Did you know that one oral-history interview can be both a firsthand record and a retelling of someone else’s experience? A student hears a grandparent describe an old neighbourhood and writes down every sentence as if the speaker witnessed every event. The interview may be valuable, but the wording quietly erases the difference between ‘I was there’ and ‘someone told me.’ Good history begins by noticing that difference.
The immediate answer: firsthand testimony describes something a speaker personally witnessed, did or experienced; a secondhand account repeats information learned from another person or source. An oral-history interview is a recorded recollection and itself a primary source for what the interviewee said. But individual claims within it may be firsthand memories, secondhand stories, interpretations or uncertain recollections. Attribute each claim to its actual source before presenting it as a historical fact.
- Listen for evidence words: ‘I saw’, ‘I did’, ‘I remember’ versus ‘my neighbour said’, ‘I later read’, ‘people told us’.
- Record the interview date separately from the date of the remembered event.
- Preserve important exact wording and a timestamp or page reference, especially for qualified memories.
- Do not treat a firsthand memory as automatically exact, or a secondhand story as automatically worthless.
- Check important historical claims against other relevant independent records when possible.
How oral history works as evidence
Oral history is the practice of recording recollections through a planned interview so that a person’s voice, perspective and experience can be studied. A survivor may tell us what a flood felt like, how the family adapted and what neighbours discussed. A second speaker may recount a story heard from a parent. Both can tell researchers something important; they answer different questions.
The Smithsonian Institution Archives guide to primary sources explicitly distinguishes first-person testimony from second-person or hearsay testimony and explains that one document can mix both. Its example of a diary containing the writer’s observations alongside stories supplied by a relative captures the same challenge. ‘Primary source’ describes a document’s relationship to a research question; it is not a stamp saying that every claim inside it is firsthand or true.
The Smithsonian guide on conducting oral histories recommends seeking the interviewee’s personal recollections and preserving the full interview rather than only an interviewer’s interpretation. This matters because a short edited quotation can hide words such as ‘I think’, ‘I was told’ or ‘perhaps’. Those words are not clutter. They tell us how the speaker knows, or believes they know, the story.
Worked example: classifying a fictional neighbourhood interview
Invented classroom example, not an actual historical interview. In 2026 a student records an interview with an adult called Ms Lim about a made-up neighbourhood called Willow Quay. At timestamp 02:14 Ms Lim says: ‘In 1984 I walked to the old market with my father before school.’ At 03:09 she says: ‘My aunt told me the market had first opened in 1952.’ At 04:32 she says: ‘I think the old building was demolished sometime after I moved away.’
We can now classify the statements without altering them. Statement one is Ms Lim’s claimed firsthand recollection of walking to the market in 1984. It is evidence of what she remembers, not independent proof of the date. Statement two is secondhand testimony about 1952 because the information comes from her aunt. It needs further records if the research question asks when the market actually opened. Statement three is a qualified inference or uncertain recollection, not a precise demolition date. The phrase ‘I think’ and the vague ‘sometime after’ must survive any summary.
A careful three-sentence historical note could say: In a fictional 2026 class interview, Ms Lim recalled visiting Willow Quay market before school in 1984 (02:14). She repeated her aunt’s claim that it opened in 1952 (03:09), a date not independently verified in this exercise. She was unsure when the building was demolished (04:32). This note distinguishes memory, hearsay and uncertainty. It does not invent an archive citation or present the imaginary town as Singapore history.
Notice the two timescales. The interview took place in 2026; the remembered visit is dated 1984; the hearsay opening date is 1952. Those are not competing dates for one event. They are dates attached to different acts and claims. Separating them prevents the common mistake of writing ‘2026: market opened’ because a modern recording mentioned an older event.
A diagnostic: how does the speaker know?
Offer three new lines without labels: ‘I helped paint the sign’, ‘My sister says her teacher drew the sign’, and ‘The sign was probably there for everyone.’ Ask the learner to identify firsthand experience, secondhand information and an interpretation. Then ask which can be checked with a contemporary photograph, dated receipt, council record or another independent witness. The right answer includes a willingness to say ‘not yet verified’.
For a stronger diagnostic, ask whether ‘I remember’ makes a date certainly correct. It does not. An eyewitness might misremember a year or blend two similar occasions. Equally, a person might truthfully report that a particular rumour was common at the time; that claim about what people believed is different from claiming the rumour itself was true.
Progressive practice: from listening to cautious historical writing
1. Guided listening to one minute
With permission, use a suitable short sample interview or a teacher-read fictional script. Pause after each sentence. Ask learners to colour-code firsthand observation, information learned from others and personal interpretation. Explain the speaker’s actual words before naming the categories. Let a child finish the last two statements independently.
2. Supported evidence table
Give an invented four-line interview transcript with reference numbers: line 1 ‘I heard the bell’; line 2 ‘My neighbour later told me a bridge had closed’; line 3 ‘The closure must have been a protest’; line 4 ‘I cannot recall the month.’ Have the pupil fill columns for exact claim, how the speaker knows, uncertainty and possible corroborating record. Accept ‘unknown’ when no evidence supports a proposed cause. A parent or teacher can demonstrate line 1 and observe how the learner handles lines 2–4.
3. Independent mini-research task
Ask the learner to use an approved, consented interview or publicly available archival oral history and identify three statements with exact timestamps. They should produce a paragraph that names which claims are eyewitness, which are secondhand and which remain uncertain. For a historical factual claim, require at least one additional relevant source where available, and explain whether it corroborates, challenges or simply discusses the claim. The goal is justified attribution, not a forced conclusion.
Why a firsthand recollection is valuable but not infallible
A remembered smell, sound, feeling or decision can reveal aspects of daily life missing from official statistics. Yet an interview records a person speaking at a later time, influenced by recall and the questions asked. Memory and documentary evidence need not agree neatly. Contradictions can become worthwhile research questions rather than reasons to silence a witness or rewrite the account.
Different questions require different classifications. An interviewee’s statement that ‘people in our neighbourhood believed the bridge was unsafe’ may be firsthand evidence of a remembered belief if the person heard discussions. It is not automatically engineering evidence that the bridge was unsafe. The same recording can be a primary source for studying memories in 2026 and a secondhand source for an older event the speaker never witnessed.
Misconceptions worth correcting
- ‘Recorded interview’ means ‘everything in it is eyewitness evidence’. The document and the individual claims require separate classification.
- ‘Firsthand’ means ‘undoubtedly true’. Experience may be genuine even when dates or details are uncertain.
- ‘Hearsay’ means ‘useless’. A repeated account may reveal the spread of a story, and it may direct the researcher toward other evidence.
- ‘Interview date’ equals ‘event date’. A retrospective interview usually describes earlier events.
- ‘A polished paraphrase’ is always better than the original words. Smoothing away qualifications can make the evidence misleading.
Practical guidance for teachers, parents and young researchers
Before interviewing someone, explain the purpose, obtain appropriate consent and ask what may be quoted or shared. Use respectful open-ended prompts such as ‘What did you see?’ and ‘How did you learn about that?’ rather than questions that supply the expected answer. Distinguish sensitive personal testimony from public records, and never pressure a relative to disclose private events for an assignment.
For progress monitoring, save three versions of a short evidence note: the original unsorted claims, a guided classification and an independently revised paragraph. A useful rubric asks whether a learner (1) identifies the firsthand/secondhand distinction, (2) preserves the speaker’s qualifiers, (3) distinguishes interview and event dates and (4) names what would verify an important claim. Improvement is visible when the pupil applies these tests to a new speaker without prompts.
Limitations and respectful uncertainty
A recorded conversation can include translation, editing, omitted sections, unclear audio and changing memories. Some speakers cannot give precise dates; others use terms or place names that have changed. Different witnesses may disagree sincerely. Preserve provenance and exact references, note the limits of the available material and resist turning ambiguity into a catchy but unsupportable headline. This classroom method is about source criticism, not legal rules on admissibility of hearsay.
Frequently asked questions
Is an oral-history interview always a primary source?
It is primary evidence of the interviewee’s recorded words and perspective. Whether it directly supports a claim about an earlier event depends on what the speaker witnessed and what the research question asks.
What should I do if the speaker says ‘my father told me’?
Keep that attribution. It is a report of the father’s story, not proof that the speaker personally observed the event. Look for additional evidence if you need the event itself established.
Can I correct a date that seems wrong?
Do not silently change the transcript. Preserve what was said and add a clearly labelled editorial note or explanation supported by the record you consulted.
What if two eyewitnesses disagree?
Record both accounts faithfully, ask what each could see or know, and seek independent corroboration. Do not invent certainty merely because a school answer seems to demand one date.
Further education and source-checking
The Core Aim lesson about photographers and subjects shows the parallel skill of separating who created a source from what it depicts. Find other careful research methods in the Core Aim learning hub and eduKate Education Hub. The Clementi teaching reference likewise emphasises diagnosing the problem, explaining the reasoning and checking it independently.
A good oral historian does not dismiss a story or transform it into a certainty. They listen closely, preserve the speaker’s words, ask how the information reached that person and write conclusions only as strong as the evidence allows.