Sentence analyzer

• US 4,864,502 A
• Filed: 10/07/1987
• Issued: 09/05/1989
• Est. Priority Date: 10/07/1987
• Status: Expired due to Term

• Current Assignee: Vantage Technology Holdings LLC
• Original Assignee: Houghton Mifflin Harcourt Publishing Company (Houghton Mifflin Harcourt Co. (Private Equity))
• Inventors: Carus, Alwin B., Kucera, Henry
• Primary Examiner(s): Fleming, Michael R.
• Application Number:
  US07/106,535
• Time in Patent Office: 699 Days
• Field of Search: 364/419, 364/900, 364/300
• US Class Current: 704/8
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique

Abstract

An apparatus for the grammatical anlysis of digitally encoded text material receives encoded text, annotates each word of the text with a tag, and processes the annotated text to identify basic syntactic units such as noun phrases and verb groups. A clausal analyzer then operates on the identified nominal and predicate structures to identify clause boundaries and clause types. During processing, feature agreement between parts of successively larger entities--noun phrases, predicates, and clauses--are successively derived. When an error is detected, an error maessage identifies the error and displays a suggested correction.

Claims

1. Apparatus for the analysis of digitally encoded natural language, comprising

means for receiving encoded data representative of a body of natural language words such encoded data including for each word at least a provisionally-assigned syntactic tag of the word,

noun group identifying means operative on the encoded data for identifying noun groups and producing noun data structures representative thereof which augment the encoded data,

verb group identifying means operative on said encoded data augmented by the noun data structures for identifying verb groups and for producing predicate data structures representative thereof which further augment the encoded data, and

clausal analysis means operative on said encoded data augmented by said noun data structures and said predicate data structures for identifying well formed clauses of said body of words.

2. Apparatus according to claim 1, further comprising concordance means for determining agreement among syntactic parts of a clause identified by the clausal analysis means.

3. Apparatus according to claim 1, further comprising

means for identifying a syntax error, and

means for displaying an error message indicative of the identified error.

4. Apparatus according to claim 3 wherein the noun group identifying means identifies the words of a noun phrase from the ordering of syntactic tags of words.

5. Apparatus according to claim 4, wherein the noun group means comprises means for identifying noun phrase boundaries and a feature agreement property.

6. Apparatus according to claim 5, wherein the verb group identifying means comprises a finite state automaton operative on words not lying within the boundaries of a simplex noun phrase.

7. Apparatus according to claim 6, wherein the clausal analysis means comprises means for identifying clause type.

8. Apparatus according to claim 7, wherein the means for identifying a syntax error further comprises concordance means for determining agreement among parts of a clause identified by the clausal analysis means.

9. Apparatus according to claim 8, wherein the means for identifying a syntax error further comprises means for displaying an error correction message keyed to a detected error type.

10. Apparatus for the grammatical analysis of digitally encoded natural language text, such apparatus comprising

code annotation means for annotating each word of a text with data codes, such code annotation means including a stored dictionary of words together with associated data codes, wherein the data codes include tag codes representing possible grammatical or syntactic used of a word, and feature codes representing agreement properties of a word and such code annotation means also including means for looking up words of the text in the dictionary to identify said tag and feature codes for annotating the words, and

noun phrase means operative on the identified tag codes of successive words for identifying noun phrases from the ordering of the tag codes.

11. Apparatus according to claim 10, further comprising

means for processing the feature codes of the annotated words of a noun phrase identified by the noun phrase means to determine feature agreement consistency.

12. Apparatus according to claim 11, further comprising

complex phrase means for identifying complex noun phrases from the noun phrases identified by the noun phrase means.

13. Apparatus according to claim 12, wherein the complex phrase means include means for processing feature codes of constituent noun phrases to determine feature codes for a complex noun phrase.

14. Apparatus according to claim 13, further comprising

predicate means for identifying verb groups of the text.

15. Apparatus according to claim 14, further comprising

means for identifying feature agreement inconsistencies between verb groups and noun phrases.

16. Apparatus according to claim 15, further comprising

means for determining an error message associated with a feature agreement inconsistency and for displaying the error message.

17. Apparatus according to claim 16, further comprising

means for determining a correct replacement word for a word of text having feature agreement inconsistency,

said means for determining and displaying an error message including means for receiving the correct replacement word so as to display the correct replacement word together with the error message associated with the feature agreement inconsistency.

18. Apparatus according to claim 15, wherein the means for processing the identified predicates and noun phrases comprises means for checking subject-verb concordance.

19. Apparatus according to claim 15 wherein the means for processing the identified predicates and noun phrases comprises means for checking subject-verb-object concordance.

20. Apparatus according to claim 10, further comprising predicate means, for identifying verb groups in the annotated text, and

clausal analysis means for operating on identified verb groups to determine the structure of clauses in the annotated text.

21. Apparatus according to claim 20, wherein said clausal analysis means includes means for identifying types of clauses.

22. Apparatus according to claim 21, further comprising means for identifying clause-based errors of syntax.

23. A method of analyzing a digitally encoded body of natural language words for grammatical well-formedness, such method comprising, in order, the steps of

(I) annotating the words with candidate syntactic tags to produce a data structure,

(II) inspecting the ordering of the candidate tags to identify noun groups and producing noun data representative thereof thereby augmenting the data structure,

(III) processing the annotated words and the noun data to identify verb groups and producing predicate data representative thereof further augmenting the data structure, and

(IV) processing the further augmented data structure to identify well-formed clauses of said body of words.

24. A method according to claim 223, further comprising the step of displaying an error message when one of said steps of inspecting or processing detects probable errors.

25. A method according to claim 24, wherein the step of displaying an error message includes the step of displaying a message identifying a probable error together with a correction therefor.

26. An automated parser for processing digitally encoded natural language text, such parser comprising

means for annotating words of text with possible grammatical tags,

noun means operative on the annotated words of text for identifying noun phrases and for further annotating the text with noun phrase data,

predicate means operative on said annotated words of text and noun phrase data for identifying verb groups and for further annotating the text with verb group data, and

error detection means included in said means for annotating, said noun means and said predicate means, for identifying text errors during operation of the aforesaid three means and displaying an indication thereof,

whereby errors in text are identified for correction during processing stages of ascending structural complexity thereby permitting in-process correction of errors and parsing of unedited text without breakdown of the processing operations or requiring excessive processing time.

27. A method for grammatical analysis of digitally encoded natural language text, such method comprising the steps of

providing an automated means for annotating each word of a text with data codes, such means including a stored dictionary of words together with associated data codes, wherein the data codes include tag codes representing possible grammatical or syntactic uses of a word, and feature codes representing agreement properties of a word and such automated means being operative for looking up words of the text in the dictionary to identify said tag and feature codes thereby annotating the words, and

providing an automated noun phrase identifier which successively inspects the order of the identified tag codes of successive words to identify noun phrases of the text.

28. A method for an automated parsing of digitally encoded natural language text, such method comprising the steps of

annotating words of text with possible grammatical tags,

identifying noun phrases and further annotating the text with noun phrase data,

identifying verb groups and further annotating the text with verb group data,

identifying text errors during the aforesaid three steps and displaying an error indication thereof, and

correcting the displayed errors in text during processing stages of ascending structural complexity thereby permitting automated parsing of unedited text without breakdown of the processing operations or requiring excessive processing time.