THE LATE Peter Drucker claimed to have invented the term ‘the knowledge economy’ in the 1960s. Whereas industrial-age workers toiled with their hands and produced ‘stuff’, he posited, today’s employees work with their brains and produce ideas and knowledge. Even in the material world, success has increasingly come to depend on knowledge: an iPod (and Apple) or Yaris (and Toyota) win by delivering more, and better, crystallised ideas per buck than rivals.

But in other domains ideas about managing knowledge still seem stuck in the industrial age. In 1994, an influential publication distinguished two main approaches to the creation of knowledge. ‘Mode 1’ regards it as an economic commodity – dematerialised ‘stuff’ – and treats it as a production-line issue: ramping up production of ideas in university research departments (increasingly funded by industry), handing them on to company development labs through spin-outs and knowledge transfer activities, where they are translated into marketable products or services.

‘Mode 1’ is ‘hard’, linear, and simple. That’s why government and many business people like it. It has become pervasive, particularly in the public sector. University research and teaching assessments are based on this approach to knowledge, of which Dickens’ oppressive schoolmaster Thomas Gradgrind would have been proud. Even in the humanities, whole areas of study are increasingly specified by the funding research councils.

However, there is increasing evidence that Mode 1 is not just simple: for any of the messy, boundary-spanning issues we face – the ageing of the population, climate change, sustainable energy and global security, funding for which the Department for Innovation, Universities and Skills (DIUS) proudly announced last week – it is simplistic. For these kinds of problems, we need ‘Mode 2’.

Whereas ‘Mode 1’s’ vision of a lab-to-market supply chain is a closed innovation model, Mode 2 is open, diffuse – and much more difficult to command. Knowledge is not produced simply by researchers but by the collision of many scientific, professional, managerial and societal influences, each with its own distinctive angle. Mode 2 is not simply about producing knowledge: it is about absorbing, applying and creatively modifying it into often wholly unpredictable innovations. The iPod, Linux and Google are Mode 2 innovations, products of a melting pot rather than a machine, and of social ties rather than instrumental relationships. This is why it often happens in spontaneous clusters, such as Silicon Valley or Silicon Fen.

How incompatible Mode 1 mechanisms are to Mode 2 initiatives emerges from a study of the government’s Genetic Knowledge Parks (GKP) initiative, by the Evolution of Business Knowledge programme of the Economic and Social Research Council. This had brave beginnings. It emerged in around 2001 from concerns about scientific directions in the wake of BSE and GM foods. Genetic knowledge was obviously important across the biomedical field a correspondingly wide range of stakeholders were co-opted into the arrangements for the parks.

Unfortunately, with no clear lead from the centre, different departments imposed conflicting agendas and priorities. Then the funding process for the new organisations led to fierce competition between existing research centres that had previously co-operated as part of a dispersed, informal research network spread over a number of universities and institutions. Worse, under the new rules the groupings had to be regional, cutting across organic national and even international ties.

The centre’s desire for ‘accountability’ led to classic Mode 1 monitoring and control arrangements, complete with targets and standards (‘name the major scientific advances you have made in the past 12 months…’). Not surprisingly, these measures created fear and loathing among researchers and a crisis of governance and quality control at the centre, leading to pressures for even more monitoring and control. So the GKPs turned inward to concentrate on their own survival, further undermining collaboration. After five years and pounds 15m, the government pulled the funding plug.

This tale has powerful implications for many other areas of public and private endeavour – indeed for every domain where part of the object is to discover and learn from ‘what works’. That might include, for example, the NHS as well as the ‘big issues’ singled out by the DIUS. In all these, the challenge is not so much creating brilliant new science but getting different groups to work together to shape the ideas and bring them to practical fruition. Using the Gradgrind model in such circumstances is worse than ineffective it can actually destroy the capacity for innovation. The knowledge economy cannot be run on 19th-century management lines. It’s not just knowledge creation that urgently needs a Mode 2.

The Observer, 23 December 2007