There is very little knowledge of where the words Se-Tswana and Bo-Tswana come from, only speculation exists:
The name Bechuana seems derived from the word chuana – alike, or equal – with the personal pronoun Ba (they) prefixed, and therefore means fellows or equals. Some have supposed the name to have arisen from some mistake of some traveller, who, on asking individuals of this nation concerning the tribes living beyond them, received the answer, Bachuana, “they (are) alike;” meaning “They are the same as we are,” and that this nameless traveller, who never wrote a word about them, managed to engraft his mistake as a generic term on a nation extending from Orange River to 18° south latitude…. The Bechuanas alone use the term to themselves as a generic. (Livingstone 1857:200/201)
Whether we accept Livingstone’s explanation of the etymology of the Tswana (as coming from tshwana, in current Setswana orthography) stem is not of major significance to our enquiry at this point (see Ramoshoana in Cole 1950:xx-xxii for criticism of this view). But many historians and grammarians agree that all the Tswana tribes identify themselves as members of a larger national unit – that of the Tswana cluster (Cole 1950:xx). They therefore call themselves Batswana and their language Setswana. Here I differ with Andersson and Janson in their claim that:
The idea of Setswana as one unified language, with one written form and only dialectal variation between the spoken forms is fairly recent… this was hardly the general view fifty years ago: at that time, a Mokwena might well argue that his language was Sekwena, not Setswana (1997:26).
There is enough compelling evidence dating back to the time before the missionaries codified the language, that the term Setswana included various “Setswana” dialects. This is adequately captured in the grammar books that I have quoted above, some travels records of explorers, and it is in no way a development of the past fifty years as Andersson and Janson would have us believe. But I would like to provide additional evidence that the concept of Setswana language is not as recent as it is claimed.
As far back as November 1806 the German, Hinrich Lichtenstein in Ueber der Beetjuans, published in 1807, which was later translated into English (see Lichtenstein, 1973:63), considered the Batswana a single linguistic group and wrote of ‘Beetjuana words’. He also lists in Upon the Language of the Beetjuans (1815:478-488) a vocabulary of The Beetjuan Language. And around the same time, Henry Salt (1814: appendix, xxvii) records A few words of the Mutshuana language copied from a manuscript journal of Mr Cowan. These words included, ‘sun’ let chãchi (letsatsi in current orthography); ‘moon’, werri (ngwedi); ‘much’ too na (tona) and ‘morning’ kom mo shu (kamoso). Campbell (1815:221) also lists Bootchuana Words in his Travels. Schapera records that,
The people among whom Moffat laboured at Kuruman were called Batlhaping (sing. Motlhaping). They were the southern-most tribe of the Bantu-speaking group collectively known as Batswana (usually written Bechuana or Bechwana) and the first group to come into contact with white people (Schapera, 1951:xv).
In 1857, Livingstone writes that Moffat had just completed translating ‘the Bible in the language of the Bechuanans, which is called Sichuana’. Lichtenstein (1930:407) also notes that,
Under the name of Betjuana, Sihtjuana, Muhtjuana, are to be included all the tribes that inhabit the country that extends from the river Kuruhman, as its most southern boundary thirty or forty days journey northwards; several tribes inhabiting this latitude extend quite to the eastern coast of Africa.
The weight of evidence that illustrates that Setswana has been regarded as a single language is weighty. The fact that a Mokwena might argue that he is speaking Sekwena and not Setswana as Andersson and Janson (1997) argue is a rather unconvincing way of attempting to cast doubt on the commonly known linguistic facts of this linguistic group. It may still be the case that a Mongwato man may argue that he is speaking Sengwato and not Setswana. Such statements should be interpreted as assertion of one’s identity with their tribe (or smaller group), to propagate and retain group pride. This is a well-known sociolinguistic fact that an attachment to a certain language or dialect may be used to signal group solidarity. But faced with the current data, we can only conclude that the compelling body of evidence points to the fact that Setswana has for a long time, at least two centuries, been the body label that includes the various Setswana dialects.
 Anderson and Janson are however right in alluding to the fact that tribal names are reflected in the names of the languages with the prefix change. Thus Bangwaketse speak Sengwaketse; Bangwato, Sengwato; Bakwena, Sekwena; etc.
This an extract from my Oxford University M.Phil thesis in General Linguistics