An Integrated Semasiological and Onomasiological Presentation of Semantic Information in General Monolingual Dictionaries as Proposed in H . E . Wiegand ' s Semantics and Lexicography

Herbert Ernst Wiegand is a very important figure in international metalexicography. A large part of his research has up to now been unavailable to the majority of the English-speaking world, because it has mainly been published in German. A new book Semantics and Lexicography seeks to break through this obscurity by providing English translations of a selection of articles spanning thirty years (from 1976 to 1996), which trace the development of Wiegand's views on semantic information. This book offers a valuable insight into the theoretical and corresponding terminological development that has already had such a remarkable impact on the "practical science" metalexicography. This article focuses on Wiegand's theories on the integration of the semasiological and onomasiological presentation of semantic information in the microstructures of general monolingual dictionaries. The theories are explicated and illustrated by examples from two Afrikaans general monolingual dictionaries and evaluated in order to establish their relevance. The conclusion is reached that the semasiological presentation of meaning is most appropriately located in the item giving the meaning paraphrase, which should form part of the semantic comment in an integrated microstructure. In the semantic comment, this item must be followed by dearly distinguished items giving onomasiological information, especially about synonymy.


1_
Historical context The last three decades have seen a tremendous growth in the academic study of the dictionary as a linguistic artefact.The publication of Ladislav Zgusta's Manual of Lexicography in 1971 proved to be a watershed in metalexicography.
His detailed synthesis of insights from lexicology and the systems behind the secondary lexicographic process (Le. the actual compilation of the dictionary) in particular~ provided a foundation for other academics to build on.One such academic was Herbert Ernst Wiegand.
Wiegand has, however, not only expounded on existing theoretical frameworks, In 1984 he published his General theory of lexicography (Wiegand 1984: 15).The focus of his subsequent research has rested heavily on the second subcomponent of his theory of the lexicographic deSCription of language: the textual theory for lexicographic texts.In this textual theory those texts which contribute to the transfer of semantic information in printed dictionaries play an important role, as these are critical to the user perspective.
A large part of his research has up to now been unavailable to the majority of the English-speaking world, because it has mainly been published in German.A new book Semantics and Lexicography seeks to break through this obscurity by providing English translations for a selection of articles spanning thirty years (from 1976 to 1996), which trace the development of Wiegand's views on semantic information, This development is plotted from early, pretheoretical groundwork on metalexicography (in which discussions take place largely within the confines of linguistic jargon) through to lucid, metalexicographical formulations and analyses of items that contain semantic information.The book therefore provides a valuable insight into the theoretical and corresponding terminological development that has had such a remarkable impact on the "practical science" of metalexicography., However profitable a thorough discussion of the historical value of these articles may be, such a discussion does not fall within the scope of this article.Instead, it is the aim of this article to offer a synthesis of some of the most important threads running through these articles and to illustrate the practical application of the theoretical postulates.The medium for this illustration will be Afrikaans general monolingual dictionaries, with the Verklarende Handwoordeboek van die Afrikaanse Taal (henceforth HAT) as representative of standard dictionaries and the Woordeboek van die Afrikaanse Taal (henceforth W AT) as representative in the comprehensive dictionary class.The focus of the discussion will be on the integration of semasiological and onomasiological presentation of semantic information in the microstructures of general monolingual dictionaries.As not all articles in the book deal with general dictionaries, the focus will be on those that do, but the author also considers relevant postulates from those articles dealing with dictionaries for languages for special purposes.

2.
The semasiological and onomasiological presentation of semantic information Wiegand initially draws a distinction between semasiological and onomasiological presentation in monolingual dictionaries.On a macrostructural level, semasiological presentation corresponds to a strict initial alphabetical listing and on a microstructural level, to an approach emphasizing the denotative meaning of a word.On a macrostructural level onomasiological presentation corresponds to a thematic listing and on a microstructural level, to an approach that emphasizes the semantic relations a word has with other lexical items in a specific language, e.g.synonymy, hyponymy, hyperonymy and antonymy.
Wiegand points out that current general-purpose monolingual dictionaries focus too much on the semasiological presentation at the expense of the onomasiological presentation.He proposes a more integrated approach on a microstructural level which will see the semasiological and onomasiological presentation of semantic information in separate article positions within the semantic comment in a dictionary article.

3.
The semasiological transfer of meaning

3.1
The term "definition" The predominant view in metalexicography is that the denotative meaning of a lexical item in a dictionary is represented by a lexicographic definition.Defi-nition types have been identified and qualitative criteria have been set as part of lexicosemantic theory.Yet the term "lexicographic definition" has gone largely unchallenged in metalexicographical circles as well as in editorial style manuals.
In several of the articles in this book Wiegand questions whether "lexicographic definition" is "an adequate meta lexicographical term" (Wiegand 1994: 241).He attempts to offer terminology with which one can describe the different elements in the lexicographic transfer of meaning.Initially, he replaces "lexical and lexicographic definition" with "lexical and lexicographic paraphrase".This shift seems to be motivated by a lack of faith in the seamless appropriation of the philosophical term "definition" for use in lexicography.By avoiding "definition" several of the terminological potholes, such as the difference between a nominal and a real definition, can be avoided as well.On this point Wiegand (1977: 93) states the following in his 10 theses on lexicography: Written, from an extra-communicative point of view, the leXicographic paraphrases should not be regarded as lexicographic nominal definitions, but as rule formulations for the use of the lemma in conformity with semantic rules.Only in this way can it be adequately explained why and in which way lexical paraphrases, when being read, become texts-in-function, thus 'enabling the reader to infer the meaning of the lemma.
A lexical paraphrase represents the denotative meaning of a lexical item which is "regarded as the rules of reference and predication for the use of (the lexical item) in habitual texts for (that specific lexical item)" (Wiegand 19~3: 148).The lexicographic paraphrase seeks to abbreviate these rules.
These early thoughts are clarified and systematised later in a seminal article "Elements of a Theory Towards a So-called Lexicographic Definition", which was published in German in 1994.In this article the "So-called lexicographic Definition" is not viewed in isolation, but rather as part of an inte-gr~tedmicrostructure within the paradigm of Wiegand's comprehensive textual theory for leXicographic texts, specifically as part of the integrate core of the semantic comment or sub comment (in the case of different senses).Therefore the terminology which is given preference is "an item giving the meaning paraphrase" (Wiegand 1994: 253).

3.2
The traditional approach to the semasiological presentation This rejection of the term "definition" as not being an "adequate metalexico graphical term" (Wiegand-1994: 241) underlines a more fundamental disagreement with its use in both meta-and practical lexicography.Numerous academic studies have highlighted the variety of "definition types" which can be employed in general monolingual dictionaries.In most of these studies, 123 preference is given to the so-called definition using genus and differentiae (see Gouws 1989: 115-116).These "true definitions" (Svensen 1993: 120), w~ch Svensen further describes as being "intensional", expressing "a generic conceptual relationship whereby concepts are arranged in classes according to similarities and differences noted between them" (1993: 122), can be used for most of the syntactic classes (parts of speech) to which a lexical item can belong.nus definition type is firmly rooted in a traditional approach to semantics where componential analysis is the starting point in the construction of a definition.In the introduction to Wiegand's book, Wolski makes it clear that Wiegand does not regard this theory of meaning as entirely suitable for the needs of lexicography.These articles therefore refleCt the development of an independent theory of meaning, which is in accordance with Wiegand's textual theory of lexicographic texts.An actional-semantic approach is taken to the expression of the denotative meaning of a lexical item.Denotative meaning is not abstracted from componential analysis, but, as was noted in paragraph 3.1, is "regarded as the rules of reference and predication for the use of (the lexical item) in habitual texts for (that specific l.i.)" (Wiegand 1983: 148).

3.3
The actional-semantic approach to the semasiological presentation In his later articles Wiegand clarifies why a definition based on componential analysis and presented as a so-called definition using genus and differentiae can never adequately reflect the denotative meaning of a lexical item (see Wiegand 1994: 250-252).The item giving the meaning paraphrase needs to be based on more pragmatic principles and a different perspective on the constitution of meaning knowledge.Within the actional-semantic approach Wiegand identifies types of meaning knowledge central to the lexicographer's efforts to represent semantic information in the dictionary.Wiegand (1994: 262) works mainly with the syntactic classes which contain the most words and in this context states: In the case of lexicalized predicators (adjectives, nouns, verbs) the actional knowledge for the linguistic acting consists of (i) the non-encyclopaedic meaning knowledge (the fact that a predicator is habitually used for performing the partial act of predicating the encyclopaedic, object-constituting meaning knowledge (the fact that a predicator is habitually used to refer to something defined by this very predicator or by the habitual use of other predicatorsi.e. by language and its use), -(iii) in some cases, by another kind of meaning knowledge (Le.that a predicator is used to evaluate the reference object, to express an attitude, etc .... ) ... Meaning descriptions ... have to impart the actional knowledge necessary to infer from these meaning descriptions (i.e. from items giving the meaning paraphrase) this specific knowledge and consequently the rules.
The rules referr~d to in this statement are the rules of reference and predication in Wiegand's interpretation of denotative meaning, of which the item giving the meaning paraphrase should be an abbreviation.
How does one determine what should form part of such an item giving the meaning paraphrase?Componential analysis is replaced with a pragmatic approach which focuses on the needs of the target user of the dictionary.Whereas the definition, which is a result of componential analysis, is a static entity, the item containing the meaning paraphrase is a dynamic entity which can be different for different types of dictionaries, in each case being adapted to the needs of the target user.
This innovative view on the construction of the item containing the meaning paraphrase can lead to the compilation of far more user-friendly dictionaries.It comes to terms with the fact that a dictionary article is a "simulated question and answer dialogue" (Louw: 1998: 105) between the lexicographer and the dictionary user.Wiegand (1994: 254) reiterates this point by stating that, "producing a dictionary article ... is the formulation of a number of potential answers in a condensed and standardised form to anticipated types of search questions which are integrated in types of use situations".Earlier in the book he states that "only by anticipation can the lexicographer hope to meet the expectations the user has of the codified results of the lexicographer's written acts, the dictionary excerpts" (Wiegand 1976: 14).An item giving the meaning paraphrase should then best be constructed according to a list of "user questions" (Wiegand 1994: 265).These user questions must be in line with the lexicographic needs assessment done as part of the dictionary plan and should be based on detailed empirical research to determine the user perspective.
If an item is constructed within these parameters "one can be reasonably sure that (against a presupposed background of a language community which has reached a preliminary understanding on the use of the lemma-sign in habitual texts) a competent user can infer from a correct item giving the meaning paraphrase for a designative lemma-sign ... which he understands correctly, the actional knowledge which as a rule is necessarry and sufficient to correctly make a habitual and correct reference with this lemma-sign" (Wiegand 1994: 265).
This shift in emphasis can be illustrated by some examples from Afrikaans monolingual dictionaries.If one employs, for example, componential analysis in order to determine the denotative meaning of the lexical item olifant (elephant) one would find an entry similar to the one in the HAT, but different to the one in the WAT.
It could be argued that user's needs have played a decisive role in the construction of the item containing the meaning paraphrase in both these articles.In the desk dictionary a somewhat cryptic, yet distinguishing paraphrase is given, while the paraphrase in the comprehensive dictionary is more detailed and specific.Both items provide more information than a definition constructed by means of a strict componential analysis and presented in the form of a definition using genus and differentiae.A probable question from meta lexicographical circles would be: is the additional information provided encyclopaedic, rather than semantic?In order to answer this question within the terms of reference of this article, one would have to return to Wiegand's theory of the kinds of meaning knowledge.

Semantic and encyclopaedic information
In Wiegand's semantic theory the boundaries between semantic and encyclopaedic knowledge are not as discrete as in traditional semantic theories.The relationship between the two types of knowledge are visually represented as follows in Wiegand (1994: 268): The item containing the meaning paraphrase must not only reflect the "nonencyclopaedic meaning knowledge" but also the "encyclopaedic object-constituting knowledge".In the two articles above this is certainly the case.
In the article from the HAT, very little information is added to what would be gained from astrict componential analysis.A supplementary component is included in the item giving the meaning paraphrase, viz. the fact that the ivory tusks are valuable.Taxonomy is also given.This is a procedure which Wiegand opposes, preferring taxonomic details in a separate article position.One can agree that the target user group of the HAT probably does not need these details, as thes~ users only require a very broad perspective on the object that is described.In general though, this item from the HAT succeeds in its goal by conveying the necessary object-constituting knowledge.
The needs and expectations of the W A T's target users differ greatly from those of the HAT's target user group.As a comprehensive dictionary, it has to provide enough information for the user to gain a detailed and specific perspective on the object described.Whereas the HAT would for example suffice with "very big" as a size description, the WAT gives a more detailed account of the size and weight of an elephant.The HAT only gives a cursory physical description of an elephant, enough (one• could argue) to distinguish it from other "very big mammals", but the W AT paints a. detailed picture not only of an elephant's distinguishing features, but also of the differences between the two main species of elephant.
Both these dictionaries have.transferred object-constituting knowledge by means of the item giving the meaning paraphrase.Yet judgements of adequacy in this regard depend solely on the context of the dictionary search which was conducted.
Neither dictionary stops here in their attempts to adequately describe the animal.The HAT proceeds to use own examples to enhance the meaning transfer.In the WAT's case, additional encyclopaedic information is presented by means of citations, which help to further improve the user's world knowledge.

Frames
In order to systematise this user-oriented approach to the transfer of denotative meaning in monolingual dictionaries, the concept of "frames" has been introduced by Wiegand (1989: 573) and expounded upon by Konerding (1993).This concept refers to a set of "thematic question groups" (Wiegand 1994: 272) which are set up for a specific type of lexical item.In the case of natriumlamp, such a set could constitute more or less the same as the set for tools proposed by Wiegand (1994: 272-273) and the one for musical instruments quoted by Smit (2000: 178-179).
Five broad groups of questions can be identified: 1.
How can the (outer) form of the object be described?(Wiegand 1994: 272) 2.
What are the "predicates characterising the hierarchically superordinate whole of which the object is an ingredient/part"?(Smit 2000: 179) 3.
What are the "predicates characterising activities in which the object functions/plays a role"?(Smit 2000: 179) 5.
Which other objects is the object under consideration similar to and in which respects are they different?(Wiegand 1994: 273) According to Smit (2000: 178-179), who discusses the groups of questions set up by Konerding (1993), question groups 1 and 2 fall under the subheading '~con stitutive relationships and features of the object", question group 3 under "phases of existence and distribution" and question groups 4 and 5, together with a separate single question on "other names for the object", under "the meaning of the object for people".
In the WAT's article for natriumlamp the object-constituting knowledge (as part of the denotative meaning) is imparted effectively by providing answers to some of the above questions.Its outer form and the "activities in which the object functions/plays a role" (Smit 2000: 179) are disclosed, but not the way the object is produced, as this is not considered to be truly relevant to the target user of the dictionary.Some idea is given as to "the hierarchically superordinate whole of which the object is an ingredient/part" (Smit 2000: 179) by means of the traditional first element of a "genus and differentiae definition".In this case, however, the procedure has some value, because the superordinate chosen is not hierarchically far removed from the lemma (as "lamp" would be for example)."Gasontladingslamp" is a direct superordinate and as such is a valuable point of departure to the users of the dictionary.By scientifically constructing such an item giving the meaning paraphrase, the W AT succeeds in conveying the necessary semantic information.Further encyclopaedic information about the object is available in the examples for those users who need or wish to know more than is required to identify the object.

3,6
Qualitative criteria In Wiegand's opinion, the use of frames for construction of items containing the meaning paraphrase also seems to obviate the need to identify traditional qualitative criteria.This includes criteria meant to encourage "good defining practice" (Landau 1984: 132) such as "priority of essence", "substitutability", "brevity", "simplicity", etc. (see Landau 1984: 132-138 for a more comprehensive discussion of these criteria).Wiegand (1994: 235-237) is especially harsh on substitutability, questioning its "relevance" and usefulness as a criterion.He (Wiegand 1994: 271-272) further re-emphasises an earlier call for "a re-orientation away from the traditional concepts of definition and, therefore, away from such fruitless alleged problems which also appear in their wake, e.g.circularity, incompleteness, vagueness, and redundancy".On this point Wiegand (1994: 272) adds the following in summary: ' These albeit difficult problems with respect to definition theory are of oruy marginal importance for leXicography as a scientific practice because every good lexicographer knows how to handle, for instance, circularity.Neither do these tasks relate to any tasks in dictionary research that are still of importance today.
These problems should, if the correct procedure is followed while defining/not arise.This once again shows the value of a well-designed dictionary plan in which scientifically sound structural guidelines (drawn up with the target user's needs in mind) form an integral part.

The onomasiological transfer of meaning
Not all the semantic information inferable from the frames needs to be presented in the item containing the meaning paraphrase.Several of the questions pertain to the semantic relations existing between the lemma and other lexical items in the lexicon.TItis wotild include questions from question groups 2 and 5, as well as questions pertaining to "other names for the lemma".In these cases Wiegand prefers segregation.The semantic relations should be given by items in other article positions which are separated from the position of the item giving the meaning paraphrase by clear, unambiguous structural markers.Different article positions should preferably be allocated to the various items, reflecting the different semantic relations.
As was mentioned in paragraph 2, the semantic relations under discussion include hyponymy and hyperonymy, synonymy and relations of semantic opposition.Of these, synonymy is primarily emphasized in Wiegand's book, with three of the nine articles dealing almost exclusively with either synonymy in general monolingual dictionaries or synonymy in dictionaries for special purposes.

Synonymy in general monolingual dictionaries
The term "synonymy" does not escape redefinition.More so than "definition" perhaps, the term "synonymy" has been a bone of contention in metalexicographical circles.The common misconception of synonymy as referring to "two words meaning the same thing" has been severely criticised.Firstly, it is usually pointed out that synonymy is a relation between lexical items and not merely between words.Secondly, the inherent vagueness of the phrase "meaning the same thing" is criticised.It does not address the complexity of synonymy as a semantic relation.In fact as Louw (1998: 176) points out, "there are few if any absolute synonyms in a language" (on this point see Louw and Nida 1988: 15).Contextual and other differences which exist between two lexical items with similar reference, have necessitated the distinction between absolute and partial synonymy.Rather than viewing these terms as polar opposites, it is useful to view synonymy as a scale with these two types of synonymy at various ends.One could argue that even though there are few absolute synonyms in a given language, there are usually many that are close to them on the scale of synonymy and can therefore be Classified as near-or near-absolute synonyms.In this .class,the contextual differences are relatively slight and the interchangability of synonyms would not cause substantial embarrassment to speakers, as for example with a standard language and vulgar pair.It is mainly the class of near-synonyms that concerns Wiegand.
In this context Wiegand (1983: 146) redefines lexical synonymy as occurring between two lexical items when "the rules of usage ... (are) similar to such an extent that the rules of reference and predication are the same in habitual texts".The lexicographer must represent this "tolerance relation" of substitutability effectively when presenting leXicographic synonyms.

4.2
On the so-called "synonym definition" Even in his earlier articles, Wiegand criticizes the treatment of synonyms in general monOlingual dictionaries.The presentation criticized is mainly the socalled "synonym definition".This type of definition has generally been accepted in metalexicographical circles, though with some reservations.These reservations have mainly targeted the inadequacy of lexicographic articleexternal cross-referencing between the "synonym definition" and corresponding target entry, as well as that between the article where the "full definition" appears with the synonym mentioned and the article with the synonym definition as target entry.
With regard to standard or smaller general monolingual dictionaries, the prevailing view is that "synonym definitions" are necessary.Landau (1984: 270) states that "synonym definitions are not necessarily a mark of lazy or inept editing; they may be the best solution to the problem of too little space".Svensen (1993: 119) concurs, but also explicates the major reservation as follows: Using synonyms and near-synonyms as definitions saves space, and the method is entirely valid when the need for semantic precision is not too great, but one has to be on guard against synonyms that have multiple meanings.If a synonym is polysemous, it must not stand as a complete definition, but must be disambiguated ... Unfortunately this is often not done.Consider, for example, the treatment of the near-synonyms kyker and oog in HAT.
, Both these lemmas are polysemous and in' one sense of each, they are used to refer to the same object.At kyker 2 a so-called synonym definition is given, i.e. "oog".It is not specified for which sense ofpog this reference is valid.Only an experienced speaker of Afrikaans would kn,ow that the first sense of oog is the relevant one and if this speaker already knew this, there would be no need to consult the dictionary.The problem is compounded by two other factors.Firstly "kyker" is not even mentioned under oog 1 as a possible synonym.Sec-ondly,-a further "synonym definition" ("oogappel") is inserted into the same article position under kyker 2 and separated from "oog" by a comma.The conclusion to which a dictionary user would probably come is that these two are synonyms of kyker and of each other (since commas are usually used to separate so-called synonyms in the semantic comment).Yet "oogappel" (in its literal use) refers to the pupil of the eye whereas "oog" refers to the whole eye.They cannot be listed together without further semantic and contextual guidance.Furthermore, oogappel is itself a polysemous lemma and no attempt is made at kyker to disambiguate the reference.This confusing use of so-called synonyms and so-called synonym definitions is an obstacle to the effective transfer of semantic information in a single-volume general monolingual dictionary.The problem is even greater in comprehensive general monolingual dictionaries, where cross-references are not only article-external.Comprehensive dictionaries often comprise several volumes and cross-references are often also volume-external.The user is then expected to combine the correct pieces of information by making use of remote article-external and volume-external nonlemma tic addressing procedures, which are very difficult manoeuvres.Furthermore, a comprehensive dictionary often takes a long time to complete, with any number of years passing between the publication of its first and last volumes.Cross-references are therefore often made to lemmas which will only appear in much later volumes.These "synonym definitions" are therefore not easily "disambiguated" by means of a reference to a specific sense of a lemma, because the article for that synonym has not been constructed yet.Comprehensive dictionaries are also only revised after all volumes have been published, which means that the .userwill have to contend with imprecise "synonym definitions" for a long time.
As with "definition", Wiegand rejects both the term "synonym definition" and its application.He (Wiegand .1976: 25) seeks to redefine this item (which was traditionally referred to as a "gloss" at that time) in the very first article, referring to "dictionary excerpts in the form of lemma word" instead.According to Wiegand, this type of dictionary excerpt is not suitable for use in dictionaries, as a synonym cannot fulfil the same function as a lexicographic paraphrase given as the item containing the meaning paraphrase, i.e. an abbreviated rule formulation.
In the later articles this view is systematically supported by distinguishing, on the grounds of separate functions, separate article positions for items givmg the meaning paraphrase on the one hand and for word synonyms on the other.With the textual theory for lexicographic texts as broad framework, word synonyms are to be presented as part of the integrated microstructure as follows: Word synonyms ... belong into the SK (semantic commentary) yet not in the semasiological, but in the onomasiological part.Here, they are best listed under an explicit cross-reference by synonyms (e.g., "Sy" or "~"), which can be understood as a semantic commentary symbol, functioning as across-reference ... ; the synonyms should be printed in such a way that the font differs from the one used for the ME (meaning explanation) ... In each SK, the ME is therefore obligatory and word synonyms are listed if there exist any (Wiegand 1983: 150).Wiegand (1983: 150) adds that this separation of semasiological and onomasiological information is important, because "word synonyms are not 'hidden' in the ME, instead, semantic knowledge of the lexical structure is made explicit".This is illustrated in the treatment of onbewerk and onverwerk from the WAr.
The WAr does employ so-called "synonym definitions", but with accurate specification of the sense to which the synonym applies, in this case "Onverwerk (ONVERWERK I)" at onbewerk 2. The synonym definition is met by a "synonym mention" in the other article.This mention is close to Wiegand's proposal for the treatment of all synonyms in general monolingual dictionaries.It is positioned close to the item giving the meaning paraphrase and separated from it by means of a typographical structural marker, here a semi-colon.A second structural marker, "sin.", is added for more clarity and this is followed by the synonym, in this case onbewerk, which is presented in italics, whereas the item giving the meaning paraphrase is given in roman font.It is also separated by means of typographical structural markers from its super<;eding information categories.To complete the picture, some field labels in •the microstrucfural treatment of the synonym, such as "(minder gebruiklik)" in this case, are also included in the synonym mention, in order to indicate the restrictions on use which exist for that specific synonym.A criticism of the WAT is that sense specification is not presented at this item, but only at the so-called synonym definition.The ideal would be a repetition of the full item giving the meaning paraphrase, with synonyms presented in separate article positions in the format displayed by the WAT, but with an accurate sense specification as employed in the WAT's "synonym definitions".This would also be the most effective way of curbing the problems created by synonym references between volumes, Consequently, one can agree that, in keeping with the user perspective, Wiegand (1983: 150) is correct in remarking that "the 'semasiological principle of the item giving the meaning qy means of word synonyms' ... should ... be substituted by an 'onomasiological principle of supplementing the lexicographic meaning explanation by word synonyms"'.
This would benefit users of comprehensive dictionaries, who would probably also benefit most from a repetition of a full meaning paraphrase and two-way synonym references in well-demarcated article positions.This can be an effective answer, if the lexicographer, as part of the planning of the semantic component of an article (preferably by using frames), anticipates as many of the pOSSible near-synonyms as possible within the dictionary's macrostructural scope.This is a demanding brief, but seems to be in keeping with Wiegand's (1976: 24) view of the lexicographer as "an empirical scientist".The corpus, combined with other lexicographic sources and the lexicographer's own linguistic intuition, should provide an adequate picture of the lemma's synonymous relations with other lexical items.
Whether the repetition of "meaning paraphrases" is a viable option for smaller dictionaries, is a different matter.In commercially published standard, desk or smaller paper dictionaries, the maximum level of textual condensation is pursued.In this regard "synonym definitions" have always been an important space-saving (and therefore money-saving) tool.It is hard to see Wiegand's vision of a separate slot for synonymy in this predatory milieu.Perhaps the best one can hope for is that a suitable "synonym definition" with a precise cross-reference will be employed and met by an equally precise mention of that lemma as a synonym in the article of the more frequent lexical item, which should also contain a full meaning paraphrase.This mention, in the form of an item, should be separate from the item giving the meaning paraphrase and should be introduced by a unique structural marker.An adequate distinction of the contextual differences between the synonyms should then be drawn by means of labels.The main concern is that the decision be based on the needs of the target users as revealed by a detailed empirical study.

Other semantic relations
One of the most challenging tasks of the lexicographer is to adequately underpin the semantic relations between a specific lexical item and others in the macrostructural scope of the dictionary.Corpus material is often inadequate and decisions of this nature usually rely on "the language intuition, language competence and therefore world knowledge as well as world conception of the lexicographers, taking into account the empirically ~stablished use of language" (Wiegand 1976: 36) and on secondary sources such as thesauri and antonym dictionaries.In order to make the lexicographer's task easier, it is therefore important that there be a scientific system of presenting semantic relations, which can be applied consistently throughout the dictionary.
In the articles selected for publication, Wiegand focuses on presenting such a system for the lexicographic treatment of synonymy, but he also provides useful guidance for the treatment of:other semantic relations.In a onevolume semasiological dictionary (such as the HAT) Wiegand (1982: 133) finds the indication of synonymy sufficient, but for "a multi-volume integrated dictionary" (such as the WAT) he proposes the same treatment for hyponomy, antonomy, etc. as he did for synonymy, i.e. an item in an article position separate from the item giving the meaning paraphrase and well-demarcated by a unique structural marker.
Attempts have been made in the WAT to reflect relations of semantic opposition in this way.A separate article position is employed and this item is separated from the preceding definition or synonym by means of a semicolon.No further distinction is, however, made as to whether the item denotes antonymy or shows that the user is dealing with a complementary pairing.Both types of item are introduced by the structural marker "teenoor".
Unfortunately the W AT does not have a consistent system for dealing with hyponymy and hyperonymy which is in line with the one proposed by Wiegand.As pointed out in 3.5 under the discussion of natriumlamp, superordinates or hyperonyms are usually given as the points of departure in items giving the meaning paraphrase, but there are no separate article positions containing hyperonyms as independent items.
With regards to hyponymy there is a tendency not to present hyponyms or cohyponyms, because it is not, for example, seen to be in the interests of the target users to present "lemoen", "pomelo", etc. as cohyponyms at nartjie.Only in what the lexicographer believes to be exceptional or difficult cases where the two objects are, for example, closely related and therefore easily confusable, are cohyponyms given.This would then be presented near the end of an article and introduced by "Vgl."("Cf.").There are two problems here.Firstly, this placement causes a separation of this type of item from the other items giving onomasiological information.In fact, it separates the hyponym or cohyponym from all other items in the semantic comment, thereby making the user's search path harder and less predictable.Secondly, the article position introduced by "Vgl."does not only contain cohyponyms.Partial synonyms and other confusable words (even though a separate article position for these does exist) find their way into the "Vgl."-item.The resulting loss of uniqueness is a further stumbling block on the user's inner search path.
In this regard, Wiegand has a valid point of view.The onomasiological content of these items is an integral part of the minimal frame for each lexical item, and the item itself should be a crucial part of an effective, integrated microstructure, specifically as part of the semantic comment presented in close proximity to the item giving the meaning paraphrase.

Conclusion
In this article I have attempted to clarify, apply and also bring together two major threads running through this selection of articles.The semasiological presentation of meaning has been seen to be located most appropriately in the item giving the meaning paraphrase, which should form part of the semantic comment in an integrated microstructure.In the semantic comment this item must be followed by clearly distinguished items giving onomasiological information, especially on synonymy.This article has, however, merely begun to show the potential of Wiegand's theories (such as the frame theory) for the improvement of dictionaries' user-friendliness.Lexicographers working on dictionaries for specific languages need to explore this potential for the specific types of dictionaries they wish to compile.We are entering an exciting new lexicographic dispensation in South Africa and the lexicographers of both the new and the existing National Lexicographic Units would do well to take cognisance of Wiegand's unique and powerful theories when planning the secondary lexicographic processes of their respective dictionaries.