The problem is basically that any mechanical process of cause and effect may be reinterpreted as a "goal-directed" process along Taylorian lines by simply supposing that whatever happens to be its outcome has all along been the "end" toward which the events comprising the process have been "directed".