Machine Translation: The Languages Network 
(versus the intermediate language.) 
P.C. ROLF 
Dpt. of Computational Linguistics 
Literary Faculty 
Nijmegen University 
NL6500 HD Nijmegen 
The Netherlands 
Abstract 
Jonathan Slocmn /Slocum, 1985/ divides MT techniques from 
a linguistic point of view into three two-way perspectives which are 
not quite disjunct: "direct versus indirect; interlingua versus trans- 
fer; and local versus global scope.". 
in this paper we present a research paradigm which, in fact, does not 
exactly match any of these perspectives: The Languages Network. 
In this paradigm each pair of languages will he treated as within 
a transfer application hut with tbe characteristics of indirect trans- 
lation: analysis of the source language and synthesis of the target 
language are not totally dependent on each other. 
The proees must be split up into a large number of pieces which can 
be connected into a huge network performing MT from and into sev- 
eral languages. 
hnplementations of this paradigm are |)eing carried out by the author 
hy means of the translator generator SYGMART (see /Chauch6/, 
/Chauch~, 1974/ and /Rolf, 1985/), which pernfits the linguist to 
implement whatever he wants in the field of MT in an efficient way 
on a wide range of computers (t¥om Atari1040STf via SUN's to IBM 
VM/CMS mainframes). 
1 The goal of the intermediate language. 
In discussions on translation systems, the question is often asked 
whether the system is based on direc| translation, whether it works 
according to |he transfer method, or whether it uses an intermediate 
language. "\['his question suggests that translation systems can he 
defined exactly by dividing them into these three categories. 
It is our conviction that solving problems in translation is more com- 
plex than is suggested by this question. In this paper we will show 
precisely where this question is inadequate, by looking at some as- 
pects of the translation process. 
1.1 Why an intermediate language? 
The idea which leads to the definition of an intermediate language 
originates from the wish that, in translation from one source lan- 
guage into several target languages, no completely different transla- 
tion should have to he made for each pair of languages. 
Brandt Corstius /Brandt Corstius, 1978/expresses this idea as fol- 
lows (the citation has been translated from Dutch into English): "In- 
stead of making 90 programs in order to translate ten languages 
into one another (from each language into each of the nine others), 
it would be sufficient to have 20 translation programs (from each 
language into Maehinish, and fi'om Machinish into each of the lan- 
guages). It is even conceivable that eighteen progrmns would be 
sufficient, if one of the ten languages is given the role of this inter- 
mediate language." 
This standpoint implies an "efficient" method in terms of the amount 
of work, without indicating whether this method solves also any prin- 
cipal problems with respect to machine translation. 
On the contrary, Brandt Corstius remains sceptical about this. 
1.2 What shall be defined? 
Ideally the intermediate language will have to be an unambiguous 
representation of the meaning of (each of) the source language(s). 
544 
This implies that a language should be found in which it is possible 
to represent all possihle meanings in an unanfl)iguous way. 
If Brandt Corstius is followed in this, then in the translation h'om 
and into natural language this would have to he one of the natural 
languages. But one of the main characteristics of natural language 
is precisely that it is efficient, which implies that with few words and 
constructions a lot can be expressed in very many different circum- 
stances. 
The way in which each individual natural hulguage is etFlcient differs 
from language to hmguage: ambiguities and vagueness in a source 
language cannot usually be projected in a one-to-one col'respondenee 
onto a target language. 
So it would seem to be not entirely plausihle to select an interme- 
diate language from the languages to he translated, seeing that tt~ 
demand of unambiguity is too heavy precisely for natural language. 
An intermediate language should not only be unaml)iguous, hut it 
should also be able to represent all possible meanings. We, are con- 
vinced that for every sentence of a natural language an imfinite set 
of meanings is possible, since meaning depends on the universe of 
discourse and the set of possible universes of discourse is infinitely 
large. 
All in all there is enough cause for a fundamental approach to the 
prohlem: what is to be achieved in defining an intermediate language 
in the maclfine translation of a set of natural languages from and into 
each of the menfl)ers of that set? 
The need for an intermediate language originates, on the one hand, 
from the idea that the analysis of a source language will be largely the 
same, irrespective of the target language into wieh it is to be trans- 
lated, and on the other hand, from the need to analyse the source 
language in such a way that all ambiguities have been solved, and 
that therefore the generation of the target language can take place 
without any further prohtems. 
With respect to the former, we, too, believe that the idea that the 
analysis of a source language is partly the same, irrespective of the 
selected target language, is entirely correct. 
When we plot the translation process on a line from source language 
to target language, this will be the part which is close to the source 
language: to put it in rather more linguistic terms: the morphologi- 
cal analysis and that part of the other syntactic analysis that can be 
smmned up in the term surface grammar, so in any case the NP and 
VP detection, for instance. We shall return to this below. 
The second need, viz. completely disambiguating the source lan- 
guage, would seem to be too heavy a demand, as was formulated 
before, with reference to ~all possible universes of discourse'. The 
two needs that have been mentioned cannot be fulfilled, but it can 
be maintained that there is no need for the entire analysis of the 
source language and the entire generation of the target language to 
be done over and over again for every pair of languages. It has to be 
deternfined what the two paths, analysis and generation, will look 
like. Parts of these two paths will renmin the stone, for the source 
language irrespective of the target language, and for the target lan- 
guage irrespective of the source language. 
2 The analysis of the source language. 
Since there is no reason to adopt an intermediate language as has 
been argued, the problem facing us is the analysis of the source lan- 
guage, as well as the generation of the target language and the pro- 
cess between analysis and generation, which will be discussed in the 
following two sections. 
The point of departure for both these sections, and, in fact, for this 
paper is the hypothesis that tbe meaning of the source language de- 
pends on the objective one has in mind. In the case of machine 
translation the meaning, expressed in the translation, depends on 
the target language defined. 
In this line ot arguing translating i.'~ therefore always a matter of 
a specific relat:ionship between two lauguages. When we plot the 
process of translation on a line, however~ we can distinguisb three 
phases, which can he referred to as analysis, translation and gener- 
ation. The present and the next following two sections have been 
divided on the basis of this principle. 
2.1 What is analysis? 
The analysis or a source language can be defined, in a very abstract 
way~ as the addition of information to the input. It may seem triv- 
ial, but this st;u'ting point implies that no information must be lost 
in the analysing stage, lnlbrmation may only be added. This also 
implies that the input order must not be changed. So, in our view, a 
dependency grammar is not suitable for the analysis because of the 
loss of the input order. 
Changes in the input order can only be brought about on the basis 
of requirements posed by the target langnage. 
This brings us to a second aspect of the analysing stage: in this stage 
only source language inherent data are worked with, to be subdivided 
into static (lexical) and dynaufic (granmmtical) data. 
\]'he fact that in {:he analysing stage solely data inherent to the source 
language are used, does not mean that the target language has notb~ 
ing to do with the nature of the analysis. That would clash with our 
starting point, viz. that the meaning of the source language depends 
on the target language. 
The influence of the (set of) target language(s) extends over the way 
it1 which the analysis is carried out, in other words, what type of 
information has to be added to the text of the source language. 
Let us take, by way of example, the translation Dutch-English. We 
will assume (for convenienee's sake) that in Dutch the word order in 
subelauses is S(ubject)-O(bject)-V(erb), while in English the stan- 
dard word order in subclauses is S(ubject)-V(erh)-O(bject). 
The change fr()m S-O-V into S-V-O does not belong to the analysing 
stage of Dutch, lbr it implies lnss of infornmtiou because of the word 
order cbange. ~\]owever, the translation into English requires fron~ the 
analysing stage of Dutch that, among other things, the categories S~ 
0 and V are a~;signed. The assignment of S and 0 implies that NP's 
have to be ibn:od, etc. 
The analysing stage comprises all stages which belong to morphof 
ogy~ surface grammar and possibly a large number of matters that 
belong to the field of semazatic interpretation (see /Bakel, 1984/). 
This last category, semantic interpretation, is close to the translation 
stage and will possibly be different for groups of target languages. 
in the section headed 'Prospect' we will indicate schematically how 
this semantic interpretation has to be situated in the whole of the 
translation process. 
2.2 Algorithmic consequences. 
.~;tar ting ii'om the assumption that the addition of information (com- 
mitting abstractions) is brought about, anmng other things, by the 
application of some sort of dependency structure, what is needed is 
a form of graphic representation. 
For many languages, and certainly also for Dutch, the traditional tree 
structure clashes with our demand formulated earlier, that informa- 
tion should be retained: the original word order nmst be nmintained 
during the anMysing stage. (In Dutch a postmodifier in an NP is 
often extraposed, e.g. "Ik heb de nmn gezien met de bril." \]¥ans- 
\[ated word by word: "I have the man seen with the spectacles.") 
\],hn'thermore linguists should have the possibility of expressing lin- 
guistic notions in a way which is adequate to them. For this purpose 
a distinction has been made, in the SYGMART system, between 
the morphololgical analysis, which operates on words (the subsystem 
OPALE) and a tree transformational part (the subsystem TELESI), 
whicil ol)eral.es on nndti-dimensional trees over text(s) (the notion 
sentence does not exist in SYGMART). 
The surface gremlmar and the semantic interpretation cannot there- 
fore be algorithmically distinguished. 
This multi--dimensionality enat)les the linguist to establish relation- 
ships between sentence constituents which are far apart, without 
having to extract them out of their original order. This multi- 
dimensionality has to be looked upon as'the definition of graphs 
more complex than trees over tbe input. For a more detailed discus. 
sion of the mnltl-dimensionMity the reader is referred to/Chauch6, 
1984/. 
Our arguulents for not using the traditional grmmnatical types are 
given in /Roll, 1986/. Part of the analysis of Dutch is shown in 
Appendix A. 
3 The t'eanslation. 
The translation stage is the stage between the source language inher- 
ent analysis and the target language inherent generation. This stage 
can be roughly compared to a transfer component, as suggested in 
the beginning of section /. 
Two features arc characteristic for the translation, viz. word order 
change and the addition of target language features. 
Word order change(s) (better: conlpouent or category nlovement) is 
(are) not per definition carried out separately for all possible target 
languages. If in the example of the word order change in subclauses, 
presented in the previous seclion, the rule SOV ---) SVO has to be 
al)plied to a subset of the target languages, tbis can be done tbr the 
entire sul)set prior to the introduction of target language specific fea- 
tures. 
This introduction of target language specific features is clone by the 
lexical translation, or the translation of the words. We will assume 
here that the analysing stage has provided all the necessary informa- - 
tion to gel: the correct translation for every word. 
Because of the information added in the analysing stage the correct 
translation of a word implies the translation of a complex data strut-' 
ture into another complex data structure, in which the written base 
forin in both cases is hut one wdue of that data structnre. 
On the basis of information that comes in after the lexical transla- 
tion, further word or.der changes will generally have to take place, as 
well as the generation of grannnatical structures. 
A simple exaulple in this connection is the folh)wiug: the I)utrh verb 
hlijven is translated into English keel) , but in Dutch blijven is 
completed by an infinitive (blljven wachten), whereas in English a 
gerund is expected (keep waiting). 
If on the basis of the new lexical information further grannnatieal 
rules have to be applied, such as nmving the verb in the gerund con- 
strut(ion (Dutch "ik blljf op hem waehten" into Englisb "l keep 
waiting for hiln'), these rules also belong to the translation stage, 
not to the generation stage, unless the rules apply to all possible 
source languages with respect to English. 
As in the previous section, here, too, the demand made of the algo- 
rithmic procedures and the possibility of building and manipulating 
conlptex datastructures is heavier than in traditional gramlnatical 
types. Within SYGMART the subsystem TELESI is used for the 
translation stage, which does not imply that for each pair of lan- 
guages a separate TELESI implementation has to be made after all: 
SYGMART provides for the application of different TELESI gram-- 
n|ars one after another. 
4 The generation of the target language. 
15-om the previous sections it has already beconle apparent that the 
generation of the target language does not come into play, until only 
target language inhdrent matters are at issue s matters which hold 
irrespective of the source language that is used. 
They are in arty case all matters of a morphological nature which 
in the entire translation process are the last to be dealt with. In 
the translator generator SYGMART the generating morphology is 
treated by the sul)systen) .AGATE. 
5~5 
If there are grammatical rules which also have to be applied inde- 
pendently of the source language, they also belong to the generating 
stage, in our set-up. These rules will not be many, for that would 
imply that in a target lauguage certain constructions should occur 
for which in no (source) language an analogous constrnction was to 
be found. 
In our set-up the technical three-way division of SYGMART (OPA- 
LE, string into tree, TELESI tree into tree via network, AGATE, tree 
into string) cannot be measured in a one-to-one correspondence onto 
the three-way division of the translation process, viz. analyses, Irans- 
lation and generation. The morphological analysis always takes place 
in OPALE and is a, rather small, part of the entire analysis. The 
greatest part of the analysis, consequently, takes place in TELES1. 
Everytlfing belonging to the translation hapI)ens in TELESI. As far 
as the generation is concerned, a small part is possibly carried out 
in TELESI, but the morphological generation, naturally, takes place 
in AGATE. 
5 Prospect: The Languages Network. 
From the foregoing it can be deduced that as far as we are concerned 
the question formulated in the beginning disregards the complexity 
of the translation of natural language from and into one another. In 
general the analysis of the source language is the most important 
component: once the analysis has been carried out on all possible 
levels in all possible details, generating the target language is 'rela- 
tively' siml)le: at that stage word meanings have, of course, been dis- 
ambiguated, semantic interpretation have been assigned, references 
have been determined, etc., but all this has been done in relation to 
the meaning, viz. the translation. 
Seeing that we analyse on the basis of the requirements of the target 
language, analysis is only partly an unambiguous notion: not all the 
abstractions will be equal for all the target languages, nor will they 
be required for all the target languages. 
If for each pair of languages the entire process is plotted on a line 
from source language to target language, it will be possible to point 
to a number of points on that line, where analyses will go into dif- 
ferent directions (depending on the subset of target languages) and 
Where translati6ns and generations merge. These lines together form 
the languages network. 
On the hasis of such a network it will be possible, in the future, 
to formulate relationships with reference to the affinity of languages 
mutually. This may sound speculative, but we are convinced that 
what has been presented cannot be reduced to the definition of a 
single intermediate language, unless it is done for subsets of natural 
languages which have been defined precisely. Our objections • to this 
are formulated in/Roll', 1986/. 
/ 
A An example of the analysis of Dutch. 
The following is an extract of a running implementation of a part of the 
analysis of Dutch in the SYGMART system. It can be regarded as one of 
the nodes in the proposed Languages Network. 
This node constitutes a network in itself, with four possible entry points, 
eharacterised by "&ENTREE:", and several end nodes, eharaeterised by 
"-- >%STOP.'. 
~GRAHMAIRE. 
/* Basisgrammatica's */ " 
&GRAH: NP(E). 
FINDKERN: 0(1(*)) / 
1: ((NRDS00RT=SUBST)I(PRON=PERS))&(CATEGORY'=NCKERN) 
=> x(Y(1)) / 
x: (CATEGORY=NBAR) ; 
Y: (CATEGORY=NCKERN). 
FINAPQP2: 0(I(2),*,3) / 
2: (NRDSOORT=ADJ)I(NRDS00RT=TELW); 
3: CATBGORY=~BAR 
=> X(*O<,1>*,Y(l(2),*3<,>*),*O<3,>*) 
X: 0 ; 
Y: 3 ; 
1: I(SUBCAT(2)). 
546 
RESTADJ: 0(1(2(0)),*,3(4)) / 
2: WRDSOORT=ADJ; 
3: CATEGORY=NBAR; 
4: CAYEGORY=AP 
=> X(*O<,l>*,Y(Z(2,*4<,>*),*3<4,>*),*O<3,>*) / 
X: 0 ; 
Y: 3 ; 
Z: 4. 
FINDEOPI: 0(1(2),*,3) / 
2: (WRDSOORT=LIDW)I(WRDEOORT=TELW) 
LOCDE~PI(0,2,3) 
=> X(*O<,l>*,Y(l(2),3),*O<3,>*) / 
X: 0 ; 
Y: (KENCAT(3)) ; 
1: I(KENCAT(2)). 
RESDEQPI: 0(%1(Z2),*,3(4)) / 
O: (Vt\]RM='SENTENCE*****~)I(VORM='BIJZIN*****')I 
(CATEGORY=PVWWCL); 
2: (NRDSOORT'=LIDW)&(WRDSOORT'=TELW) / 
CATRES(3,4) 
=> X(*O<,I>*,ZI(Z2),¥(3(4)),*O<3,>*) / 
X: 0 ; 
Y: (KENCAT(3)). 
ONLYAD3: 0(1(2(*))) / 
O: (VORM='SENTENCE*****')I(VORH='BIJZIN*****')I 
(CATEGORY=PVNNCL); 
1: CATEGORY'=AP; 
2: WRDSODRT=ADJ 
=> X(*O<,I>*,Y(I(2),KERN(EMPT)),*O<I,>*) / 
X: O; 
Y: (CATEGORY=NBAR); 
1: I(SUBCAT(2)); 
KERN: (CATEGORY=NCKERN). 
-->ZNUL. 
&GRAM: NPCONS(E). 
PREPCON: 0(1(2),*,3) / 
2: NRDSOORT=PREP ; 
2: CATEGORY=NP 
=> X(*O<,l>*,Y(2,3),*O<3,>*) / 
X: 0 ; 
Y: *PkEP. 
-.-> ~NUL. 
~GRAM: WWCI.US(E). 
FINDWWCL: TOP(I,*,2) / 
TOP: VORM='SENTENCE*****'; 
1: (WERKW'=WERKN.->O)&(WERKW'=PV)&(WERKW'=PV\[IMPERAT); 
2: (WERKW^=WERKW->O)a(WERKW'=PV)~(WERKW^=PV\]IMPERAT) 
WWCL(I,2) 
=> NEWTOP(*TOP<,I>*,NEW(1,2),*TOP<2,>*) / 
NEWTOP: TOP; 
NEW: (CATEGORY=WWCL;VERBUM=VERBUM(1)); 
2: (KENWW(I,2)). 
FIBYPVCL: TOP(I,*,2) / 
TOP: VORM='SENTENCE*****~; 
1: (WERKW~>=PV); 
2: (WERKW'=WERKW->O) / 
PVWWCL(1,2) 
=> NEWTOP(*TOP<,I>*,NEW(1,2),*TOP<2,>*) 
NEWTOP: TOP; 
1: I(WERKW:PV); 
NEW: *PVWWCL; 
2: (KENWW(I,2)). 
FBYPVCI,2: TOP(2,*,I) / 
TOP: VORM='SENTENCE*****'; 
I: (WERKW~>=PV); 
2: (WERKW'=WERKW->O)&(CATEGOR¥'=WWCL) 
PVWWCL(1,2) 
=> NEWTOP(*TOP<,2>*,NEW(2,1),*TfP<I,>*) 
NEWTOP; TOP; 
1: I(WEREW=PV); 
NEW: *PVWWCL; 
2: (EENWW(I,2)). 
FINDPVCL: TOP(I,2) / 
TOP: VORM='SENTENCE*****'; 
I: (WERKW~>=PV); 
2: (WERKW~=WERKW->O) / 
PVWWCL(I,2) 
=> NEWTOP(*TOP<,I>*,NEW(1,*TDP<I,2>*,2),*TDP<2,>*) / 
NEWTOP: TOP; 
1: I(WERKW=PV); 
NEW: *PVWWCL; 
2: (KENWW(I,2)). 
FINDPV: TOP(1) / 
TOP: VORM='NENTENCE*****'; 
1: (WERKW©>=PV) 
=> NEWTOP(*TDP<,I>*,NEW(NEWI(1)),*TOP<I,>*) / 
NEWTOP: TOP; 
NEW: (CATEGORY=PV); 
NEWt: (WERKW=PV; 
<(VERBSORT(1)'~>=COPULA): VERBSORT=ZELFST # 
VERBSORT=COPULA >). 
--->%NUL. 
hGRAM: BYZIN(E). 
BYZBEG: TOP(*) / 
TOP: (PRON=RELPRON) I(VOEGWRD=ONDER) 
=> NEWTOP(NEWR(TSP)) / 
NEWTOP: (VORM='BIJZIN*****';P~ON=PRON(TOP)). 
BYZREST: 0(I(2),*,3) / 
2: VORM='BIJZIN*****'; 
3: (CATEGORY'=WWCL)k(CATEGORY'=PVWWCL)~(CATEGORY'=PV) 
:> NEWTDP(*O<,1>*,I(NEW2(*2<,>*,3)),*O<3,>*) / 
NEWTOP: O; 
NEW2: 2. 
BYZEND: 0(1(2),*,3) / 
2: VflRH='BIJZIN*****'; 
3: (CATEGORY=PV) i(CATEGORY=PVWWCL)i(CATEGDRY:WWCL) 
=> NEWTOP(*O<,I>*,I(NEW2(*2<,>*,3)),*9<3,>*) 
NEWTOP: O; 
NEW2: 2. 
RELBYNP: TOP(I,*,3(4)) / 
1: CATEGORY=NP; 
4: (VORM='BIJZIN*****')h(PItON=RELPRUN) 
=> NEWTOP(*TOP<,I>*,I(3(4)),*TOP<3,>*) / 
NENTOP: TOP. 
--> ZNUL. 
aENTREE: TELWRD(1). 
DELTIGEN: 0(1(*),*,2(*)) / 
2: ((VORM='tig')\[(VORH='on')) 
:> o(1) I 
I: I(<VORH(2):'tig':SOORT=SOORT(2)>). 
KEERSM: 0(1(*),*,2(*)) / 
1: SOORT:EENHEID ; 
2: SOORT=TIENTAL 
=> 0(2,1). 
DEL_ENHO: 0(1(*),*,2(*),*,3(*)) / 
i: (SOORT=DUIZTAL); 
2: (SDORT=EENHEID); 
3: (SOORT=HONDTAL) 
=> 0(1,2) / 
3: 3(REPR=REPR(2)). 
VULOP: 0(1(*),*,2(*)) / 
O: (WRDSDORT=TELW) / 
AANW(I,2) 
=> 0(1,2,2) { 
3: (KENTEL(J)). 
VULAAN: 0(1(*),*) / 
1: (SOORT:DUIZTAL) I(SOORT=HONDTAL)I(SOORT=TIENTAL) 
:> o(1,2) / 
2: (KENTEL(1)). 
547 
DEL_DU(*(*);O(2)): 0(1(*),*,2(*)) / 
I:(SOORT=EENEEID) I(SOORT=COMBI); 
2: (SOORT=DUIZTAL) J(SOORT=HONDTAL) 
=> 0(2) / 
2: 2(REPR=REPR(1)). 
--> ~STfP. 
aENTREE: ENGTLWD(I). 
TranstelSHLT(Transtel): blad(*) / 
blad: (TRANS=TRANS->O) 
=> blad / 
blad: blad(TRANS=TRANS(DICT(*))). 
CENGORD: O(een(*),*,en(*),*,tien(*),*,tig(*)) 
een: SODRT=EENHEID ; 
on: VORH='en'; 
tien: (S00RT=TIENTAL))(SDDRT=EENEEID); 
tig: VORH='tig' 
=> O(tien,tig,een). 
__> ZSTOP. 
&GRAM:REsTNP(1). 
RESTADJ. 
ONLYADJ. 
FINAPQPR(*(*);Y(2)). 
FINDEQPI(*(*);¥(2)). 
RESDEQPI(*(*);¥(4)). 
--> ~STOP. 
&GRAM: HOUDOP(E). 
--> ZSTOP. 
/* Netwerk van grammatica'S */ 
&ENTREE: SEPARATE(1). 
ONBEKEND(*(*);O(1))$TRF(HDUDOP): 0(1) / 
i: DICT~>=ONBEKEND 
=> 0(1). 
LEESTSHLT(ONBEKEND): 1(2(4(*),*,3(*))) / 
3: LEEST-=LEEST->O 
=> NEWl(*I<,2>*,4),NEW2(3,*l<2,>*) / 
NEWI: I. 
LEEST2$ELT(ONBEKEND): 1(2(4(*),*,3(*))) / 
4: LEEST'=LEEST->O 
=> NEW1(*1<,2>*,4),NEW2(3,*1<2,>*) / 
NEW1: 1. 
SPLITB(*(*);X(P2,R))$HLT(LEEST,LEEST2,DNBEKEND): 
O(1(P1),2(P2)) / 
PI: LEEST=EOSENT; 
P2: LEEST=EOSENT 
=> Y(*O<,I>*,1(P1),X(*O<I,2>*,2(P2)),*O<2,>*) / 
X: (VORM='SENTENCE***** ') ; 
Y: O. 
SPLITOP$HLT: O(I(P)) / 
P: LEEST=EOSENT 
=> Y(X(*O<,I>e,I(P)),*O<I,>*) / 
Y: O; 
X: (VORM='SENTENCE*****'). 
--> PREPROC. 
&ENTREE: PREPROC(U,PREPROC,PREPROC). 
WW2(*(*);O(2)): 0(1(3(2(*)))) / 
O: VORM='SENTENCE*****'; 
I: WRDSOORT=WRDSOORT->O; 
2: WRDSOORT=VERB 
=> 0(1(3(2))) / 
1: I(VERBUM=VERBUM(2)). 
548 
WWl(*(*);l(2)): 1(2(*)) / 
I: WRDSBORT:WRDSOORT->O; 
2: WRDSOORT=VERB 
=> i(2) / 
I: I(VERBUM=VERBUM(2)). 
TEINF(*(*);TOP(4,2)): 0(1(2),*,3(4)) / 
2: VORH=Jte~; 
4: WERKW@>=INF 
=> TOP(*O<,I>*,NEWTOP(X(2,4)),*O<3,>*) / 
TOP: O; 
NEWTOP: (WRDSOORT=WRDSODRT(4);WERKW=TEINF; 
VERBSORT=VERBSORT(4)); 
2: 2(WRDSOORT=WRDSOORT->0); 
4: 4(WERKW=INF); 
X: (WERKW=TEINF;VERBSORT=VERBSDRT(4)). 
--> TELW: 0(*) / O: (WRDSDORT=TELW)a(SODRT^=CYFER) 
&(SOORT'=SOORT->O) . 
--> FINDNPCL. 
~GRAM: TELW(E). 
TELWDRD(@TELWRD;O): 0(1(*)) / 
1: (WRDSOORT=TELW)&(SOORT'=CYFER)~(SOORT'=SDORT->O) 
=> o(1), 
--> FINDNPCL. 
--> ZNUL. 
&GRAM: FINDNPCL(I). 
FINDKERN(*(*);X(1))$HLT(FINDKRRN). 
FINDWWCL(*(*);NEW(2,1)). 
FIBYPVCL(*(*);NEW(2,1)), 
FBYPVCL2(*(*);NEW(2,1)). 
FINDPVCL(*(*);NEW(2,1)). 
RESTADJ. 
FINAPQP2(*(*);Y(2)). 
FINDEqPI(*(*);Y(2)). 
RESDEQPI(*(*);Y(4)). 
F!NDPV(*(*);NEWTOP(NEW)). 
UNLYADJ(*(*);X(EMPT,KERN,2)). 
--> FIBYZ: 0 / O: (PRON=RELPRON)\](VOEGWRD=ONDER). 
--> RELBYZIN: 0 / O: (WRDSODRT=PREP). 
--> TRANSLAT. 
~GRAM: FIBYZ(I). 
BYZBEGSHLT(BYZBEG). 
BYZEND(*(*);NKWTOP(S,NEW2)). 
~"B~'ZREST(*(*);NEWTOP(3)). 
--> RELBVgIN: 0 / O: (PRON=RELPRON)((WRDSOORT=PREP). 
--> TRANSLAT. 
&GRAM: RELBYZIN(I). 
RELBYNP. 
PREPCON(*(*);X(3)). 
--> TRANSLAT. 
&GRAM: TRANSLAT(U,TRANSLAT,TRANSLAT): 
<STAM^=STAM->O :TRANS=TRANS(DICT(*))>. 
--> ZSTOP. 
~FIN. 

References

\[Chauehd, 1982\] 
Chauchd, J., V. Cheboldaeff, R. Leseoeur, M. Jatteau, Rapport de 
prdsentation du project relatif b l'6crlture des specifications 
d'un syst$me de tradnctlon assistde par ordinateur, SLIGOS, 
Conseil, assistenee et rdalisations, produits informatiques, services de 
gestion, Avril 1982. 

\[Chauclff, 1984\] 
Chaueh~, J.,Un outil multidimensionnel de l'analyse du discours, 
in: Proceedings of Coling84, Stanford University, California, pp. 11 - 15, 
1984. 

\[Landsbergen, 1982\] 
Landsbergen, J., Machine Translation based on logically 
• isomorphic Montague grammars, in: Horecky, J. (ed.), COLING 82, 
pp 175-182, North-Holland Amsterdam, 1982, 

\[Landsbergcn, 1984\] 
Landsbergen, Jan, Isomorphic Grammars and their use in the 
Rosetta Translation System, Pbilips Research Laboratories Eindhoven 
- the Netherlands, (presented at the Tutorial on Machine Translation, 
April 2 - 6, 1984, Lngano), 1984. 

\[Rolf, 1983\] 
Rolf, P.C., Vertalen van getalsnamen, Verslagen Computerlingu~'stiek 
No. 3 - 1983, KU Nijmegen. 

\[Roll, 1985\] 
Rolf, P.C., Beschrljving van het vertaalsysteem SYGMART, 
versie 4, intern verslag KU Nijmegen, 1985. 

\[Rolf, 1986\] 
Rolf, P.C. and J. Chaucht4, Machine Translation and the 
SYGMART system., in: Computers and the Hmnanities, 1986, Vol.20, 
no 4,pp 283-288. 

\[Slocum, 1985\] 
Slocum, Jonathan, A Survey of Machine Translation: its History, 
Current Status, and Future Prospects. in: Computational 
Linguistics, Vol 11, Number 1, January-March 1985. 

\[Witkam, 1983\] 
Witkam, A.P.M., Distributed Language Translation, Feasibility 
Study of a Multilingual Facility for Videotex Information Networks, BSO, 
Utrecht, 1983. 

\[Brandt Corstius, 1978\] 
Brandt corstius, H, Computer-Taalkunde, Muiderberg, 1078. 

\[Bakel, 1984\] 
Bakel, Jan van, Automatic Semantic Interpretation, Foris 
Publications, Dordreeht, 1984. 

\[Barr, 1981\] 
Barr, Avron and Edward a. Feigenbanm , The Handbook of Artificial 
Intelligence, Stanford, 1981, VoI.L 

\[Chauchd, 1974\] 
Chauchd, Jacques, 'Pransducteurs 8z Arborescences, Etudes et 
r~alisations de syst~mes appliqu~es aux grammaires 
transformationnelles, Thbse d'Etat, Grenoble, 1974. 

\[Chauchd\] 
Chauchdl J., Systb.me de manipulations algorithmiques et 
r6cursives de texte, SYGMART, Labotatoire de traitement de 
l'information, Le Havre, France. 
