? A companion video is hosted at https://youtu.be/xOpFB0r0WPg.
The occasion of the BIR conference has prompted me to say a few words about research possibilities in information science that are suggested by my background in the history and philosophy of science. I do not represent the constructivist viewpoint and am more traditional in my belief that science is or should be evidence-based and is not predominantly an interest-driven and socially constructed activity.
The rst research problem information scientists might address is the nature
of scienti c discovery. We know that from an information perspective, discoveries
are often associated with high citation rates after the fact, and that many
discoveries are not recognized as such for many years, although some enjoy immediate
recognition. We do not have a clear understanding of the reason for this di
erential. Nor do we understand how major discoveries can be distinguished from
more modest but important research ndings and advances. From a prospective
point of view, it has been argued that discoveries are novel associations of facts
or ideas that had not been previously connected. For example, Don Swanson [
Another type of \discovery" that needs attention is the invention of new methods. The importance of methods in contemporary science is revealed by an analysis of the most cited papers in almost any eld of science. It might be argued that methods are now driving science. We know very little about how new methods are invented or how old ones enhanced. Do methods emerge from basic research or do they represent a separate evolutionary path more akin to technological developments? Finally, we are interested in the applications of methods in the conduct of basic research. Obviously, they are a source of evidence to test theories, but also data is collected for the sake of collecting and stored in computer databases. Another question is whether theories rely on existing methods for testing or require the development of new methods?
If we look closely at some historical cases, for example, the discovery of the
neutron in 1931, we see that scientists had initial inklings prior to the discovery
that an electrically neutral massive particle existed. This takes us to the next
historical process that needs more research, namely con rmation [
Implicit in my discussion of the problems discussed above are the application
of methods for clustering and mapping scienti c communities and the ability
to delineate structures of leading ideas and concepts at the specialty level [
Fortunately, very e ective computation methods have been devised for
detecting community structures using bibliographic databases most notably citation
indexes. More research is needed, however, into studying how these structures
evolve over time. Do research areas go through a lifecycle and how do we identify
their starting and ending points? Where do discoveries and methods t into the
cycle and can we nd evidence of con rmation or discon rmation occurring? In
what sense do these clusters or communities de ne what Thomas Kuhn called
\paradigms" [
Finally, I should mention various approaches to analysis of scienti c texts, and particularly citation context analyses. I do not think that the analysis of negative sentiments in citation contexts will be that fruitful because scientists are reluctant to engage in public criticism of their colleagues in print. More productive are studies of the degree of concept uncertainty as indicated by hedging.
However, much broader in scope is the use of what might be called \epistemic labeling" in scienti c contexts. If we look, for example, at highly cited papers we
nd not only consensus in the citation contexts on the meaning of cited texts, but also consistency in their labeling as \discoveries", \advances", \methods", \reviews", \databases", etc. and the use of other terms that indicate the cited Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
BIR 2020, 14 April 2020, Lisbon, Portugal.
119 concepts' epistemic role, which are often expressed in relational terms, such as \causing", \explaining", \predicting", \con rming", etc. Linguistic methods will be required to delineate, for example, what is \explained" by what, or what is \caused" by what. Obviously, such vocabulary analyses dovetail with some of the questions raised above. Whether the study of such epistemic terms and relationships will take us closer to understanding the \logic of science", as was long the objective of empiricist philosophers of science, remains to be seen.