Introduction

Why are policies so often poorly constructed and badly implemented? An increasingly popular answer points the finger at politicians and civil servants. Full-time politicians, and today most politicians are full-time, run errands for their constituents, stopping from time to time to gibe at their opponents. As a result, the few who make it to ministerial rank have had scant opportunity to acquire the knowledge or develop the judgement to oversee and advance a department’s policies. The civil servants have little incentive to develop the competencies in policy-making and management that they cannot deploy in conversations with their ministers. What use is a servant who cannot be understood?

Is this obvious answer the correct one? Behind it, is a judgement that attainable standards of rationality have not been reached. What are these standards? We cannot assess the answers without, at least, some general understanding of what these are. The best place to start to find this understanding is in the tradition of enquiry into decision-making established by Herbert Simon (1916-2001). The work of Simon and his followers helps us explicate the assumptions that underlie our appraisals of policy. Drawing on this I present an outline of the ‘standard model’. The model exposes the limitations of most critiques and their prescriptions for reform (including, alas, those of the government). The renewed interest in pragmatism does provide guidance, allowing me conclude with an account of the roles ministers and civil servants play in carrying forward policy-narratives.

The standard model

The economist’s lesson for the manager is clear: to maximise profits make marginal costs equal to marginal revenues. The relevance of the lesson is less certain. It hardy answers the manager’s question ‘how can I maximise profits?’ Neither marginal costs nor revenues are readily found in the here and now when the manager must make decsions that impact on his firm’s profitability. The theory is clear. The gap between it and practice is clear. Simon was intrigued by this gap and how it might be closed.

He recognised that individuals navigating their way through the uncertainties and complexities of everyday life behaved differently from ‘homo economicus’ whose rational decision-making delivered a corner stone of economic theory. It is not that ‘homo adminstratus’ did not want to follow the ‘rational model’ that guided ‘homo economicus’; nor did the model lack rationality. It was just that circumstances frequently frustrated its deployment. A rational decision results in (scarce) resources being deployed to optimum effect. To accomplish this, the rational model instructs us to identify the ends we wish to attain, find the all means that might be deployed to reach them, identify their consequences, and finally to pick those means that maximise the achievement of our aims at minimum cost. That is all. There are circumstances where the model finds ready application. The owner of a number of oil well and refineries dispersed over a wide area could use it to work out the optimal deployment of his fleet of tankers. An Urban District Council could use it to plan the replacement of light bulbs to meet the standard of illumination at minimum cost.

However such circumstances are uncommon. Most of the time our ends are competing, or conflicting, and we find it difficult to prioritize them. Given the complexities and uncertainties of the world great efforts and resources are required to find all the means and identify their consequences. To evaluate the consequences, we must imagine a world in which they are realised and the human imagination is an unreliable instrument. The mundane decision of choosing a holiday destination illustrates the difficulties. For example, how can you square the desires for solitary enjoyment of a wild and rugged landscape and fine meals in Michelin starred restaurants? How many brochures should you peruse and how much effort should you put into verifying them? Finally, how much trust should you put in your imagination as it helps you escape from bleak mid-winter into sunnier times?

Given that the rational model is no guide, how do we proceed to make decisions? Simon proposes that as we pursue our ends we aim to ‘satisfice’ rather than optimise: we are satisfied with what suffices. Rather then search the haystack for every needle it contains so that we can find the sharpest, we pick the first one we find that is sharp enough for the job. Simon points out the central role that prejudice plays in our decision-making. We make most decisions on the basis of unquestioned assumptions. We take as certain what is uncertain and ignore complexities while we attend to those that we believe to be important and which we can manage. In Simon’s terms, we act with ‘bounded rationality’.

Simon’s major contribution to organization theory was his explication of the role organizations play managing the ‘buzzing, blooming confusion that constitutes the real world’. Organizations helped close the gap between theory and practice. He invites us to see organizations as systems of decision-making. Consider the modern (machine) bureaucracy. The tasks assigned to the workers who produce the goods or services are tightly circumscribed. There is little variety in the decisions they make. These present few options and require little thought. The programming of decisions in this way has several advantages. It reduces the uncertainties that would follow if workers had scope to manage their interdependencies with other workers. It allows the implementation of quite complex technologies in large chains of highly interdependent tasks with highly predictable outcomes. The programming of decisions sets the bounds to the workers’ rationality. We can think of the way an organization sets these bounds as the ‘stipulating of facts’. The worker does not question the wisdom of his routine. He takes it for granted that it is the best guide. However circumstances can disrupt it: power fails, inputs to the task do not arrive, outputs from it do not move on. The stipulated fact is no longer a guide. The responsibility for correcting it to match the new reality assigned to levels higher up the hierarchy. The levels immediately above the base monitor the execution of the programme’s routines and accommodate them to minor changes in circumstances. Higher levels, who see more of the big picture have the authority to modify, or adapt, programmes to handle more significant disturbances and solve problems of coordination among programmes. The highest levels are responsible for adjusting programmes. They decide on the allocation of resources and when existing programmes should be changed or abandoned and replaced with new ones. These strategic decisions align the organizations capacities, embedded in its operations, to its interest in acquiring resources from outside the organization.

The Garbage Can

This is not the place to trace out how Simon’s propositions have been refined and developed to make a major contribution to organization theory. However, one episode is instructive for the light it throws on the civil service. In the 1970s researchers working in the tradition turned their attention to organizations that resisted bureaucratic structuring. ¹ Clarity of purpose and well understood technologies are necessary conditions for the programming of decisions and the hierarchy of decision-making that ascends from supervision to strategy. They called organizations that lacked these characteristics ‘organized anarchies’ and they developed a model – the garbage can – to explain how decisions were made within them. A University was their prime example of organised anarchy. It is easy to see why. Universities have competing aims with little or no prospect of fixing preferences among them. They are the guardians of our civilisation responsible for teaching its fruits to students. They contribute to the expansion of knowledge for its own sake. They serve the needs of a knowledge economy. How can the value of one discipline be balanced against another? Research for its own sake against research for industry? Time spent teaching against time spent researching? While we know how to manufacture cars, there is no technology to programme decisions to execute what ever answers are found to these questions.

In the garbage can model decisions result from the conjunction of four flows- problems, solutions, participants, and choice-opportunities. A choice-opportunity occurs whenever a decision has to be made. For example, a meeting is held to decide on next year’s timetable. The participants who assemble for the meeting bring with them both problems and solutions. The problems are a mixture of personal and organizational. The solutions are individual capacities. Whatever the ostensible purpose of the meeting, the participants will see it as an opportunity to find solutions for their problems and problems for their solutions. So my solution, a course on decision-making models- solves your problem of finding a course for students of public management (organizational) that allows you keep Fridays free (personal). While both of us leave the meeting happy, other problems depart unsolved, other solutions leave unattached.

On reflection, it is clear that the four flows feature in all organizational decision-making. When preferences are clear and technologies well understood, participants, problems and solutions can be brought into fixed relationships. As we have seen, in machine bureaucracy, each level in the hierarchy has responsibility for solving problems that are beyond the bounds of the stipulated facts of the inferior level.

This level of regulation is not possible in an organized anarchy. Consequently, the flows are more or less random and the outcome of their conjunction in a decision-situation can not be guaranteed to solve all problems or deploy all solutions.

Lindblom and policy-making

Simon and his colleagues were interested in how we could improve the rationality of our decision-making. They examined how individuals made decisions and how organizations structured individual decision-making to achieve a rational direction and control over endeavours that were beyond any one individual’s rational capacities. Lindblom, a political scientist, was interested in a particular type of decision: those made by policy-makers in a democracy. Like Simon, he supposed that that the analysis should begin with the rational model. The realities of policy-making were to be gauged by their distance from this model. The distance was quite far. It had to be. The number of interest groups (stakeholders), each with their own criteria for judging policy and the complexity of the problems posed made it so. Lindblom ² provides a (rational) strategy for deciding how the multiple criteria could be satisfied and means found. He advises the policy-maker to avoid any attempt to spell out clearly the objectives of policy. These are contentious. Even if the stakeholders could be brought to agreement on the objectives, they will not agree on how priorities should be assigned among them. The chances that some synoptic view could be found that brought all the competing values into harmony is remote. Lindblom reminds the policy-maker that agreement on specific proposals is far easier to find than agreement on abstract principles. This leaves the question as to where the specific proposals are to be found. The chance that some theory will provide specific guidance is slight. Certainly, the policy-maker in Finance can turn to economists, in Justice to criminologists, in education to psychologists/sociologists… However, he will find that the theorists, even if they subscribe to the same theoretical framework, will disagree in their prescriptions. Furthermore a policy problem seldom falls neatly into the domain of one discipline. It will have economic, social, psychological aspects. Rather than look to theory, Lindblom advises to look to existing policy. He points out that policy is seldom made de novo. Policy-making is more often than not, about solving problems that arise within the existing resources, commitments and organizational arrangements. Proposals for resolving these should be sought in incremental changes to these. This ‘incrementalism’ has several advantages. In an uncertain, complex, world we can never be sure that what appears as a plausible adjustment will not have negative unintended consequences. Incremental changes can be easily reversed and the damage mitigated. Problems rarely concern the policy per se. They occur within a course of action which is generally satisfactory. No one knows why the policy works as well as it does. The knowledge is not captured by theory but is implicit in the action. The search for incremental changes taps into this implicit knowledge.

The roles of the civil servant in the standard model

The standard model provides not one but three roles for the civil servant: The bureaucratic manager, the bearer of solutions and problems, and a canny incrementalist.

The bureaucratic manager

This is the role all efforts to reform the civil service have unquestionably assigned to the civil servant. Completely ignoring the characteristics of the Court in which he operates as a servant of democracy, these efforts draw on the experience of the private sector where some lessons of the standard model have been learnt with considerable profit. Viewed from this perspective, the civil servant is a lamentable figure. He lacks a sense of professionalism and the education in management that would sustain it. He lacks the leadership skills to put shape on the haphazardly coordinated and poorly controlled organization that he supervises. His failure to implement the good advice that he receives from business consultants exasperates all right thinking people. This advice is straightforward and aimed at converting government departments into machine bureaucracies. Strategy statements with clearly defined aims allow analysis identify purposes and plans actions to achieve them. These plans can be elaborated into divisional, then unit, and finally, individual plans. Performance measurement at every level, not only controls, coordinates and motivates, it also allows the methodical correction of the facts stipulated in the plans. A far higher level of rationality in both the formulation and implementation of policy is thus achieved. And, just as important in a democracy, far greater levels of accountability and transparency are attained.

The bearer of solutions and problems

While the standard model teaches us the characteristics of machine bureaucracy and its merits, it teaches us more. This more is, surprisingly, absent from discussions of civil service reform. The standard model was part of a wider development in organizational studies. This was the emergence of ‘Contingency Theory’ in the nineteen-sixties. Contingency Theory ³ challenged the view that there was one best way to manage and structure organizations. Its researchers showed that styles of management that led to success in one set of circumstances were irrelevant, or misleading, in others. It was a matter of ‘horses for courses’. The task was to identify the salient features of an organization’s circumstances (contingency factors) and find the structural features best matched to them. Through its focus on decision-making, the standard model contributed to this effort. Machine bureaucracy was not the ‘one best way to organize’. It was the best way when aims were clear and well understood technologies could guide the purposes that pursued them. In the absence of these ‘contingency factors’, Machine Bureaucracy is not the solution. Civil service departments do not have clear aims, nor can they command well understood technologies. They are, like Universities, ‘organised anarchies’. The role of the civil servant is manifest in those ‘choice-opportunities’ where the problems and solutions flowing from the political system meet those originating within the Department. The former, of course, have more authority than the latter. In this context, the civil servant does not fail for want of leadership skills or management know how. He fails because of the quality of the solutions/ problems that he brings to the meeting. Lindblom’s examination of policy-making instructs us where these are found, and therefore, perhaps indicates how their quality might be improved

The civil servant as canny incrementalist

The standard model provides us with a ‘contingency theory’ of decision-making. The salient features of a decision are how well both means and ends were known. Decision-making procedures where both ends and means are certain are quite different from those required when neither is well understood. Fig 1 presents a typology of decision-making situations.

Fig 1

When commentators (and academics) criticise policy-making, they almost invariably assume that the policy under examination is either A or C. In both cases the call is for more expertise, evidence –based policy-making, cost-benefit analysis and rigorous evaluation of outcomes makes evident sense. Civil servants are, accordingly, criticised for not bringing problems and solutions couched in these terms to the meeting. But is this reasonable? Certainly, there are policies, perhaps a substantial number, which fit into these categories and would greatly benefit from this advice. However, the majority of policies with which the public sector is concerned fall into B or D and these are not assisted in any way by the advice. The question is how we improve civil servant’s performance in identifying what problems are occurring within the flow of policy and the incremental changes might mitigate them.

Unfortunately while the standard model is helpful in indicating the limitations of most of the discussion of civil service reform, it is less helpful in pointing the direction forward. It tells us that we need canny incrementalists working in successful organized anarchies but not how to get them. Pragmatism does address this limitation of the standard model. I turn now to discussing how it complements, or corrects, each of the three contributions reviewed above and the role that it assigns to the civil servant.

The pragmatic approach and the rational–model

The pragmatic approach ⁴ invites us to step back from decisions, how they are made, how they should be made, and pay attention to the context in which they are made. This context is action. And when we consider our actions, we quickly realise how little decisions figure in it. Most of our actions are conducted as routines. Certainly, the execution of routines requires decisions. However, most of the time, these are barely conscious accommodations we make to align our regimes with changing circumstances. So, for example, when we are stopped in the course of action and asked why we are doing what we are doing, we find it difficult to track back to the decision that started us on this course. I cannot readily identify the decision, or decision processes, that lead me to write this. I can, however, identify how it serves the purpose of achieving some of my aims while attending to my interests. The point is that actions do not originate in decisions but decisions originate in the actions. They are provoked by obstacles encountered in the course of action. I hope to show that this pragmatist perspective adds valuable lessons to those that we have learnt form the three contributions of the standard model that I have discussed. Before proceeding to each contribution, one general advantage that Pragmatism has over the standard model is worth mentioning. This is its superior success in handling the role of values in decisions.

Decisions cannot be understood without attention to values. This is clearly seen in the standard model where decisions are understood as the relating of means to ends that are given by values. However, while values are indispensable to the standard model, they are not intrinsic to it. They come in from the outside. They are a given and not subject to decision themselves. When others figure in our decisions, we are interested in their values not because we are concerned why they think this or that good, but because they indicate their interests.

For Pragmatism, a serious obstacle prompts the question ‘why am I doing this?’ The taken for granted aims, purposes and interests come into reflection as we ponder how the handle the obstacle. We want to hold on to what we have achieved in the action, and apply what we have learnt in the course of it, to overcome the obstacle. So we frame decision on what to do next in a narrative. The question is how we are to ‘write’ the next ‘sentence’. A new understanding of our aims, a revaluation of what purposes serve them and a rebalancing of our interests may be required. The shift to Pragmatism is to move from decisions as scientific calculus to decisions as hermeneutic judgements in which aims, purposes, and interests are the heuristic, that tentative understanding of the whole, that guide our decisions on what to do next. As we move backwards and forwards from the whole to the action to be taken, our understanding of the whole changes. We find that the givens –interests, values- that the standard model brings to its analysis are variables within the pragmatist account. Furthermore, importantly, a coherent account can be given of why we change them in our engagement with the world.

Now to examine what Pragmatism adds to each of the contributions. 

Pragmatism and the rational model

The rational model fits circumstances where limited resources are the obstacle. There are no conflicts among our aims and our purposes are in order. The decision to be made is concerned solely with our interests as we ask how we can maximise the achievement of our aims and purposes at minimum cost?

As Simon points out these circumstances are more the exception than the rule. In pragmatist terms, our actions typically involve competing aims and purposes that, from time to time, encounter obstacles with no readily available solutions. Family life provides an example. What is required to procure the material needs of the family can conflict with the care and attention that are the childrens’ due. Parents find it difficult to balance what they owe their children with what their own well-being appears to demand. There are no well understood technologies to guide their actions. In these circumstances, bounded rationality and stipulated facts are unavoidable. These come with the practices the parents inherited from their parents. It is in these that they learnt the genre of family life with which they navigate towards competing goals through poorly charted waters. How often do parents find themselves saying to their children, in the same tone of voice, things their parents said to them and which they swore they would never say to their children? What parents have not encountered obstacles that prompted a radical change in script accompanied by a new understanding of their roles and the good it aimed at that was quite different from their parents?

What advantages come from this rewriting of Simon in pragmatist terms? The first has already been mentioned: it shows how values are brought into the picture and are opened to question when obstacles are encountered. The second is the manner in which the social setting that shapes actions and decisions is brought to our attention. Of course, Simon is aware of the social setting. Unlike many economists, he recognises the importance of organization. He has made major contributions to organization theory. However, his unit of analysis is the individual. This individual is a fallen angel banished from a paradise where the rational model is the rule, not the exception. With his capacities for rational action diminished, he has no option but to work in organizations, accepting the bounds that they place on his rationality and taking for granted the stipulated facts he might otherwise have good reason to question.

Pragmatism offers a more realistic account. The individual emerges within action. It is as an actor in its genres that his individuality emerges. Bounded rationality is a horizon of possibilities and its facts are provisional rather than stipulated. This individual may, indeed find himself in an organization whose scripts are structured as Simon prescribes. This is not his fate qua human being but as a subject of the exigencies of capitalist production whose considerable material benefits come with a high cost.

The garbage can model and pragmatism

Consider the meeting that the garbage can invites us to observe. The rational, logical, discussions that clear objectives and well understood technologies allow are not possible. In absence of this narrative, no narrative is possible: only anarchy with its flows of participants, solutions, problems, and choice-situations. If collective rationality is not possible individuals can be individually rational in advancing their interests. The model shows how these relate to the organizational setting by presenting them as ‘solutions’ and ‘problems’. Participants as we have seen bring some combination of personal and organizational problems to the meeting. They also bring ‘solutions’, expertise, or capacities that come with their organizational roles.

How realistic is this? The absence of a narrative built around the rational pursuit of clearly stated ends with clearly understood means is realistic. However, how plausible is it to suppose the rational/instrumental narrative is the only one? Revisiting the meeting from a pragmatist perspective, we find participants who are not haggling over what one will do for another. They are conversing. Each has a different perspective on the agenda. This is given by his role within the organization. He finds his aims/purposes and interests within it. To contribute to the meeting he must understand what the others are saying; know where they are coming from. He interprets what they are saying by extending his perspective to incorporate some of theirs. His achievement in this regard is tested as the conversation flows. This merging of perspectives is possible because the meeting is an episode in a narrative in which individuals with different, interdependent roles, discuss and argue as they protect their interests, propose purposes or defend both in the light of their aims. The narrative unfolds as the pursuit of broadly defined competing aims. As it proceeds the aims are reformulated and the balance among them adjusted. The broad aims are what the participants have in common. They are the possibility of the merging of perspectives in which differences can be identified and conflicts pursued and resolved.

This, it could be protested is far too idealistic a picture. Certainly no narrative is guaranteed success. Factors in the realm of interests (power) or values (ideas) can, separately or in combination derail it. Inequalities in power are the inevitable consequence of the division of labour. They allow participants play the ‘power’ card and demand that their aims and purposes take precedence. However, the ‘power’ card is not always the ‘trump’ card and a clear view of one’s interest can lead to an acknowledgement of interdependencies that can only be handled in the realm of aims and purposes. The effort to find the new interpretation of aims that will overcome obstacles can also fail. The available intellectual resources may not be up to the task. As efforts to adjust aims become more difficult, interests come under threat and participants believe that they have no option but to resort to power to defend them. The narrative disintegrates.

Does the garbage can model account for what then happens? It is unlikely that satisfactory solutions to the problems of the allocation of resources, the coordination of activities, and the generation of income, could be found by the random play of interests marked by inequalities of power. The most likely outcome is that coalitions will form as like-minded individuals with congruent interests band together to block initiatives advanced by their opponents. In time, a dominant coalition will emerge. Its first task will be the dismantling of the blocking coalitions that have left too many problems unresolved. It does this by centralising decision-making to reduce the flow of ‘choice-opportunities’ that gave the coalitions their chances to veto. The coalition must find a statement of aims and purposes and a technology that allows a hierarchy replace the unworkable anarchy. A ‘vision’ statement whose aspirations are so high-minded and general that they provide no grounds for challenging the dominant coalition’s favoured purposes accomplishes the former. An array of performance measures provide for the latter. While these have no more than a tangential relationship to the purposes pursued, they do manage the problem of competing interests. They do this by providing a non arbitrary basis for the distribution of resources that can be accepted by all interests as fair, even if it is not reasonable.

Pragmatism and incrementalism

The convergence between Lindblom’s incrementalism and Pragmatism is clear. Lindblom’s recognition that policy is generally made in response to problems encountered in the progress of policy mirrors the priority pragmatism gives to action. Lindblom’s prescription for incremental change translates easily into Pragmatism’s hermeneutics. Indeed, this translation throws further light on the merits of that prescription.

The rational model that guides Lindblom’s analysis directs his attention to interests. The obstacle is the dissatisfaction of one, or more interest groups in the outcome of policy. What small adjustment, the policy-maker, asks himself is most likely to buy off the dissatisfied without causing still greater dissatisfaction? The aims and purposes of the policy are left out of the calculation. This is unreal. No one has interests apart from aims and purposes. It is the failure of policy to address these aims and purposes that frustrates interests. Lindblom takes this step away from reality because he supposes that values belong to a realm separate from the world of facts, we deal with when we attend to interests and the consequences of action. Should we move into this realm, we would find it almost impossible to find agreement on aims and purposes. What would amount to a winning argument?

The decisions made in incrementalist policy-making do shape how interest groups understand their aims and purposes. However, whatever changes do occur are the unconsidered by products of discussions that evaluate proposed increments in terms of interests. The protagonists have no option but to settle on solutions that are fair to all interest (reflect the balance of power among them) even if this frustrates the aims and purposes of all concerned and deflects them from the goods that they seek among competing aims and purposes.

However, pragmatism sees no reason to consign values to a distinct realm. They are integral to action. Viewed in the pragmatic frame, the ongoing policy is a more or less successful narrative. The effort is to find a move that will solve the problem while holding on to that success. In some cases this can be easily done as adaptations are found that all can accept. The policy-maker need only attend to finding a new balance among competing interests. However, if that prove impossible without damage to the coherence of the policy, then purposes come into question. If a solution cannot be found in finding new ones, abandoning some or adjusting others, then aims come into view.

Pragmatism and the role of the civil servant

The typical western state has long term commitments in three domains. The first, its concern with external and internal security, is essential. The other two, economic development and social welfare are optional. Pragmatism invites us to see the policies that pursue these commitments as narratives. Responsibility for one or more of these narratives (and sub narratives) is assigned to a Department. What role do civil servants play in these narratives as they engage with the other actors – Ministers and stakeholders?

One way of delineating their role is to consider the various responses that can be made to obstacles the narratives encounter. Some obstacles are minor. They can be handled within the existing arrangements and resources. It is a question of accommodating ⁵ them within the narrative. Other obstacles are more challenging. These do affect the interests of one or more stakeholders whose dissatisfaction demands a response. It is now a question of adapting the narrative to these demands. The narrative can encounter obstacles that call into question its aims or purposes. These must then be adjusted. As a first approximation we can say that senior civil servants make the accommodations that keep the narrative moving forward (implementing the policy). They prepare proposals for the (incremental) changes that adapt the narratives to the demands of the stakeholders. They advise Ministers on the steps required to give effect to adjustments to aims and purposes. While this is a useful way of stating the responsibilities of the role, we need to go further if we are to grasp what its execution entails.

Obstacles do not always present themselves conveniently labelled ‘accommodation’ ‘adaptation’ or ‘adjustment’. Judgement may be required and the senior civil servant must be capable of making it. And to make it, he must be conversant with the worlds of implementation, interests and aims and purposes. How otherwise could he manage those tricky circumstances where what appeared as a simple accommodation had unanticipated consequences for the balance of power among contending interests? How otherwise could he recognise when efforts to resolve conflicts of interests amounted to ‘squaring the circle’ and that adjustments in aims and purposes was required? While the senior civil servant may not be responsible for sanctioning new compromises or authorising adjustments in aims or purposes, he should be a competent contributor, bringing his knowledge of the narrative’s history and his appreciation of the stakeholders’ interests and limited perspectives, to bear in the conversations on these matters.

At first sight this might seem to suggest that the civil servant was no more than an honest broker. He is more than that. Each of the contributors to the conversations that carry the narrative further has his own narrative. His contributions acknowledge that this narrative depends upon the cooperation of others. The narratives of interest groups have their roots in the economic and social activities of civil society. The Minister’s narrative clearly belongs to the politics with its stories of elections won and lost, promotion achieved or denied. What is the source of the civil servants’ narrative? The achievements of the policy depend on how successful its conversations are in merging horizons, or bringing the interlocutors to identify their interest in the context of the good for all. While all the other interlocutors must move into this horizon, the civil servant already orientates his actions within it: his aim, the good for all, his purpose, making democracy work. He moves towards his aim by assisting the other interlocutors locate their particular interests in the common advantage as the conversations of democracy proceed.

As I noted in discussing the ‘garbage can’ model, narratives are not guaranteed success. They can be subverted by power, or frustrated by inadequate knowledge. Given the great disparities of power in our society and the complexity of the state’s ambitions (eg solve the drugs problem, reduce crime, abolish poverty, sustain a capitalist order…) we should not be surprised at the number of stumbling narratives. The consequences of failed narratives are readily seen in our politics. Parties compete with visions so general that they commit them to nothing. (‘Ireland the best small nation in the world by 2016’). Aims/purposes drop out of political discourse. Demands for greater accountability and transparency increase as interests, deeply suspicious of their competitors, seek assurances of fair treatment. Civil servants are cast as bureaucratic managers. As the possibilities for successful policy-making diminishes, frustration at failed policies increase. In Part 4, I examine what steps might be taken to improve matters, given the constraints of audience democracy.

audience democracy, Contingency Theory, Garbage Can, Herbert Simon, incrementalism, Lindblom, Machine Bureaucracy, policy, Pragmatism