Legal Institutional Knowledge
In this chapter I describe the different types of institutional knowledge that exist in law, how they interact with each other and how they may be seen to be founded on different features of the legal institutional context. Following Tsoukas (1996, 1998b, Tsoukas and Vladimirou 2001), I will argue that while it is true that the propositional structure of legal knowledge is fully realised within formal legal contexts this tells us only part of the story. As well as being institutions, formal legal contexts are also practices; that is, shared traditions in and out of which legal practitioners live and work. In this latter sense, legal knowledge has a narrative structure, maintained by story, anecdote and example. However, these two features of legal institutional knowledge sit uncomfortably alongside each other: there is an uneasy tension between the propositional form of legal knowledge fundamentally associated with law as an institution and the narrative form of legal knowledge associated with law as a shared tradition or practice. In order to survive as a practice, law requires its institutions to be strong but these same institutions, by their very nature, as they strengthen and become more autonomous, begin to act as a corrosive influence on the shared tradition; nonetheless, without its foundation in a shared tradition law as an institution is weak and unproductive and incapable of functioning. Thus, some equilibrium must be achieved, its tension negotiated and maintained.
Here are two examples of propositional statements in statute:
If a person–
(a) drives or attempts to drive a motor vehicle on a road or other public place, or
(b) is in charge of a motor vehicle on a road or other public place,
after consuming so much alcohol that the proportion of it in his breath, blood or urine exceeds the prescribed limit he is guilty of an offence.
(Road Traffic Act 1988, s.5(1))
(a) has or has recently had in his possession any tool or other object from the possession of which it May reasonably be inferred that he intended to commit theft or had committed theft; and
(b) is unable to demonstrate satisfactorily that his possession of such tool or other object is or was not for the purposes of committing theft, shall be guilty of an offence and liable, on summary conviction, to a fine not exceeding …
(Civic Government (Scotland) Act 1982, s.58(1))
At common law, the following statement by Lord Sutherland exhibits a similar structure and has been taken, for the purposes of decision-making in Scots Criminal Law, to provide a working definition of ‘wicked recklessness’:
If you act in such a way as to show that you don’t really care whether the person you are attacking lives or dies, then you can constitute this degree of wicked recklessness which is required to constitute murder. It may, in the end of the day come as a considerable surprise to you, and indeed a matter of regret too that your victim dies, but that doesn’t alter the fact that you have committed murder.
(Lord Sutherland in H. M. Advocate v. Hartley 1989 SLT 135 at 136)
In each of the above examples, the preceding conditional statements operate to identify as significant recurring events or behaviours that serve to provide a basis for the formulation of rules to guide future adjudication. Such recurring events are assumed to be patterned, ordered and non-random, made up of elements that are objectively available and which can be re-presented in an abbreviated form. That is, the elements are seen to be ordered according to a pattern that can be replaced with a rule to capture its information content, doing away with the need to list repeatedly the whole contents of that pattern. In this way, the mass of observed events and the statements made about them can be compressed into a small number of propositional statements with the same informational content, permitting economy of effort, transferability, and remote control.
However, for social reality to permit its abbreviated representation in this way, and for propositional knowledge to be possible, the world must first be capable of being understood in such regular, patterned and non-random terms. In what sense might we affirm this to be the case? According to Peter Berger and Thomas Luckmann (1996):
[a]ll human activity is subject to habitualization. Any action that is repeated frequently becomes cast into a pattern, which can then be reproduced with an economy of effort and which, ipso facto, is apprehended by its performer as that pattern. Habitualization further implies that the action in question may be performed again in the future in the same manner and with the same economical effort.
(Berger and Luckmann 1996: 70)
For Berger and Luckmann, institutionalisation provides the context for linking habitualisation and typification: it occurs ‘whenever there is a reciprocal typification of habitualized action by types of actors’. Within such contexts, intentions and purposes are assigned to actors and, when certain actions recur, these intentions and purposes are also held as recurrent. Reducible to role and rule, behaviour becomes to a large extent routine and predictable and ‘[t]he institution posits that actions of type x will be performed by actors of type x’ (Berger and Luckmann 1996: 72). In this way, the social world is seen as submitting to an ordering and regularity that makes it possible for us to arrest from it patterns and routines and to represent these formally, in an abbreviated way.
Clearly, the more that human social life becomes institutionalised, the more concentrations occur, then so the more accessible to regular pattern and ordering it becomes and the easier it is to represent this in an abbreviated form as propositional statements. In this way, rules become a means for the prescriptive ordering of human behaviour in specified circumstances. As Twining and Miers put it, a rule ‘prescribes that in circumstances X, behaviour of type Y ought not to be, or may be, indulged in by persons of class Z’ (Twining and Miers 1991: 131). Rules, as generalisations, connect types of behaviour by types of actors to types of situations (Schauer 1991). To affirm that a rule exists is to generalise, to institutionalise behaviour is to affirm the existence of rules (MacCormick 2007). Between rules and propositional statements there exists a sort of mutually dependent relation.
Legal Rules and Facts
What, then, is the relation between general categories and the particular instances they seek to relate? Obviously, any particular object, action or event is subsumable under a whole range of separate, though not mutually exclusive, categories; for example, I am white, married, middle-aged, bespectacled, driver, ex-army officer, dog lover, keen gardener, and so on. However, not all of these or my other attributes are always necessary in order to offer a full and relevant description of me in every situation. Only a very limited set of descriptors will often be all that is required and my choice of action in any given situation will not depend so much on any of these generalisations as on the type of situation I find myself in or the discursive context in which I am described. Indeed, we might go even further and say that I am not always the same person in all of my different situations in life: there is not one over-aching community to which I belong and by which I am adequately or completely described but I belong to a number of different communities and I am a different person in each one of them. Within these situations, I as I am in these situations make my choices, and my possibilities for future choosing are shaped by and depend on the possibilities that I choose to actualise, make concrete, in each present moment of choosing.
This is important. Through my ability in any given situation to generalise in one direction, to choose A and thus to actualise its possibilities, I not only accept the consequences for my future choosing that are given by my choosing A but also, by default, I choose not to actualise other possibilities. Therefore, in this sense, we can say that particular situations or discursive contexts make institutional action possible, for saying that I am a white married male is quite different from saying that I am a bespectacled driver, and the presence of these different particulars will assume greater or lesser significance depending on the context chosen. In some contexts, the fact that I share particular characteristics with other persons will have significant consequences; in a different context, and depending upon the context, they might assume greater, lesser or no significance at all.
Put differently, generalisations as category descriptors are necessarily selective: inclusive as well as exclusive, suppressing as well as revealing. But, what determines which generalisation will constitute a given rule’s factual predicate is its purpose: the goal prescribed (or permitted) or the evil proscribed by it. What is significant about institutional rules is that while their consequents are forward-looking (meant to be applied to future instances) their factual predicates are backward-looking (in the sense of having been derived from regularities viewed retrospectively) or forward-looking in the limited sense of being based on current assumptions about future behaviour. But there is a difficulty here. While propositional knowledge can provide an explanation retrospectively as to why a social system functions as it does, it cannot prospectively inform actors as to how to apply any given set of rules or how to create new ones.
The reasons for this are famously stated by Hart and developed by others. First, there is the inherent instability of language and representations of meaning. Any illusion of stability is only temporary, and new definitions and new symbolic representations are forever emerging to overtake, overshadow or erode old established ones. Thus, while on the one hand a social system such as law tries to fix its definitions and representations with regard to its purposes, inevitably, at some point, definitional control passes over to the context in which it is set; if we affirm the inherent and ultimate instability of knowledge representations in institutional contexts then we must also acknowledge the same in respect of its functional rules. Second, if it is to be suggested that a rule must always be followed in the same way repeatedly in the future, then what determines this cannot itself be a rule (Taylor 1985: 57). Rules, as guides for social action in open social systems, are fundamentally imperfect tools: the definition of a rule cannot itself determine how, on every occasion, it is to be applied, and there is no point in pursuing the argument ad infinitum by formulating ever more new rules to determine the use of the first rule, so we can only conclude that the application of rules cannot itself be determined through a rule; that is, it must be rule-less.
To put this more concretely, suppose that law books containing codified rules were issued to all judges. The abstract representations of actual situations imagined behind these codified rules will be only weakly related to the actual situations that later confront the individual judges. The application of rules falls to take place in social contexts the details of which cannot possibly be known in advance and fully to the rule codifiers. Moreover, just because some generalisations are selected does not mean that those that have been suppressed are irrelevant. This will depend on the circumstances within the context. In certain combinations of circumstances they may become central, but the point is that there is no way of knowing beforehand what particular combination or type of combination of circumstances will make a certain feature salient. Only the decision-maker faced with making the decision in respect of the actual circumstances of the case before them will be able to make that decision and make judgement accordingly. So not only is what is going on in an institutional context not static, and indeterminate, but the rules governing situations are bound to be to some extent of limited utility: all sorts of things are going on at the same time that cannot be described in advance, and can only be known at the time from the particular perspective of the observer as an involved participant. There is no escaping the difficul-ties that arise with new circumstances and even the most informed and imaginative codified system will come up against this problem in the end.
MacCormick argues that if we can identify regularities of behavioural patterns then it should be possible to state these in the form of conditional, ‘If, Then’ propositional statements, which will be valid under certain stated conditions: ‘Whenever OF [operative facts], then NC [normative consequence]’ (MacCormick 2007: 24–25). This accords well with the idea that propositional knowledge is necessarily concerned with generalisations, connecting types