Published in European Journal of Law, Philosophy and Computer Science 5 (1995), 41-46

(Note: This is a manuscript version. Please quote only the published version.)

NORMS AND AGREEMENT

by
Raimo Tuomela and Maj Bonnevier-Tuomela
University of Helsinki



I TWO KINDS OF SOCIAL NORMS

The main aim of this is to consider the normative strength of different kinds of social norms as far as philosophical and conceptual considerations go. We will start by summarizing the account of social norms given in Tuomela and Bonnevier-Tuomela (1992) and Tuomela (1995). This theory of social norms will be taken for granted below.
By social norms we here mean community norms, and they can be of two kinds (see Tuomela, 1995, Chapter 1, for a proper discussion). A norm in the sense discussed in this paper can be taken typically to have the form "An agent of the kind F in group G ought to perform task T in situation C". Instead of 'ought to' we sometimes have 'may' (or 'is permitted to') such as in "One may sell one's property whenever one wants". Social norms in the sense meant can be divided into rules (or r-norms ) and proper social norms (s-norms ). Rules are created by an authority or body of agents authorized to represent the group (this body can also be the entire group). While r-norms can be taken to be based on (explicit or implicit) agreement-making, s-norms are based on mere mutual belief. (As to the notions of agreement and mutual belief we assume a standard common-sense view of them - see Tuomela, 1995, Chapters 1 and 2.)
Rules are either formal or informal . Formal rules are articulated and written norms with formal (viz. legal-like) sanctions. They have been brought into existence by a proper authority (e.g., group members with authority - often all members form such an authoritative body). Laws and regulations (except perhaps "secondary" ones) are examples of formal rules. Informal rules are articulated norms with informal sanctions (see below), which are also brought into existence by an authority or an authoritative body of group-members (viz. by group-members' mutual, authority-involving explicit or implicit agreement), but articulated informal rules generally are not in written form, but only presented orally (if by the use of natural language at all).
Proper social norms can roughly be divided into conventions or conventional norms (which are general in the sense of concerning a whole society, a community, or social class) and group-specific norms (which are specific to a group or some - but not all - groups in a society). Social norms of obligation - be they r-norms or s-norms - are typically of the form "One ought to do X in C"; and of course it is required that for a prescription of this kind to be a social norm in force, it must be accepted (to a large extent) in the group. In the case of r-norms the acceptance is based on agreement, in the case of s-norms on mere mutual belief. Conventions and group-specific norms are typically unarticulated norms with social sanctions (approval, disapproval). However, they are generally articulable and verbalizable "upon request" and are often presented in verbal form to children and strangers.
In light of the preceding discussion, we will now give a more detailed analysis of obligating social rules (r-norms) and proper social norms (s-norms). The "oughts" involved in norms that we will consider express, in the first place, what the group prescribes (or proscribes) or permits concerning the persons to whom the norms apply. These prescriptions and permissions in turn can have various contentual grounds or sources, e.g., legal, moral, or prudential.
Our proposal for a noncontingently true analysis for social ought-to-do rule-norms, summarizing the central aspects, is as follows, G being a social group:

(RN ) A norm N of the form "Everyone in G ought to perform task T when in situation C" is a social ought-to-do rule (or ought-to-do r-norm ) motivationally in force in G if and only if
a) i) N (or at least a prescription whose logical or conceptual consequence N is) has been properly issued by an authority (either an authority external to the group G or by an authoritative body of members of G), and ii) the members of G can acquire the belief that they ought to perform T in situation C from linguistic (typically verbal) information made available by the aforementioned authority to the members of G;
b) i) many members of G perform T in C (or at least are so disposed), and ii) at least some of them sometimes perform T because of their believing, in part due to the factors specified in clauses a) and c), that they ought to perform T in C;
c) there is in G some pressure - mutually believed by at least those members referred to in clause b)ii) - which is at least in part due to rule-sanctions, against deviating from performing T in C (and this may in part account for both b)i) and b)ii)).

Our analysis of proper social ought-to-do norms (conventions and group-specific norms) is this, G being a social collective but not necessarily a social group:

(SN ) A norm N of the form "Everyone in G ought to perform task T when in situation C" is a proper social ought-to-do norm (or an ought-to-do s-norm ) motivationally in force in G if and only if
a) there is a mutual belief in G to the effect that the members of G ought to perform T in situation C;
b) i) many members of G perform T when in C (or at least are so disposed), and ii) at least some of them sometimes perform T at least in part because of their believing that they ought to perform T in C, and are, in accordance with a), expected by other members of G to perform T in C; and there is a mutual belief approximately to the effect that i) and ii) in G;
c) there is in G some pressure, at least in part due to social sanctions, against deviating from performing T in C; and there is a mutual belief to this effect in G (and this mutually recognized pressure may in part account for both b)i) and b)ii)).

In these analyses the conditions can accordingly be named as follows: Clause a)i) of (RN ) can be called the promulgation condition, and its clause a)ii) the accessibility condition; clause a) of (SN ) is the acceptance condition. Clause b)i) in both analyses can be called the pervasiveness condition, and b)ii) the (normative) motivational condition. Clause c) of (RN ) is the sanction condition while c) of (SN ) is the recognized sanction condition. These requirements have been discussed and argued for in Tuomela (1995), and I will here only take up some aspects needed for the purposes of the present paper.
As to the promulgation clause, the point about a norm's being properly issued by an authority relies in part on the exercise of the authority system of the group (in the sense of Tuomela, 1995, Chapter 5.) We will below solely consider the simple but common case of a collective in which the laws and other r-norms are produced by some "operative" members (such as legislators) authorized by the members of the collective to create r-norms for them.
The pervasiveness condition of (RN ) and (SN ) accounts for the supposed fact that people tend to obey norms and the motivational condition for their at least in some cases obeying them in terms of action performed for the right reason, viz., because "that's the social norm in force". This reason-decription basically refers to the the agent's belief that it has has been agreed or is mutually normatively expected among the members - and is appropriately sanctioned - that the norm content is to be obeyed. The pervasiveness condition says that many people in G indeed behave in accordance with the norm: the norm pervades the group in this sense. The motivational requirement that at least some agents on some occasions C obey the norm for the right reason is "built into" the norm. If nobody obeyed N for the reason that it is a social norm, it would not count as a norm in force even if all people somehow accidentally or for some other reason performed T when in C. Norm-obeying in its full-blown sense means acting for the right normative reason. In the case of an r-norm the belief that one ought to perform T in C may be due to the reason that "it's the law" or "that is our agreement" (or is our rule created by our agreement, or something analogous). It may in part be due to r-sanctions. In the case of an s-norm the belief that one ought to perform T in C is, at least in part, grounded in the agent's belief that he is normatively expected (required) by the others to act thus, that there is mutual belief about this, and that s-sanctions will (otherwise) be applied (about which there is mutual belief). In the case of r-norms also s-sanctions may apply for not obeying the r-norm in question, but there are no r-sanctions related to proper social norms.


II DO AGREEMENT-BASED NORMS OVERRIDE OTHER SOCIAL NORMS?

In addition to social norms in force there are other kinds of norms, e.g. idiosyncratic purely personal norms (e.g. "Always do some gymnastic exercises before breakfast"). In between social norms - based on "responsiveness to social factors" - and purely personal norms there are what can be called potential social norms, norms which are normally widely obeyed but which are not in their essence based on "social responsiveness" and which could in principle be personal only. Among such norms there are - as we all know - moral norms (e.g. "One shall not steal in normal circumstances" and prudential ones (e.g. "One ought to maximize one's expected utility" or "One ought to act so as not to seriously endanger one's life" or even, in some collectives, "Free-ride - e.g. cheat - whenever you can get away with it"). We will accordingly speak of r-norms, s-norms, m-norms (moral norms), and p-norms (prudential norms). Respectively, reasons for obeying these norms involve responsiveness to authorities (r-norms: e.g. I obey the r-norm because it is an agreed-upon rule), mutual social expectations (s-norms: e.g. I obey the s-norm because it is normatively expected of me by the fellow members), responsiveness to one's moral view of what is right or - or colloquially put - one's "conscience" (m-norms: e.g. I obey the m-norm because my conscience demands it), or responsiveness to one's reason (p-norm: e.g."X is the rational thing to do").
As a matter of conceptual possibility, both m-norms and p-norms could be merely personal - even the respective conscience- and reason-related sanctions involved could be merely intra-psychological. But they have a claim to universality: typically some kind of desideratum of universalizability is imposed on them making them at least potentially widespread (and in this sense social in the wide sense). These norms can - but need not - also at the same time be social norms in force in our narrow sense. In accordance with this, m-norms usually in our Western societies are social - be they teleological or consequentialist in their nature; and the same goes for p-norms. We can say in general terms that an m-norm becomes social if some group of people in one's collective starts expecting and demanding obeyance of a certain moral norm. Then, if the person in question obeys it not merely on the ground of his conscience but in part also on the ground of social expectations, then the moral norm is on its way to becoming an s-norm. And when sufficiently widely accepted (and in normal circumstances followed and sanctioned) in the collective it indeed becomes an s-norm for the collective. Analogously, some r-norms can be moral: A body of operative members (legislators, rulers) can accept (accept in the agreement-sense) a moral norm for the collective; and as a factual consequence of this the m-norm becomes an r-norm.
In an analogous fashion one can also comment on p-norms: the set of them normally will overlap with both the set of r-norms and the set of s-norms in a collective; and they can become s-norms or r-norms in the same way as m-norms can.
What can now be said about overridingness? We shall here be concerned only with r-norms and s-norms (and the m-norms and p-norms whose contents also are contents of either r-norms or s-norms) Here is a defensible two-sided thesis: a) R-norms override s-norms in a sense of having more "agreement-based" objective social obligating power than the latter, but b) which one or ones of these kinds of social norms wins "empirically" (viz. in terms of actual obedience or, which we also allow here, mere acting in accordance with the content of the norm) is a factual issue not to be decided a priori. This thesis is argued to hold for all norms concerning the same topic. Without attempting to give a full analysis of what this involves we mention two central cases in which there can be rivalry between r-norms and s-norms: i) parallel norms in which both the r-norm and the s-norm have the same content (e.g. "Everyone ought to perform T in C"); ii) conflicting norms, norms with conflicting content (e.g. "Everybody ought to do X in C" versus "Everyone ought to refrain from performing T in C").
As to a), it is true because r-norms are - in the case of the internal authorization we are discussing here - based on the members' transferral of their relevant rights (viz. in the authorization) to the body of operative members having the power to create group-binding r-norms. What is the structure of internal authorization? It suffices here to consider the matter in idealized terms, assuming the setting of a sufficiently small collective in which the members can really be thought to intentionally make agreements and to transfer their rights to others. We want to have the result: Everybody r-ought to do X, and this is mutually or generally known (in the collective). A simple argument for this result can be construed as follows (see Tuomela, 1995, Chapter 6):
i) We decide about our affairs either directly or transfer our relevant rights to our representatives, and this is mutually or generally known.
ii) We are represented by O (the set of operative members), and this is mutually or generally known.
iii) Representation entails an r-obligation for everybody to do what the members of O jointly decide or agree for our group - at least to obey by tacitly accepting its decisions, and this is mutually or generally known.
iv) O decides or agrees something for our group entailing that everybody r-ought to do X - at least ought tacitly accept something, and this is mutually or generally known.
Therefore, Ever generally known.
The operative members have also been given the right to establish a system of sanctioning for the collective, and this system is of the kind that in principle it can be made so strong that it forces people to behave according to r-norms and indeed obey them (rather than e.g. a conflicting s-norm, if it happens to exists). However, s-norms can be strongly sanctioned as well (in the sense of purported production of shame) and what actually happens in our existing social collectives is a different matter which cannot be decided a priori. (It can be noted here that the case of external authorization with r-norms imposed or coerced on the collective is somewhat different but the crucial point also in this case is the members' acceptance in the transferral-of-rights sense of the r-norms in question.)
As to internal authorization, we can also put the basic issue of the tranferral of one's relevant rights to the disposal by the collecgive by saying that the members thereby intentionally "give up their wills" with respect to the common issues at hand. In accordance with clause (iii), the kind of giving up of one's will taking place in the agreement-based authorization is accompanied by prima facie obligation to accept what the operative members decide for the collective (e.g. what laws are passed), including the sanction-enforced carrying out of those decisions. Suppose an r-norm with the content "One ought to do X in C" is decided for a collective and indeed becomes an r-norm in force. In view of our (RN ) and (SN ) it is still compatible with this that there be an s-norm with the content "One ought to refrain from doing X in C" in force at the same time. Now the r-norm is taken to override the s-norm in the sense that the people have indeed intentionally given up their wills in favor of it and thus have, loosely speaking, invested "the rational part of their personhood" to it while their "hearts" may still go with the s-norm. A close analogue is promising to do something which one still finds e.g. morally objectionable or suspect. As emphasized above, which of these norms empirically wins in the sense of evoking more obedience and following is another matter.
As a summary, we have been dealing with the following potentially overlapping sets of norms:
(A) Social norms in force (norms based on "social responsiveness"):
R = the set of r-norms
S = the set of s-norms
(B) Potential social norms (norms potentially social in the sense of wide-spreadedness but which need not be social norms in force):
M = the set of moral norms (m-norms);
P = the set of prudential norms (p-norms), e.g. so-called "technical" norms.
Let us concentrate on ought-to-do norms, for simplicity (there are also ought-to-be-norms and analogous may-norms, and there are "enabling" or "constitutive" norms, but space does not allow a discussion of them). On our account, the central factor serving to individuate norms is the underlying reason for obeying it: r-reason (e.g. "that is the agreed-upon, sanctioned rule"), s-reason (e.g. "that is mutually normatively expected of me at the threat of mutual disapproval"), m-reason (e.g. "that is the right thing to do" or "that is what my conscience tells me"), p-reason (e.g. "that is the most rational thing to do"). As far as our earlier analysis of r-norm and s-norms goes, one and the same norm-content can be the content of both an r-norm in force and an s-norm in force. Analogously, m-reasons and p-reasons can consistently coexist with each other - and with r-reasons and s-reasons. The corresponding norms can analogously coexist - consider the clauses of (RN ) and (SN ).
Going back to the problem of overridingness, we have objective overriding (in our a priori sense of agreement-basedness) of s-norms by r-norms in the case of conflict (viz. cases such as the r-norm having the content "Everyone ought to perfrom T when in C" and a related s-norm having the content "Everybody ought to refrain from performing T when in C"). What about norms with the same content? The thesis of the greater objective obligating power of r-norms over s-norms hold true here as well. Which norm is more often obeyed in actual practice is still basically an empirical matter. What about social prudential norms and social moral norms (be they r-norms or s-norms)? Here the r-norm overrides the s-norm as presented above.
As an empirical hypothesis we claim that a) r-norms generally tend to empirically win over the other norms, still it is true that b) instances of each kind of norm can on some occasions empirically win over all the other kinds of norms. Thus, for instance, there are cases of p-norms (viz. some p-norms and some application situations) such that people obey (or tend to obey) them rather than another kind of norm (especially m-norm or s-norm) in the case of conflict. M-norms and especially s-norms need not be very rational, as we all know (the same applies, however, to norm-followers!). When discussing empirical overridingness, we need not actually require norm-obeyance in the full sense of acting for the right reason (r-reason or s-reason, as the case may be): behaving in accordance with the norm suffices.
Order in society seems to require norms in force, but order may be effected by means of norms of different kinds. S-norms and m-norms of the s-kind are norms for "high-trust societies" (e.g. Japan, Finland) whereas r-norms are norms for low-trust, business-like societies (e.g. USA). There seems to be the kind of compensatory "mechanism" here that the lack of s-norms in force can - to some extent at least - be compensated by the creation of r-norms (to the enjoyment of the law profession). As is well known, due to various factors (e.g. the increase of secularity and social mobility in all directions) old moral norms (especially Christian norms) and values expressed through them are currently losing ground in Western societies. In terms of our conceptual framework, we can say that moral s-norms are losing ground and the question arises whether r-norms can replace or compensate for them or whether it would be better to work toward the re-establishment of important moral norms. This is an especially important problem in multicultural societies such as the U.S. and to substantial extent also in the EU, for instance. What is needed here is a new kind of interaction between the laws and the s-norms so that laws as much as possible reflect the contents of s-norms; when they do not do or cannot do that, s-norms supporting the laws will have to be ingrained in the minds of people. This opens new areas of cooperation between the law and the psychological professions. The task of social policy and of education should be to bring the people within the reach of this new cooperative effort.


References

Tuomela, R. and Bonnevier-Tuomela, M. (1992), Social Norms, Tasks, and Roles, Reports from the Department of Philosophy, University of Helsinki , N:o 1, 46 pp.
Tuomela, R. (1995), The Importance of Us: A Philosophical Study of Basic Social Notions , Stanford University Press, Stanford.pa