Lindblom et al.’s biteblock experiments show unequivocally that the formant structure of vowels is essentially unchanged when the speaker has a biteblock between his teeth. This must involve compensatory positioning of the tongue since formant structures would be substantially different in the biteblock condition. This compensation is there on the very first trial: thus, it cannot be based on auditory feedback.
Lindblom et al. (1979) assume that the brain is informed about the state of the vocal tract by proprioceptive feedback. This tells the brain what the shape of the tract is like. The brain can then compute the sensory effect when a tract of that shape is used as a resonator. If this sensory representation is critically different from the sensory target, the speaker can adapt the shape of the vocal tract and compute the sensory representation that goes with this new constellation.
Lindblom, B., Lubker, J. & Gay, T. 1979. Formant frequencies of some fixed-mandible vowels and a model of speech motor programming by predictive simulation. Journal of Phonetics 7, 147-161.