I’m going to talk about three different examples of useless classification today, but I guess in order to do that, I need to say what would constitute useful classification, so I can show how these examples don’t fit that mold. That, of course, is the main question of today’s readings. For my purposes, I’ve been using the term ‘useful’ to express any system that allows us to look for and hopefully find the information we are seeking, whether it be through specific or general searching, or simply browsing.
Each of my examples has a different relationship with usefulness. The first is useless because it organizes based on criteria that one would never use to search. The second because it invokes too many criteria and conflates what is sought after with what is found. The last is not useless, but is used in a manner that often defies use on its original ideas, though certain of its principles remain the same.
Some Pros and Cons of Organizing Your Bookshelves by Color
The first example defies use as a classification system by sorting based on principles that could never assist anyone in seeking and finding the information they were looking for. Organizing books by color is functionally useless because we cannot use that as an identifying feature when looking for a specific book. The organizing principle is based on criteria that are unsearchable. It would take another index, another layer of metadata, to make sure we knew what book – by criteria we find useful to have about a book, such as author or title – was what color, and only then could we make use of the color-sorted books. The author of this essay admits that you will never find your books again, and suggests that there is also the need to “articulate this particular shame every time a new guest appears in your apartment.” And on top of this, if you insist on keeping this system of classification, you will need to spend an incredible amount of effort maintaining it, because adding new books will require the entire system to be rearranged.
But there are several ‘pros’ offered for this bizarre system. Aesthetic value is one – the arrangement of books in this way is pretty. But in order to reap this aesthetic reward, you need to sacrifice functionality. Or at least, you sacrifice one functionality that we think is necessary for shelves of books: the ability to find them. But in classifying books this way, we are just making it apparent that we are willing to give up this function if we want to make art a priority. A simultaneous loss and gain of control is another pro. The author suggests you can surrender book organizing decisions to the “tyranny” of this system. At the same time, in another point, she suggests you can spend hours making “banal” yet “delightful” decisions about how or whether to split up the brightest oranges, or where beige, white, and black will fall on the spectrum.
But another, more telling pro from the author is listed as number three:
3. You will learn important life lessons. For example, a book that is both red and blue may be placed successfully in either the red or blue section, and look as if ’twas ever thus.
It was not ever thus. All the banal and delightful decisions you made earlier are what contributed to the system, but by committing to its tyranny, it comes to appear natural, as though it could never be any other way. Just as Perec decries the ‘logical sequence’ of the Exposition, and Weinberger argues with Plato’s logic of ‘separating the joints,’ this example serves as another instance where ‘natural’ classification is exposed as fictional and constructed by human biases or preferences.
Even with the infallibility of classification reduced to a fiction, it’s worth noting that the books in the photo are arranged by a hierarchy – Roy G Biv. Sticking to this established pattern, which has resonance when talking about light but no real reason to exist in this context, makes this instance not unlike Weinberger’s planetary example, wherein organizing principles were created to include an existing group into a newly formed category, not for any scientific but to give them a reason to remain grouped. It is also comparable to Perec’s questioning the logic of the alphabet. There is no real reason for the letters to be in their order, but the simple fact that they are privileges some letters over others and encourages other decisions to be made surrounding that false hierarchy (company names changed to fall higher in alphabetically organized directories). If the order of the books is already unusable by our practical standards, what reason is there to impose another hierarchy onto them, if it has nothing practical to add to the system’s validity?
23 Oddly Specific Netflix Categories That Only Have One Show You Can Watch
This next example is the most Borges-like (Borgesian?), and comes courtesy of Buzzfeed. A while back an intrepid individual counted the number of possible Netflix categories. There are over 70,000 of them. Many films and shows appear in more than one, but the opposite is also true: some categories contain only a single suggestion.
So how is this useful? Are we likely to ever browse so specifically (and with Netflix’s code words to guide us) that we will find these categories? Will we come through a Cartesian classification system, first selecting out the simplest and most easily knowable preferences, then slowing parsing out the most complex until they are layered and leave only one obvious option? Are we at the point yet where Netflix can know use well enough to suggest these categories? If we boil our way down to these specific categories, does this mean that Netflix truly knows what we want to watch, since it must fit this category and there is only a single choice?
A few theoretical concepts apply to this example. It seems as though Netflix does not know what Perec knew, which is that “there is a difference between classification and its subsequent use” (Boyne).
The categories are not unlike Borges quoted in Perec: “a) belonging to the Emperor, b) embalmed, c) domesticated, d) sucking pigs…k) drawn with a very fine camel-hair brush, l) etcetera, m) which have just broken the pitcher, n) which look from a distance like flies.” Which is to say, they are descriptive, but not overtly useful. If you don’t know anything about Orange is the New Black, “raunchy TV comedy featuring a strong female lead” is not an unhelpful descriptor. But you can’t use this category very effectively to look for something to watch, because it’s buried so many levels deep. When you’re browsing Netflix, you’re searching through keywords because you don’t know exactly what you want. If you were able to search for the exact category, since it only represents one item, you may as well look for the item itself. You could probably find Orange is the New Black in some of the higher level categories, like “raunchy TV comedy” or “TV featuring a strong female lead,” but at the point that you merge those two categories to leave only one show left, you’ve effectively merged the category with the thing itself.
The Netflix categories also toy with the idea present in Perec that “thinking is not in control of what is thought,” as Boyne expresses it. Do we think and then classify, or classify and then think? This is brought to the fore when you consider the creation of categories like “Gory Canadian revenge movies.” Is this a category that existed before something existed to fill it? If all Netflix categories are just amalgamations of certain key words with increasing specificity, then it’s possible that yes, the classification existed before the content, and perhaps it let us know that a very niche market existed to be filled. But it could also be context dependent – the presence of a film prompts the creation of a hyper-specific category, refining the system but simultaneously making it obsolete.
I’m not suggesting that these films would not exist without these categories, but that the 70,000+ categories of Netflix films begin to defy practical functionality. They make us ask whether there is use in a category of one. Perhaps with an item level finding aid, the answer might be yes. But in a library of films, perhaps this is more like the infinite aphorisms Perec discusses – a mere list of possible outcomes of permutations – than a system of classification.
In my early ‘teaser,’ I referenced Borges, because his stories explore hypothetical taxonomies or systems in need of classification, and I think two of his stories are most applicable in to this example. In The Library, the titular structure contains nearly infinite information (so, practically infinite) and the search for a guide is both paramount and also ridiculous, because what would a guide for infinite information look like? Just another layer of infinity that would render itself useless. And the other story is that of the infamous map referenced in Baudrillard, where the scale is a mile to a mile. With a map detailed enough that it becomes so big as to blanket the entire empire, the use disappears and the function collapses.
Hashtags
Lastly, I want to talk about hashtags, because I love them. David Weinberger discusses the phenomenon of tagging as user-directed, as part of our acceptance of growing miscellany which has developed ways to make what’s important to us available to all and a possible, valid system of classification. It’s a very populist, participatory version of the faceted classification he demonstrates on a library website, where you direct the information hierarchy based on categories that matter you you.
Weinberger’s talk is from a few years ago, so it’s quite possibly pre-hashtag explosion. But a hashtag is, like the examples he uses such as Digg, anything that lets users categorize for themselves or publicly for others (or both). The primary function of a hashtag is to collect/aggregate content, most often tweets, but now Facebook or Tumblr posts. They are clickable and searchable, so you can view content in categories that interest you. And they are hierarchical – some trend, some don’t – and can be standardized – some events/organizations/institutions have official hashtags. So they bring together alike content and sift out irrelevant content. This can be useful for conversations, or for tracking, if you’re using hashtags in your social media marketing, for instance.
But what about utterly useless hashtags? If the point of a hashtag is to allow us to seek out like, why does the practice prevail of using hashtags in ways that defy this purpose? Like the Netflix example, what use is a category of one? Or a category that no one would ever look up because it’s of no use as a collective?
An extremely anti-hashtag article in Gizmodo titled “How Hashtags are Ruining the English Language” states
Hashtags at their best stand in as what linguists call “paralanguage,” like shoulder shrugs and intonations. That’s fine. But at their most annoying, the colloquial hashtag has burst out of its use as a sorting tool and become a linguistic tumor—a tic more irritating than any banal link or lazy image meme. The hashtag is conceptually out of bounds, being used by computer conformists without rules, sense, or intelligence, a like yknowwwww that now permeates the internet outside of the tweets it was meant to corral. It pervades Facebook, texting, Foursquare—turning into a form of “ironic metadata,” as linguist Ben Zimmer of The Visual Thesaurus labels it.
In other words, hashtags are out of control and moving out of place. The tagging metadata is moving beyond a useful place, where it would sort tweets into categories that can help us find information or actually be searched for in the first place. But Susan Orlean of the New Yorker explains that the hashtag is undergoing a “mission creep.” In 2010, she wrote
…you could tweet,
“I would rather have a moose for President than Sarah Palin!”
Or, making good use of a hashtag,
“I would rather have a moose for President! #SarahPalin”.
The second example imposes a category on the tweet that is useful. Sarah Palin is a searchable hashtag that someone might conceivably look at or look for. But the “mission creep” is that paralanguage that the Gizmodo article references – a s(n)ide commentary on the tweet:
A typical commentary-type hashtag might look like this:
“Sarah Palin for President??!? #Iwouldratherhaveamoose”
This usage totally subverts the original purpose of the hashtag, since the likelihood of anyone searching the term “Iwouldratherhaveamoose” is next to zero. But that isn’t the point. This particular hashtaggery is weirdly amusing, because, for some reason, starting any phrase with a hashtag makes it look like it’s being muttered into a handkerchief; when you read it you feel like you’ve had an intimate moment in which the writer leaned over and whispered “I would rather have a moose!” in your ear.
Orlean ends her piece by remarking on the richness that can be imparted into a 140 character tweet with the use of a hashtag as this “paralanguage.” If we take this meaning instead of the original purpose of a hashtag, these tags are not “useless” classifications, just useless in the classification of the tweet’s content. The use the provide has changed, as they instead classify the tweet itself as a joke or an aside, or they openly identify the tweet’s subtext or intended meaning while keeping the text itself intact (another of Orlean’s examples: “I just made out with your husband! #kidding #hewishes #likeIwouldadmititanyway”). Functionally, we are typifying ourself and our speech, claiming ownership as we first obscure and then clear up our intended meaning.
Users of the hashtag in this way are coding a number of layers onto some simple, short text, but it’s not its intended use. Maybe that’s why it makes people so upset (an Urban Dictionary user defined the hashtag as “Yet another social network fad that teaches young Americans to not write in full, understandable, proper sentences.”). But then again, its intended use was a measure of taking control away from those who sought to classify for us and give us the freedom to classify based on our own preferences and designs. And in both cases, it’s used to add new layers of complexity and nuance to existing systems and communication. Both uses stem from a desire to be understood and to understand, and the knowledge that there’s nothing inherent or essential about knowledge. Nor is any attempt to organize and make meaning clear ever truly complete. To summarize with a quote from Perec:
The sheer number of things needing to be arranged and the near-impossibility of distributing them according to any truly satisfactory criteria mean that I never finally manage it, that the arrangements I end up with are temporary and vague, and hardly any more effective than the original anarchy…
In short, I muddle along.