INFORMATION FLOWS AND HUMAN CHANNEL CAPACITY
The forum of public interaction may be thought of as a cultural marketplace where senders and receivers seek each other out. There must be on element of the unexpected in the messages delivered there, a bit of surprise, for most of the receivers or their time is wasted. Even the most static culture requires that the new generation must learn the traditions that have been preserved by the old. The primitive ritual that is performed correctly is as much high culture as grand opera and the swearing in of presidents and it is at least as instructive to the onlookers. The potential for surprise in the messages, as measured by various adaptations of Shannon's technique, may serve intuitively as an index of cultural value. The measure says, in effect, that for a receiver with a given vocabulary and lexical experience who remains alert and attentive a certain quantity of information can be apprehended. From small-scale experiments conducted by educators it is evident that an amount between 30-60% is actually retained, and even that is slowly dissipated overtime due to forgetting.
Methods of measuring the novelty in communications other than print, such as maps, graphs, sketches, photographs, views of the environment and any other visual material, even of music and texture, can be constructed in a manner analogous to the method of Shannon. The identification of elementary symbols is of necessity cruder than for print, but, as will be seen later, this imprecision does not disturb preliminary estimates of overall non-redundant information flow too much because the other channels are less used for public interaction.
The novel impressions acquired as a consequence of communication and observation may be expected to influence choices immediately, particularly when the receiver switches to the role of sender and asks a question or adds a comment, or they may be stored because they promise to be of value for guiding future behavior. This cultural index enables the weighing of many different messages in terms of the same standard unit, the bit, so that one transaction may be asserted to have more bits than another. The measure can be manipulated as a scientific instrument enabling independent investigators to agree with each other within the limits of experimental error. Then the cultural quality of various transactions would not rest upon any man's opinion, or even the collective opinion of a dominant social class. The judgment of a third party with such a tool is less arbitrary, less susceptible to the inflations and deflations of faddism, and more comprehensive in application than that other ubiquitous measure of value--market price. The aggregates of bits should be far more stable, and much more free of paradox, than the comparable accounts constructed by economists.
Shannon's method for empirical measurement of the non-redundant (and therefore "significant") information in prose was stimulated by much more fundamental theories for which he himself is responsible to a large degree. The history of the mathematical theory of communication illuminates both the need for rigor and the basic limitations upon application in the present formulation.
Classical information theory can be traced back to a paper published by Leo Szilard when he was attempting to resolve some apparent ambiguities in statistical mechanics and