Are responses in avoidance procedures "safety" signals?

Dinsmoor's (2001) position has the advantage of parsimony in that it relies on well-established principles rather than a separate process--shock-frequency reduction--to account for avoidance. Other advantages are that it blends well with what is known about the effectiveness of momentary contiguities in the study of positive reinforcement and that it might provide an account of why different response forms seem to condition at different rates. Despite these advantages, the view needs elaboration about the temporal characteristics of response-associated stimuli, the functions that "warning'' stimuli may have, and especially about how ''safety" is established.