Second language acquisition (SLA) researchers measure linguistic complexity to assess pedagogical effectiveness and depict development (Norris & Ortega, 2009). Yet, from linguistic and cognitive perspectives, commonly used approaches oversimplify complexity. Furthermore, such approaches do not consider the morphological complexities of a highly inflected L2 like Spanish. Norris and Ortega (2009) encourage SLA researchers to develop empirically and theoretically valid measures of linguistic complexity through an organic (i.e., iterative and data-driven) investigative process; resulting models should be multidimensional and developmentally sensitive. This study delineates three multidimensional models of Spanish L2 linguistic complexity based on a principal component analysis of a corpus of learners participating in a task-based activity at the beginning, intermediate, and advanced levels of university instruction.