ON THE SEMANTICS OF FOCUS PHENOMENA IN EUROTRA 
Erich H. STEINER Jutta WINTER-THIELEN 
IAI Eurotra-D 1A1 Eurotra-D 
Martin Luther Str.14 Martin Luther Str.14 
D-6600 Saarbriicken D-6600 Saarbrtieken 
FRG FRG 
ABSTRACT 
In this paper, we discuss issues connected to the phenomenon of 
linguistic FOCUS or INFORMATION DISTRIBUTION in the 
sentence in the context of the multi-lingual machine translation 
project EUROTRA. We shall present some of the arguments 
why a consideration of FOCUS phenomena is important for the 
determination of linear order and for semantic interpretation. 
We shall proceed, in sections 2 and 3 of the paper, to mention 
the main lines of development in the dicussion of FOCUS 
phenomena in Computational Linguistics and in Linguistics 
respectively. Section 4 contains an illustration of a pilot 
implementation covering some aspects of FOCUS phenomena in 
EUROTRA-D. 
1. Problem description 
1.1. The determination of linear order in the clause 
An initial, but faulty assumption, underlying a possible theory 
of language is that linear order of syntactic constituents in the 
clause is fully determined by a combination of the following 
types of information: 
- basic type of language (SVO,SOV,etc) 
- clause modus (indicative, interrogative,etc.) 
- information on verbs in the lexicon (lexical forms). 
This paper will present some of the arguments why this 
assumption is wrong for the clause, in how far this fact is 
relevant for EUROTRA, and initial suggestions towards a 
treatment within our framework. We shall not be able here, for 
reasons of time, to deal with linear order within NPs, ADJPs, or 
ADVPs (cf. for English on these questions Halliday.1985 : 158ff, 
and many others). 
At least the following hierarchy of factors can be said to 
contribute to linear sequence (cf. Allen.1987 : 51): 
(1) The familiarity hierarchy 
(2) The topic < comment, given < new hierarchies 
(3) The universal sequencing conventions 
(4) The definitieity and referentiality hierarchies 
(5) The personal, social status, and role hierarchies 
(6) The dominant descriptor hierarchies 
(7) The formal hierarchies 
Of these, this paper will deal with focus phenomena, covered in 
(2) above. Different aspects of this area are covered in the 
linguistic literature under the headings of "topic-comment", 
"focus-presupposition", "theme-rheme", and "given-new". The 
difficulty with these pairs of terms is 
- that they reflect a difference in orientation (rhetorical vs. 
logical vs. psychological); 
- that they cover the related, though different aspects of linear 
order and intonation. 
630 
Rather than go into an extended theoretical discussion at this 
point, we shall in 1.2. discuss a range of examples to show 
where exactly these phenomena would seem to be relevant for a 
multi-lingual MT-system like EUROTRA. The system within 
which the implementation of the ideas suggested here were 
implemented is the Eurotra-D system as described in 
Steiner,1986 and Steiner et al.forthcoming and in other places. 
1.2. The relevance of linear order for semantic interpretation 
1.2.1. The scope of negation 
Cf. the following pair of sentences (capitals standing for stress): 
(1) The eee is not controversial, because it is a multi-natlonal 
ORGANIZATION. 
(2) Because it is a multi-national organization the eec is not 
CONTROVERSAL 
(1) and (2) are identical, except for the position of the Adverbial 
Clause, and the assignment of primary information focus 
realized by main stress and symbolized here by capitals. 
However, their semantic interpretation differs with respect to 
the scope of negation: (1) is ambiguous with respect to the 
queston of whether or not the propostion expressed in the 
matrix clause is true, cf. (3): 
• (3) The eee is not controversial, because it is a multi-national 
ORGANISATION. It is controversial because it is EXPENSIVE. 
The second sentence in (3) is not possible with (1), whereas it is 
with (2). In (1), the scope of negation includes the Adverbial 
Clause, whereas in (2), it does not. In general, it seems to be the 
case that the scope of negation may include everything in the 
clause to the right of the negation particle and including the 
constituent which has a lexical item carrying focus. Therefore, 
the truth values of (1) and (2) are different, which implies that 
they are not acceptable translations of each other in a truth 
value oriented semantics, such as we are having it at the moment 
in EUROTRA. Yet, in the present framework, (1) and (2) would 
receive one representation at ERS, which means that ECS-ERS 
translations are not translationally equivalent. The implications 
for other pairs of levels of representation are obvious. 
We shall not go into the question here of how marked intonation 
in such cases would influence truth values - as we ar0 dealing 
with written texts exclusively, we shall assume unmarked 
intonation for each syntactic structure, which implies that in the 
case of thematizing constructions, we shall assume main stress in 
the constituent functioning as a marked theme, and stress within 
the rightmost clause constituent carrying a fully lexical item 
otherwise. All this, for the time being, applies to English. 
1.2.2, The scope of quantification 
Cf. (4) and (5): 
i4) Everyone in this room speaks two languages. 
(5) Two languages are spoken by everyone in this room. 
This is a classical example which has been discussed in the 
literature again and again, so we shall merely re-state the 
problem: 
In (4), it is not implied that everyone speaks the same two 
languages, whereas this does seem to be the implication in (5). 
The difficulty seems to be that type and scope of underlying 
quantifiers may change depending on linear order. 
This problem may be even more important in languages with a 
somewhat freer word order like German, cf. (21)-(24): 
(21) Einit~e yon uns haben eine Serge. 
Some of us have a worry. 
(22) Eine Serge haben einige von uns. 
A worry some of us have. 
(23) Wit alle arbeiten an einer L~sung. 
All of us a working towards a solution. 
(24) An einer L6sung wird von alien von uns gearbeitet. 
A solution is worked on by all of us. 
In (21) and (23), the indefinite article in the object NP does not 
necessarily imply "one and the same for all of (Subject)", 
whereas the implication may be given in (22) and (24) (changes 
of syntactic constituents considered). Note that (24) includes 
passiviz~ttion, whereas (22) includes thematization without 
passivizalion. This shows that what is at issue here is not 
VOICE, but, as we have been emphasizing all the time, FOCUS. 
Howevel, it seems to be type, rather than scope of 
quantification which is at issue here. Furthermore, the 
judgements of native speakers here arc influenced by intonation 
and ethel factors. 
1,2.3. Acceptable question-answer pairs 
For an illustration of this problem, cf, (6) to (11): 
(6) When did Parliament decide on tile budget? 
(7) Parliament decided oil the budget last WEEK. 
(8) It was last WEEK that Parliament decided on the budget, 
(9) ?Lasl week Parliament decided on the BUDGET. 
(10) ?It was on the BUDGET that Parliament decided last week. 
(l 1) ?What Parliament DID last week was decide on the budget. 
As it is well known from discussions in tile literature, (7) and 
(8) are fine as answers to (6), yet (9)-(11) are at least 
questionable. Observe that thematizing devices like in (8)-(11) 
do not tlanslate in simple syntactic transfer between different 
European languages. German, for example, can simply change 
linear order where English needs clefting, pseudo-clef ring, 
extraposition, or similar devices. Observe furthermore that, as a 
first rule, focus goes with tile thematized constituent in cases of 
marked theme assignment, wbereas the remainder of the clause 
would seem to function as presupposition, which is important to 
observe ia the case of written language, wbere we do not have 
any encoding of intonation directly. In the case of wh-questions, 
focus usually goes with tile wh-elemeut. In cases of unmarked 
theme a~;signment, focus, in English, falls on the last fully 
lexical item in the clause. 
Back now to the question of the translation of sentences with 
certain types of marked themes, cf.(12) to (15): 
(12) Zu elnem groBen UNGL~CK wurde die Entscheidung erst 
dutch die weitere ENTW1CKLUNG (D) 
(13) ?A big DESASTER the decision only became because of 
further DEVEI,OPMENTS 
(14) It was only through further DEVELOPMENTS that the 
decison became a DESASTER. 
(15) ?Es war nut dureh die weitere Entwickluug, dass 
die Entscheidung zu einem Ungliiek wurde. 
(13) is a questionable translation of (12), yet the present 
framework would either dictate so, or else produce the 
"canonical" form underlying (16) as a translation of (12): 
(16) The decision became a desaster because of further 
DEVELOPMENTS. 
(16), however, misses the double focus of (12) and is thus 
problematic, also in view of what we have said in 2.2.1. 
On the other hand, (15) is a questionable translation of (14), 
because the syntactic device of "clef ring" does not carry over 
from English into German in a non-complex way. 
We shall not go into other areas where FOCUS is important at 
this point, such as, for example, the scope of certain classes of 
Adverbials like "only, just,..." in English (we would like to 
express our gratitutde to Auand Syea, Manchester, for alerting 
us to this area.) 
It should be obvious that: 
a: These problems occur not just in isolated examples, but with 
thematizing devices in general. 
b: The problems are magnified greatly between languages wbicb 
are less closely related than English and German, such as, for 
example, Germanic and Romance languages etc. 
As we hope to have indicated in this section, a disregard for 
questions of focus may lead to translations which are not truth 
value equivalent, or else to translations which are grammatically 
wrong or at least hardly acceptable. One could, of course, choose 
tile approach of introducing complex syntactic transfer to handle 
such problems. It should be obvious that the complexities 
involved would be considerable. Even in a syntax based transfer, 
though, one would need a systematic account of focus 
phenomena. 
The alternative would seem to be to include into EUROTRA 
Interface Structure semantic information about focus phenomena 
and let the rest be handled in target language generation. This 
seems to be a solution which is theoretically more desirable and 
practically far superior in view of the fact that with the number 
of languages involved in EUROTRA, introducing a major 
source of complex syntactic transfer is always extremely costly 
and should be avoided as far as possible. 
The following two sections will give references to places in the 
literature where focus phenomena are discussed more 
extensively. 
2. Issues in the discussion of focus phenomena: 
Computational Linguistics 
The relevance of focus phenomena and topicalization for free 
word order languages is discussed in Karttunen/Kay.1985. 
A criticism of the general neglect of focus phenomena and 
topicalization can be found in Kay.1985 : 252, as well as in 
Winograd.1983 : 280ff. 
McDonald.1985 discusses the relevance of such phenomena for 
the production of coherent text in MT. 
Mann/Matthiessen.1985 emphasize the need for having a theory 
and implementation of topicalization in text generation, 
illustrating the particular strength of implementations of 
Systemic Functional Grammar in this area, just as do 
Winograd.1983 : 280ff, Patten.1984, and 
Houghton/Isard.forthcoming. 
Recent contributions of Prague School representatives on the 
issues connected with our topic can be found in 
Hajicova/Sgall.1986 and in Sgall/Hajicova/Panevova. in press. 
631 
Within the field of Machine Translation, it would seem to be 
clearly the Prague School approach which has the strongest 
tradition with respect to our topic. Functional Unification 
Grammar and Systemic Functional Grammar both have a known 
history in this respect in Computational Linguistics, even if not 
in MT directly. 
3. Issues in the discussion of focus phenomena: 
Linguistics 
We shall only be able to give an indication of the main lines of 
thought, because the literature on issues of 
focus/presupposition, theme/rheme and topic/comment has 
become very extensive. 
Within the paradigm of Generative Grammar and Government 
& Binding, we find, roughly from the early seventies onwards, a 
discussion of focus phenomena in so far as they contribute to a 
semantic structure of focus and presupposition (cf. Chomsky. 
1970, Jackendoff. 1972 , Jackendoff. 1983, Rochemont.1986). 
The emphasis is on the implications for scope of negation, 
Adverbials, and quantification, as well as on rules for 
generating acceptable sequences of sentences (cf. 1.2. of this 
paper). 
The Prague School has perhaps 'the longest tradition of 
investigating focus phenomena under the heading of "Functional 
Sentence Perspective" organization of the clause (of. 
Danes.F.ed.1974.). The notions of "Theme" and "Rheme" are 
refined in order to overcome a purely binary division of the 
clause, using the notion of "Communicative Dynamism". Also, 
the notions of "Theme and Rheme" are extended to cover aspects 
of text-structure, rather than only sentence-structure. 
Lexical Functional Grammar, to my knowledge, has not given a 
central place to consideration of focus phenomena, while not 
denying their importance (cf. Kaolan/Bresnan. 1982 : 255). 
However, in the area of LFG, my knowledge of the literature is 
far from comprehensive. 
Systemic Functional Grammar has a long tradition of giving a 
central place in their theory for considerations of the "textual 
organization" of sentences into "Theme-Rheme" and "Given- 
New". These two aspects are conceived of as related, though 
distinct. The "Given-New" dimension expresses the assignment 
of "old vs new" information in the clause, whereas the "Theme- 
Rheme" dimension is assumed to account for "Point of 
departure"-assignment to constituents of syntactic structure (el. 
Halliday, 1967-68./ Halliday.1985:38ff/ Fawcett.1980:157ff/ 
Steiner.1983 : 228ff). 
Outside of certain "Schools" of Linguistics, it is certainly worth 
mentioning Chafe. 1970:210ff, Prince.1981 for an interesting 
treatment of focus phenomena within the perspective of 
Pragmatics, Sperber/Wilson.1986 for a definition of "focus" with 
respect to hierarchies of ordered entailments between different 
syntactic and semantic constituents of sentences, and Jones. 
1977. for a very useful treatment of the notion of "Theme" as it 
extends from the sentence to the text. 
4. Towards a treatment of the semantics of focus phenomena in 
EUROTRA: some preliminary suggestions 
In this section, we shall outline an implementation of a 
treatment which will at least begin to cover some of the 
translationally relevant phenomena which we have outlined in 
• the previous sections. This implementation is entirely within the 
EUROTRA framework as described in Arnold et a1.1986. 
For the time being, we shall assume that each clause constituent 
is assigned to either the focus or the presupposition of a clause. 
Later on, it may become important to distinguish more precisely 
between "clause" and "sentence", and between primary and 
secondary focus. Also, there are reasons to assume that focus is 
assigned to lexical categories in the first instance, and that 
upwards percolation to the phrase is not unrestricted. However, 
at present, we shall only try to account for movement of phrasal 
constituents as wholes, and therefore, we shall assume that focus 
is automatically percolated upwards from lexical categories to 
their phrasal mothers. Note that in the future this might create 
special problems as it is not necessarily the head element of a 
phrase which is marked for focus. 
For work in the immediate future, it seems important to 
investigate a somewhat more constrained interpretation of the 
percolation of the feature \[+FOCUS\]. We shall in the following 
use~ome of the insights to be found in Roehemont.1986:84ff: 
a. If some constituent A is \[+focus\] and A is X °, then X n is 
\[+focus\]. 
b. If A is \[+focus\] and A is an argument of X ° contained in X n, 
then X ° is \[+focus\]. 
e. If X ° is \[+focus\] and A is an Adjunct of X °, then A is 
\[+focus\]. 
From a.-c. above it follows that focus percolation operates 
differently on arguments and on modifiers: \[focus\] percolates 
upwards from arguments to heads (b.) and reverse (a.),. 
However, \[focus\] percolates downwards from heads to modifiers 
(c.), yet not in the reverse direction. Now, while this is certainly 
an interesting hypothesis, it should be investigated before we 
base any implementations on it. 
From a linguistic point of view, the semantics of focus and 
presupposition determines linear sequence at ECS in interaction 
with intonation (stress assignment). For an illustration of this 
fact, ef. (17) to (20) (Jaekendoff.1972:321): 
(17) Did Maxwell kill the judge with a HAMMER? 
(18) Was it with a HAMMER that Maxwell killed the judge? 
(19) No, he killed him with a ZAPGUN. 
(20) No, it was SAM who killed the judge. 
Jaekendoff's point here is that in a polar interrogative it is the 
focus rather than the presupposition which is questioned~ and 
that therefore (20) is odd as a response to (17) or (18), both of 
which presuppose that Maxwell did the killing. 
Our point here is that (171 to (20) exhibit unmarked stress 
assignment relativ to their syntactic structures, and that marked 
assignments are possible, even if they are unlikely. As we shall 
argue shortly, the knowledge of unmarked stress assignments for 
the major syntactic patterns of a language is important in our 
context, because when dealing with written texts, stress itself is 
not accessible to us. Observe that "focus" and "stress assignment" 
are not the same phenomenon: in the case of wh-questions in 
English and German, for example, focus goes with the wh- 
element rather than with the element carrying primary stress. 
For each of the EUROTRA languages, there would seem to be 
an unmarked stress assignment to ECS (Eurotra Constituent 
Structure) translations of one ERS (Eurotra Relational Structure) 
representation, e.g. in cases of fronting," dislocation, cleft 
sentences, pseudo-claft sentences, extraposition, actives vs. 
passives etc. Assuming unmarked stress assignment, we could 
then predict the elements of sentences which carry focus. In the 
case of wh-questions, focus will always be assigned to the wh- 
element. 
Within our framework, we can introduce an attribute-value pair 
"FOCUS=yes/no" which is included into the feature description 
632 
of syntactic constituents in the course of translating from ECS to 
ERS. Thus, we translate the information which we have 
available at ECS (linear sequence, whether we have a wh- 
question etc,) to ERS. The linguistic basis of this knowledge is 
the knowledge about unmarked stress assignments on ECS for 
each of thn EUROTRA languages plus tim assumption of 
unmarked intonation throughout, which is an assumption that 
we could modify in a controlled way at a later stage. 
From ERS to IS, our FOCUS feature and its value would simply 
be carried over, which is a natural consequence of the fact that 
information about focus and presupposition is semantic 
information based on information from the constituent structure 
level. Our sentence (1) 
(1) The eec is not controversial, because it is a multi-national 
organization. 
thus receives two representations both at ERS and at IS, whereas 
sentence number (2) 
(2) Because it is a multi-national organization the eec is not 
controversial. 
receives only one represemation. In such cases, the difference is 
not only in the area of FOCUS assignment, but is also due to 
differences in scope assignment in tbe case of (1), which is 
dependent on the ECS position of the element carrying FOCUS. 
So far, we have sketched an approach to the problems of 
FOCUS, yet not automatically to problems of scope of negation 
and quantification. The reason is that in our representations at 
ERS and at IS, each sister constituent of one level has to be 
either a predicate or an argument/modifier. A negative particle 
like English "not", however, does not seem to be one or the other 
at clause level. Semantically speaking, "not" correspondes to a 
logical operator rather than to a predicate or an argument. In 
other words, our IS or ERS representations do not correspond to 
a "logical form", which would represent problems of scope in the 
usual way, e.g. (roughly): 
(NOT a(X)) AND b(Y) vs. NOT((a(X))AND(b(Y))). 
Within our present framework, a possible approach would seem 
to be the following: 
We define a feature SCOPEN=yes/no (Scope of negation) which 
is assigned to clause constituents in the translation ECS->ERS. 
As we assume that tbe scope of negation may include all the 
constituents right of the negative particle at ECS up to the 
constituent dominating the element carrying FOCUS, we could 
assign a SCOPEN value to constituents in the course of 
translation ECS->ERS and then simply carry it over to IS. The 
isssue we have to decide on in this context is whether we want 
to introduce this additional source of creating ambiguities into 
the framework. Note that scope is said to extend "up to" the 
FOCUS constituents, yet this may or may not be the case, which 
exactly is the problem in (1) above. (2) is not ambiguous, 
because the adverbial clause is definitely out of the scope of 
negation. Ambiguities are created in all those cases where we 
have constituents between the negative particle and the FOCUS, 
and in the case of longer clauses, this my cause a high number 
of readings which only differ in the scope of negation.lt is not 
obvious at the moment whether the magnitude of the problem 
created by questions of scope is sufficient to warrant such a 
step. Our treatment of FOCUS, it should be said, does not create 
additional ambiguities at this stage, it merely adds to the feature 
descriptions in representations which we are generating anyway. 
Our Appendix contains representations showing the operation of 
the FOCUS feature in ECS and ERS structures. 
633 
APPEND\]X 
A. 
ECS-representaiione 
1. Letzie Woche enischied die Kommission ueber den Haushalt 
(Last week the Commiemton decided on the budset.) 
eced/l 
s 
sbar 
comp matz 
npp : npp 
np : detp np 
app : ....... : ..... | ........ : 
ap 
a 
adj 
letzt 
n vrb de( n 
woche entscheiden d kommission 
ppP 
- : npp 
; ..... :___ 
: detp np 
: : : 
prep del n 
ueber d haushalt 
t~ee ,{cat = eep I 
focu~ : yes)) I 
.\[(ppp,\[cat = ppp 
asr = {case = acc} 
prep = ueber)) 
.\[{prep,{lex : ueber 
lu = ueber 
cal = prep 
agr : {case =acc})) 
(npp,(cat = npp 
per = 3 
asr : (case :acc 
send : mas 
nb = s9} 
Jnflec = def 
n-iyp = sub)) 
.\[(detp,(cat = detp 
aSr = {case = ace 
send = mas 
nb = sg} 
inflec = def)) 
.\[(det,\[le× = den 
lu = d 
de~-typ = arl 
cal = de( 
|nf\]ec = def 
a9r = {case = a¢¢ 
9end = ma$ 
nb = *9)))\] 
(np,(cat = np 
(nfle¢ = def 
aSr = {case = ace 
Send = mas 
nb = mS) 
n-typ = lub 
per = 3)) 
• \[(n,(lex = haushalt 
lu = haushalt 
n-typ = oub 
cat = n 
aSr : (tale : ace 
Send = mal 
nb = 89) 
per : 31)\]\]\]\]\]\]\]. 
63~ 
B. ER$-repreeenta(tonn 
ermd/! 
under 
.......................... ~ .......................... 
: compO 
: : mod 
9ov 9or 9ov 
entlcheiden komm|•m|on d 
comp4 
: comp 
: : mod 
: ! : 
9or ~ov SoY 
: : ; 
ueber haumhal( d 
mod 
: mod 
9or 9ov 
woche letz( 
History : erudl! <= ec•dll 
(••def.(ca( = • 
voice =&cilve 
(enme = pa•()) 
.\[(9ov,(sf = 9or 
C4( = vrb 
lu = en(scheiden 
frame = (CompO = (cat : np 
ca~e = nom) 
I co'mp4 = (cat = PP I 
focus = yes 
prep =ueber) 
-prep ='ueber " 
compl = no 
Comp2 = no 
comp3 = no 
Comps = no 
Comp6 = no 
comp7 = no 
Comps = no 
Comp9 = no} 
vrb-(yD = mlin 
per = 3 
nb = •9}) ..... 
(compO,(ca~ = hp 
c&se = flom 
sf = co.pO)) 
.\[(9ov,(•f = 9or 
Ce( = n 
lu = kommission 
n.typ = lub 
per = -209 
19 
prep = -21~ 
frame = (ComP2 = -2\]1 
comp4 = -212\])) 
(mode\[cat = detp 
inflec = def 
det-typ = art 
lu = d 
sf = rood)) 
.((9ov,(mf = 9ov 
cat : de( 
lu = d 
det..typ = art 
inflec = def))\]\] 
635 

REFERENCES 

Allan.K.19g7. Hierarchies and the choice of left conjuncts (with particular attention to English). in: Journal of Linguistics. 
23.1.51-78 

Arnold.D.J./ Krauwer.S./ Rosuer.M./ Des Tombe.L./ Varile. G.B. 1986. The <C,A>,T Framework in Eurotra: A 
theoretically committed notation for MT. in: Proceedings of 
COLING. 1986. 297-304 

Chafe.W.1970. Meaning and the structure of language. Chicago : The University of Chicago Press 

Chomsky.N.1970. Deep structure, surface structure, and semantic interpretation, in: Jakobson.R./ Kawamoto.S. eds. 
1970. Studies in general and oriental linguistics. Tokyo : T.E.C. Corporation 

Danes.F.ed.1974. Papers on functional sentence perspective. The 
Hague : Mouton 

Fawcett.R.P.1980. Cognitive linguistics and social interaction. 
Heidelberg : Groos 

Halliday.M.A.K.1967-68. Notes on transitivity and theme in English. in: Journal of Linguistics. 3.1.,3.2.,and 4.2. 

Halliday.M.A.K.1985. An introduction to functional grammar. 
London : Edward Arnold 

llajicova.E./Sgall.P.1986. Degrees of understanding, in: COLING proceedings. 1986. 184-186 

Houghton G/Isard.S. forthcoming. "Why to speak, what to say and how to say it: modelling language pr~luctiou in 
discourse, in: Morris.P.ed.forthcoming. Models of Cognition. 
New York : Wiley 

Jackendoff.R.S.1972. Semantic interpretation in generative grammar. Cambridge/Mass.: The MIT Press 

Jackendoff.R.S.1983. Semantics and cognition. Cambridge/Mass.: 
The MIT Press 

Jones.L.K.1977. Theme in English expository discourse. Lake 
Bluff/Ill.:Jupiter Press 

Kaplan.R.M./Bresnan.J.1982. Lexical Functional Grammar: a formal system for grammatical representation, in: 

Bresnan.J.ed.1982.~I'he mental representation of ~mmatical relations. Cambridge/Mass.: MIT Press. 173- 

Karttunen.L./Kay.M.1985. Parsing in a free word order language, iu: Dowty.D.R./ Karttunen.L./ Zwicky.A.M.eds. 
1985. -Natural language parsing. Cambridge : Cambridge 
University Press. 279-306 

Kay.M.1985.Parsing in functional unification grammar, in: Dowty/ Karttunen/ Zwicky. eds. 1985. Natural language 
parsing. Cambridge: Cambridge University Press. 251-278 

Mann.W.C.~Matthiessen.C.M.I.M.1985. Demonstration of the NIGEL text generation computer program, in: 
Benson/Greaves. eds. 1985. Systemic perspectives on discourse. Vol.l. Norwood,N.J.:Ablex 

McDonald.D.1985. Recovering the speaker's decisions daring mechanical translation, in: Colgate Univers ty. Proceedings 
of the conference on theoretical and met mdologica ssues in Machine Translation. 183-199 

Patten.T.1984. Planning and Systemic text generation: a , 
conflatlon. Edinburgh : Department of Artificial Intelligenee Working Papers. no.166 

Prince. E.1981. A taxonomy of given-new information, in: Cole.P.ed.1981. Radical pragmatics. New York : Academic 
Press.223-256 

Sgall.P./Hajieova E./Panevova.J. in press. The meaning of the sentence in its semantic and pragmatic aspects. Dordrecht : 
Reidel 

Sperber/Wilson. 1986. Relevance: Communication and cognition. Oxford : Basil Blackwell. 

Steiner.E.19g3. Die Entwicklung des Britisch Kontextualismus. Heidelberg : Gyoos. 

Steiner.E.1986. Generating semantic structures in Eurotra-l). in: COLING-proceedings 1986 

Steiner.E./ Schmidt.P./ Zelinsky-Wibbelt.C. eds. forthcoming: 
From syntax to semantics: insights from machine translation. London : Frances Pinter 

Winograd.T.1983. Language as a cognitive process. VoI.I : Syntax. Reading, Mass.: Addison Wesley 
