The present paper is an exercise in the description of actant structures in Spanishpossessor raising constructions in the framework of the Meaning-Text Theory (MTT). We argue that there is a significant difference between constructions in which the possessor is incorporated into the meaning of the main verb and constructions in which it is not. This difference explains the availability or lack of certain alternations for specific constructions ; it also leads to different semantic and syntactic actant structures of the main verb. We show that the multilevel linguistic model of the MTT allows for a distinct representation of each type of construction, and that the concepts of Government Pattern and transduction underlying MTT facilitate an adequate representation of the projection of the structures of possessor raising constructions between the different levels of the MTT model. A particular role in these projections is assigned to another central feature of the MTT : the semantic communicative structure.