In the past two months my opinion has been sort by at least three Members of Parliament and at least five prominent members of the Botswana society on one specific matter. The matter of interest is: Is Radio Botswana right in saying molaomothao instead of molaomotheo in their news? At first I was bemused by the persistent questioning from different corners on a matter which in the first place I had never encountered on the Radio Botswana news. Language questions of what is right usually lean on prescriptivism. In other words, they require a linguist to be a referee or a judge of linguistic propriety. They therefore must be handled with supreme care. We must admit that we have multiple dialects of the Setswana language spoken by various Batswana groups in Botswana, Namibia, South Africa and Zimbabwe. The question of which dialect is right therefore can be incredibly subjective and can arouse dialectal defences from speakers of various dialects. Modern linguists in general bother minimally about what is right or wrong. Their principal occupation is descriptive. This descriptivism is usually based on language usage and linguistic patterns that hold in a language. For instance, English plurals are formed by the attachment of the plural morpheme [s]. This is not a law which language must obey. Instead, it is a description of English plural regularity. Needless to say there are exceptions to this rule as in words such as ox > oxen, fish > fish, man > men, phenomenon > phenomena. The question of whether it is right to say molaomotheo or molaomothao can only be fairly and accurately addressed on the basis of linguistic description and theory and not on dialectal bias and individual preferences. In this column I will therefore address this question on the basis of linguistic evidence, linguistic patterning and a bit of corpus evidence.
Let us start with corpus evidence. A corpus is a large collection of texts designed to be used for linguistic investigation. It is a collection of naturally occurring text which is large enough to be queried with fairly representative results. The corpus (body of text) is also diverse in its composition such that it could be considered representative of the speech community from which it was abstracted. It contains novels, plays, newspaper text, business text, sports, Christian literature, news reports, spoken text and hundreds of other varieties of text from diverse sources. The corpus we use is about 19 million words. The test is a simple check of whether molaomotheo and/or molaomothao exist. If they do exist, we wish to check their frequency of occurrence. Additionally, we wish to test if the nouns theo and thao exist in the corpus as well. This is a fairly straightforward test conducted using lexical software. Our findings are that molaomotheo occurs 34 times in the multi-million database and molaomothao does not exist. The structure theo appears 126 times. It occurs as part of formations such as molao-theo, tlhatlhobo-theo, dipopego-theo, lenaaneng-theo, theo (establishment), including the personal name Theo. On the other hand thao does not exist in the corpus. That’s basic corpus evidence.
The next, perhaps more compelling evidence, is to consider the Setswana morphology, in other words rules that govern how words are structured. In my discussion I will try and avoid as much linguistic jargon as possible and yet aim to include as many examples as to make matters clear. The word molaomotheo or molaomothao is what linguists call a compound word, that is, it is a word that is made of two words: molao and theo or molao and thao depending on which side of the debate you stand. The first part of the compound word, molao is not a subject of debate. The debate lies in whether we say theo or thao. To answer this question properly we must first accept that the name/noun theo/thao comes from a Setswana verb thaya ‘to start, begin or set a foundation’. Theo/thao is therefore derived through a linguistic process from thaya. To determine which one is right we need to observe linguistic evidence to see how Setswana works. In particular, we need to consider verbs which when derive nouns which end with [-eo] or [-ao], just as the current debate is whether we have theo or thao. Let us consider the evidence in favour of an [-eo] ending: baya > peo not pao; tsaya > (mai)tseo not tsao; thaya > theo not thao; naya > neo not nao; itaya > kiteo not kitao; tlhatlaya > (di)tlhatleo > not tlhatlao; apaya > kapeo not kapao; raya > theo not thao. There is also evidence that nouns that end in [-ao] do exist. For example: laya > tao not teo; kaya > kao not keo; bolaya > polao not poleo; otlhaya > kotlhao not kotlheo. What is obvious is that all the verbs above end in [-aya]. For the nouns that end with [-eo], the [-aya] in the verb form is preceded by one of the following consonants [b, ts, th, n, t, r, tl, p]. For the nouns that end with [-ao], the [-aya] in the verb form is preceded by one of the following consonants [l, k, tlh]. The results are therefore attractive since they demonstrate that there is no overlap in the sounds that precede [-aya] in the determination of whether to end a noun with [-ao] or [-eo]. We must therefore return to the question that has been posed: do we say molaomotheo or molaomothao? On the basis of linguistic evidence and rules the answer is we say molaomotheo and not molaomothao. This is because theo of molaomotheo is derived from the verb thaya and any very that has [th] before final [-aya] always ends with theo. There are also clear uses of the noun theo in various contexts in Setswana while thao doesn’t exist. For instance, we say theo ya ntlo (the start of house building) or theo ya mafoko (word formation). Confusions similar to the one above are therefore better handled by appealing to linguistic evidence and not to individual preferences. It is my hope that in the future, Radio Botswana would appeal to linguistic expertise to resolve language problems such as these, in particular that the country has a number of capable language experts who could offer invaluable advice.