Recent Advances in Computer Science
Visual Semantic Networks of Bulgarian Possessive and
Reflexive-possessive Pronouns Inflection
VELISLAVA STOYKOVA
Bulgarian Academy of Sciences
Institute for Bulgarian Language – BAS
52, Shipchensky prohod str., bl. 17
1113 Sofia
BULGARIA
[email protected]
CHAVDAR LOZANOV
Sofia University “St. Kliment Ohridsky”
Faculty of Mathematics and Informatics
5, J. Bourchier blvd.
1164 Sofia
BULGARIA
[email protected]
Abstract: - The paper presents approaches to formal interpretation of Bulgarian possessive and reflexivepossessive pronouns inflectional morphology. The analysis is based on grammar features and semantic
representation of possessive and reflexive-possessive pronouns inflectional morphology and uses semantic
networks. The visual representations of Bulgarian possessive and reflexive-possessive pronouns inflectional
morphology is based on their programming encoding using DATR language for lexical knowledge
representation. It converts orthogonal relations as semantic offering a space interpretation.
Key-Words: - Semantic Networks, Knowledge Representation, Natural Language Processing, Visualization.
1 Introduction
2 The semantics and grammar
features of possessive pronouns
The semantic networks are widely known formalism
for knowledge representation tasks. They have been
used to represent various types of knowledge among
which representing knowledge of a particular
domain mostly by building a related domain
ontology. At the same time, semantic networks were
successfully used in Natural Language Processing
(NLP) to represent different types of grammar
knowledge - morphological, syntactic, etc.
Also, there exist some already established
systems using semantic networks for NLP
applications among which are the DATR language
for lexical knowledge representation, Universal
Networking Language (UNL), WordNet, etc. They
differ with respect to their principles of design and
use, semantics, scope and scale of applicability, etc.
At the same time, the inflectional morphology of
possessive and reflexive-possessive pronouns in
Bulgarian language have been developed in both
DATR [5] and in UNL [7]. A comparison of both
formal semantic network applications is given in
[10] and it relates both application with respect to
the expressive power of the formalisms used.
In further description, we are going to present
analysis of formal representation of Bulgarian
possessive and reflexive-possessive pronouns and
we will propose a visual representation of their
semantic networks encoding.
ISBN: 978-960-474-311-7
The semantics of possessive pronouns in Bulgarian
includes various relationships like: possession
(depending whether it is an object or a subject of
possession [3]), part-of-whole, relational, etc. Only
full forms of possessive pronouns have inflection
[2]. The full forms of possessive pronouns are:
’moj’ (my), ’tvoj’ (your), ’negov’ (his), ’nein’ (her),
’nash’ (our), ’vash’ (your), ’tehen’ (their). They
have grammar features of person, number, gender,
and definiteness.
The grammar feature of person is not inflectional
and expresses information both at the level of syntax
and at hypertext level through agreement. The full
forms imply information both about the possessor
and the object being possessed using agreement in
number and gender.
The grammar feature of definiteness implies
information about possession at syntactic level using
agreement and is expressed by a formal
morphological marker which is an ending
morpheme [2]. It is different for genders however,
for masculine gender two types of definite
morphemes exist – to determine a defined in a
different way entities, which have two phonetic
alternations, respectively. For feminine and for
neuter gender only one definite morpheme exists,
138
Recent Advances in Computer Science
pronouns is given at [5] and presents an inheritance
semantic network consisting of different inflectional
type nodes which uses a rule-based formal grammar
and a lexical database (the pronouns). The particular
queries to be evaluated are related inflected word
forms.
The interpretation is based on the adjectives
encoding [4] and takes as a starting point linguistic
motivation, in particular, the priority of one or
another grammar feature. Thus, the feature of
gender is accepted as a specific trigger to change the
values of inflected forms for the features of number
and definiteness. The DATR account of Bulgarian
inflectional morphology offers space semantic
networks representation for nouns [6], adjectives [8]
and numerals [9] as well. The encoding is as
follows*:
respectively. For plural, two definite morphemes are
used depending on the ending vocal of the main
plural form.
The features of gender and number of the
definite article are different from gender and
number features of the possessive pronouns,
themselves. The former are inflectional whereas the
later are not inflectional, even both they can express
agreement.
3 The DATR language for lexical
knowledge representation
The DATR language is a non-monotonic language
for defining inheritance networks through path /
value equations [1]. It has both an explicit
declarative semantics and an explicit theory of
inference allowing efficient implementation, and at
the same time, it has the necessary expressive power
to encode lexical entries presupposed by the work in
unification grammar tradition. In DATR information
is organized as a network of nodes, where a node is
a collection of related information.
Each node has associated with it a set of
equations that define partial functions from paths to
values where paths and values are both sequences of
atoms. Atoms in paths are sometimes referred to as
attributes. DATR is functional, it defines a mapping
which assigns unique values to node attribute-path
pair, and the recovery of these values is
deterministic.
The semantics of DATR uses non-monotonic
inference and default inheritance, and allows
generalization-capturing
representations
of
inflectional morphology. DATR has expressive
power which is capable to encode and process both
syntactic and morphological rules and allows
representation of grammar knowledge by using
semantic networks. The DATR language has a lot of
implementations however the analyzed application
was made by using QDATR 2.0 [11] (see related
file bul_det.dtr). This PROLOG encoding uses
Sussex DATR notation. DATR allows construction
of various types of language models (language
theories), and the implementation allows to process
words in Cyrillic alphabet.
The encoding presents inflectional rules for
generation of all related inflected forms of
4 The DATR encoding of Bulgarian
possessive
pronouns
inflectional
morphology
*
Here and elsewhere in the description we use Latin alphabet
to present morphemes instead Cyrillic used. Because of
mismatching between both some of typically Bulgarian
phonological alternations are assigned by two letters, whereas in
Cyrillic alphabet they are marked by one.
The available published DATR encoding of
inflectional morphology of Bulgarian possessive
ISBN: 978-960-474-311-7
139
Recent Advances in Computer Science
possessive pronouns. Node DET defines definite
inflectional morphemes and all other nodes define
inflectional rules for 4 related inflectional types.
Thus, node Adj defines the rules for the pronouns
’negov’ and ’nein’; node Adj_2 defines the rules for
the pronoun ’tehen’; node Adj_4 defines the rules
for the pronouns ’nash’ and ’vash’, and node Adj_5
defines the inflectional rules for the pronouns ’moj’
and ’tvoj’. The pronouns are given as a different
node defined by its person, number, gender and
inflectional roots. Thus, the pronoun ’moj’ is given
as follows:
7 Visual representation
The orthogonal structure of DATR and the related
encoding of possessive pronouns inflectional
morphology allow visual space representation.
Generally, the visualization techniques are based on
estimation of statistical models. However, for our
representation we interpret orthogonal relations as
semantic as it was presented in [6, 8, 9].
The visual representation of possessive pronouns
inflectional morphology is presented at Fig. 1.
Its generated inflected forms are given at the
Appendix. The DATR interpretation of possessive
pronouns uses inheritance hierarchical formal
representation to interpret inflectional morphology
rules and uses 4 inflectional rules most of which
were defined for adjectives. It accounts for sound
alternations and for irregular inflected forms. It also
uses semantic hierarchical representation of the
inflectional grammar features of gender, number
and definiteness and concise encoding.
Fig. 1. Visual representation of possessive pronouns
inflectional morphology.
6 The semantics and grammar
features
of
reflexive-possessive
pronoun
Similarly, Fig. 2 presents visualization of reflexivepossessive pronouns inflectional morphology.
The semantics of reflexive-possessive pronoun
combines semantics of possession relationship and
that of reflexivity. That means it expresses
possession relationship between the possessor
(defined by the subject in the sentence, and agreed
with it in gender and number) and the thing being
possessed (to which the pronoun is referred, and
agrees in gender and number). The reflexivepossessive pronoun is one and it has a full and a
short form, and both they can be used with respect
to the agreement. However, only its full form ’svoj’
(-self) has the inflectional grammar features of
gender, number, and definiteness which are similar
to that of the adjectives and of the possessive
pronouns.
The DATR formal account of reflexivepossessive pronoun inflectional morphology uses
inflectional rules already defined at node Adj_5 and
uses the same principle as for the possessive
pronouns.
ISBN: 978-960-474-311-7
Fig. 2. Visual representation of reflexive-possessive
pronouns inflectional morphology.
140
Recent Advances in Computer Science
[8] Stoykova, V., Visual Representation of
Bulgarian Adjectives Inflectional Morphology,
In Advances in Applied Information Sciences,
2012, WSEAS Press, pp. 91-96.
[9] Stoykova, V., Space Representation of
Inflectional Morphology for Numerals in
Bulgarian Language, In Latest Trends in
Information Technology, 2012, WSEAS Press,
pp. 232-236.
[10] Stoykova, V., Formal Representations of
Bulgarian Possessive and Reflexive-possessive
Pronouns, In Recent Researches in Information
Science and Applications, 2013, WSEAS Press,
pp. 144-149.
[11] The DATR Web Pages at Sussex. 1997.
http://www.cogs.susx.ac.uk/lab/nlp/datr/
8 Conclusion
The presented analysis of encoding inflectional
morphology of Bulgarian possessive and reflexivepossessive pronouns using DATR language for
lexical knowledge representation is accepted as a
starting point for space representation. The related
encoding define grammar rules for generation of all
related inflected forms and concise encoding of 4
inflectional types represented in orthogonal
semantic network structure.
Further, the orthogonal relations are interpreted
as semantic and are visualized in space
representation.
Appendix:
References:
[1] Evans, R. and Gazdar, G. DATR: A Language
for Lexical Knowledge Representation.
Computational Linguistics 22(2), 1996, pp.
167–216.
[2] Gramatika na suvremennia bulgarski knizoven
ezik. Morphologia, tom. 2 , BAS Publishing
house, Sofia., 1983. (in Bulgarian)
[3] Nicolova, R. The Bulgarian Pronouns, Nauka i
izkustvo, Sofia, 1986. (in Bulgarian)
[4] Stoykova, V., The Definite Article of Bulgarian
Adjectives and Numerals in DATR. In C.
Bussler and D. Fensel eds., Artificial
Intelligence: Methodology, Systems, and
Applications. Lecture Notes in Artificial
Intelligence 3192, Springer-Verlag, 2004, pp.
256–266.
[5] Stoykova, V., Bulgarian Possessive and
Reflexive-possessive Pronouns in DATR. In R.
Trappl (ed.) Cybernetics and Systems 2010,
Vienna: Austrian Society for Cybernetic
Studies, 2010, pp. 426–432.
[6] Stoykova, V. and Lozanov, Ch., The Geomety
of Language – a Space Semantic Network of
Bulgarian Nominal Inflectional Morphology. In
N. Mastorakis, V. Mladenov, Z. Bojkovic and
D. Simian eds., Latest Trends on Computers,
vol. 2, 2010, WSEAS Press, pp. 721-725.
[7] Stoykova, V., The Inflectional Morphology of
Bulgarian Possessive and Reflexive-possessive
Pronouns in Universal Networking Language,
In A. Karahoca and S. Kanbul eds.,
Proceedings of the 1st World Conference on
Innovation and Computer Sciences (INSODE),
Procedia Technology, vol. 1, 2012, Elsevier,
pp. 400-406.
ISBN: 978-960-474-311-7
141