If the dictionary is intended to help to generate German texts, the lexical meanings of the German equivalent will have to be specified, for example in the following way:

ch 'u-chia, hairaten ('to get married'),

sich verheiraten ('to get married')

A combination of these intentions would lead to an entry of the type

ch 'u-chia (von Frauen), heiraten ('to take in marriage'),

sich verheiraten ('to get married').

It is probably not necessary to describe the different possible entries of a German-Chinese dictionary.

The German equivalent of Cninese hsien-hsieh can be regarded beinahe, fast 'almost, nearly'. An examination of the Chinese contexts (which we shall quote here in approximative English translations only) shows that all of them are roughly of the type: 'he nearly stumbled, fell, starved, died, knocked down someone, poisoned someone, crushed someone'. We shall not repeat the different forms the entry could have, according to the lexicographer's intentions, beginning with the simplest form:

[Chinese] hsien-hsieh [German] beinahejast.

Let us, however, suppose that the lexicographer intends to describe in his dictionary the source language (Chinese), at least to some extent. He can then, add a restrictive, specifying gloss of the type, say

hsien-hsieh (bei negativen Ereignissen) beinahe, fast (about negative events) almost, nearly.

It will not always be easy to construct such glosses, especially if the bilingual lexicographer does not have the support of a good monolingual dictionary; therefore, he may prefer to add one or more Chinese contexts as examples. If he chooses them well, he can convey the necessary information concerning the restriction without accepting the commitment that the restriction is absolutely valid. Or, in a bigger dictionary, he can give both the gloss and the contextual examples. Let us now consider the equivalents; for the beginning let us discuss the English ones (which could serve as equivalents of the Chinese lexical unit, though the example is based on the German material), Both the English lexical units have multiple meaning. If we accept Hornby's description of it, we see that almost has two senses: (I) as in He almost fell; {almost is replaceable by nearly); (2) as in Almost no one believed her; (almost is not replaceable by nearly). On the other hand nearly has three senses: (1) as in It is nearly 1 о 'dock; (replaceable by almost); (2) as in / have £20, but that will not be nearly enough for my journey (not replaceable); (3) as in nearly related (hot replaceable).

If we quote almost, nearly together as equivalents of the Chinese lexical unit, they disambiguate each other, because it is obvious to the user of the dictionary that only that sense applies which is common to both of them.

If we consider the German equivalents beinahe, fast, we see that they are as close synonyms as one can find, the difference in their meaning being almost imperceptible. The question arises, then, why both of them should be quoted, what purpose does it serve. There are two reasons: First, the indication of synonyms in the target language helps the user to find variant possibilities of expression, if only for purely stylistic variation. Second, imperceptible and hard to state as they are, there usually are some slight differences between such synonyms so that if more are indicated, the information conveyed is richer. On the other hand, if the dictionary is to be only a small one, synonyms of this type can be omitted; and they should not be indicated too lavishly, even in big dictionaries.

We see that we can discern three types of indications of synonymous equivalents which are not differentiated formally in the absolute majority of dictionaries:

1. heiraten, sich verheiraten: a rule (semantic or grammatical) of the target language makes it predictable which of the two equivalents will be used;

2. almost, nearly: both can be used, but only in the senses of their multiple

meanings which overlap, which are disambiguated;

(3) bemahe,jasV. any of the two can be used, and the two taken together make the information somehow richer.

These are only the polarized types; in the concrete situation there is much overlapping and uncertainty over which category a case belongs to: it is above all types (2) and (3) which are frequently hard to discern. The lexicographer should, however, try to know to which type a single case belongs: if the dictionary must be short, he can reduce type (3) by the omission of one of the synonyms, without a great loss; type (2) can be thus reduced only if the meaning of the remaining equivalent is disambiguated by another means, i,e. if it is clear, e.g. from an example quoted, in which sense it is to be taken; the reduction of type (1) usually causes considerable mistakes or difficulties on the part of the user of the dictionary. Synonyms both of type (2) and (3) are usually joined by the comma. In type (1), either comma or semicolon is used; in bigger dictionaries, the latter is preferable.

The German equivalent of Chinese jiu can be regarded as alt 'old⁷. When the lexicographer collects and analyzes the contexts of (he source language (we use again an example of the more difficult situation, where the lexicographer does not have the support of a standard-descriptive monolingual dictionary of the source language), he will perceive that they belong roughly to the following three groups:

(1) old edition of a book, and old malady makes its appearance again, old society, old ideology, old dwelling, old job;