Coding Control: Governance and Contingency in the Production of Online Worlds
First Monday

Coding Control: Governance and Contingency in the Production of Online Worlds by Thomas M. Malaby

Approximately 11 million people worldwide regularly participate in persistent, graphically–realized, online virtual worlds (including EverQuest, Ultima Online, Second Life, and Lineage). Originally built on computer game platforms, these worlds are in many respects the most viable online arenas for broad–based social action; their participants pursue lasting social relations as well as globally consequential economic activities that elide the boundary between offline and online experience. Amidst this startling growth it is the producers of these worlds who are confronting in practical terms unprecedented challenges of governing what are in many cases fundamentally open–ended, yet architected, environments. How are they doing this, and in particular how are they developing their own position as those theoretically (if not effectively) in ultimate control? This article, based on ethnographic research at Linden Lab, the makers of Second Life, considers this question with respect to competing and continually changing ideas of Second Life’s content, a particularly unruly yet central concept in this virtual world’s ongoing governance.


Governance and Values
Perspectives on Content: Multiple Rotations
First Rotation: Content as Developer’s Product
Second Rotation: Content as Marketing Resource
Third Rotation: Content as Commodity
Fourth Rotation: Content as Aesthetic Aspiration




4.2 Service Provider. You acknowledge that Linden is not a traditional game provider; instead Linden acts as a venue and a service provider that may allow people to interact virtually regarding almost any topic, at any time, from anywhere, in a variety of formats. In addition, the Service may allow some users to alter the gaming environment on a real–time basis. Linden is not involved in actual communications between users or even in users’ interactions with the virtual world of the Service. As a result, Linden has very limited control, if any, over the quality, safety, morality, legality, truthfulness or accuracy of various aspects of the Service. — Second Life Terms of Service agreement, 1.2006

Linden Lab, the maker of the online virtual world Second Life, is indeed not a traditional online game provider. Second Life is a rapidly growing online world (launched in June 2003, by mid–2006 it had 350,000 “residents”) that is graphically realized, open–ended, and which lacks many of the typical elements of massively multiplayer online games (MMOGs), such as foundational game objectives. Most obviously and importantly, Linden allows and encourages Second Life’s users to build, program (“script”), and paint (“texture–map”) simple to complex 3D objects in world, and these users then own the intellectual property rights to these creations. In contrast to many online world companies, Linden Lab has thus ceded to a great extent its governance of the content of Second Life by forgoing its own property rights claims to what its residents make. In its going against the grain in these and other respects, Linden Lab has put itself in an unusual and constantly changing position vis–à–vis its creation, and understanding this position is important for anyone interested in how ethics and governance are both implemented and emergent in, to use Edward Castronova’s term, “synthetic worlds” (Castronova, 2005).

Two issues appear in this brief segment of the Terms of Service Agreement (ToS hereafter) that are central for this inquiry. First, Second Life is presented here as a place that can generate the unexpected; that is, as an environment of heightened contingency, in which users may interact “regarding almost any topic, at any time, from anywhere, in a variety of formats,” and in which, furthermore, some users may “alter the gaming environment on a real–time basis.” This is not to say that Second Life is utterly chaotic; as a persistent world it generates many regular and observable patterns and trends as well. But contingency is central here because it is only if we recognize Second Life’s potential to generate new practices, meanings, and social forms that we can grasp the full extent of its possibilities and regularities. Second, and correspondingly, Linden Lab’s governance, or control, is presented here as fundamentally limited; Linden Lab’s position is one of “very limited control, if any” over a wide (one is tempted to say total, given the list) range of the world’s attributes. As complex and open–ended social venues which give rise to communities, economies, and meaning, all synthetic worlds exhibit both expected and unexpected outcomes which may collide with their producers’ efforts to control them, but Linden Lab is unusual in the extent to which it acknowledges and even embraces this element of contingency.

As Linden Lab seeks to foster Second Life’s stability and growth it must confront governance issues in deeply practical terms, with a pragmatic awareness of this tension between its attempts to control and the world’s unpredictability. The most obvious and fundamental manifestation of this approach is Linden Lab’s turning over the production of Second Life’s content to its users. The term is used not only in synthetic worlds but in computer game design generally to connote, loosely speaking (see below), the “stuff” that is in synthetic worlds and which their users encounter, use, or with which they can interact. It is at the center of discourse and practice about synthetic worlds, for both their makers and their users, because the “quality” of a world’s content is often seen as an index of its appeal. But this concept has, in fact, multiple understandings in practice, and here, based on ethnographic research with Linden Lab, I examine these various dimensions of content for Linden Lab and consider to what degree this illuminates its ethical position as willing and able to govern in only limited fashion.

The nature of governance in synthetic worlds is a pressing question, at root driven no doubt by the now familiar tension between the imagined transformative capacities of technology, on one hand, and the countervailing suspicion that digital society simply replicates offline social processes. The truth, of course, lies somewhere in between, but the issue of synthetic world governance is urgent in a more immediate and practical sense because of the demonstrated viability of these online environments as arenas for broad–based social action — the social activities that emerge even within traditional MMOGs and which transcend their game objectives. I have elsewhere discussed how the persistence and contingency of synthetic worlds accounts for the generation within and beyond them of “real” economic value, durable social effects, and meaning, and how this suggests that it is untenable to view them as ontologically separable from other domains of human action, and I will not replicate that argument here (Malaby, 2006). Instead, I assume for the purposes of this discussion that we can safely take the reality of the social activity in these worlds as a given, leaving the real vs. virtual and work vs. play distinctions behind. This puts us in the right position to explore, as essential background to any effort to understand human experience that involves these environments, the position of those who wield, at least in relative terms, the greatest amount of control over these worlds — those with access to the largely inaccessible software code of the program itself, the stuff that Lindens (employees of Linden Lab) say is “under the hood”.



Governance and Values

In the case of Linden Lab, governance is complex, given the degree to which it has turned content creation over to its users. Linden increasingly limits its own creation of content to providing more landmass — bringing additional land (and with it, more servers) online to accommodate the growing population of users (“residents”). Residents can buy this land and not only build on it, but literally re–shape it themselves (raising hills, adding lakes and trees, etc.), meaning that the production of content in Second Life does not stop at ground level. But more broadly than this, governance in and of Second Life is complex because it is a world that generates, out of the collective actions of those connected to it, unexpected social patterns and phenomena, emergent qualities that outstrip the capabilities of anyone — however well–positioned — to observe or to predict with complete confidence. Governance in Second Life is at least four–fold, produced by the meeting point of Linden Lab’s attenuated vertical power, its employees’ access to and competence in manipulating code, the nature of Internet control, and the social conventions generated around and within Second Life itself [1]. Allow me to discuss this multiform governance in more detail.

How powerful is Linden Lab, then? Linden Lab can, of course, pull the plug on Second Life; this is what prompts Richard Bartle to argue that MMOG developers are best understood as “gods” (Bartle, 2006). But this would not necessarily lead to the shutdown of the large number of third party sites that also constitute part of Second Life. The near collapse of another virtual world,, led to a massive influx in Second Life of “refugees” from There (many of which chose the last name “Therian” to signal their membership in this group), and they set about recreating places and activities from their “homeland”. Something similar could of course happen if Second Life were ever shut down. With even the effects of exercising this ultimate form of control potentially thus attenuated, we are forced to acknowledge the limited, though still dominant, nature of Linden Lab’s control of Second Life, and it is possible that this position of limited control is coming to be an undeniable feature of MMOGs, who are under pressure to relinquish some control (or sovereignty) over their creations (see Taylor, 2006).

For Second Life, convention and architecture are both important ways in which governance shapes activity in the world ...

This embrace of a position of limited sovereignty on Linden’s part is recognized within the company, and is marked by the circulation and debate of ideas about collective action and emergence, primarily through book recommendations such as James Surowiecki’s The Wisdom of Crowds, New York: Doubleday, 2004). On the whole, the opposition that they develop about this issue is between vertical, top–down control, and free–form social action that is governed (like Surowiecki’s actors) only by the aggregation of individual actions. This opposition, which serves as a touchstone for practice at Linden Lab, will not suffice for this article’s attempt to understand the nature of Linden Lab’s governance, however. This is because in what is now a popular rush to exalt emergent collective phenomena we are in danger of missing the forms of governance that exist even in the absence of vertical control — what Hardt and Negri (2000), and others have called decentralized control — but which is not simply the aggregate of individual decision–making in an “invisible hand” sense. This form of governance tends to be implicit, for it is constituted by (cultural) convention — practices and expectations so taken for granted that they rarely (if ever) are discursively acknowledged, otherwise glossed in this special issue as governmentality — and by architecture in the broad sense — constraints of the material circumstances that are (for most) practically impossible to manipulate. Both of these forms of governance are typically decentralized; they operate without top–down authority. For Second Life, convention and architecture are both important ways in which governance shapes activity in the world, but which tend to be forgotten in Linden Lab practice in favor of the two types of control introduced above.

In synthetic worlds, then, we have shared expectations and practices (governmentality) both at work and in the continual process of formation, and for the users these are rooted in their historically situated past experiences. But we have an architectural decentralized mechanism at work, and that is the code itself, which, like cultural convention, governs at least in part implicitly, behind the scenes of everyday interaction (despite the fact that many of its particulars were explicitly planned and implemented), but which also, like a given material condition in the Marxian sense, is relatively unchangeable for those users enmeshed in it. Digital environments are themselves built, and like culture they are built on shared practices. In the case of the Internet these are the shared protocols by which computers communicate, and in Second Life, which is situated within these, there are further coded protocols that govern the relationship between the user’s computer and the Second Life servers, inter–user communication, real–time changes to the synthetic world itself, and so forth.

So, while Second Life users can manipulate their environment, and do so even in unexpected and powerful ways, to a certain degree the most fundamental constraints of the environment in which they operate are both ineffable and inaccessible. Not so for Linden Lab’s developers, who can and do tinker “under the hood.” And while they know full well that they cannot perfectly anticipate the effects of these changes, nor can they transcend the broader protocols of the Internet over which Second Life operates, they are in a fundamentally different position of agency relative to Second Life’s users nonetheless. This brings me to the topic which animates this essay: how does the concept of “content” reveal how values circulate among these forms of governance in the ongoing design and construction of Second Life by Linden Lab?



Perspectives on Content: Multiple Rotations


Working with a prim in Second Life

Figure 1: Working with a prim in Second Life.


In Figure 1 you see a user in Second Life working with a prim, the basic unit for building within the world, and which, when initially made, appears as a plywood–covered shape; such as a box, or cone. If you look closely you can see the three axes that appear as soon as you begin to work with a prim, allowing you to rotate the prim and change your perspective on it, the better to view and make changes. Prims are somewhat fetishized — in the anthropological sense — around Linden Lab, appearing amidst jokes and other comments on the dry erase boards used for common concerns and light–hearted jokes in the Linden office, and this reflects their fundamental presence as the protean substance for building, the most prominent of Second Life’s forms of content creation, followed by scripting (programming) and texture–mapping (the wrapping of objects in complex “textures” or printed designs).

All the more reason, then, for me to draw an analogy from this interaction with prims in order to structure my discussion of content here. In what follows I will rotate through four perspectives on content in Second Life as constructed through practice and discourse by those at Linden Lab, showing how content is seen, variously, as developer’s product, marketing resource, commodity, and aesthetic aspiration. This treatment will also help us to recall that these different views, here presented as somewhat distinct, are in reality situated perspectives of what is an emergent totality — an ongoing negotiation of what content is in Second Life among those at Linden Lab.



First Rotation: Content as Developer’s Product

Beginning with this image of prim manipulation has another useful effect, and that is to demonstrate a common tendency around Linden Lab throughout the period of my field research (the calendar year 2005): to view content as that which is scripted, built, or mapped. On this view, prims are content, par excellence. Interviews with Linden employees suggest that the roots of this conception lie in the practice of computer program development and game development specifically, where content is the graphics, maps, and objects that are made for a game and which are in some sense independent of the game engine, which itself remains “under the hood” — out of view, technically involved, of a different order, and largely inaccessible. The general emphasis around Linden Lab in invocations of content in this way is on that which users can make without getting their hands dirty by messing with the engine. Instead, by using the building, scripting, and texture–mapping tools residents can create content in what is, effectively, its principal sense.

Despite this suggested distinction, the implicitly coded engine and the explicit content creation tools that users use are nonetheless interrelated, and this is helpfully represented in Figure 1 as well. Note how the avatar of the resident in the foreground is gesturing toward the prim on the right as it is manipulated. Every time a resident works with the tools of Second Life, whether doing something as common and simple as engaging in text–based chat, or changing clothes or appearance, but most obviously and grandly when scripting or building, the resident’s activity is represented in gestures and actions of the avatar itself. What is more, if an object is being worked on, the changes to that object are observable by others in the world in real time, even down to click–to–click changes in, say, the object’s color as the user tries different points on the building tool’s color wheel. Here you can see another resident, relaxing on the chairs in the background, watching on.

This publicly visible creation was an intentional decision by the developers of Second Life, and, according to one engineer, it ran counter to a number of users’ preferences. It was, however, consistent with, as he put it, the promotion of “shared experience,” the idea that while “in world” the users would be able to be in touch with what others were doing [2]. We must note, however, that the activity that the code highlights for others’ observation is the making of content in the conventional programmer sense. This engineer commented that one of the most remarkable experiences he had in world was watching a master builder (who before participating in Second Life had had no prior 3D modeling experience) sculpt a dinosaur in real time. The emphasis, deeply inscribed in Second Life’s code, is on representing content creation activity to others, but again with a conception of content that gives pride of place to the technical activities of content creation, especially building and scripting. Contributing other kinds of content to Second Life, such as the organization of a regular meeting group for victims of domestic abuse, or establishing oneself as a charismatic socialite in dance clubs, are not distinctively represented in one’s avatar’s actions, and this creates ongoing challenges for Linden Lab’s efforts to recognize and promote the creation of content in this broader sense.

Contributing to this view of content as developer’s product is the organization of time itself within Linden Lab. Lindens often acknowledged the emphasis on development in Linden Lab’s work practices, and in many cases pointed specifically to Linden Lab’s beginnings as a company, when it was almost entirely developers who focused on creating a streaming technology that would make a real time–manipulable synthetic world work. The most obvious way in which the development of Second Life is shaped by practices of computer program development is in the schedule of version releases, a form of temporal organization, which fundamentally shapes for many Lindens the experience of working at Linden Lab. Like most software development practice elsewhere, version releases at Linden Lab had, through most of 2005, a target date, and everything from brainstorming to coding to testing to bug fixing was temporally determined by this date. What is more, a major focus throughout this calendar was on added capabilities to the Second Life client code, such as streaming audio, often characterized as adding “tools” for the users. The emphasis here, again, was on those technical tools which allow for new possibilities in scripting, building, and texture–mapping. There is thus an important presumed relationship between Linden Lab’s developers and the developers of content in world.

Content in this sense governs users’ (and Lindens’) expectations and practices through its pervasive inscription in the software code. Inescapably, to act within Second Life in certain respects (such as building) generates distinctive aspects of shared experience. Over time, this becomes self–legitimizing, as social practice often does. The powerful connection is between the software architecture as a form of governance, and those practices themselves. Together they contribute to governmentality about content in Second Life, most obviously seen in the celebrity status which some master builders, scripters, and texture–mappers (most often clothing, skin, and avatar designers) enjoy in the world.



Second Rotation: Content as Marketing Resource

Another aspect of how content is conceived in and around Linden Lab is its position within the effort to market Second Life. Shifting from a focus on the technical orientation of Linden Lab as a company, this reminds us of the business–oriented component of Linden Lab which, like any company, looks to make money from its creation. In order for it to grow its Second Life subscriber base, this means that the content users make needs, on the whole, to be a draw for new users. (This position itself rests upon an assumption that the percentage of potential users who want to create something with Second Life’s tools is limited, a point to which I will return.) For marketing purposes, this creates a circumstance that is unusual in an important way: something that is absolutely central to the appeal of a product, the content, is not wholly or even mostly controlled by the product’s maker. When any given content is created within Second Life it can surprise Linden Lab, whether as a boon or a disappointment, and as a result much of the marketing imagination of Lindens is occupied with trying to find a potentially electrifying match between resident–created content and the drive to find and appeal to potential users.


SLAP: Second Life Avatar Products

Figure 2: SLAP: Second Life Avatar Products.


A relatively straightforward example of this was in an online ad campaign that parodied an early 2005 Gap campaign (see Figure 2). Here the marketing team at Linden Lab identified residents who had created striking custom avatars (a form of content creation that emphasizes texture–mapping), and inserted them into a clickable ad that poked fun at the Gap’s pretensions while pointing to the possibilities and distinctiveness of Second Life as a comparable world. Here it is not the meeting point of Linden Lab developers and Second Life residents as developers that generates expectations about content. Instead, the meeting point here is between Second Life content and an imagined public desire, with the marketing team of Linden Lab playing matchmaker. The text of the ad, which in every phrase was a slightly altered version of Gap’s original ad copy, is revealing: “enjoy being whatever you want in Second Life”; “there’s more at more avatars. more fashion. more possibility.” The marketing imperative to reach an audience of potential users pushes one aspect of the values that, in Linden Lab’s view, characterize Second Life — its limitless possibility — into the realm of marketing discourse, as the ad makes an explicit link between the contingency of the world and individual freedom of expression. The appeal of Second Life, in this presentation, rests in its potential as a place where you, the user, can express your uniqueness, your individuality, and be unlike anyone else. (The extent to which this possibility may rest upon the user’s own ability to script, build, and texture–map is left implicit, and I will turn to this in the next rotation.)

Another case that illustrates the position of Linden Lab’s marketing efforts as situated between the in–world content creators and the imagined potential audience of new subscribers is that of the creation of two in–world games: Tringo and Chinatown. Tringo is a game invented in Second Life by a user, and it went on to become very popular in–world. Combining bingo and Tetris, Tringo brings users together in–world where they compete over multiple rounds of the game for a jackpot. The gameplay is simple and compelling, and the success of the game in Second Life led in late February 2005 to the first ever licensing deal for the use of Second Life content outside of the world, a story significant enough to be covered by the Wall Street Journal, and promoted by Linden Lab on the Second Life Web site and via a trading card (see Figure 3; for an extended discussion of the trading card, see Malaby, 2006).


Kermitt Quirk's Tringo

Figure 3: Kermitt Quirk’s Tringo.


The case seems to be an ideal demonstration of what Linden Lab’s project can generate: Second Life has the sophisticated building and scripting tools to allow a resident to create content that he owns the IP rights to, and he can then license that content, and Linden Lab has not missed the opportunity to point to this example. I was researching at Linden Lab around the time that Tringo was beginning to make a splash, and what I noticed at the time was something perhaps a bit surprising: Tringo was not the focus of Linden talk about content that would lead to an influx of new residents. In the early part of that year another effort to make an in–world game preoccupied and involved Linden Lab, and that was Chinatown (or more properly, USL: Chinatown). Whereas Kermitt Quirk, the maker of Tringo, worked on his own and to a certain extent below the radar, a development team of residents within Second Life was working to build an elaborate first–person shooter (FPS) game on an island in Second Life (see Figure 4).



Figure 4: Chinatown.


This team, Bedazzle, involved a number of well–known builders, scripters, and mappers, and was run by a user with a strong reputation as a project manager, as demonstrated in several of Bedazzle’s prior projects. Chinatown was to be the culmination of Second Life’s technical capabilities: an actual real–time FPS, with its own homemade combat system, multiple weapons, fast and responsive action, and a gorgeously rendered environment — a few blocks of a mythical “Chinatown” — in which to play. It was also to be a culmination of Second Life’s promise as an environment for creativity, especially creativity that was collaborative. Chinatown would testify to the world’s viability as a development environment, one where a globally far–flung team could coordinate and produce a sophisticated product.

Before Chinatown’s opening, it was featured on Second Life’s Web site, and the office was abuzz with excitement about it. Again, for a company with many employees whose background was in computer game development, this anticipation was understandable: this was content, this was what Second Life could do, and this would bring in at least a portion of the potentially huge audience of gamers and game developers out there. Tringo, meanwhile, continued to grow, and one heard it mentioned around the office as something fun, and worth checking out. By February 2005 roughly one in four in–world events was a Tringo competition — a staggering number. Tringo’s success and subsequent licensing shouldered the highly touted Chinatown project aside, and this demonstrates the moving target that is the Second Life marketing effort, which shifted gears to promote Tringo more prominently.

The relationship between what happens within Second Life and the broader out–of–world landscape is central, then, to the marketing efforts of Second Life, but it also relates to broader efforts to realize Linden Lab’s intention that Second Life be increasingly connected with other aspects of experience, rather than being set apart from them. Currently, this involves increasing the connections between in–world and out–of–world activities; supporting third party Web sites for among other things, the sales of in–world items; and, markets for its currency, blogging, and forums.

The challenge here again is related to how content is coded within Second Life. Because its developers wanted content within Second Life to be sui generis, made within the world and by its tools uniquely (and demonstrably to other users, as noted), residents correspondingly have had very little ability to import 3D models rendered in external programs or scripts written outside the Second Life tools.

This decision (which contrasted with some efforts by other synthetic worlds that allowed for the importing of user–created content, such as Active Worlds) reflected the aforementioned imperative to make the experience of making content within Second Life a shared experience (and thereby, I would further suggest, to elevate human activity in world, to make it ontologically on a par with out–of–world activity). But as activities within Second Life, in a variety of ways, increasingly cross over to out–of–world arenas, such as third–party Web sites, the demand to make these connections easier by, for example, adopting common formats, bumps up against the specificity of Second Life’s content creation tools. So whereas the first perspective on content that I described highlighted the relationship of Linden Lab’s developers to the segment of its residents who themselves are in world developers, Linden Lab’s marketing and other broader efforts highlight a different relationship, that between the content created within Second Life and users outside of the world, who (the marketing team hopes) may become interested in Second Life through encountering its products and other content outside of the world. Among other challenges, this brings them into closer contact with the protocols and common file standards that govern the Internet more generally, and thus the obstacles that having a world full of unique tools creates.

While the presence of a potential audience for Second Life across the Internet is crucial for understanding the ongoing efforts to attract new users, we must not lose sight of the fact that much of the activity in these Web sites is by Second Life residents themselves, who purchase items, post images, and otherwise engage Second Life while outside the environment. This brings another aspect of content into view, and that is the way it serves to divide the residents of Second Life.



Third Rotation: Content as Commodity

In both its development background and its marketing efforts, Linden Lab has been invested in the content creation of its users, and the tools of the world reflect an emphasis on content creation of a particular sort, while the efforts to demonstrate Second Life’s appeal to those outside the world makes a different claim about content. While all of this has been happening, however — while more powerful scripting and building tools have been added, and while selling its content outside the world has itself been expanded — more and more residents within Second Life have joined and begun to engage the world without building, scripting, or texture–mapping.

... more and more residents within Second Life have joined and begun to engage the world without building, scripting, or texture–mapping.

One of the most notable events in this transformation of Second Life’s community of residents was an unexpected occurrence that followed the release of v.1.4 of Second Life. This version, like most releases, included new capabilities for the users. Linden Lab had worked hard to include avatar animations in 1.4, the ability to write script that would animate your avatar performing an activity. According to recollections of those working at Linden Lab at the time, there was a great deal of anticipation amongst them about this new capability, and they expected to see users scripting a wide variety of animations for everyday activities (smoking a cigarette, leaning casually, etc.). At the last minute, Linden Lab realized that they could also easily add (“throw in”) another capability, one which allowed users to stream audio content from their own computers into the world, audio which others users could hear when on that user’s property. What happened next was a remarkable and entirely unanticipated in–world phenomenon: dance clubs. Users built dance floors, sometimes just on bare land, sometimes amidst huge structures, and streamed dance music from their PCs, while other users scripted dance after dance and sold these animations to users (or clubs, which made them available for free).

Dance clubs continue to be a significant part of Second Life social activity, and have taken their place in Linden Lab’s marketing efforts for the world, but as content they posed a challenge to prevailing expectations at Linden Lab. Just what kind of content were the dance clubs? They involved little to no technical skill to create, either in scripting or texture–mapping, but by late 2004 they were by far the most popular in–world places, and continued through 2005 to vie with Tringo as the most popular sites for interaction.

Dance clubs forced a recognition of a new kind of resident for Linden Lab, one which, by some off–the–cuff estimates, may account for up to 80 percent of the users: consumers. These are residents who do not make their own content, and instead are happy to purchase objects, avatars, skins, accessories, animations, and the like. This has generated a significant economy within and beyond Second Life. By mid–2006, approximately US$2 million in trade was happening in Second Life every week. By their very actions these residents have shaped another aspect of content, rendering it commodified. Second Life makes production and distribution of the content created by users nearly costless, so selling multiple copies of items could not be easier (for an extended treatment of the implication of this for market capital and other forms of value, see Malaby, 2006). These items can be coded to be, after sale, not copyable, not modifiable, and nontransferable, so the profit margin of the maker is assured.

By mid–2006, approximately US$2 million in trade was happening in Second Life every week.

Competition, then, is over design, and fashion rules much of the market in Second Life. What, then, are the implications for the values of technical skill and individuality that we have already seen are closely identified with content in its other aspects? Potential new users are told that they can “enjoy being whatever they want in Second Life,” but for most of them this seems to involve buying clothes and other items that thousands of others have bought as well. The consumers of Second Life seem largely untroubled by their lack of technical engagement with the world’s tools, although a vocal minority have raised concerns that, in comparison with content creators, they feel ignored by Linden Lab. But Linden Lab is not in any clear position vis–à–vis these two types of users, seeking instead ecumenically to support both uniquely inspired and technically created content and its commodified versions. The desire to make the selling of Second Life objects on third–party Web sites easier, for example, is an instance of Linden Lab looking to support this commodification.



Fourth Rotation: Content as Aesthetic Aspiration

It may come as a surprise, given the above, to learn that Linden Lab has a “content team,” a group of employees whose purview is Second Life content, but this group’s continued role and changing perspective on content fills out this picture, because content also has an aesthetic dimension, again inextricably linked to the values inscribed in Linden’s practice and the code itself.

Linden Lab has always needed a team to provide new land for the world that is Second Life. They make islands (available for purchase to residents, where they can control access and other features), provide extensions of the existing (“mainland”) landmass, and construct whole new continents (theoretically, at least, there is no limit to how much new land could be introduced; the universe, or “grid,” of Second Life is ever–expanding). But early on in Second Life’s existence, when the number of its users was relatively small (less than 10,000), the content team made more than just land. They built infrastructure (roads, which they continue to provide for the mainland), and, most importantly to them, pieces of content in the first sense above: a chalet high in the mountains, the welcome area for new residents, public meeting spaces (such as amphitheatres and the like), and other things. One member of the content team described the intent of this creation: “We’re just trying to provide the residents with some ideas, to prompt them to build around a theme, and work together.” That is, the content team saw its role for much of Second Life’s history as seeking to inspire collaborative creation by residents through well–designed evocative content. This manifested itself in the landscape — as in the case of the “snow sims”, regions of the mainland with high altitudes, snow, and other wintry features — and in buildings like the mountain chalet itself.

Why was this involvement by Linden Lab, this effort to inspire residents, seen as necessary? Early on, it was often the Linden employees (many of whom had been beta or early release residents themselves) who knew how to get the most out of the Second Life tools, so building content and placing it in world (and making it, often, copyable, modifiable, or transferable) was an effective way to prime residents in what they could do themselves. What is more, Linden Lab would then provide the textures, scripts, basic building shapes, and other elements that would be needed to build similar content, at “dispensers,” kiosks in world where residents could click and grab them. As Second Life grew, however, a second consideration became more pronounced. Second Life’s landscape had a problem. It was, as various Linden employees put it, “ugly,” “trashy,” and “junk–filled.”

Since residents can buy land and build permanent structures on them, but (on the mainland) are under no “governmental” requirements to build consistently either in zoning (residences next to residences, businesses next to businesses), or in terms of style, the Second Life built landscape is remarkably variant, with giant advertising signs next to enormous modernist skyscrapers next to medieval castles next to five–story dinosaurs next to perfect recreations of art deco gas stations. With no gravitational limits on building, the skyline is correspondingly unconstrained, and the word that springs to mind, as one’s avatar flies up and surveys the horizon is, indeed, “trash”. The dissonance of colliding styles is deafening.

But Second Life was intended, as we have seen, to be a place for collaborative creation, where residents not only made whatever they wanted, but worked together to make content better than they could ever make themselves. Thus, Chinatown was a triumph–to–be in the eyes of many at Linden Lab because, again, it testified to the ability for a group of residents to make great content together, and the aesthetic appeal of those alleys, the reflections in the puddles on the ground, and the neon lights blurred in a seeming fog, did more than testify to the collaborative technical prowess of Bedazzle, it suggested that residents could make something beautiful together. So why wasn’t, if not the whole Second Life landscape, at least more of the mainland seeing this kind of beautiful work?

This question was a lingering one for Linden Lab, and I heard it raised regularly as a kind of lamentation: remarks that Second Life was, well, ugly brought a rueful smile, and comments about its impact on — corresponding to the perspectives on content above — new features (once we increase the draw distance [how far one can see] the world will look worse, not better!), consumer appeal (we’ve got to direct our new users to good content, so that they’re not put off by the trash), and Second Life’s market (how can we stop the ugly billboards if that’s how they get business?).

Thus in late 2004 through early 2005 the content team embarked on an ambitious plan. Needing to be ready to bring on a lot of new land as growth increased, but wanting to do more than just attach more ground to the mainland already in existence (which was beginning to look more and more amorphous), they decided to design an entire new continent, an atoll, which they would bring online in pieces as needed. Thus, they sought to bring an interesting and aesthetically coherent shape (it resembled that of the Greek island Santorini, formed by a collapsed volcano) to Second Life on a grand scale, one with a dramatic topography and correspondingly scenic vistas. But they decided to go further, and build content on this atoll that would itself be aesthetically coherent, hoping to prompt residents to explore and expand on a design style together. This content would be archeological — the remains of an imagined past civilization that itself moved slowly up the atoll in altitude as it advanced technologically. This civilization had an iconography (built around the totem–like presence of the moth), and distinct building styles (such as the corrugated steel and whitewashed colors that typified areas that extended out on platforms over the water which the atoll surrounded).

As the atoll was brought online, the content team watched interestedly to see what the residents would do, and indeed in some places residents built with an eye toward the content already in place. Making use of the textures and other things Linden Lab provided, neighborhoods with a compelling consistency of style emerged here and there, but on the whole the engagement of the “mythology” of the past civilization was not extensive, and in many cases, especially commercial buildings, there were no common stylistic elements at all.

By late 2005 Linden Lab’s content team appeared disillusioned with the prospect of prompting aesthetically compelling, collaborative content within Second Life on a large scale, and at the same time there was a significant shift in residents’ relationship to land itself. More and more private islands were being sold (at approximately US$1,000 apiece) and these owners, typically powerful residents with lots of real estate holdings, were “renting” space to other residents. Making use of the sovereignty these islands afforded them to build neighborhoods of a particular style, much like offline housing developments, these island owners made spaces that appealed to consumers who wanted a place to live in Second Life, ready–made, and with clear zoning restrictions (contained in every renter’s agreement with the island owner, much like a covenant in real estate). It seemed, again, that like other aspirational values of Second Life, like creativity and collaboration, beauty had found purchase only through a market mechanism, and this development was treated with ambivalence by the content team and throughout Linden Lab. As always, the world confounded their best–laid plans, but seemed somehow to thrive nonetheless.




Across these perspectives one sees a common trend — content seems in Second Life particularly vulnerable to commodification, and seems to stand in most marked contrast to the exaltation of collaborative, technically–skilled creativity that underwrites Second Life’s code and design. The reason for this may be as simple as pointing to Second Life’s growth. As its community is less and less a set of early adopters, who saw something different and possible in Second Life’s affordances, and instead is a community that expects more and more to find in Second Life what it finds elsewhere, this transformation is perhaps inevitable. Thus the emergent values that characterize Second Life are pushed toward those which are already quite familiar out–of–world, where individuality and freedom of expression are largely constituted by and through commodity choices, which become themselves creative acts. We are now on familiar territory, and rightly so, if we believe that there is nothing unreal about what is at stake in synthetic worlds. The governmentality that emerges from the relationships between Second Life’s code, its makers, and its users is of course not itself isolated from the broader modernist governmentality that obtains throughout much of offline life.

What we must be ready to examine, however, is the way in which the synthetic nature of these worlds generates important effects, both through the artifactual nature of their code and through the emergence of shared conventions within and across their communities. Content in its various aspects reveals the value–laden quality of the relationships of control between developers, marketers, resident content creators, potential new users, and resident consumers, and how what is valued by these different groups collides and transforms their expectations. Second Life demonstrates how an open–ended, and extremely user–manipulable virtual environment nonetheless evinces the position of its makers — not in a monolithic or determinative way, but in a way that suggests a constant state of mutual remaking, where an asymmetrical relationship leaves both parties nonetheless transformed. This suggests that even as commodification exerts an enormous influence on Second Life, it is not necessarily the end of the story. Second Life continually surprises its makers, and it may surprise us too. End of article


About the author

Thomas M. Malaby is Associate Professor in the Department of Anthropology and co–coordinator of the Modern Studies Program at the University of Wisconsin–Milwaukee. He has published articles and essays on virtual worlds, practice theory, risk, and mortality, and his book, Gambling Life: Dealing in Contingency in a Greek City (Urbana: University of Illinois Press, 2003) explores how games reveal human attitudes toward contingency.
E–mail: malaby [at] uwm [dot] edu



1. Beyond the immediate settings of Linden Lab and Second Life are other forms of governance which are beyond the scope of this discussion, most notably national and international law.

2. It is important to note that over the course of 2005 changes to Second Life made it easier to import and export content between Second Life and their party content creation programs, such as Poser. This change itself represents a broader transformation in Second Life’s relationship to other domains of user experience and as well Linden Lab’s efforts to accommodate them.



Richard A. Bartle, 2006. “Why Governments aren’t Gods and Gods aren’t Governments,” First Monday, Special Issue number 7: Command Lines: The Emergence of Governance in Global Cyberspace, at

Edward Castronova, 2005. Synthetic worlds: The business and culture of online games. Chicago: University of Chicago Press.

Michael Hardt and Antonio Negri, 2000. Empire. Cambridge, Mass.: Harvard University Press.

Thomas M. Malaby, 2006. “Parlaying Value: Forms of Capital in and Beyond Virtual Worlds,” Games & Culture, volume 1, number 2, pp. 1–22.

James Surowiecki, 2004. The wisdom of crowds: Why the many are smarter than the few and how collective wisdom shapes business, economies, societies, and nations. New York: Doubleday.

T.L. Taylor, 2006. “Beyond Management: Considering Participatory Design and Governance in Player Culture,” First Monday, Special Issue number 7: Command Lines: The Emergence of Governance in Global Cyberspace, at

Editorial history

Paper received 15 August 2006; accepted 25 August 2006.

Contents Index

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial–NoDerivs 2.5 License.

Coding Control: Governance and Contingency in the Production of Online Worlds by Thomas M. Malaby
First Monday, special issue number 7 (September 2006),

A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2020. ISSN 1396-0466.