Search

To smoothly interact with other people requires individuals to generate appropriate responses based on other’s mental states. The ability we rely on is termed mentalizing. As humans it seems that we are endowed with the abilities to rapidly process other’s mental states, either by taking their perspectives or using mindreading skills. These abilities allow us to go beyond our direct experience of reality and to see or infer some of the contents of another’s mental world. Due to the complexity of social contexts, our mentalizing system needs to address a variety of challenges which put different requirements on either time or flexibility. During years of research, investigators have come up with various theories to explain how we cope with these challenges. Among them, the two-system account raised up by Apperly and colleagues (2010) has been favoured by many studies. Concisely, the two-system account claims that we have a fast-initiated mentalizing system which guarantees us to make quick judgments with limited cognitive resource; and a flexible system which allows deliberate thinking and enables mentalizing to generalize to multiple targets. Such a framework provides good explanations to debates such as whether preverbal young children can process mentalizing or not. But it is still largely unknown how healthy adults engage in mentalizing in everyday life. Specifically, why it seems easier for some targets to activate our mentalizing system, but with some others, we frequently fail to consider their perspectives or beliefs? To give an explanation to this question, I adopted a different research orientation in my PhD from the two-system account, which considers the dynamic interactions among three key elements in mentalizing: the self, agent(s), and object(s). I put forward a mentalizing triangle model and assume the interactions in these triadic relationships act as gateways triggering mentalizing. Thus, with some agents, we feel more intimate with them, which makes it easier for us to think about their minds. Similarly, in certain context, the agent may have frequent interactions with the object, thus we become more motivated to engage in mentalizing. In the following chapters, I first reviewed current literatures and illustrate evidence that could support or oppose the triangle model, then examined these triangle hypotheses both from behavioural and neuroimaging levels. In Study 1, I first measured mentalizing in the baseline condition where no interaction in the triangle relationships was provided. By adapting the false belief paradigm used by Kovacs, Teglas, & Endress (2010), I imported the Signal Detection theory to obtain more indices which could reflect participants mentalizing processes. Results of this study showed that people have a weak tendency to ascribe other’s beliefs when there is no interaction. Then, in Study 2, we added another condition which included the ‘agent-object’ interaction factor while using a similar paradigm in Study 1. Results in the noninteractiond condition replicated our findings of Study 1, but adding ‘agent-object’ interactions didn’t boost mentalizing. Study 3 and 4 tested the ‘self-agent’ interaction hypothesis in visual perspective taking (VPT), another basic mentalizing ability. In Study 3, I adopted virtual reality approach and for the first time investigated how people select which perspective to take when exposed to multiple conflicting perspectives. Importantly, I examined whether the propensity to engage in VPT is correlated with how we perceive other people as humans, i.e. the humanization process. Congruent with our hypotheses, participant exhibited stronger propensity to take a more humanised agent’s perspective. Then in Study 4, I used functional near-infrared spectroscopy (fNIRS) and investigated the neural mechanism underlying this finding. In general, the ‘selfagent’ hypothesis in the mentalizing triangle model was supported but not for the ‘agentobject’ hypothesis, which we consider may due to several approach limitations. The findings in this thesis are derived from applying novel approaches to classic experimental paradigms, and have shown the potentials of using new techniques, such as VR and fNIRS, in investigating the philosophical question of mentalizing. It also enlights social cognitive studies by considering classic psychological methods such as the Signal Detection Theory in future research.