The Hijack (Proper)

Below are all the unedited notes I kept during this project. I meant to turn them into an essay, but instead I'm just dumping them here. Xanadu.


 

Unlike FAT Labs / Aram Bartholl speed projects and speed shows, it's a slow project. It took a long time to actually implement, and it will likely take a long time to disperse, if it even disperses at all. It's not to get it to blow up, but to to track the ways in which it disperses or fails to disperse.

Brute force attack (so to speak), not an algorithmic attack (algorithms vs. algorithms) -- because google algorithms will just route around algorithms meant to game it, and I don't have the ongoing software development budget for that. But google algorithms are actually supposed to value human brute force activity and attention. So this is a much more robust, sustainable, and legacy-minded (albeit herculean and quixotic) mechanism of dispersal. Plus I am less interested in merely hijacking Google Images than I am in hijacking natural human language use of the word Xanadu and having Google Images reflect the results. Google and all of these social media companies can go out of business, but the logo/sigil will still live on and continue dispersing.

Difference between a logo and a sigil. It's more like a sigil.

Everything about google image search is geared toward the idea of a user wanting to see a discrete image for whatever personal reason. And their instructions for developers are geared toward corporate marketing. But since they champion the user experience over the corporate marketing, we are not spamming or stuffing, since we are not selling anything. Indeed, we are more interested in the user experience than someone trying to sell the user something.

But the hijack consists of no single image. The logo will wind up being identified by google as the "object" of the hijack, the discrete digital image, but the hijack is actually an event that seeks to subvert the inherent dependence that object-selling corporations have on discrete images.

The google images system doesn't filter out watermarks. recognizing them as a legitimate means within capitalism of protecting one's ownership of an image. And there is no single "subject" (signified) meant to be associated with the signifier "Xanadu". So a sophisticated image recognition algorithm can tell that the textual signifiers "anarchy.jpg" or "alt="power to the people" or a caption that reads "stick it to the man" are incongruously associated with a jpg image of a cat. But any image may be associated with any proper noun, particularly with one as mythical as Xanadu. Indeed, the histories of the reappropriations of this noun reveal multiple uses. A major theme they have in common is they are overly ambitious and ultimately fail. The Xanadu Hijack is all those things too. So, according to Google Images filtering and ranking system, the Xanadu Hijack is a perfectly legitimate use of the network.

Nathaniel Stern and Scott Kildall's Wikipedia Art is related. Initially, it didn't meet Wikipedia's criteria for a legitimate entry. But once enough press was written about the project itself, suddenly it became a legitimate event. The Xanadu Hijack is similar, but more gradual. The more "actual" mindshare (and press, and repostings) the project gets, the more the images will begin to appear at Google Images. This won't be gaming the google algorithm, it will be "gaming" the minds of people on the internet. The Google Images algorithm will only be accurately reflecting a topical cultural interest that has already occurred. In other words, the project will just accomplish what other art press releases and marketing campaigns accomplish all day long. But there is no product (painting, tennis shoe). There is only the event. The logo brands the rebranding event itself. "All advertising advertises advertising" (Marshall McLuhan, 1963). In this case, literally. I know, rite!

The story about Howard Finster's tools for his house (used his hands to build his tools, used his tools to build his house, used his tools to build a box, placed his tools inside the box, used his hands to bury the box under the house, went inside the house). the idea of placing a word under erasure. The X of the logo. Camouflage via data saturation. The only way to return the word back to itself is to dilute the potency of every particular co-option of the word. But then won't the hijack wind up having co-opted the entire word? Perhaps, so then it will be your duty to hijack he hijack. Perpetual deferral. Enacted deconstruction.

--------

Google policy punishes sites that try to game their algorithms without adding any "new" and "valuable" "content" to the net (and by extension, to the world). But what about a site explaining clearly how one might game Google algorithms (without itself trying to game those algorithms)? Such a site is indeed adding new and valuable content to the net, but content which implicates google as being part of the net itself (and by extension part of the world itself). The only way, then, for a company (like Google) to assume a detached, meta-perspective on the world (wide web) is to implicitly deny the problematic but obvious fact that said company is itself a part of the word (not apart from the world). Such denial de facto leads to the implicit policing and punishing of any site that would expose google as just one more part of the net's "content," rather than a special meta-entity outside of the net's "content." Thus a site that tells people how to game google's ranking algorithms would be treated similarly to a site attempting to game google's ranking algorithm.

Furthermore, how does google treat a site meant to incite an action or an event rather than a site merely meant to provide "content." The Xanadu Hijack is such a site, although in order to incite its action it must necessarily provide some content. But the "content" is just kindling for the fire. The first of Google's "Quality guidelines - basic principles" is, "Make pages primarily for users, not for search engines." But the goal of the Xanadu Hijack site is to enable users to modulate search engines.

J.L. Austin would distinguish between the action-inciting site and the content-providing site. He would call the action-inciting site "illocutionary" (uttered commands that perform what they command, like "I now pronounce you man and wife."); and the content-providing site "declarative" (an utterance that merely declares an opinion or fact, like "the sky is blue.") Derrida would problematize this distinction, noting that even "mere" declarations are always uttered with at least some performative intention in mind (I declare the sky blue because I want to be sociable or appear erudite or sensible or whatever). So even a hobbyist blog about waterfalls is never merely declarative (as evidenced by its ability to sell in-line banner ads to travel companies once the blog gains enough traffic). Every piece of "uttered" "content" on the web is always already performative, or it wouldn't have been posted.

Like Google, linguistic sign systems are also part of the world they describe. And, like Google, language itself is forever at pains to defend the myth of its detachment from the world it purports to "merely" describe. So, for instance, language is infamous for its inability to talk about nothing. Language presumes and preferences presence and objecthood. The statement, "this is nothing," must first make absence a present thing, then negate this thing (no-thing), then equate its existence (via "is") to another present thing ("this"). Such convolutions (resident in such a simple, common statement) reveal the wildly ingrained prejudices of language. Such is the tyranny of language.

But the Xanadu Hijack is not merely content, nor is it merely a declaration of fact. It is an event waiting to be enacted, an event booby-trapped to foreground the problematic relationships between capital, objects, images, algorithms, search engine corporations, and language itself. (Plus you get to reblog a bunch of funky picture of Olivia Newton John.)

----

The fact that Google's algorithms are currently sophisticated enough to distinguish between bot retweets and human retweets is great. Because then it is not just a battle of software vs. software. In order to hijack Google Images search results for Xanadu, we actually have to hijack the way a bunch of humans think about the word Xanadu. And then the software will report this change in mindshare.

////////

"some people add copyright text, watermarks, or other information to their images. This kind of information won't impact your image's performance in search results, and does help photographers claim credit for their work and deter unknown usage. However, if a feature such as watermarking reduces the user-perceived quality of your image or your image's thumbnail, users may click it less often in search results."
- from https://support.google.com/webmasters/answer/114016

The difference between this project and branding is that we are not branding anything other than the process of branding. We are (re-)branding the word to nothing, to return the ownership of the word to the word itself. By making Xanadu a common noun rather than a proper noun, it becomes its own word again. Once a noun becomes a proper noun, it is owned by a particular immanent instantiation in the historical world. But lowercase "water" is just water.

//////////////

"Critically, the semantics here are in the users, not in the system. This is not a way to get computers to understand things. When del.icio.us is recommending tags to me, the system is not saying, 'I know that OSX is an operating system. Therefore, I can use predicate logic to come up with recommendations -- users run software, software runs on operating systems, OSX is a type of operating system -- and then say, Here Mr. User, you may like these links.'

What it's doing instead is a lot simpler: 'A lot of users tagging things foobar are also tagging them frobnitz. I'll tell the user foobar and frobnitz are related.' It's up to the user to decide whether or not that recommendation is useful -- del.icio.us has no idea what the tags mean. The tag overlap is in the system, but the tag semantics are in the users. This is not a way to inject linguistic meaning into the machine."
- Clay Shirky

(related to the idea of not using software to fool software, but using humans to actually gain cultural mindshare. The google image collage results will indicate something that has already happened, and not "falsely," algorithmically, or whatever. But it's happened via humans.)

not a name associated with an object, but a name associated with the event of appropriating the name itself. So it returns the name to itself, free of the historical duties of having to name things. But of course this can never happen. All that will happen is that this failed hijack project (the desire to give the name back to itself) will be just one more additional historical thing that the name will be burdened with pointing to.

Still, the project is not entirely conceptual. We really mean for people to attempt to hijack the image associated with the name. It can't actually fail without first attempting to succeed, or it will have succeeded right off the bat (and thus will have failed to actually fail).

----

Include the Thom Yorke image dispersion.

-----

Mention pantry war -- over time the inlined image that remains wins. Net project over a long period of time. Slow net art.

-----

discuss the "queries" document [pdf] and the idea that Google analytics allows you to collaboratively write a poem via a honeypot trap and watching footprints leading to the honeypot, but the footprints are shaped in the form of the honeypot. Because we are all using words. So I don't write the queries, but the queries combined with algorithms and databases lead to things I have posted. A kind of oblique collaborative distributed durative poetry.

The Xanandu Hijack will be that, but with a collage. So to add to the collage, tag and include the watermark/sigil.

---

The thing where it got press in the new york times, but that was kind of an accident, but when you put things out onto the network they can begin to connect. So try to lure those connections, to seed them. Like a "reverse big house" in Oceans 13 -- you don't have to win, as long as the house loses. The project doesn't have to achieve a particular result, as long as something happens. And even if nothing happens, the fact that it is Xanadu (an epic fail) means that something has happened.

Talk about etoy's 1996 hijack. But that could not happen today. So then, a minor viral phenomenon, akin to a minor literature. Which may just be like people sharing things with each other.

Mention Ulmer/Derrida's idea of feeding things forward, speculative etymologies, applied grammatology. It happens anyway, so to try to put some intentional inflection on it, for reasons other than making money.

-----------------------------------

Bio

Curt Cloninger is an artist, writer, and Associate Professor of New Media at the University of North Carolina Asheville. His art undermines language as a system of meaning in order to reveal it as an embodied force in the world. By layering, restructuring, hashing, eroding, exhausting, and (dis)splaying language, he causes language to perform itself until its "meaning" has less to do with what it denotes and more to do with how it behaves. His work has been featured in the New York Times and at festivals and galleries from Korea to Brazil. Exhibition venues include Centre Georges Pompidou (Paris), Granoff Center for The Creative Arts (Brown University), Digital Art Museum [DAM] (Berlin), Ukrainian Institute of Modern Art (Chicago), Black Mountain College Museum + Arts Center (Asheville), and the internet.

Curt has been published on a wide range of topics, including new media and internet art, installation and performance art, experimental graphic design, popular music, network culture, and continental philosophy. Recent topics have included glitch art, the "new aesthetic," electronic voice phenomenon, bodily affect, object oriented ontology, process philosophy, and artistic lying. His articles have appeared in Intelligent Agent, Mute, Paste, Tekka, Rhizome Digest, A List Apart, and on ABC World News. He is the author of four books, most recently a collection of his essays spanning 14-years (aptly) titled One Per Year.

He maintains lab404.com, playdamage.org, and deepyoung.org in hopes of facilitating a more lively remote dialogue with the Sundry Contagions of Wonder.