Network Surplus Value

In Googlearchy, Matthew Hindman elucidated the inequalities of PageRank through the lens of traffic, which measured the visibility of any site based on its search result ranking and the number of links pointing to it:

Links do not just provide paths for surfers…If links help determine online visibility, how links are distributed tells us much about who gets heard on the Web…The importance of links challenges notions that online equality is easy or inevitable (ibid:132).

By the mid 2000s, political discourse was already filtered thanks to Googlearchy, thus

‘deliberative democracy’ was prohibited by the infrastructure itself––‘the social, economic, political and even cognitive processes that enable it’ (ibid:130). Googlearchy purported that

‘niche dominance’, where only a small portion of websites receive most of the traffic, is self-perpetuating––the sites with more links receive greater traffic whereas those with few links are harder to find and require better searching skills (2009:55).⁶⁹ With hyperlinks continuously being added and Google collecting data ad infinitum, the bias in search engine results simultaneously became more noticeable––‘bias that invites users to click on links to large websites, commercial websites, websites based in certain countries, and websites written in certain languages’ (Van Couvering 2010:3).

outsourcing the decision and labour to the community of the web (users) as a whole (2006:242).

It is this ‘social networking’, where

people on the web decide what counts literally, with whatever they like for whatever reason, vanity, pleasure, and they produce links, which are connected to each other, and these are counted and eventually monetised (Benkler 2008).

It was PageRank that introduced the ‘original notion of quality’ or relevance on the web, determined by the ‘collective intelligence’ of those on it and the ‘opinions of the millions of people that populate this universe, is exploited to determine the importance, and ultimately the quality, of that information’ (Franceschet 2010:6.0). In this sense social relations become

significantly more important than they ever were as an economic phenomenon, where the labour of users clicking willingly on links that interest them adds to ‘social networking’ (Benkler 2006) yet the criteria remain invisible. Matteo Pasquinelli elucidates the exploitation of users’

cognitive intelligence through the hidden ‘immaterial factory’ of Google, with each link that is clicked on adding value: ‘If commodity is traditionally described by use value and exchange value, network value is a further layer attached to the previous ones to describe social relations’

(2009:156). Although PageRank determines the ranking of these links, Google itself does not produce content; rather it is expropriated in the form of ‘cognitive rent’ (ibid). ‘Accordingly, Google can be described as a global rentier that is exploiting the new lands of the internet with no need for strict enclosures and no need to produce content too’ (ibid:159).

By determining its own attention economy with these already ‘trusted’ links, PageRank captures

‘living time and living labour time and transforms the common intellect into network value’

(Pasquinelli 2009 cited by Bilic 2017:13). The re-appropriation of network value through the labour of user interaction and engagement is the common intelligence that Benkler (social production and networking) and Pasquinelli (network value) articulate. ‘Google’s implicit design decision, is “an intricate form of populist hypermedia” as Kleinberg put it’, where websites make themselves found, thereby reading ‘the web as the embedded intelligence of millions of users’ (Peters 2015:327). With this ‘cognitive capitalism’ (Boutang 2012), PageRank was able to restructure links to such a degree as to create a ‘link economy’, turning the hyperlink into the ‘currency of the web’ (Pasquinelli 2009; Helmond 2013). Value is commensurated by the amount of hits, or traffic to the site and how the ‘network surplus value’ of the nodes is redistributed (Pasquinelli 2009).

Indeed, PageRank produces what Deleuze and Guattari described as machine surplus value referring to the surplus value accumulated through the cybernetic domain, or the transformation of a surplus value of code into a surplus value of flux (ibid:156).

The value of these social relations is then measured and PageRanked, making the amount of attention received by a particular text or website visible thus adding even more value or trust to the user.⁷¹ Along these lines, Brin and Page mention that PageRank carries out judgments that reflect the personal choices and affiliations of the ‘user’ who became central to their

terminology of human computer interaction. The ‘number of new users inexperienced in the art

71 Produsage is a portmanteau of production and usage coined by Axel Bruns. Whether the (prod)user should be remunerated for the production of content that is difficult to produce and the further questions surrounding the terms of the digital labour involved in regard to search habits is beyond the scope of this thesis.

of web research’ (1998:107) increased and eventually became ‘trusted users’, which I will further discuss in Chapter 8. In their ‘Future Work’ section, Brin and Page explain that they foresee applying the ‘trusted user’s’ search history to PageRank, which ‘can be personalized by increasing the weight of a user’s home page or bookmarks’ and ‘result summarization’ text’

(ibid:6.1).⁷² Moreover they also ‘plan to support user context (like the user’s location) and wish to ‘extend the use of link structure and link text’ (ibid). Over the years, by pinpointing

locations, search histories, IP address and tracking their ‘usage’ with the collation of data, Google created personalised search results, which I will return to in Chapter 5.

However, Brin and Page mention that companies whose business model is co-opting users’

attention and manipulating the ‘unseen’ metadata contained in search results for profit is a

‘serious problem’, yet they are keenly aware of data produced from usage.

Usage was important to us because we think some of the most interesting research will involve leveraging the vast amount of usage data that is available from modern web systems. For example, there are many tens of millions of searches performed every day.

However, it is very difficult to get this data, mainly because it is considered commercially valuable (Brin and Page 1998:109).

Already presaging the coming era of ‘dataism’ (Harari 2001), the more people used PageRank the more it improved and Google simultaneously constructed a proprietary database. With the building of ‘possibly the most lasting, ponderous, and significant cultural artefacts in the history of humankind’, Google’s ‘Database of Intentions’ constantly captured users’ search queries and histories thereby enabling ‘a new culture to emerge’ (Battelle 2005:7 cited Noble 2018:148).

Subsequently search data became a commodity, along with monitoring and influencing user behaviour in an era of ‘surveillance capitalism’ (Zuboff 2015), which I will address in more detail in Chapter 9.

In their Appendix A, Brin and Page explain that web search ‘remain[s] largely a black art and to be advertising oriented’ (1998) and that they ‘believe the issue of advertising causes enough mixed incentives that it is crucial to have a competitive search engine which is transparent and in the academic realm’ (ibid:107). Referring back to the intertwining of research and search through relevance, academic research was determined by a few experts or peers, where

‘[c]itations generally witness the use of information and acknowledge intellectual debt’

(Franceschet 2010:6.0). As shown previously, the shift to hyperlinking URLs (Uniform Resource Locator) is a reflection of the authoritative pages found when surfing the web, showing ‘preferential attachment’ (Halavais 2009). This linking paradigm in turn reflects the analysis of attribution via the hyperlink as the essential structure or ‘mesh’ of the web (Berners-Lee 1989) indicated by associative linkages between the various parties, as mentioned in Chapter 2.

Google treats its search algorithm […] like a happy-go-lucky pragmatist willing to crawl the snail trails of associations wherever they lead […] A page is valued by how other actors in the system value it, and their power to value it is determined by the value that others place in them’ (Peters 2015:330-331 emphasis mine).

Nowadays the importance of a contribution is based on the ‘collective evaluation’ of

bibliometrics––how citations from the academic community are measured and commensurated through attention.

Additionally, Brin and Page impart that besides being a ‘high quality search engine’, Google is

‘a research tool’, and they close with the ‘hope Google will be a resource for searchers and researchers all around the world and will spark the next generation of search engine technology’

(1998:112). They point out that ‘[a] Web search engine is a very rich environment for research ideas’ and they also state their goals of ‘pushing more understanding into the academic realm’

(ibid). The authors also compare the increasing amount of documents available due to the growth of the web with the lack of research documents about search engines, lamenting that so few are willing to discuss the details of the technology in academic papers. ‘According to Michael Mauldin (chief scientist, Lycos Inc) [Mauldin], the various services (including Lycos) closely guard the details of these databases’ (ibid:107). This is quite contradictory in regard to their own ‘visibility management’ (Flyverbom et al. 2016) regarding the PageRank algorithm as a black box, or for some a black art similar to an alchemical formula that performs for its

creators, in this case a corporation that maintains its ‘secret recipe’.

Lastly, in their ‘Future Work’ section Brin and Page state that much ‘remains to be done’––

‘updating’ is an area that needs attending to, along with ‘smart algorithms’ that discriminate what should be recrawled and new ones to be crawled (1998). This foreshadowing of artificial intelligence applied to search has now become a reality with the shift from PageRank to RankBrain, a ‘machine learning’ algorithm, which I will explain at the end of Chapter 5. But first a reflection on methods where I tested out Google Search and PageRank with specifically chosen keywords in a self-designed experiment, gathering data on myself in the hope to glean insight into its proprietary and hidden algorithmic ranking procedures and processes.