Wikipedia’s classifications and its consequences

Wikipedia continues to demonstrate its importance to social and political relations beyond the platform. This month, Elon Musk admitted that Twitter’s classification of NPR (National Public Radio) in the United States as “state-affiliated media” was drawn from Wikipedia. The relevant page is actually a category page that automatically collates articles that share the category of “publicly funded broadcasters”. Wikipedia’s categories constitute a collection that is much more loosely adjudicated and widely distributed than a traditional article. They are more like tags (multitudinous and not instrumental to search and retrieval). Categories are barely noticed… until they become controversial (like when a list of American novelists was found to include hardly any female authors). 

“Wikipedia’s classifications”, as wikihistories Advisory Group member, Brian Keegan noted on a LinkedIn thread, “have consequences” (mirroring the subtitle of a classic book in Science and Technology Studies by Geoffrey Bowker and Susan Leigh Star). It is yet another example of the power of Wikipedia’s content to influence meaning in the world (even if that meaning was badly translated when it moved to Twitter’s infrastructure). Wikipedia is authoritative at least partly because it is an encyclopedia and because it aims to represent everything.  

In this first newsletter of the wikihistories project, Michael and I reflect on a key text about the nature of encyclopedias and of encyclopedic-ness called Encyclopedic Discourse by the literary scholar, Hilary Anne Clark. Hilary Ann Clark is a poet and literary scholar, who published a pathbreaking study of The Fictional Encyclopaedia in 1990. This article summarises the key theoretical findings of her study.

Ping us for a copy if you don’t have institutional access. 

In project news, we’re gearing up to our first (online) project symposium on June 8 and 9. The theme for wikihistories 2023 is “Wikipedia and its implications for memory (and forgetting)”. We have an excellent line-up of speakers showcasing top previous research like the very first research about Wikipedia in its role as documentor of historic events, as well as some of the exciting current approaches that are shaking up Wikipedia and the world beyond Wikipedia, not to mention some other thought pieces that perhaps should be shaking up Wikipedia a little more than they have.

We’ve also set up an incredible international Expert Advisory group made up of representatives from the academy, civil society and Wikipedia governance. All our members have significant expertise in the issues that our project will raise and the international network of social science and humanities scholars that we want to both connect to one another and to some of the excellent projects on the ground that would benefit from their expertise.

That’s it for now. If you’re interested in Wikipedia’s role in history-making, read on!

What makes a text encyclopedic?

Encyclopaedias can seem dull, boring, explicable. There was a time when they were literally part of the furniture. Hilary Anne Clark, ever the poet, tries to reveal the sensuousness and strangeness of this hoary literary genre.

The key to her argument is her shift in focus from ‘encyclopaedias’ to ‘encyclopaedic discourse’. The term ‘encyclopaedia’ is typically used to refer to large reference works such as L’encylopédie, Encyclopaedia Britannica or Wikipedia. Yet, as Clark observes, many texts that hardly resemble a work of reference are labelled ‘encyclopaedic’: Dante’s Commedia, Ezra Pound’s Cantos, or (I would add) Alexis Wright’s Carpentaria and The Swan Book. This raises Clark’s central question: what precisely do Britannica and the Commedia have in common?

Her answer is comprehensiveness. ‘Encyclopaedic discourse’ attempts to encompass the world. All discourse is selective, argues Clark, but encyclopaedic discourse is ‘special’, because

it selects from the entire domain of human knowledge, arranging selections according to specific orders—thematic and encyclopedic—that have developed historically, and representing its own discursive process in tropes such as the mirror, the tree, the labyrinth, the circle, and the network.

p. 98

The comprehensiveness of the encyclopaedic discourse opens a dialectic between inclusion and exclusion (p. 96). On the one hand, the encyclopaedia aspires to include all knowledge. On the other, it aspires to organise, define, to make accessible this knowledge (to ‘counter entropy’ as Clark puts it, p. 104). To organise knowledge, the encyclopaedia needs to define what knowledge is, and this will inevitably exclude certain kinds of discourse which will then clamour for entry into the encyclopaedia.

Perhaps the most counter-intuitive ramification of Clark’s theory is that there is really no such thing as ‘an encyclopaedia’. ‘Encyclopaedic discourse’ is not ordinary discourse, but rather a kind of ‘metadiscourse’ (p. 98). Any ordinary discourse can ‘become encyclopaedic’ when it begins to reflect upon its own totalizing operations (p. 107). Thus any ‘encyclopaedia’ is already something else. Wikipedia, for example, comprises encyclopaedic articles, just as Finnegan’s Wake is an encyclopaedic novel or Dante’s Commedia is an encyclopaedic poem.

As I see it, Clark’s theory can benefit Wikipedia Studies in two main ways.

First benefit: Clark’s theory can provide a viewpoint from which to critique Wikipedians’ use of the term ‘encylcopaedic’. Wikipedia’s own articles on encyclopaedias and encyclopaedism are remarkably narrow in focus. When Wikipedians complain that an article or revision is ‘unencyclopaedic’, the problem generally seems to be with the article’s ‘articleness’ rather than its ‘encyclopaedism’. Indeed, the infamous ‘Wikipedia Art’ article was arguably far more encyclopaedic in Clark’s sense than any typical article, and yet was rejected precisely for its ‘unencyclopaedic’ qualities (see Tkacz, Wikipedia and the Politics of Openness, chap. 2).

Second benefit: Clark’s theory can provide a richer comparative framework for scholars. Wikipedia scholars habitually compare Wikipedia to ‘traditional encyclopaedias’, by which they apparently mean L’encylopédie, Encyclopaedia Britannica and so on. Compared to such ‘traditional encyclopaedias’, Wikipedia can often seem highly innovative, for instance because of its frequent revisions, inclusion of ‘low-brow’ content, rejection of original research and so on. In the wider field of ‘encyclopaedic discourse’ identified by Clark, Wikipedia may seem more typical. Wikipedia scholars may find Clark’s analysis of Finnegan’s Wake especially interesting (pp. 105-7). Many of the points she makes about Joyce’s ‘recycling’ and ‘revision’ will sound eerily familiar to inhabitants of the Wikiverse.

Encyclopedicness is political 

It is so wonderful to read something about encyclopedias that I would never have come across had I not been working with such a great interdisciplinary team (it was Michael’s choice for a reading this week). Hilary Ann Clark has studied not only encyclopedias but fictional encyclopedias, trying to figure out what connects Dante’s Commedia to Encyclopaedia Britannica in their “encyclopedicness”. As Michael recounts above, her conclusion was that the primary feature of encyclopedias and encyclopedic fictional works is that they are comprehensive, and that “the comprehensiveness of the encyclopaedic discourse opens a dialectic between inclusion and exclusion”. 

The encyclopedia is limited by the need to organise knowledge and the process of
leaving things out is inevitable, Clark argues. This makes sense. In order to produce a summary article about “Tibet”, for example, we need to summarise everything that’s ever been written about Tibet, its geographic, history and politics. In doing so, we necessarily leave things out: we leave out detail, we merge themes according to our own perspective of what is most important, we select particular images among multiples. We produce a representation of Tibet that is shaped by the time in which we are living.

And yet, what is striking about Wikipedia to me is how its claims to completeness are the starting point for its politics, and that things are left out of Wikipedia, not only because of the necessity of summarising, but because certain types of representations fit particular ideologies and others sometimes readily oppose those ideologies. Wikipedia is limited, then, not only because it encircles knowledge, “categoriz(ing) while amassing, exclud(ing) while including” (p96) but because it is political. Clark’s view seems to suggest a kind of encyclopedic practice that is messy but disinterested. My view is that on Wikipedia there is no disinterested practice, even though there can be skilled and reflexive drawing together of sources. Wikipedia reflects the needs, desires, interests and knowledges of those who participate in its production. Wikipedia is peopled. It constitutes emotional labour. It is embodied. Its practice involves reducing or summarising in order to fit particular ideologies or narratives about what happened to whom and to what effect. If the encyclopedia is “all we know of the world, and we cannot know beyond it” (p93), who is the “we” here? And what is stopping “us” from knowing beyond it? 

In addition to the theme of comprehensiveness, Clark introduces another key feature of encyclopedias that I had never really thought about. “Any text (whether fictional or not) that we would call encyclopedic must speculate on its own discursive processes of discovery or arrangement, and on the limitations of these processes, given the fact of time and change” (p105), she writes. Encyclopedias, in other words, are reflexive. Although Wikipedia has long been criticised for reflecting a dominant worldview (that in turn reflects the interests of North America and Western Europe, white men, recent history), it has reflected on those skews from very early on in the history of the project. “WikiProject: Countering Systemic Bias” is a project whose history starts in 2004. Wikipedians and Wikipedia researchers have been thinking about what Wikipedia doesn’t, won’t and can never include for almost two decades. If Wikipedia is the biggest encyclopedia in history, it also reflects the largest documented attempt to reflect on the knowledge it produces. 

It is this reflexivity that I think is the most useful element of Clark’s work on encyclopedias. Just looking at the first version of this page reminds me of how much we have learned about what it takes to make a global encyclopedia. We may no longer imagine that bias can be overcome but we still agree that countering systemic bias is worthwhile and necessary.

References

Slashdot, July 28, 2004 “Wikipedia founder Jimmy Wales responds” https://slashdot.org/story/04/07/28/1351230/wikipedia-founder-jimmy-wales-responds 

In other news

Advisory Group Member, Dr Kirsten Thorpe with Nathan Sentance and Lauren Booker have recently released their anticipated report that considers how Wikipedia and Wikidata editors might describe Australian Aboriginal and Torres Strait Islander content in a more self-determined and culturally appropriate way. They write that “Wikimedia platforms offer significant visibility of referenced content to a vast global audience however First Nations content from Australia is under-represented and in some cases, culturally inappropriate or inaccurate. As an open knowledge system, it also enables usage of that content by other third-party platforms (for example, in Google search boxes), amplifying content across the internet, along with any inaccuracies, existing bias and knowledge gaps.” (p2) 

There are a lot of really important insights in the paper about the possible application of current Aboriginal and Torres Strait Islander protocols, for example, as well as recommendations for the production of style and terminology guides and mechanisms for a “right of reply”. One of the key priorities, according to the authors, might be to consider articles relating to Australia’s “Country [places], events and people” (p9). Wikipedia, they write, has an opportunity to highlight Aboriginal and Torres Strait Islander peoples, histories and cultures in geographic locations throughout Australia and to remedy the erasure of Indigenous histories in the representation of historical events. 

We at the wikihistories project think that complementary to these goals would be to understand a) the scope of articles in this set of Australian people, places and events and b) the current representation of Aboriginal and Torres Strait Islander in articles about Australian people, places and events. At the wikihistories project we’ll be playing a small role towards supporting this priority in the coming months.  

Contact us

Any news for us to share about Wikipedia and its role(s) in history making? Contact us!