Equivalence in Bilingual Lexicography : Criticism and Suggestions

A reminder of general problems in the formation of terminology, as illustrated by the German Äquivalence (Eng. equivalence) and äquivalent (Eng. equivalent), is followed by a critical discussion of the concept of equivalence in contrastive lexicology. It is shown that especially the concept of partial equivalence is contradictory in its different manifestations. Consequently attempts are made to give a more precise indication of the concept of equivalence in the metalexicography, with regard to the domain of the nominal lexicon. The problems of especially the metalexicographic concept of partial equivalence as well as that of divergence are fundamentally expounded. In conclusion the direction is indicated to find more appropriate metalexicographic versions of the concept of equivalence.


1.
Aspects of the problem of equivalence: a sketch The problem of equivalence lies in a fact about which there exists interdisciplinary consensus: the lexical-semantic structures of the lexicon of a particular language are language-specific and therefore partly unique.This implies that the lexical-semantic structures of two (or more) languages are not isomorphous.The non-isomorphism (Zgusta 1971: 294) within the area of the lexicon (which of course also applies to the area of grammar) forms the pre-theoretical, observable empirical circumstances, the study of which led to discipline-specific manifestations of the problem of equivalence.Accordingly, the problem in question here has a long scientific history.Knowledge of this history and its most important basic outlines is a useful presupposition by which one can structure the problem of equivalence.My -by all means limited because of annoying language barriers -knowledge of this history of the problem enables me to distinguish the following dimensions: (1) a dimension of the problem, in which general problems of scientific conceptualization are embedded; (2) a contrastive lexicological dimension of the problem; and (3) a genuine metalexicographical dimension of the problem.
The last-mentioned has two sub-dimensions, namely: (3)(a) one which relates to the subject of the dictionary (in the sense of Wiegand 1998: 302), and (3)(b) one which relates to the dictionary form.
In what follows, I shall briefly give attention to the specific dimensions.According to my theme, the focus will be on the metalexicographical aspect.I have to leave out the sections on dictionary form touched on by the theme, hence specifically the questions in connection with the particular dictionary type and therefore also the presentation of equivalent items and items indicating differences in equivalence determined by the functions of the dictionary, as well as the question of addressing (cf. in this regard, for example, Meyer and Wiegand 2000, and Wiegand 2000, 2001, 2002).

General problems of concept building: a reminder
Problems of concept building and conceptual vagueness can have different causes.A typical constellation that can easily lead to inaccurate ways of thinking and argumentation exists when language expressions -in our case the nouns equivalence and equivalent as well as the adjective equivalent -are used in both general language and in various scientific disciplines, and in addition to this, also in practical professional fields belonging to the scientific disciplines.
Both the nouns equivalence and equivalent as well as the adjective equivalent are used -apart from their use in general language -in, amongst others, the following disciplines: in the theory of science, logic, mathematics, physics, jurisprudence, ethics, translation theory, in dictionary research as well as in the practical field of lexicography.Thus one easily finds illustrations for equivalent in bilingual dictionaries of which the circle of addressees consists of educated potential users.For example, in the user instructions of the Wörterbuch Deutsch-Sanskrit (Mylius 1988: 10), there is a paragraph headed "Das Sanskrit-Äquivalent" (The Sanskrit Equivalent).The use of the particular words in question is not at all uniform in the mentioned fields.Especially diverse and relatively unclear is the use of equivalent and translation equivalent in translation theory (cf., for example, Kade 1973, 1975, Wilss 1977: 156 ff. and Koller 2001: 159 ff.).
Although the use of the three expressions in question is semantically really different, there are however some common traits.These can be found in the origin of the generally used loan word equivalent.Middle Latin aequivalentia belongs to the Latin aequus meaning "same" and to the Latin valere meaning "having value".An appropriate general correspondent expression for equivalent would therefore be "having the same value" (cf., for example, GIWDS 1995, s.v. Äquivalenz).Accordingly, the adjective equivalent would mean, in general usage, the same as "having the same value", and equivalents are in accordance various items which have the same value with regard to at least one other item.
"Having the same value" must clearly be distinguished from identicalness.Two items are -roughly speaking -identical when they correspond with regard to all their distinctive features.They have, on the other hand, the same value when they have the same purpose on the basis of at least one identical feature in the context of a thought or action.Identical items have the same value with regard to a similar context of thought or action; the opposite is, however, not applicable.In my view, almost all subject-specific concepts of equivalence have something to do with "having the same value" in this general sense.The given similarity of concept building in the various subject fields does however not automatically lead to an appropriate understanding of this particular theory-specific concept of equivalence in question.One can rather only reach, on the basis of its use in general language, a more or less vague preconception.If somebody, for example, states: an expression in language A is equivalent to an expression in language B when they have the same meaning, then such a proposal is, even though not untrue and quite understandable from the viewpoint of the general language use of equivalent, so hopelessly vague from a linguistic point of view that one can hardly do anything with it in the scientific sense.A concept of equivalence specifically for dictionary research should not be built counter-intuitively from its use in general language, but should be conceived more precisely and should also be differentiated from the concepts of equivalence from neighbouring disciplines, especially those of contrastive linguistics and translation theory.

A brief look at contrastive lexicology
The terms in question only have a marginal role in lexicology, when lexicological investigations are only concerned with one language.In this way, one will, for example, refer to lexical synonyms within the designative lexicon (such as Orange and Apfelsine) which are extensionally equivalent, meaning that they have exactly an equal number of denotations.In addition, one will also debate whether the relationship of lexical synonyms can be valid as equivalent relationships in the area of scientific theory (cf., for example, Fischer 1973).
The concepts of equivalence, on the other hand, have a crucial role in contrastive or confrontational lexicology.There are also different lexicological manifestations of the problem of equivalence.Concepts of equivalence in contrastive lexicology also had an effect on metalexicographical publications (cf., for example, Karl 1982).In what follows, I can only give a limited critical perspective on a few variants of the concept of equivalence.
Contrastive lexicology is seen as a partial discipline focusing on langue.Its concepts of equivalence accordingly focus on the language system, but are mostly relatively vague.There are several reasons for this.The designative lexicon has as its basis a polysemous concept of language signs.Noun items in the lexicon can therefore be n times polysemous (with n ≥ 2).When contrasting a source language noun item with a corresponding target language item, the denotative relationship is usually taken as basis for the comparison.Accordingly, equivalence, which is then usually called semantic equivalence (cf., for example, Karl 1982: 34), is present in a polysemous noun item precisely when, firstly, the number of sememes in the source language are equal to those in the target language and, for example, have the value m, and when, secondly, their denotation correspond in m pairs of sememes with regard to a source and target language sememe in each pair.
In Figure 1, this type of equivalence is illustrated.Figure 1: System-related semantic equivalence in threefold polysemous lexical items (m = 3).Abbreviations: SL = source language; TL = target language; Se = sememe; D = denoted item; Notational convention: " " means is a sememe in; " " means is semantically equivalent to; "x y" means x has as denoted item y.
The above-mentioned concept of equivalence (using the structuralist terms of its advocates) has various weak points.I can only briefly go into two of them.Firstly, the relationship of semantic equivalence is defined with sets of which the elements are not semantic units at all, but bilateral linguistic signs; the second weak point exists because the point of departure is a concept of equivalence based on semantic units, namely, sememes.Sememes have to "correspond" with reference to the denotation; this means, however: they have to be denotatively equivalent.In contrast to semantic equivalence, one therefore also refers here to sememic equivalence (cf., for example, Karl 1982: 35).Besides, it is not always clear at all whether, for sememic equivalence to be present, one only needs an equal structure of denotative semantic markers, or whether this also applies to non-denotative semantic markers.
Already on a systemic level, one consequently works here with two different, but related, concepts of equivalence, which must necessarily lead to complications.Accordingly, the concept partial equivalence already causes considerable confusion in many publications, a fact that I cannot demonstrate here with citations.For, in the first instance, one talks about partial equivalence when the polysemy structure of a source language lexical item does not correspond with that of the target language on the systemic level (cf., for example , Sternemann 1983: 43f).This case is illustrated in Figure 2 as a first case example.A second case example (1b) is illustrated in Figure 3. http://lexikos.journals.ac.za In the second place, one uses the expression partial equivalence when there is either a hyperonym void or a hyponym void in one of the contrasting partner languages.For example, in Russian, there is no word which corresponds to the German word Kirschbaum.In Russian, only matches for the German Süßkirschenbaum and Sauerkirschenbaum have been lexicalised.Kirschbaum is then a partial equivalent to both the Russian čerešnja and višnja.The reason for this type of partial equivalence on the systemic level is based incorrectly on factors that have to do with parole, because one argues that a hyperonym expression could refer to the same referential object as one of the hyponym expressions in the text (cf., for example, Sternemann 1983: 44).The second type of partial equivalence is illustrated in Figure 4.In the third place, one uses the expression partial equivalence when the set of classes of items referring to a collective noun in the source language (for example, German Vieh) either do not totally correspond with a collective noun in the target language, or, if there is correspondence of at least one class, it only occurs in one set in both languages (cf., for example, Vietze 1981: 78f).In German, for example, pigs also belong to Vieh, but not in Mongolian.On the other hand, in Mongolian, camels, amongst others, also belong to the class denoted by Vieh in German.An illustration for case 3 is seen in Figure 5.It is not possible to remain on the topic of contrastive lexicology here any longer.It is clear, however, that concepts not treated here, such as mono-equivalence, poly-equivalence, non-equivalence (cf., for example, Karl 1982: 34ff.), approximate, facultative, communicative and total equivalence (cf., for example, Kade 1968Kade : 1973;;Wotjak 1982: 113ff.)as well as numerous other postulated types of equivalence (cf., for example, Wilss 1977: 156ff; Scholze-Stubenrecht 1995: passim), http://lexikos.journals.ac.za cannot be conceived sufficiently exact if they are conceptually related to concepts of equivalence such as the ones criticized above.The same applies for terms such as, amongst others, congruence, divergence, poly-divergence and convergence (cf., for example, Vietze 1981: 78f;Hausmann 1977: 54f.;Rettig 1985: 98ff;Gouws 1996) as well as for terms for the most diverse types of equivalence, such as quasi-equivalent, assisting equivalent (Karl 1982: 40f) and numerous others (cf.altogether, for example, also Koller 2001: 159ff).

Describing concepts of equivalence in dictionary research more precisely
In the past three decades, many important insights were gained on equivalence in bilingual lexicography in numerous publications.I mention only a few of these publications: Atkins 1996; Baunebjerg Hansen 1990;Duval 1991;Gouws 1996Gouws , 2000Gouws , 2002;;Hartmann 1994;Hausmann 1977Hausmann , 1988Hausmann , 1995Hausmann , 1997;;Jarošová 2000;Meyer and Wiegand 2000;Petkov 2001;Scholze-Stubenrecht 1995;Sinclair 1996;Tognini-Bonelli 1996;Werner 1999;Wiegand 2000;Zgusta 1971Zgusta , 1984.The insights are concerned, amongst others, with the following aspects, into which I cannot go into detail: evaluation of equivalents, -the microstructural presentation of equivalent items, -addressing equivalent items, -discriminating between several presented equivalents by means of "items discriminating between equivalents", -distinguishing between types of equivalents and types of equivalence, and distinguishing between various functions of equivalent items depending on the dictionary type and dictionary function.
However, it is remarkable that the concept of equivalence -in contrast to contrastive lexicology and translation theory -has hardly been discussed in metalexicographical publications.In my view, there exist, in the meantime, grave differences of opinion which have led to a whole range of misjudgments about the features of equivalent relationships in bilingual lexicography.In what follows, I will try to verify these critical statements.Accordingly, I will now concentrate on the concept of equivalence, but can only single out some aspects.It should in the first instance be immediately clear: the langue-related concept of equivalence of contrastive lexicology is inappropriate for bilingual lexicography, because bilingual dictionaries are not conceptualised as aids for contrastive studies of language systems (even though some advocates of contrastive lexicology use them in this way).They are rather meant, in the first place, as a means to understand and produce foreign texts and to make translations in both directions.Because the systemic level can also play a part in dictionary research and lexicography, I suggest that one speaks, with reference to Koller (2001: 216ff), of correspondence instead of equivalence when one deals with lan-guage systems.It would of course also be necessary precisely to determine the concept of correspondence.In the theory of bilingual dictionaries, equivalence would then be a term reserved for parole phenomena.Another possibility would be permanently to distinguish between systemic equivalence and parole equivalence.The parole-relatedness brings a metalexicographical concept of equivalence close to most of the concepts in translation theory, so that here one has to draw a clear distinction.The concepts of equivalence in translation theory thoroughly differ.They have, however, the following in common: they refer to whole texts and their translations.Without taking proverbs into account, this is not at all the case in lexicography.Here one deals with the equivalence of meaning-bearing units below the level of sentences.It therefore also deals with the equivalence of word formation devices, words, free syntagmas, and with equivalence of various items that consist of several words which do not form sentences, especially idiomatic expressions and collocations.One should already point out here that a metalexicographical concept of equivalence should not be reduced to lexical items.Otherwise, several cases cannot be taken into account, for example, when a source language item is equivalent to a non-lexicalised target language item.Thus the French espacer is equivalent to the German Zwischenraum lassen zwischen.The German equivalent is, however, not lexicalised (cf. in this regard Rettig 1995: 93ff versus Hausmann 1977: 53ff).
Up to now, it has only very roughly been outlined which language items from the dictionary contents can feature in the pre-and post-domain of a bilingual equivalence relationship.The following should of course be very clear: independent of which concept of meaning is used in detail to interpret language expressions below the level of sentences, each polysemous item can only belong to an equivalence relationship, based on parole conditions, with one of its meanings.
In what follows, I limit my consideration to noun items.Bilingual equivalence is a relationship between a source and target language item which is present when the threefold predicate x is equivalent to y with regard to z is true; "z" is the variable for the criterion of equivalence.All features of noun items can in principle function as equivalence criteria.For instance, the following statement is true: (1) German Scheiße is equivalent to British arse with reference to the pragmatic label "vulgar".
Even though both the expressions Scheiße and arse are obviously in a relationship of equivalence, they are just as obviously not equivalents in a lexicographical context.This is because the essential requirement for lexicographical equivalents to be present in the area of nouns consists of the source and target language item denoting the same object in usual texts (in the sense of Wiegand 1996a).This means: they have to be referentially-semantically equivalent, in short: semantically equivalent.To form the concept lexicographically equivalent, one therefore has to evaluate the possible equivalence relationships: the most important one being semantic equivalence.One can accordingly only speak of lex-http://lexikos.journals.ac.za icographical equivalence when semantic equivalence occurs.Two items which are semantically equivalent can also be in other equivalence relationships. (2) The German Arsch and British arse are semantically equivalent.
However, it is also true that (3) Arsch and arse are equivalent with reference to the label "vulgar".
Both expressions therefore have two equivalence relationships.If one speaks of an equivalence relationship of which the equivalence criterion is a dimension of pragmatic labels, of pragmatic equivalence, then Arsch and arse are semantically as well as pragmatically equivalent, in short: semantically-pragmatically equivalent.If one has at one's disposal a system of labelling, then one can distinguish several pragmatic equivalence relationships.If one dimension of labelling, for example, is "style", one can state: (4) Arsch and arse are semantically-stylistically (or semantically and stylistically) equivalent.
Two expressions which are semantically equivalent, can also be pragmatically non-equivalent.This is the case when one of the expressions is pragmatically labelled and the other not.In this case, one can refer to quasi-equivalents, with reference to Karl (1982: 40f).One can then state: a pair of quasi-equivalents is to a lesser extent equivalent than a pair of semantic-pragmatic equivalents.In bilingual lexicography, one therefore strives for semantic-pragmatic equivalence.
To introduce some necessary further distinctions, we can subsequently look at a dictionary article from Neubert and Gröger 1991 3 .Only the first part of the article in which the noun has been treated lexicographically is of interest here.
In dictionary article da1, the German equivalent item "Bett" is addressed to the English item giving the form of the lemma sign bed.Both expressions are pragmatically non-labelled.This is also a type of pragmatic equivalence.The lexicographical statement therefore is: the standard language noun bed used in its usual sense is semantically-pragmatically equivalent with the standard language noun Bett used in its usual sense.This means: if one forms a usual text with bed in English (that is, a text which is in accordance with habitual contexts of designation for bed), the German word Bett can be used in the German translation.The English bed could then -as we can furthermore conclude from dictionary article da1 -occur in the following standard language co-texts (≡ stands for is semantically-pragmatically equivalent with): bed and breakfast ≡ Übernachtung mit Frühstück single bed ≡ Übernachtung für eine Person to be brought to bed of ≡ niederkommen mit bed of roses ≡ leichtes od.unbeschwertes Leben bed of thorns ≡ Schmerzenslager In addition to these standard language co-texts mentioned in the dictionary, there are others which are domain-specific and which are referred to in the dictionary by means of other items.One either presents specific classes of referential objects, for example "(Tier) Lager" (this means it is Lager in German when the English bed refers to the "bed" of animals), or one presents specific subject fields, for example "Bergb Flöz".
The article da1 is consequently constructed in such a way that the general case of semantic-pragmatic equivalence is dealt with in the first instance, then followed by the specific cases.
Let us now look at the co-text item "bed and breakfast" from da1.Of course, this item was presented because the word equivalence of bed and Bett does not occur.Rather, an equivalence of syntagmas is present.The English syntagma bed and breakfast and the German syntagma Übernachtung mit Frühstück, in which the word Bett obviously does not occur, are pragmaticallysemantically equivalent.
In both cases discussed above, the elements in the pre-and post-domain of the equivalence relationship are on the same hierarchical level.They can, however, also be on different hierarchical levels.For example, the German compound Schwarzmarkt is semantically-pragmatically equivalent to the English syntagma black market.This is a case of word-syntagma equivalence.And the French marché noir is also semantically-pragmatically equivalent to the German Schwarzmarkt; here also, syntagma-word equivalence is present.
When no equivalent is given in the target language, one refers to nonequivalence.One may ask here: exactly when does non-equivalence occur?When considering this question, one is faced with the fact that each language item of a particular language always has a paraphrase in this language which can be more detailed or less detailed.If there is no semantic equivalent for an item in another language, one can translate its paraphrase in the language lacking an equivalent.Accordingly, another question arises: is this translation then an equivalent?Let us look at an example.For the German Amtsgericht, there is no equivalent in French, neither on the word level nor on the syntagmatic level.If one wants adequately to explain to a French person what Amtsgericht means, one has to form a longer French text, which one can also consider a translation of a corresponding German text.In my view, it makes no sense to let such a text pass as equivalent in dictionary research.One should rather state the following definition criterion for equivalents: Only items below the sentence level which can be used in target language sentences can feature as equivalents of the target language in lexicography, to which can be addedas already mentioned -that, for equivalent syntagmas in the target language, no stipulations regarding lexicalisation should be formulated, so that equivalence can also be free from the limitations imposed by the lexicon.In my opinion, one should either give up the distinction between so-called "translational equivalents" and "explanatory equivalents" or, on the other hand, define it more precisely, so that one can only refer to "explanatory equivalents" when target language syntagmatic equivalents in sentences are applicable.Furthermore, it should be clear that non-lexicalised equivalents should only be presented when no lexicalised ones can be found.Non-equivalence is therefore present when no word or syntagma which is at least semantically equivalent can be found in the target language.Non-equivalence should explicitly be marked in the dictionary article (for example, by means of "0") (cf., for example, Wiegand 1996: 228ff in this regard).
At present, since Hausmann (1977: 54f), one refers to divergence and convergence in dictionary research.Examples of divergence can be found in Figure 7.  mann (1977), Koller (2001: 228ff) and Kromann et al. (1991).In Hausmann (1977: 54), divergence and convergence belong to the "basic types of equivalence relationships".One can interpret this in such a way that, for example, in the case of divergence, one can speak of several equivalence relationships.Hausmann (1977: 55) formulates this as follows: "Two or more equivalents in the target language corresponds to a source language word: bedauern = regretter, plaindre."Divergence and convergence are not to be understood extensionally in Hausmann.It is not concerned with relations representing sets of ordered pairs.There are, consequently, two equivalence relationships, namely: bedauern is equivalent to regretter and bedauern is equivalent to plaindre.In metalexicographical context, only one type of equivalence relationship is now formed out of divergence and convergence in each case, that is, out of phenomena both consisting of several equivalence relationships (cf., for example, Gouws 1996: 17 andGouws 2000)!When divergence, for example, is present, then consequently one relation (of a specific type) should exist.This is only the case when one equivalent relation is understood to be a set (cf. Wiegand 2002).When, as is the case with Hausmann, one understands divergence and convergence intensionally, this is conceptually not correct.In the mentioned intensional sense, divergence and convergence should not be interpreted as one relationship, even though one would have liked to establish an independent concept of relationships for metalexicography which, in my view, would be totally unsuitable.While the technical analysis in Gouws, which is performed in the context of the concepts divergence and convergence, is completely acceptable, another line of reception is muddled.Kromann et al. (1991Kromann et al. ( : 2718) ) writes under the heading of the second version "Partial equivalents": "[…] there is divergence when a lemma, contrasted with the lexical units of the target language, must be divided into several 'sub-meanings' […]".This means: when, for example, somebody contrasts the German bedauern with the French regretter and French plaindre, then bedauern suddenly receives (as if from a higher authority) two "sub-mean-ings".It should no doubt be clear that one cannot argue like this.
Already in Hausmann, it is not clear (to me) whether the sentence quoted above refers to langue or parole.This remarkable carelessness when equivalence is dealt with, can be found in a large portion of the metalexicographical literature.Since this carelessness with regard to the consideration of an important distinction does not have an effect on congruence (cf. Figure 9), because here the reference to the langue and parole levels are analogous, only congruence is, in my view, a sufficiently clear concept.

German Kalenderjahr
French année civile English two

German zwei German die Schweiz
French la Suisse A solution for the above-mentioned conceptual discrepancy may be sought in the direction as indicated below.
Semantic and pragmatic equivalence are, for example, types of equivalence.Following Scholze-Stubenrecht (1995), one can distinguish additional types of equivalence relationships.Convergence and divergence are then, for example, not types of equivalence relationships.Here one is rather dealing with entities on the systemic level which could, for example, be called a correspondence network.Koller (2001: 228ff) calls them correspondence types (Entsprechungstypen). Convergence corresponds with the correspondence type "manyto-one-correspondence"; divergence corresponds to "one-to-many-correspondence".The threefold relationship statement (x corresponds with y with reference to z with "x" and "y" as variables for noun items of the source and target language and "z" as variable for the correspondence criterion) belongs to the correspondence relationship.This is given in the relationship of denotation: "x corresponds with y" should be read as "x is the systemic correspondence for y".
If there are n systemic correspondences (with n ≥ 2) in a target language for a lexical item in the source language, so that the correspondence network of divergence is present, the source language item can only be semantically equivalent with the n target language items if the n target language items are totally synonymous on the lexical-semantic level.If such synonymy is not present in the target language items, then co-text independent semantic equivalence is impossible.On the contrary, every corresponding target language item determines by means of a specific set of features (which can, in comparison to the source language, be considered as a semantic restriction of use) a source language co-text class as condition for equivalence for the parole, to which those usual texts belong in which the target language item, when used in a usual sense, is semantically equivalent to the source language item, when also used in a usual sense.For example, the Swedish morfar determines, as condition for equivalence, the class of all co-texts with the German Großvater in which Großvater refers to a grandfather on the mother's side.If one finds in a dictionary article of a German-Swedish dictionary the entry (5) Großvater … (mütterlicherseits) morfar; (väterlicherseits) farfar …, then the meaning of the lemma sign Großvater is not split up into two "submeanings" -as Kromann et al. (1991Kromann et al. ( : 2718) ) incorrectly believe -but rather, with "mütterlicherseits" and "väterlicherseits", the co-textual conditions for equivalence are presented which have to be fulfilled in order for Großvater to be semantically equivalent either to morfar or to farfar.Großvater is consequently not partially semantically equivalent to morfar and farfar in the sense that a "part" or "sub-meaning" of Großvater is partially semantically equivalent with morfar, and another "part" or "sub-meaning" with farfar.The postulation of such "sub-meanings" makes no sense.It is rather more appropriate to say that Großvater can be co-text-specifically equivalent with both morfar and farfar.It is therefore indeed better motivated and more appropriate to speak of co-text-specific semantic equivalence than of partial semantic equivalence.The other examples from different language pairs presented in Figure 7 can also be treated in the same way.The distinction introduced here -even if only very roughly -enables a more appropriate understanding of items differentiating between equivalents of a specific type, such as, for example, "mütterlicherseits" and "väterlicherseits" in (5).With these items specific co-text classes are determined referentiallysemantically.
Here I have to end my reflection for now.

Concluding remarks
If my exposition gives rise to a new consideration of the problem of equivalence within the framework of dictionary research, it has fulfilled its purpose.The suggestions made here will be further discussed in Wiegand 2002.http://lexikos.journals.ac.za