Coffee Cupping Translation and CVA Scores

Translating a coffee cupping form correctly means preserving how a particular evaluator described a particular coffee, not making the tasting note sound more luxurious. Readers searching for how to translate coffee cupping notes, Coffee Value Assessment (CVA) forms, sensory descriptors, coffee defect terminology and cupping scores need a method that separates observations from opinions, maintains sample identities and leaves recorded values intact. A fluent translation that turns a descriptive citrus note into a claimed ingredient, or a quality preference into a laboratory measurement, changes the evaluation.

Specialty coffee evaluation now uses more than one information layer. The Specialty Coffee Association describes its Coffee Value Assessment framework in terms of physical attributes, descriptive sensory findings, affective judgments and extrinsic information. These dimensions can inform one another, but they are not synonyms. In multilingual coffee trade, Q Grader training, roasting teams and buyer communications, a translator must identify the function of each field before deciding on the target-language phrasing.

This guide explains the canonical job of translating coffee cupping records and tasting assessment forms. It is not an article about brewing a better espresso, converting roast profiles, translating product marketing or assigning an official grade to a coffee. The approach begins with first principles, develops a repeatable review method, tests many original fictional examples and ends with practice, failure analysis, FAQs and sources readers can examine themselves.

The visible problem: a tasting note can look precise while meaning something else

Suppose an evaluator writes “jasmine-like aroma, medium intensity; affective impression: pleasant.” A poorly translated catalogue turns that into “contains jasmine and is certified excellent.” There are three separate errors: a similarity descriptor becomes a physical ingredient, a description of intensity becomes a proof of composition, and an evaluator’s preference becomes an institutional certification. None can be repaired by spelling correction. The translator must preserve what sort of assertion each field makes.

A useful mental model has five columns: sample identity, assessment domain, observed or selected value, evaluator or method, and status. Sample identity tells us which coffee was examined. Assessment domain distinguishes physical, descriptive, affective and extrinsic fields. Value holds the actual words, numbers or tick boxes. Evaluator/method explains how the statement arose. Status says whether the record is complete, disputed, preliminary or final. Losing any column can produce a plausible but false translation.

First principles: twelve distinctions that keep cupping evidence intact

Separate description from liking

A descriptor such as citrus, cocoa or herbal tells a reader what the evaluator perceived; an affective judgment tells how that evaluator valued the perception. Neither can be inferred automatically from the other. A coffee may be strongly acidic without the evaluator liking that acidity. Write the two fields separately, even where the target language offers an attractive adjective that seems to cover both.

A practical test for this principle is to cover the source column and ask an independent reviewer which original evidence could justify the translated sentence. If the reviewer would need an unrecorded measurement, a different assessor’s opinion or an assumption about product ingredients, the translation has crossed its evidence boundary. The repair is to restore the narrower field meaning and retain any unresolved question in an editorial note rather than in the public result.

Distinguish the sample from the commercial product

A cupping table may use a blind code, a roast batch code and a farm lot code for one coffee. Those identifiers have different jobs. A blind sample is an assessment handle, not proof of origin; a producer label is not proof that every tested cup came from the same shipment. Keep identifiers intact and ask the record owner to resolve any mismatched cross-reference.

Keep physical assessment separate from tasting

Moisture, green-bean defects, size and other physical measurements do not turn into flavour scores merely because they appear on adjacent pages. A sensory “defect” and a physical green-bean defect may be governed by different methods. Name the layer before interpreting the number, and do not make the translated sheet imply that a flavour comment came from a laboratory instrument.

Protect scales and anchors

A value means little without knowing its scale, endpoint labels and whether higher numbers mean stronger intensity or greater preference. Never assume that a number from one historical cupping form can be mapped directly onto the Coffee Value Assessment’s current form. Preserve the original scale, capture the form version, and explain rather than silently normalise.

Do not turn flavour associations into ingredients

Tasting language routinely compares sensations with familiar foods, flowers and textures. “Chocolate-like” does not mean chocolate was added; “peach” is not automatically a traceable additive. A reliable translation signals association, perception or character when needed. Ingredient and allergen assertions require different evidence from a sensory note.

Preserve uncertainty and individual voice

“Possibly,” “slight,” “faint,” “not confirmed” and “one assessor noted” are not embarrassing filler. They convey how certain the result is and who made it. Removing uncertainty raises the apparent evidential strength. A useful translator makes the report readable without pretending that a subjective and contingent observation has become universal truth.

Keep dates and conditions attached

Roast date, cupping date, sample preparation and rest period can alter interpretation. A calendar date without timezone or locale may be ambiguous, and a preparation temperature is not a tasting score. Retain labels and units; use unambiguous date formatting only if the publisher’s conventions allow it and the original date remains recoverable.

Distinguish negative preference from defect status

Not enjoying a flavour is not the same as identifying a defect under a defined protocol. A sourness disliked by one evaluator may be an expected bright acidity to another. Translators must not upgrade “I disliked this aftertaste” to “the lot failed inspection.” The source must establish a defect, its method and its status.

Maintain category boundaries across languages

Some languages use one everyday word for strength, intensity, power and quality. A form cannot. A descriptive intensity item names magnitude; an affective item records impression or quality. Choose a phrase long enough to protect that distinction, and prefer explicit labels to a shorter ambiguous adjective.

Respect official names and evolving methods

The Specialty Coffee Association’s Coffee Value Assessment is a named framework. Its four assessment categories must not be replaced casually with an earlier scoring vocabulary. The official current forms and standards, rather than a remembered social-media table, establish what a label means. If the source uses an older form, identify it honestly instead of pretending that every historical score is CVA.

Translate absent data as absent data

An empty box, not assessed marker or illegible note is not a zero and not a normal result. Zero may be a valid numerical observation in a field that permits it. The translated record should preserve the source status or explicitly indicate missing information. Filling a blank to make a spreadsheet complete can distort purchasing or quality decisions.

Make the translation auditable

A reviewer should be able to reconstruct each translated statement from the original field and its evidence. Keep a field-by-field trail, retain raw sample IDs, record terminology choices and flag uncertain source readings. An elegant prose paragraph is not enough if the reader cannot identify which field produced its conclusion.

A repeatable twelve-pass workflow for multilingual cupping records

1. Establish the document’s job

Decide whether the text is a laboratory-facing record, educational worksheet, internal purchasing note or public tasting description. A lab record needs exact columns and status markers. A consumer-facing explanation can define difficult terms but cannot quietly rewrite the source result.

The pass is complete only when the reviewer can point to a concrete source field, a translation decision and a verification action. Where the record lacks a necessary legend, the correct output is a clearly flagged question, not a guessed interpretation. This is a deliberately slower standard for critical cells; routine descriptive prose can move more quickly once the terminology map is stable.

2. Identify the governing form

Record the form name, year or version and the assessment category. Check the current SCA Coffee Value Assessment resources where appropriate, but do not impose them on a different vendor’s historical format. Treat each form’s scale and labels as a closed system until its legend is understood.

3. Map the fields before words

Create a table with source header, target header, data type, allowable values, evidence origin and required reviewer. Keep check boxes separate from free-text observations. Mark each line as numeric, identification, descriptive, affective, physical or extrinsic so that a persuasive adjective never replaces a technical category.

4. Freeze identifiers

Copy sample codes, lot references, barcodes and batch IDs character for character. Check zero against letter O, one against letter I, and hyphen or slash placement. If a form contains conflicting identifiers, hold the record for confirmation rather than choosing the one that looks most plausible.

5. Preserve units and scales

Record units, decimal marks, upper and lower scale anchors and applicable score definitions. Convert a decimal separator only under a documented locale convention. Do not invent totals, interpret unlabelled scales as percentages or replace an assessor’s selected value with an apparently equivalent word.

6. Build a small sensory lexicon

List the exact words the project uses for aroma, fragrance, flavour, aftertaste, acidity, sweetness and mouthfeel, noting differences between similar target-language terms. Attach concrete examples and permitted context, rather than a bare bilingual word list that ignores field boundaries.

7. Translate every hedge

Underline uncertain wording: possibly, resembles, faint, inconsistent, one cup, pending retest. Translate those modifiers before polishing the main adjective. A record with qualified observations may feel less commercial, but it is more accurate and therefore more useful to buyers.

8. Review descriptors against claims

Check whether a consumer might read a translated note as a statement of ingredients, origin, processing certification or health effects. If the source merely records sensory resemblance, make resemblance explicit. If the source includes verified extrinsic evidence, cite that evidence separately.

9. Rebuild tables with alignment

Before approval, compare target rows with the source at the cell level. Check row order, merged headings, table footnotes and score legends. A perfectly translated column placed under the wrong heading creates a more dangerous error than an awkward but correctly aligned word.

10. Test with two readers

Ask a competent target-language reader to explain what the table says. Ask a domain reviewer whether that explanation is supported by the original cupping evidence. Where they disagree, examine category and scale mismatches before debating stylistic preference.

11. Reconcile revisions

Track whose note changed, when it changed and whether scores were corrected or merely copied to a fresh layout. Keep the previous version where a buyer or quality manager may need it. Do not conceal source changes by silently replacing translated entries in a published PDF.

12. Release with boundaries

Publish the final text with its sample date, assessment type, known limitations and the appropriate source references. If the final audience needs marketing language, write it as a distinct editorial summary attributed to the published assessment, never as if the evaluator recorded words that did not exist.

Twenty-eight worked cases: from misleading translations to recoverable meaning

Every short source below is an original fictional teaching example. It is not a real SCA form entry, a certification result or a recommendation for a particular coffee. The contrast is between a misleading English rendering and a safer interpretation of what a multilingual assessor’s record might mean.

Case 1: Aroma, not fragrance

Source field: “Wet-cup note: aroma, orange blossom, medium.” Misleading rendering: “This dry coffee contains orange-blossom perfume.” Evidence-preserving rendering: “Wet-cup aroma described as medium-intensity orange-blossom-like.”

Fragrance and aroma occupy different preparation moments in some protocols; a floral likeness is not a formula. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Confirm which phase was actually smelled and that the result remains a description. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 2: Dry fragrance field

Source field: “Dry grounds: fragrance, hazelnut, low.” Misleading rendering: “Hazelnut ingredient detected.” Evidence-preserving rendering: “Low-intensity hazelnut-like fragrance in the dry grounds.”

The sheet records an association before water, not a laboratory identification of nuts. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Check preparation stage and remove unsupported ingredient assertions. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 3: Acidity intensity

Source field: “Acidity intensity 4 on this form’s scale.” Misleading rendering: “Acidity quality 4 out of 5.” Evidence-preserving rendering: “Acidity intensity: 4, using the stated form scale.”

A magnitude rating is not a quality judgment and its denominator must come from the actual form. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Read the source legend and display the same scale. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 4: Preference score

Source field: “Affective impression: 7, evaluator A.” Misleading rendering: “Seven measures of citric acid detected.” Evidence-preserving rendering: “Evaluator A’s affective rating: 7 on the form’s scale.”

A preference value belongs to an evaluator and assessment instrument. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Keep assessor identity and scale anchors. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 5: Aftertaste duration

Source field: “Aftertaste: short, cocoa-like.” Misleading rendering: “Long-lasting cocoa aftertaste.” Evidence-preserving rendering: “Short aftertaste with a cocoa-like character.”

Changing short to long alters the observed duration even when the flavour noun survives. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Compare duration and character as separate fields. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 6: Mouthfeel descriptor

Source field: “Mouthfeel: silky, moderate weight.” Misleading rendering: “A thick, oily coffee.” Evidence-preserving rendering: “Moderate weight with a silky mouthfeel.”

Silky concerns texture and does not prove high thickness or oil content. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Ask another reader to state texture and weight separately. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 7: Jasmine association

Source field: “Jasmine note recorded by panelist two.” Misleading rendering: “Produced with jasmine flowers.” Evidence-preserving rendering: “Panelist two noted a jasmine-like sensory character.”

The source does not establish processing ingredients, and not all panelists may agree. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Preserve attribution and resemblance. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 8: Citrus ambiguity

Source field: “Citrus impression; fruit not specified.” Misleading rendering: “Clearly lemon and grapefruit.” Evidence-preserving rendering: “Citrus impression, specific fruit not identified.”

A hyper-specific descriptor may seem vivid but invents observations. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Retain the level of specificity. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 9: Processing metadata

Source field: “Producer reports washed processing.” Misleading rendering: “Washed flavour confirmed by cup.” Evidence-preserving rendering: “Processing method reported as washed by producer.”

Extrinsic or supplier information is distinct from a sensory conclusion. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Keep source of production claim. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 10: Roast date

Source field: “Roasted 06/07/26, date order unknown.” Misleading rendering: “Roasted on 7 June 2026.” Evidence-preserving rendering: “Roast date recorded as 06/07/26; date format needs confirmation.”

Month-day ambiguity can reverse chronological interpretation. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Query the origin locale before normalising. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 11: Cupping date

Source field: “Assessed 2026-08-10.” Misleading rendering: “Produced on 10 August 2026.” Evidence-preserving rendering: “Cupping conducted on 10 August 2026.”

Assessment and production are different events. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Keep date label attached. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 12: Blind code

Source field: “Sample B07, origin hidden during testing.” Misleading rendering: “Brazil lot B07.” Evidence-preserving rendering: “Blind sample code B07; origin not disclosed at assessment.”

A code must not be decoded from guesswork or stereotype. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Cross-check with the authorised key after evaluation. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 13: Lot number

Source field: “Lot 04-031B.” Misleading rendering: “Lot 04-0318.” Evidence-preserving rendering: “Lot 04-031B.”

Letters B and numbers 8 are distinct identifiers despite similar appearance. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Compare against the master file character by character. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 14: Physical moisture

Source field: “Green sample moisture: 10.5 percent.” Misleading rendering: “Sweetness score: 10.5.” Evidence-preserving rendering: “Green sample moisture measured at 10.5%.”

A physical percentage is not a sensory or preference score. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Preserve measurement method and unit. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 15: Bean defects

Source field: “Physical assessment: two counted defects.” Misleading rendering: “Two taste defects in every cup.” Evidence-preserving rendering: “Two defects counted in the physical assessment.”

The count does not establish a corresponding number of perceived off-flavours. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Check the defect category and its measurement method. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 16: Single negative comment

Source field: “One evaluator: ‘earthy, not pleasant to me’.” Misleading rendering: “Certified defective earthy flavour.” Evidence-preserving rendering: “One evaluator described an earthy note and personally disliked it.”

Preference is not an official defect classification. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Retain evaluator attribution and hedging. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 17: Unscored blank

Source field: “Affective field left blank.” Misleading rendering: “Affective score zero.” Evidence-preserving rendering: “Affective rating not recorded.”

Blank is missing information and may reflect incomplete administration. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Request clarification before calculation. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 18: Not applicable

Source field: “Physical size field: N/A.” Misleading rendering: “Particle size below detection.” Evidence-preserving rendering: “Physical size field marked not applicable.”

N/A must not be reinterpreted as a failed test or a zero reading. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Check the form’s permitted statuses. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 19: Retest underway

Source field: “Second assessment requested; result pending.” Misleading rendering: “Second assessment passed.” Evidence-preserving rendering: “Second assessment requested; result pending.”

A process decision does not predict the result of that process. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Keep pending status visible. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 20: Conflicting descriptors

Source field: “Evaluator A: floral; evaluator B: herbal.” Misleading rendering: “Panel unanimously identified flowers.” Evidence-preserving rendering: “Evaluator A noted floral; evaluator B noted herbal.”

Disagreement is part of the actual dataset and must not be erased by a single polished summary. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Preserve individual evidence. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 21: Decimal separator

Source field: “Recorded score: 8,5 in a comma-decimal locale.” Misleading rendering: “Eight and five separate ratings.” Evidence-preserving rendering: “Recorded score: 8.5, after verified decimal-format conversion.”

Punctuation may encode a decimal and must not create two values. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Check spreadsheet type and scale. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 22: No official certification

Source field: “Internal note: interesting specialty profile.” Misleading rendering: “Official SCA-certified top quality.” Evidence-preserving rendering: “Internal assessor described the profile as interesting.”

Specialty is contextual language; it does not automatically confer institutional certification. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Remove invented authority and grade. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 23: Flavor wheel category

Source field: “Descriptor family: dried fruit.” Misleading rendering: “Contains dried fruit puree.” Evidence-preserving rendering: “A dried-fruit-like sensory descriptor.”

A sensory taxonomy does not identify recipe ingredients. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Ensure analogy remains an analogy. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 24: CVA category heading

Source field: “Descriptive assessment sheet.” Misleading rendering: “Quality approval certificate.” Evidence-preserving rendering: “Descriptive sensory assessment sheet.”

The title defines what kinds of conclusions a reader may draw. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Match official terminology. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 25: Extrinsic certification

Source field: “Supplier provided a certificate ID; not verified.” Misleading rendering: “Independently certified organic.” Evidence-preserving rendering: “Supplier supplied certificate ID; verification status not established.”

Reporting receipt of evidence is not verifying its authenticity. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Preserve provenance and validation status. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 26: Questionable flavour word

Source field: “Source says ‘bright’, context notes acid character.” Misleading rendering: “The cup is physically luminous.” Evidence-preserving rendering: “Bright acidity as described by the evaluator.”

Figurative vocabulary needs a domain-level interpretation rather than literal word substitution. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Check the tasting field and nearby qualifiers. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 27: Computed average

Source field: “Average shown by source: 7.2; individual scores attached.” Misleading rendering: “All tasters scored 7.2.” Evidence-preserving rendering: “Source reports an average of 7.2 across the listed assessments.”

An aggregate conceals distribution and is not an individual consensus. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Keep mean label and underlying scores accessible. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Case 28: Quality-control correction

Source field: “Original sheet corrected sample C12 to C21.” Misleading rendering: “The sample was always C21; no correction.” Evidence-preserving rendering: “Correction noted: sample code changed from C12 to C21.”

A correction history helps buyers reconcile circulated versions and sample records. The reviewer should not judge this pair by whether the second English sentence sounds more attractive. The question is whether a person consulting the translation would make the same decision, identify the same measurement or understand the same uncertainty as someone who read the original. Coffee assessment tables often look compressed, so expanding a label into a clear phrase can improve readability without importing a new claim.

Verification drill: Preserve the revision note and authorised final identifier. Then perform a reverse explanation: tell a second reader what was observed, who reported it, and which evidence was not supplied. Where the safe version preserves an unresolved question, label that question for the record owner. Do not let the translation solve it by invention. Transfer this discipline to another field on the same sheet; the goal is consistent judgement rather than memorising the example’s wording.

Failure modes and how to repair them

Marketing language enters the data column

The translator improves every descriptor into a confident selling adjective. Repair by splitting literal assessment text from separately labelled marketing commentary. A buyer should never have to guess whether praise came from the assessor or the copywriter.

A useful release test asks three questions. What claim would a reader make after seeing the target sentence? What exact source field supports that claim? Would a different assessor or buyer reach the same conclusion without extra assumptions? If the answers diverge, keep the item in review rather than smoothing it over.

A number loses its anchor

A score is copied without its form version or scale. Repair by supplying the full field label, scale and date from the source. Do not derive a percentage or quality grade merely because readers prefer familiar numbers.

Subjective and physical fields are blended

Acidity preference is displayed as a chemical acidity measurement. Repair by marking whether the data came from a sensory assessor or physical testing. The evaluation system matters more than a superficial similarity between words.

Tables shift during localisation

A translated header wraps onto two columns and moves ratings under the wrong category. Repair by table-by-table alignment checks, exported cell comparisons and a final rendered-page inspection, not only a spellcheck.

Historical forms are silently modernised

A legacy scorecard is relabelled as current CVA. Repair by preserving the actual originating form and version. Explain current practice in a contextual note, if relevant, without retroactively changing the earlier document.

Translation states certainty where source has doubt

One uncertain note becomes a category-wide conclusion. Repair by restoring words like possible, faint, disputed and pending, along with individual assessor attribution.

Origin or certification is guessed

Coffee flavour is used to infer country of origin, processing or certified claims. Repair by citing only verified extrinsic fields, and by quarantining uncertain source information.

Missing fields are silently filled

An empty box becomes zero, a dash becomes no defect or a sample code is invented to complete a column. Repair by distinguishing blank, not applicable, not tested, illegible and genuinely measured zero.

Source revisions are lost

A correction reaches the buyer in a later translated file without history. Repair by recording version identifiers, changed rows, correction date and who approved the release.

Translation is reviewed by language skill alone

A fluent reviewer accepts a technically false rendering. Repair by pairing linguistic review with a domain reader and testing a sample of difficult fields against the source.

Four-week practice and transfer programme

Week 1: field identification

Print a blank fictional cupping sheet and label every field as identifier, physical, descriptive, affective or extrinsic. Translate only headers first. Ask a partner to sort them independently, and discuss where one everyday word conceals two assessment meanings.

Score this practice by fidelity rather than elegance: one point for the right field, one for the right value, one for correct evidential status, one for keeping source attribution, and one for stating what remains unknown. After correcting a mistake, retry it on a different sample instead of merely reading the answer. Transfer happens when the learner can protect meaning in a new record with unfamiliar terminology.

Week 2: faithful descriptive language

Create twenty original tasting notes with soft qualifiers, comparative words and sensory associations. Translate them into a second language, then back-explain their meaning in plain English. Flag every moment a resemblance becomes an ingredient or an intensity becomes praise.

Week 3: scores and record state

Build a spreadsheet with assessor IDs, deliberately ambiguous dates, empty cells, decimal commas, pending retests and corrections. Perform a field-by-field translation and have a separate reviewer find five planted errors without being told where they are.

Week 4: full professional handoff

Prepare a bilingual assessment summary with a source ledger, glossary, field map, reviewer notes, version number and a limited publication statement. Present it to a fictional buyer and answer follow-up questions only from the recorded evidence.

Frequently asked questions

Is a sensory descriptor the same as an ingredient?

No. In tasting notes, coffee may be described as citrus-like, chocolate-like or floral because those terms help communicate perceived sensations. An ingredient declaration is a separate factual claim supported by composition and production information. A translator must preserve the distinction.

Can I replace an old cupping score with a current CVA score?

Not without an authorised and methodologically justified reassessment or conversion explicitly supported by the governing source. A legacy field does not automatically correspond to a current assessment dimension. Name the actual form in use and retain historical scores with their original scale.

Why separate descriptive and affective assessment?

Description addresses perceived sensory character, while affective assessment captures an evaluator’s impression of quality or preference. Treating those judgments as identical makes it impossible to distinguish a shared observation from a difference in taste.

Should origin and process labels be translated?

Translate explanatory words where useful, but preserve traceable lot IDs, registered names and authenticated producer information. Do not infer origin or process from a flavour note. Indicate whether extrinsic facts were supplied by a producer or independently checked.

How do I handle a blank score?

Keep it blank or mark it as not recorded according to the form. Do not insert zero unless the source contains zero and that value is valid for the field. An unknown result is not an unfavourable result.

Are coffee tasting terms universal across languages?

No. Language communities may divide taste, flavour, texture and intensity differently, and metaphors vary. Build a project-specific sensory glossary with examples and have bilingual domain readers test important contrasts.

Does a polished translation make a cupping note more trustworthy?

Clearer wording can improve access, but trust depends on source evidence, method, attribution, revision handling and accurate field boundaries. Fluent prose cannot supply a missing measurement or resolve a contradiction in the original record.

Who should review a translated cupping sheet?

Ideally a target-language specialist and someone qualified to interpret the specific cupping protocol. The domain reviewer checks whether the technical fields lead to the same reading as the source; the language reviewer checks whether target readers can recover that meaning.

What if one evaluator disagrees with another?

Record the disagreement and preserve attribution. Do not force a unanimous panel conclusion for smoother copy. Aggregate results should be labelled as aggregates, and raw observations should remain traceable where publication permits.

Where should beginners study the wider language foundations?

Use domain resources for assessment methodology and the eduKateSG language hubs for grammar, precision and vocabulary. Technical translation requires both a reliable model of the subject and the ability to write distinctions readers can understand.

Sources and contextual next routes

The authoritative starting point for the current framework is the Specialty Coffee Association Coffee Value Assessment, which describes physical, descriptive, affective and extrinsic assessment. For educational background, the SCA introduction to cupping education explains beginner sensory learning. Read primary documentation before interpreting a particular official form; this guide does not replace current standards or professional evaluation.

For a different canonical problem, see translating coffee roast labels, processing terms and brewing ratios. That article concerns consumer-facing roast and brewing information, not the evidential status of cupping assessments. For the language behind precise distinctions, continue to Vocabulary Learning Hub and How English Works. The related translation lane on coin auction catalogue grading demonstrates the parallel danger of replacing recorded classification with advertising language.

The durable rule is modest: say exactly what the assessor recorded, under exactly the field and method that produced it, and no more. A translation succeeds when a different-language reader can reproduce the scope, uncertainty and identity of the evidence without borrowing facts that the source never supplied.

Top Ways to Translate Correctly | Translate Coffee Cupping Forms, CVA Sensory Notes and Scores Without Changing Evaluation