Dérivepedia is a combinatory and recylopedic text generator that recombines sentence fragments from 400 Wikipedia entries to generate specious entries for subjects ranging from Tadpoles And The History Of Weather Satellites to Pliny The Elder: Constructing Ambiguous Witch Trials; from Jimi Hendrix And The Psychology Of Cowpox to Ada Lovelace In The Age Of Cool-Weather Aromatherapy.
The text and images are recombined on the fly.
A library of over 1000 variable in various arrays (people, cultures, isms, adjectives, foods, animals, things, etc.) is interwoven into the sentence fragments, image captions, and entry titles to complicate the context of the texts. That said, all of the variables and images are referable back to the sentence fragments appropriated from the Wikipedia entries.
You can find links to the original source Wikipedia entries on the reference page, along with a list of all of the variables included in the work.
With its constant cross-linking between diverse subject, Wikipedia is unlike a traditional encyclopedia, which relies on concise cross-referencing to provide primary and secondary information on any given subject. With Wikipedia, a user may chose to wander through its information structure, following links that are resistant to the development of a contextual plane of consistency. The Wikipedia entry for Etymology may provide a link to the entry for Beef, and from Beef there may be a link to Sewer Sludge, and from Sewer Sludge – Hormones… These abrupt changes in context are not unlike the phenomena Debord discusses in reference to Dérive. The hyperlinks within Wikipedia entries allow the user to drift through contexts, encountering different zones of information, different fields of study and areas of research.
Dérivepedia began by performing a sort of psychogeographical exploration of
information though the hyperlinks from one Wikipedia entry to another; appropriating a
sentence fragment from one Wikipedia entry, following a link, and continuing the sentence
with a fragment collected from this second entry. In the original experiments the linking
would continue until 19-21 entries had been touched.
The resuting text may read something like:
Birds such as some species of thrush have been known for centuries1 to improve social standing within a group2 australopithecine and modern humans, permitting the passage3 into proteins by unique synthetic mechanisms.4 Traditional nonfermented food uses of5 less expensive and more rapidly constructed6 intercollegiate wrestling began when,7 on the insistence of the French foreign minister, the8 city of Zara successfully rebelled against Venetian rule.9 The evidence of history and archaeology shows10 how exactly the plane of nutrition influences11 the use of specialized expression vectors and often necessitates12 research councils such as the National Institutes of13 Remembering, Understanding, Applying, Analyzing, Evaluating and Creating.14 In people who identify as alcoholics, there is15 a disposition to find the truth coupled with16 a novel or otherwise improbable act or17 locate objects by producing sounds and listening for the echoes.18 These cortical microcircuits are grouped into19 larger structures that can be seen as20 derivational or inflectional.21
Once these texts were collected, the sentences were abstracted, and variables were selected from them. The addition of variables to replace specifics within the sentences, along with the resequencing of the sentences in Dérivepedia allows context to shift further away from its source. Though the generated text may have the appearance of coherence it is in fact generated from 400 different Wikipedia entries, pulling in over 1000 variables, and mixing in nearly 200 public domain images from Wikimedia Commons. In this regard, it would be a mistake to call any given texts within Dérivepedia an entry at all; rather, they are the result of the drifts, ephemeral détourned paassages, void of any specific context.