The question of how human intelligence evolved has been the focus of a vast amount of research. To investigate the evolutionary trail of our cognition and to better estimate when different cognitive capacities appeared through time, we must look to other species using the comparative approach. Theory of mind is one such cognitive capacity that has a fundamental role in the proper functioning of our social interactions. To possess a ‘theory of mind’ implies the ability to attribute mental states (goals, intentions, knowledge and beliefs) to one’s self and to others. It is suggested that individuals in possession of this capacity can understand the reality of others, at least implicitly, even if that reality differs from one’s own, and use mental states as a basis of interpreting and predicting the behaviour of others. Research efforts over the past 40 years have produced an increasingly complex picture of theory-of-mind related capacities in great apes, suggesting that a cognitive capacity previously thought to be unique to humans may be at least partially shared with animals. My research aimed to investigate, through a variety of approaches, whether bonobos are able to maintain different perspectives of reality; their own and that of others. To address this, I attempted to shed light on three questions. First, how do bonobos perceive themselves in different kinds of self-images and how does experience modify such perception? Second, do bonobos socially learn from others in the context of feeding, even if they have contradictory knowledge about reality compared to what is being demonstrated to them by others? Third, are bonobos able to use previously acquired knowledge about others’ behaviours to make pragmatic inferences when attending to their communication signals? To answer the first question, I investigated how mirror-naïve bonobos reacted to a variety of contingent and non-contingent self-images in the form of video and mirror images, as well as footage of known and unknown conspecifics. I then investigated how the bonobos’ perception of themselves in non-contingent video footage was influenced by prolonged mirror exposure. My findings provide evidence supporting the idea that bonobos do indeed have a sense of self but that this ability develops with experience. To answer the second question, whether bonobos could learn from others through observation despite own contradictory experience, I designed an ecologically relevant social learning experiment that demonstrated that bonobos memorised others’ food preferences and adhered to them, even if they knew viable alternatives. Importantly, socially acquired food preferences remained stable even after subjects became aware that both food types were palatable. To answer my final question, I assessed whether bonobos could use previously acquired social knowledge about others’ food preferences as premises for pragmatic inference when attending to their vocal behaviour. To do this, I ran a playback experiment that demonstrated that bonobos reacted differently to food calls of a demonstrator with a known food preference, compared to the food calls of a control individual with no known food preference. While subjects were not able to integrate all the information made available to them, they were able to integrate enough to draw inference from the two call providers’ underlying difference in knowledge. In sum, my findings provide ground to suggest that our last common ancestor was already in possession of a number of cognitive faculties that form the foundation of human intelligence. These findings, that range from perceiving one’s own reality through a sense of self, to understanding the reality of others and incorporating their knowledge of reality, to finally using their knowledge of others’ reality to make inferences of their communication signals, have provided, I hope, a further glimpse into the bonobo mind and, by extension, that of our common primate ancestors.