Saturday, January 30, 2010

The Decline of Western Alphabetization

by Drew Martin

Alphabets are peculiar creations. They are not essential to language but are very common and some cultures have grown quite attached to them. The northern regions of North America were populated for at least twenty thousand years with people that did not even possess a written language and the world's most populous country, China, has a written language based on pictographs and does not use an alphabet to create words.

It is hard not to think of a first grade classroom without an alphabet running along a wall or little kids singing what we know as the ABC's. Even after one accepts the role the alphabet plays in language and grasps that letters are simply marks for consonants and vowels, it is still a bit disorienting to realize that the order of our alphabet seems random. A, B, C, D, E, F, G might as well be F, E, A, G, D. Oral language precedes written language, both historically and developmentally, per individual, so I thought that perhaps the order of the alphabet has something to do with a reflection of usage. This kind of makes sense when you consider the rank of XYZ but then RST should be before Q and O should be up front, although English inherited its alphabet from Latin, which does favor the Q. The order of the alphabet, like the QWERTY keyword, is most likely a previously determined and purposeful structure, which we simply hold on to.

I am sure the set arrangement of the seminal alphabets took awhile to shake out and the idea of alphabetization beyond the order of an alphabet did not actually catch on until fairly recently. There were instances in ancient Greek culture of subject alphabetization but its usage was not more common until the middle ages. As I recall from reading The Professor and the Madman: A Tale of Murder, Insanity, and the Making of the Oxford English Dictionary, the OED, created in the mid 1800s, was the first English dictionary to fully commit to alphabetization and previous English dictionaries were more relational. It sounds odd and dysfunctional but the reign of alphabetization has already peaked and is quickly declining because it is not always efficient and actually causes a lot of problems. Search engines are relational and the more information you have the more troublesome alphabetization is.

I recently joined facebook and was surprised to find that friends, by default, are sorted in alphabetical order by first name. That works fine and such a list is easy to browse when someone has 40 friends but the more friends someone has, the more cumbersome it gets. So facebook has various ways to sort, including thumbnails of everyone as well as by mutual relations, recent updates, college friends, work friends and by city. Postings are chronologically stored, ranking the most recent posts at the top of the wall. Google works quite sorts based on popularity or forced paid ranking. The only true way to sort is numerically.

Since the dawn of computers, letters moved from being an arrangement of strokes on a given medium to a numerical code. Even keywords are really just an arrangement of numbers, not very different than how a date is structured. This digitization of text changed everything and is patiently waiting for humans to think a little differently: more creatively. Currently we sort by alphabet, date, category, keyword and popularity and these can be quite frustrating. Alphabetical sorting is especially frustrating in some software packages, such as Adobe Illustrator, where I have to open up several panes of font lists to get to fonts "down" on the list...such as Trajan. If I need Trajan, I need it as quickly as I need Helvetica or Arial and there is no reason it should take longer. For that matter, I need to know what Trajan looks like otherwise I am shooting in the dark. I recall Corel Draw having a much better sort, which altered the text highlighted to take on the font characteristic of the font name you were scrolling over.

Keyword searching is even more frustrating when the results are ranked by popularity, as with Google. I was recently trying to find images from Gray's Anatomy, the classic medical illustration tome but all the results I got were pictures of cast from the television show Grey's Anatomy. Searching for images is an interesting situation because language is the limit. In text searches we want that limit. Let's say I want to learn more about women, so I would Google "women". My results should therefore be in the English language. If that is my native language and possibly the only language I know then that's perfect because there is no reason for me to attempt to read an untranslated essay in a language that is foreign.

With images, however, my return is severely limited by the word I use to search. If I search for women then I will only get a return of images with women in the file name or keyword...but then I am missing out on the uploads from all over the world in other languages for women. Image searches should be tied to a translation program so if I type in woman, I also get zeny, mujeres, damen...etc. What is also missing is a way to search for images without words...via images. This kind of happens now with thumbnail views but the driver behind that is a word search. At a local level I can bypass words and do an image search on my computer just by viewing thumbnail properties but then I am simply looking at thumbnails and not actually using images to search.

Since images in the world of computers are actually numbers and colors are simply R,G,B number combinations, it should be quite easy to search by colors directly from the file properties and bypass search words altogether. There could be a grid of color swatches in a search engine and by clicking on an orange swatch my return would show me images with large swatches of orange in them: a glass of orange juice, Tibetan monks, a beautiful sunset, etc. The programing behind this might simply read every pixel in an image and give me a return for any image that has more than 50% of its pixels in a certain orange range. The next step would be shape based searches...if I click on a triangle I get pictures of pyramids, dunce caps, roofs, tents, and yield and safety signs. Likewise, if I click on a circle my return would include images of the sun, a full moon, planets, a soccer ball, a clown nose and the Google logo.

As with the color search, the numeric code behind images would express that there are triangular or circular elements in an image, which a computer program would recognize, the same way a trained graphic artist knows the R,G,B code: 0,0,50 is navy blue. Since humans and animals are graphically just shapes, more advanced form searches could be created for identifying, for example, a woman riding a horse that would not require those words or related words, such as "equestrian". What I am suggesting is indeed distancing us further from the "literal turn" but it does not abolish language. If anything, it embraces a more universal experience of communication and does not favor a particular language and the culture tied to it.

Beard alphabet by TimYarzhombeckLink