Monday, June 22, 2009

Open Access and the A-Bomb

Many have wondered why the first scientists to embrace Open Access (OA) were physicists.

That physicists were the OA trailblazers is not in doubt: it was, after all, theoretical physicist Paul Ginsparg who in 1991 created the seminal physics preprint repository arXiv; and today physicists as a matter of course deposit their preprints in arXiv before sending them to a publisher.

This has seen arXiv grow to the point where it currently hosts over half a million papers, with roughly five thousand new ones being added each month.

Importantly, by putting their papers into arXiv physicists ensure that they are freely available to anyone who wishes to access them – assuming they have an Internet connection – regardless of whether they or their institution has a subscription to the journal in which the paper is published. Indeed, some papers in arXiv are never published in a journal at all.

Economists and computer scientists aside perhaps, no other discipline can claim to have embraced OA as enthusiastically as physicists.

But why?

Long before the Internet

Maybe because physicists have been sharing paper preprints with one another for decades? OA advocate Eberhard Hilf tells me that this began as long ago as 1932, when the Italian Nobel Laureate Enrico Fermi started to routinely mail preprints of his papers to colleagues prior to publishing them.

Fermi is noted for his work on the development of the first nuclear reactor, and for his contributions to the development of quantum theory, nuclear and particle physics, and statistical mechanics.

In this light, arXiv was simply a digital manifestation of a practice that began long before the Internet.

On the other hand, as Annette Holtkamp – an information professional at Germany's largest particle physics research centre DESY – pointed out to me last year when I interviewed her, the preprint culture is principally a habit of particle physicists.

Photon scientists, for example, still rely entirely on journals, she told me; and even some within the particle physics community – e.g. accelerator physicists – remain much more dependent on conference proceedings.

But why particle physicists?

If asked the question, particle physicists will usually say that their field is too fast-moving for them to want to wait for a paper to go through the lengthy process of being published in a journal before sharing it with colleagues.

The world's first mandate?

Holtkamp offers a different explanation. Particle physicists, she told me, are “a small closely-knit community and the authors and readers are practically identical. The community has also always been part of a very international enterprise, and worldwide collaboration has long been the norm. This has meant that the rapid long-distance exchange of information has always been crucial — something that is not as important in many other disciplines, which becomes obvious when you compare citation histories of typical papers in different disciplines.”

That makes sense. But recently I have begun to wonder if there could be an additional reason.

A couple of weeks ago, for instance, I had an email conversation with Jens Vigen head of the Scientific Information Service at the world's largest particle physics laboratory CERN, as a result of which Vigen sent me a copy of a memo dated 17th March 1955 and signed by the then Director General of CERN. The memo instructed CERN employers to send copies of all the technical papers they wrote to the library, with a view to making them publicly accessible.

I asked Vigen if he thought the memo could be classified as the world’s first OA mandate. He replied, “It’s hard to say. I guess at the time no universities would have had such mandates in place (I think it would have been considered a threat to their academic freedom), but there could be other labs that would have similar things in place.What about Los Alamos?”


So I emailed Miriam Blake head of the library at the Los Alamos National Laboratory (LANL).

Based in New Mexico, LANL was one of the secret laboratories and production sites that jointly developed the atomic bomb during World War II – as part of what was known as the Manhattan Project.

Blake was kind enough to ask one of her colleagues – LANL librarian Michelle Garcia – to see if she could find any reference to an OA mandate in the Los Alamos archives. Blake cautioned me however. “With the history of Los Alamos, I suspect it may be less lucid than the CERN directive. But obviously at some point scientists here began to openly publish.”

A few days later I had an email from Garcia. There was no mention of a mandate in her message, but she did send me something of greater inherent interest: a link to the 1945 Smyth Report.

The Smyth Report, Garcia explained, is “the earliest example of any kind of acknowledgement on the need for public release of information specifically on the development of atomic energy by the US government. Following the Smyth Report, there was a declassification program headed by a committee of senior scientists that led the Manhattan Project, which came up with the declassification guidelines in 1946.”

Sufficient information

Essentially the Smyth Report is an official account of the development of the atomic bomb in the US. Its stated aim was to give US citizens sufficient information to enable them to understand enough about atomic weapons to make sensible policy decisions about their use, and the use of the technology underlying them.

As the preface to the Report puts it, “The ultimate responsibility for our nation’s policy rests on its citizens and they can discharge such responsibilities wisely only if they are informed.”

One could perhaps argue that the Smyth Report was an acknowledgement of what in today’s OA parlance might be referred to as “Taxpayer Access”.

As the US-based Alliance for Taxpayer Access (ATA), puts it, for instance: “American taxpayers are entitled to open access on the Internet to the peer-reviewed scientific articles on research funded by the US Government.”

On the other hand, as we have seen, the Smyth Report stressed that scientific information should be released to the public not because its creation had been funded by taxpayers, but because it would enable them to make informed decisions about how the science should be used.

The Preface to the Smyth Report did however acknowledge: “The average citizen cannot be expected to understand clearly how an atomic bomb is constructed or how it works but there is in this country a substantial group of engineers and scientists who can understand such things and can explain the potentialities of atomic bombs to their fellow citizens.”

We should note that the Report was released to the public on August 12th 1945 – just days after the atomic bombs were dropped on Hiroshima and Nagasaki (August 6th and 9th).

Did the public want to be informed on the topic? It certainly seems so: Wikipedia reports that when it was published as a book the Smyth Report sold 127,000 copies in its first eight printings and was on the New York Times best-seller list from mid-October 1945 until late January 1946.”

At the time, some US politicians complained that the Report had “given away the secret of the A-bomb”.

Since then the Smyth Report has been translated into 40 different languages.

Furthering the national welfare

There was, however, a further argument used to justify the release of the information in the Smyth Report. This too is interesting in the context of OA.

In an article published in the November 1946 issue of Bulletin of the Atomic Scientists (Pages 14/15), for instance, Lt. Col. Hutchinson explained how an army officer had described the value of the declassification process in the following way:

This war was won by teamwork and the two and one-half ton six wheel drive. This makes the cargo truck a military weapon. Suppose that after World War I the government had declared that all future research directed toward improving cargo trucks was to be conducted in strictest secrecy. If this had been done, maybe we wouldn’t have had such a good cargo truck by the time World War II came around. In science the more contributors there are, the more progress will be made. We can’t release everything concerning the atomic bomb, but we can release a lot of valuable information that will help American science to advance from here, particularly in applying wartime developments to peacetime use.

Hutchinson concluded: “Thus the Manhattan Project is furthering the national welfare by releasing scientific information where this can be done without danger to the national security.”

This argument would doubtless be endorsed by OA advocates like Stevan Harnad, who insist that the only valid reason for making research OA is that doing so will allow scientists to maximise the impact of their work.

This view says that OA has nothing to do with providing the public with access to scientific information, but other scientists – so that they can build on one another’s work, to the benefit of themselves, to the wider research community, and to the world at large.

Default position

Back to the question of why physicists were the first to embrace OA: Could it be that the US atomic weapons declassification program helped create the preprint culture characteristic of the particle physics community?

In other words, in being asked to think through the reasons for and against making their research freely available, could it be that physicists became acculturated into assuming that the default position should be one in which scientific information is made as widely available as possible, as soon as possible – on the assumption that in most cases the benefits far outweigh the disadvantages?

Far fetched? Maybe. But it reminds us that we are entering a world in which it will be expected that more and more information will be freely available – be it scientific research, or be it the details of British MPs' expenses – unless someone can give a damned good reason why it shouldn’t be.

Of course the tragedy is that in the wake of the declassification program a growing amount of the world’s research was nevertheless to become inaccessible to a growing number of people – for a very different reason.

That is, the subscription costs that have to be paid to buy access to scholarly journals have constantly risen, to the point where today even the world’s richest universities are struggling to afford them, depriving both scientists and the public of access to more and more of the research corpus.

This of course is not a consequence of government restrictions, but of the financial restrictions arising from scholarly publishers seeking to maximise their revenues. It was partly in an attempt to find a solution to this problem, of course, that the OA movement emerged.


waltc said...

A slight clarification: Gold OA journals go back well before arXiv. New Horizons in Adult Education began in 1987. Public-Access Computer Systems Review began in 1989, as did Journal of Technology Education. There are others...

Stevan Harnad said...

Except that the first ones to make their research papers open access were not physicists, in 1991, but computer scientists, thoughout the 80's, ever since they invented unix, UUCP and FTP archiving...

Stevan Harnad said...

PS: citeseerx , the computer science paper harvester, although it only began harvesting in the late 1990s, is already almost three times as big as arXiv and, because computer scientists have been self-archiving longer, citeseerx also reaches further back than arXiv (as Google Scholar does, likewise by harvesting local deposits, rather than waiting for direct central deposits)...