«Andrés Gregor Zelman The University of Amsterdam 2002 ii Mediated Communication and the Evolving Science System: Mapping the Network Architecture of ...»
Four relevant communicative domains of the SOEIS were isolated: print writing, electronic writing, journal publication, and the mailing list environment. Arguably, oral modes of communication as exchanged during project meetings, telephone calls, conferences and the like could have been selected for analysis, and likewise, we could have chosen to look at any number of other combinations between the domains than the ones selected. However, of crucial interest to this project is the creation of a method through which one can compare the residual traces of project communications – hence written communications were biased. Those communicative dimensions that have left residual traces were selected specifically; oral communication (unless audiotaped or inscribed via ‘participant observation’) exists only as it is happening – it leaves no traces for further analysis. Print and electronic writing (as a noun not a verb) are precisely traces of previous communications and this renders writing an extremely important research site not only for understanding the symbolic dimensions of communication, but its networked aspects as well. By comparing residual traces left by communicating parties we can gain insight into the fundamental differences between media forms used.
44 Print and electronic communication were selected as the embodiment of the content of the intellectual exchange of SOEIS members. Since the nature of the information communicated along both channels of information flow are expected to be very similar, they were selected with precisely the intent to compare them for notable similarities and differences. Together the print and electronic communications contain all of the relevant information concerning the internal (not interpersonal) workings of the SOEIS research project.
The publications of SOEIS members and the mailing lists they use to communicate their findings comprise the external workings of the project selected for analysis. The publication patterns of this particular group of actors are examined over five years, accounting for the citing and cited Journal publication environments before, during and after the research project. The two SOEIS project mailing lists have been combined for another level of analysis; together they embody the electronic communications described above. Here the aim is to compare the thread dynamics of these mailing lists with a selection of others from the fields of Science & Technology Studies and Self-Organization Theory.
The networked relationships of the external workings of the SOEIS can be conceptualized in a similar way to the internally networked dimensions of the print and electronic communications described above. Overall, the concern is limited to the Architectural – Network – Systemic dynamics involved. The content of the information exchanged is addressed only secondarily; that is, for the purpose of interpretation. To be clear, we do not necessarily address how the content has changed – the primary concern is how the networked relations between people communicating
can be shown to be changing. The theoretical bodies outlined in Chapter II:
Theoretical Grounding largely reflect this approach, and they help frame the types of questions we can ask about the dynamics of knowledge production. Clearly there is no necessary relation between the theories and methodologies employed in the analysis;
the methodologies are best perceived as operationalizations of specific modes of theorizing.
Print Communication The print documents of the SOEIS research project were collected for the period between January 1997 and November 2000.3 They include the initial TSER application and corresponding Annex, the project milestones, and the final report to the European Union. The documents themselves are not publicly available, but were obtainable through professional association.4 The texts include all print data submitted internally to the group throughout the two-year project. Collectively these documents are considered to house the print component of the SOEIS, and by extension they are conceived to be the ‘location’ of the print knowledge produced in the project.5 For a detailed outline of the texts used in this analysis see Appendix A.2.
3 Note that two texts from the final report to the European Commission were received as late as November, 2000, despite the project’s completion at the end of 1999; see Appendix A.
4 The author would like to thank Peter van den Besselaar and Moses Boudourides for making the documents available.
5 Importantly the print documents required no filtering beyond saving the documents as plain text files instead of Microsoft Word document files; this speaks to the nature of print communication itself. In the case of the electronic communications, several levels of filtering were employed, precisely because
The next analysis aimed to illustrate the networked patterns of relation between keywords in order to identify developments in the concepts being exchanged. The keyword-searches were performed for each time-period, and for the entire document set. Changes in the position of the keywords indicate if they are new, constant, increase or decrease in importance, or disappear altogether.
In the first of three keyword analyses the analysis of keyword distribution is restricted to the top 50 keywords. This data can be used to create networks of related scientific topics (Whittaker 1989, Callon & Rip 1986). The differing patterns of top keyword use are expected to reveal central topics being exchanged over the time periods.
However, since the most commonly occurring words would include title words and other words significant to the research project itself, other analyses were pursued that emphasized strong relations and de-emphasized weak relations, thereby demarcating the central and non-central topics being discussed. By comparing the texts a perspective was gained on the relative importance of terminology. The second keyword analysis compared the text for each time period with the total document set in order to highlight the relationship between each time period and the whole project (stock concept). The final analysis compared each consecutive text; thus, the first is compared with the second, the second with the third and so on. In this way the flow of terminology is highlighted. Finally, in order to enhance our understanding of the networked relationships between those words isolated as key, a collocate analysis was performed on selected keywords to show the density of other keywords in proximity to the search term.
The systemic features of SOEIS print communication were examined by testing the respective time periods for points of critical transition indicative of a path dependency in the communication. The analysis will reveal whether or not particular ways of communicating information were integral for the development of the project’s final output. The analysis includes an assessment of the specificity of the datasets (a measure of the degree to which some words occur differently across the dataset), and of transmission as a measure of the flow of mutual information across the dataset.
Chapter IV: Analysis Print Communication examines the print exchange between members of the SOEIS research group using varied bibliometric techniques to determine the architectural, networked, and systemic patterns of word use.
of the way the information is exchanged and categorized. Please see Chapter V for more detail.
6 The Wordsmith program is the textual analysis program of the Oxford University Press; it was selected for its simplicity, ease of result exportation, and sophisticated indexing logics.
46 Electronic Communication In October of 1996 the participants of the EuroCon-Knowflow mailing list used the forum for establishing relations between research groups to formulate a joint research project within the framework of the TSER program of the European Union. It facilitated the joint collaboration and creation of a research project that eventually became the SOEIS research project. The Self-Organization of the Information Society (SOIS) mailing list was added later to enhance the communication of the group within the larger scientific community. The SOIS was introduced midway through the SOEIS project to incorporate the larger framework of the global Information Society.7 Both mailing lists served as the central bulletin boards for administrative messages, theoretical discussion, general inquiries, and the maintenance of the lists themselves during this time period. The term electronic communication, used here, refers to the communication amongst members using these two mailing lists. The EuroConKnowflow list participation died with the finalization of the project, and the SOIS list is currently being used to facilitate the organization and communications of another European Research Project.
The textual component from both mailing lists for the period between the beginning of the project in October 1997 and the end of the project at the end of 1999 was used.
All relevant emails from both lists were collected and divided into four 6 month chunks in order to provide an initial basis for comparison. This initial process outlined the architectural dimensions of the SOEIS electronic communications. As with the print analysis, the texts themselves were collated into four groupings to be analyzed along a time axis to discern emergent network properties. The total number of individual emails included in electronic database (from both EuroCon-Kowflow and SOIS lists) was 1261. The emails were collected from respective archives of the two mailing lists.8 As with the print analysis the first keyword analyses was restricted to the top 50 keywords. The second keyword analysis compared the text for each period with the total document set, and the third compared each consecutive text, resulting in three keyword lists that highlighted the transmission of keywords between the time periods.
A collocate analysis was also performed on a selection of important electronic keywords to reveal patterns of keyword association. The analysis of the electronic communications will identify the network relations between keywords, both at different time periods and between time periods. The difference between print and electronic writing will also be addressed: how often certain keywords occur in the body of the individual time periods, and how they compare with each other and the full document set reveals patterns that will enable us to discern properties or biases specific to each medium.
Relevant here is that keywords provide descriptors of the content of emails, and can be valuable indicators of the exchange of concepts within the group of active 7 Interestingly, the attempt to enhance the communication resulted in an almost total participant shift to the new list. By contrast, when the EuroCon-Knowflow mailing list was divided into several distinct
lists in an attempt to enhance the communications within each task, there was the opposite effect:
communications became fragmented and the participants eventually organized themselves back around the original list.
8 See: Appendix D: Archive of Mailing List Analysis for details of each list.
47 participants.9 Importantly print communications are meant for publication and thus require a formalized method of communicating results, whereas electronic communications in the form of email tend to have a distinctive set of linguistic features. Most notably, the latter include an informal manner of writing, a particular situational context, and a triadic relation between the addressor, the addressee, and an assumed audience (Collot & Belmore, 1996). With respect to the SOEIS we must assume a common referent, which we identify here as an affinity of intellectual pursuits with a common aim. By definition, the common referent is the collective writing that constitutes the print database, as the SOEIS research project was necessarily aimed toward a final print report to the European Commission.
Finally, the third analysis examined the systemic features of the electronic communications by testing the respective time periods for points of critical transition, thereby indicating whether the information necessarily followed certain pathways to achieve the communication. This information will prove most useful when juxtaposed with the results of the systemic dimensions of the SOEIS print communications. It should be noted that the latter are collectively written documents (task / group specific) whereas the former are specific to individuals (irrespective of task / group).
Chapter V: Analysis of Electronic Communication examines the communication between members using the EuroCon-Knowflow and Self-Organization of the Information Society (SOIS) mailing lists to determine their architectural, network and systemic communicative patterns.
Journal Publication The names of the 74 individuals examined in this analysis were obtained from the participant list of the EuroCon-Knowflow mailing list. The publications of both the SOEIS members and the group are compared with those that cited the SOEIS material. The central journals from both groups will be compared – that is, the journals that the SOEIS members used to communicate their findings, and those that in turn are used to cite the SOEIS relevant material. The aim was both to describe the publishing relationships between these SOEIS members and associates, and to situate the SOEIS within the scientific community. The analysis highlights the architectural, network, and systemic features of the research project by emphasizing different features of the publication environment.