JPH03248264A - Japanese-language sentence elaborating device - Google Patents

Abstract

PURPOSE:To standardize the style of Japanese-language documents by subjecting an input Japanese sentence to morpheme analysis processing to detect a clause including a predicate and referring to a style expression table in accordance with this predicate to detect a clause of a designated style and adjusting the clause in the case of difference as the comparison result. CONSTITUTION:Morpheme analysis results 11 of the input Japanese sentence and style designation 12 are inputted to a Japanese-language sentence elaborating device 10, and a style detecting part 2 detects a clause including a predicate. The predicate is discriminated by a predicate discriminating part 3, and the detecting part 2 refers to a style expression table 5 based on this predicate and the designated style to extract a clause of the designated style. When the style of the extracted clause and that of the clause in the input Japanese sentence are different as the comparison result, the clause is supplied from the detecting part 2 as the clause requiring style adjustment and passes a style adjusting part 6 and a clause converting part 7 to output a clause whose style is converted in accordance with the conversion operation instruction from the table 5, thus standardizing the style of Japanese-language documents.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a Japanese sentence editing device, and particularly to a Japanese document editing device that inputs a Japanese document composed of a plurality of Japanese sentences and processes the Japanese document. In a processing system, there is a Japanese sentence editing device that performs processing to unify the writing style of Japanese documents.

[Conventional technology]

Conventionally, technology related to document editing support in Japanese document processing systems has mainly focused on word-level editing techniques, such as detecting typographical errors and unifying terminology. There are few processing systems, but there are some technologies that have already been published in the literature regarding document editing functions, such as [Nakatama.

Kogen, Daikoku, Tsutsumi: Japanese processing in text creation support systems (3), Information Processing Society of Japan 35th National Conference 4S-7,
(1987) J (Reference 1), “Fukushima.

Otake, Oyama 2 Document rewriting function in the Japanese document creation support system COME, Information Processing Society of Japan 34-cap National Convention 6X-6, (1987) J (Reference 2).

However, in terms of the document editing function up until now, Document 1
As can be seen in the examples in ``■'', only the style of document expression can be unified. In addition, as seen in the example described in Document 2, there are problems such as ■ In order to deal with a large number of expressions, all expressions that cannot be processed using the basic principles must be registered in the rewriting table in advance, and There are problems such as it may not be possible to adjust parts of inappropriate writing style that exist in the text, or it may not be possible to correctly adjust writing style for expressions that have not been registered in advance.

[Problem to be solved by the invention]

As mentioned above, many of the conventional document editing techniques support the correction of word-level errors such as typos and omissions, and there are also a few published document editing techniques that involve unifying writing styles. Although it is possible to deal with expressions at the end of a sentence such as "... deru." For expressions that cannot be handled with , there is a problem that it cannot be processed unless the rewriting pattern for all expressions is registered in the table in advance.

The present invention has been made to solve the above problems.

An object of the present invention is to provide a Japanese editing device that provides an editing function of unifying the style of Japanese documents that have already been input in a Japanese document processing system that processes Japanese documents as input. .

[Means to solve the problem]

In order to achieve the above object, the Japanese sentence editing device of the present invention inputs a Japanese document consisting of a plurality of Japanese sentences and processes the Japanese document in a Japanese document processing system. This is a Japanese sentence elaboration device that performs processing to unify Japanese sentences, and includes an input section for inputting morphological analysis results for each Japanese sentence making up a document and a designation of the writing style to be unified, and a section that extracts clauses containing predicates from the morphological analysis results. A predicate judgment unit that detects clauses that exist at the end of a sentence, a sentence-end judgment unit that detects clauses that exist at the end of a sentence, a stylistic expression table that registers stylistic expression rules, and a sentence-final judgment unit that detects clauses that exist at the end of a sentence. a writing style detection unit that detects a clause that requires stylistic adjustment; and a clause conversion unit that adjusts the writing style of the clause detected by the writing style detection unit to a designated writing style using the writing style expression table. , a writing style adjustment unit that adjusts the writing style of the Japanese sentence to be processed by the phrase conversion unit to a designated writing style, and an output unit that outputs morphological data of the Japanese sentence whose writing style has been adjusted.

[Effect]

According to the above-mentioned means, the Japanese sentence elaboration device includes an input section 9 that inputs the morphological analysis results for each sentence constituting the Japanese document to be processed and the designation of the writing style to be unified; The Japanese sentence elaboration device is equipped with 9 sentence ending determination units 9 stylistic expression tables, 9 sentence style adjustment units 9 phrase conversion units, and an output unit that outputs morphological data of Japanese sentences whose writing style has been adjusted. A designation of the writing style that is desired to be unified with the document resulting from the morphological analysis performed on each Japanese sentence constituting the Japanese language document is input. The predicate determination unit detects clauses containing predicates from the morphological analysis results of each Japanese sentence constituting the Japanese document to be processed.1 The sentence-end determination unit detects clauses at the end of sentences.0The style detection unit first The writing style of a clause including the predicate detected by the predicate determination unit is detected using a writing style expression table. Thereafter, the writing style of the phrase including the detected predicate is compared with the writing style designation, and phrases having a writing style different from the specified writing style are detected as phrases requiring style adjustment.

If the writing style requires adjustment, the detected phrase is sent to the writing style adjustment section. The writing style adjustment department.

First, the phrase converting section converts the phrase sent from the writing style detecting section that requires stylistic adjustment based on the conversion operation command written in the stylistic expression table. Thereafter, the converted clauses are incorporated into the morphological analysis results of each Japanese sentence forming the original Japanese document to be processed, and rewritten morphological data is generated. The generated rewritten morpheme data is outputted via the output section.

In this way, according to the Japanese sentence editing device, the writing style of Japanese documents can be unified by specifying the writing style. Also, here:
The final nine sentences are distinguished, and stylistically inappropriate expressions that exist within the sentences are also unified. The stylistic expression table registers stylistic expression rules as conversion operations based on the connection relationship with the immediately preceding word, and by using the conversion operation rules of this stylistic expression table, even somewhat special expressions can be written without a special rewriting table. It is possible to unify the writing style.

Therefore, in a Japanese document processing system that has a style unification function, it is possible to handle stylistically inappropriate expressions that exist not only at the end of a sentence but also within the sentence, and even for somewhat special expressions, the style can be unified without a special rewriting table. Elaboration equipment can be provided.

〔Example〕

Hereinafter, one embodiment of the present invention will be specifically described using the drawings.

FIG. 1 is a block diagram showing the configuration of a Japanese text editing device according to an embodiment of the present invention. As shown in FIG. 1, the Japanese sentence elaboration device 10 includes an input section 12, a writing style detection section 2. It is composed of a predicate determining section 3, a sentence ending determining section 4, a writing style expression table 52, a writing style adjusting section 62, a clause converting section 7°, and an output section 8.

FIG. 2 is a diagram showing an example of the structure of a stylistic expression table. As shown in FIG. 2, the stylistic expression table 5 in this Japanese sentence elaboration device 10 has a relational rule 5 a between the part of speech and the conjugated form of the connection source for the stylistic expression 5 b, and the stylistic expression 5 c t and stylistic conversion. This is a table in which operation 5d is registered in advance.

Continuing the explanation with reference to FIG. 1, the input unit 1 inputs the sentence morphological analysis result 11 of the document that has already undergone morphological analysis processing for each Japanese sentence that makes up the Japanese document to be revised. , writing style designation 12 is input. Morphological analysis is a process that divides Japanese sentences written in kanji and kana into word units. During this process, the part of speech and conjugation of words are simultaneously detected, and clauses are also identified. .

For example, when morphological analysis is performed on the Japanese sentence ``This is the only area that the system has constructed.'', the analysis results are broken down into word units and phrase units, as shown in Figure 3. Parts of speech and conjugations are detected, and clauses are also recognized. In other words, the Japanese sentence input in step 21, ``This is the only area that the system is building.''

” is decomposed into word units and phrase units in step 22, and the system is broken down into word units and clause units to create the system that is constructed by the area
This/Only/Desuhachi” (/ indicates a word boundary,
// indicates a clause boundary). Then, the part of speech and conjugation form of the word are detected for this decomposed word, and in step 23, a morphological analysis result is obtained.

The Japanese processing system is equipped with such a morphological analysis processing function (not shown), and this morphological analysis processing function is used to analyze each Japanese sentence that makes up the Japanese document to be revised. On the other hand, the sentence morphological analysis result 11 of the document is input from the input unit 1 after the morphological analysis process is performed.

In addition, writing style specification 12 is input data that specifies the writing style of the entire Japanese document to be revised, and is specified by determining the input method specified by the user himself and the writing style that is frequently used in the document to be processed. There is an input method. A writing style designation 12 obtained by these input methods is input from the input unit 1.

The writing style detecting unit 2 detects the mid-sentence 2 obtained by the predicate determining unit 32 and the sentence ending determining unit 4 for each Japanese sentence that constitutes the document to be processed.
From a clause containing a predicate at the end of a sentence.

According to the designation of writing style and the writing style expression table 5, clauses requiring adjustment of writing style are detected. The predicate determining unit 3 detects a clause containing a predicate from the morphological analysis result 12. The sentence-end determining unit 4 detects the phrase at the end of the sentence among the clauses containing the predicate obtained from the predicate determining unit 3, and adds identifier data for determining the end of one sentence in the sentence to the clause containing the predicate. The adjustment unit 6 transfers the phrase detected by the writing style detection unit 2 to the phrase conversion unit 7, and rewrites the morpheme data using the rewritten phrase obtained there.

Furthermore, using the rules for writing style conversion operations in the writing style expression table 5, the phrase conversion unit 7 regards phrases that require stylistic adjustment, such that phrases that occur in the middle of a sentence are always treated as normal, and phrases that occur at the end of a sentence are treated as regular phrases. Rewrite as specified. The rewritten morphological data is outputted via the output unit 8.

The details will be explained below. Here, the following two examples of Japanese sentences are included in the Japanese document to be processed, and a process of adjusting the style of the example sentences of the Japanese sentences will be described as an example.

Example sentence: ``Please wash your hands before reading the book.''
(Designation of writing style: Regular) Example sentence ■ “Do not run the system immediately after it starts up.” (Designation of writing style: Respectful) These example sentences ■ and ■ have been revised and rewritten as follows. shall be treated as such.

Example sentence ■′ “Wash your hands before reading the book.” Example sentence ■″ “
Do not operate the system immediately after it boots up. ” In this case, first, the two example sentences (■) and (2) are each subjected to morphological analysis processing, and the results of the processing are input to the Japanese sentence elaboration device 10.

FIG. 4 is a diagram showing an example of morphological analysis processing for example sentence (■), and FIG. 5 is a diagram showing an example of morphological analysis processing for example sentence (■).

That is, for example sentence ■, as shown in Figure 4, for example sentence ■ Japanese sentence "Please wash your hands before reading a book" input in step 24, the words are input in step 25. Break it down into units and phrases and write ``hand/wo//
Wash / Mashi / Te / Kara / / Book / / Read / De / Please 8” (note that / indicates a word boundary and // indicates a clause boundary), and further, for this decomposed word, , the part of speech and conjugation of each word are detected, and in step 26, the morphological analysis results of the part of speech and conjugation of each word are obtained as shown in the figure.

Similarly, for the example sentence (■), as shown in Figure 5, the Japanese sentence of the example sentence (■) input in step 27 is ``Do not start the system immediately after it starts up.''
, in step 28, it is broken down into words and phrases, and is written as ``system/ga//rising/mashi/even/.
/Immediately,/is/acts/makes/nahachi' (/indicates a word boundary, // indicates a clause boundary), and furthermore, for this decomposed word, the part of speech of the word, Conjugation is detected.

In step 29, as shown in Figure 1, the morphological analysis results of the part of speech and conjugation form for each word are obtained.

In this way, each Japanese sentence constituting the Japanese document to be revised is recognized in clause and word units through morphological analysis, and at the same time, the part of speech and conjugated form of the word are recognized.

The input unit 1 generates morphological analysis results 11 as shown in FIGS. 4 and 5 for each example sentence ■ and example sentence ■.
At the same time, the writing style designation 12 is inputted and transferred to the writing style detection section 2.

6a is a flowchart showing the processing flow of the writing style detection section 2, FIG. 6b is a flowchart showing the processing flow of the predicate judgment section 3, and FIG. 6c is a flowchart of the processing flow of the sentence-end judgment section 4.

In the writing style detection unit 2, as shown in FIG. 6a, step 3
In step 1, the morphological analysis result and writing style instructions are input, and in the next step 32, the morphological analysis result is sent to the predicate determining section 3 to determine the predicate. Then, in the first step 33, the predicate determining unit 3 determines whether or not there is a detected predicate. As a result, if there is a result output of the determined predicate, the process proceeds to step 34, where the writing style of the clause including the predicate is detected using the writing style expression table.

Next, in steps 35 and 36, the writing style of the six detected clauses is determined.If the writing style among the six clauses is a middle-of-the-sentence clause, whether or not it matches the regular style (step 35), and if it is a clause at the end of a sentence, determines whether it matches the specified writing style. If they do not match, proceed to step 37 and output the clause containing the predicate. If they match, that is fine, and return to step 33 to output the clause containing the predicate. Step 33: perform the process, and repeat this process as many times as there are results from the predicate determination unit.
If it is determined that there is no result output from the predicate determining unit, the process ends.

On the other hand, in the predicate determination unit 3, as shown in FIG. 6b, the morphological analysis result is input in step 41, and in the next step 42, the predicate is detected from the part of speech and conjugated form of the word based on the morphological analysis result. Perform the processing to do. Then, in the first step 43, it is determined whether a phrase including a predicate has been detected. As a result, if a clause including the predicate is detected, the process proceeds to step 44, where the clause including the predicate is sent to the end-of-sentence determining unit, and processing is performed to determine whether the clause is the end of the sentence.

Then, in the next step 45, a clause containing a predicate determined as ``in the middle of a sentence'' or ``at the end of a sentence'' is output.Next, the process returns to step 43 and the first-order predicate (a clause containing the detected predicate) is output. ) is repeated for the number of phrases that include the predicate. If it is determined in step 43 that there is no phrase that includes the predicate, the process ends.

Furthermore, the end-of-sentence determination unit 4 performs end-of-sentence determination processing as shown in FIG. 6c. First, in step 51, a clause containing a predicate is input (from the predicate determination unit 3), and in the next step 52, a sentence-final symbol is input.

Performs sentence-final expression detection processing. Next, in step 53, it is determined whether there is a sentence-final symbol or a sentence-final expression. If there is a sentence-final symbol or sentence-final expression, it is determined that the sentence is at the end, so in step 54 outputs clauses containing predicates. If there is no end-of-sentence symbol or end-of-sentence expression, it is determined that the sentence is in the middle of a sentence, so the process proceeds to step 55, where a mark of ``0 in a sentence'' is added and the clause containing the predicate is output, and the process ends.

In this way, the writing style detection section 2 receives the morphological analysis result and the designation of the writing style, and transfers the morphological analysis result to the predicate determination section 3 (steps 31 and 32). Predicate determination unit 3
Now, the morphological analysis results are input, and based on the morphological analysis results, the predicate determination section 3 detects the predicate from the part of speech and conjugation of the word, detects a clause containing the predicate, and sends it to the sentence-end determination section 4 (step 41 to Step 44) The 0-sentence-end determination unit 4 inputs a clause containing a predicate, detects a 1-sentence-end expression and a sentence-final symbol, and determines whether the clause containing the predicate exists in the middle of the sentence or at the end of the sentence (step Steps 51 to 53) mark the phrase as "in the middle of the sentence" or "at the end of the sentence" and return it to the predicate determining unit 3.

(Steps 54 and 55), the predicate determination unit 3 returns the morphological analysis result and the clause containing the predicate determined to be “in the middle of the sentence” or “at the end of the sentence” to the writing style detection unit 2.

FIG. 7a is a diagram showing an example of processing by the predicate determining unit for example sentence (■), and FIG. 7b is a diagram showing an example of processing by the predicate determining unit for example sentence (■).

In the case of the example sentence ■, as shown in Figure 7a, the predicate determination unit 3 determines the phrase [钑-ii/mashi/te/kara] that includes the predicate in the example sentence ■ (/ is a word boundary, and is a predicate). In contrast, the in-sentence-final judgment is determined as “in-sentence”, and for the phrase “Nezutomo/De/Pleaserehachi” (/ is a word boundary, and is a predicate) that includes a predicate, the in-sentence-in-sentence final judgment is judged as “in-sentence”. It is determined that it is the “end of sentence” and is output.

Similarly, in the case of the example sentence ■, as shown in FIG. For predicates), the in-sentence-end judgment is judged as “in-sentence”,
For the phrase ``Kamimujo/Seru/Nan'' (/ is a word boundary, and is a predicate) that includes a predicate, the end of the sentence in the sentence is determined to be ``the end of the sentence'' and is output.

The writing style detecting unit 2 detects the writing style of the clause including the predicate by associating the clause including the predicate obtained from the predicate determining unit 3 with the writing style expression in the writing style expression table 5 (step 34). As described above, the stylistic expression table 5 (FIG. 2) is table data in which the relationship between stylistic expressions, the parts of speech of their connection sources, and conjugations, and the rules for conversion operations in stylistic conversion are registered in advance. The example of the stylistic expression table shown in FIG. 2 particularly corresponds to the above-mentioned example sentences (2) and (2) and shows an example of the table data of the portion used in the writing style conversion operation.

For example, in the example of the stylistic expression table 5 in FIG.
is a respectful form, and the conversion operation 5d for converting into a normal form instructs to perform a conversion operation to ``remove ``masu'' and its conjugated forms.''

The writing style detection unit 2 further applies the rule that phrases in the middle of a sentence are always regular sentences and sentences at the end of the sentence follow the designation of writing style to the writing style of clauses that include predicates, and detects phrases that require stylistic adjustment. is detected (Step 34), and transferred to the writing style adjustment unit 6.

FIG. 8a is a flowchart showing the processing flow of the writing style adjustment section 6, and FIG. 8b is a flowchart showing the processing flow of the one-phrase conversion section 7.

In the style adjustment unit 6, as shown in FIG. 8a, step 8
In step 1, input the output of the 1 writing style detection section, the morphological analysis result, and the designation of the 2 writing styles. Next, in step 82, the output of the writing style detection section and the designation of the writing style are sent to the phrase conversion section to convert one writing style. Namely.

The clauses requiring style adjustment and the designation of the writing style are transferred to the clause conversion unit 7 to convert one writing style.Next, in step 83, the converted clauses are replaced with the existing clauses in the morphological analysis results. Then, in step 84, the replaced morpheme data is output.

On the other hand, in the phrase conversion unit 7, as shown in FIG. 8b, in step 85, a phrase including predicates marked "in the middle of the sentence" and "at the end of the sentence" is input, and in step 86, the phrase is inputted from the stylistic expression table 5. In step 87, the phrase is converted according to the conversion operation in order to rewrite the phrase that requires stylistic adjustment. Then, in step 88, the converted phrase is output, and the rewritten phrase is output to the writing style adjustment section 6.

FIGS. 9a and 9b show an example of the processing performed by the clause converter 7 on the example sentence (■), and FIGS. 10a and 10b show an example of the process performed by the clause converter 7 on the example sentence (■).

For example, in the case of the processing example shown in FIG. 9a, the phrase conversion unit 7 performs the following:
For the phrase "Arai/mashi/te/kara" which is input in step 91 and requires stylistic adjustment, a phrase conversion operation is performed in step 92 from keitai to jotai. In this clause conversion operation, the respectful expression ``Mazu'' and its conjugated form are removed, so here the respectful expression ``Mashino'' is removed, and in step 93, ``Arai/te/kara'' is output.

Furthermore, in the case of the processing example shown in FIG. 9b, the phrase converter 7 converts the phrase "Yomi/
In step 95, a phrase conversion operation from keitai to jotai is performed for ``de/please eight''. In this clause conversion operation, the particle + "please" is removed and the independent word is converted into the imperative form, so "deskeep" is removed, and in step 96, "Koryu/. (imperative form)" is output. be done.

Similar clause conversion processing is also performed on the example sentence ■, as shown in FIGS. 10a and 10b. Chapter 10a
In the case of the processing example shown in the figure, the phrase conversion unit 7 converts the phrase “rise/mashi/
In step 98, the phrase conversion operation from keitai to jotai is not performed for ``mo.''. In this phrase conversion operation, the respectful expression "masu" and its conjugated forms are removed,
The respectful expression "mashi" is removed, and in step 93, "
``Rise up/temo'' is output.

In addition, in the case of the processing example shown in FIG. 10b, the phrase conversion unit 7
In step 101, a phrase conversion operation from normal style to respectful style is performed for the phrase ``moto/suru/nahachi'' input in step 100 that requires stylistic adjustment. In this clause conversion operation, ``na'' is removed and ``nai-'', ``de'', and ``please'' are added. As a result, in step 96, "operate/enable/not/do/please" is output.

After such processing, the writing style adjustment unit 6 generates adjusted morphological data for one writing style by incorporating the rewritten clauses into the morphological analysis results of each Japanese sentence that is the original processing target (step 83). , and output the exchanged rad morpheme data (step 84).

FIG. 11 shows an example of processing the morpheme data after rewriting the example sentence (■), and FIG. 12 shows an example of processing the morpheme data after rewriting the example sentence (■).

The style-adjusted morpheme data generated as described above is outputted via the output unit 8, and converted into a style-adjusted Japanese sentence by an inflection conversion device that converts existing inflections.

In other words, in the case of one example sentence, as shown in FIG. //Book l is changed to //l!A/ (imperative form) 8'', and the conjugated form is converted in step 113, the revised Japanese sentence is 1 hand and then book 11 is written. " is output.

In addition, for the example sentence ■, as shown in FIG. /soon/to/is//works/dan3
α unnatural form) / not / de / please 8” and step 1
When the conjugated form is converted in step 13, the refined Japanese sentence ``Please do not start working immediately after the system starts up'' is output.

The present invention has been specifically explained above based on examples, but
It goes without saying that the present invention is not limited to the embodiments described above, and can be modified in various ways without departing from the spirit thereof.

〔Effect of the invention〕

As described above, by equipping a Japanese document processing system with the Japanese sentence editing device according to the present invention, it is possible to easily unify the writing style. Here, stylistically inappropriate expressions that exist other than at the end of a sentence can be given the same function as the end of a sentence to unify the writing style, and there is no need to prepare a special rewriting table for this purpose.

In this way, a Japanese document processing system with a stylistic unification function can handle stylistically inappropriate expressions that exist not only at the end of a sentence but also within the sentence, and can unify the writing style even for somewhat special expressions without the need for a special rewriting table. A Japanese sentence editing device can be provided.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a Japanese sentence editing device according to an embodiment of the present invention, and FIG. 2 is a diagram showing an example of the configuration of a stylistic expression table. FIG. 3 is a diagram illustrating an example of a morphological analysis process for a Japanese sentence, and FIG. 4 is a diagram illustrating an example of a morphological analysis process for an example sentence (■). FIG. 5 is a diagram showing an example of morphological analysis processing for example sentence (2). FIG. 6a is a flowchart showing the processing flow of the writing style detection section. FIG. 6b is a flowchart showing the processing flow of the predicate determination unit. FIG. 6c is a flowchart of the processing flow of the sentence-end determination unit; FIG. 7a is a diagram showing an example of processing of the predicate determination unit for example sentence ■; FIG. 7b is a diagram showing an example of processing of the predicate determination unit for example sentence ■; Figure 8a is a flowchart showing the processing flow of the writing style adjustment unit; Figure 8b is a flowchart showing the processing flow of the clause conversion unit; Figures 9a and 9b show an example of the processing of the clause conversion unit for example sentence ■. Figures 10a and 10b are diagrams showing a processing example of the phrase conversion unit for example sentence ■. Figure 11 is a diagram showing a processing example of morpheme data after rewriting example sentence ■. It is a figure which shows the processing example of the morpheme data after rewriting example sentence (■). In the figure, 1...input section, 2...writing style detection section, 3...
Predicate determination unit, 4... Sentence end determination unit, 5... Stylistic expression table. 6... Writing style adjustment unit, 7... Clause conversion unit, 8... Output unit. 10... Japanese sentence elaboration device, 11... Sentence morphological analysis result, 11... Writing style specification, 13... Morphological data after rewriting.

Claims (1)

[Claims] (1) A Japanese sentence editing device that performs processing to unify the writing style of the Japanese document in a Japanese document processing system that inputs a Japanese document consisting of multiple Japanese sentences and processes the Japanese document, An input section that inputs the morphological analysis results for each Japanese sentence that makes up a document and the designation of the writing style that you want to unify, a predicate determination section that detects clauses that include predicates from the morphological analysis results, and detects clauses that exist at the end of sentences. A sentence ending determination unit, a stylistic expression table that registers stylistic expression rules, and a stylistic expression table that detects clauses that require stylistic adjustment to the specified writing style in the Japanese sentence to be processed. a writing style detection unit that adjusts the writing style of the clause detected by the writing style detection unit to a writing style specified using the writing style expression table; and a clause conversion unit that specifies the writing style of the Japanese sentence to be processed. 1. A Japanese sentence elaboration device, comprising: a writing style adjustment unit that adjusts the writing style; and an output unit that outputs morphological data of a Japanese sentence whose writing style has been adjusted.