Quick answer: How do you translate a radiocarbon dating report correctly? Preserve the laboratory’s conventional radiocarbon age, measurement uncertainty, sample identity, pretreatment, calibration curve and probability ranges as separate facts. Translating radiocarbon dates, BP and cal BP values, BCE/CE intervals or archaeological dating results requires careful distinction between measurement, calibration and historical interpretation. This guide explains the first principles, gives fictional worked report excerpts and shows how to avoid claiming a precise calendar date that the evidence does not provide.
The common shortcut is to read ‘2,500 ± 30 BP’ as a plain calendar date 2,500 years before today. For conventional radiocarbon results, BP uses a 1950 reference point, and calibration is needed to relate radiocarbon measurements to calendar chronology. A 95.4% probability range is not the same as an evenly likely period or an exact known year. A sample date may also differ from the date of an archaeological event because the material had its own history.
The IntCal20 calibration family distinguishes Northern Hemisphere atmospheric, Southern Hemisphere atmospheric and marine cases; calibration choices and reservoir corrections must come from the scientific source, not a translator’s intuition. Historical work also uses different date conventions and specialised archaeological terms. This is a language and evidence-navigation guide, not a laboratory calculation service: it does not invent calibrated intervals or replace an archaeological or radiocarbon specialist’s interpretation.
Reader’s map: Begin with the visible error, understand the technical distinctions, follow the twelve-stage process, practise with twenty-five contrasting examples, diagnose predictable failure modes, then test a transfer exercise. This method is designed for study, drafting and careful review.
The visible translation problem
An exhibition label says a charred timber sample has a conventional radiocarbon age and a modelled calendar range. The translation says the building was constructed in one specific year. Three things have disappeared: the measurement uncertainty, the distinction between timber age and construction, and the probability range. The exhibit feels clearer, yet it claims certainty that no method supplied. Better translation keeps the measurement and inference on separate lines.
Where readers may act on the translation, a sensible quality test is not ‘Does it sound fluent?’ but ‘Would a reader make the same permitted choice, at the same time, for the same reason?’
Here is the decision gate: before translating a sentence, identify the action the sentence enables or prohibits. Ask what evidence it relies on, whether it is already effective, who is responsible, which date or quantity constrains it, and what a reader might do differently if one word changed. A useful translation protects the original answer to those questions.
First principles: what the document actually carries
Every technical document has at least two layers: a human-language explanation and a system of relationships that the text describes. A confident translator reads both. The principles below show where an apparently harmless paraphrase can change the operational relationship.
1. Radiocarbon age versus calendar age
A conventional 14C age in BP cannot simply be read as a calendar age in years before the present. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: 2,500 ± 30 radiocarbon years BP is not a calendar conversion instruction. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Keep the 14C label and uncertainty. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
2. 1950 reference
Conventional BP uses a reference year of AD 1950 rather than the current year. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: Before Present in this context is not before this morning or the publication date. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Explain the convention once, precisely. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
3. Measured uncertainty
A result such as ±30 is not decoration; it conveys measurement uncertainty under a specified convention. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: 2,500 ± 30 BP cannot be shortened to exactly 2,500 BP. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Retain symbols, unit and uncertainty. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
4. Calibrated range
Calibration reports may express one or more calendar ranges at specified probability levels. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: 95.4% reported probability interval is not a fixed construction date. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Keep interval, label and probability. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
5. Curve choice
Calibration depends on the relevant dataset and treatment, such as IntCal20, SHCal20 or Marine20. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: Marine shell cannot automatically use a northern atmospheric curve. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Copy curve identifier and method. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
6. Reservoir effects
Aquatic samples can require reservoir considerations and locally relevant corrections. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: An unverified marine correction cannot be invented to narrow a range. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Preserve ΔR and status. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
7. Sample versus event
The material dated may precede, follow or only indirectly relate to the archaeological event. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: Old timber can predate its later use in a building. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Name what was actually sampled. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
8. Uncalibrated versus modelled
Laboratory results, calibration output and Bayesian modelled estimates represent different processing stages. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: Modelled phase range is not simply the original measurement. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Label every stage. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
9. Multiple ranges
Calibration can yield disjoint intervals and uneven probability distributions. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: Two intervals must not be collapsed into a single continuous certainty. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Keep each reported interval. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
10. Comparability and version
Calibration versions, dating conventions and reporting standards affect how results should be read. That distinction matters because readers use a translation to recognise status and plan their next action, not simply to admire polished language. Keep the exact source role of each condition: a description of evidence must remain evidence, while a commitment, recommendation, prohibition or target must stay in its original category.
Small example: Legacy report and newer recalibration are different analyses. A careful rendition names the distinction plainly. Where the source uses a specialist term, an unobtrusive explanation can help, but explanation must not replace the source classification or add a new entitlement, calculation or interpretation.
Verification: Retain provenance and software/curve version. Read the original and translation side by side, then ask a second reader to explain the rule or claim without consulting the source. If the paraphrase changes the task, rework it before release.
The repeatable twelve-stage translation method
This procedure is intended to be repeated, not merely read once. It keeps a translator from switching between linguistic fluency and factual accuracy at random. Write down the result of each stage. If a fact cannot be verified, mark it for review instead of creating an answer by instinct.
Step 1: Identify specimen and context
Record laboratory code, sample material, excavation or collection identifier and context statement before any date conversion. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
At this point, test the audience’s likely question: ‘What should I do next, and what would make that action wrong?’ A translation that cannot answer because it hid a condition is not yet ready. Do not use a more attractive sentence to compensate for missing data.
Step 2: Copy the measured result
Preserve the exact conventional age, ± measurement error, unit and notation exactly as printed. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
Use a simple three-column review: source fact, target-language rendering, evidence location. When the same object appears in a heading, table and footnote, reconcile those occurrences before approving the section.
Step 3: Classify the dating form
Separate radiocarbon BP, calibrated BP, cal BCE/CE, uncalibrated expressions, modelled ranges and other chronologies. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
If two reviewers disagree, isolate the disputed phrase and ask the original document owner to clarify it. Do not resolve a technical contradiction by majority vote, assumed convention or a convenient machine-generated explanation.
Step 4: Record the reference year
Keep the correct BP convention explicit; do not interpret years-before-present relative to the current calendar. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
At this point, test the audience’s likely question: ‘What should I do next, and what would make that action wrong?’ A translation that cannot answer because it hid a condition is not yet ready. Do not use a more attractive sentence to compensate for missing data.
Step 5: Record calibration method
Extract curve family, calibration software, correction assumptions and any specialist notes. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
Use a simple three-column review: source fact, target-language rendering, evidence location. When the same object appears in a heading, table and footnote, reconcile those occurrences before approving the section.
Step 6: Preserve probability
Copy each numerical range and its probability level without smoothing or selecting the most convenient interval. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
If two reviewers disagree, isolate the disputed phrase and ask the original document owner to clarify it. Do not resolve a technical contradiction by majority vote, assumed convention or a convenient machine-generated explanation.
Step 7: Respect geographic and material assumptions
Keep atmospheric and marine assumptions separate and flag missing or contradictory reservoir information. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
At this point, test the audience’s likely question: ‘What should I do next, and what would make that action wrong?’ A translation that cannot answer because it hid a condition is not yet ready. Do not use a more attractive sentence to compensate for missing data.
Step 8: Separate material history
Describe the dated sample, not automatically the date of construction, settlement, burning or deposition. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
Use a simple three-column review: source fact, target-language rendering, evidence location. When the same object appears in a heading, table and footnote, reconcile those occurrences before approving the section.
Step 9: Preserve formatting
Protect minus signs, superscripts, Greek letters, en dashes, parentheses and BCE/CE notation. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
If two reviewers disagree, isolate the disputed phrase and ask the original document owner to clarify it. Do not resolve a technical contradiction by majority vote, assumed convention or a convenient machine-generated explanation.
Step 10: Check narrative claims
Compare interpretive prose with the laboratory numbers and verify that the museum or author is not exceeding the reported evidence. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
At this point, test the audience’s likely question: ‘What should I do next, and what would make that action wrong?’ A translation that cannot answer because it hid a condition is not yet ready. Do not use a more attractive sentence to compensate for missing data.
Step 11: Seek specialist review
Ask an appropriate archaeologist or radiocarbon scientist to resolve scientific ambiguity, not a generic machine translator. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
Use a simple three-column review: source fact, target-language rendering, evidence location. When the same object appears in a heading, table and footnote, reconcile those occurrences before approving the section.
Step 12: Publish traceable chronology
Keep sample ID, report date, references, version and explicit uncertainty visible in the finished text. This is an explicit task, not a stylistic preference. Work from a copy of the source and preserve a revision trail so that a reviewer can see where a number, date, status, role or exception came from. The result should be a checkable record of what was translated and what was left unresolved.
If two reviewers disagree, isolate the disputed phrase and ask the original document owner to clarify it. Do not resolve a technical contradiction by majority vote, assumed convention or a convenient machine-generated explanation.
Twenty-five worked translation cases
These are constructed teaching examples, not reproductions of authentic personal records or live instructions. Each demonstrates one specific translation risk. The improved wording is an illustrative English rendering; for actual work, fit the same meaning accurately into the target language and its writing conventions.
Worked case 1: Conventional age
Fictional source wording: “2500 ± 30 14C yr BP” Misleading shortcut: “2500 years before today” Meaning-preserving English: “2500 ± 30 conventional radiocarbon years BP.”
The result is not a directly converted calendar date. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Retain 14C and ±30. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Worked case 2: BP reference
Fictional source wording: “BP = years before AD 1950” Misleading shortcut: “Years before 2026” Meaning-preserving English: “BP uses the AD 1950 reference convention in radiocarbon dating.”
Present is a fixed reference under this convention. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Clarify once in a glossary. Try the reversal test: describe the target statement back in the source language without looking. If the returned claim differs in scope, fix the specific missing qualifier.
Worked case 3: Uncalibrated label
Fictional source wording: “Uncalibrated radiocarbon age” Misleading shortcut: “Verified historical age” Meaning-preserving English: “Uncalibrated radiocarbon age.”
Measurement and calibrated interpretation are different stages. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Check reporting header. Check the exact reader action that would follow. If the source and translation lead to different decisions, the wording remains unfinished even if it is grammatically smooth.
Worked case 4: Calendar label
Fictional source wording: “cal BP” Misleading shortcut: “Radiocarbon BP” Meaning-preserving English: “Calibrated calendar years before AD 1950.”
Calibrated values must remain distinguished from 14C ages. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Keep ‘cal’ prefix. Compare the line with its neighbouring heading, table and footnote. The corrected sentence must agree with the document as a whole, not just with a disconnected phrase.
Worked case 5: Calendar era
Fictional source wording: “cal BCE” Misleading shortcut: “Uncalibrated BC” Meaning-preserving English: “Calibrated calendar BCE date.”
Era and calibration state are independent distinctions. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Retain both. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Worked case 6: Confidence level
Fictional source wording: “95.4% probability range” Misleading shortcut: “95.4% accurate date” Meaning-preserving English: “Reported 95.4% probability range.”
A probability interval does not assign certainty to one year. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Preserve reported level. Try the reversal test: describe the target statement back in the source language without looking. If the returned claim differs in scope, fix the specific missing qualifier.
Worked case 7: Multiple intervals
Fictional source wording: “Two separate calibrated ranges” Misleading shortcut: “Single exact middle date” Meaning-preserving English: “Two distinct calibrated calendar-date ranges.”
The shape of a distribution can yield separated ranges. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Copy both bounds. Check the exact reader action that would follow. If the source and translation lead to different decisions, the wording remains unfinished even if it is grammatically smooth.
Worked case 8: Calibration curve
Fictional source wording: “IntCal20” Misleading shortcut: “Generic dating software” Meaning-preserving English: “IntCal20 Northern Hemisphere atmospheric calibration curve.”
The chosen reference curve has scientific meaning. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Verify curve name. Compare the line with its neighbouring heading, table and footnote. The corrected sentence must agree with the document as a whole, not just with a disconnected phrase.
Worked case 9: South curve
Fictional source wording: “SHCal20” Misleading shortcut: “Use IntCal20 by default” Meaning-preserving English: “SHCal20 calibration curve as reported.”
Hemisphere-specific treatment should not be silently changed. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Check sample region. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Worked case 10: Marine curve
Fictional source wording: “Marine20 for sample” Misleading shortcut: “Standard terrestrial curve” Meaning-preserving English: “Marine20 calibration curve for the specified sample.”
Marine conditions require separate assumptions. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Keep sample material. Try the reversal test: describe the target statement back in the source language without looking. If the returned claim differs in scope, fix the specific missing qualifier.
Worked case 11: Reservoir correction
Fictional source wording: “Local ΔR assessment pending” Misleading shortcut: “Correction applied and verified” Meaning-preserving English: “Local ΔR correction assessment is pending.”
Pending assumptions must not become completed calculations. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Preserve status. Check the exact reader action that would follow. If the source and translation lead to different decisions, the wording remains unfinished even if it is grammatically smooth.
Worked case 12: Old wood
Fictional source wording: “Charcoal from long-lived timber” Misleading shortcut: “Exact year building was built” Meaning-preserving English: “Radiocarbon result for charcoal from long-lived timber.”
Sample carbon age may predate the construction event. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Include sample relationship. Compare the line with its neighbouring heading, table and footnote. The corrected sentence must agree with the document as a whole, not just with a disconnected phrase.
Worked case 13: Shell sample
Fictional source wording: “Shell from coastal deposit” Misleading shortcut: “Date of coastal settlement” Meaning-preserving English: “Radiocarbon measurement of a shell from coastal deposit.”
Aquatic reservoir context and deposition interpretation remain separate. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Review context. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Worked case 14: Lab reference
Fictional source wording: “Lab code AB-04126” Misleading shortcut: “Site code AB-04162” Meaning-preserving English: “Laboratory identifier AB-04126.”
A single transposed number can detach results from a specimen. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Character-by-character check. Try the reversal test: describe the target statement back in the source language without looking. If the returned claim differs in scope, fix the specific missing qualifier.
Worked case 15: Pretreatment
Fictional source wording: “Acid–base–acid pretreatment” Misleading shortcut: “Sample was radiocarbon calibrated” Meaning-preserving English: “Acid–base–acid sample pretreatment.”
Preparation steps and calendar calibration are different processes. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Keep workflow stage. Check the exact reader action that would follow. If the source and translation lead to different decisions, the wording remains unfinished even if it is grammatically smooth.
Worked case 16: Stable isotope
Fictional source wording: “δ13C reported” Misleading shortcut: “Carbon-14 calendar date” Meaning-preserving English: “Reported δ13C value.”
Stable isotope reporting is not itself a calibrated 14C date. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Retain symbol. Compare the line with its neighbouring heading, table and footnote. The corrected sentence must agree with the document as a whole, not just with a disconnected phrase.
Worked case 17: Below detection
Fictional source wording: “Result beyond reporting limit” Misleading shortcut: “Exact age found” Meaning-preserving English: “Result is outside the reportable dating limit.”
A bound is not a point estimate. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Preserve inequality. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Worked case 18: Modern carbon
Fictional source wording: “Modern radiocarbon indication” Misleading shortcut: “Artifact made this year” Meaning-preserving English: “Result indicates modern radiocarbon in the tested material.”
Modern carbon signal need not identify an exact manufacture date. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Keep specimen context. Try the reversal test: describe the target statement back in the source language without looking. If the returned claim differs in scope, fix the specific missing qualifier.
Worked case 19: Bayesian model
Fictional source wording: “Posterior phase boundary estimate” Misleading shortcut: “Direct radiocarbon measurement” Meaning-preserving English: “Modelled posterior phase-boundary estimate.”
Model results incorporate assumptions beyond an individual measurement. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Name model stage. Check the exact reader action that would follow. If the source and translation lead to different decisions, the wording remains unfinished even if it is grammatically smooth.
Worked case 20: Stratigraphic order
Fictional source wording: “Sample from layer 4” Misleading shortcut: “Layer 4 is definitely oldest” Meaning-preserving English: “Sample recovered from archaeological layer 4.”
Context number is not itself a chronological conclusion. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Link stratigraphy. Compare the line with its neighbouring heading, table and footnote. The corrected sentence must agree with the document as a whole, not just with a disconnected phrase.
Worked case 21: Date conversion
Fictional source wording: “cal BP and cal BCE appear in parallel” Misleading shortcut: “Subtract BP from the current year” Meaning-preserving English: “Report both original calibrated conventions as supplied.”
Arithmetic shortcuts can mishandle calibration and era boundaries. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Verify existing converted pair. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Worked case 22: Precision
Fictional source wording: “Interval spans several centuries” Misleading shortcut: “Construction in one year” Meaning-preserving English: “A reported calendar range spans multiple centuries.”
False point precision misleads readers. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Use range language. Try the reversal test: describe the target statement back in the source language without looking. If the returned claim differs in scope, fix the specific missing qualifier.
Worked case 23: New calibration
Fictional source wording: “Recalibrated with newer curve” Misleading shortcut: “Same result, no version needed” Meaning-preserving English: “Recalibrated result using the specified curve version.”
Different analysis versions should not be silently merged. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Retain report version. Check the exact reader action that would follow. If the source and translation lead to different decisions, the wording remains unfinished even if it is grammatically smooth.
Worked case 24: Context identification
Fictional source wording: “Short-lived seed sample” Misleading shortcut: “Charcoal from large beam” Meaning-preserving English: “Radiocarbon measurement of a short-lived seed.”
Material type affects how closely its age may relate to an event. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Verify specimen. Compare the line with its neighbouring heading, table and footnote. The corrected sentence must agree with the document as a whole, not just with a disconnected phrase.
Worked case 25: Probability partition
Fictional source wording: “58% in interval A; 37.4% in interval B” Misleading shortcut: “Equal likelihood in two periods” Meaning-preserving English: “58% and 37.4% probabilities for the two reported intervals.”
Probability mass is not necessarily distributed evenly. The danger in the shortcut is not merely that a phrase has a different dictionary definition. It changes a condition, evidence category, actor, time, numerical relationship or available action. A reader relying only on the shorter wording could form a materially different understanding.
Review move: Use reported figures only. Then ask the reviewer to identify the original source line that justifies every operative word in the target. Unsupported words should be removed or clearly marked as commentary.
Failure modes that deserve a second read
A good revision is not a generic spelling check. It is a search for the specific failure that would misdirect a reader. The following patterns are common enough to deserve their own explicit checkpoint.
Failure 1: BP mistaken for today’s past
Dates shift because ‘present’ is taken as the publication year. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Explain the 1950 convention. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 2: Uncalibrated age passed off as calendar
A measurement becomes a historical interval without calibration. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Maintain the 14C/uncalibrated label. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 3: ± uncertainty deleted
The report appears more precise than the experiment. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Retain the uncertainty and its unit. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 4: Probability collapsed
Two calibration intervals become one chosen year. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Preserve interval structure and percentages. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 5: Curve replaced
Marine or southern samples are silently calibrated by another curve. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Copy reported curve and specialist assumptions. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 6: Sample equated to event
An old wood date becomes a construction date. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Describe the dated material and separate interpretive argument. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 7: Laboratory codes drift
A report is assigned to the wrong sample. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Check identifiers against the lab sheet. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 8: Model result mislabelled
Posterior date is treated as raw measurement. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Name analysis stage and assumptions. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Failure 9: Calendar era mistranslated
A BCE label is turned into a CE label. This type of error can be difficult to notice because the text may sound confident and internally coherent. It is precisely the kind of sentence that slips through when a reviewer reads only the target version.
Repair: Verify all bounds and era notation. After repair, compare source and target again for actor, quantity, condition, evidence state, date and exception. Make a short note of which element originally moved so the same class of error is easier to detect next time.
Practice workshop: transfer the skill to a new document
Treat the next exercises as a miniature quality-assurance practicum. Cover the suggested response before trying each task. Mark the exact information that must remain invariant and identify one question that cannot be answered from the excerpt alone.
Practice 1: Conventional age
Task: A bilingual draft renders “2500 ± 30 14C yr BP” as “2500 years before today.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “2500 ± 30 conventional radiocarbon years BP.” The result is not a directly converted calendar date. Retain 14C and ±30. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 2: Uncalibrated label
Task: A bilingual draft renders “Uncalibrated radiocarbon age” as “Verified historical age.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Uncalibrated radiocarbon age.” Measurement and calibrated interpretation are different stages. Check reporting header. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 3: Calendar era
Task: A bilingual draft renders “cal BCE” as “Uncalibrated BC.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Calibrated calendar BCE date.” Era and calibration state are independent distinctions. Retain both. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 4: Multiple intervals
Task: A bilingual draft renders “Two separate calibrated ranges” as “Single exact middle date.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Two distinct calibrated calendar-date ranges.” The shape of a distribution can yield separated ranges. Copy both bounds. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 5: South curve
Task: A bilingual draft renders “SHCal20” as “Use IntCal20 by default.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “SHCal20 calibration curve as reported.” Hemisphere-specific treatment should not be silently changed. Check sample region. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 6: Reservoir correction
Task: A bilingual draft renders “Local ΔR assessment pending” as “Correction applied and verified.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Local ΔR correction assessment is pending.” Pending assumptions must not become completed calculations. Preserve status. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 7: Shell sample
Task: A bilingual draft renders “Shell from coastal deposit” as “Date of coastal settlement.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Radiocarbon measurement of a shell from coastal deposit.” Aquatic reservoir context and deposition interpretation remain separate. Review context. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 8: Pretreatment
Task: A bilingual draft renders “Acid–base–acid pretreatment” as “Sample was radiocarbon calibrated.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Acid–base–acid sample pretreatment.” Preparation steps and calendar calibration are different processes. Keep workflow stage. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 9: Below detection
Task: A bilingual draft renders “Result beyond reporting limit” as “Exact age found.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Result is outside the reportable dating limit.” A bound is not a point estimate. Preserve inequality. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 10: Bayesian model
Task: A bilingual draft renders “Posterior phase boundary estimate” as “Direct radiocarbon measurement.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Modelled posterior phase-boundary estimate.” Model results incorporate assumptions beyond an individual measurement. Name model stage. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 11: Date conversion
Task: A bilingual draft renders “cal BP and cal BCE appear in parallel” as “Subtract BP from the current year.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Report both original calibrated conventions as supplied.” Arithmetic shortcuts can mishandle calibration and era boundaries. Verify existing converted pair. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
Practice 12: New calibration
Task: A bilingual draft renders “Recalibrated with newer curve” as “Same result, no version needed.” Explain the altered meaning, then produce a defensible revision using the source only. Consider what the receiving reader would do differently if this remained uncorrected.
Suggested response: “Recalibrated result using the specified curve version.” Different analysis versions should not be silently merged. Retain report version. An independent learner should also mark the original identifier, scope, date or numerical condition wherever one is present; this is what makes the exercise transferable beyond a single sentence.
A practical quality-control checklist
- Can the intended reader tell which institution, source or process issued the original?
- Does every identifier, person, place, code, date and unit still match?
- Are prohibitions, permissions and duties equally strong in both versions?
- Are source observations distinguished from targets, proposals, estimates and decisions?
- Are exceptions, uncertainty, applicability and version status visible?
- Have tables, footnotes, graphs and attachments been checked along with prose?
- Would the reader take the same next step after reading either language?
- Are unsupported interpretations flagged rather than quietly blended into the translation?
- Has a qualified domain reviewer seen high-stakes or contested terms?
- Can the published version be traced to the exact source and revision date?
Use the checklist with another person. Ask that person to identify a deliberately misleading line from the worked examples. The purpose is to train error recognition, not to perform a ceremonial sign-off. If a reviewer cannot explain the source-target relationship, slow down and check the particular missing condition.
Frequently asked questions
What does BP mean in a radiocarbon report?
For conventional radiocarbon dating it references years before AD 1950; check the report’s precise usage. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Is 2,500 BP the same as 2,500 years ago?
Not as a reliable calendar statement. Conventional radiocarbon age requires calibration and its BP reference is not the current year. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
What is cal BP?
It describes a calibrated calendar age relative to AD 1950, as distinct from conventional radiocarbon BP. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Can I subtract radiocarbon BP from 1950 to find BCE?
No. Radiocarbon-to-calendar calibration is not generally simple subtraction; use scientifically supplied calibrated results. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Why might a report list two date ranges?
Calibration curves can produce separate probability intervals; the translator must preserve both ranges and their stated probabilities. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Is ±30 the size of a calendar-date error bar?
Not automatically. It normally describes the measurement uncertainty at the reported stage; calendar calibration produces its own distribution. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Do radiocarbon dates tell us when a building was built?
They date tested carbon, which may or may not relate directly to construction. Interpretation needs context. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
What are IntCal20 and Marine20?
They are calibration reference curves for different sample/environmental contexts; curve choice belongs to scientific analysis. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Should I translate the lab code?
No. Preserve the identifier exactly so a reader can locate the correct result. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Can AI calibrate a report from the translated words?
Do not invent numerical calibration. Use original laboratory data and an accepted analysis reviewed by qualified specialists. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Why is a result described as modelled?
A model may combine radiocarbon dates with stratigraphic or other assumptions; the resulting estimate is not identical to raw laboratory measurement. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
What makes a strong museum label?
It names the dated material, the stated calibrated range where available, the uncertainty and the limits of inference. In consequential work, always identify the precise source document and intended audience. A translation can be easy to read without adding facts; when the material is outside your expertise, obtain clarification rather than filling the gap from memory.
Where to go next
If your weakness is vocabulary rather than procedure, read the eduKateSG Vocabulary Learning Hub and build a small set of verified terms from the original document. For broader sentence meaning, ambiguity and reading skill, follow How English Works. More precise language helps when it is paired with an evidence check.
The wider translation series has a separate canonical role for survey wording, response scales and skip logic: that guide concerns how choices and questions shape answers. In contrast, this article owns radiocarbon laboratory dating results and chronological language, including the distinction between a conventional measurement in radiocarbon years and calibrated calendar-date probability ranges. Do not combine these jobs merely because each includes the word ‘translation’.
Two nearby specialist routes are Archaeological context and provenance translation and Scientific results and uncertainty translation. They are complementary cases, not alternative versions of this article.
Sources and currency notes
The following original authorities and technical references inform the distinctions used above. They should be consulted at the original URL for updates. The worked examples are fictional and do not purport to quote these sources.
Editorial review date: 30 October 2026. The document owner, court, educational authority, collection service or laboratory remains the source of truth for an actual current case. Rules, technical interpretations and local procedures may change after this date.
The final principle
The purpose of accurate translation is not to make every document sound the same. It is to make the right meaning available to someone who cannot comfortably read the source language. Preserve what is definite, identify what is uncertain, keep the exact conditions that affect action, and let the responsible authority decide what only that authority can decide.
When you finish, put the source and final version together and ask one last question: would the same person, reading either version, know the same facts and the same limits of those facts? If the answer is yes, the translation has done the job that matters.