This paper lends support to Richard E. Clark's position that media make no difference in student achievement, but also argues that alternative ways of thinking about media--i.e., about the differences media do make along with the differences they can make--are necessary and important. Two major areas of study are examined. In discussing the first area--cognitive and affective learning outcomes of the use of different media for the delivery of instruction--it is pointed out that there may be no differences in learning achievement from a cognitive point of view, but there could be large differences in attitude and motivation that might interfere with learning of cognitive tasks from an affective point of view. In considering the second area--the differences between common usage and the potential of a medium--it is argued that each medium is a combination of attributes capable of carrying various combinations of different symbol systems, and that only to the extent that media carry identical symbol systems relating to "critical" content can they be interchangeable. It is suggested that neurological theories of how the brain functions promise an immense contribution to the psychological study of how people learn, and that direct exploration of new cognitive constructs along with psychophysics might strengthen the bond between cognitive psychological research, neurological brain research, and instructional design theories. This approach, it is argued, would lead to more accurate descriptions of the learning process, new constructs for research, and answers not to whether singers or songs are better, but why. (6 references) (CGD)