back next next
Table of Contents
Index
Glossary
















































































Table of Contents
Index
Glossary
back next next

Next: Conclusions Up: Representing scientific discourse relations Previous: Relations based on the

  
Content relations

 The relations pertaining to the dependency in the problem-solving process could be identified quite easily. In appendix A we have given clear rules for the identification of these relations between particular modules. In A08, we identified 47 of these relations, for example between the measured experimental cross sections represented in A08-m4bi1 and the molecular beam method given in A08-m3a (see table 5.4).

 The subtype of dependency relations, transfer relations, was also identified easily and frequently. Often, a particular value obtained in another module, in the same article or elsewhere, is used as input. For example, the value for the vertical electron affinity that is obtained in A05-m5bii is used as input in the discussion of the Theoretical methods in A05-m3c. Figures are imported as well. For instance, the figure of the potential curves is generated in the Interpretation module A08-m5bi. It is copied to A08-m4bii1, which is expressed in that Treated results module by a link carrying the label `input'.

   About half the links in A08 express either a general elaboration relation, or a more specific resolution relation or context relation, or a combination of these. The general elaboration relation, where the one relatum provides more details as well as a wider context of the other relatum, is not very common in A08. Only 17 links are labelled as `is elaborated in', and no link carries the opposite label. The other combination of the resolution relation and the context relation, where the details are more focused, is identified 75 times. More than a quarter of the links lead to a target that provides more details about a particular issue addressed in the source. For example, resolution relations can be identified between, at one end, most mesoscopic and macroscopic  modules, and at the other end, the  microscopic modules connected to them. Usually the authors of a microscopic module refer to these higher level modules, because they don't want to provide all relevant details in the microscopic module. Therefore links to other articles are also often labelled as `is detailed in'.

The opposite, where the target relatum of the relation summarises the source, is expressed explicitly in about a eighth of the cases. Between module summaries in complex modules and constituent modules, both ends of this combination of resolution and context relations are made explicit. A link is made explicit in the constituent modules, leading to the module summary, which provides more context and less details. The opposite of these relations results in an explicit link from the module summary, to a more detailed and focused account in the constituent modules.
 We have identified very few similarity relations in A08. Most of them are associated to the Treated results. The experimental and theoretical results that are compared in the interpretation, and that in the original version were presented in a single figure, are connected by a link expressing that comparison. In addition, the results are compared with the results of others, in the context of the discussion of the reliability of the results, by means of `external' links. We have not labelled any links with the more explicit labels `agrees with' or `disagrees with'.

 Of all relation we have made explicit, 16% were synthesis relations. Most of them were identified between  complex modules and constituent modules. For example, the specific experimental results for collisions between ${\rm KBr_2}$, ${\rm LiBr_2}$ and ${\rm KI_2}$, represented in the elementary module A05-m4bi1 and A05-m4bi2, are grouped into a  cluster module A05-m4bi, which deals with alkali atom-halogen molecule in general.

We have also expressed synthesis relations in links between microscopic  and  mesoscopic modules (which therefore also express range-based relations). For instance, the microscopic modules Experimental methods, in which the set-up used in a particular experiment is presented, refer to various constituent modules of the mesoscopic Experimental methods, in which the components of molecular beam experiments in general are presented.

 A link expressing a segregation often also represent an elaboration relation: the target module provides a more focused and detailed account of a particular component of the central concept of the source module. These types do not necessarily coincide, as the target can focus on a concept that is not a component of the aggregate central concept of the source module.  

In the modular versions of A05 and A08, causality is expressed in the reasoning in the text within the modules. However, we have hardly expressed any causality relations in links.  Causality relations can be identified between relata of an explanation, because explanation by definition involves causality. However, we have not labelled links expressing an explanation as causal by default, but we have only made causality explicit between relata that have a clear causal relation. This does not happen very often, as generally causality is intermingled with less strict reasoning, e.g. a justification of the choices that were made.

  Summarising, we see that the methods Positioning and Experimental methods are quite easy to write. They also fulfil the needs of readers who either want to avoid unnecessary details (by avoiding these modules) or to get a fully detailed account (by consulting the modules in full detail).  The module Results provides a focused account of specific results, that fulfils the readers' requirements, and it is also feasible, if the guidelines are followed.  The module Theoretical methods was somewhat complicated in the case where the theoretical aspects were emphasised, but in fact the account of the theoretical methods in the original version even more complicated, because the modular version is more explicitly structured and it contains module summaries that clarify the line of reasoning. The  Interpretation module is also a difficult one to write, because it is a very complex module, just like the corresponding Discussion section.

   Organisational relations, as well as dependency relations based on the problem-solving process and similarity relations, are easy to provide and quite useful. Other scientific discourse relations were more difficult to identify in the modularisation process, but may be easier to deal with by authors writing a new modular article.



Next: Conclusions Up: Representing scientific discourse relations Previous: Relations based on the