20010830: Shannon, The mathematical theory of communication

Contact:cdent@burningchrome.com

Shannon, C.E., & Weaver, W. (1963/1949). The mathematical theory of
     communication (p. 31-35). Urbana: University of Illinois Press.

Introduces the basis for information theory wherein a communication
system consists of five parts which work together to deliver a
message: an information source, a transmitter, a channel, a receiver
and a destination. Each of these parts can be represented as
mathematical entities and thus empirical studies can be made of the
transfer of information through the system.

-=-=-

There's no doubt that Shannon's work has had massive impact, good and
bad, on both the practical/technological and theoretical sides of
information science. Information transmittal, between electronic
systems and between human brains, can be modeled with the five parts
of Shannon's system. That modeling can help break down a problem into
solvable pieces, improving information uptake.

There is, however, an unfortunate side effect to the model: any system
which is predicated on the presence of a single piece which transmits
to a single piece which receives implies that at any given moment in
time a message goes in one direction. While this may be true in
electronic circuitry[1] it does not appear to be the case in the
exchange of ideas. When a human reaches out to a source of information
to learn, that reaching is accompanied by a wealth of preconceptions
that color the transmittal of information from the source.

Presumably an adherent to Shannon's theory would suggest that the
preconceptions are in fact feedback noise fed into the channel from
the receiver. Again, electronically this has appeal, but from other
angles the simple act of calling the preconceptions noise degrades
their value and the importance of the experiences of the information
seeker.

So, like so many of these theories, it is instructive and helpful, a
good one for the toolbox, but incomplete without the salt shaker.

[1] Full duplex traffic is of course possible in some network
topographies, but there the bi-directional traffic is of two
different messages, passing like ships.


Back to the Index