Thursday, July 09, 2009

What if I had called them "genre networks"?

Just thinking aloud here: What if, instead of calling them genre ecologies, I had called them genre networks?

First, some background. When I was researching software development for Schlumberger Well Services during a summer internship in 1997, I cast around for ways to describe the clusters of information resources that the developers strung together in localized and often idiosyncratic ways. The information resources themselves were easy to characterize - I had been studying genre theory, so it was natural to characterize them as genres. But as the developers used these genres in conjunction, often in nonlinear ways such as juxtaposition and annotation, the clusters yielded sums greater than their parts. And the existing ways of talking about these - genre sets, systems, and repertoires - seemed to assume different things about such linkages, generally more linear things, without the interesting substitutions I had seen there.

I hit on the term "genre ecology" and it stuck. Independently, others had hit on the term at the same time, including Tom Erickson as well as Aviva Freedman and Graham Smart. The ecology metaphor did the work I was hoping it would do, describing the emergent clusters of genre resources as dynamic, localized systems whose components might have developed in other environments before interacting in their current systems. It also emphasized the point that such systems didn't necessarily have to be that way - that these systems were emergent, not following predetermined paths.

The notion of the genre ecology helped me immensely as I began my dissertation research, which I reworked in my first book Tracing Genres through Organizations. In one instance, for example - which became so key that I used it to open the book - a police officer had introduced an improvised, unofficial genre into the ecology, using it to mediate between two official genres that embodied different representational systems and logics. The improvised genre actually substituted for one of the official genres - a map - so much so that she didn't even bother using the map. She actually avoided it. And as I looked at the other incidents in which participants had improvised genres, I could see that these generally introduced more flexibility into the system. The police officer had substituted the unofficial genre for the more cumbersome official one - but she still had both resources. She had built flexibility into the system, but she had not simply traded the old genre for the new; she had redundancy in case she needed it. Yet since each genre has its own "logic," that redundancy isn't simple substitution, it's a redundancy in "logics" as well. The more genres you introduce into the ecology, the more complex the task of managing genres becomes.

By the time I was visiting campuses for job talks in 1999, I was able to talk coherently about how I was mapping these genre relationships via genre ecology diagrams. At one campus visit, I tried to explain these diagrams to a professor from another department. "Oh," he said. "Network diagrams."

I was hesitant to say yes, because I had only a lay understanding of networks at that point. But, yes, the genre ecology diagram is a form of network diagram. It maps out linked relationships among nodes. And a genre ecology diagram, like a network diagram, can show which node is most densely linked and to which others.

So that idea rattled in the back of my mind as I continued developing my scholarship. My next study was a larger one, of a telecommunications company, and although I used an activity theory framework to plan it out, I kept seeing incidents that didn't fit neatly into activity theory. I turned to other frameworks, most importantly actor-network theory, in a detour that took several years and produced my second book - titled Network. Here, I brought together ideas from activity theory, genre theory, actor-network theory, distributed cognition, netwar, and knowledge work, among others. I'm happy with the result, but ever since, I have been wondering: what if I had characterized genre ecologies as genre networks?

This question occurred to me again yesterday, as I read Arquilla and Ronfeldt's collection In Athena's Camp. In netwar analysis - as opposed to actor-network theory - networks are conceived as composed of interlinked nodes that provide alternate arrangements and resources to accomplish a purpose. Netwar is mostly concerned with organization, so these nodes tend to be seen as military units or troops. Such networks can be interconnected in various ways, but the most yield comes from an all-channel network, in which every node is connected to every other. The more interconnections, the easier it is to reroute around a node; to achieve localized efficiencies; to minimize steps in communication; to push power to the edge, which is to say, delegate discretion to more localized levels, resulting in faster reaction time. But this interconnection comes at a price: more connections and more communication yield a higher information processing load, so nodes must be able to adapt to that load. Although the networked form of organization has arguably been with us for millenia, it can scale appreciably only with information technologies that have lowered the cost of communication transactions - and allowed us to aggregate the flood of information in useful ways.

I'm intensely interested in work organization, but when I conceived of genre ecologies, I was looking at how people used textual resources individually and in small groups. Yet the analogy can take us some distance. Like organizational networks, genre ecologies provide alternate resources to accomplish the same thing, resulting in easy rerouting and potentially localized efficiencies. Depending on the organization, unoffical genres can introduce ways to localize the work rapidly, allowing more discretion. The more genres, the more potential interconnections, yielding more flexibility and exploration - but also more instability, incompatibility, and information load. That last problem is compounded by the fact that genres tend to come from different activities and reflect different "logics"; their users must either learn and handle each of these logics - soaking up genre knowledge from the genres' originating activities - or work around them. Related is the fact that the network nodes - the genres - don't produce clean outputs; they're not conduits, they're alternate problem-solving approaches with different assumptions or warrants. Different logics and outputs spell big trouble when the cost of communicating and transforming information is high; so in more hierarchical organizations, unofficial genres are either coopted or squeezed out, reducing flexibility and local discretion, reducing information load, boosting regularity and centralizing "topsight" over the activity. (This is the kernel of my critique of Contextual Design in my first book.)

But in knowledge work - and more specifically, in net work - more flexibility is needed because more discretion is needed. As Drucker argued, knowledge workers must manage themselves and must be given enough flexibility to solve their problems. That is to say, discretion over one's work is pushed to the edge (organizationally), resulting in more idiosyncratic work arrangements and thus more unofficial, localized genres. Even when knowledge work is implemented in a more hierarchical organization, individuals will tend to route around hierarchical constraints because those constraints can hinder the localizations they need to enact in order to handle unique problems. That's part of what I began to see in my studies, from Schlumberger to the Iowa DOT to Telecorp to my current studies in progress, all of which can be characterized as knowledge work environments.

Let's bring this back to genre ecology models, because although it's tempting to stop there, the classic network concept is not a perfect fit. For one thing, these nodes (genres) are unique, so routing around a particular node means changing the character of the network. For another, genres aren't atomistic: they are more rhizomatic, trailing lines of association into other activities. And of course when we talk about genres, we aren't talking about concrete things: genres are "text types," or logics or traditions that we might use when composing or using particular text instances, so a given text could be perceived as belonging to different genres or a hybrid genre. Finally, like actor-networks, genres can be unpacked to reveal other genres - see my discussion of the PC-ALAS interface and its interface elements in Tracing Genres through Organizations - and that unpacking goes beyond simple decomposition because these other genres each trail associational linkages as well.

So the metaphor of network, like the metaphor of ecology, only gets us so far. But it does point us toward some interesting analytical approaches. For instance,
  • What if we see genre ecology models as network diagrams that help us to identify possible points of information overload?
  • What if we see communicative event models as records of how individuals traverse the nodes of the genre ecology/network?
  • What if we see sociotechnical graphs as ways to map alternate nodes and connections in a network, as well as to monitor the creation of new nodes?
Maybe these will constitute a good starting place for my next book. Meanwhile, I'd love to hear your comments.

Wednesday, July 08, 2009

Reading :: The Concept of the Corporation

Concept of the Corporation
By Peter F. Drucker

This is actually Peter Drucker's second book, but it's the classic study of GM that began his long string of management publications. I read the 1964 edition, with a 20-year retrospective - the original 18-month study was completed in early 1945, and the manuscript was completed just days after Germany surrendered. The 1964 edition - written three years before The Effective Executive, which revisits many of the concepts of the earlier book - has an introduction and epilogue wrapped around the 1946 text, making it - from my perspective - a time capsule wrapped in a second time capsule. Personally, I found it disorienting to read about the successes of 1945 GM as the GM of 2009 headed into bankruptcy.

The events of 1945 weighed heavily on Drucker's mind, and he repeatedly discusses GM's wartime production as well as those of Nazi Germany and the USSR. He is quite concerned about democracy and capitalism, which, he emphasizes, are certainly not the same thing. But he argues that since the US has cast its lot with the free-enterprise system, "The central questions of American statesmanship must thus be: how does the free-enterprise system function and what are its problems; what can it do, what can it not do; and what are the questions yet to be answered?" (p.15). So he brings the tools of political science to bear on the corporation, the linchpin of the US free enterprise system (at the time). In particular, he attempts a three-level political analysis of GM: (1) the institution as autonomous, "judged in terms of its own purpose"; (2) the institution in terms of how it relates to "the beliefs and promises of the society which it serves"; and (3) "the institution in its relationship to the functional requirements of the society of which the institution is a part" (p.24). He argues against elevating one level over the others, because that's what the Nazis did (p.29); rather, "both our statesmen and our business leaders have to find solutions to the problems of the industrial society which serve at the same time equally the functional efficiency of the corporation, the functional efficiency of society, and our basic political beliefs and promises" (p.29).

So Drucker begins his work of understanding the corporation in political, particularly organizational, terms: the corporation is not raw materials or gadgets, but organization, and even its technical problems are primarily problems of human organization (p.31). And as he was to argue in The Effective Executive, Drucker emphasized that "no institution can possibly survive if it needs geniuses or supermen to manage it" (p.35). The organization, he says, must make average or even under-average people succeed by leveraging their talents, capacities, and initiative (p.36). It must develop independent leaders and balance power and responsibility between central management and local leadership (p.37). In fact, modern industrial enterprises need many more leaders than institutions normally do, he argues (p.37).

GM, he says, achieves this aim partially through decentralization. But Drucker is not talking about the sort of radical decentralization we might see in, for instance, terrorist networks. Rather, he means that GM's divisions were organized as autonomous units (p.47), in a sort of federalist arrangement (p.49) in which division heads "work through and with two closely coordinated committees, one on policy, one on administration" (p.48). This form of decentralization works out to be a set of smaller hierarchies comprising a larger one. Indeed, decentralization "can be applied only where there is at least a rudiment of genuine executive functions" (p.149). Decentralization at GM was not part of a master plan, but it didn't represent muddling through either; it developed through some basic principles and practices (pp.69-70). It resulted in efficiencies at each level (p.109) as well as developing more leaders in each division (p.110). He goes on to argue that decentralization reaches its potential only in a decentralized big business: these have the flexibility to develop, but also more opportunities and resources than small businesses (p.189).

Drucker becomes very interested in meaning as well, meaning as an aspect of autonomy. He retells the experiments in the late 1920s at Western Electric that led to the concept of the "Hawthorne Effect" - but whereas qualitative researchers think of the Hawthorne Effect as demonstrating that any observation can change the work, Drucker argues that the real moral is that productivity and satisfaction increase as long as the worker believes her or his work was receiving attention and recognition. He argues that mass production has largely leached away the positive recognition that workers seek: mass production leads to no completed product (from the worker's perspective), therefore no meaning, no satisfaction, and no "citizenship" (p.135). And, he adds, solutions such as company paternalism and unionism are superficial (p.136). Higher pay doesn't translate into higher satisfaction or lower absenteeism - only reinjecting meaning into work can do that, he suggests.

Drucker goes on to take on Marx, arguing that the contradiction of use-value and profit-value makes no sense (p.190). I'm not sure I grasp the argument he makes here, frankly. But in this discussion, he makes a claim that arrested me simply because it's far less true now than it was: "competition is usually severely limited by geography and communication" (p.204).

In his 1964 epilogue, Drucker amplifies some of his themes. He adds that decentralization, in the sense he uses in the book, makes sense only if the units are distinct businesses within a company; it makes no sense for materials businesses, such as a paper company, to decentralize. He also notes that the book didn't address knowledge workers, which in the intervening 20 years had become the most important group: "the 'knowledge workers' of the professional middle class are doubling in number about every eight years or so. They are rapidly becoming the representative and most important group both in business and in our society" (p.241). He argues that GM was the first large-scale knowledge organization and that "the modern corporation is a knowledge organization" (p.241). And he says that knowledge work has a broad impact: "Automation, for instance, is the substitution of knowledge work for manual work and of perceptual skills for manual skills" (p.242).

Would I recommend this book? Certainly not as your first Drucker book. The concepts are too unformed, the marks of the Second World War are too deep, and the corporation is too new to Drucker at this point. But if you've read some of his other books, this one provides a fascinating view into the study that began forming some of Drucker's most important contributions.


I'm way behind on my book blogging, but once this semester is put to bed, I'll be reviewing these books:
  • Roberts-Miller, Fanatical Schemes
  • Clark, Supersizing the Mind
  • Drucker, The Concept of the Corporation
  • Arquilla and Ronfeldt, Swarming and the Future of Conflict (second reading)
I'll probably also work through several RAND publications, including Arquilla and Ronfeldt's collection In Athena's Camp. Great stuff to read in preparation for my summer study on coworking.