Source Search for a Historical Research Question
The Vorort and the Accession to the Council of Europe in 1963
This exercise was translated automatically from the German original and can contain errors. Consult the original version in case of doubt.
Overview and Didactic Goal
This exercise trains transferable research competencies: from the precise operationalization of a research question, through systematic and unsystematic (snowball) search strategies, to the critical evaluation of sources (provenance, context, bias, relevance). (Putnam 2016) You use generative AI here as a heuristic aid (clarification of terms, expansion of keywords, search paths). (Milligan 2019)
The case study is this research question:
What role did the Swiss Trade and Industry Association (Vorort) play in the decision process on the accession to the Council of Europe in 1963?
The exercise follows a circular understanding of research: search results change the keywords, the hypotheses and the delimitation of the research design.
For work in an archive, note the protection periods, the terms of use and the digital availability of each holding. Plan the workload for requests for access, for consultation on site and for reproduction rights.
At a Glance
| Item | Value |
|---|---|
| Duration | several days, of which about 6 to 8 hours at the screen |
| Level | Advanced |
| Core path | 1 → 2 → 3 → 4 → 6 → 7 → 8 |
| Optional deepening | 5 (secondary literature and reference works) |
| Hand-in (minimum) | Search log: 15 entries. AI log: 3 entries. |
This exercise is not an exercise for one afternoon. You open five portals, you run at least ten documented searches, and you build a shortlist of ten archive units.
Plan the work in three stages:
- Screen work (steps 1 to 6): about 4 to 5 hours, which you can split over two sessions.
- Finding aids and shortlist (step 7): about 2 to 3 hours.
- Documentation (step 8): about 1 hour.
You need no visit to an archive for the hand-in. The shortlist comes from the finding aids.
If you really want to consult the dossiers, plan for additional weeks. A request for access, a shortening of a protection period and an appointment in the reading room all have long lead times. This calendar time belongs to the exercise, but not to the hand-in.
What You Need
Subject knowledge
- A basic understanding of historical research methods
- Basic knowledge of work with generative AI, above all prompting
If prompting is new to you, we recommend that you work through the exercise Prompt Engineering first.
Capabilities of the AI system
The system must be able to do the three following things:
- Search the web and support each finding with an address. You need this in step 3 and in step 5.
- Read a file that you upload, for example a finding aid as a PDF file.
- Deliver an answer in a format that you prescribe, for example as a table with search terms.
Other tools and access
- An account for the library catalogue of your university.
- An account for the archive portals that you use in step 7. The registration is mostly free, but it takes one day.
- A reference manager for the short annotated bibliography.
- One file for your search log and one file for your AI log.
You can do every exercise with AI systems from different makers. What a system can do depends on the product and on your account. Settle the five following points before you start.
Cost
A free account is enough for short text tasks. For some capabilities you need a paid account. This mostly concerns the search on the web, the reading of long files and the built-in reasoning mode. Check first whether your university provides access.
Tools
Check which tools your account unlocks. Can the system search the web? Can it read a file that you upload? Can it call external services? Each exercise names the capabilities that it requires.
The section Check and classify capabilities explains five newer capabilities. It also describes what each capability means for your work with sources.
Memory
Some systems carry over statements from earlier sessions or from a project context. Check the settings for memory, projects and chat history. A new session therefore guarantees no empty context.
Account: institutional or private
Check which account you work with. The rules differ in three points.
| Point | Account of your institution | Private account |
|---|---|---|
| Contract | Your university concludes it and shares the liability. | You conclude it yourself and carry the liability alone. |
| Training | A contract often excludes the use of your inputs. | Many providers use inputs for training. |
| Permitted material | The university rules which data you can enter. | You have no such rule and carry the risk yourself. |
Ask the IT office or the library of your university when you cannot find the rules. Record the answer in your AI log.
Do not upload material that you have no right to pass on. An upload leaves your computer and you cannot take it back.
These materials stay outside the AI system:
- Personal data of living persons, for example from interviews, from private papers or from holdings under a protection period.
- Digitized items whose terms of use forbid a disclosure.
- Unpublished texts of other persons, for example manuscripts or reviews.
- Licensed full texts, when the license of your library forbids a disclosure.
In case of doubt, work with a short excerpt or with an anonymized version. You can also use a local model, because the material then stays on your computer.
Law: data protection and disclosure
Two areas of law concern your work directly. This section gives an orientation and no legal advice.
Data protection. In the EU the General Data Protection Regulation applies. In Switzerland the revised Data Protection Act applies (Verordnung (EU) 2016/679 Zum Schutz Natürlicher Personen Bei Der Verarbeitung Personenbezogener Daten (Datenschutz-Grundverordnung) 2016, Bundesgesetz Über Den Datenschutz (DSG) 2020). Both laws protect personal data of living persons. The text of the law itself is binding. Article 6 GDPR requires a legal basis for every processing operation, and an upload is a processing operation.
The role of the provider decides the classification. A provider can act as a processor and process your data only on your behalf. Article 4 point 10 GDPR expressly excludes a processor from the third parties. This role needs a contract under Article 28 GDPR. Without such a contract, a provider uses your inputs for its own purposes. An upload is therefore not always a disclosure to a third party, but without a contract you must assume that it is. Check the contract that applies to your account.
Research is subject to safeguards, but it gets no exemption. Article 89(1) GDPR requires appropriate safeguards, for example data minimisation and pseudonymisation. Article 89(2) GDPR permits the legislator to derogate from single rights of the data subject. The information of the data subjects can also fall away, namely under Article 14(5)(b) GDPR. That applies only when the information proves impossible or would involve a disproportionate effort. Research therefore does not release you from the information duty in general. Settle your individual case with the data protection office of your university.
The European Data Protection Board presented a guideline on research in 2026 (European Data Protection Board 2026). This guideline was in public consultation until June 2026. It is therefore not applicable law, but an aid to interpretation. For deceased persons, the protection periods and the rules of use of the archive apply.
Disclosure. The AI Act of the EU governs the transparency of generated content (Verordnung (EU) 2024/1689 Zur Festlegung Harmonisierter Vorschriften Für Künstliche Intelligenz (Verordnung Über Künstliche Intelligenz) 2024). Article 50 applies from 2 August 2026. It separates two addressees: the provider of an AI system and the deployer. You are a deployer as soon as you use an AI system professionally. For purely private use you are not a deployer, under Article 3 point 4.
As a deployer you carry two separate duties from Article 50(4).
- Deep fake. If an AI system generates or manipulates an image, an audio recording or a video, and the content is a deep fake, then you disclose that. Under Article 3 point 60, a deep fake resembles existing persons, places or events and falsely appears authentic. A generated image without this resemblance to reality does not fall under this duty.
- Text on public matters. If you publish generated text to inform the public on a matter of public interest, then you disclose that. The duty falls away when a person reviewed the text and holds the editorial responsibility for the publication.
Article 50(4) therefore requires no disclosure for every generated image and for every generated text. Article 50(2) addresses the provider instead. The provider must mark generated content in a machine-readable format. This duty does not lie with you, and it does not replace your own disclosure.
The legal disclosure and the academic disclosure are two separate duties. Your university requires a declaration even when the law requires no disclosure. The exercise Citing shows how you declare your AI use.
Record your answers to these five points. They belong to the description of your work environment.
Learning Objectives
After the exercise you can:
- tell source types apart (primary, secondary, tertiary) and give reasons for their epistemic function,
- translate a research question into sub-questions that you can search and into search indicators,
- use the systematic search and the snowball search in combination, and document both,
- develop multilingual search strategies (DE/FR/EN, including historical terminology),
- use library catalogues, bibliographic databases, finding aids and portals for a purpose,
- document findings reproducibly and evaluate them with source criticism,
- treat answers as heuristic suggestions, check them systematically, and reflect on them.
Structure of the Exercise
- Operationalize the research question
- Source types and the logic of evidence
- Systematic search I: a context frame from tertiary sources
- Systematic search II: catalogues, databases, official publications
- Systematic search: secondary literature and reference works
- Unsystematic search: snowballing
- Archive search: finding aids, holdings, access
- Iteration, reproducibility and the AI log
The evaluation of the documents with source criticism is not part of this exercise. You practise that step in the exercise Source Criticism.
1. Operationalize the Research Question
Goal
You break the research question into sub-questions. From them come concrete search paths and sub-hypotheses that you can check.
Task (Without AI)
Sketch a working model of the decision process in 10 minutes (actors, arenas, phases, document types).
Guiding questions:
- Which state arenas matter (Federal Council, departments, parliament, consultation procedure, diplomacy)?
- Where could the Vorort have intervened (position papers, contacts, media, expert commissions)?
- Which analytical meaning do agenda setting, expertise, lobbying, coalition building, legitimation and gatekeeping carry?
Task (AI Exploration)
Break the research question into 4–6 sub-questions that I can check against sources.
For each sub-question, state which document types could serve as evidence.Critical Comparison
Mark in the answer:
- the statements that you can check against the vague statements,
- the missing arenas (for example the consultation procedure, or the federal departments),
- the implicit assumptions in the terms (for example “lobbying” as a potentially anachronistic category).
2. Source Types and the Logic of Evidence
Goal
Clarify which evidence can support which claims, and which systematic limits exist.
Task: Matrix of Source Types (Without AI)
Build a matrix with:
- rows: the sub-questions and sub-hypotheses
- columns: the source types
Source types:
- Primary: association minutes, correspondence, internal reports, submissions, memoranda; government records; parliamentary minutes; media reports
- Secondary: research literature on foreign policy, European policy and the policy of associations; organizational history of the Vorort and economiesuisse
- Tertiary: catalogues, bibliographies, dossiers, registers, databases, finding aids
For each cell: What would count as convincing evidence? and Which biases are likely?
Task (AI Exploration)
Name suitable primary, secondary and tertiary sources for the research question.
Assign each source to an arena (Federal Council / department / parliament / association / media / international).Reflection
Where does an AI typically “invent” sources, for example minutes or letters that look concrete but carry no record? Note three fields of risk.
3. Systematic Search I: A Context Frame from Tertiary Sources
Goal
Build an orienting frame of reference before the deep search starts.
Core Resources
- Dodis: Switzerland and the Council of Europe
- Dodis: Accession to the Council of Europe 1963
- Année Politique Suisse: Council of Europe
- Official Publications of Switzerland (Federal Archives)
- chgov.bar.admin.ch (Federal Archives)
- e-newspaperarchives.ch (ETH Library Zurich)
- Archives Online (State Archives of Switzerland)
- e-periodica.ch (ETH Library Zurich)
- IMPRESSO Project
- Memobase
- swisscollections.ch
Task (AI as a Search Assistant, Not as a Source)
Build a list of 15 search locations (portals, catalogues, editions, archives, databases)
that could matter for the question about the Vorort and the accession to the Council of Europe in 1963.
For each search location, state 1–2 typical types of hit (for example a message, a minute, a dossier, a newspaper article).Validation
- Open at least 5 of the named search locations and test 2 queries in each.
- Document the null results (wrong arena, wrong term, wrong period, missing digitization).
4. Systematic Search II: Catalogues, Databases, Official Publications
Goal
Develop search terms as hypotheses that you can test, including historical terminology and variants of names.
Task A: Fields of Terms
Build three fields of terms (DE/FR/EN):
- Actor: Vorort; Schweizerischer Handels- und Industrie-Verein; SHIV; (FR) Union suisse du commerce et de l’industrie; economiesuisse (mark this as a projection backwards)
- Event and institution: Europarat; Council of Europe; accession and membership; 1963; the European Convention on Human Rights and its court (as context paths)
- Process and mechanisms: consultation procedure; interest association; submission; statement; expert opinion; economic interests; foreign economic policy
Task B (with AI): Controlled Lists of Synonyms
Build 12–20 search terms for each of the three fields of terms (actor / institution and event / process).
Conditions:
- Give DE, FR and EN separately.
- Take historical spellings and abbreviations into account.
- For each term, state where it appears more often (parliamentary minute / internal records / press).
- No invented proper names or archive references.Quality Check
- Remove the anachronistic terms, or mark them as search anchors in hindsight.
- Add anchors for persons and places (for example the Vorort presidency, delegates; Bern and Zurich; the relevant departments).
5. Systematic Search: Secondary Literature and Reference Works
Goal
Build a reliable starting base of secondary literature, including bibliographies.
Task A: Catalogue Strategy
Search with combinations such as:
- (Vorort OR SHIV) AND (Europarat OR Council of Europe) AND 196*
- (Verbände OR Interessenpolitik) AND (Aussenpolitik OR Europapolitik) AND Schweiz AND 1950–1960
- Schweiz AND Europarat AND (Beitritt OR membership OR accession)
Document:
- the search terms
- the filters (year, language, material type)
- the numbers of hits
- the selection criteria
Task B (with AI): Optimize the Search Terms
Replace the example lines in the following prompt with your own search terms and hit problems.
Here are my search terms so far and my hit problems:
- "Vorort AND Europarat": 3 hits, all after 1990.
- "SHIV AND Beitritt": 0 hits.
- "Interessenverband AND Europapolitik": over 800 hits, hardly any relevant.
Optimize the search terms for catalogues (Boolean, phrases, truncation).
Give 6 variants and explain briefly which problem each variant addresses.Deliverable
- A short annotated bibliography of 8 to 12 titles with:
- the relevance to the question (1 to 2 sentences)
- the expected source value (which primary sources does the title cite?)
6. Unsystematic Search: Snowballing
Goal
Start from a small corpus and reconstruct debates and references to sources.
Task: Define the Starting Set
Choose:
- 1 dossier or edition (as a navigator),
- 1 overview account,
- 1 in-depth study (close to an actor or to an institution).
Search Log, Section on Snowballing
- the central references (10 to 20),
- the primary sources that the texts cite (archives, references, document types),
- the actors and commissions that appear again and again.
Task (AI Exploration)
Replace the two example entries in the following prompt with your own eight entries.
I have the following 8 literature or source references:
1. Altermatt, Bernhard and Gilbert Casasus (eds.): 50 Jahre Engagement der Schweiz im Europarat 1963–2013. Zurich 2013.
2. Dodis document 5020: memorandum Petitpierre, Bern, 7.7.1949.
3. …
Order them by (a) primary or secondary, (b) arena, (c) potential value for
the question about the role of the Vorort.
Suggest one follow-up search for each.Reflection
Snowballing reproduces canons and blind spots. Note the missing perspectives (for example trade unions, parties, international actors) and possible counter-strategies.
7. Archive Search: Finding Aids, Holdings, Access
Goal
Identify primary sources so that you can order and analyse them.
Starting Points (Orientation)
- The holdings of the Vorort and economiesuisse at the Archives of Contemporary History (ETH Zurich)
- Records in catalogues (for example SWA/UB Wirtschaft Basel)
- Documents on the Vorort in contexts of diplomatic history (for example Dodis)
“Dossier found” is not the same as “evidence secured”. Finding aids are tertiary instruments of access: they structure visibility, and not truth.
Task A: Analysis of the Finding Aids
Reconstruct for one archive:
- the types of series and dossiers,
- the logic of cataloguing,
- the restrictions on access.
Task B: Shortlist of Dossiers (1960–1963)
Build a shortlist of 10 units with:
- the reference or permalink,
- the title and the time span,
- the expected relation to a sub-hypothesis,
- the access (online / reading room / request).
Task C (with AI): Plan the Search Paths
Suggest three search paths:
1) centred on the actor
2) centred on the process
3) centred on the event
Per path: 8–12 search terms (DE/FR/EN), including variants of names.This exercise leads you to holdings with restrictions on access. Do not upload a digitized item whose rules of use forbid a disclosure. Do not upload records with personal data of living persons either. The section “What you need” names the complete list and the legal basis.
In case of doubt, work with a short excerpt or with a local model. With a local model, the material stays on your computer.
In the archive you meet typescript, carbon copies and handwriting. An AI system can read a digitized item and produce text from it. The error rate of handwritten text recognition depends strongly on the script, the layout and the language (Hodel et al. 2021; Romein et al. 2025).
The most dangerous error is the smooth reading. A general AI system does not stop at an illegible passage. It inserts a plausible word instead (Díez García et al. 2025). Such errors travel into every later analysis (Strien et al. 2020).
Note in the search log which texts a machine recognized. Compare every citation with the digitized item. The section Check and classify capabilities explains this point in detail.
8. Iteration, Reproducibility and the AI Log
Search Log, Section on the Search
At a minimum:
- the date, the tool or portal, the query, the filters, the number of hits
- the selection criteria
- notes on bias (optical character recognition, handwritten text recognition, cataloguing, language)
Questions for Reflection (250 to 400 Words)
- What did the AI concretely accelerate?
- Where did it mislead you (false specificity, anachronisms)?
- Which validation steps did you introduce?
- Which decisions stayed necessarily human?
AI Log (Appendix)
At least 3 entries. Each entry holds:
- the prompt in its exact wording,
- the shortened answer,
- your reasoned correction or your rejection.
Learning Outcome
At the end you hold these artifacts:
- A matrix of sources (sub-questions × source types)
- A search log (15 entries or more)
- A shortlist of 10 units of primary sources
- A short annotated bibliography (8 to 12 titles)
- A reflection on the AI use, with an AI log (3 entries or more)
One methodological insight comes on top: you can give reasons for which search paths deliver reliable hits, and where an AI answer only sounds plausible.
Further Resources
- Ranke.2: Lessons Overview – Learning modules on the web, social media and digital change.
- Knowing Machines: Models All the Way Down – Critique of training data and its parallels to source criticism.
Bibliography
Citation
@inreference{mähr2025,
author = {Mähr, Moritz},
title = {Source {Search} for a {Historical} {Research} {Question}},
booktitle = {Critical AI Literacy for Historians},
date = {2025-12-29},
url = {https://maehr.github.io/critical-ai-literacy-for-historians/en/exercises/source-search-vorort-council-of-europe-1963.html},
langid = {en}
}