Aftereffects of azimuthal auditory motion may have two components. A sensory component is inferred from strong aftereffects, because they are spectrally dependent and have shallower response functions than those for non-adaptation. Neither property applies to weak aftereffects, suggesting a cognitive component. Two experiments determined whether changing-loudness aftereffects (CLA) might be understood similarly. In a single-interval forced-choice procedure, listeners responded "growing softer" or "growing louder" to test stimuli changing in intensity. In Exp. 1, adapting and test stimuli were diotic and had the same 1-kHz sinusoidal carrier. Although response functions following adaptation were displaced from response functions for non-adaptation -indicating CLA -their slopes were broadly similar. In Exp. 2, stimuli were monotic; adapting frequency was 1 kHz and test frequencies were between 0.5 and 2.0 kHz. CLA was present in most adaptation conditions, but was strongest when the test frequency was 1.0 kHz; functions' slopes again evinced no systematic variation. The two-component hypothesis for CLA is supported by spectral dependence alone. It is argued that the slope of response functions is due to the nulling procedures for measuring auditory aftereffects. The slope depends on whether the adapted property is processed by "direct" and "indirect" mechanisms; aftereffects tap "direct" mechanisms alone, which may affect sensitivity during measurement.