Research on engagement has exploded in recent years. It is of interest in almost any domain of human interaction from workplaces to schools, to business, to games, and many more besides. However, measuring engagement is difficult. For the most part researchers must ask participants about their engagement (which interrupts the engagement) or bring them into the lab. Here we develop a domain-neutral approach to engagement, drawing on philosophical phenomenology and complex embodied approaches to cognition. We also develop a new paradigm for analyzing mouse traces from this perspective. A phenomenological theory of engagement is presented, and two experiments are described, both of which provide evidence for that theory.

In Chapter 1, a definition for problem is introduced, along with a model of the problem cycle, which is used to conceptualize the relationship between normal cognition, problem finding and problem solving. Initial experimental efforts to explore problem finding are discussed. In Chapter 2, previous work on problem finding and problem solving is reviewed. The few existing accounts of the stages of problem solving are surveyed, and the older literature on different kinds of problems and how they are solved is reviewed in detail. Chapter 3 builds upon the first two chapters by introducing a new taxonomy of problems, highlighting four dimensions that differentiate problem types and explaining each type at length. Chapter 4 describes an experiment developed to test the idea that problems can be organized along the lines of the taxonomy introduced in Chapter 3, and that these differences determine how they try to solve a specific instance of a problem. A simple game is developed and behaviors in the game are tracked, confirming the hypothesis that different problem types are solved in different ways, that are specific to the type of information available to the problem solver. The last chapter summarizes the dissertation and describes avenues for future research.

In this thesis, I examine how human experience (thoughts, feelings) has been treated in archaeology and how phenomenology is traditionally employed in archaeological research. Phenomenological approaches have been heavily critiqued as lacking in rigor. Many argue that the differences between past and present human experience is too vast to allow for valid interpretations. I propose that humans in the past and present are similar enough to allow for plausible inferences. Further, I argue that phenomenological interpretations of the past are special forms of analogies, which are accepted by many within the archaeological community. To bolster these interpretations, new theories in cognitive science support the validity of phenomenological inference. Based on those insights I elaborate on methodology for constructing and evaluating phenomenological analogs in archaeological interpretation.