by  Steve Bayne
August 2002

Whether it is truth 'under a description', the nature of incontinence, the primitivity of what have been called basic actions or any one of a number of other important philosophical issues, Elizabeth Anscombe's ideas have been an important influence on Donald Davidson. But, perhaps most significant has been Anscombe's influence on Davidson's theory of "pure intentions." This theory is almost entirely a response to Anscombe's discussion of the nature of practical reasoning and in particular the nature of the premises of what has been called the "practical syllogism." In this paper, while I will examine the historical lines of force that determined the shape of Davidson's treatment of pure intentions my ambition is to do more than this. It is to evaluate the Davidson/Anscombe theory of pure intentions from the point of view of a long neglected theory deriving from F. H. Bradley based on the understanding that intentions are "ideas," and that volitions which are guided by such ideas transform the world as integral to the satisfaction of our intentions. I will begin by spelling out a crucial observation made by Anscombe and, then, proceed to examine how Davidson makes use of it. I will allege three objections sufficient to abandon Davidson's account. The first two will be dealt with briefly while the last, possessing as it does a bearing on what I shall call the "classical theory," will be considered in greater detail.

When we act intentionally is there an act of will in addition to the act itself that serves either as its cause or something that gives it direction? At the heart of the controversy is the status of what have been called "pure intentions," that is, intentions which are neither acted upon nor of any consequence. If it held that an act of will is required in addition to the action willed then we are open to the charge of introducing mysteries rather than solving them. The challenge, then, is to explain the possibility of "pure intentions" without invoking a special category of mental acts. Davidson believes that this can be done if we carefully attend to the nature of practical reasoning, and it is in this regard that Anscombe's position proves critical.

Anscombe contributed significantly on more than one occasion to our understanding of Aristotle's syllogism of which, according to her (_Intention_ p. 60),there are three forms: the idle practical syllogism, the practical syllogism proper, and the theoretical syllogism. Let's take a brief look at Anscombe's examples of the first two, beginning with the idle practical syllogism.

Dry food suits any human being
Such-and-such food is dry
I am a human
This is a bit of such-and-such food
Therefore: This food suits me.
This syllogistic form is called "idle" because the conclusion, while suggestive of action, is not an action nor does it specify an action per se. Next consider an example of the practical syllogism proper.
Do everything conducive to not having a car crash
Such-and-such will be conducive to not having a car crash
Therefore: Do such-and-such.
In fact, Aristotle never suggests a syllogism of precisely this form and Anscombe acknowledges Hare as the one who engineered it, but it illustrates issues at the core of practical reasoning since the conclusion may be an action or an imperative to perform a certain action. The conclusion of this example (of a practical syllogism proper) is an imperative, but the active content could be made more explicit by leaving blank what follows "therefore" and using an arbitrary symbol to represent the action. Further, the conclusion of the practical syllogism proper, since it refers to an action, has a subject matter that may be referred to demonstratively. This is a consideration figuring highly in Davidson's own account. Although she acknowledges that unlike the idle syllogism the practical syllogism proper is suggestive of action, Anscombe calls the premise of such syllogisms "insane."
But this syllogism suffers from the disadvantage that the first, universal, premise is an insane one, which no one could accept for a moment if he thought out what it meant. For there are usually a hundred different and incompatible things conducive to not having a car crash... (op. cit. p. 59).
What Davidson will do is introduce two categories of premises; those which are "insane" (which he will call "prima facie judgments") and those which are not (which he will call "all-out judgments"). He will apply this distinction to an understanding of the nature of pure intentions, intentions which are neither acted upon nor having any effects. Having identified pure intention as intention without an ensuing action, it is only natural that we ask what such an intention might be.

The difficulty is that an intentional action seems to be a pure intention plus its actualization, but what then is a pure intention which is not actualized? Is pure intention just an extra something that in fact we may do without in the treatment of ordinary intentional behavior? Davidson remarks:

But it would be astonishing if that extra element were foreign to our understanding of intentional action. For consider some simple action, like writing the word 'action'. Some temporal segments of this action are themselves actions; for example, first I write the letter 'a'. This I do with the intention of initiating an action that will not be complete until I have written the rest of the word. It is hard to see how the attitude towards the complete act which I have as I write the letter 'a' differs from the pure intention I may have had a moment before. (Actions and Events. Harvard. 1980. p.88)
There is an important distinction that defines the difference between Anscombe and Davidson on the one hand and what I have called "classical action theorists" on the other. It is a distinction that there is reason to believe both Anscombe and Davidson reject although there is no explicit rejection to be found in the literature.

The distinction I have in mind is that between "resolve" or "mere intention" and the intention of a completed action or action in progress. The latter is not a "mere intention" because it is being or has been realized. This latter manifestation of intention is associated with volition, where volition is to be regarded as the actualization of a certain disposition to act (Locke). With this distinction in mind we can look at Davidson's remark, above, in a somewhat different light. The activity he describes in his example of writing the word 'action' is in fact not one associated with "mere intention" but rather with the intention which is the guiding idea (James, Bradley) of an action in progress, what Bradley described as an "incomplete volition." Let's take a brief look at what Bradley has to say about incomplete volition before returning to Davidson's example and what he does with it.

Unlike the philosophers criticized for introducing mysterious inner actions and inner tweaks and twinges, Bradley requires no act of will in addition to behavior itself. He does, like James, require attention to an idea, but he does not, like, say, Prichard - whose views appear to weigh heavily on the minds of philosophers such as Ryle - invoke a special class of acts of will that serve as "springs of action." For Bradley it is ideas that serve as springs of action. In fact, Bradley's reason for taking this approach was owing to his realizing very early on that introducing acts of will in addition to behavioral acts would lead to a vicious regress of acts of will. Bradley thus avoids the criticisms of both Anscombe and Ryle having to do with a regress of acts of will associated with conceiving of the will as some sort of "executive" authority. This difference between certain "classical action theorists" coupled with the distinction between intentions and mere intentions carries with it a lesson for contemporary action theory.

For Bradley there are no pure intentions which are neither mere intentions or mere resolve. To be clear on this consider that a mere intention or resolve has to do with an entirely future action and may be codified as a mere state of mind, whereas the intention associated with an actual volition pertains to a present action or episode, such as, to follow Davidson's example, being in the process of writing 'c' while in the process of writing 'action', following one's already having written 'a'. For Bradley there is no pure intention connected with an act, whereas for Davidson every action has some relation to what amounts to being a pure intention, which is evident from the above quotation. For Bradley it is the "prevalence" of an "idea" that is the spring of action not an act of will. Furthermore, whereas Davidson is interested in the relation of pure intention to actions themselves, Bradley is interested in action that falls short of completion. In other words, in Davidson we have a preoccupation with pure intentions associated with actions in no wise complete or even initiated, whereas in Bradley we have a preoccupation with "incomplete volition" where 'volition' is itself a voluntary action. But let's be clear on what, exactly, Bradley means by a "volition".

A volition, I have said, is the self-realization of an idea with which the self is identified, and a volition is a whole, in which we may go on to find the following aspects. There is (1) existence, (2) the idea of change, and (3) the actual change of the existence by the idea to (4) the idea's content. And (5) in this change the self feels itself realized. ("Def. of Will" C.E. p. 477).
For Bradley once an idea "prevails" we are prevailed upon to act in accordance with that idea - and here, once again, there is a close tie to James.
An actual volition may certainly be involved in the prevalence of an idea, but that volition at the same time will be incomplete. But there will not be in any case even an incomplete volition, unless to some extent the idea carries itself out beyond itself. ("The Definition of Will (I)" MIND. 1902. Collected Essay. p. 492)
It is clear that for Bradley there are no pure intentions associated with real action; at best a pure intention in Davidson's sense is a mere resolve.

Davidson's analysis of intending, and his views on pure intention in particular, depend on the distinction between two kinds of judgment: judgments which may serve as premises in a practical syllogism and judgments which may serve as conclusions of "practical syllogisms proper" (Anscombe _Intention_ p. 60). It is only if we are given an action that we might refer to demonstratively that we have anything which we can evaluate (A & E. p. 96-97). When we evaluate an action on the basis of its being of a single sort, e.g., that it satisfies our desire for sweet food, such a judgment is "prima facie." But prima facie judgments only concern actions in light of a single sort. However, a particular action, such as satisfying our pro-attitude towards eating sweets, may also be an action of self poisoning as some sweet things are poison. As long on we are concerned merely with judgments about an action as an action of a single sort, we may engage "insane" premises (Anscombe), premises that would direct us to eat sweets whenever they are there for the having, even if they are poison. But in relation to individual actions, basing such actions on prima facie judgments leads to disaster. But when we take as basic the notion of an "all out judgment," a judgment which is unconditional, having the form "This action is desirable," then as long as we have acted on this judgment what we in fact act upon is a pure intention, i.e., a member of a subclass of all out judgments. While I will have more to say about prima facie and all-out judgments shortly, a couple of questions immediately surface.

Two considerations enter. First, that all out judgments in principle may be actions when they are conclusions of what Anscombe calls 'practical syllogisms proper'(_Intention_ p. 60). Because no prima facie judgment is "directly associated with actions" (A & E p. 98) it would appear at first that no prima facie judgment can be the referent of a demonstrative, something characteristically associated with the conclusion of a "practical syllogism proper." But why, one may ask, cannot prima facie judgments ever be all out judgments? The answer is merely that "it is unreasonable to perform an action because it has a desirable characteristic" (ibid). What this ignores is that not all actions are reasonable; as Anscombe pointed out, there are intentional actions which not only are done for bad reasons, making them unreasonable, but there are intentional actions done for no "reason" at all. So it seems arguable at the very least that Davidson is wrong if he is suggesting that prima facie and all out judgments represent nonoverlapping classes. There is a second possible point of disagreement: Because Davidson's example of prima facie judgments ("My eating something sweet is desirable" (op. cit. 97)) involves something commonly regarded as desirable, what is missed is that whether a sort of action (eating sweets) is desirable itself may be made to depend on all-out judgments. One man's prima facie judgment may be another's all out judgment. In order to avoid this dilemma, Davidson must do something which is at least equivalent to assuming an initial desirability of sorts of actions; but if so, then how do we make such an assignment when we are talking about nailing two boards together or standing atop a hot tin roof? If it is just that prima facie judgments are what we begin with, then Davidson is doing psychology and nothing of philosophical significance need follow - although it might. It is central to Davidson's theory of pure intentions that an all-out judgment may not be an action, although there is nothing to exclude it.

Having mentioned two points of possible disagreement, I will now pass to the main area of disagreement - a Disagreement which illuminates the main ontological difference between Davidson, who is in fact arguing against philosophers such as Prichard, and the classical action theorists represented here by Bradley. Davidson neatly dispatches one objection to his view that pure intentions are all-out judgments, viz. that because intentions cannot single out actions, owing to the fact that they are future directed, it cannot concern particular, demonstratively identifiable, judgments but only judgments of a sort. He attempts this by introducing intentions conditioned by beliefs rather than conditional intentions; but we need not go into this, as our discussion henceforth concerns the ontology of action.

Notice this much about the ontological implications over and above the formalities of the practical syllogism. We must deal with what may be either pure intentions or completed actions, actions to which demonstrative reference is possible. What is lacking in Davidson's account, and what is crucial to the classical action theorists, lies within an area of difference between judgments of a sort (prima facie judgments) and "all out judgments," which concern completed actions. What is missing is the notion of an incomplete action; and this is quite important if for no other reason than it includes the potential for adaptivity in performance, a subject not to be pursued here. Yet, with this in mind let us return to Davidson's example, since it places both the classical position and his own within a single purview.

When I write 'a' what is involved is in the way of intention more than mere resolve. It is a matter of volition. What changes by my writing 'a' is not my resolve to write 'action'. What is changed is that having written 'a' the intention I now have is to write 'c'. I write 'a' then, intending to write 'a' but *with* the intention not only of satisfying my initial resolve but also with the intention of writing 'ction'; I write 'c' intentionally but *with* the intention of writing 'tion' and so on. In other words what changes is not my "attitude" towards my resolve; what changes is the volition that is the writ-ing of the word 'action'. So Davidson is wrong to puzzle over whether

...the complete act which I have as I write the letter 'a' differs from the pure intention I may have had a moment before.
Bradley's way of putting it would be to say that my intention ("idea") is becoming "particularized" as it becomes more real and less "ideal." The difficulty Davidson will have in distinguishing resolve and volition is owing to his notion of pure intentions as "intendings abstracted from normal outcomes" and his concomitant but erroneous identification of volition as a mental action which precedes behavior (A & E. p. 89). Davidson collapses the distinction between resolve and intention and along with it that between pure intention and volition.

A resolve, we have been maintaining, following Bradley, is concerned with the future, no action need be present. But for Davidson there is no temporal hiatus separating the intentionality of resolve and that of volition; it is merely a matter of variation in time between the willing and the acting. The mechanism is simple enough: "we may reduce the delay to a moment" (A & E p. 96).

By reducing the delay to a moment we narrow the distance between resolve and volition without affecting our intentions, but if, as Bradley maintains, there is a fundamental difference between resolve and volition, tinkering with the time module will not effect a transition from resolve to volition. But because Davidson has failed to recognizes a difference in our intention as that intention becomes satisfied, that is, as I write the word 'action' he fails to see need to acknowledge a difference between resolve and the intention associated with volition. Indeed, he rejects volition because he thinks of it as did Prichard, whose views were the real targets of Ryle and Anscombe before him, rather than thinking of Bradley who never viewed volition as a mental action independent of behavior in the first place.

So we see that contrary to what Davidson alleges our pure intention changes as the action progresses, where that action is more than a simple or "basic" action, such as clenching one's fist or rotating one's shoulder.

In conclusion: we have seen that Davidson's idea of a pure intention developed as an effort to make sense of practical reasoning, while accepting Anscombe's conjecture that some premises of the practical syllogism proper are "insane." Davidson creates a distinction between two sort of judgments, prima facie and all out, which he employs in fashioning a theory of pure intentions, one that identifies them with all out judgments, while avoiding commitment to volitions as independent mental acts. We have offered three arguments against Davidson's position, citing one in particular as vindicating certain classical theorists such as Bradley and affording us access to a theory which make distinctions which before were not possible without falling into the usual criticisms directed against volitions as independent acts of will.