Information Technology Reference
In-Depth Information
Therefore, the definition of the word meets the rigorous test for
knowledge as “an ideal reproduction of the external world serviceable
for cooperative action thereon” (Childe 1956, 54). For instance, specific
cooperative actions could include isolation of words from written dis-
course for full-text indexing conducted either by human clerical labor or
machine process. Clerical labor effectively instantiated an understanding
of the word strongly analogous to Shannon's definition, for instance for
producing biblical concordances or nineteenth-century indexes for news-
papers (Palmer 1885) before its formal definition. Clerical and automatic
computational processes have used a similarly implicit understanding of
the word and also have recognized the significance of the space; how-
ever, such understandings seem to have developed largely independently
of the relevant theory. Thus, practical understanding has occurred partly
in advance of theoretical formulation and diffusion, but practice could be
informed by the encompassing theory.
The definition of the word as “a cohesive group of letters with strong
internal statistical influences” is consistent with the already adopted con-
ception of the message for selection as the individual characters of the
Roman alphabet, and of the message as compounded from messages for
selection. It also incorporates qualities specific to the message: cohesive-
ness and strong internal statistical influences. The definition of
word
can
also provide a basis for a consistent understanding of the multiword
sequence.
Multiword Sequence
As revealed by testing human subjects, prediction possibilities offered by
the written-language message connect the word to and differentiate it
from the multiword sequence (Shannon 1951/1993; Moradi, Grsymala-
Busse, and Roberts 1998). From prediction testing, Shannon concluded
that statistical effects extending up to one hundred letters (and therefore
beyond the separate word) reduced entropy, the amount of information
conveyed, to the order of one bit per letter for ordinary literary English.
Understood as the negative of entropy, redundancy was roughly 75 per-
cent; subsequent studies have yielded similar but not identical estimates
(Moradi, Grsymala-Busse, and Roberts 1998). The space was revealed as
almost entirely redundant in sequences of one or more words (Shannon
1951/1993, 198). Real-world coding systems have displayed a practical
