Motion Verbs and Their Modifiers (Prepositions and Adverbs)
By action verbs, we mean verbs that are used to specify actions that a human being, an animated agent (e.g., a game player) or a robot can perform. Examples of usage of action verbs can be seen in sentences such as the following.
- Open the door.
- Put the cup on the table.
- Turn the switch to position 6.
- Slowly roll the ball across the table.
I had performed some initial work on the representation of meanings of such verbs and their modifiers quite a while ago. Here are a couple of papers: AAAI paper1991 and Computational Intelligence paper 1997. Here is a recent paper by Zoost along the same line.
There has not been much work on the representation of such verbs and their modifiers, from a computational point of view.
- We are also interested in understanding the meanings of sentences that use such verbs and modifiers at a level that can be "executed" by a computational agent. In addition, with the movement to create computational lexicons such as verbnet and framenet, the work reported in these papers have become outdated. We would like to look at a large number of motion verbs and create lexical entries that fit their requirements.
- We are also interested in developing a lexicon like verbnet or framenet for prepositions (especially spatial ones) and adverbs that go with motion/action verbs.
- A project that I am very interested in is disambiguating the meanings of spatial prepositions. A simple spatial preposition such as on or in may have many different meanings as explained by Herskovits in Language and Spatial Cognition: An Interdisciplinary Study of the Prepositions in English. Is it possible to use a corpus and machine learning techniques to disambiguate the meanings?