This dissertation investigates the issue whether we can ascribe beliefs to non-human animals that do not possess a language. First, I consider a well know argument by the philosopher Donald Davidson (1917-2003) for the conclusion that we cannot ascribe beliefs to non-human animals, insofar they do not master a natural public language, such as, for example, Croatian or English. Adopting the objections of some philosophers and ethologists, I argue that Davidson’s argument is not convincing. Then, I consider the possibility that be can ascribe beliefs to animals, insofar as they possess an internal language of thought of the type hypothesized by the philosopher Jerry Fodor (1935-2017). I argue that this approach to the ascription of beliefs to animals that do not possess a public language is problematic. While both Davidson and the supporters of the language of thought relate the issue of the ascription of beliefs to non-human animals to their linguistic capacities, the supporters of minimalism offer a different approach. These authors maintain that we can ascribe to animals who lack a language certain form of beliefs that involve certain minimal non-linguistic capacities. However, I will show that there are good arguments for the conclusion that minimalism is not a satisfactory account of animal thought, given that it stresses that animals can only have thought that concern the here-and-now. The last chapter of the dissertation describes José Luiz Bermúdez’s approach to the ascription of thoughts to non-linguistic animals that is based on success semantics. I argue that this theory appears to be supported by certain experiments and offers a viable framework for ascribing certain forms of beliefs to nonlinguistic animals.