The purpose of this study is to investigate the flow of nonsymbolic auditory information in humans. To achieve this goal the observer adjusts the loudness of a Gaussian noise (GN) signal to match the perceived intensity of an electrodermal stimulus (EDS), a cross-modality matching task. The loudness of the GN and the voltage of the EDS mere maintained within the observer's (O's) sensory probabilistic zone. The sensory probabilistic zone is a range of stimulus intensity from somewhere above "no detection" to "100% detection". The independent variable is a binary condition in that it consisted of catch trials of the time and the remaining time of an EDS at a given intensity (previously determined at an intensity that elicited a response 50% of the time). Consequently, the O's adjustment of the GN reflected his decision-making processes, analysed by signal detection theory and cybernetics. Experiment one demonstrated that neither practice nor the experimenter influenced the O's performance. Also, there are no residual carry-over effects. The 0 has difficulty assigning relative value to the auditory signals and has little or no difficulty in combining, organizing or co-ordinating the auditory information. The second experiment demonstrated that the O's difficulty in assigning relative value to the signals is due to a short memory retention of the nonsymbolic auditory signals (half-life of less than a second). However, verbal confidence ratings (VCRs) add persistency to the auditory memory (echoic memory), its half-life is extended to about 3.1 seconds. The VCR has a half-life of about four seconds. Experiment three demonstrated that nonsymbolic auditory information processing is not affected by interfering signals. From a cybernetic model of the results it is concluded that nonsymbolic auditory information is processed by a filtered, single channel, series processor. Persistency is added to the system's memory component by an interaction between the auditory (echoic) memory and the conceptual symbolic system (VCR).