|Table of Contents|
|Table of Contents|
The modular model for experimental sciences has been based on an iterative empirical analysis of the corpus of publications on a research project in experimental molecular dynamics. We have analysed a set of ten articles, in which an account is given of a coherent part of the research. We have focused on the standard journal articles on original research.4.3 From the perspective of physics, we have examined the different types of information that are presented in physics articles and the different types of relations that can be identified in them. The analysis is focused on the information at the conceptual level, abstracted from the actual representation of the information in terms of a language.4.4 This implies that neither the choice of the language, nor the formulation is addressed in detail. We investigate how the information can be organised in a new way that allows for effective and efficient communication, given the requirements of the interactants in the process of scientific communication via articles, which have been described in section 2.2.2 and will be further specified for the domain of experimental molecular dynamics in section 5.1.2.
To enrich the analysis from the physics perspective, the point of view of `speech
act theory' 4.5 is adopted as well. The speech act analysis is focused on the textual representation of the physics content of articles. The purpose of the empirical analysis from this
point of view is to identify the different communicative functions of an article and to reveal the text structures designed to fulfil those functions. We have taken as a starting point the fact that scientific articles have reports aspects and argumentative aspects: they are not only aimed at
informing readers, but also at convincing them of standpoints about the
research presented in the article, in accordance with the requirements for
rational scientific communication by way of articles.
In the iterative process of the development of the modular model, firstly an initial model has been hypothesised as a starting point. The first ingredient of a modular model is the typology that determines what is considered to be a simple concept and thereby determines what information unit is to be represented in a module (satisfying definition 3.1.1). Following standard practice, characterisations of the information by its physics contents and by its bibliographic information were taken into account, although these classification types have not been elaborated in terms of specific classifications. We have also taken into account a characterisation of the information by its role in the problem-solving process. For that characterisation, the starting point was a classification directly reflecting the prototypical sections of scientific articles: Introduction, Methods, Results, Discussion, Conclusions (abbreviated as IMRDC) .
Guided by the initial model, we did an explorative analysis of corpus articles. The analysis pertained to the nature, the completeness and the coherence of the scientific information: we studied the structure of the article, identifying the different types of information contained in them, in terms of information units and relations between them. We also identified those types of information that could contribute to a more effective and efficient communication, but were left implicit or were entirely absent from the articles. In this manner, we reconstructed the information underlying the article that we assumed to form the complete message to be communicated. Doing so, we played a role similar to that of the actual physicist reading the article.
|Figure 4.1: Linear and modular representations of information: mappings from the information space.|
In the next step, we confronted the initial modular model with the corpus. For that purpose, we used the model to `modularise' the corpus articles we explored. This modularisation did not entail a direct translation of the original `linear' ones. The information underlying each original article has been represented in a different way, by recasting it in an explicit modular structure, as is illustrated in figure 4.1. In order to facilitate comparison between the versions, the modularisation entailed mainly restructuring, but as little as possible actual rewriting of the text. Writing the modular version, we tried to put ourselves in the position of the author, in order to gain insight in the feasibility of a new modular structure and in its advantages and disadvantages for the author as compared with the existing linear structure. Of course, the difference is that a `real' author aims to directly present his own work, and we endeavoured to recast the work of others in our own mould. On the one hand, this confronted us with problems familiar to translators. On the other hand, it is difficult for authors to organise novel work that is not entirely crystallised into its final shape, whereas we could organise the information in hindsight. Those differences had to be taken into account in the evaluation of the modularisation process.
The resulting modular articles and the modularisation process itself were subject of an evaluative analysis, to determine whether all information could be recast in modular form and whether the resulting article was adequate in the light of the communication criteria specified in section 2.4. At the level of the individual article, the original articles were compared with the modularised versions. We also compared the modules of the same type from different, related articles.4.6
On the basis of this evaluation, our initial modular model has been adjusted. The adjusted model has in its turn been used to analyse and modularise other corpus articles, further adapting the model to scientific practice. In this process, we tried to keep the modular model itself general. The aspects that are particular to the domain of experimental molecular dynamics were taken into account in the formulation of domain-specific rules. As we stated in section 3.3, domain-specific rules must be specified to allow for the adequate application of a general modular model to the writing and evaluating of actual modular articles. The better the modular structure is tailored to the domain and genre, the better the resulting domain-specific modular articles will satisfy the particular needs of the interactants in that domain.
The final modular model, in which the different types of modules and links are defined, is presented in this chapter.4.7 The final domain-specific rules for the applicatio of these definitions is given in appendix A. In appendix C, we give the modularised versions of two standard articles, as well as several connected modules providing the necessary background to the modules that are part of the modularised articles.