|Table of Contents|
|Table of Contents|
As an illustration of how our modular model can be adapted to other domains, we consider four examples of publications on: 1) a predominantly experimental study of high temperature superconductors; 2) the development of a process for the synthesis of a particular compound in bioorganic and medicinal chemistry; 3) an empirical study in the domain of argumentation theory; and 4) abstract research on logic for artificial intelligence.
In the domain of experimental molecular dynamics, both experiment and theory play an important role: experimental methods are used to perform measurements, and the results of these measurements interpreted are using the theory. In some domains of science, research is concentrated on experiments in which, for example, different samples are studied using the same experimental method, or where different types of measurements are performed on the same sample.
In appendix D.1, an example is given of a predominantly experimental publication: [Ihara et al, 1997], a short contribution to a major conference on materials and mechanisms of superconductivity. In this publication, a sample of a potentially good superconductor is prepared and studied using standard methods.
In a modular environment, a publication of this type does not have to include an Interpretation module. The core of the publication is a Results module . It is supplemented with a module Experimental methods about the sample preparation and further details about the experimental methods can be made available by links to mesoscopic or macroscopic modules. Modules Central problem and Situation can be included to explain what the authors tried to do and why. A module Findings can be included to allow reader to search for the main result under that heading. Of course, a module Meta-information has to be included as well.
In some publications in experimental science, the result reported is a newly developed method: a particular apparatus, technique or procedure, which can be used in subsequent research. A straightforward example of such a publication is described in appendix D.2: the result of the work presented in [Hutchison and Brouillette, 1998] is a procedure for the synthesis of a particular bioorganic compound. In a modular environment, the new procedure can be presented in a Results module , so that the reader can locate the original module in which the method was reported when it was first developed. When the new method is used in later work, it can be made available by means of a link in a module Experimental methods.
The main modules we distinguished for experimental science may be directly applicable to publications about empirical studies in the humanities and social sciences that traditionally have an IMRDC structure.
In appendix D.3 we consider [Van Eemeren et al., 1995], an article in the domain of argumentation theory. In this example, hypotheses are explicitly formulated and tested. In a modular environment, these hypotheses can be presented in a module Central problem, which is more extensive than the modules Central problem we have encountered in our analysis in the domain of experimental science. The methods pertain to the test design (experimental methods), statistical methods (numerical methods) and a theoretical framework (theoretical methods). The interpretation of the empirical results amounts to a discussion as to what extent the hypothesis has been confirmed.6.1
Our model must be more thoroughly adapted in order to be of service in more abstract domains, like mathematics. As an example, we discuss in appendix D.4 a contribution to a book published as a consequence of a conference on game theory and epistemic logic, in the domain of logic for artificial intelligence.
Whereas the research reported in our corpus was concentrated on measurements and the interpretation of the experimental results in terms of a theory, the main activity in [Van der Hoek and Meyer, 1997] is proving theorems about a logic, consisting of a deductive system, a model and a language. Since the logic (with its three components) plays a central role in publications in this domain, it can be presented in a module Description of the logic (containing three constituent modules) that lends itself to multiple use. The counterpart of our module Methods may be a module Proving consisting of the constituents Proof techniques (which is suitable for multiple use) and Proofs. The result, a proven theorem, may then be presented in a module Theorems that the reader can directly locate, retrieve and consult.
In short, we see that the modular structure developed for experimental science can be adapted to other domains. The bibliographic component of the typologies may be standardised for all scholarly articles. The range-based component of the typology for modules and the typology for the links might be standardised for all domains in which articles are published in the context of research programmes. The domain-oriented component of the typology for modules must be customised for the domain in question. Concerning the component of the conceptual function, the modules Meta-information, Positioning (with a Central problem and a Situation) and Findings can probably be used in all domains. The distinction of the modules Methods, Results and Interpretation, or their specific equivalents, is less clear cut. For instance, in abstract domains there are no experimental results to be interpreted, and for research in which a new method is developed, that method is the result.