Cooperation between Transfer and Analysis 
in Example-Based Framework 
Osamu FURUSE and Hitoshi IIDA 
ATR Interpreting Telephony Research Laboratories 
2-2, Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan 
e-mail: furuse {iida} %atr-la.atr.co.jp@uunet.uu.net 
Abstract 
Transfer-Driven Machine Translation (TDMT) is 
presented as a method which drives the translation 
processes according to the nature of the input. In 
TDMT, transfer knowledge is the central knowledge of 
translation, and various kinds aml levels of knowledge 
are cooperatively applied to input sentences. TDMT 
effectively utilizes an example-based framework for 
transfer and analysis knowledge. A consistent 
framework of examples makes the cooperation between 
transfer and analysis effective, and efficient translation 
is achieved. The TDMT prototype system, which 
translates Japanese spoken dialogs into English, has 
shown great promise. 
1 Introduction 
Many applications dealing with spoken-language, 
such as automatic telephone interpreting system, need 
efficient and robust processing. The system must be 
capable of handling many idiomatic expressions and 
spoken-language-specific expressions which deviate 
from conventional grammar. 
Also, spoken-hmguage has both easy and difficult 
expressions to translate. In human translation when 
translating an easy sentence, tile translated result is 
produced quickly using only surface-level knowledge. 
When translating a complex sentence, a more elaborate 
process is performed, using syntactic, semantic, and 
contextual knowledge. Thus, many strategies and 
various levels of knowledge are used to effectively 
translate spoken-language. 
This paper proposes a method called Transfer-Driven 
Machine Translation (TDMT), which carries out 
efficient translation processing by driving the necessary 
translation processes according to the nature of the 
input sentence. An example-based framework can 
achieve quick processing and consistently describe 
knowledge. The integration of transfer and analysis in 
an example-based framework is i)roposed as a method 
for achieving TDMT. In this method, transfer and 
analysis proceed autonomously and cooperatively. Also, 
a well-balanced load on each process can be achieved by 
employing this integrated processing mechanism. 
Sectiou 2 explains the idea of TDMT. Section 3 
explains distance calculation and transfer in an example- 
based framework. Section 4 explains analysis in an 
example-based framework. Section 5 reports on the 
TDMT prototype system, and Section 6 reports on tile 
experimental results. 
The explanations in the following sections use 
Japanese-to-English translation. 
2 Transfer-Driven Machine Translation 
TD/vlT performs efficient and robust spoken-language 
translation using various kinds of strategies to be able 
to treat diverse input. Its characteristics are explained in 
the following sub-sections. 
2.1 Transfer-centered cooperation mechanism 
Translation is essentially converting a source 
language expression into a target language expression. 
In TDMT, transfer knowledge consists of various levels 
of bilingual information. It is the primary knowledge 
used to solve translation problems. The transfer module 
retrieves the necess~u-y transfer knowledge ranging from 
global unit like sentence structures to local unit like 
words. The retrieval and application of transfer 
knowledge are flexibly controlled depending on the 
knowledge necessary to translate the input. Basically, 
translation is performed by using transfer knowledge. A 
transfer module utilizes analysis knowledge 
(syntactic/szmantic infonnation) which helps to apply 
transfer L,~owledge to some part of the input. And 
generation and context knowledge are utilized for 
producing correct translatiou result. In other words, 
TDMT prodaces translation results by utilizing these 
dift'cK,~ut kinds of knowledge cooperatively and by 
centering on transfer, and achieves efficient translation 
according to the nature of the input. 
ACTES DE COL1NG-92, NAN'rES, 23-28 ^otrr 1992 6 4 5 l'~oc, oF COLING-92, N^rcrl~s, AU . 23-28, 1992 
2.2 Utilization of example-based framework 
Transfer knowledge is the basic data which is used for 
totally controlling the translation process. 
Most of the transfer knowledge in TDMT is described 
by the example-based framework. An example-based 
framework is useful for cortsistenfly describing transfer 
knowledge. The essence of the example-based 
framework is the distance calculation. This framework 
achieves the best-match based on the distance between 
the input and provided examples, and selects the most 
plausible target expression from many candidates. The 
distance is calculated quickly because of its simple 
mechanism. Through providing examples, various 
kinds and levels of knowledge can be described in the 
example-based franlework. 
2.3 Multi-level knowledge 
TDMT provides multi-level transfer knowledge, 
which correspoods to each translation strategy. In the 
transfer knowledge of the TDMT prototype system, 
there is string-, pattern- and grammar-level knowledge. 
TDMT achieves efficient translation by utilizing multi- 
level knowledge effectively according to the nature of 
input. 
Some conventional machine translation systems also 
provide multiple levels of transfer knowledge for 
idioms, syntax, semantics, and so on, and try to apply 
these levels of that knowledge in a fixed order to cover 
diverse input \[Ikehara et al. 87\]. However, this method 
proceeds with the analysis lot deciding which level of 
knowledge should be applied for any given input 
sentence in a fixed order, placing heavy load on the 
analysis module. Also, the knowledge description is 
ratber more complicated than that of the example-based 
framework. Therefore, the lrauslation of a simple 
sentence is not always quick because the system tries to 
cover all translation strategies. 
3 Example-based Transfer 
TDMT utilizes distance calculation to determine the 
most plausible target expression and structure in 
transfer. 
3.1 Word distance 
We adopt the distance calculation method of Example- 
Based Machine Translation (EBMT) \[Sumita and lida 
91\]. The distance between words is defined as the 
closeness of semantic attributes in a thesaurus. Words 
have certain thesaurus codes, which correspond to 
particular semantic attributes. The distance between the 
semantic attributes is determined according to the 
relationship of their positions in the hierarchy of the 
thesaurus, and varies between 0 and 1 (Fig. 1). The 
distance between semantic attributes A and B is 
expressed as d(A, B). Provided that the words X and Y 
have the semantic attribute A and B, respectively, the 
distance between X and Y, d(X, Y), is equal to d(A, B). 
d(A,D~ 
Figure 1 Distance between thesaurus codes 
The hierarchy of the thesaurus that we use is in 
accordance with the thesaurus of everyday Japanese 
\[Ohno and Hamanishi 84\], and consists of four layers. 
when two values can be abstracted in the k-th layer 
from the bottom, the distance k/3 (0 -< k _< 3) is 
assigned. The value 0 means that two codes belong to 
exactly the same category, and 1 means that they are 
unrelated. The attributes "writing" and "book" are 
abstracted by the immediate upper attribute "document" 
and the distance is given as 1/3. Thus, the word 
"ronbun{technical paper}" which has thesaurus code 
"writing", and "yokoushuu{proceedings}" which has 
the thesaurus code "book", are assigned a distance of 
1/3. 
3.2 Description of Transfer Knowledge 
Transfer knowledge describes the correspondence 
between source language expressions (SE) and target 
language expressions (TE) in certain meaningful units, 
preserving the translational equivalence \[Tsujii and 
Fujila 91\]. The condition under which a TE is chosen 
as a translation result of an SE is associated with the 
TE. Transfer knowledge in an example-based 
framework is described as follows: 
SE => TEl (Ell, E12,...), 
TEn (Enl, En2,...) 
Each TE has several examples as conditions. Eij means 
the j-th example of TEi. The input is the SE's 
environment, and the most appropriate TE is selected 
according to the calculated distance between the input 
and the examples. The input and examples comprise a 
set of words. 
Let us suppose that an input I and each example Eij 
consist of t elements as follows: 
Au~s DE COIANG-92, NANTES, 23-28 ^otn" 1992 6 4 6 PROC. OF COLING-92, NANTES, AUG. 23-28, 1992 
I : (I1,...,It) 
Eij = (Eijl,...,Eijt) 
Then the distance between I and Eij is calculated as 
follows: 
d (I, Eij) = d ((I1,...,It), (Eijl,...,Eijt)) 
t 
= Z d (Ik, Eijk)*Wk 
k=l 
The attribute weight Wk expresses fire importance of 
the k-th element in the translation 1. The distance 
from the input is calculated for all examples. Then the 
example whose distance to the input is least, is detected 
and the TE which has the example is selected. When Eij 
is close,st to I, TEl is selected as file most plausible TE. 
The enrichment of examples increases the accuracy of 
determining the TE because conditions become more 
detailed. Further, even if there is only one TE, but there 
is no example close to the input, the application of the 
tc, msfer knowledge is rejected. 
3.3 Wide application of distance calculation 
Distance calculation is usexl to deternfine which TE 
has the example that is clo~st to the input, ,'rod can be 
used in various abstract level expressions tlepending on 
how the input words are provided. 
Various levels of knowledge can be provided by the 
wide application of distance calculation. TDMT 
achieves efficient translation by utilizing multi-level 
knowledge effectively. 
In the transfer knowledge of Ihe TDMT prototype 
system, the string-, pattern- and grammar-level 
knowledge, the latter two of which can be described 
easily in an example-based framework, are now adopted. 
String-level knowledge is the most concrete, while 
grammar-level knowledge is fire most abstract. 
3.3.1 String-level transfer knowledge 
Since this kind of knowledge has a condition outside 
the SE, the cooperation with such as context module is 
sometimes necessary. 
In some cases rite conditions can be dcseribed by the 
examples of the most closely related word in which the 
SE is used, as follows: 
sochira => this ((des'u {be}2)...), 
you ((okuru {send})..), 
it ((mira {see})...) 
1 Wk is given for each lk by TE's distribution that sematic 
attribute of Ik brings \[Sumita and lida 91\]. 
2 { w l'""Wn } is the list of corresponding English words. 
Applying this knowledge,"you" is selected as the 
word correspondiag to the "scehira" in 
"sochira ni{particle} tsutaeru" because of the small 
distauee between "tsutaeru{convey}" and "okum{send} ". 
3.3.2 Pattern-level transfer knowledge 
Pattermlevel transler knowledge has variables. The 
binding words of the variables are regarded as input. 
For example, "X o o-negaishimasu" {X particle will- 
ask-for} has a variable. Suppose that it is translated into 
two kinds of English expressions in the example 
lyclow: 
X oo-negaishimasu => 
may 1 speak to X '3 
please give me X' 
((jimukyoku{office}), ...), 
((hangou(number}),...) 
In the translation of "X o o-negaishimasu", the TE is 
detennined by the calculation below: 
if Min (d((X), (jimukyoku)),....) 
< Min (d((X), (bangou)),....) 
then tile TE is "may I speak to X' 
else the TE is "pl-ease give me X' " 
Tire following two sentences have the pattern "X o o- 
negaishimasu": 
(1) "jinjika{personnel section} o o-negaishimasu." 
(2) "daimei {title} o o-negaishimasu." 
The first sentence select,; "may I speak to X' " because 
(jinjika) is close to (jimukyoku). The second sentence 
selects "please give me X' " ,because (tlaimei) is close 
to (bangou). Thus, we get the following translations: 
(1') "may I speak to the l)ersnnuel section." 
(2') "please give me file title," 
3.3.3 Gramntar-level transfer knowledge 
Grammar-level transfer knowledge is expressed in 
terms of grammatical categories. The examples consist 
of sets of words which are concrete instances of each 
category. The following transfer knowledge involves 
sets of three common nouns (CNs): 
3A' is the transferred expression of A 
ACRES DE COLING-92, NANITS, 23-28 AO~t 1992 6 4 7 t'ROC. OF COL1NGO2. NANTES, AU6.23-28, 1992 
CNI CN2 CN3 => 
CN3' of CNI' 
(("kaigi, kaisai, kikan 
{confereltce, opening, time} "),...), 
CN2' CN3' for CNI' 
(("sanka, moushikomi, youshi 
{participation, application, form } "),...), 
This transfer knowledge allows the following 
translations. 
kenkyukai kaisai kikan {workshop, opening, time) 
o> file time of the workshop 
happyou moshikomi youshi 
{presentation, application, form} 
-> the application form for presentation 
The above translations select "CN3' of CNI' " and 
"CN2' CN3' for CNI' " as the most plausible TEs, as 
the result of distance calculations. 
3.4 Disambiguation by total distance 
When there are several ways to apply transfer 
knowledge to the input sentence, structural ambiguity 
may occur. In such cases, the most appropriate structure 
is selected on the basis of total distance. The least total 
distance implies that the chosen structure is the most 
suitable input structure. For example, when the pattern 
"X no Y" is applied to the sentence "kaigi no touroku 
hi no waribiki {conference, particle, registration, fee, 
tkar ticle, discount} ", there are two ppssible structures: 
(1) kaigi no (touroku hi no waribiki) 
(2) (kaigi no touroku hi) no waribiki 
The pattern "X no Y" has various TEs, such as in the 
following 
XnoY => Y'ofX' (Ell, El2,...), 
Y' for X' (E21, E22, ... ), 
Y'atX' (E3I, E32 .... ), 
X' Y' (Eel, E42 .... ), 
The respective TE tree representations constracted from 
structures (1) and (2) are shown in Figs. 2 and 3. 
The structure of (1) transfers to "Y' of X' " with the 
distance value of 0.50 and "Y' of X' " with the distance 
value of 0.17, and generates (1') with a total distance 
value of 0.67. In structure (2), "Y' of X' " with the 
distance value of 0.17 and "Y' for X'" with the distance 
value of 0.17, generates (2') with a total distance value 
of 0.34. The latter result is selected because it has the 
least total distance va~ue. 
(1') "discount of the regiswation fee of the conference" 
(2') "discount of registration fee for the conference" 
discount of registration fee of the conference 
(total distance-0.67) 
Y' of X' (distanee~0.50) 
t- 
X' 
the conLrence I Y'of X' (distance=0.17) 
X' Y' I I 
registration fee discount 
Figure 2 Translation of 
"kaigi no (touroku hi no waribiki)" 
discount of registration fee for the conference 
(total distance-0.34) 
Y' of X' (distancn=O. 17) 
X' Y' I 
discount 
Y' for X' (distance~O.17) 
W ¥' I I 
the conference registration fee 
Figure 3 Translation of 
"(kaigi no touroku hi) no waribiki" 
4 Example-based Analysis 
For some structurally complex sentences, translations 
cannot be performed by applying only transfer 
knowledge. In such cases, analysis knowledge is also 
required. The analysis module applies analysis 
knowledge and supplies the resulting information to the 
transfer module, which then applies transfer knowledge 
on the basis of that information. When no analysis 
knowledge is necessary for translation, the application 
of only transfer knowledge produces the translation 
result. The analysis described in this paper is not the 
understanding of structure and meaning on the basis of a 
parsing of the input sentence according to grammar 
rules, but rather the extraction of the information 
ACq'ES DE COLING-92, NANfES. 23-28 AOfrr 1992 6 4 8 PROC. OF COLING-92. N^.,wrEs, AUG. 23-28, 1992 
required to apply transfer knowledge and to produce the 
correct translation from the input sentence. 
4.1 Description of analysis knowledge 
Analysis knowledge is described by examples in the 
same way as transfer knowledge, as follows: 
SE => RevisedSEl (Ell, El2 .... ), 
Revised SEn (Enl, En2 .... ) 
Although the form of knowledge description is virtually 
the same, transfer knowledge descriptions map onto 
TEs, whereas analysis knowledge descriptions map onto 
revised SEs. 
4.2 Cooperation mechanism 
The transfer and analysis processes operate 
autonomously but cooperatively to produce the 
translation result shown in Figure 4. 
Analysis 
application of 
analysis 
kn0wlege 
Input 
f Transfer \] 
\[application of 
information| transfer 1. knowlege 
1 output 
Figure 4 Relation between transfer and analysis 
At present, we are providing analysis knowledge for 
normalization \[Nagao 84\] and for structuring with 
TDMT. In the following sections we will explain the 
cooperation mechanism between transfer and analysis 
based on these two kinds of analysis knowledge. 
4.2.1 Analysis knowledge for normalization 
Normalization is putting together minor colloquial 
expressions into standard expressions It leads to robust 
translation and efficient knowledge storage. Analysis 
knowledge for normalization is utilized to recover the 
ellipsis of function words such as particles, and to 
normalize some variant forms such as sentence-final 
forms into normal forms. Such knowledge helps the 
application of transfer knowledge to the input sentence. 
The sentence "Watakushi wa Suzuki desu {I, particle, 
Suzuki, complementizer}" is uanslated into " 1 am 
Suzuki" by applying transfer knowledge such as the 
following: 
XwaYdesu =>X'beY' 
However, in spoken Japanese, particles are frequently 
omitted. The sentence "Watakushi Suzuki desu" is 
natural spoken-Japanese. It is normalized to "Wataknshi 
wa Suzuki desu", which has the omitted particle "wa" 
recovered, hy applying the following analysis 
knowledge: 
Pronoun Proper-Noun => 
Pronoun wa Proper-Noun (a set of examples) 
The analysis module sends the information about tile 
application of the analysis knowledge to the transfer 
module. The transfer module receives the information 
and applies the transfer knowledge to produce the 
English sentence " I am Suzuki" 
By examples, tbis kind of analysis knowledge cau 
also classify the particles to be recovered as shown 
below: 
CN Verb => 
CN o Verb 
((hotem {hotel}, yoyaku-snra {reserve\]), ,..), 
CN ni Verb 
((kaigi{confemnce}, sanka-suru {i)articipate}),...), 
This analysis knowledge allows the recovery of various 
particles such as, 
"hoteru yoyaku-suru" -> "hotern o yoyaku-suru" 
"kaigi sanka-suru"-> "kaigi ni sanka-suru" 
Analysis knowledge for nomlalization also has the 
advantage of making file scale of knowledge more 
economical and the translation processing more robust. 
4.2.2 Analysis knowledge for structuring 
Structuring is recognition of structure components of 
by insertion of a marker in order to apply transfer 
knowledge to each structure component. Analysis 
knowledge for structuring is applied to detect special 
linguistic phenomena such as adnominal expressions, 
wh-expressions, and di~ontinuities, so as to assign a 
structure to the SE. 
Adnominal expressions appear with high frequency in 
Japanese, corresponding to various English expressions 
such as relative clauses, infinitives, pronouns, gerunds, 
and subordinate clauses. They can be detected by means 
of inflectional forms. Three components of adnominal 
expressions must be considered in the translation 
process: the modification relationship, the modifier, and 
AcrEs DE COLING-92, NANTES. 23-28 ho(rr 1992 6 4 9 PROC. OF COLING-92, NANTes, Ate. 23-28, 1992 
the modified. Analysis information for structuring is 
used to insert a marker at the boundary between the 
modifier and the modified. The following analysis 
knowledge can be constructed. 
Adnominal-inflection CN => 
Adnominal-inflection Adnominal-marker CN 
(a set of examples) 
This knowledge identifies adnominal relationships and 
separates the modifier from the modified so that transfer 
knowledge can be applied. When the transfer module 
receives the information about the application of this 
analysis knowledge, it applies the transfer knowledge 
needed to translate each component of the expression: 
the adnominal relationship, the modifier, and the 
modified. The scope of the modifier and the modified is 
determined by the total distance of each structure in 
which transfer knowledge is applied. 
The following transfer knowledge about the 
adnominal relation determines the English expression 
by distance calculation with examples before and after 
the marker as follows: 
XAdnominal-mark Y => 
Y' that X' ((iku{go} , basu{bas} ), ...), 
Y' when X' ((deruIaueod} , hi{day} ), ...), 
For example, analysis knowledge is applied to "Kyoto 
eki e iku basu{Kyoto station particle go bus}", and the 
revised SE "Kyoto eki e iku Adnominal-marker basu" is 
produced. Then, by the application of the above transfer 
knowledge about the adnominal relation and the 
following transfer knowledge about the modifier and 
modified, the translation result "the bus that goes to the 
Kyoto station" is produced. 
XeY => Y'toX', 
Kyoto eki => Kyoto station, 
iku => go, basu => bus 
5 TDMT Prototype System 
A prototype Japanese to English system constructed 
too confirm the feasibility and effectiveness of TDMT is 
running on a Genera 8.1 LISP machine \[Furuse and lida 
92\]. 
Due to the restriction of the sequential mechanism, a 
method for driving the necessary process at the required 
time has not been completely achieved. However, the 
following control mechanism is used to obtain the 
most efficient processing possible. 
• As much as possible, translation is attempted by first 
applying only transfer knowledge; when this fails, the 
system tries to apply analysis knowledge. 
• Transfer knowledge is applied at the most concrete 
level as possible, that is, in the order of string, pattern, 
and grammar level. 
In order to achieve flexible processing which 
exchanges necessary translation information, a parallel 
implementation is under study based on the results from 
the prototype system. 
The knowledge base has been built from statistical 
investigation of the bilingual corpus, whose domain is 
inquiries concerning international conference 
registration. The corpus has syntactic correspondences 
between Japanese and English. We have established 
transfer and analysis knowledge as follows: 
• string-level transfer knowledge (about 500 items) 
• pattern-level transfer knowledge (about 300 items) 
• grammar-level uansfer knowledge (about 20 items) 
• analysis knowledge (about 50 items) 
6 Evaluation 
We have evaluated the TDMT prototype system, with 
the model conversations about conference registration 
consisting of 10 dialogs and 225 sentences. The model 
conversations cover basic expressions. Table 1 shows 
the kinds of knowledge that were required to translate 
the model conversations. 
Table 1 Knowledge Necessary to Translate 
Model Conversation 
( total number of sentences - 225) 
sentences rate 
string only 73 32.4% 
pattern and string only 90 40.0% 
grammar-level 21 9.3% 
transfer knowledge needed 
analysis knowledge needed 41 18.2% 
At present, the prototype system can produce output 
quickly by the example-basod framework. 
200 of the sentences are correct, providing a success 
rate of 88.9%. The coverage by string- and paaem-level 
knowledge is wider than expected. 
Table 2 shows the main causes of incorrect sentences. 
ACT~ DE COLING-92, NANTES, 23-28 AOt~r 1992 6 5 0 PROC. OF COL1NG-92. NANTES, AUG. 23-28, 1992 
Table 2 Causes of Incorrect Sentences 
(total number of incorrect sentences - 25) 
oct urrence~ 
(1) inability to get such TEs 9 
as elided objects 
(2) selection of incorrect TEs 8 
(3) error in adverb position 4 
(4) incorrect declension 1 
(5) incorrect tense 1 
(6) etc 2 
The second factor shows that an elaboration of 
distance calculation and an enrichment of examples are 
needed. The first, third, and fourth factors are caused by 
the shortage of generation knowledge. The fifth factor is 
caused by the shortage of analysis knowledge. These 
facts show that the cooperative control that flexibly 
communicates various kinds of knowledge including 
context mid generation knowledge, and various kinds of 
frameworks such as a rule-based and a statistical 
framework are useful to improve the translation 
performance. 
7 Related Research 
The example-based approach was advocated by Nagao 
\[Nagao 84\]. The essence of this approach is (a) retrieval 
of similar examples from a bilingual database and (b) 
applying the examples to translate the input. Other 
research has emerged following this line, including 
EBMT \[Sumita and Iida 91\], MBT \[Sate and Nagao 
90\], and ABMT \[Sadler 89\]. EBMT uses phrase 
examples and will be integrated with conventional rule- 
based machine translation. MBT and ABMT use 
example dependency trees of examples and translate the 
whole sentence by matching expressions and by a left- 
to-right search of maximal matching. TDMT utilizes an 
example-based framework for various process as the 
method of selecting the most suitable TE, and 
combines multi-level transfer knowledge. On the other 
lmnd, MBT and ABMT utilize uni-level knowledge only 
h~r transfer. 
8 Concluding Remarks 
TDMT (Transfer-Driven Machine Translation) has 
been proposed. The prototype TDMT system which 
translates Japanese to English spoken dialogs, has been 
constructed with an example-based framework. The 
consistent description by example smoothes the 
cooperation between transfer and analysis, have shown 
the high feasibility. Important future work will include 
the achievement of flexible translation which effectively 
control the translation process. Also important is the 
implementation of TDMT in distributed cooperative 
processing by a parallel computer and incorporating 
various kinds of processing such as rule-based and 
statistical framework into the cooperation mechanism. 
Acknowledgements 
1 would like to thank the members of the ATR 
Interpreting Telephony Research Laboratories for their 
comments on various parts of this research. Special 
thanks are due to Dr. Kohei Habara, the chairman of the 
board of ATR Interpreting Telephony Research 
Laboratories. Dr. Akira Kurematsu, the president of 
ATR Interpreting Telephony Research Laboratories, for 
their support of this research. 

References 

\[Furuse and lida 92\] Furuse, O., and Iida, H. : An 
Example-based Method for Transfer-driven Machine 
Translation, Prec. of the Fourth International 
Conference on Theoretical and Methodological Issues in 
Machine Translation, (1992). 

\[Ikehara et al. 87\] lkehara, S., Miyazaki, M., Shirai, 
S., and Hayashi, Y : Speaker's Recognition and Multi- 
level-Translating Method Based on It, Trans.IPS Japan 
Vol.28 No.12., IPSJ , pp.1269-1279, (1987), (in 
Japanese). 

\[Nagao 84\] Nagao, M. : A framework of a mechanical 
translation between Japanese and English by analogy 
principle, in Artificial and Human Intelligence, ed. 
Elithorn, A. and Bmrerji, R.. North-Holland , pp.173- 
180, (1984). 

\[Ohno mid ltamanishi 841 Ohno, S. and Hamanishi M. 
: Ruigo-Shin-Jiten, Kadokawa, (1984), (in Japanese). 

\[Sadler 89\] Sadler, V. : Working with Analogical 
Semantics, Foris Publications (1989). 

\[Sato and Nagao 90\] Sato, S. and Nagao M. : Toward 
Memory-Based Translation, Proc. of Coling '90, 
(1990). 

\[Smnita and lida 911 Sumita, E., and Iida, 1t. : 
Experiments and Prospects of Example-based Machine 
Translation, Proc. of the 291h Annual Meeting of the 
Association for Computational Linguistics, (1991). 

\[Tsujii and Fujita 91\] Tsujii, J. and Fnjita, K. : l,exical 
Transfer based on Bi~Lingual Signs -Towards 
Interaction during Transfer, :Proc. of the 5th 
Conference of the European Chapter of the Association 
for Computational Linguistics, (1991). 
