A consistent finding of recent research on respiratory muscle training (RMT) in healthy humans has been an attenuation of respiratory discomfort (dyspnoea) during exercise. We argue that the neurophysiology of dyspnoea can be explained in terms of Cambell's paradigm of length-tension inappropriateness. In the context of this paradigm, changes in the contractile properties of the respiratory muscles modify the intensity of dyspnoea predominantly by changing the required level of motor outflow to these respiratory muscles. Thus, factors that impair the contractile properties of the respiratory muscles (e.g. the pattern of tension development, functional weakening and fatigue) have the potential to increase the intensity of dyspnoea, while factors that improve the contractile properties of these respiratory muscles (e.g. RMT) have the potential to reduce the intensity of dyspnoea. In patients with obstructive pulmonary disease, functional weakening of the inspiratory muscles in response to dynamic lung hyperinflation appears to be a central component of dyspnoea. A decrease in the intensity of respiratory effort sensation (during exercise and loaded breathing) has been observed in both healthy individuals and patients with obstructive pulmonary disease after RMT. We conclude that RMT has the potential to reduce the severity of dyspnoea in healthy individuals and in patients with obstructive pulmonary disease, and that this probably occurs via a reduction in the level of motor outflow. Further work is required to clarify the role of RMT in the management of other disease conditions in which the function of the respiratory muscles is impaired, or the loads that they must overcome are elevated (e.g. cardiorespiratory and neuromuscular disorders).