---

Chapter 9: Postcognitivist Theories in Interaction Design

Contents

Postcognitivist Theories in Interaction Design
Common Ground
Sorting Out the Theories
Three Fissures: Creativity, Reflexivity, and Resistance
Creativity
Reflexivity
Resistance
Conclusion

 

---

 

Postcognitivist Theories in Interaction Design

In chapter 2 we argued that theory is important for interaction design. We observed that the development of multiple theories provides opportunities for multivoiced conversations, each theory contributing a unique set of perspectives and concepts. This chapter compares postcognitivist theories in interaction design—activity theory, distributed cognition, actor–network theory, and phenomenology. We call the theories “postcognitivist” because they have been brought into interaction design to remedy perceived shortcomings of cognitivist theory.

Distributed cognition is a well–known approach in HCI (Hollan, Hutchins, and Kirsch 2000) and CSCW (Rogers and Ellis 1994) that has highlighted the importance of tools in cognition. Actor–network theory is beginning to have significant impact (Berg 1997; Hanseth and Lundberg 2001; Holmström and Stalder 2001; Kling, McKim, and King 2003; Mahring, et al. 2004; Lamb and Davidson 2005), having made vivid the power of technologies, their "push–back" in activity. Phenomenology, in insisting that everyday experience be a focus of inquiry, inspired approaches that have shaped interaction design (Winograd and Flores 1986; Suchman 1987; Dourish 2001a). Distributed cognition, actor–network theory, and phenomenology share significant common ground with activity theory. They also diverge in critical ways. These commonalities and differences between the theories will be examined. We discuss actor–network theory as developed by Callon (1986) and Latour (1992), and distributed cognition as developed by Hutchins (1995), Clark (1997), and Ackerman and Halverson (1998). Dourish’s (2001a) account of phenomenology is our primary resource for that tradition. Dourish’s discussion is useful for our purposes because it focuses on interaction design rather than phenomenology in its entirety, which is beyond the scope of this discussion.

 

Common Ground

A major point of agreement among the postcognitivist theories is the vital role of technology in human life. Each theory incorporates technology in its own way. In activity theory, a key principle is tool mediation. Distributed cognition views cognition as distributed across people and their tools (Pea 1993; Salomon 1993; Hutchins 1995). Actor–network theory specifies the agency of technology — the way things (such as machines) are agents in their own right, interacting with humans in actor–networks (Callon 1986; Latour 1994). Phenomenology (most particularly Heidegger 1962) suggests that we understand thinking as derived from being, that is, “being–in–the–world,” including the tools in the world.

In activity theory, consciousness is seen as a result of practical activity with other people and with tools, so mind necessarily goes beyond the individual.

Another commonality among the theories is the claim that individuals are not defined by the boundary of their skin. In activity theory, consciousness is seen as a result of practical activity with other people and with tools, so mind necessarily goes beyond the individual. And the idea of functional organs asserts a tight binding between human and tool (Ukhtomsky 1978, cited in Zinchenko 1996; Leontiev 1981). Heidegger’s (1962) concept of the use of tools as “equipment” is similar to activity theory’s framing of tools as mediating object–oriented activity. Equipment is a tool for some task, for a purposeful activity. Thought is bound up with the use of equipment (see Dourish 2001a). Distributed cognition defines cognition as distributed between people and tools where both are equivalent “media” in a system (Hutchins 1995). Actor–network theory places people and tools in a larger network and claims that a boundary cannot be drawn at the individual.

Postcognitivist theories are highly critical of mind–body dualism. Vygotsky’s ideas of internalization and interpsychological functions explicitly sought to provide an account of the dynamic relationship between internal cognition and the external world of artifacts and people. Leontiev’s criticisms of the “postulate of immediacy,” in which a simple duality between stimulus and response was said to describe behavior, pointed to the need for more complex concepts: “the effects of external influences are determined not immediately by the influences themselves, but depend on their refraction by the subject” (Leontiev 1977).

Phenomenology draws attention to the unity of mind and the world through humans’ necessary relationship with tools. The very notion of distributed cognition (Salomon 1993; Hutchins 1995; Salomon and Perkins 1998) asserts that cognition cannot be captured with a concept of an isolated mind. Actor–network theory places all activity within the network, where nodes of any type can transact with nodes of any other type.

All of these concepts are examples of concrete analytical tools developed to move beyond dualism and to explain the fundamental unity of the mind and the world. Moving away from the cut–off Cartesian individual, postcognitivist theories came to focus on the physical and social distribution of phenomena, such as agency and cognition, that were traditionally considered to belong to the separate reality of the individual mind.

Postcognitivist theories provide an important alternative not only to cognitive science, but to authoritative theories in biology, neuroscience, and key areas of philosophy and psychology in which technology is nearly invisible. Such theories aim to describe universal human capabilities and experiences. Technology, a culturally variable phenomenon, is not foregrounded in these theories, if it is mentioned at all.

Without reviewing theory in biology and the other disciplines, we can give a few key examples regarding the invisibility of technology. In neuroscience, the brain, brain processes, and genes are seen as the locus of cognition (Edelman 1992; Crick 1993; Searle 1995). Crick (1993), for example, declared that emotions, memories, ambitions, identity, and free will are “no more than the behavior of a vast assembly of nerve cells and their associated molecules.” Sociobiologists (e.g., Wilson 1975) depict genes as determinants of social activity, selected in early humankind’s struggle for survival. The psychological theories of Freud, Jung, Maslow, Pavlov, and Skinner all posit panhuman processes. For example, Freud asserted universal processes such as repression, denial, and reaction formation, while Maslow suggested a universal hierarchy of needs.

Such universal formulations simply bypass the highly variable designed world and the diversity of practical activity taking place within that world. But the design and deployment of technology is a quantum leap in human evolution, involving the mobilization of vast cultural resources that do not reduce to universal biological or psychological processes. Precedents for the use of simple tools in nonhuman primates such as chimpanzees suggest how tools may have been introduced in early human communities (Köhler 1925; Goodall 1964). It appears that technology use preceded language in human evolution (Donald 1991). With increased sociality and the ability to talk about their tools as language developed, early humans made rapid advances in creating unique artifacts of immense value to human life (see Wells 2000).

The primacy of tool use in the human experience requires theorizing.

The primacy of tool use in the human experience requires theorizing. Postcognitivist theories address this problem. The texture of the theories is not that of the sweeping reductionist statements about genes, neurons, dreams, neuroses, or goal hierarchies of the universalistic theories, but rather detailed depictions of the complexities of human activities. Even the most mundane activities turn out to require elaborate description (and still, it often seems there’s more to say). Kirsh (1995), for example, described the surprisingly complex activity of cooks making salads. Scribner and Sachs (1990) studied training in a stockroom at a manufacturing plant. Bellamy (1996) detailed the function of technology in a middle school classroom. If postcognitivist theories have taught us anything, it is that the closer we look, the more complex ordinary activities appear, and the less amenable they are to reductionist accounts.

The project of the postcognitivist theories, then, is to provide nuanced analyses of the microgenesis of organized activity. The theories share the practice of delivering fine–grained historical accounts of activity in rich settings, because, as Gherardi and Nicolini (2000) observed, “human action can only be explained in terms of the specific conditions in which it takes place ... .” Postcognitivist theories help us understand technology as central to human experience. They share common ground in depicting people as beings enmeshed in our own technological creations. These theories set the stage for explorations of invented worlds in which designed technical artifacts imbue human life with its distinctive character.

 

Sorting Out the Theories

Despite significant common ground, there are important differences among postcognitivist theories. We analyze these differences with respect to two related issues: (1) intentionality, and (2) fissures in routine activity, namely, creativity, reflexivity, and resistance. These issues engage the question of whether or not to theorize individual subjects. Activity theory is committed to developing a concept of the individual as a technologically empowered and socially contextualized subject. This commitment is questioned or jettisoned in distributed cognition and actor–network theory. Phenomenology stands with activity theory on this issue, but because phenomenology is not intended to be a psychological theory, its principles do not go as far in describing the nature of the subject as do activity theory’s.

We will argue that while analysis at the level of system, network, organization, institution, culture, or other supraindividual entity is fruitful for many purposes, there are important issues of interaction design and technology use that require a notion of the individual subject. Drawing from diverse theoretical and empirical work, we argue that supraindividual entities cannot mirror, represent, stand in for, or subsume the particular ways in which individual subjects may act. We examine intentionality, a property of individual subjects that systems–level analyses do not deal with, and disruptive activities of creativity, reflexivity, and resistance that involve individual subjects as they respond to social conditions. We highlight these issues of the individual subject because some of the postcognitivist theories downplay the individual, a being whose activities we wish to keep visible in interaction design. We also discuss collaborative activity, considering a scheme developed by Raeithel (1996) that provides some ways of breaking collaboration down into a less monolithic concept.

Characterizing People: Intentional Beings or Nodes in a Network?

As we saw in chapter 3, activity theory begins with the idea of a purposeful subject. Only living things have needs. These needs can be met by acting in the world, by bringing together the subject’s need and an object. When a need meets its object, the object becomes a motive and directs the subject’s activity. For humans, needs are, in significant measure, culturally shaped. The most fundamental notion of activity theory is the motivated activity of a subject enacted in culturally meaningful ways.

Vygotsky pointed to the purposefulness of activity as the defining characteristic of the higher mental functions of mature adults. In Thought and Language, he observed, “A small child draws first, then decides what it is he has drawn; at a slightly older age, he names his drawing when it is half–done; and finally he decides beforehand what he will draw” (Vygotsky 1986).

Activity theory posits a “specifically human type of orientation and consciousness” that emerged as humans evolved (Miettinen 1999). Informed by the work of Luria, Leontiev, and Vygotsky, Miettinen observed that there was

a gradual breaking of the direct, immediate, impulse–based relation to the objects of the environment. With cultural development — characterized by communication and the construction and use of tools — a specifically human type of consciousness emerged. [Such a consciousness] also implies the capability of imagining and planning what the future may hold; that is, intentionality. (Miettinen 1999)

Miettinen (1999) also noted that only humans can “tak[e] the initiative in the construction of [new assemblies of humans and materials].” We take Miettinen’s concepts of intentionality as our resource for this discussion. (For more on intentionality with an interaction design slant, see Button, Coulter, and Sharrock 1995; Dourish 2001a.)

Phenomenology also posits an intentional subject. Dourish (2001a) traced the changing concept of intentionality in phenomenology, culminating in Heidegger’s subject who encounters the world through purposeful practical tasks. Dourish observed that for a subject “the orientation is fundamentally a practical one; it is purposeful and active.” In both activity theory and phenomenology, an active human subject engages in meaningful activity mediated by tools. Asymmetry between subject and tool gives rise to concepts such as tools that are ready–to–hand or present–at–hand (Heidegger 1962), functional organs (Zinchenko 1996), tool mediation (Vygotsky 1986; Wertsch 1998), and breakdowns and focus shifts (Bødker 1996; Koschmann, Kuutti, and Hickmann 1998).

Phenomenology and activity theory have strong links in seeing tools as mediators of human experience. Phenomenology, especially Heidegger’s work, identifies tools as an important part of the way we encounter the world to accomplish our goals (Dourish 2001a). Heidegger noted that a tool is tied to a meaningful task; it is “something in–order–to” (Heidegger 1962). Heidegger distinguished tools that are ready-to-hand and those that are present–at–hand. When a tool is ready–to–hand, we have no conscious awareness of it; we simply use it to accomplish a task. This corresponds to the operational level of the activity hierarchy. When there is a problem or a breakdown, we become consciously aware of the tool and must attend to it; then the tool is present–at–hand. This sequence is the same as moving from the operational to the action level in activity theory.

Actor–network theory and distributed cognition begin from an entirely different standpoint. Where activity theory and phenomenology develop an asymmetrical notion of tools mediating human experience, actor–network theory and distributed cognition define a network or system in which symmetrical nodes can be human or nonhuman, and are treated alike.

In actor–network theory, nodes in “heterogeneous” networks are “actors” that can be people, machines, or other things.

In actor–network theory, nodes in “heterogeneous” networks are “actors” that can be people, machines, or other things. As Callon (1986) said, “The rule which we must respect is not to change registers when we move from the technical to the social aspects of the problem studied.” Latour (1992) stated that this principle of generalized symmetry was the “most important philosophical discovery” in actor–network theory.

Actor–network theory describes networks as heterogeneous (because they contain both people and things) while at same time proposing a symmetry between nodes. Caspar (1994) suggested that there is a muddle in this formulation. Along which dimensions does the heterogeneity lie? In Caspar’s words, “to argue that nonhumans should be analyzed symmetrically does not get us any closer to determining how entities are configured as human and nonhuman prior to our analyses.” (See also Collins and Yearly 1992.)

Actor–network theory’s principle of generalized symmetry delimits a network in which no entity has purpose or intentionality in Miettinen’s sense. Technologies do not possess needs, motives, or intentions. So, in actor–network theory, we have a theory that cannot say how a system develops through cultural processes of “imagining and planning what the future may hold.”

In the midst of actor–network theory’s textureless networks of generalized nodes, we unexpectedly encounter colorful princes — Machiavellian princes!

Or can it? In the midst of actor–network theory’s textureless networks of generalized nodes, we unexpectedly encounter colorful princes — Machiavellian princes! It seems that actor–network theory cannot, after all, do without agents with intentions. The princes are innovators, managers, and politicians (Callon, Law, and Rip 1986). It is they who assemble the networks, enroll other actors, drive events, and push their own agendas. Specific technologies affect the course of events, but they do not instigate the networks, the projects enacted in the networks, or the technologies whose agency is felt in the networks. Miettinen (1999) observed that despite generalized symmetry as the core principle of actor–network theory: “the principle of Machiavellianism seems to dominate. These principles [generalized symmetry and Machiavellianism] contradict each other.”

Distributed cognition also depicts a system of like nodes where human and nonhuman are of the same type. Distributed cognition constructs a system in which tasks are performed as representational state passes from one node or “medium” to another in a system (Hutchins 1995). As in traditional cognitive science, cognition is said to be the same as computation. Computation is defined as “the propagation of representational state across representational media” (Hutchins 1995). People, tools, systems, and so on are all “media.” Rather radically, Hutchins (1995) said he hoped to develop a theory that would require no change to “cross the skin.” Hollan and Hutchins (2003) observed that “Distributed cognition theory is still willing to lump together cases that activity theory long ago split.”

In distributed cognition, there is some slippage with respect to the strict definition of media because of a desire to provide an account of human cognition. While Hutchins (1995) clearly stated that distributed cognition constructs a system in which the analyst is to trace state across a variety of media without prejudice as to the type of media, it is also true that distributed cognition has a serious interest in human cognition. As Hollan and Hutchins (2003) said, “distributed cognition emphasizes cognitive aspects of human performance.” Thus its theoretical underpinnings may sometimes appear to change key, especially in lengthy accounts such as that of Hutchins (1995). When the discussion turns to tool mediation and human performance, distributed cognition is more in tune with activity theory (Hollan and Hutchins 2003). When a cognitive system of like nodes is proposed, distributed cognition has more in common with actor–network theory.

Is there a difference between “distributed cognition” and activity theory’s concept of tool mediation? Both suggest that cognition requires tools and that tools do important work that humans cannot do on their own. Both extend the analysis of cognition to insist on going beyond the individual mind.

One difference is that in activity theory, people mediate their relationship to reality with tools. The tool is there for purposeful, intentional activity. (Using different terms, phenomenologists have said the same.) It is reasonable in activity theory to speak of cognition as being distributed in that tools are used for human cognition. So, for example, a doctor decides whether a patient has a particular ailment partly on the basis of a reading taken from a thermometer. The cognitive work is distributed because the doctor uses the thermometer for a precise reading. However, the thermometer mediates the doctor’s object–oriented activity; the taking of the temperature is part of an activity driven by a human object: it is the doctor’s intention to find out what is wrong with the patient.

In distributed cognition, because of the flat system in which “media” are uniform nodes, the idea of tool mediation, which requires asymmetry, is inconsistent. In a flat system, cognition is a concept that can be applied equally well to the system or to any of its nodes. If one node can exhibit cognition, then by definition, so can any other node. Tools are no longer mediators of human experience. It is here that the plausible idea of recognizing the importance of tools in cognition becomes skewed in such a way that the basic definition of cognition is changed to apply to nonhumans (such as a system). The notion of distribution, instead of acting as a helpful mnemonic to incorporate tools into the analysis, now wipes out the very concept of human cognition and replaces it with the idea that anything can potentially cognize. This is of course not inconsistent with traditional cognitive science which equates cognition with intelligence with computation (Simon and Kaplan 1989), but distributed cognition was supposed to be a challenge to traditional cognitive science (Hutchins 1995). Without a notion of asymmetry, the cognizant system of distributed cognition is equivalent to an intelligent machine.

However, even within distributed cognition as proposed by Hutchins, there are suggestions that systems and people are not really the same. Hutchins (1995) spoke of the “cognitive properties” of systems. He admitted that “the larger system has cognitive properties very different from those of any individual.” So at least there was an awareness of differences between what a system does and what a person does. However, questions remain: why is the system said to have cognitive properties instead of functional properties or culture? Human cognition, long recognized as involving attention, awareness, and judgment, is now, somehow, related to a different kind of cognition that has no capacity for attention, awareness, and judgment (see Nardi 1998). We believe it is more coherent to discuss systems as having a culture or a set of functional properties, rather than cognition. This avoids the problem of deleting attention, awareness, and judgment from theories of cognition or setting up parallel systems of cognition related to one another in unspecified ways. Rose, Jones, and Truex (2005) observed that humans have “self–awareness, social awareness, interpretation, intentionality, and the attribution of agency to others” — which are not available to nonliving things.

In activity theory, the idea of the distribution of cognition goes back to Vygotsky’s universal law of psychological development: new psychological functions are first distributed between the individual and other people or tools (as discussed in chapter 3). One of Vygotsky’s greatest discoveries was that subjects use mediational tools to solve problems, but over time, they stop using them and yet maintain high levels of performance. Vygotsky (1983) attributed this finding to the process of internalization. Individuals appropriate new functions which become intrapsychological.

Distribution, then, is an aspect of development — not a separate kind of cognition as it may sometimes appear when people speak of distributed cognition. Of course, individuals may continue to make use of other people and tools. That does not mean that developmental transformations are not occurring or that a different kind of distributed cognition takes place in parallel with individual cognition. The doctor taking a patient’s temperature is doing much more than simply getting a reading from an instrument (i.e., receiving state from the thermometer medium).

She is observing the relationship of the temperature to the symptoms of various diseases she sees in different patients. She is noticing a particular temperature and the way the patient with that temperature looks and acts. She is remembering what she learned in medical school about the diseases that exhibit raised temperatures. The doctor uses the thermometer as a tool mediating her understanding of the patient’s illness, an understanding that changes over time as she gains experience in her practice. After considerable practice, she may have appropriated the function of the thermometer pretty well herself and can judge a temperature without actually taking it. She may continue to use the thermometer for social reasons — the patient thinks it is proper medical practice, the insurance company requires it, her attorneys advise her to do so. The doctor transforms the role of the tool in her work, responding to social and cultural realities, rather than passively receiving state. It is important to examine developmental transformations, rather than assume that reliance on tools means we turn all functions over to them and cease to develop and transform practice. Internalization is a developmental redistribution of internal and external components. Because the internal plane is not a carbon copy of the external, its form introduces the potential for change.

The individual subject and the world   Postcognitivist theories differ in how far they go in pushing the boundaries between the individual and the world. Activity theory and phenomenology retain a commitment to the individual subject. For them, contextual analysis is a way to reach a deeper understanding of individual human beings. Distributed cognition and actor–network theory define their focus of analysis as larger–scale agentic entities that include individuals as just one type of component. Within these approaches, individuals are deprived of their unique (or even privileged) status and are seen as components of a system or network as a whole. Hollan and Hutchins (2003) observed, “From a distributed cognition perspective, goals may be properties of institutions, but need not necessarily be properties of individuals.” In distributed cognition analyses, there is usually a large unchanging systemic goal such as piloting a boat safely into harbor (as in Hutchins 1995). The institutional goals that are part of a larger culture are acknowledged, but not the goals of individual subjects.

Wartofsky (1979) provided a vivid image of the nature of human intention. Observing the ubiquity of handprints in paleolithic art, he asked what these handprints could possibly mean:

The simplest answer seems to be that such prints are the deliberate marks of presence: ‘I was here.’ ... The handprint entails a hand pressed against a wall, with the express intention of leaving its imprint there. ... [T]he handprint records a gesture, an action, an intention ... . The marks by means of which a species represents itself to itself are thus distinctive. (Wartofsky 1979)

Wartofsky compared the handprints to Robinson Crusoe’s footprints. Crusoe’s footprints in the sand on the beach were indeed useful for Friday as a sign of Crusoe’s presence, but they were not an intentional gesture.

Without a concept of intention, of purposefulness, the handprint and footprint may be confused. Hutchins compared the development of culture to the accumulation of chemical trails left by ants. The trails enable the ant that follows its predecessors to find food more readily. “Is this a smart ant?” asked Hutchins.

No, it is just the same dumb sort of ant, reacting to its environment in the same ways its ancestors did. But the environment is not the same ... . Generations of ants have left their marks on the beach, and now a dumb ant has been made to appear smart through its simple interaction with the residua of the history of its ancestors’ actions. (Hutchins 1995)

Leontiev (1974) argued that activity does not consist of reactions to an environment: “Activity is thus not a reaction or a totality of reactions, but rather a system possessing structure, inner transformations, conversions, and development.” Human culture results from intentional human activity — someone raises a hand to make a mark. Crusoe’s footprints on the beach were not culture any more than the ants’ trails on the beach are culture. Human activity constantly changes; we do not “react” the way our ancestors did. As Luria (1972) observed, “Cognitive processes are not independent and unchanging ‘abilities’ ... ; they are processes occurring in concrete, practical activity ... .” These processes grow directly from the specificity of particular kinds of practical activity, variable across cultures. Sandberg and Wielinga (1993) noted that one problem of distributed cognition theory is that a human is taken to be “a simple organism interacting with its environment and producing complex behavior through the application of simple behavioral rules.” The ant analogy bears the marks of this misconception.

Stetsenko and Arievitch (1997) observed of socio–constructivist theories that even though they all provide a “relational, contextualized, account of the evolving self,” there are two distinct approaches. The first dissolves the self (understood as a human agent) in linguistic or social reality, while the other strives to provide a nonreductionist account. This distinction is relevant to our discussion of postcognitivist theories. Actor–network theory and distributed cognition seem to dissolve the individual by considering the individual a part of a system or network and by equating humans with nonhumans. Activity theory and phenomenology aim to construct the individual as a technologically empowered and socially contextualized subject.

Even though human motivation is profoundly influenced by culture and society, each individual has her own hierarchy of motives.

Our position in this debate is unambiguously pro–individual. We believe that analysis of human embeddedness in the world, of communication and collaboration within the social context, is of fundamental importance for understanding our relationship to technology. Human beings cannot be reduced to parts of larger–scale entities determining the nature and the meaning of the parts. It is true that individuals taking part in collective activities often have to abide by the rules, roles, and routines typical of collective work. However, it is also true that individuals do not always follow the rules, roles, and routines, and sometimes they even change them. As Middleton and Brown (2002) observed, agency is “the sense that someone or something is not following some pre–established programme of action, is not simply expressing some pre–existing structure.” Even though human motivation is profoundly influenced by culture and society, each individual has her own hierarchy of motives. Although both are parts of a system, individuals are different in kind from artifacts, because (a) they have their own needs and reasons to do things that go far beyond the specific activity they are involved in, and (b) they reflect on and make sense of the collective activity and their own actions.

Even within the context of a system/network/organization, the individual is acknowledged as different in kind from the organization’s nonliving things. Dealing with individuals and their needs and capacities becomes part of the activity in an organization. So, for example, when an organization recruits workers, a manager not only describes future work assignments but tries to make the position attractive by stressing its advantages for an individual worker, such as a creative work environment, good salary, comprehensive insurance, desirable location, and so forth (see Foot 2006). This is necessary because the individual has his own reasons for choosing to participate or not participate in the organization, his own motives and intentions. The organization’s artifacts, on the other hand, have no intentions, no motivation. No one in the organization need entice the artifacts to participate by appealing to their personal motives. The humans, however, must be dealt with as motivated individuals as long as they are employed — providing raises or bonuses to good performers, offering incentives for continuing education, encouraging workers to grow and develop.

Characteristic of Leontiev’s view of the individual was his opposition to understanding the individual as merely a component part of social systems and processes. When discussing the idea that an individual could be completely described through a set of roles defined by the social context, Leontiev (1978) called the notion “monstrous.”

In the next section, we look at the potential impact of individual subjects, with their motives and intentions, on activities central to technology and culture change.

 

Three Fissures: Creativity, Reflexivity, and Resistance

As designers, we are aware that technology is always changing; indeed it is the job of many of us to create such change. But how does culture change come about? Without attempting to answer such a large question, we can point to three “fissures” or disruptions that are part of the culture change of technology: creativity, reflexivity, and resistance. Creativity, reflexivity, and resistance have been identified as crucial to understanding technology by a broad range of scholars including Mumford (1934), Ellul (1964), Illich (1980), Winner (1986), and Postman (1993).

By creativity we mean imaginative activity directed toward an object in which an original product emerges. By reflexivity, we mean reflection that leads to a change in practice, or potentially can lead to such change when obstacles do not prevent it. By resistance we mean opposition to a technology, or to a practice associated with a technology, that is evaluated as detrimental. In this section, we discuss creativity at length and sketch the importance of reflexivity and resistance in technologically mediated activity.

Zinchenko (1996) argued that human activities go beyond culture; indeed, they must do so, or culture would never change:

[People] do not behave passively, and in due course they even become themselves the source and drive that is capable of developing culture and civilization, generating new ... forms and overcoming old ones. ... [W]e should recognize the relations between ... human and culture to be mutually active, communicative, and dialogical. The dialog may be friendly or hostile; it may become aggressive. (Zinchenko 1996)

Such culture change manifests itself through activities of creativity, reflexivity, and resistance. To account for the very invention of technology — which undeniably involves acts of human creativity — we need a theory in which creativity follows logically from basic principles. The theory must also account for the varied responses to the deployment of a technology: a technology may be embraced, rejected, or altered by its users to better meet their needs as they reflect on its use (Mackay 1990; Orlikowski and Tyre 1994; Foot 2001; Béguin 2003; Geisler 2003; Spinuzzi 2003; Gay and Hembrooke 2004; de Souza 2005; Stahl 2006). Finally, the theory must acknowledge that people may resist technology; having reflected on its uses and consequences, the dialogue may even become hostile, as Zinchenko put it.

The Machivellian princes of actor–network theory provide ample scope for activities of culture change. Who better to scrutinize, plot, and plan than such resourceful fellows? We shall leave them to their intrigues, assuming them fully capable of any necessary acts of creativity, reflexivity, or resistance.

Phenomenology posits an active subject much like that in activity theory, though as we will briefly discuss there are some differences. We use this discussion of fissures primarily to challenge some aspects of distributed cognition. (In the next chapter we do the same for certain aspects of actor–network theory.) We draw on empirical work in the activity theory tradition from Engeström (1990), Miettinen (1999), DeCortis, Rizzo, and Saudelli (2003), Spinuzzi (2003), Gay and Hembrooke (2004), and others.

 

Creativity

Creativity manifests itself in insights. Without a subject to have such insights, there is no creativity. The individual subject is crucial because creativity may require an instant restructuring of the whole representation of a problem (Wertheimer 1961; Csikszentmihaly 1996), a “global spontaneous reorganization of the whole body of knowledge about the problem situation” (Kaptelinin 1996b). This reorganization happens internally, not in resources distributed externally to the individual. Thus an internal process is necessary to creativity.

Even everyday activity of the most mundane sort often involves unexpected bursts of creativity in which a subject can reframe a representation to successfully solve a problem. In the well–known cottage cheese incident (Lave, Murtaugh, and de la Rocha 1984), a dieter following the Weight Watchers program was at first stumped when he had to measure three–quarters of the two–thirds cup of cottage cheese allowed by the program:

The problem solver in this example began the task muttering that he had had calculus in college, and then, after a long pause, suddenly announced that he had “got it.” From then on he appeared certain he was correct, even before carrying out the procedure. He filled a measuring cup two thirds full of cottage cheese, dumped it out on a cutting board, patted it into a circle, marked a cross on it, scooped away one quadrant, and served the rest. (Lave, Murtaugh, and de la Rocha 1984)

The dieter suddenly saw the measurement problem in a new way. He had not learned the technique he used to solve the problem at Weight Watchers, nor had he learned it in school. He used no external artifact to create the solution to the problem. While Lave et al. saw the dieter’s measuring of the cottage cheese as emerging from the “situation,” we note that the dieter experienced a sudden insight — he “got it” — followed by flawless enactment of the measurement. The dieter himself brought something to the situation — the subjective refraction of which Leontiev spoke. He used the resources at hand, but he was able to perform the measurement without trial and error, without interacting with the environment. He seemed to simply reassess the problem and then act, in a single smooth move. This calls to mind Köhler’s chimp who “realized” it could get the bananas with a stick — no trial and error needed!

The dieter’s solution was not obvious, nor was it a foregone conclusion given a “situation” of cottage cheese and a measuring cup. We can imagine other dieters confronted with exactly the same problem who would not come up with a solution — a less motivated dieter who might simply consume the extra cottage cheese, an elderly dieter having forgotten about fractions, a dieter with less confidence to find a solution than someone who “had had calculus in college.”

The measuring cup, the cutting board, and the dieter would be considered symmetrical nodes in distributed cognition and actor–network theory, even though there is no sense in which the cooking utensils could have committed the act of creativity the dieter did. The limits of symmetry are met rapidly with an even simpler example like this one. Spinuzzi (2003) reminds us of the “inventive, wily, sly, cunning, and crafty” — descriptors of the creative human caught in the act of object–oriented activity. In considering the internal process of creativity, classical activity theory, which resolves to the individual, provides a firm footing. Other variants of activity theory that begin with a collective subject are less adapted for explaining creativity (Kaptelinin 1996b). The individual subject, with his or her unique potential to instantly restructure a representation in the context of meaningful activity, is essential for a theory of creative activity.

But perhaps the cottage cheese example is too simple? What about the fact that under many circumstances, an individual cannot accomplish as much as a group? Is there some kind of “group cognition” or “group creativity” that happens when a group works together? Is “collective creativity” qualitatively different from individual creativity? Some insight might be gained here by considering Vygotsky’s concept of the zone of proximal development. While this concept was developed to explain differences in learning, it might also apply to the creative process in groups.

In the zone of proximal development, the subject learns according to the specific mediation provided by a teacher or more experienced peer. The difference between what the learner can do alone and what she can do with teacher or peer is the “zone.” This zone varies according to both the individual abilities the learner brings to the learning task and the kind of mediation provided by the teacher or more experienced peer.

This zone will be different for Learner A and Learner B even when each is provided with the same mediation because each learner begins with a different set of understandings and abilities. We cannot explain differences between different pairs of learners and teachers without an appeal to learners’ differences. For example, although one of us spent many hours observing neurosurgeries for a study of video monitoring in a teaching hospital (Nardi, et al. 1993), in no way can she claim the faintest expertise in removing brain tumors. The medical residents in the operating room, on the other hand, also observed the surgeries at the same time, but they picked up a good deal more surgical expertise than a social scientist would. What was effective mediation for medical students learning neurosurgery could not have the same impact for a nonspecialist.

The same relation between individual learner and group found in the zone of proximal development may hold for creativity. Imagine a group has convened to design new educational technology. The group consists of Lev Vygotsky, Aleksey Leontiev, Maria Montessori, and Don Norman. The designs created by the group would be different from those of a group of, let us say, first–year graduate students. And why? For two reasons: first, the designs would depend on the specific individuals present. As Leontiev observed, “the effects of external influences are determined not immediately by the influences themselves, but depend on their refraction by the subject” (Leontiev 1977). Second, the designs would be different because of the mediation provided by this group’s members. Through the mediation of group conversation (and perhaps artifacts such as diagrams), members would come up with new resources for reframing problems and developing creative designs (see Fischer 2004 on the importance of conversation for “social creativity”), and these resources would vary depending on the group. We can well imagine the fruitful mediations each of these group members could provide through his or her conversation.

What we wish to point out with this thought experiment is that activity theory’s concept of mediation, combined with understanding creativity as the internal restructuring of a problem representation, helps us conceptualize creativity in groups. In a group setting, the mediation of conversation from other insightful people may help individual group members to frame problems in new ways and then contribute those new insights to the group. Creative insights take place in concrete activity in which specific individual subjects converse, communicate, and respond to one another (Stahl 2006).

The designs for new technologies from our fantasy group would likely be more than merely the sum of the ideas contributed by each member if he or she were working individually. But the designs cannot be explained simply by reference to the fact that they emerged in a group. The specific individual members of the group count for a good deal in terms of both their own creative abilities to restructure problems and the mediation they potentially provide to other group members. Indeed, we look to people such as those in our imaginary design group for inspiration not because of the groups they represent, but because of their fresh insights, new ways of framing problems, and surprising perspectives. We probably do not have to convince anyone in the interaction design community that an hour with Don Norman is an experience unlike any other!

We should not forget that many artists (and other creative people) work mainly independently. Their material comes from the social and physical worlds around them. They may be part of intense social worlds that shape their creativity, but they simply do not, for example, write poems or paint pictures in a group. (Countless biographies clarify the circumstances under which artists produce their art. Vincent van Gogh, for example, was more attuned to the physical than social world because of his mental illness. Sylvia Plath moved in rarified intellectual circles, but wrote her best poems in solitude at 4:00 in the morning before her family awoke.)

While it is important to foster groups that enhance creativity, and this is increasingly important in science and industry, we believe that a theory that begins analysis at the group level leaves out much of importance, namely, the fissures an individual subject is capable of causing, fissures that may lead to new ways of framing problems (or to less useful disruptions, which must also be considered). From a practical standpoint, it is important to understand the kinds of mediation that can be provided by specific group members (and the need to convene groups with particular kinds of individuals), the training that may be needed for individuals to work together well in groups, and the technologies and practices (such as facilitation) that will underwrite the mediations that enable individual group members to contribute new insights. The particular expertises chosen for the group must be the right ones to produce optimal results. One person’s expertise cannot be traded for another, at random. In everyday practical activity, we act accordingly, choosing this doctor and not that one, Colleague A and not Colleague B. Skilled managers observe that their job consists primarily in finding the right people and providing the resources for them to get a job done.

This insistence on positioning an individual intentional subject as a creative agent may seem to leave us open to charges of cognitivism or mentalism. But such charges would only arise from a conflation of social activity with group activity. As we hope is clear by now, activity theory constructs all activity as social, inevitably involving other people, artifacts, and culture, in various ways. Sometimes a subject acts independently (though never in social or cultural isolation) and sometimes with others. The individual subject may have agentic power, under some circumstances, to set in motion, or encourage, social processes leading to change. For example, Vygotsky was able to begin a process of framing a new theory of human consciousness that is a distinctive way of looking at the human mind. Vygotsky’s work was informed by significant historical influences, but the work is unique and powerful in its own right. Precisely what Vygotsky was able to do was to take concepts from Marx, Engels, Piaget and others, and restructure them in ways that led to new theoretical understandings.

Creativity is by definition social creativity (just as all activity is social activity). There is no lone genius.

Creativity is by definition social creativity (just as all activity is social activity). There is no lone genius. In acknowledging Vygotsky’s great contributions, we in no way mean to place him outside of, or above, his particular historical context. Vygotsky delivered his seminal lectures to audiences capable of appreciating them and no doubt capable of helping him work through difficult concepts. He had access to theoretical writings essential for his own theoretical development. As Seitz (2003) observed, “[A]ny creative product emerges from a unique coincidence of individual intellective abilities; the nature and relative sophistication of a scientific, artistic or entrepreneurial domain; the complexity and structure of the field of legitimization; and the distribution of power and resources within a group, community, or society.” Seitz acknowledged “individual intellective abilities,” situated in complex social settings involving the resources and relations of distinctive domains.

Like most members of social groups, we feel it is fitting to recognize individual contributions. We believe this penchant for recognition is social in origin. We name contributions after people — Newton’s laws, Rubik’s cube, Engeström’s triangle. Paradoxically, the recognition of individual creativity seems a way to make the creative product a part of culture and society, to absorb it into the larger milieu by personifying it. We might just as well speak of “developmental work research triangles.” Or “Rubik family cubes,” as Rubik gave credit for his inventiveness to the combination of influences from his father who was an engineer and his mother who was an artist. We personify creativity for many reasons, one of which may be the uptake of new ideas in our rapidly changing culture through the drama of cultural characters.

Phenomenology and Creativity

Dourish (2001a) observed that “Embodied interaction is the creation, manipulation, and sharing of meaning through engaged interaction with artifacts.” While this definition is broad, it points particularly to activity that is “grounded in everyday, mundane experience” (Dourish 2001a). In this emphasis on everyday, mundane experience, phenomenology and activity theory differ somewhat. Activity theory does not necessarily exclude activity that is beyond the scope of everyday, mundane experience.

It seems quite easy to come up with examples of human creations not rooted in a particular everyday, mundane experience. In the design of computer artifacts, for example, there are formal structures such as database tuples or programming languages that do not have clear antecedents in embodied experience. Walt Disney is said to have based designs for Disneyland in part on conversations with his father about his work on the Chicago Columbian World Exposition held in 1893 (Larson 2003). This dreamlike extravaganza, an enormous city–within–a–city painted snow white, with ethereal landscaping, attracting hundreds of thousands of visitors per day, did not embody mundane experience. Discussions between father and son about the Exposition would have been of a mythic, heroic character, not a reference to prosaic activity.

Fantasy is precisely the effort to get beyond the everyday. A modern example of fantasy play is computer gaming. Participating in online games calls forth acts of social and technical creativity that stretch past mundane experience. MMOGs such as World of Warcraft, Everquest, Lineage 2, and Shadowbane take users into social worlds where they cannot be said to have had embodied experience. A MMOG is a persistent computer game in which players develop a character and then advance by slaying dragons, healing the sick, performing magic, or any number of high fantasy activities. Players regularly “use” artifacts with which they have had no real–world experience. For example, in Everquest, a player could kill an enemy with a slime–coated harpoon while wearing an emerald dragonscale tunic.

Both the fantasy nature of these games and the kinds of social interaction they promote are new to users who must create not only characters for the games, but new ways of interacting. One player said:

I don’t know why I am such close friends with my EQ [Everquest] buddies. I do know that my EQ relationships are better than most of my relationships in [real life]. I think this is because when you are talking to someone online it’s easier to talk about certain things since you don’t have to look at a person face to face. (Yee 2002)

Such a statement about sociality defies what is learned from everyday experience about looking at others while speaking. Another player observed that in MMOGs people are “good friends due to the fact that you must throw away ... everything related to physically meeting a person” (Yee 2002). Here the player reflected that forming new kinds of social relationships was different from forming those of his embodied experience. The relationships were “better” and were constructed as not embodied; the player said he “threw away everything related to physically meeting a person.” It is evident that new forms of social interaction were taking place. An overemphasis on “everyday mundane experience” could make it difficult to interpret virtual experiences such as those that take place in MMOGs.

MMOGs also give rise to experiences that cannot be embodied. One player said,

I feel totally free in Shadowbane. I do what I want, I say what I want, and I kill who I want .. I feel at home here. Don’t get me wrong, generally I’m a nice guy, but I love just letting loose in there. (Ardisson, et al. 2004)

In the phenomenological tradition, “Embodiment is the common way in which we encounter physical and social reality in the everyday world. Embodied phenomena are the ones we encounter directly rather than abstractly” (Dourish 2001a). By contrast, experience in virtual games does not seem to have clear “direct” relationships to embodied experience.

Wartofsky’s Typology of Artifacts

We have found Wartofsky’s typology of artifacts useful for thinking about creativity. Wartofsky began with the observation:

By contrast with non–human animals, human beings create the means of their own cognition. That is to say, we create cognitive artifacts which not only go beyond the biologically evolved and genetically inherited modes of perceptual and cognitive activity, but which radically alter the very nature of learning and which demarcate human knowledge from animal intelligence. The cognitive artifacts we create are models: representations to ourselves of what we do, of what we want, and of what we hope for. The model is ... a putative mode of action, a representation of prospective practice, or of acquired modes of action. (Wartofsky 1979)

A philosopher of science, Wartofsky was concerned with many issues of interest to activity theory. His work is widely cited by modern activity theorists (e.g., Cole 1996; Engeström and Escalante 1996; Hedegaard 1999; Miettinen 1999), and his ideas are consistent with much of activity theory. Wartofsky’s emphasis on the activity of representing (as opposed to representations, emphasized in cognitive science), the concern with purposefulness, the use of Marx and Vygotsky, and the focus on artifacts will feel familiar to those conversant with activity theory.

Wartofsky proposed a scheme of primary, secondary, and tertiary artifacts. Primary artifacts are used directly in production. Secondary artifacts are “those used in the preservation and transmission of the acquired skills ... by which production is carried out.” In other words, secondary artifacts tell us how to use primary artifacts; they mediate and support the development of tool–related competencies. A manual might explain how to use a software system. The software system itself might explain how to use it (in help screens or animated paper clips). So, a single artifact can have a dual character as both primary and secondary artifact. (It should be noted that neither a user manual nor a good help system automatically results in the skills necessary to use a primary artifact successfully).

Kaptelinin (1996b) proposed that metafunctional competencies enable our understanding of how to use functional organs (such as knowing tricks and work–arounds), recognizing their limitations, and knowing how to maintain and troubleshoot them. Metafunctional competencies and secondary artifacts work together to provide expertise about the actual use of functional organs. Keeping in mind that functional organs involve both internal and external resources, we can see that metafunctional competencies assess both in determining how to deploy tools in a given situation. So a physician who is a generalist knows that when confronted with a patient with an unusual allergy, he should consult his online help system which may direct him to a secondary artifact to find a particular database of interest

Tertiary artifacts emphasize creativity. A tertiary artifact “transcends the more immediate necessities of productive praxis,” giving freer rein to imagining “possible worlds” (Wartofsky 1979). Such possible worlds occur as

models [which] are the highly specialized part of our technological equipment whose specific function it is to create the future. ... [M]odels are embodiments of purpose and, at the same time, instruments for carrying out such purposes. (Wartofsky 1979)

Creativity, then, is embodied in such models, or tertiary artifacts, that specify future object–oriented activity. We can create representations of “prospective practice,” as Wartofsky said, to go beyond the immediate environment. Creative activity means we are not limited to “reacting” to the environment as in distributed cognition; we have the possibility to design the environment through the creation of tertiary artifacts.

Miettinen (1999) used Wartofsky’s notion of tertiary artifacts to analyze changes in ethanol production in Finland over a period of several years. Scientists attempted to produce ethanol from birch chips (such chips being plentiful in Finland). The materials proved resistant to the scientists’ plans, but a simple diagram for the production of ethnanol published in 1981 guided the scientists’ efforts as they worked to discover whether the birch chips could be used economically. The diagram was a tertiary artifact that “synthesize[d] and generalize[d] the modes and results of actions, carrie[d] and transmit[ed] purposes, and oriente[d] to the future” (Miettinen 1999). Thus a simple tertiary artifact, shared among the scientists, embodied a creative idea about a new form of energy production and represented a “working hypothesis and a research plan” (Miettinen 1999). The ethanol production was not successful, but the knowledge gained during experimentation turned out to be useful in other arenas as the scientists made fundamental discoveries about the chemistry of the wood chips (Miettinen 1999).

Using activity theory, Miettinen gained theoretical purchase in several ways. His analysis began with self–aware human subjects in a collective object–oriented activity, not simply generic actors (or actants) in a system or network. A human subject created the diagram. Fellow subjects (scientists) could use the diagram as a mediating artifact projecting a future model of ethanol production. Sensitive to contradictions, the analysis highlighted the material resistance of the wood chips as they defied the scientists’ efforts. The wood chips were a major problem, not something that could be managed through small adjustments. In finally facing the fact that the wood chips did not have the chemistry for economical ethanol production, the scientists co–constructed the knowledge gained for other areas of application.

In a study of organizational innovation in a Finnish hospital, Engeström (1990) described the importance of tertiary artifacts in creatively changing doctors’ practices. The doctors accepted a new organizational model for their work only when its larger future implications were made clear through the presentation and discussion of tertiary artifacts. The artifacts “creat[ed] a perspective for the future of the entire activity system” that motivated the doctors to move forward and clearly depicted these new directions (Engeström 1990).

Miettinen (1998) emphasized the importance of tertiary artifacts for social change. He observed that, for example, artifacts of scientific research may “connect research activity to society, integrating ecological, economic, and societal values in a vision of the future ... .”; In a flat network, where human and thing are not distinguished, what entities could there be to have such visions?

Coordination, Cooperation, and Co–construction

The postcognitivist theories differ in the extent to which they emphasize creative fissures in activity. Activity theory looks for the creative possibilities of breakdowns, conflicts, and contradictions. Béguin and Clot (2004) observed that “Obstacles, disagreements [and] conflicts encountered in activity generate tension ... and invite the subject to mobilize and develop [new forms of organization].” Distributed cognition tends to focus on the smooth operation of a system as it reproduces its function. For example, Hutchins (1995) emphasized such a function by describing the means by which navigation on a U.S. Navy vessel was faithfully reproduced to pilot ships safely.

To give some structure to those places where we might seek creative fissures at different levels of collaborative activity, we refer to Raeithel’s (1996) three–part scheme based on earlier work by Fichtner (1984):

• coordination
• cooperation
• co–construction

Coordination is the most basic form of collaboration. It refers to cases in which people work toward a common goal, but carry out individual activities basically independently. Even though the outcome of each individual activity contributes to the outcome of a collective activity as a whole, these contributions are coordinated in a way external to activities of other individual participants (see also Engeström et al. 1987). In coordination we might find creativity in very small actions. As Béguin and Clot (2004) noted, “[E]ven the most repetitive movement of a production line worker is always unique.”

Cooperation is a more advanced form of collaboration. Collaborating individuals need to relate their goals to the overall objective of a collective activity, be aware of the actions of other participating individuals, and adjust their actions to the actions of other people. Typically these are small to medium adjustments, in line with the constancy of the overall object of the collective activity. As people work together to solve problems, there is considerable scope for creativity, though changes support the established object of the activity. In cooperation, we would expect to find people changing actions to overcome disruptions that threaten the instantiation of the shared object (as in the biotechnology research described in chapter 7).

Co–construction takes place when collaborating individuals not only cooperate to accomplish a prespecified common object but can also collectively redefine the object — and the collective activity — itself. Here creative activity may lead to major alterations in the very activity, as the object is no longer agreed on and is a source of contention and reevaluation. The object may then be constructed anew, that is, co–constructed.

Coordination and Creativity

Distributed cognition has been the theory that places the most emphasis on coordination. The method of distributed cognition — to trace representational state as it changes in moving across nodes in a system — makes visible and salient coordination among nodes (Hutchins 1995; Clark 1997; Kirsh 2004). While Raeithel considered coordination to be part of human activity, distributed cognition expands the notion to include both people and tools. Hutchins (1995) noted, “Representational states are propagated from one medium to another by bringing the states of the media into coordination with one another” (emphasis in original). Consistent with Raeithel’s scheme, distributed cognition highlights the external or distributed nature of the coordination.

In distributed cognition, the analysis of coordination tends to emphasize stability and the reproduction of a system (opposite in spirit to the point made by Béguin and Clot). Some of the emphasis on the smooth functioning of large systems may derive from the fact that Hutchins’s (1995) theoretical work was based on a study of a U.S. military vessel, a setting where precision and predictability were sought. Halverson (1993), herself a proponent of distributed cognition, said that the approach works best for “highly rationalized systems” in which elaborate coordination is likely to be a focus. Bazerman (1996) made a similar point, observing that such systems are not necessarily typical and should not be taken to be representative. Hutchins’s (1995) work is a beautifully detailed ethnography of the use of modern navigational tools on a military craft, but claims regarding a general “cognition in the wild” require considerably wider ethnographic reach. A focus on coordination is but one part of the story as Raeithel’s scheme indicates. And, there are differing ways of looking at coordination, as Béguin and Clot’s statement suggests. Activity theory positions coordination within a larger scheme that has scope for creativity at different levels of coordination, cooperation, and co–construction. Rather than locking down activity in predictable flows of state from one medium to another in a stability–seeking system, activity theory suggests that within the enactment of actual activity there is the potential for movement and change, even in what appear to be highly regulated activities such as those of factory workers.

Cooperation and Creativity

The second level of Raeithel’s hierarchy concerns everyday practical activity that does not require radical change, but does require some cooperative problem solving. Here we consider analyses of group activity in which subjects cooperated to advance their activity.

Spinuzzi (2003) analyzed the way automobile traffic engineers in Iowa transformed a system of accident narratives (official reports of accidents) into quantitative data. The quantitative data revealed accident patterns the engineers needed to understand to plan for better traffic management. The workers “turned thousands of narrative and diagrammatic descriptions into visual data display genres such as pie charts, bar graphs, and tables, each of which could have more impact on public policy than individual narratives could” (Spinuzzi 2003). With new ways of viewing accident data, the workers could “label dangerous stretches of road, dangerous behaviors, and dangerous classes of drivers.” Spinuzzi reported that at each step in the process “workers had to decide what to include in the transformation ... and what not to include ... .” Such human judgments were needed to advance the task.

In Spinuzzi’s analysis the workers were active subjects. Traffic safety was the collective motive in the activity. The workers “made decisions,” “transformed data,” and otherwise acted as aware subjects. Spinuzzi’s discourse was not of small adjustments or the application of simple rules. He described how the workers faced a mountain of narrative data — filing cabinets full of thousands of accident reports. It is a tribute to their creativity that they figured out how to transform such voluminous data. To do so, they employed specialized domain knowledge and conceptual tools including mathematical and statistical analyses to create more revealing representations of the data (Spinuzzi 2003).

Traffic safety is an apt illustration of the importance of a concept of motive and its relation to cooperative actions. Without the diligence of committed individuals who care about safety, accident rates in the United States would be higher (as they are in many other countries). Safety is not residua of prior function, or a by–product of tokens circulating in a network. In Iowa, active subjects — the workers — labored to transform accident data from one format to a very different format so that they could better fulfill the motive of improving traffic safety. The creativity that led to a cooperative transformation was motivated by a shared object.

Hutchins (1995) provided examples of sailors cooperating in response to breakdowns. Because of the social nature of cooperation, analysis at the abstract level of state propagation across representational media was inadequate at times in this analysis. For example, in describing how team members might pitch in to help one another, Hutchins (1995) began the discussion saying, “We can think of the team as a sort of flexible organic tissue that keeps the information moving across the tools of the task. When one part of the tissue is unable to move the required information, another part is recruited to do it.” But then he seemed to mentally step back from this formulation as he acknowledged that the sailors’ activity involved human sociability and could not be depicted simply as abstract information moving in a tissue. An anecdote entered the narrative, a shift in rhetorical tone from systems abstraction to personal reminiscence. Recalling the words of Roy D’Andrade, a well–known anthropologist, Hutchins wrote:

As I worked with the data, something that Roy D’Andrade once said kept coming back to me. A student was making a point about what people do at work, saying that in an auto factory people mostly make cars. Roy said something like: ‘How do you know what they are doing? Maybe what they are making is social relationships and the cars are a side effect.’ (Hutchins 1995)

D’Andrade’s sharply worded question suggested workers engaged in object–oriented activity — living out their own social lives as they saw fit and not just fitting into the constraints of the auto factory. With D’Andrade’s admonishment in mind, Hutchins (1995) observed that shipboard workers were “also constructing social relationships” as they filled in for one another. They conveyed subtle social messages by the extent to which they offered help when others needed it. A sailor would observe another’s level of effort in covering for a coworker and make a judgment about what it meant: “Doing the absolute bare minimum required when others know that one has the time and resources to do more is a clear statement,” Hutchins reported.

The sending and receiving of these subtle messages was not the function of a system on autopilot. Human judgment entered in, and awareness of what a worker’s actions might mean — nuanced cultural meanings. Here we have human cognition, not mechanical function or abstract information flow in a system of representational media. A different level of analysis — a social level, and thus an intersubjective level — was introduced. Social processes that could not be ascribed to the ship’s charts, plots, and protractors figured in the analysis (introducing asymmetry). It seems that at times distributed cognition analyses veer away from the purer forms of the theory, and bring in concepts more closely related to activity theory and phenomenology.

Co–construction and Creativity

Let us return to Spinuzzi’s traffic engineers. The transformations in the accident data the workers created altered the larger activity. Once the workers got a taste of the quantitative data, they wanted more. They wanted to know not just how many accidents occurred at rural intersections, but at which particular intersections. What kinds of accidents? Under what conditions? The labor–intensive system of going through the narratives manually was not a practical means by which to answer such questions (Spinuzzi 2003).

In actor–network theory, the manager would have been visible as a prince (or princess, in this case), but the state college student who engineered the system might never have appeared.

In the 1970s, the Iowa State Highway Commission hired a contractor to automate the process of analyzing accident reports. A mainframe system was built. It greatly enhanced accident analysis and transformed work at the agency in many ways (Spinuzzi 2003). But it was slow and difficult to use. In 1989, a manager hired an eighteen–year–old computer science student from Iowa State University to develop a PC–based solution. The student created a system that was used for several years and yet again transformed work at the agency. In actor–network theory, the manager would have been visible as a prince (or princess, in this case), but the state college student who engineered the system might never have appeared.

In distributed cognition, such a radical act of creativity would ill fit the focus on the smooth function of a coordinated system, nor would the transformative artifact created by a single individual. Distributed cognition analyses often provide especially fine descriptions of sophisticated technologies while at the same time suggesting that these technologies exist merely to reduce cognitive effort, to allow for the application of simple rules (e.g., Hutchins 1995; Halverson and Ackerman 2003). How such clever technologies themselves are invented is not discussed. Clark, a proponent of distributed cognition, approached this contradiction with a joke: “Our brains make the world smart so that we can be dumb in peace!” (1997). How “our brains” do this was left to the imagination. (And what is the difference between our brains and the rest of us that is left in peace?) Activity theory, by contrast, points to the creativity of the human subject in responding to cultural challenges, such as those described in Miettinen’s study analyzing creativity in scientific activity, or Engeström’s description of the doctors’ response to tertiary artifacts that enabled them to improve their practice. Spinuzzi’s historical analysis of an activity stretching over decades, detailing the contributions of individual subjects and groups, in a specific context, showed how the reflective and creative work of members of the activity led to the co–construction of new forms of data analysis and the larger social organization of the agency.

Spinuzzi (2003) studied the use of artifacts among technical communicators, highlighting the “idiosyncratic, divergent understandings and uses of artifacts and the practices that surround[ed] them as they develop[ed] within a given cultural–historical milieu.“ These divergent understandings and practices are a source of variation needed to provide material for change. Squashing individual variation forecloses the opportunity to study the microgenesis of change. As Béguin and Clot (2004) said, “[W]e need to situate action in the development of activity to retain the virtues of creativity and inventiveness ... recognize[d] in action.” Such divergent understandings, creativity, and inventiveness are not foregrounded in distributed cognition analyses. As we saw in chapter 7 in the discussion of the biotechnology researchers, what scientific research conceals may be as important as what it reveals. “Selecting what others see” is a social process that shapes scientific research. Examining what a theory leaves out may be just as important as understanding what it brings in.

Activity versus System

In attempting to understand creativity, we may ask: Where is the locus of control in culture? Does it lie in the system or in the enactment of activity? If it is solely “in the system,” then where could creative change come from? These questions expose a central tension in social theory of the last several decades: is the real question how to explain durable, stable structure, or is it to explain development and change? In approaches such as ethnomethodology and distributed cognition, stability is the phenomenon to be explained. Some interesting empirical accounts in distributed cognition concern the way stability is preserved in the face of technical change (Licoppe 2005). Actor–network theory’s case studies have examined change, especially in studying the activities of “princes.” Activity theory is also focused on development and change, attempting to position them in a cultural matrix assumed to have some stability (see, e.g., Blackler 1993).

In distributed cognition, control is seen as being primarily in the organization or system, especially in its management of coordination. Hutchins observed:

[M]uch of the organization of behavior is removed from the performer and is given over to the structure of ... the system. This is what it means to coordinate: to set oneself up in such a way that constraints on one’s behavior are given by some other system. (Hutchins 1995)

Here the locus of control is outside the “performer,” which has the effect of downplaying — even making invisible — acts of creativity and the uniqueness of an actual activity. But, as Béguin and Clot (2004) said, “[E]ven the most repetitive movement of a production line worker is always unique.” This view puts at least some control in the enactment of activity (even at the level of coordination). Béguin and Clot argued that the organization of activity is stable and predictable, but not the activity itself (see also Vergnaud and Récopé 2000). Even in the tiniest movement of an assembly line worker, there is scope for the new, the unpredictable. Such reasoning situates the locus of potential change in activity rather than in organizational structures. We saw in chapter 7 that the managers at Ajaxe dealt with the reality of organization versus creative scientific activity. As one manager said, “No matter how much structure you have in an organization, still the only way to get people to do work is to get them excited about something.”

Béguin and Clot observed that in practice, activities are enacted with judgment and interpretation, so the subject is always capable of something new (see also Taylor 1985). This way of thinking opposes that of distributed cognition, which looks for constraints on activity rather than possibilities for change. Béguin and Clot’s statement is interesting precisely because it directs us to look for the uniqueness of even the most apparently regular activity, and to be cognizant of the actual enactment of activity, not only to structures within which an activity is played out. In a similar vein, Emirbayer and Mische (1998) argued that it is essential to understand how “structural environments of action are both dynamically sustained by and also altered through human agency — by actors capable of formulating projects for the future and realizing them, even if only in small part ... ."

Ilyenkov (1982) observed:

In reality it always happens that a phenomenon that later becomes universal originally emerges as an individual, particular, specific phenomenon, as an exception from the rule ... . Thus, any new improvement of labour, every new mode of ... action in production, before becoming generally accepted and recognised, first emerges as a certain deviation from previously accepted and codified norms. Having emerged as an individual exception from the rule in the labour of one or several [workers], the new form is then taken over by others, becoming in time a new universal norm. If the new norm did not originally appear in this exact manner, it would never become a really universal form, but would exist merely in fantasy, in wishful thinking.

Hedestig and Kaptelinin’s (2005) study of the video technician illustrated Ilyenkov’s idea. The technician’s job was technical support for an educational videoconferencing system but in actual practice, he went well beyond his formal duties. Rather than giving over his activity to the structure of the system, the technician proactively and independently took on a wide variety of cooperative roles, including cameraman, systems administrator, teaching assistant, and supervisor. Initially, the technician’s extra work was not even visible in the system, and management did not understand how necessary his varied services, which exceeded the constraints set by the system, actually were to the success of the educational experience. Hedestig and Kaptelinin (2005) observed that an individual’s ability to transcend the immediate situation and carry out “suprasituational” activities is key to both personal development and to the larger development of an activity. Over time, the changes made by an individual subject may become part of the way the activity is normally done, or they may be the catalyst for other changes.

 

Reflexivity

Leontiev (1974) pointed to reflexivity as “a rise of consciousness,” “a reflection by the subject of reality, of his activity, of himself.” The aware subject is capable of reflecting on his or her use of technology in an activity. A technology could be imagined to be better than it currently is in the context deployed; it could be used differently than intended; it could change in certain ways. As with creativity, individual reflexivity always takes place in a social context. It is never isolated from activity. Gay and Hembrooke (2004) studied the development of learning communities and the role of technology in fostering community in an investigation of the introduction of wireless laptop computers at Cornell University. Wireless access was campus–wide, so students could use the laptops “anywhere, anytime.” Gay and Hembrooke examined students’ reflections on the technology through student journals and interviews with students in a communications class.

Some students reported succumbing to the “temptations” of reading e–mail, sending instant messages, and browsing the Web during class. They said that this behavior at times became “addictive” or “obsessive.” Students reflected that they were often “distracted” by the laptops, and that they had short attention spans as they switched from listening to a lecture, to reading e–mail, to instant messaging. While interactions within the communications class were enhanced in some ways with the technology, the wireless laptops opened the door for students to participate in other activity systems having nothing to do with class, during class time. Instead of listening to a fellow student’s question in a class, for example, a student could send instant messages to friends. Though the student physically sat in the classroom, with the new technology he could participate in unrelated activities. One student reported,

I felt in some ways that [the wireless computer] acted as a gateway to many activities that brought me away from my studies ... . I fell victim to searching the Web for unrelated nonacademic information that took my focus off the class. (Gay and Hembrooke 2004)

Gay and Hembrooke observed that

Students were ... participating in ... multiple divergent ... activities that were not class related (such as day trading, communicating with family and friends, and searching for personally relevant information) ... . Even when carrying out group assignments in class, many groups worked individually, coming together at the last moment to compile individual contributions into a collective product.

A concept of individual subject was crucial to Gay and Hembrooke’s investigation. While reflection is a social process, it is an individual subject who assesses experience, sometimes reformulating its meaning, and communicating that meaning to others. The students’ reflexivity was prompted by the researchers’ asking them questions and reading their journals. But it was the individual student who switched out of the classroom activity to another activity, and who could reflexively see that happening and report it. It was an individual who could, simultaneously, be “in a classroom” and, at the same time, be part of another activity, say, instant messaging with friends. It was an individual who was aware of this activity shifting, the need to coordinate the activities, and the costs such coordination imposed on each activity.

In reflecting on their own distracted behavior, some students came to the conclusion that social solutions were needed to curtail the negative aspects of the wireless technology. They suggested rules or restrictions on laptop usage. “I like the suggestion made in class the other day: the class as a whole should make a list of rules, and everyone should abide by them,” said one student. Students in the disrupted activity attempted some co–construction as they recognized that the technology that was supposed to form a learning community was in fact weakening that community as other activities intruded during class time. On reflection, the students recognized that unfettered access to competing activities would divide them, rather than unite them in a cohesive learning community (Gay and Hembrooke 2004).

DeCortis, Rizzo, and Saudelli (2003) approached the design of digital tools for elementary school children with activity theory’s notion of tool mediation. They designed a set of educational tools for narrative construction for children ages 6 to 8 to encourage creativity and cooperation. The tool environment, “Pogo,” was found to “offer new resources to encourage thought and choice.” In particular, the Pogo tools

support[ed] both personal reflection and intersubjective comparison, help[ing] children think about and analyze their own experience. This confers a role of ... reflexive mediation to instruments ... . (DeCortis, Rizzo, and Saudelli 2003; emphasis in original)

The children’s creativity was seen to be tied to the sharing of their reflections with one another. The analysis relied on strong notions of tool mediation and asymmetry, as the tools were analyzed to show how they inspired reflection and how the children then shared their reflections to advance the creative activity of constructing the narratives (DeCortis, Rizzo, and Saudelli 2003).

Through a process of reflection, people often change the way they use technologies. They go beyond what the designers had envisioned and cast tools as mediators in new activities.

Through a process of reflection, people often change the way they use technologies. They go beyond what the designers had envisioned and cast tools as mediators in new activities. Geisler (2003) provided a detailed account of the way she used her Palm Pilot to attain work–home balance, a function unforeseen by Palm designers. She noted that such usage was outside the cultural frame established for the Palm Pilot, which emphasized professionalism and efficiency at work. Reflecting on her own needs, she devised creative ways to use the device to fulfill her objects. She had used other tools in similar ways in the past. Using activity theory, Geisler’s account emphasized development over time:

Activity theory calls attention to the trajectory of personal motives that lead to technological adoption. The balancing of the competing demands of work and family shaped, almost from the beginning, my choices in time management technologies from the Day–Timer through HyperCard to the Palm. (Geisler 2003)

Geisler (2003) was able to look at her own activity, assess the shifts in her activities over time, and create effective ways to use a technology designed for other purposes. Theories in which people and things are equivalent, or “lumped together,” as Hollan and Hutchins (2003) put it, delete investigation of asymmetric relations such as a human reflecting on a tool and creatively changing its use, or a human developing new functions by building on the use of prior tools, as in Geisler’s evolution from Day–Timer to Palm Pilot.

Actor–network theory has its princes, but in distributed cognition, the special medium is no aristocratic alpha male ...

In distributed cognition, reflexivity is downplayed to preserve symmetry. There is a “special medium” but not an active intentional agent as realized in activity theory or phenomenology. Actor–network theory has its princes, but in distributed cognition, the special medium is no aristocratic alpha male, but rather a humbler entity “that can provide coordination among many structured media — some internal, some external, some embodied in artifacts, some in ideas, and some in social relationships” (Hutchins 1995). People are seen as having more advanced abilities to coordinate than other media, but to speak of the special medium (referred to as “that,” not “who”) as an intentional being capable of creating, reflecting, or resisting would be to miss the essential character of the special medium which is its ability to coordinate.

And yet, distributed cognition analysts find that they encounter activities of creativity and reflexivity in the arenas they study. Ackerman and Halverson (1998) investigated a telephone hotline group. They traced the propagation of cognitive state through the hotline’s organization, illuminating the use of various kinds of “organizational memory” to solve callers’ problems. The research is an elegant microanalysis of a single call, providing insights into the nature of organizational memory.

But the analysis went beyond distributed cognition in drawing attention to the hotline staff’s need to reflexively reconfigure work practices. The authors reported that the staff created a set of flexible informal routines to cover a wide range of hotline problems. They noted, “Indeed, the [hotline] manager repeatedly mentioned during the study that he was trying to balance flexible diagnosis and service with transaction efficiency” (Ackerman and Halverson 1998). There would be little room for such reflexivity and creativity in a truly symmetrical account. While people reflect, technologies do not. A pencil will never look at itself and say, “I could be sharper.” Only people reconfigure and redesign their practices and tools. Ackerman and Halverson (1998) moved away from symmetry toward a position more akin to activity theory. They remarked that their account had “a certain social twist” — allowing distinctly human social activities of reflexivity and creativity to enter the discussion.

 

Resistance

Resistance to technology is hardly a lively topic in the interaction design community. We are the designers and purveyors of new technology. Economically, intellectually, and spiritually, technology is at the heart of many of our favorite activities. Nonetheless, as experts, we have a responsibility to assess the impact of new technologies and to voice concerns when that is appropriate. We have a responsibility to craft theories that allow for activities of resistance, placing them within the scope of the human relationship to technology. The theories that posit flat networks of uniform nodes have the narrowest scope for describing or explaining resistance. An active subject is a necessary component of theorizing resistance, to explain how people act with technology.

Wertsch observed that an individual’s

stance toward mediational means [may be] characterized by resistance or even outright rejection ... . A focus on resistance and rejection leads one to consider a host of issues that do not arise when one assumes that cultural tools are friendly helpers. (Wertsch 1998)

Like creative activity, resistance is a rupture in the smooth flow of routine collective activity. It upsets the culture; it introduces conflict and discoordination. Wertsch’s making visible rejection and resistance is important because it addresses the need to “consider a host of issues” that do not arise in discussions of technology that focus only on its utility and functionality. Such issues do not arise when a theory lumps together people and things, since things themselves do not offer resistance by reflecting on something and then rejecting it. (In chapter 10 we discuss material resistance as conceptualized in actor–network theory, but that is a different kind of resistance.)

In the modern context, resistance is not so much the resistance of Luddism as the refusal to accept new technologies without broad–based analysis and discussion. Forms of such analysis have been pioneered in Europe. In Denmark, for example, the Danish Board of Technology organizes panels of citizens to study important issues regarding technology and society (Sclove 1995). The citizens study the issues and prepare an advisory report which is then consulted by government decision makers. The media report also on the panels’ findings so that wider discussion can take place. Could such an approach work in larger, more diverse countries? Sclove (1995) observed that in the United States, juries routinely reach consensus on difficult legal matters. He argued that though representation on citizens’ panels is not statistically meaningful, it is better to have the informed opinions of a group of citizens than to have no opinion at all, except for the opinions of those who stand to gain economically from a technology.

The Danish citizens’ panels have been influential throughout Europe and in the United States. Langdon Winner, an American political scientist, testified before Congress that citizens’ panels should be utilized in the United States to discuss new technologies such as nanotechnology (Winner 2003). He pointed out that the possible benefits of nanotechnology are enthusiastically proclaimed by scientists studying nanotechnology and business owners who wish to promote it, while those who have concerns are derided as anti–progress. But the possibility of invisible machines tiny enough to be inhaled should certainly give anyone pause. Such activities of resistance seem to fit most comfortably within phenomenology and activity theory in which technology mediates the activities of an active, purposeful subject.

While we see the incorporation of multiple voices in discussions about the societal uses of technology as the key issue of resistance in the twenty–first century, it is also true that some outright refusals of technology take place and must be analyzed. For example, Zambia refused to accept genetically modified corn, even when it was offered as charity. No new nuclear plants have been constructed in the United States for three decades, owing to the resistance of a coalition of scientists and ordinary citizens (see Bazerman 2001). European countries do not buy genetically modified food from the United States (Winner 2003). Such social activity involving complex negotiations and discussions among diverse groups of humans requires a theory that allows for intentional activity enacted by beings capable of imagining and planning the future.

 

Conclusion

In this chapter, we discussed two topics about which postcognitivist theories have distinctive perspectives: (1) intentionality, and (2) fissures in routine activity arising from creativity, reflexivity, and resistance. Examination of these topics reveals the extent to which a technologically empowered and socially contextualized subject is theorized. Activity theory and phenomenology construct such a subject, while actor–network theory and distributed cognition focus more on the level of network or system (though they bring in empowered humans through Machiavellian princes or special media). Activity theory and phenomenology position humans as the most powerful agents, beings with purpose and intentionality. In activity theory, the aware, intentional subject, engaged in object–oriented activity, acts to fulfill specific motives, rather than reacting, applying simple rules, or just doing what the system says to do. In phenomenology, Dasein is a human orientation of purposeful activity in a world that includes “tools for our use” (Dourish 2001a).

By contrast, in actor–network theory’s symmetrical networks, actors are equivalent agents. The agency of things, a particular emphasis in actor–network theory, is evident in that they have their own “timings, tempos and properties”; they push back in their interactions with other agents (Latour 1996a). In a contradictory move, actor–network theory introduced princes, who are subjects with intentions of the kind found in activity theory and phenomenology (Latour 1992). Distributed cognition proposed a system of state–passing media, but brought in the special medium, people. As we saw with ethnomethodology in chapter 2, formulating a theory around a radical, vivid, but circumscribed principle may result, if we were to invoke Freud, in a sort of reaction formation. That which is most ardently desired is turned on its head and its opposite put in motion. Thus actor–network theory, though it hoped to avoid “changing registers,” violated its most important discovery in asserting a principle of Machiavellianism. In admitting the “specialness” of people, distributed cognition gave a nod to a principle of asymmetry, but at the cost of the same theoretical contradiction Miettinen (1999) noted for actor–network theory: the humans end up being different from things after all.

In examining creativity, reflexivity, and resistance, we introduced Wartofsky’s hierarchy of artifacts and Raeithel’s hierarchy of collaborative activity. These conceptual tools structure analyses of the different kinds of tools and the way individuals and groups orient to collective activity. We found it necessary to make arguments for the importance of a concept of the individual subject. These arguments do not mean that we wish to reinstate a Cartesian isolate, a noble cognitive savage, as it were. Activity, always social in character, is never a closed system of cogito ergo sum. What we propose is that changes in both the subject and the system cannot be fully explained at the system level. The individual subject develops through mediated activity in observable, measurable ways (e.g., the zone of proximal development). Systems change under the influence of observable individual actions. Deleting the individual subject deletes important sources of change, fissures that lead to transformations, and responses to fissures. An important challenge for activity theory is to better theorize transformations between individual and collective levels.

We discussed both commonalities and differences among the postcognitivist theories. There are many reasons for the differences, among them temporal and cultural. Activity theory developed in Russia, phenomenology primarily in Germany, actor–network theory primarily in France, and distributed cognition in the United States. The latter two theories are of recent vintage, the other two stretch back decades. But perhaps more important than these differences are the distinctive intellectual challenges each theory took on in its historical context, what each theory responded to as it struggled to put forth an original perspective. Activity theory sought an alternative to behaviorist notions of stimulus–response and the shortcomings Vygotsky perceived in Piaget’s work. Vygotsky felt that a concept of culture was essential to psychology. Phenomenology responded to the overformalization of science and mathematics in which the everyday world of practical experience had little place. Actor–network theory was frustrated with the privileging of the social over the technical, or vice versa, depending on the account. Actor–network theory developed the notion of generalized symmetry to make visible the agency of the material world. Distributed cognition aimed to overcome the limitations of cognitive science where mental representations were the primary focus, ignoring the importance of tools. The varied intellectual predecessors of the theories are in part responsible for the distinctive “look and feel” of each.

As a group, the postcognitivist theories provide an important voice in the dialogue with universalistic theories in biology, psychology, and elsewhere. Although they do so in different ways, postcognitivist theories argue the need to analyze the microgenesis of change in particular historical contexts and to make visible the importance of technology.

It remains to be seen how each theory will influence interaction design in the future, and how the theories will influence one another. Some of the theories reproduce one another in key ways (such as activity theory and phenomenology), others contradict each other in some dimensions (such as activity theory and distributed cognition), while still others provide “probes“ that can be deployed in the development of a neighboring theory. In the next chapter we take up one such probe from actor–network theory, discussing its perspectives on agency and how they might be used in the development of activity theory. We build on actor–network theory’s emphasis on material agency, formulating new notions of agency from activity theory’s principles of asymmetry and mediation.

 

About the authors

Victor Kaptelinin is Professor in the Department of Informatics at Umeå University, Sweden.

Bonnie A. Nardi is Professor of Informatics in the Donald Bren School of Information and Computer Sciences at the University of California, Irvine.

 

Return to the Acting with Technology: Activity Theory and Interaction Design Table of Contents

 

---

Copyright ©2007, First Monday.

Copyright ©2007, Victor Kaptelinin, Bonnie Nardi, and MIT Press.

Acting with Technology: Activity Theory and Interaction Design: Chapter 9 by Victor Kaptelinin and Bonnie Nardi
First Monday, volume 12, number 4 (April 2007),
URL: http://firstmonday.org/issues/issue12_4/kaptelinin/chapter9.html