The Problem of Normativity in Kant s Philosophy of Logic Rebecca Victoria Millsop April 16, 2010 Submitted in partial fulfillment of the requirements for Honors in Philosophy at the University of California, Berkeley I want to express my gratitude to Professor John MacFarlane for being my advisor for this project; working with him has been an invaluable philosophical experience. I also want to thank the following individuals for their sustained support, advice, and comments on previous drafts of this essay: Gus Holcomb, Eugene Chislenko, Yuan Wu, Laura Davis, Justin Bledin, Nathana O Brien, Devin Rusky, Lara Krisst, Paolo Mancosu, Branden Fitelson, Hannah Ginsborg, Daniel Warren, Markus Kohl, and my mom. I would also like to thank the audience in Branden Fitelson s Logic, Formal Methods, and Epistemology seminar, as well as Berkeley s Undergraduate Philosophy Forum.
Contents 1 Introduction 1 2 A Law as a Norm 4 3 Kant s General Logic 7 3.1 Abstraction and Necessity............................ 7 3.2 The Constitutive Interpretation......................... 10 3.3 The Normative Interpretation.......................... 12 4 The Moral Law and the Laws of Logic 22 4.1 Both Normative and Non-Normative...................... 22 4.2 The Two Standpoints and Relational Normativity............... 26 4.3 Analogy...................................... 30 4.4 Pure and Applied as Noumenal and Phenomenal................ 36 5 Conclusion 38 A Appendix: Kant s Transcendental Idealism 40 A.1 Cognitive Faculties and the Discursivity Thesis................ 40 A.2 The Two-Aspects Interpretation......................... 42
1 Introduction Throughout his discussions on logic, Immanuel Kant claims that the laws of logic are normative for thought; in other words, logic as a science does not tell us how the understanding is and does think and how it has previously proceeded in thought, but rather how it ought to proceed in thought (Jäsche Logic, 14), where the understanding is taken to be our cognitive capacity for thought in general. If the laws of logic provide us with the correct way to think, then it should be possible to think incorrectly, or deviate from those laws in some way. Christine Korsgaard makes this claim clear when she states, There is no normativity if you cannot go wrong (Korsgaard (1996b, 162)). Thus, Kant must be able to explain how logical error is possible. This is a problem as he also claims that the laws of logic are in some way essential to, or constitutive of, the activity of thought, and essential laws cannot be violated in the way a norm can be violated. He admits this problem in the following passage: It is hard to comprehend how a power can deviate from its own laws, since it acts only according to certain laws. If these laws are essential, then the power cannot deviate from them[;] if, then, among the formal laws of the understanding that logic expounds an essential one is possible, then the understanding cannot deviate from it. (Vienna Logic, 824) Error is only possible when the sensibility, which is the capacity (receptivity) for receiving representations through the mode in which we are affected by objects, (Critique of Pure Reason, A19/B22) somehow interferes with the activity of the understanding. The understanding does not err in its own right, and our thought must, in some way, adhere to the laws of logic. Otherwise, it cannot be considered a product of the understanding: A man can never err completely and utterly... Total error would be a complete opposition to the laws of the understanding. But then it could not arise from the understanding, which can produce nothing that conflicts with its nature... (Vienna Logic, 825-6). Thus, any product of the 1
understanding must comply with its own laws, which just are the laws of logic. But now it seems that the laws of logic do not possess the desired normative force; it seems that there are at least some laws of logic that describe how we do think because they are essential to the act of thought itself. We have run into what seems to be a serious problem for Kant s claims regarding the nature of logic. As described above, Kant makes the following claims: (N) The laws of logic are normative for thought. (Jäsche Logic, 14) (ES) There are laws of logic which are essential for, or constitutive of, the act of thought. (Jäsche Logic, 13) (V) A law that is essential for the activity of the understanding (or any faculty by itself) cannot be violated. (Vienna Logic, 824) And we also want to hold that (ER) For a law to be normative, it must be possible to violate it in some way. In other words, an ought implies can err. (V) and (ES) together imply that there are laws of logic which cannot be violated. This along with (ER) implies that there are some laws of logic which are not normative for thought, which contradicts (N). How are the laws of logic supposed to be normative if there is no possible way of violating them? The goal of this essay is to propose and evaluate an interpretation of Kant s logic that explains all four claims above while also explaining away the contradictory consequence that apparently follows from the conjunction of these claims. My interpretation will focus on the taxonomic distinctions that Kant makes throughout the Critique of Pure Reason and his Lectures on Logic regarding the nature of the different kinds of logic. 1 It is generally assumed that the laws of logic that Kant describes as both normative and essential for thought are 1 The Lectures on Logic are a collection of four texts from different periods of Kant s career as a lecturer on logic. The first three texts are transcripts from his lectures, and the fourth, titled the Jäsche Logic, is the textbook that he had a student compile based on the lecture notes he had accrued throughout his career. The last two parts, the Vienna Logic and the Jäsche Logic, are the two I will use in this essay, as they were written near or during the periods of his more mature critical philosophy. 2
those laws Kant defines as pure general laws of logic; however, I claim that this assumption is not textually based and leads to the contradiction described above. Instead, I will argue, the laws that Kant describes as normative for and essential of thought are the laws of general logic; it is because we can then represent these laws as either pure or applied that these laws have the dual nature described. The laws under the broader description of general can be thought of as both normative and essential. In the end pure general and applied general logic do not consist of two separate kinds of logic, rather they are two distinct ways of representing the same laws. The pure general laws are strictly essential for, or constitute of, thought, while the applied general laws are those that we can violate and those that Kant takes as normative for thought. In the end, I claim that the normativity of the pure general laws of logic is at best externally conferred, rather than essentially inherent (Tolley (2006), 375); our nature as thinking human beings is such that the laws of pure general logic are essential for thought yet our cognitive capacities allow for empirical-psychological rules to override the laws of pure general logic while retaining the illusion of thought. It is because we are beings with these different cognitive capacities which are in relation to one another that the laws of logic are viewed as normative. I will begin by explaining what is to be expected of a normative interpretation of such laws, how a constitutive interpretation of Kant s laws of logic does not meet these expectations, and then go on to discuss how different normative interpretations of Kant s laws of logic attempt to meet these expectations. I then go on to consider how the normative interpretation may be justified in light of an analogy with Kant s description of the moral law as constitutive and normative for action. In making this analogy I will emphasize the two standpoint view that Kant puts forth in the Groundwork of the Metaphysics of Morals, and conclude that the moral law is not essentially normative but rather normative in relation to certain kinds of beings, or relationally normative. My goal from that point on will be to show how the two-standpoint view in the moral case is analogous to the distinction between 3
pure and applied general logic, and explain how we can make sense of the normative claims that Kant makes in light of this distinction. I will end the essay with a discussion of how this interpertation allows for all four of the claims mentioned above to coexist while eliminating the contradictory conclusion. 2 A Law as a Norm If any normative interpretation of Kant s laws of logic can be made sense of, the requirements for a law being normative instead of descriptive must be set forth. A law that is descriptive, for example, the law stating that the speed of light is 186,282 miles per second, simply describes the way light is. The law doesn t make any normative claims about how fast light should travel, it states how fast light does travel. In an attempt to get at what it is for a law to be a norm on Kant s account I will consider the two conditions that Clinton Tolley, in his essay Kant on the Nature of Logical Laws, (Tolley (2006, 375)) suggests: 2 (1) Error Any being subject to the law must both be able to succeed and be able to fail to act (or be) in accordance with the law. (2) Identity Even when a being subject to the law does not act in accordance with the law, the subject still identifies itself as being subject to the law regardless. These are the conditions that he claims, on Kant s account, give a law normative force; such laws prescribe correct action to a being that is subject to the law. To illustrate these conditions I will use the example of fidelity as a norm imposed on a subject in a monogamous 2 Tolley provides a third condition, Bindingness: The laws must retain their validity or bindingness over their subjects regardless of the (lack of) actual adherence to the norms by their subjects though, to be sure, there must be the possibility of such adherence (to uphold the traditional formula that ought implies can ) Tolley (2006, 375). I will not be using this condition because I do not see how it adds significantly to the set of conditions, as much of what is important in this third condition is captured by the second condition. And there are several counterexamples to its validity. For example, consider the following law: all residents of Berkeley cannot leave Berkeley. But this law is only binding on people in Berkeley, thus if you leave you are no longer bound by the law. I believe there are ways of addressing this counterexample; however, the best way to address the issue is by referring to the second condition, identity. Thus, I will exempt this third condition from my discussion. 4
relationship, that typically imposed by the bond of marriage. The cases I am discussing are those in which two subjects have agreed that the law of fidelity is binding over their relationship. The first condition, Error, is obviously met: either of the two subjects are capable of an act of infidelity, while also capable of adhering to the law. A being who never engages in a monogamous relationship will not be subject to the law of fidelity, just as inanimate objects can never be subject to this law. This condition is the same one discussed in the introduction; we can think of it as saying that ought implies can err. The second condition, Identity, is quite clear in the case of fidelity; if one of the subjects fails to act according to the law of fidelity, they feel guilt (to some degree) because they still identify themselves as subject to the law. Although this condition clearly holds in the case of fidelity, there are other cases where this is less clear; for example, the case of traffic laws. Perhaps I am unaware of a specific traffic law; assume this is the law that says that I must drive under 30mph on the city streets. It is difficult to say that I identify myself with this law given that I am unaware that it is binding on me. It is clear that this law is still normative for my action, as the content prescribes how fast I ought to drive, not how fast I am driving. Thus, this seems to be a counterexample to the second condition. However, even though I do not identify myself as being subject to this law, it is the case that I identify myself as being subject to all of the pertinent traffic laws. As Tolley notes, this condition is important, as it implies that evaluative ascriptions in light of norms (e.g., x as in or out of accord ) institute a division within some otherwise well-defined class (Tolley (2006, 375)). Given that I take myself to be subject to the traffic laws, when I disobey some law, even those I am unaware of, my action is evaluated as being out of accord with the laws I identify myself with. There is a stronger response against the second condition, however. Instead of being unaware of the law, I may consider myself above the law. In this case I am aware what the law expects of me, but I do not identify myself with the law whatsoever. The law in this case 5
is still binding over my action as a driver, and thus it seems like the law is still normative despite my lack of identification with it. This seems like a clear case where a law is a norm yet does not satisfy the second condition. This examples brings out an important distinction between normativity in general and normativity on Kant s account; this distinction is further brought out by considering the differences in the fidelity and traffic laws cases. In the fidelity case the subject has bound themselves to the law and in that way the subject has internalized the law as a norm. In the case of the traffic laws the laws are externally conferred on the subject, which is why one can opt out of identifying with the law despite the fact that other people still view the law as normatively binding on that subject. In a discussion of Kant s account of the moral law being normative, Korsgaard makes the following claims: Here we come to an important distinction, between norms that are constitutive of, and so internal, to the activities that they claim to govern, and norms that are external to those activities... External norms give rise to further questions, and space for skeptical doubt. But if we can identify something as an internal norm, the question why you should conform to the norm answers itself. (Korsgaard (2008, 61)) Given this description, our example of traffic laws can be seen as external norms; the activity of driving is not constituted by the traffic laws. Whereas the law of fidelity is constitutive of a monogamous relationship. To be in a monogamous relationship is to identify yourself with the law of fidelity, and in this way the law is internally normative. Thus, it seems that the condition of identity is a condition of a law being a constitutive norm. And the goal of this essay is to interpret Kant s laws of logic such that they can be taken as both constitutive of and normative for thought. So although the condition of identity is not a condition on normativity in general, it is a condition for the kind of norm that we are considering given Kant s account. I shall therefore refer to these conditions throughout the rest of the essay when considering the different interpretations of Kant s laws of logic. 6
3 Kant s General Logic Before addressing the constitutive and normative interpretations I want to briefly describe the nature of Kant s logic, as well as describe one of the two important taxonomic distinctions that I will be focusing on throughout this essay: the distinction between general and particular logics. The second distinction, the distinction within general logic between pure and applied logic, will be discussed in detail in later sections of this essay. It is clear that the laws of logic that Kant describes as both normative and essential for all thought are to be of the general kind, not of the particular kind, as in all the relevant texts the important claims he makes are all in regard to general logic. However, it is generally assumed that when Kant discusses the laws of logic he is discussing pure general logic only, and that we should set applied general logic aside because it is not really logic. As we shall see below, I claim that the distinction between pure and applied is the most important for making sense of a constitutive and normative interpretation of Kant s logical laws. However, before delving into that topic, I will first examine the general/particular distinction and try to get at what is definitive of general logic. Through this examination the ambiguity in Kant s description of general logic as necessary will be illustrated; does he mean the laws of general logic are necessary in an essential way, or in a normative way? It is not clear because he never spells it out for us. I will walk through several passages in the Critique of Pure Reason and the Jäsche Logic in which he describes his distinction between general and particular logics, emphasizing the ambiguity of the term necessary. 3.1 Abstraction and Necessity Kant claims that general logic is distinguished from particular logic in that the former abstracts away from all objects of thought. It is because of this abstraction that logic 7
provides the necessary laws of thought. 3 After having abstracted away from all objects of thought whatever is left will be something that all acts of thought share without exception. The only exception would come about because of a particular kind of content, which would result in the use of the understanding that is particular. Kant states there is, for example, a use of the understanding in mathematics, in metaphysics, morals, etc. (JL, 12). When thinking about any one particular area, then, there are laws of a particular logic that are applicable. But general laws of logic are applicable to all thought. His descriptions of general logic are similar in both major texts: Now logic in turn can be undertaken with two different aims, either as the logic of the general or of the particular use of the understanding. The former contains the absolutely necessary rules of thinking, without which no use of the understanding takes place, and it therefore concerns these rules without regard to the difference of the objects to which it may be directed. (Critique of Pure Reason, B76/A52) If now we put aside all cognition that we have to borrow from objects and merely reflect on the use just of the understanding, we discover those of its rules which are necessary without qualification, for every purpose and without regard to any particular objects of thought, because without them we would not think at all. Thus we can have insight into these rules a priori, i.e., independent of all experience, because they contain merely the conditions for the use of the understanding in general, without distinction among its objects, be that use pure or empirical. (Jäsche Logic, 12) Given his descriptions, there are two main features of general logic: 1) Abstraction from all content of thought. 2) Because of this abstraction, we are left with the necessary rules of the use of the understanding. Thus, we must understand these two features in more detail, and how they are related to one another. The first thing to understand is exactly what Kant means by abstracting away from the objects of thought. He describes the abstraction as leaving only the relations between 3 Note that I will be using laws of thought synonymously with laws of general logic, as Kant does, unless otherwise stated. 8
thoughts, or, as he describes this, the mere form of thinking (Critique of Pure Reason, B78/A54). For example, if we abstract away from the content of the following inference: All objects on the table are green. This cup is on the table. Therefore, this cup is green. we are left with the relation of these thoughts to one another. This leaves us with a skeleton, so to speak, which can be classified as a specific form of logical syllogism that can then be analyzed. The content of these thoughts are the objects related to the concepts object on the table, green, and cup. The understanding has related these contents to one another according to the following form: 1) All F s are Gs. 2) S is a F. S is a G. Premise 1 is the major premise; in this case it is a universal proposition; premise 2 is the minor premise; in this case it is an affirmative proposition. This is an instance of the first figure of inference, from this the following rules for the three latter figures emerge (Jäsche Logic, 127). Much of his discussion of general logic is put in terms of the formal aspects of thought, and thus it is important to explain the relationship between Kant s logic being general and its being formal. This is discussed in detail in John MacFarlane s paper Frege, Kant, and the Logic in Logicism, in which he argues that Frege and Kant can be considered as having the same view of the nature of logic because they both claim that logic should be general; where general here is of a normative sort discussed below. In the process of providing his argument, MacFarlane shows that the formality of Kant s logic is actually a consequence of the generality of logic, and not an independent feature. Thus, whenever we discuss the formal 9
nature of logic, it is imperative to realize that this formality stems from logic s generality. Although we can separate them conceptually, it is not possible to have one without the other. At this point we understand that general logic abstracts away from the content of thought, and the laws of logic are in some way derived from this abstraction. However, this does not further clarify our understanding of the way Kant uses the term necessary in his description of general logic. Given different pieces of textual evidence we can interpret Kant as intending to mean different things. Thus, at this point there are two ways to understand what Kant means by absolutely necessary in the phrase absolutely necessary rules of thinking. There is a stronger interpretation of necessary which suggests that every act of thought adheres to these laws of logic; this coincides with the constitutive interpretation. If this is correct then every thought is logical. As we shall see, this interpretation is supported by sections of the texts where he claims that the faculty of the understanding, when considered by itself, cannot think in opposition to its own rules. The weaker, normative interpretation takes this necessity to imply that the laws of logic are normatively binding on the rational individual, and given the conditions for a law being a norm this interpretation attempts to make sense of logical error. 3.2 The Constitutive Interpretation The constitutive, essential interpretation claims that thought must adhere to the laws of logic if the action is considered thought at all. The term constitutive can be understood as, in a sense, descriptive. Just as there is a law which states how fast the speed of light is, the laws of logic, on this interpretation, describe how the understanding works. Thus, if the understanding is in use it must be the case that the thought produced by the understanding abides by, or adheres to, those general laws of logic. This is one way of explaining what it means for rules to be necessary without qualification; in other words, this interpretation takes the laws of logic to be necessary conditions for every act of thought. This interpretation 10
takes Kant s laws of logic as failing to meet the first condition of normativity, Error. Although this interpretation may seem a bit extreme, and seems to go against what we ve seen Kant says about the normative nature of logic, it is supported throughout both the Critique of Pure Reason and the Jäsche Logic when he discusses how the understanding cannot err because it necessarily adheres to its own laws:... a total error would be a complete opposition to the laws of the understanding and of reason. But how could that, as such, in any way come from the understanding and, insofar as it is still a judgment, be held to be a product of the understanding. (Jäsche Logic, 54)... it is hard to comprehend how error in the formal sense of the world, i.e., how the form of thought contrary to the understanding is possible, just as we cannot in general comprehend how any power should deviate from its own essential laws. (Jäsche Logic, 53) hence neither the understanding by itself (without the influence of another cause), nor the senses by themselves, can err; the first cannot, because while it acts merely according to its own laws, its effect (the judgment) must necessarily agree with these laws. (Critique of Pure Reason, B350/A294) The constitutive interpretation draws from these quotes that there are laws of logic that cannot be violated in thought. The understanding acts according to the laws of logic, and the resulting judgment must agree with them. This last quote is especially strong: the effect is not just assessable by the laws, the result does agree with these laws. 4 This interpretation requires an explanation of what it is for thought to be untainted by the influence of other faculties, and how it is that this purity relates to the strong necessity claim. The constitutive interpretation of necessity thus makes sense of the textual claims that Kant makes regarding error. However, it leads us to the unsatisfactory conclusion that logical error is impossible. And, as we discussed above, if logical error is impossible the logical laws 4 I draw this conclusion from Kant s usage of the phrase must necessarily in the passage. It may be argued that this phrase could be taken deontically, so that the passages implies that judgment ought to agree with these laws. This would imply some sort of normative interpretation. However, if this were the case, then there should be a deontic interpretation of can err in the preceding sentence. This does not seem correct at all, which is why I am taking must necessarily agree with these laws as a metaphysical necessity. 11
do not meet the requirement for being a normative law. As the nature of logical error is important to our discussion, I will briefly comment on Kant s claims regarding the existence of error. As pointed out above, the understanding itself cannot err, and neither can any faculty alone err. Instead, error is the product of some kind of interference with the use of the understanding. In the discussion of truth and error in the Jäsche Logic Kant claims that the ground for the origin of all error will therefore have to be sought simply and solely in the unnoticed influence of sensibility upon the understanding, or to speak more exactly, upon judgment (Jäsche Logic, 53-4). Sensibility is our capacity for taking in the world, and all cognition requires input from sensibility and conceptual activity from the understanding. Thus, the laws of logic only govern the use of our higher cognitive abilities, and our ability to think badly stems not from any fault of the understanding itself, but rather from its interaction with objects in the world. Another way of saying this is that error stems from how our passive faculty of sensibility comes together with our active faculty of understanding. It seems, then, given the connection we have been discussing between the ability to err and normativity, it follows that the laws of logic are taken to be normative only insofar as there is this possibility of this external interference. It could be that the seemingly normative nature of the laws of logic stems from some relation between the different cognitive faculties that beings like us possess. 3.3 The Normative Interpretation Despite Kant s claims about the laws of logic being essential for the activity of thought, throughout the Lectures on Logic he makes several explicit claims regarding the normative nature of these laws of logic. If this is the case then the laws of logic should meet the two conditions for normativity. The most oft-cited quote is from the Jäsche Logic, where in describing the laws of general logic he makes the following claim: 12
As a science of the necessary laws of thought, without which no use of the understanding or of reason takes place at all, laws which consequently are conditions under which the understanding can and ought to agree with itself alone the necessary laws and conditions of its correct use logic is, however, a canon. (Jäsche Logic, 13) This is the first place the word ought shows up in this text, and suggests that this ought is a consequence of the necessity of the laws of thought. Here Kant also talks of the laws of logic as also being conditions for the use of the understanding, and it is in terms of conditions that he speaks of the correct use of the understanding. Thus, general logic consists of the necessary laws of thought and from these laws the conditions for correct thought follow. Unless Kant is using the term ought in some non-normative way, it must be possible to make sense of the laws of general logic in some way to fulfill the two conditions provided above for a law to be a norm. Thus, we cannot rule out the possibility that normativity is somehow essential to the nature of general logic; the question at hand, however, is whether or not normativity is definitive of generality. At this point it is worth noting that in the Critique of Pure Reason Kant does not mention the normative nature of general logic explicitly, neither in the section On Logic in General nor anywhere else in the text. 5 If the generality being discussed were inherently normative, meaning that this generality could be explained by normativity, it could be expected to see use of the term ought in this discussion. However, in the Critique he explains generality only in terms of abstraction, not normativity. This will only make 5 By explicitly I mean the explicit use of the word ought. However, it could be argued that the normativity is explicitly referenced in both texts on when Kant talks about the laws of logic being rules for thought. For example, in the Critique of Pure Reason Kant describes these laws as the absolutely necessary rules of thinking (Critique of Pure Reason, B76/B52). And in the Jäsche Logic the laws of logic are those rules which are necessary without qualification (Jäsche Logic, 12). However, at the very beginning of the introduction to the Jäsche Logic Kant claims that Everything in nature, both in the lifeless and in the living world, takes place according to rules... Water falls according to the laws of gravity, and with animals locomotion also takes place according to rules. The fish in water, the bird in air, move according to rules (Jäsche Logic, 11). The laws of gravity are not normative, they are descriptive, and the fact that Kant is suggesting that these laws are to be considered rules shows that we should not take Kant s description of the laws of logic as rules to be indicating any sort of normativity. 13
sense if we find that this abstraction is itself inherently normative. Although it is left unclear exactly how these laws are also essential conditions for the activity of thought, the normative interpretation views Kant as going on to further forge the connection between necessity and normativity as inherent, as he claims in the Jäsche Logic: In logic, however, the question is not about contingent but about necessary rules; not how we do think, but how we ought to think... In logic we do not want to know how the understanding is and does think and how it has previously proceeded in thought, but rather how it ought to proceed in thought. Logic is to teach us the correct use of the understanding. (Jäsche Logic, 14) Here we can interpret Kant as drawing a parallel between descriptive laws and contingency, and normative laws and necessity. Thus, there is some connection between the laws of general logic and how we ought to think, although it is not described explicitly any further in his works. The laws of general logic, then, are normative for thought in some way, yet it has not been shown that this normativity is definitive of why general logic is to be thought of as general. Why should the mere abstraction from all content of thought result in normative rules? This normativity seems to be adding something to the idea of general logic, but Kant never specifies what. Thus, in order for the normative interpretation of necessity to hold, it must be able to explain how the nature of general necessity as abstraction from all content of thought entails the normative claims that we find in the Jäsche Logic. 3.3.1 MacFarlane s Normative Interpretation John MacFarlane, in his paper briefly discussed above, titled Frege, Kant, and the Logic in Logicism, advocates a version of the normative interpretation of Kant s general logic. He claims that normativity is definitive of Kant s general logic: The generality of logic, for Frege as for Kant, is a normative generality: logic is general in the sense that it provides constitutive norms for thought as such, regardless of its subject matter (MacFarlane (2002, 35)). In this 14
section I will present and evaluate MacFarlane s claims regarding the normative generality of logic, concluding that the outcome of his explanation of logical laws as constitutive norms is attractive, but that it is does not fully account for the second condition of normativity because it does not explain how the laws of logic are active in thought. MacFarlane first presents and explains his distinction between descriptive laws and normative laws, and then goes on to explain that Kant s logic must be of a normative generality because Kant s laws of logic are to be thought of as constitutive norms given the textual evidence just discussed above. He distinguishes between the two types of laws as follows: A normative law prescribes what one ought to do or provides a standard for the evaluation of one s conduct as good or bad. A descriptive law, on the other hand, describes certain regularities in the order of things typically those with high explanatory value. (MacFarlane (2002, 35)) Given this distinction, then, we can draw a parallel with the constitutive and normative interpretations of necessity associated with generality. A descriptive characterization of generality is going to align with the constitutive interpretation, wherein the laws describe some aspect of the process of thought which is necessary in that this description is of something which always occurs. A normative characterization of generality is one in which the laws are standards for the evaluation for some act in general; in other words, in order to perform such an act one must necessarily be assessable in light of those laws. The laws of logic are general in this way, MacFarlane claims, in that to count an activity as thinking at all is to hold it assessable in light of the laws of logic (MacFarlane, 37). The laws that are normatively general in this way are to be taken to be constitutive norms. MacFarlane provides an example of how the laws of physics are constitutive norms for making judgments about the physical world: [The laws of physics] have prescriptive consequences for anyone engaged in the game of thinking about the physical world: such a thinker ought not make 15
judgments that are incompatible with them. Indeed, insofar as one s activity is to count as making judgments about the physical world at all, it must be assessable for correctness in light of the laws of physics. In this sense, the laws of physics provide constitutive norms for the activity of thinking about the physical world. (MacFarlane (2002, 37)) MacFarlane claims that the laws are constitutive because they are unconditionally binding on the activity itself; in other words, the activity cannot be performed without those laws as standards for evaluation of the activity. For example, if I begin to posit the existence of flying pigs it is either that I have opted out of the activity of thinking about the physical world, and thus am no longer assessable by the laws of physics, or I am having incorrect thoughts about the physical world despite the fact that I am attempting to think correctly because I consider my thoughts to be assessable by the laws of physics. Thus, constitutive norms are also normative because it is possible to think incorrectly; it is only necessary that the thought be assessable in light of those norms. MacFarlane s goal in his essay is to show how both Frege and Kant think of logic as general in the normative sense, and most of his discussion of the normative characterization of generality is in light of Frege s claims. However, in the section titled Kant s Characterization of Logic as General, he quotes from Kant to support this claim. He quotes the distinction in the Jäsche Logic between general logic being necessary and particular laws being contingent, where Kant describes the laws of general logic as those without which no use of the understanding would be possible at all (Jäsche Logic, 12), a quote which we are now quite familiar with at this point. And, as we have seen, generality and necessity go together on Kant s account, although the nature of this necessity has been ambiguous. This is the interesting move: MacFarlane assumes that Kant s general necessity is normative. The necessary rules are necessary, not in the sense that we cannot think contrary to them, but in the sense that they are unconditionally binding norms for thought norms, that is, for thought as such. (Compare the sense in which Kant calls the categorical imperative necessary. ) Similarly, the contingent rules of 16
the understanding provided by geometry or physics are contingent, not in the sense that they could have been otherwise, but in the sense that they are binding on our thought only conditionally: they bind us only to the extent that we think about space, matter, or energy. (MacFarlane (2002, 47)) Surely we can make sense of this interpretation, as the normative interpretation is attractive for explaining the claims that Kant makes regarding the ought in the Jäsche Logic, but it is unclear still why we should interpret the necessity as normative. The main problem is that MacFarlane does not go on to mention the other more explicit normative claims that Kant makes about the laws of logic, nor does he go on to say anything else to support his interpretation. If the necessity under discussion is normative, then this will make sense of Kant s claims in the Critique of Pure Reason as well as the Jäsche Logic regarding the normative nature of general logic. But further justification is not provided. MacFarlane s idea of a constitutive norm is certainly an attractive one for explaining the nature of Kant s laws of logic. The laws have a constitutive component that makes them unconditionally binding on the thinking subject, which, it could be argued, meets the second condition of normativity, identity. His account definitely allows for error, which is the first condition for a law to be a norm, because his concept of a constitutive norm is defined in terms of assessibility. As mentioned above the laws of physics are normative in that in thinking about the physical world the subject s thought is assessable in light of the laws of physics. The laws of logic are unconditionally binding in that in order to be considered thinking at all, one must be assessable by the laws of logic. Thus, MacFarlane s interpretation assumes that the laws can be adhered to and not adhered to, which allows for error. However, it is not so clear exactly how MacFarlane s explanation accounts for the second condition. At first, it seems he could meet this condition through some explanation of the unconditional binding on the thinking subject; the laws of logic constitute the act of thought and thus are internally conferred laws. Yet MacFarlane s explanation of this binding is in 17
terms of assessability, which is a passive notion. It is unclear how the notion of assessability is supposed to make such a strong binding claim on the subject. Consider the case of the law of fidelity; my action is surely assessable in light of the law of fidelity, but that is because I identify myself as a being in a monogamous relationship. It is not that because my action is assessable in light of the law of fidelity that I identify myself as a being in a monogamous relationship. Assessability is a consequence of the two conditions of normativity, not the other way around. These norms should have the feature of being used in the assessment of whether or not a logical law is abided by, but it is unclear how this assessment can be definitive of the constitutive nature of the laws. In his explanation MacFarlane notes that the necessity inherent in the laws of logic is to be taken as analogous to the necessity inherent in the moral law. This suggests that the moral law is also unconditionally binding on action insofar as action is assessed in light of the moral law. However, a being subject to the moral law is not merely bound because their action is assessable by the moral law; the moral law actively guides the subject. The way in which the moral law is active will be discussed in more detail below. Given that MacFarlane s notion of assessability is a passive one, it does not account for the active force that the necessary laws of logic should play in Kant s account. On MacFarlane s interpretation, every act of thought is to be assessable by these norms, but Kant s description of the use of the understanding requires that these laws be active. These laws are constitutive of the activity of the understanding itself, thought, and the use of the understanding is a spontaneous activity which imposes conceptual form onto all objects of experience. Kant describes the spontaneous nature of the faculty of the understanding throughout his works, but most heavily in the Transcendental Logic of the Critique of Pure Reason: If we will call the receptivity of our mind to receive representations insofar as it is affected in some way sensibility, then on the contrary, the faculty for bringing 18
forth representations itself, or the spontaneity of cognition, is the understanding. (B75/A51) Concepts are therefore grounded on the spontaneity of thinking, as sensible intuitions are grounded on the receptivity of impressions. (A68/B93)... the combination of a manifold in general can never come to us through the senses, and therefore cannot already be contained in the pure form of sensible intuition; for it is an act of the spontaneity of the power of representation, and, since one must call the latter understanding, in distinction from sensibility, all combination, whether we are conscious of it or not, whether it is a combination of the manifold of intuition or of several concepts, and in the first case either of sensible or non-sensible intuition, is an action of the understanding... which can be executed only by the subject itself, since it is an act of its self-activity. (B130) In all of these passages the understanding is contrasted to the sensibility, in that the faculty of sensibility receives sensory input, which is why it is considered a passive faculty. The understanding, on the other hand, is spontaneous; it is an active combination of representations into concepts, judgments, and inferences. If the laws of logic are constitutive of the use of the understanding, just as the moral law is for the use of practical reason, then these laws must be guiding the act of thought. Thus, although this notion makes sense of the normative claims that have been discussed, there is something missing from MacFarlane s description; that something is an explanation of the active normative force that Kant s laws of logic seem to impose on all use of the understanding. This notion of normative force is itself not clear, but it can been seen how the necessity involved with Kant s laws of logic, if they are to be considered constitutive norms, must do more than act as passive standards from which all thought is assessed. 3.3.2 Korsgaard s Normative Interpretation Christine Korsgaard, in the chapter The Metaphysics of Normativity of her book Self- Constitution: Agency, Identity, and Integrity, discusses a somewhat similar notion to Mac- Farlane s constitutive norm, except her notion brings out the active nature of the laws of logic that MacFarlane s discussion lacks. She describes the constitutive norm as uncondi- 19
tionally binding because all actions under the power of these norms are necessarily guided by the constitutive norms; not merely assessable in light of them. Korsgaard uses the concept of a constitutive principle, which she describes as follows: In these cases what we say is that if you are not guided by the principle, you are not performing the activity at all. In the case of essentially goal-directed activities, constitutive principles arise from the constitutive standards to which they are directed.... To use a controversial example, it is a constitutive principle of thinking that you swerve when you see a contradiction looming ahead in your path. And in these cases, we can say that unless you are guided by the principle in question, you are not performing that activity at all. (Korsgaard (2009, 29)) On this account the principles are unconditionally binding, except we make sense of this in terms of a kind of self-imposed intention towards always acting, or thinking, according to those principles. In terms of Kant s general laws of logic, we can see this as saying that every act of the understanding strives towards adherence with the laws of logic. It is not just that our thought is assessable in light of these laws, which is also certainly the case on this interpretation, it is that the activity itself is guided by the intention of adhering to those laws. Korsgaard provides an intuitive example of how to further understand this notion of guided by: The presence of both a noun and a verb in an English sentence is constitutive of its being a sentence, that is, of its expressing a complete thought. Yet those of us whose work includes grading papers have all encountered the verbless string of words that wants to be a sentence and fails, and yet is not mere gibberish. There is such a thing as speaking English badly, and it is not quite the same as not speaking at all, although importantly it tends in that direction. For if you ignore the rules of English altogether, what you speak will simply not be English. (Korsgaard (2009, 30)) The act of writing the incorrect yet not gibberish sentence involves striving towards a correct sentence; when I am writing I am not actively aware of the grammatical rules of English, but my action is guided by those rules in that I am writing a sentence in English. The rules of 20
English grammar are not normative in that the product of the writing is assessable in light of these rules, but that the activity of writing an English sentence is compelled by those rules. If the laws of logic are constitutive principles in Korsgaard s terms then they should fulfill the two conditions for being norms for thought. It is possible to think incorrectly, yet one is compelled by the laws and is therefore identifying oneself with the laws while also being bound by them. The advantage that Korsgaard s account has over MacFarlane s is that is provides the sense that the laws of logic are active in relation to thought, they play some sort of role in creating the thought instead of being those laws that thought is merely assessable in light of. Despite this advantage over MacFarlane s interpretation, Korsgaard does not fully explain how it is possible to err on her account. It is the case that to think you must be guided by the laws of logic, but this notion of guided by is not made clear and this does not seem to explain why the laws of logic meet the first condition, Error. The laws of logic could constitute thought insofar as these laws guide the use of the understanding, while the understanding always succeeds in adhering to the laws. Thus, this interpretation succeeds in providing an account of the active nature of the laws of logic, but does not succeed in explaining how and why the laws of logic meet the first condition for normativity. My goal in this section has been to provide an example of the normative interpretation of the general nature of Kant s logic, show how it is inadequate as it stands, and go on to suggest ways which this description could be improved upon. Towards this goal I presented MacFarlane s account of Kant s general logic as definitively normative, and then showed how his interpretation, although plausible, is unsupported as it stands. I pointed out that MacFarlane s notion of a constitutive norm is the kind of notion we need to interpret Kant s laws of logic so that the problem described in the introduction can be avoided. Thus, the goal at hand is to find a way to further justify the existence of such constitutive norms for logic, 21
and how they stem from the general logic. I also pointed out that the notion of constitutive norm we are searching for should be of an active sort, somehow explaining the normative force that necessarily compels every thought towards logical adherence. MacFarlane draws an analogy between the categorical imperative in Kant s moral philosophy and the constitutive necessity inherent in the laws of logic, and on Korsgaard s account we find a similar analogy as well. In the next section I will explore how the moral law on Kant s account is to be interpreted as normative, constitutive, or both with the goal of drawing the proper analogy with the laws of logic. 4 The Moral Law and the Laws of Logic 4.1 Both Normative and Non-Normative Kant s moral philosophy is taken as a paradigm account of presenting a normative law. The categorical imperative states that one must act only in accordance with that maxim through which you can at the same time will that it become a universal law (Groundwork of the Metaphysics of Morals, 4:421). 6 The categorical imperative is absolutely general in that it is objectively necessary for all action without reference to another end. The normative nature of the categorical imperative is that it is an imperative, which is defined by Kant in the following way: All imperatives are expressed by an ought and indicate by this the relation of an objective law of reason to a will that by its subjective constitution is not necessarily determined by it (a necessitation). They say that to do or to omit something would be good, but they say it to a will that does not always do something just because it is represented to it that it would be good to do that thing. (Groundwork, 4:413) 6 Throughout the rest of the essay I will abbreviate the title of this work as Groundwork. 22