Semiotics has a rich heritage in pre-computer European scholarship, where it is associated with names like Barthes, Greimas, and Eco, building on the structural linguistic tradition of de Saussure, Hjelmslev, and Halliday. Semiotics goes beyond classical linguistics by including in its domain not only phonological signs constructed and interpreted by people in the process of speech, but also graphical markings, gestures, cultural artifacts, and indeed anything in human society that can convey a meaning. This book is a systematic development of the thesis that semiotics, “the science of signs and their life in society” (p. 3), provides a useful methodological framework for studying systems that include both people and computers.
As computers increasingly dominate human society, it is inevitable that they should come under the scrutiny of the disciplines that study that society. A long tradition, epitomized in the work of Noam Chomsky, associates computer science with linguistics; in the light of that tradition, linguistics has seemed at least superficially adequate for describing systems in which people communicate with computers through keyboards and text displays. The recent proliferation of graphical user interfaces forces the study of human-computer interaction (HCI) beyond linguistics to consider the structure and use of a wider set of signs, and Andersen quite reasonably proposes that semiotics is a ready-made framework in which to consider these issues.
After an introduction surveying the objectives and overall line of the argument, the book has three main parts. The first reviews the major ideas of linguistic structuralism and semiotics. The second applies these concepts to computers, developing the theory proper referred to in the title. The third section broadens the scope of analysis further to include not only computers but the linguistic and work context in which they are used; this section draws heavily on an empirical study of a large systems development project at a business office in Stockholm.
The work’s major contribution may be that it exposes computer scientists to the rich structuralist and semiotic heritage. This extraordinarily fertile set of ideas has had considerable impact in linguistics, anthropology, sociology, and philosophy, and will bring much to the study of human-machine interfaces. In particular, Andersen’s work provides a language and methodology for talking about and designing human-computer interfaces. Consider these examples.
The structuralist tradition suggests defining the object of study as neither human language nor interface symbols, but the computer-based “register,” the union of a computer interface (with computer-generated signs) and a human work language (which introduces the discourse-linguistic concept of a “register”). This approach avoids getting mired in the full AI problem of natural language recognition, yet can still deal intelligently with people and their constraints on a computer system.
Andersen proposes an extension from object-oriented programming (OOP) to sign-oriented programming. OOP focuses attention on computer-generated artifacts and their relations among themselves. Sign-oriented programming includes the relation of these objects to the people who manipulate them. They become elements in a play that gives the audience experience of and insight into some real or imaginary part of the world.
The notion of a “story grammar” permits the analysis of the computer-based register to move beyond the individual task to a complete system of tasks, and provides a tool to model not only what the computer does or what people do, but the integrated human-computer system.
The overall thesis is that semiotics is formal enough to handle the computer side of HCI, and human enough to manage the human side. Andersen’s work will be of interest to educators, researchers, and implementors of HCI. The book does not pretend to survey the entire field of HCI or summarize alternative approaches, and so will not be widely used as a text except by those who share the author’s orientation to semiotics. It provides excellent support material for such a class, however, particularly since the author has been careful to indicate material that can be skipped on a summary reading. Because it systematically develops how semiotics can be applied to the study of human-computer interactions, HCI researchers will find the book stimulating, even if their own approach is different. Finally, systems builders who are languishing for lack of a methodology could do worse than to review Andersen’s approach.
