LINGUISTIC PROCESSING USING A DEPENDENCY STRUCTURE GRAMMAR 
FOR SPEECH RECOGNITION AND UNDERSTANDING 
Sho-lchl MATSUNAGA 
NTT Human Interface Laboratories 
Musashino, To;¢yo, 180, Japan 
and Masaki KOHDA University of Yamagata 
Yonezawa, Yamagata, 922, Japan 
Abstract 
This paper proposes an efficient linguistic 
processing strategy for speech recognition and 
understanding using a dependency structure grammar. 
The strategy includes parsing and phrase prediction 
algorithms. After speech processing and phrase 
recognition based on phoneme recognition, the parser 
extracts the sentence with the best likelihood taking 
account of the phonetic likelihood of phrase 
candidates and the linguistic likelihood of the 
semantic inter-phrase dependency relationships. A 
fast parsing algorithm using breadth-first search is 
also proposed. The predictor pre-selects the p}~.ase 
candidates using transition rules combined with a 
dependency structure to reduce the amount of phonetic 
processing. The proposed linguistic processor has 
been tested through speech recognition experiments. 
The experimental results show that it greatly 
increases the accuracy of speech recognitions, and 
the breadth-first parsing algorithm and predictor 
increase processing speed. 
1. Introduction 
In conventional continuous speech recognition and 
understanding systems\[1~4\], linguistic rules for 
sentences composed of phrase sequences are usually 
expressed by a phrase structure grammar such as a 
transition network or context free grammar. In such 
methods, however, phoneme recognition errors and 
rejections result in incorrect transition states 
because of the strong syntactical constraints. 
These erroneous transitions cause the following 
candidates to be incorrectly chosen or the processing 
system to halt. Therefore, these errors and 
rejections can be fatal to speech understanding. 
Furthermore, a complete set of these grammatical 
rules for speech understanding is very difficult to 
provide. 
To address these problems, this paper proposes a 
new linguistic processor based on a dependency 
structure grammar, which integrates a bottom~up 
sentence parser and a top-down phrase predictor. ~lis 
grammar is more semantic and less syntactic than 
phrase structure grammar, and, therefore, syntactic 
positional constraint in a sentence rarely occurs 
with this parser. This effectively prevents extreme 
degradation in speech recognition from errors and 
rejections in phoneme recognition and greatly 
increases the accuracy of speech processing. This 
grammar only has two syntactic rules, so this parser 
is free of many cumbersome grammatical rules that are 
indispensable to other grammars. This grammar 
particularly suits phrase-order-free languages such 
as Japanese. 
4a~ 
rule I 
rule 2 
For the parser of this grammar, a depth-first 
parsing algorithm with backtracking which guarantees 
the optimal solution was devised\[5\]. However, parsing 
long sentences composed of many phrases with this 
algorithm can be time-consuming because of 
combinatorial explosion, since the amount of 
computation is exponential order with respect to the 
number of phrases. Therefore, a fast parsing 
algorithm using breadth-first search and beam search 
was developed. This algorithm is based on flmdamental 
algorithms\[6,7\] which only take account of the 
dependency relationships of the modifier and 
modificant phrases, and it handles higher linguistic 
or semantic processing such as case structure. The 
processing ability of this breadth-first algorithm is 
equivalent to that of the depth-first algorithm. 
To effectively recognize speech, the amount of 
phonetic processing must be reduced through top-down 
prediction of hypotheses. However, top-down centre\] 
using the principal dependency structure is 
impossible. To solve this problem, this novel phrase 
predictor was devised. This predictor pre-selects 
hypotheses for phoneme recognition using prediction 
rules, and then it reduces the amount of phonetic 
processing. Prediction rules are created by 
integrating connection rules and phrase dependency 
structures. 
The effeetivenessof this linguistic processing 
was ascertained through speech recognition 
experiments. 
2. Linguistic processor 
2. 1 Dependency ntJcueturegrammar 
This grammar is based on semantic dependency 
relationships between phrases. The syntactic rules 
satisfy the following two constraints. First, any 
phrase, except the last phrase of a sentence, can 
modify only one later phrase. Each modification, 
called a dependency relationship or dependenc 2 
structure, can be represented by one arc. Second, 
modification arcs between phrases must not cross. 
These rules are illustrated in Fig. I. In two 
unacceptable sentences, one sentence is unacceptable 
because one phrase modifies the former phrase, and 
Acceptable sentence ~~a~le sentence 
A B C D 
A B C D 
× 
A B C D 
× 
A B C D 
Fig. 1. Examples using a dependency structure grammar 
A,B,C and D are sentence phrases. 
the ether sentence is unacceptable because arcs cross 
in its dependency structures. 
2. 2Parser 
After phonetic phrase recognition, recognition 
results are represented in a phonetic score matrix 
form as shown in Fig. 2. When analyzing dependency 
relationships, the parser extracts the most likely 
sentence in this matrix by taking into account the 
phonetic \]_ikelihood of phrase candidates and the 
linguistic likelihood of semantic inter-phrase 
dependency relationships. The parser also obtains the 
dependenqy structure that corresponds to the semantic 
structure of the extracted sentence. 
2. 2. 1 Objective function 
This parsing is equivalent to solving the 
following objective function using the constraints of 
dependency structure grammar. For simplicity, the 
following linguistic formulation is described for 
speech uttered phrase by phrase. The process for 
sentence speech is described in section 4. 
N 
T -- max\[ 2;'c(xi ,p) + maxZdep(w I j_1,Xj,piY1,j,p)\] (I) p j::l " Y j=l ' 
where I~j.<N, 1Kp~M, N is the number of input phrases, 
M is the maximum number of phonetic recognition 
candidates for each phrase, Xj,p is a candidate of 
the j~th input phrase with the p-th best phonetic 
likelihood, and c(xj,p) is its phonetic likelihood 
(positive value). Also, Xi,j, p is a phrase sequence 
with one phrase candidate for each i-th to j-th input 
phrase and whose last phrase is Xj,p. Yi,j,p is one 
of the dependency structures of Xi,j,p, wi,j_ I is the 
set of phrases that modify Xj,p in the sequence 
Xi,j, p. Here, dep(w,xlY) is the linguistic likelihood 
(negatiw! value) of dependency relationships between 
w and x taking Y into account. Namely, the first item 
of the teem on the right in Eq. (I) is the summation 
of phonetic likelihoods of the hypothesized sentence 
composed of its phrase sequence, and the second item 
is the summation of linguistic likelihood. Maximizing 
Eq. (1) gives the sentence and the dependency 
structure of it as speech recognition and 
t~dersta~ling results. 
Because dependency structure grammar is compatible 
with case grammar\[8\], the linguistic semantic 
likelihood(dep) of the dependency structure is easily 
provided using case structure. The following are 
example:~ of items for evaluating dependency 
relationships: the disagreement between the semantic 
primitives of the modifier and that requested by the 
modificant, the lack of the ob\].igatory case phrase 
requested by the modificant, and the existence of 
differenb phrases with the same case and modifying 
the same phrase. The likelihood values for these 
items are given heuristically. 
To so\]re equation (I), a fast parsing algorithm 
using breadth-first search and beam search was 
developed. This algorithm deals with higher 
linguistic or semantic processing such as the case 
structure.. Although this algorithm offers sub-optimal 
solutions, it is practical because it requires less 
processing than depth-first search. 
2. 2. 2 l~ce~t~z-firstpe~csingalgorithm 
The breadth-first algorithm is formulated as 
order of candidates 
1 2 ...... M 
x\[,l Xl, 2 Xl,~4 I 1 
2 x2,1 x2, 2 
/ 
x3, I xs, 2 x3,MJ B 
Fig. 2. A matrix of phrase candidates 
input 
utterance : 
number 
N-1 
follows. 
First, dep(w,xlY) can obviously be divided into 
two terms. 
dep(wl,j-1 ,Xj,plYl,j,p) = 
E dep1(x,Xj,p) + dep2(Y1,j,p,Xj,p) (2) 
x Cw I ,j-1 
where depl is the likelihood associated with 
dependency relationships of only the modifier and 
modificant phrases, and dep2 is the likelihood 
associated with Y1,j,p. An example of dependency 
relationships is shown in Fig. 3. 
Eqs. (I) and (2) give the objective function's 
wtlue S(1,Xj,p) of a phrase sequence including the 
top phrase to Xj,p in the sentence as: 
S(1,Xj,p) = 
J J J 
Z C(Xh,p) + ~ ~ dep1(x,xh,p) +Edep2(Y 1,h,p,xh,p) (3) 
h::1 h=1 xCw I ,h-1 h=1 
On the other hand, the value of a phrase sequence not 
including the top phrase (i¢I) is defined as: 
D(i,Xj,p) = 
j j j-1 
_Z C(Xh,p) + Z Z depl (X,Xh, p) + Z dep2(Y i,h,p,xh,p) 
h: :i h=i xCwi,h_ I h=i (~) 
Tim main difference between Eqs. (3) and (4) is that 
phrase 
sequence 
wh,F 
I 
i 
I 
I 
E (A, E, F) 
depl (F, G) 
depl (E, G 
depl (A, G 
', '.-t dep2( YA,G 
dep(wA,p. G i YA,G =dep2( YA.G ' G) +dspl (F, G) 
+depl (E. G) +depl (A, G) 
Fig. 3. Illustration of dependency relationships 
403 
dep2(Yi,j,p,Xj,p) is not evaluated in Eq. (4). 
Using notation S and D, the recurrence relation 
among the objective functions are derived. This is 
shown in Fig. 4. The recurrence relation are 
transforms into the following equations using beam 
search. 
S(1,Xj,p,r) = rth-max\[S(1,Xk,q,rl) + D(k+1,Xj,p,r2) 
k,q,rl,r2 
+ dep1(Xk,q,Xj,p) + dep2(Y1,j,p,Xj,p) , if i=I (5') 
D(i,xj,p,r) = rth-max\[S(i,Xk,q,rl) + D(k+1,Xj,p,r2) 
k,q,rl,r2 
+ dep1(xk,q,Xj,p) 4 dep2(Yi,k,q,Xk,q), if i~I (6') 
where i~k~j-1, 1~q~M, and 1~r, rl, r2<_L. Here, r, rl 
and r2 indicate the rank of beam, L is the maximum 
number of beams, S(1,Xj,p,r) and D(i,xj,p,r) are the 
r-th value of the element whose phrase sequence is 
Xi,j, p and the dependency structure is Yi,j,p. 
ilere, rth-max\[ \] is a function for deriving the r-th 
best value. When Eq. (5') or (6') is calculated, 
Yi,j,p is stored for use in the later stage of 
evaluating dep2. 
Initial values are given as follows. 
S(1,Xl,p,1) = C(Xl, p) + dep2~1,1,p,Xl,p), 
if i=1(top phrase) (7) 
D(i,Xi,p,1 ) = e(Xi,p) , if i~I (not top phrase) (8) 
After calculating the recurrence relation, the value 
of the objective functions is obtained, 
T = max\[S(1,XN,p,1)\] , (9) 
P 
where 1~p~M. The best sentenc~ and its dependency 
structure are given through YI,N,p where p maximizes 
Eq. (9). The parsing table is shown in Fig. 5 and the 
parsing algorithm is shown in Table I. In Fig. 5, 
the first row corresponds to S, and others correspond 
to D. The phrase sequence for first to N-th phrase 
corresponds to the right-most top cell. Each cell is 
composed of ML sub-cells. Arrows show the sequence of 
calculating the recurrence relation. The processing 
amount order for this algorithm is O(N3M2L2). 
Comparing the theoretical amount of processing for 
these two parsing algorithms, the breadth-first 
parsing algorithm clearly requires much less 
processingthanthe depth-first parsing algorithm. 
The amount of processing for each parsing algorithm 
is shown in Fig. 6. 
2. 3 Predictor 
To pre-select the phrase hypotheses for the speech 
recognition, the predictor is devised\[9\], using 
prediction rules created by integrating connection 
rules and dependency structures of phrases. These 
rules are described with rewriting rules: 
(Xi,j)->(Xi,k)(Xk+1,j) 
where Xi, j is the phrase sequence for the i-th to j- 
th phrase. (Xi, j) is the sequence with a Closed- 
dependency-structure where the tail phrase xj has the 
dependency relationships with phrases out of Xi,j, 
and other phrases in Xi, j have dependency 
relationships with phrases within Xi, j. (Xi, j) is 
divided into two phrase sequences with the closed- 
dependency-structure by modifying x k by xj, and 
following Xi, k by Xk+1, j. A single phrase x i is also 
4O4 
(J) if i=l 
1 
F" 
Phrase no. 
k ... j 
depl( Xk, q, Xj,p) ~ 
dep2(Y.. , x. 
S(I, Xk, q) D(k+l, Xj,p 
S(1. X.p. 
) =S(l, x k q) +\]J(k+l, x. ) 
S(I, Xj,p +depl( Xk,q', Xj,p) +dep2t'P~l,j,p , Xj,p) (5) 
® if i ¢ 1 Phrase no. 
i ... k '" j \[ .....  xUU 
___~kdepl ( , x j ) Xk.q ,p 
dep2( Yl,k,q ' Xk,q ) 
D(t, Xk, q) D(k+l, X i,p) 
I.-- " >1 F" 
D(I, x.p) 
) =D(t, x k q) +~(k+l, x,j _ ) 
D(I, Xj,p +depl( Xk,q', Xj,p ) +dep2('~/i,k, q , xk, q ) (6) 
Fig. 4. Illustration of deriving recurrence relatiun among 
objective functions 
N 
1 2 ' j o N 
order of 
candidates bea~ 
(1, x N • .) 
Fig. 5. Configuration of a parsing table 
regarded as a phrase sequence with a closed- 
dependency-structure. These rules are described for 
the sequence, the i-th phrase to j-th phrase modified 
by the i-th phrase, as 
(Xi,j)->(xi)(Xi+1,j) 
The hypotheses are predicted as follows. 
<I> x i is detected as a reliable phrase recognition 
result. If there are no reliable phrase candidates in 
the i-th phrase recognition results, the following 
procedure is not carried out. 
<2> The rules whose left term is scanned are such as 
(Xi+l,j)->(Xi+1,k)(Xk+1,j) 
After the left-most derivation is repeated to detect 
hypotheses for i+1~thphrase speech recognition, xi+ I 
is detected in the following form. 
(Xi+1,j)->(xi+1)(Xi+2,h)----(Xk+l,j) 
Generally, there is more than one (Xi+1,j) , so xi+ I 
is a setof phrases. 
<3> The phrase recognition is carried out for the 
i+l-th ~\]case utterance whose hypotheses are elements 
of the set xi+ I. 
<4> If the reliable phrase recognition result was 
detected in operation <3>, the rules which derived 
elements of xi+ I are scanned again and hypotheses for 
the next utterance are derived using same procedure 
as <2>. 
<5> Thes~ operations, namely hypotheses derivation 
and its phonetic verification, are carried out until 
xj is detected. 
<6> The detected phrase sequence Xi~ j and its 
dependency structure Yi,j is passed to the parser. 
During bhese operations, if the phrase recognition 
results are unreliable in operation <3>, the 
detectioa process of Xi, j is halted and phrase 
recognitSon for all hypotheses is carried out. 
Althe~,gh Japanese is a phrase-order-free language, 
there are some relatively fixed please-order parts in 
a sentence. These rules are applied to these parts. 
The number of hypotheses and the amount of acoustic 
processing can thus be reduced, maintaining the above 
characteristics of the dependency structure grammar. 
By linking the predictor to the parser, parsing can 
be accomplished using the dependency structures 
detected in operation <6> of the prediction 
procedure. This linkage method greatly increases 
parsing speed. 
3* Speech recognition exporimen%~ 
3- 1St~echrecognibion system 
The speech recognition and understanding system 
is shown in Fig. 7. The system is composed of 
acoustic and linguistic processing stages. The 
acoustic processing stage consists of a feature 
extraction part and a phoneme recognition 
part\[10,11\]. The linguistic processing stage consists 
of a phrase recognition patti11\], a parsing part (a 
dependency relationship analysis part), and a phrase 
prediction part. The linguistic processing stage uses 
a word dictionary, word connection rules for intra- 
phrase syntax, dependency relationships rules and 
phrase prediction rules. The word dictionary is 
composed of pronunciation expressions, parts of 
speech and case structures. Dependency relationship 
rules produce negative eva%~ation values that are set 
to the dependency relationships contrary to case 
structure discipline. 
3. 2 Speschreeognitionprocess 
For separately uttered phrases, acoustic feature 
parameters are extracted and bottom-up phoneme 
recognition is carried out. The phrase hypotheses for 
top-down phoneme recognition are pre-selected by the 
Table 1. Parsing algorithm 
{ 1 } Loop for the end phrase of the partial sequence 
DO {2} to {5} for j = 1,2,---,N 
{2} Loop for the candidate 
DO {3} to {5} for p = 1,2,---,M\] 
{3} Setting the initial value 
SET S(1,Xl,p,1) or D(j,Xj,p,1) (Eqs. (7),(8)) 
If J = 1, go back to {2}. 
{4} Loop for the beginning phrase of the partial sequence 
DO {5} for i = 1-1j-2,--,1 
{5} Calculation of reccurence relation 
< Loop for tlle end phrase of the former sequence > 
{5-1} DO {5-2} to {5-4} for k = i,i÷1,-~-,1-1 
{5-2} DO {5-3} to {5-4} for q = 1,2,--~,M k 
< Loop for the beam width > 
{5-3} DO {5-4} for rl = 1,2,---,L 
{5-4} DO for r2 = 1,2,---,L 
* Evaluation of S(1,x 1 n,r) or D(j,x~ n,r) taking 
account of YtZ,p or (Eq,. 
* Store of Ytj,p 
{6} Acquisition of the parsing results 
* Detection of value p maximizing Eq. (9) 
* Acquisition of the phrase sequence and its dependency 
structure using Y1,N,p 
I@ ° 
F-- 
5 Io 2o 
mr   ,olo 
NUMBER OF PHRASES 
0 enumerative method 
A depth-first method 
V breadth-first method(L=oo) 
ill breadth-first method(L=8 ) 
# breadth-flrst method(L=l) 
Fig. 6. Comparison of processing amount 
10 20 30 
predictor. The pre-selection is also carried out 
using bottom-up phoneme recognition results\[12\]. 
Next, top-down phoneme verification is carried out 
and phrase recognition results are generated. Phrase 
recognition results are represented in the form of 
score matrix with phonetic recognition scores 
averaged for each hypothesized phrase. When the end 
of a sentence is detected, the parser extracts the 
phrases with t~e best sentence likelihood by scanning 
this matrix, and determines the dependency structure 
of the extracted phrases. 
3- 3 Performance 
The effectiveness of the proposed linguistic 
processor was tested in speech recognition 
experiments. The experiments were carried out using 
20 sentences containing 320 phrases uttered by 2 male 
speakers. These results are shown in Table 2. 
4O5 
recognition and understanding results 
I wg\[dd~c~l._o+ary__~--~\] dependency relationships analysls~--~idependency 
, I case J ~ /\ I relatlonshlps 
Lstructure ~ hypotheses prediction ~----7 \[\[~2~ ......... J i~ ............ 1 i l 
linguistic I \[pronounciatlon\[ + M / 
I I processing , expression ~ hypotheses derivation | L_~predic(lon ..... ? 
.............. i+J- t, \[-- L~uJS~ ......... J ~tntra-phrase \[pre-selectto 
L~Y~ ........ J ~ phrase@ (word) recognition 
...................... ............................................................................................... bottom-sp phoneme ~honeme\[ 
acoustic recognition recognition __\] 
processl ng ~ 
\] feature parameter extraction \] 
input speech 
Fig. 7. Speech recognition and understanding system 
,~ 10~ 
102 
W =8 10 0 
13.. 
0 
0 
0 
O 
O O 
O 
lb 1~i' "< 2b' '25 
NUMBER OF PHRASES IN EACH SENTENCE 
lO 4 
F- 102 
tl,I 
~10 o 
0 0 
0 0 0 0 0 
0 0 0 ~ ~Y~ 
i0 1~i '2'0 
NUMBER OF PHRASES 
IN EACH SENTENCE 
25 
(a) speaker KI 
O depth-first method 
~7 breadth-first method(L=8) 
(b) speaker KS 
O depth-first method 
breadth-first method(L=8) 
Fig. 8. Comparison of processing time for dependency relationships analysis 
Table 2. Speech recognition results 
phonetic recognition S 
number of i phonetic 
hypotheses i processing 
i tlme 
I without predictor 542 i 1. 0 
':~i~hp~i~i~i:__ ........ i~~ ....... i6:~h ........ 
without parser 
phrase recognition 
rate \[~\] 
0: within 
top 8 candidates 
57 (87) 
60 (89) 
with parser 
depth-flrst parsing breadth-first 
parsing 
phrase i parsing 
recognition i time 
rate \[~\] 
* Recognition of predicted phrases ( 33~ to the total input phrases 
The proposed parser using the depth-first parsing 
algorithm increased phrase recognition rate by 
approximately 2OZ (from 57Z without the parser to 77~ 
with the parser). This result shows the effectiveness 
of a parser using a dependency structure grammar. 
The processing time with the breadth-first 
algorithm was reduced to approximately IZ of that 
with the depth-flrst algorithm for sentence parsing, 
while keeping the same level of speech recognition 
77 il. 
phrase i parslng 
recognition i tlme 
rate \[~3 i 
77 ~1. .............................. } ...................... 
78 i O. 89 
rate as with the depth-first algorithm. This result 
shows the great effectiveness of the breadth-first 
parsing algorithm. This result is shown in Fig. 8 for 
each speaker when M is 3 and L is 8. 
Next, using 26 rules, the prediction was carried 
out for 33~ of the total input phrases. It reduced 
acoustic processing time to 60Z at these parts in a 
sentence, and it increased speech recognition speed. 
Finally, linking the predictor to the parser reduced 
406 
parsing time to less than 10% of the time for the 
depth-first parser, and to approximately 90% of the 
time for the breadth-first parser. This shows the 
usefulness of the linkage. 
4- Breadth-firs% parsing algorithm for sentence 
speeehreco~tion 
The breadth-first parsing algorithm for the 
sentence speech or connected phrase speech is 
devised\[13\] by the same procedure as in section 2. 2. 
Based on basic expansion algorithms\[14,15\] from 
phrase-wise to sentence speech, the speech 
recognition and understanding accuracy using the 
proposed algorithm is greatly increased compared to 
the accuracies using the basic algorithms. In the 
sentence speech, phrase recognition results after 
phonetic processing are represented in a score 
lattice form withphonetic recognition scores 
averaged. The parser extracts the best sentence 
composed of a phrase sequence by scanning this 
lattice. The processing order is O(NSM2L2), which is 
practical amount of computation, where N is the 
number of detected phrase boundaries in the uttered 
sentence, M is the maximum number of phonetic 
recognitio~ candidates for each phrase segment f~em 
one boundsry to the next boundary, and L is the 
maximum number of beams. 
The effectiveness of this parser was tested 
through sentence speech recognition with one speaker 
uttering i0 sentences containing a total of 67 
phrases. This parser increased phrase recognition 
performance in the sentences by approximately 49Z 
(from 27~ without the parser to 76~ with the parser). 
5- Conclusion 
This paper proposed an efficient linguistic 
processing strategy for speech recognition and 
understanding using a dependency structure grammar. 
This grammar suits processing of phrase-order-free 
languages such as Japanese and processing the result 
of front-end speech recognition, which is usually 
erroneous~ This linguistic processing strategy 
includes bottom-up parsing and a top-down phrase 
hypotheses predictor. In particular, the bottom-up 
parser, taking account of the phonetic and linguistic 
likelihood, greatly increases the accuracy and speed 
of speech recognition. The predictor reduces the 
amount of phonetic processing by pre-selecting the 
phrase hypotheses. The effectiveness of this parser 
and predictor was shown in speech recognition 
experiments. 
Future development is include the statistical 
likelihood of dependency relationships, integration 
with the statistical phonetic method like Hidden 
Markov Models, and higher linguistic processing using 
the semantics and context knowledge. 
Aaknowlmlgment 
The authors would like to express their 
appreciation to Dr. Kiyoshi Sugiyama and Dr. Sadaoki 
Furui, for their invaluable guidance. The authors 
would also like to thai( to Dr. Kiyohiro Shikano and 
Shigeki Sagayama for their useful suggestions. 

References 

\[I\] Leser V.R., et al(1975)WOrganization of the 
Hearsay II speech understanding system.',IEEE Trans. 
ASSP., 23, I, pp. 11-24. 

\[2\] Woods W.A.(1976)'Speech understanding system - 
Final Report',Tech. Rep.,3438. 

\[3\] Levinson S.E.(1985)'Structural Methods in auto- 
matic speech recognition',Proceeding of the IEEE, 11, 
pp. 1625-I 650. 

\[4\] Ney H.(1987),'Dynamic programming speech recogni- 
tion using a context-free grammar.'Proc. 1987 ICASSP, 
pp. 69-72. 

\[5\] Matsunaga S. & Kohda M. (1986)'Post-processing 
using dependency structure of inter-phrases for 
speech recognition.'Proc. Acoust. Soc. Jpn. Spring 
Meeting, pp. 45-46. 

\[6\] Hidaka T. & Yoshida S.(1983)'Syntax analysis for 
Japanese using case grammar.'Natural Language Pro- 
cessing Technique Symposium,pp41-46. 

\[7\] Ozeki K.(1986)'A multi stage decision algorithm 
for optimum hunsetsu sequence selection.' Paper Tee. 
Group, IECE Japan,SP86-32,pp.41-48. 

\[8\] Filmore C.(1968)'The case for case.' in Bach and 
Harms (eds.), 1-88 

\[9\] Matsunaga S. & Sagayama S.(1987)'Candidates 
prediction using dependency relationships for minimal 
phrase speech recognition.'Paper Tec. Group, IEICE 
Japan, SP87-29, pp. 59-62. 

\[10\] Aikawa K., Sugiyama M. & Shikano K.(1985)'Speken 
word recognition based on top-down phoneme segmenta- 
tion.'Proe.1985 ICASSP,pp33-36. 

\[11\] Matsunaga S. & Shikano K.(1985)'Speech recogni- 
tion based on top-down and bottom-up phoneme recogni- 
tion.' Trans. IECE Japan,J68-D,9,1641 -I 648. 

\[12\] Matsunaga S. & Kohda M.(1987)'Reduction of word 
and minimal phrase candidates for speech recognition 
based on phoneme recognition.'Trans. IEICE Japan,J70- 
D, 3, pp. 592-600. 

\[13\] Matsunaga S.(1988)'Dependency relationships ana- 
lysis for connected phrase recognition.' Nat. Conv. 
IEICE Japan, 84-I-3,pp.345-346. 

\[14\] Ozeki K.(1986)'A multi-stage decision algorithm 
for optimum bunsetsu sequence selection from bunsetsu 
lattice.'Paper Tec. Group, IECE Japan, COMP86-47, 
pp. 47-57. 

\[15\] Kohda M.(1986)'An algorithm for optimum selec- 
tion of phrase sequence from phrase lattice.' Paper 
Tech. Group, IECE Japan,SP86-72,pp.9-16. 
