This paper presents a computational model of visual association- making. Our model focuses on the interaction between the processes of representation and matching in association. Re-interpretation of the objects being associated changes the landscape in which the matching process is searching for potential mappings between those objects. We call this process interpretation-driven search. We demonstrate the capabilities of our system through some examples of previous work in simple shape domains, then discuss ongoing research into applying this system to design domains.