LINGUISTIC MEANING AND KNOWLEDGE REPRESENTATION 
IN AUTOMATIC UNDERSTANDING OF NATURAL LANGUAGE 
Eva Haji~ov~ and Petr Sgall 
Charles University 
Prague 
Czechoslovakia 
S ummary 
The necessity of and means for 
distinguishing between a level of lin- 
guistic meaning and a domain of "factu- 
al knowledge" (or cognitive content) 
are argued for, supported by a survey 
of relevant operational criteria. The 
level of meaning is characterized as a 
safe base for computational applicat- 
ions, which allows for a set of infer- 
ence rules accounting for the content 
(factual relations) of a given domain. 
1. Linguistic ~eaning 
and Factual Knowledge 
1.1 The results gained in theoret- 
ical linguistic research as well as the 
experience coming from the domain of au- 
tomatic understanding of natural langua- 
ge have convinced us that it is neces- 
sary to distinguish between two domains: 
(i) one of them has been called "the 
form of content" by F. de Saussure and 
L. Hjelmslev, "Bedeutung" or "(linguis- 
tic) meaning" bF Coseriu and others 
from European structural linguistics to 
David Lewis; (ii) the other domain - or 
set of' domains - concerns other than 
linguistic structurings of "cognitive 
(ontological) content" or "l actual 
knowleGge" (i. e. beliefs, assumptions 
and other attitudes). 
From a linguistic point of view 
the i ormer laser is understood as be- 
longing to the system of' language (lin- 
guistic competence); it consists in a 
patterning of semantic ann pragmatic 
issues b3 the given language; though 
in this level of meaning (or tectogram- 
matics) languages do not aiffer to such 
an extent as in other levels, there are 
such differences present here as those 
of the verbal aspects and tenses, of 
the restrictions of certain syntactic 
• ° constructions concerning expresslve 
power • in the sense of Keenan I or those 
of the systemic ordering of particip- 
ants, cf. Sgall and Haji~cv~ 2. The lat- 
ter layer is not immediately structured 
by the system of language, though there 
are certain types of regular correspond- 
ence, which we want to discuss later. 
From a viewpoint of research in 
logic the layer of (linguistic) meaning 
can be identified with that of Frege's 
sense , and with certain reservations 
or extensions it can be regarded as a 
counterpart of Carnap's 3 intensional 
structure4; the aspects of the layer of 
(cognitive) content studied by logic ap- 
pear there in the shape of intensional 
units (concepts, propositions, truth 
conditions, etc.). 
Other viewpoints from which the 
dichotomy should be studied systematic- 
ally are those of psychology, of arti- 
ficial intelligence, and, of course, of 
the siences studying the individual do- 
mains of (factual) knowledge. In the 
present paper we concentrate on the re- 
lationship between the linguistic view- 
points and those of artificial intellig- 
--67-- 
ence (automatic understanding). 
The necessity of distinguishing 
between meaning and content is well 
substantiated both from the viewpoint 
of theoretical linguistics, as well as 
from that of linguistic computation: 
(a) Without distinguishing the le ~ 
vel of meaning it is difficult to imag- 
ine an integrated description of lang- 
uage, since the linguistic structuring 
of semantic and pragmatic issues has to 
be described independently on what we 
assume to be the "reel" or "actual" 
structure of the world. The study of 
combinatorial properties of linguistic 
units without taking account of the au- 
tonomous level of meaning leads direct- 
ly to the skepticism known from Postal, 
as well as to what Bar-Hillel called 
"excluding canibaiism by linguistic 
means": the selectional restrictions of 
such a verb as eat would then be des- 
cribed by some framework including a 
relation defined on the set of eaters 
and on the set of eated objects, assign- 
in~. grass to horses, mice to cats, but 
not mice to horses or grass to cats,... 
In linguistic writings pursuing this 
line we find such arguments as those by 
Kuno 6, according to whom with The chic- 
ken on hi s farm is healthy it is lingu- 
istically relevant that on a farm there 
is usually moire than one chicken at a 
given time point; similarly Fillmore 7 
argues that "the wind.., is using its 
own energy", or "the wind is the direct 
cause of the door's opening". However, 
it is not directly linguistically rele- 
vant whether a horse could (or would, 
under some conditions) eat mice, whet- 
her there are usually more or less chic- 
kens on farms, whether ~an will once be 
able to use himself the force of wind 
(also for closing doors, if not only 
for driving wind mills), etc. We have 
just implicitly shown that mice can be 
used as object of eat with horse as sub- 
ject, and it is possible to find many 
such examples in the literature on 
structural linguistics. The structure 
of language itself is certainly condi- 
tioned to a large degree by the world 
we live in (as well as 03 the innate 
properties of our species), but there 
are Do immediate connections of this 
kind between individual features of the 
world (or our image of it) and individ- 
ual features of the language structure. 
An insufficient account oi the linguis- 
tic structuring of meaning has misled 
even some of the best specialists in 
linguistic semantics, as we have seen, 
and thus we consider it worth while to 
look for a more precise boundary bet- 
ween meaning and content than that 
which could have been given in the clas- 
sical structural linguistics. 
(b) In the domain of automatic an- 
alysis of natural language it is al- 
ways necessary to work with a level 
Z'unctioning as the output language of 
the analysis procedure. I1 we are 
speaking about understanding natural 
lan~uage (i'or such purposes as question 
answering, machine translation, man- 
machine aialogue or other aims within 
the area of artilicial intelligence) 
rather thanabout mere surlace parsing, 
then the output of the analysis is re- 
quirea to Dear a aisambigmatea inIor- 
mation; a language must be defined for 
this purpose which can get a semantic 
interpretation (in Carnap's sense); 
this must be a language the (element- 
at# and complex) units of which are un- 
ambiguous. However, they cannot be ful- 
ly relieved of vagueness or indistinct- 
ness (this concerns not only hedges or 
fuzz# units, but also the indeterminacy 
of reference, which is removed in human 
- 68-- 
discourse by mechanisms some of which 
are of a linguistic nature, but all of 
which are pragmatically based, cf. § 2 
below). @he distinction between ambig- 
uity and vagueness (indistinctness) 
belongs to the distinction between me- 
aning and content: a linguistic expres- 
sion is ambiguous iff it has more than 
one meaning; a linguistic unit is vag- 
ue iff it is a unit of meaning corres- 
ponding to two or more units of a rel- 
evant structuring of the layer of con- 
tent. In most systems of automotic un- 
derstanding the domains of meaning and 
content are not distinguished, and this 
fallacy leads to two ma~or difficulti- 
es: First, no clear criteria could ha- 
ve been found for a classification o£ 
units of the "cognitive" domain, be it 
described in a form of nets, frames, 
scripts, or by another' means of "know- 
ledge representation'; only for the 
classes of texts belonging to one of 
the "exact sciences" it i8 possible to 
use the structuring elaborated within 
the competent science (mathematics, 
chemistry), but even there this does 
not cover consistentlF the requirements 
of the analysis of those parts of texts 
which are concerned with motivation and 
background analysis. Second, and most 
important, the structuring of the layer 
of content that is made by the method 
of trial and error, in the experiment- 
sl systems, often leads to the neces- 
sity to postulate more and more subtle 
structuring; thus e. g. for the Fill- 
morean case roles it appears that every 
small group of verbs (of saying, of 
perception, of movement, of simple phy- 
sical actions, of puz~chase, etc., etc.) 
has its own set of roles: no element 
of the set {buyer, seller, goods, pri- 
c~ is identical with any element of 
the set ~speaker, addressee, object 
spoken of, type of messag~ , etc. (see 
e. g. Fillmore 7 quoting Cole). Thus 
it seems there is no boundary that 
would ensure the possibility to descri- 
be the structuring(s) of content by fi- 
nite means. Every system of natural 
language understanding then has to be 
restricted to a certain domain and the- 
re i8 no guaranty that the oasis of the 
system would have to be rebuilt if the 
time comes to apply the system to an- 
other area. 
On the other hand, when the struct- 
uring of meaning inherent to natural 
language itself i8 well understood and 
appropriately used, then the universal- 
ity of natural language (which allows 
its users to express everything they 
can think of, with the necessary degree 
of precision) gives at least a common 
basis for the most divergent domains of 
cognition (or types of texts), frcm 
science to pop-music, and that only the 
mechanisms accounting for the reldtion- 
ships between the (comm~on, general) 
linguistic meaning and the (specific, 
more or less ad hoc) factual knowledge 
of the given area will be to a certain 
degree specific to this area. 
1.2 However, is i~ actually pos- 
sible to find a clearly specified boun- 
dary between meaning and content, to 
find operational criteria showing what 
distinction belongs to the level of me- 
aning? As. H. Putnam's account of lex- 
ical meaning has shown, there is a cer- 
tain "division of labour", connected 
with individual and temporal differen- 
ces of the boundaries between meaning 
and content. However, some basic lay- 
ers of terminology (e. g. the kinship 
terms) may serve as an evidence that 
even in the lexicon there are clear 
cases in which the knowledge of a ~iven 
• 69 
meaning (within linguistic competence) 
is not intermingled with requirements 
concerning factual knowledge. Also the 
possibility to find fully synonymous 
pairs of' words (connected with a mere 
stylistical, non-semantic difference) 
and distinguish them from others cor- 
roborates the view that Putnam's "divi- 
sion of labor" is a symptom of individ- 
ual difference in a language community 
rather than of an absence of a differ- 
ence of principle between meaning and 
content. :In any case, with respect to 
grammatical relations (expressed by 
function words, endings, word order, 
etc.), the distinction between meaning 
and content can be established on the 
base of criteria that have been elabor- 
ated and explored in the classical pe- 
riods of European structuralism, as well 
as more recently by Keenan 8, by Zwicky 
and Sadock 9 and b~ others. 
None of these criteria can be cla~ 
med to have an absolute validity: the 
old maxim according to which only phra- 
ses of the same syntactic value can be 
coordinated does not hold 1'or such ex- 
amples as here and now or for the sake 
of A and in spite of B; the tes~ used 
to distinguish topic and focus by means 
of question or negation are not immedi- 
ately useful for interrogative senten- 
ces; the requirement that the speaker 
must know which of the two meanings of 
an ambiguous expression he "had in mind" 
meets ~ifficulties in connection with a 
first person subject (having said I rol- 
led down the hill the speaker of course 
knows whether he acted as a conscious 
agent, or only as an "experiencer" or 
passive objective, but with John rol- 
led.., the situation certainly is not 
the same); Panevov~'s 10 "dialogue test 
works better with aaverbials than with 
the "inner participants" (having saia 
John is coming the speaker is expecteu 
to know where to; but with He has paid 
~lre~ it is not clear whether fo_.r 
w~a__t has to be understood as deleted, 
i. e. known); the systemic ordering of 
the modifications of verbs 2 may be used 
as a useful means to distinguish bet- 
ween different types of modifications, 
but the results are not always of the 
same degree of certainty, etc. ~very 
such - or another - type of a "diagnos- 
tical context' may be considered highly 
useful, even if in some cases it does 
not give clear results. It has been 
possible to use such criteria to estab- 
lish clear notions of the obligatorin- 
ess of adverbials (see the "dialogue 
test" just quoted), of the topic/focus 
articulation, of presupposition and 
allegation, of the scope of negation~ I'12 
Also an operational criterion for iden- 
tifying strict synonymy of grammatical 
constructions has been formulated, which 
makes it possible to combine empirical 
research in linguistic semantics with 
the theoretical framework of truth con- 
ditions and possible worlds 4, though 
many linguists doubted the possibility 
to connect these two domains (cf. the 
notions of "internal" and "external" 
semantics in Fillmore. 7 
2. ~ethods for s General 
Account of Linguistic Meaning 
2.1 According to the criteria cha- 
racterized in 1.2 a repertoire of units 
of the level of meaning (structured more 
subtly than truth conditions are) and of 
relations between them has been establ- 
ished. A generative specification of 
this level was discussed in our paper 
at Coling 1978 in Bergen. 13 
The meaning of a sentence can be 
represented by a rather simple tree (in 
-70- 
accordance with the traditions of Eu- 
ropean linguistics we prefer dependency 
to categorial or phrase structure gram- 
mars) with the following properties: 
(a) the tree has a single root, is 
finite, connected and projective (cf. 
Hays 14, NarcuslS); 
(b) the edges are labelled by the 
types of modifications, which are lis- 
ted partly in the lexical (not only 
verbal) frames of the "governing" lex- 
ical unit, and partly in a list of free 
modifications (adverbials), common to 
all the units of a given part of speech; 
besides the Actor/Bearer (rather than 
Agentive, see Haji~ov~ 16) the verbal 
frames may contain the Patient (Goal), 
and, if these participants both are 
present in the frame, then also the Ad- 
dressee, the Origin and/or the Zffect 
may be included there; Instrument, 
~mnner, Measure, various types of Loc- 
ative, Duration, Cause, Condition, etc. 
belong to the list of free modificati- 
ons; they can occur with every verb - 
at least in principle, i. e. are not 
excluded linguistically - and they may 
occur even more than once with a single 
verb token; they have to be listed in 
individual frames only if they are ob- 
ligatory with the given verb; with 
nouns, the General Relation is a typic- 
al free modification, while the Patient 
has to be included in the frame of such 
nouns as directQr, etc.; 
(c) the nodes are labelled by com- 
plex symbols corresponding to leXical 
and morphological meanings (the latter 
comprise tense, aspect, modality and 
others with verbs, numoer and delimit- 
ire features with nouns, degrees of 
comparison with adjectives); 
(d) the "left-to-right" order of 
the nodes is interpreted as the "deep 
word order" or communicative dynamism, 
which corresponds to the order of quan- 
tifiers in formal languages; on this 
scale the boundary Oetween topic and 
focus can be established (primarily 
just before or just after the verb); 
the scope of negation i~ identical with 
the focus in the unmarked case. 
2.2 The level of meaning contains 
semantic and also pragmatic units (in- 
dexical elements, modalities, tenses, 
etc.). Also the topic/focus articulation 
and the hierarchy of dynamism are prag- 
matically based: only such items can be 
used as contextually bound that have 
been activated by the context, i. e. 
have a great degree of salience in the 
stock of "knowledge" shared by the spe- 
aker and the hearer; also definite NP's 
in the focus meet such a requirement, so 
that their referents may be identified 
by the hearer on the base of the state 
of his model of the world in the given 
time point of the discourse (cf. Barba- 
ra Grosz" "shifting of focus ''17). In 
these questions the study of the struct- 
ure of natural language should be con- 
nected with psjc\[\]ologically oriented in- 
vestigations into the structure of human 
memory. 
Also the connections between mean- 
ing and intensional logic are being 
studied. The linguistic (s~ntactico-se- 
mantic) analysis translating sentences 
to their semantic (tectogrammatical) re- 
presentations is combined with a proced- 
ure translating these representations to 
a formal language based on the theory of 
types; meaning postulates are used in 
this procedure, which also converts the 
patterning of obligatory and optional 
modifications (dependent words) into 
structures connected with the arity of 
predicates; furthermore, communicative 
dynamism is transferred here to the 
--71-- 
usual form of denoting the scope of the 
quantifiers. 
The linguistic description itself 
has a generative power moderately ex- 
ceeding that of context-free grammars, 
according to a scale constituted by a 
18 sequence of pushdown transducers . 
3. Approaches 
to Knowledge RePrgsentatiog 
(compared with the expressive 
power of natural language) 
3.1 There are many different de- 
grees of complexity connected with the 
representation of data (information, 
knowledge), from simple data bases 
through more sophisticated ones to cogn- 
itive networks and other kinds of appar- 
atus. Most of these approaches are bas- 
ed on experimental research in a rest- 
ricted domain and thus connected with 
different kinds of rules of thumb and 
ad-hoc devices, so that it can be never 
taken for sure whther a broadening of 
the investigated domain would not re- 
quire a radical change of the approach. 
• c 19 In Schank s conceptual dependen y , 
lot instance, there are five "cases" 
(actor, objective, recipient, directive, 
instrument) as conceptual categories of 
relations between actions and nominaL, 
in addition to another set of relations 
of' specification, understood as a rel- 
ation between action and modifier and a 
yet another set of relations between 
the concept categories of nominals and 
moailiers. There are 13 types of prim- 
itive acts (e. g. physical transaction, 
mental transaction, etc.), each oZ which 
has a frame ascribed to it, which spe- 
cifies the set of "cases" necessarily 
present in the conceptualization, even 
if not in the surface structure of the 
sentence identifying the given action. 
An actor of a physical transaction is 
understood to be a relation different 
from that of an actor of a mental 
transaction; thus one arrives at a num- 
ber of 80 different case relations in 
addition to a great number of specific- 
ation and state relations. 
Such a system in fact works with 
different case roles for very small 
groups of verbs; in other words, the 
roles sre then "word-specific". 
Simmons "20 conviction that John, th__~ 
machine and the brook in the sentences 
John ran to school. John ran the machi- 
ne. The machine ran. The brook ran all 
are instances of "causal actants" and 
that the specification of "instruments" 
with which the action of "running" is 
performed should be specified in the 
lexicon for di\[Iferent meanings of the 
verb "run" rsther than regarded as a 
matter of the semantic representations 
(semantic networks) of these sentences 
may be quoted as a support for the view 
that even when attempting at a classific- 
ation of units and relations within a 
system of knowledge representation, one 
should carefull$ observe the properties 
of language structure itself. 
Also Schlesinger 21 duly recalls 
that one should distinguish the domain 
of cognitive structures (with a great 
variety of distinctions) and the level 
of semantic deep structures, with a li- 
mited number of deep structure relationa 
In the domain of cognitive structures, 
he speaks about a conceptual continuum 
(or even about a multidimensional cognit- 
ive space) which each language segments 
in its own way (a reminder - without a 
specific reference - of the well known 
Hjelmslevian approach). 
It might be objected that more can 
be inferred from a representation inclu- 
ding the (word-speoifio) roles. The 
72 
object of such verbs as mak_.e, build re- 
fers to something which comes into exi- 
stence through the action denoted by 
the verb, and this fact is not captured 
by a notation handling the object of 
these verbs simply as 8 Patient, i. e. 
in the same way as that of see, bit, 
etc. However, in many cases the infer- 
ence that the o00ect exists after the 
action cannot be based immediately on 
the verb itself (we have not only 
a. picture, but also aint~, not 
to speak about painting a fence); with 
such outspoken cases as build the lex- 
ical meaning of the object noun perhaps 
is not relevant, but the modality of 
the verb is (if one wants to build a 
house, one may fail to do it). It ap- 
pears that in any case the formulation 
of adequate inference rules of this 
kind requires a classification of lexic- 
al as well as grammatical meanings. 
3.2 Under the given conditions a 
certain amount of trial-and-error work, 
analyzing one lexical unit after anot- 
her without an explicit statement of 
general criteria, is more or less ine- 
vitable in'the domain of content(ex- 
cept for the regulerities known by the 
science concerning the studied domain). 
It seems, however, advisable to use a 
relatively complete analysis of the 
structure of natural language as a base 
from which these or other parts may be 
chosen for a given application, ensur- 
ing that such simplifications can be 
replaced by a fuller specification if 
this becomes necessary, since the base 
is universal, in the same way as natur- 
al language is. The empirical investig- 
ation of natural language syntax and 
semantics, using a formal framework, 
thus appears to belong to the most im- 
portant preconditions of natural langu- 
age understanding. 
Also the use of such notions as 
topic and comment or focus in connecti- 
on with question answering 22'23 sup- 
ports the view that a deep understand- 
ing of the structure of natural langu- 
age is of crucial importance for the 
linguistic aspects of artificial intel- 
ligence. 
~ Ru!eg_of Inference 
(as a means to account for factual 
knowledge in automatic understanding 
of natural language) 
4.1 The above considerations have 
led us to the conclusion that rules of 
inference operating on the representat- 
ions of the meaning of sentences (cf. § 
2.1 above) may be useful to handle the 
relationships between meaning and con- 
tent ("factual knowledge"). Such rules 
of inference are used in connection 
with the method TIBAQ (Text-and-Infer- 
ence Based Answering of Questions) by 
the linguistic group of Charles Univers- 
ity, Prague (Fac. of Mathematics and 
Physics); 
An experiment has been prepared in 
the form of algorithms that are being 
implemented now (in Q-language, PL-i, 
and Assembler), the aim of which is to 
show that a relatively rich analysis of 
natural language makes it possible to 
construct fully automatic question ans- 
wering systems based only on input texts 
and factual questions in natural langu- 
age. A set of inference rules is includ- 
ed, operating on the representations of 
the meanings of the input sentences. 
These rules range from general ones to 
more or less idiosyncratic cases con- 
cerning the relationships between spec- 
ific words, as well as modalities, h~po- 
nym,y, etc. 
--73-- 
A rather general rule changes e.g. 
a structure of the form (V-act (l~Acto r) 
...) into (V-act (DActor)(Rinstr)...), 
where V-act is a verb of action, D is 
a dummy (for the general actor) and N 
is an inanimate noun; thus The negative 
feedback can_servo the voltage to zero 
is changed into One can servo the volt- 
a6e to zero by ..... A rather specific 
rule \[connected with a single verb) is 
that changing (us_..~ (Xpatient)(YAccomp) 
...) into (use (XRegard)(ZPatient)...), 
e.g. An operational amplifier can be 
.... , .. ,, 
used with a negative feedback = ~fith an 
o~erational24uplifier a negative feedback 
can be used. Other similar rules concern 
the division of conjunct clauses, the 
possible omission of an adjunct under 
certain conditions (i.a. if not being 
included in the topic), as well as the 
connection of two sentences one of which 
can be embedded into another. 
In the look-up for an answer in 
the set of assertions (enriched by the 
rules of inference) we have formulated 
substitutions, some of which again are 
general (e.g. I~!anner is considered as 
substitutable by Accompan~nent or by 
Effect, Place by Regard), others being 
restricted to individual verbs: use how 
may be answered by use for (purpose),etc. 
4.2 Besides the kinds of rules 
illustrated above it is also necessary 
to study (i) rules standing closer to 
inference as known from logic (deriving 
specific statements from general ones, 
etc.), (ii) rules of "typical" (unmark- 
ed) consequences as given e.g. by a 
"script" , and (iii) rules of "probable 
consequence", e.g. if John worked hard 
in the afternoon and he is tired in the 
evening, then the latter fact probably 
was caused by the former (if no other 
cause was given in the text). In our 
experiment of question answering we do 
not use these types of inference, but 
they will be useful for more general 
systems. 
A broadening of the scope of the 
experiment (which now concerns a sub- 
domain of electronics) will require con- 
siderable modifications of the inference 
rules, since in the domain of the relat- 
ionship between meaning and content we 
are entering a new domain, the regtllar- 
ities of which have to be studied joint- 
ly by logicians, psychologists, ling- 
uists and specialists in artificial in- 
telligence. However, the linguistic proc- 
edures will have to be enriched mainly 
with respect to the lexicon, where new 
questions of principle would- not arise, 
if the questions of gra,~nar have been 
handled adequately. Technical texts 
written with a necessary niveau of clear 
formulations, carefully defining newly 
introduced terms and distinguishing de- 
finitions from assertions, can be well 
"understood" by such a linguistically 
based system. This means that Karlgren's 
systems 24 of the third type, using the 
usual human expression as input and pre- 
senting their output in natural language, 
are already feasible° It is possible to 
attempt seriously at a solution of one 
of the main tasks of linguistics: to 
conform the automatic information syst- 
ems to the usual way of life of human 
beings. 
The structure of natural language, 
including its patterning of the units of 
meaning, has to be empirically studied 
and explicitly described. The ambiguit- 
ies and irregularities inherent to nat- 
ural language may then be removed, while 
its flexibility (com~ected with a necess- 
ary ~nount of vagueness) is retained. In 
such a way natural language understand- 
ing can be given a sound general basis. 
This view is supported by ~Jilks 25, who 
74 
duly argues that the formulae in know- 
ledge representations should represent 
the meanings of words, and nothing else: 
a man knows about the real world nothing 
more than can be expressed in natural 
language (giving examples of the verbs 
to break - which need not have an Instr- 
umental, and to~, which should have 
one), 
Also RitchJe 26 comes close to this 
standpoint saying that hypotheses "must 
be based more on the actual patterns 
within language, rather than on current 
dogmas ..."; he is also right in point- 
ing out the usefulness of choosing sem- 
antic categories generally applicable, 
thus avoiding the risk of having to con- 
struct "as many different analyzing gram- 
mars as there were domains of discourse". 
It seems to be justified to comb- 
ine the study and description of langu- 
age with those of the domain(s) cover- 
ing the given area of content and of 
psychological phenomena, to be able to 
construct systems of general applicab- 
ility as well as to reach a better under- 
standing of cognition. 

References 

i E.L.Keenan, Logical Expressive Power 
and Syntactic Variation in Natural 
Language, in: Formal Semantics of ~L 
(E.L.Keenan, ed.), Cambridge 1975, 
406-421. 

2 P. Sgall and E.HajiSovg, Focus on Foc- 
us, PBML 28, 1977, 5-54; 29, 1978,23-42. 

3 R.Carnap, Meaning and Necessity, Chic- 
ago 1947. 

4 P.Sgall, E.Haji~ovg and O.Prochgzka, 
On the Role of Linguistic Semantics, 
Theor.Ling. 4, 1977, 31-59. 

5 P.M.Postal, The Best Theory, in: Goals 
of Linguistic Theory (P.S.Peters,ed.), 
Englewood Cliffs, 1972. 

6 S.Kuno, The Structure of the Japanese 
Language, Cambridge, Mass., 1973,P.43. 

7 C.J.Fillmore, The Case for Case Reopen- 
ed, in: Kasus Theorie, Klassifikation, 
sem. Interpretation (K.Heger,J.S~Pet~fi, 
eds.),Hamburg,1977,3-26,§4. 

8 E.L.Keenan, Some Logical Problems in 
Translation, in: Meaning and Translati- 
on (F.Guenthner, M.Guenthner-Reu~er, 
eds.),London 1978,eSpop.173. 

A.M.Zwlck y and J.M.Sadock, Ambiguity 
Tests and How to Fail Them, in:Syntax 
and Semantics 4 (J.P.Kimball,ed.), 
New York 1975, 1-36. 

lO J.Panevovg, Inner Participants and 
Free Adverbials, Prague Studies in 
Math.Li~g~ 6, 1978, 227-254. 

ll E.Haji~ovg, Negation and Topic vs. 
Comment, Philologica Pragensia 16~1973, 
81-93; also in P.Sgall, E.HajiSova, 
E.Bene~ovg, Topic, Focus and Generative 
Semantics, Kronberg/Taunus 1973. 

12 E.Haji~ov~, Meaning, Presupposition and 
Allegation, Philologica Pragensia 17, 
1974,18-25;reprinted in Functional 
Generative Grammar in Prague (W.Klein, 
A.v.Stechow, eds.), Kronberg/Ts.,1974. 

13 E.HajiSovg and P.Sgall, A Dependency- 
-based Specification of Topic and Foc- 
us, SMIL 1980. 

14 D.G.Hays, Dependency Theory, Language 
40, 1964, 511-525. 

15 S.Marcus, Sur la notion de projectivit~, 
Ztschr.f.~ath.Logik ll, 1965, 181-192. 

16 E.Haji~e~g, Agentive or Actor/Bearer? 
Theoretical Linguistics 7, 1980. 

17 A.E°Robinson et al., Interpreting Nat- 
ural Language Utterances, TN 210, SRI 
International, Menlo Park, 1980. 

18 M.Plgtek and P.Sgall, A Scale of Context- 
Sensitive Languages, Information and 
Control 38, 1978, 1-20. 

19 R.Schank, Identification of Conceptual- 
izations Underlying NL, in:Computer 
Yodels of Thought and Language, San 
Francisco 1973, 187-247. 

20 R.Simmons, Semantic ~etworks, in:Comp. 
Models of Thought and Lang,dtto,63-113. 

21 I.M.Schlessinger, Cognitive Structures 
and Semantic Deep Structures, Journal 
of Ling. 15, 1979, 307-324. 

22 H.Karlgren and D.Walker, The Polytext 
System, Stockholm 1980. 

23 K.R.McKeown, Paraphrasing Using Given 
and ~ew Information, Philadelphia 1979. 

24 D.Walker and H.Karlgren, paper submitt- 
ed at Visegrad meeting on Question 
Answering, 1980. 

25 Y.Wilks, Preference Semantics, Stanford 
A.I.Memo 206, 1973. 

26 G.D.Ritchie, Predictions and Procedures 
in Semantically-Based Grammar, in:Proc. 
of AISB Conf., Hamburg 1978, 273-282. 
