Text Simplification

Text simplification is an emerging technology whose goal is to make text easier to read without significantly impacting the content. There are several aspects of text that are potential candidates for simplification, specifically lexical choice, syntactic simplification, and conceptual simplification. Current work focuses on syntactic simplification, which uses such heuristics as making embedded clauses or other modifiers into independent sentences. For example, instead of "The big dog that chased the white cat is sleeping", could be simplified to "The dog is big. The dog chased the cat. The cat is white. The dog is sleeping". Lexical choice simplification would substitute easier words for more difficult words in the original text. For example, "car" could be substituted for "automobile", or "gas pedal" for "accelerator". This relies on the availability of good lists of synonyms in the languages of interest, which may be hard to obtain. Conceptual simplification, or simplifying the concepts expressed in the text, is the most difficult task technically. It is also the type of simplification that's most likely to change the meaning of the text in undesirable ways.

One example of an online simplification demo can be seen at http://homepages.inf.ed.ac.uk/kwoodsen/demos/simplify.html

At this time, although the technology is potentially promising, it doesn't seem sufficiently advanced to rely on for routinely simplifying texts on the web in real time.