Friday, May 20, 2005

Reading :: Pandora's Hope (Supplemental Notes)

Originally posted: Fri, 20 May 2005 09:22:28

Pandora's Hope: Essays on the Reality of Science Studies

by Bruno Latour

Pandora's Hope was one of the first books I reviewed on this blog. It's a fairly short review, partially because I was still trying to absorb the basics of actor-network theory at the time. This second reading comes after several other books and articles on ANT. So I'll be a bit more detailed in my comments this time.

Latour begins with the question of what science actually is, framing it with the story of an encounter with a scientist who nervously asks him if he believes in reality. Of course he does, Latour says. The rest of the book is a discussion of what that means -- a difficult proposition, Latour maintains, because of the way the world has been divided between subjects and objects. This modernist settlement has led us away from commonsensical statements and to a point at which we must discuss whether we believe in reality in the first place!

To illuminate, Latour takes us in Chapter 2 to Brasil, where he follows an interdisciplinary set of scientists who are examining the dividing line between a savannah and a forest. How is their work done? Through a set of circulating references, Latour says: representations that themselves become represented by other representatives. "A word replaces a thing," he notes, "while conserving a trait that defines it" (p.63). The inscriptions that science mobilizes together constitute long chains of representation that typically converge into something about the size of a tabletop -- whether the thing that is represented is a galaxy, a microbe, or the soil underneath a savannah. And this chain renders the subject-object distinction useless -- and a bit ridiculous, in Latour's view. "Knowlendge, it seems, does not reside in the face-to-face confrontation of a mind with an object, any more than the reference designates a thing by means of a sentence verified by that thing." Rather, it resides in the long cascade of re-representations.

An essential property of this chain is that it must remain reversible. The succession of changes must be traceable, allowing for travel in both directions. If the chain is interrupted at any point, it ceases to transport truth -- ceases, that is, to produce, to construct, to trace, and to conduct it. The word "reference" designates the quality of the chain in its entirety, and no longer adequatio rei et intellectus. (p.69)

The principle is nicely illustrated on p.71, in which "the transformation at each stage of the reference ... may be pictured as a trade-off between what is gained (amplification) and what is lost (reduction) at each information-producing step." In the diagram, successive steps lead to narrower locality, particularity, materiality, multiplicity, and continuity, but greater compatibility, standardization, text, calculation, circulation, and relative universality -- that is, the transformation leads us to representations that are less rich but more calculable and transportable. Transformations lose some properties but gain others. Conducting a phenomenon in this way, in successive layers, makes it more real (p.76). It is the erasure of these intermediate steps that causes us to believe in the subject-object distinction.

With that in mind, we go to Chapter 3, in which Latour discusses translation by using the historical case of Joliot attempting to develop an artificial nuclear chain reaction. Such a feat required a long set of translations: "when their goals are frustrated, actors take detours through the goals of others, resulting in a general drift, the language of one actor being substituted for the language of another" (p.89). Dautry's goal, national independence, can be reached only by taking a detour through Joliot's goal, being the first to master a chain reaction. Joliot's goal can only be reached by building a laboratory that can help to reach both goals. These detours link together very different sorts of enterprises, resulting in an interesting drift as the chain of translations expands, a drift from the exoteric to the esoteric. Some of the actors are humans, like Dautry and Joliot, but others are nonhumans, such as neutrons that must be persuaded to slow enough to hit a uranium atom. "For Joliot, it wasn't very difficult. In the morning he dealt with the neutrons and in the afternoon he dealt with the minister. The more time passed, the more these two problems became one" (p.90).

Based on this work, Latour proposes that scientific facts are created and maintained through a "circulatory system" with five loops: mobilization, autonomization, alliances, public representations -- and the loops that knot them all together, the "links and knots" that are typically considered the scientific concept. In the modernist subject-object understanding, this last loop is conceived as an inner core surrounded by context; in the amodernist understanding being promoted here, it is one loop among others, moving and changing rather than transcendant and static. "A concept does not become scientific because it is farther removed from the rest of what it holds, but because it is more intensely connected to a much larger repertoire of resources" (p.108).

This leads us to Chapter 4, where Latour returns to Louis Pasteur's work to suggest that we stop worrying about whether facts are constructed or real. Of course they are both -- and the shifting out of one plane of reference to another is what makes this doable (p.130). Latour suggests that instead of thinking of the world in terms of statements that allow words to correspond to the world (i.e., bridge the subject-object divide), we turn to a model of propositions in which we examine which propositions are articulated with one another (p.142).

Chapter 5 takes up a common criticism of ANT, which is its lack of historicity. Following his standard pattern, Latour outlines two understandings of historicity, using the illustration of the microbe. "Did ferments exist before Pasteur made them up?" (p.145). If the question is in terms of representations, the answer is no: the notion of the microbe didn't come into being until Pasteur. If the question is in terms of evolution, the answer is yes: microbes were really there, but hadn't been named or theorized. But this is the words-and-world, subject-and-object distinction all over again: "the divide between what pertains to human history and what to natural history would not have been bridged in the slightest" (p.146). To overcome the distinction, Latour turns back to the notion of circulating reference from Chapter 2. Every change in the series of transformations makes a difference (p.150); "Associations of entities have a history if at least one of the articles making them up changes" (p.152). So an amodern account of historicity is grounded in changes in associations. And such a historicity is reversible; no inertia is enough to keep up the reality that has been so difficult to produce" (p.155). Settlements can unravel. Scientific facts have relative existence (p.156), an existence that takes hard work to keep up. Victories always require further actions (p.168). And now we get to another of Latour's diagrams (p.171), this one illustrating an amodernist historicity of Pasteur's work. The X axis is the "linear succession of time," while the Y axis is the "sedimentary succession of time" -- the year 1864 as it is represented in subsequent years. The 1864 of 1864 was not the same as the 1864 of 2005: in the earlier sedimentation, spontaneous generation was a serious theory that had to be addressed with careful methodology, whereas in the later sedimentation, we don't consider spontaneous generation seriously. The concepts become rearticulated and reread in later history. (From my own research: the concept of universal service became rearticulated in the 1970s, then was read back into Theodore Vail's 1907 statement on universal service as well as the Communications Act of 1934, neither of which originally shared the 1970s-era understanding of universal service.)

Now for Chapter 6, in which Latour talks about mediation. How do we deal with technical mediation and the questions it poses for agency? Do guns kill people or do people kill people? Latour sees the subject-object dichotomy in this question as well, and characteristically proposes a middle ground through his account of technical mediation. Mediation, he says, involves mutual transformation of the assemblage (in this case, the gun-human), and this transformation has four parts:

First, translation: "we can portray the relation between two agents as a translation of their goals which results in a composite goal that is different from the two original goals" (p.179).

Second: Composition. "Action is simply not a property of humans but of an association of actants" (p.182). This composition of different agents in an assemblage is enacted through a series of "subprograms" that together achieve the composite goal of the assemblage.

Third: Reversible blackboxing. An assemblage (composition) of agents can be "blackboxed" or regarded as a single entity. But if the settlement breaks down, the black box can open; the composition can unravel as easily as it was raveled. In Aramis, for instance, Latour demonstrates how engineers labor to turn a field, a set of motors, a set of citizens, and innumerable other humans and nonhumans into a public transportation system. If blackboxing works, you can say, "I'm taking the train to the meeting." If the black box is opened, you might say, "the drivers are on strike" or "there's something wrong with the engine." Technical and scientific settlements need to be renegotiated and supported as much as political ones do. (pp.183-185).

Fourth: Delegation. The illustration here is the "sleeping policeman" or speed bump. "The driver's goal is translated, by means of the speed bump, from 'slow down as not to endanger students' into 'slow down and protect your car's suspension.' ... The driver's first version appeals to morality, enlightened interest, and reflection, whereas the second appeals to pure selfishness and reflex action. ... But from an observer's view it does not matter through which channel a given behavior is attained" (p.186). The nature of "meaning" is itself modified through delegation (p.187). "The speed bump is a meaningful articulation within a gamut of propositions" (p.187).

Given these four meanings or aspects of mediation, Latour attempts to trace "how a collective of one given definition can modify its makeup by articulating different associations": through translation, then crossover (i.e., humans and nonhumans exchange properties), then enrollment, mobilization, and finally displacement ("the direction the collective takes once its shape, extent, and composition has been altered by the enrollment and mobilization of new actants" (p.194). This part confuses me because according to one of Michel Callon's most quoted articles, translation occurs in four moments: problematization, interessement, enrollment, and mobilization. Is Latour using a more narrow meaning for translation here? Or different meanings for the other terms? I really wish he had traced the genealogy of the term here and related it more strongly throughout the chapter, since Callon's piece is much more established.

In any case, we now get to Latour's two-chapter discussion of the Gorgias, Plato's famous Socratic dialogue in which Socrates gets to beat up on rhetoric. As a rhetorician, I am not very happy with the two-dimensional portrayal of rhetoric. Neither is Latour. In the dialogue, Socrates represents Right: the lone man with knowledge and virtue on his side, who invokes absolute truth and declares that he will be vindicated in the afterlife. Callicles represents Might: the strongman who commands the dumb masses through force of personality, ruthlessness, and deception. Callicles is the very image of the Prince, at least the Prince that we have come to expect from the many slanders leveled against Machiavelli. But 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: the citizens in the marketplace, bargaining, negotiating, persuading, and trading.

It's worth noting that just as he did in Aramis, Latour takes a little time here to distance himself from Machiavelli. He warns against accepting "the Machiavellian definition of politics as being unconcerned with morality" and disapproves of "what Machiavelli will later esteem as a positive definition of political cleverness --- although Machiavelli's position is, of course, not a wholly immoral one" (p.253). Later, he says that "Machiavelli fell into Socrates' trap and defined politics as a cleverness entirely freed from scientific virtue" (p.263). Notice two things here, though. One is that Latour is not repudiating Machiavelli entirely, just Machiavelli's acceptance of the Socrates-Callicles settlement -- and in that way Machiavelli is no different from any modernist. The other is that Latour is talking about Machiavelli in the context of the Socrates-Callicles debate after emphasizing that the Callicles in this dialogue is a straw Callicles, distinct from the historical Callicles, who would have surely said something different. Is this a clever way to get across the difference between a straw Machiavelli and the actual Machiavelli, as he did in Aramis? Or am I trying too hard here?

In any case, Latour says that "power and reason are one and the same" -- that Might vs. Right is an illusory dichotomy and that "the Body Politic built by one or by the other is shaped with the same clay" (p. 262).

This brings us to Chapter 9. Latour talks about "factishes." "Fact" and "fetish" come from the same root word but connote opposite meanings, Latour contends, because of this modernist settlement that has split subject and object (p.267). He tries to reunite them as a "factish," something that, if fabricated well, allows reality to be autonomous. The question is not whether facts are real or constructed, "the Procrustean bed in which the modernist settlement wants us all to slumber" (p.275). They are both! And here's where Latour does another one of his hit-and-run jobs against dialectic:

When a fact is fabricated, who is doing the fabrication? The scientist? The thing? If you answer "the thing," then you are an outdated realist. If you answer "the scientist," then you are a constructivist. If you answer "both," then you are doing one of those repair jobs known as the dialectic, which seem to patch up the dichotomy for a while, but only hide it, allowing it to fester at a deeper level by turning it into a contradiction that has to be resolved and overcome. And yet we have to say that it is both, obviously, but without the assurance, certainty, or arrogance that seems to go with the realist or the relativist answer or with a clever oscillation between the two. (p.281)


Am I simply restating the dialectic? No, there is no object, no subject, no contradiction, no Aufhebung, no mastery, no recapitulation, no spirit, no alienation. But there are events. I never act; I am always slightly surprised by what I do. (p.281)

This is a fairly abbreviated critique of dialectic, isn't it? But at least it's a little more coherent than his previous attempts. On the other hand, Latour's critics have been guilty as well, and that has agitated him:

If I have not answered the science warriors' arguments term for term -- or even cited their names -- it is because the science warriors too often waste their time attacking someone who has the same name as mine, who is said to defend all the absurdities I have disputed for twenty-five years: that science is socially constructed; that all is discourse; that there is no reality out there; that everything goes; that science has no conceptual content; that the more ignorant one is the better; that everything is political anyway; that subjectivity should be mingled with objectivity; that the mightiest, manliest, and hairiest scientist always wins provided he has enough 'allies' in high places; and such nonsense. I don't have to come to the rescue of that homonym! (p.300)

That Latour sounds familiar too. I recognize him from Miettinen's criticisms, and Engestrom's, and others. >

Blogged with Flock

Wednesday, May 18, 2005

(Reading Roundup: de Laet and Mol)

Originally posted: Wed, 18 May 2005 18:05:47

de Laet and Mol's "The Zimbabwe Bush Pump: Mechanics of a Fluid Technology" represents still another example of how actor-network theory (or, I suppose, one postvariant) has pushed from science studies to studies of technologies. Like Madeleine Akrich's study of the gazogene, this study examines how technologies are not contextualized but actually constituted by their relationships with local conditions, terrains, practices, and so forth. I like it very much. The ZBP, in this account, is articulated with a number of different actants to make it successful. One striking example is that the ZBP's documentation stresses the importance of consulting with local water diviners before siting a water hole (p.234) -- a beautiful illustration of how multiplying associations can produce a more durable and far-reaching settlement than otherwise.

And the ZBP settlement is quite durable. Unlike Akrich's gazogene, the ZBP is a runaway success: "A national standard, the Zimbabwe Bush Pump is a nation-builder that gains strength with each illustration" (p.236). Unlike Engels, who insisted that 100,000 steam engines did not prove Carnot's principle more than the one, de Laet and Mol boldly insist quite the opposite. The more ZBPs, the more real the ZBP is. Whereas Engels thought the steam engine could be reduced to a "germ-cell," a minimal abstract representation of the essential principle behind the steam-engine, de Laet and Mol insist that the ZBP has fuzzy boundaries:

Even if nothing can be taken from it, it is not clear where this pump ends. For what is the Zimbabwe Bush Pump? A water-producing device, defined by the mechanics that make it work as a pump. Or a type of hydraulics that produces water in specific quantities and from particular sources. But then again, maybe it is a sanitation device -- in which case the concrete slab, mould, casing and gravel are also essential parts. And while it may provide water and health, the Pump can do so only with the Vonder Rig -- or some other boring device -- and accompanied by manuals, measurements, and texts. Without these it is nothing, so maybe they belong to it too. And what about the village community? Is it to be included in the Pump --- because a pump has to be set up by the community and cannot be maintained without one? But then again, perhaps the boundaries of the Bush Pump coincide with those of the Zimbabwean nation. For in its modest way this national Bush Pump helps to make Zimbabwe as much as Zimbabwe makes it. (p.237)

If this discussion sounds familiar, perhaps you're thinking of Gregory Bateson's discussion of the blind man with the cane. Engels was searching for a way to generate abstract statements about essences while remaining materialist; de Laet & Mol, Bateson, Deleuze & Guattari, Latour, etc. were looking for a way to define technologies relationally and nonessentially. Partially this is because their different projects demand different tacks. Engels and his activity theory descendants were interested in learning and development, so they were concerned with how to describe concept formation in all its manifestations; they were interested in how abstractions were developed over time and made into mediators that guided further activity. Bateson and his actor-network theory associates were much more concerned with ontology, and found abstract essences to be quite untenable as ways to address that problem.

Back to de Laet and Mol. Although most of the article was quite good, I was disappointed to see that the Machiavellian slur was leveled (or at least entertained) here as well. Critics have argued that in invoking Machiavelli, actor-network theorists have created an incompatibility between the principle of symmetry that they hold and the extreme asymmetry of a Machiavellian viewpoint. de Laet and Mol give those critics comfort here:

For even if Latour's work shifts Pasteur out of the centre by pointing to the network he needs, it also suggests (or has been read as suggesting) that innovation, even if it turns out to be the work of a large army, does need a general in order to spread out. This Machiavellian reading of Latour says that technologies depend on a power-seeking strategist who, given a laboratory, plots to change the world. (p.227)

But neither Latour nor Machiavelli subscribe to this view! With friends like these, who need enemies?


Blogged with Flock

Tuesday, May 17, 2005

(Reading Roundup: Engestrom on Expansive Learning at Work)

Originally posted: Tue, 17 May 2005 18:49:15

Engestrom's "Expansive learning at work: Toward an activity theoretical reconceptualization" (Journal of Education and Work 14.1, 2001) is very similar to some of the other Engestrom articles and chapters I've reviewed recently. In fact, it shares some passages with "Developmental work research as educational research." But it covers some different ground as well.

For instance, in its brief history of activity theory, it acknowledges the "deep-seated insensitivity of the second generation activity theory toward cultural diversity" originally pointed out by Michael Cole (p.135), and says that the internationalization of AT meant that "questions of diversity and dialogue between different traditions or perspectives became increasingly serious challenges" (p.135). Ah. So is this why Bakhtin's dialogism has become so interesting to third-generation activity theorists?

Engestrom then lays out the working of an activity network, including a diagram on p.136 in which the objects of two activity systems form a third object -- a Venn diagram portraying the intersection of the separate objects. He differentiates among

- "object 1" ("an initial state of unreflected, situationally given 'raw material'" such as "a specific patient entering a physician's office")

- "object 2" ("a collecively meaningful object constructed by the activity system" such as "the patient constructed as a specimen of a biomedical disease category and thus as an instantiation of the general object of illness/health")

- "object 3" ("a potentially shared and constructed object" such as "a collaboratively constructed understanding of the patient's life situation and care plan")

"The object of activity," he concludes, "is a moving target, not reducible to conscious short-term goals.

Given the understanding of the object, Engestrom summarizes AT in terms of five principles:

- The activity system "seen in its network relations to other activity systems" becomes the unit of analysis (p.136).

- Activity systems are multivoiced: "the activity system itself carries multiple layers and strands of history engraved in its artifacts, rules, and conventions. The multi-voicedness is multiplied in networks of interacting activity systems. It is a source of trouble and a source of innovation, demanding actions of translation and negotiation" (p.136).

- Activity systems have historicity: they "take shape and get transformed over lengthy periods of time." "History itself needs to be studied as local history of the activity and its objects, and as history of the theoretical ideas and tools that have shaped the activity" (pp.136-137).

- Contradictions are central to change and development. "Contradictions are historically accumulating structural tensions within or between activity systems" (p.137).

- Activity systems undergo expansive transformations in which "the object and motive of the activity are reconceptualized to embrace a radically wider horizon of possibilities than in the previous mode of the activity" (p.137). Engestrom links expansive transformations to the zone of proximal development.

Engestrom uses these five principles as the columns in a matrix; the rows are the questions:

- Who are learning?

- Why do they learn?

- What do they learn?

- How do they learn?

The matrix turns out to be a useful heuristic for examining learning, which is AT's main focus, at least as it has tended to be conceptualized by Vygotsky and Engestrom. In fact, the question of learning is so important for Engestrom that he begins to see it everywhere, and assert it with such assurance that I begin to doubt my own memory. Does Engestrom really say that "Latour's (1987, 1996) actor-network theory recommends that we locate learning in a heterogeneous network of human and non-human actors"? Why, yes he does, here on p.140. Does Latour ever mention the term "learning" or a related concept? No, not really, not unless you locate learning as central to collective activity. Latour does not. In fact, from what I can tell, Latour remains agnostic about individual learning, and is much more interested in the forming of settlements -- which does not necessarily include learning or subscribe to a theory of learning at all!

To Engestrom, studying knowledge formation or the sociology of scientific knowledge implies a theory of learning; to Latour, it does not. That's because the two have very different projects with very different emphases and origins. For instance, look back at the history of activity theory and you'll see a progressive outgrowth from the study of individual cognition (Vygotsky) to larger structures. As you can see from the matrix mentioned above, the individually oriented concepts have turned into concepts for studying overall activity. Expansive visibilization, for instance, is simply the zone of proximal development applied to activities rather than individuals. Compare that with actor-network theory, which has its roots in political and economic theory and sociology, with an assist from philosophy and rhetoric -- all of which are concerned with group or societal relations rather than individual relations. ANT doesn't build on the notion of an individual to reach its understanding of human activity; it tracks movements, not learning.

Let's end with something that has been interesting me quite a bit lately: The concept of the germ-cell. In Learning, Working, and Imagining, Engestrom discusses Frederich Engels' example of Carnot's theorization of the steam engine. From Engels' Dialectics of nature, p.229: "the steam engine provided the most striking proof that one can impart heat and obtain mechanical motion. 100,000 steam engines did not prove this any more than the one?>

Blogged with Flock

Monday, May 16, 2005

(Reading Roundup: Engestrom on co-configuration work)

Originally posted: Mon, 16 May 2005 19:22:33

Last week I read a few more articles meant to help me understand the relationships between activity theory and actor-network theory. Things have bogged down a bit as I've been trying to digest these many articles as well as well as rereading Latour's Pandora's Hope, one of the first books I reviewed on this blog. I'll have to post supplemental notes on it soon, since the first review really doesn't do it justice. But in any case, let's look at the article in today's reading roundup:

Engestrom, Y. "New forms of learning in co-configuration work."

This isn't really an article -- it's a paper that Engestrom presented at a seminar in 2004. Although it's only ten pages long, it's full of insights.

Engestrom starts with Victor and Boynton's "historical framework for the reintegration of organization, work, and learning" in which they "identify five types of work in the history of industrial production: craft, mass production, process enhancement, mass customization, and co-configuration" (p.1). (I haven't read Victor and Boynton's book, but it sounds of a piece with Zuboff and Maxmin's.) Engestrom's Figure 1 helpfully depicts these types of work, along with descriptions of each. Under co-configuration, the description is "Dialogical configuration knowledge."

The question, then, is one that Engestrom has become increasingly interested in addressing: how do we account for work-based learning in what some gloss as the "new economy" or the "knowledge economy" or "distributed capitalism" -- or if you prefer, the "control society" or the "informatics of domination"? Engestrom is clearly positioning third-generation activity theory for the challenge, and the term "dialogical" signals that he sees Bakhtin's dialogism as continuing to be a key part of adapting this third-generation AT appropriately.

What is co-configuration work? Engestrom quotes Victor and Boynton:

The work of co-configuration involves building and sustaining a fully integrated system that can sense, respond, and adapt to the individual experience of the customer. When a firm does co-configuration work, it creates a product that can learn and adapt, but it also builds an ongoing relationship between each customer-product pair and the company. ... a living, growing network develops between customer, product, and company. (Victor & Boynton 1998, p.195)

So the customer becomes a "real partner" with the producer; producers form "a strategic alliance, supplier network, or other such pattern of partnership" (p.2). And here Engestrom firmly pegs this work to knotworking. (I'm starting to see not just a trend but also a strategy on Engestrom's part, a strategy to reposition developmental work activity as a player in the new economy. I'm also starting to see why Engestrom wants to pick off actor-network theory as a potential rival in this area.) And Engestrom adds:

A precondition of successful co-configuration work is dialogue in which the parties rely on real-time feedback information on their activity. The interpretation, negotiation, and synthesizing of such information between the parties requires new, dialogical and reflective knowledge tools as well as collaboratively constructed functional rules and infrastructures. (p.3)

That reflective work happens across as well as within activities (p.4), meaning that the next generation of expansive learning "puts the horizontal and inter-organizational dimension of learning in the center" (p.4). Yet expansive learning is still taken as "a sequence of learning actions ascending from the initial abstract 'germ cell' to the concrete whole of the system to be mastered" (p.4) -- what Engestrom has often called "ascending from the abstract to the concrete." The notion of a germ cell is a bit questionable in my mind given the apparent reliance on linearity of development (common to dialectic), but I need to do much more thinking on the subject.

In any case,

learning in co-configuration settings is typically distributed over long, discontinuous periods of time. It is accomplished in and between multiple loosely interconnected activity systems and organizations operating in divided local and global terrains and representing different traditions, domains of expertise, and social languages. Learning is crucially dependent on the contribution of the clients or users. Learning is embedded in major transformations, upheavals, innovations, implementations, and movements. It takes place in heterogeneous patchworks and textures of small and large, unnoticeable and spectacular actions, objectifications, trajectories, and trails. (p.5)

Engestrom proposes that "the expansive learning required and generated by co-configuration work" has three features: It is transformative (broadening the shared objects of work with new and adapted tools); horizontal and dialogical ("crossing boundaries and tying knots between activity systems operating in divided multi-organizational terrains" (p.5)); and subterranean (blazing "cognitive trails" that constitute the "invisible, rhizomatic infrastructure of new forms of expansive learning at work" (p.5)). The last one is a bit puzzling, since "rhizomatic infrastructure" seems to be a contradiction in terms, but it's interesting to see how Engestrom is pullling other perspectives into activity theory. Engestrom is a great synthesizer, and sometimes I think his reach exceeds his grasp, but he certainly has developed activity theory quite a bit through this synthesis.

Brilliant stuff here, but then Engestrom has to follow up with a blatantly incorrect critique of actor-network theory: "there is a risk in focusing exclusively on actors. The professionals and their discursive interactions may appear as somewhat omnipotent constructors of their activities and social worlds. From the point of view of activity theory, this would mean that the material grounding and stubborn system dynamics of practical activities are lost or ignored, the resistance of objects is forgotten" (p.6). But this is a description of constructivism, not actor-network theory, and Latour makes a nearly identical argument against constructivism in Pandora's Hope! How often Latour and his colleagues have made this same point, clarifying that "actors" are not just humans or professionals, but also nonhumans or objects that continually resist! How often actor-network theorists argue that we are not omnipotent constructors, not by a long shot! Not to mince words: This is just a sloppy critique, not up to the quality of the rest of the piece.

So let's get back to the rest of the piece. Engestrom argues that rather than following the actors, we should follow the objects of work activity. Professional work is socio-spatially distributed, "forming long chains of interconnected practical and discursive actions" (p.6). The objects of work are "heterogeneous and internally contradictory," enduring, and constantly reproduced (p.6). The object is the key to understanding change and learning, and "expansive learning is above all stepwise explanation of the object" (p.6). so Engestrom suggests that we

(1) follow the objects of co-configuration work in their temporal and socio-spatial trajectories, (2) give the objects a voice by involving the clients or users in dialogues where the object is negotiated, (3) expand the objects by organizing intervention sessions where the producers and clients construct new shared models, concepts and tools to master their objects. (p.7)

All right, although "give the objects a voice" sounds more like giving the participants a voice; I wish Engestrom would be more careful to delineate what he's doing from ANT's related project.

Blogged with Flock

Reading :: Learning Challenges in Organic Vegetable Farming

Originally posted: Mon, 16 May 2005 18:13:01

Learning Challenges in Organic Vegetable Farming: An Activity Theoretical Study of On-Farm Practices

by Laura Seppanen

A while back I planned to read a set of dissertations that had come out of the Center for Activity Theory and Developmental Work Research at the University of Helsinki, starting with Mervi Hasu's. Other readings intervened, so I've only now come to the second. Like Hasu's, Seppanen's dissertation is a "wrapper" that contextualizes and pulls together several published single-author and multi-author articles. The text itself is only 74 pages, but it refers to the six other articles, indicating an impressive amount of work.

I confess that I skimmed quickly through the dissertation. Much of it involved simply applying activity theory to the activity of organic farming -- a worthwhile exercise, but not one that generated many new insights into activity theory itself. The insight that "learning in organic vegetable farming is a continuous, dynamic process" (p.5) doesn't seem to be terribly surprising (although I thought it was a bit broad of a claim given the scope of the study).

On the other hand, there were some surprises in the way the activity was visualized. For instance, on p.13, Seppanen provides a diagram that visualizes the levels of activity in organic farming. Levels of activity are typically represented textually or in tables, but here Seppanen depicts the levels of activity, action, and operation in what looks like a hierarchical or "tree" graphic; each element on one level is decomposed into many elements on the next level. In addition, she provides an intermediary step between activity and action: "Practices (groups of actions) such as onion planting." The activity of the particular farm is depicted as "embedded" in the general activity. I think that this diagram is useful for thinking about the relationships among these different elements, but it's also misleading in a few key ways.

First, activities are not simply embedded as depicted here, they interpenetrate other activities simultaneously. This is a key tenet of the "third-generation" activity theory being developed at the Center for Activity Theory and Developmental Work Research, and it's also reflected in Seppanen's discussion of the particular organic farmers she studied -- who entered the activity of organic farming late and brought in tools and practices from other activities.

Second, the intermediary step of practices may or may not make sense, but it's really not orthodox activity theory and Seppanen doesn't provide a discussion or cites to justify this change. Perhaps this notion of practice has been picked up by others at CATDWR, but I don't see much evidence of that. I'm not sure how grounded this intermediary step is in that research, but it seems like trying to leverage a notion that is often discussed in related literatures (e.g., Lave and Wenger).

Third, the portrayal of activities decomposing into practices which in turn decompose into actions, etc. is one that, although more or less consistent with activity theory since Leont'ev, overmodularizes the relationships and underestimates the interconnectedness of these relationships. For instance, the same action (say, planting an onion) could help to move along several activities simultaneously; the same operation (say, lifting an onion sack) could intersect more than one action; and so on. I'm reminded of the family farmers in Iowa who often take up other jobs so that they can support the family farm, which remains unprofitable in itself. Each corn seed that is planted is supposed to eventually bring a monetary yield, but it also allows the farmer to continue participating in the family farm, to continue being a farmer in her community of family farmers, to figuratively put her thumb in Monsanto's eye, to (perhaps) draw farming subsidies, to feel closer to the land. These are not only significances but social-material activities, and the modular relations depicted in this figure don't really do them justice.

As you can tell, the dissertation is thought-provoking. My sense is that it was written just as much for agricultural readers as for activity theorists, so it's not surprising that the dissertation doesn't go to heroic lengths to generate new theory; it is primarily meant to recontextualize organic farming with the help of activity theory. That it tends to do quite well, providing another example of how learning-by-expanding can provide insights into developmental work activity.

Blogged with Flock