Theory of Mind (ToM) presents perspective-taking as an ability to infer another person’s mental state. Researchers in various fields have searched for evidence of ToM in both human and non-human subjects; however, this approach is highly “mentalistic,” and rarely scrutinized in the relevant literature. Attempts to identify the behavioural functions of the stimuli involved in perspective taking are still in their infancy. My primary aim was to evaluate and further refine existing behavioural approaches to perspective taking. I focused on the environmental settings required for a person to engage in deriving a perspective of another person, and how the acquired stimulus functions are generalised to novel contexts. First, I examined the integrity of the concept of “deictic framing,” which proposes that deictic expressions (I-You, Here-There, Now-Then) are the core components of perspective-taking behaviour, as defined within relational frame theory (RFT). Participants were first trained with the existing deictic-framing protocol with and without the deictic expressions. I then measured response generalisation using two different tasks designed to measure perspective taking: a visuospatial and an implicit relational assessment procedure (IRAP). No evidence of response generalisation to the other tasks was observed. Additionally, there was no difference in the performance of the participants in the two experimental groups, suggesting that deictic expressions do not have their claimed core status in perspective taking. Secondly, I examined the Relational Triangulation (RT) framework to investigate whether a specific stimulus function (i.e., perspective) would be derived in accordance with contextual stimuli bearing same and oppositional relational properties. The findings supported the broader applicability of the RT framework; however, there are many perspective-taking tasks that are beyond its scope. After considering the findings from these two experiments, I developed the proposition that we can treat perspective taking as problem-solving behaviour because there are considerable overlaps between the core behaviour relevant to various perspective-taking and problem-solving tasks, including stimulus generalisation of relevant stimuli, simple and complex discriminative stimulus functions, and relational responses. Based on this conceptual analysis, I hypothesised that a common false-belief task, the Sally-Anne test, can be represented and evaluated purely by using the concept of stimulus control. I used a conditional discrimination task in which reinforcement for identifying an individual’s false belief depended on a stimulus change in a given setting. Participants completed a non-verbal computer task and selected a particular stimulus dependent on certain stimulus changes, replicating the conditional discriminative stimuli available in this false-belief task. Approximately 61% of participants in Experiment 1 and 88% in Experiment 2 successfully discriminated the pattern of stimulus changes. In conclusion, analysis of the stimulus functions involved in deriving another’s perspective suggests that mind-reading is in fact a type of behaviour-reading. With this functional definition, we can help to address the practical challenge of resolving perspective-taking deficits in a range of contexts.