Knowledge transfer without knowledge? The case of agentive metaphors in biology

Are scientific metaphors dispensable shortcuts that encapsulate knowledge but can always be translated back? Or do they constitute cases of knowledge transfer, even if seemingly based on scientifically underdeveloped domains? This paper defends the latter view. By drawing on the linguistic theories of metaphors, we assess a variety of agentive metaphors that pervade biology. Intentional metaphors are found unsatisfying because their use is either rigid or too widely flexible. By contrast, rational agent metaphors constitute good scientific metaphors, displaying flexible use and heuristic fruitfulness. Their range of application constantly evolves because they provide guidelines that permit the exploration of the applicability of source domain in a target domain. Their unique heuristic value makes them akin to research programs and allows for knowledge transfer, because they are based on a proper scientific source domain rather than on a folk or underdeveloped one.