close

Вход

Забыли?

вход по аккаунту

?

Презентация

код для вставкиСкачать
Semiotic Models for Cognitive Processing of
Language Information about Translation
Difficulties
Alexander A. Durnovo, Igor M. Zatsman
Institute for Informatics Problems of RAS, 44/2
Vavilova Street, Moscow, 119333 Russia,
iz_ipi@a170.ipi.ac.ru
1
We are describing two semiotic models for cognitive processing of
language information about translation difficulties.
They are based on Wierzbicki – Nakamori approach which divides
personal, collective (group), and conventional knowledge to pose
the problem of knowledge creation and verification.
We are working on the project for creation of goal-oriented
knowledge system (GOKS) about translation difficulties.
The models display evolution stages of a GOKS as a computer
systems in the time domain.
The models are illustrated with our investigation of the asymmetry
of the French and Russian languages.
2
We plan to use the paper corpus of literary texts in Russian
available in implementing the project as a source of translation.
In these texts the fragments are marked which most clearly
manifested the asymmetry of the French and Russian,
and which have caused difficulties for French translators.
The corpus of texts was formed during the period of 19 years
(1992 – 2010) while analyzing texts of about 70 prose and drama
works of different sizes.
3
One of the objectives of the project is creation of a
computer model of the translation difficulties (TD) typology.
Two semiotic models, the stationary and the non-stationary
(based on the first one), are supposed to be used for this pupose.
4
The stationary semiotic model
The stationary semiotic model supposes all composing
entities do not change in time.
The model is based on the Frege’s triangle, where the three
vertices are the sign-meaning (the concept), the sign-form,
and the denotatum of the sign.
5
Each vertex is assigned a computer code of its own. The set of three codes assigned
to the vertices of a triangle we call a digital semiotic triangle.
The model includes five components. These are:
the mental sphere, the material sphere of physical objects and effects,
the social and digital spheres;
the denotatum, the corresponding concept (sign-meaning) of the mental
sphere the information object (sign-form) of the social sphere,
the computer codes for the denotatum, the concept,
and the information object respectively, of the digital sphere;
the sign, joining the concept as sign-meaning and the information object
as sign-form, being an element of the sign system of a computer
system;
the formcode, joining the information object as sign-form and the computer
code, being an element of information codes for
sign-forms of a computer system;
the semcode, joining the concept as sign-meaning and the computer code,
being an element of semantic codes for concepts of a computer
system.
6
Semiotic model for description of relations between
denotata, concepts, sign-forms, and their computer
codes (identifiers)
M aterial sphere of
ph ysical objects and
effects
D en o tatu m (as p h ysical ob ject
o r effect)
S ign m ean in g
(co n cep t)
M ental
sphere
S ign
S ig nfo rm
S em co d e
F orm co d e
S phere of
social
co m m unications
ID fo r
d en otatu m
ID fo r
co n cep t
ID fo r
sign fo rm
D igital sphere
7
The main idea of computer coding of the TD typology as a GOKS is
that each TD classified unit is analyzed and fixed from three points
of view: 1) as a denotatum, 2) as one of the GOKS concepts,
and 3) as a verbal designation of the concept.
In case of TD typology, we consider the denotatum to be a pair of
corresponding fragments of Russian and French texts, fixing a
TD unit.
As concept we consider a meaning of the definition of a TD unit,
assigned by linguists-expets.
Finally, the verbal designation is the name of the unit.
Concept definitions and names are to be assigned by the
project’s participants.
8
Гоголь
- Здравствуй, Петрович!
- Здравствовать желаю, судырь, - сказал Петрович и покосил свой глаз на руки Акакия Акакиевича,
желая высмотреть, какого рода добычу тот нес.
- А я вот к тебе, Петрович, того...
Нужно знать, что Акакий Акакиевич изъяснялся большею частью предлогами, наречиями и, наконец,
такими частицами, которые решительно не имеют никакого значения. Если же дело было очень
затруднительно, то он даже имел обыкновение совсем не оканчивать фразы, так что весьма часто,
начавши речь словами: "Это, право, совершенно того..." - а потом уже и ничего не было,
И сам он позабывал, думая, что все уже выговорил.
Denotatum B
Denotatum A
Léon Goldschman et Ernest Jaubert (1896)
– Bonjour, Petrovitch.
– Soyez le bienvenu, monsieur, répondit le tailleur dont le regard s’arrêta sur
la main du conseiller titulaire pour reconnaître ce qu’elle tenait.
– J’étais venu... pour... Je voudrais...
Nous ferons remarquer ici que le timide conseiller titulaire avait pour règle de
n’exprimer ses pensées que par des bouts de phrases, verbes, prépositions, adverbes
ou particules, qui ne formaient jamais un sens suivi.
L’affaire dont il s’agissait était-elle d’un caractère important, difficile,
jamais il ne parvenait à achever la proposition commencée. Il s’embarrassait
dans ses formules. Ce fut le cas cette fois : il resta court.
En même temps il demeura debout, immobile, oubliant ce qu’il avait voulu
dire ou croyant l’avoir dit.
Henri Mongault (1938)
« Bonjour, Pétrovitch !
– Je vous souhaite le bonjour, monsieur, répliqua Pétrovitch en
reportant son regard sur les mains de son visiteur pour voir de quelles
dépouilles elles étaient chargées.
– Eh bien, voilà, Pétrovitch, n’est-ce pas… »
Il faut savoir qu’Akaki Akakiévitch s’exprimait le plus souvent
au moyen d’adverbes, de prépositions, voire de particules absolument
dépourvues de sens. Dans les cas embarrassants, il ne terminait pas
ses phrases, et fort souvent, après avoir commencé un discours de
ce genre : « C’est vraiment tout à fait… n’est-ce pas… »,
il s’arrêtait court et croyait avoir tout dit.
9
The coding strategy must take in mind that in the process
of development of TD typology as a GOKS all its concepts
and their forms of expression may change in great extent.
Moreover, a key stage of the creation of a GOKS is
the coordination of personal concepts of participants
of the project to form collective (group) concepts.
10
The non-stationary semiotic model takes into account
the variability of personal concepts and forms of their
presentation in the time domain.
The model based on the stationary semiotic model
and is designed for computer coding concepts,
information objects and denotata under the condition
that they change over time.
11
To construct the model we define a set of points named the
“Frege’s space”.
This space is built to display the stages of evolution of concepts,
as well as corresponding information objects and denotata.
12
We fix the evolution of concepts of a GOKS at the successive
moments of time (ti, i = 1,2,...), where ti is the ith moment since
the computer system started to describe vertices of Frege’s semiotic
triangles, i.e. evolving concepts of the GOKS, as well as
corresponding information objects and denotata.
The starting moment of describing vertices is then t1.
13
It is supposed, that at ti moments ( i = 1, 2, ...) the computer system
generates digital semiotic triangles, reflecting corresponding
Frege’s triangles.
The Frege’s space is constructed according the following
requirement: any digital semiotic triangle must correspond
to just one point (t, n, m, k), consisting of the following elements:
14
t – moment when the Frege’s triangle was described in
the computer system and the corresponding digital
semiotic triangle was generated;
n – semantic code for the concept as the sign-meaning;
m – information code for the information object as
the sign-form;
k – object code for the denotatum.
15
Reflection of five digital semiotic triangles on
the Frege’s space
Time parameter can vary in this space from one to infinity. As for the other
axes, any code can be an integer number.
16
At any discrete time moment ti, the description of relationship
of any Frege’s semiotic triangle with its concept,
information object and denotatum, is fixed with the help
of the stationary semiotic model.
17
The model supposes neither deletion nor editing of their
objects.
That is why, for any ti (i = 1, 2, ...), it is posssible to see
the state of any TD concept existing at the time, as well as
to see the state of the TD typology in whole.
In other words, the change of the TD concept A does not
imply editing of its description.
Rather, it implies the creation of a new TD concept B
(which becomes an inheritor of the TD concept A) and
the description of this new TD concept.
18
Moreover, sometimes the evolution of the TD typology requires
to combine the existing TD concepts A and B into a single TD
concept C, which becomes, in this case,
the inheritor of both A and B.
So it is necessary for the system to store the information on
inheritance (possibly multiple) relations between TD concepts.
That why speaking of “the TD typology evolution”
we mean the set of chains of TD concepts, linked by the relation
of inheritance.
19
In the course of the project realization it is expected to develop a prototype
of the Russian-French TD typology corpus, formed with an experimental
array of pairs of fragments of parallel texts in Russian and French,
fixing Russian-to-French translation. It is supposed that the prototype can
contain several translations of same works made by different translators.
Each classified translation difficulty in the typology will have some hypertext
links to corresponding pairs of fragments in Russian and French, included
in the prototype, as well as back links from those pairs to difficuties
descriptions.
20
Thank you!
21
Документ
Категория
Презентации по английскому языку
Просмотров
10
Размер файла
182 Кб
Теги
1/--страниц
Пожаловаться на содержимое документа