What About Open Access to Books?

Les Carr (U.Southampton) wrote:

"Do ERC (or other short-term funders') research projects result in books? I am only an engineer who gets a bit lost outside STM, but I thought that books were written independently by researchers and that funded research projects had papers (and similar low-investment texts) as explicit research outputs?

NOTE: I am not asking whether books count as research outputs (they do) but whether they are the outputs of funded projects. I'll confine the scope of the question to single-author books, rather than multi-author books or edited collections."

(1) I would be very surprised if it were not the case that (in some disciplines at least) books count as the outputs of funded research. (Book citations certainly redound to an author's research credit as surely as article citations do.)

(2) Insofar as OA (and Green OA self-archiving mandates) are concerned, however, the relevant question is not whether books count as the outputs of funded research. (OA is for the outputs of research, whether or not the research is funded. And Green OA self-archiving mandates apply to the research output of a university's salaried academics, whether or not their research receives external funding, just as the university's publish-or-perish mandate applies to publications irrespective of whether they are the result of external funding.)

(3) Another way to put this is that even an academic who receives no external funding is institutionally funded to do research, inasmuch as research and publishing are part of his job-description.

(4) So the relevant variable is not funding but whether the research publication is an author give-away, written purely for the sake of research uptake, usage and impact -- the way all peer-reviewed articles are written -- or whether it is also written in the hope of royalty income (as many books are -- even though their hopes are usually not realized!)

(5) Perhaps trumping even the impact vs royalty question is the question of the cost of publication, and with it the question of whether a print run of the publication is desired.

(6) For better or worse, books are still preferentially published and read as conventional print-runs, rather than online-only, plus local print-offs by users.

(7) As long as that is true, the essential costs of producing and distributing a print-on-paper book will differ from the essential costs of producing and distributed a journal article (which can all be done online).

(8) Those essential costs of book publication need to be recovered regardless of whether the author hopes for royalty revenues over and above them.

(9) Some have suggested that making a book OA online will not hurt but help the sales of the print edition, but this is far from empirically established as the general rule (although it has happened in a few cases).

(10) Hence, although funders and institutions can and should mandate the self-archiving of peer-reviewed articles, they cannot and should not mandate the self-archiving of books.

(11) If it were proposed to extend Green OA self-archiving mandates to books, there would be (justified) resistance from both authors and publishers, and that would needlessly reduce the chances of adoption of what would otherwise have been an articles-only mandate.

(12) Once the 2.5 million articles published annually in the world's 25,000 peer-reviewed journals have been made OA by universal Green OA self-archiving mandates, the number of books and publishers that show an interest in pursuing a similar option will no doubt increase -- but that's not the same as subsuming books under the Green OA self-archiving mandates themselves.

Klaus Graf wrote:

"Stevan Harnad wrote: '(9) Some have suggested that making a book OA online will not hurt but help the sales of the print edition, but this is far from empirically established as the general rule (although it has happened in a few cases).'

"Looking at the evidence, it is far from empirically established as the general rule that a book OA online will hurt the sales. Please [cite] any valid articles or research confirming your prejudice. If you are stating a general rule you have to [prove] it."

Much as I may wish it were otherwise, Klaus Graf unfortunately has it exactly backwards:

The burden of proof is most definitely not on those who think that making books free online will hurt sales, to provide evidence that it is so. The burden is on those who think it is not so, to show that it is not.

The default or null hypothesis -- not just in this instance, but in the much more general one, of which books are just a special case -- is that, ceteris paribus, yes, if you make a digital version of a product available free for all online, you will hurt its sales (digital and analogue). There may be exceptions, but they have to be demonstrated.

And evidence is a tricky matter, especially for a negative hypothesis. It is not sufficient evidence that a platypus does not lay eggs, to show photos of some platypuses, not laying eggs.

For journal articles, although the evidence so far seems to be that Green OA self-archiving has not caused cancellations, it remains a possibility that it eventually will. However, as has been pointed out repeatedly, in that very special case, it does not matter, either way:

i. because research usage and impact is far more important than sustaining the current journal-publishing cost-recovery model;

ii. because authors, their institutions and funders are unanimous in insisting on the priority of research usage and impact; and

iii. because Gold OA journal publishing is there to take over if and when the subscription model should become unsustainable.

But none of that carries over to books and book publishing -- neither to books in general, nor to scholarly/scientific books in particular.

(a) It is not true that books are mostly written only for research usage and impact. (Most are also written in the hope of royalty income.)

(b) It is not true that book authors all or even mostly want to give them away free online.

(c) It is not clear that book authors or readers no longer desire a paper edition.

(d) It is not clear that there is a viable, sustainable Gold OA publishing model for books yet, even if authors did want to give their books away and no one wanted the print edition any more.

(e) And, most important, it has not been shown that giving away the online edition will not hurt print sales (or even make them unsustainable); it has only been shown not to hurt some books' print sales, so far.

Neither for journal articles nor for books can free online access be attained through wishful thinking and righteous indignation alone. Fortunately, for journal articles, it needn't be.

Let us not, then, needlessly handicap the strong special case for OA -- which covers all of peer-reviewed journal articles and authors without a single exception -- with the unneeded extra baggage of the uncertain, untested and equivocal case of books. Some books may eventually go the way of OA too; but right now, when the research community is finally on the verge of successfully inducing its funders and universities worldwide to adopt Green OA self-archiving mandates, this is definitely not the time to try to change the rules, raise the stakes, and insist on mandating book-deposit too.

Leave book-deposit as an author option, like access-setting itself, and give OA, already so grotesquely overdue, the chance to come into its own at long last.

Klaus Graf wrote:

"Is there any empirical proof for this default hypothesis? There is only empirical evidence for the contrary. It's purely nonsense to state a "default hypothesis" if empirical facts should be given."

I am afraid that Klaus Graf has not understood what is meant by the default or null hypothesis: To demonstrate [proof is possible only in mathematics] that the null hypothesis is false you need to provide compelling evidence to the contrary. The null hypothesis is that vitamin C does not cure cancer. The evidential burden is on those who think it does, to provide compelling evidence that it does, not on those who think it doesn't, to provide evidence that it doesn't. (Most things do not cure cancer, by default. No one has to give evidence that vitamin C does not cure cancer.) The occasional report that some people who were diagnosed with cancer took vitamin C and are still alive is not compelling evidence. The jury is still out on vitamin C. It has been out for a long time.

(Apologies for the shrillness of the example. I did not want to use miracles or telekinesis, because that would be too exacting and dismissive. I just wanted to give an illustration of the evidential burden on the null hypothesis with a case where -- some believe -- closure has not yet been reached. Another example would be the legal presumption of innocence until compellingly "proved" guilty: There is no burden to "prove" innocence: it is the default hypothesis.)

As with vitamin C, with book sales it is not sufficient merely to cite self-selected positive evidence; it is necessary to do systematic controlled comparisons, on sufficiently large and representative samples: N new books, selected at random, half of them then made freely accessible online, half not, and then a comparison of their sales for several years. Perhaps even a replication or two...

Pablo Ortellado wrote:

"You may call it OA or not, but books in many cases should be mandatorily made available on the Internet..."

Mandatorily "in many cases"? How many? And which? and who decides, how?

Please take a moment to reflect.

The substantive issue is not what we do and don't call "OA."

The issue is what we can and cannot consensually mandate (and what a mandate is).

After much too long a delay, the momentum is finally gathering for funders and universities to adopt Green OA mandates to deposit all peer-reviewed research journal articles in OA repositories. (Not "in many cases": all, without exception: that's why/how it's a mandate.)

The reason those mandates proved possible was that all the authors of all those articles (as well as their universities and their funders) without exception, wanted to give away those articles for free online.

None of them sought royalty revenues or print sales -- they sought only maximal research impact.

None of this is true without exception, or even majoritarily, of books. It is true of some authors of some books. And those authors are all free to deposit them in their OA IR if they wish.

(Not only are the same OA IRs there, ready for books to be deposited into them too; they can even be deposited IDOA, if the author wishes: Closed Access but with the option of emailing one copy to any requesters the author approves.)

And as with Gold OA journal publishing, book publishers are more than welcome to experiment with offering free online access, as National Academies Press does, extremely successfully and valuably.

But if you insist on including books in the deposit mandates, you will simply prevent the adoption of the mandates themselves, because they were predicated on consensus among authors, their institutions and their funders that their articles were intended as give-aways all along (even before the online era). There is no such consensus on books (in fact, I suspect, far from it).

Hence on no account should we needlessly jeopardize the spread of Green OA self-archiving mandates today, just when adoptions are at last beginning to gather speed, by raising the goal-posts, this time to a height for which the research community can no longer sustain its natural consensus, by now declaring that book deposit is to fall under the OA mandates too.

Stevan Harnad
American Scientist Open Access Forum