All linear accelerometers, including the otolith organs, respond equivalently to gravity and linear acceleration. To investigate how the nervous system resolves this ambiguity, we measured perceived roll tilt and reflexive eye movements in humans in the dark using two different centrifugation motion paradigms (fixed radius and variable radius) combined with two different subject orientations (facing-motion and back-to-motion). In the fixed radius trials, the radius at which the subject was seated was held constant while the rotation speed was changed to yield changes in the centrifugal force. In variable radius trials, the rotation speed was held constant while the radius was varied to yield a centrifugal force that nearly duplicated that measured during the fixed radius condition. The total gravito-inertial force (GIF) measured by the otolith organs was nearly identical in the two paradigms; the primary difference was the presence (fixed radius) or absence (variable radius) of yaw rotational cues. We found that the yaw rotational cues had a large statistically significant effect on the time course of perceived tilt, demonstrating that yaw rotational cues contribute substantially to the neural processing of roll tilt. We also found that the orientation of the subject relative to the centripetal acceleration had a dramatic influence on the eye movements measured during fixed radius centrifugation. Specifically, the horizontal vestibuloocular reflex (VOR) measured in our human subjects was always greater when the subject faced the direction of motion than when the subjects had their backs toward the motion during fixed radius rotation. This difference was consistent with the presence of a horizontal translational VOR response induced by the centripetal acceleration. Most importantly, by comparing the perceptual tilt responses to the eye movement responses, we found that the translational VOR component decayed as the subjective tilt indication aligned with the tilt of the GIF. This was true for both the fixed radius and variable radius conditions even though the time course of the responses was significantly different for these two conditions. These findings are consistent with the hypothesis that the nervous system resolves the ambiguous measurements of GIF into neural estimates of gravity and linear acceleration. More generally, these findings are consistent with the hypothesis that the nervous system uses internal models to process and interpret sensory motor cues.