Synthesizing a large concept hierarchy from French hyperonyms 
Jean Fargues, Adeline l'errin - IBM Paris Scientific C.enter 
3 et 5 Place Vendome, 75021 Paris Cedex 01, France 
INTRODUCTION 
The KAI,II'SOS prototype we have developed at 
the IBM Paris Scientific Center is able to analyze 
texts written in l;reneh and to produce a semantic 
representalion of these texts expressed as a set of 
inter-related Conceptual Graphs \[6, 1\]. It uses a 
semantic lexicon which contains, for each word to 
be defined, one or more Conceptual Graphs corre- 
sponding to one or more meanings of the word. 
The KAI,IPSOS questkm/answefing system ana- 
lyzes a Natural l.anguage query, translates it into 
Conceptual Graphs, performs pattern matching 
and deduction on these text graphs to select the 
answer, and finally generates a Natural Iangmage 
answer fiom the selected answer graphs. 
We do not detail this KA1JI'SOS system here 
because many papers have already been published 
on it (see the references). We have chosen to 
present recent work \[5\] which has been done on 
building a large concept hierarchy from an existing 
published dictionary. More precisely, we have syn- 
thesized a large semantic network by selecting 
hYtx'xonyna definitions from th.c "l)ictionnaire du 
vocabulaire esscntiel - \[,es 5000 roots fondamen- 
taux" (G. Mator(:, Iarousse, Paris 1963) and 
coding them as a set of Prolog clauses awtilable for 
the KAI,IPS()S system. 
Concept type hierarchy and 
hyperonymy 
First, we must remind you how the concept type 
hierarchy is the necessary basis for any use of the 
Conceptual Graph model. The reader may of 
course refer to \[6\]. In the Conceptual Graph 
model, the concept types are not supposed to be 
words but abstract symbols (atoms) used to denote 
a concept. Ia'or example, we could have COMMU- 
NICATION-I~I1OCESS as the concept type that 
occurs in the definition of the verbs "to say", "to 
communicate", "to discuss", etc. In the same way, a 
polysemic word like %at" should point on distinct 
graphs containing STICK and ANIMAl, as primi- 
tive concept types. It appears that the concept type 
hierarchy must also contain chains like: 
EI,I;,PlIANT < MAMMAl, < ANIMAl, < 
I,IVING-III¢ING < F, NTITY 
112 
Such a concept type hierarchy is necessary to 
define the patiern matching algorithms on Concep- 
tual Graphs which are used to build a graph from a 
Natural Language sentence by joining the Concep- 
tual Graphs of its parts. It is also necessary 1o 
encode and to verify the preference semantics con- 
straints in the semantic lexicon. The concept hier- 
archy is the basis for the join and projection 
algorithms \[3, 6\] which provide a way to disam- 
bignate the Natural l~anguage complex sentences 
and to perform query/answering on Conceptual 
Graphs. 
In the present work, we consider that concept 
types may generaUy be identified to word senses. 
Thus, the word "bat" poinls on concept types 
BAT. 1 and BAT.2 and BAT. 1 < STICK, BAT.2 
< ANIMAI, is stored in the hierarchy (STICK 
and ANIMAL being the concept types associated 
to the main meanings of the words "stick" and 
"animal"). This implies that the synthesis of a large 
concept type hierarchy is related to the seleclion of 
correct hyperonyms. We give here the logical inter- 
pretation of the hyperonymy relation between the 
words with meanings wl and w2, derived from the 
one given in \[4\]: 
w I is hyperonym of w2 ill', for every sentence S 
true\[S(w2)\] ~ true\[S(w2/wl)\] 
where: 
S(w) stands for a sentence containing an occur- 
rence of w, 
S(w2/wl) stands R)r the sentence S(w2) in which 
the occurrence of w2 is replaced by w 1. 
is the usual logical implication. 
l:or example, ANIMAl, is hyperonym of I)OG 
because all assertions about a particular dog remain 
true when we substitute "the animal" for "the clog". 
Of course, this criterkm is not "always verified in a 
such formal way. It is only a guideline. 
In a Natural Language dictionary, the Natural I,an° 
guage definitions may be classified hlto a typology, 
as in \[4\]. For example, all the definitions of the 
form NP VP may be hyperonym definitions, as in: 
l 
'elephant': a very large animal with two tusks and 
a trunk with which... 
But NP VP definitions may also be meta def- 
nitions, as in: 
'beget': old use to become the father of 
or, as an example in l:rench: '~tre': mot qui ddnote 
la facult6 d'exister. 
In this paper, we have tried to translate the defi- 
nitions into English to make it easier to read, but 
our French dictionary (5,00(1 entries) uses simpler 
definitions than the l,ongman dictionary. This is 
the reason why the reader will nut find a perfect 
match when referring to the Ixmgman. Further- 
more, this work depends on the particular dic- 
tionary (Mator6 l.~mmsse) we adopted but the 
important fact is that the result we have built is 
coherent and con'ect. 
The method 
The method was mainly empMcal: it was not so 
clear that the information contained in the dic- 
tionary would be useful for synthesizing a large and 
coherent concept hierarchy. We will return to this 
important point later. But we must add that the 
building of a large concept hierarchy from natural 
langmat,e definitions has limits. For example, it 
cannot be a simple hierarchy but a hierarchy in 
which the links are labeled by conceptual relations 
like part-of, set-of, etc. Another limit is that tile 
theoretical transitivity of the hyperonymy relation 
can only be verified on a chain of word senses if 
the chain is not too long. It should be noted that 
we were particularly interested in the top part of 
the hierarchy, i.e. in the list of the basic concepts 
from whicla all the others may be derived. The 
method, a bottom-up one, was carried out in tile 
following stages: 
1. The hyperonymy definitions were selected from 
the dictionary (by hand). 
2. The meanings of the words, in the entries and 
in the definitions, were distinguished by intro- 
ducing a coherent subscript notation for the 
current word and the main noun of its deft- 
nifion (by hand). 
3. The relation between the current meaning of 
the word and its hyperonym were encoded as a 
l'rolog clause (by hand). 
4. Ixmps were suppressed by the application of 
Prolog consistency checking programs that 
introduced an additional syuonymous relation 
between concepts. We mean here that when 
wl < w2 and w2 < wl are found, we declare 
as a l'rolog clause that SYNONYM(wl,w2). 
5. t'rolog programs were applied to the result in 
order to display it in a suitable way (see appen- 
dices A and B), and to have associative access 
to this data from Prolog. 
There is a difference between simple hypemnym 
definitions and compound hyperonym definitions. 
A simple hyperonym definition has the syntactic 
pattern N + VP..., or N + 
RF, I,ATIVF,-CI,AUSE... In this case we choose N 
as the hyperonym of the current word, if it is a 
correct hyperonym. A compound hypemnym deft- 
nition has the syntactic pattern: 
NP VP..., where NP has the form: 
1. N AI),IL;CTIVI~ 
:2. N PRF, PD N (I'I~,EPI) stands for 'de' 'du', 'de 
la') 
3. Absence de N (absence of N) 
4. Manque de N (lack of N) 
5. Action tie V (action of V) 
6. I~aSsultat de N (resull of N) 
7. Ensemble de N (set of N) 
8. Masse de N (mass of N) 
9. Groupe de N (group of N) 
10. R&nfion de N (urlion of N) 
i 1. Fair de V (fact consisting in V) 
12. Fawm de V (way of V) 
13. Mani~re de V (maturer of V) 
14. Possibilit6 de V (possibility of V) 
15. l~,tal de N (state of N) 
16. Art de V (,art of V) 
17. Quantitd de N (quantity of N) 
18. l,iste de N (list of N) 
19. Suite de N (sequence of N) 
20. Pattie de N (part of N) 
21. Morceau de N, pi~',ce de N (piece of N) 
22. UNITE. I de N (unit. l of N) 
23. l)ivision de N (division of N) 
24. Element de N (element of N). 
In all these cases, we keep the informalion con- 
tained in the NP and we code it into Prolog as 
follows: 
• Case 1: we include the adjective in the frst 
hyperonym and we derive a secondary hyper- 
onym, h)r example: 
F, lephant: A large animal .... 
1;,I ,F, PI IA NT < I,ARGE-ANIMAI, < 
ANIMAl, 
• Case 2: we keep the compound noun as tile 
first hyperonym and we generate its secomla W 
hyperonym, for example: 
l)oute: Etat d'esprit .... 
(Doubt: Stale of mind ...) 
t)OUTE. I < I~,TAT-I)F,-I~,SPRIT.2 < IiTAT 
2 113 
.lcudi: Jour de la semaine ... 
(Wenesday: day of the week ...) 
.lli~\[J\]l)I < JOUR.2-I)E-SEMAINE < 
JOUR.2 
• Cases 3 to 24: we consider that the relation is a 
primitive conceptual relation which labels tile 
hyperonymy link in the concept hierarchy, for 
example: 
lnfanterie: ensemble de troupes ... 
(Infantry: set of troops ...) 
INFANTERIE < 
ENSEMBI,E-I)E(TROUI~E. I) < 
ENSEMIH,E 
Manche: 
1. Partie d'instrument ... 
Part of an instrument ... (handle) 
2. l'artic de v6tement .., 
Part of an item of clothing ... (sleeve) 
MANCHE. I < PAR- 
TIE. I-I)I3(INSTRUMENT) < I~AR'I'iE.I 
MANCIIE.1 < PAR- 
"I'IE.I-I)E(VETEMIr, NT) < I'AP, TIE.1 
l)6cision: Action de choisir ... 
(decision: act of choosing ...) 
I)!!,CISION < AC.T1ON-DE(CHOISIR) < 
ACTION-DI_ ~, < ACTION 
Tiffs last case implies that the result is more than a 
simple hierarchy: from a formM point of view it is 
a semantic network because of the use of primitive 
relations ACT-OF, PART-OI;, SET-OF etc. 
The result shows that there are 57 main hierarchies. 
We give the corresponding table containing the top 
concepts and the number of sons they have. In all, 
more than 3,600 word meanings have been coded 
into the network. Please see Table 1 Before the 
Appendices. 
This restflt is not homogenous: some hierarchies 
contain many nodes and tile others a few nodes. 
We can consider that: 
• Some hierarchies correspond to fundamental 
types: 
-- Etre (being) 
- Chose (thing) 
- Fait (fact) 
- Action (action) 
- Substance (substance) 
- Quantit6 (quantity) 
- Mani~re (manner) 
- l;orce (strength) 
- Son (sound) 
- Feu (fire) 
Other hierarchies correspond to logical types: 
- Motif (motive) 
- Cons6quence (consequence) 
- Fonction.2 (function) 
- Lien (link) 
- Manque (lack) 
- Absence (absence) 
- El6ment (element) 
- Nombre (number) 
- degr6 (degree) 
Other hierarchies correspond 
concepts: 
to topological 
- Borne (boundary) 
- Bordure (edge) 
- Direction (direction) 
- Dimension (dimension) 
- Espace (space) 
- Intervalle (interval) 
- Contenu (content) 
- Volume (volume) 
• The remaining hierarchies correspond to other 
isolated types and contain fewer concepts than 
the preceding ones. They are also pertinent but 
it is surprising to obtain some of them as basic 
genetic concepts. 
Another remark must be made on tile transitivity 
of the ' <' hyperonymy relation. It appears that in 
a chain w l < w2 < ... < wn, it is possible to con- 
sider that each relation wi < wi+ 1 is justified. 
Nevertheless, it is more difficult to justify wl < 
wn. For example, consider the chain: 
TOMBE < FOSSE < TROU < OUVER- 
TURE.2 < PASSAGE.2 < LIEU < PAR- 
"1"II~,. I(ESPACE) 
grave < pit < hole < opening < passage < 
location < part-of(space) (we give tiffs translation, 
but it is very difficult to keep the exact nuances of 
the French chain). 
In this chain, each contiguous relation is justified, 
but to justify tile link between 'grave' and part- 
of(space) requires specifying the point of view that 
is taken. In fact, we have reached the limits of the 
process of building a concept hierarchy from 
existing dictionaries. 
Appendix A contains an extract of the hyperonym 
dictionary we obtained (for the meanings of words 
beginning with G) and Appendix B contains an 
extract of the hierarchy whose top concept is 
CIIOSE (thing). 
i14 3 3 
Tops 
'6tre' (being) 
'action' (action) 
'autorit6' (authority) 
'bien' (good) 
'bordure' (edge) 
'coiffure' (hairstyle) 
'coup. 1' (blow) 
'dimension" (dimension) 
'effort' (effort) 
'feu.2' (fire) 
"habilet6' (skill) 
'lien' (link) 
'manque' (lack) 
'nourriture' (food) 
'quantit6' (quantity) 
remarque' (remark) 
Tops nb 
of 
'fige' (age) 
'acc616ration' (acceler- 
ation) 
'affairc' (business) 
'b~.timent. 1' (buikling) .55 1 
_'b_°)sss' fl'!,_lrin_k\[) ....... 
'chose' (tiring) __221 _\[ 
'contenu' (content) 
'demandc' (request) 4 I 
i. 
',to,n.,~ge 2' (,tamago) \[ 
'tait' (ram -7)---| 
'tbrce' (strength) -42-- 
'intervalle' (interval) 
'mani$re' (manner) 545 / 
'nombre" (number) -7----|t 
'organisation' (organiza- 2 
tion) --5-- 
'religion' (religion) ------\] 
'reste' (remains) 2 I 'rfcit' (tale) 
'r6gle.2' (rule) 11 I 'son' (sound) 
--'v;i~q;;;?i ~ (v,--,~7.\]e) ....... U--\[ 'w,~.3" (v~w) 
....................................... \[ 
Table I. l.ist of top conccpk~ and number of sons 
nb 
of 
SOilS 
557 
637 
n____ 
5 
2 
2 
7 
3 
15 
5 
2 
3 
33 
13 
14 
51 
w_ 
4 
7 
14 
3 
Tol~ 
'absence' (absence) 
'activit6' (activity) 
"avantage' (advantage) 
'blessure' (wound) 
'borne' (boundary) 
'cons6quence' (conse- 
quence) 
'degr6' (degree) 
"direction. 1' (direction) 
'espace. I' (space) 
'fonction.2" (function) 
"habitude' (habit) 
'ligne. 1' (line) 
'motif' (motive) 
"obstacle" (obstacle) 
'relief' (relief) 
'renseignement' (infor- 
mation) 
'r6ponse" (reply) 
'substance' (substance) 
'616ment' (element) 
nb 
of 
sons 
11 
57 
7 
5 
2 
714 
6 
5 
93 
5 
3 
17 
4 
7 
3 
13 
2 
177 
663 
4 115 
Appendix A 
G 
gain (gain) 
1. somme.2 (sum) 
2. quantit6 d argent (amount of money) 
3. quantit6 (quantity) 
gala (gala) 
I. frte.l (commemoration) 
2. manifestation.1 (event) 
3. action (action) 
galerie (gallery) 
1. passage.2 (passage) 
2. lieu (place) 
3. PARTIE1 DE espace.l (PARTI OF space) 
4. partie. 1 (part. 1) 
galon (stripe) 
I. bande.l DE tissu (strip OF material) 
2. bande.l (strip) 
3. morceau. 1 (piece) 
4. PARTIEI DE objet.1 (PART OF object) 
5. patlie.1 (part) 
galop (gallop) 
1. aUure (gait) 
2. MANIERE DE aller (MANNER OF to go) 
3. mani~re (manner) 
gamin (kid) 
1. enfant (child) 
2. personne (person) 
3. 6tre humain (human being) 
4. 6tre vivant (living being) 
5. 6tre (being) 
gamme (scale) 
1. SUITE2 DE sons (SERIES OF sounds) 
2. suite.2 (series) 
gant (glove) 
i. vrtement (item of clothing) 
2. objet.l (object) 
3. chose (thing) 
garage (garage) 
1. bfitiment.l (building) 
garagiste (garage owner) 
1. homme.3 (man) 
2. 6tre humain de sexe masculin (human being of 
male sex) 
3. 6tre humain (human being) 
4. 6tre vivant (living being) 
5. Etre (being) 
garantie (guarantee) 
1. responsabilit6 (liability) 
2. obligation. 1 (obligation) 
3. devoir (duty) 
4. travail, l (work) 
5. aetivit6 (activity) 
garc, on (boy) 
l. enfant de sexe masculin (child of male sex) 
2. enfant (child) 
3. personne (person) 
4. 6tre humain (human being) 
5. 6tre vivant (living being) 
6. 6ire (being) 
garde (surveillance) 
1. ACTION DE surveiUer (ACTION OF to look 
after) 
2. action (action) 
116 5 5 
Appendix B 
chose (thing) 
chose nouveUe (new thing) 
I nouveaut6.2 (novelty) 
chose vraie (true thing) 
I v6rit6.2 (truth) 
enigme (riddle) 
merveille (marvel) 
objet. 1 (object) 
objct.1 crcux (hollow object) 
moule (mould) 
panier (basket) 
I corbeille (small basket) 
r6cipient (container) 
r6,cipient petit (small container) 
I lasso (cup) 
baignoire (bath) 
bol. 1 (bowl) 
bouteille. 1 (bottle) 
I carafe (decanter) 
I flacon (flask) 
cendrier (ashtray) 
cuvette (basin) 
I lavabo (wash bowl) 
I 6vier (sink) 
pot (pot) 
poubelle (trash can) 
r6servoir.2 (tank) 
seau (bucket) 
tonneau. 1 (barrel) 
vase. 1 (vase) 
objct.1 fabriqu6 (manufactured object) 
instrument (instrument) 
I instrument I)E m6tal (instrument made OF metal) 
\] \] cloche (bell) 
\] instrument de musique (musical instrument) 
I I guitarc (guita_r) 
I \[ orgue (organ) 

REFERENCES 

Natural Language Analysis Using Conceptual 
Graphs A. Berard-l)ugourd, J. Fargues, M.C Landau 
Proceedings International Computer Science Conference 
'88 llong-Kong, Dec. 88, pp 265-272. 

Natural Language Information Retrieval from 
French Texts A. Berard-Dugourd, J. Fargues, MC. 
Landau, JP Rogala Proceedings Third Annual Work- 
shop on Conceptual Graphs St Paul, Minnesota, Aug. 
88 

Conceptual Graphs for Semantics and Inlormatiou 
Processing J. Fargues, MC I,andau, A. Dugnurd, !,. 
Catach IBM Journal of' P, esearctl and l)evelopment Vol 
30, No 1, pp 70-79. 

Pour une loglque du sens R.. Martin, IJinguistique 
Nouvelle, Presses Universitaires de France, 1983. 

Elaboration d'un dictionnaire s6mantique 
infbrmatis~: un r6seau d'hyperonymes pour la 
compr6hension du langage naturel A. Perrin, R, apport 
de I)EA, Centre Scientilique IBM de Paris, Sept. 89. 

Conceptual Structures: Information Processing in 
Mind and Machine J.F. Sowa, Systems Programming 
Series, Addison Wesley, Keading, 1984. 

Synonymy and Semantic Classification K. Sparck 
Jones, Edinburgh Press University, 1986. 

Discovering Relationships among Word Senses R_I. 
Byrd, IBM Research Report Number RC 14799, 1989 
