Historical evidence: from a surviving record to a defensible claim
A readable document can still be poor evidence for your particular question. This route teaches you to separate what an item says, how it reached you, and what you may infer from it. Finish with a short evidence memo whose uncertainty another reader can inspect.
Course topic: Historical Methods and Historiography
Study route and exercise: University Reading List editorial project. How we curate · Sources checked:
Scope and responsibility
Tosh is the general methods spine; Bloch is a demanding second voice on criticism; From Dust to Digital is a free collection of archive projects. The free book supplies cases about preservation and access, not a complete replacement for an introductory methods textbook. This is a source-based reading route, not a claim to have tested the books in a class.
Choose your starting point
I need a first source-based essay
Choose Tosh first. Use the source and interpretation chapters to turn a large subject into one answerable question; add one archive case only after you can explain what counts as evidence for that question.
I already have sources but trust them too quickly
Pair Tosh with Bloch. Read observation and criticism against your own evidence table. Buy or borrow Bloch for the argument about testimony, not for current digitisation instructions.
I need a free route into digital archives
Start with Kominko’s introductory essay and the metadata case. Practise documenting what survived and what was catalogued. Use library access to Tosh if you also need essay construction and a broad methods introduction.
Compare the three books
The Pursuit of History
John Tosh · Edition: 7 · 2021
Main methods text
- Choose it when
- You need a connected introduction to sources, interpretation and historical knowledge.
- Use another route when
- Your immediate need is a technical guide to scanning or repository software.
- Selected reading
- Seventh edition: chapters 4–7; return to chapter 1 when checking present-day assumptions.
- Limits and trade-offs
- It offers the general framework, but your chosen archive still needs local language, catalogue and contextual expertise.
Try before choosing
With Tosh’s chapters 4–7, turn this guide’s three-item packet into a claim/source/dependence table. Separate the question of authenticity from what an authentic item could establish.
Check your work: The copied notice is not counted as independent implementation evidence. Each row records origin, date, purpose and the narrow proposition it can support.
The Historian's Craft
Marc Bloch · 1964
Critical companion
- Choose it when
- You can bring a specific source and test the difference between testimony and inference.
- Use another route when
- You need a step-by-step modern research workflow before you can identify an author or date.
- Selected reading
- Selected 1964 English edition: the sections on historical observation and historical criticism.
- Limits and trade-offs
- Its unfinished, historically situated argument rewards slow reading; it cannot settle every contemporary question about consent, access or digital reproduction.
Try before choosing
Use Bloch’s discussions of observation and criticism to ask two different questions of a hostile account: what does its author assert, and what might its production reveal unintentionally?
Check your work: Attribute both readings to identifiable features of the document. Hostility neither makes every detail false nor licenses an unsupported reconstruction of motives.
From Dust to Digital: Ten Years of the Endangered Archives Programme
Maja Kominko (editor) · 2015
Free archive case laboratory
- Choose it when
- You want concrete reasons why the available digital collection may differ from the past population.
- Use another route when
- You expect a single beginner textbook covering an entire history degree’s methods.
- Selected reading
- 2015 volume: Kominko’s “Crumb trails, threads and traces”; chapter 2 on Ahom metadata; chapter 16 on a photographic studio.
- Limits and trade-offs
- The projects expose selection and description problems. They are distinct case studies, not representative samples of every archive or a promise that all source images share one reuse licence.
Try before choosing
Compare the Ahom-metadata and photographic-studio cases in the 2015 archive collection. Draft one question each collection could help answer and one missing selection detail that could change the answer.
Check your work: Name the collection and the relevant selection/description process. A digitized subset is not silently treated as a representative population, and image access is not called a blanket reuse licence.
Check your preparation
Do I need to know a historical period first?
A modest amount helps. Define one place, date range and institution, then list unfamiliar terms to check. You do not need an encyclopaedia of the period before making a provisional source inventory.
What should I record before interpreting a scan?
Creator, intended audience, production date, event date, repository identifier, original or copy status, and the route from object to scan to transcription. An unknown field remains unknown; a catalogue label is evidence about description, not automatic proof of the object’s contents.
Does agreement between two items count as corroboration?
Only after checking dependence. A newspaper repeating a committee report adds another publication event, not necessarily another observation. Ask whether sources share a witness, earlier text, institutional interest or later transcription.
Six steps with a concrete output
Outputs marked ready: 0 / 6
These are your own study marks, not an assessment. They are not sent to this site.
-
1. Bound the historical question
Selected reading: Tosh, seventh edition, chapters 1 and 4.
Make this: Write a question about one decision in one institution within a stated date range. Make an inventory of three possible source types and explain one thing each could fail to record.
Check your work: A 120-word question note distinguishes the event from the later records about it.
-
2. Build a provenance ledger
Selected reading: Tosh, chapter 5; Bloch, historical observation.
Make this: Create columns for creator, date of event, date of record, audience, copy status, access chain and missing context. Apply them to the fictional packet below before using real documents.
Check your work: Every item has a production date or an explicit “unknown”; no scan is silently treated as an original.
-
3. Test apparent agreement
Selected reading: Bloch, historical criticism; Tosh, chapter 5.
Make this: Draw arrows for copying and quotation between sources. Write two competing explanations of an apparent contradiction and name the document that would help discriminate between them.
Check your work: You count independent observation chains, not URLs or files; both explanations remain possible until tested.
-
4. Inspect catalogue selection
Selected reading: From Dust to Digital, chapter 2, “Metadata and endangered archives: lessons from the Ahom Manuscripts Project”.
Make this: For this chapter’s described collection, note one issue concerning description, identification or access. State how a search relying on a single catalogue term could miss relevant evidence; label your proposed search strategy as your own.
Check your work: One paragraph separates the chapter’s documented case from your hypothetical search problem.
-
5. Read an image as a made object
Selected reading: From Dust to Digital, chapter 16, “Archiving a Cameroonian photographic studio”; Kominko’s introductory essay.
Make this: Make a checklist for photographer, client, intended use, selection, caption and digitisation. Identify which questions require evidence outside the image rather than answers guessed from appearance.
Check your work: At least three unanswered questions are preserved as research needs, not turned into confident captions.
-
6. Write and audit the claim
Selected reading: Tosh, chapters 6–7; revisit Bloch’s criticism.
Make this: Write a 350-word memo with a narrow claim, source identities, one dependence warning, one plausible alternative and a next archival check. Revise a sentence that previously overstated what the evidence proves.
Check your work: A reader can trace each factual assertion to an item and can tell what would make you change your conclusion.
Worked exercise
Worked evidence memo: was the reading room open on Sundays?
Entirely fictional practice packet: A is a signed committee minute dated 4 June 1890 authorising Sunday opening from 15 June, “subject to staff availability”. B is a newspaper notice dated 6 June that explicitly copies A. C is a clerk’s 1905 handwritten transcription of a now-unlocated July 1890 duty ledger, listing Sunday shifts on 6 and 13 July. A catalogue calls C “Sunday attendance, 1890”. Decide whether these establish regular Sunday public access in June–July 1890.
Assumptions
All people, documents and dates in this packet are invented. Assume the supplied descriptions are accurate but do not assume that a transcription is complete, that scheduled shifts occurred, or that staff presence implies public admission. There are no genuine archival identifiers or quotations to cite.
- Separate claims: A supports an authorised plan with a condition. It does not establish implementation. B supports public announcement and depends on A, so A+B are one chain for the opening decision.
- Trace C: proposed chain is a July 1890 ledger → 1905 transcription → current description. The ledger’s whereabouts, omissions and copying accuracy are unknown. The catalogue’s “attendance” label is broader than the supplied shift entries.
- Check dates and coverage: the proposed start is 15 June; C names only 6 and 13 July. Neither June opening nor every July Sunday follows. Even genuine duty entries concern staffing, not necessarily public use.
- Compare explanations: the room may have admitted readers on those two Sundays; alternatively staff may have worked behind closed doors. Seek access rules, an independently produced visitor register or a contemporary closure notice before choosing between them.
Answer and interpretation
The committee authorised conditional Sunday opening from 15 June, and the proposal was announced. A later transcription reports two July Sunday staff shifts; it offers limited, indirect support for implementation. The packet does not establish regular public access through June–July. A visitor register covering the relevant dates would be the next discriminating source, with its own provenance checked.
Test the result
There are three items but only two described production chains, and only one chain about staffing. Remove B and the evidence for actual access does not weaken. If C were a complete original staffing ledger, its provenance would improve but staffing would still not logically establish admission.
Change an assumption
Revise the conclusion when a counterexample arrives
Keep the original fictional packet. Add D, an independently produced original visitor register: 6 July records nine public readers; 13 July explicitly records “closed for repairs”. For this exercise assume D is authentic and reliable for those two dates, but supplies nothing about June or other Sundays. Does it establish regular public Sunday access throughout the period?
- The 6 July entry now supports actual public use on one specified date, going beyond the earlier indirect staffing evidence. Nine readers is a count for that date, not an average or a representative sample.
- The explicit 13 July closure is a counterexample to “open every Sunday throughout June–July”. It is not merely a blank page being interpreted as closure. Reliability for these two dates is stipulated, not inferred from D being an original.
- June and the other Sundays remain unestablished. Record the dependency of B on A as before; D supplies a new production chain rather than making the copied announcement independent.
Answer and interpretation
Revise to at least one documented Sunday with public readers and one documented closure. The universal every-Sunday claim is false under the exercise assumptions. A weaker claim such as “usually open” still needs a defined denominator and evidence covering more dates.
Test the result
Remove the explicit closure wording and leave only a blank 13 July entry: the universal claim would then be unproven rather than refuted by that blank. Explain why absence and an affirmative negative record have different evidential roles.
Common mistakes
Primary does not mean neutral
A contemporary document can be accurate about one administrative act and silent about lived experience. State the question it can answer before ranking it.
Digital survival is not prevalence
Ten searchable images may reflect a donor, scanning budget or catalogue choice. Do not turn the size of an accessible subset into a population estimate.
Uncertainty is not an excuse to stop
Give a bounded conclusion and a concrete next check. “We can know nothing” discards real distinctions between a plan, a copied notice and evidence of implementation.
Before moving on
- A provenance table keeps event, recording, copying and access dates separate.
- A dependence diagram prevents copied testimony from being counted twice.
- A qualified memo identifies an alternative explanation and the evidence that would distinguish it.
Sources and edition checks
Sources checked:
- Routledge: The Pursuit of History, seventh-edition chapter resources
The publisher’s companion identifies the seventh edition and chapter sequence. Chapter resources are not free access to the whole textbook.
- Vintage / Penguin Random House: The Historian’s Craft
The publisher identifies the 1964 Vintage mass-market paperback, ISBN 9780394705125. This selected reissue year is distinct from the French original and the first English translation; the guide uses named sections rather than unstable page numbers.
- Open Book Publishers: From Dust to Digital, formats and contents
The publisher verifies the 2015 edited volume, paperback ISBN 9781783740628, free online formats and chapter identities. Book-level CC BY-NC-ND terms do not erase separately stated chapter or image rights. The exercise here is original and unrelated to any real collection.
These are proposed study activities, not official course assignments or a claim that every book has been read cover to cover.