Book review of: Context and Consciouness: Activity Theory and Human-Computer Interaction

Edited by: Bonnie A. Nardi

Review by: Shilpa Shukla

Context and Consciousness, edited by Bonnie Nardi, is an intriguing collection of papers by various proponents of activity theory. With the upcoming CHI '97 conference and its theme "Looking into the future" a book such as this one could not have come to the foreground of the CHI community at a more appropriate time. What this book offers for the HCI student and researcher is a new perspective and a conceptual framework for HCI tool design. Perhaps Activity Theory will assist HCI researchers to shape this future.

There is a growing consensus among HCI researchers that the current frameworks for HCI research adapted from information processing cognitive psychology and situated action models have serious shortcoming. Activity Theory presents an alternative framework for HCI research and design which addresses these shortcomings.

Rooted in deep Soviet psychology from the 1920s, this theory proposes the view that tools mediate thought. Activity Theory is based on work by Vygotsky and was developed by Leont'ev. Much like HCI, Activity Theory concerns itself with practice problems and focuses on practice in context with emphasis on the work process level of problems. This theory is concerned with understanding the role of artifacts in terms of how these artifacts exist in everyday situations and how theses artifacts are integrated into social practice. The purpose of Activity Theory is to understand the coupling of consciousness and activity. What Activity Theory offers is a set of conceptual tools applicable in various situations.

Victor Kaptelinin, the author of Chapter 5, presents principles pertaining to: (1) consciousness and it's relation to activities, (2) the social and cultural properties of the environment (referred to as the object-orientedness of activities). This principle contrasts with the cognitive approach which believes that the human mind contacts reality only through low level input-output processes [p. 108]. (3) Understanding the hierarchical level of activities. The levels in this structure help to differentiate between processes at various levels. The processes distinguished are activities, actions and operations. The process of making these distinctions comes from the need to understand and predict the changes in peoples' behavior in different situations. In Chapter 2, by Kuutti, the reader will find a more detailed explanation of the hierarchical structure of an activity, (4) understanding the internalization-externalization of artifacts. As discussed in Chapter 3, in regards to the IPA concept, the same reference is made by this principle. That is, it describes the mechanisms underlying the origin of mental processes, (5) understanding the mediation of human activities through tools (which are considered to be carriers of cultural knowledge and social experience), (6) the development of complex phenomenon.

In Chapter 4 (Studying Context: A Comparison of Activity Theory, Situated Action Models, and Distributed Cognition), Bonnie Nardi provides a comprehensive set of comparisons between the various concepts from activity theory, distributed cognitive approach and situated action models. In previous chapters the emphasis was placed on teasing out the more subtle highlights of activity theory by presenting general frameworks and principles. The difficulty in understanding the differences among the various approaches may not be due to the principles of the approaches themselves but due to the abstract and subtle issues involved in HCI.

While Situated Action is not considered to a be theory unworthy of application for design use, in isolation this theory as well as the cognitive approach fails to address many interesting and problematic issues.

The view of Situated Action's approach is "that the structuring of activity is not something that precedes it but can only grow directly out of the immediacy of the situation [p. 72]." In the Situated Action model emphasis is placed on the emergent, contingent nature of human activity, the way activity grows directly out of the particularities of a given situation. While cognitive science approaches within artificial intelligence view human activities in a highly rationalized fashion with overemphasis on plans in shaping behavior [p. 72], Situated Action Theory grew out of the need to consider the necessity of particular situations and the emergent, contingent nature of actions.

According to Nardi, the major difference between Activity Theory and Situated Action is the way in which the structuring of an activity is determined. In Activity Theory the structuring of an activity is determined by human intentionally before the unfolding in a particular situation, whereas in Situated Actions, activity "can be known only as it plays out in situ [p. 82]." The goals and plans, according to situated action theory, are determined after the activity has taken place. These goals and plans, then become "constructed rationalizations" that can be applied to particular situations.

By contrast, in Activity Theory and in Distributed Cognition Theory, an "object-goal" is decided upon prior to an analysis. An object is said to proceed and motivate an activity. An object "is (partially) determinative of activity [p. 80]," where as in situated action, each activity is created by particular factors that converge to form a situation. Thus Situated Action rejects the notion that objects engage activity. Further more, in Situated Action theory objects (goals) and plans are "retrospective reconstructions" and post hoc "artifacts of reasoning about action," after the action has taken place [p. 80]."

Hence, although situated action theory provides insight for the need to shift from the traditional cognitive science model of inflexible plans and goals and asks us to view activity as dynamic, it does not provide the researcher with a vocabulary for describing activity. It fails to recognize the "conscious" decisions made by a person to arrive at a situation. Nardi points out that the "situation" that one finds themselves in is created, in part, by the desires of a person to carry out some activity.

Nardi ends the chapter with a set of methodological implications of activity theory for HCI studies. These implications are based on the key ideas presented by authors in previous chapters. The implication for activity theory are: 1. a research time frame long enough to understand user's objects, 2. attention to broad patterns of activity, 3. the use of varied set of data collection techniques, 4. a commitment to understanding things from the users' point of view.

What activity theory instructs is that the practitioner start an analysis with the object made consciously by the subject. This point made by Nardi is unfortunately difficult to understand. This difficulty can be attributed to the new jargon, the abstractness of the issues discussed and the subtlety of the message conveyed. Regardless, as mentioned earlier the reader must be patient in reading these chapters.

Overall Part 1 is quite thorough and enlightening. It can certainly be a source for inspiration and foundation for future HCI researchers. Yet the theory is not without its flaws, as mentioned succinctly at the end of Chapter 3. According to Kaptelinin the primary limitations of activity theory can be numerated as: (1) Activity theory was originally developed for understanding individual activity. Yet, "the user" should include the group, the organization and the individual. (2) Activity theory has adopted a very narrow view of culture. Yet, activity theory "cannot completely substitute for an anthropology that defines and understands culture [p. 64]." (3) Activity theory's perspective on tools mediation has not fully anticipated all of the representation problems in virtual reality realms. Thus, this theory must adapt new ideas from the cultural-historical traditions and other related approaches [p 64]. (4) Activity theory is not operationalized enough. This field still lacks sufficient methods and techniques that can be utilized directly to solve specific problems.

The value of the book appears to lie in Part 1. The reader will find in the remaining two parts topics pertaining to the practical application of activity theory in design and theoretical development of activity theory.

Part Three of the book consists of five chapters. Chapter 6 (Designing Educational Technology: Computer Mediated Change) focuses on practical application in the field of education with some insightful analysis of the application of activity theory framework as applied to K-12 education. Chapter 7 (Applying Activity Theory to Video Analysis: How to Make Sense of Video Data in HCI), Chapter 8 (Tamed by a Rose: Computers as Tools in Human Activity) and Chapter 9 (Joint Attention and Co-Construction of Tasks: New Ways to Foster User-Designer Collaboration). These chapters provide examples for the application of activity theory in HCI research ranging from activity theory techniques for video analysis of various complex activities involved in observing the use computer applications, activity theory techniques for managing joint attention and techniques for eliciting common view of a domain, and activity theory techniques applied in combination with traditional ethnographic approaches.

The significant contributions provided by Chapter 6, by the author R.K.E. Bellamy, are the "three principles for the design of educational environments" from Vygotsky's work. The three principles display the need for: (1) authentic activities, (2) constructive environments, (3) collaborative environments in education. The reader interested in field of educational technology will find a great deal of discussion on Vygotsky's concepts and its application to design problems. The central argument made by Bellamy is that technology can serve as a catalyst for educational change. Two examples are provided by the author to demonstrate how technology-based educational environments can be designed based on principles derived from Vygotsky's theories of child development.

Chapter 7 and 8 are useful in that they provide interesting examples for the applicability of activity theory, yet the style of presentation of theoretical ideas in these chapters is highly dense. The contribution of Chapter 7, written by Susanne Bodker, lies in the introduction of the concepts "breakdowns" and "focus shifts" as pointers for understanding what role applications play in successfully or unsuccessfully mediating work activity. With the help of these two concepts from activity theory, one can hopefully identify problems of mediation in designing an application.

Chapter 8, by Ellen Christiansen, is rich with activity theory applications in practical or real life scenarios (namely, case studies of computer use by Danish detectives). Unfortunately the reader may find frustration in grasping the argument, which is highly abstract and hidden amidst the use of analogies. The author, through case studies, attempts to generalize a hypothesis about the possible key factors that lead an artifact to move into an activity as a tool [p. 178]. The argument made by the author is basically that activity is both culturally and socially formed by the individual in a community of practice and by the role. The subtlety of this concept makes it difficult to understand the role of tool mediation.

In Chapter 9, the authors Arne Raeithel and Boris M. Velichkovsky, discuss the use of representations to assist the design process. The authors shifts focus from the operational level to the actions level of design. The significant contribution of the chapter is the presentation of the idea of co-construction. The authors through use of examples explain how co-construction of a task and goal structure of an activity can be done between domain experts and designers. The unique contribution of Chapter 9 is a table that presents a possible optimal match between evaluation methods and concrete research goals.

Chapter 10 (Some Reflections on the Application of Activity Theory), by Bonnie Nardi, is a chapter in which the author reflects on how the use of activity study would have been a better choice in analyzing her data from field study of slide makers. The use of the slide makers case study is useful in gaining an intuitive understanding of the applicability of activity theory in field studies of software applications. The goal of the study was to understand if end users prefer "task-specific" or "generic" application software. The author provides a lucid comparison between analysis of data from the field study before the application of activity theory and after. The study indicates that activity theory provides some of the following analysis support: 1. a unified structure useful for describing what is being observed, 2. a conceptual vocabulary to describe concepts such as sub tasks and end-to-end task specificity, 3. a way to describe the conscious intent of a goal (through activity theory's notion of goal-action relation) which is formulated to represent the action necessary for the fulfillment of the object, 4. a notion of subject as an individual or a group.

Part Three of the book consists of three chapters. Chapters 11 (Activity Theory and the Views from Somewhere: Team Perspectives on the Intellectual Work of Programming), Chapter 12 (Developing Activity Theory: The Zone of Proximal Development and Beyond) and Chapter 13 (Mundane Tool or Object of Affection ? The Rise and Fall of the Postal Buddy). These chapters investigate possible theoretical extensions to activity theory. The contribution of Part Three can be fully appreciated by the reader who wishes to augment their knowledge of activity theory, however the reader may find frustration in understanding the abstruse theoretical concepts.

Chapter 11, written by Dorthy Holland and James R. Reeves, discusses the issue of studying group perspectives. The authors in this chapter focus on the role of collective activities in public arenas for the joint construction of a social reality. Chapter 12, written by Vladimir P. Zinchenko, is based on the activity theory notions of functional organs and the relation of internal and external forms. The author approaches the above mentioned issues with an interesting angle, that is understanding activity through a spiritual dimension. A word of caution for the reader here is that the term "spiritual" is not referring to a belief in the "supernatural (though it could be) but is rather a property of human consciousness that exhibits the means to resist and to create." The author connects the "spiritual" approach to technology. The contributions made are quite fascinating. Chapter 13, written by Yrjo Engestrom and Virginia Escalante, explores what is referred to as "actor-network" theory for possible enhancements to activity theory. The primary focus in this chapter is on exploring a means of dealing with interactions among multiple activity systems. The authors add some new concepts to activity theory based on the an analysis of a system, the "Postal Buddy," which was based on the actor-network theory.

What the authors of this book do not advocate is that activity theory is the "silver bullet" for the HCI community. Rather, a combination of cognitive science's theories of distributed cognition, situated work theories and activity theories can be used to support a design practitioner in their design and field research endeavors. Furthermore, authors of this book (specifically in Part Three) emphasize the need for expanding activity theory and integrating other useful theories with activity theory.

>From the perspective of an HCI student, the essence of the book lies in Part One. For it is in Part One that the novice student exploring activity theory becomes familiar with the theory's jargon, and gains insight of activity theory principles and frameworks.

Part 1 of this book consists of five chapters. In these chapters the authors offer principles and frameworks as an introduction to Activity Theory. These chapters also offer comparisons between activity theory and theories such as the information processing theories of Cognitive Psychology and Situated Action models. The motivation to study context stems from the need to understand the relations among individuals, artifacts, and social groups [Nardi, p. 69].

In Chapter 1 (Activity Theory and Human-Computer Interaction), Nardi poses the obvious question of whether HCI needs to worry about consciousness. Her opinion is that HCI has been concerned with consciousness all along (as evident in HCI areas such as "intelligent agents", "direct manipulation", and "novice and expert behavior").

Nardi indicates in this chapter that "consciousness" is not an entity located in the brain. Rather it is located in everyday practice. Consciousness is recognized to be embedded in a social matrix. The basic tenant of Activity Theory is that the notion of consciousness is central to the depiction of activity.

What this argument suffers from is the possibility of being prematurely dismissed by the critical reader as a philosophical argument about the spiritual state of one's mind. The reader may feel it lacks obvious practical application for HCI research. Yet the subtle existence of terms and concepts based on one's consciousness is evident in HCI research areas and as further reading will indicate, conceptual frameworks of Activity Theory are applicable in various design situations. Activity Theory thus "extends the concept of consciousness past an idealistic, mentalistic construct in which only cognitive resources and attention "inside the head" are at issue, to a situated phenomenon in which one's material and social context are crucial [p. 13]."

Activity Theory places particular emphasis on how artifacts and social practices are integrated in a social matrix. The unit of analysis is centered around the development and function of a subject (that can be an individual or a group), the artifacts and their roles in the subject's environment and the social matrix in which the subject and artifacts interact. Conceptually artifacts and subjects are not ontologically equal. The artifacts exist as mediators of human thought.

Where cognitive science is concerned with information and its propagation, activity theory is concerned with practice (that is, doing and activity). Activity Theory views artifacts as extensions and transformers of the mind and body.

In Chapter 2 (Activity Theory as a Potential Framework for Human-Computer Interaction Research), the author, Kari Kuutti, presents the types of problems and pressures (as described by Liam Bannon), that are causing a movement away from mainstream information processing psychology. First of all, there is a need to view humans as active actors (a view that does not describe humans as a set of attributes of cognitive processors -- which is a common in mainstream cognitive psychology). There is also a need to move away from using predetermined fixed requirements for product design and a serious need to diminish reliance on laboratory experiments. The third problem indicated is the lack of recognition that actual use of systems is a long term process that cannot be adequately understood by studying just the initial steps of usage. The final critical problem mentioned is the need for more emphasis on design. Evaluating features in isolation with judgments on existing systems applicability in various situations leads to future problems that can snowball into larger problems.

Three solutions offered by Bannon are: 1. the need for active actors, 2. contextuality, 3. a constructivist relation between users and systems.

Kuutti, in this chapter offers a broad definition of Activity Theory and Activity Theory principles (which although dense in its description is tight in its description of an applicable framework), that offer potential solutions to the above mentioned problems. Activity theory is described as "a philosophical and cross-disciplinary framework for studying different forms of human practices and development processes, with both individual and social levels interlinked at the same time [p. 25]."

In Chapter 2 and 5, the reader will find an explicit set of Activity Theory principles. The abstract nature of the principles may cause some frustration in recognizing the applicability of the frameworks presented by activity theory. To that reader I suggest patience. The usefulness of activity theory principles and framework becomes more clear in part 2 of the book, where the practical applications are discussed.

According to Kuutti, the author of Chapter 2, the key principles of Activity Theory are: (1) activities are considered to be the basic unit of analysis. The solution offered by this theory is that "a minimal meaningful context for individual actions must be included in the basic unit of analysis, (2) activities are not static. Thus an activities dynamic nature and the consideration of the history of an activity is pertinent. (3) Artifacts have a mediating role. Relations between the elements of an activity are considered to be directed and not mediated. For example, "an instrument mediates between an actor and the object of doing". (4) Activities have a basic structure that can be described by the mediated relationship at the individual level. According to this structure activities have the following characteristics: (a) activities are a form of doing directed towards an object, (b) activities are distinguished from one another according to their object, (c) the motivation for the transformation from an object to an outcome comes from an activity. The object can be either a material thing or an intangible entity (such as a common idea). (d) mediation during an activity is carried out by a tool (which can be anything in the transformation process -- which includes both material tools and tools for thinking), which carries with it the history of the relationship of the subject and object.

The "relationship" between subject and object is mediated by tools, the relationship between subject and community is mediated by rules, and the relationship between object and community is mediated by the division of labor (which refers to the explicit and implicit organization of an community as related to the transformation process of the object into the outcome).

In Chapter 2, Kuutti provides a thorough description of the levels of an activity. There are two types of activity processes: a long term and a short term process. In a long term process outcomes are realized from objects through a process that generally consists of several phases. Short term processes are activities that consist of actions, which in turn consist of operations. An activity can be "realized using different actions, depending on the situation." And one particular action can "belong to different activities [p. 31]."

Kuutti also provides three perspectives which justify why Activity Theory contribute significantly to the field of HCI. The first perspective describes the benefits of the "multilevelness" of activity theory as an enabler for discussing the different levels involved in an activity within an integrated framework. The second perspective considers the "interaction context." The third perspective considers the issue of "development." This perspective considers change and the dynamic nature of an activity.

In Chapter 3, 4 and 5, the authors continue the stream of comparisons among the three theories which describe contexts (that is, activity theory, cognitive psychology and situated action theory). In Chapter 3 (Computer-Mediated Activity: Functional Organs in Social and Developmental Contexts), Victor Kaptelinin describes how the nature of any theory of HCI faces a challenge due to an "increasing units of analysis [p. 47]." The same solution is offered by cognitive science and activity theory to amend this problem: the researcher should consider human interaction with computers as a multilevel hierarchical structure. Yet the meaning of a hierarchical structure is different among these two approaches. In the cognitive science approach "the major theoretical task is to develop a conceptual scheme that can give a coordinated description of multilevel information processing in both human beings and computers [p. 47]." By contrast, the activity theory approach is such that "the hierarchical organization of human computer interaction is determined by its embeddedness into the hierarchical structure of human activity that mediates the user's interaction with reality [p. 47]."

In Chapter 3, the author Victor Kaptelinin, describes a type of three dimensional "explosion" in the expansion of HCI subject matter. This "explosion" framework is similar to the three perspectives provided by Kuutti in Chapter Two on the contributing role of activity theory for HCI. Although one of the dimensions, the "levels of interaction", is considered in the cognitive approach, the activity theory shifts focus from interactions between the user and the computer to a "larger context of interaction of human beings with their environment. The second dimension considers development. According to Kaptelinin, understanding the use of computers in a particular case requires an analysis of the computer's history and its potential developmental transformations. The third dimension considered is the individual and social dimension. The author states that the term "user" encompasses not only an individual but also groups and organizations [p. 48]. The HCI field needs to be recognized as an open system consisting of the meaningful context of the user's goals, environment, available tools, and interactions with other people.

According to Kaptelinin one of the salient features that distinguishes activity theory from cognitive approach is that activity theory "considers computers as a special kind of tool mediating human interaction with the world [p. 49]." Such a "tool mediation perspective" holds useful implications for HCI. These implications pertain to HCI in regard to: 1. questioning the very literal name of the discipline. There seems to be a need for more emphasis on stressing the tool nature of computers and a name that implies that, 2. questioning the existence of one interface when the tool mediation perspective implies the need to deal with two interfaces instead of one, with two borders, separating: (a) the user from the computer and, (b) the user and the computer from the outside world. What this latter implication poses is the need to understand where the boundary between the individual who uses a tool and the external world lies. HCI researchers need to answer the question on whether this "boundary" coincides with the individual-tool boundary or with the tool-world one.

Activity Theory attempts to answer the above questions posed by the tool-mediation perspective based on the "functional-organ" concept. There are many types of functional organs based on computer tools since computer tools do not possesses one particular function. One of the most important functions according to Kaptelinin is defined as "an extension of the internal plane of actions [p. 51]." IPA is a unique activity theory concept that refers to the human ability to perform manipulations with an internal representation of external objects before starting actions with these objects in reality. Activity theory considers the idea of functional organs to be of particular interest to HCI studies.

While the concept of IPA is similar to cognitive concepts of working memory and mental models [p 51], this concept differs from the cognitive approach in that it refers to the general ability to create and transform internal models. Although computers are not the only type of tools used as an extension of the IPA, the author explains how an understanding of "the mechanisms underlying the use of computer tools as extensions of the IPA is directly relevant to the development of useful and usable systems [p. 53]."

Another concept that the tool mediation perspective brings to the HCI field is the issue of culture. Tools not only transmit culture but in fact shape the goals of the people using the tools.

Although the concepts presented by authors in Chapter 2 and 5 remain fairly consistent, the principles in Chapter 5 (Activity Theory: Implications for Human-Computer Interaction) are a summary of most of the principles covered in Chapter 2 and Activity Theory concepts mentioned throughout Part 1 of the book.

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