autonomy she can hope for. Why is there any need for an extra step that aligns the agent's ends with the ends of reason?


In a way the question is badly put: there is no need for the extra step. It is rather the other way round: the morally indifferent account of self-governance is an artificially limited account, that is, with an artificially thin sense of the kinds of valuing that informs what to do (the intention or policy formation or 'motives'). On the Kantian view of value, at least as I see it, value is a unified domain, which is certainly not to say that all value is moral, but it is to say that we can entertain conditional goods by virtue of being able to entertain the possibility of an unconditional good. Be this as it may, it is still to be shown that giving


34 The full story is more complicated than this thumbnail sketch; see Deligiorgi 2017.

CON-TEXTOS KANTIANOS

344 International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347

ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

The 'Ought' and the 'Can'


their proper weight to the considerations that are properly most weighty is expressive of individual autonomy.


A hint towards an answer is given in the account of formal modality. The thinker as agent considers the modal status of a disjunctive proposition about courses of action. In doing so, she employs her understanding, the faculty of cognition of rules (see e.g. Anth 7:197-8); she is not simply navigating among given rules, she is making rules for herself in thinking for herself. Thinking is an activity. This is relevant to moral autonomy because confronted with the moral ought and having a set of options in view, which are not just options about how best to apply the ought, but also about whether to obey its command, the agent chooses to obey. Just like any other assignation of modal values to a judgment of practical reasoning this is unrestricted, it is something she does as thinker full stop.


The agential involvement, which underwrites the negative freedom of the unrestricted 'can', means that whatever she decides to do she is in a recognizable, albeit non-Kantian, sense autonomous; that is to say, the agent is not deciding capriciously what to do. Call this 'psychological' autonomy in acknowledgement of the contemporary debate, which only deals with such types.35 From a Kantian perspective, a theory that only gives us this type of agential involvement must be considered incomplete, because the assignment of modal values to options, that is, the considerations that inform her decision about what she should do are all in a deep sense optional, reasons weigh this way and that but at a different juncture the same reasons might weigh differently. There is only one disjunct of the problematic judgment of formal modality that carries weight unconditionally: the moral 'ought'.


Recognition of the 'ought' is not mere rationomy that leads to a morality of systematic self- alienation, as Schiller worries, because the uptake of the 'ought', by an agent who sees its force and uses it to shape her ends, reveals her to be enabled as agent in a way that no other consideration does (including no other 'oughts'). The unconditional moral ought implies can, the unrestricted can of practical thought. The structure of practical thought, revealed through our analysis of OIC, shows a relation of necessity and possibility from which the agent as thinker is ineliminable. If we now add normative content, that is, consider a case of moral deliberation with a view to action, the 'uptake' of the 'ought' is just a restatement of moral autonomy: the agent does the right thing because -and just in case- she judges it the right thing to do. This is how taking the long route, through an analysis of OIC, Schiller's worry may be addressed and moral autonomy shown to have a decent claim at being expressive of individual autonomy.

Works cited



35 Psychological autonomy models are models of self-governance tend to be vulnerable to the question why the mental item expressive of the controlling self is in fact expressive of one's true self. There is extensive literature on the topic see e.g. Anderson and Christman 2005. For my own critical take see Deligiorgi 2016.

CON-TEXTOS KANTIANOS

International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

345


Katerina Deligiorgi


Anderson, J. and J. Christman, eds. 2005. Autonomy and the Challenges of Liberalism: New Essays. Cambridge: Cambridge University Press.

Allison, H. 1990. Kant's Theory of Freedom. Cambridge: Cambridge University Press. Baldwin, T. 2002. 'Kantian Modality'. Proceedings of the Aristotelian Society, Supplementary Vol. 76: 1-24.

Barcan Marcus, R. 1980. 'Moral Dilemmas and Consistency'. Journal of Philosophy

77:3:121-136.

Blecher, I. S. 2013. 'Kant on Formal Modality'. Kant-Studien 104:1:44-62.

Bok, H. 1998. Freedom and Responsibility. Princeton N.J.: Princeton University Press. Brown, S. M. Jr. 1950. ‘Does Ought Imply Can?’. Ethics 60:4:275-284.

Chituc, V., Henne, P., Sinnot-Armstrong, W., De Brigard, F. 2016. ‘Blame, Not Ability, Impacts Moral “Ought” Judgements for Impossible Actions: Toward an Empirical Refutation of “Ought” Implies “Can”’. Cognition 150: 20-25.

Deligiorgi, K. 2017. 'Interest and Agency'. In: Gabriel, Markus and Rasmussen, Anders Moe, eds. German Idealism Today. De Guyter Verlag. 1-25.

Deligiorgi, K. 2016. 'Autonomy in Bioethics'. Symposion 3:2:177-190.

Donagan, A. 1984. ‘Consistency in Rationalist Moral Systems’. Journal of Philosophy

81:6:291-309.

Double, R. 1996. Metaphilosophy and Free Will. Oxford: Oxford University Press. Frankena, William K. 1958. ‘Obligation and Motivation in Recent Moral Philosophy’, Essays in Moral Philosophy, ed. A. I. Melden. Seattle: University of Washington Press, 40–81.

Griffin, J. 1992. ‘The Human Good and the Ambitions of Consequentialism’, The Good Life and the Human Good, ed. E. F. Paul, F. D. Miller and J. Paul. Cambridge: Cambridge University Press.

Leech, J. 2010, 'Kant's Modalities of Judgement' European Journal of Philosophy

20:2:260-284

Lemmon, E. J. 1956. 'Review of G. E. M. Anscombe, Aristotle and the Sea-Battle'. Journal of Symbolic Logic 21:4:388-9.

Lewis, D. K. 1973. 'Causation'. Journal of Philosophy. 70: 556–67.

Martin, W. 2007. 'In Defense of bad Infinity'. Hegel Bulletin 55/56:168-187.

Mattey, G. J. 1986. 'Kant's Theory of Propositional Attitudes'. Kant-Studien 77:423-40 Nelkin, D. K. 2011. Making Sense of Freedom and Responsibility. Oxford: Oxford University Press.

O'Neill. O. 2003. The Inaugural Address: 'Autonomy: The Emperor's New Clothes'.

Aristotelian Society Supplementary Volume 77:1 - 21. Osborne, N. 2016. [thesis University of Lille]

Sellars, W. 1956. "Imperatives, Intentions, and the Logic of 'Ought'". In Hector-Neri Castañeda and George Nakhnikian, eds. 1963. Morality and the Language of Conduct. Wayne State University Press.159-214.

Sellars, W. 1966. 'Ought and Moral Principles' Box 36, Folder 8. Wilfrid S. Sellars Papers,


CON-TEXTOS KANTIANOS

346 International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347

ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006

The 'Ought' and the 'Can'


18999-1990. ASP. 1991.01. Special Collections Department. University of Pittsburgh. Sinnott-Armstrong, W. 1984. ‘Ought’ Conversationally Implies ‘Can’’. The Philosophical Review 93:2: 249-261.

Schiller, F. 'Über Anmut und Wurde'. In Lieselotte Blumenthal and Benno von Wiese, eds. 1943- Schillers Werke Nationalausgabe XXb. Weimar: Hermann Böhlhaus Nachfolger.

Slote, M. 1982. 'Selective necessity and the Free Will Problem'. Journal of Philosophy 79: 5-22.

Stern, R. 2004. ‘Does Ought imply Can? And does Kant think it does?’. Utilitas 16 (1): 42- 61.

Swank, C. 1985. ‘Reasons, Dilemmas and the Logic of ‘Ought’. Analysis 45:2: 111-116. Talbert, M. 2016. ‘Symmetry, Rational Abilities, and the Ought-Implies-Can Principle’. Criminal Law and Philosophy 10:283-296.

Timmermann, J. 2003. ‘Sollen und Können: “Du kannst, denn du sollst” und “Sollen impliziert Können” im Vergleich’. Philosophiegeschichte und logische Analyse 6:113–22. Timmerman, J. 2013. 'Kantian Dilemmas? Moral Conflict in Kant's Ethical Theory'. Archiv für Geschichte der Philosophie 95: 36-64.

van Fraassen, B. C. 1973. 'Values and the Heart's Command'. The Journal of Philosophy

70:1: 5-19.

Watkins, E. 2014. 'What is, for Kant, a law of Nature?'. Kant-Studien 105:4: 471-490. Williams, B. 1973. Problems of the Self. Cambridge: Cambridge University Press.

Wolf, S. 1980. 'Asymmetrical Freedom'. The Journal of Philosophy 77:1


CON-TEXTOS KANTIANOS

International Journal of Philosophy

N.o 8, Diciembre 2018, pp. 323-347 ISSN: 2386-7655

Doi: 10.5281/zenodo.2384006


347

Enlaces refback

  • No hay ningún enlace refback.
Utilizamos cookies propias y de terceros para realizar análisis de uso y de medición de nuestra web para mejorar nuestros servicios. Si continua navegando, consideramos que acepta su uso.


Creative Commons by-nc 3.0 Logo

ISSN: 2386-7655

URL: http://con-textoskantianos.net

DOAJ LogoErih Plus LogoCitefactor logoredib Logo
LatIndex LogoISOC Logo MIAR Logo
SHERPA/RoMEO Logo
MLA LogoZenodo Logo
ESCI LogoEBSCO LOGOWeb of Sciencescopus logo