In the context of remote medical monitoring of swallowing, the authors investigate the potential of non-invasive respiratory inductance plethysmography (RIP) technique to automatically detect swallow events in a wide context of respiration and phonation. Signal acquisitions were carried out on 5 healthy volunteers equipped with RIP and electroglottograph as swallowing reference. They were asked for spontaneous breathing, speaking and diverse bolus ingesting. The RIP signal was then segmented into cycles, each cycle being annotated according to one of the three class of interest, respectively ventilation i.e. spontaneous breathing (1257 cycles), swallowing (221) and phonation (216). Automated classification was performed using quadratic discriminant analysis. Focusing on swallowing class, the authors achieve an accuracy of 79% from the full wide protocol. It increases up to 86% with prior removal of vocalizations. These preliminary results in healthy subjects make RIP a promising candidate as a non-invasive and convenient technology for medical follow-up of swallowing.