According to the NSF (5), “Today's grand challenges will not be solved by one discipline alone. But the integration of knowledge, methods and expertise from across science and engineering is not simple or automatic”. This is not news. What is new, is an increasing emphasis on addressing grand challenges with non-academic partners, such as industry or community organizations. We argue that such partnerships can put significant stress on core scientific commitments regarding the kind of knowledge making (KM) that is most valuable. That stress is better managed if the commitments of the competing KM standards are made explicit so that they can be negotiated.

While the exact nature of science’s core commitments regarding KM is subject to debate (6) we can draw on the work of Karl Popper (2) for a simple model that captures a set of commitments that are widespread in the sciences. For Popper, scientific knowledge consists of testable generalizations arrived at by a process of bold conjectures and refutations. That is, scientific KM involves practitioners generating unlikely hypotheses which are tested as stringently as possible in order to generate claims that apply widely (e.g. to many different times and places) and which significantly advance our understanding of the world.

It is convenient to think of KM with non-academic partners as falling into two main categories; partnering with industry and partnering with communities. Either possibility can give rise to tensions with the commitments of the Popperian model of KM.

An important component of KM with industry partners is ‘de-risking’ (3,4). That is, roughly, good industrial KM can require pursuing conservative strategies that raise the likelihood of successful commercialization at the cost of missing out on the most innovative possible solutions. This is in tension with the Popperian norm of ‘bold conjectures’.

Working with community partners can often involve addressing ‘wicked problems’ (1), such as how to route an interstate through an urban area. Such problems involve contested framings of the issues, incompatible stakeholder values, and high sensitivity to local context among other features. As a result any ‘solution’ to such problems will be extremely specific and untestable. Again this sort of work is in tension with Popperian norms, in this case the norm of creating general testable knowledge.

If these tensions in norms of KM are not addressed, collaborations between scientific and non-academic partners can become incoherent with the parties striving for incompatible forms of knowledge.

We conclude our presentation with a discussion of mechanisms for uncovering these sorts of tensions (e.g. the Toolbox Dialogue Initiative) and suggestions about how to manage them. For example, scientists cannot propose ‘solutions’ to the wicked problems communities face, rather, they must negotiate how they contribute resources (e.g. models, data sets, etc.) and use trained judgment (7) in co-identifying ‘solutions’.. This allows both groups to operate in ways that are mutually beneficial but also consistent with their differing norms.