Translation between natural languages

• US 4,502,128 A
• Filed: 06/01/1982
• Issued: 02/26/1985
• Est. Priority Date: 06/05/1981
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Nitta, Yoshihiko, Okajima, Atsushi, Yamano, Fumiyuki
• Primary Examiner(s): Smith, Jerry
• Assistant Examiner(s): Fleming, Michael R.
• Application Number:
  US06/383,952
• Time in Patent Office: 1,001 Days
• Field of Search: 364/900, 364/419, 434/167, 434/157, 434/169, 434/176
• US Class Current: 704/8
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/45 Example-based machine translation; Alignment
• CPC Class Codes: G06F 40/55 Rule-based translation
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation

Abstract

An input sentence described by a first natural language is sectioned into individual words. Parts of speech corresponding to the individual words are retrieved from a lexical word storage, whereby the input sentence is described by a corresponding string of the parts-of-speech as retrieved. A translation pattern table is previously prepared which defines correspondence between patterns of strings of parts-of-speech for the first natural language and those for a second natural language by which corresponding output sentence is to be described. By referring to the translation pattern table, the string of the parts-of-speech of the input sentence is transformed into a corresponding string of the parts-of-speech for the second natural language. The output sentence described by the second natural language is generated by sequencing target words in accordance with the sequential order of the parts of speech of the string pattern obtained after the transformation.

Claims

1. A method of performing translations between first and second natural languages comprising:

a first step of sectioning a source text described by a first natural language word by word and storing the words in storage;

a second step of deriving information regarding the parts-of-speech of the stored source words through retrieval of the parts-of-speech information from a lexical dictionary storage in which words are registered together with information regarding the parts-of-speech thereof;

a third step of deriving target words in the second natural language which are equivalent in meaning to the respective stored source words in the first natural language;

a fourth step of translating a first pattern of a string of parts-of-speech formed by gathering the parts-of-speech for each of the stored source words derived in said second step into a pattern in which the order of the parts-of-speech is in accordance with said first natural language, into a second pattern of a string-of-parts of speech with an order in accordance with said second natural language by referring to a translation pattern table which registers therein the correspondences between first patterns of parts-of-speech strings defining sequences of parts-of-speech of the first natural language and second patterns of parts-of-speech strings defining sequences of parts-of-speech of the second natural language; and

a fifth step of generating a sentence in the second natural language by sequencing said target words in accordance with the sequential order of the parts-of-speech in said translated second pattern of the string of parts-of-speech of said second natural language.

2. A method of performing translation between natural languages according to claim 1, wherein the source words are extracted from the source text through a character-based examination with the aid of a space and/or a punctuation serving as a delimiter symbol at said first step.

3. A method of performing translation between natural languages according to claim 1, wherein the information regarding each of the parts-of-speech registered in said lexical dictionary storage includes at the end thereof a mark representing the end of said information.

4. A method of performing translation between natural languages according to claim 1, wherein when a source word from the source text is absent in said lexical dictionary storage, said source word is assigned the parts-of-speech of a noun and is itself used as the target word.

5. A method of performing translation between natural languages according to claim 1, wherein information of the parts-of-speech includes specifically prepared parts-of-speech intended to be used for designating special words as opposed to grammatical parts-of-speech.

6. A method of performing translation between natural languages according to claim 1, wherein said second parts-of-speech string pattern registered in said translation pattern table includes information of word sequence and suffix words in said second natural language.

7. A method of performing translation between natural languages according to claim 1, wherein said translation pattern table includes an area for indicating whether a plurality of the sentences in said second natural language which correspond to a single source sentence in said first natural language are present or not.

8. A method of performing translation between natural languages according to claim 1, wherein when said translation pattern table has registered therein no first pattern of a string of parts-of-speech which corresponds to the string of parts-of-speech of the source text, a message indicative of inability to translate said source sentence is displayed.

9. A method according to claim 1, wherein said lexical dictionary storage also stores therein target words in the second natural language corresponding in meaning to a plurality of words in said first natural language, and wherein said third step comprises retrieving said target words from said lexical dictionary storage for each of said stored source words.

10. An apparatus for translating a sentence described by a first natural language into a sentence described by a second natural language, comprising:

a keyboard means for inputting the sentence described by said first natural language on a character-by-character basis;

keyboard interface means for converting the inputted characters from said keyboard means into corresponding codes to be supplied to a processing unit;

fixed storage means storing therein a dictionary of words including words in the first natural language, parts-of-speech thereof and corresponding words in the second natural language and a translation pattern table defining correspondences between patterns of strings of parts-of-speech relevant to said first natural language and patterns of strings of parts-of-speech relevant to said second natural language;

a random access memory having buffer memory areas for temporarily storing an input sentence and an output sentence;

processing means for retrieving said dictionary of words to obtain information of the parts-of-speech and a corresponding word in the second natural language for each word of an input sentence in the first natural language, storing the information in said random access memory, retrieving said translation pattern table to obtain a pattern of a string of parts-of-speech in the second natural language corresponding to the pattern of a string of parts-of-speech of the input sentence in said random access memory, arranging corresponding words in the second natural language according to the retrieved pattern of string of parts-of-speech in the second natural language to compose an output sentence and storing the output sentence in said random access memory; and

display means for displaying the input sentence and the output sentence both stored in said random access memory.

11. An apparatus for performing translations between first and second natural languages comprising:

means for sectioning a source of text described by a first natural language word by word and for storing the words in a storage;

a lexical dictionary storage in which words are registered together with information regarding the parts-of-speech thereof;

means for deriving information regarding the parts-of-speech of the stored source words through retrieval of the parts-of-speech information from said lexical dictionary storage;

means for producing target words in the second natural language which are equivalent in meaning to the respective stored source words in the first natural language;

means for translating a first pattern of a string of parts-of-speech formed by gathering the parts-of-speech for each of the stored source words derived in said second step into a pattern in which the order of the parts-of-speech is in accordance with said first natural language, into a second pattern of a string-of-parts of speech with an order in accordance with said second natural language by referring to a translation pattern table which registers therein the correspondences between first patterns of parts-of-speech strings defining sequences of parts-of-speech of the first natural language and second patterns of parts-of-speech strings defining sequences of parts-of-speech of the second natural language; and

means for generating a sentence in the second natural language by sequencing said target words in accordance with the sequential order of the parts-of-speech in said translated second pattern of the string of parts-of-speech of said second natural language.