This thesis seeks to answer a basic question of the philosophy of art: What is art? After examining the answers provided to this question by several art theories, it firstly proposes that the definition of art offered by the philosopher Arthur Danto is the one that most conform to the reality of art, although it suffers from certain imperfections and insufficiencies. On the basis of this critical conclusion as well as of a careful analysis of what is categorized as art in the present time, it attempts to modify Danto’s definition and to transfigure it into a new one, capable of better conforming to the reality of art phenomena. It consists of five chapters. The first chapter introduces briefly the contemporary situation of art and the existing art theories which didn’t success to do sufficiently justice, in my opinion, to the totality of what is considered as artistic. Through a critical examination of the definitions of art proposed by these theories, it tries to show that the sufficient and necessary conditions for being art that they propose are either not sufficient or not necessary. And consequently, that the problem of defining art remains unsolved. The second chapter tries to clear up the obstacles to the possibility of providing a definition of art that have been put forward by anti-essentialist theorists of art, who claim that art cannot be defined in a classical way, and to show accordingly that that a definition of art in the form of necessary and sufficient conditions is still possible. The main argument of the anti-essentialists is based on the concept of “family resemblance” proposed by Wittgenstein, according to which there is no common feature to all artworks, only similarities of various degrees between them. This argument is refuted by using Danto’s claim that the artistic character of an artwork lies in the relationship in which it is involved and cannot be captured in its appearances. The third chapter is a comprehensive presentation of Danto’s definition of art. Danto starts by comparing a pair of works sharing the same appearances, while one of them only is an artwork and the other is not. He concludes that the artwork is the representation of a subject while its non artistic twin is not. Danto then compares a mere representation like a diagram with the artwork so conceived, and proposes that the specificity of the artistic representation consists in being a representation with a metaphorical ellipsis, while the diagrammatic representation fails to be metaphorical. He claims furthermore that the artwork is a representation endowed with style. He also introduces the concept of a “style matrix”, which consists in a set of predicates about art. Danto’s idea is that each artwork represents the art history by consciously manifesting or rejecting these predicates, or even adding new ones, in order to locate itself in the historical evolution of art. The fourth chapter presents the main criticisms proposed to this day against Danto’s definition. Some of these criticisms are judged to be unfair and consequently rejected, leading to a defense of the validity of several key aspects of Danto’s theory; others are considered as correct. In addition, the chapter confronts Danto’s theory with the reality of the present art scene and formulates additional criticisms that have not been put so far forward in the Danto scholarship. The most important one is that his concept of style reveals inconsistencies throughout the various presentations of his theory. Another one, of lesser although still substantial importance, is that representing the art history is neither necessary nor sufficient for being an artwork, since there are plenty of artworks that don’t acquire historical significance by consciously representing the art history. The fifth chapter, which is the most important one of the whole thesis, endeavors to elaborate the basic elements of a new definition of art based on the acceptable parts of Danto’s definition: that is to say, the claim that art is a metaphorical kind of representation. This definition considers the artwork as the outcome of an artistic type of action, this outcome being sometimes the transitive product of these actions, sometimes these actions themselves. An artistic action is itself analyzed as an action intending to metaphorically represent something. Through the study of different examples, it is shown that artworks are artistic in varying degrees: some works are more artistic than others. This gradational phenomenon is the result of the possibility that the specific intention of an artistic action interferes with other non artistic intentions, as well as additional contextual conditions, such as the lack of skill to fully realize one’s artistic intention. This action-based definition of art provides a new perspective onto the entire field of art research by transforming the art-related problems into specific types of action-related problems. This is the reason why it also mixes elements of Danto’s definition with elements of Searle’s theory of action, although it explicitly acknowledges that its further refinement demands a thorough grounding on a more detailed action theory. Finally, strongly consonant with Danto, this definition breaks away with him by defending the idea that it is not insuperable, because a theory of art should always keep up with the evolving reality of art and that this evolution is not over, even though it has reached a fairly stable phase.