Translating Lexical Semantic Relations: 
The First Step Towards Multilingual Wordnets* 
 
Chu-Ren Huang, I-Ju E. Tseng, Dylan B.S. Tsai 
Institute of Linguistics, Preparatory Office, Academia Sinica 
128 Sec.2 Academy Rd., Nangkang, Taipei, 115, Taiwan, R.O.C. 
churen@gate.sinica.edu.tw, {elanna, dylan}@hp.iis.sinica.edu.tw 
 
                                                 
* An earlier version of this paper was presented at the Third Chinese Lexical Semantics Workshop at Academia Sinica in 
May 2002. We are indebted to the participants as well as colleagues at CKIP for their comments. We would also like to thank 
the SemaNet 2002 reviewers for their helpful comments. It is our own responsibilities that, due to the short revision time, we 
were not able to incorporate all their suggestions, especially comparative studies with some relative GWA papers. We are 
also responsible for all remaining errors 
 
Abstract 
Establishing correspondences between 
wordnets of different languages is essential 
to both multilingual knowledge processing 
and for bootstrapping wordnets of 
low-density languages. We claim that such 
correspondences must be based on lexical 
semantic relations, rather than top ontology 
or word translations. In particular, we define 
a translation equivalence relation as a 
bilingual lexical semantic relation. Such 
relations can then be part of a logical 
entailment predicting whether source 
language semantic relations will hold in a 
target language or not. Our claim is tested 
with a study of 210 Chinese lexical lemmas 
and their possible semantic relations links 
bootstrapped from the Princeton WordNet. 
The results show that lexical semantic 
relation translations are indeed highly precise 
when they are logically inferable. 
 
1. Introduction 
A semantic network is critical to knowledge 
processing, including all NLP and Semantic Web 
applications. The construction of semantic 
networks, however, is notoriously difficult for 
‘small’ (or ‘low-density’) languages. For these 
languages, the poverty of language resources, 
and the lack of prospect of material gains for 
infrastructure work conspire to create a vicious 
circle. This means that the construction of a 
semantic network for any small language must 
start from scratch and faces inhibitive financial 
and linguistic challenges. 
In addition, semantic networks serve as 
reliable ontolog(ies) for knowledge processing 
only if their conceptual bases are valid and 
logically inferable across different languages. 
Take wordnets (Fellbaum 1998), the de facto 
standard for linguistic ontology, for example. 
Wordnets express ontology via a network of 
words linked by lexical semantic relations. Since 
these words are by definition the lexicon of each 
language, the wordnet design feature ensures 
versatility in faithfully and comprehensively 
representing the semantic content of each 
language. Hence, on one hand, these conceptual 
atoms reflect linguistic idiosyncrasies; on the 
other hand, the lexical semantic relations (LSR’s) 
receive universal interpretation across different 
languages. For example, the definition of 
relations such as synonymy or hypernymy is 
universal. The universality of the LSR’s is the 
foundation that allows wordnet to serve as a 
potential common semantic network 
representation for all languages. The premise is 
tacit in Princeton WordNet (WN), EuroWordNet 
(EWN, Vossen 1998), and MultiWordNet (MWN, 
Pianta et al. 2002). It is also spelled out explicitly 
in the adaptation of LSR tests for Chinese 
(Huang et al. 2001).  
Given that LSR’s are semantic primitives 
applicable to all language wordnets, and that the 
solution to the low-density problem in building 
language wordnets must involve bootstrapping 
from another language, LSR’s seem to be the 
natural units for such bootstrapping operations. 
The rich and structured semantic information 
described in WN and EWN can be transported 
through accurate translation if the conceptual 
relations defined by LSRs remain constant in 
both languages. In practice, such an application 
would also serve the dual purpose of creating a 
bilingual wordnet in the process.  
In this paper, we will examine the validity 
of cross-lingual LSR inferences by bootstrapping 
a Chinese Wordnet with WN. In practice, this 
small-scale experiment shows how a wordnet for 
a low-density language can be built through 
bootstrapping from an available wordnet. In 
theoretical terms, we explore the logical 
conditions for the cross-lingual inference of 
LSR's. 
 
2. Translation Equivalents and Semantic 
Relations 
 Note that two translation equivalents (TE) 
in a pair of languages stand in a lexical semantic 
relation. The most desirable scenario is that when 
the two TE’s are synonymous, such as the 
English ‘apple’ to the Mandarin ‘ping2guo3’. 
However, since the conceptual space is not 
segmented identically for all languages, TE’s 
may often stand in other relations to each other. 
For instance, the Mandarin ‘zuo1zhi5’ is a 
hypernym for both the English ‘desk’ and ‘table’. 
Suppose we postulate that the LSR’s between 
TE’s are exactly identical in nature to the 
monolingual LSR’s described in wordnets. This 
means that the lexical semantic relation 
introduced by translation can be combined with 
monolingual LRS’s. Predicting LSR’s in a target 
language based on source language data become 
a simple logical operation of combining 
relational functions when the LSR of translation 
equivalency is defined. This framework is 
illustrated in Diagram 1.  
 
CW2 a0    ii a1 a2a4a3 2 
 
     y           x 
 
CW1 
a5
   i  a1 a2a4a3 1 
 
 x = EW1 - EW2 LSR 
 y = CW1- CW2 LSR 
 i = CW1 - EW1 Translation LSR 
 ii = CW2 - EW2 Translation LSR 
The unknown LSR y = i + x + ii 
Diagram 1. Translation-mediated LSR Prediction 
(The complete model) 
 
CW1 represents our starting Chinese lemma 
which can be linked to EW1 through the 
translation LSR i. The linked EW1 can than 
provide a set of LSR predictions based on the 
English WN. Assume that we take the LSR x, 
which is linked to EW2. That LSR prediction is 
mapped back to Chinese when EW2 is translated 
to CW2 with a translation LSR ii. In this model, 
the relation y, between CW1 and CW2 is a 
functional combination of the three LSR’s i, x, 
and ii. 
However, it is well known that language 
translation involves more than semantic 
correspondences. Social and cultural factors also 
play a role in (human) choices of translation 
equivalents. It is not the aim of this paper to 
predict when or how these semantically 
non-identical translations arise. The aim is to see 
how much lexical semantic information is 
inferable across different languages, regardless of 
translational idiosyncrasies. In this model, the 
prediction relies crucially on the semantic 
information provided by the source language (e.g. 
English) lexical entry as well as the lexical 
semantic correspondence of a target language 
(e.g. Chinese) entry. The translation relations of 
the relational target pairs, although capable of 
introducing more idiosyncrasies, are not directly 
involved in the prediction. Hence we make the 
generalization that any discrepancy introduced at 
this level does not affect the logical relation of 
LSR prediction and adopt a working model 
described in Diagram 2. We only take into 
consideration those cases where the translation 
LSR ii is exactly equivalent, i.e., EW2 = CW2. 
This step also allows us to reduce the maximal 
number of LSR combination in each prediction 
to two. Thus we are able to better predict the 
contribution of each mono- or bi-lingual LSR. 
 
    a1 a2a4a3 2 = CW2 (ii = 0) 
 
          y       x 
 
a5
   i  a1 a2a4a3 1 
 The unknown LSR y = i + x 
Diagram 2. Translation-mediated LSR Prediction 
(Reduced Model, currently adopted) 
 
2.1 LRS Inference as Relational Combination 
 With the semantic contribution of the 
translation equivalency defined as a (bilingual) 
LSR, the inference of LSR in the target language 
wordnet is a simple combination of semantic 
relations. The default and ideal situation is where 
the two TE’s are synonymous.  
 
 
 
 
a0a2a1a4a3
2 = EW2 
 
          y   x 
 
a5 i 
a6  
CW1 = EW1 (i = 0) 
 The unknown LSR y = x 
Diagram 3. Translation-mediated LSR Prediction 
(when TE’s are synonymous) 
 
In this case, the translation LSR is an identical 
relation; the LSR of the source language wordnet 
can be directly inherited. This is illustrated in 
Diagram 3. 
 However, when the translation has a 
non-identical semantic relation, such as 
antonyms and hypernyms, then the LSR 
predicted is the combination of the bilingual 
relation and the monolingual relation. In this 
paper, we will concentrate on Hypernyms and 
Hyponyms. The choice is made because these 
two LSR’s are transitive relations by definition 
and allows clear logical predications when 
combined. The same, with some qualifications, 
may apply to the Holonym relations. 
Combinations of other LSR’s may not yield clear 
logical entailments. The scenarios involving 
Hyponymy and Hypernymy will be discussed in 
section 3.3. 
 
3. Cross-lingual LSR Inference: A Study 
based on English-Chinese Correspondences 
 In this study, we start with a WN-based 
English-Chinese Translation Equivalents 
Database (TEDB)1. Each translation equivalents 
pair was based on a WN synset. For quality 
control, we mark each TE pair for its accuracy as 
well as the translation semantic relation. 
 For this study, the 200 most frequently used 
Chinese words plus 10 adjectives are chosen 
(since there is no adjective among the top 200 
words in Mandarin). Among the 210 input 
lemmas, 179 lemmas2 find translation 
equivalents in the TEDB and are mapped to 497 
                                                 
1 The translation equivalence database was hand-crafted by 
the CKIP WordNet team. For each of the 99642 English 
synset head words, three appropriate translation equivalents 
were chosen whenever possible. At the time when this study 
was carried out, 42606 TE’s were proofed and available 
2 The input lemmas for which TE’s were unable to find are 
demonstratives or pronouns for nouns, and aspect markers 
for adverbs 
English synsets. The occurring distribution is as 
follows: 84 N’s with 195 times; 41 V’s with 161 
times; 10 Adj’s with 47 times; and 47 Adv’s with 
94 times. 441 distinct English synsets are 
covered under this process, since some of the 
TE’s are for the same synset. This means that 
each input Chinese lemma linked to 2.4 English 
synsets in average. Based on the TEDB and 
English WN, the 179 mapped input Chinese 
lemmas expanded to 597 synonyms. And 
extending from the 441 English synsets, there are 
1056 semantically related synsets in WN, which 
yields 1743 Chinese words with our TEDB.   
 
3.1. Evaluation of the Semantics of Translation 
Six evaluative tags are assigned for the 
TEDB. Four of them are remarks for future 
processing. The LSR marked are 
 a7
Synonymous: TE’s that are semantically 
equivalent. 
a8 Other Relation: TE’s that hold other 
semantic relations 
 
The result of evaluation of TE’s involving the 
210 chosen lemma are given in Table 1. 
 
 
 Syn. Incorrect Other Relation Total 
148 32 15 195 N 
75.90% 16.41% 7.69% 100% 
113 29 19 161 V 
70.18% 18.01% 11.8% 100% 
39 8 0 47 Adj 
82.98% 17.02% 0% 100% 
83 8 3 94 Adv 
88.3% 8.51% 3.19% 100% 
382 78 36 496 Total 
77.02% 15.73% 7.26% 100% 
Table 1. Input Lemmas (Total subject =496) 
 
Illustrative examples of our evaluation are given 
below: 
 
1a) Synonymous: a9a11a10  qi4ye4 (N) // enterprise: 
an organization created for business ventures 
1b) Incorrect: a12a12a14a13a13  biao3shi4 (V) // ‘extend’, 
‘offer’: make available; provide 
1c) Other Relation: a15a15a17a16a16  shi4chang3 (N) // 
‘market, securities_industry’: the securities 
markets in the aggregate 
 
Table 2 indicates the relations between the 
synonyms of an input lemma and the same 
English synset. Recall that our TEDB gives more 
than one Chinese translation equivalent to one 
English WN entry. Hence we can hypothesize 
that the set of Chinese translation equivalents 
form a synset. It is natural, then, to examine the 
semantic relations between other synset members 
and the original WN entry. Table 1 and 2 show a 
rather marked difference in terms of the 
correctness of the synonymy relation. This will 
be further explained later. 
 
 Syn. Incor. 
Other 
Rel. Others Total 
114 51 25 19 209 N 
54.5% 24.4% 11.0% 9.1% 100% 
104 46 18 14 182 V 
57.1% 25.3% 9.99% 7.7% 100% 
37 8 2 10 57 Adj 
64.9% 14.0% 3.5% 17.5% 100% 
119 20 4 6 149 Adv 
79.9% 13.4% 2.7% 4.0% 100% 
374 125 49 49 597 Total 
62.6% 20.9% 8.2% 8.2% 100% 
Table 2. Synonyms of Input Lemma  
(Total Subject=597) 
 
From the data above, we observe two 
generalizations: First, polysemous lemmas have 
lower possibility of being synonymous to the 
corresponding English synset. In addition, we 
also observe that there is a tendency for some 
groups, i.e., groups with polysemy and with 
abstract meanings, to match synonymous English 
synsets. These findings are helpful in our further 
studies when constructing CWN, as well as in the 
application of TEDB. 
 
3.2 Cross-lingual LSR predictions with 
synonymous translations 
The next step is to take the set of English 
LSR’s stipulated on a WN synset and transport 
them to its Chinese translation equivalents. We 
evaluated the validity of the inferred semantic 
relations in Chinese. In this study, we 
concentrated on three better-defined (and more 
frequently used) semantic relations: antonyms 
(ANT); hypernyms (HYP); and hyponyms 
(HPO). Here, we limit our examination to the 
Chinese lemmas that are both translation 
equivalents of an English WN entry and are 
considered to have synonymous semantic 
relations to that entry. The nominal and verbal 
statistics are given in Table 3 and Table 4 
respectively. 
 
 Syn. Incor. 
Other 
Rel. Others Total 
7 3 0 2 12 ANT 
58.3% 25% 0% 16.7% 100% 
117 33 15 20 185 HYP 
63.2% 17.8% 8.1% 10.8% 100% 
284 119 66 256 725 HPO 
39.2% 16.4% 9.1% 35.3% 100% 
408 155 81 278 922 Total 
44.3% 16.8% 8.8% 30.2% 100% 
Table 3. Nouns (Total Number of Inferable 
Semantic Relations=922) 
 
 Syn. Incor. 
Other 
Rel. Others Total 
8 6 0 9 23 ANT 
34.8% 26.1% 0% 39.1% 100% 
61 18 6 2 87 HYP 
70.1% 20.7% 6.9% 2.3% 100% 
118 81 19 74 292 HPO 
40.4% 27.7% 6.5% 25.3% 100% 
187 105 25 85 402 Total 
46.5% 26.1% 6.2% 21.1% 100% 
Table 4. Verbs (Total Number of Inferable 
Semantic Relations=402) 
 
 From the 148 nouns where the English and 
Chinese translation equivalents are also 
synonymous, there are 357 pairs of semantic 
relations that are marked in English WN and are 
therefore candidates for inferred relations in 
Chinese. On average, each nominal RC 
translation equivalent yields 2.41 inferable 
semantic relations. The precision of the inferred 
semantic relation is tabulated below.  
 
 Correct Others Total 
ANT 8 100% 0 0% 8 100% 
HYP 70 79.5% 18 20.5% 88 100% 
HPO 238 91.2% 23 8.8% 261 100% 
Total 316 88.5% 41 11.5% 357 100% 
Table 5. Precision of English-to-Chinese SR 
Inference (Nouns) 
 
The study here shows that when no additional 
relational distance is introduced by translation 
(i.e. the 75.9% of nominal cases when TE’s are 
synonyms), up to 90% precision can be achieved 
for bilingual LSR inference. And among the 
semantic relations examined, antonymous 
relations are the most reliable when 
transportabled cross-linguistically. 
For the 112 verbs where the English and 
Chinese TE’s are synonymous, there are 155 
pairs of semantic relations that are marked in 
WN and are therefore candidates for inferred 
relations in Chinese. In contrast to nominal 
translation equivalents, each pair of verbal TE 
only yields 1.38 inferable semantic relations. The 
precision of the inferred semantic relation is 
tabulated in Table 6. 
 
 Correct Incorrect Total 
ANT 14 100% 0 0% 14 100% 
HYP 35 70% 15 30% 50 100% 
HPO 75 82.4% 16 17.6% 91 100% 
Total 124 80% 31 20% 155 100% 
Table 6. Precision of English-to-Chinese SR 
Inference (Verbs) 
 
Similar to the results of nouns, antonymous 
relations appear reliable in the behaviors of verbs 
as well. As to the other types of relations, the 
correct rates seem to be slightly lower than nouns. 
The precision for English-to-Chinese semantic 
relation inference is 80% for verbs.  
The observed discrepancy in terms of 
semantic relations inference between nouns and 
verbs deserves in-depth examination. Firstly, the 
precision of nominal inference is 8.52% higher 
than verbal inference. Secondly, the contrast may 
not be attributed to a specific semantic relation. 
Both nouns and verbs have the same precision 
pattern for the three semantic relations that we 
studied. Inference of antonymous relations is 
highly reliable in both categories (both 100%). 
Hyponymous inference is second, and about 12% 
higher than hypernymous inference in each 
category (the difference is 11.64% for nouns and 
12.42% for verbs). And, last but not least, the 
precision gaps between nouns and verbs, when 
applicable, are similar for different semantic 
relations (9.55% for hypernyms and 8.77% for 
hyponyms). All the above facts support the 
generalization that nominal semantic relations 
are more reliably inferred cross-linguistically 
than verbal semantic relations. A plausible 
explanation of this generalization is the 
difference in mutability of nominal and verbal 
meanings, as reported by Ahrens (1999). Ahrens 
demonstrated with off-line experiments that verb 
meanings are more mutable than noun meanings. 
She also reported that verb meanings have the 
tendency to change under coercive contexts. We 
may assume that making the cross-lingual 
transfer is a coercive context in terms of meaning 
identification. Taking the mutability into account, 
we can predict that since verb meanings are more 
likely than nouns to change under given coercive 
conditions, the changes will affect their semantic 
relations. Hence the precision for semantic 
relations inference is lower for verbs than for 
nouns. 
In the above discussion, we observed that 
the three semantic relations seem to offer clear 
generalizations with regard to the precision of the 
inferences, as shown in Table 7. 
 
 Correct Incorrect Total 
ANT 22 100% 0 0% 22 100% 
HYP 105 76.1% 33 13.9% 138 100% 
HPO 313 88.9% 39 11.1% 352 100% 
Total 440 85.9% 72 14.1% 512 100% 
Table 7. Combined Precision of 
English-to-Chinese SR Inference (Nouns+Verbs) 
 
Two generalizations emerge from the above data 
and call for explanation: First, inference of 
antonymous relations is highly reliable; second, 
inference of hypernymous relations is more 
reliable than inference of hyponymous relations. 
 The fact that inference of antonymous 
relations is highly precise may be due to either of 
the following facts. Since the number of 
antonymic relations encoded is relatively few 
(only 22 all together), they may all be the most 
prototypical case. In addition, a pair of antonyms 
by definition differs in only one semantic feature 
and has the shortest semantic distance between 
them. In other words, an antonym (of any word) 
is simply a privileged (near) synonym whose 
meaning offers contrast at one particular 
semantic dimension. Since antonymy 
presupposes synonymous relations, it preserves 
the premise of our current semantic relation 
inference.  
 The fact that hyponymous relations can be 
more reliably inferred cross-linguistically than 
hypernymous relations is somewhat surprising, 
since they are symmetric semantic relations. That 
is, if A is a hypernym of B, then B is a hyponym 
of A. Logically, there does not seem to be any 
reason for the two relations to have disjoint 
distributions when transported to another 
language. However, more careful study of the 
conceptual nature of the semantic relations yields 
a plausible explanation.  
 We should take note of the two following 
facts: First, a hyponym link defined on an 
English word Y presupposes a conceptual class 
denoted by Y, and stipulates that Z is a kind of Y 
(see Diagram 4).  
 
 
 
 
 
 
 
 
 
Diagram 4. class vs. member identity (HPO) 
 
Second, a hypernym link defined on Y 
presupposes an identity class X which is NOT 
explicitly denoted, and stipulates that Y is a kind 
of X (see Diagram 5). Hence, it is possible that 
there is another valid conceptual class W in the 
target language that Y is a member of. And yet 
W is not equivalent to X. 
 
 
 
 
 
 
 
 
 
 
Diagram 5. class vs. member identity (HYP) 
 
Since our inference is based on the synonymous 
relation of the Chinese TE to the English word Y, 
the conceptual foundation of the semantic 
relation is largely preserved, and the inference 
has a high precision. The failure of inference can 
in most cases be attributed to the fact that the 
intended HYP has no synonymous TE in Chinese. 
To infer a hyponymous relation, however, we 
need to presuppose the trans-lingual equivalence 
of the conceptual class defined by HPO. And 
since our inference only presupposes the 
synonymous relation of Y and its TE, and says 
nothing about HPO, the success of inference of 
the hyponymous relation is than dependent upon 
an additional semantic condition. Hence that it 
will have lower precision can be expected. 
 To sum up, our preliminary evaluation 
found that the precision of cross-lingual 
inference of semantic relation can be higher than 
90% if the inference does not require other 
conceptual/semantic relations other than the 
synonymy of the translation equivalents. On the 
other hand, an additional semantic relation, such 
as the equivalence of the hypernym node in both 
languages when inferring hyponym relations, 
seems to bring down the precision rate by about 
10%. 
 
3.3. When Translation Introduces an additional 
LSR 
 In this section, we study the cases where 
translation introduces a hypernymous/ 
hyponymous LSR. These cases offer the real test 
to our proposal that TE’s be treated as bilingual 
LSR’s. The LSR inference here refuses 
non-vacuous combinations of two LSR’s. For 37 
Chinese input lemmas that hold other relations 
with English synsets, 57 semantically related 
links were expanded. First, we investigated the 
situation when the English synset occurs as a 
hyponym of the Chinese input lemma (Diagram 
6). 
      a0 a1a3a2 2 = EW2 (ii = 0) 
 
          y       x 
 
CW1
a4
   a0a6a5 a2 1 
    
(a) IF x = HPO 
y = HPO + HPO = HPO (Hyponym is 
transitive.) 
(b) IF x = HYP 
y = HPO + HYP = a7  
Diagram 6. Predicting LSR, when English is the 
hyponym of Chinese translation 
 
33 inferable relations satisfied above description. 
Among them, 8 falls in the entailment of figure 
Y= class identity 
class 
Tclunknown 
LSR 
Z 
HPO Member set identity 
is entailed. 
Y=member identity 
X W 
Class identity is 
NOT entailed. 
HYP HYP 
  i = HPO 
6(a). Manual evaluation confirms the prediction. 
The other 25 cases are not logically inferable and 
do indeed show a range of different relations. 
The logically entailed HPO relation is 
exemplified below: 
 a0
a1a1 chi1 HPO
a2 a3a5a4a7a6a9a8a11a10a13a12a14a4a15a4a7a16a18a17a19a8a20a6a22a21a23a4a25a24a26a27a12a29a28a18a28a18a16
HPOa30a32a31a34a33a36a35a22a37a23a38a22a39a41a40a43a42a14a38a7a38a7a44a46a45a47a33a36a38a7a38a7a44a18a48a5a49a50a52a51a54a53a55a30 a56a58a57a56a58a57a46a59a59
a60a60  lang2tuen1hu3yan4 
SO, a61a61 chi1 HPOa62 a63a63a65a64a67a66a64a67a66 a60a60  lang2 
tuen1hu3yan4 
Next, when an English synset is marked as a 
hypernym to the Chinese input lemma, logically, 
hypernymous relation is transitive (Diagram 7). 
 a68a70a69a72a71
2 = EW2 (ii = 0) 
 
          y       x 
 
CW1
a73
   
a68a75a74a76a71
1 
    
(a) IF x = HYP 
NOTE: y = HYP + HYP = HYP 
(Hypernym is transitive.) 
(b) IF x = HPO 
NOTE: y = HYP + HPO =a77  
Diagram 7. Predicting LSR, when English is the 
hypernym of Chinese translation 
 
We found 2 cases (actually expanded from the 
same synset) under this condition.  
 a78 a79a79
 shi3 HYPa80a82a81a84a83a86a85a7a87a22a88a23a85a7a89a91a90a7a87a22a92a94a93a95a85a97a96a20a98a100a99a41a85a102a101a103a104a87
specified state HYPa105a82a106a5a107a109a108a22a110a11a111a47a112a114a113a22a115a23a108a7a116a91a107a18a117a118a23a108
certain properties to something TEa105 a119a119  
shi3 / a120a122a121  zhi4shi3 
 
Note that the same Chinese word a123  shi3 is 
used for both the head word and its hypernym. 
Hence, there are two possible interpretations of 
the data. The first possibility is that Chinese 
simply has a coarser-grain sense distinction in 
this case and the hypernym relation is incorrect. 
The second possibility is that the relation is 
self-hypernym (Fellbaum 1999). Since a 
fine-grain sense distinction is beyond the scope 
of the current paper, we will not decide on either 
interpretation.  
 In sum, our lexical semantic relation model 
makes correct distinctions among inferable and 
non-inferable LSR’s. More specifically, it has a 
100% prediction for hyponymous relations. For 
hypernymous relations even though the logical 
entailment could not be verified due to 
sparseness of data; it did correctly predict the 
portion of data that was logically non-inferable. 
We expect future studies with a wider set of data 
to confirm this prediction. 
 
4. Conclusion 
In this paper, we proposed to treat the translation 
equivalents relations as a set of bilingual lexical 
semantic relations. This proposal allows us to 
process bi-lingual inference of LSR’s as simple 
functional combinations of semantic relations. 
The process itself greatly reduces the complexity 
of bootstrapping wordnets from a different 
language. We empirically supported our proposal 
by successfully applying it to the inference of 
Chinese LSR’s from English WN.  
The proposed approach requires bilingual 
TEDB’s that are marked with translation 
semantic relations. Although such TEDB’s are 
not widely available yet, they are necessary for 
cross-lingual language processing such as MT 
and IR, as well as for any type of knowledge 
processing. We hope that our approach can 
promote the construction of LSR-marked TEDB 
as well as multilingual wordnets. 
 
References
Ahrens K. 1999. The Mutability of Noun and Verb 
Meaning. Chinese Language and Linguistics V. 
Interactions in Language, Y. Yin, I. Yang, & H. 
Chan (eds.), pp. 335 – 548. Taipei. Academia 
Sinica. 
Fellbaum, C. (ed.). 1998. WordNet: An Electronic 
Lexical Database. Cambridge, MA: MIT Press. 
Huang, Chu-Ren. 2000. Towards a Chinese Wordnet 
and a CE/EC Bi-Wordnet. Chinese Language 
Sciences Workshop: Lexical Semantics. October 
9, 2000. Department of Chinese, Translation and 
Linguistics. City University of Hong Kong. 
Huang, Chu-Ren, D. B. Tsai, J. Lin, S. Tseng, K.J. 
Chen, and Y. Chuang. 2001. Definition and Test 
for Lexical Semantic Relations in Chinese. [in 
Chinese] Paper presented at the Second Chinese 
Lexical Semantics Workshop. May 2001, 
Beijing, China. 
Pianta, Emanuel, L. Benitivogli, C. Girardi. 2002 
MultiWordNet: Developing an aligned 
multilingual database. Proceedings of the 1st 
International WordNet Conference, Mysore, 
India, pp. 293-302. 
Vossen P. (ed.). 1998. EuroWordNet: A multilingual 
database with lexical semantic networks. 
Norwell, MA: Kluwer Academic Publishers. 
