Spinuzzi, Clay. Network:  Theorizing Knowledge Work in Communications. New York: Cambridge UP, 2008.

Chapter One:  Networks, Genres, and Four Little Disruptions

In this introductory chapter Spinuzzi does a lot of definitional work in order to inform the rest of his study.  S. is an activity theory researcher at heart; however, he also sees that AT has severely limited itself.  To understand and work past these limitations, Spinuzzi recommends adopting some of the tenants of Actor-Network-Theory.  To put his two theories to the test, S. will research a mid-sized telecommunications company in Texas.  He aims to answer the following questions in his study:

  1. How on earth does this company function when its right hand doesn’t know what its left hand is doing?
  2. How do such knowledge work organizations function and thrive, and how can we develop a better theoretical and empirical account of this sort of work?

In sketching out the two theories in the introduction, S. draws some interesting distinctions:

  1. ANT is considered political and rhetorical because it is in effect a politics and rhetoric of symmetry, one in which no Cartesian lines are drawn between humans and non-humans (7).  This is important because it implies that ANT has non-human actants.
  2. Activity networks are “linked activity systems – human being laboring cyclically to transform the object of their labor, drawing on tools and practices to do so.  These activities themselves are nodes, nodes that are constituted by, but transcend, the humans and nonhumans who participate in them.  The links in the nodes of an activity network are often portrayed as supply lines:  Activity A labors to produce an artifact for Activity B; Activity C labors to develop practices that then serve as rules for Activity B; and so on (7)
  3. Activity theory adopts distributed cognition as an explanation for knowledge making (8).
  4. For activity theory, “contradiction” is the fundamental disagreement about how two activity systems should relate; this contradiction motivates innovation and changes the activity systems themselves so that the link can remain. (12)
  5. Multiplicity – or as ATs call it “polycontextuality” is what happens when “people with different sets of expertise tend to use different frameworks, techniques and tools, apprehend shared phenomena quite differently, and still manage to discuss these phenomena as more or less coherent” (13).
  6. Networks, while they seem really big, are actually “vanishingly small” – especially if we take the ANT definition of an actor-network.
  7. Spinuzzi is interested not so much in networks, but more so in “net work” – “the ways in which the assemblage is enacted, maintained, extended and transformed; the ways in which knowledge work is strategically and tactically performed in a heavily networked organization (16).  To achieve this he relies on AT to provide a cultural-historical developmental view of networks.  He uses ANT to provide a political and rhetorical view of networks that foregrounds the continual recruitment of new allies to strengthen the network.
  8. Boundary objects – “objects that are both plastic enough to adapt to the local needs and constraints of the several parties employing them, yet robust enough to maintain a constant identity across sites (21).
  9. Mediations – involves controlling one’s own behavior “from the outside,” as it were, through physical and psychological tools.  This is self-regulative work is transformative:  by mediating their own work, human being transform themselves, finding that they can do things that they could not do in an unmediated way (21).

To understand the ways that networks work, S. relies heavily on his own definition of “genres.”  For S., a genre creates stability and strengthens connections (18).  Because ANT conceives of mediation in terms of transformation and because AT takes mediation as a “black box” act of transference, S. employs the genre as a means to frame the stability/instability of meaning more productively.  By relying on Bakhtin, S. demonstrates how the “social language” or the sociolinguistic belief system that defines a distinct identity for itself within the boundaries of language. . . . No words are neutral: ‘Language has been completely taken over, shot through with intentions and accents.  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 (26).  By using Bakhtin’s conception of social language, S. can explain how multiple genres work together to compose the networks he so wishes to trace.

Chapter Two:  What is a Network?

In this chapter Spinuzzi illustrates the differences and similarities of ANT and AT by tracing the evolution of a problem in a telecommunications company.  There are some great moments in this chapter where the difference between the two is well stated and the similarities are well compared.  First, S. illustrates the two different kinds of networks: AT & ANT.

For AT, networks are comprised of interwoven activities.  Drawing philosophically from Marx, AT demonstrates how when actions/labor is divided into specific units to enhance and increase the speed and method of production, we become specialists.  Marx called this process “organic” or the process “in which the same materially is progressively transformed, allowing the different stages to be isolated and to yield a chained division of labor” (34).  Activity theory uses this metaphor to describe the ways that networks are comprised of interwoven historio-cultural elements.  What this conception of the network also implies is that the larger the network becomes the less strong or robust it will remain.  In AT’s conception of the network, the contradictions and disjunctions that occur because one laborer is unfamiliar with the network as a whole creates weakness.

In contrast to AT’s definition of the network as interwoven, ANT conceives of the network as spliced.  According to S., “Whereas woven networks grow through development, spliced networks grow through opportunistic alliances, through unpredictable jumps and sideways connections.  They do have a history, a history of translations, but that history is one of contacts and negotiations and compromises (35).  Because of all of the new contacting and negotiating that occurs in an A-N, the network itself is expanded and made more solid.  In other words, the commitments of the actants involved in the network are built upon one another, consolidating power and making the network stronger through the process of negotiation and contact.

For Telecorp’s network in S.’s study, there are three networks always in play.  First there is the actual technological network.  This is the network composed of wood, plastic, fiber, etc.  These networks are all about “coverage,” but what they actually connect is quite small (in spatial terms).  The network is multiply linked and easy to disrupt, but almost impossible to immobilize/take offline because of its multiplicity.  This network is also “black boxed” – a concept we’ll revisit in a moment.

Telecorp’s spliced actor network is the long train/chain of associations between humans and non-humans that join humans together.  As S. notes, “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 are enrolled in them, the stronger and more durable they become.  An actor-network is, of course, spliced; the actants intersect (39).  The A-N works to work AROUND the problems rather than address them specifically (because of the multiplicity of routes) and strengthens itself by enrolling the work of allies through such strong mediators as texts.  Texts transform and persuade allies (and traitors) to return to the network and act right.  In other words the “actor-networks expand through intersecting, enrolling, and translating other actants.  They consolidate through the ties that bind – the ever-tightening mutual enrollment of intersected actants” (41).

Telecorp’s woven activity network is also present in the company.  While AT is also interested in the spaces that occur between the micro and the macro (43), the existence of distributed cognition and the independent activity of human agents, AT does not account for the nonhuman agents as intermediaries or objects of labor rather than as actants (43).  The AT view of Telecorp accounts for the cultural-historical (training, apprenticeship, socialization as well as the development of new and different technologies, etc.) in the development of the event/connection being studies.  ANT doesn’t.  For AT, the question becomes “Where did this rule, artifact, subject, etc., come from?  These connections are lines of development – of sorts” (45).

In tracing networks, S. identifies 4 characteristics:

  1. Heterogeneous – networks are composed of assemblages of humans and non humans – really composed of assemblages of assemblages.  The assemblages are woven over time; however, sometimes the woven assemblages (telephone, internet, cell) need to be spliced together to function (the entire multimedia company).
  2. Multiply Linked – Networks are made of innumerable parts that are linked to one another.  Because of this multiplicity, bad connections or difficult actors can be routed around, spliced and replaced fairly easily.  This is the nature of work in the transnational capitalist age.  Interestingly, texts provide a HUGE amount of linkages because they weave the relationships between individual actors and collectives.  In other words, texts attempt to demonstrate or translate how the author of the text (management in this case) wants to produce smooth, predictable relations.  Splicing – as an act – always involves a negotiating and rearticulation – transformation – of texts (48).
  3. Transformative – Because networks transfer information, the information they transfer often undergoes transformations at each node on the network.  Each node, according to S., “has its own logic, its own connections, its own texts, and its own scales of space and time” (49).
  4. Black Boxed – This metaphor – originally developed by Latour – argues that most nodes on the network we “black box” to hide their complexity.  We do this so that we can more easily understand the simple interface for a complex set of workings. . . in other words we truncate the network to a unified entity in the interest of efficiency.

Because of the new nature of the networked information workplace, it is impossible to have a single point of contact or a universal set of operators (cross-training in total – Latour discusses this in terms of Pandora’s Box – what happens when you open all the black boxes of an entity like an organization at once.  It’s overwhelming).  Activity theory’s accounting of weaving is also not adequate because of the rapid changes inside each of black boxes of the organization (departments).  This is why ANT must be supplemented.  In closing, S. notes that a network is a “woven and spliced, divergent and convergent, culturally-historically developed and anachronistically associated” state of events/actions/affairs” (60).

Chapter Three:  How Are Networks Theorized?

Spinuzzi attempts to tease out the theoretical differences between AT and ANT in this chapter.  For ANT, the first move for an analyst is a splice – and splicing is rhizomatic (66).  According to S., “existence is achieved through accretion rather than development, associations rather than evolution” (66).  But before I go any farther on ANT, let’s consider the theory behind AT.

Spinuzzi begins by noting that “Activity networks consist of developing sets of activities anchored by a common object toward which people strive.  The root of this developmental view is dialectics, the interactionist understanding of change that permeated Marxist accounts of activity” (68).  In this way, AT is truly concerned with what happens when actors come into contact with interconnections.  Much like ANT, this is a key focus of study; however, for AT the change or action that result from dialectic encounters is unchangeable, irreversible.  This implies an arborescent or genealogical tracing of development – linearity in the action of actors.  Further, AT doesn’t take into account the agency of non-humans; rather, they are conceived of as tools of labor or objects as themselves.  This account – with its linearity and non agency of non-human objects – reads like a modular conception of work.  Every person on the assembly line performs a specific task and that accretion of tasks leads to a completed whole.  This assembly line metaphor is what happens – through and over time – in a dialectical activity theory model.

The contradiction is also key to AT.  The contradiction is firmly embedded in a dialectical accounts of development.  Though this does reconceived of essentialist and cause-effect explanations of things-in –themselves, it doesn’t account for processes that undo the objects they once constructed.  More on that in a bit.  Anyhow, contradiction. . . Contradictions are “engines of change:  they provide the impetus for the sorts of reorganizing, reconceiveing, and reworking that characterizes a living activity system or network” (73).  In other words, the contradiction is the site where the forces of dialectic come to one another and eventually push past both original positions into newer, higher forms (73-4).  After realizing that “activity networks” as originally theorized had problems: 1) there must be internal, not only external contradictions and 2) boundaries and paths aren’t clear cut (there might be Hannibal’s passes that allow for bypassing the traditional in-place network), ATs started to rethink their activity networks.  Eventually a model that valued overlapping activity systems was adopted.  This model realizes the importance of polycontextuality (working on tasks from different activities or frames of work simultaneously) and boundary crossing (agents who can mediate between two separate activity networks through the “hidden passes”).

Now, on to ANT.  Spinuzzi identifies a lot of components of ANT, so I’m just going to list them:

  1. Ant is pragmatic because it relies on the individual actants to define the network and to provide their own explanation of said network.
  2. ANT allows for the reversal of relational interactions – in other words, whereas AT’s dialectical interactions created a one-way path into history, ANTs relational interactions can be undone  – unraveled when the allies of the AN betray one another.
  3. There is not an underlying structure dominating all things – power 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).  This is the reason that ANT is often called Machiavellian.
  4. All actants are also actor-networks in that actants are comprised of assemblages of networks of relations.  There is no organic unity, only recomposition and reanimation – think Haraway’s cyborg.
  5. I love this quote – Assemblages make sense of a heterogeneous jumble of infinitely recombinable parts, not just semiotically but functionally.
  6. Mediation – Mediation is the transference of anything that happens when information is exchanged between ANs.  In this sense, agency then becomes distributed because the informant for that agency is the result of the infinite chain of mediators that assemble that agency – otherwise known as distributed cognition.
  7. Translation – This is the tracing of how power applies to change.  In explaining an actants ontology, Spinuzzi notes that, “We can see why translation means transformation:  cascades of intermediaries, including representations, transform actants in ways that facilitate this compromise work.  ANs, then, represent standing sets of transformations – although in a rather different sense than activity networks do (88).  Translation is comprised of four parts:
    1. Problematization – What must be accomplished or negotiated?
    2. Interessement – What stakeholders are involved in the negotiation?  Interessement knots together the different actants.
    3. Enrollment – How do the stakeholders relate and negotiate in order to be mobilized?
    4. Mobilization – How can stakeholders be persuaded to link up and accomplish objectives?
  8. Composition – This is the result of a successful act of translation on a network of A-Ns.
  9. Reversible Black Boxing – This is what happens when the walls of the box are torn down in an actor network.  It is through the process of black boxing that associational/relational dialecticism goes where material/cultural dialecticism cannot – opening up the Black Box into a multiplicity of Pandora’s box.

Despite the differences, Spinuzzi identifies a lot of commonalities  between the two theories.  He notes:

They are both monist, materialist approaches to understanding activity.  They both applied to technical mediation.  They both theorize mediated activities in terms of networks, posit multiplicity within those networks, and allow for different operant social languages in different parts of these networks.  They both see networks as heterogeneous, multiply linked, transformative, and black-boxed.  They do represent points at which the approaches can inform each other. (95)

Chapter Four:  How Are Networks Historicized?

In this chapter, Spinuzzi traces the historical development of a central clause in mass telephony: Universal Service.  First S. traces the “history” of the term.  It’s a bit confusing because the history seems a lot like the activity theory account.  Anyhow, the main premise is that the universal service clause has undergone three iterations:

1.  Universal Service as the Principle of Interconnection – This iteration was the first.  It stated that universal service was necessary because the negotiation of all the different networks (can’t call A if you’re on Bs network) was unwieldy for the consumer.  In its stead, one large long-distance network should unify the disparate systems.

2.  Universal Service as Total Market Penetration – This iteration posited that Universal Service was a right that should be extended to all human beings.  Much like electricity or water, the telephone underwent a democratization.  In other words, using appeals to the government and the citizenry, MaBell turned itself from a marketable service to a public utility.  This also coincided with a belief in the natural monopoly of telephony.

3.  Universal Service and Universally Obtainable Slates of Services – This current iteration turned the simple right of telephony to a right to access numerous communication technologies.  What is scary here (or not) is the idea that access to up-to-date telecommunications services is “now the material basis for individuals’ effective participation in a democratic society” (108).  Hence the DANGER alluded to in Chomsky’s work on the media.

After tracing these three iterations of Universal Service, Spinuzzi performs an Activity theory and an ANT account of how this clause underwent it’s many iterations.  I won’t get into the analysis in depth here, but suffice it to say that the AT account traced the development of the universal service clause as a series of contradictions whereas the ANT account traced the associations that negotiated the universal service clause.  Here’s the breakdown that mirrors the historical development provided above (118-131):


1.  Contradiction 1 : Exclusivity or Interconnection?

2.  Contradiction 2:  Business or Public Utility?

3.  Contradiction 3:  Competition or Public Good?


1.  Translation 1:  From Disunity to Unity

2.  Translation 2:  From Unity to Universality

3.  Translation 3:  From Universality to Rising Tide

The value of ANT in this case is, according to S., “ANT decouples actants from an evolving object and opens the possibility of seeing new actants emerge through negotiation – even overlapping actants.  By keying actants to interests rather than objects ANT embraces the interlinked (interpenetrated, overlapped, multiroled, spliced) activity rather than trying to put it back in the box, as activity theory tends to do.  And by seeing history in terms of political-rhetorical settlements and negotiations, the translation account opens up the possibility of examining phenomena nondevelopmentally” (131).

Chapter Five:  How Are Networks Enacted?

Before S. attempts to answer this question, he describes the transfigurations that have occurred in the period between the waning of the industrial age and the development (or negotiation) of what Castells has called “informational capitalism.”  This new brand of capitalism is different from the modular form that Marx envisioned in that the “deskilling” that occurs when tasks are “broken down into easily learnable and repeatable components” is challenged.  No more assembly lines and workers who can’t see the final products.  Rather, in information capitalism the complete net work is interpenetrated, deeply rhizomatic: “it has multiple, multidirectional information flows” (137).  Because of this characteristic, some folks claim that capitalism will move toward a more distributed form.  Distributed capitalism will come to look a lot like shareholders in companies – distributed, desires for “unique support” from vendors, and trustworthy relations among consumers (think’s comment function).  This process of co-configuration – whereby producer and consumers configure one another at all times reciprocally – will disrupt supply chains and create “advocates” or “professional relationship workers” who “assemble temporary ‘federations’ of suppliers for each transaction or service.  In effect, the layer between producer and consumer will have an individualized shim.  While these new ways to describe capitalist paradigms in the information age could be positive, they also have a negative side.

Net Work and the Informatics of Domination:

In this section, S. argues that while the 20th century was a hundred years of Foucauldian discipline, the next 100years (21 century) will be a period of control.  In a particuarly illustrative and LOVELY meditation on this subject S. notes that

In shifting from monodirectional to multidirectional information flows and from limiting to proliferating links among heterogeneous entities, net work has shifted from the Panopticon to the Agora, which is to say, from surveillance by an authority figure to mutual, distributed surveillance and critique” (142).

Drawing on Sless, S. also characterizes this new capitalism thusly:

In net work, digital technologies play a vital role in forming, interconnecting, and even dispersing nodes; consumption is individuated as the desire for unique identities and unique experiences [thinking of Thomas Frank here]; relationships between customers and businesses become more important, even  as the distinctions between them become unclear; and customers look for stable beneficial relationships among consumers, and producers that support these individual experiences” (143).

Three Senses of Texts:

  1. Texts are inscriptions – According to Callon, this means that texts are “relatively immutable media that resist transport” (145).  In other words, to reference Latour, texts are immutable mobiles – “referential inscriptions that can circulate from one locale to another while resisting deformation” (146).  This is why they are so great for codifying.  Even if a text is inscribed in different ways, if those inscriptions are consistent, the multiplicity can be stabilized. . . the text can essentially be a consolidator.
  2. Texts belong to genres – If inscriptions are a way to fix record and dominate phenomena by capturing representations, then genres are the woven or developed inscriptions that  that consistently respond to specific situations.  They are also the “spliced or hybridized” texts that adapt to local conditions and intersecting activities (146).  In this sense then, S. is defining genre as a behavioral instead of a structural construct – a “tool-in-use” (147).  Finally, genres provide a “stability with flexibility.”  Because they are developed among associations, the long-lasting genres provide stability to the network; however, because they are substitutable in different situations, they are really flexible.
  3. Texts function as boundary objects.  This is directly related to genres being both flexible and stable.  As S. notes, boundary objects are often an assemblage of related texts (inscriptions, genres) that collectively plays different roles in overlapping activities.  In this sense then, the boundary object acts as ANTs negotiator – the boundary object is the mobilizer for some actions.

After tracing how net work works by following an: order, money, substitutions, and workers, S. notes how the process of following these actions reveals how “genres circulated as boundary objects that both wove and spliced functional units” (171 – for more detailed information, see examples 148-170).  As workers picked up the genres and social languages of other areas (because Telecorp IS a boundary), they were able to learn. . . which is the focus of the next chapter.

Chapter  Six:  Is Our Network Learning?

In this chapter S. discusses how the nature of work has changed fundamentally in the age of informational capitalism.  By referring to workers as “deskilled” (Haraway), “dividuals” (Deleuze), “reskilled” (Castells), and “lifelong learners” (Zuboff and Maxmin), S. points out that the worker in the informational age will be in a constant state of negotiating different tasks and demands.  This argument is laid out in more detail at the beginning of Chapter 5.  Anyhow, after illustrating that Telecorp’s problem is that they are training (if and when at all) on a modular production model, S. argues that they must do more associational – or rhizomatic – training in the future.  Instead of vertical integration models of learning (it all comes from the tops down), what needs to happen more is horizontal learning practices – learning predicated on associations.  While S. criticizes the formal training, apprenticeship and self-learning at Telecorp, he recognizes that learning did occur – to some extent.  He categorizes that learning in the same way that he characterizes networks in chapter 2.

  1. Heterogeneous – Workers at Telecorp were able to juxtapose different things – humans, individuals, nonhumans, groups, tools, belief, etc. – into assemblages that collectively perform activities.  They even learned in a heterogeneous manner through heterogeneous genres and tools.
  2. Multiply linked – Because of all of the hidden passages and non-official pathways at Telecorp, learning happened through multiple links to multiple people and technologies.
  3. Black-boxed – There was a problem with black-boxing at Telecorp.  Because the various assemblages wherein work occurred had so many facets and were changing so often, the information to be translated between actants became too idiosyncratic and specific.  What was really needed to get folks working associationally in order to traverse the assemblage effectively was more horizontal learning in confidence-building and negotiation.
  4. Transformative – Transformations are central to any network because they allow for new pressures (AN) to be assembled.  While Telecorp did a good job transforming texts, they did not provide for a way to transform their workers so that the workers themselves could also be more capable of working through the associations and away from modularities.

Conclusion:  How Does Net Work Work?

S. frames this question around the tension that he has been riffing on throughout the book: dialectics vs. rhizomatics.  In this conclusion he identifies the four components of the network that have been illustrated throughout the book:  heterogeneous, multiply linked, transformative, black-boxed.  In trying to account for what to DO with all of this information,  S. says there are implications for three groups: workers, managers, and researchers.

a.  Workers:

  1. 1.  Rhetoric – they need a better handle on rhetoric to negotiate the boundary territories.
  2. 2.  Time Management – Because the work is so heterogeneous, workers must develop new time management skills – integrating genres and rules to create stable transformations.
  3. 3.  Project Management – Because of the border work, groups must become more porous and collaborative.
  4. 4.  Adaptability – “Being on the border means having to learn horizontally as well as vertically, having to understand others’ work and social languages and genres, having to forage expertly for information” (202).

b.  Managers:

  1. 1.  Black-boxing – Managers need to open up black boxes to make communication associative and rhizomatic.  To do this, they should use liaisons (folks that create connections across the boundaries), APIs (routines, protocols, and tools that allow simple interactions to generate complex effects), and Aggregations (bottom-up characterizations of large sets of data – like tagging).
  2. 2.  Strategic Thinking – Provide a persuasive vision for workers
  3. 3.  Training – Horizontal as well as vertical.

c.  Researchers:  don’t bind the case and don’t look for a context or frame of reference, embrace the multiplicity and allow the researched to characterize themselves and the boundaries of their work.

S. closes by noting how Bakhtin’s dialogism could be just the thing that AT is looking for.  To compensate for ATs dialectical shortcomings (developmental emphasis – north to south), dialogism could reveal that rhetoric is always at work in net work.  It also provides a way to account for the multiplicity of research subjects ontologies.

3 Responses to “Network – Spinuzzi”

  1. Dias et al & Spinuzzi « Queer/strokes/

    […] Justin’s blog post. […]

  2. Missy

    These notes are incredibly helpful. I find myself indebted to you again and again :).

  3. justin lewis | justin lewis

    […] genres reads thusly:  the levels of abstraction dictate the level of activity vis-a-vis Spinuzzi (see here).  As such, this reading of genre might correlate to the different kinds of […]

Leave a Reply