INTRODUCTION
Metaphysical analysis goes beyond conceptual or semantic inquiry, and also beyond necessary biconditionals, which can fail to provide the metaphysical explanation of special interest to the philosopher.
Consider for example the metaphysics of persons. In the broad domain of persons, we find our threefold divide among: (a) words, such as the word ‘person’, (b) concepts, such as the concept of a person, and (c) extra-linguistic, extra-conceptual entities, the living persons. Concerning the latter we find metaphysical options such as substance dualism, animalism, and so on. According to an Aristotelian view, a person is never identical with, but only constituted by a body, which needs to be alive, and in possession of certain powers and abilities, in order to (thereby) constitute a person.
This Aristotelian view in the metaphysics of persons involves metaphysical dependence. One thing exists or is actual dependently on certain other things and their properties. The dependent thing then exists or is actual dependently on the other things and on how they are propertied or related.
Turning to epistemology, we can now discern three problems that are quite distinct, however closely they may be interrelated. First is a problem of semantic analysis: What is the semantic analysis of the linguistic expression ‘S knows that P’? Second is a problem of a certain sort of conceptual analysis: With concepts understood as psychological entities, what is involved in someone's possession and/or deployment of the concept(s) of knowledge? This problem thus concerns people's minds, their psychology. Third is a problem of metaphysical analysis. Here our focus is on an objective phenomenon that need be neither expression nor concept. Our focus is rather on a state that people host, or an act that they perform. This is the phenomenon whose ontology we now wish to understand. What is the nature of human knowledge and how is it grounded? In virtue of what is it actual when it is actual?Footnote 1
Performance whose success manifests the relevant competence of the performer avoids thereby a kind of luck. According to competence virtue epistemology, knowledge is a special case of that. Knowledge of a sort is belief whose correctness is attained through epistemic competence, belief that is thus “apt.” A serious problem affects the metaphysics and ideology of perception and action, and a similar problem affects the metaphysics and ideology of knowledge. This is the problem of deviant causation.
We shall consider a solution for the problem in its three varieties. But first we examine Davidson on action, Grice on perception, and the account of knowledge as apt belief, as belief that gets it right through competence rather than luck. We take up the opposition between such traditional accounts and “disjunctivist” alternatives. And we explore how our take on the point and substance of metaphysical analysis bears on the problem and on competing reactions to it.
What follows divides into three parts. In a first section, the main lines of the view are laid out, and it is shown how it applies with the same basic structure in all three domains. A second section then considers how our account goes beyond Davidson and Grice. A final, third section reflects on a methodology that will fit our approach.
1. THREE MIND-WORLD RELATIONS
1.1 Action
What is it to act intentionally? As a first approximation, you might think, to act intentionally is to succeed in a certain intentional aim, where the success is owed to the agent's intention.
But that has counterexamples, such as the following.
A waiter intends to startle his boss by knocking over a stack of dishes right now, which makes him so nervous that he involuntarily staggers into the stack and knocks it over, thus startling the boss. But this is not something he does intentionally.
So, we should require that the agent's intentions must bring about the success in the right way, with “the right kind of causation.” Or so Davidson advises repeatedly in his long struggle with the problem, and in his parting thoughts on the matter. Here is how he puts it (with minor variations): What is it for an agent to F intentionally on a particular occasion? There must be some G such that the agent's intending to G must cause “… in the right way, the agent's particular act of F-ing.”Footnote 2 The waiter's knocking over of the dishes is not caused in the right way by any such intention. But no account of “the right way” has won consensus.
1.2 Perception
1. What is it to perceive an entity? The account of perception defended in Paul Grice's “Causal Theory of Perception” (Reference Grice1961) is an early, influential answer. Grice begins with a view drawn from H. H. Price's Perception (Reference Price1932):
X perceives M iff X has a sense experience that is causally dependent on some state of affairs involving M.
This, he argues, is subject to counterexamples. Our visual sense experiences while in the sunlight, for example, are causally dependent on the sun even when we look away from it. Nor do we perceive our eyes even when our visual experiences are highly dependent on the state of our eyes.
2. The account is then revised to say that an object is perceived if and only if some condition involving it is a differential condition that affects some but not all of the perceiver's relevant sense experience at the time of perception. The sun is not seen when we look away from it; on the revised account, this is because no condition of it affects only some and not all of one's visual sense experience.
However, the revised account too has counterexamples. Torches can shine respectively on statues viewed concurrently, each torch thus affecting the perceiver's visual impressions differentially, though only the statues are seen, with the torches blocked from view.
3. Grice eventually arrives at approximately the following view:
X perceives M if, and only if, X hosts a sensory experience for which M is causally responsible in the right way.
This is what Grice's view comes to, given how he thinks the “right” way is grasped, to be considered below.
1.3 Knowledge
As a first approximation, propositional knowledge can be understood as belief that attains its aim (truth) and does so not just by luck but through competence. Such knowledge is then a special case of performance that is not just lucky but apt: i.e., performance whose success is owed to the performer's relevant competence. The aptness of a performance is thus supposed to block an important sort of luck, the sort that precludes Gettiered subjects from knowing what they believe both correctly and competently. A belief falls short of knowledge when its truth is owed to luck and not to the believer's competence.Footnote 3
That view of the apt (not just lucky) performance has ostensible counterexamples. Take an archer's competent shot that (a) would hit the target absent intervening wind, and (b) does hit the target because, although a first gust diverts it, a second gust puts it back on track. Here the agent's competence yields the early orientation and speed of the arrow, and this combined orientation and speed, together with the two compensating gusts, results in the bull's eye. So, why is this shot not apt after all? A performance is apt when it succeeds because of the agent's competence. But our archer's wind-aided shot does seem to succeed because of his competence! If the agent's competence had not resulted in the right orientation and speed upon release from the bow, then the arrow would not have hit the target.
Taking a leaf from Davidson and Grice, we might judge success to be apt only if it derives causally from competence in the right way. Success essentially aided by lucky gusts of wind would not derive in the right way from the archer's competence.
1.4 Assessing the three accounts
1. All three accounts may be rejected as unsatisfactory until we are told what it is for success to derive “in the right way” from the relevant causal sources.
2. We are considering accounts of phenomena that are broadly “factive,” such as perceiving x, killing x, perceiving that P, intentionally øing, and knowing that P. These involve relations spanning mind and world, relations between the subject/agent's mind and her environing world. Philosophical analyses of these various relations are then proposed. And the analyses repeatedly appeal to some essential causal relation.
Thus we reach the nub of the problem. The problem is often posed by deviant causation, wayward causation that gives rise to counterexamples whether the analysis targets action, perception, or knowledge. Time after time, a kind of “luck” derives from the deviant character of the causation, incompatibly with appropriate success and relevantly creditable perception, action, or knowledge.
3. For all such “factive” phenomena, there is a good case and a bad case. In the good case the agent fully succeeds. In the bad case she fails in some way or other.
All three accounts – that of intentional action owed to Davidson, that of perception due to Grice, and the account of knowledge as apt belief – are analyses of the good case into factors, and in all three there would be a highest common factor shared by the good case and the bad case. But in none of them would this highest common factor figure as a conjunct in a conjunctive analysis of the good case into independent conjuncts. Why is there no such analysis? The reason is uniform across the three cases, as they all involve a causal connection said to be present in the good case and absent in the bad case. In each case, the causal connection is just not relevantly detachable from other factors conjoined with it, and with each other, in the analysis.
No state of affairs that comprises relata related by causation will have a metaphysical analysis such that no two factors constitutive of the whole are necessarily related by logical or metaphysical necessity. X's causing Y thus comprises X and Y as relata related by causation, but the whole causal state of affairs is not fully resolvable by analysis into logically and metaphysically independent factors. Even if factors X and Y are logically and metaphysically independent, there is no way of adding a further factor independent of these two that will secure the required causal connection crucially involved in X's causing Y.
4. The upshot is that if causal accounts of perception, action, and knowledge are to be rejected in favor of disjunctive or X-first views, the objection will need to go beyond any assumption that proper analyses must be conjunctive analyses into logically independent factors. In order to clinch their case, opponents of traditional analyses must argue more fully than has been done to date. They must show not only that there is no factorizing analysis of the relevant phenomena into independent conjunctive factors. They must show also that there is no acceptable causal analysis.
Suppose even that essential appeal to “the right way” spoils semantic and conceptual analysis. Suppose firsters are thus right in thinking that there is no such linguistic and conceptual analysis in any of our three domains: that of perception, that of action, and that of knowledge. Whether we are in a position to give it in full detail or not, however, there might still be a metaphysical analysis, even if the formulation of the analysis must make use of “in the right way.” Any formulation we could give might then have to be partial, not complete. Recall in this connection our cat-on-mat example. And compare Leibniz on “infinite analysis.”
Firsters thus owe further argument that there is no metaphysical analysis of perception, action, or knowledge into phenomena metaphysically more fundamental. Such metaphysical analysis is not precluded even if there is no interesting, non-circular, informative semantic or conceptual analysis of the words or concepts in the relevant domains (that of action, that of perception, and that of knowledge).
1.5 An approach through performance theory
What follows aims to turn the tables on objections to traditionalist causal analyses. The use of a concept of “manifestation” will enable causal analyses in all three cases. Appeal to manifestation helps to develop a better solution to those problems. The notion of aptness (success that manifests competence) promises to be helpful not only in the theory of knowledge, but also in the theory of action, and in the philosophy of perception.
Both Davidson and Grice make a crucial move in defending their respective accounts. Even though their formulations are different, the move is essentially the same. They both in effect require a particular sort of causation, while ostensibly assuming that no verbal formula can non-trivially define it. Davidson then says that no such formula is needed, and Grice adds that a grasp of the right sort of causation can be attained through examples. Let us have a closer look.
Recall the waiter who intends to knock over a stack of dishes right now, but does so only through an attack of nerves caused by the nervy intention. Why is this not a way in which a doing can relate to an intention so as to constitute intentional action? What is the required causal relation? Can it be defined so as to reveal why the waiter's doing does not qualify? Davidson claims that we need no armchair analysis of this matter. In his view intentional actions are analyzable as doings caused by intentions in the right way, and no further analysis of the right way is possible or required. We might ask: “No further explication is required for what?” And here is one plausible response: We need not provide a further explication (of what that “right way” is) in order to make any progress. We can at least partially formulate an analysis of intentional action through appeal to appropriate causation “in the right way.”
Still, it would be nice to be able to make further progress, by going beyond such “appropriateness,” beyond invoking “the right way.”
Let us try an account in terms of competences and their manifestations. Consider:
Knowledge is apt belief.
Perception (propositional perception, perception that such and such) is apt perceptual experience, experience whose success manifests competence. A perceptual experience succeeds when it is veridical or accurate. An apt experience is one whose accuracy manifests the relevant competence of the subject's perceptual systems.
Action is apt intention.
In all three cases, the following factors come to the fore:
Success, the attainment of the aim.
The competence of the performance.
The aptness of that performance: whether the success manifests competence.
And it is not just an accident that competence is the key to “the right way.” Again, all three human phenomena involve aimings, performances with an aim. Perception involves functional, teleological aimings, through the teleology of our perceptual systems. Intentional action involves aimings that are full-fledged intentions. Knowledge divides into two sides: a functional perception-like side, and a judgmental action-like side.
The sort of causation essentially involved in all three phenomena is hence the causation of aptness. It is not enough that the success be just caused by competence, for it may be caused deviantly, by luck. Rather, the success must be apt. It must manifest the performer's competence.
1.6 Conclusion of the first section
We find unity across action, perception, and knowledge. All three are constituted by aimings, by performances with a constitutive aim. In perception the aim is functional, through the teleology of our perceptual systems. The aim of an intentional action is obvious in its constitutive intention. Knowledge comes in two sorts. One is functional, so that its aim can be teleological, like that of perception. The other is rather like that of intentional action, and that is because judgment is a kind of action, with judgmental belief the corresponding intention.
When causation figures in the right way in all three of these phenomena, it is hence the causation of aptness. It is not enough that the success be just caused by competence, for it may be caused deviantly, by luck. Rather, the success must be apt. It must manifest the performer's competence.
2. HOW DO WE GO BEYOND GRICE AND DAVIDSON? WHAT SORT OF ACCOUNT IS OURS?
2.1 Understanding and Ineffability
1. Our account goes beyond Grice or Davidson by specifying, in a performance-theoretic way, what the “right way” is in which causation must join together the relevant items: intentions with doings in intentional action, sense experiences with objects in perception, and beliefs with truth in knowledge.Footnote 4
We would like to understand the metaphysics and epistemology of action, perception, and knowledge, which we must attain through certain concepts, even when these are not helpfully expressible through verbal formulas. Nor need they be thus expressible even when widely shared among us.
Really? How do we understand those oracular claims?
2. Just compare how we manage to grasp what politeness is, what it requires. No verbal formula can fully convey or determine (by explicit convention) what is or is not polite conduct. Polite face-to-face conversation sets limits to the proper distance between the partners, and limits the volume of voice and the tone. How is any of this to be captured non-trivially through verbal formulae? It seems quite hopeless. Yet, somehow, antecedent community convention sets those limits. Such convention requires antecedent agreement, at least implicit agreement, which in turn requires shared content, which must apparently be shared even without the sort of formulation that would be required for explicit conventional agreement.
Compare the “manifestations” of a competence. A community might similarly agree (however in the end we understand such implicit “agreement” and its content) on what are cases of “manifestation” of a given competence, even with no helpful verbal formula to cover all such cases. This is like “politeness,” in general and in specific respects. Consider the SSS structure of complete competences (skill, shape, situation), and our concepts of these, and the induced SS and S correlates. Take for example our complete driving competence on a certain occasion, including (a) our basic driving skill (retained even when we sleep), along with (b) the shape we are in at the time (awake, sober, etc.), and (c) our situation (seated at the wheel, on a dry road, etc.). Drop the situation and you still have an inner SS competence. Drop both shape and situation and you still have an innermost S competence: i.e., the basic driving skill that one retains even asleep (in unfortunate shape) in bed (in an inappropriate situation).
Such concepts are broadly shared with no benefit of linguistic formulation. What counts as manifestation seems also graspable only in implicit ways, as with etiquette, and not through explicit (and non-trivial) verbal formulation.
2.2 Competences, Dispositions, and Their Manifestations
1. It is plausible that driving competence comes in three varieties: Skill (basic driving competence), Skill+Shape (skill plus being awake, sober, etc.), and Skill+Shape+Situation (skill plus shape plus being at the wheel of an operative car while the road is dry, etc.) Only with the SSS competence are we fully competent to drive on that road. What determines whether we have the innermost S competence? It is presumably a modal matter: very plausibly what is implicated is that if we tried to drive safely we would reliably enough succeed. But in any conditions? Surely not. It is not at all likely that we would drive safely, even if we tried, when drunk or on an oily road. But this does not bear on our competence to drive safely on a given road. There is an array of SSS conditions that would likely enough crown our attempts to drive safely on that road with success. This would involve certain ranges of the shape we need to be in, and certain ranges of how we must be related to the road, including the condition of the road. We communities who use cars and roads are interested in certain particular combinations of Shapes and Situations, and we are pretty well implicitly agreed on what those are. Innermost driving Skill is then determined as the basis for our likely enough succeeding if we tried in those Shape+Situation combinations.
2. We have a large and varied array of commonsense dispositional concepts: fragility, flammability, malleability, etc. These all can then perhaps be understood in terms of our SSS structures, along with relevant triggers and outcomes. For an outcome behavior of an object to manifest a given disposition, then, is for it to flow causally from its triggering event involving the object, when the object has the relevant Seat, and is in the relevant Shape and Situation. What are the relevant shape, situation, trigger, and outcome associated with a certain dispositional concept. This may simply not be formulable in full explicit detail by humans who nonetheless agree sufficiently in their grasp and deployment of the concept. A particular disposition, then, will have a distinctive SSS profile, with restricted Shape and Situation. Not every disposition to shatter amounts to fragility! Zapper-dependent dispositions, for example, do not count.
3. But why should we have all this implicit agreement on how to categorize dispositions, and their special cases, such as abilities, and in turn competences? Why do we agree so extensively on whether an entity's output is to be attributed to a disposition hosted by that entity, as its manifestation, and by extension attributed to the entity itself?
When the output is good, it is then generally to the entity's credit, when bad to its discredit. The entity might be an agent who manifests a competence, or it might be a lifeless patient manifesting a mere disposition. Why do we agree so extensively on these dispositions, abilities, competences, and on the credit and discredit that they determine (whether this be to the credit of a moral agent, or to the credit of a sharp knife), and on the sortals that they help constitute?
Is that not all just part of the instrumentally determined commonsense that humans live by?
Such commonsense helps us keep track of potential benefits and dangers and how the bearers of these are to be handled. As a special case of how properly to handle things and agents that manifest dispositions and competences, we have propriety of encouraging praise or approval, or discouraging blame or disapproval, which in turn helps to fix the relevant dispositions, abilities, and competences in ourselves and in our fellows.
Such an instrumentally determined commonsense must of course be structured against background implicit assumptions about what is normal or standard, either in general or with respect to the specific domain of performance that may be contextually relevant. Many are the domains of human performance that allow and often require degrees of expertise beyond the ordinary: athletic, artistic, medical, academic, legal, etc., etc. Expert perception, agency, and knowledge would be determined proportionally to the respective levels of competence set for the specific domain. This is often set largely by convention, or, for more basic competences and dispositions, by the requirements of success in our evolutionary niche. After all, how we credit, discredit, trust, and distrust, has a large bearing on human flourishing, individually and collectively.
Manifestation determines credit and discredit, and also is attributable causally to the host of the manifest disposition in a way that is projectible, though this is no more amenable to formulation than seems the projectibility of greenness (or ‘green’) by contrast with grueness (or ‘grue’). When something shows its true colors through manifestation, we can take notice and revise our view of what to expect from the host of the manifest disposition. This is in contrast to when the disposition is only mimicked, so that the correlated trigger prompts the correlated ostensible manifestation, but only through the trumping action of the mimic. Such fake manifestation is not to the relevant credit, causal or otherwise, of the disposition. And the host in turn acquires no credit or discredit thereby.
4. Consider the mimicking of fragility when a fine wine glass shatters because it is zapped upon hitting the hard floor. By hypothesis the causal action of our zapper trumps the inner structure of the glass, whereby it normally shatters on impact. Still that inner structure can be causally operative, as it is through the agency of the zapper (who hates the impact on the hard floor of the fragility that he spots in the fragile glass). Despite being causally operative in that way, through the knowledge of the zapper, that inner structure is not causally operative in the right way. And this is why the fragility that we normally attribute to the glass is not really manifest on that occasion.Footnote 5
That is why, although the competence-theoretic account goes beyond the proposals of Grice and Davidson, it may itself eventually need to rely on our shared grasp of causation in “the right way.” We may need to rely on our shared grasp of how competences are constituted and manifested: for example, of when the relevant causation is not deviant or wayward.
That requires an SSS-joining of seated Skill, Shape, and Situation, so as to cause the manifestation upon the onset of the trigger. And this must occur in an appropriate way. Consider for example what is required for a true manifestation of fragility as a fine wine glass shatters upon hitting the hard floor. By common consent such true manifestation excludes the zapping of the glass by someone who hates fragility meeting hardness, even if, through his knowledge and action, he does manage to link causally the fragile structure of the glass with the shattering upon impact.
As we focus on examples such as that of fragility zapped, it emerges that we may still need reference to “the right way,” by requiring that a disposition can be manifest in a certain outcome only if it accounts for the outcome in the right way.Footnote 6 This requires a joining of seated skill, shape, and situation, so as to cause the manifestation, upon the onset of the trigger. And this must take place in the normal way, which by common consent excludes the action of a zapper, even if he does manage to link the fragile structure of the glass with the shattering upon impact.
2.3 How We Go Beyond Appeal to “The Right Way”
1. That is to present our account with a certain modesty. More boldly we might claim that when the zapper shatters the glass because he knows it to be fragile, the shattering does not intuitively manifest the glass's fragility. Anyone who joins me in finding that plausible enough can make the following bold claim:
Manifestation enables us to go beyond the need to appeal to “the right way.” The manifestation of competences and other dispositions then provides a solution to the problem of specifying “the right way” as it pertains to action, perception, and knowledge.
That includes the problem of causal deviance. But it also includes the problem faced by Grice in his analysis of causation, which is not exactly a problem of deviance, though it is closely related. The causal bearing of our eyes on our visual experience is not really deviant, nor is that of the sun even when our eyes are open to the daylight with the sun out of view. In any case, Grice too must rely on an assumption that the causation linking an object and one's sensory experiences must be causation in a particular way to be brought out through examples.
Both problems of specifying the right way are solved through a primitive relation of manifestation that has outcome manifestations (successful performances) on one side, and competences (perceptual, agential, epistemic) on the other.
2. Some may well remain skeptical of the powers alleged for our primitive concept of “manifestation.” To such skeptics we can offer, as a fallback, a more modest option whereby, perhaps through examples, we can still explain what is required for proper manifestation. And we could even disown any ambition to rely exclusively on explicit verbal formulation (as by invoking “manifestation”).Footnote 7
Even on this more modest option, we will have made progress. We will have specified more fully the sort of causation involved. And we will have seen that it is the same sort of causation in all three cases: that of perception, that of action, and that of knowledge.
3. METHODOLOGICAL CONTEXT FOR OUR INQUIRY
We often appeal to what we would ordinarily say, and even to what we would ordinarily think, in the exercise of generally shared concepts. But our main interest in philosophy is not restricted to semantic or conceptual analysis. When we wonder about personal identity, freedom and responsibility, the mind and its states and contents, justice, rightness of action, happiness, and so on, our main focus is not, or not just, the words or the concepts. There are things beyond words and concepts whose nature we wish to understand. The metaphysics of persons goes beyond the semantics of the word ‘person’ and its cognates, and even beyond the correlated conceptual analysis.
The same goes for epistemic concerns such as the nature of knowledge and other epistemic phenomena. Consider the semantics of epistemic vocabulary, and even the conceptual structure of epistemology and its normativity. It seems an open possibility that our words and concepts are not in the best shape just as they stand for grasping and understanding the relevant domain of objective phenomena. Why not leave open the possibility of terminological and conceptual improvement in epistemology in a way analogous to what is familiar in science when we reconfigure the terminology and ideology of fish, for example, or of vegetables, fruits, and much else? If so, semantic and conceptual analysis might still remain an excellent start in epistemology. Such analysis would remain important in various ways to the epistemologist, and to the philosopher generally, but we might also be able to delineate phenomena whose importance is obscured by ordinary speech and thought.
If so, that might also, as a bonus, help throw light on pervasive and persistent disagreements so common in philosophy. Some of us may just be trying too hard for the exact, fully general analysis that will apply smoothly and directly to all thought experiments. And those of us who are willing to stick to a simple and illuminating take on some range of phenomena may just be right to “bite certain bullets,” if by so doing we can distinguish a type of phenomenon that seems plainly important in the domain of interest, such that it should also prove illuminating to consider its relations to other such phenomena. We may then reject an ostensible counterexample, while allowing that the example points to some further interesting phenomena interestingly related to the phenomena of more direct and central interest to us in our specific inquiry.
Philosophical progress might then take a form similar to the kind of scientific progress that involves conceptual innovation. We may find in the phenomena themselves differences that seem important even if there are no proprietary terms or concepts that correspond to them neatly and without exception. If so, it may behoove us to stretch close terms or concepts so that they will help us to mark the relevant phenomena, and to cut the domain more closely at the joints.
Finally, once our objective is analysis that is metaphysical, rather than linguistic or conceptual, the bullet-biting recommended does not amount to giving up on intuitions. The metaphysical project is driven crucially by intuitions concerning the phenomena themselves, and not the (proper use of) the language used to describe them, nor the content of the related concepts. After all, to bite the bullet, on the proposal floated, is precisely not just to describe or understand but to change our given language or concepts, at least by addition, but perhaps also by subtraction, or by modification. True, the relevant metaphysical intuitions will need conceptual content, but our focus on the phenomena may lead to concepts that are modified, or even quite new. We need not be restricted to concepts used when we begin our inquiry, as our inquiry may properly lead to revision.Footnote 8
