Arbitrarily Combining The Social Network Approach With The Ethnographic Approach

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Arbitrarily Combining the Social Network Approach with the Ethnographic Approach Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam. Introduction Simply combining the ethnographic approach to the structural approach of network analysis is fraught with, at the same time, dangers and potentiality (Knox et al., 2006). Using hyperlink analysis and textual data gathered during a situation on the Singapore blogosphere as a case study I ask, could a combination of the two create a ‘better’ picture or will it result in the metaphor being mistaken for the ‘real’? Lin et al. (2006) using the structural social network analysis approach have defined the Singapore blogosphere as a “community with no obvious central topic”, and stated that it was a rather closed network, or rather closed off from the wider global network of bloggers. The ethnographic approach tends to take a very different position arguing, “It is rarely that we find a community that is absolutely isolated, having no outside contact. At the present moment of history, the network of social relations spreads over the whole world, without any absolute solution of continuity anywhere (Radcliffe-Brown, 1940:224).” This paper addresses the inadequacies of using hyperlink analysis or the ethnographic approach alone when uncovering online networks. Arbitrarily combining the two approaches will highlight the theoretical problems, benefits and limitations. Using a situation in 2006, I extracted a corpus of 29 blog posts. Using the social network approach I ask, which blogs are the keyplayers? Using the ethnographic approach, I ask what discourses and styles of discourse appear in the Singapore blogosphere?

Social network analysis seeks to trace the flow of information that passes through a network of relations. As actors make use of computer networks the computing networks are “clear indicators of communication structures within society” (Garrido & Halavais, 2003). Garrido and Halavais argue, “A map of the communication network is roughly isomorphic to the structure of the relationships among the users (2003).” Creating a Website or blog, the blogger ties their own efforts to those with Draft – Not for Citation

1

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

similar interests using hyperlinks. Designing and placing a hyperlink is an act that requires a certain level of hypertext mark-up language (html) knowledge and as Adamic and Adar (2001) state, a form of cognitive, social or structural connection between the blogs. Jackson (1997) and Kling (2000, cited in Garrido & Halavais, 2003) indicate, “Hyperlinks represent reasonable approximations of social relationships.” I have targeted blogs using hyperlink network analysis uncovering the keyplayers of the Singapore blogs with higher levels of ‘closeness centrality’ and ‘betweenness centrality’ (de Nooy et al., 2005) to assess which blogs are more ‘important’ to the flow of information. A blogs position indicates whether it has access to information and better opportunities to spread information.

Social Network Analysis Deleuze and Guattari (2004) present us with a model of knowledge and perception known as rhizome. The rhizomatic model of knowledge according to Cavanagh (2007:43) results in a network model that appears to be chaotic. The rhizomatic network works on the principles that any point in the network can be and is connected to every other point in the network. The logic of the connection in the rhizomatic network is movement. A connection is the sprouting off in a new line. The lines may appear to be random as they do in hyperlink analysis but they do have a purpose. Hyperlinked culture has as its main aim ‘intertextual evolution’ (Dreyfus, 2001) whereby all possible associations and linkage is enabled regardless of how tenuous they may appear. Resulting in a disordered knowledge and enabling a new form of knowledge to emerge. The main point with Deleuze and Guattari’s rhizomatic network is that there is no hierarchy; no node takes precedence over another. The order is in constant flux with total inclusiveness. The flow of information however pre-dates the existence of the nodes. The nodes are interruptions in the flow. The nodes merely channel the flow of information. Cavanagh (2007:47) argues that the main concern regarding the utility of Deleuze and Guattari’s rhizomatic network is that it is a philosophical position rather than a method for studying hyperlinks. In Deleuzian terms, the Internet itself is the node in the flow of information and knowledge exchange.

Draft – Not for Citation

2

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Knox et al. (2006) argue that American Social network methods map roles comprehensibly and this results in the incorrect assumption that they have delineated the ‘real’ social structures. Although it provides a counter measure to the encroachment of rational choice theory in American social sciences the problem is that it ends up reinforcing a view of relations that are very far removed from the everyday experiences of people. Knox et al. argue that Social Network Analysis’ focus on structuralism has in recent years shifted to attempts at developing a cultural approach. This shift in focus has come about because of the study of social movements by researchers such as Ansell (1997), Bearman (1995), Gould (1995), and Mische (2003). For Knox et al. the problem is that other methods of social research are either quantitative or qualitative. The network produces a unique situation unlike the graphs, charts and diagrams of statistical data. The network according to Knox et al. results in a position because of its use as a method, metaphor and form, of being the sign and the signifier, the referent and the representation. This they argue opens up potential benefits and new horizons but is also potentially dangerous. Similar to Knox et al. (2006) I will arbitrarily add the anthropological or ethnographic approach to the structural approach of American Social Network Analysis. I intend to conduct a twotiered approach in the hope that it sheds light on the methodological issues raised in doing so.

I chose the Singapore blogosphere as a case study as it is, according to others (Lin et al, 2006 and Hurst 2006), an isolated and distinct network. In adherence with Nadel (cited Cavanagh, 2007) the important factor is to not to arbitrarily demarcate a unit of analysis to study. By reducing the social to the network, it allows for the unit of analysis studied to materialise throughout the process of conducting the research. The network will be discovered using empirical evidence rather than imposed by the researcher at the beginning.

Social Network Analysis is a formal, mathematical technique of analysing relational data. It is concerned with the contacts, ties and connections, group attachments and meetings (Scott, 2000:3). “The relations are not the properties of the agents themselves, but of systems of parts; these relations connect pairs of agents into larger relational systems.” (Scott, 2000:3). The appropriate method for the analysis of Draft – Not for Citation

3

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

relational data is network analyses that are qualitative measures of network structures. Emphasis is on the ‘structure’ of social action. “Structures are built from relations” (Scott, 2000:4). Social meaning constructed by the group members of the network based on the perceptions and experiences of the context in which they are operating. Paths of connections run between the groups and these paths divide the groups into distinct regions. Lack of paths separates regions from each other (Scott, 2000:11). Paths run within the regions but not between the regions. These regions are constraints or boundaries. These boundaries are the ‘forces’ that determine group behaviour.

American Social Network Analysis on the other hand prefers to work with the concept of the ‘network’ as a geographical metaphor instead of a structural metaphor. Garrido and Halavais (2003) argue that, “A map of the communication network is roughly isomorphic to the structure of the relationships among the users.” Emphasis is on the communicative aspect of the network therefore seeing the ties as facilitators and not constraints or boundaries. The communicative basis of the network leads to the erosion of the distinction between organisational networks and interpersonal ones (Scott, 2000:33-36). It is more concerned with ‘intensity’ and ‘strength’ not ‘reciprocity’ and ‘durability’. The focus is on a form of social capital that facilitates action while at the same time arguing that individual components retain their preexisting identities (Cavanagh, 2007).

Actor-Network theory defines the social as a series of conduits or bridges that facilitate the flow of information, resources and ideas around a society. ActorNetwork theory is primarily concerned with an investigation of power. It is concerned with the way that ideas and beliefs form and develop in a society. Power emerges from interaction, which, similar to society and the social order is something that is always in process. This is different to the anthropological concentration on structures that are fixed and rigid over time. For Actor-Network theory society is temporarily situated, it can be activated and de-activated and is in a state of constant creation and re-creation. Actor-Network theory is not so much a theory rather it is a set of procedures to aid in the investigation of the social. It is a methodology not a method. The points or nodes in a network are potentially equal in terms of power. The ‘forces’ Draft – Not for Citation

4

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

that shape the interaction and the network are themselves generated by the network in the interplay of the component parts that comprise the network (Cavanagh, 2007:3334). The power of a network resides in the interactive and generative mechanism of the network itself. The network according to Actor-Network theory behaves as if it is a thing, a separate phenomenon from the component parts that make it up. ActorNetwork theory networks are not only comprised of people and personal connections but also texts, objects and a diverse array of material. The network as a whole, the materials, the nodes, the links, and the images do not possess fixed properties so agency is a property of the initial goal that resulted in the creation and formation of the network (Cavanagh, 2007:37).

According to Castells (2000), networks are comprised of personal contacts, places, technologies and functions. Those within the network select the elements and therefore there is exclusion. Networks are exclusionary and particular, therefore not universal. Again, the network transcends or goes beyond the component parts that comprise the network. The network is an actor in its own right and yet it is still dependent on all parts according to Castells (2000). The heterogeneity of all the parts is vital for its continuation. Networks operate globally and global imperatives make them globally orientated in order to maintain the dominant interests of the elite.

For Hardt and Negri (cited Cavanagh, 2007) the network form is the dominant form of power in modern society. They argue that the social network is – plural, inclusive and yet always contested and it enables action by the component parts rather than the network as a whole. They posit that the network has no central power relation and that “one essential characteristic of the distributed network form is that it has no centre. Its power cannot be understood accurately as flowing from a central source or even polycentric, but rather as distributed variably, unevenly, and indefinitely.” (Hardt & Negri, cited Cavanagh, 2007:42) The network behaves somewhat like a swarm, which may appear to be uncoordinated. However, communication is not the top-down decision-making model but is communication between the various component parts that comprise the network. This leads to the problem of Hardt and Negri’s ‘swarm’ seemingly arguing that the network acts and yet it is agency without an agent (Cavanagh, 2007:43). Draft – Not for Citation

5

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Knox et al. (2006) argue that network methods map roles comprehensibly and results in the incorrect assumption that they have delineated the ‘real’ social structures. It ends up reinforcing a view of relations that are unlike the everyday experiences of people. Knox et al. argue that Social Network Analysis’ focus on structuralism has in recent years shifted to attempts at developing a cultural approach. This shift in focus has come about because of the study of social movements by researchers such as Ansell (1997), Bearman (1995), Gould (1995), and Mische (2003).

Hyperlink Network Analysis

No network is ever truly ‘isolated’. However, Lin et al. (2006) have conducted attempts at defining a core group of users who continually return to a particular site of online interaction. They used various techniques to visualise and extract these communities. Lin et al. (2006) used blog ranking and their social connections, via hyperlinks of various types to devise a visual representation of blog communities. They established communities by assessing the level of mutual awareness through the various actions of bloggers, such as commenting on each others sites or using trackbacks to inform the writer of an article that they have linked to it. Lin et al. (2006) have defined the Singapore blogosphere as a “community with no obvious central topic”, and stated that it was a rather closed network, or rather closed off from the wider global network of bloggers. Hurst (2006) using the same data as Lin et al., created for the WWW 2006 Workshop, highlighted the same group of blogs as Lin et al. (2006). I collected an initial list of political blogs or seeds by engaging with the blogosphere. I started by gathering a large set of political blog addresses, or URLs, by downloading a list of political blogs1 compiled by myself and a group of Singapore bloggers and others. I used computer-assisted measurements for this research, namely issuecrawler 2 and Pajek to construct the following graphs.

1

List available http://spreadsheets.google.com/pub?key=pQcRq80yuyWXeqpakIg2ibA viewed 02/08/2007 2 http://www.issuecrawler.net/index.php

Draft – Not for Citation

6

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Figure 1. Singapore Political Blogosphere measured in July 2007. All network closeness centralisation score of 0.38627 and a betweenness centralisation score of 0.08783. Blogs grouped together by colour.

Using the same approach as Lin et al. (2006) what resulted was more akin to an egonetwork. It appears isolated from the wider global blogosphere with no links to blogs of other countries appearing. It is also important to point out that the resulting ‘network’ based on the more quantitative aspects of Social Network Analysis generates a representation that results in upwards conflation. I return to this in the discussion below.

Beyond Technological and Organisational Determinism When researching a physically decentralised social network or a network made up of individuals who form a network but are not necessarily members of a formal organisation, how do you avoid organisational as well as technological determinism? To argue that the network and hierarchy are determining the culture is to have slipped Draft – Not for Citation

7

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

into organisational determinism. Someone or something, like a law or the digital divide which is not engaged in the online network, may still retain power in shaping the online network by working as an external constraint. To avoid organisational determinism the researcher must consider broader social, culture and political phenomenon in which the blogosphere is contextualised. The intention of this research is to construct fieldwork that does not simply result in organisational determinism. The next concern is that of ‘Technological Determinism’ – that technology influences society but is not in turn affected by society. This is an idea founded on two assumptions: Technology progresses from less to more advanced configurations or rather that technological progress proceeds from lower to higher levels of development and social institutions must adapt to the ‘imperatives’ of the technological base.

Determinism is an interpretation of history, “which makes it seem as though the end of the story was inevitable from the beginning by projecting the abstract technical logic of the finished object into the past as a cause of development” (Feenberg, 1992). This is similar to Actor-Network theory that argues that the goals that initiated the formation of the network resulted in the nature of the network. Determinism confuses our understanding of the past and limits our imagination or ability to envisage different futures.

Feenberg (1992) argues that technology is a social object, technology is more than its explainable functions, and it has interpretable meaning. Technology has ‘social’ meaning and a ‘cultural horizon’. It might be argued that once the ‘object’ is fixed in its design, that it becomes a debate over the ‘goals’ and that the engineer has the last word. The focus on goals by managers, engineers, strips technology of its social context. However technology is a historically evolving phenomenon.

What the object is for the groups that ultimately decide its fate determines what it becomes as it is redesigned and improved over time. Technology then can only be studied by studying the situation of the various groups involved. In this instance the cultural situation of Singapore is important when conducting an ethnographic study of the Singapore blogosphere. Draft – Not for Citation

8

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Ethnography

The ethnographic approach enables the researcher to ascertain what the network boundaries are as defined by the process of attaining access. It also enables the classification of situations (Mische & White, 1998). The ethnographic researcher takes part in the everyday experiences as a blogger uncovering meanings pertinent to those involved.

This approach enabled me to overcome the ‘ego-network’ of my initial data collection. While engaging daily with the others online I was aware of who can gain access to the self-declared ‘socio-political blogosphere’ uncovered in figure 1 above. Self declared political groups in Singapore tend to focus on discussions framed within the dominant ideological focus of the Singaporean government.

Political talk in Singapore takes place within the ‘discourse of communitarianism’. The discourse of communitarianism is enshrined in the language of; ‘shared values’, ‘national interest’, ‘good government’, ‘survivalism’, ‘pragmatism’, ‘political stability, ‘collective interests, ‘social stability’, ‘conservative’ and, ‘economic growth’ (Chua, 1997) of the Singapore government. The discourse of globalism or anti-globalisation has made inroads into the online discourse of the Singapore blogosphere (see appendix 1). According to Fairclough (2003: 159), “styles are the discoursal aspect of ways

of

being,

identities.”

Fairclough

(2003)

goes

on

to

argue

that

styles/identification is distinct from discourse/representation and genres/action but are not independent. Fairclough (2003) highlights different types of modality, which can be associated with different types of exchange and speech functions. He highlights the characters of politicians, management gurus, experts, priests and academic styles of speech. Therefore, by limiting my initial collection of blogs to those who engaged in the above speech styles I was assigning the boundary of the network based on the discourse of the self-proclaimed political elite.

The researcher encountered blogs that are essentially political in nature and yet did not appear on the initial list of self-proclaimed political blogs. Individuals approached me via email and co-authored blogs before finally setting blogs up on their own. I Draft – Not for Citation

9

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

was then able to harvest a list for the Singapore Gay and Lesbian blogosphere from Sayoni Speaks (http://blog.sayoni.com/). The addresses of the Singapore Social blogosphere

came

from

the

Singapore

Social

Media

Directory

(http://sgsocialmediadir.wikispaces.com/).

Figure 4. From Left to right: Singapore Gay and Lesbian blogosphere: Closeness Centralisation = 0.38918, Betweenness Centralisation = 0.16183; Singapore Social blogosphere: Closeness Centralisation = 0.34469, Betweenness Centralisation = 0.08866.

I then combined the two lists of blogs to the initial list along with a Malaysian group and a Christian group. Using these blogs as seeds, I was then able to compile a list that generated the following graph.

Draft – Not for Citation

10

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Figure 3. The Singapore Blogosphere 2008: Contains 1,239 nodes. The size of the node represents the Betweenness Centrality (BC), the larger the node the larger the BC score. The black lines refer to the hyperlink connections between the nodes.

To the left of the graph are English-speaking bloggers and to the right are Malayspeaking bloggers. It is clear from the graph and confirmed by further Social Network Analysis that the Singapore blogosphere is comprised of two distinct factions (see Appendix 2 for a further breakdown of the top 50 key players in table 2, ranked according to Betweenness Centrality in the Singapore blogosphere of June 2008 and figure 4 further illustrating the two distinct factions).

Discussion

The two distinct factions comprised of English-speaking bloggers and Malayspeaking bloggers with limited cross-linking occurring is evidence of the cultural context of Singapore shaping the online network. This is however not to simply ignore the influence of other generative mechanisms. It is evident that the choice of Draft – Not for Citation

11

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

blogging technology used by the bloggers is also having an effect on how the blogosphere is breaking down into two factions. The English-speakers use technology, such as Blogspot, Wordpress, and Livejournal while the Malay-speakers tend to favour alternative providers such as Blogdrive.

The Singapore blogosphere on this occasion acted not simply as a means of reinforcing the discourse of ‘communitarianism’ or ‘globalism’ (Chua, 1997) but also managed to amplify criticisms of that discourse. It was criticism of the justification of the government to know what the public is or wants. The most important political and social issues faced by the Singapore public, as defined by the ‘public’ and not mediated by a state controlled press, will continue to find space and expression in the Singapore blogosphere, retaining the potential to engender democracy.

The Singapore blogosphere is not a “community with no obvious central topic” (Lin et al., 2006). When seen through the cultural lens of the ethnographic approach what emerges is a ‘public’ engaged in oppositional discourse. Using the same approach as Lin et al. (2006) what resulted was more akin to an ego-network. It appears isolated from the wider global blogosphere with no links to blogs of other countries appearing. The motivations or the will of those involved in the initial stages are evident at the beginning of the narrative but the boundaries are the ‘discovered’ structural elements of the network. Agency becomes structure; and then agency is a mere epiphenomenon of the network. The so-called quantitative elements of social network analysis runs the risk of reducing the network into a metaphor for ‘the masses’ when conducting hyperlink analysis; in turn reducing its own efficacy into that of an advanced polling system of online networks. The shift in focus to a cultural approach within Social Network Analysis will undermine the structural dominance. However, it will not overcome the problem of upward conflation and the reduction of the agent to that of an epiphenomenon.

If as Dewey (1927, cited Kelly & Etling, 2008) argues, “The outstanding problem of the Public is discovery and identification of itself.” The problem for Social Network Analysis is to ensure that it does not conflate that public that is appearing in the blogosphere with the wider social structural elements, be they structures resulting Draft – Not for Citation

12

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

from the technology and institutional generative mechanism or cultural forces, be they ethnicity, race or nationalism.

There is no hierarchy; no node, (be that a blog, technology, ethnicity or cultural mechanism), takes precedence over another. The order is in constant flux with total inclusiveness. Singapore society and the flow of information pre-date the existence of the Internet and specific applications within it such as the blogosphere. These nodes are interruptions in the flow of information. The nodes merely channel the flow of information. How is the Singapore blogosphere affecting that flow?

References Adamic, L. A. and Adar, E. (2001). You are what you link. 10th annual International World Wide Web Conference, Hong Kong. Retrieved June 19: 2001. Ansell, C. K. (1997). "Symbolic Networks: The Realignment of the French Working Class, 1887–1894." AJS 103(2): 359-90. Bearman, P. S. (1993). Relations into rhetorics: local elite social structure in Norfolk, England, 1540-1640, Rutgers University Press, New Brunswick, NJ. Castells, M. (2000). The Rise of the Network Society. Blackwell Publishers. Cavanagh, A. (2007). Sociology in the Age of the Internet, Open University Press. Chua, B. H. (1997). Communitarian Ideology and Democracy in Singapore. New York, Routledge. Deleuze, G. and F. Guattari (2004). Anti-Oedipus: Capitalism and Schizophrenia, New York, Continuum International Publishing Group. de Nooy, W. et al. (2005). Exploratory Social Network Analysis with Pajek. New York, Cambridge University Press. Dreyfus, H. L. (2001). On the Internet, London, Routledge. Fairclough, N. (2003). Analyzing discourse textual analysis for social research. London, Routledge. Feenberg, A., (1992) Subversive Rationalisation: Technology, Power and Democracy, Inquiry, 35: 3 / 4. Garrido, M. and Halavais, A. (2003). Mapping networks of support for the Zapatista movement. In McCaughey, M. & Ayers, M.D. (2003). Cyberactivism: Online Activism in Theory and Practice, Routledge. Gould, R. V. (1995). Insurgent Identities: Class, Community, and Protest in Paris from 1848 to the Commune, University Of Chicago Press. Hurst, M.W. (2006). Interactive Map of the Blogosphere. http://datamining.typepad.com/data_mining/2006/07/interactive_map.html. Jackson, M.H. (1997). Assessing the structure of communication on the World Wide Web. Journal of Computer-Mediated Communication, 3(1): 1997. Kelly, J. and Etling, B. (2008). Mapping Iran’s Online Public: Politics and Culture in the Persian Blogosphere. Research Publication No. 2008-01, The Berkman Center for Internet & Society Research Publication Series.

Draft – Not for Citation

13

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Knox, H. and Savage, M. et al. (2006). "Social networks and the study of relations: networks as method, metaphor and form." Economy and Society 35(1): 113-140. Lin, Y.R., Sundaram, H., Chi, Y., Tatemura, J., Tseng, B. (2006). Discovery of blog communities based on mutual awareness. Third Annual Workshop on the Weblogging Ecosystem: Aggregation, Analysis and Dynamics, at the 15th Annual World Wide Web Conference-WWW: 2006-03. Mische, A. (2003). "Cross-talk in Movements: Reconceiving the Culture-Network Link." Social Movements and Networks: Relational Approaches to Collective Action: 258–280. Mische, A. and White, H. (1998). Between Conversation and Situation: Public Switching Dynamics across Network Domains. Social Research, Vol.65, No.3. Radcliffe-Brown, A.R. (1940). ‘On Social Structure.’ Journal of the Royal Anthropological Society of Great Britain and Ireland 70(1-12). Scott, J. (2000). Social Network Analysis: A Handbook, Sage.

Draft – Not for Citation

14

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Appendix 1. Discourse and Styles of Discourse Discourse Discourse of Communitarianism

Occurrences (Frequency %)

Anti-Social Cohesion Questions the Discourse Disagrees with Social Cohesion Sarcasm Total

50 (25%) 24 (12%) 18 (9%) 94 (47%)

Total

19 (9.5%) 8 (4%) 5 (2.5%) 3 (1.5%) 35 (18%)

Total

12 (6%) 11 (5.5%) 6 (3%) 4 (2%) 4 (2%) 4 (2%) 1 (0.5%) 42 (21%)

Total

6 (3%) 6 (3%) 5 (2.5%) 4 (2%) 4 (2%) 3 (1.5%) 28 (14%)

Social Cohesion Agrees with Social Cohesion Democracy Accountability Refers to the Greater good

Styles of Discourse Occurrences (Frequency %) Politician 49 (27%) Personal 39 (22%) Citizen 28 (16%) Academic 22 (12%) Journalist 19 (11%) Activist 10 (5.5%) Expert 9 (5%) Lawyer 2 (1%) Priest 1 (0.5%)

Discourse of Globalism Globalism Supports Globalism Progress and Technology Inevitable Knowledge Based Economy Competing on a Global Scale Consumerism Processes Anti-Globalism Disagrees with Globalisation Social Justice Negative Interpretations Unfair Welfare Rights

Table 1. The occurrence and frequency of the nine styles of discourse and two forms of discourse with subsets.

Table 1. shows that a politician style of discourse occurred on 49 (27%) occasions, personal 39 (22%), citizen 28 (16%), academic 22 (12%), journalist 19 (11%), activist 10 (5.5%), expert 9 (5%), lawyer 2 (1%) and priest 1 (0.5%). The most dominant form of discourse of communitarianism and discourse of globalism was that of anti-social cohesion which scored 94 (47%) occurrences followed by a discourse of globalism (42 (21%)), social cohesion (35 (18%)) and anti-globalism (28 (14%)). Draft – Not for Citation

15

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam

Appendix 2. Top 50 Key players of the Singapore Blogosphere June 2008 by Betweenness Centrality. ID blogdrive.com tomorrow.sg photobucket.com wordpress.com technorati.com sgblogawards.omy.sg kuncup.blogdrive.com flickr.com bloggersg.com nus.edu.sg creativecommons.org klinikminda1.blogspot.com ads.blogdrive.com kennysia.com apple.com singabloodypore.rsfblog.org azlina.net rinaz.net friendster.com clappingtrees.com hendri.squoar.com yesterday.sg youtube.com sabrina.sg mr-endoh.com sgblogs.com design.blogdrive.com blog.dk.sg chaosdingo.lah.cc sgentrepreneurs.com i-speak.blogdrive.com limetouch.com toomanythoughts.org mrbrown.com sixapart.com jason.sg ping.sg nadnut.liquidblade.com feedburner.com bleongcw.typepad.com imageshack.us thunderstorms.blogdrive.com advertlets.com blogshopr.com facebook.com vantan.org

Degree 174 216 91 111 120 75 130 96 81 48 102 100 153 100 71 75 54 133 44 92 86 57 64 90 102 100 54 110 38 102 29 44 59 89 60 78 123 95 57 100 33 45 42 32 68 69

Betweenness Centrality 53533.926 51178.59 43716.617 33384.738 29589.938 29244.604 28501.156 25962.582 19397.098 19138.492 18930.762 18157.41 17577.324 16561.314 16197.214 15155.418 12439.063 11686.452 10979.605 10529.274 9688.821 9657.044 9583.163 9443.527 9430.799 9320.283 9299.348 9147.646 8939.661 8720.266 8460.372 8008.385 7892.979 7697.419 7671.539 7591.144 7559.005 7376.924 7187.758 7164.742 7018.768 6981.098 6706.255 6681.252 6643.85 6615.124

Draft – Not for Citation

Closeness Centrality 4279 3929 4096 4074 3919 4471 4396 4016 4058 4624 4020 4418 4554 4133 4056 4192 4324 4103 4216 4134 4166 4282 4069 4109 4163 4163 4394 4166 4322 4232 4414 4233 4326 4215 4321 4227 4226 4168 4273 4217 4558 4457 4346 4425 4268 4249

Factions2 1 2 1 2 2 2 1 2 2 2 2 1 1 2 2 1 1 2 2 2 2 2 2 2 2 2 1 2 2 2 1 2 2 2 2 2 2 2 2 2 1 1 2 1 2 2 16

Paper presented by Steven McDermott ([email protected]) at the Communication Networks on the Web 18 - 19 December 2008, Amsterdam School of Communications Research, University of Amsterdam fastonlineusers.com exampaper.com.sg mizi.blogdrive.com feedjit.com

41 64 81 32

6577.517 6052.812 5958.695 5957.418

4473 4197 4596 4469

1 2 1 1

Table 2. Faction 1 = Links within the Malay-speaking blogs, Faction 2 = Linked to English-speaking blogs. Total number of blogs ranked is 1,239.

Figure 4. Singapore blogosphere June 2008 as two factions

Figure 4. The list of 1,239 blogs is in the graph above. The red nodes to the left represent faction 2 in table 2 and are the English-speaking linked blogs. The blue nodes to the right are faction 1 from table 2 and represent the Malayspeaking blogs. The red lines between nodes are hyperlinks between the two factions. Size of node represents Betweenness Centrality score.

Draft – Not for Citation

17