Part I: On the Concept of Legitimacy: Two Questions for Authority
1. The Service Conception of Authority: An Answer to the Question of Content
As we saw in the last chapter, for Enoch's second condition for authoritative commands to be successful an independent reason must be triggered. This amounts to Raz's 'Normal Justification Thesis' (NJT). Let us now look more closely at Raz's account of authority.
Raz starts by helpfully putting forth the following moral question: “how can it ever be that one has a duty to subject one's will and judgement to those of another?” (Raz 2009,
p. 135; cf. Raz 1979 Ch. 1) The centrepiece of Raz’s explanation of how this is possible is what he terms the NJT. His justification reads, with a slight modification, as follows:
the normal way to establish that a person has authority over another person involves showing that the alleged subject is likely better to [conform] with reasons which apply to him (other than the alleged authoritative directives) if he accepts the directives of the alleged authority as authoritatively binding and tries to follow them, rather than by trying to follow the reasons which apply to him directly (Raz 1986, p. 53).*
This is an explanation of 'the ideal exercise of authority' (ibid., p. 42), i.e., Raz's explanation of when authority is legitimate. Concisely put, authority is justified if and only if it enables subjects to act in conformity with what is required by background reasons better than the subjects could on their own. These are background reasons because, as NJT suggests, they are not followed directly; rather, they are pre-empted by the authoritative command which is to be followed. Raz calls his conception the ‘service conception of authority’, being that its main justification is that it serves to benefit those subject to it: “It is to help them act on reasons which [already] bind them” (ibid., p. 56).
The NJT is one of a duo of mutually reinforcing theses (ibid., p. 55), the second part being what Raz calls the ‘dependence thesis’:
all authoritative directives should be based on reasons which already independently apply to the subjects of the directives and are relevant to their action in the circumstances covered by the directive. (ibid., p. 47)
We can see how close this is to the NJT. The 'dependence thesis' states what kind of reasons the justification of authority should be based on, viz., the subject's background reasons. A
* I have, following Scott Hershovitz, as well as Raz's more recent formulation, replaced the original “comply” with “conform” which is more accurate to Raz's original intention (Hershovitz 2003, pp. 206- 207). Raz recently reformulated the thesis in “The Problem of Authority”, it now reads: “that the subject would better conform to reasons that apply to him any way (that is, to reasons other than the directives of the authority) if he intends to be guided by the authority’s directives than if he does not.” (Raz 2009, p 137-138). This is important because Raz makes a technical distinction between “complying” and “conforming” to reasons. Hersovitz nicely sums up the distinction in the following way: a person
conforms with a reason if she does what the reason requires, whereas a person complies with a reason if she acts for it (Herschovitz 2003, p. 202; Raz 1999, pp. 178-179).
command is to be justified by reasons for action that exist independently of the command. Hence, the NJT is an evaluative judgement on how well an authority accomplishes this task. If someone claiming authority successfully gives commands based on background reasons, then, according to NJT, they are a legitimate authority.
In a recent restatement of the service conception of authority, Raz adds a further condition, something he calls 'the independence condition' to his account, reinforcing his overall point. It reads as follows:
that the matters regarding which the first condition [NJT] is met are such that with respect to them it is better to conform to reason than to decide for oneself, unaided by authority. (Raz 2009, p. 137)
This added condition reinforces an important part of the NJT, one intended to avoid a problem which Enoch's account is liable to encounter. Enoch's account would make it impossible for authorities to make mistakes; authoritative commands would only count if they match the reasons that already exist. This is a high price to pay and it is doubtful that any authority could ever live up to such a high standard. It conflates legitimate authority with a completely just authority so the concept of legitimacy becomes redundant. Surely what we are after when it comes to legitimacy is something less than justice; a standard or criterion which authorities should meet even if they are unable to be fully just. It would leave no room for authorities to make mistakes. Raz avoids this complication (and Enoch could adopt a similar strategy) by qualifying his account with 'likely better to conform'; this makes the justification turn on a counter-factual rather than the necessary connection that Enoch seems to favour.
We must be careful with what Raz means by claiming that authorities have to serve their subjects. On Raz’s account, the reasons for action are not synonymous with individual interests. His claim that authoritative commands should reflect, and help subjects conform
to reasons they already possess should be clearly distinguished from the similar sounding claim that authoritative commands should advance the interests of their subjects (Raz 1986, p. 48). There are many instances within Raz’s account in which an authoritative command helps its subjects conform to reasons they already possess but are also against a subject's individual interests. Raz's example is that of a military commander giving orders against the individual interest of her soldiers. The command may still help the soldiers conform to reasons, viz., the defence of their country, even though it goes against their personal interest and so their survival (ibid.).
One thing that is troubling about NJT and Enoch's success condition is that they rely on the existence of separate, distinct reasons that the subject already has and that the command must match up with. I do not think the right way to account for the triggering aspect is by enumerating conditional duties. That is, Enoch thinks that A can only give authoritative commands if and only if B already has a specific conditional reason to conform to what A commands. In the case of the parent telling a child to go to bed, Enoch says that the child perhaps had the reason all along 'to-go-to-bed-if-you-tell-him-to': “a reason that you successfully triggered precisely by telling him to go to his room” (Enoch 2012, p. 11). I therefore have trouble making sense of this example by Enoch.
This particular example does, for the most part, seem to be a better described case in terms of epistemic authority, insofar as there does not seem to be a reason created. Remember that the difference between epistemic authority and practical authority turns largely on whether a new reason is created. Is it really plausible that, in this case, a new reason is created for the child to go to bed? Is this really the best way to think about the relationship between a parent and a child? I would suggest not. Rather, the child does not (yet) have capacity to properly recognise the reasons that apply. What the ideal parent, I
presume, would do in these cases would be to attempt to explain to the child why she should go to bed (cf. Edmundson 2013, p. 53). Part of the reason for thinking that this is the case is that a parent is not always interested in enforcing compliance. Rather, they are interested in being a model for the child; a source of advice and counsel.
Even in bracketing this worry about epistemic and practical authority, there is still a problem with any account that attempts to justify commands based on specific reasons that an agent already possesses. For a determinate answer to the question of content, such an account relies on a consensus about what reasons exist. Granted, Raz and Enoch will object claiming that consensus is irrelevant; for them it is a matter of fact whether there are these reasons to be triggered or not. However, this seems to have little practical relevance for concerns about the actual legitimacy of authorities because of the deep but reasonable disagreement amongst people about what these reasons are. Indeed, it is not clear how we are to determine which reasons actually exist without a full blown account of normative reality, one that will be highly controversial.
What I will argue for in the following chapters is opposed to Raz and Enoch. What we should be looking for is general reason, rather than specific reasons; that which authorities trigger to justify the content of their command. For one, this is a more parsimonious answer and will allow us to create a critical standpoint internal to a particular authority to judge the content of a command. This internal standpoint will give us greater determinacy in the question of content and our judgements about the legitimacy of particular authorities. However, we will have to wait for this positive story.
Raz's and Enoch's point is that this ability of authorities to create reasons needs justifying. It is not enough for the authority to have the intention to give reason; we must have a clear criterion as to the normative success of reason giving. We want to have a
criterion in regard to which reasons are acceptable for authorities to create and to determine those which reasons are not acceptable. That, in barest outline, is Raz's conception of authority. Let us now turn to what I take to be its shortcoming, viz., Raz answering the question of standing.