Checking Sources in Multilingual Regional Research

Keep original-language sources, translations and interpretation distinct so regional research conclusions can be checked and appropriately qualified.

Checking sources in multilingual regional research comes down to answering three questions for every claim: which passage supports it, who produced the translation being relied on, and what the issuing institution's role does to the meaning of its terms. A statement that cannot be traced to a passage in an identified document, written in an identified language and issued by an identified body, is a lead rather than evidence. The discipline lies in recording those answers at the point of collection, while the document is still open and the uncertainty is still visible.

Locating the passage

Begin with the text itself, not the summary of it. Record the document title, the issuing body, the publication date, the article or paragraph number, and the language and script exactly as published. Where a claim appears in a secondary source, news report or partner briefing, the passage to check is the one in the underlying document, not the sentence that repeats it.

Three distinctions repay attention. First, the difference between operative text and surrounding material: a recital, preamble or ministerial foreword may explain intent without creating an obligation, and citing it as though it were the operative clause misstates the instrument. Second, the difference between an original and a consolidated version: a consolidated text incorporates later amendments and may not match the version a decision-maker actually applied. Third, the difference between a document and a page about a document: a ministry news item announcing a change is not the instrument that makes it.

Quoting the source in its original language alongside any translation is usually worth the space, because it lets a reviewer check the inference rather than trust it.

Classifying the translation

A working answer to "is this official?" needs more than two categories, because the risks differ.

An official translation is published by the issuing authority, or by a body with a mandate to produce authoritative versions. It still requires a date check: the translation may predate amendments to the original, and an official version in one target language says nothing about the reliability of a version in another.

A published but unofficial translation (from a court, an international organisation, a reputable legal publisher or an established research institute) carries weight without carrying authority. Attribute it to its publisher and note that status.

A researcher-produced translation is the normal case in regional work and is legitimate, provided it is labelled as such and the reasoning behind contested terms is visible.

A machine translation is acceptable for orientation and risky as a citation base, particularly where word order, negation, formal register or legal terminology carries the argument.

Confusing these categories is the most common way a sound finding acquires a false footing. A translation being fluent is not evidence that it is accurate, and a translation being official is not evidence that it is current.

Reading the institution

Meaning in regional research is institutional before it is lexical. The same word can describe a binding status in one system and an informal label in another, and the difference is settled by who issued the document and what competence they hold.

Ask whether the issuing body decides, advises, coordinates or reports. A decision by a competent authority, guidance from the same authority, and a statistical release about the authority's caseload are three different kinds of evidence, and a term that appears in all three may not carry the same force in each. Ask what kind of instrument is in hand (statute, regulation, decree, circular, guidance note, press release, minutes) and whether it creates rights, sets expectations or merely describes a situation. Ask, finally, what the term does in that system: whether it names a status with legal effect, an administrative stage, or a descriptive shorthand used by officials.

Where these questions cannot be answered from the document, the honest output is a recorded uncertainty and a named review question, not a confident paraphrase.

A translation note format

The following worksheet is a suggested working format for capturing the reasoning behind a translated term. It is a documentation habit, not a validated instrument, and no institution requires it in this form. Its purpose is to make the chain from original term to analytical claim inspectable by a colleague who does not read the language.

Field What it records Why it matters
Original term The term in its original script, with the sentence it sits in Keeps the evidence anchored to text rather than to a gloss
Context Document type, issuing body, date and the surrounding passage Meaning shifts with instrument and author
Working translation The rendering used in analysis, with alternatives noted Shows the choice was made, not inherited
Authority Whether an official translation exists, and who produced the one used Separates authority from fluency
Ambiguity The plausible readings, stated as a range Preserves uncertainty instead of resolving it silently
Review question The specific question to put to a translator, local specialist or document Converts doubt into a task with an owner
Analytical effect What changes in the conclusion under each reading Connects the language question to the research question

The last field does most of the work. If two plausible readings produce the same analytical conclusion, the ambiguity is worth a footnote. If they produce different conclusions (a different procurement timeline, a different compliance obligation, a different risk rating), the ambiguity belongs in the finding itself.

Worked example (fictional)

The following example is invented to show the format in use; it does not describe a real jurisdiction, document or client.

A fictional regional development agency publishes a notice stating that a named facility holds a status rendered literally as "permission under consideration". Two readings are plausible: that an application has been lodged and remains under review, or that a conditional approval has been granted subject to outstanding conditions. The note might read:

  • Original term: the invented term verostak, quoted with the full sentence from the notice.
  • Context: a notice on the agency's own site, undated in the visible header, referring to a numbered article of a regional statute.
  • Working translation: "permission under consideration", with the alternative "conditional authorisation" recorded.
  • Authority: no official translation located; the rendering is researcher-produced.
  • Ambiguity: the two readings above, distinguished by whether the status follows a decision or precedes one.
  • Review question: does the cited article treat this status as pre-decisional, and does a public register list the facility's position separately?
  • Analytical effect: under the first reading, commissioning is at least one procedural stage away; under the second, operations could begin subject to conditions. The procurement timeline in the draft analysis would move accordingly.

The note does not settle the question. It makes the unresolved question portable, so that a translator or regional specialist can answer it without reconstructing the whole file.

Conflicting versions and unresolved terms

Where two renderings conflict, state both and attribute each. Where a term carries legal or financial consequence, the working translation should be confirmed by someone qualified in that language and that legal system; general research support is not a substitute for that check. Where access is restricted (a paywalled official gazette, a document available only in person), record the constraint rather than substituting a secondary account without saying so.

Document retrieval conditions matter too. Note the date and URL consulted, and keep a copy where copyright and access terms allow, because official pages are revised without notice and a quotation that cannot be re-found is difficult to defend.

Evidence standards across a team

External guidance addresses the wider evidence chain. The UK government's Magenta Book, through its supplement Quality in Qualitative Evaluation, sets out appraisal questions for judging qualitative evidence, including how clear the links are between data, interpretation and conclusions, a test that applies with equal force to the step between an original term and its English rendering. The European Code of Conduct for Research Integrity, published by ALLEA, sets expectations for accuracy in reporting and transparency about methods and data. Neither framework prescribes how to handle a translated source; the worksheet above is a practical response to that gap, offered as a working convention.

Limits

A translation note documents reasoning; it does not certify a language, guarantee accuracy, or substitute for qualified translation and legal review. Coverage varies by language, region and document type, and some terms will remain contested after every available check. The appropriate response is to keep the uncertainty visible in the deliverable and to state which reading the analysis relies on and why.

A practical next step is to adopt the seven-field note for one live workstream and test whether a colleague can follow a translated term from original passage to analytical conclusion without asking you. If they cannot, the note is doing too little. Researchers preparing to take on project-based regional work may find the guidance on what to establish before accepting an expert project a useful companion on scope and evidence expectations.

No advertising or analytics cookies are used. Your preference and application receipt may be stored in this browser; CV files are not.

Read the storage explanation