See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/230557481

Complexity, Configurations and Cases

Article  in  Theory Culture & Society · October 2005

DOI: 10.1177/0263276405057194

CITATIONS READS

151 266

1 author:

David Byrne
Durham University
84 PUBLICATIONS   1,636 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

the politics of taxation and paying for welfare. View project

post-industrial class View project

All content following this page was uploaded by David Byrne on 17 June 2014.

The user has requested enhancement of the downloaded file.

 Theory, Culture & Society
http://tcs.sagepub.com

Complexity, Configurations and Cases

David Byrne
Theory Culture Society 2005; 22; 95
DOI: 10.1177/0263276405057194

The online version of this article can be found at:

http://tcs.sagepub.com/cgi/content/abstract/22/5/95

Published by:

http://www.sagepublications.com

On behalf of:
The TCS Centre, Nottingham Trent University

Additional services and information for Theory, Culture & Society can be found at:

Email Alerts: http://tcs.sagepub.com/cgi/alerts

Subscriptions: http://tcs.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 95

Complexity, Configurations and

Cases

David Byrne

. . . a practice of social and historical explanation, sensitive to structure but

aware of contingency, is not yet at hand. We must build it as we go along, by
reconstructing the available tools of social science and social theory. Its
absence denies us a credible account of how transformation happens. (Unger,
1998: 24)

‘A
PRACTICE OF social and historical explanation, sensitive to
structure but aware of contingency’ – Unger is demanding a social
science as a basis for social action – a programme which explains
how things are as they are so that we might understand how to make differ-
ent futures, and futures which are radically different – transformed. His is
not a lone voice. Latour (2001, n.d.), in a programme of civic and cultural
engagement, has not only remarked on the breakdown between the domains
of scientific knowledge and experimental practice on the one hand, and
political action informed by value systems on the other, but has also called
for simulation as a basis for informing public understanding and political
practice:

The third type of event, much more risky and difficult to express but indis-
pensable, is a Simulation [original emphasis] at various scales of real
debates on pending issues to present to the public the various possible solu-
tions of what could be due process in matters of scientific democracy, or what
Sheila Jasanoff has proposed calling civic epistemology. (n.d.)

Of course Latour is a constructionist – a practitioner of what Pawson and

Tilley call ‘Hermeneutics II’ (1997: 21) – and, in relation to the field of
scientific knowledge, ends up with a conventionalist story of the products
of science. Desrosières (1998), an explicit follower of Latour, has taken this

■ Theory, Culture & Society 2005 (SAGE, London, Thousand Oaks and New Delhi),
Vol. 22(5): 95–111
DOI: 10.1177/0263276405057194

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 96

96 Theory, Culture & Society 22(5)

account into the realm of knowledge in the form of official statistics –

perhaps the most significant of all knowledge forms for the practice of
politics in post-industrial post-democracies,1 where social being might be
considered to be constituted through practices of consumption and politics
is acutely concerned with the representation of the standard of public
services, through statistical measures of standards and targets achieved, to
the public as consumers.
Where Latour is an epistemological constructionist; Unger’s position
is characteristically realist. And yet even Latour is asking for a science – a
knowledge set – which informs the discussion of alternatives in policy
formation. Unger goes further. He recognizes not the conventionalism of
knowledge – an inherently epistemological position – but the contingency
of the world – a world both social and natural – which that knowledge must
describe. His is a plea for a different form of science founded on this onto-
logical recognition – for knowledge that is local rather than universal – for
an abandonment of the dichotomy between nomothetic and ideographic.
The purpose of this article is to argue first, in a general sense, that
complexity science provides a systematic – and that word is used here
absolutely literally – foundation for the kind of social science that Unger is
seeking as the basis for a knowledge-based politics at every level, but
particularly that of popular participatory practice. Complexity provides a
foundation for one side of the debate specified by Paulo Freire as the foun-
dation of dialogical social research as part of a programme of participatory
politics. Freire’s ideas and practices have been taken far too little notice of
in the North, outside the extra-academic ghettoes of community develop-
ment and adult education. In the South they are fundamental to some of the
most radical institutional and processual developments in popular politics,
through the Brazilian Workers’ Party.2 The lack of attention to Freire reflects
both the contemporary weakness of Philosophy of Education,3 particularly
in the UK, and a continued incapacity on the part of Northern intellectuals
in the post-communist era to engage outwith the academy and liberal
educated elites.
The argument will then proceed by arguing against a specific, and in
the Anglo-Saxon scientific community currently dominant, mode of
complexity in which general simple rules which drive complex emergence
are seen as appropriate replacements for universal laws. Instead we should
engage with complex complexity, and do so through methods which enable
us to delve into complex causal processes through a reconsideration of the
nature and potential of that most distinctive of sociological techniques – the
comparative method. It will do so for three reasons. First, complexity must
become more than a metaphorical apparatus in social science and this can
only happen if the complexity frame of reference shapes the actual tools of
investigative social science themselves. Second, the comparative method is
particularly well suited for explication of the character of such shaping, both
intrinsically – since recognition of complex causation has always been a
foundation of the comparative approach – and in relation to the limitations

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 97

Byrne – Complexity, Configurations and Cases 97

of many social scientists who do not possess the vocabulary necessary to

engage with serious quantitative investigative tools. Finally, the compara-
tive method employed at the level of neighbourhood and city region has very
considerable potential for informing the participatory process in policy
formation and implementation – for serving as a basis for what is actually
an ongoing process very little noticed by social theorists but one with very
considerable implications for the nature of politics in post-democratic
societies.
Complexity Defined
Complexity – the word is everywhere in the field of knowledge. There are
complexity institutes, complexity journals, complexity popular science
books, complexity management, complexity art, this issue of this journal –
and the list could roll on ad nauseam. Let us begin with a working definition
of the term by using it as an adjective – by defining complexity theory as:
the interdisciplinary understanding of reality as composed of complex open
systems with emergent properties and transformational potential. A crucial
corollary of complexity theory is that knowledge is inherently local rather
than universal. Complexity science is inherently dynamic. It is concerned
with the description and explanation of change and one of its most powerful
terms, imported from general dynamic theory, is trajectory – the account of
the actual pattern of change of a system. As social scientists, as we shall
see, we have the advantage of being able to deal with multiple trajectories
– to use dynamic theory’s useful expression – with ensembles of trajectories.
Complexity theory seems to offer social science a way of resolving the diffi-
culties expressed in the common examination question of my undergradu-
ate days: Is sociology a science?4 It does so not by reconciling sociology
and the other social sciences to the ‘hard’ sciences and the nomothetic
project of the establishment of universal laws – although a substantial part
of this article will be concerned with defining and dismissing a version of
complexity theory, ‘simplistic complexity’, which seeks to do something very
much like that. Rather, complexity theory challenges the nomothetic
programme of universally applicable knowledge at its very heart – it asserts
that knowledge must be contextual. Moreover, it breaks down the bound-
aries between natural and social as objects of knowledge and action, and
necessarily places human social agency as of crucial historical and poten-
tial significance for the constitution of planetary reality as a whole, precisely
because human agency can change system trajectory.
The general significance of complexity theory and its potential as a
frame for science has been asserted frequently since the early 1990s (see
Byrne, 1998; Cilliers, 1998). Despite assertions that complexity theory
represents a foundation for postmodern science, the complexity project
necessarily confronts the subjective relativism of postmodernism with an
assertion that explanation is possible, but only explanation that is local in
time and place. Complexity science addresses issues of causation with
cause, necessarily, understood as complex and contingent. The project of

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 98

98 Theory, Culture & Society 22(5)

establishing how things have come to be as they are is valid in the complex-
ity frame of reference. Urry’s recent Global Complexity (2002) exemplifies
the use of the language of complexity theory as a method of describing a
current system state through his account of the character of global systems
at the level of the macro-social.5
Complexity is an idea whose time has come. And it has come because
the complexity frame of reference enables us to do exactly what Unger speci-
fies as the necessary task of social science and to do it in just the way he
suggests for us – through reconstruction. We can go forward not by import-
ing a whole new set of tools and meta-theoretical specifications from some-
where else, but by reconstructing the tools and theories that we already have
in complexity terms.
Let us pay attention first to the necessity of the task. This article is
written in the spirit of an unregenerate progressive modernism – in the belief
that social science in a dialogue with collective experience has something
to contribute to the resolution of the human condition. It is written in support
of the importance of empirical investigation and against both the sterility of
structuralism, magisterially dismissed by E.P. Thompson in The Poverty of
Theory (1978), with its anti-empirical claim to a deductive and general
account of social process, and the disengagement of relativist post-
structuralism as the meta theory of postmodernity from any practical engage-
ment at all of social science with processes of social transformation. It also
rejects the elitist turn taken by Flyvbjerg (2001) towards ‘phronesis’ – the
virtuoso performance of the social scientist as elite gentleman. If we are to
use Aristotelian terminology, then what follows is an argument for techne as
the basis of praxis, but for a techne that does have a specific episteme to
underpin it. This is an argument for the complexity frame of reference as a
way of informing half of the dialogue of reflexive participatory social
research, and a proposal for a reconfigured version of the comparative
method as a key tool in the toolbox of the social scientist as craft worker.
And this is a task of reconstruction – not of importation. Hayles (1991:
7) made a very important point about the complexity project in its prelim-
inary form of ‘chaos theory’ when she denied the primacy of ‘scientistic’
chaos and argued that the theme of chaos was everywhere in the episteme.
Complexity theory is not a matter of importing ideas from ‘the hard sciences’
into the consideration of the social, although some of the terminology of non-
linear dynamic theory can be rather useful to us. Rather it involves thinking
about the social world and its intersections with the natural world as involv-
ing dynamic open systems with emergent properties that have the potential
for qualitative transformation, and examining our traditional tools of social
research with this perspective informing that examination. The idea of
complex systems in its essential form is not a new thing. If we adopt the
Gulbenkian Commission’s very useful definition of science as ‘systematic
secular knowledge about reality that is somehow validated empirically’
(1996: 2), then the complexity science programme is at least as old as
modernity, although we have not had that name to give to it. McKeown

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 99

Byrne – Complexity, Configurations and Cases 99

(1979) got it absolutely right when he identified the understanding of the

relationships among human social organization, the urban environment and
infectious diseases as the greatest contribution of medical science to human
progress.6 That understanding, developed in its essentials in the second
quarter of the 19th century, was essentially a complexity-based account of
how the natural world, social structure and social action interacted.
Moreover, it constituted a complex science which informed human actions
that changed the character of urban natural/social7 environments and
engendered the health transition8 which has transformed human demogra-
phy and is transforming human history.
The history of epidemiology is one in which the complexity-based
understanding of early modernity was challenged by the analytical and
simple understanding made possible by the empirical validation of the germ
theory of infectious disease – the main support to this day of medicine’s
commitment to specific aetiology. The analytical programme of simple
science triumphed both over the case knowledge of pre-modern medicine
and the complex social understandings of early modern9 public health. The
prime purpose of this article is to argue for a complex version of case-based
methods derived from casting recent developments in the comparative
method in complex terms. In so doing it will argue against simplicity and
against ‘simplistic scientistic’ complexity. At the same time it will be
unremittingly hostile to the ‘post-explanatory’ turn of much of contempor-
ary ‘social science’. Complexity specifies knowledge as local but not as
relative. It always understands knowledge as socially constructed but not as
reified. Reed and Harvey (1992, 1996) have indicated the congruence in
this respect of complexity with the programme of ‘scientific realism’ and that
position is endorsed here. We are in the business of elucidating how things
change, even if our methods are radically different from the dominant form
of scientific explanation through analysis and reduction.10 Moreover, we can
use these methods as part of a dialogical engagement with human agents
rather than the decomposition of the complex social through analysis based
on the reification of variables. Complexity understood in this way validates
the engagement of science with reflexive social action.
Abbott, who has directly engaged with the case-based methods which
will be the basis of the substantive discussion in this article, has argued
that:

. . . the larger, universal framework for social science is by no means the

standard, often-parodied axiomatic structure. Rather it resembles what the
Romans called the law of peoples (ius gentium), a law that they applied to
diverse groups at the edges of empire and that they distinguished from the
formalized civil law (ius civile) that applied specifically to Roman citizens.
There is no universal social scientific knowledge of the latter kind – system-
atic, axiomatic, universal in a contentless sense. There is only universal
knowledge of the former kind, a universal knowledge emerging from
accommodation and conflict rather than from axioms, a universal knowledge
that provides tentative bridges between local knowledges rather than

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 100

100 Theory, Culture & Society 22(5)

systematic maps that deny them, a universal knowledge that aims, like the
ius gentium, at allowing interchange among people who differ fundamentally.
(2001: 5)

That is a good basis on which to proceed.

Putting Social Science to Work
As always with complexity, context is everything and the present context for
social science, particularly in the UK, is one in which considerable demands
are being made on the social scientific project – in effect to put up or shut
up. The recent Academy of Learned Societies for the Social Sciences report,
Great Expectations (2003), addressed the potential of social science as a
basis for policy formation and social practice. That report is banal, defen-
sive and utilitarian in the extreme. Nevertheless, the requirement that social
science speaks to audiences beyond the academy is appropriate, although
the tone of a complexity-informed dialogue may be more critical than the
authors of Great Expectations would welcome. Moreover, within the academy
itself, complexity-informed assertion will disturb those who are complacent
about some rather self-confident disciplines. Great Expectations explicitly
rejected the Gulbenkian Commission’s wholly accurate assessment of the
failure of nomothetic social science to deliver any sort of foundation for
prediction in relation to social action. The Gulbenkian report was informed
by the complexity frame of reference, through the participation on the
commission of Prigogine. It did not conclude for an abstract and disengaged
project of relativist reflection, but rather sought to move things on beyond
the sterility of positivist science and linear modelling. It proposed, inter
alia, a complexity turn.
This has very considerable salience. Policy-makers and practitioners
perfectly reasonably expect science – defined in the Gulbenkian fashion as
above – to deliver some sort of answer to the pertinent question: what works?
Indeed, contra Weber, most social scientists do hanker for a role in the
informing of the democratic process, the essential politics of modernity.
However, given the relativist turn in social meta-theory, it often seems as if
the task of informing policy elites has been willingly ceded to a revived
crude positivism in which the randomized controlled trial is hailed as the
‘gold standard of evaluation’. The task of informing democratic debate seems
to have been ceded to mass media dominated by corporate interests! Inter-
estingly, the main challenge in terms of evaluation, and so far the only one
that has had any impact on the policy process, has come from a critical
realist turn which is wholly compatible with the complexity perspective (see
Pawson and Tilley, 1997). The engagement of social scientists with policy
is not a straightforward process. In particular such engagement poses
dangers for the critical function of social science as commentary and as
practice. Complexity does have potential for engagement, although one
version of complexity would deliver an engagement which would replicate
the worst technocratic elitism of early attempts to inform policy through

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 101

Byrne – Complexity, Configurations and Cases 101

systems-theory-based approaches. One of the tasks being attempted here is

to argue that a different fix on complexity and a different approach to method
has a radical and emancipatory potential. It certainly can inform evaluation,
but it can go beyond that to inform democratic political practice as social
action.
It is necessary in writing about complexity to move beyond the prelim-
inary specification of the general character of complexity science and the
delivery of glossaries of its terminology. Let us agree that we are involved
in a project of discovering complexity, that like the scientists studied by
Bechtel and Richards, we are:

. . . not primarily interested in general laws; [but rather are seeking] to

identify causal components that explained how various systems produced
specific phenomena. (1993: xi)

In other words as Cilliers (1998) has demonstrated, the essential character

of complexity-based knowledge is that it is local. The big question is how
can we interrogate the local to understand how things have come to be as
they are and how they might be made different. Here social science has
something to it which is singularly lacking in the ‘hard sciences’ other than
implicitly in ecology. It understands how to compare. Much of ‘scientistic
complexity’ is concerned with the dynamic trajectories of individual
systems. We as social scientists can deal with – to use the terminology –
ensembles of systems. We can deal with lots of cases and see how the config-
urations they represent can help us to understand the various ways in which
things have come to be as they are, the various ways in which they might
be different, and – with luck and the wind in the right quarter – how social
action might produce one possible future rather than another.
So we need a comparative complexity method – and the central
argument of this article is that we have one in the form of a complexity-
based interpretation of Charles Ragin’s ‘fuzzy set’ qualitative comparative
analysis. However, before we come to that we need to clear the ground a
little. We need to examine simplistic and rule-based complexity and distin-
guish it absolutely from case-based empirical investigation. Having done
that, we can examine case-based methods and see if they do offer us a new
way of seeing how things have come to be. Then we can return to the ‘big
issue’ – to the assertion of complexity approaches as precisely the foun-
dation of an engaged social science which can help us to understand how
transformations happen and how we might, dialogically, engage in making
them happen.
Simplistic Complexity
Before getting into the assault on simplistic complexity, which is the
substance of this section, there is something that needs to be stated very
clearly. Arguments for complexity are not arguments against simplicity.
Some things can be understood by the analytic and reductionist programme

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 102

102 Theory, Culture & Society 22(5)

and where that programme works it has done great service in elucidating
causality. The problem is that it works where it works and it does not work
everywhere. Indeed in a natural/social world the range of its applicability
is rather limited. The problem is that, instead of the application of the simple
model being understood as something that always has to be justified by
showing that what is being dealt with can be analysed, the simple model is
taken as ‘the scientific model’, which is always applicable. The dominant
contemporary modes of statistical reasoning in the social sciences are a
particular example of this (see Byrne, 2002).
Before trying to put complexity to work we need to delimit a very
special sort of complex reasoning which takes the simple model as its
template and which, like the simple model itself, asserts general applica-
bility. Whereas the long-term project of the simple model as ‘the scientific
method’ is the establishment of universal laws which apply always and
everywhere, the project of what I am going to call – in a deliberately
oxymoronic fashion – simplistic complexity, is the establishment of a
general set of rules from which emergent complexity flows. There are import-
ant differences between the simple model and simplistic complexity. In the
simple model, knowledge of the law and of initial conditions of the system
enables prediction of future states of the system. In simplistic complexity
knowledge of the rules does not, in principle, allow us to predict in this
way. Rules are not laws. Laws describe the behaviour of the system as a
whole. Rules define the range of action of components of the system.
The difference between simplicity and simple complexity can be illus-
trated by comparing a system described by a set of simultaneous equations
and a system modelled in a simulation. A mathematical model cast as a set
of equations enables, in principle, absolute prediction. A simulation can be
set running to see what happens and what happens can be used as a predic-
tion of what might happen in reality, but there is no established law to
validate that prediction. Of course the key issue for both mathematical
models and simulations is the degree of isomorphism with the real system
of which either is a representation. Indeed, those mathematical models that
go beyond the trivially simple are in some respects more like simulations
than they are like the simplest mathematical models. Only the simplest
mathematical models can be validated by hypothesis testing. Nonetheless
the equation set does fully describe a mathematical model. In a simulation
the rules simply describe how things get done.
Simple complexity does permit emergence – emergence is the whole
point of it and Emergence is the title of Holland’s (1998) canonical text. The
point is that the rules in simple complexity are themselves simple. For
example, in a cellular automaton such as the Game of Life, very simple rules
lead to complex emergent forms over a number of iterations. Simple
complexity is very attractive to the scientistic programme. It cannot estab-
lish laws but it can establish rules which seem to describe lots of systems
and the establishment of the rules has the same general character and, very
important this, status, as the establishment of scientific laws. Hayles, in her

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 103

Byrne – Complexity, Configurations and Cases 103

review of the history of cybernetics, employed the archaeological term

skeuomorph – a design feature which no longer has any function but served
some purpose in earlier periods. For cybernetics she argues, skeumorphs
served as threshold devices smoothing the transitions in understanding and
(in consequence) method. Using this terminology we might see ‘simple
complexity’ as a kind of meta-skeuomorph which keeps simulation as
acceptable scientism despite the profound anti-simplification content of the
whole idea of complexity, the inexorable character of the real as opposed to
the artificial.
This is why simplistic complexity is so attractive to the worst sort of
evolutionary psychology11 and contemporary ideologues of market models.12
Write a few rules – the selfish gene, the territorial imperative, profit maxi-
mization, rational choice, or, preferably, a combination of all of these, and
away we go. Simplistic complexity does deal with a kind of complex emer-
gence but it remains reductionist.
Simulation as a method, at the current level of development of simu-
lation tools, is based on simplistic complexity. It may be that simulation has
the potential to be developed beyond this but for that to happen software
developers will have to find a means of representing not just agents but
social structure – with all the complexities and contradictions that that
metaphor implies. Agent-based simulations can generate emergent struc-
ture but in the real social world structure is the product not just of contem-
porary actions but of history. It is there as we act although our actions
constitute it and change it. In other words, a simulation that would be
adequate as a representation of the social complex cannot start from agents
alone.
This is not to dismiss simulation, or indeed to dismiss simplistic
complexity as a mode of reasoning. Rather, as with simplicity, the appli-
cation of simplistic complexity must always be justified by at least an
argument to the effect that the system(s) being represented can be described
in simplistic complex terms. For Holland and those who follow his approach,
simplistic rule-based complexity is everywhere and science is just a matter
of establishing, or more often asserting, the rules. This is very different from
the careful empirical work of some sociologists, anthropologists and ecolo-
gists who are using simulation approaches. For example Chattoe and Gilbert
(1995) used interviews to establish budgetary behaviour among a particu-
lar set of consumers before formulating rules for agents in a simulation.
Likewise Doran and Palmer (1995) based their simulations of upper palaeo-
lithic social change on archaeological evidence about human groups in that
period. In other words, these ‘empirically based’ simulators did look at the
systems they were concerned with and, implicitly at least, ascertained that
simplistic complexity could serve to describe them. In general simplistic
complexity is appropriate for real micro-social phenomena which may
nonetheless have macro-social implications. Schelling (1978) produced a
convincing cellular-automaton-based simulation of segregation in US cities
which at least raised the issue that an important part of the mechanism that

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 104

104 Theory, Culture & Society 22(5)

generated observed segregation was the desire by families, both black and
white, to live in areas in which their own ethnic group was the majority,
although they did not want to live in completely segregated areas. To say
this is not to deny institutional racism and other determinants, but it is to
say that micro-social actions may have unintended macro-social conse-
quences.
So simplistic complexity is fine as a modest basis for exploratory
empirically grounded simulations but when it is asserted as the basis of
everything, then turn off the sound. If the sound is not turned off, then
simplistic complexity, with an apparatus of technical model-building and
the potential for close association with deterministic ‘bio-social’ theorizing,
could very well acquire the kind of status in relation to policy formation and
development currently occupied by simplistic science. Indeed simulations
are already in use, although the foundations of such simulations often lack
the empirical foundation which careful social science simulation brings to
rule specification. The simulation of foot and mouth in the UK was a perti-
nent example. Bizarrely, but of course not surprisingly in relation to the
findings of science studies, accuracy of prediction is not a requirement for
status in terms of the relationship between scientific technique and policy
processes.
More directly, and in particular in urban planning, simplistic complex-
ity is already demonstrating considerable potential for supporting a combi-
nation of technocratic elitism and market determination in the setting of
urban futures. There was a good deal wrong with the old ‘evangelistic
bureaucrats’ who drew on simple systems theory in the days when planning
was conceived of as a social-democratic activity. Simplistic complexity
informing urban simulations of catalysed markets would be a good deal
worse. But there is an alternative!

Comparing Cases – A Complex Fix on the Comparative

Method
Ragin, in his recent book Fuzzy-set Social Science, refers to:

. . . an idea widely shared by social scientists – that it is possible to derive

useful empirical generalizations from the examination of multiple instances
of social phenomena (i.e. from comparable cases). (2000: 332)

That is a pretty good specification of the essentials of the comparative

method. Let us see what complexity might tell us about cases and their
trajectories, and then compare the implications of a complex account of the
case with Ragin’s understanding of cases as ‘configurations’. This will offer
us an actual method for exploring complexity and complex causation, which
not only corresponds to one of the foundational techniques of social science,
but can be the basis for dialogically engaged action-research for social
change. In other words, we might be able to go beyond understanding social
transformations towards a programme for shaping those transformations.

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 105

Byrne – Complexity, Configurations and Cases 105

The essence of the case in a complex frame is that cases are in them-
selves complex systems which are nested in, have nested within them, and
intersect with other complex systems. So, for example, a city-region is nested
within global and national systems and has nested within it neighbourhoods,
households and individuals. Nesting is not hierarchy. Determination runs
in all possible directions, not just top down. All these levels potentially have
implications for all other levels. However, as well as these elements which
might be described by a hierarchical data system, there are other systems
which intersect with them, cut across them, constitute part of them and are
constituted by them. What is a system of interest at any point in time is
defined by observation and action. Boundaries depend on what we are
looking for and at. This is not to say that boundaries are arbitrary, relative
or unreal. At one level the geography of city region is real. With another
focus of attention the global socio-scape of an immigrant household is real.
Each matters for the other.
The trajectories of systems are the histories of cases. The trajectories
of complex systems have histories that are a mixture of ‘much the same’
and change. For much of the time complex systems remain the same sort of
thing. There are changes in them over time but these do not constitute
changes of kind. To use attractor labels, the systems are in a torus attrac-
tor. The changes that matter are in complex systems’ terminology phase
shifts, and what social science generally, after Polyani, calls transform-
ations. The language of transformation is essentially dialectical. We see
changes of quantity become changes of quality. The accumulation of con-
tinuous change leads to the crossing of a threshold of kind. So, for our
example, a city-region might move from an industrial to post-industrial char-
acter. However, the forms available to post-industrial cities are multiple.
There is not one kind of post-industrial city. Rather, there is a range of
possible post-industrial forms, even within a particular global system.
Moreover, and this is the point about the non-hierarchical character of
nesting and intersection, different trajectories attained by city-regions (the
active form of the verb is very deliberate) have the potential to influence the
form of the global system as a whole. It should be evident that, at the very
least, we will be interested in what sorts of actions might generate a particu-
lar outcome from a post-industrial transformation locally. Beyond this, the
emergent consequences of varied local trajectories for the global system as
a whole, will matter for us.
This project is necessarily one of seeking to establish causes.
However, it is evident that causal processes in complex systems cannot be
accessed by simple analysis. The trajectories of complex systems will always
be directed by complex and contingent cause. History will matter. There
will be path dependency. Context will matter. Agency will matter. This
account of causation corresponds exactly with the essentials of the scien-
tific realist description of cause as Reed and Harvey (1992, 1996) have indi-
cated.
Compare this with Ragin:

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 106

106 Theory, Culture & Society 22(5)

For causation, the main contrast is between the conventional view of causa-
tion as a contest between individual variables to explain variation in an
outcome and the diversity-oriented view that causation is both conjunctural
and multiple. In the conventional view, each single causal condition,
conceived as an analytically distinct variable, has an independent impact on
the outcome. In the diversity-oriented view, causes combine in different and
sometimes contradictory ways to produce the same outcome, revealing differ-
ent paths. (2000: 15)

This is the same thing said in a slightly different way, with one very inter-
esting additional twist. Ragin points out that the same outcome might be
produced by different causal combinations. There are different ways to the
same future.
To develop this point let me elaborate on Ragin’s notion of configura-
tion. Ragin seems to have developed his arguments in isolation from the
critical realist tradition and with limited (up to now) explicit reference to
complexity theory and the vocabulary of open systems, but his essential
account is substantially identical. For him;

. . . cases are viewed as configurations – as combinations of characteristics.

Comparison in the qualitative tradition thus involves comparing configura-
tions. This holism contradicts the radically analytic approach of most quan-
titative work. (1987: 3)

This resembles Elias’s view,13 when he remarked that what is required is to:

. . . investigate the nature of this range of possible transformations and the

configuration of factors responsible for the fact that, of all of the possibilities,
only this one is materialized. (1970, cited in Noves, 2002: 5)

However, Elias was dealing with a single trajectory. Ragin, although dealing
at the same macro-social level as Elias in empirical work, deals with
multiple cases. Nonetheless configuration here is used as a way of express-
ing complex and contingent cause.
This is not the place for a full elaboration of Ragin’s approach, although
some colouring is appropriate. Basically Ragin has developed and refined a
‘small N’-based method in which Boolean truth tables are constructed as a
way of representing specific configurations which have engendered particu-
lar trajectories towards an outcome. This is work in the US macro-social
tradition of seeking to delineate the causal processes in social transformation
and it has had most influence where sociology and political science intersect
in the study of ‘international relations’. In his most recent book, Ragin has
become ever more assertive in his dismissal of an analytical and variable-
centred mode of causal reasoning and now argues that:

Fuzzy sets offer researchers an interpretive algebra, a language that is half

verbal-conceptual and half mathematical-analytical. . . . Most theoretical

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 107

Byrne – Complexity, Configurations and Cases 107

arguments, as verbal formulations, deal with set-theoretic relationships, so

they offer the opportunity for creating a close correspondence between theory
and data analysis. (2000: 4)

For me this resonates absolutely with understanding complex systems as

best explored by taxonomic procedures – what Ragin would call the estab-
lishment of differences of kind. But he also argues that we must pay atten-
tion to:

The second aspect of diversity [which] involves the varying degrees of

membership that instances may exhibit in the categories and types used
to categorize them. . . . The key to understanding the power of fuzzy sets is
to see that it is possible to specify qualitative breakpoints on continua and
to incorporate these qualitative breakpoints directly into the analysis of
evidence that varies by level. With fuzzy sets, researchers can analyze set-
theoretic relationships while still attending to phenomena that vary by level
or degree. They do not have to forfeit the study of variation by level in order
to study cases as configurations or to explore causal complexity. (2000: 16)

This is very interesting because it provides us with a way of understanding

measurements, taken as variate traces of the character of open systems (see
Byrne, 2002), which points us towards the significance of threshold values
as indicators of system change – indicators, not causes. The crossing of
thresholds indicates a difference of kind. In complexity terminology we have
markers for phase shifts.
Ragin’s is not the only approach that has this character but it is
probably the most clearly articulated. What is most interesting is that ‘post-
analytical’ research strategies, grounded in processes of comparison, are
emerging across the quantitative and qualitative research programmes in
social science. This is particularly so where the quantitative and qualitative
intersect and the development seems to be in considerable part driven by
the potential of computer technology and its capacity to serve us as ‘macro-
scope’ (see Byrne, 2002). Cilliers (1998) was making a very important point
with profound implications when he noted that we might be able to do with
technology what we cannot do with science. That is to say we can scan,
describe and represent using computing technology without a law-based
model of the processes being represented. Here the commercial develop-
ment of data-mining tools using neural net technology is extremely interest-
ing. In any event the point is, that as Hayles noted for the ideas themselves,
complexity-based reasoning and tool-building is popping up all over the
place, whatever name we care to put on it. It is in the episteme, in the
culture, and in the theory. So, what can we do with it?
Complexity and Social Practice – Regaining Confidence
If we are going to have an engaged social science then the character of the
process of engagement will, in the best complexity fashion, be inseparable
from the modes of reasoning which inform that engagement. Complexity

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 108

108 Theory, Culture & Society 22(5)

theory, particularly simplistic complexity theory, looks like the property of

an intellectual elite who claim technical expertise can engage with real
human systems in processes of social engineering. That is one way things
might go. It is not what is being proposed here.
Instead – assertively and at the risk of pretentiousness – it is argued
that complexity does offer an historically grounded frame of reference which
can be used to inform the organic practice of intellectuals engaged in par-
ticipatory research. Action-research has developed as a research practice
in a range of applied fields in the social sciences, but the implications of
the process have not informed meta-methodological debate. Original inter-
ventions in action-research argued for strict experimental designs, but,
under the influence of practice and of the pedagogical theories of Paulo
Freire, there has been a general shift towards dialogical modes of engage-
ment with social actors in the research field. This means that scientists –
in the Gulbenkian sense of this role – do not confine their engagement with
social actors to the elicitation of information but, in the process of research,
feedback in which there is a dialogue both as to the validity of account and
as to its implications for practice. This has much in common, although there
does not seem to be much if any cross-referencing, with integrative method
as proposed by Lemon (1999), which is explicitly complexity-based. The
point about action-research is that it takes the process of reflexivity out of
the epistemological ghetto and into social life. It is reflexivity that leads not
to relativistic passivity but rather to engagement in action. In this respect
it meets Paul Cilliers’ requirement that those who wish to engage with
complex social systems must, ethically, do so from within those systems.
The actual scale of evaluative and action-research in contemporary
post-industrial and post-democratic societies is enormous. The UK is an
extreme case but here almost all the policy initiatives of New Labour since
1997 have involved both action-research and proposals for systematic
evaluation. Most of the ‘applied funding’ for UK social science comes in
relation to these modes of intervention.14 Little of this work is dialogical and
virtually none of it is informed by a complexity frame of reference. It just
gets done without any real debate about its nature, purposes and conse-
quences. An important objective of this piece is the raising of questions
about this form of practice as part of the general debate about complexity
in social science as a whole. The intention has been to get beyond simple
rejection. Social science will be applied. The intention here has been to use
complexity as a mode of understanding and on that basis develop some ideas
about how a meaningful practice of applied research might be developed.
Notes
1. See Crouch (2000) for an explication of this term.
2. Ironically, in Brazilian politics Unger has been associated with a Bonapartist
figure, Gomes, from the traditional north-eastern elite. In the 2002 election Gomes
was obliged to support Lula – the Workers’ Party candidate and eventual victor, but
Unger’s political practice has not been connected in situ with Freirian radicalism.

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 109

Byrne – Complexity, Configurations and Cases 109

3. Durkheim’s first Chair was in Philosophy of Education. For much of the 20th
century radical pedagogy was the most important social practice informed by any
philosophical notion of ‘critique’.
4. This question is of course not meaningful in other languages where the word
commonly used to translate the English word ‘science’ includes all organized
secular knowledge.
5. One important point about complexity theory is that it resonates with other
accounts which use a somewhat different, although often overlapping, vocabulary.
In terms of the macro-social processes with which Urry is concerned, an important
resonance is with figurational sociology as proposed by Elias in a series of works.
6. This word is not used innocently as we shall see.
7. It is wrong to say natural and social. The point is that the environment is
emergent from the natural and social and is both and more than both.
8. This expression refers to the sudden and qualitative change of the pattern of
death in human societies which began in the West in the late 19th century and is
now almost global. Before this, in urban and agricultural societies most deaths were
due to infectious diseases and happened across the age range, but particularly in
infancy and childhood. Now most people die in old age of non-infectious conditions.
9. A teacher at the University of Durham has to specify that here early modern
here does not mean 6th century but rather early modernity.
10. The argument for the congruence between complexity as a scientific ontology
and realism as a philosophical ontology is well made by Reed and Harvey (1992).
In brutal summary we might say that both derive from a concern with the nature of
what is. This means that epistemological arguments about how we know have to be
answered in a way which derives from that specification of the nature of the objects
of knowledge. This approach resonates with the arguments advanced by Lakoff and
Johnson (1999) for an embodied theory of mind and an evolutionary understanding
of perception and cognition.
11. There is a much more subtle and interesting evolutionary psychology which
takes the Darwinian programme but deals in terms of cognitive, linguistic and
cultural emergence rather than genetic imperatives.
12. Upon whom a reincarnated Adam Smith would not make water were they spon-
taneously to ignite.
13. The revived interest in figurational sociology, which derives in very consider-
able part from a return to Elias’s work, is significant in two senses. It reinforces the
validity of comparative method at the level of the macro-social. Indeed, little macro-
social work outside this tradition gets beyond simple generative determinism
typified by Castells’ techno-deterministic version of his original Althusserian
project. It also reflects a desire by sociologists for an explanatory framework that
does not reduce to epistemological rhetoric.
14. Interestingly, explicit adherents of generally postmodernist meta-theoretical
positions often seem quite happy to take money for research projects in which evalu-
ation is cast in an entirely scientistic mode.

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 110

110 Theory, Culture & Society 22(5)

References
Abbott, A. (2001) Chaos of Disciplines. Chicago, IL: University of Chicago Press.
Academy of Learned Societies for the Social Sciences (2003) The Social Sciences
in Britain. London: Academy of Learned Societies for the Social Sciences.
Bechtel, W. and R.C. Richardson (1993) Discovering Complexity: Decomposition
and Localization as Strategies in Social Research. Princeton, NJ: Princeton
University Press.
Byrne, D.S. (1998) Complexity Theory and the Social Sciences. London: Routledge.
Byrne, D.S. (2002) Interpreting Quantitative Data. London: Sage Publications.
Chattoe, E. and N. Gilbert (1995) ‘A Simulation of Budgetary Decision-making
Based on Interview Data’, paper presented at the workshop Simulating Societies
’95, Third International Symposium on Approaches to Simulating Social Phenom-
ena and Social Processes, Boca Raton, Florida, 15–17 September.
Cilliers, P. (1998) Complexity and Postmodernism. London: Routledge.
Crouch, C. (2000) Coping with Post-democracy. London: Fabian Society.
Desrosières, A. (1998) The Politics of Large Numbers. Cambridge, MA: Harvard
University Press.
Doran, J.E. and M. Palmer (1995) ‘The EOS Project: Integrating Two Models of
Palaeolithic Social Change’, pp. 103–25 in N. Gilbert and R. Conte (eds) Artificial
Societies: The Computer Simulation of Social Life. London: UCL Press.
Flyvbjerg, B. (2001) Making Social Science Matter. Cambridge: Cambridge
University Press.
Gulbenkian Commission on Restructuring the Social Sciences (1996) Open the
Social Sciences. Stanford, CA: Stanford University Press
Hayles, K. (1999) How We Became Posthuman. Chicago, IL: University of Chicago
Press.
Holland, J.H. (1998) Emergence. Reading, MA: Addison-Wesley.
Lakoff, G. and M. Johnson (1999) Philosophy in the Flesh: The Embodied Mind and
its Challenge to Western Thought. New York: Basic Books.
Latour, B. (2001) ‘What Rules of Method for the New Socio-scientific Experi-
ments?’, URL (consulted June 2005): http://www.ensmp.fr/~latour/poparticles/
poparticle/p095.html
Latour, B. (n.d.) ‘From Realpolitik to Dingpolitik: An Introduction to Making Things
Public’, provisional introduction to B. Latour and P. Weibel, Dingpolitik Atmos-
pheres of Democracy. URL (consulted June 2005): http://www.ensmp.fr/~latour/
articles/article/96-DINGPOL-INTRO.html
Lemon, M. (ed.) (1999) Exploring Environmental Change Using an Integrative
Method. Amsterdam: Gordon and Breach.
McKeown, T. (1979) The Role of Medicine: Dream, Mirage or Nemesis? Oxford:
Blackwell.
Noves, A. (2002) ‘Policy in the Era of Globalization: Diversity and Intercommuni-
cation from the Perspective of “Complexity”’, Revista de Sociolingüistica, URL
(consulted June 2005): http://cultura.gencat/llengcat/noves
Pawson, R. and N. Tilley (1997) Realistic Evaluation. London: Sage Publications.
Ragin, C.C. (1987) The Comparative Method. Berkeley, CA: University of California
Press.

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.
 06_byrne_057194 (jk-t) 20/9/05 8:58 am Page 111

Byrne – Complexity, Configurations and Cases 111

Ragin, C.C. (2000) Fuzzy-set Social Science. Chicago, IL: University of Chicago
Press.
Reed, M. and D.H. Harvey (1992) ‘The New Science and the Old: Complexity and
Realism in the Social Sciences’, Journal for the Theory of Social Behavior 22:
356–79.
Reed, M. and D.H. Harvey (1996) ‘Social Science and the Study of Complex
Systems’, pp. 295–324 in L.D. Kiel and E. Elliott (eds) Chaos Theory in the Social
Sciences. Ann Arbor, MI: University of Michigan Press.
Schelling, T.C. (1978) Micromotives and Macrobehavior. New York: W.W. Norton
and Co.
Thompson, E.P. (1978) The Poverty of Theory. London: Merlin.
Unger, R.M. (1998) Democracy Realized. London: Verso.
Urry, J. (2002) Global Complexity. Cambridge: Polity.

David Byrne teaches in the School of Applied Social Sciences, University

of Durham. He is the author of Complexity Theory and the Social Sciences
(Routledge, 1998) and Interpreting Quanitative Data (Sage, 2002). His
interests include action and the use of social science in participatory democ-
racy research.

Downloaded from http://tcs.sagepub.com at PENNSYLVANIA STATE UNIV on April 17, 2008

© 2005 Theory, Culture & Society Ltd.. All rights reserved. Not for commercial use or unauthorized distribution.

View publication stats