The Microstructural Treatment of Sub lexical Lemmas in Afrikaans Descriptive Dictionaries

Arguments in metalexicographic literature on the status of subword and multiword lexical items resulted in a more comprehensive lemma tic treatment of these lexical Hems in the latest editions of Afrikaans descriptive dictionaries, e.g. Woordeboek van die Afrikaanse Taal Volume IX (WAT), Verklarende Handwoordeboek van die Afrikaanse Taal (HAT), Nasionale Woordeboek (NW), Verk/arende Afrikaanse Woordeboek (V A), and Basiswoordeboek van Afrikaans (BA). This article seeks to investigate whether sub lexical lemmas are microstructurally treated the same as lexical lemmas, or whether lexicographers still distinguish between these various types of lemmas in some way or other. If differences exist in the treatment of the various types of lemmas, Afrikaans lexicographers are still word-biased: although subword lexical items are sometimes microstructurally treated in the same way as lexical items in that they are given lemma tic status, they are nevertheless distinguished microstructurally. For example, in WAT IX and HAT, sublexical lemmas are treated in almost the same manner as word lemmas in that their pronunciation is ,consistently given. However, in most of the dictionaries mentioned, grammatical information on sub lexical lemmas is found as part of the definiens. Except for BA, no Afrikaans dictionary illustrates the contextual use of sub lexical lemmas by means of examples or quotations; only their formation products are provided. More labels, and even etymological information, should also be provided more frequently as part of the microstructure of sublexicallemmas. On the basis of the precediflg it can already be concluded that Afrikaans lexicographers have not nearly done enough in removing traces of word-bias in descriptive dictionaries. In addition to macrostucturally treating subword lexical Hems as lemmas, these Hems should, where possible, be dealt with on a microstructural level in exactly the same way as lexical lemmas.


Introduction
Arguments in metalexicographic literature (e.g.Gouws 1989, 1990, 1991, Rettig 1989, Stein 1985, Zgusta 1971) on the status of subword and multiword lexical items resulted in a more comprehensive lemmatic treatment of these lexical items, especially subword lexical items, in the latest editions of Afrikaans descriptive dictionaries, e.g.Woordeboek van die Afrikaanse Taal Volume IX (WAT), Verklarende Handwoordeboek van die Afrikaanse Taal (HAT), Nasionale Woordeboe1c (NW), Verklarende Afrikaanse Woordeboe1c (VA), and Basiswoordeboek van Afrikaans (BA).Gouws (1991: 75) states: "To ensure a sound treatment of all these lexical items the traditional word-based lexicography should be replaced by a broader lexicon-based approach that offers a more comprehensive reflection of the lexicon by listing and treating multilexical and sub lexical lemmas" (own italics).This article seeks to examine whether this broader lexicon-based approach in more recent Afrikaans descriptive dictionaries has indeed resulted in a more sound treatment of sub lexical lemmas, or whether the macrostruc-tural changes merely ascribed to them the same lemmatic status as that of lexical items, while still distinguishing between them; on a microstructural level.
The types of information to be discussed follow the system of Hausmann and Wiegand (1991: 343) for the specification of microstructural information in monolingual descriptive dictionaries.It includes synchronic identification dealing with the form of the lemma, e.g.pronunciation and lexical category, diachronic identification that provides the etymology, labelling, descriptive i~ormation such as the definiens, syntagmatic information that covers collocations and examples, and paradigmatic information on the semantic relationship between the lemma and other lexical items, e.g.synonymy, opposition, homonymy and polysemy.
Dictionaries generally treat subword lexical items more satisfactorily than multiword lexical lemmas, since in form subword lexical items resemble words more closely than multiword lexical items.For example, some stems are used both lexically and sublexically, the only difference being the hyphen attached to the sublexical stem.This inevitably leads to a freer inclusion of subword lexical items in the macrostructure, but does this necessarily mean the same microstructural treatment as that of lexical items?
In this paper the only distinction regarding sublexicallemmas will be the distinction between affixes and stems.Stems here also include techno-stems.

Affixes
According to Muller (1989: 876), suffixes in German dictionaries generally show an inconsistent lemmatization.The same applies to Afrikaans dictionaries in that suffixes are lemmatized even more inconsistently than prefixes.The reason for the more frequent lemmatization of prefixes is that they are easier to alphabetize, and at first sight also appear more word-like than suffixes~ In German dictionaries lexicographic treatment of suffixes is less intensive than that of prefixes.Muller adds that this is also true for French and English dictionaries.This is not the case with Afrikaans dictionaries.When affixes are lemmatized, the microstructural treatment of prefixes and suffixes is very similar.See _inC and in-' in HAT: _inC Onproduktiewe agtervoegsel-tans in die baie gevalle ongebruildik waarmee vroulike persoons-en diemame gevorm is van manlike persoons-en diemame, bv. in Iodin, koningin, gemalin, vriendin, waardin, eselin, leeuin.Consequently, no distinction will be made between prefixes and suffixes ir\ the following discussion.

Stems
Sublexical stems, especially techno-stems, are more consistently lemmatized than affixes, and in particular suffixes (see Carstens 1995: 148-149).This could be ascribed to two factors: stems resemble words more closely than affixes, and since many of the techno-stems are of Latin or Greek origin, they require elucidation.As a result of the closer resemblance between words and stems, one would expect that if discrepancies in the microstructural treatment of words and suffixes occur, the inconsistencies between that of words and stems would be less.

Pronunciation
Although it is sometimes provided inconsistently, pronunciation is one type of information seldom neglected in sublexical lemmas.WAT IX is the only descriptive dictionary to provide phonetic transcriptions, and all sub lexical items are consistently transcribed.However, they consistently lack an indication of main stress.Compare the information for word lemmas and sub lexical lemmas: Stress guidance is in fact required for sublexical -logie, since the main stress differs from that of logies.
In HAT, where main stress and syllabic divisions are given for lexicallemmas, the same applies mostly to the sublexical items: gal.va'no-, -Lseer', Lo'ne- Carstens (1995: lSI), however, identifies the following inconsistencies in HAT: for some techno-stems neither main stress nor syllabic division is given: chiro-, filo-, fono-, hid'ro-, intro-, mega-, meta-, tele-, while for.others no main stress is given, although the syllabic division still occurs: e.lek.lro-,ga.mo-, he.te.ro-,ho.mo-, xe.no-In VA, where only the main stress is given, inconsistencies are also found: for in'tra-, io'ne-, he'tero-and me'de-stress is given, but not for iso-and kouewater-, kouwater-.
In other articles no stress pattern is given for any product, while the sub lexical lemma itself shows the main stress: 'anti-: 1. ... anti-Christelik, anti-republikeins, '" anti-militarisme, ... 2 .... anti-kritiek, ... 'hetero-: ... Heteroseksueel, -sentries.These inconsistencies do not point to an underestimation of sub lexical lemmas; they merely show lexicographic carelessness.The complete omission of information such as the indication of the main stress in W A T IX seems to reveal a more fundamental problem: sublexical lemmas are treated differently than lexical lemmas, which could indicate a negation of their lemma status.

Lexical Category
One of the most noticeable anomalies in the treatment of sublexical lemmas is the statement of the lexical category both with regard to the slot in which it occurs and the terms used for this classification.In every Afrikaans dictionary except BA, the lexical classification for all types of sublexical lemmas is.presented as part of the definiens, while the lexical categories for lexical lemmas occur in a specific slot, namely directly after the lemma (HAT and VA), or after the phonetic transcription of the lemma (W AT IX).BA is the only descriptive Afrikaans dictionary to consistently categorize sublexicallemmas in the same slot as lexical lemmas, namely directly after the lemma (henceforth the appropriate part in the various articles used as examples will be underlined): -toe agtervoegsel.As for WAT IX and VA, there is no justification for devaluating the semantic status of the sub lexical lemma by including grammatical and functional information in its definiens.The definientia of sublexicallemmas should and could be reserved mainly for semantic information.(VA) kwasi-.Asof, kamtig, kastig, veral gebruik as voorvoegsel (woorddeell wat aan die grondwoord die betekeniswaarde gee van: ... (VA) In considering the embedded lexical category of kwasi-(V A) in its definiens, Carstens (1994: 259) stresses that the user would be able to retrieve the correct grammatical information far more easily if only the lexicographer used the expected slot more consistently.However, VA treats its sublexicallemmas in more detail than it does its lexical lemmas, since no lexical categories are given for lexical lemmas, except when they are multifunctional.
According to Stein (1985: 43), a well-argued theoretical distinction is needed between the different types of sub lexical lemmas, e.g.prefixes, suffixes, technostems, and other bound stems (see also Gouws (1989: 85-96) for a classification of sublexicallemmas).Unfortunately, such a well-argued theoretical distinction does not exist in Afrikaans descriptive dictionaries.The most appropriately termed, with the least deviations, are the affixes.Interestingly enough, all the lexicographers, except those of BA, use the Germanic terminology voorvoegsel and agtervoegsel for the Latin-derived prefix and suffix respectively.This in itself creates problems in the distinction, since the Germanic terms are usually more vague than their counterparts of Latin origin.BA terms the lemma -toe (as in agtertoe and huis toe) an agtervoegsel, but it is not a suffix.
Sometimes affixes are termed more descriptively, but this is usually the exception to the rule.See HAT for: -iesZ, ook -etjies, -jies, -kies, -tjies Affektiewe verkleiningsuitgang ... wat meestal voorkom by (a) adjektiewe en bywoorde waarin dit die betekenis het van ... Further inconsistencies are found in the designation of more foreign sub lexical items.Van Niekerk (1991: 289) shows the variety of terms used for technostems in the 1979 2 HAT.Nothing much has changed in the 1994 edition of HAT, since, according to Carstens (1995: 150-151) In the above examples, the treatment is identical to that of lexical lemmas, i.e. with the etymological information separate from the actual definiens.Unfortunately these examples are less common: for most of these sublexicallemmas etymological references are incorporated in the definientia.Stein (1985: 40) criticizes this deficiency in the articles of sub lexical lemmas, where etymologicalor diachronic criteria apply instead of the required synchronic criteria such as functional and semantic aspects.
It can be added in HAT's favour that the etymology is also given for those sub lexical lemmas which the average user would not consider foreign: -isme l Taamlik produktiewe agtervoegsel ontleen aan F.< G. -ismos, en gebruik om benamings te vorm ...
-ier 2 , ook -enier, -nier, -r Agtervoegsel ontleen aan F. -ier < L. -arius met die betekenis '" _ies l , soms ook -iek, -ities, -ties Adjektief-en bywoordvormende agtervoegsel, afgelei van F. -ique < L. -icus met die betekenisse ... HAT seems to be the only Afrikaans descriptive dictionary to recognize the important encoding function of etymological information, especially in sublexical lemmas (see also Van Niekerk 1991: 288).For example, the lexical isme does not have etymological information, but as can be seen above, the sublexical -isme has.The etymology enables the user to interpret unexplained lemmas or lexical items not included in the dictionary.The user can also create his or her own lexical items accurately and correctly if he or she knows the original meaning obtained from the etymology.

Labelling
According to Stein (1985: 39), restrictive usage labels are not applied as consistently in sublexical as in lexical lemmas.This is definitely the case in Afrikaans dictionaries.Where the user would expect a label, there is none, and where there are labels, their application is inconsistent and differs from that of lexical lemmas.Compare the following entries in HAT: lui l mv.(w.g.: veroudJ Mense: ...

Reproduced by Sabinet Gateway under licence granted by the Publisher ( dated 2011)
http://lexikos.journals.ac.za The Microstructural Treabnent of Sublexical Lemmas

3S
The sublexical -lui is nowadays used just as infrequently, and is therefore becoming equally archaic as the lexical lui, but only the latter is labelled.In WAT IX, however, 3-lui is microstructurally correctly labelled: In NW, the label for by-is also included in the slot reserved for labels in the articles of lexical lemmas: by-: ... 3. (krieket> een van verskillende posisies van veldwerkers aan ... bybreekbal (krieket) bal wat so geboul word dat dit ... It has already been mentioned that etymological information is given for most of the foreign techno-stems in HAT.However, only a few are provided with labels to show in which scientific or technical field they are the most produc~ tive.Labelling would to an eVen greater degree enhance the encoding facility effected by the etymological information.When HAT does label sublexical lemmas, the method is as follows: in_ 3 t ~ Eerste lid van 'n aantal weinig gebruikIike samestellings ... in-I Onproduktiewe agtervoegsel-tans in die baie gevalle ongebruiklik -waarmee ... -iee Agtervoegsel ontleen aan F ... , (c) in die chemie -'n sout van 'n suur met minder as ... by_3 ... 3 In krieket, posisie aan die rugkant van die kolwer; ... Of these four examples, only the first is labelled in the same way as lexicallem~ mas.The others all have the restrictive usage indicated as part of the definiens.Even in the lexicographically correctly labelled example (in-\ the information of the label is repeated in the definiens.

S.t Definiens
According to Stein (1985: 39), "(a)ffix definitions ... are usually extremely brief and therefore often unsatisfactory".Unfortunately the definiens for regter-in BA is not semantically as explicit as the one for linker-.This is easily observed, since these two lemmas are crossindexed.
In WAT IX, for instance, circular definitions are used for some sublexical lemmas.The sub lexical lemma is defined with reference to the lexical item: lede-... la Aanvangskomponent van komposita waarin ledein bet.parallel is aan lede CLEDE), by ....This is possible only if the user has access to the definiens of the lexical item.
For lede-this is lexicographically sound, but for laeweerstands-thiS can be confusing, since lae weerstand is not referred to the section for specialized expressions of 21aag.where it is defined: laeweerstands-... Aanvangskomponent van selfst.komposita, wat aandui dat dit wat m.d.tweede komponent benoem word, 'n lae weerstand besit of bied, of vir of van 'n lae weerstand is .... Stein (1985: 39) also misses information on collocation restrictions in the definientia of affixes.Collocation restrictions are found in two Afrikaans dictionaries, namely BA and W AT IX, but then only with sub lexical stems: aand-... 2 Aand-beteken vir 'n besonder deftige geleentheid wat gewoonlik in die aand plaasvind.Oit word veral ten opsigte van kleredrag gebruik: ... (BA) laat-...I 1 a ... Aanvangskomponent van selfst.komposita wat 'n laat of latere stadium van 'n geologiese tydperk.'n tydperk i.d.kultuur-of kunsgeskiedenis.e.d.aandui; ... (W AT) Collocation specifications should be common lexicographic practice, not only for lexical items, but particularly for all types of sub lexical lemmas.This will provide the user with specific encoding instructions regarding the type of stems with which a specific sublexicallemma can be combined.
If lexicographers consistently and adequately define sub lexical lemmas, they would perhaps not feel the need to fill the definiens with nonsemantic baggage such as grammatical and etymological information.

Collocations and Examples
The absence of contextual guidance in the form of eith~r collocations or example sentences is one of the greatest lexicographic shortcomings in the articles of sublexicallemmas, affixal and others.Syntagmatic information takes the form of lists of one-word citations illustrating formation potentialities of the sublexical lemmas as initial or final components.According to Rettig (1989: 645), this is "Information zur Wortbildungslehre" rather than actual syntagmatic information.The word-generating processes of sub lexical lemmas should be fully recorded, since the formation product as such is the environment within which the full value of the sub lexical lemma is realized.The formation product, however, is not used in isolation.It should be placed in a broader context in order to demonstrate and complement all the other microstructural information within the article of the sub lexical lemma, e.g.labels, definientia, etc.This would also lead to greater cohesion within the article.BA is the only Afrikaans dictionary to not only record, but also illustrate word formations by means of collocations or example sentences: on-prefiks.

Synonymy
According to Muller (1989: 880), the articles of sublexical lemmas could also include synonyms and semantic oppositions.Since all Afrikaans descriptive dictionaries supply these lexical relations for lexical lemmas, there is no reason why they should be omitted in the articles of sublexicallemmas.Naturally the synonym or antonym for sublexical lemmas need not be sublexical as well, since sublemmas could have lexical or multilexical items as synonyms or oppositions.Synonyms are frequently used as definientia for sub lexical lemmas (see paragraph 5 on descriptive information).In addition to other semantic information in the definientia of sublexicallemmas, synonyms are only found in BA Reproduced by Sabinet Gateway under licence granted by the Publisher ( dated 2011) http://lexikos.journals.ac.za and WA T IX.In BA, a synonym occurs in the article of aand-.In W AT IX, synonyms are found in the articles of linker-, links-, and laevet-in the same slot as that for the lexical lemmas.Usage labels are sometimes given for the synonym in WA T IX, which points to good lexicographical practice.Lexicographically, WAT IX handles the synonymy of this paradigm very well in that cross-references to e.g.linker-are made in the articles of linkerkantse and linkerhand-.Even hot-in W AT IV refers to links-in volume IX.Unfortunately there is no reference to aand-in the article of formeel in BA.

Opposition
The lexical relation of opposition is provided in most of the Afrikaans descriptive dictionaries, but rather infrequently for sublexicallemmas.One of the biggest shortcomings here is inconsistency, not only regarding the frequency or the slot used, but also, in most cases, regarding the incomplete cross-referencing of the two opposing terms.Most of the dictionaries except HAT and VA provide the opposite terms for linker-I links-and regter-I regs-.NW, however, gives the opposite term for regter-very explicitly (the opposition relation to linker-is used as the definiens), but at linker-the opposite term is absent, making it impossible for the user who refers to linker-first, to find its opposition: regter-: teenoorgestelde van linker.Regterarm, -hand, -oog, -voet.linker-: aan die kant van die liggaam waar die hart is.Linkeragterbeen, -arm, -been, -hou, ... The same incomplete lexicographical treatment is found in the article of the cricket term by-, where wegis indicated as opposition, but for weg-no opposition (or label) is given: by-: ... -BA certainly treats linker-and regter-well regarding opposition, but its treatment of particles shows inconsistencies: for two polysemous distinctions of onder-, oppositions are provided.However, for the first one oor-, there is no cross-reference in its article to its opposition onder-, and for the second one bo-, no sub lexical lemma is recorded at all.onder-... 1. Wanneer onder-vooraan 'n woord staan, '" ondergewig.onderproduksie; onderbelig, ... Die dorp is reeds onderbevolk, ... Inconsistencies are found here as well, since sublexical lemmas on the same page as laevolume-, such as laeweerstand-, laevesel-, laevet-and laevlak-do not have their opposites indicated.The only inconsistency regarding the use of the slot of the opposing term in the articles of sub lexical lemmas, compared to that of lexical lemmas, occurs in HAT.In the article for anti-, it is incorporated as part of the usage, information.Incidentally, anti-is the only sublexical lemma found in HAT for which an opposition term is recorded.anti-Voorv.voor b.nw.en s.nw.met die betekenis "teen", by.antikerklik, anti-Russies: .... antibakteries, ... Dikwels ook gebruik (teenoor pro-) by geleentheidsvorminge, by.antiaUes, antirook, ' " in which these indications are done, is not always lexicographically way lly sound.W AT IX records three separate homonyms for the sublexical ~lq~a as well as three homonyms for the lexical lui (see Feinauer (1996».In ~T, however, sublexic~ -lui is indicated as pa~ of the ho~onymic paradigm f lexical lui.HAT conslstently records sublexlcal and lexlcallemmas incor-~ctlY as homonyms: they c~ot be d~ssified ~s homonyms since.the hyphen of the sublexicallemma shows It to be different m form from the lexlcallemma.Stein (1985: 40) mentions that in the case of affixes, the etymological principle often overrides functional and semantic aspects.This leads to the indiscriminate grouping of affixes of the same form.What should synchronically be listed as homonyms are still treated as various polysemous senses on the grounds of diachronic semantic resemblances.WAT IX and HAT, e.g., both record -logie as a polysemous sublexicallemma, whereas their various meanings have drifted so far apart that they should have been recorded as homonyms: -logie '" 1 Slotkomponent van selfst.komposita m.d.bet."leerstelling.teorie of wetenskap", bv.antropologie, biologie, ekologie, ...

(HAT)
This treatment of homonymy in Afrikaans descriptive dictionaries, however, is not unique to sub lexical lemmas, but has been criticised often with regard to lexical lemmas as well.
, 5 are termed voorvoegsel(s) (prefix(es», and 6 are described as eerste lid van samestellings I woorde (first part of compounds / words).This does not, however, occur in HAT alone.BA also terms hiper-and super-respectively as prefiks (prefix) and eerste deel van samestelling (first part of compound).The compilers of VA, in their tum, give two possible terms, leaving It to the user to choose the more appropriate one: As was argued above, this might also be the reason for the difficulty in finding the correct terminology to describe sub lexical lemmas.It varies from first / second part / segment of compound, initial/final (word) element, initial / final component of compound, to word part / element.Afrikaans grammarians, and subsequently lexicographers, find it difficult to create and establish new classificatory terms for lexical items smaller or larger than words unless woord forms part of the terms.Lilliput.denkbeeldigeland in 1. Swilt se Gulliver's Travels.waarinalles.ookdie inwoners.baieklein is) ...HAT frequently provides etymological information in the articles of sublexical lemmas.Most of the foreign techno-stems are provided with etymolOgical information such as: super-... !LJ Eerste lid van ss.met die bet....