Create Change

Shouldn't the way we share research be as advanced as the Internet?

Scholars Speak


David R. Morrison
University of California, Santa Barbara

David R. Morrison
David R. Morrison

David R. Morrison is a professor of mathematics and physics at the University of California in Santa Barbara. He received his undergraduate degree from Princeton University and his master’s and doctorate at Harvard University. Before joining the faculty at Santa Barbara, Dr. Morrison was the James B. Duke Professor of Mathematics and Physics at Duke University. Recent honors have included a Mathematical Sciences Research Institute Research Professorship in 2006 and a John Simon Guggenheim Memorial Foundation Fellowship in 2005-2006. He served on the editorial board of Communications in Mathematical Physics from 2002-2003. Currently, he serves on the editorial board of Advances in Theoretical and Mathematical Physics and the New York Journal of Mathematics. He is also on the Advisory Board of Project Euclid, a Cornell University Libraries initiative to offer digital editions of independent and society journals in mathematics and statistics.

Your fields—math and physics—were among the first to recognize the possibilities of the Internet. The original online archive, arXiv, was born here. How did arXiv come about?

Through the 1980s, the field of high-energy theoretical physics had developed this system of circulating preprints among researchers, copying them and mailing them to departments. It was a way of more rapidly communicating scientific results than waiting for a journal publication.

In the early 1990s, people preparing preprints electronically started to realize: Hey, I can get it even faster by email. So in the first phase, they emailed preprints to their colleagues. Then, a system was set up by which people who had new preprints gave them to one person, who then emailed them out. Instead of having the emails flying all over the place, there was this one central source. In the summer of 1991, Paul Ginsparg supposedly said ‘This is very inefficient. What we need is something like a server. I can develop one in an afternoon.’ It took him a couple of afternoons, but he got one running and it started in August 1991.

Personally, I had become interested in some problems in string theory about five months before this. I was trying to read this new literature and get current with it. I was incredibly enthusiastic about arXiv because it provided a way for people who were not central figures in the field to stay in touch with what was going on. It’s a really democratic way of communicating information. Instead of having to be on the right list, you are able to access whatever anybody was doing. In the first few months, it really caught on.

How has its role evolved over time?

ArXiv was an email-based server in the beginning. The early years featured many technical advances in how it functioned, including a move to the Web. In the beginning, you emailed the server a request for a paper and it emailed the paper back to you. There wasn’t a Web. Otherwise it hasn’t changed its basic operation since the mid-1990s.

What’s really changed over the years is the number of fields in which arXiv has caught on and become a major part of scientific communication. A number of people approached Ginsparg with requests to expand arXiv into their own field. I was one of those. I set up a branch of arXiv in algebraic geometry in February 1992. At the same time, in different fields of mathematics there were various subject areas with other preprint servers. In math, we convinced a lot of these different preprint servers to join forces and create a unified math branch of arXiv. That went into operation on January 1, 1998. In math there have been increasing participation levels ever since them.

How have the Internet and arXiv changed the way mathematicians and physicists research and communicate? What have been the benefits?

The Internet itself has really enabled a kind of long-distance collaboration that was difficult to do before. I wrote a paper in 1989 in which my collaborator was on sabbatical in Germany. We were so pleased to email drafts back and forth. This notion of making progress by instant communication of long technical documents, which could be revised by either person—that was really a major advance.

It’s more than just the availability, it’s being able to search and find things quickly.

At the beginning of my career, I often wanted to know something about some area of math that I did not know too well. I was good at using the library reference tools. I could spend an entire afternoon tracking down something that now I can find really easily. The very fact that a lot of important bibliographical tools have become electronic is a big advance and really streamlined the way we work. These days I can sit at my desk and access—often with arXiv or electronic versions of journals—much of the literature that I need to look at with a couple of mouse clicks.

Has arXiv had an impact on journal publishing?

The main effect so far as been to keep the publishers on their toes. There were people who believed in the beginning when arXiv was catching on that this was going to completely replace the journal model of publishing. And that hasn’t happened. The reasons are complicated.

Journals saw this kind of activity as a threat best countered by improving their own electronic services. They really have done this. In high-energy physics, the major publishers have come out with electronic products which help their published products compete in the electronic playing field. Of course, they are not free. Most researchers using them are not aware that their university library is paying huge bucks to the publisher for the use of these resources. I think if that hadn’t happened, then people might have stopped using the journals.

The other thing that made the transition difficult is that, when academics try to evaluate each other, publication lists and the reputations of the journals in which the scholar is publishing are quite important. Although a field such as physics or some parts of math may decide that there are better measures of scholarly productivity, the university as a whole has to reach a consensus on that point before it becomes possible to replace the traditional evaluation methods. For example, to convince my colleagues in the English department that even though my papers weren’t published in journals, they are still meritorious, can be quite difficult when the standard in the humanities is “getting that book published.” I think eventually that the academy will come to recognize many different ways of evaluating scholarly productivity, which will decrease the necessity of journals.

What do you think is the outlook for journal publishing?

I’ve been predicting the demise of the journal system for 15 or 20 years and it hasn’t happened. So I’m maybe not the best person to ask. The publishing industry really has adapted in creative ways.

What challenges do publishers face?

The scholarly enterprise does need to jump ahead. When the act of publishing was putting ink to paper and mailing out the results and I would have to walk to a building to pick up a paper, the publisher was definitely a necessary component. But it’s not clear that an independent entity as a “publisher” is necessary for putting electrons to disk.

What is still important is the filtering process. It should be possible for a group of scholars to set up an operation that doesn’t involve a publishing house, but does involve soliciting independent opinions, evaluating those, and stamping their approval on some collection of scholarly work. We don’t have a good mechanism to do that outside of publishing houses and we don’t have a good formal mechanism for establishing the reputation of such an enterprise.

What is the future of peer review of scholarship in your field?

People who are actively working in my field don’t really need a formal peer review mechanism because we are doing this sort of thing informally all the time. We recognize which are the important papers and can make those evaluations quickly. But if someone is trying to evaluate me for my next promotion, they need more than just letters of recommendation. They need some kind of quantitative measures to assess how things are going and the traditional journal system provides them with that.

Maybe we can move to the point where we can think about peer review in a new way. Instead of being tied to the traditional process—submit the paper, have it reviewed, revise it, have it accepted—maybe there is some other way that is not tied to the document. Some have experimented with blog-style comment pages. I’m not sure how a tenure committee would evaluate a blog comment page.

In your field, are scholars generally concerned about barriers to sharing information digitally?

The fact is you can only get the attention of about one percent of working academics on this issue—maybe that’s even optimistic. People are using the existing system, and working within it. If you explain to them what some of the problems are, they will express shock and horror. They may even go so far as to send in these copyright forms once or twice, but then they sort of go back to not thinking about it.

I do think the passage of time will go some way to changing these attitudes because for the younger generation everything has been digital their whole lives. So it’s not a change from one way of doing things to another.

People pay attention when there is a crisis. If your library tells you there is a big budget crisis, they have to cut 20 percent of our journal subscriptions—that gets people to pay attention. The form of paying attention may simply be to bemoan what’s happening rather than to participate in change.

What recent improvements or changes in this area are encouraging?

For those of us at institutions that can afford a lot of important, electronic library subscriptions—through a combination of the open access and paid subscription—it’s made our lives very easy. I’m very encouraged about what I can access.

One of my concerns is that the library is only renting access. If they give up the annual fee, then they would lose access to this backfile that they were counting on.

In many cases, that’s the way the contract is written. Unlike the old days, when you bought the journal and put it in basement and it was something that you owned, with the big electronic archives, you don’t own it outright.

What should be the role of funding agencies in scholarly communication change?

My view is that when the federal government is investing large money in scientific innovation, the scientific community as a whole has a right and duty to make the information available to as broad of an audience as possible.

So, the current model of taking the results of that research and giving it for free to a publisher who will then lock it up and only allow access after paying a fee is incompatible with that free communication of ideas. I really applaud the [US National Institutes of Health’s] PubMed Central initiative and I understand why Nature and Science are so upset about it, but it’s the right thing to be doing.

On the other hand, there are costs associated with publication. And the funding agencies have not been willing to provide money to pay for these things. It’s understandable that funding agencies want to focus their funding on the science, but finding a way to facilitate this kind of open communication is really important.

What are the benefits of more open sharing of research?

The nature of the scientific enterprise is founded on free and open communication of results. It’s true that there are places where people engage in science, like industrial labs or people carrying out classified research, where perhaps wide communication is not appropriate. But outside of those contexts, it’s fundamental to the enterprise that people be able to freely communicate what they have done.

In the old version of the system, where the only way to communicate your work was to put ink to paper and all major research libraries were subscribing to all the major research journals, the way to be open about communicating was to go through the journal system. Today, it’s not clear at all that is the way to be open about communication in research. To reach a broad audience of other scientists may require new techniques of communication that people may not have adopted as strongly as they ought to.