Theory in Informatics

My contribution to a panel debate in Göteborg, May 4, 2023

1. A theory gives a first answer to the question “What is this?” (Asplund Om undran inför samhället, 1970). The theory uses concepts to define your subject matter and says something interesting about it. The theory guides your further examination. It gives you a language to speak about your subject, indicates aspects to stress, questions to ask. I once heard research in psychology described in this way: “Most of us are digging in one place when suddenly someone walks away and starts digging in another place. Sometimes we all follow him and then psychology changes direction.” When you decide where to dig you use a theory. If most of us are digging in the same place, we have the same theory, a paradigm in Thomas Kuhn’s sense. Informatics is a preparadigmatic discipline so there are many theories to choose from. Often we borrow from other disciplines, from economics, organization theory or philosophy. There are many such borrowed theories: activity theory, transaction cost theory, actor network theory, infrastructure theory, structuration theory, institutional logics theory, discourse theory, sociomateriality theory, and so on. I believe that informatics would do well to develop its own theories, using concepts from the practice, developing those concepts.

2. But what is your motivation for digging in a certain place. How do you motivate your research? The subject matter of the discipline we call informatics is the use of digital technology. It is not the technology, not the human beings using the technology or the way they are organized, but the hybrid combination of people, organizations and technology. Of course, “use” is not the perfect concept to capture this hybrid, but it seems to be the best there is. The motivation for this choice of subject matter is the role technology plays in the lives of human beings. The use of technology is what makes us human. Without the use of technology we were like the other big apes, the chimpanzees, gorillas, orangutans. We use technology in everything we do and we have built societies around our use of technology. Technology is the backbone, circulatory system, infrastructure, medium of our societies. To study human beings and what they do, to study societies, you have to study the use of technology. And in our time digital technology is the most important technology, making informatics the most important discipline. That is why it is not such a good idea to borrow theories from other disciplines. Other disciplines tend not to take technology use seriously enough.

3. Digital technology develops and changes, and changes its name, and our discipline follows wherever it goes. This is a very important characteristic of the discipline and it gives it a changing nature of much consequence. Digital technology is only the current name of the subject matter (and not a very good name either). The identity of our discipline is genealogical. So far we have followed digital technology from mainframes with numerical computing and information systems to work stations to personal computers with desktop publishing to client servers with workflow management and cscw to Internet with the world wide web to mobile phones with apps to digital platforms to artificial science and large language models. Since technology develops and changes, the discipline and its theorizing has to change too. When the discipline was first founded in Scandinavia by Börje Langefors in the 1960s, he developed a theory of information systems (published as Theoretical Analysis of Information Systems in 1966), defining information systems as social organizations, people-technology complexes, distinguishing between data and information. Most of his work is now irrelevant. The technology and its use have changed so we are digging in other places today. The discipline is still often called “information systems”, the name appears in journals, conferences and organizations, and this is sometimes confusing.

4. Technology is made by us, it is artificial, and therefore informatics is not a natural science. It is an artificial science and therefore normative. It wants to make a contribution to the development of technology and its use. It has to follow where the use of the technology goes in order to stay relevant and it has to develop its theorizing. The changing nature of its subject matter makes theorizing in informatics different from the natural sciences. Theories trying to conceptualize the relations between people and technology will change when technology and its use is changing. The theory has to be close to the way the profession speaks about the technology in order to be relevant and be able to influence the development. It should add depth and draw distinctions, but it should stay relevant to the professional practice.

5. The normative nature gives to informatics a future orientation. Since technology use is changing we are curious about what it will become and we want to influence the direction of change. Theories of the hybrid people-technology will always be tentative because when technology develops and technology use changes, so will our relations to technology, and relations between organization, work and technology. There may be some deep, general truths about people and technology, but they are not really our business. We are not philosophers. The important questions for us are how to make the best of the new use of technology.

6. One may wonder if our discipline is really contributing to the development of digital technology and its use. Maybe most of that contribution is indirect through undergraduate education rather than through research. The research has a tendency to deal only superficially with technology, sometimes getting stuck in theorizing the hybrid people-technology turning into philosophy. This was even true of Langefors. And yet, Langefors stressed that the discipline, he called it infology, should be a bridge between engineers and users. The MIS Quarterly editors have recently (March 1, 2023) introduced what they call “significance statements”, asking contributors to include such statements, indicating the relevance of the submitted research to practice. They exemplify the need for more relevant research by pointing to the many pressing policy issues that are raised by digital technology. The commotion created by artificial intelligence is only the most recent example.

7. A surprising example of how digital technology can change things is the role of big data. In his theory of information systems Langefors stressed the importance of interpretation, how people add theory to data. Since then digital technology has made unbelievably large data sets available. The large language models are all about data and very little theory. Artificial intelligence is in effect data science. So why do we have this panel discussing theory? We should have had a panel on data. (Or perhaps not. Even if data plays a greater role today, we cannot do without theory.)

Bo Dahlbom