THOMAS A. SEBEOK
Speech communication maybe studied as an information system and natural language as the code used in the system.1 From this point of view it seems useful, on the one hand, to investigate the analog and digital characteristics of human communication and animal communication systems as to the manner in which they code information.
Communication engineers often distinguish between two kinds of control machines: those for counting and those for measuring. The former, which are "all-or-none" devices, are sometimes called digital; they are of a "yes/no" type. The latter, which operate on the basis of connections between measured quantities and the quantities they represent, are, by contrast, known as analog; they are of a "more-or-less" type. The prototype of a digital system is the abacus, that of the analog, the slide rule. The chief limitation of analog systems relates to their accuracy; those physical systems, on the other hand, which are able to make yes/no decisions can achieve any desired precision, given sufficient capacity and time.
A further observation--developed by N. Wiener, among others--is that the functioning of the nervous system is prima facie digital:2 If a combination of incoming messages will not cause an outgoing fiber to fire, it is said to be below threshold; otherwise, it is said to be above threshold. Nervous pulses can thus be viewed as two-valued markers: the presence of a pulse (release of acetylcholine?) above threshold represents one value, say, the binary digit 1, and the absence of a pulse (release of cholinest-____________________