Is Education Rational? (a sketch for an argument)

This is what I've been working on for a couple of weeks now. It's chapter 6 of my book, and in many ways it forms a central part of the argument for the whole thing. At the moment it's a bit disjointed, and it's changed quite radically over the last 2 weeks partly because I was sent a remarkable paper by Robert Ulanowicz (thanks to Loet Leydesdorff) on Shannon's information theory. Ulanowicz is an ecologist who uses information theory to characterise ecological health, diversity, and so on. The paper I was sent dealt with Shannon's formulae of information, but contained an important critique. In particular, Ulanowicz identified a central confusion that surrounded Shannon's use of Boltzmann's statistical thermodynamic equations, together with confusion over the application of the term 'entropy' to information theory (which, Ulanowicz asserts was the result of a joke by John von Neumann). At the heart of the problem is the assumption of probabilities of 'successful communication' against a fixed base in the context of noise, redundancy and so on; in place of this, Ulanowicz argues that information needs to be considered 'relativistically'. My immediate thoughts were (having wrestled with Shannon a lot in the last few years), "Dammit! Of course it does!"

More importantly, however, was the fact that Ulanowicz talks about the importance of 'not information' as opposed to information. Mirroring my own obsession with absence (which I got from Bhaskar and later Deacon), he characterises 'not information' as flexibility, and contrasts it with 'mutual information', which (derived from Shannon) refers to common codes within a communications system. As has been the case with Leydesdorff's work (which I have played a small role in) the negative view simplifies the equations. In Ulanowicz's case, it means that the overall complexity of the system, measured as 'uncertainty' equals the sum of the mutual information and the flexibility:

H=A+θ

Now, flexibility looks rather like the 'excess energy' that Bataille talks about in his economic theory. 'Giving' flexibility becomes the fundamental driving force in education, not the establishing of mutual information (this, I think is the most powerful argument against the simplistic cybernetic-conversation models of education like the Laurillard/Pask model, or the less sophisticated Siemens/Downes connectionist model). There are implications of this which then apply to the way we organise and manage education - which I will deal with later. One in particular is worth drawing attention to: FEARFUL INSTITUTIONS EMPHASISE MUTUAL INFORMATION AT THE EXPENSE OF FLEXIBILITY. Consequently they become brittle and unadaptive. Since there's a lot of fear in the sector, I'm interested in this angle particularly.

Here's the paper in it's half-finished state...

Reason, Value and Education

Introduction: Reasoning

Is education rational? Is it possible to use scientific methods and logical tools to reach conclusions about how to make education better? The question mirrors a broader debate about the possibility of naturalism in the social sciences more generally. On considering the great scientific advances of the enlightenment, the importance of reproducible experiment, the processes of “conjecture and refutation” and the closing of gaps between theory and practice, to what extent is the naturalistic scientific investigation of education possible?

Comte imagined a ‘social physics’, where causes could be established with a certainty that mirrored the certainty in the physical sciences. Yet it wasn’t long before reflexivity, the role of language, the nature of discourse created loops in processes of discovery and explanation whose complexity contrasted with what appeared as the "solidity" of reproducible experiments and event regularities in the physical sciences. Indeed, the discovery of the importance of discursive processes provided an important contribution to the appreciation of physical sciences, as awareness of discourse more generally left nothing untouched: from Kuhn’s paradigms to Feyerabend’s highlighting of the irrationality of scientific method itself, issues of ego, value, identity, responsibility, obligations, roles, attachment and personal history all appear as inseperable from the processes of scientific thought. 

The social science response, and particularly the response in education, is to work towards consensus. Constructivist thinking, facilitating discussion among learners, and guarding against the arrogance of experts and sage-like practices, influenced the practice and thinking about education itself. The way forward, the rational way forwards is to regard everyone involved as a learner. The way forward was to find ways in which people could talk to each other. Unfortunately, it is very hard to reach consensus as to what to do about educational problems. Whilst the ‘hermeneutic circle’ has a democratic appeal, and space is given to self-expression, its permissiveness does not mirror the constraints of the everyday political world. If consensus was the last hope for rationality and consensus about educational policy is challenging, how is any rational path possible?

There is a problem with consensus and conversation: no account of conversation considers individuals as ‘hysteretic’ – organisms whose dispositions, values and utterances have been shaped by things that have happened to them in their personal history. Conversation becomes an 'ideal' type. The British-based Austrian émigré philosophers Karl Popper and Isaiah Berlin (together with the economist Friedrich Hayek) criticised this kind of idealism as tending towards the utopianism which they saw at the heart of the processes that led to the destruction of their country. As Popper identified, “idealism” was a trend of thought in European culture prevalent in Germany, but with its roots in Plato. Berlin felt consensus to be impossible on the grounds that core values sitting behind each individual’s position need not in any way be commensurable with one another, irrespective of their merits as core values in each case. Quoting Goethe, “Out of the crooked timber of humanity, nothing was ever built straight” Berlin’s argues against the possibility that human values could never be agreed: some values were fundamentally rooted in deep experience and could at the same time be fundamentally true in their context, and at the same time incommensurable.

The belief that core values can be in some way broken-down and differences healed and practice changed tends to typify the positions of those who crave a better education system. There are plenty of examples of attempts to create a better system: for example, one based on ‘transformative learning’ (Mezirow), or one based on learning conversations (Harri-Augstien and Thomas), or of Inquiry-based learning, or of deschooling (Illich), and so on. The proposition that this should be possible sits on the central proposition of any educator: that lives can be utterly transformed with powerful conversations. However, the central difficulty and problem is that what is meant by ‘conversation’, what is meant by ‘consensus’, what actually happens in teaching and learning is never specified in a way which sits outside the particular value-set from which transformative change is advocated. Value sits beneath everything, and whilst personal change does emerge through powerful conversations, it is not clear that changes to personal values relates to this.

The question about whether education is rational depends on a deep examination of the ways that individuals communicate, express their values, experience changes to their values, and make decisions. Fundamentally, it is a question about how we think, and how we learn to think.

Irrational debates?

What would rationality look like in education? One view is presented by Jon Elster who represents Rational Choice Theory with the diagram below:

Here, humans are constituted through desires and beliefs which result in actions (or perhaps preferably, decisions). At the heart of this model of rationality is information. Information informs beliefs about the state of the world: the states of other people, the likely impact of decisions made, the desires and beliefs of other people in the world. Where desires are articulated, they too become information. Furthermore, desires may indirectly influence beliefs: desiring the love of somebody doesn’t mean that we believe it to be true; but the desire may constrain the way that beliefs are formed, for example, by precluding other beliefs which are not compatible with a particular desire.

The basic rational choice diagram is a way of representing the situation portrayed by Game Theory. Players in a game have information from which they form beliefs about other players, and they have desires to attain one goal or another. Game trees can be drawn to show the options available to a particular player. Game theoretical choices have a number of different representations. The simplest is a binary tree, where each node has two options:

Each option in the tree carries pay-offs or penalties to a particular route chosen. Evaluating these payoffs and penalties  becomes central to the rational process. A decision can be taken as to that node which the most advantageous pay-off for the player making the choice.

A different way of representing a game tree is to use a table in what game-theorists call 'normal form'. The situation where players A and B have 2 options (as in the diagram above) can be written:

	A	B
A	2,2	1,2
B	2,1	1,1

Real “games” can be harder to deal with than this simple binary example. The payoff function, represented as an ordinal value (say, 1,2) in reality can be harder to calculate. The differences in types of game - whether they are cooperative, whether they involve learning between the different moves, or whether the choice between winning and losing is always balanced where winning is proportional to losing (the so-called zero sum game). Attempts to model each of these situations involves making assumptions about the behaviour of individuals, institutions, and so on. It involves making assumptions and desires and beliefs and the ways in which information is interpreted. Sometimes, it seems, game theory can work: it was instrumental when the UK government held an auction to sell bandwith for fast mobile internet connections for the maximum price. Such economic examples, usually involving maximising profit have made game-theory popular among economists.

Sometimes it doesn't work. Game theory was also used to model the different effects that changing the funding regime of universities to £9000. It was envisaged that rational behaviour by institutions would ensure that the top fee of £9000 would only be chosen by elite institutions, and that other institutions would set their fee according to market conditions. Unfortunately, what actually happened didn’t bear this out.

The assumptions and complexities of the game theoretical model are apparent if we try to model the decision-making process of a student that is trying to decide which institution to study at. Students must decide which institutions and courses are available to them. Having done this, each choice carries various risks: the risk of failure, the risk of dropping out, the risk of not getting a job at the end of it, and so on. With the change in funding regime, each risk carries varying levels of financial penalty. So, for example, if a student could be certain of a good job without a degree, then there is no sense in getting one. However, statistically, it appears that having a degree improves their employability, although it depends on the institution. It they study a course at an institution where their job prospects are relatively low, then how are they to calculate the risk-benefit of the course? Furthermore, this is not to factor-in the chances of failure or dropping-out. In some cases, a student would be better-off to drop out as soon as possible before incurring more fees - particularly if other employment opportunities become available to them. The changes to the funding regime have ushered in fundamental changes to the game that must be played by students.

This example reveals a deeper problem with the extent to which decisions about education can be rational: the extent of the divergence of opinion about education and its value are related to unequal distribution of information and an unequal distribution of the capacity to interpret information. Decisions about education would be equal if people had equal information about it. Yet the real problem with information and decision is not its quantity but its interpretation. How is it that individuals might interpret information differently?

For students, education is a very emotional decision. The anxieties of worrying about the future cause decisions to be made that sometimes do not appear rational. Elster suggests that the role of emotion in this situation is of great importance. His adapted rational choice model takes this into account:

Given this, however, drawing the line between feelings and thinking is difficult. Can it account for intuition or instinct? Here emotions exist in individual instances in ways which inform desires, beliefs and information in such a way that different decisions can look irrational.  Emotions act as ‘disruptive signals’ to processes that would be otherwise rational. Within this guise, unemotional, unempathic decision-making should appear to be the most rational. Yet it isn’t.

Metagames and Rationality

Rational choice theory, and its related elaboration in game theory has been challenged in various ways. Among these criticisms are:

1. Failure to account for collective action
2. Failure to account for social norms
3. Individualistic account of rationality

These challenges are addressed by Howard who advocates 'meta-rationality' in the form of a consideration of ‘metagames’. Meta-rationality is formed with knowledge of possible reactions by those around us to the choices we make. This is calculated on the basis of what is known about other people: in fact, it is calculated on the knowledge of the constraints that bear upon other people. For Howard, it is in the interaction of constraints where communication can take place.

Considered from the metagame perspective, the relationship between rational choice and inspection of metagames produces some significant challenges:

1. The impossibility of calculating ordinal payoffs
2. The need for constraint in making any kind of decision
3. The importance of the ‘unthinkable’ in determining a course of action

The metagame model demands that a person considers not only the available options, but the likely actions of others, including the different games and the different strategies that they might employ, in the light of each of those options being taken. Furthermore, it demands a deeper projection of the actions that each other person might take in the light of each of the choices that a person might take from that point – and so on. It quickly becomes overwhelming. How can we talk of rational choice in these circumstances? What kind of model of other people is required in order for this to work? How is this model constructed?

In this perspective, the ordinality of pay-offs and the assertion that rational pay-offs can be made appears as the imposition of a constraint on the complexity of the situation, rather than the upholding of a rational principle. Indeed, the rational principle produces the most complexifying situation. Howard’s argument is that rational choice – as it is commonly upheld as the calculation of pay-off functions within a context – is the irrational application of constraint on rational thought. This simple switch from an orientation based around the implications and comparisons of rational judgement to the implications of particular kinds of constraints raises new questions.

Before delving into this complexity, we should inspect what is considered when we think about a particular option. We might assume that we have an expectation about the probable actions of an individual if we behave in a particular way. To form an expectation, it is necessary to have some kind of model of the other person. We entertain many models of different people: person x is very organised, and so we might expect them to behave in one kind of way; person y is disorganised and so will behave differently, and so on. Any kind of ‘model’ is effectively a function which generates what we imagine to be likely behaviours. These complexities increase when we consider whether people will cooperate with us or not. Expectations are a set of projections as to what might happen if a particular action is taken.

Expectations in different circumstances are not unique. Indeed, the principle characteristic of expectations is that they overlap: the same set of expectations may apply for different kinds of options. Other expectations will be less common, or will demand more inspection of deeper motives.

The unevenness of expectations with regard to choices creates distinctions between pay-off choices. In this way, we might consider a negative relationship between redundancy of expectations and the ordinality of pay-offs which then contributes to the decisions to act in particular ways. Rationality and irrationality are asymmetric in the same way that beauty and ugliness, truth and falsehood: in human history, beacons of rationality shine out against a background of excess and waste. The rationality behind the feats of human engineering and endeavour for the construction of pyramids, palaces and cities pales before the senseless waste, exhuberance and excess of it all.

Expecting waste: An Anthropological Interlude

The relationship between rationality and waste as it manifests in culture over history presents compelling evidence to which anthropologists have drawn attention to. Human history is replete with wasteful acts: art, love, play, the quest for knowledge and the struggle for power. In the execution of wars or in in lavish expenditure on castles and palaces, or the story of love throughout the centuries, we see extravagant and often ridiculous human behaviours. Education is no exception.

How might rationality relate to this? Howard’s dissolution of rationality through the complex web of the metagame suggests that the relation between ordinary thought and the ordinal, countable or calculable disappears with the application of formal logic. Modern educational obsessions with the science, technology, engineering and mathematics (STEM) overlooks the profound role the irrational plays in the lived experience, where rationality gives way to meta-rationality. Bataille argues that the metagame of rationality is avoidance of taboo. Drawing on anthropological evidence including Mauss and Levi-Strauss, Bataille argues that rational behaviour grows as a veil that is drawn over more basic primitive and animalistic tensions. Bataille’s friend Roger Caillois followed a similar argument in presenting game-play as different forms of wasteful activity: “an occasion of pure waste: waste of time, energy, ingenuity, skill, and often of money.” (Caillois, 1961)

For Mauss and for Bataille, the ‘gift’ is the dominating feature of primitive societies. Bataille characterises “giving” is an act of sovereignty, highlighting the relationship between social requirements for giving, and the emergence of social taboos. Departing from Mauss, Bataille sees gifts not so much as a proto-exchange rituals, but rather serving the purpose of expending creative energy with which the human organism is endowed and through the exercise of which is an expression of the sovereignty of the individual. If the capacity to give must be preserved, then the resources with which one might give must be protected. It is in the protection of the resources to be given that Levi-Strauss sees the beginning of basic taboos. For example, the taboo of incest is precluded because it “damages” the gift of the daughter in marriage. Prohibition of murder is an affront to the sovereignty of high priests to make sacrifice.

What emerges from anthropology is the deeper metagame of rationality: a game of primal, existential forces which drive forms of behaviour which appear barbaric and irrational from our perspective of rationality, yet about which a deeper understanding of the rational processes can reveal a richer understanding of their logical structure. Whilst rational thought about education within the lens that education itself defines is always likely to break down, the characterisation of deeper metagames behind that lens of rationality reveals new possibilities for articulating the rationality of human action.

The connection between constraint and rationality can adapt to account for different levels of rationality. Since the surface of rationality appears to be based on connection between concepts, it is unsurprising that it breaks down. However, if the light is switched on to focus on the context – the connstraints of thought – then a picture emerges of rationalities as forms of constraint, where agency is steered in different ways.

Howard’s theory of metagames suggests that the thinkable options reveal a negative image of the patterns of thinking and non-thinking which produce constraints which lead to rationally-grounded decisions. This search for the negative becomes a question of information. Yet it is not, as the rational choice theorists would argue, the information that bears upon desires and beliefs which matters – for such information remains positive in being articulated as a defence for decision-making. It is the patterns that sit as background to this information which matters. Whilst the ‘game of education’ involves the gaining of qualifications, a deeper motivation is the exercise of sovereignty and giving as a means of relieving creative energy. The rational choices of education presented to young people are constrained within the economic fiction of a capitalist society. It is an artificial and imposed constraint. Yet the release of creative energy may find many forms – it is in reconciling the different forms of creative expenditure with the rigid constraints that the system presents that the deep paradoxes and enslavement of the education system manifest itself.

Emerging from the anthropological account is a picture of modern humanity balanced between the primal forces that seek social expression in numerous ways of giving, social expectations and norms which seek to constrain primal forces, eroded structures within which the deep rationality of humankind can be situated against. Whilst we might expect most of those around us to be bound by social convention, each of us is aware of the ‘dark’ (invisible and possibly unattractive) forces that pull us in directions away from those conventions. The message from anthropology is that the dark forces have their own logic and their own ethics which is an inversion of the norm. It is in understanding how the unthinkable shapes thinkable that a deeper awareness of rationality might emerge: but this requires looking for evidence of the unthinkable, and establishing new theories that balance the impact of the unthinkable to the thinkable.

Sense and Absence

Any sense-making process involves involves both accounting for the available information and for the absence of other information. Whether we adopt a simplistic model of information processing from rational choice theory, or we adopt a more sophisticated metagame method, information about the world bears on the desires and beliefs of agents eventually produces sufficient certainty to act. We now need to understand how this reduction in uncertainty occurs.

Information theory as described by Shannon presents information as a measure of uncertainty in the transmission and reception of messages. The machine-to-machine communication which Shannon concerned himself with is very different from human situations, although there is sufficient similarity for useful parallels to be drawn. In particular, Shannon’s concept of ‘mutual information’ is particularly powerful. For Shannon, mutual information, or “transmission information” amounts to the reduction in uncertainty produced through the combination of different sources of information. Mutual information, or ‘interaction information’ is useful in describing the ways that causes are ascribed to events. For example, we might consider:

T (blood on floor; missing person; bloodied knife)

= (blood on the floor; missing person | bloodied knife)

= (blood on the floor; bloodied knife | missing person)

= (missing person; bloodied knife | blood on the floor)

If we consider three ‘clues’ as representing a level of uncertainty about the conclusion, we might say that

HA=blood on the floor

HB=bloodied knife

HC=missing person

HA, HB and HC overlap as in the Venn diagram below:

So the reduction in uncertainty with the combination of the three clues can be given by:

T=HA+HB+HC-HAB - HBC- HAC+HABC

This means that if there is no correlation between the variable A,B and C, then the equation would produce a high value of uncertainty, since the values HAB and HBC which are subtracted would be very low. If there is a high correlation between HA and HB, then the values for HAB and HBC might be high, producing a negative value.  The problem is that the switching of the sign in the calculation of the information variable seems counter-intuitive: a positive uncertainty is a strange phenomenon. Indeed, this is a situation that doesn’t occur in the dealing with information interaction between two sources.

However, the information contained in HA or HB sits against a background of redundancy. Overlapping of information (measured as uncertainty) is equivalent to an overlapping of redundancy. This simplifies the information interaction equation: rather than examine the information content of each source, we look at the relationships between what might be excluded by each source of information:

What is excluded by blood on the floor? What is excluded by a bloodied knife? What is excluded by a missing person?

For Shannon, redundancy takes the form of added bits of information: repeating messages, grammar and so on all constrain messages. This is done so as to facilitate the transmission of messages successfully across a noisy medium. In human communication, the nature of noise itself is not separable from redundancy: a noisy room might cause somebody to repeat themselves, but it also reveals something about themselves in their unwillingness to move to a quieter room to talk! Noise contributes to the metagame – whilst noise is ‘not information’ in the same way that redundancy too is ‘not information’.

Robert Ulanowicz has recently argued for a deeper focus on ‘not information’ as a whole, suggesting that the measure of information should be seen relative to its context and not measurable against absolute probabilities of successful or unsuccessful transmission. Redundancy in Shannon’s sense is a value which sits relative to the circumstances of electronic communication and the need to transfer messages over a noisy medium. Is it necessary to present a different picture of ‘not information’. Ulanowicz suggests that an overall relative measure of information must take account of the mutual information plus what he calls Flexibility, which is the sum total of possibilities of interpretation, redundant pathways. Both the uncertainty of information and the uncertainty of flexibility need to be contextualised relativistically: any probability that determines the ‘success’ of transmission of a message must be related to things that have happened before.

Information results from negative feedback – whether it is the Watt governor attenuating the speed of the engine, or the different mechanisms in Ashby’s homeostat feeding back on each other to create stabilities in the mechanism. However, in nature we also see positive feedback: from the growth of plants, populations, cities, economies through to global warming. Ulanowicz points out that negative feedback is easier to study since it results in states that are stable within boundaries; positive feedback on the other hand quickly results in systems that fly out of control. Yet as a natural phenomenon positive feedback has important properties.

1. It results in transformation of an entity which creates new contexts within which information is situated
2. It operates as an open system rather than a closed system (with negative feedback)
3. In transforming and growing, it creates new possibilities for development
4. It avoids atomism in its explanations

Processes of positive feedback interact with one another. Although positive feedback mechanisms can grow out of control, two systems with positive feedback can interact so as to exhibit both negative feedback on each other and positive feedback within the system. This produces what Kauffmann calls an ‘autocatalytic set’, which Ulanowicz draws:

In one direction, there is positive feedback, in the other there is negative feedback.

Ulanowicz presents his equation:

H=A+θ

as a way of summarising this process of ‘ascendency’. Total uncertainty is equivalent to the mutual information between entities, together with the autocatalytic ‘flexibility’ between those domains. This is not just the coherence of utterances within a ‘code of communication’, which would characterise ‘negative feedback’, although utterances are constrained within a code of communication (where such a code of communication is measurable by A). F, however, is everything else that is potential but unrealised.

We can speculate as to what happens with increasing processes of flexibility. An autocatalytic growth of the system causes interaction with other autocatalytic systems. The overlapping of redundant pathways between autocatalytic sets creates new domains for negative feedback, whereby the twin system becomes itself an information system which also exhibits its own autocatalysis. Another way of thinking about this is to consider the metagame model where the redundancies which emerge are redundancies of expectation of the agency of other people. Whilst the multiple presence of an expectation may create an “information pathway” which facilitates the making of a decision (so it is a blinding of decision pathways), autocatalysis of another entity transforms expectations such that new weights of probability emerge with new redundant pathways which overlap.

The redundancy of expectation amounts to the flexibility to think in different ways: when flexibility is reduced, it will seem that there is ‘no alternative’ to thinking in a particular way; when flexibility is enhanced, free thinking and open expression of ideas might be encouraged. Each individual will differ in the extent to which they possess flexibility to think differently: since ‘uncertainty’ and ‘flexibility’ are relative measures, each individual will be affected by their personal habits of communication. High levels of mutual information (Ulanowcz’s ‘A’) will lead to the creation of fixed theories about the world which themselves will constrain ways of thinking which do not comply with a particular collective world-view. Very severe negative feedback in an organisation can inhibit the growth of flexibility.  The consequence of this are not just for the moment within the organisation: they have impacts in the future. Inhibition of growth and flexibility creates an organism which deals with uncertainty by adhering to a code of communication, and that code of communication contains within it the demand to inhibit growth: an organism produced through an environment of inhibited growth becomes an organism which inhibits growth. The organism produced becomes an organism for the manufacture of fear: the result is a bad habit.

Habitualisation and Information

In understanding the ways in which a system changes through processes of negative and positive feedback, we now need to characterise the emergent properties of this system in terms of the feedback processes which produce it. Fundamentally, the question concerns the ways that learned behaviour resulting from growth in an environment result in propensities to reproduce feedback patterns. In other words, behaviour becomes habitualised, and the establishment of habits constrains growth in other environments.

The relationship between reasoned behaviour and argument and habitual behaviour is analogous to the relationship between foreground and background. Habits constitute the environment of flexibility within which rationalised utterances are made. We give different words to different kinds of flexibility: for example, the habit of play is autocatalytic – it increases flexibility. On the other hand, habits of worry and anxiety accentuate mutual information (conformity) and tend to constrain flexibility to the reproduction of fears which further drive conformity. Fear is an instance of positive feedback enhancing negative feedback.

Discourses result from habitualised practices. A successful discourse must balance the reproduction of the conditions under which utterances are made (so they must conform with a code) whilst at the same time ensuring that sufficient flexibility is generated to ensure that the discourse can adapt to changing environments. With regard to the informational function of a discourse, there must be a declaration of boundaries: what is and is not relevant. The discourse reinforces the differentiation between those who are ‘in’ the code and those who are outside the code. Given that participants in a discourse come to that discourse gradually, and that the habits of participation grow from historical patterns of adaptation of each individual, is it possible to identify the origins of a discourse in the emergent discursive patterns within an individual?

The coping strategies of a baby contain both information and flexibility. The relations of a baby with carers – particularly the mother – result in emergent patterns of flexibility and mutual information: flexibility which may be expressed through play, whereas mutual information gradually emerges through the establishment of boundaries of behaviour and conditioning. In common with the ethological work of Lorenz and the attachment theory of Bowlby, we can identify that patterns of engagement in early life lead to the establishment of practices which gradually find discursive expression in communities. In later adolescence, subjects, groups, political agencies all form part of the background against which habits may be reinforced and accepted. Within those groups, there are processes of acquiring rights and responsibilities within groups serve as vehicles for the production of flexibility within a context that is increasingly dominated by mutual information. Political strategies are born in play; transformations are effected through flexibility of thought and the collision of different worlds. At the same time, ritualised codes of communication help to establish the coordinating forces within which everybody is assumed to operate: awareness and understanding of the constraints that bear upon everyone helps with the selection of utterances that may be successful.

Values from the Cradle

Politics and flexibility begin in the cradle. Early life, family life have a powerful bearing on the development of individuals. Given this, what then for the provision and nature of schooling? The emergence of habitualised practice presents a number of questions:

1. How do specific discourses arise: why does knowledge acquire a particular ‘form’ in terms of academic subjects?
2. Why is there such diversity in opinion and value about how education should proceed?
3. Why do individuals tend to different political orientations?
4. Why can’t we all be good people?

Diversity in the population emerges from early childhood. The patterns of parenting and attachment cause repeated patterns of practice in terms of habit. Habits tend towards propensities to engage in discourses in particular ways: values emerge as codifications of habitualised flexibility which encodes mutual information and flexibilities. Suppressing play in early childhood will result in a loss of flexibility in adulthood which in turn may result in behaviours which themselves lead to the suppression of play. In addressing question 1, subjects differ in terms of the respective emphasis on mutual information and the necessity to reproduce mutual information, and the extent to which they afford free play. This is not to say that mathematics can not be taught  in a playful way, but that the requirement to reproduce codes of communication – particularly if supported by parenting which similarly emphasises the reproduction of codes of communication – can result in different ways in which flexibility is expressed.

Schooling produces flexibility through exposing diversities of encoded discourses which exhibit varying degrees of flexibility and opportunities for playfulness. Such an approach helps children from varied backgrounds explore their own flexibility and background through their exposure to the discursive model. The curriculum represents institutionalised flexibility. Similarly, universities are places where the different paths that may be trodden. In codifying discourses, some disciplines struggle with paths which they don’t tread because they lack flexibility. Some disciplines cross everyone else’s path: Cybernetics is such a discipline.

With the flexibility and redundancy provided by schooling, together with the social mixtures, school contrive to create conditions where children may explore each others’ flexibility and mutual information. The contribution that teachers make to schooling is to reveal their own flexibility and capacity to participate in socially established discourses, whilst maintaining their own flexibility. In this way, there remains an essential hope within education that lack of flexibility resulting from early childhood and habitual practice might be remedied with a combination of exploitation of the tendencies of habits together with furnishing of new ways to play and acquire flexibility. This potential of schooling demands its political prominence. Yet the situation concerns not only the development of children within the system, but the determination of the system by adults, themselves who are products of the system. The nature of dependency relations that set in to various positional statements about what should be done in education each have their origin in the manner of attachment situations from early childhood through to schooling and higher education. Ministers of education are no different, and it has become unfortunately common that a political discourse and direction in education appears to result from lack of flexibility. Most importantly, this position results from values that become encoded within a partisan political discourse. Values in this sense appear as emergent from early childhood, and the result of coping strategies; just as individual human histories are varied, so too are the values which emerge from them.

Values, Flexibility and Forgetting

The central idea here is that the social world, as Elster (REF) and others have commented, exhibits “hysteresis”. Hysteresis is a term from engineering which refers to the property of elastic matter whereby deformations to a material can result in permanent changes to its natural state: an overstretched spring never returns to the state it was in before being stretched. In Ulanowicz’s account of information in biological systems, autocatalytic processes and mutual information contribute to co-evolution of ideas and discourses which fundamentally change the direction of development of social entities. Values become encoded and communities which uphold them embraced as the means by which individuals maintain communicative habits. Values emerge as combining mutual information and flexibility: values can be both ‘ways of forgetting’ and ‘ways of embracing’ that which lies beyond them.

The emergentist account of values suggests that Berlin’s concept of ‘value pluralism’ is correct: values are at once fundamental to a community, since they have their roots in the ontogeny of each individual’s habits, and yet they are likely to be incompatible with one another. The balance between a value balanced with flexibility (the consideration of what does not fit a value) shapes choices of action and utterance, but also the diversity of value positions within an ecology of values. Here the difference between those value positions which exhibit little flexibility (‘hard line’ values) – for example, championing exclusive education for the elite – have the knock-on effect that they increase the coordinating powers of the communities that sustain them. Value positions which encompass greater flexibility can be a source of novelty in their interactions with less flexible positions but which will exhibit less cohesion in the communities espousing them.

When Berlin talks of value pluralism, and the inevitable conflict of values, this is itself dependent on those different histories of individuals and the emergent social conditions within which they operate. The principle question is to what extent a value is a rationalisation of irrational forces which have their origin in the cradle and of which we are barely aware. Reasoned value positions are a coping strategy: values are a meta-strategy. Reasoned positions autocatalyse their mutual information: when confronted, behaviour becomes unstable (even in the most mild-mannered of people). Because of its ontogeny, value depends on irrationality, and consequently the demand that agreement and consensus must be the way forward to the organisation of a rational society cannot be right. This is not to say that agreement and consensus are impossible – clearly something happens when people agree things – but rather the privileging of codified linguistic mechanisms of mutual information misses the autocatalytic mechanisms which feed the flexibility of discourses.

Teachers believe (on the whole) that the lives of others can be transformed through talking to them. However, this is not the same thing as assuming that the transformative agency of conversation is the mutual information which is shared between teachers and learners. If the information dynamics of Deacon, Ulanowicz and other is right, then what happens in ‘conversation’ between teachers and learners is the encounter with flexibilities of thought and  the transformation of values. What occurs is a gift of flexibility. This amounts to transformations not of the game of the conversation of the topic of conversation, but also changes to the metagame, whose impacts on the learner are considered by the teacher as a means of addressing the deeper needs for flexibility of thought. The coordination of consensus is a meta-rational calculation that takes account of the ontogeny of individual values and flexibility. The coordination of relationships has to account for irrationality and absence.

Education and the Form of the Social World

In critiquing the simplistic rational-choice theoretical model of the social world, I began by highlighting the importance of meta-rationality in the processes of decision-making. The question then arose as to how one meta-rational option might be preferred over another. In a world of non-ordinal outcomes, absent options can present a way of making distinctions between what is chosen and what is not. The absent has a corollary in the form of the relationship between information and ‘not information’, the relationship between figure and ground, and the relationship between messages, noise and redundancy. In drawing on Ulanowicz’s ideas of information ecology, the balance between what is and what is not can be presented as a balance between mutual information and flexibility, where flexibility is characterised as the ‘redundant pathways’ which are as yet unrealised. It turns out that these redundant pathways are of fundamental importance in the emergence of discourses and values. Their autocatalytic dynamics creates opportunities for creative adaptation and co-evolution with the environment. Yet autocatalytic events constrain the development of mutual information.

The tragedy of education today is that the forces that reinforce the talking about things that can be talked about are not just the naturally occurring forces within the institution; increasingly they are economic and political forces which have acquired new forms of manipulation. These are the brittle dynamics that result from too much mutual information and too little flexibility. Education’s tendency is towards greater mechanisation and measurement of what can be codified and what is seen to be common (mutual information). Yet, we see that the ‘not information’  - indeed, the irrational parts of education – play a more fundamental role in the creativity and diversity of economic and social development. Ulanowicz’s ‘flexibility’ measure, and maybe even Shannon’s ‘redundancy’ measure offer clues as to how we might go about measuring the flexibility within the ecology of education. Yet, as long as we merely focus on codified mutual information – which is the principle target of current efforts in data analytics – education seems doomed to cast itself in a plaster casing which is likely to break with the slightest environmental perturbation.

Conclusion

I began by challenging the idea of rationality in decision-making. However, I have indicated how our thinking about what amounts to ‘rational’ decision-making is unsophisticated and inaccurate. It demands a deeper causal model of the way decisions are made. The simplistic presentation of rational choice theory presents decisions as being made in the light of information. More accurately, decisions are made in the light of uncertainty, and uncertainty may be seen to comprise two components: mutual information and flexibility. Mutual information presents positions which we can evaluate; flexibility represents those positions which cannot be thought and so are not information, but necessarily are also causal in the thinking process. As far as metarationality is concerned, what is forgotten has a more powerful bearing on decisions that are made than what is thinkable.

The fact that this order of unthinkable or absent ‘not information’ may be measurable, and the extent to which the degree of this unthinkable ‘not information’ has a bearing on the capacity to develop new lines of thought presents the promise that a revolution in information theory might usher in a revolution in thinking about the relationship between educational processes and social structure.