Human social interaction critically depends on theory of mind, the ability to reason about behavior in terms of mental states. The goal of the current research program is to assess what kinds of special-purpose neurocognitive mechanisms support theory of mind. Three theories are considered. The first theory proposes that mechanisms are specialized not for theory of mind in particular but more generally for metarepresentation, the ability to reason about representations, including mental (e.g., belief, desire, and intention) and non-mental representations (e.g., maps, signs, and notes). The second theory proposes that mechanisms are specialized for theory of mind, selectively processing mental states and leaving the processing of non-mental representations to other mechanisms. The third theory proposes that mechanisms are specialized for theory of mind, but are co-opted to process non-mental representations owing to non-mental representations sufficiently resembling mental states, allowing them to also activate the theory of mind system.

A fundamental distinction between the proper and actual domain of a specialized mechanism reveals how the three theories are related. The proper domain refers to the inputs a mechanism is designed to process; the actual domain refers to the inputs a mechanism is capable of processing (Sperber, 1994). The first theory proposes the mental and non-mental representations belong to the proper domain of a single mechanism. The second theory proposes that mental states belong to the proper domain of one mechanism and that non-mental representations belong to the proper domain of a distinct mechanism. The third "co-option" theory claims that mental states constitute the proper domain of a theory of mind mechanism but that non-mental representations, because they are also capable of triggering the system, belong to the actual domain of that same mechanism.

The following chapters explore the extent of domain-specificity for mechanisms underlying theory of mind along the proper and actual domain dimensions. Chapter 1 sets the stage for the subsequent chapters, reviewing some preliminaries, highlighting the importance of theory of mind, and providing theoretical background. Chapter 2 presents four reaction time experiments that compared reasoning about beliefs to maps, signs, and notes (under two conditions), testing the scope of the proper domain for mechanisms governing theory of mind. Chapter 3 presents two reaction time priming experiments and one repetition suppression fMRI study assessing the scope of the actual domain of mechanisms supporting theory of mind specializations. Finally, Chapter 4 summarizes the findings, weighs them against the theories under consideration, and considers several outstanding issues.