The development sector really does manage to get its knickers in a knot at times. While agreeing that ‘it’s all rather complex’, they continue to hold on like shipwrecked sailors to linear models of planning, monitoring and evaluating, scratching their heads about why ‘it isn’t working’. And the sharks circle ever closer –dissatisfied donors who want to see value for money, frustrated activists having to provide data, overworked staff trying to chase up and make sense of the information that does trickle in. Innovation is not allowed to fail and indicators must be set years ahead of time that guarantee ‘poverty alleviation’ and ‘citizenship strengthening’ impacts. The sector partly needs to shake off unrealistically ambitious expectations. But it needs to start understanding more fundamentally the nature of the beast with which they are dealing.

Two things for now. The first is more of a comment on Irene's 'Aha! Moment in the Development Sector'. I can really recommend Narrative Policy Analysis by Emery Roe (Duke University Press, Durham/London,1994). This book elaborates on the benefits from applying contemporary literary theory to policy analysis using case studies in environmental and development policies. On page 5 it promises: "By establishing and fixing the assumptions for decision making, these narratives enable governments and donor agencies to make and implement developing country policy under conditions of extreme uncertainty and complexity". The second is what has been puzzling me since I attended the accreditation course in London this year. Educated in social sciences and in linguistics, I recognize a lot in the course, but wonder why the vocabulary is so different. Have I missed the point somewhere? Is there a difference between an ambiguated filter and a semantic differential? Between tagging and coding? And if so, what is the difference? Or is it all just semantics?

Floor Basten

I garden. More in theory than in practice. But I garden. For years, I’ve harboured hopes of bountiful harvests from my tiny vegetable patch. Several summertime pregnancies and a PhD have led to less than impressive vegetable harvests, though I can’t be accused of not having been productive in other areas. Well, this weekend, chaos happened – a hailstorm struck.  Hailstones as large as pingpongballs redecorated our car bonnet and wiped out the vegetable patch. Depressing. Everything flattened, berry branches snapped, fruit knocked from trees. Clearly not ‘complex’ – no amount of safe-fail experimentation could have helped me see this emergent phenomenon. The crisis task force is me, my husband, a broom, a rake and a secateurs. So far, so good. A one-off chaotic event. But how would I manage if exposed to a series of hailstorms, locust plagues, droughts and rampant mole attacks? Cry, give up, persist?

On May 26-27 2008, Wageningen International and Learning by Design organised an innovation dialogue on ‘navigating complexity’. It sought to articulate insights from the complexity sciences and systems thinking. And examine how these could reshape assumptions about the design, monitoring and evaluation of development work. This event was born out of a growing sense of urgency  about the dilemmas facing the development sector. Development organisations are guided by protocols that are essentially based on a linear approach to meeting objectives and on processes that are assessed in comparative isolation of the dynamic context. Many efforts work with limited and out-of-date insights on their operational contexts. The consequence for those managing development efforts is a reduced capability to guide strategic engagement, which affects effectiveness. Any planning process is based on many assumptions. Some of these assumptions will be quite predictable, while others may be but wishful thinking. Presentations, participant video interviews and other materials can be found here:  (click on Innovation Dialogue for the presentations and video interviews).

Irene Guijt