The Challenge Network

   back   menu   next   

Social science, its prospects and its enemies.

Social science, its prospects and its enemies.


Understanding and getting the best from people, avoiding conflict and managing complexity are likely to be important elements of success in the decades of unprecedented change which lie ahead of us. The promise of social science is that it will bring insight into what is currently an art; and an art which is, for the most part, poorly performed.

Unhappily, the social sciences seem to lack bold new 'architectural' ideas. The fragmented landscape is, perhaps, best understood as oversupply of bakers of bricks rather than builders of houses. The many excellent results which are obtained from social science are seldom elaborated into structures from which new initiatives can be developed. Structural developments seem confined to where social science interacts with disciplines such as, for example, psychology, neurophysiology and information theory. The implications of even these developments are, however, profound, and the second part of this paper discusses where these may take us.

The first section of the paper is, however, given over to a discussion of why social science appears to have particular difficulties in suggesting bold schemes. In part, it is suggested that this is due to attacks from a style of critique to which the social sciences are particularly vulnerable. Ideas that have spread from Deconstructionism have been used to attack less the content of social science than the supposed cultural assumptions which underpins it.

However, whilst the methods of some of these critics are not tenable, the vulnerability of a body of knowledge that deals with largely intangible systems is nonetheless real. Social science deals with issues which are innately laden with values, and challenge to cherished values usually attracts attack. Social science also deals with complex and intangible systems, and this requires its assertive structures to rest on apparently frail foundations.

This paper argues that these are false vulnerabilities, and that social scientists need to take their courage more firmly in their grasp. The issues of social, political and economic-resource integration in the world of nine billion will demand sweeping social solutions to social problems.


The social sciences offer us a way to perceive some of the most significant issues that face us. They offer a set of tools with which to manipulate these issues. "Social technology" is used in every sphere of life, from management to political policy, dispute settlement to the conduct of daily life. Although the practical use of these insights has never been stronger, the academic core of social science – and in particular, economics – is under attack.

The attacks come from a number of sources. Some, launched by practical people, find the progressive mathematicisation of some aspects of economics both daunting and obscurantist, and the potential source of crises such as those recently affecting banking. Mathematics has its role to play, of course, but its use is only ever as good as the underlying model which it embodies. It can be argued that sophisticated symbol manipulation has, on occasions, serve to camouflage minimal or dubious content.

Technical virtuosity may also serve as both a union card and, coupled to opaque language aimed at a specialist audience, insulate the discipline from a more pervasive style of critique. This comes from those who see social science as in some way embodying and reinforcing the values of their least favourite –ism: capitalism, consumerism, liberalism.

Intricately wound into this last group are the post-modernist critics. This group 'deconstruct' what they call a 'text1' less for the ideas which it may contain than for the assumptions about power, status and other issues on which they feel strongly. Social 'texts' are thus dismissed or (rarely) applauded not for what they say, but for the supposed stance from which they speak. This is simultaneously a comfortable stance from which the adolescent and the half-intelligent can snipe at the core community, and a way of dismissing the views of those with whom you disagree.

Unhappily, such rot – in both senses of the word – has infected much of the academic communities of the US and continental Europe, from whence the infection grew. Formal deconstruction is rare, but the tropes that derive from it are near-universal: to ignore an argument and address the putative motives of those advancing it; to change the terms of reference to something that the discussion was never intended to address; or to reject unwelcome conclusions out of hand and without consideration as being elitist, patriarchal or somehow insensitive.

Clarity and its enemies: on destructive and constructive criticism.

Analysis in the social sciences is always positioned between two styles. One of these is prescriptive: in economics and politics, for example, it describes how people who fit with the writer's notion of virtue or rationality ought to behave; and seeks to explain deviations from this. The other is descriptive, in the sense of taking observations and attempting to find mechanisms that show how these observations arise.

The first of these, deeply rooted in religion and Romanticism, predominated in the Nineteenth century and, it has to be said, gave rise to some of the most damaging political movements that humanity has experienced. It had been on the back foot since the mid-Twentieth century but, from the mid-1970s, began to mount a counter-attack. Its insights are rooted in such philosophers as Nietzsche, Husserl and Heidegger, none of whom can be praised for their clarity of expression. However, a more approachable synthesis was set out by the French academic Jacques Derrida. This was intended as a critical method of 'deconstructing' literature, but expanded beyond this role when it was adopted by the radicals during the student turmoil of the 1960s: 'a bas la société d'ancien regime' and never mind the details.

Deconstruction argues that social science is a construct into which values have been built; and as such, it is biased in favour of the views of whomsoever the writer is set against: for example paternalists, capitalists, market liberals, social conservatives.

As we noted above, however, social science can be addressed in a normative sense, where the writer takes it as read that an issue – environmental sustainability, for example – is of overarching importance. The author then discusses social systems in terms of why they deviate from this prescription. This style can easily be married to elements of Deconstructionism. The author may criticise a position not for what it says, but for the supposed motives of the people who are saying it. Equally, the critique may assert that there are embedded and unacknowledged values which mean that the argument out of hand. The author may assert that it illegitimate even to discuss the topic because the terms of debate are themselves tainted. An example of this is the debate about race and intelligence; but it is not uncommon to hear 'conventional economics' airily dismissed as being devoid of social justice, or lacking environmental credentials.

The following is taken from an article written for a United Nations University seminar in 2006, by Ha-Joon Chang of Cambridge University.

Despite the miserable failures of radical policy experiments through various structural adjustment programmes in developing countries and the big-bang ‘transition’ programmes in the former communist countries, the orthodox economists have refused to draw the most obvious conclusion, namely, that the orthodox policies, and the theories underlying them, are flawed.

At first, they tried to argue that the policy reforms need to be more extensive in order to succeed. When that happened and the good results still did not materialise, they started saying that policy reforms need time to work. However, after 15, 20 years of reform, even this line of defence has become difficult to maintain. So, now the orthodox economists use institutions to explain why ‘good’ economic policies based on ‘correct’ economic theories have so consistently failed. By talking about deficient institutions, they can argue that their policies and theories were never wrong, and did not work only because the countries that implemented them did not have the right institutions for the ‘right’ policies to work. In other words, the institutional argument is being mobilised as a means to protect the core tenets of orthodox economics in the face of its inability to explain what is happening in the real world.

It is worth noting two things about this analysis. First, the class of "orthodox economists" are treated as a superceded elite which is fighting to retain its influence against new, unspecified sources of insight. This throws the critic in a comfortable light, of course, and allows any response from those under criticism to be dismissed as being special pleading. Lenin said much the same his opponents: by opposing the 'proletarian process', they declared themselves as class enemies. If they argue with you, then by that act they declare themselves as a part of the benighted class of old authority, and can be dismissed as such.

Second, it is taken as read that there is no merit in the orthodox analysis – the "excuse" – as to why some countries do not respond to otherwise sound economic policies. This is not a valid position. There is an enormous body of evidence for the adverse consequences of thieving bureaucracies, of arbitrary law and land tenure, of banditry and armed insurrections. That economic adjustment policies may also fail in such environments is unsurprising. Well-structured economic policies are, of course, a necessary but not sufficient condition for success in national development.

Development economics – in the sense of seeking an active, rational set of policy levers with which to initiate or accelerate the ability of groups or nations to add value – is perhaps seventy years old. This experience has led to an increasingly complex way of thinking about development. Our knowledge and body of experience has become greater and with this, so have ambitions for these interventions increased. Early interventions tended to take a managerial stance. From the 1980s onward, however, the emphasis has been to create functioning institutions – in education, finance, communications, law and energy supply – such that the natural engines2 of economic interchange can function effectively.

Industrialisation creates many material benefits, but it also disrupts the social status quo. Stimulated economic performance is often connected to increased environmental degradation, to urban sprawl and to increased – or to more evident - social inequality. These evils attract criticism of the economic model which created the stimulus, and are said to be "built into" the economic analysis.

It is worth remembering, however, that these were unforeseen and unintended consequences. They were not issues which the intervention package was asked to address. The rational response has been, of course, to extend the models yet further in order to capture these new variables. It is in this way that explanatory tools are developed to tackle problems, usually after the problems become evident and not before because for reasons which range from funding support to the natural tendency to solve the closest problems first.

Disciplines advance through constructive criticism. Most start with one or two small patches within which they can say useful things and grow out from these. At its origins, for example, economics was more concerned to explain how markets came about and did their work than to address issues of market failure, externalities to markets and more recent concerns. Constructive criticism has causes the young tree of knowledge to extend new branches into areas such as regulatory economics, which do deal with exactly these issues.

'Deconstructive' criticism is neither much intended nor likely to achieve these fruitful ends. It delights in subverting anything that appears conceptually stable, and will employ more or less dadaist critique in order to achieve this. The social sciences are extremely vulnerable to this approach, for reasons which are explored in the following section. The partial solution is discussed in the section which follows, which discusses the phenomenon of emergence.

Social science: different, but still a science.

Deconstructive critique has limited impact beyond the social sciences, which have a number of vulnerabilities which we have begun to explore. The physical sciences are able to simplify the structures that they are attempting to explore. They are able to repeat their experiment over and over.

This is not usually true of the social sciences, which also tend to operate in innately complex environments. A system which has many free dimensions will generate data that are embedded in a great deal of noise. It may be hard to pin down the key variables or we may find that there are a myriad of small forces which together achieve a large affect. Standards of proof cannot be as high as those in, for example, physics.

There are, however, much stronger lines of attack on the social sciences. The anchor points for debate within the social sciences are often intangible systems, such as markets and preferences. We cannot always access the prime elements of these systems directly. One cannot yet measure what is captured in the concept of 'consumer satisfaction' directly, only watch what consumers say or do.

We discussed the way 'truth' is assigned to a model through the network of supporting validity that surrounds it. Social science relies on a network of abstractions, indirectly supported by evidence. The networks of ideas that allow complex explanatory structures to be erected are consequently weaker and less resilient to critique than they are in the physical sciences. There is inevitable self-reference and circularity. It is this circularity that the more philosophical of the deconstructionists tend to attack3.

We need to dive into some deep epistemological waters to see why this vulnerability is also a strength, and that what social scientists seem to lack is the courage to assert structural hypotheses. There are arguably too many bakers of bricks and far too few builders of houses; let alone speculative architects in the field. Indeed, it is the fledgling architect on whom the deconstructionists most prey.

Philosophers have, of course, spent centuries discussing what can be said and what can not. The main thrust of philosophy in the Twentieth century was, perhaps, that there is an absolute division between the possibility which we have for direct understanding and what it is that exists. We can model what exists in ways which are extremely homologous to it, but these models remain distinct from it. The map is not the terrain, however finely drawn and at whatever scale.

Logical positivism held that the agents within a system interact only through observables, tokens of their presence. They are known to each other only by virtue of these tokens. Modern quantum mechanics is founded largely on these principles. A free electron is whatever an electron is, but it presents itself to its peers as distributions of vectors, mass-momentum, charge, spin and nothing more.

This view led to some intellectual despair (and to the seeds of one particular flavour of obscurantism.) It has also led to a revolution in how we think about complex – and specifically "emergent" - structures. I am going to begin with the issue of how we name and test intangible or ambiguous agencies, and then move on to the critical issue of emergence in a subsequent section.

Early philosophy held that there was a one-to-one correspondence between the language of description and the thing described. Logic came from parsing such statements: ravens are black, that bird is not black; ergo, that bird is not a raven. In a long and choppy history, this view sank in the early years of the Twentieth century under many guns, including those of the later Wittgenstein. The intellectual pessimism that this induced proved a fertile ground for the phenomenologists such as Heidegger, drawing on Hegel. Things are thrown (geworfenheit) into our experience without (possible) explanation, and we knit what meaning we can around this. This is the compost heap from which the deconstructionists were spawned.

This pessimism was, however, dispelled as much by discoveries in psychology as in philosophy. We do not perceive things in the manner of a camera, but as a transaction. We see a blob, and from details of it, the context in which we know it to exist and so forth we assign meaning to it. It is a face, a horse's face, the brown attentive face of my gelding George. Our brain has many "engines" arranged in arrays and layers which undertake this sequence of assembling qualia, those elemental percepts that are presented to us by these deeper engines.

The acceptance of this – and the host of reinterpretations which have been drawn for it was foreshadowed by Wittgenstein: "When we first begin to believe anything, what we believe is not a single proposition, it is a whole system of propositions. (Light dawns on the whole.)"

The later school around Quine said, in effect, that we evaluate the truth of an interpretation by its harmony with a web of insight that we have. If we are standing by a duck pond and hear a quack we interpret this in a quite different way from the same sound, heard in a store selling shooting goods. Formally, data are turned into information by context: "42" is a hotel room number, a chest size in inches or a salary in thousands of pounds (or the meaning of the universe, in Douglas Adams units), all depending on context.

This finding deserves more weight. The network of insight matters more than the atom of data. The role of the atom of data is interpreted probabilistically. Separate and parallel strands of evaluation lead to strong statements that are reinforced by existing experience: by knowledge about the world. The collective weight of data filtered through the contextual interpretation is stronger than any one datum, and it is a mistake to believe that 'beautiful theories die at the hand of a cruel fact'. Weak theories may die in this way, but robust insight requires many such blows before it fails. Few robust structures fail utterly, but their adherents learn, change, adapt to new insight and new data.

Philosophers such as Rorty have tried to lead this idea of the network off into a space of complete relativism, in which there is no such thing as absolute truth or validity, only consensus 'construction' of a way through which to describe observables. Post modernists have developed Derrida's ideas until they feel able to say (after Jean Baudrillard4) that the 1991 Gulf War was a 'text' in which media and event were indistinguishable, and a report of a bomb falling on a reported city was equivalent to the event in itself.

Telephone networks operate at (usually) eight formal layers of abstraction. That is, the bottom layer – the hardware and the switches – support progressively more and more abstract patterns of ordering until, on the uppermost layer, we have the traffic which is being carried. Sense data, too, can be seen as being processed in layers, both in the sense of a simultaneous hierarchy in the manner of the telephone system, and also sequentially, as "George my horse" emerges from sequential filtering and fusion.

Knowledge structures are equally distributed networks in the brain. For example, when we imagine and action – throwing a ball, for example – the many parts of the brain which would be involved in actually taking this action are excited. Less physical tasks are represented in much the same way. If we are asked to imagine something disgusting, then specific and highly circumscribed parts are excited. These parts, when directly stimulated with an electrode evoke an inappropriate sense of disgust. When suppressed, we are unable to experience disgust. If we are shown a picture of a dog covered in slime from a pond, both the parts of the brain that represent "dog" and "disgust" will become stimulated. The act of perception and experience are probably one and the same.

One rather remarkable insight that derives from this is the following. Imagine a system (a part of the brain, a piece of electronics) which has been trained to identify round things. Any thing that is put into its sense field causes it to emit a signal which is proportionate to the roundness or otherwise of the entity. Now, imagine another identical system that has been trained to do the same for the quality of redness. The two signals can be seen as vectors, perpendicular axes that together make a square on a plane.

This plane divides percepts into innate sets: things which are red and not round, round red things, those things which are red without being round, and dull objects that have neither quality. In other words, two simple detectors have split the visible universe into four classes. One can see this as the root of a kind of language, as the machinery by which to alter behaviour or as a 'toy' example of how perception probably occurs. And indeed, almost exactly this procedure, but using much higher dimensioned spaces than 'round' and 'red', appears to be a feature of the sensory hierarchies of our own brains.

Any knowledge that we have is, therefore, relational: it is valid to the degree to which it coheres with other knowledge that we may have. Networks in which we place great trust arise from repeated experience, as with our use of language, for example. We speak, people respond; it seems to work. How it works, however, is obscure; and how we 'know' things about abstract and intangible structures is not at all clear. We think about these at the peak of a hierarchy of processes and systems that we do not understand.

This is a limitation, but it does not mean that all knowledge is "socially constructed" or that it is otherwise an arbitrary social artefact. At worst, it warns us of the possibility of bias, inertia and resistance to change; and at best it tells us how to structure the enterprise of science. The more a model - way of thinking - is exposed to feedback, the faster it will evolve to track the source of that feedback. Indeed, information theory tells us that without feedback, learning cannot occur. Networks of insight are calibrated against experience, where the most telling form of experience is the controlled experiment.

The box, immediately above, discussed how round, red objects can be identified from all others by a pair of relatively simple system. This is far from a thought experiment: devices called neural networks are easily able to undertake this task, and are widely used in activities such as data mining, share trading and credit assessment. Related structures, called genetic algorithms, compete to pass their 'genes' on to the next generation. These are widely used in engineering applications. Both rely on error feedback in order to accomplish their tasks, the first by iteration, the second through competition across generations.

Knowledge is, then, represented in structures which we cannot describe very well, but which it is helpful to call 'models'. Two separate networks may draw on the same model, testing its validity. Well tested models contribute to strong networks; and well-established networks become models in their own right. On occasions, largely separate agglomerations of this sort interact and form new and useful patterns.

Social science is undergoing a quiet revolution in precisely this way. One might consider the recent impact on economics of neurophysiology, psychology, psychometrics and opinion polling, game theory, anthropology, sociobiology, statistical theory for examples of this. Experimental economics sets out to test fundamental assumptions about risk, the value of information – both directly and as embodied in reputation, tacit institutions and the like – and has succeeded in calibrating hypotheses such as utility curves, hedonic choice models and the like in real populations.

Such processes do not force a group of people to arrive at perfect models; but it does serve to drive change towards improvement in the match between theory and observation. Beautiful theories, event when slain by ugly facts, tend to reincarnate as the even more beautiful. What, however, calibrates the extremely assertive notions of a Rorty, a Baudrillard or a Derrida; and indeed, against what would they agree to be calibrated? There are no observables to which their models point, no means of assessing quality, for they assert that even a failed experiment is a text of no more or less value than anything else.


Earlier, I mentioned the concept of "emergence". This is an idea – an awkward datum - with which all sciences are having to come to terms. It presents formidable problems to many of them.

Systems can have more degrees of freedom than do the component parts that make them up. This is deeply mysterious, but is everywhere apparent: there is nothing in the individual component lepton and bosons – component parts - of a carbon atom that say "carbon", and if there was, then they would be unable to say "silicon" or "sulphur". There is nothing in the carbon atom that says "dragonfly". Yet dragonflies are supposedly made of leptons and bosons, so where to the additional degrees of freedom – the observed state space – come from? We say that they are emergent, which is name a mystery but not to capture it.

Dimensionality and degrees of freedom

The state of any system can be represented with a set of orthogonal (independent) dimensions. For example, a cup of coffee can be hot or cold and, independent of this, it can be strong or weak, sugared or plain; and so forth. That specific cup, the one sitting there, can be represented – modelled - as a point in a space made up of these dimensions. Using the three dimensions just described, this space can be visualised as a cube. The cooling cup traces a line across the interior of this, with sweetness and strength remaining constant but the temperature falling.

However, suppose that we were testing an idea that people who drank strong coffee tended to use more sugar than others, and we found that this was supported by the data. That would mean that sugar use and coffee strength were no longer independent variables. These two dimensions would have to be replaced by a mathematical construct that combined this interaction. This would combine the two measures observables, the dimensions of sweetness and strength with a mathematical construct.

Systems often generate precisely these interactions, which serve as ways (degrees) in which the system is free: degrees of freedom. The state of the system as a whole at any one instant is then represented as a dot (a locus) in the space made up of these degrees of freedom, a space which is usually called the state space of the system. Over time, the locus within the state space of the system may vary, thereby tracing a line.

Most state spaces are not fully explored by events: that is, the locus and its trajectories do not enter some parts of the space and are very much concentrated in others. Volumes which repel the locus are called repellors, those which attract are called attractors.

A the locus which represents the instantaneous state of a system such as an consumer market, a colony of insects or an oil refinery may settle at an attractor, or may orbit around it, showing quasi-regular behaviour. In a space with several attractors, the locus may park near or orbit around one of them for a time, and then flip to the volume of influence of another. This can lead to sharp changes in systems behaviour – from recession to boom, perhaps, or from plant dormancy to growth – doing so a result of small perturbations.

One can, of course, see a colony of insects (for example) as a single point in state space, or grant a locus to each member of the colony. The workings of the colony are then described as a dynamic engine, with these individual points interacting through their "orbital dynamics". Indeed, one can show that many of the attractors and repellors of the system arise from – and only from – these dynamics. Markets – whereby simulated ants trade 'sugar' for 'spice', for example – arise spontaneously and show properties which are closely similar to those observed in human affairs.

The state space can also be assessed, point by point, for various characteristics. One approach, which was first used in biology but which has now spread widely, it the 'fitness surface'. Here, the point-by-point value of the state space is assessed for some observed and consequent quality – usually called 'fitness' - that such positioning gives to an organism, advertising message, engineering specification or company. (Often, such a measure relates to the survival or more qualitative long term state of the system, often with regard to competition, evolution or relative success.) The fitness surface has peaks and troughs which representing the quality emitted by the system at that point.

This approach is widely used in fields as separate as engineering and evolution theory, security systems and data mining. Systems which can learn and which have 'values', things which they strive to maximise or minimise, evolve repertoires of behaviour which optimise their relationship to the fitness surface: they are said to "hill climb" up it. Telephone systems have been taught to optimise their network usage in much this way. Darwinian evolution can be considered a paradigm of this underlying dynamic.

Unhappily for human social systems, it is a not always the case that the fitness surface maps onto the attractor-repellor dynamics that we have already discussed. That is, systems have pathologies built into their dynamics. Their 'natural' behaviour is not the same thing as behaviour which maximises their fitness. A significant aspect of social science is concerned with mapping behaviour. Other areas are concerned with pathologies of this sort where, for example, aggression negates the potential for collaboration. Yet more work looks at optimisation: at mapping the fitness surface. There is relatively little work under way that is explicit in matching these three things together, a possible area of fruitful development.

Consider a trout. Imagine that we have a near perfect model of one example of this species. This model, with minor modifications, will also exactly describe the behaviour and being of another trout. Let us use this model to generate a shoal of simulated trout. These begin to interact and emit trout social behaviour: they school, they evolve stable patterns of what we then term social interaction.

These behaviours are emergent. None of then are built into the elementary trout model. This describes just one fish. It says nothing of social behaviour, only tropisms specific to the individual trout. Our model of the trout is thus 'brokene' from within. The emergent system has more degrees of freedom than the individual component of it.

The concept of emergence is deeply significant for social science, because systems, and the emergent properties of those system, are precisely what it sets out to study. However, that fact of emergence prevents5 the disciplines which make it up from growing "fundamental" roots. Instead, they must rely on self-validating networks of insight, and it is these networks that are growing in density and interconnectivity as a theory matures.

A simple rustic market for tomatoes is, therefore, an emergent system that is describable only in terms of what it does, of its observables.

High in the Andean hills, Maria Antezana packs her crop and heads for the village. Jaime Abusado jingles cash in his pocket and dreams of an evening lomo saltado: beef strips with braised potatoes and tomatoes tossed amongst them. A large number of other potential traders set out with similar ambitions. In the course of a few minutes, tomatoes acquire a price which is valid across the entire community. However, there is no Platonic price, no Father figure who is setting the rules; only some tacit institutions that allow tomato buyers to find their sellers, and sufficient security that allows vendors to proceed without fear of theft. A set of interactions spring into existence when traders meet – a system of rational 'game' responses to the situation is triggered – that is calibrated against what is going on elsewhere.

The structure self-assembles, in the manner of a shoal of trout. We can say how this happens and what forces drive it to this or that configuration, but these descriptions are predicated on the emergent that they are trying to describe, not the isolated component parts that make up the system. Indeed, the system has deeper ramifications, again established without design. High prices encourage more tomato supplies in future, and low prices usually cause consumers to eat more of the fruit, taking up additional supply. Homeostasis emerges as a property of the system. In the jargon of systems theory, it shows "negative feedback", and pursues stable equilibria. The same mathematics which describe homeostasis also explain why the system can be volatile, leading to cycles of over and under supply, show sudden extreme behaviour in the face of scarcity and other properties of the fitness surface of the phase space, concepts that are described in the footnote on page 7.

Models of emergent systems are extremely vulnerable to destructive criticism. To give an absurd example, and extreme reductionist could insist that music is "nothing but" the rarefaction of compression of molecules in the air. Which, of course, it is; but it also encodes information that does interesting things to at least some human brains, and that is a far more complex thing to explain. The reduction is valid but not useful, as the model to which it reduces matters has no points of contact with the thing being described. When we are certain of the validity of music, we call this absurd. When we are less certain, and the frontiers of enquiry are always uncertain, then the debate may be one-sided and to the detriment of enquiry.

This is where interpretation of meaning as described in the preceding section becomes so important. The emphasis is placed on calibrated judgement, rather than absolute proof. The grounds for such judgement lie in the overall insight that we have as to the working of the system under discussion and the calibration is based on the weight of evidence that surrounds each of these. Where evidence is weak, we can get more. Where the issue under discussion affects only a part of the system then we can dissect out this element and address it within the context of the overall system. There are weaknesses – of scholastic isolation of denial – but these are risks of which we can make ourselves aware through the well-established machinery of critique and enquiry that characterise open societies.

Such a network of insight can be addressed from many angles. The question: "Why is that flower red?" can have many answers, each in their own way entirely correct. It is red because of biochemistry and evolution, through the workings of local social customs, as best understood as the consequence of the horticultural industry or the attempt of the restaurateur to position his or her brand. Context tells us which of the many (correct) answers to the question is germane to the issue at hand.

This, too, is vulnerable to distortion by fashionable causes, which have a way of being intruded into such debate. The answer that we get when we demand that the answer to the question about the redness of the flower must include reference to gender, or sustainability, or inclusion or some deism or other will be a twisted, often inappropriate thing.

An architectural renaissance: What's new?

The "demand side" call on social science could not be stronger. The very core of the knowledge economy revolves around the effective use of people and what they know. The acceleration institutional evolution which is demanded by of global integration and development is architecturally dependent on the social sciences. The role of intangible systems in international commerce – in creating financial stability and mutual confidence, in supporting intellectual property and collective identity – present formidable and mounting challenges. Non-coercive mechanisms by which to create mass movements that help people to help themselves – in health, training, in coping with old age – all underpin many billions in expenditure and the lives of hundreds of millions of people.

The "supply side" of the social and allied sciences are also advancing very rapidly. Butz and Torrey published a review (Science 312 pp.1898 - 1900 2006) in which they tried to assess the new horizons for social science. They suggest six factors which are likely to drive change: access to huge cross sectional and longitudinal data, the growth in laboratory experimentation, improved statistical methods, the use of geographic information tools, the growth of interdisciplinary "biosocial" science and the fact of inter-cultural and international replication of social science work.

These are, of course, tools and not goals. However, they do offer substantial potential. I want to emphasise three aspects of this in order to support what follows.

First, the mass collection of high quality data is virtually omnipresent in the industrial world. Analogies have been drawn with astronomy, which has moved from scrutiny of individual suns by single astronomers to automated systems in which huge teams comb the data from hundreds of millions of observations.

Enormous sums are spent on commercial social science – in financial services and trading, economic forecasting, advertising, polling, generating customer typologies and studying how goods and services are used. The relatively sweeping ideas employed by the private sector are surprisingly slow to penetrate into the academic world, which also generally fails to access the immense volumes of data now available.

Second, there is an elision between some of the many wings of biology and various aspects of social science. Those most directly relevant ask how it is that we are able to think, what influences that thinking, why we have preferences and what sets these, how we identify friends, allocate trust and flag enemies. The notion of quantitative rationality sets many aspects of economics on an experimental footing.

Other disciplines are also beginning to impinge; for example, in the use of mathematical tools to examine network topologies and flows, the dynamics of systems and systems optimisation, the diffusion of information and epizootic events such as epidemics.

Third, access to huge data sets has changed significantly the approach which is taken to model building. Classical modelling erects hypotheses, and the data are processed in ways which tries to identify support for this. An alternative that is well-suited to vast data sets treats these data as an abstract mathematical manifold. It looks for structure within this, without any prejudice as to what may be found. The principle components are extracted, and it is only after these are shown to have intrinsic mathematical validity that the experimenter tries to identify which they may mean.

Massive data sets do indeed show such structure. In the financial services environment, the dimensionality is mathematical, abstract and not at all open to being reduced to human terms. However, this dimensionality may describe a state space6 in which the various points representing traders, securities, combinations of assets orbit in orderly ways.

Social science goals seldom head the many research road maps that foresighting activities deliver. The people who are drawing these up may have a clearer idea of what investment into, for example, 'new materials' will deliver, at least as compared to potential social science goals.

It can be argued that the social sciences are poor at explaining what they might be able to do when considered as tools, as a technology. Indeed, social science often obfuscates its potential. Symptoms of this include the extreme mathematicisation of some aspects of economics, and the arcane post-modernist references which scatter subjects such as anthropology. Whatever the motive behind this, it segregates the academics from the much larger body of commercial practitioners and lessens the ability of the academics to bid for funds against the other sciences.

This said, let us look at what may be possible if the resources are made available to develop it and, in some instances, if ethical qualms can be laid to rest. Deconstructive criticism is swift to accuse social science of ethical short-comings, In addition to these vapourware concerns, there are practical issues that emerge from the use of large databases, and from the potential that social science has to change lives. It has even been suggested that datasets that have the potential to erode individual privacy should be accessed only in specific information-secure facilities, by analogy with biohazard facilities!

Areas of extraordinary potential change.

There are, therefore, some remarkable tool and data in play which will create new engines of understanding. Tools are not goals, however, and the remainder of this section is concerned with something much grander. It is to be read whilst clad in the armour against deconstructive critique which has been supplied, one hopes, by the first major section in this paper. That said, this section is unashamedly assertive, and hopes to stimulate thought rather operates under any tenuous dream of being definitive.

The topics which I am going to address are as follows.

First, it is becoming increasingly clear that our values are both learned as a part of a social process but, to a very great extent, hard wired into us. That is, what disgusts us may be culturally-determined, but the fact of disgust is not. It transpires that many inter-personal values are predicated on such 'hardware'. The implication is that dissonance in values is innate to the individual and to the social group. Negotiation amongst these drives is inalienable. The "social narrative" as a body of, in effect, case law that determines how a culture will deal with such dissonance is therefore important. However, consensus narratives are under threat from external influences, fast change and political populism.

Second, deconstruction has suggested that the implication of value-laden debate is that we can never have an absolute view on what constitutes a fine or weakened social structure. The argument that is advanced in this section disagrees with this. It achieves this by taking a 'systems view' of what is necessary for a group to understand and react to their operating environment, and to learn from doing so. The required structures may be absent, may fail in characteristic ways and all of these generate sub-optimal adaptive systems. Additional characteristics of adaptive systems are explored.

Equally, given the importance of the social narrative, and the fact of both individual diversity and plural value systems, the essential role of arbitrage and information exchange are plainly essential for a functioning, integrated social system. The conclusion is that whilst we cannot design perfect systems, we can surely identify degrees of sub-optimality. It is fairly straightforward to pick potential winners – and certainly, sure losers – in the likely operating environment in which social systems will collaborate and compete.

Third, every life create a mass of information. Every transaction does the same. This information allows typing of individuals, assets, companies in ways that have not been possible in the past. This creates a number of possibilities. It will become economic to trade assets that have not hitherto been accessible to mass markets: the "E-bay" effect, but applied to the great bulk of capital which is not currently marketised. It will allow intervention in individual lives in many new ways, One example has been called solicited advertising, providing information which you do not yet know that you want that you want. This has been called "psychic Google", and is the target for enormous investment in both the consumer and business markets. Another is the micro-management of careers, such that individuals are mentored and developed at all stages and aspects in their life. Far more pervasive potential exists, however, and this is discussed in this section.

Fourth, a brief section which suggest that as we come closer to defining excellence in social systems – companies, state organisations, families – so we will face a world that drives each of these to what might be called "customised normative" solutions. In the medical world, for example, it is becoming possible to identify predispositions to ailments well ahead of their development. Children who are very likely to develop diabetes as adults can, for example, be identified with some success when very young and, in the decades ahead, with near-certainty. A normative solution would be to apply policy levers to the child's circumstances – to the parents, for example – that lessened exposure to predisposing factors. The "customised" aspect of this flows from the issues just described, those of targeted intervention.

Values, motives and goals.

The second major implication for the future lies is consequential on our understanding of the roots of motives, abilities and cognitive processes. These impinge on extremely deep social issues. Here are some examples:

There are, of course, many more of these; and new knowledge and technology will generate more. Examples of these might include, for example, collective structures – corporations, military bodies, interest groups – which are supported by IT that takes on an independent role in modulating communications, processes and initiatives. The IT would take on many of the roles of an ideal strategic management, of support infrastructure such as finance and HR and would mentor and model the progress of individual careers and projects.

Individuals and teams would operate in a totally distinct manner, passing everything that can be specified or defined to automation, lightly supported by human staff, whilst the human workforce concentrated on the unspecifiable: new ideas, better contact with stakeholders, the endless pursuit of clarity. Indeed, the nexus around which activity was ordered would then focus on the new limiting factor – the capacity to organise amidst complexity – rather than the current strengths of access to capital, brand or scale. The boundaries of such an organisation and the nature of participation in it would look less like a conventional corporation and more like a voluntary body. Universal transparency of action and process will anyway make remuneration dependent on direct contribution to whatever goals the organisation has set for itself.

We justify a course of action because it avoids something bad, or approaches something desirable. When questioned about the nature of "good" or "bad", we may go through a long discussion or point to something immediate, but ultimately we ground our choices in the desirable and the undesirable, which have these qualities entirely because we feel that we do.

There are a class of values which we have no difficulty in seeing as essentially physiological: we are cold, we want to get warm; we see the hungry wolf pack and we feel fear. We are happy to build models that predict behaviour based upon these stimuli. Experiment and experience tell us that these models tend to work as good predictors of our behaviour. Hungry people eat more than satiated ones. Food that tastes good usually commands a higher price than food that does not.

There is another class of values which we are less happy to see as purely physiological. Some of these are concerned with interpersonal conduct or with our relationship with the collective identity or with sources of power. Yet others relate to the numinous: to religion, aesthetics or to the natural world. However, it is in precisely this area that rapid process is being made in mapping the nature and neurological basis of such values.

There is strong evidence for a physiological origin for many of our deepest values. Studies of interpersonal values have developed strongly as a result of work that has been conducted on two fronts. What has been called "experimental economics" has found it possible to identify regularities that underpin abstract ideas such as fairness. Arriving somewhat later on the scene, the neurophysiologists have used the techniques that were developed by the economists to trigger the same arousal in people whose heads are accessible to FMR and related scanners. It has been possible to identify the small and discrete areas of neural tissue that support the processing which is going on when people confront issues of equity.

An example may help to make this concrete. One of the best-studied areas in experimental economics explores perceptions of fairness. For example, the so-called "Ultimate Game" has been tested in at least thirty cultures, across social classes and educational attainment. The outcome remains constant. In this process, the experimenter offers one of the two players a sum of money. That player divides the money into some which he or she retains, and offers the rest to the other player. If that person accepts, then both keep the results of the division. If the division is rejected, however, then neither keep any money.

The rational choice for the first player is to offer the other a minimal sum: after all, something is better than nothing and they should be happy to accept. In fact, offers of under a third of the total is almost always rejected. The second player "fines" the first for being unfair, at considerable personal cost. It is this decision which is handled by the neural tissue that was discussed above.

There are many related games which measure, for example, the foundations of trust (how much information is needed to maintain trust?) and the related question of the (literal) value of reputation. What characterises these findings is that in addition to the solipsistic and personal tropisms – that is, to physical comfort and satiation, towards a predictable social and physical environment, to personal status and power – we have to add a series of hard wired "social" emotions.

Research shows that there seem to be at least five deep value systems7 that govern these interpersonal emotions. Naturally, it is these emotions which in turn govern how we organise our societies and our politics.

These five value clusters are:

Contemporary Western societies tend to base public policy with reference only to fairness and the risk-reward dimensions. The other three are much less represented or, where they are present, are actually seen as negatives: one should be for the whole society, not the group; for abstract truth and not for arbitrary authority; against 'prejudice' such that no group can be more or less 'pure' than any other and no behaviour that does not generate absolute harm ranked over any other.

This is very far from the case in the nations where 70% of the world's populations live. For example, group affiliation is seen as a major moral requirement in many tribal cultures, as it was in nationalism. If one does not put the nation, tribe, family ahead of the general mass of the population, one is derelict. If one does not promote one's relatives, one is immoral.

These societies also tend to extend unqualified respect for authority, such that established sources of guidance are to be respected, even if they are obviously defective. Western mores delight in iconoclasm, by contrast, and any source of authority will be satirised, debunked, investigated until only the thin truths that are ultimately distilled from it are valued, not the source of these.

Billions around the world disagree with this, profoundly. They also disagree with the Western tendency to dismiss their respect for purity. A person's worth is measured in part by the purity that they have attained, by pure associations and purgative acts, and by avoidance of sources of the impure. There is also a strong sense of innate impurity, something that has to be purged by religious rights, charitable acts, abstinence and the like. Catholic Europe retains a dilute form of the strongly similar views which it once held.

People are being brought into contact as never before, and stirred into a homogenising pot that is strongly favoured to Western tastes. It is not surprising that this engenders resistance. As the connections intensify – and as friction also increases – we will need insight into how to manage this integration.

This brings us to another major social science theme, which is that of the group narrative. Why do societies differ in their espoused values? Is this because the people who make them up have innately different drives, or because they have dropped into one of the many attractors in their value state space?

Psychometric analysis shows us that people differ considerably in the degree to which individual value dimensions are expressed, both in their physiology and in their personal outlook. Nevertheless, there is no trend to this variation across nations. The variation is greater within social groups than between them.

Why is it, then, that groups amplify or repress one set of values more strongly in public debate and private conduct? Why is the West different from the developing world, and why is Sweden different from, let us say, California in its public pronouncements, actions and implicit choices? Is this an emergent accident of history or something different? More to the point, however, how do groups come to have common values and a common outlook, such that an agreed set of goals and procedures get set into motion. Why was Sparta so different from Athens, in rhetoric, action and aspiration?

The causative mechanisms are still obscure, but sociologists have at least acquired a name for the phenomenon: the group narrative. Narratives are the stories which groups tell themselves about who they are, how pay-offs should be calibrated and how an individual of this or that situation within the group should respond to a particular kind of challenge. "This is the code of the Warriors, and that of the Peasant: you, as a middle aged working woman should respond – bargain, marry, react to rivalry - as follows."

Political movements are the embodiment of a social narrative, but much-simplified, less reliant on tacit rules and altogether cruder than the 'real thing'. Societies which have a strong central narrative achieve great works when the narrative matches the demands of the operating environment, but make can make terrible mistakes when this alters, or when they take the narrative beyond its bounds of validity. Societies which have more than one strong – mutually incompatible - narrative have to negotiate a variable outcome at almost every level, or fight until one narrative has quashed the other.

Such narratives are found even in animal groups, whereby groups of chimpanzees will behave consistently and coherently – but differently – in response to similar challenges from geographically remote peers. They may structure their dominance systems differently: in other words, they have customs that are transmitted socially, occupying different locations in an underlying space.

Towards some absolute standards of fitness and resilience

We have already discussed the idea of the fitness surface. It is a generalisation of the entirely obvious point, which is that choices affect our fitness. However, as we shall see, digging into this takes us into some interesting and values-related terrain.

If we are faced with a choice, therefore, then exercising this moves the locus that represents our current position this way or that, doing so in a space that describes our situation. That is, an individual, an organisation or a nation can be seen as a point or a distribution – a blob - within a space. The dimensions of the space are set by the degrees of freedom of the system that is describes. One troop of primates is located "here" in social space, another over there. The totality of the space in which we can move is called the "state space" of the system in which we exist. When we make a choice, we move in this state space.

However, the fitness surface is pasted overt the state space. That is, some movements – some choices, perhaps, or some accidental changes – take us to less or more fit destinations. When we think about a choice, therefore, we are thinking about two things: possible movement ("What can I do?") and possible outcomes. The options are defined by our perception of the state space. The outcomes are defined by our perception of the fitness surface.

How do we see the fitness surface? In some cases – the raging fire, the circling wolf – we can see it very clearly. In others, the view is less clear and we tend to extrapolate, replay past experience and – above all – call on the narrative. This leads to conflict, as when the Warrior plunges into the fire to rescue the Helpless Maiden because that is his role in life, as defined by the social narrative. (Please feel free to exchange gender roles as appeals to you, the reader.) The narrative is then good for Maidens and less so for Warriors.

That is not to say that all narratives are arbitrary. Rather, the dominant narratives have been through an extremely intensive evolutionary process, in which unfit narratives have been eliminated. (Cults that invent their own narratives have a high mortality rate. A few of these have, of course, found an extremely fit area in social space within which to propagate, changing and adapting as the social space itself alters.)

One "deep" question for social science is whether such fitness surfaces have a single optimum that holds for all societies everywhere. Alternative views would say that there is a broad optimum but that this changes with the state of development (in today's world) and has changed over history as human capabilities changed. Yet another view would be that there are indeed optima, but that the surface has many of these, and that no one is fitter than another. And a deconstructive view would deny that it was even possible to discuss the issue.

This question is likely to be answered in the next few decades. If we were able to generate such an answer, we would have a model of development that would serve as a benchmark to every country: at this stage, you should expect this from the law and that from politics, your value systems will be predicated on the following likely balance. All of this would be predicated on a migratory path that societies would recapitulate through the changing hills and valleys of the fitness surface overlaying the equally shifting model that represented practical choice.

We can even grasp for the model that would underpin the industrial economies. Societies which embody many varying sub-narratives – nevertheless held together by habit, common identity, a central body of shared assumptions – seem best able to adapt to change. Somewhere in them, constant churning will mean that an experiment will have been carried out that blossoms into the new environment: other learn, and the way of thinking and acting spreads with minimal resistance to it. Monocultures are efficient when everything suits their requirements; but a many-niche ecology is more robust when the environment is subject to shocks and changes.

Complex societies need complex forms of brokerage, in which incommensurable values are paid off against each other in extremely complex ways. Resources are directed where the emerging consensus deems them best allocated. The tools for delivering this range from simple price and market mechanisms to micro-politics, from social networks and domain-spanning journalism to non-violent coercion, from tort and litigation to statute law and extra-legal regulatory mechanisms. All of this is grounded on the social narrative, and that narrative itself evolves in adaptive societies and is armoured, rejectionist or worse in societies with do not accept that conditions alter. The primary agencies of this act at many levels and respond to many masters. Some, but by no means all or even a majority of this is conducted at national level.

The fastest adapters – perhaps due to their stripped down objective function, perhaps due to the habit of analytical thought directed at competitive positioning – is commerce. Society at large trails this pace setter, often descending into primitive responses – A strong leader! Find the malefactors! – when things go wrong. Public institutions lag both of these by a considerable margin. This imbalance cannot continue, of course, as the pace of change is set to increase. At issue is, then, the way in which public institutions could become more agile and "aware" of their operating environment.

The "deep" question which I posed was whether social science could expect to point to a universal set of quality parameters for the robustness of social groups. It seems likely that at least one dimension of this revolves around their ability to navigate, to use information. What values they deploy – their goals, their choices in the light of this insight – is almost secondary to the fact that they are able to perceive the structure in which they are embedded.

This approach takes us to an interesting hierarchy. This is expressed in deliberately abstract terms because, as we will see, each of these gives rise to characteristic pathologies which makes the society less fit, in the sense of less goal seeking, error correcting or sensitive to stimulus and opportunity.

A society which has all of these mechanisms in place will be more likely to get what it wants than one which does not. In essence, this is the minimal suite of behaviours which allows for purposeful adaptability. If the group has three of the four in place, then it will show characteristic pathologies, of which the following is a sketch:

One can, of course, ask what tools (democracy, open enquiry and scholarship, market liberalism…) predispose a society or organisation to solve or avoid particular pathologies, but these are in a sense secondary issues to the basic analysis. A fit society is one which can navigate in a purposeful manner, and which makes the most use of the information that is available to it. In other words, a systems view takes one to testable hypotheses about "good" and "bad" social systems. Much in the spirit of Popper's falsification principle7, we cannot define an absolute state of excellence for a society or for the governance of a company, but we can certainly point to systems failure, where the system in question is the ability to act purposefully and coherently.

A second self-evident quality parameter is the ability to act coherently – as defined above – but whilst also retaining flexibility and resilience. Companies or societies which are equipped with seemingly perfect information and equally excellent insight can, nonetheless, walk straight off cliffs which they could not have foreseen. Those which survive have the quality of resilience.

Resilience has many roots, major aspects of which are identical with the issues that we have just discussed: insight, purpose, tools. However, it also consists of having a portfolio of activities which are not perfectly aligned, of viewpoints which are somewhat skew to the consensus, of experiments that are warehoused against changing circumstance. The ecology of the soil is an example of this. A gram of fertile soil will contain something in the order of a billion micro-organisms, made up of tens of thousands of species amongst which only a few tens will dominate the biomass. As soil conditions change, however – through temperature, plant roots, rain, an influx of nutrients – some the hitherto minority species bloom and the formerly dominant ones die away. Nevertheless, the state of the soil – the biomass it supports, the metabolism that is hosts – remains the same.

Homeostasis requires both integrity from the information processing system and also the ability to alter the component parts of it as circumstances demand. If our model is 'broken' by events, we need an alternative way of seeing quickly. If one is sitting on a shelf – as one of the lesser micro-organisms is waiting its moment – then so much the better. If it is not, then the society needs to be able to recognise the gap, and set work in train to fill it. That is, the old model must not be defended when it becomes indefensible, and new models must be adopted on merit, with due caution, under provisional approval. This capacity to experiment purposefully, to learn how to see and to interpret, is a fundamental quality parameter which social groups display or do not. Once again, testable hypotheses present themselves as to an absolute measure of merit.

Micro-focus and mass intervention.

Well over 80% of the trades in the US stock markets address the assets of firms that hold less than 15% of American private capital. It is, however, almost as much effort to analyse a small company which is not widely traded and the standard Fortune 500 organisations. However, micro-analysis, in which increasingly automated systems match organisations to models – in the sense used earlier, rather than necessarily numerical or financial models – has the potential to bring this immense pool of assets into the mainstream. Much the same is true of emergent economies which reach the required level of reporting transparency and probity.

Individual UK supermarkets hold terabytes of information on the purchasing habits of individual consumers. Predictive software is able to estimate very closely what an individual will consume, given as little information as their approximate age, gender, martial status and post code. More detailed information allows exquisite data mining, noting and identifying the source of changes in individual behaviour due to, for example, financial stress, job change, bereavement or pending child birth. Whilst this is not yet done, such knowledge can trigger offers of assistance that are exactly appropriate to the individual. More broadly, pro-active information supply can identify what we need to know in a company, as a consumer, a pupil or a person developing a career, and supply this to us before we even know that we need it.

Similar potential exists in the medical field – where latent disease can be identified for prevention or palliation, for example – or in a host of areas where better models and more effective tools will help individuals or groups to identify their options. Extremely complex and messy structures can be rendered tractable in a high enough dimensioned space, and we are, increasingly, acquiring enough information to define that space.

These tools have great potential for good, and equally enormous capacity for intrusion and active harm. Saddam Hussein and the Stasi of East Germany both attempted to "know everything about everybody". They failed, not through under-resourced effort but because the capability had not yet been put in place. A truly effective dictatorship would operate with such insight into individuals that the carrot would always outweigh the stick, and the population would do what the elite wished because they wanted to, because it offered real advantages. Indeed, the only criticism against such a system would be that it sought an overall optimum that favoured the elite by offering local optima to each individual.

Understanding this potential for misuse, we should not shy away from the positive aspects of such a technology. Consider the child who has a permanent companion, perhaps embodied as a book, as in Neal Stephenson's The Diamond Age, A Young Lady's Illustrated Primer (Bantam 1995). This artificially-intelligent artefact explains events and people around them, and helps them explore and learn at their own pace, adjusting lessons for their relevance to their immediate context. It finds friends and diversions, jobs to do and people to meet. It stands guard, and has the judgement to know when to admonish, when to analyse quarrels and events that could have gone better, and when to call for help.

Extend this to the adult, who has much the same problems but in a broader context. This is the career advisor and relationship counsellor, communications filter and connection-maker; it is what every mobile telephone wants to be when it dies and goes to silicon heaven. Extend this to the company, drawing at least partially on this infrastructure. As discussed earlier, artificial intelligence and a deep understanding of how groups of people create value allows the system to arrange connections, start human processes running, guide and manage threads of enquiry, development, implementation and routine delivery. The collectivity becomes, in one sense, an integrated organism, debating values, understanding context, exploring the fitness surface, gaining resilience and capability.

Deeply intractable social situations may occur because the sides are monolithic and immovable. This is truly rare, however, and most opposing structures are highly plural. The way of expressing and tailoring a compromise that would fit this person will not be right for that one. Vastly plural and micro-focused communications can, therefore, arrive at a market-like solution which monobloc confrontation cannot.

Equally, if zero-sum politics consists of trading a loss here for an advantage there, then machinery that does this at a very local or individual level will allow intransigence to be bypassed. Consider two polarised communities that have to occupy the same geography, a not-uncommon phenomenon in our crowded world. There are three ways to settle such a situation: to allow one side to crush the other and re-write the narrative; to wait until the communities grow out of their antagonism; or to broker a myriad of micro-accommodation.

As with the discussion of the Stasi and Saddam above, a suitable package for everyone would defuse – or anyway, diffuse – the situation. Military planners are much-focused on enforced recovery in failed states and the stabilisation and restructuring of post-invasion societies. As one senior figure put it in a recent speech, '"Sergeant, install democracy" is not an order that you can give as a commander, but is an order which you must expect to receive.' The relevant tool kit may well consist of a micro-focused understanding of the society and its power structures, its economic flows and dependencies, its formal and tacit institutions. Its narratives and habits of thought, all distilled to the individual level such that many, many small interventions can be made to steer individual and community behaviour. This is a deeply dangerous technology, for the reasons rehearsed above, but one that will become possible as investment is made in it.

A normative, expert world

All of this offers extraordinary implications. The first and most obvious of these is that as social science becomes predictive, and able to make unambiguous statements about best practice, cost benefit and the like, so the implication for policy becomes increasingly coercive. If we know the "best" way to educate a specific individual, then can there be any justification for following any other course? If certain behaviour will ultimately lead to an individual becoming a burden on the state, then should the state act to curtail that behaviour. (One can see echoes of this in the current debate on obesity.)

One can easily imagine a world in which there is only one way to manage a company, develop a career, raise a child once one knows the details of that company, career or child. Even if states do not act in the name of competitiveness, resource efficiency, proxy parental care it is certain that scrutineers and stakeholders will do so. If a company can be proved to be badly managed on objective evidence, how long will shareholders allow such a situation to continue? If a central bank is not managing monetary policy to the prescription of provable best practice, how long will its currency retain its value?

Some closing thoughts

Pressures on government to act where action is possible and direction is established have proven irresistible. The state permeates every aspect of the life of a young person growing up in the industrial world. All of the trends and tools that have been described above will accentuate this trend. The breakneck pace of competition, the need to maintain national competitiveness and the resource pressures on government add to this. In addition, aging and demographics, the implications for public health of the new biology, the potential of semi-autonomous information technology systems all add to the tools of the state and the implication that interventions should become ever-more prescriptive.

The implications of such normative systems has already been discussed. However, one general characteristic of social systems is that any seemingly irresistible trend will eventually find a counterforce. On what grounds, and on whose authority, are such limits to be established?

Nations which are optimising are, of course, 'hill climbing' on a fitness surface. We have seen that the concept of fitness is built on broadly constant values, rebalanced by the local social narrative. The implication is that the fitness surfaces are probably broadly similar at any given resource level. That said, it is extremely unlikely that this universal system has only one peak, only one optimum. It is more likely to be characterised by an area of broad uplands, undulating with a range of peaks of similar altitude. In plain language, there are quite a large number of ways of being "fit" in the troubling world that lies ahead of us.

There are two things which can be said about this. First, competition theory tells us that agents do best when they operate under some degree of specialisation, and inactive defence of what they can do better than any other agents. (The theory of comparative advantage tells us that this is not the entire story – one does well by trading even when at a disadvantage; but this does not invalidate the point.) Consequently, it is foolish for states to pursue a global optimum as thought there was only one way of being human, being a house or being a company. It is equally foolish not to recognise advantage, however, and work to extend this into conditions which cannot yet be envisaged.

Second, external conditions are not uniform and change continually. Earlier, we noted that the ecology of the soil prospered by having "skills" – micro-organisms with capabilities for situations that did not yet exists – built into its repertoire. When conditions altered, it had the capability to react to these whilst maintaining homeostasis. Much the same has to be true of economies, in which skills and organisations, technologies and ways of organising social groups coexist in a plural way, with hopeful experiments awaiting their moment.

Third, much of the added value in the future will be derived from things which we do not yet perceive. (A perhaps-apocryphal 2004 study by the US Department of Labour was supposed to have said that half of the jobs in 2010 would use technologies that did not then exist to solve problems which we had not yet recognised.) Emergence implies that connections will generate degrees of freedom that we cannot anticipate. It is therefore essential to maintain a non-uniform, churning, stressful and complicated society whatever the expense of doing so and whatever the temptation to reach for uniformity and optimisation.


  1. Jaques Derrida is famous for saying that "Il n'y a pas de hors-texte": nothing exists beyond the text, by which he (probably) means 'the mechanism of representation'.
  2. Critics would assert that these engines are not natural but the product of Western assumptions, or if natural, then unfortunate for other reasons.
  3. The rationale of the freer spirits is less accessible, not least as they tend to use commonplace words in special and unannounced senses and scatter their texts with neologisms that defy definition. The following gives the flavour of this, and is taken from the Wikipedia article on Deconstructionism: Deconstruction identifies in the Western philosophical tradition a 'logocentrism' or 'metaphysics of presence' (sometimes known as phallogocentrism) which holds that speech-thought (the logos) is a privileged, ideal, and self-present entity, through which all discourse and meaning are derived. This logocentrism is the primary target of deconstruction. To clarify, Phallogocentrism should not, of course(!), be confused with phallocentrism, another hobby horse which some members of the movement appear to enjoy riding.
  4. Jean Baudrillard, The Gulf War Did Not Take Place, Indiana UP 1995.
  5. Kurt Gödel published the famous 'incompleteness theorem' in 1931. This statement in mathematical logic shows that any coherent system of symbols is capable of generating a statement that that the system in question can neither prove nor disprove: in effect, which had freedoms which the system could not pin down.
  6. Economics seen as the transport and partitioning of information is, perhaps, the most successful of current attempts to ground a part of the discipline in elementary components.
  7. See footnote above.
  8. For example, see The New Synthesis in Moral Psychology Haidt J Science 316 pp 998-1002 May 2007.
  9. The falsification principle says, more or less, that whilst it can be very hard to prove something, it is usually rather easy to disprove it. One white crow disproves the statement that 'all crows are black'. Billions of black crows cannot, however, prove it, only increase its likelihood of being a useful point of view. This has recently been rebranded for the mass market as the impossibility ofproving the absence of black swans by trotting out and endless sequence of white ones.
 to the top