In 1968, Nelson Goodman published his Languages of Art which became one of the most important works on aesthetics in the tradition of analytical philosophy. Goodman offered there a semiotic theory of art, the purpose of which was to explicate our concept of art in terms of different symbolic or referential functions; the theory was further developed in his subsequent book Ways of Worldmaking. Though it is a very subtle and sophisticated theory, I will argue that it is not adequate, that it doesn't even satisfy Goodman's own requirements for a theory of art. Traditionally, the aim of a theory of art has been to capture the essence of art, to describe what is common to all works of art and at the same time distinguishes them from everything else. As a nominalist, Goodman does not, however, believe that there is any essence of art that we can discover.