1. Introduction

There are a number of provocative themes that run throughout Davidson’s writings on the mind.¹ These are that the mind cannot be reductively explained in terms of the more “basic” theory of neurophysiology, that there can be no scientific regimentation of folk psychology (certainly no law-based, nomothetic, theory concerning the explanation and prediction of human thought and behavior), and that the deliverances of the most advanced physical or neurophysiological theory can neither replace nor even transform the folk psychological concepts and principles we all (implicitly) know and use. These themes are grounded in two principles underlying Davidson’s account of how we understand the thoughts, drives, and actions, of our fellow neighbors: first, that to have a mind is to have a network of intentional states that is, for the most part, rationally coherent, and second, that the content of our intentional states is holistically determined – in other words, the meaning of a mental state attributed to an individual depends on the inferential role(s) it plays within the network of mental states that constitute the individual’s mental life. The linchpin, Davidson declares, that ties together these two elements is “the irreducibly normative element in all attributions of attitude” (1974, 241). Thus, the reason why thought and behavior cannot be explained in terms of non-intentional, physical, vocabulary comes down to a certain “normative element” constitutive of our interpretation and attributions of the propositional attitudes. Clearly this normative element plays a pivotal role. But in spite of its significance, it is highly obscure and insufficiently understood. Indeed, there have been no serious attempts to systematically examine what, exactly, the normative element amounts to.

The aim of this paper is modest. I lay out several accounts of Davidson’s attempt to characterize the “normative element,” including one misleading characterization that is useful to smoke out. As we shall see, Davidson’s discussion does not lend itself to just one way of cashing out the normative component of intentionality. I end the discussion with a look at how the different manifestations collectively contribute to a more comprehensive understanding.

2. The Constraining Principles

Sprinkled throughout Davidson’s writings are the following principles that characterize the normative element:

i)

It is consistent with the descriptive nature of our belief reports and claims about meaning (1973, 254).

ii)

It is constitutive of thought and our intentional ascriptions, and is therefore an ineliminable component (1974, 237).

iii)

It is distinctive of thought and our intentional ascriptions (1970, 222).

Let us look at each in turn. Descriptive sentences have the logical form of fact-stating, truth-conditional, assertions, while prescriptive statements, under emotivist or expressivist analyses, have the logical form of non-assertoric endorsements, where “X is good” is analyzed as “Hurrah for X!” As Davidson maintains, and reasonably so, intentional ascriptions are like statements about the physical world in that they “are true or false in the same way” (1973, 254). Thus, “John believes that planets orbit around the sun,” or “By ‘sun’ John means the sun,” are just as declarative and truth-evaluable as, “Planets orbit around the sun.”² The second constraint amounts to the claim that the normative element is a necessary feature of our thoughts. Thus, as the relation of determinate to determinable is an essential feature of our application of color concepts, the normative element is an essential feature of the propositional attitudes we ascribe to our neighbors. Within the work of Davidson, this means that we must interpret our fellow neighbors according to the Principle of Charity: our attributions must make our neighbor’s thoughts and actions rational for the most part and her beliefs mostly true. The final constraint is that the normative element is unique to our ascriptions of the attitudes. Importantly, it is not a feature of physical and non-intentional ascriptions. So the claim is that "John believes that planets orbit around the sun" involves the normative element in question whereas nothing comparable is to be found in the claim, "Planets orbit around the sun.”

3. What the Normative Element Is not

With these constraints in mind, let us consider several explications of this elusive “normative element.” Here is one suggestion that Davidson himself considers but then rightfully rejects:

[W]hen we call an event an action, we are not, or not merely describing it, but are also judging it as good or bad, blameworthy or reasonable. (1973, 254)

On this view, the attribution of an intentional state is itself a normative remark or always accompanied by a conscious act of evaluating their contents. So on this view, we endorse or disapprove of the states we attribute [“‘Max believes that we should always help the poor.’ How noble!”], or laud or criticize them for being rational or unreasonable [“‘Mariel has concluded that there is no highest prime number.’ How sensible!”]. But this suggestion is precisely what the first constraint rules out. Now, we as attributors certainly judge a person’s beliefs on occasion (especially among philosophers), but it overgeneralizes what is only a fragment of our practices. Most of the time we just report with complete neutrality; indeed, we can do this for those very same attributions we color with the added value judgment.

4. What the Normative Element Is

There are four accounts that jointly compose Davidson’s normative element. Let us begin with what I call the evaluability thesis:

Psychology, if it deals with propositional attitudes … cannot be divorced from such questions as what constitutes a good argument or a valid inference, a rational plan, or a good reason for acting. (1974, 241)

I call this the evaluability thesis since the main idea is that our thoughts and actions are such that they are evaluable (as justified, valid, appropriate, and so on). This thesis does not run afoul of the first constraint for the plain reason that the normative element, under the evaluability thesis, is located in the thing that is attributed, not the attribution itself. Honoring other constraints, however, calls for emending the thesis. First, to conform to the second constraint, evaluability needs to be promoted to the status of an essential property. (Notice that evaluability would be a second-order property, for what is essential is the fact that mental states are (rationally) evaluable, not whatever first-order rational property a given state may have. In words, Max’s belief that we should always help the poor may have the property of being a valid conclusion (given Max’s other beliefs), but that belief does not have that first-order property essentially; rather it has the second-order property of being rationally evaluable as its essential property.) Conforming to the third constraint requires an explanation of what is so distinctive about the evaluability of the attitudes; after all, many non-intentional objects, such as table settings, clocks, and hearts and livers are also evaluable and, arguably, essentially so. The evaluability thesis, then, requires supplementation with other theses that can jointly honor the third constraint.

A different account is what I call the use thesis, so called because the idea is that we, as attributors, wield principles of rationality [of deductive and inductive logic, decision theory, and standards of veridical perception] in order to:

… criticize and advise others, or to modify our own beliefs and choices. (1990, 24)

Unlike the evaluability thesis, which locates the normative element in the objects of interpretation – the mental states themselves – here the normative element is located in the attributor’s demonstrable use of the various epistemic norms. Such a concrete application of the norms serves to guide (for the better) the thoughts and actions of others as well as one’s own self. Again, this thesis does not violate the first constraint, but like its predecessor, it needs further emending to satisfy the other two. As it stands, it does not explain how our practical use of rational principles is either constitutive of the attitudes or distinctive of them. But since it does not violate the constraints, we should not conclude that the use thesis is wrong; rather it indicates that it just cannot stand on its own.

The third account explicitly draws upon the two fundamental assumptions of Davidson's theory of the mind: first, that a mind is, for the most part, rational, and second, that the contents of our mental states are determined holistically. The rationality assumption is that we must interpret the content of a belief by placing it within a larger pattern of beliefs (desires and actions) whose contents rationally cohere with the content of the belief we wish to interpret. And the holism assumption is that a mental state derives the content that it has by playing its unique inferential role(s) within the network of the agent’s total mental states. These assumptions explain the following important aspect of Davidson’s normative element:

If someone believes that Tahiti is east of Honolulu, then she should believe that Honolulu is west of Tahiti. For this very reason, if we are certain she believes Honolulu is west of Tahiti, it is probably a mistake to interpret something she says as showing she also believes Tahiti is west of Honolulu. It is probably a mistake, not because it is an empirical fact that people seldom hold contradictory views, but because beliefs and other attitudes are largely identified by their logical and other relations to each other; change the relations, and you change the identity of the thought. (1990, 24)

Here, the normative element consists of the individuative role played by the norms or rationality. If it is the case that changing the inferential relations between the mental states of an individual is tantamount to changing their contents, then inferences that are correct to draw, the actions one rationally ought to perform, the beliefs that are evidentially or deductively justified, constitute the facts that serve as the individuation conditions for the content of an attributed thought. And as the conditions of individuation, these facts are essential to an individual’s mental states and their attributed contents. Given the individual’s other intentional states the content of a thought or the type of action performed is determined by its rational coherence with the totality of the agent’s mental states. It is clear that this third account fully honors the second constraint. If the norms of rationality and the norms of concept application have an individuative role, then these norms clearly have a constitutive status. That it also satisfies the third constraint will be easier to see once we have the fourth account at hand.

So let us consider the final piece of the overall picture, which I call the reflexivity thesis:

Whatever is studied, the norms of the observer will be involved. But when what is studied is the mental, then the norms of the thing observed also enter. When thought takes thought as subject matter, the observer can only identify what he is studying by finding it rational – that is, in accord with his own standards of rationality. The astronomer and physicist are under no compulsion to find black holes or quarks to be rational entities. (1990, 25)

Here, we approach the heart of what Davidson gestures at when he talks about the normative component of the attitudes. This account satisfies the third constraint, and it is crucial to understand how. As Davidson points out, the normative element ultimately has its roots in the object of the interpreter’s inquiry, which is another mind. Unlike black holes and quarks, which do not conform to norms, let alone the norms of rationality, a mind, by its very nature, has to conform to the norms of rationality. Otherwise, we are not dealing with a mind, should no or too few norms of rationality apply. Black holes and quarks certainly conform to laws – nomological principles – that support statements like “Light ought to bend in a black hole,” but such uses of “ought” have no normative implications (see Brandom 1994, ch. 1). The mental states that make up a mind, on the other hand, are such that they bear normative relations among each other, since their very contents are individuated by the norms of rationality (which is clearly stated in the third account). And the observer of a person’s mind must discern in the other’s bodily movements and vocal utterances a rational pattern that is itself a pattern to which the observer (attributor, appraiser) must subscribe. Hence, insofar as the norms of rationality are reflexive – they constrain both the mental states of the interpreted mind as well as the process of interpretation engaged by the interpreter herself – this aspect of the normative fully satisfies the third constraint.

Let us gather the four accounts together. On the view I develop of the normative element, it is not just one thing characterizable by one thesis. It involves several strands that are intricately tied together. Each of these strands are mutually supporting, so it is a mistake to try to derive a linear explanation of how they are organized. The norms of rationality that the observer uses must be the same norms that apply to the object of the observer’s inquiry (the fourth account). This both supports and is supported by the fact that mental states are essentially evaluable along rational lines (the first account), and the fact that we who interpret those states must use the norms of rationality to grasp the objects of our inquiry – mental states – (the second account). These considerations again both support and are supported by the fact that the norms of rationality play an individuative role with respect to mental states and their contents (the third account).