2. Kant on Belief and the Traditional Account

Kant situates his characterization of belief within his account of assent, or taking-to-be-true (Fürwahrhalten). By the latter term, he understands a positive attitude towards a proposition, such that we take it to truly represent some state of affairs (cf. A822/B850; JL, 9: 65–6).Footnote ⁵ In the Canon of the Critique of Pure Reason, Kant distinguishes between three kinds of assent: opinion (Meinung), belief (Glaube) and knowledge (Wissen). He claims that opinion is both subjectively and objectively insufficient, while belief is subjectively sufficient and objectively insufficient, and knowledge is both subjectively and objectively sufficient (cf. A822/B850; JL, 9: 66). It is not easy to understand what Kant means by objective and subjective sufficiency.Footnote ⁶ Since my purpose is to compare Kant’s views on practical justification with Crusius’s, however, it is enough to characterize subjective and objective sufficiency as follows: a subjectively sufficient assent is based on grounds that are sufficient to produce full approval of a proposition in a particular subject,Footnote ⁷ whereas an objectively sufficient assent is based on grounds that are sufficient to warrant the truth of the proposition to which we assent.Footnote ⁸ We can call the grounds that are at issue in the latter case ‘objective’. They identify different kinds of evidence (mathematical, empirical, etc.) that are intersubjectively recognizable as valid. The grounds that are relevant in the first case can instead be called ‘subjective’. They can be either what we subjectively take to be objective grounds or distinctively subjective grounds that do not aim at objective validity.Footnote ⁹

According to this characterization of subjective and objective sufficiency, opinion can be understood as a provisional assent (cf. JL, 9: 66) in which we have enough objective grounds to consider the truth of a proposition more probable than its falsity. Still, these grounds are insufficient to warrant the truth of the proposition to which we assent and, subjectively, to produce full approval. On the other hand, knowledge is an assent in which we have objective grounds that warrant the truth of the proposition to which we assent. Such grounds are also able, subjectively, to produce full approval. Finally, belief is an assent in which, although we do not have sufficient objective grounds that warrant the truth of the proposition to which we assent, we nonetheless have subjective grounds that not only produce but also warrant full approval of the proposition. Therefore, in the case of belief, subjective grounds have not only ‘causal’ determining force over our approval but also justificatory force, which I will clarify in a moment.

This description of belief needs specification, for it may seem that it licenses belief in cases where we lack sufficient objective grounds for assent but nonetheless take ourselves to have those grounds, thus fully approving the proposition in question.Footnote ¹⁰ These cases are in fact explicitly ruled out by Kant, however, because in order for belief to obtain, we need to be conscious that the grounds we have are not objective and, consequently, that our assent is not a candidate for knowledge (cf. A822/B850). The only kind of subjective grounds that are permissible in cases of belief are then genuinely subjective grounds (more on these below), and not what we falsely take to be objective grounds. Similarly with opinion and knowledge. That is, when our evaluation of the objective grounds we have is mistaken, our assent cannot legitimately be considered either opinion or knowledge. Both for knowledge (Log-W, 24: 852)Footnote ¹¹ and for opinion (cf. JL, 9: 66), we need to be conscious of the actual objective grounds we have.

Therefore, while the subjective grounds for opinion and knowledge seem simply to be a correct evaluation of the objective grounds we have, which produces an adequate subjective attitude toward the proposition (provisional approval for opinion and full approval for knowledge), subjective grounds for belief single out genuinely subjective grounds that are independent of objective ones. This explains why subjective grounds of belief have not only a ‘causal’ determining force but also a justificatory force over our approval. In this case, it is the subjective grounds themselves that do the justificatory work. But how are these grounds to be understood? For Kant, genuinely subjective grounds are practical in nature. They license a belief in a state of affairs because the latter is a condition for the realization of certain ends we pursue in practice. Kant’s thought seems to be the following: when the evidence we have leaves undecided whether a proposition is true or false (such that we lack sufficient objective grounds to decide the issue), we are rationally required to believe in that proposition if it identifies conditions that are necessary for the realization of ends we are pursuing. In other words, for Kant we cannot rationally pursue an end if we do not believe that the conditions for its attainment apply. For this reason, once we actually have this end and follow it in our actions, we are rationally required to believe that the conditions of its realization are satisfied even if we lack decisive evidence on the matter (cf. A823–4/B851–2).

Having characterized the subjective grounds that are at play in cases of belief in this way, Kant distinguishes between kinds of belief by making reference to: (1) the contingent or necessary nature of the ends that are at stake, and (2) the possibility, or lack thereof, of differently evaluating, in different epistemic circumstances, the conditions for realizing the ends in question (cf. A823–4/B851–2). In this respect, with pragmatic beliefs the relevant ends are contingent (we are not rationally required to pursue them, or alternatively, the fact that we have them is dependent on contingent circumstances), and we would perhaps regard other states of affairs as conditions for their realization if the epistemic situation were different in relevant respects (for example, if we had more powerful means to obtain pertinent evidence, etc.) (cf. A824/B852). As an example, Kant presents the case of a doctor who must do something to cure a patient if the latter is to survive. The doctor has no decisive evidence to determine the disease, even though there are some hints that it is consumption. He is said to be required to act on the belief that the disease is consumption because this is the only way to save the patient’s life (cf. A823–4/B851–2). Here, the doctor’s end is contingent, not because he is not required to act in a certain way in that situation,Footnote ¹² but because his being in such a situation is contingent.Footnote ¹³ Moreover, the doctor would perhaps judge the patient’s condition differently in different epistemic circumstances. Had he better instruments at his disposal, for example, he would perhaps arrive at a different evaluation, and the proposition asserting that the patient has consumption would no longer describe a condition for the realization of his end (saving the patient).

On the other hand, moral beliefs are beliefs where the ends are necessary (we are rationally required to pursue them) and we cannot but regard certain states of affairs as conditions for the realization of these ends (cf. A828/B856).Footnote ¹⁴ The beliefs that can be viewed as moral in this sense are, for Kant, the belief in God and in the immortality of the soul.Footnote ¹⁵ Such beliefs are conditions for the realization of the highest good, that is, of a world in which a particular degree of moral perfection corresponds to a particular degree of happiness (cf. A814/B842). The highest good is a necessary end for Kant since we are rationally required to pursue it as rational moral agents. Moreover, the existence of God and the immortality of the soul are states of affairs we cannot but see as conditions for the realization of this end (cf. A811/B839). The fact that we cannot evaluate these conditions differently seems here to depend on the fact that our epistemic situation cannot change in relevant ways with respect to the evaluation of these objects. We see these objects as conditions of reaching the highest good on the basis of rational reflection, which is not dependent on contingent empirical factors. Moreover, both God and the soul are transcendent objects; because they lie beyond the limits of possible experience, we cannot expect that our evidence for them will change.

Kant calls the kind of certainty that is open to us in cases of moral belief ‘moral certainty’ (cf. A829/B857). Given that the ends that ground moral belief are rationally required, and given that our viewing the states of affairs identified in those beliefs as conditions for moral ends is not sensitive to changes in our epistemic situation, our belief in these conditions cannot be viewed as rationally dispensable, and our assent can for this reason attain the status of moral certainty.

As has been shown in the literature (Fonnesu Reference Fonnesu2011, Reference Fonnesu2015; Pasternack Reference Pasternack2011), Kant’s understanding of both belief and moral certainty deviates significantly from accounts of these concepts that were widespread in eighteenth-century Germany. Belief, on the one hand, was commonly defined as assent based on the testimony of others. This is the definition that we find, for example, in Christian Wolff (Wolff Reference Wolff1713: ch. 7, §19, 204) and George Friedrich Meier (Meier Reference Meier1752: §206), whose influence on Kant is without question (see also Reimarus Reference Reimarus1766: §239; Walch Reference Walch1740: 1327–8). On the other hand, moral certainty was characterized as the certainty we can attain when absolute certainty is not within our reach. It can be understood as an inferior kind of certainty which is nonetheless sufficient for practical everyday life (Meier Reference Meier1752: §175) or as a different but equally trustworthy kind of certainty obtained not through demonstration but through a proposition’s having the highest possible degree of probability (Crusius Reference Crusius1747: §§361, 422; see also Walch Reference Walch1740: 1310; on moral certainty see Borghero Reference Borghero1983; Shapiro Reference Shapiro1983, Reference Shapiro2003; Daston Reference Daston1988; Fonnesu Reference Fonnesu2011; Perinetti Reference Perinetti2014).

In his lectures on logic, Kant himself signals these differences between his account of belief and moral certainty and the views held by the vast majority of his predecessors and contemporaries. Regarding belief, he argues that the fact that a cognition is based on the testimony of others cannot be the defining feature of belief. Testimony is just as reliable a source of evidence as experience. Consequently, there is nothing that prevents us from attaining knowledge by means of it: ‘so-called historical belief [i.e. belief based on testimony; my note] cannot really be called belief, either, and cannot be opposed as such to knowledge, since it can itself be knowledge. Holding-to-be-true (Fürwahrhalten) based on testimony is not distinguished from holding-to-be-true through one’s own experience either as to degree or as to kind’ (JL, 9: 68–9). As far as moral certainty is concerned, Kant clarifies that he does not understand the latter as having to do with probability, as many of his predecessors and contemporaries do: ‘[m]ost, almost all autores are completely unacquainted with moral certainty, and instead they take it in each case to be probability’ (Log-Bl, 24: 200).

According to the picture I have drawn in this section, it would be easy to conclude that Kant advances an account of belief and practical justification for assent that is radically new with respect to widespread views in eighteenth-century Germany. This conclusion may in fact be true as far as certain widespread views are concerned. However, it conceals the fact that aspects of Kant’s position are actually anticipated in important respects by Crusius’s account of belief and practical justification.

3. Crusius on Belief and Practical Justification

It is correct to regard Crusius’s definition of moral certainty as traditional. On his account, moral certainty is the highest degree of probability that a proposition can have and has nothing to do with morality (Crusius Reference Crusius1747: §§361, 422). Things are different as regards Crusius’s concept of belief (Glaube), however.Footnote ¹⁶ In this section, I will show that there is significant evidence that Crusius’s account of belief influenced Kant. This becomes apparent if we consider Crusius’s definition of belief and his account of practical grounds for assenting to only probable propositions.

First, Crusius explicitly rejects the traditional identification of belief with assent based on the testimony of others: ‘that clarification that some give when they say that believing means as much as giving approval to a proposition on the basis of the testimony of another person is not adequate’ (Crusius Reference Crusius1744: §344). Crusius motivates this claim by saying that we can clearly have cases of belief that are not supported by testimonial evidence (Crusius Reference Crusius1744: §344). This critical stance toward the traditional definition of belief already displays a similarity with Kant’s approach. Further similarities abound when we consider Crusius’s own definition of belief. In the Anweisung vernünftig zu leben, Crusius characterizes belief in the broad sense as ‘that state of a rational spirit (Geist), in which his will decides to let a represented proposition count as true (als wahr gelten zu lassen) and to act according to it when the occasion arises (bey Gelegenheit), as long as this is not in turn prevented by other causes’ (Crusius Reference Crusius1744: §334). A state of belief thus has two main features. It is a state in which (i) we take a proposition to be true and (ii) we are prepared to act on the basis of that proposition when the circumstances arise. This recalls what Kant says on belief in the Jäsche Logic, where he describes belief as assent which is ‘sufficient for action (genug zum Handeln)’ (9: 68n, translation altered; see also Refl 2793, 16: 513).

While Kant’s characterization of belief as ‘sufficient for action’ is essentially restricted to the specific sense he attributes to this concept, for Crusius this is a property of belief ‘in a broad sense’, which he views as a synonym of assent or taking-to-be-true in general (Vorwahrhalten) (cf. Crusius Reference Crusius1747: §447).Footnote ¹⁷ Accordingly, in Weg zur Gewissheit und Zuverlässigkeit der menschlichen Erkentniss, he describes assent in a very similar way, namely, as a ‘state in which we represent a proposition as true and decide to let it count as a true proposition (als einen wahren Satz gelten zu lassen) in our practice (Verfahren)’ (Crusius Reference Crusius1747: §444). Since assent, or belief in a broad sense, is characterized in terms of complete approval (vollständiger Beyfall) and a ‘readiness to act’, opinion (Meinung) does not count as a proper case of assent for Crusius, because with opinion our approval is incomplete (unvollständing) and insufficient for action (Crusius Reference Crusius1744: §344). We know that opinion counts as a case of assent for Kant, which suggests that his concept of assent is more inclusive. The relevant difference between Crusius’s and Kant’s views on assent appears to be that for Kant ‘readiness to act’ according to a proposition is not a defining feature of assent in general,Footnote ¹⁸ whereas it is a defining feature of belief on Kant’s specific understanding of the term.Footnote ¹⁹

While for Crusius belief in a broad sense is the same thing as assent, belief in a narrow sense is ‘mere’ belief. This is an assent that lacks either demonstrative certainty or the certainty provided by immediate experienceFootnote ²⁰ but in which we nonetheless fully assent to a proposition (cf. Crusius Reference Crusius1747: §447, Reference Crusius1744: §334). For our purposes, it is interesting to see what, according to Crusius, makes a belief in the narrow sense rational. Rational beliefs are either morally certain assents or assents in which ‘one, due to a cognized connection of a probability with one’s final ends, or also due to a legal obligation, deliberately represses one’s rising doubts and decides to accept (annehmen) as true the propositions to be believed and to act according to them as according to true propositions with complete approval (vollkommenem Beyfalle)’ (Crusius Reference Crusius1747: §447). The first kind of rational belief simply identifies those cases in which the certainty we have regarding a proposition is obtained by following ‘the way of probability’, a procedure which establishes certainty by showing that a proposition has the highest degree of probability (Crusius Reference Crusius1747: §§361, 422). Since this understanding of rational belief connects the latter to well-grounded probable cognition, it cannot be assimilated to Kant’s approach to belief. By contrast, the second kind of rational belief displays relevant similarities with Kant’s view. In this case, beliefs are rational because they are related either to ‘final ends’ or to a ‘legal obligation’ we must adhere to in our practice. We saw that a relation to ends and moral obligation plays an essential role in Kant’s justification of belief. However, this may represent only a superficial analogy. In order to determine the actual ties between Crusius and Kant, we need to determine what role ‘final ends’ play in the justification of belief and what ‘legal obligation’ means in the quoted passage. In order to do this, I will now consider Crusius’s account of practical grounds of assent.

Recall, first, that for Kant the subjective grounds relevant to belief are distinctly practical and rest on the connection between the assent and our particular ends. Kant’s insight can be briefly put as follows: when a proposition is left undecided by the evidence we have, we are rationally required to believe it if the latter describes conditions for the realization of certain ends we pursue in our practice. Secondly, Kant distinguishes between pragmatic and moral beliefs by also focusing on the contingent or necessary character of the ends we pursue, where contingent ends are so-called ends of ‘prudence’ (Klugheit).Footnote ²¹ There are striking continuities between Crusius and Kant with respect to both aspects of Kant’s approach.

As far as the first aspect is concerned, in his account of the grounds for assent in the case of probable knowledge, Crusius distinguishes between grounds of the intellect, which ‘must lie in the quality (Beschaffenheit) of the propositions and in their relation to other propositions’ (1747: §362), and grounds that ‘consist in the connection of the cognized thing with certain final ends’ (1747: §412; cf. §362). Here, grounds of the intellect seem to concern the determination of the probability of one proposition with respect to the probability of others. By contrast, the second kind of ground, which we might call practical, rests on an explicit connection with ends we pursue in practice. If we take into consideration one instance of these latter grounds, that is, those based on ends of prudence, it becomes apparent that Crusius’s account of practical justification has some elements in common with Kant’s. Thus Crusius explains the normative power that ends of prudence exert on our actions and assent as follows: ‘The obligation (Verbindlichkeit) of prudence … either originates in the fact that accepting the probable cognition and acting upon it is the only means to certain human final ends; or it is based on the consideration of the security of this means’ (1747: §414). Let us first consider the case in which we have the ‘security of the means’: when we are certain on the basis of evidence (and here it is ‘moral’ certainty that is at stake, that is, certainty based on an extremely high probability) that some state of affairs obtains and that if we act according to the proposition expressing that state of affairs we can realize certain of our ends, our ends command us to act on the basis of the certain proposition.Footnote ²² By contrast, when, on a theoretical basis, we are not certain but have a merely probable cognition that a certain state of affairs obtains, and acting on the basis of the belief that it does is the only possible way to realize an end we have, we are rationally required to believe that that state of affairs obtains and to act on that belief. In this scenario, the truth of the proposition describing the state of affairs in question identifies conditions for the success of our action. Like Kant, Crusius thus argues that our ends can command our assent to a proposition if this proposition identifies conditions that are necessary for the realization of those ends. Therefore, it seems that for Crusius ‘final ends’ play a role in the justification of our assent that is very similar to the role played by ends in Kant.

Turning to the second aspect of Kant’s approach to the justification of belief mentioned above, note that Crusius also distinguishes between two kinds of ends that can support our assent in the case of belief. These can be either ends of prudence (Klugheit) or ends of justice (Gerechtigkeit) (cf. 1747: §362), and they are linked to two corresponding kinds of obligation (Verbindlichkeit):

First, let me emphasize that both ends of prudence and ends of justice support our assent in the way I have just described. They justify our assent to a proposition that is only probable because acting on such assent is a condition for the realization of certain ends we have.Footnote ²³ Moreover, ends of prudence seem to be very close to the contingent ends that are at issue in Kant’s pragmatic beliefs. Crusius accordingly characterizes ends of prudence as ends that we desire (begehren) according to our nature and that are contingent because we would not violate any indebtedness (Schuldigkeit) toward God if we failed to pursue them (Crusius Reference Crusius1744: §162). The obligation that they issue is therefore not necessary but dependent on the pursuit of a contingent end: if I have end E, I must do A to realize it. If the success of my action depends on the truth of the probable proposition P, however, I am also rationally required (and so justified) to believe in P. Moreover, the similarity between Kant’s pragmatic beliefs and Crusius’s assents based on ends of prudence is suggested by the fact that Crusius lists the ‘evaluation of most diseases’ as an instance in which an obligation of prudence can apply (1747: §414), where this recalls Kant’s example of the doctor.

One might think that ends of justice simply identify another kind of contingent end, which depends not on our personal inclinations but on the will of a ruler. This would be a mistake, however, since for Crusius the ‘ruler’ who grounds any kind of legal obligation is ultimately God (cf. Crusius Reference Crusius1753: §481). Crusius’s line of thought can be briefly put as follows: God provides a law, which human beings are bound to follow (cf. Crusius Reference Crusius1744: §161; Crusius Reference Crusius1753: §480). It is when human beings’ free actions agree with God’s law that they partake of virtue (Tugend) (cf. Crusius Reference Crusius1744: §161). What Crusius calls ‘ends of justice’ must be clarified in this framework. These are ends that are based on our ‘indebtedness’ (Schuldigkeit) and ‘obedience’ (Gehorsamkeit) to God (cf. Crusius Reference Crusius1744: §165, Reference Crusius1753: §481). In contrast to ends of prudence, they rest on a law that we must necessarily obey. As such, they can be considered necessary moral ends. But how should we understand the obligation on which ends of justice rest? How can the latter justify our assent? Even if the ‘legal obligation’ at stake here is not categorical in Kant’s sense, it expresses Crusius’s attempt to grasp an obligation that issues necessary duties. The justification in this case works as follows: as a free agent who is ‘indebted’ (schuldig) and obedient (gehorsam) to God, I must pursue end V; that is, I must pursue virtue. An action (or a course of action) A is necessary to obtain V. If the success of my action (or my course of action) depends on the truth of the probable proposition P, however, I am also rationally required (and justified) to believe in P.

In order to see how Crusius’s approach actually works, we can take into consideration one of his proofs for the immortality of the soul. Both in the Anweisung and in the Entwurf der nothwendigen Vernunft-Wahrheiten, Crusius presents a proof with the following form: as rational beings, we have the capacity to strive toward a final end which is described as either perpetual (ewig) (cf. Crusius Reference Crusius1744: §219) or endlessly continuous (unendlich fortdauernd) (cf. Crusius Reference Crusius1753: §483). By this Crusius seems to mean an end that requires a perpetual or endlessly continuous striving on our part. Insofar as God creates no capacity in vain, we must assume that rational beings are capable of endlessly pursuing this perpetual end. But this is only possible if the souls of rational beings are immortal (cf. Crusius Reference Crusius1744: §219; cf. Crusius Reference Crusius1753: §483).

If we stop here, it is unclear how this proof matches the model of argument based on legal obligation sketched above. To make this connection, we should observe that, according to Crusius, God’s own final end is to make ‘morally active creatures’ (moralisch wirkende Geschöpfe) happy, as long as they conform to, or partake of, virtue (Crusius Reference Crusius1753: §479; cf. Crusius Reference Crusius1744: §213). God’s own final end therefore has two elements: virtue, which demands a world where every free action agrees with God’s perfection (Crusius Reference Crusius1753: §477), and happiness, which is given in proportion to the virtue that free agents have actually acquired (Crusius Reference Crusius1753: §478).Footnote ²⁴ In the Entwurf, Crusius places God’s own final end at the very basis of the perpetual ends on which he builds his argument for the immortality of the soul (Crusius Reference Crusius1753: §483). Given that virtue and happiness are the two main elements of God’s own final end and that happiness is subordinated to virtue, it is then plausible to view virtue as the main and most important perpetual end in Crusius’s argument. But virtue is a necessary end insofar as we are bound to follow God’s law. This means that, in this case, we do not simply have the capacity to follow this perpetual end but are rather bound to pursue it. Crusius’s argument can thus be interpreted as follows: as free agents who are indebted and obedient to God, we must pursue virtue. Given that this end is perpetual or endlessly continuous, we can only pursue it adequately if we have the capacity to endlessly progress toward virtue. But this is only possible if our soul is immortal. Given that the end of pursuing virtue is an end that God has set before us as necessary, and given that God does nothing in vain (vergeblich), we must have the capacity to endlessly pursue virtue and must therefore be immortal.

Notice a few things about this argument. First, it seems to build on a form of the ‘ought-implies-can’ principle. If we ought to pursue virtue, this means that we can do it, and immortality is a further condition of our capacity to do so. There is thus an inference from an ‘ought’ to a ‘can’, but this is accomplished only indirectly, through God’s nature. Since God has given us this necessary and endless end, and since he does nothing in vain, it follows that we can also adequately pursue this end. Second, the fact that Crusius here provides ‘proof’ (Beweis) of the immortality of the soul does not mean that for him we can ‘know’ that we are immortal. Rational beliefs based on practical grounds also need ‘proof’ (Crusius Reference Crusius1744: §340). What is distinctive about proof in support of practical beliefs is that it builds on obligation. The argument I have reconstructed is based on our obligation to pursue virtue. Therefore, it must be taken as only establishing the rationality of belief in the immortality of the soul. This suggests that if there is an ‘ought-implies-can’ principle in Crusius’s proof, it cannot be taken in a strong sense as implying that we can infer that we ‘know’ that we are immortal. The principle can only show that if we regard ourselves as being under an obligation and there is no evidence that the conditions for satisfying this obligation do not obtain, we are rationally required to regard the satisfaction of that obligation as possible and, consequently, the states of affairs that are conditions of that possibility as obtaining. Read in these terms, Crusius’s argument displays important similarities with Kant’s argument for the postulate of immortality in the Critique of Practical Reason (cf. 5: 122–3).Footnote ²⁵ The main difference between the two is that in Crusius’s argument the ‘ought-implies-can’ principle, understood in the modest sense I have suggested, is only obtained by means of a reference to God, whereas in Kant it is assumed as a rational principle that stands on its own.Footnote ²⁶

Let us briefly return to Crusius’s distinction between ends of prudence and ends of justice and between obligations of prudence and legal obligation. I have already suggested that ends of prudence seem to be very close to the contingent ends that are at issue in Kant’s pragmatic beliefs. I have also suggested that the obligation they issue is not necessary but dependent on the pursuit of a contingent end. From what I have just said, it seems plausible to read ends of justice in connection with the necessary ends that are at issue in Kant’s moral beliefs. Both for Kant and for Crusius, these ends rest on an obligation that necessarily binds us to act a certain way.Footnote ²⁷ There is an important difference, though. Whereas for Kant the obligation in question is grounded solely on the form of our rationality, for Crusius it is based on our ‘indebtedness’ and ‘obedience’ to God. So Crusius’s legal obligation would not be categorical according to Kant, even though it is an attempt to grasp an obligation that is necessary and independent of subjective ends. Still, there are significant similarities between the two accounts. Since we know that Kant possessed copies at least of the Anweisung and the Entwurf in his library (cf. Warda Reference Warda1922: 47) and refers to Crusius’s Logic in his lectures (Log-W, 24: 796), and since scholars have already pointed out Crusius’s influence on Kant in other respects (cf. Tonelli Reference Tonelli1964), the easiest and most plausible way to account for these similarities is to hold that Crusius’s views on belief and practical justification had a direct influence on Kant.

Let me conclude this section by considering a further aspect of Crusius’s account of practical grounds for assent. For Crusius, the strength with which these grounds support our assent is influenced by two factors: first, by the importance of the ends we are pursuing, and second, by whether we can remain neutral with respect to a proposition, or whether we are obliged to make a choice regarding its truth or falsity and to act on that choice (cf. Crusius Reference Crusius1747: §362). The more important our ends and the lower the possibility of remaining neutral with respect to a proposition, the stronger the normative power of those ends to justify our assent and motivate our action (cf. Crusius Reference Crusius1747: §414). Kant does not discuss these two issues in his account of practical justification, even though, on the one hand, his distinction between pragmatic and moral beliefs can certainly be taken to imply a scale of importance among ends – a scale that we also find in Crusius’s distinction between ends of prudence and ends of justice. On the other hand, although Kant does not explicitly mention that the possibility of remaining neutral with respect to a proposition influences our practical justification in assenting to it, this does seem to be the upshot of some of his examples. Consider again Kant’s example of the doctor. Here, the matter of whether the doctor’s choice to act on the belief that the patient has consumption is necessary or avoidable is clearly relevant to his justification.

4. The Phenomenology of Practical Belief in Crusius and Kant

According to my reconstruction, both Crusius and Kant regard practically justified beliefs as beliefs that we are ‘rationally required’ to accept given certain ends we pursue and the lack of evidence against those beliefs. What does this mean? In what sense can I be ‘required’ to have a belief? Take again Kant’s example of the doctor. Why is the doctor ‘required’ to believe that the patient has consumption? It seems plausible to say that he is required to act on the assumption that the patient has consumption, but this should not have any implications for what his psychological state must be.

Andrew Chignell has tried to make sense of Kant’s claim by suggesting that Kant could indeed mean that, in order to act rationally, the doctor is required to firmly assent to the proposition that the patient has consumption. This should, however, not imply that the doctor is required to be in a psychological state in which he ‘feels’ that the proposition is true (Chignell Reference Chignell2007b: 341). According to this approach, the attitude we are required to have in Kantian belief is similar to what in contemporary philosophy is called ‘acceptance’: ‘a voluntary species of firm assent that motivates assertion and action in a certain context, but which is justified by its nonepistemic merits rather than by objective grounds, and so lacks the characteristic phenomenology of belief (the involuntary disposition to “feel” that the proposition is true, for instance)’ (Chignell Reference Chignell2007b: 341).Footnote ²⁸ Chignell remains neutral on the issue of whether Kantian belief involves full-blown belief or only acceptance, understood in this sense (Chignell Reference Chignell2007b: 342), but he tries to establish whether Kant’s account of belief is defensible, once we understand it as only involving acceptance. Does Kantian belief only involve acceptance, or does it involve full-blown belief? If the latter, what could it mean to say that the doctor in Kant’s example is ‘required’ to believe that his patient has consumption? Approaching this question from the standpoint of Crusius’s influence on Kant suggests that Kantian belief might indeed involve full-blown belief, or at least a certain ‘feeling’ of truth.

Crusius describes assent as a mixed ‘effect’ (Wirkung) of our intellect and will (Crusius Reference Crusius1747: §445; see also Crusius Reference Crusius1744: §§335–6). This seems to mean that assent is ‘caused’ by the recognition of the grounds for assent that we have,Footnote ²⁹ grounds that lie both in the intellect and in the will.Footnote ³⁰ In this context, grounds of the intellect encompass both demonstrations and proofs through probability, whereas grounds of the will come from the relationship between a proposition and our ends. Crusius makes clear that treating grounds of the will as possible causes of assent does not open the way to the view that we are easily victims of wishful thinking. Crusius’s idea is that it is in the nature of our will to follow the intellect as a guide (Richtshnur) in our actions. Our will is naturally inclined not to act on the basis of propositions against which there are grounds of the intellect. Moreover, when grounds of the intellect are lacking, the will motivates us to search for them in order to better decide how to act.

Why are these remarks relevant to determining the phenomenology of practical belief according to Crusius? First, the claim that assent is an ‘effect’ of both grounds of the intellect and grounds of the will suggests that it is understood as a psychological state caused by our representation of those grounds. Second, grounds of the intellect and grounds of the will are described as the two elements that contribute to the occurrence of the same state. This seems to imply that an assent that is essentially based on grounds of the will (like beliefs based on ends of prudence or ends of justice) is not distinguishable from an assent based on grounds of the intellect when it comes to the features it possesses as the psychological state that it is. What distinguish these states are rather the grounds that produce and justify them. Therefore, if ‘feeling’ that a given proposition to which we assent is true is crucial to that assent’s being based on grounds of the intellect (considered as the psychological state that it is), this feature should be part of an assent based on grounds of the will as well, since the latter is not distinguishable from the former when considered simply as a psychological state.

Where does this leave us? Does this mean that for Crusius we are rationally required to ‘feel’ that a certain proposition is true when it describes conditions for realizing ends that we pursue? It is not necessary to ascribe this implausible view to Crusius. Rather, it is more accurate to view Crusius as making two different points. The first concerns the rationality of our actions and beliefs. Here, the idea is that in order to pursue an end rationally, I must believe that the conditions that make the realization of that end possible obtain. This need not mean that we are ‘rationally required’ to ‘feel’ that those conditions obtain (or that the proposition describing those conditions is true). It may simply mean that if we act as if those conditions obtain without believing that they do, we have an inconsistent system of attitudes. If this is true, it means that if we actually have the belief in question and so ‘feel’ that the conditions described by the belief obtain, then our belief can be considered rational and justified.

The second point concerns the psychology of belief based on practical grounds. It provides an explanation of how grounds of the will can cause an assent in which we ‘feel’ that a certain proposition is true. We have seen that Crusius defines assent in general as a state in which we take a proposition to be true and are ‘ready to act’ according to that proposition (Crusius Reference Crusius1744: §334, Reference Crusius1747: §444). Let us assume that Crusius thinks not only that assent implies ‘readiness to act’ according to a proposition but also that in order to be ‘ready to act’ according to a proposition we must assent to it. Understood in these terms, Crusius’s account of assent in general involves a claim regarding our psychology: we cannot be ‘ready to act’ according to a proposition if we do not assent to it, if we do not believe – in the contemporary sense of the word – that the state of affairs it describes obtains.Footnote ³¹ The psychological story involved in Crusius’s account of belief based on practical grounds might be the following. In these cases, we have an obligation to act according to a proposition that, on the basis of ‘grounds of the intellect’, is only probable. It is, however, a fact of our psychology that we cannot be ‘ready to act’ according to the proposition in question if we do not believe that it is true. Therefore, the fact that we recognize that we are under an obligation, whether prudential or legal, counts as a ‘ground of the will’ that causes us fully to assent to the proposition in question. Otherwise, we would not be in the psychological state of being ‘ready to act’ according to our obligation. Of course, saying that ‘grounds of the will’ can exert this causal force on our assent does not mean that we can ignore ‘grounds of the intellect’. The proposition we end up believing must be probable according to those grounds. Crusius can thus explain how grounds of the will can ‘cause’ us to be in a psychological state in which we are ‘ready to act’ according to an obligation and consequently also ‘feel’ that a certain proposition is true.

Since Kant claims that belief, on his technical understanding of the term, is ‘sufficient for action’ (JL, 9: 68, n.; see also Refl 2793, 16: 513), it is at least plausible that his story about the psychology of belief approaches the one I have reconstructed with the help of Crusius. This seems to be confirmed by a passage in the Jäsche Logic: ‘The greater a man’s moral sentiment, the firmer and more lively will be his belief in all that he feels himself necessitated to accept and to presuppose out of moral interest, for practically necessary purposes’ (9: 70). Note that moral sentiment can affect the liveliness (Lebendigkeit) of a belief, where this has the consequence that we ‘feel’ we are necessitated to accept the truth of a certain proposition. This passage points to a view close to Crusius’s, where ‘grounds of the will’ can ‘cause’ us to be in a psychological state that enables us to act as we ought to. In Kant’s general account of assent, this can explain how belief, when considered merely as a psychological state, can ‘feel’ like other forms of firm assent, such as knowledge. Of course, this does not mean that, subjectively, we do not have means to distinguish between knowledge and belief. We can distinguish between knowledge and belief because we recognize that the grounds we have for each are different in nature. This is compatible with saying that these forms of assent are indistinguishable if we limit ourselves to what they ‘feel like’.