Gradience, as a word and as a concept, is quite central and of a key interest within my PhD work in Computational Linguistics about Robustness and Constraint Language Processing. But is the word ``gradience'' by itself a correct, licensed english word?
I often encounter people who challenge it--fellow research students and senior researchers, referees, non-native as well as native english speakers, linguists, computational linguists among others. They mostly put forward two arguments:
They seem to be right on that I couldn't find so far any english dictionnary, which would have an entry for gradience. And strictly speaking, from a prescriptive point of view one could probably say that it is indeed not an english word. From a descriptive point of view, however, as we will see gradience is clearly an instance of the property of lexical openness, inherent to natural language, as described by Pullum ([Pullum2001]), and as such is perfectly licensed by the english language.
From a formal point of view, and also because the notion is so central to my research work, I couldn't agree more on the necessity to ``properly'' define the term and the notion it denotes. On the use of the term in the literature, however, I came to the conclusion that not only the word has been used for several decades (apparently since Bolinger in 1961) with a very well-defined semantic associated to it, but it has also been used quite consistently by a large variety of authors who otherwise are not directly connected to each other. We will in the next section what the term means in linguistics.
Historically, the present document started as an informal list of unclassified web references making use of the term gradience, which aim was to serve as a supporting evidence against the two arguments mentionned above in rather informal discussions and was sent across internet via email. Then, as the arguments turned out to be reccurrent, it seemed like a good idea to make this compilation of references publicly available, and to use it as a starting point for a web-based working document on my PhD topic. The very first list contained only links to html documents--no .ps, .pdf or other--, which is a strong limitation. In a near future this list should be completed with references in the more traditional academic sense of it, whether an html document can be associated with them or not.
Thus, the very first goal of this document is to gather evidences on the use of the term gradience. In particular, we will put some effort into addressing the two aspects raised above.
A second and more important goal is clearly to dissert about the notion of gradience in linguistics and also more precisely about how to model it within computational frameworks.
A third goal is to rely on the internet community to help improving this document and making it as accurate, up-to-date and exhaustive as possible by getting some feedback.
Please send comments or queries about this web site
to jpprost@ics.mq.edu.au
|
|