Issue #13 -- October 2001. Pp. 6-22
Non-Conceptually Contentful Attitudes in Interpretation
Copyright © by SORITES and Daniel Laurier
Non-Conceptually Contentful Attitudes in Interpretation
Daniel Laurier
Brandom's book Making It Explicit defends Davidson's claim that conceptual thought can arise only on the background of a practice of mutual interpretation, without endorsing the further view that one can be a thinker only if one has the concept of a concept. This involves (inter alia) giving an account of conceptual content in terms of what Brandom calls practical deontic attitudes. In this paper, I make a plea for the conclusion that these practical attitudes are best seen as intentional, but non-conceptually contentful. In particular, I argue that the hypothesis that Brandom's practical deontic attitudes are non-conceptually contentful wouldn't conflict with his view that non-conceptual intentionality is merely derivative. I then explore some of the implications which this hypothesis might have with respect to various forms of «intentional ascent».
§1.-- Practical Attitudes as Non-Conceptually Contentful
It is well known that for Davidson, for any individual to have conceptual thoughts at all, he must have the concept of belief, which is tantamount to saying that he must have the concept of a concept (the concept of correct and incorrect application of a concept). Since (on Davidson's account) one can only have this concept if one is an interpreter of the speech of others, it follows that only a community whose members are capable of mutually interpreting (i.e. attributing thoughts to) each other could count as a community of thinkers (and that there can be no such thing as a solitary thinker).
One useful way of reading Brandom's work is as propounding an extensive defense of Davidson's claim that conceptual thought can arise only on the background of a practice of mutual interpretation (and thus as endorsing a fundamentally interpretationist view of content, or as he would probably prefer to say, a «phenomenalist» view of content), but without endorsing the further view that one can be a thinker (and hence an interpreter) only if one has the concept of a concept (and hence the capacity to conceptually think that someone else thinks anything).
The leading intuition behind Davidson's and Brandom's common attitude towards matters of meaning and content is that an expression can mean anything only insofar as, and in virtue of the fact that, it is taken to mean it. In Brandom's favored words, this becomes the view that what one is committed or entitled to, and hence what it is appropriate or inappropriate for one to do or not to do, depends on what one is taken to be committed or entitled to. This intuition, however, is counterbalanced by the opposite intuition that one may objectively be committed or entitled to something, even when one is not taken to be so committed or entitled by anyone (and conversely, that one may objectively not be committed or entitled to what everyone, including oneself, takes one to be committed or entitled). The challenge is thus to provide a constructive account of norms that is yet capable of sustaining a real distinction between being correct or appropriate, and being treated as correct or appropriate (by anyone, including the community as a whole). It is one of Brandom's grand contention to have taken up this challenge and shown how these two conflicting intuitions can be reconciled and how social practices of a certain kind can institute objective norms and confer objective conceptual contents on expressions and performances. As Brandom puts it in his preface (1994: xviii):
A fundamental criterion of adequacy of the account [to be propounded here] is that the theorist not attach semantic contents to expressions by stipulation; it must always be shown how such contents can be conferred on expressions by the scorekeeping activities the theorist attributes to the linguistic practitioners themselves. That is, the aim is to present conditions on an interpretation of a community as discursive scorekeepers that are sufficient (though perhaps not necessary) to ensure that interpreting the community as engaged in those implicitly normative practices is interpreting them as taking or treating their speech acts as expressing the sorts of semantic contents in question.Foot note 1_1
As far as I understand what Brandom is getting at here, his aim is to specify a set of conditions such that, if a community would engage in practices meeting these conditions, then its members would, thereby, be taking some of their performances as having content, and these performances would thereby have been endowed with content. This raises two questions: (i) what must a practice look like, in order for its practitioners to count as taking or treating something as having content, and (ii) how is the fact that something is taken (in practice) to have content supposed to make it the case that it has content (in some more robust sense)? In what follows, I'm going to ignore the second question (deferring it to another paper), and focus on some of the issues pertaining to the first.
The important thing to keep in mind for this discussion is that according to Brandom, a discursive practice is essentially one in which some performances have the power to alter the normative statuses of the individuals involved in it, i.e., to alter what they are committed or entitled to do. In this «deontic» perspective, to make an assertion is to acquire (and to believe somehing is to have) a (discursive) normative status of a certain kind, which is called a «doxastic commitment». Both the fact that a given performance has the force of an assertion, and the content of this assertion, depend on the inferential role of this performance, which is itself determined by the set of further performances (of the same type) to which the participants are committed or entitled, in virtue of the fact that this performance has been issued. I cannot rehearse the details of Brandom's specific brand of inferential semantics here; it should suffice to recall that he conceives of conceptual content in terms of inferential articulation, and inferential articulation in terms of inheritance (and exclusion) relations among deontic statuses.
From this point of view, the notion of conceptual content is to be explained in terms of deontic statuses (i.e., in terms of what the members of the community are committed or entitled to do), and the notion of deontic status is, in turn, to be explained in terms of practical deontic attitudes, that is to say, in terms of attitudes of acknowledging or attributing, in practice, such and such deontic statuses.
As Brandom strongly emphasizes, for a normative account of content to avoid regress or circularity, a performance's being correct or incorrect cannot depend on its being judged or conceived to be correct or incorrect (or on anyone's having the capacity to judge or conceive it to be correct or incorrect); but since it must nonetheless depend somehow on the activities of those who produce and consume it, there arises the need to construe what it is to take a performance to be correct or incorrect (or to attribute a deontic status) in a way that doesn't equate it with any conceptually contentful state or attitude (or more generally, in such a way that the capacity to take a performance to be correct or incorrect doesn't presuppose any capacity for conceptual thought). In this respect at least, Brandom's practical deontic attitudes (taking a performance as correct, acknowledging or attributing a deontic status) are strikingly analogous to Davidson's non-individuative attitudesFoot note 1_2 of holding-true and preferring-true (i.e., those in terms of which the data of radical interpretation are to be describable), and raise the same kind of worries. These are worries concerning the legitimacy of appealing to non-conceptual intentionality in an account of conceptual intentionality.
In order to be in a position to deal with these issues, I must first be more explicit about the impact that the fact that some type of phenomenon «depends» on another is supposed to have on the way in which it can be explained or accounted for. I submit that, in the present context, the intuition behind such talk of «dependence» can be captured by taking the claim that some type of phenomenon X depends on another type of phenomenon Y as equivalent to the claim that any explanation of X involves either Y or an explanation of Y (in the sense that it contains ingredients sufficient to provide an explanation of Y). In other words, I submit that X depends on Y if and only if it is impossible to account for X without either invoking Y or being in a position to provide an account of Y as well.
Let's now ask which dependence relations could plausibly hold between conceptual and non-conceptual content. It may be useful, here, to observe that the idea that thought depends on language is naturally understood as implying, at least, that any system capable of having thoughts has mastered some system of symbolic communication. This strongly suggests that, in the same way, (i) the claim that conceptual content depends on non-conceptual content implies that no system can have the capacity to be in any conceptually contentful state unless it also has the capacity to be in some non-conceptually contentful state, and (ii) the claim that non-conceptual content depends on conceptual content implies that no system can have the capacity to be in any non-conceptually contentful state unless it also has the capacity to be in some conceptually contentful state. But then, it goes almost without saying that whoever is willing to grant that there are non-conceptual contents will be likely to hold that conceptual content depends on non-conceptual content, and that whoever admits further that non-linguistic animals and infants can be in non-conceptually contentful states, will assume that non-conceptual content does not depend on conceptual content. This implies that it must be possible to account for non-conceptual content without relying on conceptual content or being in a position to account for conceptual content, and precludes the coherentist strategy of simultaneously accounting for both conceptual and non-conceptual content.
But one must be very careful here. For the fact that some system may have the capacity to be in some non-conceptually contentful states while lacking the capacity to be in any conceptually contentful state obviously does not entail that all non-conceptually contentful states are accessible to some system lacking the capacity to be in conceptually contentful states. That non-conceptual content doesn't depend on conceptual content doesn't preclude the possibility that some non-conceptual thoughts can be had only by systems which are capable of conceptual thought, or even the possibility that some types of non-conceptual contents depend on conceptual content (i.e., on the capacity to be in some conceptually contentful states).
In other words, the claim that non-conceptual content does not depend on conceptual content should not be confused with the claim that no type of non-conceptual content depends on conceptual content. The first claim implies that it be possible to account for the capacity to be in a non-conceptually contentful state without relying on (or be in a position to provide) an account of the capacity to be in a conceptually contentful state, but not that it should be possible, for any given non-conceptual content, to account for the capacity to be in a state with this content, without relying on (or be in a position to provide) an account of the capacity to be in a conceptually contentful state.
What I want to suggest, at this point, is that there is nothing to prevent one from holding that Brandom's basic practical attitudes, if not conceptually contentful, may yet turn out to be non-conceptually contentful, and thus intentional, in some weaker sense. In particular, no incoherence or circularity would ensue from this assumption, if it is granted that admitting the notion of non-conceptual content almost forces one to assume that non-conceptual content doesn't depend on conceptual content. Moreover, even if it must be acknowledged that the capacity to practically attribute discursive normative statuses entails the capacity to be in conceptually contentful states (or in other words, that a system cannot have the capacity to practically attribute discursive statuses unless it also has the capacity to undertake such statuses), this doesn't conflict with the view that non-conceptual content doesn't depend on conceptual content. For this implies, at most, that some practical (and non-conceptually contentful) deontic attitudes (namely, those which consist in practically attributing or undertaking discursive statuses) depend on the capacity to be in conceptually contentful states, which is exactly what one would expect, if these practical deontic attitudes were «constitutive» of conceptual content, as Brandom's analysis requires. It may, however, be feared that Brandom is precluded from holding both that practical deontic attitudes are non-conceptually contentful and that they can be used to account for conceptual content, by his viewFoot note 1_3 that non-conceptual content is «derived» from conceptual content, which alone can be «original». This the worry I want to address in the next section.
§2.-- Original, Yet Dependent Intentionality?
To deal with this worry, we should first ask what the distinction between original and derivative intentionality consists in, and especially whether the fact that a kind of intentionality or contentfulness is derivative, while some other kind is original, is supposed to entail that the former depends on the latter, in the sense that any explanation of the first would involve either the second or an explanation of the secondFoot note 1_4. One obvious suggestion is that something counts as derivatively intentional when its being intentional depends on something else's being intentional, and as originally intentional when its being intentional doesn't depend on anything else's being intentional. Now, this can generalize in at least two different ways:
(1) an intentional system counts as derivatively intentional when its being intentional (its capacity to be in contentful states) depends on some other system's being intentional, and as originally intentional when its being intentional doesn't depend on any other system's being intentional
(2) a (conceptually or non conceptually contentful) state/performance counts as derivatively contentful when its being contentful depends on another state/performance's being (either conceptually or non-conceptually) contentful, and as originally contentful when its being contentful doesn't depend on any other state/performance's being (conceptually or non-conceptually) contentful.
And none of these seems to imply that if non-conceptual content is derivative and conceptual content, original, then non-conceptual content depends on conceptual content; for none implies that if non-conceptual content is derivative and conceptual content, original, then no intentional system can have the capacity to be in non-conceptually contentful states unless it also has the capacity to be in some conceptually contentful stateFoot note 1_5. However, this may not be a very significant observation, given that neither (1) nor (2) seems to capture exactly what is involved in saying that non-conceptual intentionality is merely derivative (and conceptual intentionality, original).
According to (1), to claim that non-conceptual intentionality is derivative is to say that all intentional systems which have the capacity to be in non-conceptually contentful states are such that their having this capacity depends on some other system's having the capacity to be in (conceptually or non-conceptually) contentful states, and to claim that conceptual intentionality is original is to say that no intentional system which has the capacity to be in conceptually contentful states is such that its having this capacity depends on any other system's having the capacity to be in (conceptually or non-conceptually) contentful states. But this cannot possibly be what Brandom means by «derivative» and «original»; for on this understanding, it would follow from the social character of intentionality, that both conceptual intentionality and non-conceptual intentionality are derivative, and that neither is original.
According to (2), to claim that non-conceptual intentionality is derivative is to say that all non-conceptually contentful states/performances are such that their contentfulness depends on the (conceptual or non-conceptual) contentfulness of some other state/performance, and to claim that conceptual intentionality is original is to say that no conceptually contentful state/performance is such that its contentfulness depends on the (conceptual or non-conceptual) contentfulness of any other state/performance. But on this understanding, conceptual intentionality will probably turn out not to be original, since (unless content turns out to be thoroughly atomistic) it is likely that many conceptually contentful states/performances are such that their contentfulness depends on the (conceptual or non-conceptual) contentfulness of some other state/performance. Moreover, this same reading would lead one to hold that non-conceptual intentionality is derivative even if it should turn out that the contentfulness of all non-conceptually contentful states/performances depends on the non-conceptual contentfulness of some other state/performance (and never on the conceptual contentfulness of any state/performance).
It is becoming apparent that these suggestions fail to take account of the fact that the idea behind saying that non-conceptual intentionality is derivative, is that it is somehow derived from conceptual intentionality. A better way to capture what is intended would thus be to say either that
(3) non-conceptual intentionality is derivative (with respect to conceptual intentionality) if and only if all intentional systems which have the capacity to be in non-conceptually contentful states are such that their having this capacity depends on some other system's having the capacity to be in conceptually contentful states
and
(4) conceptual intentionality is original (with respect to non-conceptual intentionality) if and only if no intentional system which has the capacity to be in conceptually contentful states is such that its having this capacity depends on any other system's having the capacity to be in non-conceptually contentful states,
or that
(5) non-conceptual intentionality is derivative (with respect to conceptual intentionality) if and only if all non-conceptually contentful states/performances are such that their contentfulness depends on the conceptual contentfulness of some other state/performance
and
(6) conceptual intentionality is original (with respect to non-conceptual intentionality) if and only if no conceptually contentful state/performance is such that its contentfulness depends on the non-conceptual contentfulness of any other state/performance.
Since it has been granted above that conceptual content depends on non-conceptual content, in the sense that no system can have the capacity to be in conceptually contentful states unless it also has the capacity to be in non-conceptually contentful states, and given the social character of intentionality, it follows that conceptual intentionality isn't original either in the sense provided by (4), or in the sense provided by (6). On the contrary (on the assumptions just mentioned), conceptual intentionality turns out to be derivative with respect to non-conceptual intentionality (when «to be derivative with respect to non-conceptual intentionality» is understood in the way suggested by (3) or (5)).
As far as I can see, however, the fact that non-conceptual content doesn't depend on conceptual content wouldn't prevent it from being derivative (with respect to conceptual content) either in the sense provided by (3) or the sense provided by (5). Actually, (5) boils down to (3), on the assumption that non-conceptual content doesn't depend on conceptual content.
On these readings, then, it would turn out that (i) conceptual intentionality is derivative with respect to non-conceptual intentionality, in virtue of the fact that it depends on non-conceptual content (and that content is essentially social), and that (ii) non-conceptual intentionality could be derivative with respect to conceptual intentionality, even though it doesn't depend on conceptual content. This may help to make sense of the suggestion that it would not necessarily be inconsistent to hold that non-conceptual intentionality is derivative with respect to conceptual intentionality while denying that non-conceptual content depends on conceptual content. But there still remains to make sense of the claim that conceptual intentionality is original.
This is easier than the preceding remarks might suggest. The only thing to do is to restrict exclusion of the dependence relation to intentional systems which are at once capable of being in non-conceptually contentful states but incapable of being in conceptually contentful states, thus:
(7) conceptual intentionality is original (with respect to non-conceptual intentionality) if and only if no intentional system which has the capacity to be in conceptually contentful states is such that its having this capacity depends on the fact that any other system lacking the capacity to be in conceptually contentful states has the capacity to be in non-conceptually contentful states.
As far as I can see, this yields an intuitively plausible sense in which conceptual intentionality probably is original, and in which its being original wouldn't conflict with the fact that it is yet dependent on non-conceptual intentionality.
To round up this discussion, let's now ask on what grounds one could be tempted to claim that non-conceptual intentionality is derivative with respect to conceptual intentionality. It seems to me that (in Brandom's and Davidson's cases at least) the motivation for this view comes mainly from the interpretationist (or phenomenalist) principle according to which any state or performance can be contentful only insofar as (and in virtue of the fact that) it is or can be treated as such (i.e., insofar as it has been conferred some content by a set of practices), together with the feeling that (i) intentional systems lacking the capacity to be in conceptually contentful states cannot have the capacity to engage in any «content-conferring» practice (i.e., to practically attribute any relevant normative status), and that (ii) such systems (e.g., non-linguistic animals) therefore count as intentional (if at all) only in virtue of the fact that other systems, which are capable to be in conceptually contentful states, treat them as capable to be in (non-conceptually) contentful states and attribute such states to them. In other words, this attitude seems to rest on the conviction that only systems which have the capacity to be in conceptually contentful states can have the practical attitude of treating anything as contentful. Thus the intuition behind this talk of «original» and «derivative» intentionality (in the context of an interpretationist perspective) may well be that an intentional system (as opposed to a kind of intentionality, such as conceptual or non-conceptual intentionality) counts as derivatively intentional when it lacks the capacity to treat anything as contentful or intentional (i.e., to attribute content or intentionality to anything), and as originally intentional when it has this capacity. At least this seems to fit nicely with the way in which I have proposed to understand these notions, in (3) and (7) above. According to what has just been suggested, if non-conceptual intentionality were to satisfy (3), and to count as derivative with respect to conceptual intentionality, it would have to be in virtue of the fact that all systems having the capacity to be in non-conceptually contentful states (whether or not they are also capable to be in conceptually contentful states) are such that they have this capacity only in virtue of the fact that some systems which have the capacity to be in conceptually contentful states treat them as having this capacityFoot note 1_6. And as for conceptual intentionality, it does satisfy (7), and this is in virtue of the fact that no system having the capacity to be in conceptually contentful states is such that it has this capacity only in virtue of the fact that some system lacking the capacity to be in conceptually contentful states (but having the capacity to be in non-conceptually contentful states) treat it as having this capacity.
Now, the idea that no system is capable of treating anything, in practice, as contentful unless it has the capacity to be itself in some conceptually contentful states (on which the view that non-conceptual intentionality is merely derivative rests, in part) strikes me both as questionable (or at least in need of further argument) and as little more than an empirical conjecture. It should at least be pointed out that Brandom's program itself requires that it be possible to practically attribute some (non-discursive) normative statuses without having the capacity to be in any conceptually contentful state. And if this is right, then it is hard to see why only systems capable of conceptual thought could have the capacity to practically attribute (non-conceptually) contentful states.
Thus, the claim that non-conceptual intentionality is merely derivative (with respect to conceptual intentionality) seems insufficiently supported, even from within Brandom's perspective. But what is more important, in the present context, is (i) that rejecting it would not involve renouncing all forms of interpretationism, and (ii) that it is in any case perfectly compatible with the claim that non-conceptual content does not depend on conceptual content. For the latter is the claim that a system could have the capacity to be in non-conceptually contentful states without having the capacity to be in any conceptually contentful state, while the claim that non-conceptual content is derivative amounts to the claim that no system can have the capacity to be in non-conceptually contentful states unless some other system has the capacity to be in conceptually contentful states. It follows that one could explain conceptual content in terms of non-conceptual content even if the latter were derivative with respect to the former. It would not be exaggerated to suggest that if interpretationism doesn't require that non-conceptual intentionality be derivative, then its opponents are thereby deprived of one of their strongest argument.
One potential problem with the suggestion that Brandom's practical deontic attitudes may have non-conceptual content may however come from the fact that, in case non-conceptual content could be accounted for in non-normative terms, one would run the risk of being forced to conclude that it is, after all, in non-normative, rather than in normative terms that content has been accounted for. In other words, Brandom's normativism would then threaten to change into its opposite (but wouldn't that be very hegelian?).
There is a sense in which any explanation presupposes conceptual intentionality, just in virtue of the fact that to explain anything involves a piece of discursive thinking. But this obviously isn't enough to make it objectionably circular to appeal to non-conceptual intentionality in accounting for conceptual intentionality (and since only systems who enjoy conceptual intentionality could set themselves the task of accounting for it, such circularity is in any case unavoidable). The main thing is that to attribute non-conceptual intentionality is not yet to attribute conceptual intentionality, and that it is conceivable that some systems be able to have non-conceptually contentful attitudes without being able to have conceptually contentful attitudes, or perhaps even that some systems might be able to have practical deontic attitudes (conceived as a special kind of non-conceptually contentful attitudes) without their capacity to have such attitudes being sufficient to make them capable of being in conceptually contentful states (e.g. because their use of this capacity doesn't exhibit the right kind of structure).
§3.-- Original Intentionality in Practice and in Theory
My aim in this section is to see how and to what extent Brandom's brand of interpretationism departs from Davidson's view that one can be a thinker only if one has the concept of a concept.
On one natural reading of Davidson's views, he holds that a necessary condition of having the capacity to be in any intentional state is having the capacity to attribute such states to others, where to attribute an intentional state involves (not just having some practical attitude, as suggested by Brandom, but) mastering the concept of a contentful, intentional state. In other words, to attribute an intentional state, on this understanding, is to be in some (higher-order) conceptually contentful intentional state whose content can be specified only by using indirect discourse (and thus amounts to thinking or judging that someone thinks that so-and-so). This might be encapsulated in the following «principle of intentional ascent»:
(8) If S has the capacity to think that p, then S has the capacity to think that S' thinks that p.
Insofar as Davidson's interpretationist perspective implies that a system is derivatively intentional only if it has the capacity to be in contentful states (i.e. to have «thoughts», in some generic sense) without itself being able to attribute such states to others, this further claim obviously doesn't leave any room for derivative intentionality, which could at best be construed as unreal and metaphorical. In this perspective, the capacity to have thoughts (or to be in contentful states) at all thus goes hand in hand with the capacity to have higher-order conceptually contentful thoughts.
But in view of the fact that it is certainly possible to think that p without thinking that anyone thinks that p, it seems that it should also be possible to attribute the thought that p without attributing the thought that anyone thinks that p. And if this is possible, then there doesn't seem to be any reason to deny that it is possible to attribute the thought that p without attributing the capacity to have the thought that anyone thinks that p. In other words, it is hard to believe that there could be systems with the capacity to attribute the capacity to attribute thoughts (in the present context, the capacity to have higher-order conceptual thoughts) without the capacity to attribute thoughts simpliciter (i.e., first-order thoughts) in the first place. But if these really are two distinct capacities, there is no reason why such systems should not also be able to attribute thoughts to systems to which they don't attribute the capacity to attribute thoughts themselves. This obviously doesn't show (8) to be false, but is enough to show that it's not necessary (and cannot be sustained as any kind of conceptual truth, as Davidson would have it), and strongly suggests that if some systems have original intentionality (which, in this context, means the capacity to attribute intentional states), then it must be possible that some systems have only derivative intentionality (which, in this context, means the capacity to be only in simple, first-order, intentional states). Hence, there seems to be a sense in which original intentionality depends on the possibility of derivative intentionality; that is to say, there would be no originally intentional systems if they didn't have the capacity to attribute derivative intentionality (i.e., if they couldn't treat some systems as having only «simple», derivative, intentionality).
The significance of this conclusion is not only that it may go some way towards doing justice to the common sense idea that non-linguistic animals might enjoy genuine intentionality even if it should turn out that they lack any capacity to attribute any intentional states, but that it may provide some ground for holding that the notion of the objectivity of conceptual norms must make sense, or at least help to make that notion intelligible. Since a derivatively intentional system is a system which can be in contentful states, but lacks the capacity to think that it (or anything) is in such states, the contentful states of such a system may turn out to be incorrect or inappropriate without its having the capacity to recognise that this is so. To admit derivative intentionality is thus to admit that some intentional states may be incorrect without the bearer of these states recognising that they are. But if it makes sense to admit this possibility, it seems it must also make sense to admit that one's originally intentional states may be incorrect without one's realising that they are, and that if this can happen, then it can also happen that no one takes (or will ever take) them to be incorrect.
I conclude that the principle of intentional ascent, as stated by (8), is untenable. But as will shortly become apparent, this is not to say that other, restricted forms of this principle may not be acceptable.
Recall that one main difference between Davidson's perspective (as just reconstructively modified) and Brandom's, is that the latter allows that to attribute an intentional state (most fundamentally, a doxastic commitment) may consist, not in being in any conceptually contentful state, but in having some practical deontic attitude. In other words, while Davidson takes all attributions of intentionality as being themselves conceptually contentful attitudes, Brandom introduces a distinction between attributions of intentionality which are somehow «implicit in practice» and consist in taking (certain kinds of) practical deontic attitudes, and attributions which are explicit in thought or discourse (and involve being in higher-order intentional states, in the form of having higher-order discursive commitments). This (as will soon appear more clearly) seems to be what permits him to draw a line separating two kinds, or grades, of original conceptual intentionality; that is to say, to allow that some systems may have only the practical ability to attribute (conceptually) contentful states, while others have both this practical ability and the capacity to make such attributions explicit (and thus not only have concepts but have specifically mastered the concept of a conceptually contentful state).
As already remarked, Brandom's project of giving an account of conceptual content (and discursive deontic satuses) in terms of practical deontic attitudes precludes one to take the latter to be themselves conceptually contentful attitudes, and thus to assume that one who practically attributes some doxastic commitment is thereby doxastically committed to anything. I suggested above that taking these practical attitudes to be non-conceptually contentful (instead of non-contentful at all) would neither conflict with the view that non-conceptual intentionality is derivative, nor be objectionably circular, if it were allowed that a system might have such attitudes without yet being able to be in or to attribute conceptually contentful states at allFoot note 1_7. But this suggestion still raises many other questions, some of which concern the status of various forms of intentional ascent which are now made available.
As can now be clearly seen, the principle of intentional ascent, as originally given by (8), can be read in at least eight different ways, depending on (i) whether various occurrences are taken to refer to conceptual or non-conceptual thoughts, and (ii) whether attributions of intentional states are taken to involve being in some conceptually contentful state, or being in some non-conceptually contentful state, i.e., depending on whether we're talking about «theoretical» (and explicit) or «practical» (and implicit) attributions. This accordingly yields four principles of theoretical intentional ascent :
(9) If S can conceptually think that p, then S can conceptually think that S' conceptually thinks that p (i.e., S can theoretically attribute the conceptual thought that p),
(10) If S can conceptually think that p, then S can conceptually think that S' non-conceptually thinks that p (i.e., S can theoretically attribute the non-conceptual thought that p),
(11) If S can non-conceptually think that p, then S can conceptually think that S' conceptually thinks that p (i.e., S can theoretically attribute the conceptual thought that p),
(12) If S can non-conceptually think that p, then S can conceptually think that S' non-conceptually thinks that p (i.e., S can theoretically attribute the non-conceptual thought that p),
and four corresponding principles of practical intentional ascent :
(13) If S can conceptually think that p, then S can non-conceptually think that S' conceptually thinks that p (i.e., S can practically attribute the conceptual thought that p),
(14) If S can conceptually think that p, then S can non-conceptually think that S' non-conceptually thinks that p (i.e., S can practically attribute the non-conceptual thought that p),
(15) If S can non-conceptually think that p, then S can non-conceptually think that S' conceptually thinks that p (i.e., S can practically attribute the conceptual thought that p),
(16) If S can non-conceptually think that p, then S can non-conceptually think that S' non-conceptually thinks that p (i.e., S can practically attribute the non-conceptual thought that p).
Just as the original principle (8) amounts to the claim that all intentional systems are originally intentional, these various principles claim either that conceptually intentional systems or non-conceptually intentional systems are originally intentional. But since an intentional system may count as originally intentional either in virtue of the fact that it can make practical attributions of intentionality, or in virtue of the fact that it can make theoretical attributions of intentionality, and since a distinction must be made between attributions of conceptual intentionality and attributions of non-conceptual intentionality, we now have four different ways in which an intentional system may turn out to be originally intentional.
Now, taking account of the fact that (i) to conceptually (non-conceptually) think that S' conceptually (non-conceptually) thinks that p is an instance of conceptually (non-conceptually) thinking that p, and that (ii) to conceptually (non-conceptually) think that S' (conceptually or non-conceptually) thinks that p is to theoretically (practically) attribute the (conceptual or non-conceptual) thought that p, it will be noted that each of (9)-(16) yields a corresponding higher-order principle as a special case :
(9*) If S can theoretically attribute the (conceptual or non-conceptual) thought that p, then S can theoretically attribute the attitude of theoretically attributing the (conceptual or non-conceptual) thought that p,
(10*) If S can theoretically attribute the (conceptual or non-conceptual) thought that p, then S can theoretically attribute the attitude of practically attributing the (conceptual or non-conceptual) thought that p,
(11*) If S can practically attribute the (conceptual or non-conceptual) thought that p, then S can theoretically attribute the attitude of theoretically attributing the (conceptual or non-conceptual) thought that p,
(12*) If S can practically attribute the (conceptual or non-conceptual) thought that p, then S can theoretically attribute the attitude of practically attributing the (conceptual or non-conceptual) thought that p,
(13*) If S can theoretically attribute the (conceptual or non-conceptual) thought that p, then S can practically attribute the attitude of theoretically attributing the (conceptual or non-conceptual) thought that p,
(14*) If S can theoretically attribute the (conceptual or non-conceptual) thought that p, then S can practically attribute the attitude of practically attributing the (conceptual or non-conceptual) thought that p,
(15*) If S can practically attribute the (conceptual or non-conceptual) thought that p, then S can practically attribute the attitude of theoretically attributing the (conceptual or non-conceptual) thought that p,
(16*) If S can practically attribute the (conceptual or non-conceptual) thought that p, then S can practically attribute the attitude of practically attributing the (conceptual or non-conceptual) thought that p.
It goes without saying that any of these higher-order principles could be true even if the corresponding first-order principle had to be rejected.
These various «combinatorial» possibilities are of course not on a par, and the purpose of making them explicit is to help to clarify Brandom's view and how it is related to Davidson's. Hence, the task before us is to determine, for each of (9)-(16), how it (and its higher-order corollary) fares with respect to Davidson's and Brandom's perspectives (on the working assumption that practical deontic attitudes are non-conceptually contentful). This fortunately turns out to be less painstaking than it might seem.
In the present larger setting, it can easily be seen that (9) is the only reading of the original principle of intentional ascent (the one given by (8)) which accords with Davidson's claim that all intentionality is conceptual (and original) intentionality, and which provides a plausible way to understand his well-known claim that no one can be interpretable unless one is an interpreter (i.e., unless one has the concept of a thought, and is a theoretical attributor of intentionality). By contrast, Brandom clearly rejects not only this principleFoot note 1_8, but all of (10)-(12) as well (that is to say, all forms of first-order theoretical intentional ascent), since he holds that it is possible to have conceptual (or for that matter, non-conceptual) thoughts without having the capacity to make any explicit, «theoretical» attribution of intentionalityFoot note 1_9.
It should however be emphasized that even though he rejects (9), Brandom could (and probably would) accept its higher-order corollary (9*), thus endorsing the claim that no one can have the capacity to theoretically attribute thoughts unless one has the capacity to theoretically attribute this very capacity (which is one version of the Davidsonian claim that no one can be an interpreter unless one has the concept of an interpreter). As far as I can see, (10*) likewise seems compatible with everything Brandom says, though it remains unclear to what extent he would be prepared to endorse it (and how plausible it really is). And as for (11*)-(12*), they clearly are excluded by Brandom's claim that one may have the capacity to practically attribute intentional states without yet having the ressources to make explicit, theoretical attributions (i.e., by Brandom's distinction between «merely rational» and «fully logical» conceptually intentional systems).
Let's now turn to the «practical» forms of intentional ascent, (13)-(16) and (13*)-(16*).
Clearly, since Brandom wishes to maintain that all conceptually intentional systems are originally intentional, while allowing that some of them may yet lack the capacity for theoretical attributions of intentional states, this leaves him no choice but to endorse (13) (and (13*) with it). This actually provides his version of Davidson's claim that no one can be interpretable without being an interpreter, one in which being interpretable is restricted to being «conceptually» interpretable, and being an interpreter is restricted to being a «practical» interpreter.
Furthermore, the argument I gave above, to the conclusion that no one could have the capacity to attribute intentional states at all unless one has the capacity to attribute derivatively intentional states, strongly suggests that (on the assumption that derivatively intentional systems, if any, must be capable of non-conceptual intentionality) (14) should also be acceptable, for what it says, in effect, is that no one could have conceptual thoughts without having the capacity to practically attribute non-conceptual thoughts, and it has already been granted that no one could have conceptual thoughts without having the capacity to practically attribute at least conceptual thoughts. But as with (10*), it is hard to tell to what extent Brandom would be prepared to endorse (14), even though it seems compatible with his views. It all depends on whether the capacity to practically attribute conceptual thoughts (which all conceptually intentional systems must have, according to Brandom) should be taken to imply the capacity to practically attribute non-conceptual thoughts.
Insofar as Brandom is committed to accept that intentional systems which have the capacity to be in non-conceptually contentful states need not have the capacity to attribute (even practically) any intentional state, he clearly must reject both (15) and (16). Moreover, it would seem that if (as Brandom claims) one may be able to practically attribute thoughts without being able to theoretically attribute them, then (a fortiori) one may be able to practically attribute thoughts without being able to practically attribute theoretical attributions of thoughts; and hence, that (15*) probably must be rejected as well. The same doesn't hold for (16*), however, which is compatible with Brandom's views.
The upshot of this discussion is that Brandom is clearly committed only to (13)-(13*), though nothing (so far) seems to preclude his acceptance of (9*), (10*), (14), (14*) and/or (16*). Note that only two of these (namely, (9*) and (10*)) are forms of theoretical intentional ascent, and then only higher-order ones. Since in the context of Brandom's attempt to explain conceptual intentionality, it clearly is practical intentional ascent that is of prime importance, they can safely be ignored here. But perhaps it is worth giving a closer look at (14)-(14*) and (16*), despite the fact that they don't seem to be strictly required by Brandom's explanation of conceptual intentionality (and that Brandom doesn't pay much attention to attributions of non-conceptual intentionality).
On the face of it, (16*) looks interestingly less compelling than its (Davidsonian) direct opposite (9*). Indeed, since practical attributions of thoughts (even conceptual ones) are not (and cannot be) conceptual thoughts, there is no obvious reason why the capacity to practically attribute (even conceptual) thoughts should entail the capacity to practically attribute practical attributions of thoughts.
Furthermore, given that in Brandom's terminology, the following special case of (16*):
(17) If S can practically attribute the conceptual thought that p, then S can practically attribute the attitude of practically attributing the conceptual thought that p,
would translate as:
(18) If S can have the practical deontic attitude of taking S' to be doxastically committed to p, then S can have the practical deontic attitude of taking S' to have the practical deontic attitude of taking S'' to be doxastically committed to p,
to endorse (16*) would imply that to be involved at all in discursive practice (i.e., to have conceptual thoughts) requires that the participants have the capacity, not only to keep deontic scores on each other, but also to keep scores on each other's scores, etc... But this raises a problem, insofar as to keep score on the score kept by S must be something different from keeping score on S's (higher-order) discursive deontic statuses (since by hypothesis, practically attributing a doxastic commitment is not having oneself any doxastic commitment). Perhaps this point could be dealt with by taking practical deontic attitudes to be commitments of a non-discursive kind (taking in practice a performance as correct might commit, and not only dispose, one to sanction it), a course which, as far as I can see, is not precluded by the non-circularity condition (and accords with some of Brandom's remarksFoot note 1_10). However, it would still be hard to see how such higher-order attitudes could nevertheless be purely practical attitudes, if this is meant to imply that they must be such that those who have them may yet lack the capacity to have corresponding doxastic commitments (as Brandom seems to require).
Another potential problem with (16*) stems from the fact that the following also is a special case of it:
(19) If S can practically attribute the non-conceptual thought that p, then S can practically attribute the attitude of practically attributing the non-conceptual thought that p,
and that to endorse it would thus commit one to the view that no system could have original non-conceptual intentionality without having the capacity to treat, in practice, other systems as themselves enjoying original non-conceptual intentionality (i.e., as practical attributors of non-conceptual intentionality); a capacity which one would perhaps not want to grant to all systems which are capable of practically attributing non-conceptual thoughts without being capable of having conceptual thoughts (if there are such systems). Of course, this would not be much of a problem for Brandom, since he denies that there could be any such intentional systems.
But even if (16*) has to be rejected (which is by no means clear), it might still be the case that all intentional systems of some interestingly restricted class are required to have the (higher-order) capacity to practically attribute practical attributions of thoughts. Indeed, it is easily seen that, on the plausible assumption that if no one can have the capacity to practically attribute the attitude of theoretically attributing the thought that p unless one also has the capacity to practically attribute the attitude of practically attributing the thought that p, (13*) (and hence (13)) entails something between (13*) and (16*), namely:
(14*) If S can theoretically attribute the (conceptual or non-conceptual) thought that p, then S can practically attribute the attitude of practically attributing the (conceptual or non-conceptual) thought that p,
(in Brandom's terminology: if S can be doxastically committed to someone's having the thought that p, then S can have the practical deontic attitude of taking someone to practically attribute the thought that p).
It would be highly interesting to be able to argue as well that only intentional systems of some interestingly restricted class can have this (higher-order) capacity to practically attribute practical attributions of thoughts, e.g., to establish the claim that:
(20) If S can practically attribute the attitude of practically attributing the thought that p, then S can theoretically attribute the thought that p,
which would restrict higher-order practical attributions to what Brandom calls «fully logical» intentional systems; or perhaps only the weaker claim that:
(21) If S can practically attribute the attitude of practically attributing the thought that p, then S can have the conceptual thought that p,
which would restrict them to conceptually (but not necessarily «fully logical») intentional systems. In either case, it would have to be granted (as expected) that some practical attitudes (namely, higher-order ones) are such that only conceptually intentional systems can have them. It has however to be observed that the weaker claim (21) would seem to be of special significance only if it could also be shown that:
(22) If S can have the conceptual thought that p, then S can practically attribute the attitude of practically attributing the thought that p,
something which doesn't follow from any of the forms of intentional ascent which have so far been considered, but to which Brandom may well be committed.
Enough has been said, I think, to demonstrate that the hypothesis that Brandom's practical deontic attitudes are to be taken as non-conceptually contentful attitudes, though prima facie coherent and appealing, raises a number of intriguing and potentially fruitful questions. Exploring them further would however require something which I don't yet have, namely, a substantial theory of non-conceptual content.Foot note 1_11