DescriptionThis dissertation starts with the premise that a well-functioning learning organization should be able to effectively and efficiently resolve both known-unknowns as well as unknown-unknowns. Furthermore it takes a bottom-up, emergent perspective on organizational learning by assuming that this is done, in large part, by organizing for and encouraging various forms of knowledge exchange interactions among its members (e.g., Argote, 1999). By reviewing a number of different literature streams (e.g,. knowledge search-transfer, advice sharing-acceptance, innovation championing-adoption), I highlight that each focuses on a different form of knowledge exchange. Furthermore, I suggest that these literatures have implicitly assumed that whether a recipient or a source initiated an exchange corresponded to whether the exchange and the knowledge it involved was solicited / sought (and thus likely to only resolve known-unknowns) or unsolicited / unsought (and thus offers the potential to resolve unknown-unknowns). However, I argue, initiation in modern, complex, knowledge-based organizations is frequently mutual or coincidental and thus may be a poor proxy for unsolicited / unsought knowledge. In order to understand instances of unsolicited / unsought knowledge across all forms of knowledge exchange, I propose that knowledge exchange interactions can be contextualized within a recipient’s overall problem-solving process. By contextualizing knowledge-exchanges within a multi-phase problem-solving process (e.g., problem formulation, problem validation, solution formulation, solution validation), I am able to examine where recipients are cognitively when they start an interaction as well as the implications for the type of knowledge provided by sources during the interaction. A survey of over 1200 respondents describing over 700 knowledge exchange interactions at four multinational Research and Development companies provided evidence of my propositions. In each of three sections / studies, I debunk what I argue are assumptions built into literature focused on either source- or recipient-initiated exchanges. Collectively my results seem to suggest that initiation is not particularly relevant for differentiating the type of knowledge exchange (or more precisely whether an exchange may resolve unknown-unknowns) and highlight unsolicited / unsought knowledge as a more relevant construct.