Article body

Joseph Raz’s and Christine Korsgaard’s accounts of practical reasons are thought to stand in sharp opposition to each other (Scanlon, 2004, p. 233).[1] Raz’s (1999; 2012) realist account tells us things like, “reasons are facts” to which agents (have to) respond. Reasons for actions are “aspects of the world.” This Razian realism about reasons, as well as similar frameworks defended by Scanlon (2014), Parfit (2011, Vol. 1, Part 1) and many others, contrasts with the Kantian constructivism advocated by Korsgaard (2008; 2009), Onora O’Neill (1989; 1996; 2013) and Sharon Street (2008; 2011). According to constructivism, practical reasons are the result of our deliberative activities that we agents engage in, in accordance with constitutive principles of practical reasoning. And indeed, there are many moments in this debate that suggest that these are mutually exclusive positions concerning the ontological nature and practical role of reasons.

I attempt to establish some common ground between these two views and focus on the relative status that Raz and Korsgaard assign to practical reasons and practical principles, respectively. This paper’s aim and scope will be limited in so far as it singles out Raz and Korsgaard as representatives of the realist and constructivist camps respectively and focuses on their work. The notions of reasons and principles are defined in sections one through three. Raz insists that reasons, broadly understood as features of actions that make them worth choosing or rejecting, have conceptual and practical priority over principles. According to Razian realism, practical principles play the mere subservient role of so-called rules that assist deliberators in systematizing the reasons that apply to them independently (of such principles). Korsgaard takes the opposite view. Her account of normativity asserts that the principles of choice that an agent identifies with are the source of her reasons. The principles that define an agent’s self-conception are at the same time always employed when she endorses a proposed action, turning the latter into an act done for a reason.

My account consists of two suggestions, defended in sections four and five. The first one is the placebo analogy: treating one’s reasons as having an agency-independent dimension along Razian lines is an attitude that agents constitutively (have to) incorporate into their self-understanding lest their own first-personal stance should end up losing its foothold in a stable normative realm. Agents generate practical reasons on the basis of their identity-constituting principles. But in order for practical deliberation to fulfill this function, deliberators must adopt the view that this process handles items that gain their justificatory force independently of their volitional apparatus. If agents were to render this process fully transparent to themselves and treated the assumption regarding the factual nature of reasons in a constructivist manner all the way down (a manner that is presumed to be solidly justified for the sake of this paper’s exercise of combining the two views), the very self-constituting activities that Korsgaard singles out as every agent’s central practical task would undermine themselves and the agents’ identities would destabilize. Like a placebo that ceases to unfold its beneficial health effects upon being unmasked as a placebo, an agent’s reasons would lose their force if they were conceptualized in the way that Kantian constructivism suggests. For the sake of their own agency, then, agents must maintain an illusion and keep on living with a threatening instability that emerges between an ontological constructivism, on the one hand, and a practical realism about reasons, on the other.

The second suggestion for overcoming the impasse between realists and constructivists focuses on the relationship between deliberation (in terms of the aforementioned principles), on the one hand, and the phenomenon of acting for a reason, on the other. Raz claims that the latter can well take place, even if no deliberation whatsoever precedes the action in question. In response to Raz’s examples of automatic, opaque, and swift actions, my account defends the view that our reasons should not be understood as analogous to causes regarding their temporal position relative to the actions that are chosen on their basis. Employing a thought by Hegel, I present a conception of reasons’ genesis and force that allows reasons for a particular action to emerge later and after the non-reasons-based features of an action are already completed. This suggestion is then spelled out in more detail to accommodate the intuition that reasons seem to motivate us before the action takes place, on the one hand, and with respect to our accountability practices, on the other. Both of these issues cause problems for my deliberation-centred account of reasons.

1. RAZ ON REASONS AND PRINCIPLES

Raz defines reasons for action (I use the term interchangeably with “practical reasons”) as “facts which constitute a case for (or against) the performance of an action” (Raz, 2012, p. 36). Raz’s realism about reasons becomes apparent in his discussion of the conceptual connection between reasons and normativity. He claims, with respect to all normative reasons (of which reasons for action are one sub category), that “aspects of the world are normative in as much as they or their existence constitute reasons for persons, that is, grounds which make certain beliefs, moods, emotions, intentions, or actions appropriate or inappropriate” (Raz, 1999, p. 67).

Raz assigns strict priority to reasons over principles. Practical principles are “general propositions, in our context general normative propositions, which are thought to be particularly illuminating, or central relative to a certain normative view or theory” (Raz, 2012, pp. 6-7). And with respect to such principles, which he explicitly thinks are occupying the foundational role in views “descending from Kant,” Raz says: “Given that what one has conclusive reasons to do […] depends on what reasons for and against the action one has, principles are conclusions from the existence of reasons, and cannot themselves occupy a fundamental role in an account of normativity” (Raz, 2012, p. 7; my emphasis).

Raz spells out the case against principles in more detail. He presents a number of intuitive cases in which agents seemingly perform full-fledged actions and do so both intentionally and for a reason without, at the same time, acting on any principles of practical deliberation. Principles rarely figure in the agent’s choice to help others: “You see children torturing a cat and you stop them. No principle figures in your decision” (Raz, 1999, p. 226). Again, rescuing the cat is an action that is intentionally performed (the agent is controlling what he or she is doing) and it is taken for a reason (the agent, if asked, would mention “prominent aspect(s) of the situation”), such as that the cat is in pain.

Raz emphasizes that intentionality and action based on a reason do not necessarily involve the active deployment of practical principles. He assigns principles the minor role of guidelines that “may point out an important consideration to be taken into account, so that being guided by such a principle is the same as giving the reason to which it points due weight in one’s reasoning” (ibid., p. 227; my emphasis). Similarly, principles “are valid if they represent the correct outcome of rational deliberation about the merit of the various reasons which bear on […] decisions” (ibid., p. 227; my emphasis). These statements highlight the priority that Raz assigns to reasons over principles. Principles get employed by practical deliberators (if at all) when they face a hard choice, in which strong reasons are present in support of two or more available courses of action. In such situations “rules of thumb” (ibid., p. 226) might provide assistance to struggling deliberators and guide them in acknowledging and organizing the reasons relevant to their options. In such cases, agents may take recourse to pragmatic rules or even more substantive moral principles and policies in order to identify the course of action that is best supported by the independently subsisting reasons. Still, according to Raz, the practical reasons are there to be recognized, and they are the fundamental building blocks of deliberation with respect to which organizing principles are then (sometimes) applied.

2. KORSGAARD ON PRINCIPLES AND REASONS

According to Korsgaard’s neo-Kantian account of action and agency, every process of practical deliberation begins with an incentive. An incentive is a perceived “good-making property of an object” of choice. Incentives are the attractive features of the objects that figure in our actions (for example, the aesthetic features of a painting that make it the proper object of us celebrating it) (Korsgaard, 2009, pp. 122-124). The Kantian approach insists that incentives in and of themselves do not provide reasons for actions, notwithstanding the fact that agents often refer to these incentives as if they were free-standing facts. In order to get to reasons, agents need to enrich incentives with another Kantian component—namely, inclinations. Inclinations are agents’ reactions to the incentives in question. They are similar to desires in that they propose certain courses of action. The controversial claim of the neo-Kantian and constructivist account is that it is ultimately we who confer value on objects of choice by being inclined towards them. Without rational agents and their inclinations, understood as reactions to attractive features of objects, there wouldn’t be any values or reasons. Inclinations are therefore a necessary condition for the occurrence of reasons for action.

However, the presence of an inclination is of course not a sufficient condition for generating reasons for action, Kantians insist. And this is where practical principles enter the picture. They provide the cognitive means that are necessary to consciously decide whether or not to identify with an inclination—that is, whether or not to endorse the suggested course of action that the inclination proposes. Practical deliberators and agents have to employ self-imposed and choice-constraining principles to get from incentives, via inclinations, to reasons for action. In her recent work, on the function of action in the course of individual self-constitution, Korsgaard (2009, pp. 23-24) clarifies the way in which our normative self-conceptions and their principles play this role of a necessary condition in the genesis of practical reasons:

We must act, and we need reasons in order to act. And unless there are some principles with which we identify we will have no reasons to act. Every human being must make himself into someone in particular, in order to have reasons to act and to live. Carving out a personal identity for which we are responsible is one of the inescapable tasks of human life.

The crucial bit in this quotation is the identification of the relationship between (identity-constituting) principles, on the one hand, and reasons for action, on the other, as one of necessity (“unless”). Without at least one principle that an agent endorses and identifies with, there cannot be any reasons for this agent. Since, however, reasons are necessary for action and, furthermore, actions are the indispensable means of constituting ourselves into agents to begin with, we (qua minimally existing agents) inescapably have to identify with some principles.[2] “To have a reason” means that one understands oneself as an agent whose deliberative stance is structured by certain normative principles, the employment of which consists in taking up incentives in a particular way rather than another. The resulting inclination, the pulling-together of one’s deliberative stance in accordance with these deliberation-guiding principles, is, together with the initial incentive, the reason for the action that one chooses. These practical and identity-constituting principles are foundational in a twofold sense: They define and determine under which normative description we regard ourselves as agents. And, at the same time, they guide our deliberation and, ultimately, our (self-constituting) actions upon encountering objects of potential choice with their specific properties.[3]

One should also briefly acknowledge Korsgaard’s important claim that the apparent contingency of agents’ specific practical identities (and principles) does not result in practical reasons’ lack of normative force. On the contrary, the principles that constitute one’s deliberative stance are normatively binding, because they instantiate one way of dealing with an inescapable problem, the task of self-constitution and agency, successfully. You have to conceive of yourself as being normatively bound in one way or the other, in order to conduct your own practical reasoning in a way that is minimally intelligible (at least to yourself). Furthermore, the reasons that spring from endorsed inclinations are capable of motivating the person who has them, because these reasons are sourced in how the agent experiences herself in responding to the necessity to act. The elegance of Korsgaard’s picture is that it accounts for reasons’ normativity and their intrinsically motivational nature by bridging the gap between independent features of the world that trigger an agent’s process of practical deliberation and action, on the one hand, and the value-conferring power of human agency, on the other.

3. PRINCIPLED REASONS AND REASONED PRINCIPLES

The comparative reconstruction of two paradigmatic accounts of practical reasons has resulted in a conclusion that seems to support the widely shared diagnosis that realists and constructivists present sharply conflicting approaches. Reasons for action and practical principles occupy the opposite priority-status in the two theories. The Kantian conception of reasons—i.e., that they originate in rational agency—favours an account that prioritizes principles over reasons, in particular when the former are considered the constitutive elements of the deliberative standpoint, from which reasons are generated (as opposed to recognized). According to the Razian alternative, reasons take priority over principles exactly because the task of the latter merely is to guide the activity of organizing and systematizing “normative facts” and “aspects of the world” in the correct manner.[4]

In the remainder of this paper I evaluate these two views by means of applying them to problems in practical philosophy and action theory. That way, the strengths and weaknesses of Korsgaard and Raz will come into view more sharply, and overcoming the weaknesses of the one will naturally point towards some strength of the other. The suggestions have already been mentioned. First, understanding one’s reasons along the lines of agency-independent facts is itself constitutive of first-personal deliberation and self-constitution. A benign analogy with the placebo effect shows that the presupposition of realistically conceived reasons is itself a principled precondition for getting the constructivist process of generating reasons going in the first place. Second, the inherent action-directedness of practical reasons too can be explained by unifying the two accounts. It is the act of identifying with certain actions that explains why the resulting reasons are necessarily capable of motivating those who have them. However, to get to this conclusion, I will need to introduce a rather unorthodox conception of how actions relate to those reasons that serve as the basis for choosing them. This conception parts company with the view that reasons always work analogously to causes, particularly concerning their temporal priority over the actions that result from endorsing them.[5]

4. THE PLACEBO ANALOGY

One of the main worries constructivist accounts of practical reasons encounter is that they seem to leave us with an unpleasant subjectivism about practical reasons (Regan, 2002). Critics of the constructivist paradigm claim that there is an arbitrariness about our reasons for action that comes with the strategy of fully internalizing their sources. If I reflectively want Anscombe’s (2000, pp. 70-71) infamous saucer of mud for its own sake, then this inclination constitutes a robust reason for preferring to have the saucer and to take the related actions. Or think of Parfit’s (2011, p. 83) recent example of someone who, after ideal deliberation, wants to be in a future state of pain, again non-instrumentally so. And if it is only the first-personal side of the encounter with an object’s features that accounts for the emergence of reasons, then an agent with the practical identity of a mud-lover successfully acts for reasons to the same extent that agents who engage in actions that are commonly regarded as more worthwhile.

Of course, Kantian constructivists have responded to this worry. Teleological varieties, for example, have recourse to normatively potent and choice-constraining conceptions of human nature and flourishing in order to block the mud-lover from generating reasons for his or her actions.[6] Closer to Kant are accounts of rational willing that put certain formal constraints (such as the categorical imperative and the principle of instrumental rationality) on that very activity. My attempt to save the constructivist approach to practical reasons from the subjectivism objection differs from both of these strategies. It is more modest (or, some realists might say, disappointing) in that it fully absorbs the objective and external constraints on the agents’ value-conferring activities into the deliberative perspective. My suggestion is that human beings cannot go about their lives as deliberating and acting agents without postulating and presupposing the independent status of their practical reasons in an objective and realist mode of reflection like the Razian one. This remains the case, notwithstanding the well-justified pull that makes agents acknowledge the dependence of their reasons’ existence on their own activities of willing, deliberating, and choosing. Below, I employ an extension of one of Korsgaard’s own central constitutive arguments to support this claim regarding the threat of an unstable agency that constructivism poses (if it is adopted practically and all the way down). The argument at stake is Korsgaard’s attack on the possibility of what she calls “particularistic willing.” This argument can be extended to cover the problematic nature of “purely constructivist willing,” as I will call it.

But before I get to that I have to introduce the aforementioned analogy in order to show that an attempt at combining Raz and Korsgaard does not result in an unstable position. Take the phenomenon of how placebos are thought to work and of what, conceptually speaking, turns a substance into a placebo drug.[7] Once patients understand that the improvement of their condition had originated solely in their volitional set up, and acknowledge that the assumed efficacious power of the drug had been merely chimerical, exactly then does that power start fading. Patients will regard it as unacceptably incoherent to find themselves in a cognitive state in which, on the one hand, they believe that the substance they have taken is causally completely inefficacious, but, on the other hand, continue to tell themselves that it is the drug that has improved their condition. It constitutively comes with the idea of an efficacious drug that it is the drug that is (at least partly but necessarily) responsible for the achieved effect. If this weren’t so, and if the effect were considered to be producible entirely from within, patients would not be able to tell themselves that they had had any rational justification to take the apparent drug in the first place. After all, if internal mental mechanisms are acknowledged to be a sufficient means for bringing about the desired beneficial effect, then taking what one knows to be a causally inefficacious substance becomes a redundant and irrational choice.

Now, what does all this have to do with the way practical reasons work? The placebo analogy is supposed to illustrate how certain, third-personal, assumptions concerning one’s own reasons’ independent standing are a non-optional presupposition for occupying the practical standpoint of agency. It is exactly in order for these reasons to exhibit first-personally perceived authority and normative force that agents (partly) conceptualize them in the form of choice-constraining standards with a realist phenomenology. In their inescapable roles as practical reasoners and agents, just like the patient who is in need of a cure, they wouldn’t be able to generate these agency-sustaining effects if they completely negated and ignored a third-personal dimension regarding the so-called nature of their reasons—even if the account of that nature turned out, upon ontologically well-informed reflection, to be an unsustainable illusion. Part of the account defended here is that an essential part of a successfully constituted deliberative standpoint consists precisely in the ability to take a step outside of that very standpoint. This ability then manifests itself in an agent asking what the correct reasons are that he or she attempts to identify in the course of practical reflection and decision making. Employing exclusively constructivist resources all the way down and in a fully self-transparent manner won’t get the process of establishing authoritative reasons off the ground.[8]

Let me defend this set of claims a bit more. In terms of the two theories that I am trying to combine, looking at practical reasons in the Razian, substantive realist, mode is the adequate one for those who are engaged in practical deliberation and reasoning. And this remains practically true even if it turns out that there are strong metaethical grounds on which this realism about practical reasons is theoretically not sustainable. In other words, as practical deliberators and agents, humans need Razian realism about the reasons they have, even if, on a different level of reflection, they might have to acknowledge them as originating in their subjective self-understanding. Part of what drives the placebo analogy is the thought that a workable self-understanding regards itself as being directed at something external to (and independent from) its own stance. It is from agents’ individual and subjective perspectives that they must treat their reasons as if they pointed towards a realm of independently subsisting normative facts that are not entirely conditional on the operations of choice and volition. This practical realism about reasons is unusual because the realist presupposition in question is not primarily born out of any epistemological or metaphysical necessities. Rather, it is a practically vindicated presupposition through and through. The presupposition is justified because it has its ultimate source in the inescapability of deliberation and action and because the task of constituting one’s agency (in the course of action) requires that agents regard themselves as located in a stable and predictable normative environment.[9]

In summary, the first suggestion of how one might go about establishing some common ground between realism and constructivism employs the placebo analogy in order to highlight a certain ambivalence that comes with the status as practical deliberator. Treating reasons realistically is beneficial relative to our needs as agents, and is a perspective that we are justified in shielding from the urge to render ourselves and the world as transparent to ourselves as possible. This antiphilosophical suggestion identifies a tension between the probable ontological truths about reasons, on the one hand, and the needs of agents and practical deliberators, on the other—needs whose satisfaction requires that agents abstain from practically acknowledging such truths in an authentic manner. The practical lives of agents are dependent on managing this ultimately irresolvable tension concerning the objective and subjective sides of practical reasons, the internal and external sources of normativity, and the realist and constructivist natures of those aspects of the world that are normatively relevant to these lives.

Now the placebo analogy allows me to return to the issue of whether to assign practical principles a foundational or merely a derivative role in an account of practical reasons. Recall that for Raz and other “reasons-fundamentalists” principles play a subservient role relative to practical reasons, whereas for many constructivists these reasons are inconceivable unless practical principles guide processes of practical reflection regarding the inclinations that humans encounter. It is at the level of those very principles that the placebo proposal highlights another, more specific, practical demand in line with its core claim that agents have to deliberate and choose as if they responded to independent reasons qua normative facts.

This way of looking at things gets me closer to one of Kant’s (and Kantians’) central theses, which is that willing an action’s end, conferring value on it, cannot take on just any form whatsoever. Willing has to be rational willing and as such it has to satisfy the minimum requirement of being of a universal form. The central idea, manifested in the most abstract version of the categorical imperative, is that the resultant practical reasons have to satisfy the requirement of having implications that go beyond the currently considered single instance of action. In order to exist as a unified person across time, an agent must conceive of himself or herself as employing principles that express the commitment that, at a different point in time, he or she would choose in the same way, were the circumstances similar enough. That’s the argument against the possibility of “particularistic willing” (Korsgaard, 2009, pp. 72-76). If, contrary to this Kantian picture, particularistic willing were an option, “it would be possible to have a reason that applies only to the case before you, and has no implications for any other case” (ibid., pp. 72-73). And, “if you have a particularistic will, you are not one person, but a series, a mere heap, of unrelated impulses. […] Particularistic willing lacks a subject, a person who is the cause of his actions. So particularistic willing isn’t willing at all” (ibid., p. 76). The Kantians’ ambitious conclusion is that this minimally universal character of all practical reasons leads straightforwardly to robust principles of morality—the moral law, the formula of humanity, and the kingdom of ends.

While these ambitious conclusions of Kantian constructivism are not endorsed here at all, there is an important insight expressed in this critique of radical particularism.[10] In order to be an agent, the principle-structured deliberative perspective, which each agent has to sustain in one form or another, has to consist of action-guiding principles that transform the incentives that the agent encounters into reasons and which do so in a temporally stabilizing manner. The placebo analogy adds that the principles that make up one’s practical standpoint have to exhibit more than just this minimally universal form. They have to align agents in a certain direction—namely, one that conceives of the chosen objects as embedded in a stable and predictable external environment, an indispensable part of which is a coherent and non-arbitrary setting of normative facts. This source of stability has to be conceived as having a foothold in an agency-independent realm, a realm that must ultimately be presumed because the stability and coherence of one’s own principle-structured and temporally-extended identity depend on it.[11]

The Korsgaardian argument against particularistic willing, therefore, gets it right, but it is not the end of the story. It must be supplemented by a thesis against what one might call “purely constructivist willing.” Such constructivist willing would be one that is incoherent in the same way that particularistic willing is. If it were possible, purely constructivist deliberation about proposed courses of action would amount to willing that takes itself to be entirely detached from any guidance by fact- and world-directed principles of choice and action. A workable self-conception and practical identity must avoid what has been called, in a different context, “frictionless spinning in the void” (McDowell, 1996, p. 50). The purely constructivist and self-certifying deliberations of a fully self-transparent and self-aware Korsgaardian agent[12] would, so the suggestion goes, amount to such a spinning in the normative void that undermines the constitution of individual agency as a phenomenon in the world and in the company of other, independent, agents.[13]

5. THE ROLE OF REASONS IN ACTION: DELIBERATION AND MOTIVATION

With respect to the issue of the motivational force of reasons for action, internalist conceptions (Humean as well as Kantian[14]) seem to be in a much better position than realist and externalist varieties à la Raz and Scanlon. Recall Korsgaard’s account sketched above: reasons for action motivate agents exactly because they emerge in the course of reflective and deliberative processes that inevitably matter to them. Agents have to decide what to do—in one way or another. And engaging their principle-structured practical identities in order to figure out whether or not to endorse any of the suggested courses of action is non-optionally exerting efficacious effects on their volitional apparatus. The inclinations that agents reflectively endorse on the basis of their identity-constituting principles are the ultimate bedrock of normative motivation. Practical reasons normatively motivate agents because of the way in which their genesis is always connected to features of their volitional experience that have the relevant potential to get them moving—to use constructivist language. Importantly, these active and quite demanding processes of practical deliberation seem to precede the execution of the actions in question. Practical reasons have much in common with causes (regarding their temporal ordering, in particular). Agents encounter a powerful incentive and the related inclination; they activate choice constraining and identity-defining principles and the resulting reason in support of the proposed action normatively motivates these agents to engage in it (in case the principles deem the proposed action worth choosing and executing). It seems as if this picture of how reasons motivate agents easily overcomes the problems that Raz’s externalist separation between reason(s) and the will leaves us with.

There are, however, problems with the Kantian/Korsgaardian account of the relationship between reasons and motivation. I should consider one that is suggested by some of Raz’s remarks and by his critique of motivational internalism. This objection has to do with the so-called “guiding problem” that Raz (1999, pp. 230-238) discusses in his evaluation of Dancy’s particularism. The relevance of the guiding problem for the discussion of motivation becomes visible when one looks at Raz’s formulation of it. He says, “the issue turns on whether there is more to people’s reasons than those factors which figure in their deliberations, and which they cite as reasons when asked, and on how much more is part of their reasons.”[15] Raz presents cases that he categorizes as “intentional actions for a reason,” but which are performed “semi-automatically,” “swiftly,” “opaquely,” etc. These cases seem not to sit well with the constructivist account of practical reasons, where properly reasons-driven action presupposes a demanding mental activity, in which, among other things, one’s proposed actions are examined concerning their fittingness and coherence with the agent’s principle-structured identity. Can there, according to this picture, ever occur cases “in which the reasons we act for are not the reasons which figure in our deliberations” (ibid., p. 231)?

There can’t, pace Raz. Also with respect to the cases that Raz presents, it remains the case that, when it comes to the reasons for which agents act, deliberative endorsement and the mobilization of one’s principles are necessary features. In order to defend this position, I need to introduce another suggestion that will assign a role to practical principles that is more foundational than even many constructivists have considered necessary. I consider a wider formulation of the role that one’s self-conception and its principles play in the process of practical reasoning and action. The question then for the remainder of this paper is the extent to which a constructivist account can accommodate the instances of acts that Raz describes not merely as intentional actions, but as intentional actions done for reasons. I introduce two options that constructivism can put forward in order to accommodate Raz’s examples.

Korsgaard herself qualifies the role that deliberation plays in her account of acting for a reason: “In claiming that such [practical] identities are the sources of our reasons, I am not claiming that thoughts about your identity need to come into your reasoning in any explicit way” (Korsgaard, 2009, p. 21). This qualification is noteworthy because it constitutes quite a significant accommodation of critics, such as Raz, who downplay the role of principles and deliberation in sufficiently accounting for reasons-based action. The problem is that Korsgaard further develops neither the qualification itself nor the implications that such a substantial concession has for her overall view. Given the dramatic and crucial function that actual deliberation in terms of principles has in the identity-constitution framework, one needs to elaborate more on her claim that even actions that do not invoke an agent’s normative self-conception “in any explicit way” nevertheless count as intentional actions done for reasons. There is a danger here of conceding too much.

And it is at this point that the first suggestion should be introduced: what I call the “indirect role of practical principles” is an alternative approach to render the overly intellectualist and rationalist picture of practical reasons and action more plausible. The apparent implausibility of this picture makes Raz’s examples of automatic, swift, and opaque actions attractive. In response to Raz (and Korsgaard’s weakening of her own claim) the first suggestion is that, even in the case of “semi-automatically, swiftly, and opaquely” chosen and performed actions, an agent’s practical identity is (though indirectly) a necessary component in bringing it about as an action for a reason. A fairly trivial semi-automatic action, such as heading to the fridge to get a drink because you’re thirsty, while not preceded by any explicit processes of practical deliberation, nevertheless remains controlled and checked by the confines set by one’s identity-constituting norms and principles.[16]

One way to look at this indirect function of agential identities is to observe how they not merely include positive commitments, aims, and projects, but are to a significant extent best conceived as negative principles that function like red lines circumscribing an agent’s options and actions. And these principles figure in the actions that Raz presents as supposed counterexamples to the constructivist requirement that all actions are principle-guided.[17] This idea of assigning principles a much wider role in an agent’s practical life actually helps to see that constructivists and realists are not that far apart from each other. Since they always set the background limits on our actions, practical principles define and circumscribe “our appreciation of our situation in the world (an appreciation of ourselves as well as of our environment)” (Raz, 1999, p. 231). At the same time, there is no need for constructivists like Korsgaard to push any of these apparent concessions too far when it comes to the three types of cases of actions that Raz presents. Even if deliberative principles are not invoked explicitly and in the form of deploying them in self-aware practical deliberation, they remain a necessary background condition for these borderline cases of action to count as done for reasons. In the case of Raz’s examples of automatic actions (the fridge example, washing your hands before having a meal), my practical principles of etiquette ensure that I don’t sit down at the dinner table with dirty hands. The principles are in play and co-constitute the reason for washing my hands. Furthermore, in the case of opaque actions, such as the case of Raz’s agent who decides to accept a job offer seemingly intuitively,[18] deliberative principles, weighing the pros and cons of the considerations involved, co-inform the decision made. Lastly, and this example will occupy us the most in the remainder of our discussion, the swift action of rescuing a toddler from stepping into the busy road is circumscribed by moral and morally neutral normative principles that are part and parcel of my perception of the situation. The option of not taking the rescuing action is checked by principles that stand ready to jump in should my reflexive response to the situation make me hesitate to take the action in question. Again, this is probably the most controversial case and I will return to it in a couple of paragraphs.

Still, there remain problems in connection with Raz’s examples for the constructivist, even if the reader accepts my argument from the indirect role of practical principles. Raz might rightly insist that there is a difference between deliberative principles passively standing ready in counterfactual scenarios, on the one hand, and actual deliberation in accordance and from these principles, on the other.[19] This problem is also highlighted by the Korsgaardian slogan “action is self-constitution” (Korsgaard, 2009, p. 25). The self-constituting function is not merely an accidental feature of actions, but an essential one. Unless an act done for an end constitutes the person who performs it into its author, it is not an action. How then can actions constitute the agent who performs them when so many actions are done semi-automatically, swiftly, and opaquely? When it comes to the identity-constituting function of actions, one surely can’t rest content with the merely indirect role of practical principles. Okay, Raz might grant, you have shown that even swift and opaque actions, in order to count as reasons-based, must somehow incorporate an appeal to an agent’s practical principles—namely, in an indirect and passive mode. But constructivists (and especially constitutivists) want more: they claim that choice and action from those principles is what constitutes the individual agent into the unique individual that he or she is. And this seems to require some kind of active deliberation. Moreover, in order for behaviour to count as actions done for reasons, the constructivist must show that these actions fully satisfy this demanding standard and not merely incorporate deliberative principles in an indirect and passive manner.

My second suggestion is supposed to explain why Raz’s borderline cases deserve to be labelled as actions done for reasons only when at some point they actively engage the agent’s deliberative stance. However, for me to do so requires a bit of controversial action-theoretical work. The controversial bit of the suggestion consists in the possibility of an incorporation of reasons-based actions into one’s practical identity after the outward manifestation of the action has already occurred. What is, for example, a reflexive action at one point (swiftly rescuing the toddler) becomes a full-blooded, reasons-based action only at a different point in time and only once the agent employs deliberative principles in the action’s conceptualization. By allowing this alternative view of how reasons-based actions come about, I can preserve the constructivist claim that it is a defining feature of such action that it does depend on active practical deliberation in terms of principles.

I am well aware that this approach to self-constituting action seems to violate the view that reasons, similar to causes, always precede action. In order to render the proposal more plausible, let us take a cue from Hegel’s remarks on the nature of action.[20] Hegel claims that an action is not necessarily completed at the time a specific empirical event that was initiated by an agent ends. An agent’s intentions fully reveal themselves to others, but also to the agent himself or herself, only upon completion of an action, and this is a temporally extended process that reaches beyond the time span in which the empirical manifestation of the action takes place. Robert Pippin summarizes Hegel’s view in the following way: “The deed and the reception and reaction to it are considered a constitutive element of the deed, of what fixes ultimately what was done and what turned out to be a subject’s intention” (Pippin, 2008, p. 152). In a similar vein, my second suggestion is that an action’s self-constituting power is not sufficiently accounted for when one ignores how it reverberates, so to speak, into the agent’s continuing incorporation of it into his or her normative self-conception. In its self-constituting mode, the action’s job is not done at a particular point in time, but continues to develop its significance in the course of practical deliberation and reflection—with oneself and in community with other agents. And with respect to the examples that Raz introduced, this suggests that it is not necessary for an “action that is done for a reason” to be performed in the deliberative and reflective mode at the point in time when it manifests itself in the world initially. What is necessary, however, is that this mode be activated at some point, with respect to the action in question.

Take again Raz’s swift action of reaching out to the toddler who is about to step into a busy road (Raz, 1999, pp. 232-233). Raz has plausibly pointed out that “you do not stop to ask yourself what you should do. You react to the situation as you see it” (ibid., p. 233). As already emphasized, he crucially adds that the agent in question does act for reasons, even if that agent does not have them in front of him or her at the time of initiating the action. Raz focuses on the forward-looking role of practical deliberation and concludes, as he does in the other cases mentioned: practical deliberation does not necessarily need to precede instances of actions for reasons. Firstly, he reminds us that the action in question is both intentional and done for reasons (“[reasons] which are available to them [the swiftly-acting agents] if asked at the time” (ibid.)). Secondly, the action is not preceded by any form of conscious deliberation and self-constituting activities. Hence, actions for reasons do not require deliberation and principles.

I would like to suggest a different account of what is going on in the toddler case. It is not the case that Raz’s account of the example is incorrect, but it’s simply not the whole story needed to make sense of the role of reasons. According to my conception of what acting for reasons amounts to, practical deliberation is very much a necessary post- (rather than pre-) condition of intelligible and reasons-based action, worth that demanding label. As emphasized above, acting for a reason is often considered to be a phenomenon incorporating a specific temporal structure, in which reasons are conceptualized analogously to “causes.” And since causes precede the effects that they bring about, it is only natural that Raz and others think that the examples of swift and semi-automatic actions are sufficient evidence for the idea that the activity of practical deliberation can be conceptually separated from counting someone as an agent who acts for a reason. If, however, the Hegel-inspired account suggests that the action of stopping the toddler is not, in all relevant senses, completed before the rescuer practically deliberates about it and thereby brings the reason for it into existence, I end up with a view about practical reasons that is not exclusively modelled within the causalist paradigm.

Let me clarify. I certainly do not want to endorse some implausible story about backward causation here (and neither does Hegel)—I am not arguing that the reason for rescuing the toddler, once it emerges in the process of continued self-constituting action, now suddenly becomes the psychologically efficacious cause that brought the rescue about. Rather, the picture is the following: action, according to the alternative view that differs from Raz’s, in order to be action done for a reason is after all always connected with practical reasoning and deliberation that employ principles of choice and action. However—and this is the controversial move—this picture works only when, firstly, actions are regarded as having a temporally extended structure and, secondly, reasons for them can present themselves in a mode that contradicts the idea that practical reasons always work analogously to causes, especially concerning their temporal ordering.

I think, then, that the story about action qua self-constitution should be read in the following way. Engagement in practical deliberation, choice, and action does not mean that agents always engage in complex thought processes before they engage in actions that constitute them into their authors; that would indeed be an implausible and empirically unsustainable claim, and Raz rightly highlights cases of actions that, at the moment of performance, are done without being engaged in any processes of practical deliberation. However, these deliberative processes have to occur at some point in order for the actions in question to satisfy their self-constituting function and, importantly, to count as done for a reason. The rescuer who continues to fail to reflectively employ any principles (and, hence, generate reasons) that had led him or her to reach out and stop the child wouldn’t count as constituting himself or herself successfully. And, contrary to Raz, that agent would not count as having performed an action based on reasons compared to an agent who does choose and act under a specific normative self-conception that incorporates that action. That this choice and genesis of reasons take place in the form of what appears to be a retroactive process need not be considered a problem once one gives up on an exclusively causalist framework. Reasons for action are not necessarily reasons before action (though very often they are). The presence and emergence of reasons for action continue to be dependent on first-personal and principle-structured deliberation. And this is true of all actions that are done for practical reasons.

There remain many issues. The first one leads us back to the question of how reasons are connected with the phenomenon of normative motivation. After all, earlier in this section, I highlighted that the constructivist paradigm, due to its internalist commitments, presents a fairly convincing account of why practical reasons do not leave agents cold. The argument from how our normative self-conceptions are implicated in the processes of practical deliberation and choice was the main part of this account. It is, however, also with respect to the issue of practical reasons’ motivational efficacy that the Hegel-inspired view about the temporal nature of practical reasons seems to run into a significant problem. After all, I certainly do not want that view to commit itself to the implausible claim that the motivational aspect of reasons can be created retroactively and after the action had been taken. The causalist intuition seems strongest at this point: the processes of practical deliberation that result in the respective reasons must be completed before the action and must manifest themselves in a resolve to engage in the action in question. In a sense, then, the Hegelian view intended as a charitable addition to the constructivist approach seems to go too far when it claims that engaging in practical deliberation is indeed always a precondition for reasons to emerge. What about the toddler case, for example? How can my version of internalism make sense of this and of the other cases that Raz presents as supposed counterexamples to the view that practical reasons are inherently connected to processes of deliberation? If it is not a reason that motivates the rescuer to perform his or her action, what is it?

The solution to that puzzle insists that even the motivating function of practical reasons is conditional on agents engaging in practical deliberation about what they do. The person stepping into the busy road to rescue the toddler is not moved and motivated to do so by a normative reason; rather, it is an immediate inclination to perform the rescuing act that does the motivating job in Raz’s cases. The inclination must be incorporable into a practical reason, though. The proposed amendment to the internalist account can stick to its guns by insisting that, also with respect to the motivational dimensions of the person rescuing the toddler, it is incorrect to refer to a supposed motivating reason being in play at the moment the rescue takes place. For the motivating inclination to turn into a motivating reason, it is again necessary that a process of practical deliberation take place. Even with respect to normative motivation, this deliberation might take place after the rescuing act was performed. That bullet I am willing to bite. I can do so because inclinations take over the role of psychological efficacious causes in Raz’s swift, opaque, etc. actions. Contrary to what Raz says, rescuing the toddler is not, at least not at all times of the action’s genesis, an action done for a reason; and it is not, at least not throughout, an action motivated by the appreciation of some reason in favour of it. Also, motivating reasons must be the result of the deliberative deployment of one’s identity-constituting principles that allow agents to endorse the inclinations with which they are confronted, and which result in the action in question.[21]

One last objection that I want to briefly address will probably emerge in response to the claim that, in cases of swift and opaque actions, no reasons, neither normative nor motivating, are present, unless principle-guided deliberation has been engaged in at some point. What about the accountability and responsibility practices of praising and blaming that agents engage in? After all, fellow agents seem to be justified when they praise the person rescuing the toddler for his or her action, aren’t they? And, the objection continues, this shows that agents do, contrary to what I have been claiming in this section, consider the rescuer to act for a reason after all—maybe a subconsciously appreciated one, but one that is there in Raz’s realist mode. And this intuition concerning holding agents responsible and praising them seems to remain stable even if one fully acknowledges the Razian assumption that no practical deliberation whatsoever has preceded the act of rescuing the child. Lastly, and more problematic for my suggestion, most of us would continue to praise the agent even if he or she were never to reflectively generate the reasons in the retroactive deliberation that the Hegelian version of internalism insists upon.

This objection from accountability-practices and attitudes can be sharpened when considering a tragic version of the toddler case. The rescuer gets hit by a bus and dies in the course of successfully pushing aside the toddler. Obviously, in this case, the deliberative engagement that lets the reason for the rescuing action emerge must remain absent, since the rescuer simply did not get the chance to reflectively endorse his or her rescuer-identity and, hence, the inclination that triggered the rescue fails to get transformed into a reflectively endorsed reason. According to my suggestion, the rescuer was directly and unreflectively acting on the basis of inclinations only; no reasons had yet entered the picture. But this claim seems implausible, given the third-personal evaluation of the case. When others praise the rescuer they presume that his or her agency was as fully engaged in the action as it is in the standard cases of action in which agents have the time and the cognitive resources to deliberate about what to do (and have been doing) and generate their reasons for action in a fully conscious manner (and without requiring the controversial Hegelian addition of retroactive endorsement).

I don’t think that the counterintuitive impression that comes with the tragic version of the toddler case suffices to undermine the claim that the rescuer’s action is indeed constituting an instance of “unfinished business” in so far as it lacks reasons. The version of internalism defended here leaves enough room for our accountability practices, even in cases of agents who are not able to engage in the task of retroactive reasons-genesis. (Interesting complications emerge in cases where agents fail to engage in that task for other reasons. Forrest Gump might be a case in which, while not completely absent, reflective endorsement is simplified and single-layered.)

In response to this objection from accountability, I should recall the first suggestion presented in response to Raz’s critique of the importance of active deliberation. According to that suggestion, the rescuer is accountable for his or her praiseworthy action when that agent’s agency-constituting principles at least passively and indirectly formed the inclinations and incentives that were the origins of his or her action. So, when we hold the rescuer accountable, at the very least we praise these normative practices and, more importantly, the kind of practical agent that these principles constitute—not the bare inclinations. Still—and this must be repeated and insisted upon—we cannot hold the rescuer accountable on the grounds that he or she rescued the child on the basis of a proper normative and motivating reason. That the unfortunate rescuer fails to actively endorse the action-triggering inclination (that he or she forcefully experiences but in a non-reason-like manner) remains a certain deficit also with respect to the extent to which one can praise that agent.

Yet another strategy out of the current problem, which cannot be pursued here, is to reconsider the notions of incentive and inclination. It might turn out that their principle-structured identities not only always at least passively constrain what incentives and inclinations agents act upon, but also shape what they perceive as incentives and how they experience their inclinations to begin with. If that were the case, then the rescuer in the tragic version of the thought experiment turns out to be worth other’s praise because his or her incentives and inclinations had been of a certain, normatively loaded, perceptual quality. Again, however, this quality must ultimately be the expression of the agent being of a certain kind—namely, one who actively endorsed good- and brave-making practical principles at some point. This explanation allows practical principles to take on yet another role in addition to their reasons-generating function—namely, in terms of principle-guided Aristotelian habituation. Obviously, these are rich issues that have to be further developed at another occasion.

6. CONCLUDING REMARKS

This paper’s starting point was the idea to reconcile two conceptions of practical reasons that are widely discussed and considered to disagree on many of the fundamental issues in ethical theory and the theory of action. The ultimate outcome of this attempt is probably something entirely different—namely, an account of practical reasons that both Razian realists, on the one hand, and neo-Kantian constructivists, on the other, will be very critical of. The resulting account draws from both paradigms and combines the diverse elements with the aim of avoiding certain inconsistencies and problems that the original two strands of theories were struggling with.

I presented two main lines of argument to illustrate this alternative conception of practical reasons: the placebo analogy and a deliberation-focused account of how normative and even motivating reasons are generated. The placebo analogy’s main claim was that agents have to (deliberately and deliberatively) uphold an inherently paradoxical conception of the relationship between themselves and their reasons. The apparent paradox emerges upon examining the presuppositions of the processes of efficacious practical deliberation and action. Reasons exist and they are there. However, they exist only because agents with their normative self-conceptions bring them about in processes of practical deliberation (I presumed that Korsgaard has presented a strong case for this throughout my argument). However, the same agents can engage in these processes in a way that is efficacious enough for them to successfully take up their roles as practical agents only when part of their normative self-conception comes with the (principle-structured) stance that negates this exclusively constructivist, agency-dependent, character of reasons. As agents, we are forced to maintain this stance on practical grounds, exactly in order to coherently uphold anything before ourselves. We are practically justified in not rendering the nature of the relationship between ourselves and our reasons fully transparent.

The second claim was introduced to deal with cases of intentional actions that seem to involve agents who do not at all engage in these principle-governed processes of identity-constituting deliberation, which my view considers to be a necessary condition for reasons to be part of an action. In response to Raz’s examples of swift, opaque, and automatic actions, I have insisted that reasons are not yet present, neither in their normative nor in their motivating varieties, because practical deliberation had not been engaged in. I have presented a noncausalist and Hegel-inspired alternative account of reasons-based action to show that reasons for a particular action do not always and necessarily precede the action that they co-constitute. If that account works, then it can successfully support the view that all practical reasons are the result of quite demanding deliberative processes in which an agent mobilizes the normative resources that he or she sees him- or herself committed to. This claim is certainly controversial regarding the normative features of our reasons, but has been especially difficult to defend in the case of reasons’ role as motivating features of our agency. The hope is that the two suggestions taken together provide some support for this alternative conception of practical reasons and that it can be further developed into a view that overcomes a set of related controversies in contemporary ethical theory.