ed a theory of 
ial generative 
oression of a 
ming of the 
'er of formal 
es,.etc.), as it 
ts substantial 
sh) should be 
ble to: 
?le statements 
ïscription that 
for a typical 
3-parts 
asal structure 
> for simple 
erivative tree 
es should re- 
g morphemes 
le sentences 
îles should, at 
tences with a 
ges, from the 
of processing 
statements is 
: order to be 
has produced 
SE 
Si PHRASE 
P-PLURAL 
NOUN 
built area 
L'*M? * ‘.-3s£2 
Æ A'.'.U / i M 
To generate the wood surrounded the built area, one 
would apply transformations mapping “TENSE 
-(-surround” to “surround +PAST”. A morphophonemic 
rule would then map “surround +PAST” to surrounded. 
The same phrase, in the passive form, has produced the 
following derivation tree: 
SENTENCE 
NOUN PHRASE 
VERB PHRASE 
NP-PLURAL 
VERB NOUN PHRASE 
DET NOUN 
AUX V NP-SINGULAR 
TENSE 
The built area 
is 
DET NOUN 
by the wood 
surrounded 
The passive sentences could be different: the verb to 
belong to has the function of passive form of the verb to 
get: 
houses belong to the class of buildings; 
the class of buildings gets the houses. 
Parsing is the delinearization of linguistic input, that is 
the usage of syntax and of other forms of knowledge to 
the aim of estimating the function of words in the input 
sentence. 
A parser (Weizenbaum, 1966, Vinograd, 1971) can be 
viewed as a recursive comparator of rules, pattern 
matcher which tries to map a string of words onto a set of 
meaningful syntactic patterns. 
The sentence a bridge crosses (over) a river could be 
matched to the pattern: 
SENTENCE 
SUBJECT PREDICAT 
VERB OBJECT 
A sequence of sentences: true, like the previous one, or 
false, like a bridge crosses (down) a river (indeed a 
tunnel crosses (down) a river), could be used in the data 
validation. 
Several types of parsers have developed; most of the NL 
programs perform parsing by matching their input against 
a series of predefined templates (template matching). 
Some programs of NL for parsing have a matching of 
their input relative to a set of previous templates. 
5. Conclusions 
This graph syntetizes , by means of a diagram picture, the 
whole of present situation, as far as Geomatics may be 
inferred from other disciplines 
Mathematics 
Statistics 
Natural Sciences 
Geodesy, Cartography 
and related sciences 
Mathematical Logic 
(Symbolic, Boolean, Fuzzy) 
I 
Hilbert 
Turing CZ) Computer Science Geomatics 
and Information Technologies 
Wiener 
Artificial Intelligence 
A 
Linguistics 
(Formalism,Structuralism) 
Politics 
Law 
Economy 
But Godel’s theorem (1931) states a limit to problem 
solving ability of computers and Wittgenstein (1921) 
said: a word hasn’t meaning, but it has its uses only. 
References 
Barr, A., Feigenbaum, E., 1981. The handbook of 
Artificial Intelligence. Kaufmann, Los Altos, Ca 
Boole, G., 1847. The Mathematical Analysis of Logic. 
Macmillan, Cambridge 
Boole, G., 11854. The Laws of Thought. Bell, London 
de Saussure. F., 1912. Cours de linguistique générale. 
Paris 
Chomsky, N., 1957. Syntactic structures. Mouton, The 
Hague 
Hadamard, J., 1945. The Psycology of Invention in the 
Mathematical Field. University Press, Princeton 
Jakobson, R., 1963. Essais de linguistique générale. Paris 
Kondratov A., 1968. Zvuki i Znaki. Mir, Moskva 
Mounin, G., 1968. Clefs pour la linguistique. Seghers, 
Paris 
Wiener, N., 1950. The human use of human brings. 
Houghton Mifflin, Boston 
Wittgenstein, L., 1939. Lectures on the foundations of 
mathematics. Cambridge 
71 
¥m 
illlllp 
¡¡rissi