The Contextual Nature of Cognition and Dancy's Moral Particularism
by Mark Lovas
... a moral principle may draw attention to a feature that is always relevant, and relevant in the same way, under certain implicit conditions... We don't normally spell out those conditions ... But the list of conditions is not open-ended, and it is knowable in advance. (D. McNaughton and P. Rawling, 2000, p. 269)
Principlism, or ... the *Leibnizian theory of inquiry, has been regarded, and is in some circles still regarded, as a live option. Thus the context problem has long remained invisible, and is still hard to grasp from certain standpoints. (Andler, 2003, p. 364)
Oh how post-Socialist countries have problems! I recently read about a Czech university student who returned home to find her apartment empty. She hadn't exactly been robbed. A court had ordered that a different woman's possessions be re-possessed; but, unfortunately, due to a bureaucratic mix-up the re-possessors went to the wrong address. The officials responsible for the mix-up were in no hurry to admit their mistake; they insisted that the student produce sales' receipts or reliable witnesses who could testify that the books, CD's and furniture really belonged to her. The whole affair was so upsetting that the student had to postpone graduation, but today she vows that she will save all of her receipts.Foot note 1
One might wonder whether saving literally every sales receipt is a realistic policy. It is, perhaps, every bit as realistic as wondering about the epistemic credentials of each of one's beliefs, or habitually doubting one's first thought about what to do next. Skeptics can make us insecure about our evidence, and in the realm of morality, Dancy's particularist arguments for the holism of reasons could make us doubt our thoughts about what to do next. Dancy claims that any given reason which is, in a given context, moral justification for an action, might lose that weight or valence in a different context. A reason might count for an action in one context, but against it in a different context. The skeptical thought is: what if the context changed without my noticing it? Or, what if I failed to notice a crucial aspect of the situation? If I am about to perform an action because I think it's just, should I scrutinize the situation to ensure that there are no hidden defeaters able to rob my reason of its usual weight?
Dancy suggests that there may be times when cruelty, rather than kindness is called for, and times when we should lie rather than tell the truth. There may even be cases where the just thing is the wrong thing to do on account of its being just.Foot note 2 More generally, so-called thick moral concepts such as «integrity, fidelity, gratitude, reparation, and so on..» (Dancy 2004, 121). can't be expected to provide reasons in every context-come-what-may, and we cannot suppose that a good reason will always be good in no-matter-what context.
Dancy is no skeptic (see, e.g., Dancy 2004, p. 1) and it would be a misunderstanding of his views if we went away doubting every reason that occurred to us, but his arguments should at least make us curious. How do we do it? If there's nothing necessary about a given reason's supporting an action, how do we recognize that it does support the action, when it does? How do we manage to change our minds when contexts change? If we have moral knowledge, shouldn't we be able to give some account of how we manage to employ the reasons which, in a particular context, are good ones?
Dancy has surprisingly little to say on these matters. In Moral Reasons (Dancy 1993, Chapter 7, Section 2) he appeals to the notion of salience. The agent doing the right thing perceives the situation as having certain saliencies. She can explain her choice by giving the right sort of description of the situation, one which makes salient what should be salient. She tells a revealing sort of story. In his most recent book-length treatment, (Dancy 2004, pp. 160-161) he tells us that our ability to grasp reasons is primitive. Without rejecting the holism of reasons, I think that it may yet be possible to say more, and I shall attempt to do so below.
In broad outline, I would like to borrow a suggestion made by Daniel Andler. (Andler 2003) We should not think about context as if it were another thing, as if it were to be dealt with by adding another variable into a principle, or as if it could be the object of a special sensitivity. What we call adjusting to the context is just the creativity and sensitivity which a person shows when rules-of-thumb or principles don't work. This insight allows us to deal with a peculiar response which Dancy's examples can provoke. They can make us wonder how it can be possible that we track good reasons. The insight here is that what we are keeping track of is, simply, the basic moral features of our lives.
The interpretation I wish to suggest is morally robust and natural. It surfaces when Dancy writes:
The particularist thinks that there are times when such ordinary considerations should be given great weight, and cases where they might be out-weighed. The ability to know when to do what is a matter of perceiving such reasons and being aware of any competitors. To speak of «context» here is a way of describing that very perception of the weight of his distress or the damage to her self-image. There is not some additional calculation of the weight-in-this-context; to perceive it as mattering is to already to have done any needed calculation.
I also propose to borrow from Kent Bach's discussions of reliabilism and default reasoning. (Bach 1984, 1985, 2005) Insofar as moral agents know when to think twice, and know when to simply trust their first thought about what to do next, their actions will tend to be morally right. This also enables us to accomodate an insight of Andler's. Andler (2003, p. 360) notices that one odd feature of a principlist view of inquiry is that when inquirers are creative, they seem to be «shooting in the dark». Bach (1984) used the pejorative phrase «jumping to conclusions» as part of his title in a paper about default reasoning, but he sketched a view which deprived the phrase of its pejorative tint. The general point is that we bend and break rules when we see fit, and that's not a failing. That's just what we do, and criticism of the process takes place when we implicitly compare it to what we don't do, invoking an ideal of inquiry which no inquirer instantiates. In the end, serious questions may remain about how we do it, but framing Dancy's particularism in this broader context should lead to a truer evaluation of its strengths. Dancy's particularism is a sustained attempt to recognize the contextual nature of moral cognition.
The Problem of Context
In a recent article, (Andler 2003), Daniel Andler provides the following characterization of the «contextual nature of cognitive processes»:
Suppose we have a principle or law which tells us that things of such-and-such a type have certain properties. Andler is suggesting that even after we've classified an object, we are, so to speak, not done. I may have blamelessly and correctly classified an object or situation as belonging to a certain category, intending thereby to understand it through the laws governing such entities, but my work is not over. There are other properties of the object or entity which I've not recognized, and those properties are not insignificant. By the same token, in the case of action, I may recognize my situation as being of a certain type, and thereby imagine a certain course of action is the right one. However, my classification does not exhaust the situation's possibilities, and there may be a «defeater» existing in the world outside my classification. It is illuminating to consider Andler's own further expansion on this remark:
(Andler 2003, p. 352)
Andler recognizes two broad reactions to the general problem of context. The deflationist thinks context is not a big problem, and is committed to using principles. The paradigmatic deflationist will be a *Leibnizian -- with the asterisk indicating that Andler is not engaged in a historical inquiry. The *Leibnizian thinks the principles can be modified, or tweaked just enough to capture the context. The key point is that inquiry is, above all, a matter of following principles. The other reaction is that of the inflationist. The paradigmatic inflationist turns out to be a kind of sceptic -- not about our knowledge of the real world, but about our ability to describe how we do it. (One example mentioned by Andler is Fodor 2000)
Andler allows that *Leibizian methods can take us far, but they have inherent limitations. There is, on Andler's account, a deep incoherence in the *Leibnizian account because all situations represent a sort of information, and the view has not got the resources to mark a principled distinction between text and context. (Andler, 2003, pp. 359-60). This point is echoed in Dancy's complaint that critics ignore his distinction between enablers and reasons. (See, especially, his complaint about Crisp in this regard in Dancy 2004, pp. 96-7) For Dancy, my decision to keep a promise does not include as a reason the fact that my promising was un-coerced. On the contrary, the fact that I was not coerced is not a reason at all. By contrast, McNaughton and Rawling (2000, pp. 271-2) claim that the fact that I was not coerced is among my reasons for keeping a promise.Foot note 3 In making this move, McNaughton and Rawling are feeling the attraction of a *Leibnizian view.
The second problem for the *Leibnizian is novelty. No investigator simply gives up because his explicit principles won't capture a particular case. (Andler 2003, pp.364, 370) What people actually do is make a guess, or use their best judgment or insight. (369-70) And, that is where Andler, who is proposing an epistemological view, wants to get some support from recent discussions of moral particularism. Andler claims that recent discussions of moral particularism show that principles alone can't be the whole story:
Andler does not discuss the ethics literature in detail, but his general observation seems correct, so far as it goes. Even Dancy's critics admit a need for «context-linked abilities», however, I am suggesting that even within the literature which Andler refers to, *Leibnizianism continues to exert an influence.
Dancy and Context
Jonathan Dancy can be classed as a context inflationist. Dancy's argument for the holism of reasons -- the claim that a supporting reason in one situation need not retain its supporting value, and might even switch value in a different situation -- evinces a lively sensitivity to the role of context. Consider the following example:
We have two situations where you've loaned me a book. So far as that classification goes, they are both tokens of the same type. But, there is a moral difference between them which brings out the difficulty in attempting to derive an action-guiding rule applying to all situations in which someone has loaned me a book. And, that is Dancy's general point: due to the holism of reasons, we can't expand particular reasons into principles; so, principles have a very limited role to play within ethics. (For an account of why principles seem important, and the limited role which Dancy recognizes for them, see Dancy 1993, 66-71)
Someone might say that the two situations are not equally of the same type. One situation involves a masquerade. The friend who «loans» me the book is actually impersonating someone who has rightfully acquired a book: you cannot really loan something which you do not own in the first place. In raising this point, however, we are making Dancy's point. There is a moral difference between the agent who loans what he owns and the agent who loans what s/he stole, but the attempt to describe the situation, as it were, in advance of complete (or fuller) knowledge runs into the danger that we have missed the very facts that are morally relevant.
One might well ask how we are sensitive to the context. After all, since the weight of reasons depends upon context, insofar as we are aware of reasons, must we not be (indirectly at least) sensitive to the context? It may well appear that Dancy simply hasn't got an answer, even if he is no skeptic.
Dancy returns to this general issue briefly in Ethics Without Principles. Part of his particularism is a denial that we can give anything like a general set of criteria which determine whether something is a reason: «... particularists tend to deny the existence of any substantial general theory of what it is to be a reason.» (2004, p. 160)
Dancy makes an important distinction between reasons and «enablers». Consider the question: Why should I keep a promise? Dancy suggests that I do what I promised because I promised. It's true that if I had been coerced into promising, my promise would not be binding, but that's not my reason: it's an enabler for my reason. (Dancy 2004, pp. 38-9) Similarly, in the example above, that s/he didn't steal it was an enabler for my friend's loaning me the book. Sometimes Dancy makes the claim that there are «too many» such potential defeaters for them to be reasons. (e.g., Dancy 1993, p. 53) To repeat what was mentioned above, this shows Dancy employing a distinction much like that between the «text» and the «context» -- something which, according to Andler, the *Leibnizian cannot find room for. Dancy has consistently claimed we don't have to think about them, or know such background conditions in order to act morally, or to know what we are doing. (Trenchantly expressed in Dancy 2004, p. 142.)
Should we be satisfied by what Dancy has to say? Once we realize that any given reason can lose its weight, we've lost the ability to think of ourselves as following action-guiding principles. At best, a principle works in a restricted area, but there's no hope of spelling out what the relevant restrictions are. Commentators have wanted to argue that not all reasons are like that, but their assertion looks like a bold-faced and stubborn re-assertion of what Andler characterizes as *Leibnizianism. Dancy allows that there may be a few reasons with constant valence -- justice may be one -- but, in general, something more than following principles is needed.
However, a puzzle does arise once we notice that no reason need stand firm as circumstances change. How are we to know if we are in a situation where a reason stands firm or not? Somehow I notice the reason, and somehow the enablers are in place, and no defeaters are present. But, doesn't it begin to seem like a miracle?
One wants to say that context is by definition (to use a notorious phrase) background. It's something like the lights in a theater or the camera angle in a movie. It has an effect upon us which enhances meaning, but it's not generally part of what we are interested in. Yet, without the right lighting the performance loses something, and camera angle can contribute much to a film's content. The effect of Dancy's examples is to foreground context, and thereby highlight the actual subtlety of our abilities.
Perhaps I can make my point in another way. Typically, Dancy's examples have a clear basic format. We imagine one example: In a given situation, S1, a reason, R1, is a reason to perform an action, A1. Then, we imagine a slightly different situation, where R1 is also present, but R1 does not support A1. A1 is not right in the second situation. It's right to lie in a game, but not in real life. In general, one should avoid cruelty, but if someone is torturing you, perhaps you should refrain from being considerate. (Dancy, 2001) In all of these examples, Dancy's point is not to raise a sceptical «what if?» question, but to emphasize the subtlety of our sensitivities. When commentators react by stressing that some virtues (honesty) or some features (justice) always count for an action, they seem to fail to notice that Dancy is stressing the subtle nature of our very real capabilities. He is, in effect, suggesting that in cases where justice (or honesty) is not what a situation calls for, we are capable of recognizing it. When Dancy describes such examples, he is presupposing that morally sensitive individuals are able to notice the difference.
One might persist in asking, «But, how do we do it?» When Dancy endorses the possibility that our ability to recognize reasons is simply a primitive faculty, he rejects the question:
It seems to me that Dancy's view is, at this point, rather hasty. On the one hand, Dancy rightly does not wish to offer an analysis of, say, «reasons» or «good reasons» in terms of a set of necessary and sufficient conditions. On the other, he is quick -- possibly, too quick -- to employ the weapon of particularism.
This is clearest in his response to the challenge that particularism leads to a kind of scepticism because there are simply too many defeaters and enablers. The objector says that one can't ever know enough. Notice how the context is expanding. What we'd like to keep in the background has now occupied the foreground, and our knowledge, it seems has disappeared.
Dancy's solution is to look for a sort of «filter» -- a principled way to limit the epistemic demands upon a moral agent. Ah, but notice; I have lapsed into speaking with the vulgar: what can it mean for a particularist to seek a «principled» limit? The particularist, it seems, must say -- and Dancy does say -- that there's no drawing a line:
Without troubling ourselves too much just now over the exact content of this subtle disjunction of ingredients -- whose complexity Dancy himself implicitly acknowledges when he later remarks upon «tortuous» epistemological problems (p. 159, 2004) -- I would like to suggest that Dancy has gotten himself focused upon the morality of the agent, and not the morality of the action. A well-meaning, sincere, good agent might do moral harm to others. (Cf. Honderich 1996) But, then, perhaps, here I assume a consequentialism which Dancy would reject.
It is enlightening to consider how Andler deals with the problem of the expanding context. Andler shares with Dancy the recognition that the potential search space can expand to such an extent as to overwhelm our resources. Thus, the problem is to limit it -- as Dancy says to find a «epistemic filter». We are able to solve problems insofar as the space of possibilites has been, in some way, cut down. If Dancy asserts (and his opponents deny) that the background conditions -- the enablers -- are too large a class to be an object of our apprehension, to be grasped as a reason, and then asserts our grasp of reasons is «primitive», how is that any better than saying that our ability to recognize good reasons and act morally is a miracle? Andler draws an analogy with treasure hunt games, textbook solutions, Relevance Theory's approach to communication (Sperber and Wilson 1995) and recent suggestions by Dan Sperber and others that many of our daily survival problems are solved for us by the software which comes as standard equipment in our brain -- the thesis of the «massive modularity» of the mind. (Defended, e.g., in Sperber 2001) Common to this varied list is the way in which the search space is limited. A textbook problem is solvable because the relevant chapter has given the student the materials needed to solve it. So, too, a treasure hunt has clues. The apparent infinity of the search space is whittled down. In the case of Relevance Theory, a communicator exploits aspects of the environment which are mutually salient to the audience. Modules are psychological features which are supposed to allow us to solve problems faced by our evolutionary ancestors: they are narrowly focused upon specific problems, and for that reason do not have a problem with context.
In the case of ethics, I would make a proposal not explicitly found in Dancy's writings -- though it would involve resurrecting and modifying the notion of salience which appeared in Moral Reasons. The search in ethical space is, I suggest, limited in somewhat the same way as the search in communicative space -- via mutual saliencies.Foot note 5 Recall Dancy's examples mentioned above: the damage to her reputation or his distress. Ethical reasoning, and the communication of ethical reasons is possible because, in general, we have such sensitivities. (The need for a qualification is suggested by research concerning psychopaths, who lack sensitivity to the suffering of others. (See Nichols 2002)
The trick is not to conceive of the moral agent's problem as knowing when a reason has been disabled. We don't perceive enablers and disablers. We directly perceive reasons. This follows if we take seriously Andler's suggestion that context is not a thing which can be the object of a sensitivity, not something we sense with a context-sensor. Context is not another thing, over and above our reasons for acting and the actions we perform. When Dancy takes enablers and disablers as though they were possible objects of knowledge, (2004 p. 158) he has taken context to be a thing. Consider Andler's words:
(Andler, 2003, p. 368)
If it is theoretically misguided to try to describe a «situation-sensor», then what looks like the absence of an epistemology in Dancy's discussions of moral particularism may, in part be, only evidence of the rejection of a *Leibinizian view of inquiry. If we conceive of context as a thing and try to adjust our principles to capture it, then we wind up with the sort of view expressed by McNaughton and Rawling in the quotation on the first page.
The deflationist thinks that by complicating principles, s/he can tame context. The problem with that solution is that it misses the fact that there's a sort of insight and creativity which we call upon when we notice that a principle doesn't apply, or that a principle needs to be tweaked. (See Andler p. 360) If context is not a thing, then what, we might well ask, is it?
Dancy's moral particularism draws our attention to our ability to proceed creatively and intelligently in a moral situation. His critics have missed this, and underestimated the depth of his thought. Dancy's silence about epistemology is largely motivated by a rejection of an implausible view of inquiry which Andler terms *Leibnizianism. However, Dancy's silence on these issues may have other causes as well. Above I suggested that we can solve the epistemological problems raised by moral reasons if we suppose those reasons are apprehensions of the effects our actions have upon other human beings. Moreover, moral communication -- the expression of and understanding of a moral agent's reasons -- is possible because we share mutual saliencies. These suggestions, once developed, may well lead to a full-blown form of consequentialism. Dancy (1993) has argued that consequentialism fails as an account of moral reasons. In order to fully defend my suggestion, I would both need to provide more flesh on the bones that I've so far offered, and to deal with Dancy's objections to consequentialism.Foot note 6 Moreover, there is need to examine the proposal that among the modules of the mind is one for reasoning about the mental states of others. Do the concrete proposals which are available support my project? That is partly an empirical question, but also a question about the current state of theoretical development. I plan to carry out that project in a sequel.
Despite a lengthy lacuna between inspiration and deed, the original impetus for this paper was provided by Simon Blackburn's National Endowment for the Humanities Summer Seminar, «Objectivity and Emotion in Practical Reasoning.» I would also like to thank Palo Cmorej and the Philosophical Institute of the Slovak Academy of Sciences, as well as the helpful librarians of the Academy's Reading Room. Without access to the Academy's resources, it would have been impossible to write this paper. Paula Gottlieb and Andy Piker were kind enough to read and comment upon the paper; time constraints prevent me from addressing all of their questions here.
Submitted: 7 March 2005
Accepted for Publication: 13 June 2005
This Revision: 24 September 2005
[Foot Note 1]
The editor points out that proof of ownership was already required in Roman times. There are, however, important differences between Ancient Rome and Central Europe of today. Friends of capitalism are fond of pointing out how industrialization brings us a much larger number of material possessions, more even than were previously available to the rich. Additionally, in the implicit background of my little anecdote lurks an especially detestable character: The Central European Bureaucrat, a person who begins with the assumption that you are guilty and need to prove your innocence, and expresses her superiority via a variety of gestures all with the subtle, indirect, but clear message that you are a disgusting inferior and that anything she might happen to do for you is a favor granted by a higher being to a lesser entity. The most perfect artistic expression of this cultural phenomenon is found in the writings of Kafka.
[Foot Note 2]
To the best of my knowledge, Dancy has not strenuously argued that justice can sometimes count against doing an action. He tends to argue in a very schematic way, saying the particularist must say this--on the basis of the general position. There may be something odd about that. In one place where there is a bit more substance (Dancy 2001), the claim is that justice sometimes might not be the strongest reason in favor, and a very particular theory of the nature of justice seems to be assumed--roughly one where property rights define justice. Arguably it is in virtue of that particular view that justice loses its weight. However, it isn't Dancy's own view that's being discussed in the relevant passage.
[Foot Note 3]
McNaughton and Rawling (2000, p. 271-2) claim that Dancy has confused what it would be odd to say with what is false. If we consider a comparison with discussions of perception where the problem is to specify what is intuitively the right object being seen, McNaughton and Rawling's view seems less plausible. Suppose that if I am to see my computer, some disjunctive set of neurons must be activated. It seems not merely odd to say that when I see, I see those neurons which are firing. The parallel is that in a given moral situation I perceive certain aspects of the situation. I do not perceive the causal background which makes that situation possible. (Cf. Recanati 2002, pp. 109-110 footnote and references therein.)
[Foot Note 4]
Cf. Dancy 1993, p. 50 where Dancy says that the virtuous agent brings to each new situation «a contentless ability to discern what matters where it matters»(p. 50) This is surely an over-statement; not all content is linguistic. I suggest that the most that Dancy could say would involve using Sperber and Wilson's notion of a «code» model of communication. We cannot encode our thoughts about moral reasons. If, as Carston (2002, Chapter 5) suggests, psychological concepts are especially prone to shifts of meaning to fit specific contextual needs, and if psychological concepts are essential to moral reasoning (as I suggested above), then we may have an independent path to particularism. The very concepts we operate with get modulated in specific contexts so that there is no guarantee that they retain their exact content across contexts. Whether the path is independent, and how exactly it is to be described is something I plan to discuss in future work.
[Foot Note 5]
Levinson (1995) argues that communication requires mutual salience: «...a notion of `natural salience', such that I can be sure for an indefinite range of phenomena or scenarios that what is salient for me is salient or you.» (1995, p. 230). The notion has it basis in research by Schelling discussed in his Strategies of Conflict. (Schelling 1960) Levinson remarks that both Schelling himself and commentators agreed that in order to explain the remarkable ability of people to coordinate their actions without explicit communication--e.g., to agree upon a meeting place or a specific number--«roughly nine times out of ten», it was necessary to appeal to a notion of mutual salience.
[Foot Note 6]
Insofar as my speculation here is based on Dancy's examples involving harm to other individuals, it might be that a form of «sentimentialism» is required--minimally, a theory giving prominence to our capacity to feel pain at the thought of causing pain to others. One such recent theory is found in Nichols 2004. (This reference may seem surprising to anyone who has read that work, as Nichols dismisses Dancy's particularism very swiftly in a footnote (Nichols, 2004, p18, fn. 6). On the other hand, Nichols never considers the sort of arguments I discuss above.)