Who really invented Google's search algorithm?

The politics of searching

Matteo Pasquinelli

To person

Author, curator, and researcher at Queen Mary College, University of London, author of Animal Spirits: A Bestiary of the Commons (2008) and editor of Media Activism (2002) and C'Lick Me: A Netporn Studies Reader (2007) ; he writes frequently on subjects of French philosophy, media culture, and Italian post-operaism.

Diagram of cognitive capitalism and rentier of common knowledge

Matteo Pasquinelli examines how Google extracts added value from individual actions and general intellect and transforms it into network value and wealth. He describes this process using the term "cognitive lease".

Small advertisement - big impact. Browser screenshot with page rank. (& copy bpb)
    The innermost core of [Google] is the PageRank algorithm that Brin and Page wrote while studying at Stanford in the 1990s. They realized that every link to another page is an expression of judgment. The person posting the link is saying that they think the other side is important. They also realized that every link on the web contains a tiny bit of human intelligence, and that all links together contain a huge amount of intelligence - far more than any human mind could possibly possibly possess. The Google search engine crawls this intelligence link by link and uses it to determine the importance of all pages on the web. The more links point to a page, the higher its value. As John Markoff explains, "[Google's software] systematically exploits human knowledge and decisions about what is important". Every time we set a link or even click on an existing one, we feed the Google system with our intelligence. We're making the machine a little smarter - and Brin, Page, and the Google shareholders a little richer.


The reversal of the panopticon: Google as a machinic parasite of the common intellect (or: the production of value)

Much of the criticism of Google focuses on the imperial nature of its monopoly: its dominant position, privacy issues, censorship, global dataveillance. In contrast, there are few studies of the molecular economy at the core of this predominance. While many critical articles on Google misuse Foucault's jargon and indulge in the idea of ​​a digital panopticon, Google's power arises from an economic matrix that is determined by the cabalistic formula of PageRank - that sophisticated algorithm that determines the importance of a website and the hierarchy of Google - Search results determined. (2) As will be shown below, the function of PageRank can be easily understood. However, a "political economy" of this apparatus is still lacking.

Even if the biopolitical dimension of Google is much discussed (often in the post-structuralist jargon mentioned), there is still a lack of a bio-economic analysis that explains how Google draws value from our lives and the common intellect and transforms it into network value and wealth. Even if it is often misused, the Foucault paradigm makes real problems visible, albeit only in part: Google's supremacy is not simply something metaphysically given, but is based on a technology platform and the business model of the search engine. As Paolo Virno thinks, in our understanding of biopolitics we should start from the potential of our living body and labor: The biopolitical structures follow as an apparatus that serves to capture this potential. (3) The metaphor of the panopticon must be reversed: Google is not just an instrument of dataveillance that looks down on us from above, but an apparatus of value creation from below. In particular, Google creates and accumulates value through the PageRank algorithm and through the transfer of collective knowledge - this is the central theme. The political economy of Google begins with the political economy of PageRank.

The first description of Google's PageRank and the starting point for the entrepreneurial adventure of Sergey Brin and Lawrence Page was included in their 1998 paper The Anatomy of a Large-Scale Hypertextual Web Search Engine. (4) The PageRank algorithm resulted in some revolutionary ones Changes in the technology of information acquisition and search engine technology of the late 1990s: The seemingly flat ocean of the Internet was transformed by Google into dynamic hierarchies according to visibility and importance. The ranking of a website is pretty easy to understand - its value is determined by the number and quality of the pages that link to it. A link originating from a high ranking node has more value than a link originating from a low ranking node.

While search engines like Yahoo were still indexing the web by hand in the 1990s and adapted it to the typical tree structure of encyclopedic knowledge, Google invented a formula that made it possible to follow a semantic value across the dynamics and chaos of hypertext. PageRank began to describe web pages based on popularity, and the search engine produced results that were ranked according to that criterion. In addition to Yahoo's trees and Google's rankings, there are numerous other techniques for gathering information. (5) The software of the PageRank algorithm is a highly complex construct that only professional mathematicians can understand. This contribution is therefore limited to those generally understandable aspects that are necessary for an initial political economy of this apparatus.

This diagram bears no resemblance in any way to the centralized structure of the panopticon described by Foucault in Surveillance and Punishment. (7) The fluid, hypertextual nature of the web (and the noosphere in general) requires a different illustration. A diagram of cognitive capitalism can be followed intuitively if every symmetrical link in the structure of the hypertext is replaced by an asymmetrical vector for energy, data, attention or value. What PageRank makes visible and measures is this asymmetrical constitution of every hypertext and every network.

The source of inspiration for PageRank was the academic citation system. The "value" of an academic publication is known to be determined by a very mathematical process in which the number of citations in other articles that refer to the relevant contribution is decisive. Therefore, the ranking for an academic journal corresponds to the sum of all citations referring to it in other publications. As Brin and Page explain:
    The academic citation process was applied to the web by counting the citations or links on a specific page. This gives an approximate value of the importance or quality of a page. PageRank takes this idea further by not counting all links on all pages equally and by normalizing based on the number of links on a page. (8th)
This genesis of PageRank, rooted in the medium of the book, should not be underestimated. The value of every cognitive object can be described in a similar way in the society of the spectacle and its dazzling brand economy. In the mass media, the value of a commodity is mainly created through the concentration of attention and collective desire. From academic publications to private labels to internet rankings, the same value concentration processes can be assumed everywhere. Just as digital colonization has given any offline appearance an online presence, the matrix of social and value relationships has migrated into cyberspace and can now be tracked and measured by search engines. In particular, PageRank describes the attention value of each object to a degree that has made it the most important source of visibility and authority even in relation to the mass media. PageRank contains a formula of value accumulation that is hegemonic and can be adjusted across media boundaries: a useful diagram that describes the attention economy and cognitive economy in general.

The concept of attention economy is useful if one wants to describe how the value of a commodity is (partially) produced today through a media-driven accumulation of social desire. (9) Other schools of thought speak of "cultural capital" (Pierre Bourdieu), " collective symbolic capital "(David Harvey) or" general intellect "(especially the more cognitively oriented post-operaism). Before the Internet, this process was described as a generic collective force; With the Internet, the structure of network relationships around a particular object can be easily tracked and measured. PageRank is the first mathematical formula that calculates the attention value of each node in a complex network as well as the general attention capital throughout the network. What is the value measured by PageRank? Interestingly, every link and attention vector is not simply an instinctive gesture, but the concretization of intelligence and often a conscious act. If it is fashionable to describe the network society as the convergence of "streams of desire", then it must be taken into account that these streams with knowledge are occupied and belong to the activity of a community intelligence.

In the quote at the beginning of this article, Nicholas Carr described how Google's PageRank works very well - how it feeds on our collective intelligence, and how value is produced and accumulated based on that collective intelligence. PageRank thus creates its own attention economy, but a large part of this attention capital is actually based on intellectual capital, since every link represents a condensation of intelligence. In this sense, Google is a parasitic device that siphons off the value produced by community intelligence. (10)