Brian J. Loasby
Organisations as Interpretative Systems
DRUID Summer 2000 Conference
15-17 June, Rebild, Denmark
Revised 1 June 2000
It is because of Herbert Simon’s influence that economists have come to associate organisations with bounded rationality. Yet the form of this association is not what Simon (1982) intended: bounded rationality has typically been restricted to organisations, instead of being treated as characteristic of all human behaviour, and the concept of bounded rationality has itself been narrowly conceived. In this paper I wish to explore a wider conception, which (following Langlois) I prefer to call bounded cognition, and to argue that organisations are a response to both the limitations and the potential of human cognition.
Since the theme of this paper is the significance of interpretation, it is appropriate to begin by observing how the interpretation of Simon’s own term has been affected by the increasingly stringent conception of rationality within mainstream economics. From an originally broad notion of purposeful and intelligent behaviour, which Simon has always adhered to, rationality has been formalised as a tautology: action is deduced from a criterion function and an opportunity set, both of which typically have to conform to strict methodological requirements. In effect, Simon’s claim that decision premises are important is conceded; but since his further claim that they are problematic is ignored this concession has no effect: decision premises are simply another label for the fundamental data. Indeed, in rational choice theory decision itself is meaningless - as is neatly demonstrated by the inability to predict or prescribe how agents should choose in situations which yield multiple equilibria. Instead of the process of choosing, we are offered existence proofs.
Simon’s argument that we should move from substantive to procedural rationality has no purchase in an analytical system which has no space for procedures to be followed by agents. Instead, the search for and processing of information is transformed into another optimisation exercise, in which it is of course necessary to specify what information is available in order to set its potential value against the costs of acquiring it. Economists such as Arrow (1974) have noted the logical difficulties associated with the notion of appraising the value of information without knowing its content, or even knowing of its existence; but effective responses to these difficulties have been confined to the problem of providing appropriate incentives to the production of information when this is a public good, and here too the goal is to prove the incentive-compatability of some particular configuration, or set of configurations, which, however, may not be Pareto-optimal.
It is no accident that attention to the implications of incomplete knowledge for incentive structures is the route by which economists have come to associate bounded rationality, in the form of incomplete contracts, with organisation. But even here rational choice reappears; for although agents are unable to write complete contracts to protect themselves against opportunism they realise that the consequences of opportunism, and of the actions necessary to avoid the threat of opportunism, give rise to potential gains from trade, and are able to realise these gains either by a redistribution of property rights which will eliminate any incentive to act opportunistically or by an agreement to
establish a hierarchy in which one party will have the right to determine, within limits which are presumably prescribed in a contract, what the other shall do. It is naturally assumed that in both cases the rights will be exercised according to the principles of rational choice.
We should note that the effect of these exercises is to isolate any explanation of organisation from the core concerns of economists with resource allocation. As Frank Halm ( 1984, p. 64) once remarked, “traditional equilibrium theory does best when the individual has no importance - he is of measure zero”; by extension it works best when the organisation is of measure zero as well. Coase (1991a, p. 51) came to recognise this pervasive concern to protect the analytic core of microeconomics, first in Robbins’ failure to discuss his explanation of the firm, then in the development of a theory of economic organisation which is intended to deduce patterns of resource allocation from market structure and which consequently “tells us nothing about the organization of industry” (Coase  1988, p. 58), and subsequently in the refusal to admit transaction costs into the theory of value, which continues to discuss “consumers without humanity, firms without organization, and even exchange without markets” (Coase 1988, p. 3). Coase ‘s own explanation of the firm, as he has made clear, does not rely on the incentive to avoid opportunism (Coase 1991b, pp. 70-2); the incentive which matters is to reduce the costs of making appropriate arrangements to take advantage of future opportunities that are not yet perceived. As with Simon, the difficulty is a combination of incomplete knowledge and the cognitive costs of dealing with this incompleteness; the link between Coase and Simon is both as obvious and as undeveloped as that between Coase and Hayek. Such undeveloped links are to be found throughout the field of heterodox economics; but to explore them is well beyond the scope of this paper. Making connections, or failing to make connections, however, is not.
The next step is to make another link, to Frank Knight. Knight (1921) was the first to identify the restrictions within which economic theory was becoming formalised, and to place entrepreneurship and the firm outside those restrictions. He did so by distinguishing between risk, which characterised a situation in which the chance of every possible outcome could be established, and in which therefore the optimal choice was no less strictly entailed by the premises than in situations that were risk free, and uncertainty, in which there was no means of determining these chances and therefore no entailment of choice. Whether Knight also envisaged the category of uncertainty which has been emphasised by Shackle, in which there is no way of definitively enumerating the set of possibilities, is not clear; but since this category is very extensive it would be surprising if Knight was not aware of it. This broader category will be assumed in most of what follows, but it is not necessary in order to demonstrate Knight’s conclusion, which was that in cases of uncertainty decisions must always go beyond the evidence, and depend on the individual - who is not, therefore, of measure zero. People “differ in their capacity to form correct judgements as to the future course of events in the environment. This capacity, moreover, is not homogeneous, some persons excelling in one kind of problem situations, others in other kinds, in almost endless variety” (Knight 1921, p. 241).
In his first published article, George Richardson (1953) took up Knight’s distinction and identified effective decision taking under uncertainty with specific performance skills, citing Ryle’s (1949) category of ‘knowledge how’. Risky decisions, by contrast, could be taken by following a codifiable procedure, which is available to all (like a production function) and an example of ‘knowledge that’. Decision theory claims to assimilate uncertainty to risk by prescribing the use of subjective probabilities, without paying much attention to the wide variation of effectiveness in decision-making that is thereby associated with the concept of rationality, and no attention at all to the implications for both industrial organisation and welfare economics. Welfare economics was Richardson’s focus in his first paper, and he concluded that whenever there is significant uncertainty it is likely to matter a great deal who is taking decisions, and consequently that an important property of any economic system is its ability to select decision makers - a property which is implicitly assumed in much economic theory, and occasionally supported by the invocation of selection mechanisms which have never been justified (Vromen 1995). Richardson is here a single thought away from his later explanation of industrial organisation in terms of capabilities, but restricts himself to the welfare implications of diverse decision-making capabilities; when he later came to write about productive capabilities (Richardson 1972) the specific problem of “deciding what to do and how to do it” (Knight 1921, p. 268) received little attention, though he subsequently returned to it (Richardson 1998b).
Knight demonstrated that risk provides no explanation for profit, except as an alternative label for wages, for it gives rise to no activity which cannot be perfectly duplicated by other people. Profit requires uncertainty, which supplies opportunities for entrepreneurship, a function which has no place in an equilibrium of rational optimisers. Entrepreneurs, Knight tells us, specialise in decision-taking under uncertainty, each choosing a range of activities in which they believe they have an advantage, and in addition offering employment to those who seek some protection from uncertainty. The firm as an organisation, therefore, is a response to uncertainty, and, not surprisingly, has no role in a body of analysis from which uncertainty (except as falsely-identified risk) is absent.
Though Coase (1937) did not emulate Knight’s precision, his account of the circumstances in which a firm is likely to be created excluded the possibility of calculating optimal future contracts; otherwise why should either the creator of the firm or its employees be satisfied (which, be it noted, is Simon’s criterion) with an imperfectly specified contract? Although at least part of the original uncertainty will be resolved in time, it is likely to be replaced by new uncertainties which will justify the continuation, perhaps with modification, of the imperfectly specified contract which constitutes the firm, and experience within the organisation may improve the decision-maker’s capabilities of dealing with particular kinds of uncertainty. Relevant knowledge may grow, and incentive structures may be improved. As Knight (1921, p. 259) noted, “[t]he problem of meeting uncertainty thus passes inevitably into the general problem of management, of economic control”. Such improved capabilities may include procedures which seem to be effective within the environment, internal as well as external, of the particular organisation, even though it may not be clear why these procedures are
effective; indeed, as has been observed, the difficulty of establishing why some ways of running a business seem to work may well provide substantial protection against actual or potential rivals.
It would be a mistake to exaggerate the tendency for organisations to converge on rational choice in the economists’ sense; not only is optimisation often inappropriate for local decisions, especially within an organisational sub-division (which will be considered later) but even when optimisation would be appropriate there may be no adequate basis from which to deduce an optimum. “Much of the error of historians, economists, and of all of us in daily affairs arises from imputing logical reasoning to men who could not or cannot base their actions on reason.” This was the view of Chester Barnard (1938, p. 306), whose work provided a major inspiration to Simon. By no means all reasonable behaviour is strictly governed by the laws of logic; indeed even behaviour which is called logical is deduced from premises which themselves cannot ultimately be validated by logical argument. What Simon calls bounded rationality, and in particular what he calls procedural rationality, includes a good deal more than logical reasoning, though logical reasoning is certainly important in showing what is entailed by particular assumptions or proposed actions. Logic is not sufficient to deal with uncertainty - which is not to claim that it is irrelevant, and organisations are a response to uncertainty. In the broader sense of the term, bounded rationality leads to organisations, and to many other phenomena of economic interest. To develop this argument, it will be necessary to consider the nature of knowledge and the characteristics of human cognition.
Deductive logic carries truth values from premises to conclusions; but if we are to rely on it we must have confidence in those premises. However, the truth of any premises that we use can never be established beyond doubt; there is no certain knowledge, except knowledge of tautologies (which can be useful, and is often not obvious). That general empirical truths cannot be derived by induction was demonstrated by David Hume, and the use of supposed empirical truths as a basis for deductive reasoning must therefore be problematic. Hume’s own response was to abandon the attempt to prove empirical truths, either by observation or argument, and to turn to the potentially manageable question of how people came to accept certain empirical propositions as true.
A notable contribution to this enquiry was made by Hume’s fiend Adam Smith in his earliest surviving writing, which is a psychological theory of the development of science, commonly known as the History of Astronomy. Smith ( 1980) postulates three fundamental human emotions, surprise, wonder (now better described as perplexity) and admiration, which provide incentives to mental activity. Phenomena which do not conform to expectations are disconcerting, and a series of such surprises leads to perplexity, and the inability to make confident decisions. The consequential discomfort stimulates a search for some principles which will connect these inexplicable phenomena in the imagination; and success in creating an interpretative system which connects many phenomena excites admiration among those who have also been perplexed, which is typically reinforced by the aesthetic appeal of the new pattern; this leads to widespread
adoption of this system, which is likely to survive until a new set of surprises stimulates a new search. Smith illustrates his theory (which is itself an example of the process which he describes) by tracing in some detail the succession of cosmological systems up to the time of Newton, whose own theories, Smith insists, should be regarded, not as a description of reality but as the creation of Newton’s imagination, and thus liable to be superseded as a consequence of future surprises.
We should note five features of Smith’s theory. Four of these have already been mentioned: the failure of existing interpretative systems as a stimulus to invention, invention as a new set of connecting principles which, in Popper’s (1963) terms, are falsifiable conjectures and not the product of purely logical reasoning, the readiness to accept other people’s systems as a solution to one’s own personal difficulties of understanding, and the primary motivation in this process as the desire for psychological comfort, including aesthetic pleasure, rather than the prospect of advantage (Smith  1980, pp. 51, 61). The fifth feature is Smith’s recognition that as scientific knowledge develops it becomes more specialised, and that this specialisation leads to the perception of detailed anomalies that would be passed over in a broader perspective, and thus to an acceleration of the growth of knowledge. This principle of the division of labour as the means of developing knowledge was to become the foundation of Smith’s ( 1976b) theory of economic growth, set out in the opening chapters of the Wealth of Nations.
The division of labour leads to the division of knowledge; knowledge grows by division. Why this is necessarily so we shall explain shortly. But first we should note some major consequences. One consequence that Smith was quick to point out is the dependence of every specialist on goods and services that are provided by other specialists; in the form of the co-ordination problem, alternatively expressed as the problem of resource allocation, this has attracted most attention by economists, who have nevertheless often ignored two central features of the problem: that there can be no single pool of knowledge which is accessible to anyone, or indeed to any group, and that the system is continually generating new knowledge and is therefore doubtfully represented by models which assume a given set of data. Hayek is the best-known critic of theorising which omits these features; but we may also note that most of what are perceived as deficiencies in Marshall’s analysis result from his attempts to incorporate them.
However, for our present purposes what matters more than the conventionally-defined problem of resource allocation is the dependence on other people’s knowledge, in production, in consumption, and in decision-making in general. In contrast to the assumption of new growth theory that knowledge, once attained, is reusable without cost, it is a necessary consequence of the means by which that knowledge is generated that the cost to the potential user of using knowledge which is openly available is often very high. Our education system is based on the presumption that the acquisition of knowledge set forth in textbooks takes several years, and requires considerable expert assistance, and no-one nowadays is expected to be capable of acquiring knowledge at first-degree level of half-a-dozen disciplines. There is no general purpose algorithm which can be applied to all fields of knowledge; each has its own connecting principles (which, in economics at least, are clustered into sets which are not entirely mutually compatible) and the
principles which give coherence to one body of knowledge often seem to impede rather than assist the comprehension of knowledge which is differently organised.
There appears to be an evolutionary explanation for all this. Conscious reasoning is a relatively recent development, and the neurological apparatus which makes it possible is necessarily an adaptation from earlier architecture, which had hitherto proved sufficient to preserve the antecedents of our species. Survival and reproduction depended on skill in identifying crucial features of the environment and taking appropriate action, and this required some means of classifying situations and actions and of forming links between them - all before conscious thought; and performance skills, including skills in decision-making under uncertainty, still remain substantially within the domain of tacit knowledge. Domain-specific networks, developed more or less independently, provided a much likelier evolutionary pathway than a general capability for serial processing (Cosmides and Tooby 1994); pattern-making outperforms logic in biological evolution. It is therefore no surprise that conscious thought should have emerged in the form of reflections on the classification systems which appeared to be in use and, in particular, in attempts to improve on classifications which did not seem to be working very well - in Hahn’s ( 1984, p. 59) terms, to devise a better theory and a better policy. As a young Fellow of St John’s College, Cambridge, Marshall (1994) postulated precisely such a sequence. Despite the fascination of economists with rational economic agents, it is still obviously true that human beings are far more skilful at recognising and imposing patterns than they are at constructing logical arguments; and such skills have consequences which are difficult to encompass by standard theory.
Now the architecture of the modern human brain is capable of sustaining any of an extraordinarily large number of alternative connections, each representing a particular structure of connecting principles. What is genetically programmed, apart from sensory perceptions and motor skills, is not primarily, as with other animals, a common set of capabilities but the capability of developing connections in response to particular experiences and actions. This is obviously much more effective in coping with relatively rapid change than the very slow process of genetic mutation and selection. But only a very small proportion of this potential can be realised within any one brain: each person becomes adapted to a particular set of situations and activities. Distinct human communities may therefore acquire distinctive and locally-appropriate shared patterns of behaviour.
However, there is another possibility. By appropriate differentiation among its members, a human community can develop knowledge to an extent that is far beyond the reach of any single member of that community. The division of labour is therefore the most important single idea in accounting for the growth of human knowledge, and also for the problems that this growth causes, in mutual incomprehension and the neglect of what become, because of the particular ways in which labour is divided, relevant interdependencies. To the effects of the division of labour in creating something roughly analogous to distinctive species of capabilities we must add, as Marshall did, the equivalent of Darwin’s principle of variation: even within a single field of knowledge or activity, differences between individuals’ environments, interacting with differences in
their responses to these environments, lead to some differences between their knowledge and skills, providing material both for a continuing selection process and, as Marshall emphasised, for interpersonal learning.
Both the ability and willingness of human beings to learn from each other deserve a smmmary explanation. The ability rests on the shared architecture of the brain which permits differentiation, but which also makes possible what Adam Smith ( 1976a) called ‘sympathy’, the capacity of imagining oneself in the situation of another, and thus making sense of another’s actions as a possible preliminary to choosing one’s own actions. This capacity for understanding is enhanced by the human need to collect phenomena and situations into categories, and to develop regularities of behaviour - what we might call routines - which are relatively easy for others to interpret, rather than tailoring actions to the specifics of each situation in ways which could only be understood through knowledge of these specifics, as would be required by the principle of costless rationality (Heiner 1983). Willingness to learn is explained by consciousness of the inadequacy of one’s own knowledge and skills, including skills in decision-making. There are even striking examples, such as the fashion and education industries, which are based on the willingness of customers to make the adoption of other people’s ideas the principle of action.
Ability and willingness to learn from others often lead to the spread of what is thought to be best practice or received wisdom, which once established may be resistant to change in the face of anomalies experienced by any one person - or even by many people. These are what we often call institutions, and it is important to remember that they emerge as aids to the solution of individual problems, which subsequently prove useful in helping to co-ordinate human interaction (Choi 1993). To begin the study of institutions by focussing on interactions is to neglect these foundations, which often mitigate the difficulties that are apparently revealed by game-theoretic approaches. Whether as aids to individual decisions or to interaction, institutions work by preventing the exploration of excluded possibilities; many possibilities have to be excluded without examination if decisions are to be made, but any means of exclusion caries an unknown opportunity cost. This is true of the exclusions which define the scope of any theoretical system, and this paper is an attempt to indicate, in the spirit of Simon, the opportunity costs which are implicit in the exclusion of cognitive possibilities and limitations from the study of economic systems.
There is a further consequence which should be noted. If we expand Barnard’s (1938, p. 163) proposition that “the decision as to whether an order has authority or not lies with the persons to whom it is addressed, and does not reside in ‘persons of authority’ or those who issue these orders” from its organisational context to all kinds of communication, we shall find that we are remarkably willing to accept the authority of communications from many different sources on many different topics. In fact, this willingness ceases to be remarkable if we reflect that it would be impossible to act intelligently if we were not prepared to take for granted an enormous amount of knowledge, including knowledge of how to decide what to do, which we cannot possibly check for ourselves. Of course we can, and do, check particular items and particular
sources, and we certainly do not accept everything that we are told; but at every point in time we remain heavily dependent on others for the premises on which we act. The quality of these premises is therefore a prime determinant of our performance, and of the performance of economies and societies. That quality depends on the organisation of the growth of knowledge. With the important exception of logical implications, knowledge can never be known to be true; but Ziman’s (1978) principle that the reliability of scientific knowledge rests on the process of generating, criticising and testing within the relevant scientific community can be extended to all kinds of knowledge. If rationality is indeed to be associated with purposeful and intelligent behaviour, it must be a procedural rationality, which is necessarily fallible even in the hard sciences, but is capable of being analysed and appraised.
We have thus arrived at two fundamental principles. First, knowledge itself is created by organisation, both within each individual, through the invention and amendment of connecting principles as explained by Smith, and by the ways in which the field of potential knowledge is divided up between individuals, through the application of Smith’s principle of the division of labour, augmented by Darwin’s principle of variation. This is the cognitive basis on which we all depend; and it must be respected by any system for the generation of knowledge, formal or informal, if that system is actually to improve our understanding and our capabilities. The second principle is that for many purposes this dispersed knowledge must be reassembled, not only in the form of artefacts (which is an extremely important means of assembling and transferring knowledge) but also in the form of organised knowledge communities; and the ways in which it is reassembled constrain what can be done with it. Neither the creation nor the assembly of knowledge can be adequately represented by models that rely on the economic concept of rational choice; but that does not imply that either process is irrational. Our primary focus in this paper is on the latter principle, though without losing sight of the former; and so it is to these knowledge communities that we now turn.
A firm is a response to human cognitive limitations. As Marshall (1920, p. 138) observed, it is one of the forms of organisation that aid the growth and use of knowledge, two processes which are “hard to separate, both in practice and conceptually” (Foss 1999, p. 95). All knowledge depends on organisation, the imposition of simplifying patterns on complexity, and the firm is a focussing device for the organisation and structured development of knowledge and skills within a cognitive framework which is reinforced by the emergence of locally relevant institutions. It is an interpretative system. However, because of the essential incompleteness of our knowledge, the choice of interpretative system for any firm, as for any field of knowledge, is problematic. From Knight’s analysis we can recall that the establishment of an appropriate interpretative system requires the particular skills that Knight identified with entrepreneurship; and Witt (1998) has re-emphasised the importance of this co-ordinating role in both the creation and management of firms, in opposition to the kind of co-ordination which is modelled in so much economic theory.
As in Austrian capital theory, the scope and applicability of knowledge depends on structure and orientation; there is a multitude of possibilities, and consequently room for many opinions. This is not to be regretted; as Richardson (1975, p. 359) observed, “it is of the essence of competition that the participants hold uncertain and divergent beliefs about their chances of success”, and each of these apparent profit opportunities rests on a differential valuation resulting from an idiosyncratic and at least partly tacit way of assembling and interpreting information, as Casson (1982, 1997) has emphasised. The “tendency to variation” (Marshall 1920, p. 355) between interpretative systems, which, as we have seen, is a natural consequence of the combined power and limitations of human cognition, is necessary for continued progress. But within each firm, complementarity and consistency of cognitive frameworks are necessary - though certainly not sufficient - conditions of effective performance and of effective learning. The content of learning is shaped by the framework within which events are interpreted, as well as by the events themselves; and these events are most unlikely to be a random sample from the environment, not least because the firm’s decisions will generate highly selective interactions with that environment. Therefore firms in the same industry will differ in their apparent knowledge and in their capabilities, as Marshall pointed out.
These differences will be consolidated in the process, emphasised by Penrose (1959), through which many aspects of performance, including decision-making, become substantially matters of routine, thus releasing time and mental energy for investigations of novelty. But although the internal managerial limit recedes as expertise in the management of current operations improves, the absorption of newcomers takes longer as the distinctiveness and detail of the organisation’s procedures increases, and so expansion always requires internal resources. There may also be an increasing problem of absorbing skills and ideas which rest on other cognitive structures, though it is also possible that particular kinds of absorptive capacity are among the skills that are developed - a possibility to which we will return.
The interpretative system does not only guide the development of each firm’s knowledge and skills, both productive and managerial. In relation to the conventional economic theory of the firm, Penrose’s (1959) most significant analytical innovation was her distinction between resources and inputs into production, which she called ‘productive services’. In contrast to the firm of price theory, therefore, a change in resources has no necessary direct implications for a firm’s choice either of products or of productive techniques. These choices are conjectures, and require a context, which is provided by the interpretative framework that guides the perception of opportunities. Resources have an option value, and may be consciously developed for this reason, as Penrose (1959, p. 77) emphasises; but the options, or in Penrose’s phrase ‘productive opportunities’, have to be perceived or created within the firm. We are still in Knight’s world of uncertainty, in which presumed knowledge and skills may be applied in new combinations to conjectured business opportunities, where these opportunities are, in Shackle’s (1979, p. 26) phrase, “the imagined, deemed possible”; and both what is imagined and which of these imagined prospects are subsequently deemed possible is conditioned by the enacted environment, internal and external, of the particular business. An integrated set of actions for realising credible prospects must then be devised by an appropriate managerial
process; it cannot be deduced. As has been argued above, if there were known and guaranteed procedures for doing all this, there would be no firms. Of course, any firm may fail in any or all of these respects, and all firms experience failure.
There is an entrepreneurial element in all interpretation, and the chief executive or the board are not the only people who determine the cognitive framework of the enterprise. Penrose, like Schumpeter, distinguished between managerial and entrepreneurial services, and assigned the discovery of new opportunities to the latter; but she explicitly differed from him in arguing that entrepreneurial services were often associated with managerial expertise, in an unconscious (but later acknowledged) echo of Marshall’s recognition that experimentation was both stimulated and guided by differentiated interpretations of differentiated experience. Marshall (1919, p. 270) even declared that “[c]onstructive speculation is inherent in nearly every business decision”. This does not mean that entrepreneurship is irrational, for there is plenty of room for care in working out the logical implications of the perceptions or assumptions on which a strategy is to be built, or in examining the compatability between this strategy and what appears to be reliable knowledge; but it does mean that no entrepreneurial insight can be arrived at by deduction from premises that are known to be true, just as no novel scientific hypothesis can be deduced from experimental data.
Because knowledge, whether true or false, is achieved by imposing patterns, organisational coherence depends on the use of compatible patterns of interpretation. We should not, however, assume that to be compatible these patterns must necessarily, or even desirably, be very similar; as we have seen, distinctive skills within an integrated set of activities result from different ways of organising knowledge. Therefore an effective firm must constitute not just an interpretative system, but (especially if it is a firm of any size and diversity, in skills or markets) a system of interpretative systems. However, what does need to be similar is the orientation of the business, and the perception of the environment to which that orientation is meant to correspond. This may be what Simon (1982, 2, p. 329) had in mind when he argued that it is “more important, in some circumstances, to have agreement on the facts than to be certain that what is agreed upon is really fact”; since he has also declared that reality is very different from what is perceived (Simon 1982, 2, 306) it is apparent that he does not expect economic agents to converge on the correct model. Competing visions between firms are necessary features of an evolutionary or experimental economy. But competing visions within firms, unless very carefully managed, and limited in scope, cause trouble.
Therefore decision premises matter, and the major premises have to be supplied by someone who is necessarily acting entrepreneurially. The ambiguity generated by uncertainty has to be resolved by the construction of beliefs, and by persuading the members of the organisation that these beliefs are credible within the context of the business. We might suspect that in the absence of any definitively correct model there would normally be much conflict of interpretation, and although differences of interpretation between firms contribute to the growth of knowledge, we might also expect that such differences would provide even greater scope for opportunism - within as well as between firms - than is envisaged in transaction cost and property-right models.
However, we must remember the cognitive limitations of every individual, and the consequent willingness to accept help and guidance from others in developing procedures for deciding what to do. Simon has duly noted the readiness of some people to join organisations is in order to benefit from the procedures and decision premises that they provide, and Witt (1998) has argued that entrepreneurial visions may provide sufficient cognitive stimulus to preempt thoughts of opportunism. There is more to the control of opportunism than is to be found in the analysis of incomplete contracts, and more to the analysis of organisation than the control of opportunism.
Within a firm as within a single brain, cognitive processes require structure; and the choice of structure matters. Penrose (1959, p. 149) recognises this with an appropriate definition: “a firm is essentially a pool of resources the utilisation of which is organised in an administrative framework”. The creation of an organisational structure allows the costs of decision-making to be spread over a range of simultaneous activities or a sequence of activities; it allows, as Coase (1937) observed, some decisions to be postponed to a later time, when both knowledge and skills would be improved; and it allows cognitive repertoires, including knowledge both of how to do things and of how to get things done by other people, to be developed within what is believed to be a locally-appropriate framework for each. In specifying the division of both productive and managerial labour, the administrative framework not only influences the ways in which resources are directed to the provision of productive and managerial services, and consequently the possibilities of modifying and enhancing those resources through non-random exposure to events, but also helps to shape the perception (which is necessarily fallible) of new productive opportunities, the exploitation of which in turn helps to shape the future possibilities of learning.
Organisational design is an option of difficulties, because it entails a necessary but fallible abstraction from an integral system, a way of framing problems which preempts attention and suppresses complexity. This practice is made necessary by our cognitive limitations, and made fruitful by our ability to develop new maps of knowledge; but it is also potentially dangerous, because the designed organisation, like all cognitive models, exists in the space of representations rather than in the space of real phenomena. It not only prescribes where the business will work on its problems, but even what will be defined as a problem and what will be acceptable as a solution. (For an explontion of the constraints imposed by acceptability see Loasby 1995). Complexity is suppressed by the dispersion of decision premises among sub-units in order to restrict agendas by excluding consideration of possible interdependencies - which makes optimisation at these levels inappropriate, as was noted earlier; and in those areas of managerial responsibility where interdependencies are explicitly targeted complexity is suppressed by excluding possibly crucial detail. Information routines and management control and incentive systems preempt attention; as Simon long ago observed, programmed activity tends to drive out non-programmed activity, and information systems tend to suppress and distort information - a tendency which is regularly exploited by management consultants.
Barnard (1938) emphasised the need for any formal organisation to be supported by an informal organisation if it is to work effectively; and the formal organisation provides the setting within which the informal organisation, or local institutional arrangements, develop. The pattern of specialisation and the relationships within and between specialised groups generates evidence on which each member of the firm may decide who may reasonably be regarded as an authority on what topics, and whose intentions - and competence (Casson 1997, p. 94) - may be trusted. Ménard (1994) has pointed out that the co-ordination of activities within an organisation cannot be achieved by hierarchical authority alone, but requires a complex flow of communications, based on other people’s interpretations, that their recipients accept as authoritative. These evolving relationships may allow for the codification of some tacit knowledge, which is necessary for the design of appropriate technology and information flows, but, as is less often remarked, it also permits codified knowledge to be expanded into highly-specific tacit knowledge, or heuristics. It is a matter of common experience that the informal organisation of a firm does not always support the formal structure; however, when it does, the tacit knowledge that it embodies, if appropriately oriented, may provide a basis for competitive advantage that cannot be quickly or easily imitated because, unlike the non-specific routines to which Heiner (1983) drew attention, the connections between situation and action cannot be identified, and also because of the interdependence between shared norms and shared cognitive perceptions which is a feature of Smith’s ( 1976a) Theory of Moral Sentiments.
Because the quality of any interpretative system can be judged only in relation to what it is intended to interpret, it is not surprising that changes in the activities of a business, or in the environment in which it operates, are often accompanied by organisational change. New patterns of organisation are likely to be needed either to develop or respond to new patterns of knowledge; that is why Schumpeter (1934) was justified in including organisational innovation among his examples of major ‘new combinations’. The movement from functional to product based organisation is the most famous example; the distinction between mechanistic and organic systems, introduced by Burns and Stalker (1961) is also familiar to specialists in organisational studies. Indeed, since knowledge is itself a matter of organisation, neurologically and cognitively, all innovation may be interpreted as some kind of organisational change.
Successful change may not be easy. There are always costs of switching to a new cognitive system, especially when the switch entails the creation of new linkages with other people’s cognitive systems which are simultaneously being restructured; every well-established organisation has its own familiar institutions, which may have become almost part of its identity. Unlearning can be both cognitively and emotionally difficult. In addition, the shared desire to preserve the organisational coalition, or truce (Nelson and Winter 1982), may reinforce the tendency for organisational members to preserve a degree of managerial independence by respecting the independence of other managers, and to resist any attempt to redefine the boundaries of that independence. The capability, as well as the willingness, to develop new interpretative schemes and new skills may be inadequate, and even the acquisition of ready-codified knowledge requires an understanding of the appropriate codes, which embody, but rarely display, the premises
of other interpretative systems. Companies that have been highly successful may have especial difficulty in introducing innovations which disrupt their existing business. Either managers are distracted from what they have learnt to do well, and the existing business suffers, or the unfamiliar and initially unprofitable requirements of the new business are slighted. A major source of error is neglect of the importance of absorptive capacity, or an overestimate of the absorptive capacity available, because of a failure to appreciate the differences between interpretative systems. The division of labour leads to greater knowledge in total precisely because every person and every organisation does not try to learn or do everything.
Knowledge is organisation imposed on phenomena. Each business is a self-organising system, both formal and informal, for the creation of knowledge through the generation and selection of skills, processes and products; and though it is natural to think of markets as the arena for economic selection we should not forget that there is a great deal of internal selection, not all of which reflects external factors. In businesses which are consciously striving to innovate, this internal selection may be organised by a division of labour between roles which are intended to represent alternative plausible perceptions or what are believed to be major complementary criteria for success, with the intention of using internal selection not only to reject unsatisfactory proposals but also to generate more promising alternatives (Van de Ven and Grazman 1997). However, there can be no assurance that any interpretative system is appropriate, and even less assurance that it will remain appropriate or evolve in an appropriate way; it would therefore be a mistake to commit any industry, or any field of study, to one such system. Because every organisation requires coherence - and perhaps fundamentally because any individual attempting to control an organisation requires coherence within his or her own thinking - no single business is likely to encompass the variety that is appropriate; rivals, and even new entrants, are essential. That is the basic case for competition, in all fields of human activity; it will not be developed here.
Every business can encompass only part of the total configuration of knowledge and skills that is required for its success. It is itself a more-or-less coherent system of interpretative systems, but is also an element in a much wider network of such systems, which includes both alternative and complementary assemblages of skills, orientations and activities; and it must draw appropriately on this network to succeed. The most readily understandable manifestations of such networks are industrial districts (Loasby 1999), but the equivalents of such districts are common in economic systems. Similar businesses generate variety within a shared interpretative framework; complementary activities which rest on substantially different ways of organising knowledge are better assigned to firms which embody correspondingly distinct interpretative systems (Richardson 1972).
Not only may the attempt to incorporate dissimilar though complementary activities within a single organisation tend to suppress the distinctive interpretative systems that are necessary for the continued growth of knowledge, but it may also suppress external contacts which are necessary for such growth. However, Richardson also argued that
arm’s-length market transactions are unlikely to be sufficient to manage such complementarities; and they are likely to be especially inadequate for the transfer and absorption of knowledge which is relevant to product or process improvement. In practice, not many market transactions between firms are at arm’s length; and all markets rely on a set of institutions which simplify transactions and provide a foundation for the growth of understanding between buyers and sellers which provides each party with what Marshall called its “external organization”. How such understanding may lead to technological innovation which cannot be achieved through membership of a group with its own substantial research facilities has been demonstrated by Andersen (1999).
The present interest in alliances perhaps leads to an exaggeration of the novelty of these arrangements; they are certainly much more obvious when they involve large firms. Neil Kay (1997) has argued that, although costly to manage, because of the disparities in interest and incentives between the parties, they are much less costly than full scale mergers, which would tend, not to resolve these disparities but to multiply disparities across many more quite distinct activities; in drawing justified attention to the problems posed by differentiated corporate identities he has perhaps underrated the advantages of preserving these identities, and their interpretative systems, in furthering the growth of knowledge within and between the allied businesses. The division of labour, supplemented by the variation within specialisms, which together allow human communities to make effective use of their cognitive abilities, still offers the most effective means of increasing human knowledge.
Despite the claims of management consultants, there are no guaranteed procedures for producing an optimal design for any organisation. What is being designed is an interpretative system, and the process of design relies on the interpretative system of its designer; by Knight’s definition, it is an entrepreneurial activity, and for entrepreneurial activities optima cannot be defined. We may appropriately conclude with an assessment by Schumpeter (1943, p. 83), which is quite correctly not restricted to economic systems.
A system - any system, economic or other - that at every given point of time fully utilizes its possibilities to the best advantage may yet in the long run be inferior to a system that does so at no given point of time, because the latter’s failure to do so may be a condition for the level or speed of long-rum performance.
Andersen, Maj M. (1999) Trajectory Change through Interorganisational Learning: on the Economic Organisation of the Greening of Industrv, dissertation presented at Copenhagen Business School, 8 June.
Arrow, Kenneth J. (1974) The Limits of Organization, New York: W. W. Norton.
Barnard, Chester I. (1938) The Functions of the Executive, Cambridge, MA: Harvard University Press.
Burns, Tom and G. M. Stalker (1961) The Management of Innovation, London: Tavistock.
Casson, Mark (1982) The Entrepreneur: An Economic Theory, Oxford: Martin Robertson.
Casson, Mark (1997) Information and Organization, Oxford: Clarendon Press.
Choi, Young B. (1993) Paradigms and conventions: Uncertainty, Decision Making and Entrepreneurship, Ann Arbor: University of Michigan Press.
Coase, Ronald H. (1937) ‘The nature of the firm’, Economica, N.S. 4, 386-405.
Coase, Ronald H. (1972) ‘Industrial organization: a proposal for research’, in Victor R. Fuchs (ed.) Policy Issues and Research Opportunities in Industrial Organization, New York: National Bureau of Economic Research, pp. 59-73; reprinted in Coase (1988), pp. 57-74. Page reference is to this reprint.
Coase, Ronald H. (1988) The Firm, the Market and the Law, Chicago: University of Chicago Press.
Coase, Ronald H. (1991a) ‘The nature of the firm: origin’, in Oliver E. Williamson and Sidney G. Winter (eds) The Nature of the Firm: Origins, Evolution and Development, Oxford: Oxford University Press, pp. 34-47.
Coase, Ronald H. (1991b) ‘The nature of the firm: influence’, in Oliver E. Williamson and Sidney G. Winter (eds) The Nature of the Firm: Origins, Evolution and Development, Oxford: Oxford University Press, pp. 61-74.
Cosmides, Lena and John Tooby (1994) Better than rational: evolutionary psychology and the invisible hand’, American Economic Review, 84, 327-32.
Foss, Nicolai J. (1999) ‘Edith Penrose, economics and strategic management’, Contributions to Political Economy, 18, 87-104.
Hahn, Frank H. ( 1984) ‘On the notion of equilibrium in economics’, Cambridge University Press; reprinted in Equilibrium and Macroeconomics, Oxford: Basil Blackwell, pp. 43-71.
Heiner, Ronald A. (1983) ‘The origin of predictable behavior’, American Economic Review, 73, 560-95.
Kay, Neil M. (1997) Pattern in Corporate Evolution, Oxford: Oxford University Press.
Knight, Frank H. (1921) Risk, Uncertainty and Profit, Boston:Houghton Mifflin.
Loasby, Brian J. (1995) ‘Acceptable explanations’, in Sheila C. Dow and John Hillard (eds) Keynes, Knowledge and Uncertainty, Aldershot: Edward Elgar, pp. 6-24.
Loasby, Brian J. (1999) ‘Industrial distiricts as knowledge communities’, in Michel Bellet and Corine L’Harmet (eds) Industry, Space and Competition, Cheltenham and Northampton MA: Edward Elgar, pp. 70-85.
Marshall, Alfred (1919) Industry and Trade, London: Macmillan.
Marshall, Alfred (1920) Principles of Economics, London: Macmillan.
Marshall, Alfred (1994)’ Ye machine’, Research in the History of Economic Thought and Methodology, Archival Supplement 4, Greenwich, CT: JAI Press, pp. 116-32.
Ménard, Claude (1994) ‘Organizations as co-ordinating devices’, Metroeconomica, 45, 224-47.
Nelson, Richard R. and Sidney G. Winter (1982) An Evolutionary Theory of Economic Change, Cambridge, MA: Belknap Press.
Penrose, Edith T. (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell. 3rd edition Oxford: Oxford University Press 1995.
Popper, Karl R. (1963) Conjectures and Refutations, London: Routledge and Kegan Paul.
Richardson, George B. (1953) ‘Imperfect knowledge and economic efficiency’, Oxford Economic Papers, 5, 2, 136-56; reprinted in Richardson (1998a), pp. 1-21.
Richardson, George B. (1972) ‘The organisation of industry’, Economic Journal, 82, 883-96; reprinted in Richardson (1998a), pp. 143-56.
Richardson, George B. (1975) ‘Adam Smith on competition and increasing returns’, in Andrew S. Skinner and Thomas A. Wilson (eds) Essays on Adam Smith, Oxford: Clarendon Press, pp. 350-60; reprinted in Richardson (1998a), pp. 157-67.
George B. Richardson (1998a) The Economics of Imperfect Knowledge, Cheltenham, UK and Northampton, US: Edward Elgar.
Richardson, George B. (1998b) ‘Competition, innovation, and increasing returns’, in Richardson (1998a), pp. 168-77.
Ryle, Gilbert (1949) The Concept of Mind, London: Hutchinson.
Schumpeter, Joseph A. (1934) The Theory of Economic Development, Cambridge, MA: Harvard University Press.
Schumpeter, Joseph A. (1943) Capitalism, Socialism and Democracy, London: Allen and Unwin.
Shackle, George L. 5. (1979) Imagination and the Nature of Choice, Edinburgh: Edinburgh University Press.
Simon, Herbert A. (1982) Models of Bounded Rationality, 2 volumes, Cambridge MA: MIT Press.
Smith, Adam ( 1976a) The Theory of Moral Sentiments, ed. David D. Raphael and Alec L. Macfie, Oxford: Oxford University Press.
Smith, Adam ( 1976b) An Inquiry into the Nature and Causes of the Wealth of Nations, ed. Roy H. Campbell, Andrew S. Skinner and W. B. Todd, Oxford: Oxford University Press.
Smith, Adam ( 1980) ‘The principles which lead and direct philosophical enquiries; illustrated by the history of astronomy’, in Essays on Philosophical Subjects, ed. William P. D. Wightman and J. C. Bryce, Oxford: Oxford University Press.
Van de Ven, Andrew and David N. Grazman (1997) ‘Technological innovation, learning and leadership’, in Raghu Garud, Praveen R. Nayyar and Zur B. Shapira (eds), Technological Innovation: Foresights and Oversights, Cambridge: Cambridge University Press.
Vromen, Jack J. (1995) Economic Evolution, London and New York: Routledge.
Witt, TilIrich (1998)’ Imagination and leadership - the neglected dimension of an (evolutionary) theory of the firm, Journal of Economic Behavior and Organization, 35, 16 1-77.
Ziman, John M. (1978) Reliable Knowledge, Cambridge: Cambridge University Press.