(a) Neurobiology

Among all species, man has by far the largest brain (Anderson, Reference Anderson2000: 1.13). It holds about 10¹¹ cells, with about 10¹⁴ synaptic links (Singer, Reference Singer, Klinke and Silbernagl2003: 745). This allows humans to handle an extremely large amount of information (Singer, Reference Singer and Schmidt1991:103). They do so by parallel processing (Singer, Reference Singer and Schmidt1991: 103). Little of this capacity is preconfigured in the neonate. The genes determine no more than a rough sketch of how the brain can be used (Singer, Reference Singer and Schmidt1991: 100, 104). All the rest has to be figured out by the child in her interaction with her specific environment (Singer, Reference Singer, Klinke and Silbernagl2003: 746; to a remarkable degree, modern neuro-biological thought on the interaction between mind and environment has been predated by von Hayek, Reference von Hayek1952).

Since the brain of the neonate is virtually free of information about the environment, learning is key to preparing the child for life (Singer, Reference Singer2001: 886). It is not done in any organized manner, but by way of association (Hebb, Reference Hebb1949: 127, see also 102). Life thus starts with trial and error (Singer, Reference Singer and Schmidt1991: 100). It allows the brain to engage in self-organization (Singer, Reference Singer and Schmidt1991: 103). Through a continuous stream of experiences, the brain gradually constructs models of the physical and social world (von Hayek, Reference von Hayek1952: chapter 5). The brain in principle works like a statistical machine. It keeps track of how often cells are activated at the same point of time. This synchrony then results in the stabilization of some links (Singer, Reference Singer1999).

(b) Developmental psychology

Developmental psychologists explain how the development of the brain translates itself into behavioural patterns in the child. Many developmentalists see themselves as constructivists, rather than individualists (Chapman, Reference Chapman1988). Some define development as enculturation (Astington and Olson, Reference Astington and Olson1995). Others see development as the progressive enrichment of the child's relations with others (Carpendale and Lewis, Reference Carpendale and Lewis2004). Both are valuable points of view. In contrast, this paper sees development as a problem for the developing child. From such an individualistic position, it is easier to understand the implications for institutional analysis and design.

For developmental psychology, the distinction between assimilation and accommodation is crucial (Wadsworth, Reference Wadsworth1996). Whenever possible, the child tries to make sense of a new element from the environment by assimilating it to categories it already possesses. If an experience does not fit those categories, it may be ignored for the time being (Selman, Reference Selman1984: 77). But if the child pays attention, the experience inconsistent with her earlier understanding may trigger a leap forward in her mental development (Selman, Reference Selman1984: 76). Such leaps are labelled accommodation. They make for the fact that child development occurs in discernible stages, rather than as part of a continuous process (Selman, Reference Selman1984: 71f., 74). There are overlapping, partly even competing ways of conceptualizing these stages (Piaget and Gabain, Reference Piaget and Gabain1932; Mead, Reference Mead1934; Kohlberg, Reference Kohlberg1981). This analysis builds on the typology offered by Robert L. Selman (Reference Selman1984). He takes the development of social cognition as the organizing criterion. This makes his model particularly conducive to understanding the learning of the law.

Selman discerns five stages. Stage 0 is called undifferentiated. The child sees no difference between the physical and the psychological and is egocentric. Stage 1 brings this difference. But the child remains purely subjective. Stage 2 is self-reflective and reciprocal. If the child observes another person, she is able to see the situation from the other person's angle. She assumes that the other person will do the same. At stage 3, the child attributes a personality to her interaction partners, which she assumes to be stable over time and consistent. At final stage 4, the adolescent gains an understanding of the existence of depth psychology, and of social symbols (Selman, Reference Selman1984: 50–55).

Suffice it to illustrate the stages by the unfolding ways of understanding punishment. At the original stage 0, the neonate sees punishment as an automatic reaction. At stage 1, she understands that punishment can serve different purposes. It can be a pedagogical tool, a way of protecting the child from danger, and a means of restoring the moral balance. At stage 2, punishment can also be interpreted as a form of communication between the parents and the child. The child understands that parents can make mistakes. At stage 3, it becomes conceivable that punishment is a way for parents to pursue their own interest in keeping the family under control. At stage 4, the adolescent sees that when the parents punish her, they might be driven by deep psychological forces, rather than conscious reasoning (Selman, Reference Selman1984: 134–146).

(c) Learning psychology

Learning psychologists have demonstrated the power of inference, even in the absence of any conscious cognitive processing (Domjan, Reference Domjan1998). The proverbial Pavlovian dog detects the correlation between food delivery and the sound of the bell. Its organism provisionally interprets it as causation. When it hears the bell, it produces saliva in anticipation of the goodies to come. If this effort has been futile too often, the previous expectation changes. The conditioned response is extinguished (Pavlov and Anrep, Reference Pavlov and Anrep1927; Anderson, Reference Anderson2000: chapter 2). An extension is operant conditioning (Thorndike, Reference Thorndike1898; Skinner, Reference Skinner1938; Anderson, Reference Anderson2000: chapter 3). The organism responds to the experience that some kind of behaviour is consistently followed by positive or negative reinforcement.

In their efforts to understand the world surrounding them, humans can rely on these subconscious learning tools as well. But humans also possess powerful tools for consciously handling information. This capacity enables them to engage in planned experimentation. Moreover, if the environment imposes experience on them, they can use it to learn in a reflective manner (Bandura, Reference Bandura1986: chapter 3). And humans possess language and the capacity for abstract reasoning. This opens up an avenue for explicit instruction. Instruction makes it possible to reduce or even eliminate the inferential component in learning (Anderson, Reference Anderson2000: 338). Experiences others have previously had can thereby be transmitted. The species can accumulate a stock of knowledge and hand it down from generation to generation.

Instruction is, however, not the only tool for transmitting knowledge across individuals. A powerful alternative is learning by observation (von Hayek, Reference von Hayek1982: 18f.; Bandura, Reference Bandura1986). For instance, in a Guatemalan clan, young girls never get any explicit instruction in girls’ work. They simply are given miniature copies of their mother's tools. They learn to use them by observing what their mothers do (Bandura and Walters, Reference Bandura and Walters1963). Outright imitation is an option in observational learning (Miller and Dollard, Reference Miller and Dollard1941; Messick and Brewer, Reference Messick and Brewer1983, see also Heyes, Reference Heyes2001). But the power of this learning tool does not end there. Rather, observational learning is vicarious (Bandura, Reference Bandura1977: 122–128). The individual generates mental representations (Bandura, Reference Bandura1986: 48). By doing this, she learns behavioural patterns, judgmental standards, cognitive competencies, and generic rules for creating new types of behaviour (Bandura, Reference Bandura1986: 49).

Finally, the distinction between declarative and procedural knowledge must be introduced (Anderson et al., Reference Anderson, Bothell, Byrne and Lebiere2004). Both kinds of knowledge are stored in memory, and can be retrieved. Declarative knowledge is additive. The individual, for instance, learns the characteristic features of dog races. Procedural knowledge, however, is integrative. Its object is abilities, not isolated pieces of knowledge. The individual, for instance, learns how to handle the clutch of her car.

(a) Purpose

There are many divergent strands of institutional thinking (Hodgson, Reference Hodgson1988; Powell and DiMaggio, Reference Powell and DiMaggio1991; Rutherford, Reference Rutherford1994; Hall and Taylor, Reference Hall and Taylor1996; DiMaggio, Reference DiMaggio1998; Immergut, Reference Immergut1998; Peters, Reference Peters1999). Some are quite sceptical with respect to purposeful institutional design (DiMaggio and Powell, Reference DiMaggio, Powell, Powell and DiMaggio1991; Hall and Taylor, Reference Hall and Taylor1996: 936,940, but see Rutherford, Reference Rutherford1994: 51–80). But they all agree that institutions serve a purpose.

From the sections on learning theory, a number of task characteristics can be derived. Primary learning is iterative. Later steps build on earlier ones. There must be opportunities for accommodation, not just assimilation. Consequently, in order to be effective, institutions must be present over an extended period of time. Intervention must be tailored to the respective body of knowledge. This requires a differential approach, at least with respect to cohorts, if not with respect to individual, idiosyncratic stages of development. Institutions must take calculated risks by occasionally surprising the addressee with tasks that transcend her present abilities, while hoping to trigger insight, and hence the entry into a new stage of development.

To a large degree, primary learning is associative. This holds more, the earlier the stage of development. Consequently, the socially expected final result cannot be transmitted to the addressee as a ready-made, final product. The addressee must be given the opportunity to reach the outcome on her own. Moreover, for primary learning, observation is more important than instruction. This implies that successful institutions do not teach abstract rules, but present the addressee with graphic situations.

Invasive, medical methods notwithstanding, the central processing within the brain will typically not come under institutional control. But the remaining elements of learning are good targets for institutions. Institutions can manipulate attention, for instance, by pointing an adolescent to situations of great importance for society. Institutions can increase learning motivation. This cannot only be done by changing incentives. Since the adolescent must still find her place in society, helping her orient herself can be at least as powerful. Finally, institutions can have an impact on the storage of what has been learned in memory. The two critical parameters are elaboration and retrieval. Information is stored better the more intensely it has been elaborated. Specifically, the number of associative cues is critical (Anderson, Reference Anderson2000: chapter 6). Institutions can use these by offering graphic, colourful context (Anderson, Reference Anderson2000: 213,70f.) and inviting the addressee to actively process the information (Anderson, Reference Anderson2000: 197–203). Long-term memory is most reliable if the learned object has been retrieved several times at sufficiently large intervals (Anderson, Reference Anderson2000: chapter 7). Institutions do therefore work best if they provide the addressee with several occasions for rehearsal (Anderson, Reference Anderson2000: 186f.).

(b) Concrete institutions

It directly follows from the foregoing that legal interventions will not be able to channel primary learning by themselves. The effect of the law must be either indirect or subsidiary. Actually, it seems to be both.

The impact of the law on the development of an adolescent is most visible if the adolescent herself, or someone from her close environment, comes into direct contact with the legal system. This is possible since the law is not a self-enforcing governance tool. When an administrative agency or a court applies an abstract legal rule to a case, it inevitably interacts with the concrete addressee of the law. This interaction presents legal institutions with an opportunity to translate the normative expectation into language the addressee is likely to understand. It thus is an occasion for bringing the abstract rule to human scale (on the importance of occasional direct contact with legal authority see Tyler, Reference Tyler1990).

A second path is more indirect. Legal rules can serve as points of reference in an uncertain environment. They can help learners gain orientation. They can become catalysts in the learners’ attempts to make sense out of their social environment. Actually, this is how the discussion on ‘expressive law’ can be linked to the topic of this paper (Cooter, Reference Cooter1998; Adler, Reference Adler2000; Anderson and Pildes, Reference Anderson and Pildes2000). Instances of rule application can contribute to this process. Graphic representations of legal rules, as in traffic signs, may help. The diffusion of knowledge of rules via the media and other intermediaries is equally important.

While the law does thus have an influence on primary learning, the key institutional forces that affect it are different: the family (cf. Kreppner, Reference Kreppner and Lampe1997); kindergarten and school; the church; the institutions for professional training like professional schools, colleges (Newcomb, Reference Newcomb1943), and universities, or apprenticeships; institutions for non-professional training like driving schools, dancing schools, military or social service.

(a) Normative proficiency

The child is surrounded by normative expectations from the very beginning of her life. Her parents tell her what to do and what not to do. They reward the child for obedience, and they punish her for infraction. Normativity is thus among the first experiences that the child has. Along with this, the child learns to control her drives, and to do what is expected of her. Normativity is a fast track procedure for endowing the next generation with the experiences of their predecessors. Any child quickly acquires the ability to handle normative expectations. Otherwise, she could not exist in what is her world in the first years of life, i.e. in her family. Consequently, learning the law starts extremely early in life, already during the first two years (Hammer and Keller, Reference Hammer, Keller and Lampe1997).

In order for the child to accept and properly handle the law, she must acquire a sense of justice (Rehbinder, Reference Rehbinder, Gruter and Rehbinder1983: 261–274; Eckensberger and Breit, Reference Eckensberger, Breit and Lampe1997). This presupposes that she adopt a social perspective (Selman, Reference Selman1984: 45, 48 and passim). The child must be able to look at her own behaviour from the perspective of the group, or society (Selman, Reference Selman1984: 34). This ability starts in stage 2. At this stage, the child can perform the mental operation of self-observation. She considers how others will see her own behaviour. At stage 3, the adolescent is able to simultaneously handle the perspectives of actor and observer (Selman, Reference Selman1984: 34).

The law confronts its addressees with normative expectations. But these are not the same kind of expectations as those that a mother has of her child. Developmental psychology demonstrates the steps in transforming the understanding of normativity that are necessary for properly handling legal expectations (cf. Tapp and Kohlberg, Reference Tapp and Kohlberg1971; Rehbinder, Reference Rehbinder, Gruter and Rehbinder1983: 261–274; Eckensberger and Breit, Reference Eckensberger, Breit and Lampe1997: 253–340; Hammer and Keller, Reference Hammer, Keller and Lampe1997: 152–181; Teuchert-Noodt and Schmitz, Reference Teuchert-Noodt, Schmitz and Lampe1997: 134–151). A child cannot understand why the law is, in principle, a legitimate form of governance unless it abstracts from her own drives and desires. This ability improves from development stage to development stage. The understanding of punishment offers a good illustration of this. Already at stage 1, the child learns that punishment can be pedagogical. At stage 2, she apprehends the communicative dimension of punishment. At stage 3, she learns to accept that punishment is occasionally driven by self-interest, rather than the interest in the child. At stage 4, the adolescent even recognizes that her parents sometimes cannot control the drives that make them administer punishment (Selman, Reference Selman1984: 138–146).

All of this can be projected to the understanding of legal intervention. It is typically not a mere act of social revenge; instead, it aims at governance. The law is one path of communication between the citizen and the state. Occasionally, the legal system is cast under the spell of the self-interest of those administering it. And sometimes only systemic reasons can be offered for the state of the law: it can only be explained in reference to legal institutions, which act as independent forces, rather than in reference to people who have intended some concrete outcome.

The law is a fairly complex institution. Form and substance are typically separated, as are rule generation and rule application. Validity and preclusion are artificially drawn lines. Facts that cannot be proven in court do not exist for the law. Valid arguments can be taken from documents that are hundreds of years old, or that stem from entirely different branches of the law, and hence life. The law can be changed at any point of time at the legislator's will. Rule application is potentially always rule evolution. For all of these reasons, morality and legality can become separated (Eckensberger and Breit, Reference Eckensberger, Breit and Lampe1997: 254, 285). Understanding all of this is a precondition for developing generalized trust in the legal system. Such trust is paramount for the proper functioning of the law (Schlicht, Reference Schlicht1998: 26).

Normative proficiency therefore implies a fairly advanced ability for abstraction (cf. Bandura, Reference Bandura1986: 100–102). It is only in the last stages of development that the adolescent acquires this ability. At stage 3, the child is able to conceptualize groups as abstract and differentiated social systems. She also gains an understanding of internal organizational structure and of role differentiation (Selman, Reference Selman1984: 166). Only at stage 4, however, does the adolescent fully grasp pluralism and tolerance. The adolescent now senses the importance of compromise, and the need for formal validity, rather than the enforcement of universal rules (Tapp and Kohlberg, Reference Tapp and Kohlberg1971: 81f.; Selman, Reference Selman1984: 167f.). The adolescent distinguishes between an office and the individual holding it (Selman, Reference Selman1984: 132). Group structure is expected to be pluralistic and formally organized (Selman, Reference Selman1984: 132–134).

It follows from the foregoing that developing normative proficiency is part and parcel of development as such. There is no separate legal proficiency. Consequently, there is no need for separate institutional intervention aimed at directing the child's attention, at giving her a sufficient motivation for learning the law, or at seeing to it that the result of learning is adequately stored in memory. All this is done by the general institutions that affect primary learning.

(b) Normative expectations

Normative proficiency is a necessary, but not a sufficient condition for governing behaviour via the law. If the adolescent has reached the final stage of her development, she knows what to expect from the law, and how to handle its expectations. She thus possesses the necessary procedural knowledge. But on its own, this knowledge is vacuous. It must be filled with declarative knowledge about normative legal expectations. How do the law's addressees learn these expectations, without knowing the wording of statutes, and without having the professional knowledge for interpreting them? How could they, given that modern legal orders contain thousands of rules, and that law students need years before their interpretations are at least reasonable?

In answering this question, we must begin with an adequate specification of the learning object. For the governance effect of law, it is irrelevant whether the addressee knows the wording of the rule. She need not even have an idea of the abstract behavioural requirement underlying the rule. All she must know is what she is expected to do in a concrete situation to which the legal rule applies. The addressees’ knowledge of the law can therefore be tacit, implicit, and practical (Polanyi, Reference Polanyi1958).

Practically speaking, there are two possibilities. In the first case, the legal rule is grosso modo mirrored by a social norm, or by custom (Schlicht, Reference Schlicht1998: 25f.). Such mirror rules will not have the same level of sophistication as the underlying legal rule. They will often only cover standard cases, not exceptional ones. That way, what comes with a good deal of qualifications in doctrine translates into a bright line rule, such as: you may drink (no more than) two glasses of beer if you want to drive: although the legal rule is about the concentration of alcohol in the blood, and hence depends on weight; although the legal threshold is grounded in the weaknesses of the available technology for analysing blood, and will be stricter if this methodology improves; although there are different legal standards if the police stops you, and if an accident has happened. The second option pushes contextualization even further. In this event, the addressee learns no general rule whatsoever. Instead, she uses exemplars for normative orientation (Bandura, Reference Bandura1986: 103; Anderson, Reference Anderson2000: 348). She knows what the law expects her to do in a certain context. She can extend this knowledge to different contexts by way of analogy. This is particularly helpful if the legal rule asks for ‘appropriate’ behaviour, given the circumstances of the case, as negligence standards do.

In principle, any learning mechanism can contribute to the acquisition of this declarative knowledge. Cognitive learning, resulting from open instruction, is one possibility. The mother tells the child that she should not steal. But most normative expectations are learned through inference. All children provisionally overstep what they expect to be their limits. This allows them to learn the exact confines of normative expectations by trial and error. If the child is constantly punished for some kind of behaviour, she will even learn by operant conditioning to avoid it.

The predominant mechanism for learning the normative expectations of the law, however, is observation. This mechanism is precisely tailored to the learning object. By watching how others master a situation, the child gradually learns the underlying models. She for instance first learns that promises are to be kept. She later learns that it is OK to send the merchandise back to the catalogue company if this is done within the time limit specified by the respective legal order. Observational learning also allows the learner to integrate the normative expectations into her own behavioural programmes in precisely the way that is most effective for her. Finally, and most importantly, the possibility of observational learning resolves the second part of the original puzzle. The law's addressees can indeed be governed by the law without having any knowledge of legal texts.