Translate | Translation Exam Marking and Rubrics: How Accuracy, Meaning, Language and Errors Are Assessed

Translation exam marking and rubrics are easier to use when candidates stop treating “accuracy” as one vague quality. Translation assessment can distinguish source comprehension, meaning transfer, target-language mechanics, writing quality, terminology, cohesion, register, style, omissions, additions and the practical effect of an error. The exact categories and scoring rules belong to the assessment provider, but the underlying distinctions can make practice much more precise.

This world-facing guide explains translation exam scoring, translation assessment criteria, translator certification marking, translation error categories and rubric-based translation practice. It is not an unofficial answer key and does not merge different credentials into one invented global rubric. Instead, it teaches readers how to read a provider’s current assessment framework, understand what a marked error means, and convert assessment language into better translation decisions.

For a concrete current example, the American Translators Association publishes its grading procedure, standardized error framework and detailed error categories. ATA currently separates target-language mechanics, meaning transfer and writing quality; uses graded error-point values based on impact; and requires a passing result on both graded passages. Those rules belong specifically to ATA and should be checked on its current official pages. ATA: How the Exam Is Graded ATA: Error Categories

50-second route

First identify the provider and current rubric. Then separate four questions: Did the target preserve the source meaning? Is the target language correct and natural? Does the target fit the brief, audience and text type? How consequential is the problem? Use those questions to diagnose practice, but never invent provider points or pass thresholds that the provider does not publish.

This page owns translation-specific marking architecture and rubric literacy. It complements rather than replaces Translation Exam Practice, Results and Feedback, and Translation Exam Retake.

Part I — How translation marking works

1. A rubric is a model of quality, not the translation itself

A rubric turns complex judgement into named dimensions so assessors can apply standards consistently. It does not mean every good translation must use the same words. Translation permits alternatives, but alternatives still have to preserve meaning, function and target-language acceptability.

Adrian learns to ask what dimension a correction belongs to before changing his process. If a term is accurate but stylistically different from a model, he does not call it a transfer error. If a beautiful sentence changes the source claim, he does not defend it as style.

Rubric literacy therefore reduces both undercorrection and overcorrection. The candidate becomes better at identifying what kind of quality is at stake.

2. Provider rubrics are not interchangeable

ATA, NAATI, CIOL and other assessment bodies define their own tasks, standards and result systems. A category or point value from one provider should not be imported into another assessment unless that provider explicitly uses it.

A world-facing learner can still use common analytical questions—meaning, target language, brief, consequence—while keeping official scoring with the credential owner.

This separation prevents a local teacher from inventing a hybrid scoring system and presenting it as official.

3. Meaning transfer is more than word matching

Meaning includes propositions, relationships, scope, attribution, time, condition, quantity, tone and emphasis. A target can contain plausible vocabulary and still alter the source.

Ben’s quantity examples show this clearly: changing “up to four” into “four” preserves the noun and numeral but changes the permitted set. A rubric that attends to transfer catches the relationship rather than rewarding surface overlap.

Practice should therefore ask what the reader would believe after reading the target compared with the source.

4. Target-language mechanics matter independently

Grammar, syntax, spelling, punctuation and word form can be wrong even when the intended source meaning is recoverable. These problems affect professionalism and readability and may be classified separately by a provider.

Mira reviews target mechanics without assuming every awkward phrase is a meaning error. This makes feedback more precise and helps select the correct repair.

If a mechanical error changes meaning, however, its consequence can move the problem into a more serious transfer category under some frameworks. Always use the provider’s definitions.

5. Writing quality is not decorative polish

Usage, text type, register and style affect whether the translation works for its intended audience. Natural target language is part of translation quality, not an optional luxury after accuracy.

Jo can preserve every fact yet write an institutional notice in an inappropriate conversational register. The problem belongs to communicative fit, even if no proposition is missing.

Practice should therefore include target-only reading and brief-aware rewriting alongside source comparison.

6. Error severity depends on impact

Not every error has equal consequence. A comma that does not affect meaning and a reversed condition can differ enormously in how they affect usability. ATA’s current framework explicitly uses different point values and considers impact.

Local practice can use qualitative severity—low, moderate, high—without pretending those labels equal official points. The aim is to prioritise repair.

A learner should spend more attention on errors that change what the reader would do, believe or understand.

7. Addition is not the same as useful explicitation

A translator sometimes needs to make implicit information explicit because the target language requires it or because context supports it. That differs from adding unsupported content.

Aisha tests every expansion: is the detail recoverable from context and necessary for the target, or did she introduce a new claim?

Rubric-aware practice teaches the boundary instead of a simplistic rule that all additions are forbidden.

8. Omission can be tiny and consequential

Small qualifiers, negatives, exceptions and attribution markers can disappear while the sentence remains grammatical. The result may be a major meaning change.

Ryan’s source-to-target sweep maps propositions and controlling words rather than checking only content nouns.

A good rubric helps learners see why completeness is semantic, not merely word-count coverage.

9. Terminology is contextual

A dictionary equivalent can be wrong in a particular domain, institution or sentence. Terminology assessment asks whether the chosen target represents the source concept in context.

Adrian records concept, evidence and rejected alternatives rather than memorising isolated pairs.

Terminology errors should lead to concept and resource training, not merely longer vocabulary lists.

10. Cohesion is text-level accuracy

A translation can contain locally acceptable sentences that fail together. Repeated terms may drift, pronouns may attach incorrectly, or logical connections may weaken.

Ryan’s paragraph-level reference work addresses this text-level quality. Cohesion cannot be repaired fully with sentence flashcards.

Rubric literacy expands the unit of translation from isolated sentence to coherent document.

11. Literalness and faithfulness are different failure directions

A target can follow source wording too closely and become unnatural or wrong; it can also move too freely and change meaning or emphasis. Good translation is not located at one fixed point called literal or free.

The decision depends on what must remain invariant and what target-language restructuring is allowed. Mira practises multiple natural versions of the same proposition.

Assessment criteria help candidates judge freedom by effect rather than ideology.

12. Misunderstanding begins before target writing

When the source has been parsed incorrectly, polishing the target cannot repair the underlying representation. A rubric category identifying source misunderstanding should route practice back to comprehension.

Clara explains the source untimed before retranslating. If the explanation is wrong, she learns the construction first.

This prevents teachers from treating every bad target as a writing problem.

13. Indecision can become visible in the submitted target

Leaving two alternatives, slashes or unresolved options may be penalised under provider rules. ATA explicitly identifies indecision as an error category.

During practice, uncertainty should be managed with temporary marks that are cleared before submission. The learner must ultimately make a defensible choice.

The repair is not blind confidence; it is better evidence gathering plus a closure rule.

14. Unfinished work is qualitatively different

An incomplete target is not simply a collection of local errors. Some frameworks have specific rules for substantially unfinished passages.

Timed practice must therefore protect whole-task completion before optional polish. Clara’s stop-loss routine exists partly to prevent one hard sentence from consuming the rest.

Provider-specific unfinished rules should always be checked directly.

15. Text type connects the brief to language choices

The same proposition can be rendered differently for a public notice, academic report, internal memo or general reader. Translation instructions and audience can constrain terminology, register and structure.

Jo’s dual-brief exercises teach which differences are legitimate adaptation and which facts cannot change.

A rubric-aware candidate reads the brief as part of the source of assessment evidence, not as decorative front matter.

16. Register is relational

Formal and informal are not absolute labels. Register depends on audience, institution, genre and communicative purpose.

Aisha compares several target versions and asks which social relationship each one constructs. She avoids replacing every ordinary word with a longer synonym in the name of formality.

Retake and exam practice should train register through authentic target-language patterns.

17. Style is functional

Style in assessment is not simply personal elegance. It includes choices appropriate to the publication or functional use described by the task.

Mira’s style check asks whether the target reads like the kind of document it is supposed to be. A technically grammatical sentence can still be an awkward fit.

This is why target-language reading in relevant genres improves exam performance.

18. A model answer is not the rubric

Models can illustrate one defensible solution. The rubric defines dimensions of quality more broadly. Candidates who chase model wording can overedit acceptable alternatives.

Ethan compares model differences by meaning, target convention and brief. He changes only when the evidence supports a better decision.

Teachers should expose learners to multiple acceptable translations so rubric literacy does not become answer-key dependence.

19. Quality points or positive recognition are provider-specific

Some frameworks may explicitly recognise exceptional translation choices. ATA’s current framework allows limited quality points that reduce the error-point total. That is an ATA mechanism, not a global translation rule.

Local teaching can still notice excellent decisions—precise terminology, elegant restructuring, strong cohesion—without inventing official bonus points.

Positive evidence matters because a rubric should teach what to preserve as well as what to repair.

20. Marking should end in a learning decision

A correction list is not yet a training plan. Group marked issues by mechanism, consequence and recurrence, then select the smallest useful next task.

Ben’s three quantity errors become one boundary repair; Ryan’s pronoun errors become a reference-chain programme; Mira’s awkward prose becomes target-language reformulation.

Rubric literacy becomes educationally valuable when categories route the learner toward better practice.

Part II — Fifty rubric laboratories

These laboratories are independent educational exercises. They do not assign official points. Use the current rubric of the assessment you are actually preparing for whenever provider scoring matters.

Rubric Lab 1 — grammar

Identify the dimension. Build three short source-target examples around grammar: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the grammar decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the grammar problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying grammar challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 2 — syntax

Identify the dimension. Build three short source-target examples around syntax: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the syntax decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the syntax problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying syntax challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 3 — word form

Identify the dimension. Build three short source-target examples around word form: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the word form decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the word form problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying word form challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 4 — spelling and character accuracy

Identify the dimension. Build three short source-target examples around spelling and character accuracy: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the spelling and character accuracy decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the spelling and character accuracy problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying spelling and character accuracy challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 5 — capitalization

Identify the dimension. Build three short source-target examples around capitalization: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the capitalization decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the capitalization problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying capitalization challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 6 — diacritics and accents

Identify the dimension. Build three short source-target examples around diacritics and accents: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the diacritics and accents decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the diacritics and accents problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying diacritics and accents challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 7 — punctuation

Identify the dimension. Build three short source-target examples around punctuation: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the punctuation decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the punctuation problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying punctuation challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 8 — addition

Identify the dimension. Build three short source-target examples around addition: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the addition decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the addition problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying addition challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 9 — omission

Identify the dimension. Build three short source-target examples around omission: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the omission decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the omission problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying omission challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 10 — terminology

Identify the dimension. Build three short source-target examples around terminology: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the terminology decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the terminology problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying terminology challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 11 — false friends

Identify the dimension. Build three short source-target examples around false friends: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the false friends decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the false friends problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying false friends challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 12 — verb meaning

Identify the dimension. Build three short source-target examples around verb meaning: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the verb meaning decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the verb meaning problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying verb meaning challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 13 — ambiguity

Identify the dimension. Build three short source-target examples around ambiguity: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the ambiguity decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the ambiguity problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying ambiguity challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 14 — cohesion

Identify the dimension. Build three short source-target examples around cohesion: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the cohesion decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the cohesion problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying cohesion challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 15 — faithfulness

Identify the dimension. Build three short source-target examples around faithfulness: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the faithfulness decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the faithfulness problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying faithfulness challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 16 — literalness

Identify the dimension. Build three short source-target examples around literalness: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the literalness decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the literalness problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying literalness challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 17 — source misunderstanding

Identify the dimension. Build three short source-target examples around source misunderstanding: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the source misunderstanding decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the source misunderstanding problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying source misunderstanding challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 18 — indecision

Identify the dimension. Build three short source-target examples around indecision: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the indecision decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the indecision problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying indecision challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 19 — unfinished work

Identify the dimension. Build three short source-target examples around unfinished work: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the unfinished work decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the unfinished work problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying unfinished work challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 20 — usage

Identify the dimension. Build three short source-target examples around usage: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the usage decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the usage problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying usage challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 21 — text type

Identify the dimension. Build three short source-target examples around text type: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the text type decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the text type problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying text type challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 22 — register

Identify the dimension. Build three short source-target examples around register: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the register decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the register problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying register challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 23 — style

Identify the dimension. Build three short source-target examples around style: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the style decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the style problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying style challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 24 — numbers and quantities

Identify the dimension. Build three short source-target examples around numbers and quantities: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the numbers and quantities decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the numbers and quantities problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying numbers and quantities challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 25 — dates and temporal sequence

Identify the dimension. Build three short source-target examples around dates and temporal sequence: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the dates and temporal sequence decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the dates and temporal sequence problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying dates and temporal sequence challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 26 — conditions and exceptions

Identify the dimension. Build three short source-target examples around conditions and exceptions: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the conditions and exceptions decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the conditions and exceptions problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying conditions and exceptions challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 27 — negation

Identify the dimension. Build three short source-target examples around negation: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the negation decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the negation problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying negation challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 28 — attribution

Identify the dimension. Build three short source-target examples around attribution: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the attribution decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the attribution problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying attribution challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 29 — modality and certainty

Identify the dimension. Build three short source-target examples around modality and certainty: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the modality and certainty decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the modality and certainty problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying modality and certainty challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 30 — pronoun reference

Identify the dimension. Build three short source-target examples around pronoun reference: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the pronoun reference decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the pronoun reference problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying pronoun reference challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 31 — institutional roles

Identify the dimension. Build three short source-target examples around institutional roles: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the institutional roles decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the institutional roles problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying institutional roles challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 32 — repeated terminology

Identify the dimension. Build three short source-target examples around repeated terminology: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the repeated terminology decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the repeated terminology problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying repeated terminology challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 33 — headings and labels

Identify the dimension. Build three short source-target examples around headings and labels: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the headings and labels decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the headings and labels problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying headings and labels challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 34 — audience fit

Identify the dimension. Build three short source-target examples around audience fit: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the audience fit decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the audience fit problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying audience fit challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 35 — genre conventions

Identify the dimension. Build three short source-target examples around genre conventions: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the genre conventions decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the genre conventions problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying genre conventions challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 36 — sentence recasting

Identify the dimension. Build three short source-target examples around sentence recasting: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the sentence recasting decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the sentence recasting problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying sentence recasting challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 37 — paragraph flow

Identify the dimension. Build three short source-target examples around paragraph flow: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the paragraph flow decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the paragraph flow problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying paragraph flow challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 38 — cause and effect

Identify the dimension. Build three short source-target examples around cause and effect: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the cause and effect decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the cause and effect problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying cause and effect challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 39 — comparison relationships

Identify the dimension. Build three short source-target examples around comparison relationships: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the comparison relationships decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the comparison relationships problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying comparison relationships challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 40 — scope

Identify the dimension. Build three short source-target examples around scope: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the scope decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the scope problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying scope challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 41 — emphasis

Identify the dimension. Build three short source-target examples around emphasis: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the emphasis decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the emphasis problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying emphasis challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 42 — tone

Identify the dimension. Build three short source-target examples around tone: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the tone decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the tone problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying tone challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 43 — collocation

Identify the dimension. Build three short source-target examples around collocation: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the collocation decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the collocation problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying collocation challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 44 — idiom

Identify the dimension. Build three short source-target examples around idiom: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the idiom decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the idiom problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying idiom challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 45 — revision-induced errors

Identify the dimension. Build three short source-target examples around revision-induced errors: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the revision-induced errors decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the revision-induced errors problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying revision-induced errors challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 46 — research-supported terminology

Identify the dimension. Build three short source-target examples around research-supported terminology: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the research-supported terminology decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the research-supported terminology problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying research-supported terminology challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 47 — acceptable alternatives

Identify the dimension. Build three short source-target examples around acceptable alternatives: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the acceptable alternatives decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the acceptable alternatives problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying acceptable alternatives challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 48 — model-answer comparison

Identify the dimension. Build three short source-target examples around model-answer comparison: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the model-answer comparison decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the model-answer comparison problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying model-answer comparison challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 49 — error severity

Identify the dimension. Build three short source-target examples around error severity: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the error severity decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the error severity problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying error severity challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Rubric Lab 50 — positive translation quality

Identify the dimension. Build three short source-target examples around positive translation quality: one clearly acceptable, one clearly problematic and one genuinely debatable. Before correcting anything, the learner states whether the issue affects meaning transfer, target-language correctness, writing quality, brief fit, or more than one dimension. The purpose is classification before intervention. A learner who cannot say what kind of quality is at stake will often choose the wrong repair.

Measure consequence. Ask what changes for the target reader because of the positive translation quality decision. Does the reader receive a different fact, role, quantity, condition or level of certainty? Is meaning intact but the target ungrammatical or unnatural? Is the sentence individually acceptable but inconsistent with the rest of the document? Describe impact in words before thinking about any provider-specific point system. This trains severity judgement without inventing official marks.

Compare alternatives. Produce at least two defensible target versions where possible and explain what remains invariant. Then introduce one version that looks fluent but violates the source or brief. This prevents rubric practice from becoming model-answer copying. Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan can arrive at different surface wording while still satisfying the same underlying quality constraint.

Repair the mechanism. If the positive translation quality problem comes from missing knowledge, teach it untimed. If knowledge is present but execution fails, practise the decision under a modest constraint. If it appears only late in a long session, test durability. If it is a target-language issue, use target-only reformulation. The rubric category should route training toward the mechanism rather than become a label pasted onto the final sentence.

Test fresh transfer. Change topic, vocabulary and sentence structure while preserving the same underlying positive translation quality challenge. Remove the exercise label. The learner should recognise the relevant quality dimension without being told where it appears. Save the first attempt and compare it with the corrected version. When the decision survives fresh material, timing and task switching, the rubric has become part of translation judgement rather than a memorised list of error names.

Part III — Eight resident marking cases

Adrian: terminology or research? Adrian chooses the correct institutional term but reaches it through a long, unfocused search. Product marking may find no terminology error; process review still finds a training cost. His next task is resource routing, not vocabulary correction. This distinction matters because an assessment rubric describes the submitted target, while preparation can also inspect how that target was produced.

Jo: accurate but wrong for the brief. Jo preserves every proposition but writes a public notice in language suited to an academic report. The source transfer is strong; communicative fit is weak. Her repair is dual-brief practice and target-genre reading, not relearning the source vocabulary.

Ben: one numeral, different set. Ben copies the digit correctly but changes “no more than six” into “six.” A surface checker might see the same number; a transfer-aware rubric sees a boundary change. He repairs quantity logic, then retests it in unrelated topics.

Aisha: fluent overstatement. Aisha writes elegant target prose but turns “may contribute” into “causes.” The sentence sounds strong and natural, yet evidence strength has changed. Her repair preserves writing quality while strengthening modality and attribution control.

Ryan: locally correct, globally incoherent. Ryan translates each sentence plausibly but uses two target terms for one institution and attaches a later pronoun to the wrong group. The case demonstrates why cohesion and reference require text-level review rather than sentence-by-sentence marking alone.

Mira: model difference without error. Mira’s target uses a different natural construction from the teacher’s model. Source meaning, register and target grammar all hold. Rubric literacy protects her from rewriting a sound answer merely to imitate surface wording.

Clara: mechanical error becomes semantic. Clara’s punctuation choice changes which clause a condition modifies. What begins as punctuation now changes reader understanding. The case shows why classification can depend on consequence rather than the visible form of the mistake.

Ethan: positive quality worth preserving. Ethan recasts a difficult source sentence into clear natural target syntax while preserving emphasis and logical relationships. Review should record this as a strength. A marking system used educationally should teach the learner what to keep, not only what to eliminate.

Rubric FAQ

Does every different wording lose marks? No. Translation normally permits multiple defensible renditions. The provider’s criteria, source meaning, target conventions and brief determine acceptability.

Can I calculate my official score with a homemade rubric? You can use local categories for learning, but do not present them as official scoring. Provider grading can include trained assessors, passage-specific guidance and procedures unavailable in a classroom simulation.

Are meaning errors always worse than language errors? Consequence depends on the provider’s framework and the particular error. In ATA’s current system, meaning-transfer errors can carry a wider range of error points, while mechanics and writing-quality categories have specified limits. Check the live framework rather than generalising this to every credential.

Should I memorise all error codes? Only if the provider’s framework makes that useful. Educationally, understanding the distinctions matters more: what changed, why it matters, and what practice repairs the mechanism.

How should teachers use rubrics? Mark consistently, distinguish error from preference, explain consequence, group recurring mechanisms and choose fresh transfer tasks. Avoid converting every model difference into a correction.

Official-source boundary and closing principle

ATA currently publishes a transparent grading architecture: two graders initially evaluate each exam; disagreement can trigger a third grader; both graded passages must pass; the published framework defines error categories and impact-sensitive point values; and separate review and appeal procedures exist. These details are ATA-specific and can change, so candidates should use ATA’s current grading page, error-category page, and review and appeal page for operational decisions.

For any other credential, use that provider’s current published criteria. This article deliberately refuses to invent one global point scale. Its transferable contribution is the reasoning architecture: classify the quality dimension, inspect consequence, distinguish error from acceptable variation, locate the mechanism, repair it at the right scale, and test it again on fresh material.

A rubric is useful when it makes translation judgement more precise without making translation more mechanical. Preserve meaning, write real target language, respect the brief, understand consequence, and let official scoring remain with the official assessor.

Final marking checklist for independent practice

When marking a fresh practice translation, begin with the brief and source rather than the model answer. Read the target once for overall function, then compare source propositions to target propositions. Mark only issues you can explain. For every proposed error, state the dimension, the source or target evidence, the consequence for the reader and the repair mechanism. If two target versions are both defensible, record them as alternatives rather than manufacturing a winner. If you are uncertain, mark the decision for review instead of pretending certainty.

After marking, group the findings. One grammar slip, three quantity-boundary errors and two terminology inconsistencies are not six equal study tasks. The quantity errors may form one recurring transfer family; the terminology problems may share a control mechanism; the isolated grammar slip may need only monitoring. Preserve successful decisions too: a strong recast, precise register or consistent terminology tells the learner what should remain stable.

Then choose the smallest practice capable of changing the next attempt. Teach unknown knowledge untimed. Drill a recurring relationship in varied sentences. Put it into a fresh paragraph. Add representative timing only after correctness. Use a mock only when the unresolved question concerns integration across the whole session. This is how a rubric becomes a learning instrument rather than a post-mortem code sheet.

Finally, keep local marking humble. A teacher, peer or self-reviewer can build excellent diagnostic evidence without claiming to reproduce an external credential’s official score. Official assessors may use trained graders, passage-specific guidance, moderation and procedures that local practice does not possess. The educational job is different: make the learner’s translation decisions more accurate, natural, evidence-sensitive and transferable. When that happens, rubric literacy has done its work.