Don R. Swanson

Fish Oil, Raynaud's Syndrome, and Undiscovered Public Knowledge

Divide and conquer—the strategy that science uses to cope with the mountains of printed matter it produces—appears on the surface to serve us well. Science organizes itself into manageable units—scientific specialties—and so its literature is created and assimilated in manageable chunks or units. But a few clouds on the horizon ought not to go unexamined. First, most of the units are no doubt logically related to other units. Second, there are far more combinations of units, therefore far more potential relationships among the units, than there are units. Third, the system is not organized to cope with combinations. I suggest that important relationships might be escaping our notice. Individual units of literature are created to some degree independently of one another, and, insofar as that is so, the logical connections among the units, though inevitable, may be unintended by and even unknown to their creators. Until those fragments, like scattered pieces of a puzzle, are brought together, the relationships among them may remain undiscovered—even though the isolated pieces might long have been public knowledge. My purpose in this essay is to show, by means of an example, how this might happen. I shall identify two units of literature that are logically connected but noninteractive; neither seems to acknowledge the other to any substantial degree. Yet the logical connections, once apparent, lead to a potentially useful and possibly new hypothesis.

An interactive system for finding complementary literatures: a stimulus to scientific discovery

Abstract An unintended consequence of specialization in science is poor communication across specialties. Information developed in one area of research may be of value in another without anyone becoming aware of the fact. We describe and evaluate interactive software and database search strategies that facilitate the discovery of previously unknown cross specialty information of scientific interest. The user begins by searching MEDLINE for article titles that identify a problem or topic of interest. From downloaded titles the software constructs input for additional database searches and produces a series of heuristic aids that help the user select a second set of articles complementary to the first set and from a different area of research. The two sets are complementary if together they can reveal new useful information that cannot be inferred from either set alone. The software output further helps the user identify the new information and derive from it a novel testable hypothesis. We report several successful tests and applications of the system.

Undiscovered Public Knowledge

Knowledge can be public, yet undiscovered, if independently created fragments are logically related but never retrieved, brought together, and interpreted. Information retrieval, although essential for assembling such fragments, is always problematic. The search process, like a scientific theory, can be criticized and improved, but can never be verified as capable of retrieving all information relevant to a problem or theory. This essential incompleteness of search and retrieval therefore makes possible, and plausible, the existence of undiscovered public knowledge. Three examples intended to throw light on the logic of undiscovered knowledge are constructed and analyzed. The argument is developed within the framework of a Popperian or critical approach within science and on Poppers distinction between subjective and objective knowledge--the distinction between World 2 and World 3.

Two Medical Literatures That Are Logically But Not Bibliographically Connected.

This study demonstrates that certain unintended logical connections within the scientific literature, connections potentially revealing of new knowledge, are unmarked by reference citations or other bibliographic clues. Specifically, 25 biomedical articles central to the argument that dietary fish oil causes certain blood changes are compared with 34 articles on how similar blood changes might ameliorate Raynauds disease. The two groups of articles are thus connected by a chain of reasoning implicitly suggesting that dietary fish oil might benefit Raynaud patients, an hypothesis not heretofore published explicitly. By retrieving and bringing together these two literatures, that implicit, unstated, and perhaps unnoticed hypothesis becomes apparent. The more general problem is posed of whether systematic search techniques for bringing together logically connected literatures can be developed and described, in the hope of discovering other implicit, unstated hypotheses. The example analyzed shows that the problem, while solved in this case by trial-and-error search methods, may be inherently and peculiarly difficult because there are virtually no references in either literature to the other, nor are there any clues from cocitation, bibliographic coupling, or statistical association of descriptors that the two literatures are logically related.

A probabilistic similarity metric for Medline records: a model for author name disambiguation.

We present a model for automatically generating training sets and estimating the probability that a pair of Medline records sharing a last and first name initial are authored by the same individual, based on shared title words, journal name, co-authors, medical subject headings, language, and affiliation, as well as distinctive features of the name itself (i.e., presence of middle initial, suffix, and prevalence in Medline).

Information Retrieval as a Trial-And-Error Process

Recognition of the essential role of trial and error in access to scientific literature may point the way toward improved information services and may illuminate inconsistencies that have beset many retrieval experiments. This paper examines three important and well-known information retrieval experiments, with a focus on certain internal inconsistencies and on the high variability of search results. In these experiments, retrieval systems are evaluated in terms of their ability to select relevant documents and reject those that are irrelevant. It is suggested that this criterion is inadequate because of ambiguities inherent in the concept of relevance and that closer attention to trial-and-error processes may be helpful in developing better criteria. Specific examples of how one might improve document retrieval, library use, and citation indexing are offered.

SOMATOMEDIN C AND ARGININE: IMPLICIT CONNECTIONS BETWEEN MUTUALLY ISOLATED LITERATURES

The purpose of this review is to show how a synthesis of the arginine and somatomedin literatures can lead one to identify an important but neglected area of research, an area that ultimately might enhance our understanding of certain emaciating diseases and age-related degenerative processes. Hundreds of biomedical articles document the stimulatory effect of infused arginine on the release of growth hormone (GH) in humans. That GH in turn can stimulate the production of circulating somatomedin C (SmC) is equally well documented. One can plausibly infer that arginine intake may influence blood levels of SmC. In this article, I show that there are many additional reasons to believe that arginine may have such an effect and that increased SmC levels can have important health benefits. Remarkably, however, there are almost no published articles that explicitly mention the possible influence of arginine on SmC. The idea that the two literatures on arginine and SmC can be linked by implicit arguments yet have few or no articles in common may have

Historical note: information retrieval and the future of an illusion

More than thirty years ago there was good evidence to suggest that information retrieval involved conceptual problems of greater subtlety than is generally recognized. The dramatic development and growth of online services since then seems not to have been accompanied by much interest in these conceptual problems, the limits they appear to impose, or the potential for transcending such limits through more creative use of the new services. In this article, I offer a personal perspective on automatic indexing and information retrieval, focusing not necessarily on the mainstream of research but on those events and ideas over a 34-year period that have led to the view stated above, and that have influenced my perception of important directions for future research. Some experimental tests of information systems have yielded good retrieval results and some very poor results. I shall explain why I think that occurred, why I believe that the poor results merit special attention, and why we should reconsider a suggestion that Robert Fairthome put forward in 1963 to develop postulates of impotence—statements of what cannot be done. By understanding such limits we are led to new goals, metaphors, problems, postulates, and perspectives.

Linking estrogen to Alzheimer's disease An informatics approach

Epidemiologic studies suggest that estrogen protects against AD.We employ ARROWSMITH, a novel computer-assisted approach, to identify possible links between estrogen and AD that are not explicit in the biomedical literature, by searching for substances or processes that are known targets of estrogen action and that have also been separately studied in relation to AD. Several links appear particularly promising (e.g., estrogens antioxidant activity) and merit attention by neuroscientists. NEUROLOGY 1996;47: 809-810

Complementary structures in disjoint science literatures

Difficult and intriguing information retrieval (IR) problems derive from what I call complementary but disjoint (CBD) structures within the literature of science. Complementary refers to the relationship between two separate scientific arguments which, when combined, yield important inferences and insights not apparent in the separate arguments. Corresponding to the two arguments are two complementary literatures. Each literature (ideally) is the “complete” set of articles that contain the argument in question. Disjoint literatures have no articles in common, do not cite or mention each other, and are not co-cited. If two complementary literatures are also disjoint, the possibility is worth investigating that the combined arguments and the inferences to which they lead might not be made explicit anywhere within the published record of science. The ever-increasing fragmentation of science into mutually-isolated specialties probably assures a limitless supply and combinatorial growth of implicit connections, some of which may be unknown solutions to important problems. These solutions are worth seeking. I have previously analyzed and reported three examples of CBD literature pairs, each of which led to a novel and plausible medical hypothesis that appeared to merit testing [1-5]. The universe of published medical literature, not a selected subset, was the target of a process that depended not only on extensive use of online information services but also on reading the text of many hundreds of medical articles. My claim that novel inferences can bc drawn from CBD literatures requires clarification. CBD literatures of course can never attain their ideal state of complclencss as specified in the definition, but they are presumed to approach as near to completion as attainable with a diligent search of the literature, On that basis, one can smkc a