This essay is based on an interview conducted by Matt Thompson, portions of which originally appeared on the blog Savage Minds the week after Cultural Anthropology relaunched as an open-access journal.1 Herein, I elaborate on my original comments to provide a nuts-and-bolts account of what has gone into the journal’s transition. I am particularly concerned to lay bare the publishing side of this endeavor—of what it means for a scholarly society to take on publishing responsibilities—since the challenges of publishing, and the labor and infrastructure involved, are invisible to most authors and readers.
Matt Thompson: When did the Society for Cultural Anthropology (SCA) decide to go the open-access route and what was motivating them?
Tim Elfenbein: Cultural Anthropology was one of the first American Anthropological Association (AAA) journals to start its own website. Kim and Mike Fortun were responsible for the initial site. They wanted to know what extra materials they could put on a website that would supplement the journal’s articles. That experience helped germinate the idea that there was more that the journal could do itself. The Fortuns are also heavily involved in science and technology studies (STS), where discussions about open access have been occurring for a long time.2 When Anne Allison and Charlie Piot took over as editors of the journal in 2011, they continued to push for an open-access alternative in our publishing program. Although the SCA and others had been making the argument for transitioning at least some AAA journals to open access for several years, it was the Crow report and survey by the Committee on the Future of Print and Electronic Publishing that finally provided the impetus for the AAA to look for a test case (Crow 2012; Nicholas and Schmid 2012). In August 2012, the AAA put out a call to all the association’s sections to see if any would be willing to transition their journal to an open-access model. The SCA formed a task force to evaluate the AAA’s proposal and the feasibility of shifting the journal. At the time, the SCA was the only section ready to make the leap, and I think this is because it had already been thinking seriously about open-access publishing for quite some time (Weiss 2014).3
MT: What has the transition to open access meant for the production and organization of the journal?
TE: From my perspective as the managing editor, the most important change that has occurred in transitioning Cultural Anthropology to an open-access publication is that the SCA is now a publisher. The journal isn’t just open access, it is society published. In other words, in leaving our contract with Wiley-Blackwell, we have not left publishing. Rather, we have become our own publisher.
From the point at which the SCA made its decision in the spring of 2013, we had a very short amount of time to develop our publishing capacities (I took over the position of managing editor from Ali Kenner in July, 2013). We are now responsible for a greatly expanded set of tasks, processes, and relationships. In brief, they are as follows:
Production process: Production takes final manuscripts through copyediting, typesetting, proofing, and file conversion, the output of which is finished articles in several file formats (typeset PDF, XML, HTML). I do some of this labor in house (coordination, file preparation, language quality control) and external vendors take on other tasks (copyediting, page and cover design, typesetting, file conversion, document quality control). A print-on-demand service adds another layer to production, including uploading final files and proofing hard copy. I am responsible for the scheduling and coordination of authors, freelancers, and vendors during the production process.
Content management: We must create and maintain a file management system and pair it with specific management practices. While the SCA has operated its own submission management system since 2008,4 we have expanded our use of Open Journal Systems (OJS) by pushing information about published articles from the submission to the publishing side of the platform. This entails uploading all final article files into the article records and correcting metadata, and then creating issues by assigning article records and structuring a table of contents. We must create and correct records not only for newly published articles but for all of the journal’s back issues up to the November 2013 issue. Beyond our content management system, we are also responsible for working with archival services to ensure the preservation of the journal’s content.
Metadata preparation, editing, and dissemination: Once the metadata in our content management system has been edited, updated, and proofed, it needs to be packaged and fed into a number of data streams. These include our website, alerting services, indexers, identifier and citation-linking agencies, and our print-on-demand service. Our social media publicity efforts can also be considered as channels for metadata.
Publishing platform: The platform where the public accesses the journal is the Cultural Anthropology website, which was designed and developed from scratch and came online in late 2012. The website needs additional developments to improve user experience and to add new features, and it requires ongoing maintenance such as the periodic updating of style sheets, the codebase, or server configurations. We also need to create adequate documentation to prepare for the succession of managing editors and web developers.
Administration: With every new vendor or service, more work goes into administration. We are now responsible for maintaining relationships with a number of abstracting and indexing services, standards agencies, archives, and production vendors. We must also respond to customer service requests, whether from libraries with questions about the journal or from authors and readers having problems with the website or submission system.
The Cultural Anthropology staff was already responsible for more tasks than usual before we left our publishing arrangement with Wiley: as mentioned above, we developed our own submission management system, and we took control of copyediting several years ago because of the poor quality of editing provided by Wiley. However, of the areas listed above, Wiley took responsibility for the majority of the labor in production, infrastructure, administration, and dissemination.
When I started as the managing editor in the summer of 2013, I had seven months to figure out what pieces of infrastructure were needed to publish the journal and how they were going to fit together. What could we do in house and when would we need to go to outside vendors? What needed to be in place for the release of the first issue and what could wait until later? How do we decide what kind of final files we will need and how they can be designed to maximize current and future usability? In designing our XML schema, would there be later consequences if we created a customized document model rather than using the current standard document-type-definition (DTD), such as greater costs to update the files in the future? What would be the translation issues when we make HTML conversions?5 Although it would have been cheaper to forego typesetting, we know a large percentage of readers either print out paper copies of articles or use PDF files when creating their own digital libraries.6 So we want nicely designed PDFs, but also need functional, machine-readable files.
Then there is the whole publishing infrastructure that circulates metadata about the journal issues among indexers, search engines, library and retail catalogs, and other discovery systems. Tapping into this infrastructure entails establishing relationships with standards agencies and commercial indexing firms, and complying with their requirements. We are now dues-paying members of CrossRef, an organization that assigns and provides the infrastructure to cross-link digital object identifiers (DOIs).7 Our primary goal in obtaining DOIs for each article is to make sure it can be persistently located, even if its uniform resource locator (URL, what is commonly understood as a web address) changes over time.8 Having our articles cross-linked with other articles that cite them, however, also plugs our metadata into a network that enhances connections between pieces of scholarship and makes our content more discoverable. Getting our data systems up and running was a big part of the journal’s development.
Beyond the workflow and procedures for producing final documents and metadata, we had to figure out what internal developments were needed for our content management system and reading platform: our OJS installation and the Cultural Anthropology website. Open Journal Systems is not bad open-source software, but my sense is that its imagination of the basic form of journals, of who might work on them, and of how editorial labor is organized is too inflexible. The system takes some time to learn: users have to overcome its lack of documentation, and to understand its quirks and limitations. The new Cultural Anthropology website was originally designed to continue playing a supplementary role to the journal, as well as allowing us to experiment with forms of non-peer-reviewed content. When the decision was made that the journal would go open access, suddenly the website was no longer the supplement; it was going to be the reading platform and principle point of distribution. This required us to rethink what would appear online and how it would be organized. The website already contained supplemental pages for most of the articles published during Cultural Anthropology’s twenty-eight-year history. Because we had already put in the labor to produce these pages, our new article–page design could not diverge radically from the underlying data model of the supplemental page. As with any media organization—from publishers to libraries—legacy content places important limitations on what kinds of changes can be made to one’s platforms. Cultural Anthropology is in a different position than a newly launched journal because we have a large catalog of back content: for example, we must focus more of our development efforts on information architecture, and on searching and browsing functions, which would not be big concerns if we only had a few volumes worth of content.
In ten months, we were able to relaunch the journal with the release of the first open-access issue in February 2014. This was an important achievement, but it needs to be put in perspective. In the rush to get our systems and workflows up and running, we had to skimp on certain technology developments and push other important projects down our priority list. Some of this leaves us vulnerable to even predictable problems. We do not have an IT department supporting us; currently, we have a freelance web developer who has a full-time job and works on the website in his spare time. This means we have a lot less flexibility when problems crop up. For example, there was a typo in the subtitle of one of the articles in the February issue as it appeared on the website. You would think it would be easy to fix one character, but sometimes it isn’t. The way we set up our systems, I place corrected metadata into OJS and then feed that metadata to the website. But we were experiencing problems with the metadata feed from OJS, which made it impossible to make the correction. It took us several days and the intervention of our web developer to fix the typo. If the same problem had occurred before our initial metadata feed to the website, the release of the issue might have been delayed.
This is one of the hardest parts about setting off on your own. When a large commercial press such as Wiley adds another journal to its collection, it will in all likelihood already have a well-tested platform to sit on. The cost of building that platform and the rest of its publishing systems is spread across a huge number of journals, which means that the company can afford dedicated technology and infrastructure services. When a large commercial publisher creates or acquires a new journal, adding the journal’s content is relatively easy (although we should not underestimate the labor still required). For the SCA, everything is being built anew, and our systems are breaking all the time. It could take us years and a whole lot more investment to fix all the bugs in our systems, and that signals a level of vulnerability we did not have when we were publishing with Wiley. It also suggests that large infrastructure projects, like the development of a complex website/reading platform, might require the investment of more than one AAA section.
MT: You have made the argument that open-access advocates do not acknowledge that there is this full slate of possible consequences of open-access publishing, which could be good or bad. What were you getting at?
TE: We have a couple of very particular problems in mind when we look at open access as a solution. It is overwhelmingly focused on access to documents as the main problem. Discoverability and persistence are, in my opinion, as great a problem as accessibility. I can put text up on the Internet and call it published, but if nobody can find it, does it really matter?9 What if an author cites a document with a link, but then the URL changes or the document disappears altogether?10 For some types of writing and informal publishing this might be fine, but the more formal the venue, the more responsibilities a publisher must take on. Much of my job is not just figuring out how to display documents online but how information about them is going to circulate, where they will be discovered, how to make sure they have a life beyond their publication date, and the whole infrastructure that supports this. The extent to which Cultural Anthropology takes on all of these responsibilities will determine whether we are any good at publishing. Our level of accessibility to readers is only one measure of the full job of being a publisher.
Not all pieces of scholarship need the entire range of practices in formal publishing. For good reason, we usually reserve the full treatment, both on the editorial and publishing end, for scholarship that we expect big things from. As long as everyone is aware of the trade-offs, let a thousand publications bloom. My worry about the possible consequences of open-access publishing comes from a sense that our focus is too narrow and our understanding of publishing not deep enough. We are already making trade-offs and compromises, whether we are conscious of them or not. As things currently stand, Cultural Anthropology is not providing the same level of discoverability as Wiley, both in terms of indexing coverage and of our website’s search functionality, and we are more vulnerable to service outages and other significant platform glitches and disruptions. For the moment, the open accessibility of the journal trumps these measures of our quality, but how we design our publishing platforms and documents, the kinds of data management practices we adopt, and our ability to maintain and improve our systems will have long-term consequences.
Possible outcomes, and strategies for influencing them, can only come to the surface if we start to take the material practices of publishing more seriously. My hope is to encourage the more wide-eyed open-access advocates to curtail their idealism: not their ethically driven ideals for a more equitable scholarly communication system, but the notion that publishing trades in ideas instead of in documents, data, and the infrastructure through which they travel.11 A critical anthropology of scholarly communication, along the lines envisioned by Jason Jackson and Ryan Anderson (2014), would need to tackle the materiality of scholarly publishing if it is going to help fill this significant lacuna in our understanding of knowledge production.
MT: How are you approaching the problem of archival storage? One hundred years from now, print is going to be accessible, but not necessarily a PDF.
TE: The nice thing about a print book or journal is that once it is printed, it is in a pretty good archival state—the object is going to be around for a while—whereas with digital files, there is no end to maintenance. Knowing that maintenance is going to be an ongoing issue, we have to design our documents, data, and systems to be as easily upgradable as possible for the day when we update the codebase of our website, or the next version of XML comes down the line. There is an enormous amount of labor required every time you upgrade or move your data to a new system, and the labor needed grows in proportion to the size of your catalog of content. I was lucky enough to arrive at the SCA after the new website had been launched and Ali Kenner and her team of interns had moved all of the content from the old website (Kenner 2014). We will only see how well we were able to incorporate data portability into the design of the current website when the journal inevitably moves to another home. I have had more of a hand in the specifications for the designs of our documents. The whole point of using a standardized XML schema was to increase the chances that there will be more, and cheaper, options for upgrading when needed. Our archival strategy is not based on the PDF files readers see but on our XML files, because the machine-readable XML can produce new galleys or new HTML files in the future. We are going to be updating files from here to eternity.
Archiving means making sure our content lives on, no matter what happens to our organization. We are very fortunate to be partners with the Duke University Libraries, who host our OJS installation and help with archiving.12 We are already looking into what archiving has been done in the two primary archiving services for electronic journals, Clockss and Portico. These are services that accept a journal’s files and keep them either in a dark archive that would open if something were to happen to one’s organization (publisher ceases operation) or in the case of an event that affects availability (servers go down).13 Most of the content of Cultural Anthropology is in one or the other of these archives. We will be filling holes in past coverage and making submissions for our new issues. When planning for the creation of new content, one has to plan for its preservation as well.
MT: The back catalog goes back twenty-eight years, but only the previous ten years’ worth of content is going to be made available on a permanent basis. What about the rest of it?
TE: Our next big project is getting that back content ready and onto our website. The AAA owns the PDFs and the final files of everything that has been published from their section journals. What Wiley delivered to us this fall was PDF files for the entire run of the journal, and XML files back to 2008. Getting back content ready means adding it to our database in the correct way, which takes several steps. From 2008 through 2013, we were using OJS as a submissions management system, not as a content management system, which is a much bigger operation. The problem with this shift is that much of the data in our records has not been corrected. By the time an essay has gone through the review and publication process to become a published article, the title may have changed, the keywords may have changed, the abstract may have been edited, the author may have moved to a new institution. We never had a reason to go back into OJS to update this information. A significant amount of labor has been expended on perfecting that data, and it ended up on the PDF and in Wiley’s content management system. We must take that corrected information and get it back into our system. When we get to articles published before 2008, there are no records in OJS, so we will need to create these and enter the metadata from scratch. One of the problems with OJS is that it is not designed for this type of operation, as it deviates from the typical submission–review–edit–publish workflow, which means that we will need to create new submissions for articles already published. Only after we have created and corrected all the article records, uploaded the final files, and built journal issues from them will we be able to publish the content on our website.
The AAA made the agreement with Wiley that the SCA would be allowed to publish ten years of back content. That is what we are aiming for. Once the content is up on our website, we will see what we can do with the rest of the files. I do not know if we are going to be able to post more or not. Although we have the PDFs back to 1986, it is unlikely that we will be able to create XML and HTML files for all of these volumes. Depending on the cost and complexity, we may not have HTML pages for anything earlier than 2008. We would have to use old, ugly PDFs to create new XML, and the files would undoubtedly require a lot of quality control—people proofing documents—to fix.
MT: With the transition to open access, have there been thoughts about changing the copyright, maybe going to copyleft or a Creative Commons model?
TE: I know that the initial push was for a Creative Commons license. My sense is that the AAA is not quite ready for this yet. Up until the week before the relaunch, we were trying to figure out our copyright language. All of the content on our website falls under a Creative Commons attribution, non-commercial, share-alike, license. The journal, however, does not use a Creative Commons license. The AAA still wants to retain copyright.14
For me, there is an open question here: Is there a value for us, as an association, to keep our content together through copyright? I would be a little worried if each journal had a different copyright regime, but perhaps my worries are misplaced. This is a discussion for the members of the association. I was happy, however, with the language that we got for what readers can do with our articles, as well as with the expanded rights granted in our new author agreement. Both strike me as being more expansive than previously. Our move to open access is requiring the staff of the AAA to think through the implications of this transition, to rethink what might have been settled practices. But we still have a long way to go before these questions are settled.
MT: Are you going to offer print-on-demand copies of the journal, and how is this going to be priced?
TE: My research into print-on-demand (POD) services has been extremely interesting, taking me into completely new areas of the publishing business. It is worth pointing out that the emergence of print-on-demand is a hugely important development in scholarly publishing, lest anyone think the Internet is the whole story. We were looking for a vendor that could provide a certain package of services: not just a machine to print copies of the journal but a good interface for uploading our cover and PDF files; a service that would distribute our metadata to all of the important retail, wholesale, and library subscription-agent catalogs; and a service with the ability to sell and ship books to as many parts of the world as possible. The major POD vendors are overwhelmingly oriented to the needs of self-publishing book authors who have very different requirements and concerns than a scholarly society with an established journal, which added to the complexity of finding the right vendor.
The two viable candidates for POD services were Amazon’s CreateSpace and Lightning Source, which is part of Ingram, the largest book distributor in the world.15 We signed a contract with Lightning Source. Our goal is to make print issues of Cultural Anthropology as globally available as possible because many libraries in South and East Asia still want print, and we have also received some requests from European libraries. The print-on-demand agreements are quite complicated: we signed contracts with Ingram printers in the United States, United Kingdom, and Australia, plus another contract for independent (non-Ingram) printers in Brazil and Germany. The Australian printer cannot sell anywhere but Australia and New Zealand, so if a Chinese or Indonesian library orders an issue of the journal, it will be sent from the U.K. And we are set up to sell individual issues, not yearly subscriptions that would automatically be printed and shipped, because we would have had to maintain subscriber rolls, accept payments, and coordinate with Ingram—tasks that are beyond our administrative capacity. All this takes more labor. So POD gets very complicated, but we are wading in.
The goal of the SCA is not to earn a lot of money from print: we mainly want to make sure issues are affordable for those who want them. The trick is pricing issues at a level where we can sell enough copies to cover the labor and material costs of running the service, and to do so with little initial sense of demand. Our retail price will be $12.95. We hope that SCA members who liked the print version of Cultural Anthropology will be able to afford purchasing POD issues. We understand that paper has affordances that electronic documents do not, and we want to make sure people can get the medium they want.
MT: Is there going to be a move to find other external funding sources?
TE: At our spring meeting, we will be convening another committee to do some strategic planning. We need to reevaluate the costs of publishing the journal and to weigh our options for possible revenue sources. At the moment, we have our financial reserves from the money we received from the Wiley deal and two main sources of revenue: our continued share of the revenues from AnthroSource, which is not insignificant, and membership fees for our section. My sense is that the membership fees will become more important, and we stand to lose some members because print copies of Cultural Anthropology are no longer a benefit.
Another small stream of revenue comes through our submission policy: If an author is a member of the AAA but not a member of the SCA, he or she must become a member of the SCA to submit to our journal. If an author is not a member of the AAA and wants to submit to the journal, he or she must pay a submission fee of $21 (half the cost of an SCA membership). We put this submission policy in place in part because we are trying to communicate to authors that we are no longer supported as we previously were by libraries through subscriptions: authors benefit directly from the activities of the journal, and we feel it is reasonable that they should be asked to help sustain us. If authors and readers understand what we are doing as a public good, then they need to support it. I use the word support, taking this directly from National Public Radio, which is asking for private support of a public good. The SCA is essentially saying to authors, readers, institutions, everyone, “If you think this channel of communications is valuable, you need to figure out how to support it.”16 We will need to look for other sources of revenue as well, whether grants, advertisements on the website, or through more assistance from departments, libraries, and university presses. No matter what, we are going to have to find more support to make this journal viable in the future.
6 Although I have taken many of Willinsky, Garnett, and Wong’s (2012) useful suggestions for improving the design and usability of PDF articles, I do not share their incredulity over some readers’ preference for paper copies of articles. See Sellen and Harper (2003) for the relative affordances of paper and digital versions of the same documents.
9 Discoverability is nicely defined by Roger Schonfeld as “the description or measure of an item’s level of successful integration into appropriate infrastructure maximizing its likelihood of being found by appropriate users” (quoted in Somerville and Conrad 2014; emphasis removed).
10 There are at least two issues here: whether a digital document continues to exist in some form online, and whether it can be located. On the problem of URL half-lives and citation link–rot, see Lyons (2005), Dimitrova and Bugeja (2007), and Sellitto (2004). To guard against these two problems, I have created WebCite submissions for my citations to content in blogs and other web publications without clearly evident archiving practices. This is a citing-author-initiated, as opposed to publisher-initiated, method of archiving web content that can be used when a relatively more persistent type of document (i.e., formally published article) cites relatively more ephemeral documents (i.e., blog entries). See the WebCite website, http://www.webcitation.org.
11 Michael Silverstein and Greg Urban (1996, 2n1) argue that “to confuse the mediating artifact and its mode of production (‘inscription’) for a text and the sociosemiotic process that produce it perpetuates a particular fetishized substitution.” This insight is as important for the material production of scholars’ own publications as it is for their analyses of the textual practices of ethnographic interlocutors. There are a number of ways of focusing on the materiality of the artifacts and infrastructures of scholarly communication. For example, see Frohmann (2004) and Hull (2012) for methods of approaching the materiality of scientific and bureaucratic documents, respectively; Coleman and Brunton (2014) for the overlapping analysis of hardware, users, and discourse in information and media infrastructure; Downey (2014) for ways of uncovering and understanding information labor; and Elfenbein (2013) for a material semiotic approach to the creation and use of metadata.
15 This should disabuse anyone who thinks we have returned scholarly publishing to its craft origins: we are more reliant than ever on multinational corporations to publish the journal, even as we have left our relationship with one corporation, Wiley-Blackwell. Among the services we rely on are Google’s suite of productivity apps and analytics, Amazon cloud servers, and Facebook and Twitter social media platforms.
2014 “Closer to the Metal.” In Media Technologies: Essays on Communication, Materiality, and Society, edited by Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot, 77–97. Cambridge, Mass.: MIT Press.
2012 “American Anthropological Association: Publishing Program Analysis,” version 1.3, September 26. Report prepared by the Chain Bridge Group for the AAA.
2014 “Making Media Work: Time, Space, Identity, and Labor in the Analysis of Information and Communication Infrastructures.” In Media Technologies: Essays on Communication, Materiality, and Society, edited by Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot, 141–65. Cambridge, Mass.: MIT Press.
2013 “Metadata Pragmatics: Toward a Unified Semiotic Framework.” Master’s thesis. University of North Carolina, Chapel Hill. Available through the Carolina Digital Repository, https://cdr.lib.unc.edu.
2014 “Inside Baseball with Tim Elfenbein, Managing Editor for Cultural Anthropology.” Savage Minds, February 11. http://savageminds.org/2014/02/11/inside-baseball-with-tim-elfenbein-managing-editor-of-cultural-anthropology. Archived by WebCite® at http://www.webcitation.org/6NSh1PleJ.
2014 “When Professional Society Publishers Take an Independent Path.” Scholarly Kitchen, February 3. http://scholarlykitchen.sspnet.org/2014/02/03/when-professional-society-publishers-take-an-indpendent-path. Archived by WebCite® at http://www.webcitation.org/6NhrlgyPW.
2012 “Liberating Cultural Anthropology? A Thought Experiment.” Anthropologies, no. 12. http://www.anthropologiesproject.org/2012/03/liberating-cultural-anthropology.html. Archived by WebCite® at http://www.webcitation.org/6NShUUhYk.
2004 Deflating Information: From Science Studies to Documentation. Toronto: University of Toronto Press.
2012 Government of Paper: The Materiality of Bureaucracy in Urban Pakistan. Berkeley: University of California Press.
2003 “Markup: XML and Related Technologies.” In The Columbia Guide to Digital Publishing, edited by William E. Kasdorf, 65–154. New York: Columbia University Press.
2005 “Persistent Identification of Electronic Documents and the Future of Footnotes.” Law Library Journal 97, no. 4: 681–94.
2014 “Reason, Risk, and Reward: Models for Libraries and Other Stakeholders in an Evolving Scholarly Publishing Ecosystem.” Cultural Anthropology 29, no. 2: 216–35. http://dx.doi.org/10.14506/ca29.2.03.
2012 “The State of the AAA Publishing Program: What We Have Learned and Next Steps.” Memorandum to AAA members, October 12.
2003 The Myth of the Paperless Office. Cambridge, Mass.: MIT Press.
1996 “The Natural History of Discourse.” In Natural Histories of Discourse, edited by Michael Silverstein and Greg Urban, 1–17. Chicago: University of Chicago Press.
2014 “Collaborative Improvements in the Discoverability of Scholarly Content: Accomplishments, Aspirations, and Opportunities.” SAGE white paper, January 14. http://www.sagepub.com/repository/binaries/pdf/improvementsindiscoverability.pdf.
2013 “Cultural Anthropology Will Go Open Access in 2014.” Cultural Anthropology website, March 11. http://culanth.org/fieldsights/142-cultural-anthropology-will-go-open-access-in-2014.
2012 “Refurbishing the Camelot of Scholarship: How to Improve the Digital Contribution of the PDF Research Article.” Journal of Electronic Publishing 15, no. 1. http://dx.doi.org/10.3998/3336451.0015.102.