2009 : WHAT WILL CHANGE EVERYTHING? [1]

marti_hearst's picture [5]
Computer Scientist, UC Berkeley, School of Information; Author, Search User Interfaces
THE DECLINE OF TEXT

As an academic I am of course loathe to think about a world without reading and writing, but with the rapidly increasing ease of recording and distributing video, and its enormous popularity, I think it is only a matter of time before text and the written word become relegated to specialists (such as lawyers) and hobbyists.

Movies have already replaced books as cultural touchstones in the U.S. And most Americans dislike watching movies with subtitles. I assume that given a choice, the majority of Americans would prefer a video-dominant world to a text-dominant one. (Writing as a technologist, I don't feel I can speak for other cultures.) A recent report by Pew Research included a quote from a media executive who said that emails containing podcasts were opened 20% more often than standard marketing email. And I was intrigued by the use of YouTube questions in the U.S. presidential debates. Most of the citizen-submitted videos that were selected by the moderators consisted simply of people pointing the camera at themselves and speaking their question out loud, with a backdrop consisting of a wall in a room of their home. There were no visual flourishes; the video did not add much beyond what a questioner in a live audience would have conveyed. Video is becoming a mundane way to communicate.

Note that I am not predicting the decline of erudition, in the tradition of Allan Bloom. Nor am I arguing that video will make us stupid, as in Niel Postman's landmark "Amusing Ourselves to Death." The situation is different today. In Postman's time, the dominant form of video communication was television, which allowed only for one-way, broadcast-style interaction. We should expect different consequences when everyone uses video for multi-way communication. What I am espousing is that the forms of communication that will do the cultural "heavy lifting" will be audio and video, rather than text.

How will this come about? As a first step, I think there will be a dramatic reduction in typing; input of textual information will move towards audio dictation. (There is a problem of how to avoid disturbing officemates or exposing seat-mates on public transportation to private information; perhaps some sound-canceling technology will be developed to solve this problem.) This will succeed in the future where it has failed in the past because of future improvements in speech recognition technology and ease-of-use improvements in editing, storage, and retrieval of spoken words.

There already is robust technology for watching and listening to video at a faster speed than recorded, without undue auditory distortion (Microsoft has an excellent in-house system for this). And as noted above, technology for recording, editing, posting, and storing video has become ubiquitous and easy to use. As for the use of textual media to respond to criticisms and to cite other work, we already see "video responses" as a heavily used feature on YouTube. One can imagine how technology and norms will develop to further enrich this kind of interaction.

The missing piece in technology today is an effective way to search for video content. Automated image analysis is still an unsolved problem, but there may well be a breakthrough on the horizon. Most algorithms of this kind are developed by "training", that is, by exposing them to large numbers of examples. The algorithms, if fed enough data, can learn to recognize patterns which can be applied to recognize objects in videos the algorithm hasn't yet seen. This kind of technology is behind many of the innovations we see in web search engines, such as accurate spell checking and improvements in automated language translation. Not yet available are huge collections of labeled image and video data, where words have been linked to objects within the images, but there are efforts afoot to harness the willing crowds of online volunteers to gather such information.

What about developing versus developed nations? There is of course an enormous literacy problem in developing nations. Researchers are experimenting with cleverly designed tools such as the Literacy Bridge Talking Book project which uses a low-cost audio device to help teach reading skills. But perhaps just as developing nations "leap-frogged" developed ones by skipping land-line telephones to go straight to cell phones, the same may happen with skipping written literacy and moving directly to screen literacy.

I am not saying text will disappear entirely; one counter-trend is the replacement of orality with text in certain forms of communication. For short messages, texting is efficient and unobtrusive. And there is the question of how official government proclamations will be recorded. Perhaps there will be a requirement for transliteration into written text as a provision of the Americans with Disabilities Act, for the hearing-impaired (although we can hope in the future for increasingly advanced technology to reverse such conditions). But I do think the importance of written words will decline dramatically both in culture and in how the world works. In a few years, will I be submitting my response to the Edge question as a podcast?