Two Sisters, One Train: Editing Partition Testimony
The editor of a school magazine in Ludhiana has one week to publish a Partition 1947 feature. Two sisters from the same village near Sialkot were interviewed separately. The elder, aged 92, says their train was attacked near Gujranwala and that 'forty people were killed'. The younger, aged 88, says the attack was near Wazirabad and that 'many were killed, maybe twenty'. Both agree they left at night in August 1947 and that their father was separated from them for two days. A photocopy of a 1950 relief committee list records 27 persons from their village as 'missing'. The magazine has no access to archives before the deadline. The editor must decide what numbers, if any, to print and how to present the disagreement without implying that one sister is lying.
1. What can and cannot be concluded from the three sources?
Read the worked answer
All three are evidence, but they answer different questions. The interviews establish that these two women left their village in August 1947 at night and that their father was separated for two days; that is a shared memory at the level of sequence. The sisters are not independent witnesses, so their agreement is one family tradition, not two. They disagree on the place (Gujranwala or Wazirabad) and the number (forty or about twenty), which shows that place and number are the fragile parts of the memory.
The 1950 list establishes that 27 people from the village were recorded as 'missing' three years later by a committee; it cannot establish that 27 died, because 'missing' may cover separation, migration to Pakistan, or death, and the committee had its own purposes and categories. No source gives a death toll.
2. How should the magazine present the death toll?
Read the worked answer
Do not print a single figure. Write something like: 'The sisters remember an attack on their train. One recalls about forty dead, the other about twenty; neither saw a count. A 1950 relief committee list records 27 people from their village as missing, which is not the same as dead.' This gives the reader the disagreement and its reason, and it lets the witnesses be truthful rather than consistent.
Keep the numbers in the same paragraph as their source labels, so that no figure floats free as an established fact. If space allows, add a sentence that memory for place names often shifts while memory for the fact of loss persists; that is a historical observation, not a disclaimer. The editorial rule should be: report what each source says, and report what it does not say.
3. Plan two follow-up checks that could sharpen the account after publication.
Read the worked answer
First, re-interview each sister separately with a simple map of the Sialkot-Lahore-Amritsar rail lines, asking which station they had passed and which they had not yet reached. This tests place memory without pushing them to agree. Second, obtain the header page of the 1950 relief list, because the definition of 'missing' and the column headings will show what the committee was actually counting.
A third check, if time allows, is a district gazetteer or a published account of an attack on the same line, which can show whether either location is independently recorded. The aim is not to choose the 'right' sister but to locate which parts of the account are corroborated and which remain single-source.