Three sources, one sequence: weighing evidence for 1904–1908
Three sources sit on the table for a 600-word assessment. Source A is a 1985 interview with a Herero woman, then aged 84, recalling what her grandmother told her about leaving the Waterberg in August 1904: 'the shooting came after the cattle had already been taken.' Source B is a page from a German officer's diary dated 12 August 1904 describing 'a punitive expedition against stock thieves' and never mentioning women, children or deaths. Source C is a mission-station burial register for 1905–1906, damaged by damp, listing 214 entries at a camp near Swakopmund with no names and no causes. The learner has also read that estimates of Herero deaths between 1904 and 1908 vary widely and that the Nama war began in October 1904. The mark scheme rewards judgements about corroboration and about how to read silence in a source. The learner must not simply summarise the three items.
1. What can these three items together establish, and what must remain uncertain? Answer with a clear judgement about corroboration.
Read the worked answer
What holds. The three items converge on a sequence rather than on a number. Source B fixes military action near the Waterberg in the second week of August 1904. Source A, although transmitted orally, describes the same sequence from the inside: cattle taken, then shooting, then flight. Source C independently records heavy mortality in a camp in 1905–1906. Because these are three different kinds of record - a participant's diary, family memory, an institutional register - the agreement between them is worth more than any one alone. A historian can therefore say with confidence that the campaign produced mass displacement, that people were concentrated in camps, and that death rates there were high.
What does not hold. The register has no names and no causes, so it cannot show whether those 214 people were Herero, Nama or others, nor whether they died of violence, hunger or disease. Source A cannot supply dates: 'after the cattle had been taken' is a sequence, not a calendar. Source B cannot establish intent. None of the three is an extermination order, so any claim about deliberate policy must be argued from other evidence. The discipline is to state what the convergence supports and to mark the rest as inference.
2. Source B never mentions civilian deaths. Is that silence evidence that no civilians died, or evidence of something else?
Read the worked answer
The distinction that matters. There is a difference between a silent source and a silence in the source. Source B is not silent; it is loudly partial. It was written the same day by a participant, in a genre - the campaign diary - that organises experience around military categories. 'Punitive expedition against stock thieves' is a description that already excludes civilians: if the operation is defined as action against thieves, there is no slot in the account for women, children or camp deaths.
Why absence is not proof. Silence proves absence only where the source would have been expected to record the thing and had no reason not to. Here the writer had reasons: self-presentation, the conventions of the genre, and possibly awareness that civilians were present. Source C, generated by a different institution for a different purpose, records mortality in exactly the period the diary is silent about. So the silence is better read as evidence about the writer and the records he kept than as evidence about what happened.
Use of the source. Its value is documentary: it shows how the campaign was represented to itself, which is itself part of the history of 1904–1908.
3. How far should a historian rely on Source A for chronology?
Read the worked answer
What oral testimony does well. Source A is strongest on sequence, meaning and experience - the order in which things happened as the family remembered them, and the interpretation placed on them. That is not a lesser kind of evidence; for the experience of dispossession it may be the only evidence there is.
What it does badly. The account has passed through two generations and an interviewer before reaching us. Dates drift; specific numbers disappear; the story is shaped by what later generations needed it to say, and by the political climate of 1985, when Namibian testimony was being collected during intense international campaigning. 'After the cattle had already been taken' is chronologically vague: it can be reconciled with the documented August 1904 sequence, but it cannot establish it.
The judgement. Use Source A for pattern and meaning, not for dates, and only where an independent record fixes the chronology. Where it conflicts with the diary over a date, prefer the contemporary record; where it supplies what the diary omits - the experience of the retreat - it is indispensable. The historian's rule is fitness for purpose, not a ranking of source types.