(}g~ ~he X~ltc=~.-£'a¢~t:.i~}~:x of Syrrta~c and Semantics 
ir~ a S-T--nta.~<tica, i.\]:y Gxai~led (;~tsef.~-ame .Parser 
Harald Trost, Ernst Buehberger, Wolf gang Heinz 
Austrian Research Institute for Artificial Intelligence 
Schottengasse 3, A-1010 Wien 
Austria 
Abs%ract: 
in this )~aper we describe a parser for German 
based on semantic caseframe instantiation 
guided by a syntactic analyzer o Pure 
casefra,m~ parsers lack the ability to capture 
syntactic ~=egularities~ which leads to ~:edun- 
dancy :I.n the lexicon and/or poor syntactic 
coverage, By combining caseframe matching 
with an explicit syntactic analysis our parse:c 
overc',omo;:~ this problem° 
Approach~)s w~l\] su:Lted fox." l~nq'lish are not 
~\]asiiy :transported to German with its z'ich 
ltlorpholo,(\]y and its If:CO0 constituent order at 
tho c\].atlse J_e, Ve!o Our parser which :{nco:r.'.- 
~ex'ates two d:i fferen% interacting parsing 
stx'ategios is we\].\], adapted to the needs posed 
bv German ¢~rammaro 
we believe that the present understanding of 
structural differences between languages does 
not yet allow for a single parsing algorithm, 
at least if one wants both good coverage and 
efficiency° As a consequence we developed a 
parser which is specifically designed to cope 
with the peculiarities of the German language. 
Nevertheless, since our approach is based on 
sound linguistic principles, most of the solu- 
tions found could be applied to other 
languages with a similar structure as well. 
In this paper' we will focus on the core of the 
system's parsing component and neglect other 
features like spelling correction, treatment 
of anaphoric or elliptic utterances, quantif- 
ier seeping and the transformation into SQL. 
The overall system architecture is deDicted in 
\[, ..o~o~.,,: I ..... 
I: F4orphuloglCUllhiah,~!:sls :: :::':" Spellli~g correctltm 
Pnrser 
.,rot $07ut II\[\]n 
Interpretotlnn 
i?: • 
query genordlor 
Fig° I : .qysLem Architecture 
an~mer oeneratlon 
parnphrase 
\[IBM$ :F2_ 
DO=I etallons 
an~tv~r 
1 o T,n~tt¢oduc%,ion 
DB~DIALOG is a German language interface to 
relational data bases° Our objectives were: 
to design a system which has good language 
capabilities and which at the same time is 
easily portable. The system has bee~ 
developed on a SYMBOLICS 3600 and up to no~ 
has been transported to a PRIME-550II~ a DEC -i 
VAX 730, and a NIXDORF TARGON-35o 
DB-DIALOG translates user-queries given in the 
fo~m of written German sentences into struc- 
tured SQL statements° Since SQL is the de- 
facto standard query language for relational 
database systems, a wide range of database 
systems is accessible. The only adaptation to 
be done is a transformation uf the structured 
SQL output by DB-DIAI.OG into the special SQL 
used by the spscific DBMSo At the moment ver- 
sions for ORACLE, REFLEX and MIMER are imple- 
mentedo 
in some other ways the interface is also 
designed to be as portable as possible° Adap- 
tatlon -to new domains is facilitated by keep- 
in q the linguistic, coverage separate from the 
actual domain knowledge which rests solely in 
the \].exlcon. Independence from the modeling 
of the domain in the data base is achieved by 
distinguishing between a linguistically 
motivated description of the domain and a 
database~,orlented one. Tllsre is an explicit 
"translation step between these two parts. 
Language independence is not aimed at, because 
figure io For a description of the interface 
as a whole see Buchberger et ai./1987/. 
We have chosen to base our parser on semantic 
caseframe instantiation. Such an approach is 
well suited for a restricted domain parser, 
because of its efficiency (by avoiding useless 
parses in case of syntactic ambiguity) and its 
robustness in the case of ungrammaticality 
(see eog. Grishman et a\]./1986/). On the other 
hand, relying solely on that method, it would 
be difficult to capture syntactic generalities 
(Cog.active-passive transformation), because 
syntactic as well as semantic restrictions 
must be specified explicitly for each slot of 
overycaseframe. This means that for every 
different syntactic realization of the same 
statement a different caseframe has to be pro- 
vided in the lexicon. There are two severe 
drawbacks to this kind of realization: First, 
general syntactic properties of the language 
are implicitly stated in the lexicon entries 
instead of explicitly in the grammar leading 
to a possibly inconsistent and patchy syntac- 
tic coverage°' Second, the lexicon is inflated 
because for a single word (meaning) a number 
of different caseframes is needed. 
To illustrate the problem let's have a look a% 
an example: 
'liefern' (= to deliver) 
could have the following caseframe: 
(LIEFERN 
(AOENT (SYNTAX NP/NOM) (SEMANTICS COMPANY)) 
(OBJECT (SYNTAX NP/ACC) (SEMANTICS GOODS))) 
677 
This w<)u 1. dl e;~ab:l e a p a~.'se:c to a~)a\] yz~i~ 
se\[lte~zo¢~ .! :i,\]{:((~ J 
FJ:cmo WV'/, 1.iefe:ct Kuehlsoh~'aenkt~ 
(X~ZZ C.ont~fi~.t'f){ del:J.?ers wefEi{~oratoy'~ ~ ) 
Bill; "\[;hO~::'~ a:c6~ vaciotl:3 syl\]th, c~,:ie Va:CJ. 6/fJ OnA3 of 
this statemei~t; 
- passiv:kzatio~ 
DJe Kuehlsc:h:,:'aeuke we~:del~ yon dec "J.Fnia XYZ 
gel iefe:ct ~ 
(The refzi.qe:cal;c:<s are de\]ive~:ed by -the XYZ 
company. ) 
,- re:l.atiw3 clausc~ (aci:ive and passive) 
Die Kueh:l schraerlke, die die F:L ~:rlla XY'Z 
:1 ie:Le:c't, , o o 
( Th{9 :fie f \].'J.ge t'a t()z s, that XYZ rJOm:{) ~tl ly 
delivers, . • ) 
Die Fiz'ma XYZ, die, Kueh\].sehraenke 
\].iefect, o.o (XZZ company, that delivers 
x-efrite.cato:cs o o o ) 
o- attributive (actLve and passi.ve) 
Die vc)n der f, i_,:ma XYZ gel iefe:Ften 
I(uehlsch:)raenke. , o 
( The ~cefrige~'atoz's delive:ced by the XYZ 
company° . o ) 
Die Kuehischraenke l:\[efernde Fi:cma XYZo o o 
( The XYZ eonipauy de .I i.ve:.,:i(kc\] 
re:\[z'igerator<~..o ) 
AS the examp\].e shows, there are six different 
syntactic forms which *nay occuz' with the same 
verb due i:o syntactic va£iationso flaying ~;ix 
different ease fz'ames for just one word 
(meaning) :iS rJdicu:!ous,. Several 
_improvements have been proposed which enable 
caseframe pax'sets to deal with syntax in a 
less ad..heo manne~', see eog,,Hayes st 
alo/1985/ and Lytinen /1986/o 
In OU:t:' approach we wenL On~ s-hap :ffurthe~' in 
this d/-<'so i:ion by combin\[ng a oasef~:'ame 
marcher wi \[:h a ~.;yntsc:-t Lc • parsez" that i:::; 
guJ.d/.~l(~' \[:h(~ f{tle\] ,¢?: i.~3 ;'.\]:Fecess aS ~:~. whole 
tn te:c >.to i.i o'{~ with i::h<i5 c, ase ft amc:; ma tch,e~. 
J.llclttde~{ {;r{.'.nsforiiiatic.tlL{ bo:l.~ltl app!iod to the 
s}zntact7 e z'es'trietions of -the casefram~s 
/nvo Iv~.-.d That way di f :f.:?:<'c~n t s~.tizaot io 
(sonstruct i< )tY.t3 1 {.\]c~5 Oo Jr; J.v~, patn:;i.v~3 ¢ 
attributive, ~-:e 1. at i.ve clause and ~ven 
nominal J :,,at ion~:; are handled wJ.tho<l h "the 
need for different casef~.'ameSo 
2. Lantuatd~ Specl.f:i.<:; )bc, peCtS 
Using Ge:cms)n as input language to our 
interface calls fore solutior~s to p:c'obloms 
which de not ari.s<~ io:m the F, nglish \].anguag,a. 
The most prominent differences are: 
- there is a :v:i..eh mo:,pho:l.ogy, 
-. constituent order at the clause level in 
fairly free, and 
-. there :i.s the v~rb-sec, ond phenomenon in main 
clauses o 
Morpholoty is dealt with in tile mo:cphc)logical 
componerrt of the scanner /Tz'ost ai~.d Do~ffne~ 
1987/. Th:i..s scanBe:c pas~;t-_',s :!.nfc, xmai;ion abotrt 
case markers ( a.~; we\].:l, as o-ibex" ;~wntactic 
features) i;o the; pax.'~ec, but -- if interpz'eted 
locally ... this i.nfo:¢mat:t.on :!.s usually highly 
ambiguous o 
AS for word order, ther{{ ar<~ basieal\]}Z two 
phrase 'types in German: noun-dependent 
phrases, like no(:~n phrase ( NP ) and 
prepositional phrase ( !:'P ), with a :cathez" 
~isid word oz'de~r, and eZ,~uso--like phz'asss, 
Ni.ke se, ntenco (S) and adJective phrase (AP)~ 
with at best a prefer:cad ordez-in 9 of 
constituentSo Fo~" a discussion of wo~d order 
J.~ G~z'~nan of,Ho¢~hle /1982/ and~ fo~< ~ n~oro 
678 
"computationallv oriented view U~;zko:r-£1£ :f t ~ 
/i98{,/ and Ilauenschlld /.I.986/~ 
Closer inspect:Lon shows that o~\] th~ ca~o hu.ad 
part of the NP,~ namely APe embedded :in the~% 
exhibit free constftuent order~ whoreav~ on 
the othez' hand c\].ause.-like constz'uc fiords 
appear to have one fixed position: tt).,, h~ad 
(vez-bal complex and adj act.lye respaegJv(:<\[y) 
which always comes :Last. The:,-o :J.;::~ the o~!j( 
exeept:Jon that in main clauses the :I.~,.:if:tect~:M 
part of the verbal com\[31s3g f~love~q to f3k-ll.3ol-~/! 
position /Haider 1.984/~, 
In parsing a language llke Oezma~ ,,~ke 
thsrefore needs two different (co~)t;cadi~;;::L~-wi.) 
sirsreties : 
,. ene for the fixed werd order' of a:cqume~ts~ 
inside constituen-ts ( i o e o determine); ~:l 
att:r:/bute of NPs) 
-- one fe:c the free constituent o~:dez' of the 
axguments and modifiera~ o± predJ.<~a%o~\[; 
( i0 e.etls censt.t.tue~H:s of g ), 
0u!: solution to thJ.:3 pcoblem i:J i:he 
iuteract.ion el:' two different teehn.Lqu~u :\]u 
our parssr~ For processin 9 constituents with 
fixed word order we chose the Aug i oxltod 
Transition Network ( A'fN ) formalism / B al <~.s 
1978/~ because ATNs are a wall undGrstoo¢i 
algorithm with very effieient imp:l-ementatior; 
techniques awailab\]_e~ and they prov:Lde fo:c a 
relatively transparen-c notation° Si~'~s~ w~; 
use the ATN only for a part of tile synl:ast:l.¢~ 
parsing which itself inteK'acts closely with 
semantics, the known weaknesses inherent to 
ATNs do not pose a problem in the centext of 
our parser~ For free.-order eonstit:uerFt~ on 
"\[:he other hand we use a uni:\[ieation-basod 
~;t:rategy whioll makes heavy use of a ca:~e 
fz'ante matdher., We will first desc~':Lbe both 
,~omponents in some detail and ",*:he~ 
de~aoi~strato how they J.nte:cacl;o 
Our ATN consists of the ~isual subnsts fo~' 
phraseo..types (NP, AP, ' ~, etc. ). in con't~'ast 
to i:he standard approach it works on a chart 
of mox pholo~ical entries created by -th~ 
morphological component mentioned ea~'li.~x." o 
This chart may contain ambiguities which the 
ATN is extended to cope with° 
Since the ATN aims at the con~3t~'uct.lo~-~ of 
funotional dependencies (an argum~nt/nlodifi~r 
- head structux'e)which is greatly eased by 
knowing the head /Proudlan and Pollard 1985/~ 
we decided to nee head-driven analysis :in "th~ 
ATNo German Ms basically a sub, eat-oh.Jest .~ 
verb (SOV) language, that means the head of a 
phrase comes last wi~h few exc~ptJ.ons, Th¢_;~:;e 
exceptions are : 
- NPs may have posl;modlfi<-~rs (~enltive NP:~,, 
PPs, zslative clauses)~ 
J,~. PPS "the propositio~z comes i~ the fi:¢st 
po~.~it.lon ,. 
• the above men~tio:o.~d vo~b~,~co.~-~d pheJnome~o~i~ 
in ~aain clauses° 
~qith a slightly different view on ~ih\[c~H~ 
~laTacture all three of -th~s~ ~x~kJ% io~ 
disappea~'o Let' s for the moment just a~anm~ 
• that the head a\].wa~s com~s in %h® Zaat 
position o Then it proves advantageous i;© 
choose a xi~hto.to--:i.eft orde:c fo~ prooes~H..~~ 
s~nt~noos o There are ~-3evera\]. int ez~¢~ £~,t i~%~ 
{~ca~,':~e)ql~ence,,:~ of this decision~" 
there i~ no need for a separate PI-<,-u~t~bn~t~ 
PP and NP are collapsed into on~ ~Dubn~t a~d 
%h{~ px'e~oaition .~ If found at th~ ~:<d' of 
t:b\[~ ph:case ~- l,':~ simply viewed as a semantic 
c~e marke:c. 
,-~ adJu~<;t~ to the x'igl-lt of a phrase head have 
to be parsed separately° In our case: 
Postmod~fiers like PPs, genitive NPs and 
~:olativ~ clauses modifying NPs a~o not 
inclllded in the NP-.-subnet. Since 
postmodJfier attachment cannot be performed 
well using local information only, this 
pairs r.~icely with our strategy of handling 
th~ a~?9ament/modi fief attachment on the 
casefram~ level and thereby reducing 
ambiguity for the ATNo 
.. in mai~~ clauses (where the verb--second 
movemeni: /Haidex 1988/ applies ) 'this 
movemer~li has to be undone to have the 
(~omple%e verbal complex as the head of the 
sel~tenco J.n the last position. This has 
anoth~c advantage~ Although word order is 
d:\[ f forent in main clauses and dependent 
~lause~ on the surface, after this 
ret~."an~ fo~mation the same subnet can be 
used for all different sentence types, and 
the same is true for the subnet fox' the 
vez'bal comp\]~ex o 
Adopting the grammar in tile way Just 
described leads to the desired situation 
~he:£~e for every phrase type. the head comes 
I a s t o 
4. £'.as~ef~:~ue~ and ~:he Cage Frame Matcher 
Casef~:'ame~ represent both a semantic and a 
syntactic representation of a phrase. The 
~enmntic ,zontent is given by a ' semantic' 
prediea't~ and the functional dependencies and 
meanings of its arguments, and further 
red, frictions by modifiers (if any) . 
The very idea of rep~;esenting semantic 
dependencies in form of case frame~ goes back 
to tile work of Fillmore /1968/, whereas ideas 
ol~ the additional syntactic and functional 
structure we use can be traced back to 
Chomsky' ,a /1981/ Theta--rules and Bresnan' S 
/1982/ functional structures and in the 
A:ctificlal Intelligence paradigm -to the work 
of Ct-eary and Pollard /1985/o 
The easofx'ames in DB-DIALOG consist of 
sever:el parted' %'lie head predicate, a SELF- 
,~ loi for prope~ ~eferencing, so-called 
VALENCY slots containing functional 
dependencies (or deep cases ), e MOD slot 
containing modifiers, a DETERMINER slot for 
NPs, and SYNTAX and CATEGORY slots containing 
various s~ntactic information° 
VALI~NCY slots in turn consist of: 
• .- an identifier 
- a syntactic restriction (SYN) 
- a semantic ~estrietion (SEM) 
o- a filler (VALUE) 
Ce~of~'am~s are instantiated, from the lexicon 
ai~d Infox',~iation is added during the analysis 
~:~f ~'abphza~eso To do so the~e is at least 
o~:~ so.-oalled "meaning" attached to the 
ie~:Lc;al entry of each verb, • noun and 
~C\]t:~c'tive. A meaning consists of a pointer to 
a oasefz-amo plus eventual modifiers to be 
applied to the caseframe at the time of 
ins~%a~rtie.t,ton o The instantiation process 
o.~'oa%es ne w edges in the chart, representing 
,th~se pa~'tJ.ally filled caseframeSo The Case 
l,'~ame Matcher (CFM) wo~ks on that chart, 
which as passed on to it by the ATN. This 
d~art consists only of those caseframes 
'.celevant to the CFM to construct the now 
casefram~ ~ Other parts, like the 
mo:cphologfeal oha~t or already constructed 
caseframes outside the seeps of tile ph:c~a~ 
actually conside~:ed x:emain invisible to ito 
One or more of the caseframes in the chart 
passed to the CFM are marked a~J pz'ospe<rtiw~ 
heads, and the otrtput of the CFWI :L~'~ a new 
caseframe (oz" more than one .in ca~e of 
ambiguity ) spanning the whole ch~r t w:i th 
several slots filled° 
VALENCY slots may be filled if: 
-. syntactic restrictions are met, 
-. semantic restrictions are met, 
-- other restrictions stemming 
eategorN of the 
met. 
and 
from the 
head (o.g.adjaconcy) are 
The syntactic restrictions a:ce met i\[f t,h~ 
feature\[~ of the SYN-~slot arid SYNTAX of the 
filler caseframe ean be unified° ',~h<~ 
restrit:tions given are usually (>n <:ategory, 
case, preposition, etc. But they need nc, t be 
given explicitly in all cases. One can make 
use of a number of structural case,.~ like SUBJ 
( subject ) and DOBJ ( direct obj ect ) 
Transformations can apply to these ea~e~\] 
under" Get-rain circumstances and <~'og~t~'ansfo~-m 
DOBJ into SUBJ in case of passive° Tile 
realization of tile structural c a'~ c, :i .~\] 
evaluated at the i:ime of slot fi fling, 
depending on the catego:cy of the head. 
Only if a restriction is stated explicitly Lf 
is taken as it stands° But sh~-uc;tura\] ca.~Je:: 
like eog. SUBJ get different interp~-'etations: 
for an S (sentence) a nominal:lye NP with 
nu,nber agreement with the head is sought, fo:c 
an AP SUBJ has to be the head of the 
governing NP, agreein 9 in case, gerldc,,r ~md 
number, and for an NP SUBJ is realized as a 
genitive NP Or a PP with the pcepos/.tion 
~von' . 
Thi,~ way great flexibility is gained and iI: 
is possible to reduee the lexicon and the 
meanings stored ~herein to the essentia!~3o 
It is even possJ.}3\] e \[:o p-,'o C e \['.~J 
nominalizattons using the meaning of i;iJe 
corresponding 'verb. 
Tho semantic \[cestz'ictions to be met aide, gLw.~h 
by a hierarchy of predicates. Slim arLd the 
predicate of the filler caseframe mu,~\]t be 
compatible to allow slot fillingo gimJlar 
considerations apply to the construction of 
modifiers : syntactic and semanhJ.c 
compatibility must be given° 
5. Interaction 
Generally speaking, the topelogioal 
regularities of phrases are handled by the 
ATN, whereas free word order constituents are 
being taken care of by the unification 
process~ This unification proces~ works on ~ 
local chart created by the A'I'N, comprlsin 9 
only those parts of the sentence relevant to 
it~ Thus various island phenomena fa\]! out 
from the conception of the parser., 
Flow of control between the ATN and the othe~ 
components is organized in a way p~:eposed by 
Boguraev /1979/o The ATN starts p'coeessi.~Lc~ a 
sentence in the usual way~ Afte:c rocognizJ.n<\] 
a phrase boundary by reachin U a POP az'<~ 
control is given either directly to the CFM 
Or the unification process o Th~ process 
evoked serves as a test for the ?~OP arc, 
i o e 0 in case of failure the ATN ha~5 to 
backtrack. 
679 
In constituents (with strict word order) the 
CFM is invoked directly and tries to build up 
a caseffame (or more than one in case of 
ambiguity)° The result is returned to the 
ATN which makes use of it during further 
processing° 
In structures with free constituent order 
(clauses) the ATN acts solely as a collector 
The words are first processed mo~phologically 
and a chart is returned, rendering a 
canonical form for each of the words together 
with word class and syntactic information 
(e°g. case markers). At this level, some 
ambiguities a~ise, eogo that of "welche" 
which might be an interrogative pronoun or a 
relative one, and "die" which may be an 
article or e relative pronoun. 
Welct\]e ,, Wien aez~el\]t fueP die Pr'odtJl<tion t)enoetigte Stoffe von Fir'men aus dem Ausland ? 
Fig.2: Morphological chart 
of constituents. Constituent caseframes are 
merely stored in a local chart and attachment 
is postponed° The only constituent 
recognized topologically is the head which 
always comes in the last position° This 
chart of constituents Ks then given to the 
unification process when the POP arc is 
reached. In addition to relying heavily on 
the CFM, the unificator also has various 
strategies at its disposal in order to take 
into consideration restrictions of adjacency 
and category dependent of the category of the 
phrase processed° This way possible 
syntactic ambiguity is m~duced and almost no 
backtracking is needed inside the ATNo 
Generally, information passed to the CFM is 
collected while traversing the subnet: head 
caseframes are instantiated, arguments and 
modifiers are collected by pushing the 
appropriate subnets and morphological and/or 
syntactic clues trigger various informations 
on the caseframes° 
AS an example we mention the passive 
• transformation: if evidence for passive is 
gathered while analyzing the verbal complex 
(for S) or a participle (for APs), this 
information is passed on to the CFM. The CFM 
then applies the passive transformation to 
the relevant slots of the head caseframe 
before the slot filling takes place. These 
transformations are one way to take general 
syntactic information away from the lexicon 
(the caseframes) to reduce redundancy /Hayes 
et ai.1985/. 
6. An Annotated Example 
To demonstrate how the system works, we will 
conclude the paper by giving an annotated 
example of a parse. For the sake of clarity 
some of the details shall be simplified, but 
all of the essentials will be properly 
described. 
We have chosen the following example 
sentence: 
"Welche yon unseren Abteilungen in Wien 
bezieht fuer die Produktion benoetigte Stoffe 
von Firmen aus dem Ausland?" ("Which of our 
Viennese departments gets materials necessary 
for production purposes from abroadT") 
Please note that the free translation does 
not capture the grammatical subtleties 
involved in the original sentence;especially 
the adjective phrase "fuer die Produktion 
ber.oet~gte Stoffe" includes a passivizatlon 
that tis usually not expressed this way in 
English. 
68O 
There is a simple global control structure 
which works on this morphological chart° Its 
main task is to transfer control to ATN 
networks for phrase-like constituents and %o 
the unlficator for clause-like constituents° 
The control structure starts by transferring 
control to the PP/NP-ATNo The chart entry for 
"Ausland" is treated first ( remember the 
right-to-left direction of processing), i~ 
is found to be a noun, and the next edge, 
DET, is processed. The third word, "aus", 
finishes the PP/NP. Control is transferz'ed 
to the caseframe marcher (CFM). The 
caseframe for the head, "Ausland", becomes 
instantiated, and the features of the other 
components are unified with it, especially 
the feature of dative, which is derived from 
the determiner. 
After completion of this caseframe, control 
is transferred back to the PP/NP net which 
processes "yon Firmen" in a similar way° The 
CFM is called again, constructing another 
caseframe~ According to our strategy, PP 
attachment will not be performed at thls 
step, instead all the constituents will be 
collected first° 
The PP/NP ATN gets its next chance° It 
treats the chart entry for "Stoffe ~ which 
makes a perfectly suitable head for a more 
complex constituent. We start to anticipate 
this when the next word, "benoetigte" 
("necessary" - albeit not an adjectives but a 
PPP in German), is processed. In general, 
inflected PPPs trigger a PUSH AP, so does 
this one. (Uninflected PPPs form part of the 
verb complex). Next, a PUSH PP/NP is 
performed which will lead to a constituent 
embedded in the APe Hut let's see this in 
detail. The PP is processed similar to %he 
others before, the head "Produktion" becoming 
instantiated and the caseframe filled after 
the entry for "fuer" has been processed. 
This finishes the AP, since the verb, 
"bezieht", definitely cannot be part of an 
AP. As you may remember, APs t~Igger th~ 
unification component which in turn calls the 
CFM to handle the simpler tasks. Thus, th~ 
head of the AP, "benoetlgte", becomes 
instantlatedo The associated caeeframe lu 
presented below: 
(BENOETIG 
(SYN SUBJ) (SEM ORGANIZATIONAL_UNIT) 
(SYN DOBJ) (SEM MATERIAL) 
(SYN PPOBJ (FUER)) (SEM PURPOSE)) 
Before the caseframe will be filled~ a 
passive transformation is applied, due to the 
fact that the example sentence contained the 
verb "benoetigen" in its PPP re,me This 
transformation simply changes SUBJ to 
PPINP~5 
5ffr~fib~i ........... k~T 1 
Fig~3: After processing of PP/NP3 
POBJ PA,%~ t~nd DOBJ "to SUBJo 'The consequences 
of :tl,iL: t~ansformat:iion will not turn out in 
this s'te~p, but in %he next one~ when the 
PP/NP w~th th(~ head "Stoffe" will hav~ been 
flnished, l~ut let ' s stick to the correct 
ordex:. Th(~ caseframe of "be,xoe'tigen" has 
b~e~ in~;'tantiated and t_~.'ansfoz'med, and it is 
abol~t 'to be filled° Normally, the unificater 
wt:ll ~aOW dome into its own, having to decide 
for pro~er attachments. In this case, thez'e 
is on:ty one constituent lef%. at this level, 
:~am~ly ":£\]\]er d~ Prodllktion" o Six~ce x~o 
Con f\] J c%ln9 attachment a!tez'nat:J.v~s aris<~. 
tho Ct"M J.s called di~.'ectly b~ the t~niffcator~ 
"~gl<~:c d:\[c~ Produktion" fits n.-Lcely i~rto tile 
PURPOSe: slot, so it :is placed there. The AP 
~,ow has helen finished~ and POP PP/NP J~ the 
next ~dg~ "t O be taken° 
Sel~ 3 
\[yj~(," NP 
5y~|l~l ~. NomlACc Pltlr31 
Head Stoff 
l~1od 
benoetig 
P0bj-Pass ~ubdj" Organlzatlonat Unlt 
................................................. }\[ 
5ubj I~ rflaterlal / 
PPobJ (fuer) Purpose 
Fig,4: Caseframe for PP/NP3 
Let us take a little digression~ Suppose the 
PP/NP ~'fuer die Produktion" would not have 
fit into a slot of the PPP. If we had taken 
"gefaerbte" ("dyed") instead of "benoetigte" 
'this wo%~id do. In this case we would not get 
the ree~ding "materials dyed for production 
puz'poses" but instead two components, dyed 
materials" and "for production pu~'poses" o 
The sophisticated reader could argue that 'the 
first reading might also be correct. The 
argument here is that the oaseframes in our 
systems are constructed .in a wa~ to fit the 
PP/NP I 
......... :IE L Pl::tO~ ........ 
needs of the domain modelled° In ou~ " domain, 
this reading would not be appropriate, so we 
d-_i.d not provide a caseframe for it, thus 
excluding a theoretical ambiguity where in 
the practical application thez"e is none° As 
the slot filling fails, the AP..ATN will. 
backtrack. We get an AP consisting of just 
one single word ("gefaerbte") filling a slot 
in "Stoffe", making up for one PP/NP and 
another PP/NP, namely "fuer die Produktion". 
These two PP/NPs will be collected at this 
stage of processing and only attached when 
all of -the sentence will have been parsed. 
We will stop our digression here a**d come 
back to the original example° Remember, the 
AP has ,lust been flnished and the PP/NP with 
the head "Stoffe" is POPped° This means a 
transfer of control to the CFM (in PP/NPs the 
CFM is called di~ectly, whereas in an AP or S 
tile unlficator is called first in o~der to 
:find correct attachments° Afterwards, the 
uniflcator in turn calls the CFM to realize 
• the selected attachments ). The AI? is 
integrated into the PP/NP caseframe as a 
modifier predicate in the MOD sloto The SUBJ 
slot of the subordinated caseframe (the one 
of "benoetigen" ) is still unfilled. For 
syntactic reasons, its filler must be the 
head of the superordinated PP/NP "Stoffe"o 
The semantic restriction of the SUBJ slot is 
MATERIAL which is compatible with the noun 
"Stoffe", so the slot may be filled (note 
that SUBJ is the transformed syntactic 
/'estriction which had been DOBJ before the 
passive transformation had taken place ) o 
Thus, a third constituent has been added to 
the pool of collected constituents° 
The global control structure ce*rtinues by 
processing the next entry, the representation 
of the word "bezieht", which is a finite verb 
and has to be at the second position 
according to German grammar° It is set aside 
for later processing and a special state is 
entered, knowing that exactly one constituent 
has been left over° The PP/NP "in Wien" is 
processed, and a corresponding casef~ame is 
created. 
Similarly, a caseframe for "welche yon 
unseren Abteilungen" is created and "in Wien" 
is attached to it when the unificator applies 
its knowledge that there cannot be more than 
one constituent in this position. This way, 
possible ambiguities e.g° trying to fill "in 
Wien" into a slot at sentence level, are 
avoided° 
PP/NP4 
PPINP PPfNP 
PP/NP~5 
..................... -AP- ................. \[ PP/NP2 
VERB PREP I~E~P-RO N- NOUN \] PPP \[NOUN PREP NOUN .... ...... <\] ........................... 15 .... 
Fig.g: Before unification at sentence level 
PP/NP 1 
;L :IZ2 
681 
By this time w~ have finished our way f~:'o~ 
rigi~t to left th~7ough the momphologic;~I chart 
and have co13 outed many components (PP/NP\[-~ 
and the predicate) at tile sentence ievel o 
The global control structure passes contx'ol 
to the urlificator which has to find cO:c'~'<~c% 
attachment and to pez'form the slot filling at 
the sentence level o Casefi:'ame Ins tantlation 
takes place, building a frame for the vez'b 
"beziehen" 0 
,(BEZIEH 
(SYN SUBJ) 
(SYN DOBJ) 
(SYN PPOBJ (FUER)) 
(SYN PPOBJ (VON BEI)) 
(gEM ORGANYZAT.UN%T) 
(SEM MATERIAL) 
(SEM PURPOSE) 
(SEM ORGANIZAT,UNIT)) 
Next, all possible attachments are sought° 
Two conditions have to hoid for them~ 
adjacency and semantic cempatibility~ PP/NP4 
e. g o cannot be attached to an V other 
constituent, because it is adjacent on:Ly to 
the main verb° Therefore, this constituent 
has to fill a slot in "beziehen"o Foi: the 
remainJng PP/NPs thOre exist different 
possibllities o Let us denote subordinat ~ on 
by the hyphen characte:c0 From the adjacency 
point of view, the possibilities are: 
i) PP/NP3 , PP/NP2, PP/NP1 
(three constituents at sentence level) 
2) PP/NP3 - PP/NP2, PP/NPI 
3) PP/NP3, PP/NP2 .~ PP/NPI 
4) PP/NP3 - (PP/NP2 - PP/NPI) 
1 and 2 are excluded, because there is no 
slot in the "beziehen" caseframe which 
matches the syntax of PP/NPI (preposition 
"aus"), nor would there be semantic 
compatibility° 3 is the reading we p~efer. 
As for' 4, its acceptability depends on 
whether we allow a slot in ~he caseframs for 
"Stof fe" which could hold an 
ORGANIZATIONAL UNIT. If we do, we will get 
an ambiguity.-- In that case, the system will 
offer both solutions, using a heuristic which 
of the solutions to present first° The 
heuristic implemented prefers flat syntactic 
structures° 
AS for' the preferred reading, the CFM 
realizes it by filling PP/NP3 into the DOBJ 
slot and (PP/NP2 -. PP/NP1) into the PPOBJ 
slot of the caseframe for "bezlehen"o PP/NP4 
has already been filled in the SUBJ slot, se 
the parse of the sentence has been completed° 
7. Conclusion 
In this paper we have described a parser 
which is able to deal with a wide variety of 
German sentences in an efficient and robust 
wry. This is made possible by two special 
properties of the parser: 
First, syntactic analysis uses two 'different 
strategies, namely an ATN and unification "to 
deal with fixed word o~der at the eonsh:\[tu~;nt 
level and free corrstituent order at the 
clause level rsspectively° Such an app'~'oaeh 
i.s well suited fox' German, but also for other 
SOV languages like Dutch, Swedish, Danish and 
Japanese ° 
Second, the interaction between casefram~ 
instantiation and a syntactic parser leads to 
the possibility of 
- clear and precise formulation of syntactic 
and semantic rules and regularities, 
reduction'of entries in tlre lexicon, 
- flexibility and better malntalnability~ 
682 
Thus pz'oblems posed by various aspeo'ts of "tbc; 
gene:c'al ~o\],~t:l,o~'~s which may be aDpllod \[:o 
othec ian\[~uages ms we~\].lo 
Aehnow\].edge~ent 
Work on DB- .D\]\[ALOG is done Jointly with 
Softwa:~c:~ Managem~r!t GmbH~ A-1140 V:ten~s ~ 
Austr:\[a0 It has also been sponsored by the9 
Austwia~ Governmer~t withJ,z'~ the "~-~{9~w~z'::,ur~c~.: 
S? (Artificial Inte\].Iigenee ) dez l~1:~.\]¢:~:0-, 
e i ektronik foerderungsprogr arums dex' 0~ LCt,~:~ ~ 
~?eichisohen Bundesreglert~ng". 

References

Bates, Mo (11.978) ~The Theory and Pr~cti.:~:~ o:~ 
Augmented Tz'ansition Netwo)_'l.~: G-c~anm~x'~. 
in ~ Bo\]o L o ( sd o ) ~ Natura£ )\]~~hf.~u~9'<~ 
Commu:,~ication witll Computers~ Sp~:\[n~_p<,-.'~ 
Berlin 0 

Dogur_aev, B °K. (1979) 'Autoraatic, Reso\]utlon 
of Lingut.stic Amblguities~ ' Un:;.vossit~f 
of Cambridge, Comp. Laboratory, TR,-!I~ 
Cambridge. 

Bresnan, J~ (ed.) (1982) ' The Menka\] 
Represontation of Ora~mnatica\] 
Relations° ~ MIT Press, Cambridge MA~ 

Buohbex'ger Eo~ Heinz ,W,, Matiasek Jo 
'frost H. ( 1987 ) ~ The German Language 
Database Interface DIALOG° v Proc o 8th 
International Congress Data P£'ocess J n<i 
in Europe, Vienna° 

Chomsky, N, ( 1.981 ) ' Lectures on Government 
and Binding°' Foris, Dordrecht~ 

Creary L°, Pollard C.(1985) 'A Computatio~al 
Semantics for Natural Language° ~ Xn 
Proc°23rd Meeting of the ACL, Chicago° 

Fillmore, CoJo(1968) 'The Case fez ~ Ca~e~ 
In ~ Bach E o, Harms R o T o ( ads o ) 
Unlve~Tsals in Linguistic Theo~y~ Hol't, 
Rinehart and Winston Inco, New York~ 

(}rishman R., Hirschman Lo, Nhan NOT.(\]986) 
' Discove~.'y Procedures for' Sublanguage 
Selectional Patterns : Xnitlal Experi~ 
marts. ~ Computational Linguistics, 3(12)o 

Haider~ H. ( 1984 ) ' Topic ~ Focus and Ve:cb ,~- 
second° ' G~oningez ' Arbeiten zux ~ 
Germanlstisehen Linguistik 25~ ppT?~-llO~ 

Haide_~:, H~ (1988) 'Verb.-seeond in German° ' 
2n: Haider Ho, P)cinzho~'n '£h0 (ads0) 
Verb-second, Foris, Dordrech%. 

Hauenschild, Ch. (1988) 'GPSG and German Wo~°d 
Order.' In: Reyle Uo, Rohrer Ch~(edso)~ 
Natural Language Parsing and Linguistic 
Theory, Reidel, Dordrecht~ 

Hayes Ph~ ~ Andersen P., Safir 8 ~ (!985) 
' Semantic Caseframe Parsing and 
Syntactic Generality. ' In~ Prec. 23rd 
Annual Meeting of the ACL, Chicago° 

Hoehle, T.No(1982) ~Explikation fuer norma\]e 
Betonung und normale Wortstellungo' !n~ 
Abraham W. (ed.) : Satzglledo~ :\[m 
Deutschen, Namr, Tuebingen. 

Lytinen, S o ( 19\[)6 ) * Dynamically Combiuln~{ 
Syntax and Semantics in Na%~rz'ai La~gu~g~5 
Processing. ~ Pr(..ooAAAI~,86, Ph:|,ladelpbi<~ 
PA. 

Px'oudian, D,, Polla~d, C0 ( 1985 ) ~ Pax'sing 
IIead,~Driver ~ Phrase Structure Gramma:<' o 
\]?roc. 23rd Annual Meeting Of the ACL~ 
Chicago° 

T~'ost, He , Dorffneg', G° (1987) 'A System fox' 
Morphological Analysis and S~nthesi~ of 
{~er~an Texts. * X~.~ Hainllne Do(ado )~ Ne~ 
Developments in Computer ~oAs~i~rte~¢\] 
Language L~arning, C~?oom Helm, Lordotic 

'\[;szkorelt~ He (1986) ' Const~?aints on Ordex' 
gtanfo:<'d Un.lv ~ ~ C8LI-86~46, S%anford~ 
CA, 
