Tangible Algorithms of the Everyday, Marialaura Ghidini

In front of me there is a library of 100 books; 100 diverse entries organised numerically in a table showing the books’ titles and authors. I have been going up and down this index, seeking connections between them, and trying to recall if I have read this or that book. I see The Female Man by Joanna Russ, a feminist science fiction novel written in the 1970s, have I come across it before? Does it hold any relationship to After the Future by Franco “Bifo” Berardi, a theory book about the development of thoughts concerning the future from the 70s onwards? I have an additional entry point to this library at my disposal: a .gif file that shows all book covers and offers an accelerated visual display of the index in the table. The more I observe it the more it becomes evident that the method I am using to find connections between the books, the authors and the content is missing something. Most probably, you, the viewer of the physical installation of the library (on display on a Vitsœ 606 Universal Shelving System) might share the same feeling.

What is the something that is missing?

This library is a special collection that gathers together the top 100 books Amazon1 has recommended to one of its customers, the artist Martin John Callanan. It has a title, I Cannot Not Communicate, and it encapsulates 15 years of Callanan’s reading habits (and purchases). Hence, what you and I are browsing is a library created by Amazon’s Recommendation algorithm, also known as item-to-item collaborative filtering (patented in 2001); an algorithm that “determines the most-similar match for a given item […] by finding items that customers tend to purchase together”2. Therefore this library is not just representative of Callanan’s own reading taste, but of the way in which Callanan, the reader-buyer, has been ‘profiled’ by the algorithm of the giant online retailer — whose trading activity spans from consumer electronics and software to industrial supplies and tools, including groceries, personal-care items, garden tools and jewellery to mention just a few categories. Such profiling is based on understanding Callanan in relation to the inclinations, choices and recommendations of other reader-buyers, that is within a database of collective behaviours built around locating affinities and trends of groups of interests, or to use the artist’s own words “built through mediated consensus”3.

As in many previous artworks, such as I Wanted to See All of the News From Today, an online project that each day displays the front pages of over 600 international newspapers, and Wars During My Lifetime, a growing list of wars that have taken place during the artist’s life firstly presented as an official declaration by a Town-crier (2012) — works also presented in printed format — Callanan is the I, the subject through whom, we, the viewers, can observe the way in which the workings of computational systems and software programmes aimed at collecting and visualising data might impact on our everyday actions. I Cannot Not Communicate has a direct reference to knowledge construction; the library after all is an institution born to be the depositary of human knowledge accessible by many. However, here the library is presented as the 21st century variant of its precursors, which were housed in buildings and organised around subjects determined by specialised librarians. Such variation is given by the mode in which an algorithm understands the I, the reader-buyer, in relation to the ‘others’; hence the creation of ‘non-orthodox’ methods of categorisation: the recommendations. Without taking any open critical stance on the online retailer, the artwork exposes a new way of accessing and sharing knowledge (and perhaps creating knowledge legacies) that is based on trust. This trust is neither in an institution nor in the retailer itself; it is a belief in the workings of an algorithm that combines computing with human behaviours — the tastes of the anonymous groups of the other reader - buyers using the online service.

For each item in product catalog. I1
For each customer C who purchased I1
  For each item I2 purchased by customer C.
    Record that a customer purchased I1 and I2.
For each item I2
  Compute this similarity between I1 and I2

.

I Cannot Not Communicate shows the possibility of creating variable ways of sharing knowledge that are based on a collectivity that performs actions (purchases), has preferences and is understood algorithmically. And what you and I should not overlook here is that these are variables generated from within commercial systems, thus often not entirely transparent. It might be interesting to note that the attempts that Callanan has made (since 2001) to obtain clarification on the exact functioning of the algorithm have been negated by the retailer, especially with regard to practices of product placement or other weighing of recommendations.

It is difficult not to think of Siegfried Zielinski’s concept of anarcheology4 — later refined into the “more positive” one of variantology5 — an “archaeology that entails envisioning, listening, and the art of combining by using technical devices, which privileges a sense of their multifarious possibilities over their realities in the form of products” and, most importantly, as a history that “cannot be written with [...] a mindset of leading the way”. Zielinski states that even if anarcheology arises from the ‘necessity’ to give order to things — such as indexical temporal and spatial classifications — its diversity is akin to the functioning of the London Library in St. James Square. In fact, in there “you are less likely to find the book you have long been looking for without success” and more likely “to find one you did not even know existed, but is of far greater value”. Such value is given by the realm of possibilities that ‘the find’ uncovers. Similarly, the heterogeneous combination of books proposed by I Cannot Not Communicate shows us renewed possibilities of celebrating the “fortuitous find”, which is frequently not associated with computational processes and databases.

Unlike artists operating with System and Computational Arts, such as Manfred Mohr, Callanan’s focus is not on writing its own algorithm to create rules that generate an artwork — which otherwise would not exist and remain disembodied. Here the artist uses an already existing algorithmic system to make its inherent abstraction tangible to the viewers of . The artwork has a functionality, it is a library, and turns its viewers, you and I, into potential readers who enter the system exposed by Callanan the artist through Callanan the buyer-reader. From passive viewers, we become active readers, and potential subjects of further future libraries that could be created by inputting our preferences for books or even for objects and tools into the Amazon database, through filling our shopping carts. Perhaps we will be able to further expose the workings and possibilities of the retailer’s semi-transparent human-behaviour-based algorithm. For the moment let’s browse and experience the “fortuitous find” and variants that the embodied library of Callanan offers us on real shelves.

  1. The Amazon.co.uk website opened in October 1998 as an online bookseller. Since then the company “has dramatically expanded its range of products and now offers millions of items” across a multitude of categories to over 270 million active customers. The retail websites Javari.co.uk, which sells shoes and handbags, is also part of the enterprise. Christopher North, the Managing Director of Amazon.co.uk, describes the website as capable to “provide a unique buying experience” through “a raft of personalisation features ensuring that the site content and communication with customers is interesting and relevant”. Information taken from About Amazon.co.uk – Overview. URL. [Accessed: 15 April 2015]
  2. Linden, Smith and York, 2003. Amazon.com Recommendations: Item-to-Item Collaborative Filtering. In IEEE Internet Computing. Volume 7 Issue 1, pp. 76-80
  3. From an email conversation with the artist; April 2015.
  4. Zielinski, S., 2006. Reality as a Mere Shadow of what is possible. In Deep Time of Media. Cambridge: The MIT Press.
  5. See Zielinski, S., 2006. Variantology 1: On Deep Time Relations of Arts, Sciences and Technologies. London: Buchhandlung Walther König.
Marialaura Ghidini is an independent contemporary art curator and researcher currently based in London, UK. She is founder director of or-bits.com, a web-based curatorial platform devoted to supporting artistic practices and discourses that explore the relationship between the arts and technology, online and offline.

I Cannot Not Communicate is a new artwork premiering at Vitsœ, 33 Bond Street, New York City, 14 May — 14 July 2015.

Made possible with the Philip Leverhulme Prize, awarded by The Leverhulme Trust.