Since the article by Gouws (1992) on the lexicographical treatment of co-ordinating conjunctions, very little research has been done regarding the treatment of different types of conjunctions in Afrikaans dictionaries.
Within the newer approaches to grammar (like functional systemic grammar and cognitive grammar) the traditional distinction between lexicon and grammar is less pronounced. Linguistic categories such as conjunctions are to be found somewhere on a continuum between lexical morphemes (Langacker´s term) such as giraffe and encyclopedia and grammatical morphemes such as -ing¸ of and be. When a closed set like conjunctions behave both semantically and syntactically as a word class of their own, it is reasonable to expect that they too must receive a special type of lexicographical treatment – currently there is no evidence of this in Afrikaans dictionaries.
Conjunctions function to integrate clauses, and on a structural level, it is their syntactic behaviour (the way in which they influence the word order of the clause which is integrated) that forms the basis for the distinction between the following three types of conjunction in Afrikaans: coordinating conjunctions, subordinating conjunctions and conjunctional adverbs. Since subordinating conjunctions and conjunctional adverbs are syntactically marked in Afrikaans (in contrast to coordinating conjunctions such as en, maar, want, of) syntactic sub-categorization is clearly called for when dealing with these conjunctions in a dictionary (Gouws 1992:93). This sub categorisation and appeal by Gouws is not indicated in the dictionaries consulted in this research.
Of equal importance is the requirement for explicitly mentioning the function of the conjunction, not only within sentence boundaries, but also within broader textual units. Semantic relations like reason, cause, etc. can be indicated in a dictionary by making use of the various taxonomies already available in the literature on cohesion (cf. Halliday and Hasan 1976, Carstens 1997) in the definiens slot.
On a semantic level it does not really matter whether these conjunctions are coordinating or subordinating, but the internal polysemic nature of conjunctions (cf. Messerschmidt & Messerschmidt 2007) should receive systematic treatment in the dictionary.
The researchers will use insights gleaned from more than one linguistic approach or theory which among others will include structural linguistics, functional systemic grammar and cognitive linguistics - as presented by Halliday and Hasan (1976), Ponelis (1979), Langacker (1987), Matthiessen and Thompson (1988), Feinauer (1990), Gouws (1992), Bosch (1997), Bosch (1998), Taylor (2003), Verhagen (2001). This method of work is admittedly somewhat eclectic but, as such, not unusual within the lexicographic tradition especially where linguistic concepts are involved. In this manner we hope to clarify the linguistic foundations of our proposed treatment of conjunctions - which is more than what appears to be currently the case.
Limitations in the treatment of conjunctions in the three explanatory dictionaries WAT, HAT and Pharos Verklarende Afrikaanse Woordeboek and in one learner´s dictionary, Basiswoordeboek van Afrikaans, will be critically analysed.
Certain general principles and guidelines for treating conjunctions - as a separate and unique word class - will be proposed. These principles will be illustrated by discussing the current treatment of the Afrikaans conjunctions dat, as and of in some detail and by suggesting alternative treatments of these three conjunctions.
Based on the analysis done so far it is clear that the current treatment of conjunctions is far from satisfactory. In the three dictionaries, HAT, Pharos VAW and WAT the lemmata voegwoord, (voegwoord is of course not yet in the WAT), neweskikker and onderskikker appear as lemmata, but are not used to categorise and subcategorise the conjunctions in these dictionaries. The definitions and grammatical guidance provided by the lemmata neweskikker and onderskikker are, moreover, far from adequate. The terms voegende or voegwoordelike bywoord do not appear as lemmata.
A first recommendation would be that the lemmata voegwoord, verbindingswoord, neweskikker, onderskikker and voegende bywoord provide more in-depth syntactic information with enough examples (also across sentence boundaries). There should be cross references from the specific conjunction lemmata to these lemmata. The examples provided should indicate whether only hypotactic binding is possible or whether incorporation is also possible.
The part of speech can then be indicated as such:
Voegwoord; (then subcategory): neweskikker, etc.
By way of example and to illustrate our preliminary findings, we offer our dictionary article for dat in a learner´s dictionary.