The Normative Weight of Convention

Robert Bass

There is something of a convention among philosophers (and among academics more generally) that supports being critical of conventions, skeptical of tradition, doubtful of the utility of received practices.¹ A question might be raised as to what justifies such a practice, but, for the present, I shall bypass it. More precisely, I shall take it for granted that the practice is justified somehow or other. But if it is justified, it is in turn a justifier of specific actions or behaviors undertaken in accordance with it. That is, it has normative weight. How can it be the case that a convention² - which could have been different - has normative weight?

In order to address this question usefully, it must be delimited or distinguished from others with which it might be confused.

First, the question is not: How can anything have normative weight? (And do conventions qualify?) That, I take it, is too deep a question. I shall assume that some things do have normative weight and, where appropriate, will try to select uncontroversial examples.

Second, the question is not whether conventions can inherit normative weight. It is not, for example, being asked whether a convention against murder could or does have normative weight. Presumably, it does. But whatever normative weight it has stems largely, perhaps entirely, from the wrongness of murder, hardly at all, perhaps not at all, from there being a convention against murder.³

Third, the question is not whether conventions have minimal normative weight. Something has minimal normative weight if it or action in accordance with it is acceptable or permissible unless there is some good reason⁴ against it. Conventions may certainly have that degree of normative weight because anything - any desire, motive, goal, whim, habit or rule, anyhow - may have that degree of normative weight. How conventions can have minimal normative weight is no more puzzling than normative weight in general.⁵ For a convention to have an interesting degree of normative weight, it must be at least as weighty as what was earlier called "a good reason." That is, it must be weighty enough that it could sometimes provide a good reason against doing something that itself has minimal normative weight. It must also provide a good reason for doing something else instead. It would not be enough that it was simply something which was de facto incompatible with the first action - and therefore a permissible alternative. It would have to have enough weight to override (at least some) alternatives.

With these restrictions in mind, return to the example with which we began. There is a convention that supports criticism of conventions (and presumably, it thereby indirectly supports the abandonment or modification of any conventions that succumb to criticism). Suppose this convention is unjustified. Then there are some conventions that have more than minimal normative weight, namely, those that it would be wrong to modify or abandon in the face of criticism. Suppose, on the other hand, that the convention authorizing such criticism is justified. Then, at least this one convention has more than minimal normative weight.

One might be tempted to say that either way, some convention or other has more than minimal normative weight. But that would be too quick. The convention supporting criticism might have only minimal normative weight. It would then not be wrong to act on it but also not wrong to act otherwise. Then it might seem that other conventions could also have no more than minimal normative weight. If they were unjustified, then criticism of them would have overriding (not merely minimal) normative weight while if they were justified, criticism would be wrong. But there is another possibility: It might be that some of the conventions have enough normative weight that there is no danger that they would be overturned or weakened by criticism. On this assumption, it would at least be possible for conventions to have more than minimal normative weight, and there does not seem to be any way to rule out such conventions.

Is it really intelligible, though, that no convention might have more than minimal normative weight? I doubt it. Suppose that the convention authorizing criticism has no more than minimal normative weight. Criticism would be permissible. It would also be permissible to adhere to a criticized convention and ignore the criticism. This would be unproblematic if the criticism were somehow defective. But suppose that it was cogent and persuasive. It seems then that one could (justifiably) adhere to the criticized convention only if it does have more than minimal normative weight.

So, given the existence of a convention that supports criticism of conventions and given certain plausible assumptions about normative weight in general, it appears that there must be some conventions - at least one - that have more than minimal normative weight.

Now, if some conventions have normative weight - I'm now dropping but still assuming, the qualifying phrase, "more than minimal" - could it be that the weight they have is entirely derivative or inherited from something else (not a convention) that has normative weight? There are too many possibilities to be addressed here within brief compass, but I'd like to focus upon one source from which conventions might be thought to derive their normative weight. Could it be that the normative weight of conventions is entirely derivative from articulate reasons - that is, the sort of reasons that might be presented in an argument?

Consider here three cases which may overlap: (1) There could be a reason for adhering to a convention, perhaps articulable, perhaps not. It would be a reason that adherence to the convention is good, desirable, correct, appropriate, etc. There might be such a reason whether or not (2) I have a reason. A reason that I have would be one that has some motivational purchase upon me. When there is a reason but I do not have one, it may be that I am unaware of it, misunderstand it or am in some way at fault for not responding to it. But when I do have a reason, I may or may not be able to (3) give a reason. That is, I may or may not be able to explain or articulate why I adhere to the convention.⁶

So, even if all reasons for adhering to a convention are in principle⁷ articulable, the person who must act upon the reason if it is to be acted upon at all (at a particular time and place) may not - either at the time or possibly ever - be in a position to articulate it. Thus, though the convention may in some sense derive all its normative weight from an articulable reason, it is at best misleading to say that the person who acts upon it does so because of that reason. Further, if the person has a normatively weighty reason (for which she cannot, at the time, provide an articulation), it might be sufficiently weighty that it could override an articulate criticism of it.

In summary, if, as has been argued, some conventions have normative weight and granting, for the sake of argument, that all of their normative weight is in principle derivative from articulable reasons, it still would not follow that the person who acts upon one of these conventions does so because of those articulable reasons or that adherence should or must yield in the face of articulate criticism.

There is a further question to which I can offer no more than the sketchiest of answers. Though I have argued that it is not the case that convention must always yield in the face of criticism, I have assumed throughout that not all conventions are created equal and that it is sometimes appropriate for a convention to yield. Briefly: How do we tell the difference? How do we know when to adhere and when to abandon?

It might be said that we should adhere to successful conventions, but it is not clear how we can judge success if we are not sure what a convention needs to be a success at accomplishing. Or, the attempt could be made to define "success" in terms of the fact that a convention has evolved and persisted. That kind of success, however, will apply to all conventions and so will not distinguish those to be preserved from those to be abandoned or modified.

Here, we are faced with problems where the level and quality of the articulate reasons we can bring to bear seem quite inadequate to the problems upon which they must bear. Perhaps only two things in general can be said: First, the fact that we don't have an explicit algorithm for dealing with a problem does not mean here, any more than it does in other areas, that we can do no better than random guessing. Our sense of what is relevant or not, of what is problematic or not is also an evolved capacity, worthy of respect. Second, it might be well to treat conventions with a bit of humility as suggested by Orgel's Law (originally applied to biological evolution): Remember - Evolution is cleverer than you are.

Comments? I'd love to hear.

1 Perhaps the most important factors explaining this are economic. Academics function in a competitive market in which one can only rarely acquire the sort of reputation that improves prospects for employment and promotion by simply reiterating what others have said. This is not, of course, true of all competitive markets nor even of all in which the prospective employee is expected to engage in largely intellectual tasks. Clerics, for example, may be selected in part for their capacity to reiterate traditional doctrines without apparent insincerity.

2 I'm going to use "convention" as a general term to cover a wide gamut of regular and ongoing human activities that are shared by significant numbers of people. That they are shared is essential: They are shared, at least in part because they are shared. If this is opaque, an example may clarify matters. Hand-shaking as a way of greeting people would not be possible if there were only one hand-shaker. It is possible to greet people by shaking hands because there are (already) large numbers who greet with a hand-shake. Conventions may range from the relatively trivial, such as rules of etiquette, to the relatively momentous, such as "the traditions of western civilization." They do not, however, cover all regular or widespread human practices and behaviors. It is not, for example, a matter of convention that people sleep regularly.

3 Perhaps, as the qualifications suggest, there is some reason for having a convention against murder which is not simply a function of the wrongness of murder. But if there is, and if that weight cannot be cashed out in terms of other things (than conventions) that have their own normative weight, then that is the portion of the normative weight of the convention in which I am here interested.

4 In the present paper, I am using the word, "reason," exclusively to refer to practical reasons - reasons for taking or refraining from actions. I am not referring to explanations, premises of an argument or the like except where these are or are part of reasons for taking or refraining from actions.

5 I think questions about what I am calling "minimal normative weight" are in fact quite important. I am only denying that they have any special relevance to the normative weight of conventions.

6 An additional complication which I'll note but otherwise ignore, is that a person might do the right thing for the wrong (articulate) reason. Long-term monogamous cohabitation, for example, may be important to provide a stable framework for the raising of children but may be practiced because people believe that it is commanded by God - which, we may suppose, it is not. The reasons to which I refer in the text, I shall assume to be genuine, real or weighty reasons.

7 Such "in principle" claims raise a number of sticky issues. To what extent do such claims allow one to abstract from actual cognitive and informational limitations? The claim that something could be in principle articulated might have little relevance to real human beings if the articulation was only achievable by a being who had far more than human intelligence or far more time to work at it than any real person would. Additionally, there is an epistemological issue: If we cannot actually produce the articulated version of the reason, how could we know that it could in principle be produced?