Notes for Clay Spinuzzi Network: Theorizing Knowledge Work in Telecommunications

Key concepts: actant, activity theory, actor-network theory, actor theory, API, arborescent, assemblage, black-boxing, boundary object, chained activity system, communal operationalizing, cross-training, cyborgs, folksonomy, genre, inscription, logics, mediation, modular work, net work, network, overlapping activity system, paradigm, rerepresentation, rhizomatic, social language, sociotechnical graph, sociotechnical networks, spliced network, sticky threads, synchretism, syntagm, text, weaving, woven network.

Extend activity theory by learning from actor-network theory to study knowledge work in organizations, concluding a synchretism is better than a synthesis. Tracing is like tracing through schematics or wire traces, which may be useful for finding breaks but very awkward for discerning the function of complex assemblies. Latour black box hidden behind interface. Sociotechnical networks characterized by heterogeneous, multiply linked, transformative, and black-boxed; Harraway sticky threads. Describes communal operationalizing and cross-training solutions. Actant like cyborg. Good analysis of texts as inscriptions functioning as boundary objects belonging to genres. Expertise emerging symmetrically from the assemblage insists on intelligence built into the environment, foregrounding the cyborg. Uses API metaphor. Lifelong employment replaced with lifelong learning.

Related theorists: Bakhtin, Bowker, Callon, Castells, Deleuze, Gee, Guattari, Haraway, Latour, Marx, Serres, Star, Vygotsky.

Networks, Genres, and Four Little Disruptions

Research question how do genres circulate in a complex organization shifted to basic question of how the company works at all.

(2) Few of these groups actually understand each other's work. When I began researching Telecorp, my research question was: How do genres circulate in a complex organization? By the end of the project, I inflected the question somewhat differently: How on earth does this company function when its right hand often doesn't know what its left hand is doing?


Latour claims term network has lost cutting edge and meaning by series of transformations to activity systems and tool use; apply understanding of physical telecommunications network made of wires, wood, plastic, glass.

(5) Bruno Latour declares that the term network has “lost is cutting edge” and in the process has lost its meaning as “a series of transformationstranslations, transductions – which could not be captured by any of the traditional terms of social theory” (1999a, p. 15).
(7) Activity networks are
linked activity systemshuman beings laboring cyclically to transform the object of their labor, drawing on tools and practices to do so.
(8) I want to exploit the tensions among these different understandings of network, and I want to apply them to a third understanding of network: a physical telecommunications network made of wires, wood, plastic, and glass.

Net Working

Activity theory and actor-network theory provide grounding around objects and recruiting allies in net working.

(16) Activity theory provides a cultural-historical, developmental view of networks grounded in the orientation of particular activities toward particular objects. . . . Actor-network theory provides a political and rhetorical view of networks and foregrounds the continual recruiting of new allies – both human and nonhuman – to strengthen the Telecorp network.


Genres as rhetorical responses to recurring social situations function in assemblages and hold network together.

(17) Genres – which can be glossed as typified rhetorical responses to recurring social situations (Miller, 1984) – do much of the enacting that holds a network together. . . . Genres typically function in assemblages, as I've discussed elsewhere (Spinuzzi, 2004), and their compound mediation enables complex activities such as the ones we've seen in this chapter.

Texts weave networks together; inscriptions vital role in actor-network theory to transform unmanageable phenomena into mobile texts (Latour).

(17) The word text comes from the root word textere, to weave together, and I suggest that's exactly what texts do: weave together these networks. In actor-network theory, inscriptions play a vital role in constructing networks. They transform complex, unmanageable, immobile phenomena into manageable, transformable, combinable, mobile texts (Latour, 1990).

Controlling behavior from outside via mediation leads to internalization of work.

(21) mediation involves controlling one's own behavior “from the outside,” as it were, through physical and psychological tools (Vygotsky, 1978, p.40). . . . As as workers mediate their work with these artifacts, they internalize the work.
(23) Genre supplies an account of stability-with-flexibility that is more fleshed out than fluids, modes of coordination, and regimes and at the same time leverages the notion of inscription that is so important in actor-network theory.

Social languages develop around particular activities enacted by particular groups as different logics, not just lists of terms; language not abstract system but rather concrete heteroglot conception of the world (Bakhtin).

(26) The term social languages is drawn from the work of language philosopher Mikhail Bakhtin, whose work has often been paired with activity theory. Social languages develop around particular activities enacted by particular groups of people. They are not simply lists of terms; they are actually different logics (again, on the sense of logos, word). . . . For any individual consciousness living in it, language is not an abstract system of normative forms but rather a concrete heteroglot conception of the world.

Connect social languages to Gee situated, embodied language.

(27) Learning a social language means joining, sharing, and coconstructing the logos constituted by it; it means joining a community and learning (if not necessarily accepting) that community's ideology or ideologies. And, as Bakhtin says, it also means differentiating oneself from others who have not learned the social language.


Trajectory of book is to show how activity theory can learn from actor-network theory to reach potential for study of knowledge work.

(30) Activity theory is well positioned to study knowledge work, probably better positioned than actor-network theory for reasons that will become clear in Chapter 6. But in its present “third-generation” stage, it needs to develop in specific ways to reach its potential. It could learn more from actor-network theory than it has.

What is a Network?

Activity theory interested in how people work, actor-network theory how power works.

(32) Activity theory is interested in how people work; actor-network theory is interested in how power works.


Tracing is like tracing through schematics or wire traces, which may be useful for finding breaks but very awkward for discerning the function of complex assemblies.

(33) two very different understandings of sociotechnical networks: the woven, cultural-historical and developmental understanding afforded by activity theory and the spliced, political-rhetorical, negotiated understanding afforded by actor-network theory. . . . Think of this work as tracing through schematics to find design flaws or tracing through a circuit to find loose wires or damaged connections, but in this case we're looking at heterogeneous elements such as people, practices, and texts.


Woven networks exemplar is Marx organic work organization progressively transforming same material yielding chained division of labor: network becomes more attenuated as craft skill separated from labor; typical focus of activity theory with interest in developmental activity.

(34) This sort of work organization is what Karl Marx called “organic,” in which the same material is progressively transformed, allowing the different stages to be isolated and to yield a chained division of labor in which work that had been accomplished in one place, by one person, is distributed in space and time. . . . This separation of labor from craft skills introduces weaknesses into the network; the longer [woven] the network becomes, the more likely it is that disjunctions or contradictions will form among activities and consequently the more attenuated and weaker the network becomes.

Spliced networks composed of convergence of different preexisting elements, leveraging unforeseen alliances and uses: network becomes stronger as more actants brought in; typical focus of actor-network theory with interest in poltiical-rhetorical alliances and negotiations.

(34-35) In contrast, spliced networks are exemplified by the technician who splices together disparate electrical devices and existing networks to respond to unforeseen alliances and uses. . . . These sorts of networks . . . are composed of different preexisting elements that have converged.
(35) These convergences lead to juxtapositions and connections among previously unconnected activities, strengthening the network as more actants are brought in to make the network solid and durable.
(36) The woven understanding is usually taken up by activity theory, with its interest in developmental activity. . . . The spliced understanding is usually taken up by actor-network theory, with its interest in shifting political-rhetorical alliances and negotiations.


Telecorp's Technological Network

Bowker study of Schlumberger found small networks claim much power.

(37) As Geoff Bowker points out in his study of Schlumberger (1987), networks claim large areas but in practice are vanishingly small; their claim to power is that they transform the world so that things outside the network don't matter.

Latour found complexity of system black box hidden behind interface such as phone company.

(38-39) The technical, organizational, disciplinary, political and economic complexity of the system was black-boxed or hidden behind a simple interface (Latour, 1987); customers just saw the “phone company.” The black box was maintained through thousands of daily, localized, often ad hoc connections, including genres such as checklists, electronic notes, and immense stacks of marked and highlighted printouts.

Telecorp's (Spliced) Actor-Network

Actor-network composed of actants entering into alliances, enrolling others, splicing together.

(39) An actor-network is composed of many entities or actants that enter into an alliance to satisfy their diverse aims. Each actant enrolls the others, that is, finds ways to convince the others to support its own aims. The longer these networks are, the more entities that are enrolled in them, the stronger and more durable they become. An actor-network, of course, is spliced; the actants intersect.

Callon argues actants define themselves by intermediaries they put in circulation, so political-rhetorical work includes non-humans.

(39-40) Actor-network theory encourages us to see the political-rhetorical work as occurring not just among corporations or people but also among non-humans and to use the same concepts for all of these entities. . . . Such actants, Michel Callon says, define themselves by the intermediaries they put into circulation. Texts such as emails, technologies such as switches, humans such as technicians, and money in all its forms put each other into motion, mediate each other, and transform each other.
(41) But as Susan Leigh
Star argues, symmetry is not pantheism. It's an attempt to avoid Cartesian dichotomies by applying the same concepts and vocabularies across the entire actor-network. And once that has been done, we begin to see the same sorts of actions taking place whether the actants are human or not.

Symmetry between human and non-human actants weakens actor-network theory, leading away from study of human cognition, competence, therefore unable to account for temporal change, cultural-historical development (Miettinen, Pickering).

(41-42) In particular, as Reijo Miettinen points out, actor-network theory's principle of symmetry leads it away from the study of cognition, human competence . . . as Andrew Pickering charges, actor-network theory does not provide a strong account for temporal change; it does not have a way to address cultural-historical development, something that impairs its discussion of agency.

Telecorp's (Woven) Activity Network

Roots of activity theory in Soviet Union did not scale well to broader social phenomena; recent theorists of activity networks try to account for stakeholders, interaction, and coevolution.

(42-43) the activity system doesn't scale up well to study broader social phenomena. Activity theory's roots are in the Soviet Union, where “the opportunity to study social phenomena was limited for political reasons,” as Victor Kaptelinin mildly observes. . . . But later activity theorists, particularly in Western Europe and North America, have belatedly begun to develop activity theory in ways that account for these stakeholders, how discrete activity systems interact, interweave, and coevolve. They have often called these complexes activity networks.

Activity networks assume asymmetry, emphasize development, foreground human ingenuity, and exhibit structure.

(43) Unlike actor-networks, activity networks assume asymmetry, casting nonhumans as mediators or objects of labor rather than as actants. They emphasize development, foregrounding human ingenuity, learning, and individual and social changes. And they exhibit structure in the composition of the activity networks and the activity systems that compose them.

(44) activity theorists tend to see humans as actors and nonhuman artifacts as crystallizations of human activity.

Emphasis on cultural-historical development in activity networks; many ways individual workers enter.

(44) Cultural-historical development is everywhere you turn in an activity network. Individual workers enter the network through formal training, informal apprenticeship, trial and error, and socializing.



Key characteristics of sociotechnical networks: heterogeneous, multiply linked, transformative, black-boxed.

(46) Sociotechnical networks can be seen as material assemblages that enact standing sets of transformations. These networks are heterogeneous, multiply linked, transformative, and black-boxed.


Multiply Linked

Haraway sticky threads connect everything as multiply linked.

(48) As Donna Haraway (1996) puts it, “sticky threads” connect everything to everything else.

(49) These activities are standing sets of transformations, but their complex interpenetrations mean that their transformations can be idiosyncratic and unpredictable. Each node has its own logic, its own connections, its own texts, and its own scales of space and time.


Nodes as such by being black-boxed to reduce and manage complexity, hiding local transformations.

(49-50) We black-box the nodes to reduce and manage complexity; that's how they become nodes. . . . Black boxes tend to hide not just complexities but also local transformations.


Examines five events at Telecorp that demonstrate most activity systems massed at border, permeable edge of organizational black box.

(53) the activity systems were almost all massed on Telecorp's border, at the very permeable edge of the black box, and any one of them could take initial input such as a call about interrupted service. Telecorp was almost all border.

Solution 1: The Cordon Sanitaire

Solution 2: The Uniform Regimen

Splicing events hindered weaving by communal operationalizing and cross-training.

(56) Cross-training might indeed help spread understanding about how Telecorp's heterogeneous parts work, but Telecorp is becoming more complicated by the day, and local knowledge has a short shelf life.
(57) Splicing events such as turnover, acquisitions, cross-training, and regulative changes meant that workers could not close the black boxes themselves by
operationalizing (to use activity theory's term) their routine actions. The weaving activity that would allow for communal operationalizing - the formation of shared tricks, habits, and genres, the sorts of things that make black boxes possible - was difficult to sustain because the activities of the different activity systems were weaving too rapidly, diverging too quickly, converging too chaotically, splicing across heterogeneous fields too frequently, and losing too many individuals too quickly.

Garrisoning the Passes and Interrogating the Locals
(59) Even if the connections didn't show up on an organization chart, these hidden passes were there.


Activity theory and actor-network theory agree networks heterogeneous, multiply linked, transformative, black-boxed.

(59) Activity theory and actor-network theory disagree about many things, but they agree that networks are (in the terms I've defined here) heterogeneous, multiply linked, transformative, and black-boxed.

How are Networks Theorized?
(62) Activity theory is primarily a theory of distributed cognition and focuses on issues of labor, learning, and concept formation; it's used in fields such as educational, cognitive, and cultural psychology, although it's also making inroads in human-computer interaction, computer-supported cooperative work, communication, and anthropology. In contrast, actor-network theory is primarily an ontology - an account of existence - and focuses on issues of power in science and politics: rhetoric, production of facts, agreements, and knowledge. It's used in science and technology studies, philosophy, and sociology. But recently the edges of these projects have begun to meet and contend with each other.


Weaving is aborescent, branching, evolutionary.

(65) Weaving is arborescent or treelike. It emphasizes the evolution and development over time from a root (the germ-cell, the “abstract”) to branches (the various implementations, the “concrete”).

First stroke in actor-network is a splice, which is rhizomatic, growing by accretion rather than evolution.

(66) In an actor-network, the first stroke is a splice.
(66) The stronger and denser the network of relations, the more real the steam engine is - the more manufacturers of spare parts within tolerances, the more mechanics on call, the more documentation for repairing such machines, the more uses developed for them. . . . Splicing is
rhizomatic. In this perspective, existence is achieved through accretion rather than development, associations rather than evolution.

Activity theory predicates splicing with weaving, development underpins political-rhetorical interests; the reverse for actor-network theory.

(67) For activity theorists, any account of splicing has to be predicated on a woven understanding; development precedes and underpins political-rhetorical interests. The first stroke is a weave. . . . For actor-network theorists, any account of weaving has to be predicated on a spliced understanding; political-rhetorical interests precede and underpin development. The first stroke is a splice.


An Engelsian View: The Science of Interconnections

Third generation activity theorists paid attention to polycontextuality and boundary crossing (Vygotsky).

(69) in this third generation, activity theory began to come to grips with two aspects of splicing: polycontextuality and boundary crossing.

(70) Vygotsky insisted that human activity must be understood as mediated by physical and psychological tools, a stance that can be understood as distributed cognition.

Structure of Activity

Cyclical transformation of particular object by individual collaborators via mediational means for particular outcome.

(71) In an activity system, individual collaborators use mediational means (physical and psychological tools) to cyclically transform a particular object with a particular outcome in mind.


Activity Networks

Variation 1: Chained Activity Systems

Contradictions in activity networks arise from chained activity systems whose hidden passages are black-boxed by organizational chart and overlapping activity systems.

(74-75) activity systems are linked by their “corners”, and each corner is something that has been produced by one activity system to be consumed by another. . . . That variation looks a lot like a supply chain.
(75) an activity network describes a
mass production or modular configuration of work with a strict division of labor, specialization, and central coordination. It black-boxes the organization chart.
(77) organizations have “hidden passes” or informal linkages that potentially connect any part of the organization to any other.

Variation 2: Overlapping Activity Systems
(77) Multiple activity systems converge on the same object, although that object is construed in different ways.
(79) So activity theorists such as Engestrom et al. (1995) are positioning activity theory to address work multidimensionally.

(79) working on tasks from different activities or frames of work simultaneously.

Boundary Crossing
(80) One example is that of sales engineers, who learn their work by gathering knowledge simultaneously from all the different activities whose borders they cross.


Spliced network understanding follows dialectics in rejecting simple causal relationships, but assumes multiplicity rather than immanent unity; pragmatic strand tracing from Machiavelli through Deleuze and Guattari and Serres.

(81) Like dialectics, a spliced or rhizomatic understanding rejects simple cause-effect relationships; unlike dialectics, it assumes multiplicity rather than immanent unity in everything and understands change as not necessarily developmental.
(81) Actor-network theory's spliced understanding follows a pragmatic strand of thought that draws from Michel
Serres and to a lesser extent Gilles Deleuze and Felix Guattari. That pragmatist strand can be traced back to Niccolo Machiavelli.

A Machiavellian View; Or, Sympathy for the Devil

In Machiavellian antireducationism power consequence of system; Latour symmetry accounts for participation of artifacts.

(82) Rather than rely on abstract social structures, ideals, or general laws to provide explanations, Machiavelli examines the relations forged among actors and allows those relations to be their own explanation.
(82-83) This antireductionism translates to the principle of symmetry that actor-network theory espouses. In this view, power is not a possession of a prince, it is a
consequence of the system; orders are followed not because the person who issued them is powerful but because they are transformed into actions that serve the interests of those who execute them.
(83) Latour scorns both Socrates and Callicles. If they just looked out the window, he says, these two would-be rulers would see the
real rulers of Athens.
(84) As an
analytical stance (not an ethical one), symmetry provides an account of how artifacts participate in the negotiations, not least by resisting the interpretations that humans attempt to impose on them.


Actant like Haraway cyborg: decentralized, interconnected assemblage functionally and semiotically recombinable.

(85) An actant is less like an astronaut and more like Donna Haraway's (1991) cyborg: decentralized, interconnected, an assemblage with constructed, confused boundaries rather than an organic unity.
(85) Assemblages make sense of a heterogeneous jumble of infinitely recombinable parts, not just semiotically but functionally.


Agency distributed, actors and mediators emergence from assemblage via translation, composition, reversible black-boxing and delegation.

(87) If everything mediates everything else, agency must be seen as distributed. . . . actor-network theory handles the problem by seeing subjects and mediators as network effects: subjects and objects, actants and mediators emerge from the assemblage rather than preexisting it.
(87) In actor-network theory, mediation involves mutual transformation of the assemblage, and this transformation has four parts or moments: translation, composition, reversible black-boxing, and delegation.

(88) Translation is a multilateral settlement involving four moments: problematization, interessement, enrollment, and mobilization.


(89) Interessement involves defining the stakeholders and splicing them in, locking them into place, inviting them to the negotiations.

(89) If it is successful, it redefines the actors' interests in such a way as to provide a shared problem space and it therefore simplifies their world.

(90) In mobilization, stakeholders agree on a cascade of intermediaries that reduces the number of interlocutors by providing a single representation (or a limited number of representations) of their interests.

(90) the assemblage becomes an actant itself.

Reversible Black-Boxing
(91) This complication is needed, paradoxically, to ensure enough simplicity for a settlement to function: complex activities require a complexity of work practice without an increase in complexity of individual actions.


Delegation crosses boundary between signs and things, affecting tasks and morality (Latour).

(92) Delegation involves “crossing the boundary between signs and things (Latour).
(92) In situations like these, tasks are delegated, but so is

How are Networks Historicized?

How are Networks Enacted?

Differentiation between modular work of Industrial Revolution and net work holding together standing sets of transformations in sociotechnical networks of cyborgs.

(135) In such organizations, everyone is at the border; everyone can link up with anyone else inside or outside the organization. This is what I've been calling the net work, the coordinative work that weaves and splices divergent work activities and that enables the standing sets of transformations that characterize such work.

(136) By
modular work, I mean the understanding of work organization grounded in the Industrial Revolution, the understanding that has enabled industrial and managerial capitalism.
(137) Relations among activities are circumscribed, black-boxed, with links limited to maximize control and enforce the process. Workers are focused on the object of their own activity system.

(137) As indicated above,
net work is the coordinative work that enables sociotechnical networks – standing sets of transformations – to hold together and form dense interconnections.
(138) Rather, we're faced with collective subjects, composite subjects whose stability is punctuated by changes in the assemblages that constitute them,

Net Work and Informational Capitalism

Invokes Castells and other theorists to reach co-configuration in distributed work.

(140) In distributed work, the emphasis shifts from predictable, monodirectional flows of information and services to unpredictable, multidirectional flows, and services and products are constantly adjusted or “co-configured.”

Net Work and the Information Age
(140) In this latest phase, the Information Age, “relationships are the dominant reality” and links across agents increase dramatically, leading to greater horizontal complexity across and within organizations.

Net Work and the Informatics of Domination

In Deleuze control societies, Haraway Informatics of domination negotiation is essential skill for Callon polycentric border crossing.

(141-142) For instance, Gilles Deleuze (1995) argues that whereas the twentieth century was primarily marked by disciplinary societies (Foucault's terms), the twenty-first will be dominated by control societies. . . . The result is a sort of “endless postponement” rather than a defined avenue of development; workers travel in continuously changing “orbits,” they “undulate,” they find themselves switching jobs and careers and positionalities.
(142) Like Deleuze, Donna
Haraway outlines the threats of what she calls the “informatics of domination”: the systematic deskilling of workers and their resulting vulnerability; the “homework economy” in which the workday is no longer limited and work is no longer confined to the workplace thanks to new technologies; the decentralization of state power, with increased surveillance and control; and the massive intensification of insecurity (1991).
(143-144) In this shift toward net work, negotiation becomes an essential skill. . . .
Rhetoric becomes an essential area of expertise; direct connections mean that everyone can and should be a rhetor. Taylor gives way to Machiavelli.
(144) This phenomenon is described in actor-network theory as “polycenteric” actor-networks (Callon, 1992, p. 83) and in activity theory as “polycontextuality” and “border-crossing” between activity systems (Engestrom).


King and Frost broad definition of text as concrete representation in some medium of abstract symbols refering to something conrete.

(145) Let's use King and Frost's broad definition: “the creation of a concrete representation on some medium (e.g., paper) of abstract symbols that refer to something concrete – in the world of things, ideas, feelings, and so forth” (2002, p. 5).
(145) Texts weave and splice so successfully because they are
inscriptions, concrete traces that represent phenomena in stable and circulable ways. The appear in genres, regular responses to recurrent situations that can connect activities in continuous, developmental ways while accommodating changes and that function ecologically. And they are boundary objects, artifacts that serve as mutual reference points across different activities while retaining different meanings within these activities.


Latour noted centrality of inscriptions in studies of scientists, creating realities by linking phenomena to particular activities; John Law multiple realities.

(145) In his studies of scientists, Latour notes how central inscriptions are in performing their work. The key, Latour argues, in in transforming the substance to be studied into an easily manipulable representation that is usually about the size of a desktop or smaller.
(146) They also create realities – sort of. In representing phenomena, inscriptions link those phenomena to particular activities. And, John Law argues, multiplying inscriptions
multiplies the realities that they describe (2004a).


Texts belong to genres providing developmental influence on human activity, woven over time and spliced as hybrids in intersecting activities; offers methodology of tracing circulation of genres building networks of activity.

(146) Texts belong to genres.
(146) Types of inscription tend to develop over time within particular activities to meet recurrent needs. These
genres provide a developmental, stabilizing influence on human activity. . . . Genres are, then, woven or developed over time to respond to recurrent situations but also spliced or hybridized to adapt to local conditions and intersecting activities. And tracing them allows us to examine how they circulate through and build networks of activity.

Boundary Objects
(147) Texts function as
boundary objects.

Boundary objects material links between activities providing productive difference or coordinative role.

(148) Boundary objects are material links between two or more activities, functioning differently in each activity, providing a productive difference and often a coordinative role.
(148) So a boundary object is often an
assemblage of related texts (inscriptions, genres) that collectively plays different roles in overlapping activities.

Texts as inscriptions functioning as boundary objects belonging to genres.

(148) So texts are inscriptions that represent phenomena, belong to genres that construct relatively stable relationships, and function as boundary objects that bridge among different activities. Texts create circulating rerepresentations: representations that themselves become represented by other representations (Latour, 1999b).



It is obviously tempting to try redeveloping these cases of net work with examples from my own workplace: following an order, money, substitutions, and workers.

(149) Let's follow an order to get a sense of how texts circulate – as inscriptions and as genres – and to see how circulating rerepresentations, or series of transformations, weave and splice together Telecorp's activities.

There Was No “Order”

There Was No Transportation without Transformation

There Was No Single Genre

Summary: Following an Order
(155) In Case 1, such texts were cast in relatively stable genres, and thus worked as boundary objects that can be shared across functional groups. To the degree that they were stable, they wove together these functional groups.
(155) At the same time, these texts served as one way to associate historically disparate and heterogeneous activities,
splicing them together.


Following the Money in Cash Posting
(156-157) Indeed, stacks were
the central way for Carly and her co-workers to coordinate the many genres they had to progressively recombine on their desktops.

Following the Money in Credit and Collections
(160) Like other texts, money traveled through circulating rerepresentations: cash, checks, AS/400 entries, adding machine tape, bank account entries, and a variety of other genres.

Summary: Following the Money
(163) Money's rerepresentations
wove the activities together by constituting regular ways for workers to relate their activities, standing sets of transformations that developed as the activities of Telecorp developed. But they also spliced these activities together by bringing different social languages and different goals into contact.


Latour sociotechnical graphs modeling assemblages paradigmatically and syntagmatically.

(163) Sociotechnical graphs are an attempt to empirically model assemblages of humans and nonhumans as they associate themselves with each other and intermediate each other. Borrowing from linguistics, Latour et al. (1992) describe these associations as having paradigmatic and syntagmatic dimensions. By syntagm, the authors mean the associative dimension, the AND dimension: the human and nonhuman elements that are associated in order to make a statement true. By paradigm, they mean the substitutive dimension, the OR dimension: the human and nonhuman elements that can be substituted for each other while keeping the statement true.
(166) Seeing relative stability allows us to black-box actants and treat them as unifed wholes; it lets us stop what could become an endless, recursive process of unpacking those black boxes or following those “
sticky threadsthat, Donna Haraway (1996) tells us, connect everything to everything else.

Summary: Following the Substitutions
(167) By viewing genres syntagmatically and paradigmatically, we can get a sense of how these genres are being perceived, used, shared, and transformed. We can see them as black boxes and examine where they come open. We can also see how syntagms – these strings of genres associated with work, which vary widely across workers – become paradigmatically compressed and black-boxed.

(168) This high level of circulating is typical of knowledge work. In fact, Castells suggests that circulating workers drives innovation, a claim that seems to be supported at Telecorp. Workers, like the environmental activists studied by Luther Gerlach (2001), functioned as “living links” who circulated through loose networks, lending relative coherence and spreading and rearticulating genres.
(171) Like the children whose memory improved when they used cards in Vygotsky's study (1962), Bateson's blind-man-with-cane (1972), or Deleuze and Guattari's horseman-with-stirrups (1987), workers were transformed in assemblages, becoming workers-with-social-languages, workers-with-genre-ecologies.

Summary: Following the Workers


Is Our Network Learning?

Reference to Zuboff and Maxmin lifelong learning reminds me of the George Bush quote, Seldom is the question asked, Is our children learning?

(173-174) lifelong employment is replaced by what Zuboff and Maxmin call “lifelong learning.” . . . workers are having to focus far more on boundary spanning or coordination among very different activities.


Learning by associative complexes, becoming a dividual.

(176) The assumption that people generally make about learning and training is that individuals progressive accumulate skills; it's developmental. But in a network, it's far from clear that developmental models obtain: the individual herself is called into question, becoming what Deleuze calls a “dividual” (1995) whose work activities are fragmented and interleaved, whose attention is continuously partial (McCarthy & Boyd, 2005; Stone, 2006), whose work takes place in many organizational structures simultaneously (Castells, 1996; Drucker, 2003), whose “activities are increasingly infiltrated with . . . new types of working and learning – of living” (Johnson-Eilola, 2005, p. 31), and whose learning is consequently more self-directed (Senge, 1994) and continual at all levels (Castells, 1996). Under such conditions, associative complexes, which Vygotsky characterized as childish (1962), become the primary way of understanding and learning about work (Johnson-Eilola, 1998, 2005).


Apprenticeship: “You Never Ever Do a Partial Connection”

Formal Telecorp Training Sessions: “Nine Times out of Ten . . .”

Corporate Training Outside Telecorp: “Nobody Had Time to Learn from Her”

Documentation: “I Need to Do It from This Day Forward”

Computer-Based Training: “Basically It's Just a Crash Course”

Trail-and-Error: “Willing to Get Your Hands Dirty”

Stories: “There Was Nothing About a Dog on the Ticket”

Summary: Making Sense of Learning Measures at Telecorp
(186) Learning becomes a different problem when organizations are less stable, less bounded and hierarchical, more interconnected.

(189) vertical expertise was supported by a mix of formal techniques and informal techniques . . . but their horizontal expertise - the cross-boundary coordinative work that is so common and vital in net work - was supported almost wholly by informal, contingent ways of learning: workers learned through apprenticeships, stories, and trial-and-error how to perform the boundary spanning that they had to do each day across multiple activity contexts.
(189) the developmental paths or learning trajectories were less like spirals and more like

Problems with Activity Theory's Developmental Account
(190) Despite some influx of dialogic theory, activity theory still understands learning as dialectical. . . . Their assumption is that a unitary, stable self develops continually, even across discontinuous contexts, like an astronaut who orbits through and learns from worlds while remaining insulated from them.


Expertise emerging symmetrically from the assemblage insists on intelligence built into the environment, foregrounding the cyborg; tie in Gee and Norman.

(191) In contrast, actor-network theorists understand expertise symmetrically, as emerging from the assemblage rather than being driven by human action. Without her lab, the scientist is not a scientist; without the food processor's safety features, the sous chef becomes a nine-fingered fumbler; without fences, we are not such good neighbors.

Workers competencies fluctuating as function of assemblages in which they act: compare Benkler networked individual, Deleuze dividual, Haraway cyborg.

(192) Like Yochai Benkler's networked individual (2006), Gilles Deleuze's dividual (1995), and especially Donna Haraway's cyborg (1999), these workers found “their” competencies fluctuating as a function of the assemblages in which they acted.



Multiply Linked




In summary Telecorp performed net work well but not net learning and training measures, which were tactical and reactive rather than strategic and productive.

(195-196) So Telecorp, although it performed net work well, had not achieved net learning to the extent necessary. . . . Informally, Telecorp's learning and training measures were better suited to a smaller organization with less turnover and consequently stressed contingencies rather than principles. The result was an organization whose learning was tactical rather than strategic and reactive rather than proactive.

Conclusion: How Does Net Work

Seeking synchretism of activity theory and actor-network theory rather than synthesis.

(197) How to take the two perspectives and put together a reasonable settlement, a synchretism rather than a synthesis.



Multiply Linked



(200) Workers must constantly coordinate, negotiate, build trust and alliances, learn, and cross boundaries. Managers can no longer rely on hierarchical work organization to provide topsight (if they ever could) and must figure out how to support net work. And researchers must learn how to trace and conceptualize the constant boundary crossing that characterizes this work

Implications for Workers


Time Management

Project Management


Implications for Managers

The Toptech Change Request process reflects the tension between the well intentioned desire to increase agility by paying attention to decentralized, cross-functional, project-oriented work processes while confounding this effort by inaugurating successive new regimes of inflexible, form-based controls.

This engagement with business problems represents the real position of the thinker in a capitalist milieu, providing insight to managers: encourage stabilizing regimes, APIs, from which folksonomies emerge and provide persuasive vision and sufficient feedback for projects.

(202) Trying to force net work into a modular work configuration tends to sharply reduce agility.

(203) managers need to encourage stabilizing regimes. Let's call these sorts of black boxes “liaisons,” “APIs,” and “aggregations.”
(203) APIs in net work might include genres and other boundary objects.
(203) Tags start out as idiosyncratic, but a “
folksonomy” or emergent set of shared categories typically emerges as a second-order effect.

Strategic Thinking
(204) managers must provide a persuasive vision for each project and sufficient feedback for workers to see - and take ownership of - that project.


Implications for Researchers
(204) Instead of bounding the case, I repeat the advice I give in Chapter 1: follow the actors and texts, the contradictions, the disruptions, and especially the genres.

(205) Activity theory is better suited for this task, I believe, because it has a coherent, developmental account of learning and competence (see Chapter 6), and at the same time it has already begun developing accounts of splicing, such as knotworking and netWORKing.

Final judgment is that most promise with activity theory, especially underdeveloped adoption of Bakhtin dialogism to better deal with rhetoric of net work; need synchretism of activity theory and actor-network theory.

Synchretism of activity theory and actor-network theory: a gnomic formula with which to end the book.

(205-206) Some of activity theory's more forward-looking theorists have begun to leverage nondialectical ways of describing interactions in networks. The most important development in this direction, although currently underdeveloped, is activity theory's adoption of M. M. Bakhtin's dialogism.
(206) Rhetoric, as I argued in Chapter 2, has been a weak spot for activity theory due to its reliance on dialectic, so bringing in dialogism provides a way to better acknowledge and deal with rhetoric in net work.
(206) We need a more flexible, more associational activity theory with a stronger splicing account if we're going to analyze net work properly. We need activity theory to be dialogical and more rhetorical.
We need a synchretism.

(207) Only by taking a strategic stance will we able to identify objectives, set goals, take action, and retain the dynamism and flexibility necessary to cope with net work – whether we're workers, managers, researchers, or theorists.

Spinuzzi, Clay. Network: Theorizing Knowledge Work in Telecommunications. New York: Cambridge University Press, 2008. Print.