It is a commonplace that many authorities are established by convention, but there is no standard way to flesh out this everyday understanding. Here I offer an analysis of conventional authority such that a command is genuinely authoritative when conforming to it leads to a particular kind of convention. This is different from the more common suggestion that it is a convention that somebody has authority.
On my account conventional authority is a mechanism that addresses what I call strategic underdetermination problem (SUP) cases. This is where individuals share a body of principles, but don’t know what would be the best way to follow those principles in a particular situation because the principles underdetermine what they should do. In SUP cases individuals need to be able to predict what their fellows will do in order to be able to determine what they themselves should do. In response to SUP cases we need what I call limited conventions which are Lewisian conventions but where we evaluate the different options not based on how they match individual preferences but rather to the extent to which they match the pre-existing principles. This means that the options that individuals may coordinate on are limited to those that are at least minimally compliant with their shared principles. On this account, a command with conventional authority gives an exclusionary reason and should be followed because it establishes the shared expectation that everybody will respond to the SUP case in the specified way, and that response that conforms to their body of principles. If those conditions are met, not following the command undermines everybody’s ability to do as their principles require, meaning everybody has an exclusionary and pre-emptive reason to conform to the command.
In support of my analysis I give a conventionalist analysis of parental authority, which is usually not taken to be conventional. Linking the authority of commands to bodies of shared principles allows me also to offer an analysis of how different authorities may overlap, and of how conventional authority can handle moral variation within a population.
The structure of this chapter is as follows. After a preamble in §I, in §II I argue that a command leading to the forming of a limited convention is sufficient for something being genuinely authoritative. In §III I place my account of conventional authority in context within the literature. In §IV I offer my conventionalist analysis of parental authority, and in §V I discuss overlapping authorities and moral variation within a society.
To forestall confusion, I want to make a clarificatory point right away. Here I offer sufficient conditions for some command to be an instance of justified authority, but not necessary conditions. My account is compatible with there being multiple sources for authority, and conventional authority being only one source amongst many. It would be excessively tedious to continuously stress that I’m offering only sufficient and not necessary conditions for genuine authority, so I ask the reader to keep this feature in mind.
I also want to highlight one unusual feature of my account: I focus on authority command-by-command, rather than individual-by-individual. Most theories of authority try to explain why some particular individual or role can be vested with authority. My account however focusses on what specific commands may carry authority. I propose that for conventional authority at least the commands are primary, and some individual becomes vested with authority derivative on the commands that they typically issue. How we can go from specific commands being genuinely authoritative to individuals becoming vested with authority is the topic of §II.i.1
I must also address here a possible mismatch between what commands require and conventions deliver. Conventions are, at least on Lewis’s analysis, about regularities in action across recurring situations. They concern multiple instances of an action repeated over a period of time. But commands very often are one-off events: someone delivers the command, somebody else responds, and that’s the end of the matter. Conventional authority then seems to miss a very large part of what we expect from commands and conformity: that we should allow for commands to be authoritative even in one-off cases.
The most important point in response is that the ability of conventions to offer guidance is unaffected by whether they get repeated or not. Since guidance is what is at issue in my analysis, I can cheerfully allow that the convention needn’t stretch over multiple occasions. On this point my purposes diverge from Lewis’. Lewis is after a way of establishing co-operation without requiring explicit agreement or pronouncements. For him, the fact that he will only call repeated cases conventional is a stipulation to make the central features of his analysis clear.2 But of course in the case of someone being subject to commands there is something that can guide the parties: the commands in question. My claim is that these commands play the same epistemic role as a pre-existing structure of expectations do in Lewis’s analysis, by giving individuals a way to tell how to reliably navigate through SUP cases. So, the kind of worries Lewis had about how conventions could be established and maintained without an agreement or pronouncement simply doesn’t apply in this case.
In any case, while the extension of a once-off command over a range of repeated situations isn’t explicit, it is possible. It is commonplace that one-off commands end up shaping behaviour in future cases as well, because these one-off commands create a precedent to be followed in future cases.3 The existence of ‘standing orders’ in military contexts, meant to regulate behaviour over repeated instances by way of the same mechanism as one-off orders, is an indication of how this would go. We can turn any one-off command into a standing order by way of adding ‘this command also applies to future instances of this situation’. So, for my purposes there is no deep difference between a one-off command and one that applies to a recurring situation.
A criterion for authoritative commands
I propose what I call the commands-as-conventions criterion:
A command carries genuine authority if a limited convention would result if the people subject to it conform to the command.
I will refer to the above as simply ‘the criterion’. It amounts to the claim that commands with genuine authority are a supplement to the moral principles that generally guide the community in question. That is, there is some set of principles that the community subscribes to, which has various consequences about how people in that community should behave; the commands of a legitimate authority is a way to give determinate moral guidance which is not entailed directly by the shared principles, but which is consistent with it, and where it matters that the members of the community are in agreement about how to act.4
I will briefly reiterate the relevant background for limited conventions that I presented in Chapter 1. Limited conventions are instances of Lewisian conventions that arise in response to what I call strategic underdetermination problem (SUP) cases. The underdetermination in question is where the principles shared by a community are such that there are cases where they only go part of the way towards determining how to respond to the situation. In particular, there are multiple options that remain after the principles dismiss various options as unsuitable, but the principles give no way to choose between them. I call the remaining options benign outcomes, because they are ones that are not malignant (that is, not determinately worse than another available outcome). The strategic part of the problem is that if you’re in an SUP case, to reliably reach a benign outcome you need to be able to predict what your fellows will do. But since the case is underdetermined, you can’t use the principles to tell what even a perfectly conscientious person will do. So, the uncertainty individuals face in their own decisions then bleeds over to the decisions of other people, making them uncertain as well.5
The way a command becomes authoritative in an SUP case is when it picks out one benign outcome from the underdetermined range and makes it common knowledge that the chosen option is the one to be taken. This removes the uncertainty of SUP cases, which in turn gives the parties reason to follow it. In particular, the subjects have good reasons to follow the commands and the resulting conventions, because they are party to the mutual benefit that comes from conforming, and they would risk everybody’s ability to reach that benefit by failing to conform. They also have no good reason not to conform, since the other outcomes they could reach are not determinately better than the one selected by the convention.6
The above is not to deny that the different available options will have different distributions of burdens and benefits. My response is that the community’s ability to reach any benign outcome at all, and so discharge the demands of their principles, is dependent on their ability to coordinate their efforts. If some attempt to reach one option with its distribution of burdens and benefits while others try to reach a different one, they are going to work at cross purposes. The members of the community need a way to have the necessary confidence in what the other parties will do, so that they can embark on one option rather than another. The commands of some recognized authority will suffice in forming the shared expectation that those subject to it will all act in the way that counts as conforming to the command, and that way of acting also suffices to coordinate towards one of the benign outcomes. This is what I mean when I say that authoritative commands can create conventions that serve as extensions of the shared principles that guide a community.7
the criterion that I have provided evaluates instances of authority command-by-command. But typically we take authority to be a property of individuals, and often a property the individual has in virtue of some station that they occupy. For example, the paradigmatic instances of authority are those of a judge presiding over a court case, or a teacher leading a classroom. We don’t as a rule try to offer a justification individually for each of the commands issued by these authorities. I don’t deny that this is the usual form in which we find authority. To match up my criterion with this observation, I introduce the standard of a benign arbiter.
To count as a benign arbiter, an individual has to fulfil two requirements. Firstly, the salience requirement: there must be a domain where it is common knowledge that the individual’s commands are meant to be respected. This means that there is both a range of issues and a group of people who are subject to the authority on those issues. Whether this is by some personal quality or because they occupy some station to which authority attaches is immaterial. Secondly, the judgement requirement: the individual’s commands as a matter of fact always select a benign outcome. This means that in cases where the principles of the community in question determine a sole outcome to work towards, the command guides the community to do so, and, in an SUP case, the command picks out one of the range of benign outcomes.
I. Introducing the strategic underdetermination problem
II. Introducing limited conventions
III. What is covered in this thesis
IV. Who is the audience for this thesis?
1. Limited Conventions about Morals
I. Lewisian and Limited Conventions
II. Examples of limited conventions
III. The Normative Force of Limited Conventions
2. The Normativity of Limited Conventions
I. Switching the focus from conventions to regularities
II. Regularities can do more than one thing at once
III. Practice-dependent moral norms
3. Conventional Authority
II. A criterion for authoritative commands
III. How conventions provide pre-emptive reasons to conform
IV. Parental authority as conventional
V. Dealing with moral variation within societiesVI. Conclusion
4. The Virtues in Word and Deed
I. Principles, stability, and variation in the v-types
II. Virtues and conventions
III. The conventional fixing thesis
IV. The paired profile thesis
V. The functional definition thesis
VI. Tying the pieces together
5. Knowing and Unknowing Rightness
I. What we learn from action-guidance
II. First-order success despite higher-order ignorance
III. Why mere conformity is second-best
IV. Why mere conformity doesn’t undermine morality
6. Social Action without Social Attitudes
I. Regularities are neutral as to their origin
II. Rational Alienation
III. How does rational alienation persist?
IV. How widespread could rational alienation be?
V. Opaque influences on the rationally alienated
GET THE COMPLETE PROJECT
Limited Conventions about Morals