Lingvist oxirgi shakl 1-son indd

Aziza RAKHMANOVA
158

---

as mapping, semantic mapping, and choosing an equivalent word.
A corpus of parallel texts on a work of art allows, first of all, 
to analyze the similarities and differences in the grammar of two or 
more languages. The corpus of parallel texts is important in com-
paring the features of artistic style in both languages. In the artistic 
style, image tools, movements, figurative expressions are reflected 
in accordance with the traditions of each language. For example, 
“musk” (a black aromatic substance) is used for eyebrows, hair, 
“shahd” (honey) for lips or words, “tulip”, “ruby” for lips, “bow” or 
“bow” for a curved figure. Attribution to the use of camels, caravans, 
camels, horses, and dogs as artistic symbols show the image style 
characteristic of Uzbek classical texts. If the grammatical features of 
languages are expressed in the author’s speech, the harmonious use 
of literary language and dialect words in the speech of the characters 
serves to illuminate the skill of imagery. The corpus of artistic texts 
can also be classified in terms of genres: poetic texts, prose texts, and 
the corpus of folklore texts. It is possible to form a corpus of parallel 
texts even in a single section of a large work.
5. Parallel corpus based on the text of dictionaries. Explana-
tory dictionaries, homonyms, synonyms, antonyms, phraseological 
units, dictionaries of compound words can be taken as linguistic 
support of parallel corpus. A parallel corpus can be used to identify 
translation difficulties. One of the main problems of translation is 
to find the exact alternative of fixed combinations such as phrases, 
proverbs, metal in the language being translated. Although words in 
two languages have similar meanings, they may belong to different 
semantic frameworks or contexts, so complete translation equiv-
alence between the source language and the target text is rare. A 
corpus of parallel texts based on works of art clearly has problems 
with idioms. Phrases are composed of two or more words and serve 
to form a new portable dictionary meaning based on the semantics 
of these words. Therefore, if a direct machine translation is used in 
the formation of a corpus of parallel texts, problems arise in the cor-
rect interpretation of semantics. In order to solve problems, phras-
es should be distinguished from simple and combined lexical units, 
word combinations, and marked with special tags and labels. The 
translation of idioms definitely requires expert supervision. For ex-
ample, the English phraseology “have eyes in the back of one’s head” 
is literally translated by a translator program or artificial intelli-
gence. As a result, an understanding contrary to reality is formed. It 
is unusual to have eyes on the back. The meaning of the phraseolo-
The role of parallel corpora in the improvement of artificial intelligence
159

---

gy “to be aware of everything” is not reflected. In the Uzbek-English 
parallel corpus, the provision of the phrase “a snake moves under 
the ground” as an alternative to the phraseology of “have eyes in the 
back of one’s head” and as its appropriate translation ensures the 
correctness and coherence of the translation. The same concept will 
exist in the mind of artificial intelligence.
In essence, the corpus of parallel texts performs two main 
tasks:
1. Comparison of texts of the same type in two, three or 
several languages.
2. Comparison of texts translated from the source language to 
the target language.
The role of parallel texts in illuminating the characteristics of 
languages in a comparative aspect is explained [Garabik, 2006:81-
87]. Its lexicographic capabilities, its importance as a source for 
creating translation dictionaries [Belyayeva, 2002; 55-64] are 
shown. Currently, parallel corpus have been created that reflect the 
text features of Russian, English, German, Japanese, Finnish, and 
Slovak languages.
The corpus of unrelated languages, in particular, Uzbek-
English parallel texts, is important in learning a foreign language and 
culture.
The creation of a corpus of parallel texts serves to increase 
the prestige of the Uzbek language and strengthen its position. First 
of all, the corpus of parallel texts performs a communicative function 
as a database. The translation of a text in one language into another is 
aimed at identifying intercultural relations and showing differences. 
An opportunity is created for semantic analysis of lexical units 
through translation options in other languages. It will be possible 
to determine the general and specific features of information 
transmission in different languages.
The corpus of parallel texts serves to compare different 
cultures based on different languages, to learn the essence of 
lexical units representing cultural relations. Phonetic, lexical, 
morphological, syntactic features of languages can be compared and 
contrasted through the corpus of parallel texts. Such corpus are also 
notable for their ability to provide information on the categories 
specific to word groups, the expression of grammatical meaning, and 
the system of word formation.
Parallel corpora play an important role in the process of 
language learning, language teaching, translation, and serve to 
Aziza RAKHMANOVA
160

---

expand the capabilities of artificial intelligence and enrich its 
artificial language base with features of the source language and the 
language being translated.