JPH01184578A - Japanese language processing system - Google Patents

Abstract

PURPOSE:To correctly analyze a double subjective case syntax by deciding whether or not an input Japanese sentence is the double subjective case syntax, and starting up a processing for the double subjective case syntax when it is the double subjective case syntax. CONSTITUTION:At a Japanese language double subjective case syntax decision part 11 in a Japanese language double subjective case syntax analyzer 10, it is decided whether or not a syntax equivalent to the double subjective case syntax (A is the declinable word of B) exists in the analysis intermediate expression 101 of a Japanese language text. When the double subjective case syntax exists, it is sent to a Japanese language double subjective case syntax analysis part 12, and when it is not, the syntax is sent to another analyzer. At the analysis part 12, a first subjective case clause and a second subjective case clause are decided. Next, analysis is performed by a dictionary retrieval part 13 by retrieving a declinable word analysis dictionary 20. When the meaning attribute of a main noun in the first subjective case clause is not included in (time), a meaning relational table 30 is retrieved by a table retrieval part 14, then, meaning relational information between the noun of the first subjective case clause and that of the second subjective case clause can be obtained. In such a way, syntactic and meaning analysis are performed and an analyzed result is outputted.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a Japanese language processing system, and particularly to a natural language translation system that automatically translates a Japanese sentence read from an input device into another natural language sentence. The present invention relates to a technology for analyzing double nominative constructions with two nominatives contained in input Japanese sentences in Japanese language processing systems.

[Traditional techniques, techniques]

Double nominative constructions take the form of ``A is B is the term J (A and B are noun phrases), which are classified into the following five types.

(1) When the case particle ``ga'' expresses the object case Example: ``Kare f'' likes him (the object case of ``Sukida'' is the noun ``Kare'' with the case particle ``ga'') (2) 1illll r is Example: When the adverb ``wa'' substitutes for the case particle of the adjunctive type.Example) The experiment was conducted privately.(The adverb ``wa'' substitutes for the case particle ``wo''.) Example of substituting the case particle ``no'': Friction coefficient f is small (the subparticle ``wa'' is substituting the adjunctive case particle ``no'') (4) Noun phrase with the subparticle ``wa'' J When represents time and acts like an adverb Example: Sales of jujube in August are going strong. , In Japanese double nominative constructions like this, the adverbial particle ``ha''
The clause with the case particle ``ga'' is called the first nominative clause, and the clause with the case particle ``ga'' is called the second nominative clause.

In conventional Japanese processing systems, input Japanese sentences are parsed using only connective value grammars or case grammars, regardless of whether the input Japanese sentences are in double nominative constructions or not.

[Problem to be solved by the invention]

In the conventional technology, when the input Japanese sentence has a double nominative construction, types (1) and (2) above can be analyzed if there is a description in the grammar, and type (5) can be analyzed using the subparticle "
If "wa" is recognized as a "case", it can be analyzed, but for types (3) and (4), it was not possible to correctly analyze clauses with the adverb "wa".

In addition, in the input Japanese sentences, double nominative constructions may occur where the case particle ``ga'' is replaced with an adverbial particle, or the adverbial particle ``wa'' is replaced with another adverbial particle (e.g. (I like him and my girlfriend.) In such a case, the clause with either adverbial is in the first nominative construction, regardless of whether it is of type (1) or (2). Therefore, it was extremely difficult to determine which one was the second nominative clause.・In other words, in the conventional technology, it is possible to parse a part of the basic form of the double nominative construction, "A is B is predicate" (types (1), (2), and in some cases (5)) , types (3), (4), and cases where the case particle r is replaced with an adverbial particle (e.g., [A is, B is also a verb]), it has the disadvantage that it cannot be analyzed correctly. Ta.

SUMMARY OF THE INVENTION An object of the present invention is to provide a Japanese language processing system that can solve the above-mentioned drawbacks of the prior art and can correctly analyze input double nominative constructions regardless of their type.

[Means to solve the problem]

In order to achieve the above object, the Japanese language processing system of the present invention determines whether or not an input sentence is a double nominative construction with two nominatives based on the combination of particles of each clause of the input sentence. The first method uses word order to determine the sentence, and the first method determines the semantic relationship between the two nominatives, the bond valence of the predicate, and Using information obtained by searching a dictionary that describes case structure using the predicate to which each clause depends,
A second means for analyzing the syntactic role and semantic role of each clause is provided.

[For production]

By determining whether the input Japanese sentence is a double nominative construction, and in the case of a double nominative construction, activating processing for double nominative constructions, double nominative constructions can be correctly analyzed. By using the word order information and semantic relation information in addition to the information on the bond valence and case structure of words, double nominative constructions can be correctly analyzed for types (3) and (4). moreover,
Even for double nominative constructions with the form ``A is also a predicate'', by using word order information and semantic information in addition to the bond valence and case structure information of the predicate, it is possible to determine which subparticle Since it is possible to recognize which phrase is the first nominative phrase and which is the second nominative phrase, it can be correctly analyzed.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows a configuration diagram of an embodiment of a portion related to the present invention in a Japanese language processing system. In FIG. 1, 101 is the morphological analysis, the dependencies are the intermediate parsing expressions of the analyzed Japanese sentence, and 102 is the analysis result after the analysis by the double nominative parser 10.

The Japanese double nominative parsing device 10 is the core of the present invention, and in terms of hardware, it includes an arithmetic device, a program,
It consists of a memory that stores intermediate analysis results, etc., but its functions include a Japanese double nominative construction determining unit 11 that determines whether the parsed intermediate expression 101 of a Japanese sentence is a double nominative construction, Japanese double nominative syntax analysis unit 12. which analyzes double nominative construction using information such as, semantic relationship between two nominatives, and word order information. It is divided into a dictionary search section 13 that searches the phrasal analysis dictionary 20 using a predicate as a key and retrieves desired information, and a table search section 14 that searches the semantic relationship table 30 and retrieves the desired information. The term analysis dictionary 20 corresponds to each term,
The semantic relationship table 30, which describes information regarding the bond valency of predicates, describes the relationships between the semantic attributes of nouns.

These word analysis dictionary 20 and semantic relationship table 30 are generally placed on an external storage device.

FIG. 2 shows a schematic flow of the operation of the Japanese double nominative parsing device 10.

The parsing intermediate representation 101 of the Japanese sentence, which is input to the Japanese double nominative parser 10, is a morphological analysis.

The analysis of the dependencies has already been completed. Note that these analysis methods are not particularly limited here.

In response to the parsing intermediate representation 101 of the five-element Ljl rising Japanese sentence, the Japanese double nominative construction determining unit 11 analyzes the parsing intermediate representation 1 of this Japanese sentence.
It is determined whether there is a construction in 01 that corresponds to the double nominative construction "A is a predicate of B". In other words, the phrase ``r〜wa'' is followed by the phrase ``〜ga'' (other elements may be inserted between both clauses), and both clauses are combined with a predicate that dominates the double nominative construction. Whether or not there is a relationship between the dependencies, or if there are multiple clauses with sub-particles that can be converted to the case particle ``ga'', and at least two of them can be rewritten to the case particle ``ga''. Therefore, the dependency between these clauses and the predicate governing the double nominative construction is the intermediate expression 1 for parsing Japanese sentences.
01, if there is a double nominative sentence, the parsing intermediate representation 101 of this Japanese sentence is sent to the Japanese double nominative sentence parsing unit 12.
If it does not exist, the intermediate parsing representation 101 of this Japanese sentence is sent to another analyzing device.

The Japanese double nominative syntax analysis unit 12 uses the dictionary search unit 13 for the parsing intermediate representation 101 of the Japanese sentence that is determined to be a double nominative construction. and the table search unit 14 to perform the analysis as follows.

First, which clause in the intermediate expression 101 for analyzing a Japanese sentence is the first nominative clause and which clause is the second nominative clause is determined using the particle and word order information of each clause (step 202
), then the dictionary search unit 13 searches the predicate analysis dictionary 20 using the predicate governing the double nominative construction as a key,
Information about the predicate is obtained, and an attempt is made to analyze the parsing intermediate expression 101 of the Japanese sentence using this information (step 203
).

For double nominative construction types (1) and (2) (and part of (5)), both the first nominative clause and the second nominative clause can be analyzed using the dictionary information of the pragmatic analysis dictionary 20. Therefore, for types (3) and (4) (and some other parts of (5)), only the second nominative clause can be analyzed, and the first nominative clause cannot be analyzed at this point. In step 203, it is determined whether both nominative clauses can be analyzed (step 204), and if both the first nominative clause and the second nominative clause can be analyzed.

The Japanese double nominative analysis unit 12 outputs the analysis result, but if it has not been analyzed, it continues to analyze the first nominative clause as follows.

The semantic attribute of the head noun of the first nominative clause is examined, and it is determined whether the semantic attribute is included in "time" (step 20
5) If the semantic attribute is included in ``1 hour'', the first nominative clause is analyzed as adverbially modifying the predicate. Therefore, the analysis result is output with the first nominative clause as an adverbial modification (step 206), type of double nominative construction (4)
is analyzed in both nominative cases in this way. On the other hand, the first
If the semantic attribute of the noun in the nominative clause is not included in "1 hour", the table search unit 14 searches the semantic relation table 3.
0 is searched to obtain the semantic relationship information between the noun of the first nominative clause and the noun of the second nominative clause, syntactic and semantic analysis is performed based on this information, and the analysis result is output (step 207).

Hereinafter, the double nominative parsing of a Japanese sentence according to FIG. 1 will be explained using a specific example.

FIG. 3 shows an intermediate parsed representation of the Japanese sentence "He likes her" (hereinafter referred to as example sentence 1). Example sentence 1 belongs to the double nominative construction of type (1).

In example sentence 1, the Japanese double nominative construction judgment unit 11 determines that in example sentence 1, a clause with the subverb μ ``wa'' is followed by a clause with the case particle ``ga'', and both clauses are in the predicate ``r''. 3, it is determined that it is a double nominative construction, and the parsed intermediate expression shown in FIG. 3 is sent to the Japanese double nominative syntax analysis unit 12.

The Japanese double nominative syntax analysis unit 12 first determines the first nominative clause and the second nominative clause, but in example sentence 1,
It is clear from Figure 3 that the first nominative clause = ``he'' and the second nominative clause = ``she.''

Next, the dictionary search unit 13 searches the dictionary search unit 13 for the dictionary analysis dictionary 20 using the term ``suki da'' as a key, and dictionary information for the term ``suki da'' is obtained.

Figure 4 shows the contents of the phrase analysis dictionary 20 for the phrase ``I like the phrase r.'' In Figure 4, 400 is the font of the phrase, 40
1 is the grammatical and semantic information that should be possessed by a clause related to the phrase ``I like the word r'', and the part 401-1 contains a particle equivalent to the case particle ``ga'' (including an adverbial particle that can be converted to ``ga''). If the semantic attribute of the head noun of the clause with is included in the semantic attribute ``subject'' and there is a dependency relationship between the clause and the clause, then the nominative case of the clause is It means that
In the part 401-2, the semantic attribute of the main noun of the clause with a particle equivalent to the case particle ``ga'' or ``wo'' is the semantic attribute r book'' (
r-” indicates that there is no restriction on the semantic attribute),
If there is a dependency relationship between the predicate ``suki da da'', it means that the clause is the object case of the predicate ``suki da da''. 402 is the meaning of the word \"I like it," and is shown in English for convenience.

The Japanese double nominative syntax analysis unit 12 refers to the target case column of the dictionary contents (401-2 in FIG. 4 for example sentence 1),
Determine whether a clause with the case particle "ga" can be an object case. As shown in Figure 4, in the phrase ``I love you'', the phrase can be the object case, so the second nominative phrase = ``Kare ga'' is compared with the part 401-2 in Figure 4. The head noun of the second nominative clause, ``girlfriend,'' has the semantic attribute ``subject.''

Since it satisfies 401-2 in Figure 4, the second nominative clause = ``Kare ga'' is analyzed as the object case of ``I like the word r''. The first nominative clause = ``He'' has an adverbial particle ``wa'', and this ``wa'' can be converted to the case particle ``ga'', so a comparison with part 401-1 in Figure 4 is It will be done. The head noun of the first nominative clause = "he" has the semantic attribute "subject" and satisfies 401-1 in Figure 4, so the first nominative clause = "he" is used It is interpreted as the nominative of ``I like words.'' At this point, the analysis of both the first nominative clause and the second nominative clause has been completed, so the analysis results are output.

Figure 5 shows the analytic intermediate expression of the Japanese sentence ``He likes her too'' (hereinafter referred to as Example Sentence 1'), in which the case particle ``ga'' in Example Sentence 1 is simply replaced with the adverbial particle ``Mo''. This example sentence 1', which shows , also belongs to type (1) double nominative construction.

In the Japanese double nominative construction determination unit 11, example sentence 1' is
The adverbial particles ``wa'' and ``rmo'' are both adverbial particles that can be converted into the case particle ``ga'', and both clauses are ``r-like''.''
, it is determined that it is a double nominative construction, and the fifth
The parsing intermediate expression shown in the figure is the Japanese double nominative parsing unit 12.
sent to.

The Japanese double nominative syntax analysis unit 12 first determines the first nominative clause and the second nominative clause.As in example sentence 1', the case particle "ga" does not exist, and the case particle "ga" is selected. When there are multiple convertible subparticles, the case particle closest to the beginning of the sentence is
In Example Sentence 1', the clause with an adverbial particle that can be converted to ``ga'' is the first nominative clause, and the second nominative clause is a clause with an adverbial particle that can be converted from the beginning of the first sentence to the second case particle ``ga''. Fifth
From the diagram, the first nominative clause = "he" and the second nominative clause = "she also j. Second nominative clause: +: The clause determined as "she too" has the case particle "ga". Since it is considered to be the same as a bunsetsu, the subsequent analysis is the same as for example sentence 1.

Figure 6 shows the analysis results for example sentence 1, and Figure 7 shows the analysis results for example sentence 1'. The difference between Figures 6 and 7 is due to the difference between the case particle "ga" and the adverbial particle "mo". The only difference is the particle information of the second nominative clause.

FIG. 8 shows an intermediate parsed representation of the Japanese sentence ``I conducted the experiment'' (hereinafter referred to as 9 example sentence 2). This example sentence 2 belongs to the double nominative construction of type (2).

In the Japanese double nominative construction judgment unit 11, 9 example sentences 2 have a clause with the adverbial particle "wa" followed by a clause with the case particle "ga", and both clauses are related to the predicate "do". Therefore, it is determined that it is a double nominative sentence, and the parsed intermediate expression shown in FIG. 8 is sent to the Japanese double nominative sentence analysis unit 12.

In the Japanese double nominative syntax analysis unit 12, in the same way as in Example Sentence 1, first, the first nominative clause = "experiment wa" and the second nominative clause = "watashi ga" are determined. The dictionary search unit 13 searches the dictionary search unit 13 for the dictionary analysis dictionary 20 using "do" as a key, and obtains dictionary information for the word "do".

FIG. 9 shows the contents of the word analysis dictionary 20 for the word "do". In Figure 9, the term “do” is ■
It is assumed that it has three different dictionary information: , ■, and ■.

In the Japanese double nominative parsing unit 12, ■, ■,
Analyze in what sense the term ``do'' is used in ■. First, for ■ in Figure 9, we look at the object case in the same way as in Example Sentence 1. However, since the word particle r is not an object case, the clause with a case particle (in Example Sentence 2, the second nominative clause = ``I Compare ``only'') with ■ in Figure 9. The information held by “I” is the 9th.
It matches the nominative information in ■ in the diagram, so ``I'' becomes the predicate ``
It is interpreted as the nominative for ``do''. In addition, if there are multiple clauses with case particles, analyze each clause with a zero-order adverbial particle (in example sentence 2, only the first nominative clause = "experiment wa") and compare it with ■ in Figure 9. do. In ■ in Figure 9, the information on the nominative case has already been used, so the information on the target case is compared with the information on j. The adverbial particle ``wa'' can be converted to the case particle ``wo''. The semantic attribute ``survey/research'' of ``experiment'' is included in the semantic attribute ``human activity'' that the object case of ■ in Figure 9 should have, so the semantic information is considered to be consistent. Therefore, the information about "experiment" is
“It satisfies the object case information of ■ in Figure 9, and is analyzed to be the object case of the term ``do''.

The same thing is done for ■ and ■ in Figure 9, but for ■ and ■ in Figure 9, the information on the object case does not match the information in "r experiment is" (semantic attributes are not in an inclusive relationship) So,
``The experiment'' remains unanalyzed.

In the Japanese double nominative parsing unit 12, ■, ■,
Looking at the matching status with ``2'', the best result for ``2'' in Fig. 9 is finally selected.If there are multiple matches with the same matching status, multiple results are also selected. - In the result of ■ in FIG. 9, both the first nominative clause and the second nominative clause are analyzed, so the analysis results are output. Example sentence 2 in Figure 10
The analysis results are shown below.

FIG. 11 shows an analytic intermediate expression of the Japanese sentence "Ice has a small coefficient of friction" (hereinafter referred to as Example Sentence 3).

This example sentence 3 belongs to type 3 double nominative construction.

The Japanese double nominative syntax determination unit 11 determines that example sentence 3 is a double nominative construction like example sentences 1 and 2, and the parsing intermediate expression shown in FIG. 11 is determined by the Japanese double nominative syntax analysis unit. Sent to 12th.

In the Japanese double nominative syntax analysis unit 12, the first nominative clause = ``ice'' and the second nominative clause = `` similarly to example sentences 1 and 2.
The friction coefficient is determined to be ``,'' and the dictionary information shown in FIG. 12 is obtained from the phrase analysis dictionary 20 using the phrase ``small'' as a key.

From Figure 12, 1. Since the term ``small'' does not have an object case, the second nominative clause (in example sentence 3, the second nominative clause = [friction coefficient ga]) is compared with Figure 12.

Since the information held by ``friction coefficient ga'' matches the nominative information in FIG. 12, ``friction coefficient ga'' is analyzed to be the nominative for the term ``small.'' Since the analysis of clauses with case particles is completed up to 2N, next clauses with sub-particles (in example sentence 3, only the first nominative clause = ``ice'') are compared with Figure 12. However, from FIG. 12, the term "small" has only the nominative case, and since the nominative case has already been analyzed, the analysis using the contents of the predicate analysis dictionary 20 ends at step 2.

Since the first nominative clause has not been analyzed yet, it is next checked whether the semantic attribute of the head noun of the first nominative clause ("ice" in example sentence 3) is included in "1 hour". However, since the semantic attribute "solid" of "ice" is not included in "1 hour",
The semantic relationship table 30 is searched by the table search unit 14, and analysis is performed based on the information.

FIG. 13 shows the contents of the semantic relation table 30 for the semantic attribute "solid" of "ice".
01 is a key for searching the semantic relationship table 30, 50
2 is the semantic attribute “Solid J is.

It shows that there is a semantic relationship called "attribute" with "semantic attribute r quantity".

The Japanese double nominative syntax analysis unit 12 examines the semantic relationship between the head noun "ice" of the first nominative clause and the head noun "friction coefficient" of the second nominative clause based on the contents of FIG. “Ice” has the semantic attribute “solid,” and “friction coefficient” has the semantic attribute “quantity.” According to Figure 13, the semantic relationship between the two is “
At this point, the first nominative clause,
The second nominative clause is analyzed together, and the analysis result is output.

FIG. 14 shows the analysis results for example sentence 3.

Figure 15 shows the Japanese sentence “Elephants also have long trunks” (Example Sentence 3 below)
This example sentence 3', which shows the parsed intermediate expression of '', also belongs to type (3) double nominative construction.

Figure 16 shows the contents of the pragmatic analysis dictionary 20 for the word ``long'', Figure 17 shows the contents of the semantic relationship table 30 for the semantic attribute ``animal'', and Figure 18 shows the analysis results for example sentence 3'. .

In the Japanese double nominative construction determining unit 11, one example sentence 3' is
Since the subparticle ``mo'' can be converted to the case particle ``ga'' just like the subparticle ``wa'', it is determined to be a double nominative construction in the same way as example sentence 1゜2.3, and Fig. 15 The parsed intermediate expression is sent to the Japanese double nominative construction analysis unit 12. In the Japanese double nominative parsing unit 12, example sentence 3' is analyzed in the same manner as example sentence 3, as shown in FIG. In addition.

As shown in FIG. 17, even if there are multiple contents of the semantic relationship table 30 for the semantic attribute of the head noun of the first nominative clause (“zomo” in the case of example sentence 3′), the second nominative clause ( In the case of example sentence 3', the semantic attribute of the main noun of the second nominative clause "hana" is searched and the semantic relationship between the two is obtained. Since it is a ``body,'' the semantic relation ``Whole part'' is selected from Figure 17.

Figure 19 shows the intermediate expression of the Japanese sentence "Sales are good in August" (hereinafter referred to as Example Sentence 4), Example Sentence 4.
belongs to type (4) double nominative construction. FIG. 20 shows the contents of the phrase analysis dictionary 20 for the phrase “do well.”
Figure 20 shows the analysis results.

In the Japanese double nominative construction determination unit 11, example sentence 4 is determined to be a double nominative construction in the same manner as example sentences 1 and 2.3, and the parsed intermediate expression in FIG. 19 is the Japanese double nominative syntax analysis. 1 Japanese double nominative sentence sent to unit 12 In the same way as in example sentence 3, the first nominative clause = "August wa" and the second nominative clause = "Sales ga" are determined, and the second nominative clause = "Sales ga". 20
According to the content of the phrase analysis dictionary 20 for the phrase ``brisk'' shown in the figure, it is analyzed that ``sales ga'' is the nominative of the phrase ``brisk.''' Up to 25, the first nominative clause = "August wa" is unanalyzed, so the semantic attribute of the head noun "August" in the first nominative clause is included in the semantic attribute "1 hour", but it is not. Find out. "August"
Since the semantic attribute ``year, month, day'' is included in the semantic attribute ``1 hour,'' the first nominative clause -rB month is'' is analyzed as an adverbial modification.

In this way, both the first nominative clause and the second nominative clause are analyzed, and the analysis results shown in FIG. 21 are output.

Specific examples have been explained above, and as can be seen from these, the Japanese double nominative construction determination unit 11 determines whether the input Japanese sentence is a double nominative construction, and in the case of a double nominative construction, By activating the processing of the Japanese double nominative syntax analysis unit 12, the double nominative sentence can be correctly analyzed. Furthermore, by using the word order information and the semantic relationship information in the semantic relationship table 30 in addition to the information on the bond valence and case structure of the predicates in the pragmatic analysis dictionary 20, type (3) and (4) double Can correctly analyze nominative constructions. Furthermore, rA
B also uses the word order information and semantic information in addition to the bond valency and case structure information of the predicate for double nominative constructions with forms such as the predicate j. Since it is possible to recognize which clause is the first nominative clause and which is the second nominative clause, it is possible to correctly analyze double nominative constructions with forms such as "A is also a predicate of B".

[Effects of the Invention] As explained above, according to the Japanese language processing system of the present invention, when an input Japanese sentence is analyzed, if the Japanese sentence has a double nominative construction, the usage in the sentence is Because it uses not only the bond valence and case structure information of words, but also the single word order information and semantic relationship information, it has the advantage of being able to correctly analyze the syntactic and semantic structures of sentences for all double nominative construction types. be.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the Japanese language processing system of the present invention, FIG. 2 is a flowchart for explaining the operation of FIG. 1, and FIGS. FIG. 2 is a diagram illustrating a specific example of double nominative Japanese sentence parsing according to FIG. 10...Japanese double nominative syntax analysis device, 11...Japanese double nominative syntax determination unit. 12... Japanese double nominative syntax analysis unit, 13... Dictionary search unit, 14... Table search unit, 20... Pragmatic analysis dictionary, 30... Semantic relationship table, 101... Intermediate expression for analysis of Japanese sentences, 102...Analysis results. Figure 3: Japan's 48th sentence, daylight, central situation, building 11, drum number 1-
Suko,・Figure 4, 'M name"@s
, fcJ+<light jsM@M't%d'i@,. Figure 5: 411 Sentence 1'Nirijizu 3 Yone 4 Shiobun I) Kasuri Oka Dimension 1 This Precinct Gu Indignation t8 Pregnant” Isultpufu---=[Nie-11------ 25 Yoyoyoyo 7.: 111. Work, Moyoyoyo 1 ■ 17-Figure 6: 1
Toshifumi 14 vs. 1st army with fL's materials, 7 yen and Kasuri - Main side (to take 1, and Saito I goes)
Shout ite” Wandering 1; 1 total 4J size) 8
°′ Nesho Shatsu Hayashi ■To (1-? b”) Monthly 1 茅ト (Koya.tsu SljWl Figure 8: Ijyū 2 Towns and Villages 1 H)
Figure 9: ff) f, rf goradekuri 111 yen 1 N1 square cow 1 introspection ＼ Figure 10: 1 Libun 2 and λzui to be dismantled (I 5 qtq *, W] 2---
-----, -1 [Yoyoyoko]]]],]],]],]
:=Yo Sanzo Tko r Figure 11: I'm Hξ length against I's first burn ^
Figure 12: ffi? 'j”v-
J+'7@'LfQ! Mmj now F circle 13 Fig. 14: I711 length j / i oh son-in-law d8 稟下t-effect Fig. 15: fallen sister' (1) weir masu f kaku 1L still 1e-'' (riy!, (,Zmafl,lζbow "shi1"htbow゛auxiliary string 4 sweating spirit@#"ta up to order -7 bribe" Figure 16: To the east of the seam 11 introduction N cultivation body No. 17 c toni!-1-31-hand t51 force, me ξffjs,]'
, UAy-q-□-1194, Figure 1S: (l'1 sentence 3
``#1'aP@'T3fJL@h!Write゛°
Front part Shun T ≧ Ni Seihiko ■ i19 diagram: 4 risu 11 day 1 meeting matari ^ shi example leaf r
4 button Kojiru 11 2 lights on the helmet も - Obstruction % f - ' 1st addition figure: Trout m R bonito - shouting Nuha, the treetops I am ugly - 21st figure: 11 prayers in the evening according to the law ! ；Last year 11, my skin became 1″
It's on the net! DemasllAtto

Claims (1)

[Claims] (1) In a Japanese language processing system that processes an input Japanese sentence, whether or not the input sentence is a double nominative construction with two nominatives is determined by the combination of particles in each clause of the input sentence and the word order of the clauses. A first means of determining the sentence using a double nominative construction, and a first method of determining the sentence using the first method, which determines the semantic relationship between the two nominatives and the connective valence and case of the predicate. a second means for analyzing the syntactic role and semantic role of each clause using information obtained by searching a dictionary describing the structure using the predicate to which each clause is dependent; A Japanese language processing system characterized by having: