català | castellano | english home   sitemap   legal notice   credits   contacte  
home home

R. Joseph Anderson, American Institute of Physics, College Park, Maryland

"Documenting the History of Physicists in Industry, an Interim Report"

Many of the papers presented here cover subjects that are shared by archivists worldwide. Others concern issues that are more specific to individual fields or countries.   I'm going to be talking about a problem, documenting industrial research and development, that is especially challenging in America, but also one that archivists in other countries need to be aware of and address as well.  In visiting six archives in Germany and the United Kingdom, I've confirmed my initial impression that, unlike America, many European countries have a network of repositories that collect and preserve a wide variety of business records.  However, I've also found that none of the archives I've visited systematically document research and development functions, and some contain no R&D records at all.  So I believe that while the facilities for documenting business as a whole are much stronger in Europe than in the U.S., documenting industrial science and technology is a problem that we share.

The Center for History of Physics at the American Institute of Physics (AIP) is halfway through a four-year Project to Document the History of Physicists in Industry, and I want to tell you about

· why it’s worthwhile to document the work of physicists in industry,

· why the AIP Center for History of Physics has taken on the job,

· some of the challenges that are involved in documenting industrial physics,

· the work plan that we've developed for the study,

· how we selected the 15 laboratories that we're targeting, and

· describe some of our preliminary findings.

Why documenting the work of corporate physicists is important:

Speaking now only about the United States, future researchers who try to understand 20th century science by consulting archival sources are likely to picture a society that was rich in academic research and where the national government supported substantial science programs.  However, they'll find that corporate research seldom appears in archival records.  Based on documentary evidence they may be justified in concluding that corporate participation in science was infrequent and occasional, and it was mostly limited to a few large companies.  The reality is very different.  About 12,000 Ph.D. physicists–approximately one-third of the total—are employed in the private sector in the U.S today.  And many of the signal products of 20th century science and technology, ranging from the modern X-ray tube (invented in 1913) to the transistor, microprocessors, optical fibers, and integrated circuits have emerged from industrial labs in America.  In some fields, such as solid-state physics, the research done by industrial scientists has often surpassed that of their academic colleagues.

The G.E., Bell, and IBM labs together have produced more than a dozen Nobel laureates.  Using the Nobel Prize as a measure of the importance of industrial research, however, is misleading.  The American business model supports research in order to develop products that create new markets or sustain old ones.  The work that corporate physicists do is important because it's a major ingredient in the combination of invention, product development, manufacturing, marketing, and business process that has created the ongoing innovation that characterizes America's economy for the past 100 years.  To understand both the process and the mix, it's necessary to document the work that physicists do in industry.  Moreover, corporate research and development is an important part of the history of 20th century physics.

Why the AIP Center for History of Physics

The American Institute of Physics represents the physics and allied science communities, and the Center for History of Physics is the AIP division responsible for the archives and history of the physics community.  An important part of our work is to research hard-to-document areas in physics, astronomy, geophysics and related fields and develop strategies to cope with them both in the U.S. and internationally, since physics transcends frontiers.  We do ongoing work to preserve valuable papers and records created in academic physics, and we conducted a 10 year project, the Study of Multi-Institutional Collaborations, to learn about and establish appraisal criteria for the records generated in the big science projects that are conducted jointly by academic and government labs in fields like high-energy physics and space science.  We have long recognized the importance of corporate research and the need to document its history, and we decided that the time had come in the late 1990s.  In 1997 we did a preliminary email survey of major industrial labs and began background research on corporate R&D.

By coincidence a national symposium on “The Records of American Business” was held in 1997.  The symposium focused on business records in general and not on those of industrial R&D in particular, but it provided the most comprehensive look to date at archives and the private sector in the United States.  One of the primary conclusions of the “The Records of American Business” symposium is that comparatively little is known about American industrial institutions because there is very little documentation that researchers can work with.  This confirmed our own impressions and the results of our preliminary survey and research about the records of private sector research laboratories.

Challenges and Difficulties in Studying Physicists in Industry

There are no laws in the U.S. that require permanent retention of records documenting the history of industrial R&D except for those that are designed to protect intellectual property and support patent claims.  In addition there are three major obstacles to preserving corporate records.

· The first is the lack of existing archives to preserve business records.  Only a few companies have their own in-house archives, and even the business archives that do exist usually lose funding or close altogether when businesses go through normal cost cutting cycles.  Similarly no national government archives in the U.S. preserves corporate records, and most university and private archives do little in the area.

· The second obstacle is that each corporate R&D operation is unique.  Unlike much of academic and government research, every company’s R&D operation differs from others in terms of organization, source of funding, areas of concentration, and expected outcomes.

· The third obstacle is the twin issues of fear of litigation and the proprietary nature of much corporate research.   Many American companies and their legal counsels are reluctant to preserve inactive business records for fear that they will be used against them in lawsuits, as they have been in trials involving tobacco, asbestos and a variety of other product liability cases.  This may be changing with new laws like the Sarbanes-Oxley Act in the U.S. that discourage records destruction.  However, access will remain a problem, for companies don’t want to divulge secret research information that may benefit competitors.

Work Plan

We began to develop a blueprint for a Project to Document the History of Physicists in Industry in the late 1990s to address the problems that I've described. By 2000 we had established our goals and decided what we could hope to accomplish in a three or four year study.  We  estimated that the total costs of the project would be about one-half million dollars, so we also began to write funding proposals to the National Science Foundation and other organizations to help support the study.

Our overall goals were to find out what kinds of research is done by physicists in corporate labs, what kinds of records they create, how the records are used and the extent to which they’re preserved, and finally to develop recommendations and guidelines for preserving the records.  The project got underway at the end of 2002.  Our major activities include:        

· Selecting a sample of 15 companies to target in the study.

· Conducting questionnaire-based oral history interviews with approximately 75 scientists and R&D executives.  We are also doing question-set interviews with records managers, technical librarians, and where they exist, corporate archivists.  The question-set interviews average two-to-four hours in length. We are also conducting much longer career-review interviews with about 20 especially important figures.

· Studying major public and academic archives in the U.S. that preserve the records of industry and visiting selected European business archives.

· Creating a bibliography of the historical literature on industrial R&D.

· Publishing our findings and recommendations at the end of the study, including detailed guidelines for identifying and preserving historically valuable corporate records.  We'll also catalog all the products of the study, including descriptions of extant records and oral history interviews, online.

What Companies and Why

In deciding what companies to target in our study, we selected a judgement sample of 15 firms from the 27 companies in the U.S. that employ the largest number of physicists.  All together these companies hire about half of all the Ph.D. physicists who work in the American private sector.  We focused mostly on the largest companies among the 27.  In addition we selected firms that present a mix of industrial sectors including computer hardware, defense, energy , transportation, telecommunications, and photonics.  Other criteria included product mix, ownership, and research structure.  Finally, we picked a few firms that have or have had in-house archives.

Over the past two years the project historian and I have completed site visits at nine of the 15 central research labs in the study, and we have conducted interviews with a total of about 80 scientists, science managers, and records professionals.  In addition, we have completed 5 of the longer autobiographical oral history interviews with selected leaders in science and industrial science policy.  The research laboratories that we’ve visited are at IBM, Xerox, Lucent Technologies Bell Laboratories, Exxon Mobil, Eastman Kodak, Corning, 3M, General Electric, and Texas Instruments.

Preliminary Findings

We've transcribed the 57 question-set interviews with scientists and managers that we've completed so far, and we’ve edited most of them.  We're now in the early stages of analyzing the interviews, so all of our findings are tentative.  We're fairly confident about some but less sure about others. So before proceeding, I want to say clearly that we’re in the early stages of analysis, and all of our current observations are speculative.

First off, we had initially expected to be able to identify a community of corporate physicists.  By a community, I mean a group of scientists from different companies who share social contacts and professional interests and who collaborate with one another on research at least occasionally, either formally or informally.  To identify a sense of corporate culture, we were looking for an awareness on the part of scientists at one company of colleagues at other companies or other venues who share similar research interests.  We found that in a corporate lab there doesn't seem to be as much interaction with peers outside their companies as one would expect among academic or government scientists.  Their main interactions as physicists seem to be with other people in their own companies, who may be in physics, engineering or other fields. The fact that the scientists whom we've interviewed seem to be somewhat isolated from outside peers means that their observations are largely limited to their own firms. With that additional caveat in place, I’ll proceed with some general observations.

The bench scientists whom we’re interviewing are all PhDs in physics and have 12 or more years of professional experience.  The managers have an equal amount of experience and a PhD in physics or a related field like engineering.   In looking at the career paths of our target group, we’ve found that nearly all of the interviewees decided to go directly into industry from graduate school, and almost all of the bench scientists have stayed with their initial employers. Managers are somewhat more likely to have moved from company to company as they climbed the corporate ladder, but more than half are with the company where they started out.

This pattern of career stability - physicists staying at the same company over a long period of time - is especially interesting given the context of rapid change in corporate R&D as a whole during the past 20 years.  It means that the majority of the people that we have talked with have remained in the same companies, but the companies themselves have changed dramatically during their careers.  Nearly all the companies in our target group have been adversely affected by America’s economic  boom and bust cycles over the past two decades. Major reductions in R&D investment from 1985 through 1994–the first decreases since World War II—created “an unprecedented period of decline that affected nearly every major [industrial] research lab . . . .” and fanned fears that America would fall behind foreign competitors. However, the trend reversed itself during the economic boom that began in 1995 and ended in 2000.  Next came a lingering recession that began at the end of 2000.  It stopped the pattern of growth and by 2002 had reduced annual private investment in R&D by $3 billion a year.

Each of the nine companies that we've visited are in different places on the economic boom and bust cycle.  The three best known labs in our sample are IBM, GE, and Lucent Technologies, the new parent of Bell Labs.  All of them have produced Nobel prize winning research and traditionally have been considered America’s premier industrial labs.  One of the characteristics that has distinguished them from some other corporate labs is that they have traditionally received funding from corporate headquarters instead of the more familiar pattern of having to negotiate with individual business units for funding.  Where the money comes from and how it’s allocated has a very important impact on the independence of a central laboratory.  Individual business units are concerned with immediate sales and profits, and they are typically willing to pay only for development that is directed at short-term results.  In contrast, corporate headquarters may have a broader vision and greater concern for long-term growth.  As a result they may be willing to support more speculative research.  An independent  budget from corporate headquarters is likely to give a laboratory more autonomy and freedom to pursue longer term research.   That said, it’s important to keep in mind that even long-term corporate research is almost always about producing products and profits, although it means that the risk of not producing useful results is greater.

The IBM, Lucent Bell Labs, and General Electric laboratories provide good examples of the differing effects that recent business cycles have had on three great industrial research operations.  IBM went through a financial crisis in the late 1980s and early 1990s, suffering record losses by 1992.  The company’s financial problems had a direct impact on its central research lab in Yorktown Heights, New York.  The lab downsized its staff of physicists and other scientists, restructured organizationally, and began putting greater emphasis on short-term product development over longer term research.  IBM returned to prosperity by 1996 however, and when we conducted a site visit at the lab in 2003 we found a group of scientists who had survived a serious crisis and avoided the more extreme reductions in staff and budget that many other industrial labs have gone through since.  However, while the IBM lab is a stable organization today, it is a much different place than it was before the crisis.  The staff is smaller, and the lab puts strong emphasis on maintaining a balance between longer term research and work that is aimed directly at creating products and responding to customer demands.  A recent article co-authored by IBM’s senior vice president for research and the head of physics research outlines “ two imperatives for success [for a corporate lab]:

· To be a great industrial research institution, an organization must do product development.

· To be fast in bringing new technologies to market, an organization must do long-term exploratory research.”

In contrast to IBM, Lucent Technologies Bell Labs, which split off from AT&T in 1995, is an example of a company that has had difficulty adjusting to the new high-tech marketplace and may not survive.  Bell Labs moved from being the research arm of AT&T, then one of America’s most stable companies, to the edge of financial collapse in 20 years.  Before the early 1980s, when AT&T’s monopoly on American telecommunications was ended after a long court battle, most of the costs of Bell Labs research was derived from a tax on consumers.  As a result Bell Labs was well funded and became the best known American industrial lab, with a strong reputation for innovative research.  Its achievements included a succession of discoveries that by its count produced Nobel prizes for 11 scientists for work done while they were there.  This is by far the greatest number for any corporate lab.  However, when we conducted a site visit in 2003, the lab’s work force had been reduced by about a third, and managers were responsible for firing people that they had  worked with for most of their careers.  There was no apparent plan for recovery, and the lab’s future remains in question.

General Electric presents the exception to the general pattern of economic downturn.  The company became the largest and most valuable U.S. corporation in the 1990s, largely because it diversified into financial services.  But despite its overall corporate success, researchers in the central lab became less autonomous during the past two decades.  From the late 1940s until the 1980s GE’s research lab in Schenectady, New York, was mostly funded by the headquarters office.  However, as I’ve mentioned, one sure way for a company to control R&D and focus on short term product development is to turn control of the lab’s budget over to product divisions.  They will typically insist that the lab concentrate on work that promises a quick return.  And that’s what Jack Welch, the hard driving new CEO of General Electric did after taking office in 1982. He changed the lab financing structure, requiring that it obtain funding from the business units.  As a result it has come to focus on shorter term, business oriented research.  When Welch’s successor took over in 2001, he restored a small part of headquarters funding, which now represents about 10% of the laboratory’s budget, in order to pursue some higher risk, higher value research.  Most lab funding remains controlled by the business units, however.

All six of the other corporate laboratories that we’ve visited are similar to one of the three different examples presented by IBM, Lucent Bell Labs, and GE.  One and perhaps two of the companies, 3M and Exxon Mobil respectively, have remained steadily prosperous but have moved to shorter term research, presumably in order to drive profits.  Another of the companies, Texas Instruments, has recovered from its financial crisis and is focusing almost entirely on product development.  And the three others - Corning, Kodak and Xerox - are working to reposition themselves in the marketplace.  Corning appears to be recovering from the recession that began in 2000.  Kodak is moving out of its traditional film business to more profitable digital photography.  And Xerox, like a number of the other firms we’ve visited, is increasingly outsourcing its manufacturing to firms in Asia in order to cut costs.

Moving now to documentation, we have found a complex and diverse pattern of records keeping.  The  laboratories - like all other organizations - are undergoing the  transition from paper to electronic records and to new forms of communication.  We had hoped to find that some of the high-tech organizations that we’re visiting, especially those that are doing pioneering computer research, would have developed new solutions to issues like storing and preserving electronic data and would be applying them in-house.  In general we’ve found that this is not true.  It appears that high-tech industry is facing and often ignoring - again like many other organizations - the problems that electronic records pose.  As for regulating the records of individuals, the corporate physicists whom we've interviewed seem to enjoy a good deal of general autonomy within the framework of shorter term projects and product development. Among other things, this means that they can usually decide on their own to either delete their email and most other records or save them for personal use on CDs or on their own computers.  Lab notebooks are a special case; the regulations for these vary from company to company and I'll discuss them in a minute. 

As I mentioned at the beginning, most companies don't have archives, and this is true even among our target set of major industrial leaders. Three of the companies that we’ve visited, IBM, Corning, and Lucent Bell Labs, have in-house archives programs, and we have interviewed the directors of two of the three archives programs.  The IBM archives focuses primarily on administrative rather than R&D records, but it has accessioned large collections of papers of two of the company's best known physicists.

Corning recently commissioned a commercial historical agency, the Winthrop Group,  to write a company history that emphasizes its research and development operation.  However, the material in Corning's archives is largely restricted to administrative records and papers of the company's founding family.  The research materials assembled for the book, including oral history interviews with scientists and science managers, are kept in the laboratory instead of the archives.

The third company, Lucent Bell Labs, would not allow us to interview the archivist or tour the archives, which is controlled by the Public Relations division.  This, by the way, is the only division in any of the companies that we've contacted that has refused to participate in our study.  During the same visit we were able to interview six Lucent physicists, including two senior managers.  We hope to be able to visit the Lucent archives later in the study.

Two other companies have placed some records at local public archives, 3M at a state historical society and GE at a city museum with a professional archivist on staff.  However in both cases these appear to be one-time or occasional transfers of records and aren't systematic or ongoing.  For 3M, the records at the public archives are administrative files - not R&D materials - that were accumulated for an  anniversary history about 15 years ago.  GE's  public archives collection consists of photos, publications and other material from the R&D laboratory.

Most of the individual labs have what appear to be generally effective records management programs for R&D records, and many maintain a few records permanently, although only for in-house use.   Permanent records sometimes include laboratory notebooks, but not always.  Two of the labs appear to have quite weak records management programs.

To our surprise only one of the firms has converted from paper to electronic lab notebooks.  However, email, power point presentations, word processing, databases, and other digital media play at least as important a role in the way that the scientists whom we've interviewed communicate and create and use records as they do for the rest of us.  Letters are largely a thing of the past, and email has increasingly replaced phone calls and in some cases even face-to-face discussions.  Two scientists pointed out the paradoxical effects that email can have on communications.  One described in some detail how he uses email to communicate with scientists at other places, especially academic institutions, in order to develop group projects.  The other mentioned that he spends less time with colleagues in his own organization face-to-face, but instead often exchanges emails with them.

As one of the most common means of communication today, email and its preservation have become a major topic of concern.  However,  the labs we've visited don't have any effective means of identifying or preserving important email.  Like many other organizations, they generally handle email by deleting it from their servers, usually after a few months. Most staff don't report any restrictions or regulations on saving their own mail, however and quite a few say that they transfer email to their hard drive or to a CD before it's deleted by the system.  Typically their email is lost when the system changes or there's a major upgrade, although some interviewees report that they migrate email from the earlier system.

One of the curious things that we've come across is that records managers at two of the companies that we've visited are putting a good deal of effort into preserving power point presentations.  The rationale is that these represent written documentation of the laboratory's ongoing research.  In fact, however, power point presentations vary enormously in content, and the best of them preserve only a general sense of the presentation.  In most of these cases there is no written text to preserve and no recording of the presentation.  And using power point slides without a text or recording of the talk that they're designed to illustrate can easily be misleading.

Our most surprising findings concern notebooks. All the companies are looking into electronic notebook systems, but only one has introduced electronic lab notebook software.  And we’ve learned that the once ubiquitous laboratory notebook has fallen by the wayside in some labs. A few companies have requirements for completing and retaining them, others have looser guidelines, and at least two companies have given up on issuing or retaining lab notebooks.  In practice, most of the companies don't seem to have enforceable policies on notebook retention.

Among the companies that try to enforce notebook policies, many scientists say that they paste screen shots and other computer data into their handwritten notebooks.  And one physicist showed us bound volumes of data printouts that he kept along with his handwritten lab notebooks.  The decline in use of the lab notebook at some companies may result in part from computerization. For example, in commenting about the case of Jan Hendrik Shön, a Lucent Bell Labs scientist who falsified research results, a Lab manager noted "that the use of physical lab notebooks has declined in this computer age."  Thus Shön's initial argument that he hadn't recorded his results seemed plausible.  Others, however, still put a strong emphasis on notebooks, and a few have created programs to preserve inactive notebooks and make them accessible in-house.

As you'd expect, scientists regularly use databases in their work, but we don’t yet have a good understanding of how safe the data is in terms of long-term preservation.  It's unclear if the database servers are backed up on a regular basis.  This is something we'll look into further along in the project.


Despite the central importance of industrial research and innovation in modern American history, the subject languishes on the periphery of scholarly research and public understanding.  The primary goals of our Project to Document the History of Physicists are to create new resources for in-house and outside researchers, including both scholars and the public, and to develop viable recommendations and guidelines that companies and academic and public archives can use to remedy the lack of documentation and build new programs. The study will produce 1) the first published analysis and description of information needs and of records-creation and records-keeping practices in 15 leading high-tech companies, 2) oral history interviews with approximately 100 leading scientists and research executives, representing the first broad oral history program on industrial R&D, 3) an online catalog of extant sources on industrial R&D that we identify during our project, and 4) an online historical bibliography of the literature, and 5) recommendations for cost-effective approaches to preserving existing historical materials and creating new resources, including detailed guidelines and models based on American and European practices and on emerging standards and technology.

We will publish summary findings and a detailed full report.  We will send the publications to several hundred influential historians, policy makers, science administrators in business, and archivists, and provide copies gratis to others who request them.  We will also mount the text on AIP’s Web site and announce the publications in relevant newsletters and journals.  And we will submit articles to appropriate journals in archives and history of science.

The extent to which the study will affect how businesses document the work of physicists and other scientists will, of course, depend on the businesses themselves.  Only they have the authority to preserve their history and eventually share it with researchers.  However, in finishing I'll mention one additional finding.  During the project we've found that industrial physicists are interested in their own history and will go to some effort to preserve it.  Not one of the companies nor the 60 physicists whom we have invited to precipitate in the study have turned us down.  Instead, they have devoted hours to answering our questions and evinced interest in the project.  This includes top science managers as well as bench scientists.