Analyzing Japanese Double-Subject Construction 
having an Adjective Predicate 
Masahiro Oku 
NTT Information and Communication Systems Laboratories 
1-2356 Take, Yokosuka-shi, Kanagawa, 238-03 Japan 
oku@isl.ntt.jp 
Abstract 
This paper describes a mefllod for ~malyzing 
Japanese double-subject constniction having an 
adjective predicate based on the valency 
structure. A simple sentence usually has only 
one subjective case in most languages. 
However, lnmly Japancse adjeclivcs (and some 
verbs) can dominate two surface subjective 
cases within a simple sentence. Such sentence 
structure is called lhe double-subject 
construction. This paper classifies lhe Japtulese 
double-subject constntction inlo four types mid 
describes problems arising when analyzing 
fllcse lypes using orditmry Japmlese construction 
approaches. This paper proposes a method tbr 
analyzing a Japanese double-subject 
construction having an adjective predicate in 
order to overcome lhe problems described. By 
applying tiffs melhod to Jap~mese senlence 
,'malysis in Japanese-to-English machine 
~,mslalion systems, tl'm~slalion accuracy can be 
improved because tiffs melhod eml analyze 
correclly the double-subject construction. 
1 Introduction 
We have bccn developing a Japanese-to-English 
madfine lrauslation syslem called ALT-J/E (lkehm-a 
et al., 1987). In ALT-J/E, Japanese sentence analysis 
is performed based on lhe wdency slruclure for the 
predicalc of the input sentence. ALT-J/E c~m trm~slate 
several special Japm~ese constntclions such as predicate 
idioms and funclion verb expressions (Oku, 1992). 
The Jap~mese double-subject conslntclion was nol 
handled well by the original ALT-J/E. This paper 
describes a melhod lbr mmlyzing a Japanese double- 
subject construction having an ad.i eclivet)rcdicate bascd 
on Ihe valency structure. 
Simple sentences normally have only one sub.icctivc 
case in most languages. However, it is possible for 
many Japlmese adjectives (and so me verbs) to dominate 
two surface subjective cases within a simple sentence. 
Such sentence structure is called lhe double-subject 
construction. JEiDA (Ihe Japan Elcclronic Industry 
Development Associalion) has issued evaluation 
st~mdards for machine lnmslation systems (JEIDA, 
1995). ~Ihe main aim of lhese standards is to establish 
objective test sets for machine translation system 
evaluation. Therefore, they include ahnost of all 
linguislic phenomena flu,l the systems have to process. 
In lhc slandards, the double-subject eonslruclion is 
listed as one of the conslruclions lhat arc difficult lo 
process successfidly. 
Many Japmlesephilologistshavesludicdthc double- 
subject construction because it is necessary to reveal 
the synlaclic or scmmltic structure of lhis construction 
if wc are to establish a syntactic or sem~ullic lheo~, 
lor the Japanese language (lshigami, 1977) (Kuno, 
1973) (Mikami, 1961)). However, there arc a few sludies 
for processing this conslmction from lhe point of view 
of computational linguislics or engineering (Mttrala, 
1990). ~llmrelbrc, lhc aim of lhis paper is to discuss a 
melhod for ;malyzing lhc Japanese double-subject 
constnlction havhlg an adjective predicate from lhe 
point of view of engineering. 
2 Definition of key terms 
Key terms used in this paper me defined as follows: 
• case-marldng particle 
In English, each case is marked by its relative posilion 
in terms of tile predicate or by a preposition. In contrast 
to lhis, Japanese language marks each case by a certain 
sort of postposilional parlicle located next to noun 
phvases sudl as "ga" or "we". Such particles arc callcd 
"case-markiu g pm-ticl es". 
• adverbial particle 
"Adverbial particles" give a case an additional function 
(topicalizalion etc.) by their allachmenl to Ibc casc- 
markingparlicle. Some adverbial particles such as "wa" 
or "me" often stand-in for case-marking parlicles and 
give the case ;m additional fimclion. 
• sentence having an adjective predicate 
In the Japanese I~mguage, adjectives function as 
predicates in sentences as do verbs, qlmrefore, in this 
paper, lhe sentence in which m~ adjective acts as a 
predicate is called a "senlence having an Mieclive 
predicate". 
. double-sul;jeet construction 
Usually, a simple sentence has only one subjeclivc 
case. ttowevcr, m~my Japanese adjeclives (and some 
865 
verbs) can dominate two surface subjective cases within 
a simple sentence. Such sentence structure is called 
tile "double-subject constrnction". This constn~ction 
includes both adverbial particle "wa" and subjective 
case-marking p~trticle "ga". Therefore, the Japanese 
double-subject construction is also referred to as file 
"wa"-"ga" construction. 
• valency structure 
The sentence structure can be considered as a 
combination of a predicate and its modifiers. Such 
sentence stntcturc is called the "valency structure". 
The valency structure represents what surface cases 
the modifiers dominated by a given predicate 
correspond to. 
• valency pattern 
"Valency patterns" are stnlcture patterns that formulate 
possible valency structures for predicates. A valency 
pattern is defined for each usage for each predicate. 
Figure 1 shows an example of a valency pattern for a 
verb "shoukai sunl (introduce)". For vcrbs and 
adjectives, wdcncy patterns formulated as shown in 
figure 1 are collect ed beforehand in the dictionary called 
the valency pattern dictionary. 
Japanese valency pattern tbr "~,~ 9'\]" J-~" 
N1 \] SR: \[(agent)\] 
(suljective) \] JR: 73";, h "~ ~J 
N2 \] SR: \[(agent)\] 
N3 | SR: \[(agent)l 
(objective2)\] JR: ~ 
Remarks: 
Ni: Label for a valency element 
SR: Semantic restriction on a noun 
JR: Restriction on case-marking particles 
Figure 1. An example of a valency pattern. 
• valency element 
In the valency stn~cture, each relation between a 
predicate and its modifier is called a "valency element". 
As shown in figure I, valency elements are described 
using both the semantic restriction on the noun including 
the modifier and the restriction on the case-marking 
pmaicles including the modifier. 
• sentence analysis (or analyzing a sentence) 
"Sentence analysis" is the process that reveals the 
valency structure of the input sentence. ALT-J/E 
perfonns sentence analysis by binding the modifiers 
for the predicate in the input sentence to the valency 
elements in the wtiency pattern for the predicate in the 
wdency pattern dictionary. 
• semantic category 
A "semantic category" is a class for dividing nouns 
into concepts according to their meaning. For example, 
both "man" and "woman" belong to the same semantic 
category \[(human)\]. In the below, the words surrounded 
by 1( )\] are semantic categories. 
3 Japanese Sentence Analysis based on 
the valency structure 
Figure 2 outlines the translation process in ALT-J/E 
(Hayashi, 1987). Morphological analysis segments the 
input Japanese sentence into its component words such 
as predicates and nouns. Dependency analysis 
determines the dependency structure to indicate the 
association between words. From the association 
between a predicate and its modifiers in the dependency 
structure, sentence analysis tries to determine the 
wflency strncture, i.e., it determines, for the valency 
pattern for the predicate, which valency element ead~ 
modifier corresponds to. This valency stnlcture is then 
converted into the equivalent English sentence 
structure. Finally, the outpnt English sentence is 
generated from the stnlcture. 
Input Japanese sentence 
\[Moq~hological analysis\]~-\]apanese dictional"~y~ 
Dictionary about \[ \[Dependency analysis j . . \[ dependency structure 
I Sent ence analysis I'~--~Val~~~ona~ry \[ 
I Structure .1 " " \[ dictionary l 
Output English sentence 
Figure 2. Outline of translation process in ALT-J/E, 
The following describes how ordintu'y sentences 
are analyzed by ALT-J/E using the valency structure 
for the predicate of file input sentence. 
Sentence analysis is the process that converts the 
dependency structure into file valency structure by 
referring to wdency patterns. All valency patterns for 
each usage with each predicate are prepared beforehand 
and held in the valency pattern dictionary.. As shown 
in figure 1, for each predicate, both file semantic 
restriction on a noun including its modifier and file 
restrictions on case-marking particles including its 
modifier are described for each valency element. When 
the modifier in the input sentence satisfies both 
restrictions for a certain wllency element in the valency 
pattern for the predicate, the modifier is botmd to that 
valency element. The valency smlctnre is determined 
by binding all predicate modifiers to valency elements. 
An adverbial particle can correspond to more than one 
case-marking pataicle, for example, "wa" is a possible 
proxy for "ga", "wo", "hi" and so on. Therefore, sentence 
analysis first tries to bind modifiers that have case- 
marking particles, each of which represents which case 
is unambiguously marked by particle spelling like "ga" 
866 
Japanese input sentence: ~Lt:kg.L~L'~L:/)))~;~{7'~ b-l-<o (he introduced his sister to me) 
¢ morphological analysis and dependency analysis 
dependency structure 
.~ refer to valency pattern 
sentence analysis as ordinary construction 
~r 
valelley strilt'|lll'e 
(-J ~(~ Ji" "I- % (past tense) --'~') .,,- "-~_ (introduce) .... ~C 
(st,bjectivecase)} <ctivecase l)~ (iobjectivecas;~ 
Figure 3. An example of the Japanese sentence analysis. 
Japanese wdcncy mttern tbr ";i:~-') 7~,' 
N1 Sit: \[(agent)l 
(subjective) JR: /J~, /~', 
N2 SR: \[(agent)\] 
(oljectivel) JR: ~ 
N3 SR: \[(agent)l 
(objective2) JR: L< 
Remarks: 
Ni: I,abel fur a wdency element 
SR: Semantic restriction on a noun 
JR: Restriction 
on case-tam'king particles 
marks subjective case, to wflency elements in tile 
wflcncy pattern lbr the prcdicate. Thc analysis thcn 
tries to bind modifiers that have ml adverbial particle 
to the non-bolmd valency elements. 
Figure 3 shows an example of this type of Japanese 
sentence mmlysis for the sentcncc "kare wa watashi 
ni kate no imouto we shoukai shita (he theeduced his 
sister to me)". First, the mmlysis tries to bind modifiers 
with case-tam'king parlicles "we" and "ni" to tile wdency 
elements in the wdency pattern for the predicate 
"shoukai sent (introduce)" which is oblaincd IYom lhc 
valency pattem dictionary. As modifiers satisfy bolh 
restrictions on the noun memfing and case-marking 
parlicles, they ,are bound to N2 (objective case 1) mid 
N3 (objective case 2), respectively. The ~malysis then 
lfies to bind lhe modifier with adverbial particle "wa" 
to a non-bound valency element. So far, N2 and N3 
in the wflcncy paltem have already been bound. 
Therefore, as shown in figure 3, the only non-bound 
valency element is N1 which is a subjective case. The 
adverbial particle "wa" c~m stand-in for case-marking 
p~wticlc "ga", which is the non-bound wfiency clcment, 
mid the noun "km-e (he)" satisfies the scmantic 
restriction on the subjcctive case N1 i.e. \[(agent)\]. 
Therefore, file modifier with "wa" is bonnd to tim 
subjective case N1. Finally, lhc valency structure as 
shown in figure 3 is obtained. 
4 Processing the Japanese double-subject 
construction 
Many Japmmse adjective predicates domhmtc two 
subjective cases and so form the double-subject 
construction. The double-subject constmctiou having 
an Miective predicate actually has several wlrimlts, so 
no one approach c,'m be uscd to mlalyze it. Accordingly, 
this section classifies the four types mid the 
characteristics of eadl type ~'e described. 
4.1 Classification of the Japanese dnuble-subjcet 
construction 
Tile Japanese doublc-subjcct consmlction (also calicO 
the "wa"-"ga" conslruction) can be classified into lhe 
following four types based on a previous study 
(lshigami, 1977): 
• type-I 
In this vafi~mt, adverbial particle "wa" is a proxy for a 
case-marking particle such as "ni" other than subjective 
case-ram-king "ga". Example-1 shows "wa" as a proxy 
for case-marking particle "hi" in tile sentence "wamshi 
no ie wa gakkou ga chikai (file school is near my 
house)". 
Example- 1 
(myhouse) (adverbial (school) (case-marking (nea 0 
pro'title "wa") particle "ga") 
~1¢ "wa" is \[br "ni". proxy 
(myhousc) (destination (school) (subjective (hem') case-marking 
case-marking particle "ni") 
pmticle "ga") 
• type-2 
in this wu'iant, adverbial parlicle "wa" is a proxy for 
case-umrking panicle "no" representing a noun 
modifier (pre-nominal). Example-2 shows "wa" as a 
proxy for pre-nominal case-marking p~ulicle "no" in 
the sentence "zou wa ham ga nagai (elcplumls have 
long trunks)". In this wu-itmI, the case of"wa" modifying 
a predicate must be analyzed as "no" modifying the 
867 
Japanese input sentence: 7~7~h~~: t, ~o (elephants have long trunks) Japanese valency pattern for "~ 1~ v, 
dependency structure ~ (subjective) JR: 7~ /F 
@@~ Ni: Label ~br a valency element refer to valency pattern SR: Semantic restriction on a noun 
on case-marking pazticles .' men" ana y~ i: sentence analysis 
as ordinary construction as double-subject construction 
 r.,,,v encyst ct..e va,enc st.,c,,.'o 
'(presen----, tense)--"/ 
Figure 4. An example of analyzing double-suhject construction (type-2). 
case with subjective case-marking particle "ga". 
Therefore, this analysis involves re-formation of the 
valency structure. 
Examlfle-2 
(elephant) (adverbial 0rose) (case-marking (long) particle "wa") particle "ga") 
~¢ "wa" is proxy for "no". 
(elephant) (pre-nominal (nose) (subjective (long) case-marking case-marking 
particle "no") particle "ga") 
• type-3 
In this variant, the case with case-marking particle 
"ga" sometimes represents ~m objective case. Allhough 
an objective case is usually marked by case-marking 
particle "wo", some adjective predicates have an 
oNective case marked by case-marking particle "ga". 
Example-3 shows that "kanojo (she)" with "ga" is an 
objective case and "kare (he)" with "wa" is a subjective 
case in the sentence "kare wa kanojo ga snkida (he 
likes her)". As case-marking particle "ga" normally 
indicates the subjective case, binding "ga" to the 
subjective case leads to incorrect mmlysis if only 
surface spelling is considered. 
Examlfle-3 
(he) (adveNial (her) (case-marking (like) particle "wa") particle "ga") 
• ,, ,, " ,a" re rcsenLs • "r~roxytor ~a.| g ,P .. Pwa" lS ~ ~ ~V all oDjecuve case 
~ (subjective (her) (objective (like) (he) case-marking case-marking 
particle "ga") particle "wo") 
• type-4 
In this variant, the case with adverbial particle "wa" 
acts as all adverbial phrase representing time and is 
actually a special form of type-1. Representing time is 
optional for most predicates. Moreover, Japm~ese time 
expressions are often translated into English adverbial 
phi'ases. Therefore, type-4 is separated from type-1 in 
this classification from viewpoint of enghmering. 
Example-4 shows that time expression "6-gatsu wa 
(in Jmle)" acts as an adverbial phrase in the sentence 
"6-gatsu wa ame ga tot (it has much rain in June)". 
Example-4 
(June) (adverbial (rain) (case-marking (ninth) 
~(t particle "wa") particlc "ga") 
ime expression with "wa" acts as an adverbial phrase 
(June) (rain) (subjective (much) (adverbial case-marking 
phrase) particle "ga") 
4.2 Problems in processing the Japanese double- 
subject construction 
Type-I and type-4 cases can be analyzed using the 
processing flow described in section 3 because 
adverbial particle "wa" simply acts as a case-marking 
particle. However, the following problems ,arise when 
processing type-2 and type-3 cases in the normal way. 
• Problem with type-2 cases 
Fignre 4 shows the m~alysis of the type-2 example, 
"zou wa hana ga nagai (elephanls have long thinks)". 
The predicate "nagai (long)" has only one valency 
element N1 with "ga". According to ordinary sentence 
analysis, the modifier "hana ga" is bound to the valency 
element NI, which means that the other modifier "zou 
wa" is left unbound. That is, sentence analysis camlot 
be completed as shown in the left bottom of figure 4. 
This complicates the accurate Wanslation of this 
modifier into English. 
868 
Japanese input sentence: ~:1:~/&'~O~ ~'57o (he likes her) 
1¢ morphological Japanese wdency pattern lbr "~(.(" ~ 7'~('" analysis ~ln(t dependency analysis 
dcpcndcllcy structure T N1 ISR: \[(human)\], \[(animal)\] 
//~--;~-~'~ ~Q ~-\]~"~ f:/)f @ \]5'(present tense) ~ .fly,} (suhjective) \]I/JR: 75{ k... Ol~w~)..} ~h-e) (@../~...~ (like) ./ 
k..._ ---~-~.-~, ~ ~ " l(obiective,!~ R" t;- -~ ............... -~ refer to valency pattern L__" _____ " '~' 
Remarks: 
Sentence mmlysis Ni: l.ahel for it valency element 
as ordimu T constrttclion SR: Semantic restriction on a noun 
wrong valency str.cture ¢ 
(:~" ~L ~ /5" (present tensei---'L. ''~ "~,______ (like) _ / 
I-----m'--- I"><" \[-~ --N~---I 
Sentence analysis JR: Restriction on case-mm-king patlicles 
as double-subject construction 
~ correct v~n-e 
Q'-~ ~:f- ~.'{ \]5.'(p(esent tense) ") 
(~,o// "-..,(we) 
f(su,:iec.ve  ol.joc"v L.se 
Figure 5. An example of analyzing douMe-sub.iect construction (type-3). 
• ProMem with type-3 cases 
Figure 5 shows the analysis of sentence analysis of 
the type-3 example, "kare wa k~mqio ga sukida (he 
likes her)". If modifiers with a case-marking parlicle 
are preferentially processed as described in seclion 3, 
then the modifier with "ga" binds to subjective case 
N1 which is wrong in this example. This leads us lo 
lhe wrong interpretation as shown in lhe left bottom 
of figure 5. 
5 Proposed method for analyzing Japanese 
double-subject construction 
In order to overcome lhe problems described in the 
previous section, this section proposes a method for 
analyzing a Japanese double-subject conslmction 
having m~ adjective predicate. The mefl~od has fllrec 
processing phases. The first determines whether lhc 
input sentence has double-subject construction or not. 
The second determines which of tim four wtri~mls file 
sentcnce is. qlm last processes the sentence according 
Io its type. 
Figure 6 shows lhc processing flow. The input 
sentence has already undergone morphological analysis 
and dependency mmlysis, i.e., it has been tdready 
determhled whirl nouns modi\[y the adjective predicate 
wilh what sort of poslpositional p~nticle expressions. 
Earl processing phase is described below in detail 
with reference to figure 6. 
• Judgment of douMe-subject construction 
If the input sentence contains two modifiers that have 
adverbial particle "wa" mid case-tam'king particle "ga", 
it is determined as a double-subject constnmtion. 
Sentences olher them double-subject construction, as 
sentence with an adjective predicate 
no =::22di;tif~l~-suhiect constrt~- 
\[type-4 1.~___ yes ~ p~processing\[--~" 
obtain the valency pattern\] 
type-3 \] ~VeS j--~'2z~-.~ p r epr o c e s sing/"-'~-'< ~ 
--n7-- ~__ ~no _ 
~ tt\[i~-~nd each modifier to a valency 1 I\[ element i!} the valency paltern .... l 
/ ~ yus / ~->--1 
h~o bind each m~lilier I ~ro-ce~ing as type-l~ d, \[ \[ff~t wdency element \[ t~ ~~ 
tin the wdency pattern \] I ~,rocessing as tyl,e-2\] 
t'rocessing V as ordinmy sentence end 
Figure 6. Processing flow for analyzing 
a Japanese douMe-suhjeet construction. 
well as sentences with verb predicates, are processed 
normally as described ill section 3. 
• Type determination 
First, lype-4 is set if the modifier with adverbial parlicle 
"wa" represents a time expression or not. Tiffs is 
efficient because type-4 is the most specific type. This 
judgment is performed by checking whether the 
semantic calegory of the noun included in the modifier 
with "wa" is associated with \[(time)\] or not. Second, 
type-3 is set by the valency paltem for the predicate in 
lh¢ input. The predicate in a type-3 case has to cover 
869 
both subjective and objective cases in its usage. The 
valency pattern for each usage for each predicate is 
defined in the valency pattern dictionary. Therefore, 
this judgment is performed by checking whether the 
valency pattern for the predicate includes both a 
subjcctive case and an objective case. Finally, type-1 
and type-2 are differentiated according to the result of 
binding betwccn the modifiers in the input senteuce 
and thevalency elements inthe valency pattern. Type-2 
is set ff the modifier with case-marking particle "ga" 
is bound to the subjective case and the modifier with 
adverbi~d particle "wa" is not. The remaining cases 
are judged as type-1. Here, modifiers with "wa" and 
"ga" ,are often bound in type-1. Notice that the binding 
process resolves what case-marking parlicle the 
adverbial particle "wa" stands-in for. 
• Determining the valency structure 
In the type-4 preprocessing step in figure 6, the modifier 
with "wa" (time reference) is considered as an adverbial 
plwase. Next, binding the other modifiers with the 
valency elements in the valency pattern for the prcdicate 
is attcmpted. This facilitates the translation of Japanese 
time expressions into English. 
In the type-3 prcprocessing step in figure 6, case- 
marking particle "ga" has to be converted into the 
case-marking particle "wo" before binding the former 
which represents an objective case. This conversion 
allows us to correctly bind each modifier to its 
appropriate valency element, becanse predicates in 
type-3 cases have both a subjective case with "ga" and 
an objective case with "wo". See the example in the 
right bottom of fignre 5. 
In type-2 processing, adverbial particle "wa" is a 
proxy for pre-nominal case-marking particle "no" and 
the modifier with adverbial particle "wa" must be 
analyzed as the phrase which modifies the snbjective 
case with "ga". As a result, type-2 mmlysis involves 
changing the valency structure. In figure 4, the noun 
phrase "zou no hana" is fornled from both "zou 
(elcphant)" with "wa" and "hzma 0lose)" with "ga" by 
converting "wa" into "no", and "zou no hana" is bound 
to the subjective case for the predicate "nagai (long)". 
See the example in the right bottom of figure 4. 
Moreover, type-2 processing tries to determine the 
semantic relation between the noun with "ga" and the 
noun with "no" (originally "wa"). Determining the 
semmltic relation helps us translate the Japanese double- 
subject construction into the appropriate English 
construction and expression. Although several semantic 
relations are known, at the present time ALT-J/E can 
resolve only two of them based on semantic categories: 
has-a relation and is-a relation. 
As adverbial particle "wa" in type-I cases is a 
proxy for a case-marking particle such as "ni", "de" 
and so on, type-I cases can be processed in the way 
described in section 3. 
The above approach con'ectly determines die input 
sentence's valency structure which allows the maddne 
translation system to produce more accurate output. 
As a result, the method proposed here improves the 
translation accuracy of ALT-J/E. 
6 Conclusion 
This paper has proposed a method for analyzing the 
Japanese double-subject construction that includes an 
adjective predicate. This paper has classified the 
construction into four .types and described problems 
when applying the ordinary sentence analysis to the 
four types. The algorithm that overcomes these 
problems has been explained in detail. 
This method has been applied to Japanese sentence 
analysis in ALT-J/E. Because this method cau correctly 
analyze the double-subject construction, the method 
helps the translation of this constrnction into an 
appropriate English construction. 
References: 
Yoshihiko Hayashi, "On the Japanese sentences 
mlalysis based on the valency structure", Working 
group report on natural language proces.s'ing, II'S 
Japan, 62-6 (1987) (in Japanese). 
Satoru lkehara, Masahiro Miyazaki, Satoshi Shirai and 
Yoshihiko Hayashi, "Speaker's recognition and 
multi-level-translating method based on it", Trans. 
IPS Japan, Vol.28, No.12, pp.1269-1279 (1987) 
(in Jap~ese). 
Terno lshigami, "Structure of Japm~ese double-subject 
construction with adjective predicates (nijml- 
stmkaku keiyoushi bun no kouzot0 ", 1,;ssay on 
Japanese philology (nihongo-gaku shiron) , Vol.3, 
pp. 1-37 (1977) (in Japanese). 
JEIDA Evaluation standards on machine translation 
systems (JEIDA kikai hon'yakn system hyouka 
kijun), JEIDA report (1995) (in Japanese). 
Susumu Kuno, 7 he structure of the Japanese Language, 
The MIT Press, 1973. 
Akira Mikami, Elephants have long trunks (zou wa 
hana ga nagai), Knroshio-shnppan, 1960 (in 
Japanese). 
Kenichi Murata, "On the formal description of the 
me~mings of Japanese adjectives", l'roc, on the 40th 
annual meeting of IPS Japan, 5F-3 (1990) (in 
Japanese). 
Masahiro Oku, "Methods for analyzing Japanese 
predicate idioms and function verb expressions", 
Journal of lnfi~rmation Processing, Vol. 15, No.3, 
pp.425-433 (1992). 
870 
