Logical analysis of natural language is able to extract
semantic relations that follow the underlying logical
formalism. Transparent Intensional Logic (TIL) has been
designed to capture even high-order relations between sentence
elements and systematically work with all kinds of language
references, i.e. extensions, intensions and hyperintensions. In
this paper, we introduce the first version of a new tool,
called AST, for automatic semantic analysis of sentence. This
tool is based on the TIL logic processing rules as they were
implemented in the SYNT parser, in its logical analysis module.
AST thus shares lexicons and semantic rules in the same format
as in the SYNT parser, but allows to build upon the output of
other syntactic parsers. AST is designed as a universal
semantic analysis tool, which strictly separates the
application logic and input data and strives for language
independent analysis. Within the evaluation, we present
preliminary results of testing AST on selected problematic
phenomena, which were not correctly processed by the SYNT
logical analysis.
[1]
Aleš Horák,et al.
Types in Transparent Intensional Logic and Easel - a Comparison
,
2004
.
[2]
Marie Duzí,et al.
Procedural Semantics for Hyperintensional Logic - Foundations and Applications of Transparent Intensional Logic
,
2010,
Logic, Epistemology, and the Unity of Science.
[3]
Ales Horák,et al.
Linguistic Logical Analysis of Direct Speech
,
2012,
RASLAN.
[4]
Pavel Tichý,et al.
The Foundations of Frege's Logic
,
1988
.
[5]
Zuzana Neverilová,et al.
Common Sense Inference Using Verb Valency Frames
,
2012,
TSD.