"The Computer Journal"

Vol. 39, Issue 9

Book Review

Bonnie A. NARD1 (editor)

Context and Consciousness: Activity Theory and HumanComputer Interaction. The MIT Press. 1996. ISBN 0262140586. £33.95. 400pp. Hardbound.

Bonnie Nardi's (ed.) book is one of the most exciting, and at the same time frustrating, books to be published on the subject of humancomputer interaction theory in recent years. Exciting because it clearly recognises a problem in contemporary research namely that 'There is a fundamental need for a theory of practice in humancomputer interaction studies.' Exciting too because it promises that activity theory, a relatively recent import from the former USSR, may provide this long awaited theory. The book is frustrating, on the other hand, because it is difflcult to glean from its pages enough detail about activity theory to confidently assess whether or not it meets this need.

Reading the book in search of activity theory is rather like piecing together a jigsaw puzzle, but one where some pieces are missing, and others don2t quite fit together. For example, one fundamental concept in activity theory is that of the 'object' of an activity. But object is defined in several rlifferent ways in several different places by different chapter authors. Minimally, a glossary of important clefinitions could he provided. The difficult of constructing such a glossary mav reflect a fundamental problem in translating ideas expressed in the original Russian into English. Instead. Nardi and her coauthors prefer a more eclectic approach, in which varied definitions, with varied illustrations, implicitly provide an impression of what is meant by the terms. This approach, whilst perhaps less sterile than a list of definitions. mav leave some readers wondering whether the concepts can be defined with sufficient precision to support a theory of practice in humancomputer interaction.

The book structured into three sections. a basic introduction. some case studies of applying activity theory in practice, and three chapters exploring possible theoretical development. The early chapters introcluce the reader to a rich set of ideas in a stimulating form. .Nardi's own early chapter, which relates activity theory to distributed cognition and situated action theory, is extremely interesting, although I am sure practitioners in these related disciplines will have serious difficulties with Nardi's, at times. simplistic distinctions.

ln and of themselves. the case study chapters provide a wealth of interesting and innovative ideas On how to do contextual research. Quite apart from activity theory the book is worth huying for these chapters. What is slightly disappointing is that it is not always clear how the case studies align to activity theory, and precisely what part activity theory played in their design. Once again the reader gets an impression of what activity theory methods might look like, without being able to clearly identify its defining characteristics.

It may be too much to expect one book to articulate an that is activity theory and this book is certainly not a 'how to do it' guide. However, Nardi makes no such claim for the book. If the authors' aims are treated cautiously as an effort to give the reader a taste of what activity theory has to offer and, minimally, to provide a language and a new way of looking at interaction in conext, then we believe that the book achieves these aims.

We participated in a reading group studying the book and it has certainly affected our outlook on the nature of contextual factors in humancomputer interaction. Perhaps one of the best ways to read the book is to dip into chapters on an interestdriven basis. Don't expect to find activity theoretic answers to your reserach problems, but do expect to find an interesting new way of looking at interaction in context.

PETER WRIGHT AND ANDREW DEARDEN

HCI GROUP, University of York



back to ... Bonnie Nardi, Context and...