Biblioteka Cyfrowa KLF UW (Digital Library of the Formal Linguistics Department at the University of Warsaw)

Bień, Janusz S.

On standards for Computer Modern font extensions

Co to jest TEX?

BOOK REVIEWS 
NATURAL LANGUAGE UNDERSTANDING AND LOGIC 
PROGRAMMING, IX: PROCEEDINGS OF THE SECOND 
INTERNATIONAL WORKSHOP 
Veronica Dahl and Patl'ick Saint-Dizier (eds.) 
(Simon Fraser University and IRISA, University of 
Rennes 1) 
Amsterdam: North-Holland, 1988, viii + 345 pp. 
ISBN 0-444-70408-6, $84.25 (hb) 
Reviewed by 
Janusz S. Bien, 
Warsaw University 
This volume contains 18 papers and the positions pre- 
sented at two panel sessions. The papers are rather 
heterogeneous and the editors have not classified them 
in any way. 
The collection shows clearly the fast proliferation of 
new logic grammars and Prolog dialects, which is per- 
haps a necessary stage for a relatively new domain. 
Some of the modifications are small enough to be fully 
described in short Prolog listings accompanying the 
papers, e.g., the functional ogic grammar of Michel 
Boyer or the epistemic reasoning of Manny Rayner and 
Sverker Janson; in other cases we have quite large 
systems practically unavailable to other parties, e.g., 
the three parsers compared by Toshiyuki Okunishi and 
his colleagues. Some papers take a specific linguistic 
theory or formalism as a starting point, e.g., Govern- 
ment-Binding theory or Sowa's Conceptual Graphs, 
while others augment existing logic grammar or pro- 
gramming tools. A paper on morphological nalysis of 
Italian is an example of the work where it is fully 
justified to use Prolog, but its use is of little relevance to 
the problem under consideration. 
According to my understanding of logic program- 
ming, three papers are most representative of the do- 
main. Patrick Saint-Dizier describes Contextual Discon- 
tinuous Grammars, which form the next step in the 
evolution of logic grammars tarted by Colmerauer's 
metamorphosis grammar. Secondly, Edward P. Stabler, 
Jr. reconstructs in logic a fragment of a linguistic 
theory, namely, Government-Binding theory, and trans- 
forms it by formally provable steps into a logic program 
for parsing. Last but not least, the paper by Annie Gal 
and Jack Minker on informative and cooperative an- 
swers in data bases shows the advantages of using the 
same formalism for representing database integrity con- 
straints and natural anguage semantics. 
As for the panels, I sympathize most with Miguel 
Filgueiras ("logic programming can only be seen as a 
programming methodology and not as a formalism con- 
tributing to the study of natural language understand- 
ing") and Start Szpakowicz ("the belief in Prolog as a 
perfect ool for natural anguage processing may some- 
times mean putting the cart before the horse") but they 
seemed to be in a minority. 
To summarize, the volume definitely proves the 
usefulness of the Prolog programming language for a 
wide range of natural language processing tasks. It gives 
the reader a fair account of the activities in the domain, 
but leaves also some doubts as to whether the domain is 
more coherent han, say, natural anguage understand- 
ing and Lisp programming. 
Janusz S. Bien, received his Ph.D. in computer science from 
Warsaw University in 1978. He has recently worked on a 
computational description of Polish morphology. He advo- 
cated the use of Prolog for natural anguage processing from 
the moment the first implementation became available outside 
Marseilles. Bien's address i : Instytut Informatyki UW, PKiN 
p. 850, 00-901 Warszawa, Poland. 
LANGUAGE AND SPATIAL COGNITION 
Annette Herskovits 
(Wellesley College) 
Cambridge, England: Cambridge University Press, 
1987, x + 208 pp. 
(Studies in natural anguage processing) 
ISBN 0-521-26690-4, $34.50 (hb) [20% discount o 
ACL members] 
Reviewed by 
James Pustejovsky 
Brandeis University 
In this book, Herskovits proposes an analysis of loca- 
tive expressions in English. Her goal is to position the 
study of linguistic expressions (in particular, spatial 
expressions) within a broader context of language usage 
and the conventions associated with communicating 
goals, beliefs, etc. Thus the work should be evaluated 
from the perspective of cognitive science as an interdis- 
ciplinary field, and not simply as a linguistic treatise on 
prepositions or a computational model of a subset of 
natural anguage. 
The underlying semantic theory that Herskovits as- 
sumes differs from more traditional "classical" ap- 
proaches to meaning (e.g., Tarski 1943) and is most 
similar to the positions argued for by Searle (1979) and 
Winograd (1980), and previously by Wittgenstein 
(1963). According to these authors, the view that mean- 
ing derives from the literal interpretation ofthe words in 
Computational Linguistics, Volume "15, Number 3, September 1989 187 


University of Warsaw

Not a member yet

On standards for Computer Modern font extensions

Co to jest TEX?

Book Reviews: Natural Language Understanding and Logic Programming, II: Proceedings of the Second International Workshop

Koncepcja słownikowych tablic morfologicznych (na przykładzie czasownika)

Komputery a język naturalny

Bibliografia Instytutu Informatyki Uniwersytetu Warszawskiego za rok 1987

Język w słowniku : desiderata semantyczne do wielkiego słownika polszczyzny

Toward computer usable dictionary of Polish

Informator Instytutu Informatyki Uniwersytetu Warszawskiego za rok 1987

Koncepcja Tokarskiego słownikowej informacji morfologicznej

159

403