Dennis Gabor gave an excellent exposition on the limitations of
information propagation in his article, "The Theory of Communication" .
This is an article I refer to constantly in trying to understand how
real physical signals are transmitted. The theory of communication
is based on wave theory.
In terms of wave analysis we try to get a best estimate or best
description of a physical signal which exists inside of sound by creating
simple wave packets enveloped by a Gaussian function. These simple
wave packets provide the best description of a real physical wave.
It is not possible to give a better description due to Nature's
limitation on what we can measure. As the wave form increases in
complexity we can use these simple little wave packets or wavelets
to build more complex signals as in the sound waves of speech.
To reconstruct a signal, various simple wave packet objects are numbered or
tagged in an orderly fashion and then used inside of a mathematical function
to build a complex signal. These numbered tags are called coefficients.
They are used in the Gabor function which represents the construction
of a complex signal from many little wave packets.
An analogy to constructing a real signal like a sound wave from
speech is abstractly putting together parts of alphabets and words
into a sequence. This is what you do in your brain before you
speak; before you create the real physical sound waves with your
vocal cords. The abstract creation of alphabetic word sequences
is studied in computer natural language programming. The software
Prism was created from using tools in computer programming
A modestly successful methodology of encapsulating ideas in set
of word clusters is using a profile as was done in the Prism software.
A profile consists of words which represent similar characteristics
of the idea. These characteristic subsets of words and phrases are
selected in relation to the other word subsets or profile clusters in
our universal word set. In the grammatical representation using strings,
the profile cluster becomes the characteristic coefficients of a
function like a Gabor transform. (I understand that terminology like
tags and coefficients, and functions and transforms may be confusing,
they are essentially the same corresponding thing. It takes a long
time to get comfortable with this terminology. You might just intuitively
feel you way through this for now, if it's unclear to you.)
The Prism program uses these "profile coefficients" to
classify a random input signal. A group of these coefficients make up
the profile dictionary. In the grammatical representation using strings,
the profile cluster becomes the characteristic coefficient.
We know that thought processes are naturally wrapped or packaged in
language. Language elements such as words, sentences, and the
associations these language elements make up in the minds exists
because, maybe, the creation of language necessarily required
a construction built on the mind's architecture. Language,
including music, exists because of how the mind is built.
We're learning from the structure and operation of the brain how
to construct an information processor resembling or modeling how
we speak. We're trying to encode the information in spoken words
in dynamic binary sequences. These sequences exist in a temporal or
time-ordered dimension inside wave forms like simple wave packets.
Hopefully this is not an over simplification of the fundamental
processes occuring in our neural circuits, but the passing of
time will tell the truth. So be a little skeptical, and always
create the "truth" for yourself by thinking it through for yourself.
Spoken and written language enables us to associate or connect sequences
of word elements together. Spoken language in particular places special
meaning on the temporal sequence of words. Images are composed of
picture elements which are meaningful to each other in terms of the
spatial relationship they have to each other. Both verbal and visual
language elements can be described by sequences of words or picture
elements, and our brain processes this information by associating these
elements together in a sequence.
Theory of Communication, Dennis Gabor, 1946,
The Journal of the Institution Of Electrical Engineers, 93(3):429-457.