«Andrés Gregor Zelman The University of Amsterdam 2002 ii Mediated Communication and the Evolving Science System: Mapping the Network Architecture of ...»
V) was defined as representative of a Mode II type of knowledge production. It serves us here to redefine the relationship between Mode I and Mode II types of knowledge production, particularly as they pertain to print and electronic media. While this definition served to provide distinct parameters around the textual analyses in previous chapters, new challenges in appropriately describing this relationship arise when analyzing the dynamics of scientific publication. Unlike chemistry and physics journals, for example, new techno-sciences like biotechnology and computer science do not build on strong institutional frameworks of discipline formation normally identified with Mode I types of knowledge production. In the techno-sciences different types of knowledge are recombined in varied multi-disciplinary foci, and appear to be driven by both social and scientific problematics, not necessarily institutional frameworks. Thus, since publication is historically a Mode I type of knowledge production it does make sense to correlate it with print, but given the new multi-disciplinary foci, it is important to understand that Mode II characteristics operate in this academic publication environment as well.
A means of understanding this overlap is provided in Fujigaki & Leydesdorff (2000).
It is argued that quality control is organized differently in the Mode II production of scientific knowledge, and that differences between these modes can be understood through an examination of validation across boundaries. Indeed, the shift from Mode I to Mode II is largely argued in terms of changes to institutional parameters.
Leydesdorff (forthcoming, 2003) argues that the criteria for quality control can
Similar to the examination of the print and electronic communications in the previous two chapters, the SOEIS relevant publications are examined for their collective architectural parameters, networked properties, and identifiable systemic indicators.
The publication environment is first examined for its architectural parameters, and then for its networked interrelationships with respect to co-authorship activity.
Finally, the positioning of SOEIS relevant publications is examined in relation to journal horizons (as determined via the factorial structure of their aggregated mutual citations) in order to discern evident validation boundaries to better understand the dynamic interrelation of Mode I and Mode II types of knowledge production in the
SOEIS project. The central results of this analysis will be revisited in Chapter VIII:
Integration & Conclusions where the relevant results of the four analyses are integrated.
Two aspects of the SOEIS publication dimension are outlined herein: the first focuses upon those texts produced by SOEIS members and the articles they have cited. The second, in turn, explores the pool of references produced by those papers which cite SOEIS relevant papers. Cited articles (and their publishing journals) are treated as events that compile and collectively form a pool of aggregated references – these are viewed as the collective publication realm of SOEIS members. Citing articles, by contrast, are viewed as indicators of how the SOEIS project and its environment have been integrated into the larger science system; this is achieved by examining the relationship between the citation environments. Journal-journal relationships relevant to the SOEIS communications were thereby identified, and their relationships examined over the five year span of the analysis (1996 – 2000).
By specifying published articles as events surrounding the SOEIS project, a window between the project and the system of scientific publication is obtained. The analysis aims to determine the positions that these publications take with reference to the journal literature; and more specifically, how these disciplinary positions change over the five years of the analysis. The communication structure that exists at this level (academic publication) yields a network architecture that remains independent of individual actions or motivations (as authors). That is not to say that individual’s ideas and foci do not change over the course of a research project, but rather that the publication an author may contribute will not necessarily register a dramatic change in the communication structure since the mode of communication may reflect the embedded-ness of the codification of their publications. The positioning of different publications in relation to the journal horizons enables the analyst to discern evident validation boundaries. In this way the networks of journal communication that are relevant to the SOEIS research group members were examined for change, thereby illustrating disciplinary emphases their respective shifts in priority, as revealed by the 94 changes specified by journal-journal communications. The analysis thereby provides this study with a richer perspective of the role of the SOEIS project in the larger context of the science system.
Research Questions & Expectations The reader should understand publications here in the sense of events, in light of the perspectives outlined in Chapter II: Theoretical Grounding. As with the previous analyses, the theoretical triad of Architecture – Network – System frames this analysis, thereby providing means of integrating different analyses herein, and a basis for comparison with other results in the conclusions of this dissertation. Three research questions have been identified, corresponding with the three nodes of the theoretical triad. These research questions and their central expectations are described below.
First, what are the architectural parameters of the SOEIS publication environment with respect to its Cited and Citing dimensions, and to what degree does their overlap reveal a cognitive bias; and further, can our understanding of the publication architecture be enhanced by comparing the rates of co-authorship and national orientation in the cited dimension? Here the parameters of the SOEIS publication environment are shown by calculating the frequency of publication activity for each year, in both the cited and citing dimensions. It is expected that there will not be a high number of SOEIS related publications from the cited dimension near the end of the 5 year scope (1996 – 2000), since related publications would presumably be published in years subsequent to the project. The analysis will reveal certain types of journals used in both the cited and citing dimensions, and when compared they should prove to overlap considerably. It is expected that given the nature of the research group, Policy and Informatics oriented journals should take precedence. This question is informed by the Medium Theory notion of the information network, and aims to characterize publication as central to academic network formation by emphasizing the interactive nature of this mode of knowledge production.
Second, given the architecture found by comparing the cited and citing dimensions of the SOEIS publication environment, what networks of interrelationship between these dimensions can be discerned by comparing the sum of most cited and most citing referenced journals? A key focus here is to illustrate how individual publications compile and collectively form network architectures. A central concern is how these relationships change over time, and to what degree these dimensions are interrelated.
The shared information network is highlighted by comparing the sum of most cited journals and sum of most citing journals for their respective fluctuations in referencing activity.
It is expected that this analysis will reveal a cognitive overlap between the two with respect to the similarity of journals operating in each dimension, thereby revealing a cognitive overlap or similar disciplinary emphasis. If so, this will reveal a cohesion between the two dimensions, and differences between them could then be interpreted as deviations from this core. Fluctuations between cited and citing articles by year are therefore interpreted as changes in the networked architecture of the SOEIS publication environment. The question is informed by Actor Network Theory and 95 outlines the network features of this interconnectivity as a product of collective action.
Finally, using the relations among journals (as measured through the factorial structure of their aggregated mutual citations) as the baseline1, can parallels be found in journal-journal distributions over time in the Cited and Citing databases, thereby revealing changes at other levels of analysis? The survey is performed using Research Policy and Scientometrics as the seed journals for the factor analysis, and the subsequent factor loadings for each are then compared for each year (1996 to 2000). Both journals are important for the SOEIS project and have been identified (in the following architectural and network oriented analyses) as the most central journals in both the cited and citing dimensions. Given the relatively small scope of the SOEIS communications and its immediate neighbourhood, by comparing the publication datasets for crucial changes in the disciplinary focus over the five year span of the study, the relationships in the identified clusters of journals should provide a sense of changing disciplinary development when compared with the field level of the SOEIS group as the baseline.
In the context of the SOEIS, when compared for yearly development we expect to find that the two journals will have a unique relationship vis-à-vis each other. The changing nature of the network structure of this relationship over the five years, as perceived using Research Policy and Scientometrics as the seed journals, will reveal that the SOEIS project was instrumental in binding otherwise disparate research foci.
Self-Organization Theory informs the approach.
Architecture The hermeneutic units of analysis sought here are individual scientific publications and their references. Two distinct databases were created in this initial stage of the analysis. The first was the Cited, and it included all articles referred to by SOEIS members and associates over five years: 1996 through 2000. The second database was the Citing, and it included all articles which in turn have referenced the SOEIS related papers that comprise the first database; thus, the citing contains similar information to the cited, but from the alternate perspective – who refers to the SOEIS related materials?
The first step in the analysis of the publication dimension entailed a collection of all of the publication information for the 15 core publishing members of the SOEIS research project (SOEIS-Core), as well as for the participants of the EuroConKnowflow mailing list (the original mailing list of the SOEIS) which also included the core SOEIS group (SOEIS-All). In all, some 74 names were collated, and their publication information for the years 1996 through to 2000 was collected from the Science Citation Index (SCI), the Social Science Citation Index (SSCI), and the Arts & 1 See Studer & Chubin (1983) where the issue of using a baseline is raised, and Leydesdorff (1986) and Leydesdorff & Schaar (1987) for approaches which make use of the factorial structure of aggregated mutual citations as the baseline for analyzing the field level of scientific publication.
96 Humanities Index (A&HI). The publication data obtained from these sources was filtered through the BibExcel and Dialogue programs.2 The procedure resulted in the formation of a database that included all references cited by the SOEIS group in their publications over the five year time period, hereafter referred to as the Cited Database.
A second step was then performed in which the SCI, SSCI and A&HI were queried for all articles which cited the entire SOEIS publication environment (not just the core members) as established by the first database, and thereby created a second database.
The second database is hereafter known as the Citing Database. In what follows is a description of the key results from the architecture analysis; detailed results showing the variance of this publication data for both the cited and citing environments can be viewed in Appendix C.
The basic architectural parameters of both databases are shown below in Table 6.1:
Publication Architecture. Note that the cited articles are differentiated between those publications made by the SOEIS – Core (15 publishing members), and then those that are also written by the periphery group of the SOEIS research project (in total, 74 members and associates)3. SOEIS Articles refers to all publications by both the project members and the periphery group, and citing articles refers to all publications which refer to SOEIS related materials.
It was found that the total number of cited and citing articles for each year of the SOEIS shared a frequency between 20 and 40 articles, with the exception of the citing articles for 2000 in which only 8 publications referred to SOEIS relevant materials.
For the cited articles the number of articles published in each year remained relatively consistent. By contrast, the number of citing articles increased substantially in 1997 and gradually decreased in each year thereafter. Perhaps more interesting is the SOEIS-Core distribution which shows high publication rates in 1997 and again in
2000. These initial observations suggest that the cited Articles of the whole SOEIS group substantially contributed to the relative journal publication literature over the years 1996 – 2000.
While the total number of citing and cited articles revealed basic information concerning the architectural parameters of the SOEIS publication environment, a more detailed look at the journal distribution shows the disciplinary emphases of the cited and citing dimensions, and thereby reveals more about the substance of the 2 The BibExcel program written by Olle Persson assists in analysing bibliographic data by generating data files that can be exported to Excel. The Dialogue Program written by Loet Leydesdorff was adapted to fit the needs of this particular research project.
3 For reference, the larger group of 74 was determined by drawing together the membership lists of the project mailing lists. This list of 74 individuals is also used in Chapter VII: Analysis of Mailing List Environment.