Tuesday, January 1, 2008

Natural Language Processing

Sometime long ago, the grunts of some species of ape began to take on a new character, enabling the animals to communicate in ways more powerful than any species had before. This surely took place gradually, and surely changed the apes as they changed this new thing they had, which we now call language. It was born around campfires and on hunts, in reference to children, prey, crops, predators, friends, enemies, and lovers. Over a span of hundreds of thousands of years, it grew up on six continents, in many different forms. It existed in two forms: In the mind, and as a spoken medium, formed in the mouth of a speaker, carried through the air as sound, and received in the eardrum of a listener.

Just a few thousand years ago, and in particular locations, humans began to write language. Just a few decades ago, humans began to encode written language so it could be stored and manipulated by digital computers. It didn't take long for people to realize that the kinds of tricks we can all perform, almost effortlessly, in crafting and understanding language are magnificently difficult to engineer in a machine.

Workers in Natural Language Processing (which also goes by the alias Computational Linguistics, among other names) have grappled with the difficulty of this subject matter. Theoretical and intellectual curiosity has driven research, as have governmental and commercial enterprises. The history of the field has been marked by optimism, setbacks, hype, successes, more optimism, more setbacks, more hype, and the occasional white lie.

The Internet, wireless communication, and portable computing are going to stimulate more interest in and need for NLP. If the past is any guide, there'll be a steady stream of optimism, setbacks, hype, and hopefully some successes. This blog will track NLP's ups and downs through 2008 and in the years to come.

No comments: