How to bridge the Valley of Death that separates complexity/systems thinkers from decision makers?

Sometimes new ideas arrive like a bolt from the blue. More often they emerge through a series of conversations, reading and thinking.UKCDS logo An element of repetition may be necessary, provided you talk to different kinds of people about the same issue (rather than having the same meeting with the same people over and over again – we call that strategic planning).

So earlier this week I ended up in a room full of scientists and research council funders, organized by UKCDS, discussing a topic dear to my heart and blog – complexity and systems thinking. The topic was how research funding can channel complexity science to improve understanding and practice in the aid/development sector. The conversation reinforced a few points, but added a few new angles.

First was the sheer difficulty of bridging the gap between complexity thinkers in academia and busy aid activists, whether in government or NGOs. One contributor talked of a ‘valley of death’ that separates pure science from ‘operationalisation’ – using the new ideas to actually do stuff.

That gulf is further deepened by an academic system that works in disciplinary siloes that seem remarkably resistant to attempts to get them to cross fertilize – even waving research funding chequebooks often fails to get economists, ecologists, sociologists etc to really work together.

complexity you are hereI came away convinced that we need to understand these institutional barriers better, even if that does entail some navel gazing. Is it about language (economists and sociologists can’t understand each other), incentives (this won’t get me published in a peer-reviewed journal) or broader frameworks for understanding the world? It reminded me of a recent conversation at the Thinking and Working Politically seminar – unless we can understand why institutions find this so hard, and what can be done about it, we’ll probably be wasting our time.

Some suggestions for where to start: find out what research has been done on bridging similar gulfs, e.g. between pure science and industry. Study the role of individual ‘boundary crossers’, like our own Jean Boulton – is spotting and supporting the mavericks a quicker route to cross-disciplinarity than urging inertia-bound institutions and faculties to collaborate?

Second was a new (for me) way of selling complexity thinking to our political masters. This is all about making decisions in situations that are uncertain and volatile. What is ‘enough evidence to make a decision’, complete with feedback loops to detect success or failure, and adapt accordingly?

Otherwise, I banged on (as usual) about the tensions between systems thinking and the results/value for money agenda and linear planning tools; the importance of understanding the circumstances in which research influences policy; the power of case studies in showing that complexity thinking can lead to results and the barriers posed for activists by jargon and the geeks’ love of, errm, complexity. For me, the best way to discuss these issues in Oxfam is to avoid the C word altogether and ask a variant of ‘How do you plan when you don’t know what is going to happen?’ or ‘how do you campaign on a problem when you don’t know the solution?’

There was particular interest in the case study question, and how to improve our ability to capture the messy reality of change process, rather than construct an airbrushed, linear, quasi-mythical story. That means investing in real time accompaniment, and improving our ability to research and narrate how change happens.

The good news is that the research funders seem to understand the importance of it, and in many cases are already invested in supporting systems thinking. The Engineering and Physical Sciences Research Council (EPSRC) seems to have done most in the area (see here for details), but has not so far tried to link it to international development. Time for some cross-fertilization, I reckon.


Subscribe to our Newsletter

You can unsubscribe at any time by clicking the link in the footer of our emails. For information about our privacy practices, please see our Privacy Policy.

We use MailChimp as our marketing platform. By subscribing, you acknowledge that your information will be transferred to MailChimp for processing. Learn more about MailChimp's privacy practices here.


15 Responses to “How to bridge the Valley of Death that separates complexity/systems thinkers from decision makers?”
  1. Andrée Carter

    Thanks Duncan, a great summary. UKCDS is keen to continue our discussion on how complexity science can add value to research programmes on international development so please add your comments to the blog or get in touch with us.

  2. Simon McGrath

    Duncan, in the social sciences there are clearly elements of incentive systems that contribute to the problem- e.g., REF in the UK is monodisciplinary in bias. However, I think it is more about culture issues – the inability of sociologists and economists (as you say) to understand each other. This one is not simply about languages used but is often at a far deeper level of what are important as issues to study and how to study them. Running a small fund to support interdisciplinary work, I find that everyone can find something in that space to pitch me an idea but that it largely remains a sociology, economics, geography, etc. take on the issue rather than a transdisciplinary one. Disciplines remain very powerful in shaping research behaviour.

  3. Nicholas Colloff

    First step might be to create a lively (yet peer-reviewed) journal – a place to publish, win academic brownie points, and stimulate cross-fertilization in the real world. One such example (from a different space) is ‘The Journal of Consciousness Studies’ ( that has helped cross-boundary working in this field markedly since its first appearance about 15 years ago. It is written in plain English too!

    Second, following the work of Elliott Jacques on ‘capability’ recognise that people have different appetites/comfort levels around acting into uncertainty. People, Jacques argued, can be helped to identify where this sits with them, how it may change over time, and where they can make their best contribution into any given working context. Thus, imagining that everyone is going to comfortably work with complexity if only we get the language, tools right is a fool’s errand. People who ‘get it’ and critically ‘feel comfortable with it’ need to create contexts in which people who do not feel comfortable with it can make their real contributions. People who navigate uncertainty well need to create known contexts for others in which they can comfortably act.

    My favourite example of this is the eternal warfare between physicians and managers in the health service. You are a great physician because you love taking an individual case, diagnosing against a body of knowledge and effecting a cure. There may be ‘uncertainty’ in any individual case (and always more to be discovered) but essentially this is about deploying the right resources to a problem and solving it. This is exactly the model that physicians tend to apply to the service as a whole – give us the resources and we will effect the work. But as the managers rightly note, this is not how the complex system that is a health service actually works and the level of complexity is significantly greater and the unknowns greatly more variable. Do the physicians get it – not usually, in my experience (of long painful conversations with psychiatrists on developing mental health systems for example) and if they do, it is probably those who have become managers because they are drawn (and enjoy) a different level of complexity in their work! The first task of complexity wonks may be to create ‘linear spaces’ into which everyone else can act!

  4. Jean Boulton

    Couldn’t agree with you more Duncan (do you want that in hard copy and signed)!

    It seems to me that there are differing groups in the world of complexity – (a) mathematical modellers who explore particular problems using approaches (even if including non-linearities and/or agents) which inevitably simplify the real world; (b) social scientists and consultants who take the concepts from such models (strange attractors, fractals, self-organisation, edge of chaos, simple rules, equifinality, sensitivity to initial conditions) and consider what these might mean for management or leadership – which is tricky as some of these concepts are particular to certain types of models and some concepts contradict each other. Then (c) there is a much smaller group who think about how to explore actual complex situations on the ground, and then consider what such ’embodied’ research implies for management, impact assessment, programme design etc – and then go and find out if it works.

    Having spent years angsting about explaining such terms as self-organisation (which requires explanations of state space, and non-linearities and feedback and the role of variation and…) to non-scientists and concluded that it does not help much to save the world or help them to wend their way through the complexity jungle of terms, I now feel that the really important project is about what ordinary people do a bit differently if they accept that things are complex – interconnected, that change comes in fits and starts, that the history and particular features of situations cannot be ignored – in applying wisdom and methods and deciding what to do and how to show what has happened. I feel I – and others – are making some headway with this, project (c) and maybe this is what should be funded – real ‘on the ground’ research -but also development of management processes and ways to fund and assess impact that are more in tune with the complex world. Pretending things are simple and measurable and controllable when they are not either means we fund things that are simple to measure or keep our eyes down in the implementation not to be distracted by change or wider impacts or the need for adaptation. I think facing up to the complexities of the world is even more relevant for ID than for business – as business wants to dominate its context and get suppliers and consumers to ‘do it their way’ whereas ID is actually trying to work with the emerging complexities of situations and help people in their own terms.

  5. Nigel Timmins

    Thanks for this interesting post. I personally enjoy complexity as a model for understanding organisations, but one of my observations that contributes to the “Valley of Death” is many people’s emotional reaction to uncertainty combined with the transaction costs of trying to keep an organisation on the “edge of chaos”. Managers trying to use complexity approaches will need to develop excellent people skills and still be able to cast a vision; for many people emergence alone is not enough. There is also the need to recognise the time commitment in allowing emergence as an approach…. something we sometimes struggle with…

  6. Nathalie

    I completely agree on the need for cross-fertilization! Besides my work within my organization, I am lecturing at the University, dealing with “Sociology of actors in humanitarian settings”. I used business models and systemic analyses to prepare some of my lectures, in order to explain how the humanitarian system works (and don’t work), referring to the models of open/closed systems. In business model analysis and tools (I love the empathy map!) and in organizational analysis, there are some tracks to understand better our humanitarian world, but very few bridges!

  7. Ian Thornton

    Thanks Jean – interesting thoughts. I very much agree with you that “the really important project is about what ordinary people do a bit differently if they accept that things are complex”. In my world, the ‘people’ are research funders and those trying to use research to underpin development policy or programmes.

    I think (speaking personally – though I work for UKCDS who convened the discussion Duncan’s reporting on) that this is a significant challenge to the dominant model of research funding. Often real world development issues are broken down into smaller and smaller parts until these can be picked apart rigorously. But, if we accept that many interesting development problems are complex, to me this means we can’t assume that the sum of the parts (re-integrated from detailed study) behaves the same way as the whole…. Is that a correct interpretation of complexity!?

    Funders are moving towards more systems research (with health systems and education systems programmes on the go) but as Duncan and Simon McGrath note above, traditions remain powerful, and embracing complexity represents really quite a big shift.

  8. Jean Boulton

    Just replying to Ian, yes that is a good interpretation, I think – and saying the same thing the other way up – causes are multiple and interact synergistically not additively – many things taken together ’cause’ outcomes and the order of events matters and outcomes may be time-delayed. So the ‘taking apart the machine to see what caused what’ is often (usually) misleading. Programmes that give more freedom to work systemically – on more than one angle – and impact assessments that are more intelligent and look for synergies and also look for wider impacts – need more development. It is probably better to be intelligently critically nearly right than precisely, exactly wrong (which is what can happen if we get into the lego model of programme design and evaluation).

  9. Alex Ademokun

    Thanks for a great post and the really interesting discussion above. One of the big impediments to bridging the gap between thinking and doing (or academia and operationalisation), as a couple of others have mentioned, is the funding structures and incentive models (in this case I speak from the development side not academia). It tends to follow a thinking/planning/proposal phase and then an operational/implementation phase where you have effectively done all your thinking and are now simply ‘doing’.

    I note Ian’s excellent point about how the current model of development funding sometimes presumes the sum of the parts behaves as the whole. The process of breaking down problems into ‘thinkable?’ units is very helpful and makes it easier to manage but often units arrange themselves into different wholes depending on where you view a problem from (as a citizen, civil society, politician, business etc. or even various combinations). It is important that programmes be able to respond to changes in the environment beyond project activities, be able to constantly refine projects with the help of academia, policy influence windows, local expertise. The bottom line is this is scary for many organisations because it not only requires an acknowledgement of our limited knowledge and control but it requires a certain amount of letting go and reacting to events out of our comfort zone.

  10. Alistair Brett

    I agree with Jean’s (a), (b), and (c) groupings, although maybe we should not completely shy away from explaining concepts such as state space to non scientists. After all many of us scientists have learned about sociology, economics, and so forth. In my series of blogs I’m trying to use my experience in commercialising technology to apply ideas of complexity to practical, problem solving applications – especially those in developing and middle-income countries.

  11. Heather Marquette

    I can definitely echo Simon’s comment about the impact of the REF on academic incentives to work in a mono-disciplinary way. Funding won’t make a difference; there’s the issue of understanding each other’s languages, of course, but ultimately we’re under pressure to publish in disciplinary journals and use disciplinary language, and performance in the REF is our number one incentive. But even in an interdisciplinary field, which development studies is, there’s a problem in translation going the other way: getting colleagues to recognise the value of mono-disciplinary work. A giant, REF-induced mess…

    But, the REF is now past (well, maybe not for those of us assessing outputs…), and so it’s a good time to try to shape the next REF, in whatever hideous form it’ll take, in a way that creates better incentives for the sort of research you’re talking about. Inclusion of ‘impact’ in this REF had an overnight effect on how universities value policy and public engagement, so it can happen again.

  12. Ugo Gentilini

    Duncan, I look at the whole agenda about connecting complexity to development assistance with growing wariness. It is getting excessively convoluted, elusive and alarmingly boring. In order to be relevant and exciting to practtioners, lets focus on the bottom line – that is, how to devise effective feedback loops from the frontlines. It seems an obvious point, but one that is increasingly lost in discussions about change, systems and uncertainty.

  13. Søren

    To Ian’s point above – on breaking research questions into parts.
    I believe Alexander Bogdanov, a truly fascinating pioneer on complexity thinking among other things, said it quite well.
    (Quoting from memory)

    *The trouble with breaking research into small bits is that you isolate the problem from what created the problem in the first place.*

  14. Rob Levy (@aid_complexity)

    I’m working on an EPSRC-funded complexity science project at UCL in London. We’re interested in getting together with people from the development world to do a “data for brainpower”-type transaction. If you can give us juicy data to get our teeth into, we can, well, get our teeth into it. There needs to be more of this sort of “researcher wants data, NGO wants research” collaboration, don’t you think?

Leave a Reply

Your e-mail address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.