ENGLISH-JAPANESE TRANSLATION THROUGH CASE-STRUCTURE CONVERSION 
Fujio Nishida, Shinobu Takamatsu and Hiroaki Kuroki 
Department of Electrical Engineering, 
Faculty of Engineering, University of Osaka Prefecture, 
Sakai, Osaka, 591 JAPAN 
This paper reports some trials on mechanical 
translation from English to Japanese through a 
case structure constructed on Hornby~s verb patt- 
erns. Though the general theoryof case structures 
is still at the beginning of study, it provides 
partial sentential patterns with rough but reso- 
nable classification labels. After determination 
of schematic dependency relations, multi-vocal 
problems for choosing appropriate equivalents are 
dissolved using subcategories of terms and cases. 
Case structures of English are transformed into 
those of Japanese if necessary, and from those , 
Japanese sentences are generated by a Japanese 
grammar. 
i. Introduction 
English and Japanese are fairly different 
languages from each other. Accordingly, it will 
be necessary to make translation based on a kind 
of deep structures of sentences which determines 
sentential forms schematically in respective lan- 
guages. The case structure used here is a conven- 
tional one which is closely related to surface 
structures. But it was found that the case stru- 
cture is fairly useful by adding some new means 
in construction and usage of it. 
The parsing for obtaining the case structure 
from English sentences is done by category-match- 
ing on a fairly precise verb pattern given by A.S. 
Hornby. It dissolves most of multi-syntactic and 
semantic category problems of words and 
phrases, and determines global dependency 
relations among constituents of English 
sentences. 
Subsequently, it determines the app- 
ropriate Japanese equivalents and post- 
positions among several candidates by 
using subcategories of terms and indivi- 
dual prepositional words. 
The case structure thus obtained con- 
taines the modal part which represents 
various characteristics of the predicate 
part in almost independent form of respe- THINGS 
ctive languages. It also embeds idioma- 
tic expressions in itself. 
At the level of the case structure, 
the difference between both language 
expressions is not so large. Transfor- 
mations of sentential forms such as those 
of non-livlng subjects and those of prin- 
cipal clauses followed by long subordl- 
nate ones for readability are processed 
at this stage without difficulty. From 
the case structures transformed into 
Japanese, translated sentences are gene- 
rated under control of various involved 
-EVENT - 
parameters. 
2. Case structures 
2.1 Case structures of verbs 
Many English words have multi-syntactic and sema- 
ntic categories. However, these multi-categories 
of a word in a sentence should be able to be dis 
-solved in a context of the sentence, namely, in 
syntactic and semantic relations to the neighbo- 
ring words. A.S.Hornby has given comparatively 
precise verb patterns. 3The use of verb patterns 
at parsing is fairly effective to identify the 
syntactic categories of words and phrases as well 
as their syntactic roles such as complements and 
adverbial infinitives. 
On the other hand, an efficient introduction 
of semantic information about words is very im- 
portant in order to identify their semantic roles 
and appropriate Japanese equivalents as well as 
to dissolve the still remaining syntactic ambi- 
guities after a syntactic analysis. For this 
purpose, there are several approaches based on 
the case grammar. 4 It is still in an incomplete 
state for general purpose of thorough comprehe- 
nsion of sentences. But if it is used on detai- 
led syntactic rules of the conventional word 
usage with the aids of categories of an approp- 
riate level at various stages of mechanical tra- 
nslation, these approaches based on the case gra- 
mmar will serve fairly well to assign reasonable 
F EXIStence 
• ~ ATTRibute 
PHYSlcal-~POSSession 
-STate ~STate I RELatio n 
L MENTal ~PERCEPtual STate 
L STate -\]-EMOTive STate 
NATURal phenomenon 
F Physical TRANSfer 
PHYSical~POSSessive TRANSfer 
-~ATTRibute TRANSfer 
~BODily ACTion 
a-PRODuction 
ACTion~ ACTi°n 
L MENTal 
ACTion ~ 
Mental TRANSfer 
PERCEPtual ACTion 
EMOTive ACTion 
THINKing ACTion 
LIVING object ~HUMan r~ical~ <ANIMal 
| c~ \[NON,LIVING rNATURE 
OBJECT ~ object -~PRODUCTS ~ 
MENTal OBJect 
MENTal SUBJect 
Fig.l The upper part of the categorical trees. 
--447-- 
semantic roles to various constituents in a sou- 
rce sentence as well as to choose their appropri- 
ate equivalents and sentential forms in the tar- 
get language. 
From the above considerations, a case struc- 
ture is assigned to a sentence based on Hornby's 
verb patterns. 
Events of discourse are classified by catego- 
ries of predicates. Fig.l shows the upper part 
of the categorical tree used here. The categor- 
ies used for describing case structures are, in 
many cases, those appearing at the rightmost in 
Fig.l. 
Semantic roles of constituents are described 
using case labels shown in Table i. 
Table i Case labels 
PREDicate, AGent, OBJect, EXPERiencer, RECIPient, 
SOurce, GOal, INSTRument, DEGree, COMPARison, 
LOCation, TIME, RANGE, MANNer, ROLE, CAUSE, 
RESULT, MEANS, PURPOSE, CONDition, MODal ( FORM, 
TENSE, ASPECT, AFfirmation, MOOD, VOICE ), °.. 
By using several pairs of categories and case 
-labels chosen from the above the main parts of 
each event are schematically described and called 
case-frames of events. The first row of Table 2 
is the main case-frame of the PTRANS event. 
The m~in part of a case structure based on a 
case-frame is transformed into a surface structure 
using a verb which has a sufficient number of 
Table 2 Case-frame of PTRANS action and 
assignment of verb patterns 
PRED- AG- 
PTRANS THINGS 
VP2B 
VP2E SUBJ- 
ect 
VPI2A 
VPI6A 
OBJ- GO- 
PHYS.OBJ LOCAT U 
PHYS.OBJ 
Indirect 
Direct Object 
Object 
links to describe the main part and belongs to 
an appropriate verb pattern. Table 2 shows verb 
patterns used for English description of PTRANS. 
On the other hand, if a verb pattern is giv- 
en, pairs of categories and case labels of depen- 
dants are almost determined every category of the 
predicate verb as shown in Table 3. Hence, if 
the verb patterns as well as the categories of a 
verb are recorded in a word dictionary, the poss- 
ible case structure can be retrieved at parsing 
immediately. Table 4 shows a part of pairs of 
verb patterns and categories about ~take ~. 
Example i shows several examples of case struc- 
tures obtained by parsing with the aids of Table 
4. 
After determination of a case structure or a 
dependency relation of a sentence, there still 
remain several problems to be solved. 
One of them is the choice of an appropriate 
equivalent or a subcategory. The word rtake~ 
with the verb pattern VP6A and the case structure 
has fairly many Japanese equivalents. However, 
this kind of multi-vocal problems can be dissol- 
ved in many cases by looking up subcategories of 
a few designated dependants involved in the case 
structure and hence in advance by recording them 
for each equivalent in the word dictionary as 
shown in Table 4. 
The other is the choice of an appropriate 
Japanese postposition corresponding to a case 
label and a category of a dependant of a verb. 
Though a good correspondence generally holds bet- 
ween them, there are some exceptions. For these 
cases, the postposition is recorded at the item 
of the associated verb in the dictionary and 
picked up at parsing. 
DEG- PURPOSE MANN 
QUANT -ACT -ACT 
ADVer 
-bial 
Adjunct 
to-INF- 
-- initive 
Table 3 Case structures of verb pattern VPI2A 
SUBJ 
AG- 
THINGS 
IO 
GO- 
LOCAT U 
PHYS. OBJ 
DO 
OBJ- 
PHYS'0BJ 
VPI2A 
PRED- 
PTRANS 
PRED- AG- RECIP- OBJ- 
POSS'TRANS HUM HUM OBJECT 
PRED- 
MTRANS 
AG- 
HUM 
OBJ- 
EVENT U 
MENT.OBJ 
PResent 
PARtici 
-ple 
EXPER- 
HUM 0 
MENT-SUBJ 
Example i 
(i) VP6A : BODily ACTion 
~He took bread. ~ 
(PRED-BOD.ACT:eat, MOD:(TENSE:past 
~OICE:active), AG-HUM:he, OBJ- 
FOOD:bread) 
tHe took a taxif 
(PRED-BOD.ACT:get-on, MOD:(TENSE: 
past, VOICE:active), AG-HUM:he, 
OBJ-VEHICLE:a-taxi) 
(2) VP6A : Mental TRANSfer 
Cl took his speech. ~ 
(PRED-MTRANS:make-a-record-of, 
MOD:(TENSE:past, VOICE:active), 
AG-HUM:I, OBJ-EVENT:his-speech) 
(3) VPI5A : Physical TRANSfer 
~I took the children to the park. 
(PRED-PTRANS:go-leading, MOD: 
(TENSE:past, VOICE:active), 
AG-HUM:I, OBJ-HUM:the-children, 
GO-LOCAT:the-park) 
--448-- 
Table 4 A part of the item 'take' in the 
word-dictionary 
Verb patterns Categories Meanings ( and the designated conditions ) 
VP6A BOD.ACT i. get into one's hand ( OBJ-PHYS-OBJ ) 
2. eat ( OBJ-FOOD ) 
3. get on ( OBJ-VEHICLE ) 
PTRANS I. go carrying ( OBJ-NON.LIVING ) 
2. go leading ( OBJ-LIVING ) 
POSS'TRANS i. receive ( OBJ-COMMODITY ) 
MTRANS i. make a record of ( OBJ-MENT. OBJ 
: U EVENT ) 
VP22 PERCEP.ACT 1. suppose ( OBJ-EVENT ) 
(4) VP22 : PERCEPtual ACTion 
~I took her intelligent. ~ 
(PRED-PERCEP.ACT:suppose, MOD:(TENSE:past, 
VOICE:active), EXPER-HUM:I, OBJ-EVENT: 
(PRED-ATTR:intelligent, OBJ-HUM:she)) 
2.2 Optional cases 
Adverbial phrases and clauses outside verb patt- 
erns constitute optional cases. Most of idioma- 
tic phrase prepositions such as rwith respect to p 
and many subordinate conjunctions such as Cwhen' 
and Calthough * determine the case labels and the 
Japanese postpositions of the following phrases 
and clauses. On the other hand, many prepositi- 
onal phrases which respectively have only one 
prepositions such as c with ~ and Cfor' require the 
information about frame of discourse for deter i 
mination of the case label and the appropriate 
Japanese postposition. These information is the 
categories of the object term of the phrase and 
those of main verb of the governor and further- 
more, in some cases, other information about 
case structure of the verb pattern. 
For example, an adverbial phrase consisting 
of a prefixed preposition ~ in' and an object term 
with category~PHYSical LOCation' can be assigned 
a case label'LOC' or CGOal', but the appropriate 
postposition and case label are determined by 
referring to the semantic category of the main 
verb. 
The necessary information for determining 
them are recorded each preposition and subordi- 
nate conjunction in the word dictionary. 
2.3 Case structures of noun phrases and clauses 
Case structures of noun phrases and clauses have 
similar forms to those of sentences as follows; 
t (Kr-Cl :t~ ,.. • ,K6-CL:t," • • ,K~-Cn:t~ ) (i) 
where t is the main term, ~ s are the depend- 
ants and K~-~ is a pair of a case label and a 
category. 
The case structure only consisting of several 
individual objects and prepositions are deter- 
mined by referring to the categories of the 
objects and their syntactic patterns linked with 
the individual prepositions. 
If a noun phrase contains some verbal nouns, 
the case structure is constructed based on the 
verbalized words of some of them. 
Example 2 
(i) Creceipt of his letter ~ 
C-EVENT: receive(PRED-POSS. TRANS: *, 
OBJ-PHYS.OBJ:Ietter(OBJ: *, 
POSS-HUM:his)) 
(2) t punctual arrival of trains' 
C-EVENT: arrive(PRED-PTRANS: *, AG-VEHICLE: 
trains, MANN-MODAL:punctual) 
where C denotes a certain case for the outside 
governor and the symbol * denotes the prefixed 
term to the case frame containing it. 
3. Parsing 
First, processing of negative expressions 
and idioms are described. They are processed in 
a somewhat special manner. 
3.1 Negative expressions 
Negative expressions of English are fairly 
different from those of Japanese at the points 
where negative adjectives or adverbs of English 
correspond to some negative modifiers and a 
negative auxiliary verb in Japanese. Hence, it 
--449 
will be better to express the negative expression 
of an English sentence in a language-free form at 
the modal part of the case structure and to 
generate the corresponding Japanese sentence from 
that anew. The types of NEGation are classified 
to INTensified TOTal,TOTal and PARtial ones. 
At parsing, the value of type is put in the 
modal part of the case structure together with 
the terms directly negated except the case when 
the terms belong to the predicate part of the 
sentence. 
Example 3 
(i) ~No student can solve the problem~ 
(PRED-THINK.ACT:solve, MOD:(AF:( INT. TOT.NEG, 
student), MOOD:can), AG-HUM:student, OBJ 
-MENT. OBJ:problem(OBJ:*, DET:def)) 
(2) ~Converses are not always true.' 
(PRED-ATTR:true, MOD:(AF:(PAR'NEG, always)), 
OBJ-MENT.OBJ:converses, FREQuency: 
always) 
Most of terms of modal cases are omitted here 
and hereafter except important terms for simpli- 
city. 
3.2 Idiomatic expressions 
For an efficient parsing, idiomatic expressions 
are processed with some priority. They are re- 
corded near the heading of the leftmost wnrd 
among multivocal constituents of the idiom in the 
word dictionary. If the idiom has a separated 
form tl~t z as shown in Eq.(2), 
t I s I •.. s~ t z (2) 
the syntactic and semantic conditions on the 
strings of words or phrases sl, • ..,s~ lying 
between tl and t 2 are recorded together with 
the partial case structure and the Japanese 
equivalent of the idiom in the word dictionary 
as follows; 
so ~ that ~ 
(DEG-QUALity:so\[ADVerb\] very*, PRED-ATTR: ~f 
\[ADJective\], RESULT(that\[SUBORDInate CONJunc- 
tion\])-EVENT: ~\[SENTence\]), 
where ~very*~denotes the Japanese word equi- 
valent to the English word ~so ~. 
When a constituent word. of an idiom is found 
in a sentence at parsing, it is examined whether~ 
or not there exist the rest parts of the idiom 
in the sentence by reading the sentence ahead 
or by examining the partial case structure al- 
ready constructed. If there is, and if the 
intermediate part sl --. s~ satisfies specified 
conditions by parsing, the partial case structure 
and the Japanese equivalent of the idiom are 
added to those already constructed by referring 
to the word dictionary. Otherwise, ordinary 
mode of parsing restarts from term t I 
Example 4 
~The problem was so difficult that he could 
not solve it. ~ 
(PRED-ATTR:difficult, OBJ-MENT.OBJ:problem(OBJ 
*, DET:def), DEG-QUAL:so(very), RESULT-EVENT: 
(PRED-THINK.ACT:solve, MOD:(AF:TOT-NEG, MOOD: 
can), AG-HUM:he, OBJ-THINGS:it)) 
3.3 Construction of case structures 
The rewriting rules used for parsing are descri- 
bed in a Chomsky-like form as follows~ 
V l (K~-C, : tl , IK2-C2 : ~2) 
--~ V2 (K,-CI :tl ) V3 ( r~z- ~Z: J~Z ,VPi ) 
(3) 
V 3 (FKz- 62 : ~2' K3-C3 :ts ' VP£ ) 
--~ V% (\]K~-~z : $2' VP{ ) V 5 (K3-C 3 : t 3 ) 
(4) 
where V{ ( ~ =i ~ 5) denotes a non-terminal 
symbol, ~- ~ : ~ a sequent of several pai~s of 
a case label and a category followed by a term, 
and VPL a label of a verb pattern. 
Eq.(3) represents the type of rewriting rules 
of a subject and a predicate part, and Eq.(4) 
that of a predicate part and a right dependent. 
The parsing is done from left to right in a 
bottom-up manner and case structures are const- 
ructed in the order of the predicate part, the 
right part of a verb pattern of the predicate 
part, the subject part and the optional adverb- 
ial parts. This order is considered to be effi- 
cient for restricting possible construction of 
case structures of many sentences of English 
because the central structure is generally 
constructed in this order. 
At first, the partial case structure of the 
predicate part is constructed. If the main verb 
is a constituent of an idiom, parsing is done in 
a manner described in the preceding section. 
Otherwise, all the case structures for the verb 
patterns are constructed using the word dictio- 
nary and a table of case frames. 
When a main term t6 with a category CL 
comes to be included in a partial case structure 
S} (~ =1,2, ...,m) as shown in Fig.2, syntactic 
and categorical consistency is checked. 
If the consistency holds, the term with the 
category C~L is added to the partial case 
structure S~ with a designated case label K~L. 
And if the case KiL is designated for dissolving 
multi-Japanese equivalents of the main verb in 
the case structure S~ , appropriate Japanese 
equivalents are extracted by subcategory match- 
ing. 
Otherwise, if the category C~& of the main 
term t i does not match one of any partial case 
structure S~ ( ~=1,2, ..-,n), the category C~L 
is deleted. Similarly, if a partial case struc- 
ture requires a main term with a designated 
semantic category and a designated syntactic 
category or a position and cannot find the term, 
the partial case structure is deleted. 
--450 
( SUBJ ) ( PRED ) 
tl t2 t3 
KH 
I 
Cln, l , I C~31 i 
L-__4 r~-i L -- _.J. , Sm.t 
I--_--4 
t+ 
Fig. 2 Construction of case structures 
In such a manner, the number of partial case 
structures decreases or increases as a term is 
added to partial case structures and, in many 
cases, it reduces to unity at the end of parsing. 
4. Case transformation 
Some typical differences between English and 
Japanese sentential forms do not require any 
case transformation, One of them is negative 
expressions and processed as already mentioned 
in section 3.1. The others are oompound noun 
expressions containing some verbal nouns which 
are prefered to be verbalized in Japanese sen- 
tences. However, these Japanese sentences are 
directly generated from the case structure 
obtained by parsing since the case structure of 
a compound noun expression is constructed in the 
same way as a sentence when it contains some 
verbal nouns as mentioned in section 2.3. 
Hence, the main differences which require 
some case transformations are those as mentioned 
as follows. 
4.1 Case transformation for non-causative and 
state oriented descriptions 
Table 5 Case transformations 
Category of Instances of 
predicate predicate 
CAUSE cause,make 
ENABLE enable 
ENABLE allow,permit 
ENABLE prevent 
PERCEPtual look 
ACTion overlook 
PERCEPtual show 
ACTion 
EMOTive surprise 
ACTion please 
Japanese sentences avoid sentential forms in 
which non-living objects bring actions such as 
causative and perceptive ones especially when 
the actions directly affect human beings and 
some events associated with them. In Japanese, 
these sentential forms are replaced by non- 
causative sentences or by resulted-state expre- 
ssions using state-oriented verbs. This problem 
is well known as the non-living subject problems 
and has been fairly well studied by many lingu- 
ists. 5 
It is shown here that the transformation of 
the expressions can be processed in terms of 
case structures in a systematic way. 
First, the value of the voiee in the modal 
case of such a sentence is set passive, because 
in such a situation a passive form is generally 
a little bit more acceptable than an active form 
in Japanese. Second, depending on the category 
of the main verb and the agent, the agent case 
is renamed to a more appropriate case so that a 
more natural Japanese sentence is generated. 
Third, if possible, the equivalent of 
the predicate part is modified into a more app- 
ropriate form such as state-oriented descripti- 
ons. 
Some examples of this modification are shown 
in Table 5. 
Case label Modification of 
for agent predicate 
CAUSE DELetion 
MEANS INTRANS VERB 
MEANS 
CAUSE 
MEANS 
CAUSE 
REPL by'capable 
verb *~ 
REPL by ~not + 
capable verb *~ 
INTRANS VERB SOURCE 
+ capable verb 
REPL by 
MEANS Cunderstand,~ 
CAUSE INTRANS VERB 
--451-- 
The third column of this table designates some 
candidates of case labels to be renamed for an 
agent• The appropriate case label and postpo- 
sition are determined by the category of the term 
to be filled. 
The fourth column designates the version 
into which the equivalent of the predicate 
part is to be modified. The versions are class- 
ified into two kinds. One of them is REPLacement 
of the predicate part by a new intransitive verb• 
The other is some modification of a passive 
Japanese predicate part. The latter is sub- 
divided into DELetion of a causative verb, INT- 
RANSitive VERBalizatio g and addition of 
the meaning Ccapable~ to it. 
The version is coded and the code is recor- 
ded together with a new Japanese intransitive 
verb for REPLacement in the word dictionary. In 
'Fable 5, the words with a symbol * denote the 
equivalent Written in Japanese• 
At the execution of case transformation, the 
routine designated by the code is called and 
carries out the necessary processing. 
Example 5 
By case transformation using Table 5, two 
English sentences 
(i) CThe weight of snow caused the shed to 
collapse' 
and 
(2) 'Our limitted budget did not allow us to 
start a new project' 
are translated to Japanese sentences correspon- 
ding to 
(f)'Owing to the weight of snow, the shed 
collapsed' 
and 
(~)tOwing to our limitted budget, we could not 
start a new project ~ 
respectively. 
The main case transformations used here are 
as follows; 
(i) (PRED-CAUSE:cause, AG-QUANT:weight(D~), 
OBJ-EVENT:(PRED-PHYS.ACT:colIapse, 
OBJ-PHYS. OBJ:shed(D2))) 
---~(PRED-PHYS-ACT:colIapse, OBJ-PHYS.OBJ: 
shed(Dz), CAUSE-QUANT:weight(DI)) 
where ~DI :=OBJ-QUANT:*, DET:def, 
POSS-PHYS.OBJ:snow, 
D z :=OBJ-PHYS°OBJ:*, DET:def 
(2) (PRED-ENABLE:alIow, MOD:(TENSE:past, AF: 
TOT.NEG), AG-QUANT :budget(D3) , OBJ-EVENT: 
(PRED-ACT:start, AG-HUM:we, OBJ-MENT.ACT: 
project(D#))) 
--->(PRED-ACT:start, MOD:(TENSE:past, AF: 
TOT-NEG, MOOD:can*), AG-HUM:we, 
OBJ-MENT. ACT:projeet(D~), CAUSE-QUANT 
:budget(D3)) 
where D 3 :=OBJ-QUANT : *, POSS-HUM:we, 
PRED-ATTR:Iimitted, 
Di :=OBJ-MENT. ACT: *, DET:indef, 
PRED-ATTR:new 
Beside the above type of case transformations 
@here are miscellaneous types of them. A kind 
of verbs such as ~bring ~ requires reference of 
the category of a term in the object case as 
well as the one about the recipient or goal 
case. 
A sentence containing a main verb such as 
Cforce~ is transformed to a more preferable form 
in Japanese which has a new manner case filled 
with a term like ' inevitably ~ instead of the 
main verb. 
A sententlal form which includes the verb 
~have ~ as a main verb is transformed to a more 
Table 6 Miscellaneous case transformations 
Category of 
predicate 
PTRANS 
MTRANS 
CAUSE 
POSSession 
Instances of 
predicate 
bring 
take 
bring 
take 
force 
compel 
have 
Referred 
case category 
GOal- 
PHYS. LOC 
EXPER-HUMO 
MENT. SUBJ 
POSS-HUM, 
OBJ-HUM U 
NATURal- 
phenomenon 
Change or 
addition of 
case 
ADDition of a 
MANNer case 
with 
~inevitably* ~ 
CHANGE 
from POSS 
to EXHIBition 
Modification 
of predicate 
REPL by 
~arrlve ~ 
REPL by 
~attain *~ 
DELetion 
REPL by 
~exist*' 
--452 - 
preferable form in Japanese even in the case the 
POSSessor is a HUMan unless the expression is 
like that "a living POSSessor has a PHYSical OBJ- 
ect in his hand~ Some examples are shown in 
Table 6. 
ExamPle 6 
By case transformation shown in Table 6, the 
following three English sentences 
(i) eA few minutes walk brought me the place ~ 
(2) ~The fact led him the conclusion' 
(3) ~We had heavy snowfall last year ~ 
are translated into Japanese sentences corres- 
ponding to 
(f) ~By walking for a few minutes, we arrived 
at the place' 
(~) ~From the fact, he attained the conclusion' 
(~) 'There was heavy snowfall last year' 
respectively. 
In sentence (i), the agent with the category 
PTRANS action is verbalized in construction of 
the case structure, and renamed as the case 
CMEANS' by the case transformation. 
4.2 Case transformation for target-clause 
ordering 
This section describes a case transformation by 
which the order of clauses of a target sentence 
corresponds to that of a source sentence to some 
extent in the locative or temporal sense. 
In general, an English sentence is translated 
into a Japanese sentence so that dependency 
relations in surface structure such as those of 
subordinate clauses to the principal clause do 
not change so much for preserving the stream of 
the main subjects included in successive English 
sentences. 
However, there are several exceptions. One 
of them has a comparatively definite type in 
sentential forms as well as in translation. A 
certain kind of English sentences do not corres- 
pond tonatural Japanese sentences unless the 
relation of a subordinate clause to the princi- 
pal one is reversed. For example, if a 
resultant state is contained as a dependent in 
an English sentence which describes an action or 
a state, each of them is generally described in 
Japanese as a dependent in a sentence which 
describes the result as shown in Example 7. 
Example 7 
Two English sentences 
tHe awoke to find himself in a strange room' 
and 
~The problem was so difficult that no man could 
solve it ~ 
are translated to Japanese sentences correspon- 
ding to 
~When he awoke, he found himself in a strange 
room' 
and 
~Because the problem was so difficult, no man 
could solve it' 
respectively. 
The first of the above translation is done by 
the case transformation; 
(PRED-PERCEP.ACT:awake, EXPER-HUM:he, RESULT- 
EVENT:(PRED-PERCEP, ACT:find, EXPER-HUM:he, 
OBJ-HUM:hlmself, LOC-PHYS°LOC:in-room(D ))) 
---> (PRED-PERCEP-ACT:find, EXPER-HUM:he, 
OBJ-HUM:himself, LOC-PHYS.LOC:in-room 
(D), TIME-INSTANT:when(PRED-PERCEP,ACT: 
awake, EXPER-HUM:he)) 
where D :=OBJ-PHYS.LOC: *, DET:indef, 
PRED-ATTR:strange. 
The Other exceptional translations do not 
have definite type, but are desirable for 
readability of translated sentences when Eng- 
lish sentences have some long adjective or ad- 
verbial subordinate clauses or phrases after 
their rather short principal clause. These 
subordinate clauses have generally more concrete 
and essential meaning than the principal clause. 
In such cases, if translation is done in the 
ordinary way so that the relation of subordinate 
clauses to the principal clause remains unchan- 
ged, the generated Japanese sentence tends to 
have a long essential subordinate part followed 
by a short and rather introductory part of the 
principal clause. The Japanese sentences thus 
obtained will often bring about difficulties in 
immediate comprehension. Furthermore, when a 
part of subordinate clauses is referred in the 
next sentence, the immediate identification of 
the anaphora will not be so easy owing to 
increase of distance from the referring part as 
shown in Fig. 3. 
In such cases a continuative translation is 
more preferable. 6 It assigns a location succ- 
eeding to the principal clause for an equivalent 
of the subordinate clause. The subordinate cla- 
uses are classified to adverbial clauses and 
adjective ones. 
In the former case, the subordinate clause 
is qualified as a sentence or a coordinate cla- 
use in a continuative translation. For a coor- 
dinate clause, the coordinate conjunction is 
preceded which is determined by the ease label 
of the adverbial clause and more precisely by 
the preposition or the subordinate conjunction. 
In the latter case, syntactic roles of a pri- 
ncipal clause and a relative clause are reversed 
through a modified term as follows; 
s % (K I :t I ,Kz:t 2, • • .,Ka:t~ (K I :sl , • • • ,K£: ~., 
I 
.... K~:s~ )) 
I --~ (KL:t ~ (K~:tl, Kz:tz,...,, K~: * )~ 
= KI:sI , • • , K~:s~ ) 5) 
where the double-underlined case means that a 
thematic fronting and addition of a certain 
thematic postposition is done for the underli- 
ned case in generation of a Japanese sentence 
--453-- 
I l PRE0 I 
principal clause 
referred 
subordinate clause the next sentence 
(a) a source sentence 
referred 
I 
parts of subordinate principal 
principal clause clause 
clause 
(b) a Japanese sentence 
the next sentence 
Fig. 3 Order of subordinate and principal clauses 
from the transformed case structure. 
Example 8 
Two English sentences 
'Science deals with sense impressions which are 
similar to make universal agreement possible' 
and 
'That all men are all equal describes a proposi- 
tion to which no sane human being has ever been 
given his assent' 
are translated into Japanese sentences corres- 
ponding to 
'Sense impressions which science deals with are 
similar to make universal agreement possible' 
and 
'To a proposition which describes that all men 
are all equal no sane human being has ever been 
given his assent' 
respectively. 
A sentence which contains a kind of appositi- 
ons is also transformed as follows; 
(K1:tl,.,., KL:it, • •., Kk:tk, 
K~:that(K~:sl, . . , K~:sz )) 
---> (PRED:that(K~:st, . . . , K~:sb ), 
OBJ:it(K I :t I ,...,KL:*, • • , K~:t~)) (6) 
where K~ is an apposition case to a case KL. 
Example 9 
An English sentence 
'We should make it clear that they lead the 
same conclusion' 
is translated to a Japanese sentence correspond- 
ing to 
'That we should make clear is that they lead 
the same conclusion'. 
5. Conclusion 
The ease structures obtained by parsing or 
the transformed case structure are expanded in 
an appropriate term-order by a Japanese grammar 
under various control of the involved paramet- 
ers. 
Some experimental results show that this 
approach is effective to dissolve multi-syntac- 
tic and semantic category problems of words and 
phrases and efficient to transform case struc- 
tures. 
From the view point of more precise and 
multi-lingual translation, it is desirable that 
more precise, detailed and standardized defini- 
tion is established for case structures and 
hierarchic category systems and also their codes 
are attached to some word dictionaries. It will 
not be impossible as far as mechanical trans- 
lation of technical papers is concerned. 

References 

i. Wilks,Y.~An artificial intelligence approach 
to machine translation', Schank and Colby 
(eds.), Computer models of thought and lan- 
guage, San Francisco, Freeman,p.l14,(1973). 

2. Vauquois,B.,'Aspect of mechanical translation 
in 1979', GETA,Unlv.Grenoble, (1979). 

3. Hornby,A.S.,'Guide to patterns and usage in 
English',second edition,London, Oxford Uni- 
versity Press, (1975). 

4. Bruce,B.,'Case systems for natural language', 
Artificial Intelligence,Vol.6,p.327,(1975). 

5. Kunihiro,T.,'Structural semantics', 
Sansei-do,(1976), (written in Japanese). 

6. Bekku,S.,'Study of translation', 
Yashio-shuppan-sha,(1976),(written in 
Japanese). 
