Magnetic resonance imaging has been applied only recently to the study of Arabic speech production. Arabic has a relatively large number of sounds produced with constrictions in the pharynx, a part of the vocal anatomy well-suited to investigation using MRI. We show that static 3D MRI techniques can be useful in distinguishing the pharyngeal sounds of Arabic and that average pixel intensity in MR images can be used to track pharyngeal articulations as a function of time.