{"id":6000,"date":"2017-05-06T07:10:08","date_gmt":"2017-05-06T07:10:08","guid":{"rendered":"http:\/\/peerproduction.net\/editsuite\/?page_id=6000"},"modified":"2017-06-04T11:02:57","modified_gmt":"2017-06-04T11:02:57","slug":"producing-a-knowledge-commons-tensions-between-paid-work-and-peer-production-in-a-public-institution","status":"publish","type":"page","link":"http:\/\/peerproduction.net\/editsuite\/issues\/issue-10-peer-production-and-work\/peer-reviewed-papers\/producing-a-knowledge-commons-tensions-between-paid-work-and-peer-production-in-a-public-institution\/","title":{"rendered":"Producing a Knowledge Commons: Tensions Between Paid Work and Peer-production in a Public Institution"},"content":{"rendered":"
by Lorna Heaton, Patricia Dias da Silva<\/strong><\/em> This paper reports on peer-production initiated and organised by a public institution. We explore tensions around work and production in the digitisation project of the Marie-Victorin Herbarium in which a team of volunteers is working to photograph thousands of herbarium specimens, complete a database entry for each specimen and associate the digital photos with the database. Carried out in the context of a digital infrastructure, Canadensys, whose aim is to make information contained in Canadian biological collections freely accessible online, the project is realised almost exclusively with volunteer labour.<\/p>\n We suggest that this case may shed light on issues surrounding peer production in a capitalist context. The Marie-Victorin Herbarium is a public institution that benefits from the freely given work of volunteers in pursuit of a noble goal: the production of an open-access knowledge commons. It is not a capitalist firm and does not seek financial profit. Yet, this peer production is strongly framed and organised by an institution. Although the situation reproduces some facets of traditional labour organisation, such as a modular production process and \u201ctime sheets\u201d filled out by the coordinator, production output (i.e., the number of entries produced) is not the primary consideration. As might be expected in a museum situation, data quality is of much greater concern. Similarly, while they recognise what they are doing as \u201cwork,\u201d the volunteers themselves reject any idea of exploitation and regularly point to the benefits they receive from participation, as well as to producing something for the common good or for generations to come. Relationships between volunteers and the two paid employees\u2014a curator and a collections manager\u2014are not antagonistic; in fact, staff members try to maximise volunteer choice (of hours, tasks, etc.), are attentive to opportunities for learning, and integrate volunteers\u2019 suggestions for process improvement. They do, however, mobilise a pool of volunteer labour in order to carry out this colossal project. They orient and supervise the project, and are ultimately responsible for the efficiency and speed of the process (i.e., making specimens accessible online more rapidly) and for the quality of the resulting product. Based on six months of observation, a series of interviews with participants and two further months of voluntary participation in the digitisation project, we ask in what ways the public character of the institution and the goods produced, as well as the collaborative process may influence our interpretations of the relationship between paid work and peer production.<\/p>\n The paper is organised as follows: First, we review significant literature in order to situate our case. We then describe the work of digitisation in detail, with particular attention to who is involved in various stages of the process. A discussion of key themes follows.<\/p>\n Although scientists and amateurs have worked together for centuries, new sociotechnical configurations are changing scientific practice, relations between those involved, and even the very definition of what qualifies as scientific knowledge, as an increased variety of actors are connected and intervene at different stages of knowledge production and dissemination (Baker and Millerand, 2010; Hine, 2008; Waterton et al., 2013). On the one hand, the digitisation and availability of research data on the Internet enables its circulation among increasingly heterogeneous groups of actors\u2014to different research projects, across disciplines, with the public, and so on. On the other, the multiplication of web platforms facilitates the participation of amateurs and the general public in scientific research (Lievrouw, 2010).<\/p>\n Projects that invite members of the public to participate in producing scientific knowledge are often classed under the rubric of citizen science. This label masks considerable diversity. In fact, citizen science \u201cencompasses very different degrees of agency with regards to the research process, very different relationships with the professional scientists and very different degrees of influence on policy relevant scientific projects where citizens contribute to as \u2018citizen scientists\u2019\u201d (Nascimento et al., 2014: 5). To avoid misinterpretation, Heaton et al. (2016) prefer the term \u201cparticipatory science\u201d to describe the engagement of non-professionals in scientific investigation, whether by contributing resources, asking questions, collecting data or interpreting results.<\/p>\n Despite said diversity, the designation \u201ccitizen science project\u201d refers in most cases to projects that rely on large numbers of contributors to provide small contributions that are more or less independent, allowing them to be treated separately and then integrated into a coherent whole (Kelling et al., 2011). These types of projects are, thus, the scientific equivalent of crowdsourcing, in which \u201ccitizens\u201d work actively to gather and contribute data (usually observations) or to code or classify existing data (often specimens) (Lievrouw, 2010; Nielsen, 2012). Participatory science initiatives open new possibilities for engaging the public with science. However, long before the advent of Web 2.0, in his seminal 1996 article, Bucchi argued against \u201creducing the public to an external, monolithic, and taken-for-granted source of support,\u201d (which we argue a crowdsourcing approach to science tends to do).<\/p>\n Several models propose classifying citizen science projects according to the type of tasks performed by non-scientists and the benefits they may derive, usually along a continuum (Buytaert et al., 2014). For instance, Bonney et al. (2009) analyse a series of projects according to the involvement and control these participants have, namely whether they are \u201cContributory\u201d (designed by scientists), \u201cCollaborative\u201d (non-scientists not only contribute, but can also play a role in steps like design, analysis and dissemination) or \u201cCo-created\u201d (designed together and with a strong and continuous involvement of the public). Nascimento et al. (2014) note that citizen scientists have been described in the scientific literature as researchers, data collectors, observers, data processors, sensors, a conservation army, communicators and disseminators, as well as amateurs and enthusiasts.<\/p>\n Still, drawing lines between participants in citizen science projects is not always clear-cut. For Haklay (2013), all active participants are scientists, with the difference that \u201cprofessional scientists\u201d receive a salary for contributing to science, whereas he considers the other participants \u201cunpaid scientists\u201d. Furthermore, these categories are fluid since professionals often assume the role of volunteer naturalists outside their working hours (Ellis and Waterton, 2004) and volunteers may become members of paid staff (Bell et al., 2008) or vice-versa.<\/p>\n Haywood (2014) suggests examining the added value of citizens\u2019 contribution to science from two different perspectives: (1) the public understanding of science and technology tradition that is oriented towards scientific research in which the external value of projects is more salient; and (2) the public engagement in science tradition that emphasises an opening up of research and policy, and can be perceived as more focused on the internal value of such projects, namely for the participants. He identifies four main goals for public participation in scientific research projects: expanding the scope and scale of scientific research, enhancing science knowledge and understanding via interactive learning experiences for \u201cnon-scientists\u201d, increasing environmental stewardship and developing more democratic and inclusive science research and policy processes (Haywood, 2014: 65).<\/p>\n For scientists and funding agencies, cost-effectiveness is often cited as a benefit of citizen science given limited financial and human resources in biodiversity conservation initiatives (Darwall and Dulvy, 1996; Miyazaki et al., 2014). Nonetheless, these projects generally continue to require at least some financial support (Thiel et al., 2014) and others\u2019 costs\u2014such as time and effort in managing people and technological support\u2014need to be considered (Gura, 2013). Analyses of the return on investment of these projects suggest that coordination, communication with volunteers, and data checking and compilation imply costs that can be very high in the long run (Tulloch et al., 2013). Still, citizen science is often synonymous with large projects crowdsourcing information at a very low cost, and therefore, for some, an exploitative form of big science (Kinchy et al., 2014).<\/p>\n The ability to learn, to make discoveries and to teach have all been identified as motivations for participating in citizen science (Raddick et al., 2010). Indeed, knowledge exchange or mutual learning seem to play a key role, \u201cspecifically, through systems of informal mentoring, where the most experienced teach the less experienced\u201d (Bell et al., 2008: 3450). Other motivations include the desire to contribute to science, a sense of being part of a community, having fun and enjoying beauty, as well as interest in the project, in the field and in science in general (Raddick et al., 2010). The motivations of volunteers may change over time and form a complex framework of both internal and external factors (Rotman et al., 2012). Studies of volunteering not restricted to citizen science have already demonstrated multiple goals and how more than one goal may be pursued at the same time (Clary and Snyder, 1999). Clary and Snyder (1999: 157) identify six personal and social functions of volunteering: values (to express or act on them), understanding (to learn), enhancement (for growth and development), career (to gain experience), social (to strengthen relationships) and protective to reduce negative feelings or to solve personal problems. Researchers\u2019 awareness of volunteers\u2019 motivations helps contribute to ongoing participation (Couvet and Teyss\u00e8dre, 2013), while lack of such understanding and issues of mutual apprehension and mistrust constitute demotivating factors (Rotman et al., 2012). Less hierarchical projects benefit particularly from long-term participation and are associated with the development of personal relationships between scientists and volunteers, promoting trust and communicating goals, acknowledgement and attribution, as well as mentorship (Rotman et al., 2014).<\/p>\n Adler (2007) has argued that capitalist relations of production are in contradiction with the socialisation of the forces of production. As technology and organisation become more complex and production ceases to be uniquely an individual matter, the socialisation of labour produces a socialised or \u201ccollective worker\u201d (Adler, 2007: 1321). Adler and Heckscher\u2019s (2006) notion of \u201ccollaborative communities\u201d also points to the idea that social relations themselves might be \u201cproductive.\u201d The post-workerist strand of Marxist organisation studies emphasises that \u201cthe work of socialisation\u2014all that labour does without wage recompense to make this regime [of wage labour] possible both in the workplace and the community\u2014are present from an incipient moment in commodity labour\u201d (Harney, 2007: 148). This socialisation of labour forms the basis of social wealth, but also serves as a source of profit for capital. For post-workerists, the wage relation is only a part of capital\u2019s command over labour, and productive activity is moved outside the contractual employment relationship (B\u00f6hm and Land, 2012). Analyses of this type point towards a new \u201chidden abode\u201d of production, where work occupies an expanded terrain of social activity; where management moves further away from direct control of work to more complex practices of governance (Arvidsson, 2005); and where collaboration in production is increasingly the responsibility of the workers (Beverungen et al., 2015: 477).<\/p>\n The need to account for the social aspects of production, such as peer production, and for unpaid work, is all the more pressing in digital contexts. The idea of freely provided labour justified by the desire to contribute according to a gift logic is only part of the story and can be seen in the context of working relations that are also characterised by capitalist logics and strategies that co-opt peer production. Terranova\u2018s concept of free labour as \u201csimultaneously voluntarily given and unwaged, enjoyed and exploited\u201d expresses the reality of an Internet that \u201cis animated by cultural and technical labour through and through, a continuous production of value which is completely immanent in the flows of the network society at large\u201d (2004: 74). A critical view on relations of crowdsourcing and peer-production relativises the exalting discourses of fans, passionate amateurs and citizen scientists, as argued above. One source of such criticism stems from research on digital labour.<\/p>\n In the digital labour literature, the discussion revolves around two main topics: the precarious working conditions of online workers, be they micro-workers or skilled creative industry professionals, and the monetisation of user behaviour on the social web (Scholz, 2013). In both contexts, the relation of power in place is described as exploitative. According to Aytes (2013), Mechanical Turk contributors (also known as Turkers) benefit little from the value they generate, their tasks are menial, they often have no knowledge of what exactly they are contributing to given the extreme modular distribution of tasks and there is no recognition of their contribution.1 Similarly, crowdsourcing projects tend to separate worker and requester as their relationship is solely task-based and limited in time.<\/p>\n In the case of social networking sites, with the exception of Google, companies do not share their revenues with their users. This appropriation of value has, in fact, led to a class-action suit brought against Facebook regarding its Sponsored Stories Advertising programme, in which regular posts were rendered into ads featuring the user’s name and photo, without asking for permission, nor due payment (Fisher, 2015). Companies like Google prefer to downplay their influence, presenting themselves as mere \u201chosts\u201d (Gillespie, 2010), yet these services exert different forms of control: over users through data-mining and profiling; over their practices by determining preferable choices; and over their content through automatic filtering (Dias da Silva, 2014). As forms of immaterial and affective labour, the activities of social web users are often not considered work at all. However, bearing in mind the double role of user as content producer and data provider (Dijck, 2009), the concept of exploitation applies more aptly to user-generated data (Andrejevic, 2009). Users have very little control and reduced knowledge of the process, as well as less to gain from the accumulation of data resulting from their actions online, whereas the creation of content is often associated with deriving pleasure and exploring one\u2019s creativity (Dijck, 2013). Online contribution can thus simultaneously promote alienation and emancipation (Proulx et al., 2011).<\/p>\n Aigrain (2005) warns against the power of corporations to shape the law, as well as to control the sharing of information and knowledge. However, he also evokes another world, one of cooperation and solidarity, in which the commons are enablers of human development. The term \u201cCommons\u201d has become widely used in different contexts, but always tied to information sharing and the reconfiguration of current property systems. Von Hippel\u2019s book on the democratisation of innovation is dedicated \u201cto all who are building the information commons\u201d\u2014the result of innovators in a particular field making their developments freely available to all, and hence providing an alternative to information as private intellectual property (2005: 12-13). In their study of the knowledge commons, Hess and Ostrom (2007) discuss knowledge as a shared resource that is subject to social dilemmas. Knowledge commons can reside at the local level, the global level or somewhere in between. Wherever it is situated, it is characterised by multiple uses and competing interests.<\/p>\n Benkler (2002, 2006) suggests that commons-based peer production is a response to a set of changes that \u201chave increased the role of non-market and non-proprietary production, both by individuals alone and by cooperative efforts in a wide range of loosely or tightly woven collaborations\u201d (2006: 2). For S\u00f6derberg and O\u2019Neil, \u201cthe commons and peer production are two names for describing the same thing: a particular kind of labour relation\u201d (2014: 2), in which work is voluntary, tasks are self-selected and motivations are mostly intrinsic (e.g., peer recognition), rather than extrinsic (i.e., monetary compensation). Building on Raymond (1999), Demil and Lecoq (2006) define a bazaar style of governance with few control mechanisms that relies on open licenses and voluntary participation. O\u2019Neil (2015) refers to ethical-modular organisations (EMOs), which operate in a logic where motivations for participation are ethical, and oriented towards others, rather than towards financial profit. Peer-production can be carried out within a commercial infrastructure, which implies that it is partially subject to the constraints of commercially-driven cognitive work, as in the instances of digital labour discussed above. In peer production, appropriation often looms not early in the process, namely in the definition of the project, but towards the end as the community\u2019s output is appropriated by external commercial interests, thus becoming free labour (S\u00f6derberg and O\u2019Neil, 2014: 2). Bauwens argues that if nonreciprocal value is indeed captured, this is problematic given conditions of worker precarity. However, \u201c[u]nder conditions of social solidarity, the freely given participation to common value projects is a highly emancipatory activity\u201d (Bauwens, 2013: 209).<\/p>\n Herbaria are collections of preserved plant specimens, usually dried and placed on sheets of paper and classified according to family, genus and species. They serve many different purposes, from scientific research on plant taxonomy, phylogeny or evolution, to applied research in fields as diverse as climate change, agriculture, human health, biosecurity, forensics, land management, conservation biology, natural resources and control of invasive species. By providing a reliable, verifiable record of the changes to our flora over hundreds of years, they are an important source of information on our natural heritage and play a vital educational role.<\/p>\n The Marie-Victorin Herbarium in Montreal is a major collection, both in terms of size and reputation. Established in 1920, it contains close to 700,000 specimens of vascular plants primarily from the north-east of North America, particularly Quebec and Newfoundland, with a number of important specialisations, making it an important botanical reference. The Herbarium has two paid staff members: a curator and, since 2012, a collections manager. Since 2012, it has been housed in the Biodiversity Centre of the University of Montreal within Montreal\u2019s Botanical Garden. The construction of the Biodiversity Centre was part of a larger project that aims to provide Canada with a pole of excellence and a network of researchers in biodiversity, Canadensys. Canadensys\u2019 goal is to offer free, universal access to the information contained in biological collections via digital infrastructures. The project to digitise the Herbarium\u2019s collection takes place in this context.<\/p>\n Digitising a herbarium\u2019s specimens represents a solution to many of the challenges involved in curating a herbarium (Flannery, 2012; Heaton and Proulx, 2012). Dried plant specimens are fragile and subject to attack by insects, deterioration due to light or fluctuations of temperature or humidity. Curators are often obliged to restrict consultation and borrowing of specimens. In this context, digitisation is both an excellent opportunity to make a systematic inventory of a plant collection, and to greatly increase its accessibility. Numerous digitisation initiatives are underway and they adopt different models, ranging from mass, semi-automated digitisation, such as that underway at the Smithsonian\u2019s Institution\u2019s National Museum of Natural History (Rogers, 2016), to contracting out digitisation work to produce the Global Plants Initiative, a database privately owned by JSTOR (Heaton and Proulx, 2012), to using gamification strategies to crowdsource digitisation in the Herbonautes, organised by the Mus\u00e9e national d\u2019histoire naturelle in France (Pignal and P\u00e9rez, 2013; Zacklad and Chupin, 2015).<\/p>\n In early 2014, the Marie-Victorin Herbarium introduced a new process based on the photography of specimens, with the goal of increasing the speed of online publication. Since this work is very intensive in terms of resources, the Herbarium recruited and trained volunteers that are responsible for most of the phases of the process and who supply an important source of labour\u2014the equivalent in time spent to that of three or four full-time employees. In calling on volunteers, the Herbarium was acting in accordance with the long and strong tradition of volunteering at the Montreal Botanical Garden. For 40 years, through the Amis du Jardin Botanique, individuals have volunteered in various capacities, including at the Herbarium. What was new for the Herbarium was the number of volunteers recruited for this project. After a flurry of initial enthusiasm, the number of regular volunteers involved in digitising specimens stabilised at around 30 people. The majority of volunteers are retired, but there are also students, and workers with rotating schedules or without a permanent position. Their backgrounds are equally diverse: a few have degrees in biology and related areas, but did not necessarily work in their field of study. Most volunteers have some type of professional or technical training, ranging from medicine to teaching, accounting, or health management. Around three quarters are female, and the level of economic wealth varies greatly. What unites the volunteers is their appreciation of nature in general, and their love for plants in particular. Most volunteer once a week for either a full day or a half day, but a few are more regular. The work of digitisation is done on site at the Herbarium.<\/p>\n The digitisation project was never imagined as paid employment. In recent decades, natural history collections the world over have been subject to increasing budgetary restrictions (Dalton, 2003; Yong, 2016), leading Funk (2014) to denounce the \u201cerosion of collections-based science\u201d. University botany programmes are closing (more than half of U.S. programs have closed since 1988), whereas in Canada, as elsewhere, smaller herbaria are closing and donating their collections to larger ones (15% of North American herbaria have closed since 1997) (Deng, 2015). The lack of a budget for digitisation should be seen more in this larger, international context of disinvestment in natural science collections, than in relation to recent Canadian or Quebec austerity measures or economic policies.<\/p>\n Our paper is based on regular observations during the first six months of the project in 2014, followed by two months of participant observation and interviews with eight volunteers, the collections manager and the botanist of reference\/curator. Each observation period was documented with notes and photographs. The notes contained both objectivist descriptions of activities and spaces and impressions\/intuitions (Maanen, 1988; Marcus, 1995) that subsequently served as a departure point for thematic analysis. Observation notes and interview transcripts were analysed collaboratively in meetings of the research team.<\/p>\n The digitisation process implies multiple manipulations, divided into four main stages. Firstly, specimens must be mounted (or the existing mounting reinforced) on standard sheets of paper and the information contained on their labels\u2014the name of the plant and the person who identified it, the date and location of the collection, habitat and the name of the collector\u2014verified. This step is fundamental for the conservation of the specimens and is not specific to digitisation, although some changes were introduced to simplify the photography process. Volunteers take complete charge of the mounting process: An experienced volunteer trains new recruits and has prepared a reference manual to help with the process. Verification is a crucial aspect of the process, since plant identifications change regularly. Volunteers do much of the work to verify the exactitude of existing labels. They use an array of tools, other databases and online resources, as well as books. Any changes in nomenclature will be approved by the curator before new labels are printed and attached. The final result is an up-to-date record, an inventory of the Herbarium\u2019s holdings, such as that illustrated in Figure 1.<\/p>\n <\/a><\/p>\n Figure 1. Herbarium sheet<\/strong><\/p>\n Before being photographed, the specimen has to have an entry in the database. Most will be classified as \u201cpartial entries\u201d, meaning that they have basic information (the content of the specimen\u2019s label) and an identification number (a unique identifier). Volunteers create entries for each specimen and transcribe information contained on the labels. Entries can be completed at a later date with additional information, such as geo-referenced data, the history of the plant\u2019s name, etc.<\/p>\n The photography stage involves a constant to-and-fro between the image and the database. The specimen is digitised as a photo, the photo is linked to its database entry, and metadata on the photography process is added into the database. This repetitive work, performed by individual volunteers or teams of two, requires a number of small steps and demands great attention to detail. Another volunteer later verifies that the photos are recorded, that they are appropriately named and that the association with the database is functional\u2014a sort of quality control that was added as an additional step after a few discrepancies were found in the first digitised records.<\/p>\n After photography, the entry has to be completed. This is sometimes done before the entry is uploaded and made accessible, and sometimes after. The form is divided into different colour-coded sections so that it is easier to understand the different types of information required: history of determinations, location, projects that have used the specimen, etc. Completing an entry may be relatively easy or difficult, depending on the information on the specimen sheet. In every case, this work is done by a small number of skilled, or specialised, volunteers. For example, certain volunteers specialise in determining the precise geographical location (coordinates) of specimens. Georeferencing a specimen\u2019s location and habitat requires particularly intensive use of a variety of sources: satellite images, Google maps, old military maps, coordinate format conversion tools, and so on. Again, this multiplicity of sources requires constant checking and crosschecking in order to ensure accuracy and data quality.<\/p>\n The final step of the process\u2014uploading entries to Canadensys\u2014is not carried out by volunteers, but by the IT team of the Biodiversity Center. Of its close to 700,000 specimens, as of December 2015, the Marie-Victorin herbarium has 150,000 online entries, 50,000 of them georeferenced, and 7,000 images. All the information regarding the specimen, including a high-resolution photograph, is available online and can be downloaded freely. A series of filters facilitate searching the available collections. Information and images have no restrictions regarding use; their licensing as Creative Commons 0 means that not even attribution is required. They become part of the knowledge commons\u2014to be used by other scientists, hobbyists, government officials, artists, NGOs, whoever, wherever, whenever.<\/p>\n Table 1. The division of labour in the digitisation process<\/strong><\/p>\n
\nUniversit\u00e9 de Montreal<\/p>\nIntroduction<\/h2>\n
Literature review<\/h2>\n
Context and method<\/h2>\n
Description of the digitisation process<\/h2>\n