4. Selection of entries

The selection of the prospective entry-words which will be included in the bilingual dictionary should be governed by the type of the dictionary, above all by its intention and purpose. The same applies to the reduction of the multiple meanings of selected entries, hi the case of a dictionary which intends to help the user to understand texts couched in a foreign language, it will be clear that the occurrence of the lexical units in those texts is the first factor which determines the selection for the entry-words; the more 'text-bound' the dictionary is, the more powerful is also this factor. In a similar way, the selection for a bilingual dictionary with descriptive intentions is governed by principles almost identical to those of the corresponding monolingual dictionary.

Some remarks must be made with respect to the bilingual dictionary which is intended to help the user to generate texts in the (foreign) target language. The basis of the list of prospective entry-words is, of course, the lexicon and semantics of the (native) source language. There are, however, several modifications of this general principle.

First, if the planned dictionary' is not to be a big one, it is possible to leave out the less known or less used synonyms of the source language if the better known and more used are included; e.g. if a smaller English-X dictionary has die entry to mil around, it is possible to leave out the entry to circumnavigate.

Second, such a dictionary, especially if it is a smaller one, should be rather reserved in its inclusion of colloquialisms, slang expressions or even vulgarities and similar levels of language; and even the bigger dictionaries of this type should be extremely cautious in this respect, lest the user be put into a ridiculous or painful position. In a smaller dictionary of this type, it is better to omit such 'bad pennies of everyday conversation' or, in a bigger one, to label them very cautiously in order to warn the user.

Third, if the target language of a dictionary of this type is spoken in a society with a different culture and in a geographical and other extralinguistic milieu vastly

72 different from that of the source language, it will be necessary to take into consideration also the target language when the entry words of the source language are selected. Different social institutions, different plants and animals may be unimportant or non-existent in the milieu of the source language while being very important or frequent in the milieu of the target language. Because it may be legitimately assumed that the source-language speaker will use the dictionary to generate texts about the milieu of the target language (possibly also while residing in the other surroundings), the respective lexical units of the target language should be indicated. <...>

On the contrary, if the lexicographer compiles a very small dictionary of two very closely related languages, he may tend to omit those lexical units which have in both languages an identical or only negligibly and predictably different form and an identical meaning. If pairs like

Czech herec Slovak herec 'actor' hereCka herecka 'actress'

herecky herecky 'actor's; histrionic'

have the same form and meaning, a lexicographer who must save space very strictly may omit at least some of them. If this is done on purpose and by principle in all such cases the result is the so-called differential dictionary in the two languages. When compiling such a differential dictionary, the lexicographer should be extremely cautious, because even if the multiple meaning of both the lexical units coincides, there may be important differences in pltraseology. One of the dangers of the differential dictionary is that since it is necessarily small, the lexicographer will take into consideration only the dominant senses of the two lexical units: these may-happen to be identical and so the pair may be omitted, though there may be important differences in the non-dominant senses about which the user will learn nothing.

But even if it is not a differential dictionary which he is compiling, the lexicographer will tend to omit, in a very small dictionary, such internationalism as telephone, mathematics, for obvious reasons. Extreme care should be exercised, because there are not a few 'false friends' among them; i.e. lexical units with very similar form but different meaning, e.g. German Phonologie 'phonemics', English phonology ~ German 'Lautlehre'. These 'false friends' can be found not only in teclmical terminology, but generally: Cf. Russian maika 'uterus, womb', Czech matka 'mother'; Russian cerstvyj 'stale', Czech cerstvy 'fresh'. Sharp differences like those observable in these examples are not particularly dangerous, because they will certainly be noticed. Terminological intricacies (English phonology: German Phonologie) and not too dissimilar senses (English to realize, French realiser), are more dangerous because they are not always easy to perceive.

Another situation of peculiar interest is the selection of entry-words for a bilingual dictionary the target language of which is not yet fully stabilized and lacks equivalents of a considerable number of lexical items of the source language. Here again, the selection of the source-language entry-words must be done with a greater than usual respect to the target language; it is usually necessary to compile at first only a shorter dictionary, rather reduced as regards both the selected entry-words and their multiple meaning, lest the dictionary be surcharged with neologisms of all sorts as well as unslabilized attempts at filling up the onomasiologtcal gaps.

Generally speaking, it is necessary to take into consideration that it is a language more or less foreign to the user with which we have to deal in a bilingual dictionary. This is important above all when it is the target language in which the foreign-speaking user is supposed to generate texts. But the same circumstance should be taken into account when it is the source language which is foreign to the user: the 'irregular forms', the multiword lexical groups and set groups generally, all these complications should be presented in as simple and as unsophisticated a way as is consistent with material correctness.

5. THE EQUIVALENT

The bilingual lexicographer's most important duty is to find in the target language such lexical units as are equivalent to the lexical units of the source language, and to coordinate the two sets. We call equivalent such a lexical unit of the target language which has the same lexical meaning as the respective lexical unit of the source language.

We know how complex the lexical meaning is. Absolute equivalence requires, then, that the lexical meaning of the two lexical units be absolutely identical, in all components (designation, connotation, range of application). Because of the anisomorphism of languages, such absolute equivalents are rather unfrequent (outside the domain of scientific terminologies).

The usual situation is that the lexical meaning of the respective lexical unit of the target languages is only partly identical with that of its counterpart in the source language. If it is necessary to be very precise, we speak in such a case of partial identity about partial equivalents. Since, however, the absolute and overwhelming majority of equivalents (irrespective of which pair of languages we observe) belongs to the category of the partial ones, it would be cumbersome to repeat the adjective endlessly. Therefore, if we speak about 'the equivalents of a lexical unit' or about "the equivalent of a lexical unit in one of its senses', etc., it is the partial equivalents that we mean.

The logical outset of the search for equivalents is a comparative analysis of the

structures of the two languages. The lexicographer must ascertain what categories of

lexical imits (i.e., traditionally, what parts of speech) are present in both the

languages, and must decide which pairs of categories will be considered equivalent.

This is relatively easy if there are observably similar or identical categories, in the

two languages: it will be easy to decide that a French noun will be considered

equivalent to an English one, that a Russian verb will be considered equivalent to a

German one, etc. The lexical equivalents will, then, be preferably chosen in such a

75

way that they belong to the categories considered equivalent. But this is a mere preference, since if the equivalence of categories is established correctly the pairs of lexical units will, after all, come out automatically; if there is a clash between this principle and a concrete case of real lexical equivalence of lexical units which belong to different classes, the correct equivalence of lexical units which belong to different classes, the correct equivalence must take precedence over the principle. For example, the German category of nouns will be considered equivalent with the English noun. But the English equivalent of the German noun Handarbeit (in its application as a label on wares) will be hand-made, which is an adjective, because the noun hand-work is used only in reference to the process of the work itself, not in reference to its result.

It is very important to observe the points of trouble at the very beginning and to decide upon a unified treatment of them. There is a good amount of equivalence between the categones of the Czech and the English lexical units; one point of trouble is, however, the adjective which can be derived in Czech from practically any noun by a limited number of suffixes, like -sky. -ovy, which is not always the case in English. <...>

The crucial thing, however, is that all cases of one and the same type should be treated in the same way, throughout the dictionary, unless there are, in isolated cases, specific reasons to treat them otherwise.

It would be easy to find a great number of examples in different languages. Sometimes i! is possible to solve the situation in a restrictive way: for example, if there are two regular grammatical categories one of which can be conceived as derived from the other in the source language, and both of which correspond to only-one category in the target language, the single pairs of members of these categories will generally have identical equivalents in the target language; in this case, the second of the two categories can be indicated only by reference to the first category, unless the concrete lexical unit or its equivalent has some semantic properties of its own,

If there is no 'categorial equivalence' of the 'parts of speech' observable, the lexicographer has to proceed on the basis of the equivalence of the lexical meaning only. In any case, the lexicographer should, at the beginning of the work, prepare as complete as possible instructions on how to treat the single cases of categorial identity or diversity of the single morphological categories of both the languages.

It follows from the first sentence of this chapter that the bilingual dictionary should coordinate the lexical units of the two languages: therefore, the equivalent should be a real lexical unit of the target language which occurs in real sentences. We shall see later that this requirement has its limitations, but has general validity within the limits of the possible.

In order to be sure that he indicates real lexical units of the target language, the lexicographer collects contexts which illustrate the whole multiple meaning of the respective lexical unit (i.e., all its senses and many of its applications, or all of them) in the source language of finesse and detail planned for the future dictionary (the bigger the dictionary, the more numerous the contexts, with subtler semantic differences); he translates these contexts into the target language and observes whether the prospective shortest possible equivalent can be used in all the translations (producing a sentence of the target language that absolutely conforms to its rules) or only in some of them, or in none. If the prospective equivalent can be used in all the contexts and only those contexts, it is an absolute equivalent. Wit can be inserted only into a group of the translated contexts, it is only partial, and another equivalent (or more of them) must be sought for the rest of the translated contexts. If it can be inserted into more contexts, its meaning is broader, usually either because the designatum has fewer criterial features, or because there are fewer applicational restrictions, If it cannot be inserted into any context, it is no equivalent at all. Tins method of testing (the translation of) all the contexts (in the desired degree of detail) has not only the advantage that it leads to lexical units of the target language; it also helps to cope with the anisomorphism, i.e., in this case, with the differences of the lexical meaning of the two lexical units. It must be remembered that it is not possible to take the (eventually multiple) meaning of a lexical unit as described in a monolingual dictionary of the source language and seek the prospective senses in the description of the (eventually multiple) meaning of the prospective equivalent in a monolingual dictionary of the target language. Such a procedure can give a first orientation, but the real task, if it is to be done with finesse, is more complicated: the comparison of the two lexical units by the contextual method very frequently shows differences between them which consist in phenomena not always stated in monolingual dictionaries. And it must also be remembered that the way the lexicographer handles the equivalence and how he presents it varies largely with the different intentions and other properties of the planned dictionary. Let us discuss some examples.

German heiraten, sich verheiraten 'to marry" can be considered equivalents of Chinese ch'u-chia. There is, however, one difference: whereas the German lexical units are used irrespective of the sex of the person, the Chinese lexical unit is used in reference to women only. In a Chinese-German dictionary intended strictly for the comprehension of Chinese texts by native speakers of German, the indication of the equivalence [Chinese] ch 'u-chia: [German] heiraten, sich verheiraten would be sufficient: the range of application of the German lexical units is greater than that of the Chinese one and either of them is always applicable, in the translation of the Chinese contexts, so that there is no need to add a gloss; the Chinese restriction need not be stated as it will be inherent in any context met by the user; and it is not necessary to inform the German user about the semantic difference between the two German lexical units, {heiraten: English 'to take in marriage', sich verheiraten: English 'to get married').

From what we discussed above it will be clear that a bilingual dictionary of another type will need some further indications. If the lexical meaning of the Chinese lexical unit is to be described more precisely, the entry will have to contain some indication of the restriction, as

ch 'u-chia, (von Frauen) heiraten, sich verheiraten.