A new paradigm of ‘process knowledge’ has emerged in recent years, distinct from the dominant ‘product knowledge’ paradigm in social work. While the latter refers to existent knowledge, which may be applied, the former refers to the development of knowledge about the ‘methodology of practice decision making’, focusing on the processes by which judgements are made. At its heart is the emerging idea of a reflexivity for practice, but studies have been, until recently, theoretical. A very small number of empirical studies have begun to identify some key elements of process knowledge. These have developed a range of concepts relating to critical appraisal, hypothesis development and hypothesis testing which characterize social work process knowledge. These include, for example, focused attention, querying information, causal inferences, partial case, procedural, and speculative hypotheses. A process of quasi triangulation is characteristic of social work practice methodology, together with a combination of inductive and deductive thinking. This paper seeks to develop our understanding by focusing on how social workers develop and appraise hypotheses, and in particular how the substantive content of hypotheses emerges (these enable social workers to make sense of, define, and respond to, situations). To understand this, the notion of rules is used, and social workers emerge as analysts employing three types of rules: substantive, application, and practice rules. A significant link between process knowledge and the content of product knowledge is identified in the form of technical language. The concept of ‘probabilistic causation’ is identified as a key epistemological dimension in the conduct of rigorous practice.