Guest Column

Rethinking Information

By William H. Edmondson

Information is an over-used word with correspondingly diluted/diverse meanings. This account begins with a non-exhaustive but interesting list of usages – the aim is to demonstrate that the diversity does not reveal a clear underlying concept. The text moves on to consider some semiotic issues before offering a definition.

1/ Information is a label for a location whereat a person/machine can be found with whom/which a dialogue can be held, questions can be answered, instructions and explanations can be given, and so forth. The location can be physically defined or virtual.

e.g. an “information desk” in a department store;

e.g. a phone-number for a “help-desk” (cf. “information” in older US telephone systems).

2/ Information can refer to a display, leaflet, brochure, etc. on/within which a reader can find instructions, explanations, answers to anticipated questions, and so forth. Such information resources may be labelled “instructions”.

e.g. a storeguide on a wall next to an escalator;

e.g. an instruction manual.

3/ Information as a text stating conditions, obligations, ownership, and the like, which relate the reader to some institution, object, setting, event, etc. Such material is usually designed to inform rather than address queries and thus differs from 2 above.

e.g. reverse side of a reader’s ticket for the British Library informs the reader that the ticket is the property of the BL and informs them what to do with the ticket if it is found;

e.g. license agreements for computer software.

4/ Information can be used to refer to the content of books, databases, transcripts of conversations, etc. and by extension is taken to refer to collections of books etc.

e.g. libraries contain information by virtue of the newspapers, books, documents, etc., contained in them (which latter also contain information);

e.g. databases (police, medical, club membership, etc.) contain information.

5/ Information can be considered to exist in objects and events.

e.g. one can “tease information out of rocks” to learn about geological activity;

e.g. one can “read” events and objects in terms of providing evidence/information in forensic settings.

6/ Information can be said to be valid or invalid; information comes with a sense of “quality”.

e.g. forecasts can be thought of as good or bad, depending on the quality of the information used;

e.g. libraries and archives have a sense of being “valuable” because authoritative, trusted.

Standing back from these varied but unremarkable usages we need to consider any commonality we can find and take that into a semiotic context for discussion. The obvious theme running through the usages, and probably all usages of the term information, is that all involve some sort of behaviour and/or activity. In essence there is some sort of pragmatic aspect which runs through the examples.

Semiosis is the process of giving/exploiting meaning in sign systems – signification. We can recall that signs need to be considered in relation to their users (pragmatics), to referents (semantics), and to each other (syntax). This is all systematized, and semiotics is concerned with the study of such systems, just as semiosis is concerned with their deployment. But that well known three-legged stool – the semiotician’s traditional seat – is in need of transformation into something more complex and more comfortable. An extra leg, for example, is readily found when one considers the need to analyse signs in terms of their subcomponents. The structures within signs are systematically complex (e.g. phonology) and this level of analysis is not easily squeezed into any generalization of the concept of syntax (except perhaps on the basis of considering this concept recursively – see Selkirk (1999) for a consideration of morphological structures – but actually the issue goes deeper, into more abstract structures found in phonology which cannot really be considered simply a recursive application of the notion of syntax). Additionally, the semiotician’s seat can be given some padding. The existence of typological variation found in languages (linked perhaps to language change and/or evolution) cannot be ignored by semioticians (a useful point of reference here is Comrie (1989)). Obviously enough, semioticians explore the variations in semiotic systems as a way of refining their ideas about the sytematicity they study (they also need to take note of the fact that languages may themselves be semiotic objects/signs). The padding is perhaps a bit lumpy because semioticians tend not to consider language acquisition in children as a topic worthy of attention – surely an unwarranted circumscription of their domain.

However, labouring on with the metaphor, the biggest change to the semiotician’s seat is to add arms and the back(rest). In a word, contextualization is the important missing component in semiotic theory. Note, not context which, in the final analysis, always reduces to an extension of the semiotic activity involved anyway, and is thus limited in the same way (to offer “context” as explicans is wrong – it is actually part of the explicandum).

Contextualization is the process of relating states, entities or activities to other such states, entities or activities. This is ubiquitous in the sensory/behavioural existence of an organism, and is considered to be semiosis when the entities are part of a semiotic system. It is important to note the transient “nowness” of contextualization, and in consequence the difficulty encountered in recovering earlier contextualizations – such reconstructions are, in the limit, impossible in detail because experience has intervened to change the states/entities/activities involved. The semiotician, in other words, needs structural support beyond the four legs of their seat. Semiosis is essentially incomplete without proper consideration of (if not formalization of) contextualization.

Information is now simply defined – it is current evidence of current contextualization (semiosis when sign systems are involved). It too is transient, in that the ‘spoils’ of contextualization, the residue, the consequence, the result, is not information but rather data and changed states, entities, activities in the beings concerned.

We can now revisit the examples of usage listed earlier. These are, in reality, opportunities for contextualization of data. It is as if the word information is used to flag a potential for contextualization, not the fleeting activity itself, and certainly not the result, which is simply more data. So libraries are full of data. We refer to the contents as information in anticipation of their involvement in contextualization in someone, at some time. The notion of information quality becomes one of utility –the contextualization of poor quality data eventually yields more poor quality data. The notion of teasing information out of rocks (or some forensic setting, or whatever) becomes one of difficulty in contextualization. It is notable that contextualization is almost invariably inscrutable, personal, and individualistic – except where social custom and training (e.g. in mathematics, sciences) are constructed to de-personalize the contextualization (there really is no point in permitting my notion of 3 to be different from yours!).

A new example: when an aeroplane reaches a way-point in its flight from A to B the sensors on the ’plane produce data which match some preset criteria for change of course. The data can be displayed on instruments which are read by the pilot who contextualizes them as information for the fleeting moment it takes to act and change course. States/activities change and one is left with data as the instruments record a new heading. Note that the autopilot, if programmed to change course at the way-point, would also fleetingly be informed until the change of course had been achieved. If the autopilot displays the data on its dials and displays it does not display information. Even if it displays data or codes indicative of its action in changing course it is still not displaying information. It is displaying data which constitute a contextualization potential for the pilot. The black-box flight recorder, on the other hand, never does anything other than record data.

So is the pilot ‘informed’ at any stage? The pilot’s contextualization results in a state of informedness – but this is transient as the contextualizations move on to other topics (checking weather predictions, chatting with a stewardess, calling a flight controller on the ground at the way-point….). The pilot will, one hopes, remember having been being informed, and will be able to produce data about it, so that when the co-pilot wakes up the pilot can tell her that they reached the way-point and she changed course (I leave it as an exercise in analysis of contextualization for the reader to work through why they assumed the pilots were men).

The notion of contextualization in semiosis has recently been promoted in relation to word meanings by Evans (2006) who writes “… the semantic values associated with words are flexible, open-ended and highly dependent on the utterance context in which they are embedded”.

Information, as here elaborated, is not a semiotic notion but is a more general cognitive notion where data and contextualization yield transient informedness which leaves a residue in changed mental entities, physical states, activities, whatever. These become available subsequently as data for further contextualizations. We find in relation to semiotics that contextualization is semiosis. Information is a word which is used to flag the potential for contextualization.

Information, informally, can be thought of as the distinction between what is and what might have been. This captures the general sense of the word in the usage promoted here, but also more formal ‘information theoretic’ notions of information whereby probabilities are analysed to determine the significance of information in binary schemes. Much has been written on information theory and such material can readily be found (for a starting point the reader can consult Cherry (1978)). It is judged here that the semiotician needs a more cognitively/semiotically oriented account which helps to unpack what is going on when we use the word information.

It is worth noting, for completeness, that information theoretic notions of codes and message transmission (conveying information) have distracted linguists and semioticians for some while. What is essentially an engineering model was widely taken into thinking about linguistics and human communication behaviour – so much so that one finds significant figures espousing such an approach. John Lyons, in Volume 1 of his two volume work on Semantics (1977), introduces the “communications-engineering” model of communication, “based upon the model described in the now classic work by Shannon and Weaver (1949).” Lyons produces a figure based on this model, one which is found widely in the literature (transmitter on the left, noisy channel in the middle, receiver on the right). Umberto Eco, in “A Theory of Semiotics” (1976), refers to essentially the same model in discussion of the notion of codes. Yuri Lotman’s earlier work in the 1960s cites the same sort of model, attributing it to Jakobson (see “Universe of the Mind: A Semiotic Theory of Culture”, English translation 1990). Perhaps more worryingly one can still find this model in use in discussion of communication behaviour – see for example Marc Hauser’s book on the “Evolution of Communication”, (1997).

Cherry, C. 1978. On Human Communication: A Review, a Surevy and a Criticism. MITPress.

Comrie, B. 1983. Language Universals and Linguistic Typology: Syntax and morphology. Blackwell, Oxford.

Eco, U. 1976. A Theory of Semiotics. Indiana University Press, Bloomington, 1979 (Midland Book edition).

Evans, V. 2006. Lexcial concepts, cognitive models and meaning-construction. Cognitive Linguistics, 17(4), 2006, 491-534.

Hauser, M.D. 1997. The Evolution of Communication. MITPress.

Lotman, Y.M. 1990. Universe of the Mind: A Semiotic Theory of Culture. I.B.Tauris Publishers, 2001 (paperback edition).

Lyons, J. 1977. Semantics, Volume 1. CUP

Selkirk, E, O. 1999. The Syntax of Words. MITPress.

Shannon, C.E. and Weaver, W. 1949. The Mathematical Theory of Communication. University of Illinois Press, Urbana.