A Rudimentary Theory of Information:

Consequences for Information Science

and Information Systems

PETROS A. M. GELEPITHIS

Kingston University, Kingston upon Thames, KTt 2EE, England.

E-mail: petros@kingston.ac.uk.

(Received October 30, 1996; accepted November 10, 1996)

The key fundamental notions of Information Science and Information

Systems are 'information' and 'information system' respectively, with infor­
~ mation the central common notion. Nevertheless, despite the huge litera­
ture on both 'information' and 'information system' their nature remains
debatable and their conceptual nexus to each other and related notions like
'sign' and 'meaning' are, at best, fuzzy or incomplete.
The aim of this paper is to briefly review the notions of 'information' and
'information system' and to outline a theory of information. Specifically, in
section one, we provide a brief analysis of 'information system' and a
summary presentation of the major views on the nature of'information',
concluding that: (i) 'information' and 'communication' constitute the back­
bone of any theory of information; and (ii) all the relevant studies of
information are fragmented, failing to provide a unifying theory and, in
particular, to clarify the highly debatable nature of information. In section
two, we provide the elements of a rudimentary theory of information and
draw some of the consequences of this preliminary body of knowledge for
Information Science and Information Systems.

KEYWORDS: information, information system, meaning, communication,

understanding, theory

1 FOUNDATIONS OF INFORMATION SCIENCE

AND INFORMATION SYSTEMS

The Information Science and Information Systems communities

are known to stand quite apart from each other despite sharing
some key foundational problems and despite the fact that the need
World FUtUTI'5, 1997, Vol. 49. pp. 27!>-286 © 1997 OPA (Overseas Publishers Association)
Reprints available directly from the publisher Amsterdam B.V. Published in The Netherlands
Photocopying permitted by license only under license by Gordon and Breach Scieuce Publishers
Printed in India

275
276 PETROS A. M. GELEPITHIS

for foundational and interdisciplinary work has been well estab­

lished (see, for example, Machlup and Mansfield, 1983; Gitt, 1989;
Checkland, 1992; Marijuan, 1996).
The key fundamental notions of Information Science and Infor­
mation Systems are 'information' and 'information system' respec­
tively, with information the central common notion. In addition, the
considerable number of disciplines concerned with 'information'
and 'information systems' has led to the development of a whole
family of notions, closely related to that of information, (e.g., sign,
symbol, meaning), which need to be clarified and become consistent
with each other. This section provides a brief analysis of'informa­
tion system' and a summary presentation of the major views on >>
the nature of 'information', concluding that: (i) 'information' and
'communication' constitute the backbone of any theory of informa­
tion; and (ii) all the relevant studies of information are fragmented,
failing both to provide a unifying framework, let alone a theory, and
to clarify the highly debatable nature of information. We start with
our analysis of the notion of 'information system'.
It is both well established and widely accepted that an information
system is really a sociotechnical system. 1 Such a view makes clear the
three types of fundamental notions required for its study. First,
notions related to the concept of an 'information system' itself;
second, notions related to all those (e.g., designers, managers, users)
involved in the development of an 'information system' (we shall
generically call those people contributors); and finally, notions re­
lated to the tools used in the development of an 'information
system'. The following paragraphs present the particular sets of
concepts characterising each of these three types and outline their
links to the pair of backbone notions. We start with the system-related
notions. These seem to fall in the following six, related, categories: ~
Group-I: Information-intelligence. The inclusion of information
is, of course obvious; that of intelligence may be seen as less so to
some people and hence a few words of explanation may be useful. A
major category of information systems is those designed by humans.
Its majority is due to its complexity and not its ubiquity in the
universe. The complexity of an artificial information system, in tum,
is due both to its links to the human elements of the designed system
and to artificial systems processing information in ways which capture
 A RUDIMENTARY THEORY OF INFORMATION 277

aspects? of human intelligent behaviour. Intelligence, therefore, in

both its human and emergent machine form is necessary.
Group-2: Communication --+ Input-Output --+ interface (the arrow
should be read as 'brings in the notion of"), It should be noted that
'communication' is necessary, above a certain threshold of complex­
ity of the communicating entities. This should be juxtaposed with
the interacting requirements of mere interfaces or input-output
devices. Similarly, it is true that artificial intelligence systems con­
tinually approximate aspects of human systems and their number
and penetration to new areas of human concern increases. The real
challenge, then, is to interface and integrate artificial intelligence
.r> systems with human intelligence systems to develop complex human­
machine systems. Communication is a must for the design, evolu­
tion, and effective and efficient running of such systems.
Group-3: Complexity --+ Hierarchy-? --+ emergent properties; and
Group-4: Filtering --+ Hierarchy-? --+ emergent properties. Complex­
ity is not very much 3 studied despite its characteristic importance
for highly evolved natural systems and sufficiently richly-struc­
tured artificial or human-machine systems. Filtering is well ad­
vanced technologically but features pretty low in theoretical studies
of both 'information' and 'information systems'. Both complexity
and filtering bring in the notions of hierarchy and emergent
properties, each of which raises fundamental issues of its own
beyond the scope of this paper. The key link of both these two
groups is with the notion of 'information system' rather than
'information'; more specifically, with the notion of a system's or­
ganisation. The reason for not including 'organisation' in the set of
characterising notions is that it is a compound notion with compo­
nents like complexity, and filtering.
Group-5: Goals, and control (including feedback). These two
cybernetic notions remain centrally important for the study of
information although not basic in the sense that 'information' and
'communication' are. As such they should play an important role in
any full theory of information but they will not be included in our
rudimentary theory.
Group-6: Design-Formalisability-Computability. This is an inter­
esting group. Formalisability and computability are related, exclu­
sively, to artificial information systems; design to both natural and
 278 PETROS A. M. GELEPITHIS

artificial information systems. The former subgroup is closely related

to the notion of uninterpreted system in formal studies but not directly

related to information or the majority of its family notions, as they

are defined in the next section. Design, in artificial information

systems, is a process requiring communication (see next section for

definitions and brief justification).

Contributors-related notions fall into two basic categories: (i) those

involving the theoretical beliefs of a contributor; and (ii) those involv­
ing the non-theoretical beliefs of a contributor. The former category
includes issues concerning the nature of organisation, society, science,
and knowledge, as well as technical issues like computability, formalisa­
bility, and design. Essentially, we meet again here all of the notions ~
characterising an information system itself, albeit mostly implicitly.
Consider for example, 'knowledge' which requires a distinction to be
drawn between individual and collective knowledge and hence brings
in the issue of communication. Or, again, the nature of science which
brings in the issues of formalisability and computability.
Finally, tools-related notions fall into three categories: (i) accu­

racy of representation; (ii); scope; and (iii) grain size. These are

important, technical concepts which depend crucially on both the

design and overall system requirements, and thus bring us via a

third route to some of the basic notions introduced under the

concept of system.

In summary, one can see that the two concepts which cut across

all three types of fundamental notions required for the study of an

'information system' are information and communication and,

therefore, these constitute the backbone of our rudimentary theory in

the next section.

We come now to our summary presentation of the major views on

'information' and the few attempts made to provide a coherent

framework for its related conceptual nexus. Concerning the nature

of 'information' one may distinguish" seven major viewpoints. First,

traditionally, information in terms of the probability of a signal

(Shannon and Weaver, 1949).5 Second, the conception of informa­

tion as order (e.g., De Vree, 1996). Third, information in terms of

knowledge and meaning at a mentalistic level (Langefors and

Samuelson, 1976); and, more strongly, information as a mental not

a material entity (e.g., Gitt, 1989). Fourth, information in terms of

- - - - - _. .

•
 A RUDIMENTARY THEORY OF INFORMATION 279

the notion of sign as a primitive (e.g., Stamper, 1985); Fifth,

information conceived in terms, essentially, of the Popperian concep­
tion of the three worlds (e.g., Tully, 1985). Sixth, information in
terms of truth conditions (see, e.g., Israel and Perry, 1990). Finally,
information as a basic property of the Universe (e.g., Rzevski, 1985;
Stonier, 1996); or, at least, as an objective commodity, or intrinsic to
external objects (e.g., Dretske, 1981; Collier, 1990). Concerning the
nexus of informational notions, the most notable attempt is that of
the 'FRISCO group' who have set themselves the grand task of
clearing the "conceptual foundations in the information system area",
but so far" they have failed in developing a consistent framework that
,r--
would be based on notions with a truly multidisciplinary acceptance.
In summary, all the relevant studies are fragmented, failing to
provide a unifying framework, let alone a theory, as well as a clari­
fication of the highly debatable nature of information.
Taking together the above remarks on the notions of'informa­
tion' and 'information system', one is led to aim for a theory
in the traditional sense of the word, that is, of a body of knowl­
edge enabling an appropriate user to draw explanations and pre­
dictions about its subject matter as well as of controlling existing
and designing new systems within its boundaries. This is a long
list. The next section is confined: (i) to define 'information',
'communication', and the nexus of interrelated notions in a coher­
ent and, if possible, unifying way which will minimise the vagueness
of the notions involved as well as of the relations among themselves;
and (ii) to draw some of the consequences of this preliminary body
of knowledge for Information Science and Information Systems.

2 THEORY OUTLINE AND SOME CONSEQUENCES

We start with human 7 'information', generalise to 'information',

continue with the rest of the major family notions, and conclude with
some clarificatory remarks with respect to our definition of meaning.

Human information = df Expressed human thought or set of

human thoughts.
Human thought = df Set of human thought elements.

280 PETROS A. M. GELEPITHIS

Human thought element = df Selected or prevailed neural

formations.
Information = df Expressed thought or set of thoughts.
Thought of entity E = df Set of thought elements of E.
Thought element of entity E = df Selected or prevailed material
formations of E.8
Symbol == df Human sign."
Sign == df Configuration meaningful to a receiver. 10
Signal = df Propagated configuration meaningful to a receiver. II
Data == df Potentially meaningful configurations.
The linguistic or perceptual meaning M of something s in the
context C" for the entity E, at time t-symbol M(s, C" E, t)- is the
selected or understood formations of the representational material
of E, at t-symbol R,':".

To avoid potential misunderstandings, with respect to the last

definition, the following three remarks are in order. First, the ex­
pressed meanings of an information system may be of the system
itself or, equally well, those of another entity. For example, for a
human perceived as an information system the meanings are
internal to that human; for a present-day'" computer though, the
meanings it processes are those that some humans have chosen to
represent in a computer processable form. Second, processing is
very different from understanding. The former is akin to uncon­
scious thinking and, in contrast to understanding, it may lead not
to primitives (see below the definition of understanding and re­
marks on it). Finally, I have only presented here the generalised
definition of meaning. For justification and discussion the reader is
referred to Gelepithis (1989).
Now to the cluster of notions centred around communication and
its basic constituent understanding. Although there is general
agreement that 'communication' involves sharing and 'understand­
ing' (see, for example, Cherry, 1957; Ogden and Richards, 1923;
Rogers, 1986) no-one had really defined it until Gelepithis (1984).
In what follows, we repeat those definitions, introduce the basic
characteristics of the communication and understanding processes,
and present a fundamental result that is used only to support
consequences with respect to Information Systems.

 A RUDIMENTARY THEORY OF INFORMATION 281

Definition ofcommunication: HI communicates with H 2 on a topic T if,

and only if: (i) HI understands T {Symbol: U(H] T)}; (ii) H 2 under­
standsT {Symbol: U(H 2 T)}; (iii) U(H] T) is describable to and under­
stood by H 2 ; and (iv) U(H 2 T) is describable to and understood by HI.

Definition of Understanding: An entity E has understood something,

S, if and only if, E can describe S in terms of a set of primitives of
its own.

The following characteristics of understanding and communica­

tion provide the basis of the consequences drawn next. First, un­
r­
derstanding is structured. This has three aspects. One, being de­
pendent on one's own primitives makes understanding dependent
on time since such primitives do change with its passage. As an
example, compare a toddler's primitives with those of a quantum
physicist with respect to the notion of electricity. 13 Therefore, within
one and the same person, understanding is 'layered' according to
one's experience. Two, since understanding depends on one's own
primitives, its end result, that is the individual knowledge reached,
may well vary very significantly from person to person depending
on the level of primitives reached by each person on a particular
topic. Finally, understanding is structured as a consequence of the
existence of two kinds of primitives: linguistic and sense primitives
(Gelepithis, 1984; 1985). Second, understanding, if not immedi­
ate'", requires a systematic approach to reach its objective. This
follows directly from its defining characteristic of reducibility. Final­
ly, understanding is not formal. This follows from the existence of
the two types of primitives mentioned above.
On the basis of our definitions and characteristics of communica­
tion and understanding introduced above, a human, say H, and an
intelligent machine, say M, would communicate on a topic T,
expressible in language L; if and only if: either PH = PM for T (P for
primitive); or PH and PM could be described in terms of each other.
Since linguistic primitives are reducible to sense primitives except
if they are purely linguistic, one needs language to describe
the senses and senses to understand language. Hence PH and PM
could not be described in terms of each other. In other words,
human-machine communication is impossible. This is a fundamental
282 PETROS A. M. GELEPITHIS

result, with ramifications extending beyond Information Science and

Information Systems (for a full exposition of the argument and a
general discussion see Gelepithis (1991). Here, it is used only to
support consequences with respect to Information Systems.
The paragraphs of this section so far constitute our rudimentary
theory of information. It is rather obvious that this preliminary
body of knowledge is characterised by conceptual clarity, internal
consistency, and a good degree of objective standing to a good
number of the family notions related to 'information' and 'informa­
tion system'. Next, we use such a rudimentary theory to derive some
consequences for Information Science and Information Systems.
With respect to Information Science, the first consequence,
derived from the nature of human information, is radical. Since
human information is the expression of a set of selected or prevail­
ed neural formations, there is no need for any new science of
information; biology is perfectly adequate for the study of human
or animal information. For information in general, physics takes
up the role of biology. 15 Would the possible discovery of extrater­
restrial information processors call for a science of information?
I do not think so. The study of extraterrestrial information proces­
sors by humans, if possible, would only require the establishment of
appropriate communication channels and the possible modification
(including extension) of biological or physical principles. Naturally,
the multidisciplinary and unifying perspectives which the propon­
ents for a science of information advocate are laudable objectives
which need to be adopted by biology or physics in their study
of information. It is worthy of a note that the eventual, if not
interrupted that is, emergence of machine intelligence will require
the much closer cooperation of biology and Artificial Intelligence.
We turn now to some of the consequences with respect to
Information Systems. First, complex human-machine systems could
not be fully formalised except if all human elements were, eventually,
to be replaced by artificial intelligence systems. The minimum
number of human elements required to be kept in the system in
order to be able to ascribe accountability to humans is a crucial,
open question. Therefore, behaviour of such a system is in general
non-computable. It can of course be constrained to produce only
the computable aspects of its behaviour.
 A RUDIMENTARY THEORY OF INFORMATION 283

Second, since human-machine systems are non-computable, no

general (i.e., system independent) information systems methodo­
logy can be constructed.
Finally, the specific methodologies for the development of knowl­
edge-using human-machine systems are constrained by the processes
of communication and understanding and therefore, cannot be
purely formal. To design an effective and efficient information
system it is necessary to include both formal and informal elements.
This constraint and key methodological tool I call the communication­
understanding principle. This last consequence can be seen clearly by
considering the rationale of structured methodologies. It is based on
r' two assumptions. First, that user requirements can be rigorously
specified. Second, that such a specification will not include elements
which are non-formalisable. But, user requirements are not even fully
specified by the users. As a result: (i) the prerequisites for the use of
structured methodologies do not hold true; and (ii) understanding of
user requirements by the listener (be that a designer, a manager or
whatever) differs from that intended to be communicated.

3 CONCLUSION

The work presented here is only a small part of what is required

in developing a full theory of information and, equally important,
presented only in outline or even in citation form, due to the usual
paper-length restrictions. What is mostly required is to consistently
put together as many of the various strands constituting the foun­
dations of information as possible and to do that in a way that will
r--- be accepted by as many of the contributing disciplines as possible
(eventually they should be all of course). The way forward is not for
the faint-hearted.

Acknowledgements

Martin Robson has read through an earlier version of this paper

and discussed with me some of the points made, or intended to be
made. I hope I made good use of his revealing comments. Thanks
284 PETROS A. M. GELEPITHIS

Martin. I would also like to thank the anonymous reviewer who

made me rethink of my presentation on the basis of his/her
astonishing and selectively useful comments.

Notes

I. For a well presented argument the reader is referred to Land (1985).

2. The question of autonomy and genuine intelligence of their own is beyond

the purpose of this paper; the interested reader is referred to Gelepithis (1991).

3. Quite revealing in this respect is Simon's paper within the theoretical literature on

the nature of complexity.

4. We exclude from our presentation all accounts which do not explicitly tackle the

issue of the nature of information. For a full review of semantics covering the

philosophical, linguistic, formal, and biological theories of meaning the reader is

referred to Gelepithis (1988).

5. See also Kolmogorov (1968) for a common basis between probability and informa­

tion theories.

6. I would like to note that the FRISCO group's work is under development and my

criticism is based on their latest, but not final, public report (personal communica­

tion with IFIP WG 8.1 Task Group FRISCO, 1995).

7. In contrast to all other attempts, all my human-depended definitions are eventually

cast in terms of neural (not necessarily neuronal) formations.

8. It may tum out that certain entities, exhibiting intelligent behaviour. may have

'thoughts' the nature of which is not captured by our definition. In such a case a

decision will have to be made whether the scope of our definition needs to be

modified, or it is preferable the discovered or designed entities to be classified as

thoughtless entities with intelligent behaviour.

9. The most unified alternative view is Newell's (1990) based on the Physical Symbol

System Hypothesis (Newell and Simon, 1976). For a summary review of the major

views on the nature of symbols see Gelepithis (1995a).

10. Quite close to Charles W Morris' conception of sign (1939).

11. In sharp contrast to Shannon's theory (Shannon and Weaver, 1949).
12. For a discussion of intelligent machines see Gelepithis (1991).
13. For a discussion see Gelepithis (1995b).
14. That is, an intuition.-~
15. It should be noted that this does not imply a reductionist view. The issue of
reductionism is much more complicated than it might appear from a face reading
of the above sentence and although extremely interesting it is well beyond the scope
of this paper.

References

Checkland, P. (1992). Information Systems and Systems Thinking: Time to Unite? In

Challenges and Strategies for Research. in Systems Development, W. W. Cotteman and
J. A. Senn, eds, John Wiley and Sons.
A RUDIMENTARY THEORY OF INFORMATION 285

Cherry, C. (1978). On Human Communication: A review, a suruey, and a criticism. Third

edition. The MIT Press.
Collier,.J. D. (1990). Intrinsic Information. Hanson, P. P. (ed.), Informatian, Language,
and Cognuion. The University of British Columbia Press.
De Vree,.J. K (1996). A note on information, order, stability and adaptability. BioSystems
Vol. 38, No.2 and 3, pp. 221-227.
Dretske, F. 1. (1981). Knowledge and the Flow of Information. Oxford, Blackwell.
Gelepithis, P. A M. (1984). On the Foundations of Artifu:ialIntelligence and Human Cognuum.
Ph.D. Thesis, Brunel University, England.
Gelepithis, P. A. M. (1985). The Nature oiHuman Understanding: Human Primitives. School of
Information Systems Reports 1985-1988, Kingston University, England.
Gelepithis, P. A M. (1988). Survey of Theories of Meaning. Cognitive Systems, No. 22,
pp. 141-162.
Gelepithis, P. A. M. (1989). Knowledge, Truth, Time, and Topological spaces.
Proceedings of the 12'" International Congress on Cybernetics, pp. 247-256, Namur,
Belgium.
Gelepithis, P. A. M. (1991). The possibility of Machine Intelligence and the
impossibility of Human-Machine Communication. Cybemetica, Vol. XXXIV,
No.4, pp. 255-268.
Gelepithis, P. A. M. (1995a). Artificial Intelligence: An Integrated, Interdisciplinary
Approach. (completed; with publishers).
Ge1epithis, P. A. M. (1995b). Revising Newell's conception of representation.
Cognitive Systems, Vol. 4, No.2, pp. 131-139 (Special issue on Representation).
Gilt, W. (1989). Information: The Third Fundamental Quantity. Siemens Review,
Vol. 56, No.6.
IFIp WG 8.1 Task Group FRISCO (1995). A Framework of Information System Concepts.
Personal communication.
Israel, D., and Perry,.J. (1990). What is Information? Hanson, P. 1'., (ed.).Information,
Language, and Cognition. The University of British Columbia Press.
Kohnogorov, A. N. (1968). Logical basis for information theory and probability
theory. IEEE Transactions on Information Theory, Vol. 14, pp. 662-664.
Land, F. (1985). Is an Information Theory Enough? The Computer Journal, Vol. 28,
No.3, pp. 211-215.
Lan gcfors, B., and Samuelson, K (1976). Information and Data Systems. Petrocelli!
Charte, New York.
Machlup, F., and Mansfield, U. eds (1983). The study of information: Interdisciplinary
messages. John Wiley & Sons.
Marijuan, P. C. (1996). First conference on foundations of information science: From
computers and quantum physics to cells, nervous systems, and societies. Bio­
Systems, Vol. 38, No.2 and 3, pp. 135-140.
Morris, C. W. (1939). Foundations of the Theory of Signs. The lntemaiional Encyclo­
pedw of Unified Science, Vol. I, No.2.
Newell, A. (1990). Unified Theories ofCognition. Harvard University Press.
Newell, A. and Simon, H. A. (1976). Computer Science as Empirical Inquiry:
Symbols and Search. Communications of the ACM, Vol. 13, No.3, pp. 113-126.
Ogden, C. K, and Richards, 1. A. (1923* 1956). The Meaning of Meaning. Harcourt
Brace and Co. Inc. NJ, U.S.A.
Rogers, E. M. (1983* 1986). Elements of Diffusion. (Extracts from Chapter 1 of
Diffusion of Innovations, Rogers F.. M. 3rd Ed, New York: Free Press, 1983).

286 PETROS A. M. GELEPITHIS

Roy R., and Wield, D., (eds), Product Design and Technological Innovation, Open
University Press.
Rzevski, G. (1985). On Criteria for Assessing an Information Theory. The Computer
Journal, Vol. 28, No.3, pp. 20{}-202.
Shannon, C. E., and Weaver, W. (1949). The Mathematical Theory of Communication.
University of Illinois Press.
Simon, H. A. (1962). The architecture of complexity. Proceedings of the American
Philosophical Society, Vol. 106, No.6, pp. 467-482.
Stamper, R. K. (1985). Towards a Theory of Information. The Computer Journal,
Vol. 28, No.3, pp. 195-199.
Stonier, T. (1996). Information as a basic property of the universe. BioSystems,
Vol. 38, No.2 and 3, pp. 135-140.
Tully, C. J. (1985). Information, Human Activity and the Nature of Relevant
Theories. The ComputerJournal, Vol. 28, No.3, pp. 206-210.