Did you know that two witnesses can describe the same historical event differently without either one necessarily lying? A notice may record what officials planned; a diary may record what one person experienced; a newspaper printed days later may summarise the event for readers who were not there. The sentences can conflict on the surface while speaking about different times, places or kinds of evidence.
Comparing historical accounts means examining what each source actually says, who produced it, when and why it was created, and which facts can be checked independently. The immediate answer is to separate observation, intention and interpretation before deciding whether the accounts genuinely contradict each other. Do not settle the disagreement by automatically preferring the older source, the more confident writer or the account that matches a textbook.
Start by writing the precise disputed question
“Which historian is correct?” is often too vague to investigate. A better question names the disputed event and the part that differs: Was a footbridge open to the public at 9:00 a.m. on 6 May? Another question might be Who had access before the official opening? These are not the same question. Students who do not define the question often compare quotations that were never intended to answer the same thing.
A useful historical response also recognises that a source has both value and limits. Personal memories can preserve details unavailable in official announcements, but memories can be mistaken or selective. Official notices may identify intended policy, but they do not by themselves prove that the policy was carried out exactly as written.
A quick learner diagnostic before teaching
Give the pupil two statements: “The town announced that the walkway would open on 6 May,” and “When I visited at 9:00 a.m. on 6 May, the gate was closed.” Ask: “Must one statement be false?” If the answer is yes, ask what time the announced opening was supposed to occur. The first statement describes a plan; the second reports an observation at a specified time. Both might be accurate.
- Can the learner identify each speaker and the source type?
- Can the learner distinguish a future plan from a past observation?
- Can the learner identify the exact time, location or group covered by the statement?
- Can the learner quote only the words needed to support an inference?
- Can the learner name one additional source that could help, without inventing it?
Worked inquiry: the fictional Willow Quay footbridge
Everything in this worked case is invented solely for education. Willow Quay, its records, notice, resident and bridge-opening incident are fictional; the quotations are not archived historical documents. The exercise teaches comparison, not a claim about real history.
Source A, fictional municipal notice prepared on 4 May: “The Willow Quay footbridge is scheduled to open for public use on 6 May.” This states a schedule. It does not specify the hour and is not a direct observation of public access on the day.
Source B, fictional resident diary dated 6 May: “At nine this morning I reached the northern gate. It remained locked, so I returned by the road.” This describes one person’s reported experience at a particular gate and approximate time. It does not say whether the southern entrance was open, whether the bridge opened later, or why the gate was locked.
Step 1: Identify scope. A is about a scheduled public opening on a date; B is about a locked northern gate at about 9:00 a.m. on that same date. Step 2: Check the contradiction. A does not promise a 9:00 a.m. opening, so the two statements are not logically incompatible. Step 3: State the bounded conclusion. In this invented case, the sources support that an opening was planned for 6 May and that one fictional diarist reported the northern gate locked around 9:00 a.m.; they do not establish the actual opening time.
Step 4: Specify what further evidence could resolve the question. A dated opening ceremony programme, a gatekeeper’s signed shift log or an independent photograph with trustworthy time and location information might help. Their mere possible existence is not evidence. The learner should write “seek a record” rather than pretending to have found one.
What if the accounts genuinely contradict one another?
Now add Source C, another invented exercise item: “The northern gate was open to everybody at 9:00 a.m. on 6 May.” Unlike A, this asserts actual access at the same place and time as B. If “open to everybody” and “locked” are used in their ordinary senses, C and B conflict on a shared claim. You cannot make them agree merely by paraphrasing them more gently.
Even then, conflict does not tell us who is correct. Ask how closely each writer could observe the gate, whether C was an eyewitness or repeating a report, whether the text was changed after publication, and whether either source refers to a different year or entrance. If independent records remain unavailable, the correct answer may be “the accounts conflict and available evidence does not resolve them”. That is a reasoned conclusion, not a failed assignment.
A small source-comparison table students can reproduce
- Claim: write one narrow question, including time and place where appropriate.
- Creator and date: record who made each source and when; distinguish event date from document date.
- Exact statement: quote only the relevant part, preserving qualifiers such as “planned”, “about”, “reported” and “may”.
- Direct knowledge: mark eyewitness observation, official intention, later recollection or second-hand summary where the source actually supports that classification.
- Supported conclusion: distinguish agreement, unresolved disagreement and genuine logical inconsistency.
- Next check: name missing evidence and how a researcher might find it.
Why eyewitness status alone is not a truth machine
An eyewitness can be strongly placed to describe what they saw, but one viewpoint is limited in time and space. A record written at the scene can still contain a wrong clock reading, a mistaken identity or an unnoticed obstruction. Conversely, a later historian may compare several independently preserved records and offer a better-supported reconstruction than any one witness could provide alone. Evaluate evidence and method, not prestige or proximity alone.
Likewise, two websites quoting the same eyewitness do not create two independent witnesses. Trace the quotation back to its origin. This avoids false corroboration, where repeated versions of one account appear to form a consensus. The question becomes not just “How many pages say it?” but “How many genuinely distinct evidentiary paths support it?”
Guided practice: add location to a conflicting account
Invented example: One notice says “the library reopened on Thursday.” A memoir says “I found the reading room closed that Thursday.” What clarification could matter? The first might refer to the building’s lending counter, while the second reports the separate reading room. It is possible both are accurate; it is also possible that a promised reopening did not occur. Write both possibilities as hypotheses to investigate, not new facts.
The teacher can then supply a hypothetical follow-up document specifying which rooms reopened. The pupil should adjust the interpretation when new evidence arrives. This practises revision of belief rather than trying to win a debate.
Independent challenge: a school exhibition
Two fictional sources describe the same exhibition. An invitation says, “Visitors may enter from 10:00 a.m.” A student journal says, “At 10:15 a.m. the door was still closed to our class.” Write a four-sentence historical comparison stating the shared subject, the difference between promised access and reported access, the limit of each source, and an additional record worth checking.
Answer guide: Both sources concern the exhibition, but the invitation sets an intended admission time and the journal describes one class’s reported experience fifteen minutes later. The invitation does not prove that all guests entered promptly, and the journal does not prove that no guest entered. A staffed entry log, contemporaneous organiser notice or independent photograph might clarify the conditions. Do not invent their contents.
Student and parent follow-through
A parent helping with history homework can ask, “Which words make you say that?” and “Is the source showing an intention or an event?” Avoid providing a dramatic explanation before the child has examined the document. A teacher can place an eyewitness text, an official notice and a later summary beside one another, asking students to colour-code claims, observations and uncertainty before drafting a paragraph.
A helpful progression is to start with a single statement, then compare two statements at the same time and place, and only later tackle accounts whose creation dates, audiences and purposes differ. Students learn more when the cognitive difficulty increases deliberately.
How to measure learning progress
- First attempt: the pupil identifies the disputed proposition and states what each source directly supports.
- After modelling: the pupil distinguishes date of composition from date of the described event.
- Next lesson: the pupil compares a new pair without being prompted to ask who, when, where and for whom.
- Independent transfer: the pupil gives a cautious explanation, describes unresolved conflict accurately and proposes a realistic verification step.
Useful limits: source availability, translation and memory
Translation can affect nuance: words such as “might”, “shall”, “opened” and “scheduled” do not always map neatly across languages. A transcription can also omit damaged or uncertain words. Students should use reliable editions, record when the original was not directly inspected, and be careful about historical categories they may be importing from today.
Not all historical questions can be settled to certainty. Evidence can be lost, inaccessible, conflicting or produced for a biased purpose. A responsible conclusion calibrates confidence rather than manufacturing a smooth story. When writing about actual events, replace teaching examples with verifiable documents and cite them properly.
What primary-source guidance says
The Library of Congress teaching guidance encourages students to observe, question, compare, consider context and test assumptions using more sources. The US National Archives introduction to primary sources explains how first-hand records reveal perspectives as well as the need for interpretation. These are general inquiry principles, not proof of any fictional Willow Quay detail.
Frequently asked questions
Should I trust the source written closest to the event?
Often it has valuable proximity, but still examine the creator’s access, purpose, accuracy and whether the document is authentic. Closeness alone does not settle a disagreement.
What if two sources say opposite things?
First compare the exact proposition and circumstances. If the same proposition genuinely conflicts, describe that conflict and seek independent evidence rather than deciding by confidence or writing style.
Can I write “we cannot know”?
Sometimes, but be precise. Explain exactly what the available documents do establish and which part remains unresolved. “Not established by these sources” is usually more informative.
Is a later history textbook a primary source?
Normally a textbook is a secondary source for earlier events, though it can itself be a primary source when investigating the time and society in which that textbook was created.
A better answer than choosing a winner too early
The core aim of studying conflicting accounts is to replace quick judgement with disciplined reading. Students who can separate intentions from observations, quote accurately, seek independent evidence and admit uncertainty are learning a skill that matters far beyond history lessons.
Continue with how timelines can be mistaken for proof of causation and how captions can go beyond photographic evidence. For broader educational pathways, consult the eduKate Education Hub; the humanities learning guide; and the immutable teaching reference on diagnosis, explanation and independent checking.