Thursday, December 22, 2005

OA as instrumental good

Historically, Open Access (OA) has been viewed as primarily an issue for researchers in the sciences. Today, however, there is a growing debate about its relevance to the humanities. So when last week the Princeton/Stanford Working Papers in Classics (PSWP) site was launched it seemed an ideal opportunity to discuss OA in the context of the humanities with Josiah Ober, the professor of classics at Princeton who, in collaboration with Stanford's Walter Scheidel, created PSWP.

Q: I understand that the purpose of the PSWP initiative is "to make the results of current research undertaken by members of Princeton and Stanford Universities in the field of classics available in advance of final publication." Why? And who do you expect to download the papers?

A: The purpose is two-fold: first, to cut down the lag time between the completion of an article or chapter and its availability to the scholarly community (the lag can be considerable, especially in the case of edited volumes, which can proceed no more quickly than the rate of the slowest contributor), and to allow authors of work in progress to get feedback from other scholars.

My guess is that most downloads will be by students and professional classicists for use in their own scholarship.

Q: Which is the more important function of PSWP in your view: to enable authors to solicit comments prior to submitting articles to a journal, or to increase the visibility of research being done by the Princeton and Stanford departments of classics?

A: I regard soliciting comments as more important than publicity per se. I am perfectly happy to have the work of our two faculties receive extra attention, but that is not my main concern.

Q: Does the PSWP concept owe anything to the physics preprint server arXiv.org?

A: I don’t know that server. My model in initially thinking through the site was social science Working Papers (WP) sites, in the first instance the WPs of the Princeton Economics Department. Gene Grossman, a professor in that department, discussed with me the history and operations of his department’s WPs. When their series began, it was distributed in hard copy. The ease of distribution via the Internet has subsequently made it a viable experiment for a relatively small humanities department like ours too — since we have fewer resources to devote to this sort of enterprise.

Q: What issues are raised when putting up a humanities preprint site?

A: I guess I don’t know yet. It is meant as an experiment — we should have a better sense of the issues in a year or so. I don’t initially see why humanities WPs would raise issues that do not come up in social science WP series.

That said, humanities fields do not seem to have as clear a hierarchy of prestige journals. Getting papers accepted at the best journals can be extremely important in promotion and tenure decisions in social science departments; humanities departments still tend to be more book and chapter driven. But whether that will make any substantial difference I don’t yet know.

Q: You maybe saw the comment made by OA advocate Peter Suber: "More OA is better than less, so I applaud this initiative. But I must say that a classics repository for all classicists would be more useful than one limited to faculty from two distinguished departments." Does he have a point?

A: Sure he does. Our initiative is constrained by two considerations: quality control and resources. The papers on the site are not refereed. So, for example, we depend on personnel decisions at our two departments, and the advice of our departmental colleagues, with respect to posting graduate student work; and we need to provide some guidance to users of the site about the quality of work they might expect to find there.

The resource constraint is both a question of staff time and of server space: We don’t currently have the resources (or for that matter the expertise) to host a high-quality site featuring the work of the worldwide community of classical scholars.

Q: Do you envisage classics departments from other universities being invited to deposit their papers at your site in future?

A: We have left that as an open question. When we began this project we knew that we would have to control costs, yet did not have an accurate sense of how much resource it would entail. Before adding other universities we would have to look carefully at the feasibility. Based on the history of economics WP series, however, we might expect other universities to set up their own sites.

The long-term answer might be for a professional association to become the primary site for hosting classics WPs. Our goal from the beginning was to run a scalable experiment in the hope that it would lead to something of genuine value to those interested in classical scholarship, and the field of classics as a whole.

Q: Suber also commented "[A]n OAI-compliant repository would be more useful than a non-compliant repository." Was there any particular reason why you chose not to adopt an open metadata harvesting standard when developing the site?

A: This is beyond my expertise level. Basically, we went with what looked to be the best cost/benefit approach, with the goal being to get a useable site up quickly, using existing resources. We assumed that we would get feedback from various quarters after mounting the site.

Q: So you have built the service using standard web tools, rather than a specialist e-print solution like Southampton University's EPrints software, or MIT's DSpace?

A: Right, the site was designed and mounted by Princeton’s great departmental IT specialist, Donna Sanclemente, using standard web tools. Walter and I are full time teaching faculty without prior experience in hosting open access sites and Donna’s time for this project is limited. Since our goal was to get a workable site up quickly in hopes of getting helpful feedback, we felt it was more cost effective to go with tools we already knew. 

Q: Another point made by Suber was that a repository for both preprints and postprints would be more useful than a repository for preprints alone. Do you have any plans to include postprints? Would providing postprints raise any particular issues in the area of classics?

A: I agree it would be more useful. But I would be very hesitant to get involved in posting material that has been published under copyright not held by the author. This is not a classics-specific issue, but a general intellectual property issue. I have no idea of what sort of legal clearances we would need to get from copyright holders, or what sort of arrangement for payment to copyright holders we would need to make.

Q: In connection with PSWP, Brian Simboli, a librarian at Lehigh University in Pennsylvania, raised this question on the Liblicense mailing list: "What are the patterns in humanities fields regarding whether journal publishers will accept items if they have been previously published as working papers or 'preprints'"? You are perhaps aware of the Ingelfinger Rule, which has posed difficulties for those working in the life sciences who want to post preprints on the Web. Can you see similar problems for classicists?

A: Here, we are guided by the experience of the social science WP sites that were our models — in social science fields WPs have become an accepted part of the publication process. My assumption is that humanities will prove to be more like social science than the life sciences in this regard.

Q: You say the working model for you was that of the Working Papers sites pioneered by social scientists. Does that imply that you do not envisage PSWP as an "open access" initiative as such? Perhaps you are also not very familiar with the open access movement?

A: My own knowledge of the open access movement is that of a layman who has read a few news reports, and so I hadn’t initially thought of our project as an open access initiative as such. Our goal, as classics faculty, is to further classical scholarship. Creating a WP site seemed to us an interesting (because as yet untried in the humanities) and relatively low-cost way to do so. If along the way we can also promote open access — which I regard as a good thing on general democratic principles — that’s a plus.

Q: Certainly OA today is mainly viewed as a science issue. Why do you think that is so, and do you think that in the long term OA has less, the same, or perhaps even more relevance to the humanities than to the sciences?

A: Humanities scholarship is a small area compared to natural science scholarship. Humanities work tends to be comparatively low-cost (few big labs or big grants) and individually authored. OA's immediate benefits are perhaps harder to measure in the humanities than in the sciences. OA for humanities is still something of a “green field.” There are a number of OA experiments in the area of classics; see for example, the site being developed by the Center for Hellenic Studies.  I am optimistic that these and other experiments will help to keep humanistic scholarship a vital and important part of the overall intellectual scene.

Q: So is OA an inherently good thing for the humanities or just something that some researchers may find interesting to provide?

A: By calling something an inherent (rather than an instrumental) good, you set the bar pretty high. I tend to think of OA as an instrumental good, but it may be understood as an intrinsic aspect of inherent goods (freedom, democracy).

Again, I don’t see why humanities would be a special case — if humanists are producing work that is of value to others, that value is increased by its accessibility. OA is in this sense an instrument for delivering something that may be of inherent value.

Q: I get the feeling that researchers in classics have shown themselves more willing to jump into the Internet pond than those in other humanities areas (the Bryn Mawr Classical Review, for instance, was launched in 1990, making it the second oldest online scholarly journal). Do you think this is merely coincidental, or do classicists today tend to be more receptive to new ways of doing things than their colleagues in other areas of the humanities?

A: Well, it’s certainly true that classicists have been out there in the forefront of humanities computing for a long time: the Thesaurus Linguae Graecae was a revolutionary textbase (and remains among the biggest textbase projects going). Perseus began as a classics site.

So as a result, classicists have been pretty attuned to computing for a full generation now, and I would guess we are on average more computing-oriented than other humanists (though I don’t know of any actual statistics on that). It didn’t hurt that David Packard got his PhD in Classics and has been a big supporter of classics computing from the beginning.

Q: You are perhaps aware of the growing trend for science research funders to request/require that their grantees make their papers available on the Web (not least the National Institutes of Health's policy on Enhancing Public Access to Archived Publications Resulting from NIH-Funded Research). Can you see a logic for funders in the humanities adopting a similar approach, or is funding in the humanities sufficiently different that such an approach may not be appropriate?

A: Humanities funding is mostly personal grants for leave time to individuals, and humanists still tend to publish a lot of books. So I would guess that this sort of requirement would be unlikely in the near term.

Q: Certainly there are calls for this in the UK. The draft proposal from Research Councils UK (Access to Research Outputs), for instance, proposes mandating all publicly-funded researchers, including those in the humanities, to self-archive their papers on the Web. It's not yet clear what the final RCUK policy will be, but do you think this is how the future will look: that in the way that researchers have been required to "publish or perish" they will find themselves also being required to ensure that their papers are freely available to all on the Web?

A: My guess is that it will increasingly be the case that material that is not web accessible will be less likely to be consulted or cited, and that this will in turn result in a lot of pressure on authors and publishers to find ways to make academic work web accessible.

Yet the crunch is likely to come in quality control (to get back that issue). Academic publishing, personnel decisions, and indeed the entire modern enterprise of scholarship is predicated on detailed and careful peer review. That costs quite a lot (many hours of work by pretty highly paid people are involved).

The costs are now (in part) covered by the profits made by academic publishers. In current business models, those profits depend on ownership of intellectual property as defined by copyright law.

Q: Would you say that peer review is as important in the humanities as in the sciences?

A: Peer review is at least as important in humanities as in the sciences — maybe more so in that we do not run experiments that can be independently reproduced. The quality control issue seems to me to be a hurdle that needs to be got over in order for the OA revolution to be completed.

Q: Of course that assumes that OA and peer review are incompatible — a notion that OA advocates would strenuously reject. Commenting on PSWP, Margaret Landesman, a librarian at the University of Utah, suggested that the classics community maybe doesn’t need "formerly published" journals any more. As she put it: "Why not keep the preprint archives and dispense with the journals." Peer review could still take place, but the whole process could be done electronically — which is the Bryn Mawr Classical Review model isn’t it? Moreover, if authors retained copyright the need to obtain permission to archive postprints on the Web would go away. Does that make sense to you?

A: There are several issues here: one is the relative importance of journal articles as opposed to books or book chapters: As long as humanities remains strongly book-oriented (which may change, of course), shifting to electronic publication of humanities journals would be only a partial solution.

Bryn Mawr Classical Review is an important success story in electronic humanities publication, but peer review/quality control concerns are less demanding when publishing scholarly book reviews than when publishing scholarly articles.

The editorial work that ensures proper peer review for scholarly books and articles is very time-consuming. It must be done to a very high standard if the enterprise of scholarship is to be sustained. Following the Bryn Mawr model, universities (or professional organisations) might choose to provide senior scholars with appropriate incentives to do the necessary editorial work on electronic journals (or even books).

Universities work hard to cut “non-essential” costs. It is a matter of persuading universities and professional organisations that it is in their interest to assume costs that are currently paid for by the publishing market. That is a daunting, but certainly not an impossible, challenge.

Q: Does OA provide benefits to researchers that were not possible historically? If so, what benefits?

A: It seems to me that it a case of making things of value (the results of scholarship) readily available to a very wide audience of researchers (professional and amateur), rather than restricting them to the privileged elite that happens to have access to great research libraries.

Of course even those who do have access to major libraries may be encouraged to read more widely because the material is “right there” in front of them. So there is increased potential for cross-appropriation of ideas and facts.

Q: What future developments do you see for a) PSWP b) OA in the humanities in general?

A: I hope that our site will grow in the number of papers we offer and the number of downloads. Beyond that, I am keeping an open mind. If there is a substantial readership for the Princeton/Stanford Working Papers in Classics, then I would hope that it would promote similar (or better!) sites in other humanities fields — and thereby push towards the long term goal of making more and better scholarship available to bigger and more diverse audiences.

The extension of “what is potentially known” is, I believe, extremely healthy for democratic societies — indeed, that is the subject of my current research.

Q: Can you envisage value in, say, attaching blogs to specific articles, and encouraging feedback/comments etc.?

A: There might be value in that, but my very limited experience with the blogosphere doesn’t encourage me to move in that direction. Blogging at this point seems to be best suited to matters in which opinion is of greater moment than expert knowledge; in matters of scholarship, where expertise does have a big role to play, I’d prefer to leave it to individual commentators to contact individual authors.

Q: Finally, do you think that OA is "inevitable and optimal" as OA advocates often put it? If so, what are the compelling reasons for arguing that that is the case in the area of the humanities in general, and in classics in particular?

A: Inevitable is a strong word and one I tend to avoid when speaking of social phenomena, like politics, economics, or OA. I think that there will be strong pressure for more OA over time, as students and scholars become more and more used to doing their research online. Optimal must of course depend on implementation.

So I’d say that OA is a very good bet, and stands to be a lot closer to optimal than any prior information regime I know of.

Humanists, whose work can be made readily accessible across disciplinary lines (in a way that is more difficult for highly mathematical fields), are likely to benefit disproportionately from OA — and should, therefore, have every reason to support it.

No comments: