From Transcendental Freedom to the Other: Levinas and Kant

© Texas Tech University Press, 2001
In Proximity: Emmanuel Levinas and the l8th Century
, Eds. Melvyn New, Robert Bernasconi, and Richard A. Cohen, pp. 327-54.

The principle of mutual love admonishes men constantly to come closer to one another; that of the respect they owe one another, to keep themselves at a distance. [1] Immanuel Kant, The Metaphysics of Morals

That to which we are alluding here seems to us suggested by the practical philosophy of Kant, to which we feel particularly close.
Emmanuel Levinas, “Is Ontology Fundamental?” [2]

In his only essay dedicated entirely to Kant, published in 1971 under the title “The Primacy of Pure Practical Reason,” Levinas applauded the “great novelty” [3] of Kant’s practical philosophy and its unique contribution to the eighteenth-century Enlightenment. Seven years later, Levinas compiled a list of exceptional moments in the history of philosophy when “under different terms [the] relation of transcendence shows itself.” [4] Alongside references to Plato, Aristotle, Descartes, Hegel, Bergson, and Heidegger, the list included “the elevation of theoretical reason into practical reason in Kant.” [5] Six years after that, Levinas referred to the doctrine of primacy not merely as an exception to the tradition of philosophy as ontology, but as an exceptional exception, the critical inception of a “new intrigue” which “no longer amounts to bringing to light presence.” [6] Evidently, Levinas found in Kant’s practical philosophy something exceptionally close to his own ethical thinking, so much so that one is naturally led to wonder why so few of Levinas’s commentators have been disposed to remark on the fact. [7]

How does Kant’s doctrine of primacy, according to Levinas, evince a break with Western philosophy characterized by ontology, freedom and the “primacy of the same” (TI 16/45)? How does it go beyond ontology when it sanctions, as we will see, the prerogative of morality to assume as “real” (i.e., existing) what critical speculation alone designates as merely possible? In what way does it overturn the traditional priority of freedom and the Same given Kant’s insistence that reason is autonomy and that “in the end there can only be one and the same reason, which must be differentiated solely in its application” [8] ? These and related issues will be addressed in the following three-part discussion. By showing how Kant’s doctrine challenges rather than reconfirms traditional ontology, I hope to provide a clearer understanding of Levinas’s claim that the practical philosophy of Kant stands closest to his own thinking in ethics.

1. The Primacy of Practical Reason

In the Critique of Pure Reason, II. ii. 3, entitled “On the Primacy of Pure Practical Reason in its association with Speculative Reason,” Kant writes:

in the combination of pure speculative with pure practical reason in one cognition, the latter has primacy, provided that this combination is not contingent and arbitrary, but a priori, based on reason itself and thus necessary. Without this subordination [Unterordnung], a conflict in reason with itself would arise, since if the speculative and the practical reason were merely arranged side by side (co-ordinated [koordiniert]), the first would [critically] close its borders and admit nothing from the latter, while the latter would extend its boundaries to everything and, when its needs required, would seek to comprehend the former within them. [9]

The interest of pure practical reason, according to Kant, “lies in the determination of the will with respect to the final and perfect end” (CPrR 120/124). This interest is advanced by assuming that the Ideas of freedom, God, and the immortality of the soul have “objective reality” (3/3). Without freedom--the ratio essendi of the Moral Law--pure practical reason would be impossible; without the other so-called postulates, the concept of the highest good (summum bonum) would find itself in an antinomy, since virtue and happiness are only contingently related from the standpoint of nature, and virtue is seldom attainable in a single lifetime. The doctrine of primacy circumvents any possible “conflict” (Wiederstreit) of interests between theoretical and practical reason. It rules a priori that theoretical reason go beyond the critical agnosticism of the Critique of Pure Reason and make room for the “assertoric” judgment that the transcendental Ideas have objects.

Such a description of the priority of practical reason over theoretical reason is meaningful only on the assumption that the interest in the summum bonum is independent of any speculative interest in architectonic completeness. Few commentators have ever granted this assumption. [10] Even Kant himself appears to have had his doubts. Near the end of the first Critique he wrote: “I maintain that just as the moral principles are necessary according to reason in its practical employment, it is in the view of reason, in the field of its theoretical employment, no less necessary to assume that everyone has ground to hope for happiness in the measure in which he has rendered himself by his conduct worthy of it.” [11] Clearly, the interest in the highest good is here assigned to theoretical reason. We are to assume the existence of God and of the soul’s immortality because of the natural or subjective interest that human reason has in representing to itself a complete system of ends united by the concept of the highest good. Like the transcendental Ideas in the first Critique (“soul,” “God,” and “world”), the Idea of the summum bonum is regulative. It directs the understanding toward answering the question “What is to result from this right conduct of ours?” [12] and, in so doing, brings theoretical closure to the practically conditioned.

It may perhaps be thought that since the practical interest in the Ideas of God and immortality of the soul does not “conflict” with the interests of theoretical reason, and even harmonizes with those interests, the primacy that Kant attaches to practical reason over theoretical reason is altogether misplaced. “If practical reason may not assume and think as given anything further [weiter] than what speculative reason affords from its own insight, the latter has primacy” (CPrR 120/124-5) (emphasis added). [13] If this is true, and there really is no non-speculative relationship between reason and the transcendental Ideas, then Levinas’s enthusiasm for the Kantian doctrine would appear embarrassingly misinformed, and his sweeping assertion that “All Kant’s work presents itself as the primacy of pure practical reason” [14] grossly exaggerated.

In “Transcendence and Evil” (1978), Levinas makes it clear that whatever the importance the resolution of the antinomy of practical reason in terms of Kant’s overall system, it constitutes a concession to speculative reason and undermines the radical implications of Kant’s moral theory: “after a moment of separation, the relationship with ontology is reestablished in the ‘postulates of pure [practical] reason.’” [15] According to Levinas, it is not belief in the existence of God and of the soul that effects the break with ontology, but the practical interest in the Idea of transcendental freedom:

across these returns of ontology, Kant was bold enough to formulate a more radical distinction between thought and knowing. He discovers a plot in the practical usage of pure reason not reducible to being. A good will, as it were utopian, deaf to the information, indifferent to confirmations, that could come to it from being (which are important for technique and for the hypothetical imperative, but do not concern practice nor the categorical imperative), proceeds from a freedom which is situated above being and on this side of knowing and ignorance. [16] (my emphasis)

Let us attempt to clarify the importance of the Idea of transcendental freedom for practical reason. The Idea was introduced in the first Critique as a way of avoiding any incompleteness in theoretical reason: “Reason showed freedom to be conceivable only in order that its supposed impossibility might not endanger reason’s very being and plunge it into an abyss of skepticism (CPrR 3/3). [17] Theoretical reason pushes inquiry to its furthest limits by seeking the unconditioned. Without the concept of a cause that is independent of the phenomenal mechanism of nature, that is, “freedom in its strictest, i.e. transcendental, sense” (29/28), this inquiry would never end. Reason would be caught in an infinite regress as it moved from the conditioned back to its condition and back to its condition in turn. Practical reason, however, goes an important step further. It demands that freedom not only be regarded as thinkable qua “the purely transcendental and to us unknown ground of the possibility of the sensible series [appearances] in general” (CPR 564/592), but also upheld without proof as true. That demand is quite different from the alleged “practical”—actually theoretical--need to assume that the Ideas of God and the soul have objective reality. Transcendental freedom is indeed a “postulate,” but not one that reason is compelled to assert in order to resolve the antinomy of practical reason. It is not an endpoint that reason must attain but rather an assumption that reason must make to be practical: “I maintain that to every rational being possessed of a will we must also lend the Idea of freedom as the only one under which he can act” (G 448/108-9). There is no mention here of the highest good and the other postulates. The presupposition that freedom exists is indispensable for practical purposes, and thus required by morality itself.

Again, it might be objected that this presupposition is not practically necessary at all, but instead answers a merely theoretical need to the extent that looking upon ourselves as self-legislating members of an intelligible realm is “only a point of view which reason finds itself constrained to adopt outside appearances in order to conceive of itself as practical” (G 458/118). The objection has been made by Nathan Rotenstreich, who argues that we are bound to adopt such a viewpoint in every case where we use reason, not just practical reason: “Kant does not need to assume the primacy of action, or of practical knowledge over theoretical knowledge, in order to prove the possibility and the reality of freedom. Freedom is warranted by the very presupposition of the discussion, by the very activity of philosophizing” (my emphasis). [18]

However, Rotenstreich appears to have misconstrued Kant’s transcendental deduction of freedom. He is, of course, right to argue that freedom is not the exclusive property of practical reason, and is presupposed by reason in general. Freedom and reason are reverse sides of the same coin. But this is not grounds for concluding that “reason is its own guarantee and need not take refuge in the dubious assumption of the Primacy of Practical Reason.” [19] As Kant repeatedly asserts in the second Critique, it is only the Moral Law qua “fact of reason” (Faktum der Vernunft) that is capable of establishing the legitimacy of the Idea of freedom. [20] Theoretical reason is altogether insufficient here. Freedom may be a condition that reasoners in the “discussion” presuppose as having been realized, but, as Habermas would argue, it is also a practical goal to attain. For the same reason that Rotenstreich is correct to assert that it “cannot be taken for granted that [immortality, happiness, reward, and God] are not related to a given historical situation and to a particular set of convictions, beliefs, expectations, dogmas, traditions, etc,” [21] it is implausible to assume that any theoretical reasoning is free of non-discursive and “irrational” interests. The opposite is true in the case of the use of pure practical reason, which by definition excludes what Kant calls the “private use of reason,” that is, reason in the service of self-interest or some other self-styled “authority” such as the church or state. Indeed, it is because “the moral law expresses nothing else than the autonomy of pure practical reason” (CPrR 33/33-4]), being none other than the “supreme principle” (G 409/73) of following principles that can be shared by all, that Levinas is justified in making the seemingly hyberbolic claim that (Kantian) “practice . . . is the basis of the logos.” [22] Pure practical reason is basic because, as Kant says in the Groundwork, “reason must look upon itself as the author of its own principles independently of alien influences” (G 448/109). The Moral Law thus supplies not only the rule for autonomy but also grounds for believing that it exists (though of course it might not).

By making the Moral Law “the principle for the deduction of an inscrutable faculty: freedom” (CPrR 49/48-9), Kant has done two things necessary to assure the primacy of morals. First, he has established an independent interest on the part of morality in the Idea of freedom. Theoretical reason indeed demands that freedom be thinkable in its attempt to think the unconditioned in a causal series and attain completeness, but it is morality that provides the only rational interest in asserting that it has objective reality. Second, he has shown that the moral law is not only the supreme principle of the practical use of reason, but also of reason in its theoretical capacity, since “we cannot possibly conceive of a reason as being consciously directed from the outside in regard to its judgments” (G 448/109). To be sure, an appeal to morality is not a theoretical proof of freedom, which would constitute an extension of knowledge and breach the critical limits defined by the Copernican Revolution. It is, however, a vindication to the extent that freedom is an assumption we are required to make if we are to consider ourselves under moral obligation, and not merely compelled by whim. No other type of interest is capable of legitimizing such an assumption (“no one would dare introduce freedom into science [Wissenschaft] had not the moral law and with it practical reason come to the concept and forced it upon us” [CPrR 30/30]). These considerations help us understand why Levinas claims that Kantian good willing “proceeds from a freedom which is situated above being and on this side of knowing and ignorance.” As something that is a merely deduced rather than a datum, a postulate rather than a proof, it exceeds ontology and the comprehension of reality, as Kant understood when in the first Critique he wrote that “transcendental philosophy . . . takes no account of objects that may be given (Ontologia)” (CPR 845/874). [23]

2. Respect and Desire

What is it that makes the will (Wille) good? Kant’s answer to this question is its rationality, its unconditional determination to act in accordance with the Moral Law, a principle of pure practical reason. For Levinas, on the other hand, it is not rationality that makes the will good but responsibility for the Other (“to be for the Other is to be good” [TI 239/261]). This entails the adoption of maxims of nonreciprocal and non-universalizable action (“Goodness consists in taking up a position in being such that the Other counts more than myself” [225/247]), presupposing the spontaneous capacity to act independently of pure practically legislative reason. The importance that Levinas attaches to what Kant in the Groundwork called “lawless” (G 446/107) freedom is not always appreciated by readers whose understanding of Levinas has been largely informed by the chapter in Totality and Infinity entitled “Freedom called into Question.” In “The Rights of Man and the Rights of the Other” (1985), Levinas clearly suggests that spontaneity of the will has a moral connection that goes far beyond the merely “negative” (G 446/107) use that Kant made of it:

Is it so certain that the entire will is practical reason in the Kantian sense? Does the will not contain an incoercible part that the formalism of universality could not oblige? And we might even wonder whether, Kant notwithstanding, that incoercible spontaneity, which bears witness both to the multiplicity of humans and the uniqueness of persons, is already pathology and sensibility and a “bad will.” . . . The universality of the maxim of action according to which the will is assimilated to practical reason may not correspond to the totality of good will. [24] (my emphasis)

Levinas is not saying that a Kantian good will is not good. What he is calling into question is the Kantian claim that a complete account of moral willing can be given in terms of universal law. Implicit in the reference to “incoercible spontaneity,” attesting to the difference between persons, according to Levinas, is the suggestion that moral goodness resides in the capacity of the will to disregard reason no less than in its capacity to follow reason.

That the will could be motivated to act independently of reason altogether would be unintelligible to Kant. In the Groundwork we are told that the will must be determined by some principle, either an a posteriori rule (a hypothetical imperative) serving to promote an end set by natural inclination, or by an a priori law of reason (the Categorical Imperative) which determines the will immediately (G 400/65). A will (Willkür, choice) that could be determined prior to the representation of any rational principle whatsoever would be for Kant a will in name only, a mere arbitrium brutum, not just pathologically affected but “pathologically necessitated” (CPR 534/562). It would not be “bad” but inhuman. [25]

It might seem that Levinas has no alternative but to fall back on some form of Kantian rationality if the good will is not to dissolve into the rest of nature. But this is precisely what Levinas refuses to do. In the chapter in Totality and Infinity entitled “Will and Reason,” he writes: “To distinguish formally will and understanding, will and reason, nowise serves to maintain plurality in being or the unicity of the person if one forthwith decides to consider only the will that adheres to clear ideas or decides only through respect for the universal to be a good will . . “ (TI 192/217). A page later, however, Levinas makes it clear that the analytic distinction does not imply a loss of practical freedom:

if the face . . . [is] the very upsurge of the rational, then the will is distinguished fundamentally from the intelligible, which it must not comprehend and into which it must not disappear, for the intelligibility of this intelligible resides precisely in ethical behavior, that is, in the responsibility to which it invites the will. The will is free to assume this responsibility in whatever sense it likes; it is not free to refuse this responsibility itself (TI 194/218-9).

Levinas describes the face as “intelligible,” but intelligible in such a way that the will “must not comprehend [it] and into which it must not disappear.” The prohibition against understanding is aimed in particular at Kantian idealism and the famous transcendental unity of apperception. Recall the “Transcendental Deduction” in the second edition of the first Critique: “Only insofar . . . as I can unite a manifold of given representations in one consciousness, is it possible for me to represent to myself the identity of the consciousness in these representations” (CPR 133). Irreducible to the synthetic application of concepts, the face has meaning solely in the context of “an ethical behavior, that is, in the responsibility to which it invites the will.” Notice that Levinas speaks of responsibility here as a type of act. It is something that the will can be “invited” to perform. There is no contradiction between this and the claim that the will is “not free to refuse this responsibility.” Levinas is simply making a distinction between being under obligation and fulfilling it. The will is under obligation to the Other without having made any binding contract or promise. Nevertheless, it is “free to assume this responsibility in whatever sense it likes.” Earlier in Totality and Infinity Levinas wrote: “I can recognize the gaze of the stranger, the widow, and the orphan only in giving or refusing; I am free to give or to refuse” (49/77). Again, this should not be taken to suggest that obligation is optional, which would be a contradiction, but only that the will has the negative freedom not to fulfil the obligation that it owes the Other. Indeed, putting it that way is perhaps not so very different from Kant’s claim that transgression of the Moral Law is “freely assumed” (CPrR 100/103), even though the Law itself obligates “unconditionally.”

The preceding observations indicate that Levinas reserves an important place for Kantian spontaneity in his moral theory without making responsibility optional. If, however, responsibility for the Other is not determined by pure practical reason in its autonomous or self-legislating capacity as Wille, how is it different from what Kant calls mere heteronomy, which occurs when the will (Willkür) is motivated to act by natural inclination or some underlying sensuous impulse that is “pathological”? To the extent that the encounter with the Other is immediate (“the immediate is the face to face” [TI 23/52]), by which Levinas means that it is independent of reason (principles) and the understanding (concepts), what stops it from being pathologically determined? The question was raised by Jacques Derrida in his influential essay “Violence and Metaphysics,” published in 1964. Finding resources in Levinas’s work that would justify calling the ethical relation “an immediate respect for the Other himself,” though apparently not “following any literal indication by Levinas,” [26] Derrida made the following note: “An affirmation at once profoundly faithful to Kant (“Respect is applied only to persons”--Practical Reason) and implicitly anti-Kantian, for without the formal element of universality, without the pure order of law, respect for the other, respect and the other no longer escape pathological immediacy. Nevertheless, how do they escape according to Levinas?” [27]   Again, Derrida would have given us an instance of Levinas’s dependence on the logos in his very attempt to go beyond it. The assumption, of course, is that all non-rational motives would be pathological in Kant’s sense, in which case Levinas’s ethics must have recourse to the very reason that it is meant to supersede if the distinction between it and need is to be maintained. [28]

Is this assumption valid? To be sure, Levinas would deny that the ethical relation is pathologically immediate if by that is meant that the will does not just undergo pathological processes but is caused to act by them. But he does not deny that it is sensibly affected: “The role Kant attributed to sensible experience in the domain of the understanding belongs in metaphysics to interhuman relations” (TI 51/79). However, Levinas makes it clear that the affective character of the ethical relation does not preclude what he calls “disinterestedness” (5/35). Levinas’s use of this term is quite different from the use Kant made of it in the third Critique, where it referred to the entirely disinterested satisfaction involved in contemplating objects of beauty. [29] In Totality and Infinity Levinas informs us that “in Desire the being of the I . . . can sacrifice to its Desire its very happiness” (31/63). Kant, of course, would simply argue that “in such a case an action of this kind, however right and amiable it may be, has still no genuinely moral worth. It stands on the same footing [gleichen Paaren gehe] as other inclinations” (G 398/64). But Levinas calls into question this reduction of moral sentiment to mere inclination. In doing so, he doesn’t simply illustrate the limitations of Kant’s famous distinction between “pathological” and “practical love” by providing an exceptional case of a “metaphysical” desire that can be commanded and thus is irreducible to self-love; he gives us reason to believe that Kant himself ultimately fails to maintain the distinction when developing his own views on the moral feeling of “respect” (Achtung).

In an essay ambiguously situated between his confessional and philosophical works, entitled “Have You Reread Baruch?” (1966), Levinas offered an unusually sympathetic reading of Baruch Spinoza’s Tractatus Theologico-Politicus. “Spinoza’s genius,” Levinas wrote, was to have perceived the irreducibly “ethical significance of the Scriptures.” [30] Levinas acknowledged the pre-critical implications of “Spinozist” moral faith, though he went on draw an analogy between it and Kant’s practical philosophy: “As in Kant, this God of faith reflects the demands of practical reason”; “faith is the support of the Scriptures . . . independent of all philosophy while agreeing with the practical consequences of philosophical reason” (“Baruch,” 164-65/115). The analogy, however, is less than perfect, not because Kant himself regarded Spinoza as an atheist and thereby “restricted in his [moral] endeavor,” [31] but because, as we have seen, for Kant “moral belief” is altogether dependent on philosophical reason in both its theoretical and its practical employment. Furthermore, Kant maintains that without the support of pure practical reason faith would remain dogmatic (“historical”) and the Scriptures would lack imperative force:

The possibility of such a command as, “Love God above all and thy neighbor as thyself” agrees very well with this. For as a command it requires respect for a law which orders love and does not leave it to the arbitrary choice to make love a principle. But to love God as inclination (pathological love) is impossible, for He is not an object of the senses. The latter is indeed possible toward men, but it cannot be commanded, for it is not possible for man to love someone merely on command. It is therefore, only practical love which can be understood in that kernel of all laws (CPR 83/86; my emphasis).

It certainly seems, at least initially, that Levinas and Kant are very close here. Both argue that there may be no relation to God separated from the relationship with human beings (cf. TI 51/78), and both make a distinction between ethos and eros in terms of the traditional distinction between logos and pathos. In Totality and Infinity Levinas explicitly says “discourse is not a pathetic confrontation of two beings absenting themselves from things and others. Discourse is not love” (49/76). However, in “Have You Reread Baruch?” he appears to undermine the distinction altogether: “the incentives [mobiles] of obedience are not of a rational order. They are incentives of an affective order, such as fear, hope, fidelity, respect, veneration and love. Obedience and heteronomy, but not servitude . . . Obedience comes not from constraint but from an internal and disinterested élan. Commandment and love do not contradict one another, contrary to Kant” [32] (“Baruch,” 163-64/114-15). Levinas does not explain how it is possible to command love and so it is unclear to what extent he considered the assertion to carry weight in a non-confessional or philosophical context. It might be thought that Levinas is not obliged to provide an argument since that would weaken the distinction he is trying to make between moral faith and reason. However, this does not sit comfortably with the idea that moral faith “can be presented as agreeing with philosophy” (“Baruch,” 167/117). Indeed, the irony in this case is that what may look like a fundamental disagreement between Levinas and Kant as to whether love can be commanded--and Levinas certainly presents it as such--turns out to be illusory when we consider what Kant himself has to say about moral incentive of “respect.”

“Respect always applies to persons only, never to things” (CPrR 76/79). Kant’s assertion from the second Critique is frequently quoted, though it is not always understood by humanists who read it in straightforwardly anthropological terms. Strictly speaking, “respect for a person [is] really for the law, which his example holds before us” (CPrR 78/81). It applies only to Persönlichkeit understood as the predisposition to exemplify pure practical reason (87/90), something that all humans share, but which is not unique to them. [33] Nevertheless, not every “personality” has a predisposition to feel respect, which can be said truly to apply to persons only. [34] In the absence of this moral feeling persons would be incapable of exemplifying the law, which requires an “incentive” (Triebfeder, literally, “spring” or “impulse”) (76/79) on the part of a subject whose will is not naturally in accord with the requirements of duty. Even so, we might be inclined to ask, what it is that allows Kant to be so certain that “Respect is the sole and undoubted moral incentive” (78/81)? Why is respect indubitable and why are not “other” incentives possible? Three pages later in the second Critique Kant asserts “reason through the practical law commands respect pure and simple” (81/84; my translation and emphasis). How is it possible to reconcile this with the claim quoted earlier that “it is not possible for man to love someone merely on command”? Respect in Kant may not be the same as pathological love, but it is a feeling that “is of such a peculiar kind that it seems to be at the command only of reason” (76/79). Can Kant continue to deny that pathological love--or any other non-rational feeling--can be “commanded” when he does not deny that practical love can?

Not that Kant claimed to have anything like insight into respect after the Copernican revolution. In the Groundwork he made it quite clear that it is “wholly impossible to comprehend . . . how a mere thought containing nothing sensible in itself [Moral Law] can bring about a sensation of pleasure and displeasure” (G 460/120). It is impossible to comprehend because to do so would be equivalent to explaining what Kant calls “moral interest,” by which is meant “that in virtue of which reason becomes practical--that is becomes a cause determining the will” (G 460n/120n), tantamount to explaining per impossibile freedom. Here we run up against “the extreme limit of all moral enquiry” (G 122/462). For Levinas, by contrast, the limit of moral inquiry is signaled not by the impossibility of explaining how pure reason can be practical in its autonomous capacity as Wille, but by the impossibility of explaining how “obedience and heteronomy, but not servitude” is possible. By “heteronomy” Levinas does not mean determination of the Willkür by practical principles in the service of inclination. Nor by “obedience” does he mean slavish subjection to the will of another implying the loss of practical (i.e., transcendental) freedom. In Totality and Infinity he writes that “obedience precisely is to be distinguished from an involuntary participation in mysterious designs in which one figures or prefigures” (52/79). But it still does not amount to Kantian autonomy since its motive is not a rational law given by the will to itself but rather un élan désintéressé inspired by the Other.

As indicated, Kant would strenuously deny that the will is capable of being obligated by anything other than reason. Any other conative (and not merely cognitive) element in willing would, according to Kant, be governed by “pathological” motives reducible to self-interest and thus not necessarily valid for all rational beings as such. Levinas too would deny that the motive behind moral action is “pathological,” if by that is meant that it is directed by an object of natural inclination and need. But the Other who is the conative element in willing is said to go “beyond” inclination: “Desire is desire for the absolutely other. Besides the hunger one satisfies, the thirst one quenches, and the senses one allays, metaphysics desires the other beyond satisfactions” (TI 4/34). [35] We should not suppose from such statements that Levinas is indifferent to the compromises and contradictions that Kant saw arise once irrational feeling and desires were permitted to filter into ethics. Indeed, in the early fifties, so preoccupied was Levinas with avoiding any such compromise that he even questioned whether the Moral Law itself was sufficiently equipped to rule out such intrusion. This was the case in an essay appearing in 1953, entitled “Freedom and Command,” in which he attempted to show how the human will can achieve autonomy without being reducible to reason pure and simple. To this end, Levinas reversed the Kantian primacy of autonomy over heteronomy by showing how reason itself presupposes the relation with the Other.

3. Transcendental Argumentation and the Other

“Freedom and Command” opens with a dilemma. How can a will be commanded to act and remain free? Alternatively, how can a command apply to anything other than a free will? From Plato to Kant the dilemma has been resolved by arguing that the command must in advance be in accord with the will of the person commanded. A will can accept the order of another only because it recognizes the rationality of the order. [36] If the will does not see this rationality then it can refuse to obey--to the point of death. The execution of Socrates is a case in point: “Up to the last moment his thinking remains a refusal” (“Freedom,” 265/16). First his feet become insensate by the hemlock, then his legs, and thereafter “as far as his waist,” [37] but no farther; at no point during his execution does he lose his head. But Socrates is a remarkable “character” in the dual sense of the term. “We know,” writes Levinas, that “the possibilities of tyranny are much more extensive. It has unlimited resources at its disposal, those of love and wealth, torture and hunger, silence and rhetoric. It can exterminate in the tyrannized soul even the very capacity to be struck, that is, even the ability to obey on command” (265/16). Levinas is not diminishing the importance of what Plato and the Greeks called “self-control” (auton heauton archein, sophrosyne, enkrateia). [38] His claim is rather that such “inner freedom”--as Kant would call it (MM 408/207-8)--requires further guarantees if it is to escape subjective degradation and be spared the ordeals of tyranny: “We must impose commands on ourselves in order to be free. But it must be an exterior command, not simply a rational law, a categorical imperative defenseless against tyranny; it must be an exterior law, a written law, armed with a force against tyranny. Such commands are the political condition for freedom” (266/17).

Nevertheless, there is no guarantee that human freedom will not experience these statutory laws as another tyranny in turn. “The last will and testament drawn up by a lucid mind can no longer be binding on the testator who has survived” (266/17). Whatever is the cause of the will’s repudiation of impersonal reason and the law, the implication is that, pace Kant, impersonal reason is incapable of generating its own incentive. Levinas explicitly says “the individual act of freedom which decided for impersonal reason does not itself result from impersonal reason” (267/18). [39] However, we need to ask how it is possible from freedom to result from anything other than impersonal reason. Is not autonomy precisely rational self-constraint? Levinas concedes that “no one wants to force another to accept the impersonal reason of the written text, unless out of tyranny. Were it done by persuasion that would already presuppose the prior acceptance of impersonal reason” (267/18).

It is at this point in “Freedom and Command” that Levinas recalls Polemarchus’s famous question at the beginning of The Republic: “Could you persuade men who do not listen?” [40] Glaucon abruptly answered that it was not possible, though the reentry of Thrasymachus into the discussion with Socrates, despite his having made up his mind “merely to nod yes and no, as one does to old wives’ tales” [41] is interpreted by Levinas as Plato’s way of saying that it is possible:

Yet there is a sort of necessity for persuasion in favor of a coherent discourse. It is perhaps this persuasion, this reason prior to reason, that makes coherent discourse and impersonal reason human. Before placing themselves in an impersonal reason, is it not necessary that different freedoms be able freely to understand one another, without this understanding being already present in the midst of that reason? Is there not a speech by which a will for what we call coherent speech is transmitted from freedom to freedom, from individual to individual? Does not impersonal discourse presuppose discourse in the sense of this face-to-face situation? In other words, is there not already between one will and another a relationship of command without tyranny, which is not yet obedience to an impersonal law, but is the indispensable condition for the institution of such a law? Or again, does not the institution of a rational law as a condition for freedom already presuppose a possibility of direct understanding between individuals for the institution of that law? (267/18; my emphasis)

On what does Levinas base his conclusion that rational constraint presupposes moral constraint? Certainly the language of “necessity,” “condition,” and “possibility” harbors the appearance of what since Kant has come to be known as a transcendental argument. As I read it, Levinas is pushing back one stage further the deduction of the conditions of the possibility of rational self-obligation (autonomy) by revealing the condition of what according to Kant is an incomprehensible and unconditioned interest imbedded in reason itself. Nevertheless, Levinas’s deduction can never be strictly “transcendental,” which requires a direct assault on the subjective conditions of possible experience. [42] As already noted, Levinas maintains that the Other is capable of obligating “ethical behavior” prior to the representation of the law, which means there may be no straightforward regression upon the unity of the will as the practical counterpart to the “I think” in theoretical reason. Here it is not a question of bringing the manifold of desires to the synthetic unity via the category of causality, namely freedom, in order to issue in a possible moral judgment. On the contrary, as Levinas will make clear in Totality and Infinity, the movement that leads to the Other, “a deduction--necessary and yet non-analytical,” entails “the break-up of the formal structure of thought . . . into events which this structure dissimulates, but which sustain it” (TI xvii/28). This confounding of thought and ontological conditions is already operative in “Freedom and Command,” though it is not explicitly marked by Levinas.

Levinas’s “transcendental argument” may be formally presented as follows (“Freedom,” 267/18):

A1) Political freedom is possible.

A2) Political freedom presupposes the acceptance of impersonal reason.

A3) If the acceptance of impersonal reason were coerced, then it would undermine the will and political freedom would be illusory.

A4) If the acceptance of impersonal reason were rationally motivated, then this would presuppose the prior acceptance of impersonal reason.

It follows from A2, A3 and A4 that

A5) Political freedom presupposes the acceptance of impersonal reason in a way that is non-coerced and non-rationally motivated.

A6) Only the Other is capable of providing a non-coercive and non-rational incentive for accepting impersonal reason through his or her capacity “to persuade people who do not listen.”

It follows from A1, A5 and A6 that

A7) The condition for the possibility of political freedom is the Other.

The argument has a simplicity about it that is liable to lead the unsuspecting reader into a false sense of ontological security. The problem with arguing in this kind of sequacious fashion is that it dissimulates an irreducible circularity in the account. Where is the circle?

Levinas claims that Other provides the only non-rational and non-coercive incentive for adopting impersonal reason. The Other does so precisely insofar as he or she is able to obligate the will directly through speech. Hence Levinas writes “commanding is speech, or . . . true speech, speech in its essence, is commanding” (23/272). However, the relation with the Other would then appear to presuppose what it is said first to make possible. If we assume that language is never merely private, that the commanding essential to speech is constrained to traffic in universal signs (signifieds/signifiers), discursive rules, regimens of statements, etc., then any priority Levinas can be seen to assign to the face to face as the “indispensable condition” for impersonal reason is vitiated by the intelligibility of the logos itself. Certainly, the priority cannot be straightforwardly chronological or logical. As a relationship that is not merely one of “irrational contact” (271/22), the face to face cannot dispense with reason. This is not equivalent to saying that discourse is reducible to the intelligibility of the logos, even if the irreducibly ethical dimension that Levinas finds within language face to face cannot be separated from ontological categories in the concrete. Indeed, it is in order to distinguish for analytical purposes the ethical and the ontological aspects of language that Levinas will later, most notably in Otherwise Than Being, deploy the terms saying and said respectively (OB 74/58). However, it is due to a critical awareness of their irrevocable conjunction that Levinas will also assert elsewhere that “There is, it is true, no Saying that is not the Saying of a Said.” [43]

Without the sophisticated vocabulary characterizing the later work, combined with the absence of a thematic presenting the limitations of transcendental philosophy in regard to the Other, “Freedom and Command” lacks the necessary apparatus for explicitly marking the self-undermining of the argument as a function of the “otherwise than being” as distinct from what might be called a “mere” contradiction. Hence, when Levinas claims that “there is a sort of necessity for persuasion in favor of a coherent discourse” (“il y a comme une nécessité de persuasion en faveur d’un discours coherent”)--which he goes on simply to call “persuasion” (“Freedom,” 267/18)--he appears to undermine his statement a couple of lines earlier where he denied that the acceptance of impersonal reason were one of “persuasion” on the grounds that it would be question begging: “Were it done by persuasion, that would already presuppose the prior acceptance of impersonal reason.” It is in apparently self-undermining fashion that he refers to the Other’s command as “reason prior to reason” and “a discourse before discourse” (267/18). What is a speech prior to logos?

The point, however, is not that Levinas offers a circular argument. The antilogy of the account signals the transcendence of the Other, not the failure of reason. In “Freedom and Command” it derives from the fact that “any one who reaches [the face] has already denied that every past must have been present” (271/22). What Levinas means by this seemingly contradictory expression is that the Other commands without having ever been present to consciousness at the time of the commanding, notwithstanding that it is always possible for consciousness “after the event” to regress upon a condition for that commanding-conditioning itself, in this instance the intelligibility of discourse. Here we have, I suggest, a precursor of the peculiar structure of the “anterior posteriori” discussed in Totality and Infinity in connection with the analysis of separated dwelling (where representational consciousness is said to rest on what it constitutes), and Descartes’s third “Meditation” (in which the apparent isolated certainty of the cogito is found to rest on the “earlier” idea of Infinity). It is the same structure that is explored at length in Otherwise Than Being under the name of “trace.”

In “Freedom and Command” the face has what we might call--to choose Habermas’s felicitous expression [44] --a “quasi-transcendental” function. A retreat upon the conditions that make autonomy possible regresses upon a condition that is also conditioned by what it conditions. Alternatively, we could say that the face is quasi-empirical to the extent that it is not entirely separate from the field of experience it makes possible. To be sure, from the point of view of the fundamental tenets of transcendental philosophy such formulations are absurd. It is just as absurd for Levinas in “Freedom and Command” to speak of “the experience of a face” (271/22) at the same time as referring to it as a “thing in itself” (270/20), “noumenon” (270/21), “intelligible” (271/22). The least that can be said is that these descriptions are not ultimate for Levinas. They belong to an “ontological” register that has not yet been modified to the point where a different “ethical” key can be heard, no less strange from the point of view of the Apollonian logos, though more faithful to what Levinas is trying to say.

Strictly speaking, there may be no “experience” of the Other, for the Other is not given as an object of cognition and is not treated as such. It is only the idea of the Other’s infinity that is given, the sense I have of the Other as infinitely transcending my cognitive powers, and that, according to Levinas in Totality and Infinity, “cannot, to be sure, be stated in terms of objective experience, for infinity overflows the thought that thinks it” (TI xii/25). Alternatively, it is not a matter of starting with the truth of experience, the conditioned, which is then traced back in linear sequence to reveal an a priori condition necessary for that truth to be possible: “the revelation of infinity does not lead to the acceptance of any dogmatic content and one would be wrong to argue for its philosophical rationality in the name of the transcendental truth of the idea of infinity” (xiii/25). The important, though--as I have shown--implicit lesson of “Freedom and Command” is that transcendental argumentation ultimately breaks down in connection with the Other in that it is incapable of regressing upon a condition that can be thought in terms of the intentional structure of consciousness governed by the unity of the “I think,” the basis of all experience, according to Kant.


In the final paragraph of “Substitution,” the “centerpiece” (OB ix/xli) of Otherwise Than Being, Levinas makes a statement that better than any other summarizes the proximity between him and Kant. The statement concerns ethical proximity itself:

If one had the right to retain one trait from a philosophical system and neglect all the details of its architecture (even though there are no details in architecture, according to Valéry’s profound dictum, which is eminently valid for philosophical construction, where the details alone prevent collapse) we would think here of Kantism, which finds a meaning to the human without measuring it by ontology and outside the question “What is there here . . .?” that one would like to take to be preliminary, outside of the immortality and death which ontologies run up against. The fact that immortality and theology could not determine the categorical imperative signifies the novelty of the Copernican revolution: a sense that is not measured by being or not being; but on the contrary being is determined on the basis of sense (166/129; my emphasis). [45]

Ignoring the Byzantine detail of the summum bonum and the practical postulates necessary for architectonic completeness, Levinas finds in Kant’s practical philosophy “un sens” (meaning, sense, direction) that is irreducible to ontology. By subordinating the interests of theoretical reason to those of practical reason, which amounts to asserting that the Categorical Imperative has practical efficacy without any theoretical explanation as to how it is possible, Kant’s doctrine of primacy signifies for Levinas a reversal of philosophy’s traditional vocation to ground thought and action in knowledge and truth. The ontological problematic “that one would like to take to be preliminary” is in this instance subordinated to ethics as an independent and preliminary praxis. We have seen how obligation in Kant is not determined by the comprehension of being, for the transcendental ground of the Moral Law--freedom--is not given to cognition. Nor is it determined by one’s “concern for being” (Heidegger) to the extent that “not even the threat of death” (MM 483/271) qualifies exemption from duty, which is unconditional. Rather, “being is determined on the basis of sense,” which is Levinas’s shorthand way of saying what Kant said in the second Critique about dutiful precepts “not needing to wait upon intuitions in order to acquire a meaning . . . for the noteworthy reason that they themselves produce the reality of that to which they refer” (CPrR 66/68). [46] Morality, according to Kant, prescribes what ought to be the case, not what is the case. It thus is capable of providing an orientation in thinking and acting independently of ontology, and this despite the fact that after the Copernican Revolution there is no longer any “up or down,” indeed no absolute knowledge at all.

If Valéry’s “profound dictum” from Eupalinos ou l’architecte (1921) [47] is “eminently valid for philosophical construction,” then so is Kant’s statement in the “Doctrine of Method” of the first Critique: “it turned out that although we had in mind a tower that would reach the heavens, yet the stock of materials was only enough for a dwelling house--just roomy enough for our tasks and just high enough to look across the plain” (CPR 707/735). Doubtless Kant can still be accused of exaggerating the commodiousness of his critical architecture. Yet his claim for its height still stands. As a philosopher “looking across the plain” he stands shoulder to shoulder with Levinas--proximity itself. [48]

Peter Atterton

San Diego State University


[1] Kant, The Metaphysics of Morals, trans. Mary Gregor (Cambridge, 1993). Henceforth MM, pagination of Prussian Academy/translation; MM 449/244.

[2] “L’ontologie est-elle fondamentale?” Revue de métaphysique et de morale 56 (1951): 98; trans. Peter Atterton as “Is Ontology Fundamental?” Philosophy Today 33 (1989): 127.

[3] Levinas, “The Primacy of Pure Practical Reason,” trans. Blake Billings, Man and World 27 (1994): 451. This article originally appeared in Dutch as “Het primaat van de zuivere praktische rede,” trans. C. P. Heering-Moorman, Wijsgerig Perspectief op Maatschappij en Wetenschap 11 (1970-1971). It has yet to be published in the original French.

[4] Levinas, “La Pensée de l’être et la question de l’autre,” De Dieu qui vient à l’idée (Paris, 1986): 185.

[5] “La Pensée,” 185.

[6] Levinas, Transcendance et intelligibilité (Geneva, 1984): 19-20; trans. Tamra Wright and Simon Critchley as “Transcendence and Intelligibility,” Emmanuel Levinas: Basic Philosophical Writings, ed. Adriann T. Peperzak, Simon Critchley, and Robert Bernasconi (Bloomington, 1996): 154.

[7] An important exception is Etienne Feron, “Intérêt er désintéressement de la raison Levinas and Kant,” Levinas en contrastes, ed. Michel Dupuis (Brussels, 1994).

[8] Kant, Groundwork of the Metaphysic of Morals (The Moral Law), trans. H. J. Paton (London, 1983). Henceforth G, pagination of Prussian Academy/translation; G 391/57.

[9] Kant, Critique of Practical Reason, trans. Lewis White Beck (Indianapolis, 1956). Henceforth CPrR, pagination of Prussian Academy/translation; CPrR 119/124.

[10] See Lewis White Beck, A Commentary on Kant’s Critique of Practical Reason (Chicago, 1960), who writes:

The truth of the matter is that the concept of the highest good is not a practical concept at all, but a dialectical ideal of reason. . . . It is important for the architectonic purpose of reason in uniting under one idea the two legislations of reason, the theoretical and the practical, in a practical dogmatic metaphysics wholly distinct from the metaphysics of morals. Reason cannot tolerate a chaos of ends . . . But we must not allow ourselves to be deceived, as I believe Kant was, into thinking [that the] possibility [of the highest good] is necessary to morality or that we have a duty to promote it (245).

Beck is surely correct. Not only would a prima facie duty to promote the highest good jeopardize Kant’s fundamental claim about the lawful form of a maxim being the sole determinant of moral action, it introduces into the field of morals impossible demands insofar as I, as a Kantian willer, can do nothing toward apportioning happiness in accordance with virtue aside from seeking to be worthy of happiness by becoming virtuous (the first component of the summum bonum).

[11] Kant, The Critique of Pure Reason, trans. Norman Kemp Smith (London, 1929). Henceforth CPR, Prussian Academy pagination to first and second editions separated by a slash; CPR 809/837.

[12] Kant, Religion Within the Limits of Reason Alone, trans. Theodore M. Greene and Hoyt H. Hudson (New York, 1960): 4.

[13] See also Kant’s remark in The Critique of Judgement, trans. James Creed Meredith (Oxford: Clarendon Press: 1957):

if we do not bring the causality of any other means beside nature into alliance with our freedom, the conception of the practical necessity of such an end [the highest good] through the application of our powers does not accord with the theoretical conception of its effectuation (sec. 87).

[14] “The Primacy of Pure Practical Reason,” 446. Levinas is not alone in considering the doctrine of primacy fundamental to Kant’s philosophy. A. D. Lindsay in his Kant (London: Ernest Benn, 1934) likewise contended: “The primacy of practical reason is Kant’s most essential doctrine. He will have nothing to do with reality except in terms of our action upon it” (303). For a quite different reading of Kant, see Habermas’ Knowledge and Human Interests, trans. Jeremy J. Shapiro (Oxford, 1987): “Kant secretly conceived of practical reason on the model of theoretical reason” (208).

[15] Levinas, “Transcendance et mal,” De Dieu qui vient à l’idée: 190; trans. Alphonso Lingis as “Transcendence and Evil,” Collected Philosophical Papers (Dordrecht, 1987): 176. Levinas continues:

In their own way the ideas rejoin being in the existence of God, who guarantees . . . in the letter of criticism, the concord of virtue with happiness . . . and the efficacy of a practice decided upon without knowledge. The absolute existence of the Ideal of pure reason, the existence of the Supreme Being, finally prevails in an architecture where the keystone was to be the concept of freedom alone (176).

Levinas’ criticism of the “postulates” here should be read in conjunction with the two lectures he gave on Kant at the Sorbonne two years earlier (see Dieu, la mort et le temps [Paris, 1993]). In the first lecture (74-78), Levinas offers a far more positive reading of the summum bonum: “To admit the existence of God and the immortality of the soul is required by Reason, but the highest Good can only be hoped” (77). For Levinas such hope delineates a temporal structure that is very different from the anticipation of being (or non-being) à la Heideggerian being-for-death. In Kant, hope is not governed by “anticipation” (Vorlaufen) or any prescience. Nor, as is sometimes thought, does it spring from the desire for immortality qua perpetual bliss (independent of one’s worthiness to be happy). Hence Levinas’ final remark: “it is not by chance that this way of thinking a meaning beyond being is the corollary of an ethics” (78). In the second lecture (175-178), Levinas deals with the first Critique and is a little less conciliatory toward Kant. Having gestured toward a break with ontology accomplished by Kant’s “Transcendental Dialectic” and the separation of ideas and concepts, reason and understanding, Levinas ends with a discussion of the “return to onto-theo-logy in Kantian thought” (177) by its determination of God as a transcendental Ideal (qua omnitudo realitatis). What is interesting about both lectures is the way in which they manage to offer a double-reading of Kant by separating practical reason from theoretical reason. The question as to which has primacy in Kant’s system overall is not addressed.

[16] De Dieu qui vient à l’idée, 190; Collected Philosophical Papers, 176. What are we to make of Levinas’ description of the good will as “utopian, deaf to the information . . . that could come to it from being”? What does he mean when he says that it “proceeds from a freedom which is situated above being” etc? As is well known, Kant considered the good will to be determined by a purely rational law called the Categorical Imperative. The Categorical Imperative does not prescribe what actually exists in the world, but what ought ideally to exist through the adoption of principles that can be shared by all rational agents. It thus produces in us the “splendid ideal [herrliche Ideal] of a universal kingdom of ends” (G 462/122) whose members subject themselves to laws of their own making. Such agents are in a sense “deaf” because the laws (of duty) they follow do not need ”to wait upon intuitions in order to acquire a meaning . . . they themselves produce the reality of that to which they refer” (CPrR 66/68). Kant calls this “property which the will has of being a law to itself” (G 447/107) autonomy. But autonomy is only the legislative side of the will (Wille). It presupposes (“proceeds from”) “negative” (G 447/107) freedom in the sense of the capacity of the will (Willkür) to act independently of the mechanism of nature. This lawless “spontaneity” (G 447/107), which is the condition of the possibility of the good will, is “freedom in the strictest, i.e., transcendental sense” (CPrR 29/28).

It is interesting to juxtapose Levinas’ reading of Kant here with that of Nietzsche. In The Gay Science, Nietzsche writes in typically irreverent fashion: “old Kant . . . was led astray--back to ‘God,’ ‘soul,’ ‘freedom,’ and ‘immortality,’ like a fox who loses his way and goes astray back into the cage. Yet it had been his strength and cleverness that had broken open the cage!” (trans. W. Kaufmann, New York, 1974, sec. 335): 264. While both Levinas and Nietzsche view the postulates as a step back into traditional ontology, their understanding of the way in which Kant can be said to have “broken open the cage” differs enormously. For Levinas this takes place in the “practical usage of pure reason,” and thus directly in association with the Categorical Imperative and the concept of a good will grounded in transcendental freedom. For Nietzsche, the Categorical Imperative is precisely what led Kant astray, and the postulate of freedom is not considered essentially different in kind from the metaphysical postulates of God and immortality of the soul. Had Kant held firmly to the critical agnosticism of the first Critique he might have remained outside the cage.

[17] Kant writes in the second Critique:

the concept or freedom was problematic but not impossible; that is to say, speculative reason could think of freedom without contradiction, but it could not assure any objective reality to it. [Theoretical] Reason showed freedom to be conceivable only in order that its supposed impossibility might not endanger reason’s very being and plunge it into an abyss of skepticism (CPrR 3/3).

Note that the resolution of the third antinomy and the theoretical interest of reason offered only a “natural recommendation for the assertions of the thesis” (CPR 475/503). Kant goes on to say that “if men could free themselves” of their “natural” bias toward the belief in freedom for the purposes of architectonic completeness, theoretical reason itself would be unable to decide on the intrinsic force of each of their grounds between the thesis and the antithesis, and we would remain in a state of “continuous vacillation” (CPR 475/503).

[18] Experience and Systematization (The Hague, 1965), 121.

[19] Rotenstreich, 121.

[20] I realize, of course, that the meaning of the “fact of reason” passages in the second Critique is open to debate. My interpretation is guided by a constructivist reading of the type undertaken by Onora O’Neill in her Constructions of Reason: Explorations of Kant’s Practical Philosophy (Cambridge, 1989). While acknowledging that “a great deal of textual spadework” (65) has to be done in support of her claim, O’Neill argues that if the Moral Law is not to be construed as something purely given, a datum, which we lack the intuitive faculty to receive, then we would do well to understand it as a factum, as something constructed or made by reason, thereby attesting to the spontaneity of reason and its capacity to generate its own principles when everything empirical or alien is abstracted. On this reading, the “fact of pure reason” is not simply a fact given by reason, but the fact that there is pure reason, i.e., autonomy in thinking and acting.

O. Neill’s book is a convincing defense of the primacy of practical reason in Kant in that she shows that any vindication of reason must vindicate the Categorical Imperative. However, her remarkable and highly original reading of Kant’s second analogy (62-63), whereby she shows that Kant’s crucial distinction between the appearance of succession and the succession of appearances presupposes practical freedom, to an extent weakens the doctrine as Kant presents it in second Critique (which she quotes in a footnote at the beginning of her book). For we have here a theoretical interest in the assertion of freedom which is independent of moral interest. Hence her conditional claim: “if we are to have empirical and scientific knowledge, we not merely may but must be free agents” (63). O’Neill thus gives a stronger argument than Rotenstreich does to show that freedom may be deduced from theoretical standpoint, though she does so only at the expense of undermining Kant’s claim that practical reason alone provides the only assertoric interest in the Idea of freedom. It is, of course, that interest over and against any theoretical interest in the existence of the Ideas that interests Levinas.

[21] Rotenstreich, 119.

[22] “The Primacy of Pure Practical Reason,”  451.

[23] Kant was also dismissive of ontology earlier in the first Critique: “the proud title of Ontology that presumptuously claims to supply, in systematic and doctrinal form, synthetic a priori knowledge of things in general (for instance, the principle of causality) must therefore give place to the modest title of a mere Analytic of pure understanding” (CPR 246-7/303).

[24] Levinas, “Les droits de l’homme et les droits d’autrui,” Hors subjet (Paris: 1987): 184; trans. Michael B. Smith as “The Rights of Man and the Rights of the Other,” Outside The Subject (Stanford, 1993): 122; modified translation.

[25] When reason is deployed in a merely cognitive and not a conative capacity, as when it is ancillary in the achievement of some pathologically determined end, then the will can really be said from a Kantian point of view to be a “bad will” (böse Wille).

[26] Jacques Derrida, Writing and Difference, trans. Alan Bass (London: Routledge, 1978), 96. However, see TI 41;240;274;279(twice)/ 69;262;298;302;303, where Levinas calls the ethical relation “respect.”

[27] Derrida, 314, n. 26.

[28] In Totality and Infinity, “need” (besoin) is said to be in principle satiable, in contradistinction to ethics, which is “beyond satisfactions” (TI 4/34). The sum satisfaction of need is “happiness” (87/115). Note that in Otherwise Than Being Levinas recasts the distinction between ethics and need in terms of the inveterate “man versus nature” opposition: “responsibility for others could never mean altruistic will, an instinctual ‘natural benevolence’” (OB 142/111-20). “It is against nature” (OB 157n27/197n27). Kant too, of course, sought to remove ethics from the realm of nature.

[29] The Critique of Judgement, sec. 5. Disinterestedness here is also to be sharply distinguished from type of disinterested pleasure that the British empiricists called “moral sense,” which we allegedly experience when performing or contemplating a righteous action, and which, according to Kant, “reduces everything to the desire for one’s happiness” (CPrR 38/40).

[30] Levinas, “Avez-vous relu Baruch?” Difficile liberté: Essais sur le judaïsme (Paris, 1976): 169; trans. Seán Hand as “Have You Reread Baruch?” Difficult Freedom: Essays on Judaism (Baltimore, 1990): 118.

[31] The Critique of Judgement, sec. 87.

[32] The idea that “love” can be commanded, “contrary to Kant,” is one Levinas finds in the work of Franz Rozensweig. See “‘Entre deux mondes’” (Difficile liberté); trans. Seán Hand as “‘Between Two Worlds’” (Difficult freedom).

[33] In the third Critique Kant states that the “feeling of the sublime in nature is respect for our vocation, which we attribute to an Object of nature by a certain subreption (substitution of a respect for the Object in place of one for the idea of humanity in our own self--the Subject)” (sec. 27). As a homo noumenon, a person is a subject of morally practical reason (autonomy); however, as a homo phaenomenon, he or she is also subject to morally practical reason (obligation). It follows that only the “personality” of a person is an appropriate target of respect, and it would be subreptic here to confuse anything sensible with what properly belongs to reason and the understanding as such.

[34] While Kant considered every rational being as such is capable of awakening the feeling of reverentia, and also due respect in the sense of observantia (not to be treated as a means only), he certainly did not think that every rational being was capable of feeling reverentia, “which presupposes the sensuousness and hence the finitude of such beings on whom respect for the moral law is imposed” (CPrR 76/79). Human beings alone, beings which can be viewed both from the standpoint of the phenomenal order of sensible appearances and the intelligible order of rational causes, are capable of this feeling. God and “holy wills,” to choose the standard examples, as pure or disembodied wills, are exempt according to Kant (CPrR 76/79). Indeed, despite the common belief that Kant sought to eradicate all sensibility from moral affairs, it is clear from all three of Kant’s major works on morals that without this peculiar feeling, morality could not function at all for us. We would be incapable of being under moral obligation and thus “morally dead” (MM 400/20).

[35] Passages such as this might suggest that Levinasian Desire is a type of libidinal movement or assemblage of singularities that Deleuze calls “desiring production,” a machine-like desire that is “productive” in the sense that it is motivated without the consciousness of any prior lack. However, it has to emphasized time and again that Levinas is entirely opposed to the theory of the unconscious (even if it is irreducible to the Freudian-Lacanian model), which he calls a “merely provisional retreat of alterity” (“Transcendance et mal,” 199; “Transcendence and Evil,” 180). Levinas maintains that the psychoanalytic “unconscious” is simply the opposite of consciousness, and thereby intelligible only on that basis. To repeat what was said earlier, the will remains free to disobey the command of the Other even though it is not free to choose responsibility itself. Indeed, it is because of the possibility of disobedience that the will can be subject to obligation at all, to an imperative in the Kantian sense, applicable to a being whose “natural” impulse is to satisfy its own material needs.

[36] Levinas, “Liberté et commandement” (Revue de Métaphysique et de Morale 58:3 (1953): 264; trans. A. Lingis as “Freedom and Command (Collected Philosophical Papers); henceforth cited as “Freedom.”

[37] Plato, Phaedo, translated by Hugh Tredennick and Harold Tarrant, in The Last Days of Soctrates, ed. Harold Tarrant (London, 1993), 118a (185).

[38] Plato, The Republic, trans. G. M. A. Grube (London, 1981), bk. IV, 430e (110).

[39] This is the first time we encounter what will become a central claim in both the third section of Totality and Infinity, and the fifth chapter of Otherwise Than Being. However, the justification Levinas presents for making such a claim is slightly different in each case. In Totality and Infinity it is argued that for the will to have recourse to impersonal reason there must be an understanding of language first taught by the Other: “The Other is not for reason a scandal . . . but the first rational teaching, the condition for all teaching” (TI 178/203). In Otherwise Than Being, a major preoccupation of which is to justify the move from ethics to third party justice and politics, reason and ontology are said to derived from the ethical necessity to surmount the threat of complaisance compromising the ethical relationship, in danger of becoming an egoïsm à deux. In “Freedom and Command”--a work that shows the influence of Plato on Levinas’ thinking at this time--it is primarily a question of sparing the will “the ordeals of tyranny” (“Freedom” 266/17).

[40] The Republic, bk. 1, 327c (2).

[41] Republic, bk 1, 350e (28).

[42] See Förster Eckart, “How Are Transcendental Arguments Possible?” Reading Kant: New Perspectives on Transcendental Arguments and Critical Philosophy, ed. Eva Schaper and Wilhelm Vossenkuhl (Oxford, 1989). In this helpful article Förster writes: “In transcendental science . . . only a single proof is possible, namely, from the concept of the subject” (7).

[43] Levinas, “Langage quotidian et rhétorique sand éloquence” (Hors sujet, 210); trans. Michael B. Smith as “Everyday Language and Rhetoric without Eloquence” (Outside the Subject, 141). See also OB 182-3n7/198n8.

[44] Habermas, 194.

[45] This statement is an extended version of one Levinas made six years earlier at the end of “Humanisme et An-archie,” Humanisme de l’autre homme (Paris, 1972): 82; trans. Alphonso Lingis as “Humanism and An-archy,” Collected Philosophical Papers, 138.

[46] See note 16 above.

[47] Literally Valéry’s dictum runs: “Il n’y a point de détails dans l’exécution” (Oeuvres II, ed. Jean Hytier [Paris, 1960], 84). These words are spoken in the dialogue by Phaedrus, who represents the views of the architect or engineer, Eupalinos. While it is possible that Phaedrus is referring here to the execution of art in general (including music) rather than to architecture as such, the fact that the prose work was commissioned by the review Architectures, who according to Walter Putnam, in Paul Valéry Revisited (New York: Twayn Publishers, 1994), “knew of Valéry’s passionate admiration for their art” (91), suggests that Levinas is probably correct in taking Eupalinos’ guiding principle to refer to architecture, the “first of the arts” etymologically speaking.

[48] I wish to thank Graham Noctor and David Webb, for their generous help with an earlier version of this paper.