Organizational Semiotics: Evolving a Science of Information Systems IFIP TC8 / WG8.1 Working Conference on Organizational Semiotics: Evolving a Science of Information Systems July 23–25, 2001, Montreal, Quebec, Canada

Organizational Semiotics: Evolving a Science of Information Systems covers such issues as:
-Fundamental concepts such as 'information', 'data', 'message', 'communication', 'knowledge', 'organization', 'system' and so on;
-Properties of signs vital to organizational functioning, such as their meanings, the intentions they express and the valuable social consequences they produce;
-'Architecture' of organizations when they are viewed as information systems, based on their semiotics features;
-Understanding language in organizational contexts, for example, the limitations on the language used to conduct business affairs;
-The empirical study of communications for requirements elicitation;
-Applying semiotic categories (e.g. physical, empiric, syntactic, semantic, pragmatic, social) to various problems;
-Organizational knowledge representation;
-Business process re-engineering methods and the design of e-commerce systems.

• Language: English
• Publisher: Springer
• Release date: Mar 19, 2013
• ISBN: 9780387356112

Book preview

Preface

Organizational Semiotics: Evolving a Science of Information Systems

Ronald Stamper

University of Twente, NL and Staffordshire University, UK,

[e-mail]

This is the first IFIP Working Conference on Organisational Semiotics. In developing a new subject we need to consider how to proceed. Thus I choose to direct my introductory remarks towards some issues of scientific method of special relevance to our work.

Only a community of scientists can produce a Science. Individuals produce most works of art. Individuals of no particular qualifications decide whether they like or dislike an art work. Scientists must reach consensus about the theoretical ideas and empirical work of their colleagues. This is not a matter of personal taste but to decide about their truth or falsity. Good scientific methods help us to decide these issues.

That contrast between science and art should not suggest that scientists do not also need to be creative. But we should recognise that the constraints imposed by our having to produce a consistent body of knowledge can be quite inhibiting. We are still in an exploratory stage, which calls for creative thinking. It does not matter that an hypothesis is wrong provided that it is subjected to rigorous critical examination. Popper’s ‘method of bold conjectures and attempted refutations’ (ref 1, p. 53) is quite liberating in this respect. Indeed he emphasises that the bolder the theory, the greater its content and the greater its risk of refutation but should it survive intense criticism, the greater its value.

Creative thinking has combined with rigorous but considerate criticism characterise this particular scientific community. Although this is the first IFIP meeting on this topic, each of the previous six years has seen an international meeting in this or a closely related area:

Of these the last three have published proceedings (Refs 1, 2, 3). These stimulating meetings have played a vital role in building the necessary community and in establishing the commitment to the necessary balance of creativity and criticism.

But before the right scientific conduct can be established we must build the scientific community. The face-to-face meetings have played an essential role in that. Of course an established science will at least appear to be conducted in its literature. A new and rapidly evolving discipline has greater need for a constant interchange of views and, more especially of talk. We are still developing a stable vocabulary that will allow us to use our shared concepts without risk of confusion. The editing process has contributed to that exchange by arranging more than the usual amount of commentary to help the authors make their revisions. We still have more community building, more work to do on forming a stable terminology and some basic concepts still call for discussion.

These papers provide evidence of convergence within our community, especially on the problems we must address and on many of the key concepts. Let us examine some of them, with special regard for issues of scientific method, while adding to our research agenda.

Foremost is our agreement to build a science of information systems using the operationally secure, primitive notion of a sign. On that concept others can be constructed with sound operational definitions. The conceptual strength of signs is that we can demonstrate what signs do and how people use them for getting things done in organisations. This operational clarity allows signs to be studied empirically, an essential for a science of information. The Peircean semiotic triangle (a sign-token representing some thing through an interpretant) suits most authors — especially as Peirce’s ‘interpretant’ at the apex allows some ambiguity of interpretation. It also induces more elaborate models involving two or more linked triangles, a notion I shall employ below.

Departing from the mainstream of IFIP thinking, we largely accept that reality (as far as we can know it) is a social construction that depends on our use of signs. Some who take it that our social reality is socially constructed may balk at arguing that the physical reality we know is also socially created. Personally I go all the way, because I feel great discomfort adopting an objectivist position with respect to some of reality and a social constructivist for the rest — but I’m ready to justify my views any time! This key issue should remain high on our agenda because it underpins so many other aspects of our work.

We can see social construction in action within these conference papers, which examine many basic concepts. A science of information systems must stabilise its main concepts. So, methodologically we must ask how should do that.

In this meeting several authors contributed to the construction of the concepts:

No one will deny the central roles that these play in both theory and practice. Most of our colleagues tend to treat these notions as basic ones that are self-explanatory. Recognising the danger of this naïve assumption, our sponsoring Working Group, IFIP WG 8.1, created the FRISCO Task Group in 1988 to arrive at a framework of information systems concepts, including data, information and knowledge and many others (Falkenberg et al. 1998). Over a period of ten years, meeting roughly each quarter for two or three days, writing papers and disputing about them, we failed to reach unanimity: hence the continuing importance, relevance and difficulty of these issues.

Semiotic applied to Organisational Semiotics provides some methodological guidance on this problem. Definition plays a key part in our discussions. To help us reach some agreement about what constitutes a good definition, we can use the Semiotic Triangle about which we already agree with Peirce: A sign ... is something which stands to somebody for something in some respect or capacity. As Peirce states, this creates another sign, which he calls the interpretant of the first sign. A definition can play the part of an interpretant. A definition (a sign, of course) expresses the link between the term we wish to define, the definiendum, and the expression that explains it, the definiens. A good definition will substitute for the definiendum terms that are more concrete and, ideally, easier to understand. As a secure basis for theory construction and empirical studies, we need to link our terms firmly to ‘reality’ (whatever we mean by that difficult term) so that our colleagues will all make the same connection between sign and object. So what criteria must a good definition satisfy?

Definition is not just a matter of words. For a definition to work, the people using it must share a number of skills that are revealed in Figure 1, an expanded version of the Semiotic Triangle.

Figure 1.

Definition as a norm linking sign and object as the interpretant in a Peircean Triangle

The figure shows the definition at the apex as a norm, that is to say a pattern of behaviour that users of the definition must shares. The interpretant is a complex norm and the semiotic triangle shows how we build it from other norms. Henceforth I shall use the term definition to refer to this norm.

Before we can construct any definition that a community can share, the members must share certain perceptions about the world. We must recognise:

a)

the sign-vehicles used in the definition in any of their accepted shapes or forms; and

b)

the objects that the terms in the definiens stand for.

We want the users of the definition to link the definiendum to reality in an operationally reliable manner. Usually the definiens is a complex verbal expression incorporating many other terms that people may or may not be expected to know. If they do not know the terms we are using in the definiens, we can recursively define them and any terms they employ using other verbal expressions. The challenge is to escape from this chain of verbal definitions that only trade words for words. This is where the strategy of examining the norms involved can help us.

The figure shows that the definition is a norm built upon two other norms, both perceptual norms. The users of the definition must have acquired these perceptual skills. We should be able to check operationally in our discussions that other members of the community who want to use the definition actually do share these perceptual norms and can use them in a reliable way.

This is trivially easy for a) the sign-vehicles. We can produce any number of concrete examples of them. The sign-vehicle must necessarily be a sign-type so that we can use different sign-tokens of that type in any number of places at any number of times when printed in various fonts (dog, DOG, dog, dog ...) or uttered various accents. This is easy for a literate community, they have the requisite linguistic norms.

The difficulty arises with the perception of the complex of objects used in the definiens. In the simplest case, the readers can relate all the terms in the definiens to real objects directly, that is without having recourse to further definitions. If we can deal with this simplest case, we can deal also with chains of definitions by applying our solution recursively. Eventually we reach the simple case. At that point the interpretant ceases to be a purely linguistic norm. Unless it links a linguistic type to an object or type of object, fails to bridge the gap between the semiological realm and reality. Reality is the world we experience directly, the world where scientists conduct experiments and about which they find inter-subjective agreement. The difficulty arises when different parts of the community include different categories of things in the reality about which they theorise when they reach the primitive notions that recursively underpin their all their other definitions.

These privileged terms that we can understand without the aid of verbal definitions are a most precious resource: they serve as the foundation for the definitions of other concepts. To find them, we must look for terms that stand for things that we can learn to recognise, intersubjecively and operationally by a mixture of doing and talking. This we term ‘ostensive definition’ to distinguish it from the purely linguistic kind of definition.

The importance of the concept of a sign for an empirical science lies in the fact that we can define it ostensively. Thus we can introduce anyone to the verbal expression ‘a sign’ and then proceed to show them any number of diverse, concrete examples, while probably giving them a commentary on how some people use the sign to stand for something or other in some circumstances. For example, find a board with the words Low bridge — 3 m clearance at a road junction; this physical object has physical properties — it might fall on your head, in which case it would help demonstrate some aspect of physics — but its function as a sign is revealed by watching the high lorries that take the other route because the driver does not have physically to encounter the bridge to know how best to act. After every few such illustrative demonstrations of our use of one thing (the sign) to stand for another (the object), we ask our student to supply an example of her own so that we can check that she is using the term ‘sign’ in the accepted way. In this way we teach the student the perceptual norm that does not involve language. The sign ‘sign’ understood this way is not an object in the mind of any person but it belongs in a world of real actions that we can demonstrate to anyone. At least we start, in this Working Conference, with this, one, well-founded, primitive notion.

Where do we go from there? The papers we shall be considering offer a variety of routes for us to consider. To add to the material for discussion, and to illustrate my points about definition, let me outline my own treatment of the term ‘information’.

My own journey into the field of organisational semiotics began when I tried to be precise when using the word ‘information’. It has so many meanings that I was led to use sign as a primitive notion. To introduce greater precision into the everyday, intuitive understanding of ‘information’ we can begin my examining some of its uses. For example:

1.

The CIA can give us massive amounts of information — gigabyte after gigabyte!

2.

We get almost as much information about a person’s lung capacity from their age, height and weight as measuring it directly with this complicated machine.

3.

Given what we know about the disposition of the enemy’s forces, from this intercepted message sent by their general we can deduce all the information we need for our defence.

4.

As a manager, I get a surprising amount of information about the morale of the staff by taking lunch with them whenever possible.

5.

All that speculation about the case contains far less information than this short statement from the lawyer on the other side.

6.

This manual contains all the information you need on routine repairs to this engine.

In the various cases the speaker uses the sign ‘information’ in a different respect or capacity, and hence with quite different meanings.

When thinking about definitions, we have a tendency to ask What is the meaning of ‘information’ (or anything else)? as though there must be one answer only, given that we have only one word. We are even tempted to reify information, treating it as a kind of mysterious substance that ‘is contained in’ (as we say) the sentences we use.

To overcome that misconception, notice in the above examples of everyday usage a number of quite different respects or capacities, relating to our talk of ‘information’. In the six cases the speakers have the following very different concerns:

1.

The physical load of information — perhaps they are discussing how to store all the information they will be given. They are not concerned what it means to anyone or how it will be used. So we can measure their information as a number of distinct patterns, the bytes they can count. (ref 6. Fig 10, p. 387)

2.

The empiric properties of one stream of observations that may be found in the medical records (age, height and weight) that can save time making costly observations. So their best measure is mutual information, a precise statistical parameter derived from a large set of repetitive data, and based on empirical probability. (ref 6. Fig 11, p. 388)

3.

The syntactic properties relating to a unique situation which has been mathematically modelled so that every item of intelligence can limit the range of possible scenarios. So their measure of information might be (though I have never seen this applied in practice, yet) one of the Carnap-Bar Hillel measures based on logical probability. (ref 6, Fig. 12. p. 390)

4.

The semantic properties of observations that rely almost entirely on intuition. Experience enables the manager to know what they mean and he can take them into account in his decision-making, we may never know how although we might build a syntactic model (as in 3) to theorise about it. However, in respect of a given decision about a threatened strike, and in the manager’s capacity as a negotiator, we can measure how he adjusts his odds. So we can measure the information he derives from the last lunchtime encounter using the subjective probabilities he assigns to a number of possible outcomes. (ref 6, Fig 13. p. 392)

5.

The pragmatic properties of the communications in the legal dispute they are discussing concern the obligations, hopes, expectations etc that the messages generate, the odds on outcomes (as in 4) may be taken into account here. So we need but do not yet have a satisfactory measure of pragmatic information, although the Language Action Paradigm could probably provide a solution by measuring the information with respect to a limited model of a conversation. (ref 6, Fig. 14, p. 394)

6.

The social product of assembling data from many observations and theories consists of rules, advice, engineering and scientific laws as well as some ordinary laws relating to the safety of the engine. ‘Information’ is being used here as a synonym for ‘knowledge’. To measure information in this sense, we should have to find canonical representation for items of knowledge in order to begin constructing a measurement. As a first step we shall need an operationally precise definition of ‘knowledge’ and my own preference is to base it on the equality knowledge = norms. (ref 6, Fig. 15, p 395)

So we have six different meanings of ‘information’ precise enough to lead to many operationally well-defined measurements, certainly in cases 1 to 4, and with a little more effort for cases 5 and 6. Each of these definitions captures some but never all aspects of the common sense, intuitive notion of information that now confuses talk of information systems because of its seductively dangerous, slippery mixture of meanings. Information looks like a straightforward concept but is not.

Working on organisational semiotics, we are duty bound to find ways of anchoring our key concepts to ‘reality’ by operational means that anyone can test in a public arena. You may think that this is unnecessary or over ambitious. I have certainly often encountered the view that ordinary usage is good enough, and I have even been told at meetings of another IFIP working group that being very precise is dangerous because it hinders imaginative thinking.

I am all in favour of imaginative thinking. We must often introduce terms that capture the flavour of our thoughts well before we can pin upon them exact meanings. In the early development of a new subject such as organisational semiotics, I would expect this to be quite common. Nevertheless, every such imaginative usage should automatically raise the question: How do we make the meaning of this term clear and operational? My own attempt to do this I have given in ref. ?? and urge my colleagues to examine it critically. The best way of doing that is by using the more precise definitions to find whether or not they work in practice.

The underlying six levels that I introduced as SF, a Semiotic Framework (or Ladder), have been applied by several authors in these proceedings. It is gratifying to see that this simple analytical tool used to elucidate so many different issues. Finally I feel it appropriate to provide a note on this Framework to help colleagues who might be interested in using it.

The Framework is based on a few simple principles:

Anything can function as a sign, possibly in many ways simultaneously.

Certain sets of properties of signs can be studied independently of other sets of properties. These sets of properties define the levels on the SF.

On the different levels of the SF we have different basic units to investigate.

You can take someone by the hand to see these units and become acquainted with performing operations on them, counting them, testing their relationships etc. (ostensive definition).

One can also narrow down investigations by looking at the X level aspects of a Y level issue eg: the semantic aspects of illocutions (a pragmatic issue).

The cells of an SF by SF matrix can be used to narrow down discussions of sign properties even more closely.

Let me propose that we could usefully clarify our terminology by examining the meanings of our key terms on each level or even within each cell of the SF by SF matrix. This should help us to remove the mystery and semantic confusion surrounding ‘information’, ‘knowledge’, ‘meaning’, ‘communication’, ‘relevance’ and other essential but challenging words. Without their having precise, operational meanings, we shall be unable to reach agreement over any theories about them. Moreover, those theories will evade and attempts at empirical testing by the simple device of adjusting the meaning to fit our observations.

REFERENCES

1.

Popper, Sir Karl, 1979, Objective Knowledge: An Evolutionary Approach, Oxford, Oxford University Press.

2.

Paerron, Paul, Marcel Danei, Jean Umiker-Sebeok and Anthony Watanebe, 2000, Semiotics and Information Sciences, Toronto, LEGAS.

3.

Liu, Kecheng, Rodney J. Clarke, Peter Bøgh Andersen and Ronald K.Stamper (eds), 2001, Information, Organisation and Technology: Studies in Organisational Semiotics, Boston, Dordrecht, London, Kluwer Academic.

4.

Liu, Kecheng, Rodney J. Clarke, Peter Bøgh Andersen and Ronald K.Stamper (eds), 2001, Coordination and Communication Using Signs: Studies in Organisational Semiotics — 2, Boston, Dordrecht, London, Kluwer Academic,. ISBN 0-7923-7509-2

5.

Falkenberg, E, W. Hesse, P. Lindgreen, B.E. Nilsson, J.L.H. Oei, C. Rolland, R.K. Stamper, F. J.M. van Assche, A.A. Verrijn-Stuart, K. Voss, 1998. A Framework of Information Systems Concepts, IFIP Geneva, on line edition at ftp://ftp.leidenuniv.nl/fri-full.zip

6.

Stamper, RK, 1996, Signs, Information, Norms and Systems, in Berit Holmqvist, Peter B. Andersen, Heinz Klein and Roland Posner (eds), Signs at Work, De Gruyter, Berlin, ISBN: 90-14-05425-2, pp.349–397.

]>

Exploring the Explanatory Power of Actability

—The Case of Internet-based Software Artefacts

Pär J. Ågerfalk¹ , Fredrik Karlsson¹ and Anders Hjalmarsson²

(1)

Dept. of Informatics (ESI), Örebro University, SE 701 82, Örebro, Sweden

(2)

School of Business and Informatics (IDA), University College of Borås, SE 501 90, Borås, Sweden

Pär J. Ågerfalk

Phone: [phone]

Fredrik Karlsson

Phone: [phone]

Abstract

This paper is an enquiry into the empirical grounding of actability. In the paper the operationalization and application of an analytic framework based on actability and organisational semiotics is described. The analytic framework has been used as a tool in a qualitative analysis of the Internet-based software artefact. The results show that actability and organisational semiotics can effectively be used to gain understanding of specific information systems phenomena.

Key wordsActabilityWebSemioticsInternet.

1. INTRODUCTION

Actability has been proposed as an important concept for the understanding of information systems pragmatics (e.g., Goldkuhl and Ågerfalk 2002). It has been defined as (ibid.): ‘an information system’s ability to perform actions, and to permit, promote and facilitate the performance of actions by users, both through the system and based on information from the system, in some business context’. The concept draws on the body of knowledge developed within the fields of Human-Computer Interaction (HCI), Information Systems Development (ISD), and the Language Action Perspective (LAP) and Organizational Semiotics (OS). The aim has been to reconcile these different knowledge domains to gain a better foundation for the understanding of information systems, including their development and businesses usage.

In a sense, actability can be understood as summarizing a theory of information systems as information action systems—an action theory of information systems (Ågerfalk 1999). Theory should in this context be understood as consisting of a set of related concepts directing attention to certain important aspects of information systems that have, at least partly, been overlooked by other similar conceptualizations. One characteristic of the theory is that it is concerned with human action in relation to information systems and that human action also is its ultimate purpose. It represents knowledge about action, intended for action. Goldkuhl (1999) refers to such knowledge as action knowledge. According to Goldkuhl (ibid.), action knowledge should be grounded in three different ways in order for it to be considered valid. The three types of grounding of action knowledge are referred to as (ibid.) internal grounding, external theoretical grounding, and empirical grounding. Internal grounding has to do with the completeness and coherence of the concepts. External theoretical grounding relates the knowledge to other existing knowledge within the same domain. Empirical grounding has to do with the applicability of the knowledge in relation to empirical data. To date, most work on actability has focused on internal grounding (e.g., Goldkuhl and Ågerfalk 1998) and external theoretical grounding (e.g., Cronholm et al. 1999; Goldkuhl and Ågerfalk 2002). Less effort has so far been put into empirical grounding, even though some work was reported by Ågerfalk (1999).

The goal of this work is to strengthen the empirical grounding of actability with a special focus on an analytic framework for analysing information systems phenomena according to actability. More specifically, actability is used to gain a better understanding of the Internet-based software artefact (IBSA). The chapter aims to provide concrete examples of actability concepts and to establish that these concepts are useful to discuss this type of system. The results of the analysis will be used in our further research on methods to support the development and evaluation of such artefacts. However, in this chapter we will focus on the analysis per se as a way towards empirical grounding while exploring the explanatory power of actability. That is, we aim to present good, empirically justified reasons as arguments for the concept of actability.

The work was performed in co-operation with an industrial partner, Volvo IT. Volvo IT’s interest in the work is to gain a better understanding of Internet-based software in order to tailor their development process according to its needs.

The choice of IBSAs as the topic for empirical grounding of actability is not self-evident. The reason for doing so is that Internet-based systems often imply that communication between human actors (typically between supplier and customer in a business setting) is to a large extent performed through the software artefact. Therefore, the qualities promoted by actability are particularly crucial in that type of system. Hopefully, this will become clear as the concept of actability is more thoroughly discussed in Sections 3 and 4 of the chapter. Firstly, however, we will elaborate on the approach used for the empirical grounding of actability and its relation to internal and external theoretical grounding. Secondly, we will present actability in more detail, and discuss its operationalization into the analytic framework. Thirdly, we will present examples from the analysis of IBSAs performed using the analytic framework (details are presented in Karlsson et al. 2001). Finally, we will conclude the work by reflecting upon the performed analysis and the explanatory power of actability.

2. OPERATIONALIZATION AND VALIDATION OF ACTION KNOWLEDGE

As suggested above, actability can be regarded as an instance of what Goldkuhl (1999) refers to as action knowledge, i.e., ‘theories, strategies and methods governing people’s action in social practices’. Action knowledge might exist in different forms of abstraction—from ‘pure’ abstract theoretical knowledge, to knowledge directly applicable in everyday situations (Ågerfalk 2001a).

The concept of grounding of action knowledge assumes that it is possible to argue the validity of the knowledge in inter-subjective dialogues (Goldkuhl 1994). This view, with its roots in Habermas’ (1984) social-critical concept of rationality, is the key to an important distinction between true and valid—something is ‘true’ if it is accepted as valid and useful. ‘Claiming the validity of knowledge is presenting good reasons as arguments for the knowledge.’ (Goldkuhl 1999). Hence, grounding of action knowledge is to present such good reasons for it that other people will accept it as valid. This is supposed to be done by argumentatively relating the focused knowledge to other existing knowledge (external theoretical grounding), and to empirical observations (empirical grounding). Furthermore, the concepts used and their anchoring in values need to be consistent and free from ambiguities and internal contradictions (internal grounding) (Goldkuhl 1999). By ‘grounding’, we refer to the combination of the two intertwined, but at least conceptually distinguishable, activities of generation and validation of knowledge.

A particular concept can be grounded internally, externally, and empirically. The result of empirical grounding can be focused on as a concept in its own right and be exposed to its own three-way grounding process. Note that this implies that empirical grounding might also include elements of internal and external theoretical grounding. Figure 1 shows the grounding of the analytic framework used in this chapter.

The principle is that the grounding of an operationalization (i.e., of a more concrete representation of a phenomenon), in this case the analytic framework, internally, externally and empirically yields an empirical grounding of the concept of which it is an operationalization, in this case actability.

As indicated in Figure 1 (the ellipsis), there are other possible operationalizations of actability worthy of study. It is, however, beyond the scope of this chapter to go into details of those (see (Ågerfalk 2001a) for further elaboration on this topic).

For the empirical grounding of actability we have adopted a qualitative research approach that combines deductive and inductive strategies. We refer to this way of working as reflective research (cf. Alvesson and Sköldberg 2000). The adopted strategy is comparable to what Walsham (1995) describes as the approach taken when a researcher uses theory as part of an iterative process to collect and interpret data. Figure 2 illustrates the way this strategy is used.

Figure 2.

Grounding of the concepts discussed in this chapter.

Figure 2 should be interpreted as follows. Actability and the semiotic framework (Stamper 1994) have been (1) operationalized into (2) an analytic framework (see Section 3). This work also led to a deeper understanding of actability as such (3), and also served as (part of the) external theoretical grounding of actability. The analytic framework has been used (4) to direct attention during data collection (5) at Volvo IT. Collected data has then been abstracted (6) into categories describing various aspects of the IBSA (7) with a Grounded Theory-inspired approach (Strauss and Corbin 1998). During this abstraction, the analytic framework has also been used (8) as a tool to direct attention to relevant categories and phenomena. Additionally (9), categories have been related to, and inspired by, existing knowledge about the IBSA (e.g., Conallen 2000). The emerging understanding of the IBSA has had continuous repercussions (10) on the operationalization process (and thus indirectly on actability itself, as well as on the analytic framework) and on the data collection.

3. AN ANALYTIC FRAMEWORK BASED ON ACTABILITY

From the definition of actability (see Section 1) we can observe some central concepts for actability. Below we will address these concepts in turn to eventually arrive at an analytic framework that can be used to understand information systems phenomena according to actability—an operationalization of actability, as discussed in Section 2 above. Before doing that, however, we will present a general information systems framework referred to as the semiotic framework (e.g., Stamper 1994), as we will use it as an inspiration and point of reference for our analytic framework.

3.1 The Semiotic Framework

Stamper (1994) introduced the semiotic framework as an attempt to create one single conceptual framework able to capture both the social and the technical aspects of information systems. The fundamental concept for the semiotic framework is that of the sign. According to Stamper (ibid.), a sign is ‘something which stands to somebody for something in some respect or capacity, in some community or social context’. Consequently, information is ‘carried’ by signs of different kinds (ibid.). Following the semiotic framework, signs (and hence information) can be studied at six different semiotic levels. That is, we can choose to focus on different aspects of signs ranging from their physical appearance to their social consequences. The six levels are referred to as physical world, empirics, syntactics, semantics, pragmatics and social world. The first three of these are referred to as the IT platform and the latter three as human information functions (ibid.). The IT platform can thus be thought of as a medium for the human information functions, and the same information functions can exist within different media. However, within the human information functions, the distinction made between social world and pragmatics is problematic. Rather, as Goldkuhl and Ågerfalk (2002) argue, these two levels are so intertwined that distinguishing them is probably not only conceptually inelegant, but also misleading. Stamper (2001) states that ‘Semiotics that excludes norms and attitudes as forms of information would be like physics with the concept of energy but without the concept of mass’ as an argument for adding the sixth level to the semiotic framework—the social world. We adhere to his argument in concept, but not in the way it is used within the semiotic framework. To understand why, and to see how the semiotic framework has been used in this work, we must first turn to the actability centre of gravity—the performance of actions.

3.2 Performance of Action

People perform actions to accomplish changes in the world—action is about making a difference (Goldkuhl and Ågerfalk 2002). Actions can be classified as material or communicative. Material actions aim at changing the physical state of the world. Communicative actions (or speech acts) aim at changing the socially constructed reality. (ibid.; Berger and Luckmann 1989; Searle 1995).

In this chapter we will focus on communicative actions since they are the most important types of action in relation to the design of information systems. (See (Goldkuhl and Ågerfalk 2002) and (Goldkuhl 2001) for a more comprehensive treatment of both types of actions.) Communicative action means that a pragmatic action mode (illocutionary force) is attached to a semantic propositional content and formulated syntactically into an uttered sentence. The action mode represents what the speaker does in relation to potential listeners. Note that actions are considered to be multifunctional and that an action mode consequently might embrace several illocutionary forces, one per function (illocutionary point). This is a pragmatic aspect of action. The propositional content represents what is talked about and consists of references to things in the world and properties predicated to those things. This is a semantic aspect of action (Goldkuhl and Ågerfalk 2002; Searle 1969; Habermas 1984).

Actions are performed through some medium. The medium is primarily an empirical and physical aspect of action. However, the medium might also affect the possible syntax of an action, and