The expanded system of the coordinated dictionaries and its using for universal semantic coding and translation of the polysemous source text (continuation)
The translation of the encoded source text into target languages.
The process of translation is in progress in a computer of a reader by the machine translation program, which includes the basic program containing own traditional bilingual dictionary and the special subprogram containing the coordinated bilingual dictionary (c.b.d.) described above.
The basic program analyzes the text a word by word and, in case the next in turn word does not contain a semantic addition, this program carries out the translation with using own traditional bilingual dictionary;
but in case the next in turn word contains a semantic addition, the mentioned special subprogram finds in the c.b.d. the translation of this word or the word-combination, taking into account the semantic addition, namely:
- if the semantic addition to the given word contains the ill.e. characterizing the use of a word, the subprogram finds in the c.b.d. the translation of this ill.e., makes the comparative analysis of the source text and of the text of translated ill.e., chooses the translation of the word in the text of ill.e. and transfers the received translation to the basic program;
- if the semantic addition to the given word contains the word-combination, the subprogram finds in c.b.d. its translation and transfers the received translation to the basic program as a translation of all group of the words which were marked in a source phrase by mentioned boundary symbols and a special font; the wedged words, non-marked by the special font, the subprogram translates as attributes, adverbial modifiers, the parenthetical words;
- if the semantic addition to the given word contains the structural formula of word-combination characterizing lexical variability of a word-combination in the source language, the subprogram finds translation in c.b.d., translates the mentioned above group of the words, taking into account the structural formula, and transfers the received translation to the basic program;
-if the semantic addition to the given word contains ill.e. characterizing the use of the word-combination, the subprogram finds in the c.b.d. the translation of this ill.e., makes the comparative analysis of the source text and of the text of translated ill.e., chooses the translation of the word-combination in the text of ill.e. and transfers the received translation to the basic program;
- if in semantic addition to the given word the differentiated (morphological) meaning is specified, the subprogram carries out translating of this word and other words connected with him, in the following way:
- if the given word is a personal, possessive, reflexive pronoun, the subprogram carries out the translating of a part of phrase (i.e. the translating of the given pronoun and the words connected with it) with the account of the meaning/ c.w. which is specified in the semantic addition, namely:
the meaning concerning the gender and /or number, the person or an inanimate object,
or the c.w. replacing the given personal or reflexive pronoun,
or the c.w. instead of which is used the given possessive pronoun-noun;
- if the given word is verb, the subprogram carries out the translating of a part of phrase connected with this verb with the account of the meaning which is specified in the semantic addition, namely, the meaning concerning the action or the state, reached as a result of action, etc., etc.;
- if the semantic addition to the given word contains the differentiated (syntactical) meaning specifying that the word is the removed / isolated attribute to c.w. (or adverbial modifier to c.w. expressing action) in the given sentence which is specified in the addition, and then carries out translating of the given word and other words in a syntactical link with this c.w.;
in generally case, the subroutine transforms character of the sentence’s components so that to provide in the translated text a direct syntactical link between the given word and the c.w..
The special subroutine transfers each of the received translations to the basic program, including the grammatical information characterizing the use of translation in target language; then the basic program carries out the grammatical and syntactic coordination of the got translation of a part of a phrase with translation of text as a single whole.
Thus the machine translation of the encoded source text is carried out.
The scheme reflecting the process of potential translation into this or that target language is presented on Fig. 4.
Some characteristic examples explaining operations of semantic coding and translating.
1. The phrases containing polysemantic words.
a) The source text: To clear up the cause of trouble.
The author has marked in the native English c.e.d. one of two meanings of word-combination clear sth up (the choice depends on the context’s sense); analogously author has marked one of many meanings of the word trouble (the dictionary entries are presented below). The service program has transferred to the source text the meanings marked by the author; then the author has marked the borders of a word-combination and has allocated its by a bold font.
As a result, the following variants of the coded text and their machine translations, for example, into Russian are received (for brevity, the variants of semantic meanings are separated from each other by a symbol /):
b) The source text: We’d better find out how the land lies before taking any action…
The author has marked in the dictionary entry land n the word-combination see, etc. how the land lies, in which different verbs may be used; then the author has allocated by a bold font the words conterminous with an idiom, including a verb find out:
As we see, the usual program of MT has not found out a word-combination in the given text (because of variable lexical elements) and executed the word-by-word translation.
c) The source text (Deutsch): Also ich finde, da ist was dran.
The author has marked in the native Deutsche c.e.d. the word-combinations ich finde and da ist etwas dran; then author has marked the appropriate ill.e(s) of this word-combinations (the dictionary entries are presented below). The service program has transferred to the source text the all elements marked by the author; then the author has marked the borders of the word-combinations. As a result the following coded text and its machine translation, for example, into Russian is received:
As we see, the usual program of MT executed the word-by-word translation.
2. The phrases containing the morphological polysemuos words.
Also let’s explain this, using the characteristic examples which concern the mentioned differentiated (morphological) meanings.
a) the polysemy of the possessive pronoun-noun mine (for brevity, the different variants of phrase are separated from each other by a symbol //).
The source text: It is your pistol // your rifle // your gun? I was surprised by this question and I have answered: yes, it’s mine.
Your photo is lovely than others – do you want to see mine?
The author has defined that the pronoun mine (ours, yours) may express depending on the context the belonging of a person of male gender or of feminine gender either the belonging of an inanimate object or the belonging of an animal of indefinite gender or the belonging of the same, what are enumerated, but in plural number (the dictionary entry is presented below); the choice of the concrete form of a pronoun depends on a gender and / or number of the contextual noun replaced by the given pronoun.
If the pronoun expresses the belonging of an inanimate object or an animal of indefinite gender, then the gender in target language may does not coincide with a gender in the source language.
In these cases it is always necessary to mark in the source text a contextual noun.
The author has marked in the native English c.e.d. this or another meaning of pronoun mine and the contextual noun (c.w.) replaced by pronoun.
As a result, the following variants of coded text and their machine translations, for example, into Russian are received:
As we see, these translations of pronouns correspond to the gender of the contextual nouns which they replace.
For comparison, the translation using the usual program of MT:
As we see, these translations of pronouns don't correspond to the gender of the contextual nouns which they replace.
b) the polysemy of the verbal forms in English.
The source text: The bathyscaph was safely dipped on the ocean floor.
The bathyscaph was safely dipped on the ocean floor during several hours. At last moment …
The bathyscaph was safely dipped on the ocean floor during several hours.
The author has defined that in the first sentence the matter concerns a state reached as a result of actions, in the second and in the third sentences the matter concerns actions, since in the end of these sentences is pointed out an interval of time … during several hours (the dictionary entry is presented below).
Further, author has come to conclusion that in a context of the second sentence a matter concerns the action of an Imperfective and in a context of the third sentence a matter concerns the action of a Perfective.
The author, on the basis of the executed analysis, has marked in section of the differentiated (morphological) meanings the different forms of a Passive Voice.
As a result, the following coded texts and their machine translations into Russian are received:
As we see, the actual sense of a context is not reflected in the translations executed without taking into account semantic additions.
3. The phrases possessing a syntactic ambiguity / syntactic incompleteness.
At last, let’s explain this, using the characteristic examples which concern the mentioned differentiated (syntactic) meanings [12*].
a)The source text(Russian): Всё это была лишь идея, основанная на анализе исторических событий, опережающая время.
Author has defined that the word опережающая is a part of the removed, separated attribute to c.w. идея; on the basis of it the author has marked in c.e.d. a corresponding syntactic meaning of the word опережающая, and also has marked the c.w. идея.
As a result the following coded text is received:
The special subprogram has found out, that the word опережающая is a part of an independent part of the sentence - the isolated participal phrase. On the basis of it the subprogram has transformed the sentence, namely, transformed the isolated participial phrase in the second autonomous sentence with the formal compound predicate это была (i.e. it was) and with the c.w. идея (i.e. idea) and its attribute опережающая (i.e. anticipatory).
As a result, the following machine translation into English is received:
Thus the direct syntactic link between words идея and опережающая (idea and anticipatory) is provided.
For comparison, the machine translation using the usual program of MT:
As we see, the syntactic links had become ambiguous as the participle anticipatory has no characteristics of a gender and number and may be related both to word experiences and to word idea.
As already noted above, the following designations in the sections of differentiated (syntactic) meanings are accepted:
c.w.* / к.с.* - the concrete contextual word which the author should marked in the source text.
b) The source text: The syntactic links characterize the some predicative (i.e. relating to a predicate) and non-predicative (i.e. including attributive, adverbial) syntactic relations.
Author has defined:
- the words predicative, non-predicative, attributive, adverbial are separated attributes to the word relations(018),
-the word relating is separated apposition to the word predicative(07),
-the word including is separated apposition to the word non-predicative(013).
The syntactic links in the given phrase are graphically presented on the Fig. 5.
As a result, the following coded text is received:
Thus the direct syntactic links between the all attributes/ appositions and the contextual words are provided.
For comparison, the machine translation using the usual program of MT:
As we see, if the machine translation is executed without taking into account semantic additions, the syntactic links between words are disturbed, namely: the all attributes/ appositions are torn off from the word отношения.
Some considerations of general character.
The service program possesses a some enterprising functions, for example, indicates to the author on the discrepancy of the use of a word or a word-combination in the source text and in the marked element of an dictionary entry, and also indicates the words missed in process of the analysis of the source text, but which, probably, possess an one or another ambiguity (Fig. 3).
Semantic coding does not depend into which target language the translation will be potentially carried out:
the fragment of a dictionary entry executes function of a code indicating an exact way from marked element to a headword in c.e.d., and from a headword to translation of this element in any of the c.b.d(s); the source text has a wide spectrum of meanings as a signal containing a multitude of additional frequencies. The semantic coding is a filter allowing to shorten this spectrum.
Each of semantic additions can be presented as well in more compact form - in the form of the numbers of the elements of dictionary entries. The specified numbers are defining automatically by means of the service program which analyzes a way from a marked element to a headword (in c.e.d.) and a way from a headword to a translation of this element (in c.b.d.).
Creation of systems of the coordinated dictionaries demands efforts of linguists, but is not connected with the decision of any transcendental problems.
Rules of semantic coding can be stated in the form of instructions which are no more complex, than a well-known grammatical and syntactical rules, and also the rules of an using the dictionary.
The coded source texts (descriptions of inventions, scientific articles, etc.) at their publication in the electronic form can contain the "invisible" semantic additions.
Also it is expedient the texts of journal articles, descriptions of inventions, the directories published in the various countries on native languages to represent in the form of the files containing universal semantic additions; it concerns also personal documents (the vitas etc.).
Evidently, all users should have the programs of the machine translation containing respective c.b.d(s).
The examples of universal semantic coding and translation of the phrases possessing the ambiguities of various types and the corresponding fragments of the coordinated dictionary entries are presented in the Appendices to the given article. The some examples contain, for comparison, the translations made by means of traditional programs of machine translation.
The fragments of the coordinated dictionary entries are made with using of explanatory dictionaries [5,6,7,11 **, 14] and of bilingual dictionaries [8,9,10,11,11***,13,15]; the polysemous word-combinations and their interpretations are borrowed in ; the elements connected with the differentiated (grammatical) meanings of pronouns and verbs are made also with use [11*, 12]; the elements connected with the differentiated (syntactical) meanings are made with use [12*].
All fragments of the coordinated dictionary entries have a conditional character.
further see: Appendices (examples of coding and translation, fragments of coordinated dictionary entries)
References are represented in the first part of THE EXPANDED SYSTEM …
Published - July 2008
See also: "Experience of preliminary identification and translation of the text's fragments with use of the system of coordinated dictionaries" by Panich Iuli