( I will be posting my thoughts on Words and Multimedia Search. There will be a few posts on this topic. Here is the first one.)
Words are important. Words have been important in human society for the longest time imaginable. In fact, it is impossible to think of the history of the World without the notion of words. We use words to express our experiences, to represent knowledge, to communicate, and to impress and influence people. The power of words has been well established in every civilization that human society has seen. We are taught from early childhood to use appropriate words for different contexts. The quest for increasing effectiveness of the words used in our expression never ends.
Many different societies and civilizations evolved in different parts of the world. Independent of the multitude of languages used in the world, however, words are the basis of human communication and knowledge.
Words were always important, but in the last decade suddenly their status has become even more exalted. Search engines have made them more important – now they have taken on even more important role in our life. Routinely we go to search engines to get information and our primary means of communication with these important powerful know-it-all sources of the World Knowledge is a set of keywords. We find objects, people, places, concepts, knowledge sources, and everything else by thinking of appropriate set of keywords to describe our needs. Suddenly grammar is not important, only a set of keywords does the trick. A bag of words is important – the sequence and other organization are not. We were taught to express our thoughts using appropriate words in a proper sentence to precisely represent semantics. Search engines are now telling us to forget the language and the notion of sentence – just consider a bag of words.
In the historical development of languages, as well as in development of linguistic skills by each human being, initially words and language are in the oral form. It is after much maturity that this oral form becomes the text form. Text is nothing but an abstract persistent representation of words that initially evolve in oral form. The words recognized by search engines do not consider the existence of oral words – they only consider textual words. That is natural because search engines only deal with language in textual form, not in oral form. Real words in oral form have emotions and implied semantics associated with them – the textual words can only have explicit semantics associated with them.
Can textual words, in fact, have emotions? A long time ago during a linguistics course at Georgia Tech we were studying the prospects of AI and its possible interpretations of words. As the internet has become ubiquitous and a legion of uninhibited users (meaning everyone younger than me) has filed into this space, I contend that new textual words have arrived complete with emotion embedded within.
Could we not view the constructed emoticons as a new type of word? If I post ;-), is the message as clear as if I did wink at you? What about hybrid words, loosely described as 733T (leet) speak? Most people familiar with the net understand what lol means.
Just testing –for some reason, my earlier postings were not accepted– if this goes through, i would like to participate in this debate on BoW
= Jay
The idea of emoticons is interesting, Citizen Deux.
In fact, you are just a bit ahead of me in terms of my postings.
I will soon talk about textual words being a subset of words and language being an evolving language — and the points that you make are directly related to that.
Jay, it will be great to hear your thoughts.