The paper reviews theoretical arguments and empirical findings in support and against the idea that cognitive processes (thinking, reasoning, expectations, and beliefs) or emotions and motivation can directly influence and change the content of vision. According to the modularity of mind hypothesis and Marr's computational theory, vision is an informationally encapsulated module with fixed, innate architecture. This innate architecture is based on a specific set of principles which are much different from general cognitive functioning. Vision must be cognitively impenetrable because its task is to create an exact mental representation of the external environment that will enable individual to navigate through it successfully. In contrast, according to predictive coding model as an example of the modern advocate of penetrability of vision, brain constantly generates predictions which facilitate and redirect visual information processing and, consequently, alters what we see. Recently, many behavioural and brain studies revealed that vision is indeed altered under the influence of learned associations and context as well as social cognition, motivation and emotions, thus lending support to the claim that vision is cognitively penetrable. However, numerous methodological and interpretational pitfalls have been identified in aforementioned research, resulting in the opposite conclusion. At the end, we described potential avenues for further theoretical and empirical research which will bring us closer to the answer to this perplexing question.
