Decentralization-Based Typology of Peer Production
* Article: Towards a (De)centralisation-Based Typology of Peer Production. Mélanie Dulong de Rosnay and Francesca Musiani. tripleC 14(1): 189–207, 2016
URL = https://pdfs.semanticscholar.org/d578/983c224717cb29df26d9b197b289b7eef233.pdf
Abstract
"Online peer-production platforms facilitate the coordination of creative work and services. Generally considered as empowering participatory tools and a source of common good, they can also be, however, alienating instruments of digital labour. This paper proposes a typology of peerproduction platforms, based on the centralisation/decentralisation levels of several of their design features. Between commons-based peer-production and crowdsourced, user-generated content “enclosed” by corporations, a wide range of models combine different social, political, technical and economic arrangements. This combined analysis of the level of (de)centralisation of platform features provides information on emancipation capabilities in a more granular way than a market-based qualification of platforms, based on the nature of ownership or business models only. The five selected features of the proposed typology are: ownership of means of production, technical architecture/design, social organization/governance of work patterns, ownership of the peer-produced resource, and value of the output."
Characteristics
Centralization vs Decentralization
The five dimensions This section presents the core elements of the typology, analyzing how the (de)centralisation criterion can be applied to each of the five selected dimensions: ownership of means of production (Section 5.1), technical architecture (Section 5.2), governance (Section 5.3), ownership of production otput (Section 5.4) and value-generation (Section 5.5). Examples follow the definition and discussion, in order to illustrate the arguments with concrete cases.
Ownership of the Means of Productio
The platform's online means of production are its creative and networking interface as well as its distribution, storage and processing capabilities. Ownership of these can be centralised in the hands of a firm or a non-profit entity, which controls the brand, the patent, the quality. It can also be distributed among contributors structured in the semi-centralised format of a cooperative, a cloud of local networks, or informal networks where each peer or node owns and control its own servers.
A previous typology of digital media labour distinguishes among three modes of organization of productive forces (Fuchs and Sandoval 2014): ownership of labour power, of the means of production, and of work output; with various degrees of control, from centralised to decentralised: self-control, partial self-control and partial "alien" (or external) control, full alien control. In the vocabulary of the commons, the ownership of the means of production corresponds to the provision of infrastructure: this is the dimension that, according to Ostrom's Institutional Analysis Framework (1990), allows work to be performed. The concept of infrastructure traditionally includes the means of transportation (roads) or communication (telecommunications networks) and has been extended to ideas: the intellectual or cultural infrastructure (Frischmann 2013).
The concept of infrastructure provision has been applied to online creation communities in order to study their governance and the building of a digital commons (Fuster Morell 2014b). The concentration of ownership of the means of production will have an impact on the other dimensions. In a Marxist sense, “The first form of alienation refers to the product of the worker’s own work and the inability to use the product of this own work for his or her living. The second form of alienation refers to the inability to organise the process of work, which lies exclusively in the hands of the capitalist who owns the means of production” (Wittel 2012). While it is evident that free labour is leveraged when a user-generated content platform is exploited by capital, no conclusion can be drawn in the case of distributed ownership. Ownership of the platform can be centralised in the hands of a single firm or stockholders (e.g. Flickr, Facebook). It can also be centralised in the hands of a non-profit entity (e.g. the Wikimedia Foundation, or Creative Commons in the case of a legal infrastructure controlling the licensing offer), which controls the peer-production brand. In the latter case, the trademark is not managed according to commons-based principles, but controlled following a strict intellectual property policy. There can also be decentralised or federated ownership of a platform among informal networks, where each node owns and controls its own servers (e.g. Identi.ca21, Diaspora), allowing, for instance, the removal of inappropriate or infringing content. This usually involves a strong technical component, which will be addressed in the next section. At this stage of our analysis, we cannot draw a conclusion as to whether contributors are able to benefit more from their work (or to be less excluded from enjoying the benefits of their production) if the platform is owned by a non-profit, or if it is not owned at all. Neither can we assert that more centralisation of platform ownership will lead to more exploitation of contributors. Both the market (if the entity is capitalistic) and commons-based peer production (if the platform is non-profit) rely on free labour. The same platform can even switch its status from non-profit to corporation (e.g. Couchsurfing; see Schöpf 2015). The question of the ownership of the contributions produced and the sharing of the generated value is addressed in subsequent sections.
Technical Architecture
To define our second dimension, technical architecture, we borrow from the IEEE Standard for Architectural Description of Software-Intensive Systems (IEEE P1471/D5.3), which describes it as “the fundamental organization of a system embodied by its components, their relationships to each other and to the environment and the principles guiding its design and evolution” (Bredemayer and Malan 2001). Thus, the architecture of a network or an application is its underlying technical structure (van Schewick 2010), designed according to a “matrix of concepts” (Agre 2003): its logical and structural layout, consisting of transmission equipment, communication protocols, infrastructure, and connectivity between its components or nodes. It is increasingly recognized that changes, notably design choices, in the Internet’s architecture affect the economic environment for innovation, and in turn impact public policy. While policy-makers have traditionally used the law to bring about desired economic effects, architecture de facto constitutes an alternative way of influencing economic systems and as such is becoming another tool that actors can use to further their own interests (van Schewick 2010). The design of peer-production platforms can be organized according to a centralised architecture, around servers hosted locally or in the cloud, and can give the administrator full control and power of exclusion over production, contributors and output. Centralisation can subsist at several layers of the infrastructure, the most obvious being the servers on which the applications are running and the content produced is hosted (e.g. Facebook).
Other (less centralised) models may include a limited number of (super-)nodes responsible for controlling and administering the production platform (e.g. Skype24). Generally, the greater the degree of centralisation, the higher will be the degree of control given to the administrator over the process and output of production. Architectures with a centralised system of governance facilitate the coordination of individuals contributing to collaborative work and are generally easier to implement technically than the decentralised alternatives (for propositions to resolve these difficulties, see Antoniadis and Le Grand 2009).
Centralised platforms based on proprietary software and single points of control at the technical level are, however, more likely to facilitate privatization and concentration (with the possible facilitation of abuse of dominant positions that this entails). At the other end of the spectrum, technical architecture can also be distributed among peers (Musiani 2015). This is the case for many peer-to-peer networks that rely on the willingness of multiple users to share their own resources (e.g. BitTorrent for file-sharing, Diaspora* for social networking) so as to contribute to the proper functioning of the network or application26. This is also the case for mesh networking applications (e.g. Commotion27) that rely on the technical infrastructures of users’ devices to provide resources and wireless connectivity to the other users of the network (Antoniadis and Le Grand 2009). Unlike centralised models, these architectures considerably enhance the autonomy of users; however, by sharing capabilities and responsibilities and fragmenting actions and files, decentralised architectures make it more difficult for an entity, be it a firm, a collective, an individual or the state, to enforce control (Guadamuz 2011; Dulong de Rosnay 2015).
Governance
The scholarly debate on what constitutes governance of socio-technical systems, how its boundaries should be drawn, and what actors are involved is currently wide-ranging and complex (Borras and Edler 2014). For the purpose of this paper and its typology, we define governance as the set of decision-making processes, the ensemble of procedures that frame production, the process of work, and the choices subtending the organizational design of peer-production platforms, including technical, legal and value-sharing choices. Depending on their degree of (de)centralisation, these features can contribute to the empowerment of peer producers and the distribution of the tasks conducted on the platform. Intended in this connotation, the governance of peer production has been addressed in several contributions. As in the present study, Aaron Shaw (2013) uses the parameter of decentralisation to assess how relational governance mechanisms, such as gatekeeping, play a role within open online collectives in managing organizational boundaries and filtering the contributions of participants; concluding with regard to the existence of a mix of centralised and decentralised gatekeeping, he observes that the egalitarian ethos of open online collectives exists in tension with the mechanisms through which participation and status inequalities emerge among participants. In collaboration with Benjamin Mako Hill, Shaw further observes that governance in peer production may actually, and counter-intuitively, tend towards concentration (“oligarchy”) as projects grow (Shaw and Hill 2014). Building on previous work on the governance of participation (Fish et al. 2011), Currie et al. (2013) apply notions such as Hirschman's “exit, voice and loyalty” in their definition of governance in peer production projects (using the case of free software). They argue that this definition needs to include the answer to questions such as: “Is there a formal procedure for decision-making? If not, how are decisions made? Do participants have the capacity to exercise both exit (without penalty) and voice (without fear of reprimand)? […] What constitutes having a real voice, and how does it manifest in discussion forums, face-to-face meetings, and financial donations? Can participants leave without losing something, or protest and expect to be heard? What can you not exit from, such as formal or technical commitments to a website, a platform, an account, etc.?” (Currie et al. 2013).
The (de)centralisation parameter at the core of our typology can be applied to the four elements of our operational definition of governance: terms of use, production rules, work process and design choices. These are examined briefly and examples provided. Terms of use can be unilaterally imposed by the entity operating the platform (e.g. Flickr), or they can be drafted collectively or otherwise negotiated among peers in the case of self-managed projects (e.g. Lorea28). In the former case, users do not have any influence on the governance of the platform; if they want to use the service, they necessarily have to accept and comply with the terms and conditions. The latter case represents, instead, an example of distributed, commons-based governance; in this case, peers not only have their say in the terms of use, but also have the opportunity to participate in establishing the rules of production. In this regard, different degrees of decentralisation may range from the opportunity to comment or vote29 on licensing changes (e.g. Facebook) to the possibility of engaging in a real collaborative drafting process (e.g. Wikipedia; see Forte, Larco and Bruckman 2009). Wikipedia is a case of governance being semi-distributed among several nodes structured in semi-centralised clusters (administrators and editors control what is accepted or rejected in case of conflict) or decentralised local networks (chapters for choosing projects and thematics on which to focus, e.g. the Wikicheese30 project of Wikimedia France). Beyond the drafting of rules which users have to respect in order to participate in the peerproduction process, governance is also embedded in the work process and the organizational design. Distributed organizational design of product development can take place in innovative firms, which apply self-governance (Valve31). More fully distributed governance models will be based on decentralised decision-making and cooperative management. As opposed to traditional systems of governance based on hierarchical and centralised decision-making, distributed governance models are often more empowering to contributors and users. They are based on bottom-up, decentralised processes of decision-making and generally exist in innovative firms or communities, which rely on self-governance and cooperative management. Even companies with centralised ownership can embrace distributed management (e.g. Valve; see van der Graaf 2012). Finally, the governance of the design choices of the platform itself can be subjected to more or less centralised control. This will move towards more distribution of power if users have the capacity to propose new features to be integrated into the service. It will go further if they can develop such features themselves, in the form of plug-ins or add-ons. If the technology is based on Free/Libre Open Source Software (FLOSS), this feature will reach the fullest level of decentralisation and enable exit without penalty, by means of forking to create a new, parallel project duplicating previous contributions and starting afresh. Decentralised governance is here supported by, and embedded in, the openness of the technical design based on the software and data licensing, empowering peers to modify the platform, adapt it to their needs and transfer the datasets. Inversely, if the technology is closed, a central point of control will be re-introduced in order to validate the new feature (e.g. AppStore32).
Ownership of Peer Production Output
From texts and photos to comments and editing, most creative contributions on peer platforms are subject to copyright law. By ownership of peer production output, we mean the large scope of copyright licensing arrangements, which may be present in the platform’s Terms of Use, governing the use and reuse conditions of the content created by users.
A synonym for this is availability: according to the typology of Currie et al. (2013), “from GNU General Public License to Creative Commons licenses to moral economies where no formal legal structure exists, to direct corporate expropriation, where participants knowingly carry out free or underpaid labour and indirect expropriation, as when participants often unwittingly offer up their data for commercial usage”. These conditions of appropriation of the fruits of creative (in the copyright sense) labour performed by peers can be a direct consequence of features such as ownership of means of production and governance of terms of use. If a platform is centrally owned and governed, terms of use can be and will probably have been drafted by the corporation or the non-profit. Distributed governance gives users a chance to obtain more favourable licensing choices even if Terms of Use cannot generally be negotiated and have to be accepted in order to contribute on the platform. The scope of licensing options, the bundle of rights33 according to our typology oscillates between the highest and exclusive concentration of rights in the hands of the platform’s owner (thus, centralisation) and their widest redistribution and availability for others (hence, decentralisation).
It will have an impact on the value of the output generated (as we analyse in the next section), depending on the gradation in sharing of rights between the platform, the original user-contributor and subsequent users in the audience who may become contributors and may build upon the peer production. A few notable examples (e.g. Facebook) show that terms of use endowing the platform owner with the highest degree of concentration favour a high level of concentration in platform ownership and rights ownership. The platform can assert exclusivity over peer production through Terms of Use governing users' contributions. Contributors accepting the transfer of all rights and granting full exclusivity to a platform will not be able to reuse their contribution on another platform. Social media often require a full transfer for the exploitation of fan fiction (Ford 2015): Amazon Kindle Worlds, although offering royalties to contributors, requires that “When you submit your story in a World, you are granting Amazon Publishing an exclusive license to the story and all the original elements you include in that story”. Some platforms relying on user participation, even if they are owned by commercial companies, may allow modifications to be freely available in order to increase further traffic and reuse, and thereby purchases, donations or grants, leading to new forms of value-sharing (in the context of “mods”, modified video games, see van der Graaf 2012). Instead of a transfer of ownership away from the creator, the platform can require a nonexclusive licence that allows it to reuse content, sell, transfer or sub-license content, or to use it in conjunction with advertising (Instagram36 or Second Life37). The platform can claim a non-exclusive licence, but only to display the content, thereby allowing the system to function a minima, without extending the license to the right to sell (Twitter also corresponds to that standard). If information is absent, if no terms of use or licence are specified, copyright will apply by default and all rights will remain with the author; only the use of a Creative Commons or similar licence would define the scope of third-party permissions (Kopfschlag39). As is the case, for example, with Creative Commons licences,40 the range of options used among free culture and digital commons communities includes different ways of the creator retaining some rights (more centralisation) while sharing some rights over peer production with the public (more decentralisation). All Creative Commons licenses options guarantee the right of access to all (usus), the right of modification or commercialization (fructus) may be performed by others according to creators’ permissions, who can prevent exclusion or enclosure (abusus) with a copyleft, share-alike clause. Reserving rights to commercial exploitation will partially recentralize ownership, to the extent that the author becomes a central contact point for commercial exploitation. Reserving derivative rights only will partially decentralize the commercial exploitation of the original production. The copyleft option provides a distributed, shared ownership and prevents further recentralisation (or enclosure) of the production's derivatives. The latter regime requires modifications (fructus) to be licensed according to the same conditions of access and use (usus and abusus), thus distributing ownership by granting rights to the public, which members may use by creating a derivative. The attribution option and public domain dedication will fully distribute the rights’ exercise, allowing anyone even to re-proprietize a derivative of the peer production.
Value
Our fifth dimension addresses the value of the output of the peer-production process. While we acknowledge that defining value as it relates to peer production is a process in the making, and still incomplete41, we define value operationally, for the purpose of this typology, as the management of benefits, understood as both market-based economic and unquantifiable socio-cultural benefits, generated from the peer-production process, ranging from their exclusive appropriation by an actor or a group of actors (most often the owner of the platform), to redistribution to the community and even further, to society at large (thus, according to our typology, at the further end of the decentralisation spectrum). Value is heavily linked to ownership of the means of production and of the peer production itself; in this regard, Currie et al. (2013) wonder: “Is there collective control and/or individual access to the resource produced by participation? Can participants trust that what they give to a project will be returned to them in some form (credit and authorship, legal rights, access to resources)?” We build this definition on the number of recent contributions in which value in commons-based peer production has been tentatively defined. Coleman (2005)’s notion of “ethical labour”, Kelty (2008)’s analysis of “schooling”, Arvidsson and Peitersen (2013)’s contribution on “reputation” focus on the production work more than on the production output. Value assessment of public domain works (Pollock 2006; Erickson et al. 2015), which digitization and preservation can be achieved through peer production (Dulong de Rosnay 2013a), and which reuse will lead to further peer production, can serve as a useful analogy. Evaluation models developed by Pollock and Erickson et al. are powerful, but still market-based. In a broader definition of value, we want to take into account not only the commercial exploitation of the production (or its equivalent) and the social value of cooperation for the peers, but also include non-quantifiable cultural benefits. Capital-based metrics are unable to take into account the quality cultural value of peer production. Happiness metrics are also market-based, and demonstrating the political value of the commons and of sharing will require a conceptual framework, which will depart from capital-like evaluation of the public domain, open data or reputational economies. Net value, compensated value, creation costs, business models, economic opportunities for investors, commercial equivalent, cost savings, visitorship are useful indicators in Pollock and Erickson et al. Personal benefits for the producer, and advantages for the user, have been studied for free software. But there are no metrics to appreciate the advantages and “pleasure” for society derived from production and access to works and services of quality, towards intangible cultural collective development and not only economic development.
Fuster Morell et al. (2014) argue that the question of value in collaborative communities is a matter of both economy and justice, as “[t]he problem of how to regulate and reward activities that are presently without a market value […] is contingent [upon finding] a rational and transparent measure of value”, and build a definition of value in peer production that includes six dimensions: community building (dimension of the community/network of producers), objective accomplishment, monetary value, reputation, ecological value and social use value of the resulting resource. Building upon Bauwens' seminal work on the political economy of peer production (2005), Tkacz, Mendoza and Musiani (2014) also highlight the role of peer-production processes in producing “bearable hierarchies”, and in turn how these are crucial in defining value, observing how “the intensification and extension of computational processes [...] has led to a proliferation of bottom-up procedures to formalise (social) values, rendering them easily calculable and lending order to the decentralised world of peers, but without necessarily replicating capitalistic calculations of value”; if hierarchies persist in peer production, they point out, “perhaps it is the qualities of these new hierarchies and competitive forms that is novel”. This ties into Bechmann and Lomborg’s (2012) observation that value creation in social media is either addressed in economic and socio-political terms (power, exploitation and business revenues) or as “sense-making”, creative explorations of the self and management of social relationships; the authors argue that these different conceptions of value creation have consequences when it comes to conceptualize users as participatory agents—as peers in the sense adopted in this article. More specifically on the quality and the reusability of CBPP (and user-generated content distributed under open licenses at large), Dulong de Rosnay (2013b) introduces the impact of the legal framework (besides copyright licensing) on value in peer production, as the provision or disclaimer of liability for peer-production outputs will have an impact on the motivations of contributors and on the possibility to add value to the output of others by creating derivative works, as one of the motto of free culture is to build upon previous works: large amounts of open content, without a representation by the licensor that this is not infringing upon others' rights—a frequent output of peer production—cannot be safely reused and do not necessarily entail good value in the long term. If copyleft licensing potentially multiplies the value of the production, as it allows its transformation, the absence of warranty that it can be safely and legally done hampers the economic value, as authors or derivatives may face legal risks. It also potentially diminishes the social value of the commons if cooperation is discouraged. Quantitative measurement of Creative Commons licensed works is achievable, but measurement of derivatives from Creative Commons licensed works has not been performed. Along these lines, Kreiss, Finn and Turner (2011) note that, in peer production, changes in ubiquity and power relations at the level of processes do not necessarily imply a revolution in the attribution of benefits related to outputs. Kostakis (2013), building on previous work on the “crisis of value”, points out the sustainability problem when assessing value in peer production: he remarks that more and more of commons-based peer producers’ time goes into producing use value (utility as valuable to someone), but there is no substantial return of income to them: a volunteer pool can be sustainable collectively through the continuous renewal of its members, but membership cannot be exclusively composed of volunteer individuals on a permanent basis. In its variety of for-profit forms, the so-called sharing economy is an example of how value in peer production is often (re)centralised in the same way it was analyzed by Terranova (2000). Several businesses in the sharing economy, such as Über and Airbnb43, can arguably be considered as a deviation from the symmetrical, egalitarian, (perhaps utopian), collaboration-based model of the original idea of carpooling and couchsurfing, reducing or voiding some of the traditional costs associated with doing business by delegating tasks to users and co-opting their output, and concentrating the benefits derived from this. In this sense, “peer labour seems to both depend on, and, at times, extend existing concentrations of resources” (Kreiss et al. 2011, 255).
At the other end of the spectrum, emerging legal instruments such as copyfarleft licences (Kleiner 2007) can be leveraged to decentralize value in peer-production processes, as they “distinguish between commercial usages enacted by worker-owned collectives, cooperatives, or any other institution where profits are distributed (equally) amongst all workers, and those enacted by commercial entities or corporations whose businesses are exclusively based on the exploitation of wage labour” (Vieira and De Filippi 2014). Copyfarleft and reciprocity principles are emerging, and licensing schemes are still under development at the time of writing; therefore, there is no practical example of an application embedding these principles of value redistribution."
(https://pdfs.semanticscholar.org/d578/983c224717cb29df26d9b197b289b7eef233.pdf)
More information
- "Two research projects have been collecting data on peer-production platforms: in Europe the P2P value project and its directory, which is itself crowdsourced, of 375 commons-based projects, and in the United States the Birds of the Internet team, privileging the study of participation dynamics (Fish et al. 2011). Literature differentiates between several types of organization for peer production by online communities, depending on factors such as top-down coordination (crowdsourcing) or bottom-up self-organization (the commons)."
- The Birds of the Internet project: http://recursivepublic.net/