Introduction

Why should we have a mechanism which can keep some sounds which are separate entities separate from one another? The usual explanation attributes everything to the cortex, which is a superior way of saying we don't know. The standard example always described, and said to indicate the remarkable ability of a cortex to extract one of two or more received sets of sounds, is called the cocktail-party phenomenon. It enables us to overhear a conversation far more interesting than the one we are suffering from the person who has latched onto us (and today usually with some mindless music being played at the same time). It is a bad example because it confuses three phenomena: the direction of the sound, tracking, and using our knowledge of language to interpret partially heard word noises.

If we receive noises from two different directions at once, and we could not distinguish the pulse streams which they generate at the same time in the ears, we would not know that they were separate sounds. But the sensations arrive in the cortex as separate noises, each with its direction. It may not occur to everyone that direction is just as much a sensation of a noise as is its general pitch or anything else; it is a more basic feature of the hearing sensation than any other characteristic except loudness. We can track and concentrate on either of two sets of noises, and that is a cortical activity. But no cortex can re-assemble a vast stream of pulses arising from two or more complexly changing distinct noises if it does not know how the noises are going to change instant by instant. We can learn to track the contributors to chamber music from more or less one direction, but music is a special artificial kind of sound.

Many mammals can obtain the direction of a sound source in the presence of other sounds, and they do not have advanced cortical processes to track one sound sensation from the mass of pulses coming from their ears. Neither they, nor we, have a problem with a noise that has never been heard before. Sound location is automatic, performed by processing stations before nerve pulses reach the cortex. The processors can send instructions about sound direction to eyes, neck, or rotating external ears; it is a reflex response to one noise in the presence of other noises.