A belated contribution on “data” and “datum”. These are one or more items of information
taken as given for performing a task or solving a problem. So what do we mean by “item” and
“information”?
Appalled by the sloppy terminology
in the information systems field, I wrote a book (Information, 1973), which
used Peircian ideas among many others to show how the existing literature sets
some excellent examples.
That book is out of print. There
would have been a second edition but a burglar took my computer and its backup
disc, thus diverting my energy to a book on semantics. The main argument was as follows.
Computer
science has a sound basis in the formal sciences but it is too narrow for a
science of information systems, which is failing to develop because the IS
community uses language so imprecisely, talking, for example, of “Data”, “information”, “meaning”, which have
multitudes of precise meanings. One
hears the chemical engineering metaphor about a series of mystical fluids. Thus: distil data to obtain information, which you distil to obtain knowledge
and meaning. I joked one day that someone
would distil knowledge to obtain wisdom only to read that proposition in an
article a leading accountancy journal and in a pamphlet thrust into my hand by
a member of a religious sect! “DIKW” has
just yielded 159,000 hits on Google! I
could scream.
I introduced a ‘semiotic
framework’ to focus on six ways of understanding different aspects of signs. Each limited range of properties of signs builds
on the earlier ones. Within each level,
‘’data”, “information”, “meaning”, “knowledge” etc. have at least one
operationally precise meaning, often several. There’s not space enough to give
examples for each level.
Physical level – Just he physical
properties of marks treated as information; the material and energy employed in
operations on them, whether performed by machines or people. All economic questions about information
systems efficiency end up here.
(Physics, engineering, ergonomics)
Empiric level – The
statistical/probabilistic properties of streams of sign-tokens: the domain of
Shannon’s work.
Syntactic level – The structures
built of sign-types and the operations that may be performed on them: the
formal sciences and most of computer science (eg: Carnap and Bar Hillel on
info.).
LEVELS ABOVE - THE TECHNICAL
BASIS OF ANY INFO SYSTEM
LEVELS BELOW ARE INTRINSICALLY HUMAN
AND SOCIAL
Semantic level – The nature and
properties of the ‘stands for’ relationship between a sign-types and some
thing(s) in the real world (!!!???). (Philosophy
and semiotics have contributed most here.)
Pragmatic level – The properties
of sign-tokens that explicitly or implicitly express the intentions of their
users. (Language theory, philosophy,
ethnomethodology, semiotics …)
Social level – The changes to attitudes
and norms resulting from the interpretation of sign-tokens; these effects
account for the positive or negative value of information. (anthropology, social psychology,
jurisprudence etc.)
No organisational information
system can function well unless it does so on ALL those levels, a fact worth
considering when ca. 25% of IS projects fail totally and ca. 50% fail in major
respects. My industrial experience
suggested that we neglect the social levels; hence my research on ‘real world’
semantics and systems of social norms.
Ronald Stamper