Interaction between Structural Changes in Machine Translation 
Satoshi KINOSHITA 
Research and Development Center 
Toshiba Corporation 
Komukai-Toshiba-cho 1, Saiwai-ku, 
Kawasaki, 210 Japan 
John I'HILLIPS Jun-ichi TSUJII 
Centre h)r Comt)utational Linguistics 
UMIST 
P.O.Box 88, Manchester, 
M60 1QD, U.K. 
Abstract 
This paper discusses coml)lex structural changes dur- 
ing transfer within a non-destructive transfer frame- 
work. Though the description of each individual 
structural change is not difficult, special provision 
must be made when they are combined, because in- 
teraction between them sometimes causes unexpected 
problems. Transfer of coordinate structures is also 
discussed a.q this sometimes necessitates a structural 
change and interacts with other structural changes in 
a problematic way. 
1 Introduction 
Several focal issues have emerged in machine trans 
lation (MT) research as the result of recent intensive 
studies in the field. That is, 
• Declarative Frameworks for expressing Bi-lingual 
Knowledge 
• Integration of Knowledge-based Processing and 
Contextual Processing with the Translation pro 
cess 
• Effective Exploitation of Domain/'l~xt Type- 
Specificity (or Sublauguageness) in MT and Dis- 
covery Processes for such Specifieities 
Though new frameworks of MT such as Statistics- 
based MT (SBMT), F, xample-bascd MT (F, BMT), 
Analogy-based MT (ABMT), Knowledge-based MT 
(KBMT) etc. look radically difihrent from con- 
ventional liuguisties-based MT (LIIMT) anch as 
Transfer-based MT, they address role or two of the 
above focal issues and ignore the rest\[7\]\[10\]. In par- 
ticular, the new paradigms of MT tend to ignore t.he 
first issue ie. declarative representation of bi-lingual 
knowledge and the complexities involved in the pro- 
cess of combining units of bi-liagual knowledge. 
It is our contention that any MT system, whichever 
paradigm it belongs to, has to store a set of 
translation-equivalent units for a pair of languages 
and combine these to produce larger units of trans- 
lation. In EBMT, for example, a set of translation 
examples has to be stored and several of them have 
to be combined properly to produce translation. Be- 
cause of the declarative nature of translation exam- 
pies, EBMT inevitably encounters the same complex- 
itics of combining translation units in declarative rep- 
resentation o.~ I,BMT does. 
tLescarch in LBMT\[1\]\[2\]\[3\]\[ll\]\[12\]\[13\] has revealed 
that dittieulties in tim declarative representation of 
bi-lingual knowledge stem mainly from the treatment 
of idiosyncratic structural changes caused by lexieal 
items, and interactions of such idiosyncratic struc- 
tural changes when they co-exist in single sentences. 
These type of structural changes also cause problems 
when they are combined with general or ordinary lin- 
guistic phenomena soeh as coordination. A formal 
framework to cope with these matters is essential in 
other aPl)roaches, such as EBMT, as in LBMT, if the 
translation is produced in a compositional way. 
In this paper, we discuss problems caused by such 
interaction, and give solutions in our logic-based 
transfer framework\[8\]. Our transfer model views the 
transfer process as deduction; it produces the target 
linguistic description without changing the source de- 
scription. This giw:s a clear contrast to the conven- 
tional tree-traasducer model, which gradually tranS- 
forms the source description, and is therefore hard to 
control. 
Because of the logic oriented nature of our frame- 
work, it can also be extended easily to cope with the 
other local issue of MT such as the integration of 
knowledge-based processing with translation\[4\]. 
2 Formalism 
We use a simple formalism, representing a sentence 
as a network of semantic dependencies. The basic 
units of the formMism are indices, properlies, and re- 
Istwna. A logical form consists of an unordered net 
of Ierms; each term is either a property predicated 
of an it~dex, or ~t relation between two indices. The 
written notation depicts properties and relations as 
unary and binary functors, respectively, and indices 
as theiL" arguments, i.e. within brackets. A logical 
form representing 'John saw Mary' might be 
e : john(j) & see(el & mary(n 0 & tense(e,past) & 
subj(e,j) & obj(e,m) 
This representation in simplified particularly in that 
Actxs DE COLING-92, NANTES, 23-28 Ao~r 1992 6 7 9 I'~oc. ov COLING-92, NANTES, AUG. 23-28, 1992 
the man who Mary saw: x : man(x) & mary(m) & see(el & t ..... (e,past) & subj(e,m) & obj(e,x) 
the man who saw Mary: x : man(x) & nrary(m) & see(el & tense(e,past) & subj(e,x) & obj(e,m) 
I managed to painl my house quickly: m : me(me) & manage(m) & tense(m,past) & subj(m,me) & 
obj(m,p) & paint(p) & obj(p,h) & of(h,me) & house(h) & qniek(q) & subj(q,p) 
Mary looked at the brightly painted house: l : mary(m) & look(l) & tense(l,past) & subj(l,m) & at(l,h) & 
bright(b) & subj(b,p) & paint(p) & obj(p,h) & house(h) 
Mary's house was painted: p : mary(m) & of(h,rn) & house(h) & paint(p) & tense(p,past) & obj(p,h) 
Figure 1: Examples of Logical Forms 
the relation tense here stands for what should be a 
complex representation of tense, aspect, and aktion- 
sart, related anaphorieally to the preceding discourse. 
It can be seen that the representation includes the 
names of the concepts, both objects and events, which 
are mentioned in the sentence, aud the semantic re- 
lationships between them. Each of the three indices, 
j, m and e, is a unique label given to a node in the 
dependency network. The indices also serve as dis- 
course referents when the logical form is incorporated 
into a discourse representation. The root node of the 
network is specified at the beginning of the represen- 
tation, in tiffs ease e, the node representing the seeing 
event. In terms of discourse representation theory, e 
is tim discourse referent of which the logical form is a 
description. As long as the root node of the network 
is specified, it is possible to convert a logical form 
mechanistically into a structured representation such 
as a dependency graph or a feature structure such as 
tense : past 
subject : \[predicate : john\] 
object : \[predicate : mary\] 
Some other examples of logical forms are shown 
in figure 1. The particular characteristics of the 
formalism -- the types of relations and predicates 
used and the manner of their use -- are justified 
elsewhere\[9\]. We only state here that tbe forrualism 
can treat the phenomena which are treated by tradi- 
tional formalisms, along with additional phenomena 
relating to discourse structure and lexical semantics. 
3 Transfer 
We follow the standard division of a machine trans- 
lation system into stages of analysis, transfer, and 
generation. The parts of the system include algo- 
rithms for analysis, transfer, and generation, descrip- 
tions of the strneture of individual languages (used 
by the analysis and generation algorithm.s), and de- 
scriptions of equivalence relationships between pairs 
of languages (used by the transfer algorithm). It 
is a requirement for the transfer description that it 
should state all and only the equivalence relation- 
ships between expressions of a pair of languages. It 
should contain no general linguistic information, el- 
tiler universal or specific to any single language: these 
types of information belong ill the formalism and con- 
tent, respectively, of the descriptions of individual 
languages. In fact most of our transfer descriptions 
consist of direct equivalences between predicates or 
groups of predicates of the language pair. 
A transfer rule shows a particular translation 
equivalence between two languages. It consists of 
three parts: a pair of logical forms, one for each 
language, representing tile equivalence, and a third 
logical form giving the conditions under which the 
equivalence is applicable. We call these the equiva~ 
lenee pair and the condition respectively. Two logical 
forms form an equivalence pair if the natural language 
expressions they represent have overlapping denota- 
tions (using 'denotation' in a very broad sense to en- 
compass not only the referential possibilities of nouns 
but the possibility in general of applying a predicate 
to a particular index). The rule can therefore be used 
for translating in either direction: the two logical 
forcers of the eqnivalenee pair are always translation- 
ally equivalent if the condition is satisfied. The logical 
forms of the equivalence pair will be indistinguishable 
from logical forms of the respective languages, using 
the same predicates and relations. The logical forms 
of the condition use meta-predieates which allow ref- 
erence to to the logical form of the current local trans- 
lation unit (sentence or clause) and to the linguistic 
context. In practice, most transfer rules have no con- 
dition; they consist just of an eqnivalenee pair. Some 
examples of rules follow: 
English Japanese Explanation 
table(X) teeburu(X) table = teeburu 
paper(P) ronb,m(p) paper = ronbun 
paper(P) sinbmt(P) paper = sinbun 
subj(E,X) ga(E,X ) subject marker 
obj(E,X) wo(E,X) object marker 
hypothetical(El,like(El -tai(F),omou(E) would like 
iff inLeft(obj(E,F)) ~tal omou 
The following is the basic definition of transfer be- 
tween two texts (sentences, clauses, or whatever): 
A source logical form and a target logical 
form are correctly transferred if tile terms of 
each can be divided into non-overlapping sub- 
sets such that the source subsets can be placed 
into one-to-one correspondence with the target 
subsets by each corresponding pair being 'uni- 
fied' with the two halves of the equivalence pair 
AcqEs DE COLING-92, NANTES, 23-28 AOt~T 1992 6 8 0 PROC. OF COLING-92, NAI~rEs, AUG. 23-28, 1992 
of a transfer rule, and if the accumulated con- 
ditions of the transfer rules used are satisfied. 
'Unification' in this definition is rattmr similar to 
graph unilication. 
q~ansfer rules specifying larger sets of terms can 
be used to state trauslational equivalences for idioms 
and fixed expressions. Simple idioms present no par- 
ticular problem. One translation of the English idiom 
to kick the bucket into Japanese is sinu 'die', and this 
can be represented by the transfer rule 
kick(K) $z. obj(K,n) & the-bucket(B) ~ sinu(K) 
Fixed expressions, appropriate only to a specific con- 
text, may require conditions which refer to the dis- 
course representation. 
4 'Complex' transfer 
4.1 Types of complex transfer 
Lindop & Tsujii (1991) list and discuss a variety of 
examples that are alwaYs problematic for machine 
translation systems. We suggest that what makes 
these examples difficult is that different dependency 
structures are associated with otherwise synonymous 
lexical items. We break the problem down into four 
subparts. 
i) Argument-switching as in the translation of the 
German 
Der Wagen gef'~llt mir -- I like the car 
The car pleasen me 
gefallen(E) & nora(O) & dat(S) 
like(E) & subj(S) & obj(O) 
In argument-switching, the relations between the lex- 
ical item and its dependents are not translated stan- 
dardly. Although the German nominative normally 
corresponds to an English subjcct, it must in this ex- 
ample be translated as a dative. 
it) IIead-switching as hi the translatlolt of Ger- 
nlan 
tlaas sehwimmt gem -- John likes swinmfing 
John swinm ~ladly 
like(L) & obj(L,S) 
geru(L) & subj(L,S) iyshift(L,S) 
The Germau sentence is a description of a situation 
to do with swimming; the English is a description of 
a situation to do with liking. The shift predicate is 
explained below. 
iii) Decomposition as in the translation of the 
Japanese 
John-ga jisatusita -.- John committed suicide 
jisatu_suru(E) 
commit(E) & suicide(S) & obj(E,S) 
llere the single Japanese word jisatusnrn is tra~l~- 
lated as the English phrase to commit suicide. Some 
types of decomposition can give rise to special prob- 
lems when there arc modifiers or dependents involved. 
These are discussed in detail by Tsujii et al. (1992), 
Shortage of space aml tile variety and complexity of 
examples prevents their discussion here. 
iv) ttaistng as in the translation of the Welsh 
Fe ddigwydd rod Si6n yma 
happells be ~l o|ln here 
John happens to he here 
digwydd(E1) ~ subj(E1,E2) & subj(E2,X) 
. -+ happen(E1) ,~ subj(E1,X) & obj(E1,E2) 
In the English, the subject of happen is John, but in 
the Welsh, the subject of digwydd ('happen') is the 
situation of John's being here. 
The predicate shift is introduced to define a trans- 
lation equivalence where certain dependencies are dif- 
ferently arranged in the target language, as compared 
to the source language. It can be understood in terms 
of semantic type-shifting - - the types are differently 
distributed in the source- and target-language logi- 
cal forms. Shifl(H,E) means that in any instance of 
subj(H,X) or l'(X, tt) (where V stands tbr auy rela~ 
tion) in the logical form representing the English text, 
the H corresponds to an E in the logical form repre- 
senting the other language. The following example 
shows how the example under (it) above would work 
in practice. 
lch welt\] Hans schwimmt gern - i know John likes 
swmuning 
ieh(i) me(i) --~en(wi know(w) 
sut,j(w,i) subj(w,i) 
ohj(w,s) obj(w,s)' 
Hans(j) John(j) 
sehwimmen(s) swim(s) s.bj(sj) suhj(iQ) 
gern(g) like(g) 
subj(g,s) obj(g,s) 
{shift(s,g)} 
me(i) 
know(w) 
subj(w,i) 
obj(w,g) John(j) 
swim(s)' 
subj(gd) 
like(g) 
obj(g,s) 
The columns of the table show tile German source 
logical form, the English logical form as it would be 
without shift, and tile actual English target logical 
form. The horizontal divisions of the table represent 
the fields of operation of separate transfer rules. 
4.2 Interaction between types of com- 
plex transfer 
When several of these syntactic changes are combined 
in one example, the logical form resulting from trans- 
fer is incoherent. The following examples are types 
on which most, if not all, existing machine translation 
Acq~s DE COLING-92, NAN-I~S, 23-28 Aour 1992 6 8 I PROC. OF COLING-92, NANTES, AUG. 23-28. 1992 
systems will fail, as will the mechanism proposed here 
so far. 
English - Welsh; raising with argument-switching: 
John happens to like swimming 
fe ddigwydd fod yn dda gan John nofio 
happens be nice by John swim 
(swimraing-bein~-nice-in-dohn's-opinion occurs) 
Dutch - English; verb - head-switching twice: 
Jan zwemt grnag toevallig 
John Bwims gladly happeningly 
John happens to like swimming 
(also 'John likes happening to swim~ 
There are two separate causes of difficulty in 
these examples. Firstly, the long-distance movement 
caused by raising canses problems in the transfer of 
structures which have been the target of raising. The 
simplest way to get rmmd the problem is to require 
syntactically-deleted arguments to be marked explic- 
itly on all the verbs of wbich they are logical ar- 
guments, treating the phenomenon syntactically as 
an example of lexieally-triggered extraction, on a par 
with the extractions of syntactic objects in a man I 
knowand an easy man to please. Transfer rules which 
introduce raising predicates will then have to bring in 
the new subject explicitly. For instance, the rule for 
happen and digwydd, (iv) in §4.1, will be re-written 
as 
digwydd(E1) & subj(E1,E2) 
happen(E1) & subj(E1,X) & obj(E1,E2) 
iff subj(E2,X) 
The second point is that the shift predicate must 
be defined in such a way that it can cope with re- 
cursive changes in the dependency structure brought 
about by the occurrence of several interdependent 
head-switching translations. It seems that shift can 
be made to do this simply by having e.g. shifl(H,E) 
affect all instances of snbj(tl, X) or P(X,H) (including 
shifl(X,H)) not transferred by the transfer rule which 
introduced the shift(ll, E). 
Together, these two stipulations enable the trans- 
fer of examples involving head-switching, argument- 
switching, raising, and most types of decomposition. 
5 Transfer of Coordinate 
Structures 
5.1 Problems in transfer of coordi- 
nate structures 
Coordination often reqnires exceptional treatment in 
parsing and generation of natural language. Transfer 
is no exception. Transferring a coordinated structure 
sometimes produces a miserable result due to the ac- 
companying combination of structural changes or the 
fact that coordinated linguistic objects require differ- 
ent target words for a shared source word. However, 
few attempts have been reported to formalize this 
problem. 
We here divide the problem into two categories: t 
• Divergence in semantic constraints 
• Combination of complex transfers 
The first type of problem occurs when semantic fea- 
tures of the coordinated source words require a source 
word to be translated into two different target words. 
A typical example can be seen in translation of the 
following English sentence into Japanese. 
(la) Site wears a bat and shoes. 
(lb) kanojo-ga boushi-wo kaburi, kutsu-wo 
she-suhj lmt-ohj wear shoe-obj 
haku. 
wear 
As is understood from its translation, "wear" is trans- 
lated "kaburu" or "haku" in Japanese, depending on 
whether its object is something worn on the head or 
on the foot(or leg). This means that, in this example, 
coordination of objects in English should be altered 
to that of verb phrases in Japanese. 
This type of knowledge for lexical choice is very 
common in a transfer or bi-lingual dictionary, and 
plays an essential role in lexical transfer of most cur- 
rent transfer-based MT systems. The problem is that 
ueither a transfer program or a transfer-rule writer 
expects such an awkward problem caused by coor- 
dination. To translate "wear" iuto "kaburu" in the 
above example, n rule writer may usually write the 
following rule in our notationS: 
wear(X) ~ kaburu(X) iff obj(X,Y)&HAT(Y) 
But the condition part of this rule is implicitly ex- 
pected to be interpreted as follows. 
wear(X) ~ kaburn(X) iff V Y obj(X,Y)&;HAT(Y) 
The precise definition may change depending on 
how the coordinate structure is represented. But the 
1 There is eamther type of problem wlfich is baaed on the syn- 
tactic differences between coordination constructlom in source 
arid target languages. For example, "2 or 3 potuld~" in En- 
glish should be trmmlated "2 pondo ka 3 pondo'(2 poundJ or 3 
pounds} in JaF.mlese and "dwybunt ncn clair" (2 pounds or 3) in 
Welsh. (The Welsh expression is used only for a sum of money. 
Another exists for weight.) This type of problem tlaould alto 
be solved in transfer, but we do not mention it here. 
2In this section, we coltsider translation whose source and 
target logical forms are on the left and right sides of a trantfer 
rule. For the sake of simplicity, trmasfer rules hereafter are 
described as uni-directional ones. 
AC'TES DE COLING-92. NANqT~. 23-28 Ao~r 1992 6 8 2 PROC. OF COLING-92, NANTES. AUG. 23-28. 1992 
point is that "wear" may be trmmlated "kaburn" only 
if all the objects have a feature of being a "HAT". 
A simple transfer algoritlnn, for example, may 
choose the target word when it finds the first ap- 
plicable transfer rule for a source word: this algo- 
rithm may produce "boushi-to kutsu-wo kaburu" for 
the sentence (la), which memm that tile semantic re- 
lation between "wear" and "shoes" is ignored. There 
may be another type of transfer algorithm which sim- 
ply fails beeanse it cannot provide two dilfcrent target 
words to one identical source word. 
The second type of the problem occurs when one 
of the coordinated objects triggers a complex transfer 
which is described in §3. This type of problem can 
bc seen in the following translation. 
lie connnitted nmrder then suicide. 
kare-ga satsujin-wo okaslfi, jisatsu-shita. 
he~subj murder-obj conmlit commit-ed suicide 
This problem is more complicated than the previous 
one because complex ti'ansfer, in this example "many- 
to-one transfer", causes a structural change. Our sim- 
ple transfer algorithm mentione<l in tile previous sec- 
tion may produce a disastrous result for this type of 
translation. 
There are several possible solntions to this problem. 
The simple.st one is that a transfer:rule writer writes 
all the transfer rules which explicitly describe every 
possible sub-structure with coordination. This is of 
course unrealistic. 
Another solution is to make a transfer program 
which modifies the transfer result dyuanfically when, 
tbr example, a source word is given two differeut tar: 
get words. But such a dynamic modilication of tile 
result during transfer is against our policy of logic: 
based transfer, because this means gradual transfor- 
mation of the source strueturc and therefore transfer 
cannot be formalized as logical inference. 
5.2 'lYansfer with Coordinate Expan- 
sioll 
5.2.1 C~xwdlnate Expansion 
tiereafter we concentrate on a case where coordinated 
objects of a verb cause the problem, though there 
is apparently an example where other eases such a.~ 
"snbj" cause the same problem. 
The basic reqnirement in logic-based transfer is 
that coordination of objects should he reorganized 
to that of verb phrases or sentences, which is not 
supposed to cause problems in transfer. We call this 
reorganization "coordinate expansion". 
The following is a part of logical form for (ta), 
which involves a coordinate structure. 
wear(w)&obj(w,o)&coord(o,ol)&hat(ol)& 
coord(o,o2)&shoe(o2) 
Input Logical Form 
U 
Expansi ..... \[ Expandruleu \] 
,'runsler +- \[ Transfer rules \] 
g 
Reduction 
Output Logical Form 
t,'igure 2: Transfer with Coordinate Expansion 
In this form, o is a linguistic object, and a predi- 
cate coord represents a relation between the linguistic 
object and its coustituents. The following is a r~ult 
of expansion. 
coord(w,w 1)&we;tr(w 1 )&obj(w l,n l)$zhat(ol)& 
coord(w,w2)&wear(w2)&obj(w2,o2)&shoe(o2) 
The most naive and simplest strategy usiug this 
expansion is to expand cvery coordination within a 
sentence and rel)reseut it in sentcnce coordination be- 
fore transfer. This transfer result will be reorganized 
again into an appropriate representation of coordina- 
tion in tim target language. But this solution seems 
inefficient from the computatioual point of view beo 
ettase caseswhere expansion is uecessary are rare . 
Unnecessary expansion and reorganization of coordi- 
nate structures should be avoided. 
The strategy wc propose executes coordination ex- 
pansion only if it is necessary\[5\]. Figure 2 shows a 
general view of our modified approach to transfer. 
Transfer is divided into three phases; in the first 
phase, logical forms are expanded if expand rnles (ex- 
plained below) find tim necessity of coordinate expan- 
sion. This process continues as long as the necessity 
remains. In tile second phase, transfer described in 
previous sections is executed. Fiually, in the third 
phase, coordilmtion is reorganized if the target lan- 
guage has a more appropriate structure for coordi- 
nation than tile second phase result. (Consider the 
translation of (lb). Without reorganizing eoordina~ 
ti(m, the transfer result will contain two "wear"s .) 
Tile following is all expand rule which detects the 
necessity of expansion concerned with translation of 
"wearing a bat". 
wear(X)&obj(X,Y)&coord(Y,Y1)&llAT(Y1)& 
coord(Y,Y2)&YleY2&~HAT(Y2) 
--- expand(Y,X) 
ht tile rule, e~:pand(Y,X) means that coordination 
of the level Y should be expanded to that of X. This 
rule suggests that coordinate expansion is necessary 
AClES DE COLING-92, NAI, CrEs, 23-28 AoI~r 1992 6 8 3 PROC. OF COLING-92, NANTES, AtnL 23-28, 1992 
if the object of "wear" is a coordinate structure, and 
one constituent is a HAT while another isn't. 
We assume that expand rules are produced auto- 
matically from transfer rules before the actual trans- 
fer operation. The simplest way of extracting such 
a rule refers only one transfer rule at one time. This 
means that the necessity of expansion can be detected 
not by creating expand rules but by referring trans- 
fer rules in actual transfer phase. But the former 
approach seems essential if we wish to optimize ex- 
pansion detection. 
5.2.3 Discussion 
Our coordinate expansion detection works even if a 
coordinate structure has more than two constituents. 
What wc have to consider is an appropriate expan- 
sion algorithm. For example, in translating (4a), 
an appropriate expansion should keep coordination 
of "shoes" and "stockings", as shown in (4b), he- 
cause both satisfy a semantic constraint on which the 
system chooses "haku" as the translation of "wear". 
Otherwise reorganizing a coordination in the genera- 
tion phase is inevitable. 
5.2.2 Other examples 
Expand rules from a transfer rule which involves a 
structural change are little different in forms to the 
previous ease. The following are a transfer rule for 
translating "commit suicide" into Japanese and its 
expand rule. 
commit(X)&obj(X,Y)&suicide(Y) 
--~ jisatsu..suru(X) 
commit(X)&obj(X,Y)&coord(Y,Y1)&suicide(Y1)& 
coord(Y,Y2)&Y l#Y2&--,suicide(Y2) 
expand(Y,X) 
Another example is the translation of the En- 
glish "HUMAN have ADJ NOUN" construction into 
Japanese. The sentence (2E) is translated (2J) using 
the rule (3). 
(2E) She has beautiful eyes. 
(2J) kanojo-no me-ga utsukushii. 
aho-poas eye-aubj beautiful 
(3) have(H)&subj(H,S)aobj(H,X)&mod(X,U) 
poss(X,S)&subj(M,X) 
iff ItUMAN(S)&PART_OF.BODY(X) 
(4a) She wears a hat, stockings and shoes. 
(4b) She wears a hat and wears stockings and shoes. 
Reorganization of a coordination in the target lan- 
guage does not only occur as outlined in the above 
case. Since the coordinate expansion is completely 
separate from the actual transfer process, transfer 
rules which do not cause problems might be used. 
There is still a problem to be solved with regard 
to tile transfer of coordination with "expansion"; ex- 
pansion is not always straightforward. There is of- 
ten a ease where coordinate expansion is impossible 
without deep understanding of the sentences, or it is 
impossible in that it may change their semantic struc- 
tures. For example, the sentence (5b) cannot be the 
expansion of (5a) though it seems so at first glance. 
(5a) I saw a group of men and women. 
(5b) I saw a group of men and I saw a group of women. 
The apparent disadvantage of our approach with 
"expand rules" is that a large number of expansion 
rules might be created. Though it provides an effi- 
cient way of detecting the necessity of expansion, it 
consumes a lot of memory, which will raise an imple- 
mentation problem. 
This case is more complicated than the previous 
ones because the transfer rule refers to two semantic 
features. Therefore we will get two expand rules, one 
of which is the following rule. 
have(X)&subj(X,Y)&coord(Y,Y1)&HUMAN(Y1) & 
coord(Y,Y2)&Y I#Y2&-,H UMAN(Y2) & 
obj(X,Z) & rood(Z,_) & PART-OF-BODY(Z) 
expand(Y,X) 
In addition, we need another expand rule which 
checks another type of coordinate construction. This 
rule will apply when the system translates "she has 
long hair and beautiful eyes". 
have(X) & suhj(X,S) & HUMAN(S) & obj(X,Y) 
8* -,mod(Y,_) g~ coord(Y,Yl) & rood(Y1,_) & 
PART_OF_BODY(Y1) 
---+ expand(Y,X) 
6 Conclusion and future work 
In this paper, we showed how complex structural 
changes in transfer are treated within our logic-based 
transfer mode\[, in which the target linguistic descrip- 
tion is obtained from tile source description in a non- 
destructive manner. These type of structural changes 
cause a problem when they co-occur and their trans- 
fer rules are interacted. We also discussed a problem 
in transferring coordinate structures and presented an 
extended transfer model with coordinate expansion. 
Some of our solutions to these problems are rather 
simple and need further investigation when we apply 
our framework to a practical system. We also have to 
evaluate our framework by examining a wider range 
of translation problems. 
One of our current concerns is to implement our 
transfer model in a parallel processing framework. 
Our current algorithm for transfer and its implemen- 
tation have a procedural aspect of operation. That 
ACRES DE COLJNG-92. NANTES, 23-28 Ao~r 1992 6 8 4 Pgoc. OF COLING-92, NANTES, AUG. 23-28, 1992 
is, the sense of "logic" in the name of our model llas 
not been fully achieved. We think that the search 
for parallel implementation will lead to "logic-based 
transfer" in the true sense. 

References 

\[1\] H.AIshawi, D.Carter and M.Rayner: 'qYansla- 
tion by Qua.st Logical Form Transfer', in Proc. 
of 29th A CL Conference, Berkeley, 1991. 

\[2\] C.Copeland, J.Durand, S.Krauwer 
and B.Maegaard: 'The Eurotra Linguistic Spec- 
ification', in Studies in Machine Translation and 
Natural Language Processing Vol. 1, Office for 
Official Publication of the European Commmli- 
ties, 1991 

\[3\] l{.Kaplan,K.Netter,.l.Wedekind and A.Zaenan: 
'Translations by Structural Correspondences', in 
Proc. of 4th European ACL Conference, Manch- 
ester, 1989. 

\[4\] S.Kinoshita and J.Tsujii: 'l~amework for Incor- 
porating Knowledge into Machine Translation', 
in Proc. of International Workshop on Funda- 
mental Research for the Future Generation of 
Natural Language Processing, Kyoto, 1991. 

\[5\] S.Kinoshita: Transfer of Coordinate Structures, 
CCL/UMIST Report 91/13, 1991. 

\[6\] J.Lindop and J.Tsujii: Complez 7~'aasfer in MT: 
A Survey of Examples, CCL/UMIST Report 
91/5, 1991. 

\[7\] S.Nirenburg: Machine Translation, Cambridge 
University Press, 1987. 

\[8\] J.D.Phillips, K.Yoshimura, K.Dohsaka and J. 
Tsujii: 'Linguistic Context for Machine Trans- 
lation', in Proc. of SICONLP '90, Seoul, 1990. 

\[9\] J.D.Phillips: An Architecture for Machine 
7~nslation, CCL/UMIST i~port 92/7, 1992. 

\[10\] S.Sato and M.Nagao: 'Memory-based Transla- 
tion', in Proe. of COLING 90, llelsinki, 1990. 

\[11\] J.Tsujii and K.Fujita: 'Lexical Transfer based 
on Bilingual Signs: Towards interaction during 
transfer', in Proc. of 5th European ACL Confer- 
ence, Berlin, 1991. 

\[12\] G.van Noord, J.Dorrepaal et al: 'The MiMo2 
Research System', in Proc. of the 3rd Interna- 
tional Conference on Theoretical and Method- 
ological lssues in Machine Translation, Austin, 
1990. 

\[13\] If. Watanabe: 'A Model of a Transfer Process 
Using Combinations of Transfer Rules', in Proc. 
of PRICAI '90, 1990. 
