There are a number of unanswered questions about Olivier Messiaen’s use of ethnological rhythmic fragments; in particular, Śārṅgadeva’s set of deçi-tâlas—a collection of 120+ popular provincial rhythms collected in 13th century India—and Greek prosodic feet. Although the composer discusses the fragments at length in his Traité de Rythme, de Couleur, d’Ornithologie (1995), answering large scale music-theoretic and musicological questions about his use of such rhythmic forms is difficult, if not impossible, by hand. To understand the latent “langage musical” of Messiaen, we use computational tools adopted from natural language language processing and computational musicology (supported by the music21 framework). Specifically, we use pattern recognition techniques to identify the locations in compositions in which deçi-tâlas and prosodic feet are included, and to identify cross-corpus consistencies in his use of such forms (e.g. pitch contour).