Rhona P. Hellman Northeastern University
Thirty-five years have elapsed since S. S. Stevens ( 1955) showed that, over a stimulus range from about 40 to 140 dB, loudness is a power function of sound intensity with a slope (exponent) of 0.3 (0.6 re: sound pressure). Stevens determined the form of the loudness function from a histogram of 178 sets of data obtained by halving and doubling and magnitude estimation procedures. Despite the accumulation of evidence from many laboratories and countries, including the impressive predictive relations obtained from cross-modality matching ( Stevens, 1959, 1975), considerable disagreement about the validity of the 0.3 power law for loudness persists. This disagreement concerns both the procedures used to measure loudness and the underlying meaning of the data (for reviews see Falmagne, 1986; Gescheider , 1988; Luce & Krumhansl, 1988; Marks, 1974). The controversy centers mainly on two issues: (a) the specificity of the slope and (b) the relevance of the measured loudness function to the intensity characteristic of the auditory system. In light of this controversy, both issues are reassessed.
Psychophysics, unlike the natural sciences, must rely on the human observer for a response. Because the observer can easily be influenced by subtle cues, psychophysical judgments of sensation magnitude are fre-