System and method for dynamically determining the function of a lexical item based on discourse hierarchy structure

• US 7,363,213 B2
• Filed: 02/11/2003
• Issued: 04/22/2008
• Est. Priority Date: 02/11/2003
• Status: Expired due to Fees

• Current Assignee: Fuji Xerox Company Limited (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Xerox Company Limited (Fujifilm Holdings Corporation)
• Inventors: Zaenen, Annie, Polanyi, Livia
• Primary Examiner(s): Dorvil; Richemond
• Assistant Examiner(s): SAINT CYR, LEONARD
• Application Number:
  US10/364,191
• Publication Number:
  US 20040158453A1
• Time in Patent Office: 1,897 Days
• Field of Search: 704/8, 704/9
• US Class Current: 704/8
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique

Abstract

A method for adaptively determining the attitudinal function of a lexical item occurring in a natural language document, comprising determining a base valence of the lexical item; determining an actual valence of the lexical item by adjusting the base valence based on the discourse structure of the document; and wherein the actual valence denotes an attitude of an author towards an entity or event.

Claims

1. A method for adaptively determining the attitudinal function of a lexical item occurring in a natural language document, comprising:

determining a base valence of the lexical item;

identifying a contextual valence shifter (CVS) which can have a negative impact on the base valence if the consequences of a not born-out expectation are negative and can have a positive impact on the base valence if the consequences of the not born-out expectation are positive, wherein the lexical item is within scope of the CVS; and

determining an actual valence of the lexical item by adjusting the base valence based on the CVS.

2. The method of claim 1 further comprising determining:

at least one additional CVS, where the additional CVS is different from the CVS identified in claim 1, where the additional CVS is one of;

reported speech/thought, an elaboration, and a genre.

3. The method of claim 2 wherein:

the additional reported speech/thought CVS does not neutralize the base valence if the author adopts the attitude ascribed to the document entity.

4. The method of claim 2 wherein:

the additional elaboration CVS adjusts the base valence to reflect the valence of at least one elaborating lexical item.

5. The method of claim 4 wherein:

a contradiction between the lexical item and the at least one elaborating lexical item can indicate irony.

6. The method of claim 2 wherein:

the additional genre CVS adjust the base valence using knowledge of expected genre content.

7. The method of claim 6 wherein:

the genre can be a movie review.

8. The method of claim 1 wherein:

the base valence represents positive or negative attitudinal information encoded in the lexical item.

9. The method of claim 1 wherein:

determining the base valence is based on taking into account the domain in which the lexical item occurs.

10. The method of claim 1 wherein:

determining an actual valence of the lexical item by adjusting the base valence based on a plurality of contextual valence shifters.

11. The method of claim 1 wherein:

the lexical item can be a CVS.

12. The method of claim 1 wherein:

the actual valence denotes an attitude of an author towards an entity or event.

13. A method for adaptively determining the attitudinal function of a lexical item occurring in a natural language document, comprising:

determining a base valence of the lexical item;

determining an actual valence of the lexical item by adjusting the base valence based on the discourse structure of the document;

wherein the actual valence denotes an attitude of an author towards an entity or event;

wherein the discourse structure can have a negative impact on the base valence if the consequences of a not born-out expectation are negative and can have a positive impact on the base valence if the consequences of the not born-out expectation are positive.

14. The method of claim 13 further comprising determining:

at least one additional discourse structure, wherein the discourse structure is a contextual valence shifter (CVS), wherein the CVS is different from the discourse structure identified in claim 13; and

wherein the CVS is one of;

reported speechlthought, elaboration, and genre.

15. The method of claim 14 wherein:

the reported speechlthought CVS describes a context in which attitude is ascribed to a document entity rather than a document author and can neutralize the base valence.

16. The method of claim 15 wherein:

the reported speechlthought CVS does not neutralize the base valence if the author adopts the attitude ascribed to the document entity.

17. The method of claim 14 wherein:

the elaboration CVS adjusts the base valence to reflect the valence of at least one elaborating lexical item that elaborates on the lexical item.

18. The method of claim 17 wherein:

a contradiction between the lexical item and the at least one elaborating lexical item can indicate irony.

19. The method of claim 14 wherein:

the genre CVS adjust the base valence using knowledge of expected genre content.

20. The method of claim 19 wherein:

the genre can be a movie review.

21. The method of claim 13 wherein:

the base valence represents positive or negative attitudinal information encoded in the lexical item.

22. The method of claim 13 wherein:

determining the base valence is based on taking into account the domain in which the lexical item occurs.

23. The method of claim 13 wherein:

determining an actual valence of the lexical item by adjusting the base valence based on a plurality of contextual valence shifters.

24. The method of claim 13 wherein:

the lexical item can be a CVS.

25. A system for adaptively determining the attitudinal function of a lexical item occurring in a natural language document, comprising:

a base valence calculator to determine a base valence of the lexical item;

a contextual valence adjuster coupled to the base valence calculator, the contextual valence adjuster to determine an actual valence by adjusting the base valence based on a contextual valence shifter (CVS); and

wherein the CVS can have a negative impact on the base valence if the consequences of a not born-out expectation are negative and can have a positive impact on the base valence if the consequences of the not born-out expectation are positive.

26. The system of claim 25 wherein:

the base valence calculator utilizes a thesaurus; and

wherein the thesaurus provides the base valence of the lexical item.

27. The system of claim 26 wherein:

the thesaurus can use domain information to determine the base valence.

28. The system of claim 25 wherein:

the base valence calculator analyzes the natural language document to determine a domain of the lexical item.

29. The system of claim 25 further comprising determining:

at least one additional CVS, wherein the additional CVS is different from the CVS identified in claim 25; and

wherein the CVS is one of;

reported speechlthought, elaboration, and genre.

30. The system of claim 29 wherein:

the additional reported speechlthought CVS does not neutralize the base valence if the author adopts the attitude ascribed to the document entity.

31. The system of claim 29 wherein:

the additional elaboration CVS adjusts the base valence to reflect the valence of at least one elaborating lexical item.

32. The system of claim 31 wherein:

a contradiction between the lexical item and the at least one elaborating lexical item can indicate irony.

33. The system of claim 29 wherein:

the additional genre CVS adjust the base valence using knowledge of expected genre content.

34. The system of claim 33 wherein:

the genre can be a movie review.

35. The system of claim 25 wherein:

the base valence represents positive or negative attitudinal information encoded in the lexical item.

36. The system of claim 25 wherein:

determining the actual valence of the lexical item by adjusting the base valence based on a plurality of contextual valence shifters.

37. The system of claim 25 wherein:

the lexical item can be a CVS.

38. The system of claim 25 wherein:

the actual valence denotes an attitude of an author towards an entity or event.

39. A machine readable medium having instructions stored thereon that when executed by a processor cause a system to:

determine a base valence of a lexical item;

identify a contextual valence shifter (CVS) which can have a negative impact on the base valence if the consequences of a not born-out expectation are negative and can have a positive impact on the base valence if the consequences of the not born-out expectation are positive, wherein the lexical item is within scope of the CVS;

determine an actual valence of the lexical item by adjusting the base valence based on the CVS.

40. The machine readable medium of claim 39 the instructions further comprise identifying:

at least one additional CVS, wherein the additional CVS is different from the CVS identified in claim 39; and

wherein the CVS is one of;

reported speechlthought, elaboration, and genre.

41. The machine readable medium of claim 40 wherein:

the additional reported speechlthought CVS does not neutralize the base valence if the author adopts the attitude ascribed to the document entity.

42. The machine readable medium of claim 40 wherein:

the additional elaboration CVS adjusts the base valence to reflect the valence of at least one elaborating lexical item.

43. The machine readable medium of claim 42 wherein:

a contradiction between the lexical item and the at least one elaborating lexical item can indicate irony.

44. The machine readable medium of claim 40 wherein:

the additional genre CVS adjust the base valence using knowledge of expected genre content.

45. The machine readable medium of claim 44 wherein:

the genre can be a movie review.

46. The machine readable medium of claim 39 wherein:

the base valence represents positive or negative attitudinal information encoded in the lexical item.

47. The machine readable medium of claim 39 wherein:

determining the base valence is based on taking into account the domain in which the lexical item occurs.

48. The machine readable medium of claim 39 wherein:

determining an actual valence of the lexical item by adjusting the base valence based on a plurality of contextual valence shifters.

49. The machine readable medium of claim 39 wherein:

the lexical item can be a CVS.

50. The machine readable medium of claim 39 wherein:

the actual valence denotes an attitude of an author towards an entity or event.