Stretch reflex is known as one mechanism that contributes to the stability of human sensorimotor control. When a muscle is stretched by unknown disturbance, the muscle makes a quick contraction to maintain the posture (and then keep its length constant). Although this quick reflex has been quite extensively investigated in limb system, the extent to which it contributes to the motor control of the tongue remains unknown. While a previous study failed to induce stretch reflex in the tongue (Neilson et al., 1979), we have found behavioral evidence that the tongue shows relatively quick reaction of motion in response to external mechanical perturbation (Ito et al., 2020). Although the latency of this behavioral response was in the range of reflex delay, the involvement of reflex mechanism is still uncertain due to lack of neurophysiological evidence. This study aims at clarifying this issue with the recording of the muscle activity induced by a quick stretch of the tongue during speaking. In the test, the tongue stretch was produced using a precisely-controlled robotic device. Electromyographic response (EMG) was recorded from the anterior part of the mouth floor using a uni-polar surface Ag-AgCl electrode as done in Ishiwata et al., (1997). The recorded EMG signal contains muscle activation in the Genioglossus and of the Geniohyoid muscles. These muscles are involved in the production of the vowel /i/ (Miyawaki et al., 1975; Baer et al., 1988; Buchaillard et al., 2009). Since our previous study (Ito et al., 2020) also found a clear compensatory response in the production of vowel /i/, this study focuses on the production of vowel /i/. Participants were asked to sustain this vowel for 3 s. In addition to this speech condition, we also carried out a non-speech condition in which participants were asked to produce the same magnitude of muscle activation, but without articulating any speech sound, and a rest condition associated with little or no muscle activation. For comparison with the latency of voluntary reaction, we also carried out a reaction task, in which the subjects were asked to produce a strong muscle activation once they perceived the perturbation force during the production of /i/. Our preliminary result showed that the increase of muscle activation in response to the perturbation force was induced in the periods corresponding to the latency of compensatory response observed in our previous behavioral study. The magnitude of the response was larger in the speech condition than in the non-speech condition. The response was also induced earlier than the voluntary reaction task. These preliminary results are consistent with the hypothesis that the observed response is driven by reflex mechanism. Since the latency of the observed response was comparable with the cortical reflex for lip compensation in speech (Ito et al., 2005), the current reflex may also be mediated within the cortical loop.