Article Versioning: The Reality on the Ground

Column Editor:  Greg Tananbaum (Founder and CEO, Anianet)  <[email protected]

“The reality on the ground” is a phrase I have lately appropriated to separate what is actually happening in our industry from the incredibly nuanced but often largely hypothetical discussions we observe on Liblicense or the Scholarly Kitchen.  We are blessed to have any number of big thinkers in the scholarly communication space — Joe Esposito, Toby Green, and ATG’s own Sandy Thatcher spring to mind — but we don’t necessarily excel in reportage.  This means we can debate the implications of Green vs. Gold OA, for example, using incredibly complex and well-reasoned arguments, but we are less successful when it comes to talking about what these theoretical mean in practical application.  What is the reality on the ground?  What is actually happening, as opposed to what should happen or what might happen?

When I learned that this issue of Against the Grain would be dedicated to the subject of journal article versioning, my first thought was “Who cares?”  Now, to be clear, I wasn’t asking myself why anyone would be interested in this subject.  Rather, I truly wondered if this was an issue that mattered to publishers and librarians, but not to rank-and-file researchers.  What is the reality on the ground for this population?  We in the academic publishing world devote nontrivial energies to this subject, but should we?  It seems to me that if scholars themselves don’t particularly care about provenance or versions of record, then this is energy misspent.

With that in mind, I called up three researchers I know.  One is a senior professor in the social sciences.  The second is a mid-career professor in the humanities.  The third is a mid-career professor in the physical sciences.  They generously allowed me to pick their brains on the subject of journal article versioning.  The results, of course, offer no statistical significance, but they do speak, at least anecdotally, to the reality on the ground.

My first question was blunt — Do you care whether the paper you read is the version of record or some other version?  The responses ran the gamut, with the humanist expressing deep concern that any non-definitive version could include subtle errors or differences that might impact the substance of the article.  The physical scientist prefers the version of record because it simplifies the citation process, but is happy to use non-definitive copies if he is simply reading a paper for informational purposes.  The social scientist just wants to read a paper, and to ensure that anyone who wants to read it can have access to its substance.  In that sense, the version of record is not important to him.

Given the era of tight library budgets, how would they feel if their institution were to cancel subscriptions to a journal because its contents could be acquired on the Web for free in non-authoritative form?  The social scientist, consistent with his prior answer, would not care, provided he had the ability to cite the paper properly in his own writing (see next question).  The physical scientist indicates he would make due, though he might call upon colleagues at other institutions or the author him/herself to send the version of record on occasion.  The humanist would be the most resistant to this change, promising to “raise a fuss like an old grandma at a buffet when they run out of the expensive stuff.”  In his opinion, the lack of access to definitive content would be a significant disadvantage in his own research and writing.  It would make the authoring process less efficient for him compared to his colleagues at other institutions.  He fears that, in a publish-or-perish environment, such inefficiency could (not would, he is careful to note, but could) substantively damage his career prospects.

Digging a bit deeper, I next asked, “If you had access to a copy of a paper that was not definitive but was substantively the same as the published version (e.g., a postprint), would you use it for your own research or your teaching?”  Here, all three professors responded similarly.  They strive to use the definitive version of an article for outward-facing activities such as lectures, syllabi, and citations.  They do so because they hope to ensure the author gets full and proper credit for his/her work.  All three acknowledge that the current tenure and promotion system relies heavily on publishing high-impact articles.  As such they do not want to undermine the professional prospects of an author whose work they admire by mis-citing him/her.  So while they may or may not care what version of a paper they read, they all want to give authors their due by passing along the version of record.

Finally, in large part because I, like Whitney Houston, believe the children are our future, I asked the three scholars how they would advise students beginning their own research careers if asked about the acceptability of reading and referencing non-authoritative versions.  The humanist feels most strongly that the version of record is what should be both read and cited.  His belief is that this is the one copy of the article most likely to be unaltered both today and into the future, and, therefore, the one copy to which all readers can uniformly relate, as it were.  The social scientist and the physical scientist were less dogmatic.  They acknowledge that if a paper is readily accessible in non-definitive form that this convenience may outweigh the possible negatives of relying on something other than the version of record.  They also both teach their students the distinction between reading and referencing.  For them, accurate citations to the definitive version matter, a lesson which they pass along to their pupils.

My focus group of three does not put the issue of article versioning to bed once and for all.  What it tells me, though, is that three professors I trust actually know about this issue.  They have thought about it independent of the questions I posed to them.  And they care about it as it fits into larger questions of access and long-term citation trails.  The reality on the ground, if we are to extrapolate from this gang of three, is that article versioning matters not just to publishers and librarians, but to the constituents we serve.  They balance this concern with a desire to actually get at the content in a quick and painless manner.  As we continue to debate this issue at our industry conferences, in our blogs, and on our listservs, we should be mindful of the fragile equilibrium between accessibility and provenance that informs the reality on the ground.