Closest conjunct agreement in translation occurs when an agreement target—such as a verb, adjective, participle or honorific marker—matches one member of a coordinated phrase rather than using a single feature set resolved from the whole coordination. Languages including South Slavic languages, Slovenian, Xhosa, several Romance varieties and others have provided important evidence that agreement with coordinated nouns can depend on proximity, linear order, structural position or construction type. Related strategies include first conjunct agreement, last conjunct agreement, closest conjunct agreement, resolved agreement and default agreement.
People searching for closest conjunct agreement, first conjunct agreement, last conjunct agreement, coordination agreement, gender resolution, agreement with coordinated subjects are solving a controller-selection problem. A phrase meaning the mother and the father can trigger plural agreement resolved from both nouns in one language, masculine plural resolution in another, or gender agreement with whichever conjunct is structurally or linearly closest to the agreeing adjective or verb. Translation therefore cannot assume that source agreement morphology describes the whole group in the same way target agreement does.
This article owns the coordination-agreement and closest-conjunct precision layer for eduKateSG. It does not replace Grammatical Gender and Agreement, Agreement Attraction, Coordination itself, or Anti-Agreement Effects. Its narrow task is to distinguish true closest-conjunct agreement from performance errors, separate number resolution from gender/person resolution, track conjunct order, handle prenominal versus postnominal adjectives and preverbal versus postverbal predicates, preserve noun-class and honorific systems, and rebuild a target where agreement may resolve differently without changing who belongs to the coordinated group.
Quick Read
Write CONJUNCT 1 FEATURES, CONJUNCT 2 FEATURES, WHOLE-COORDINATION FEATURES, AGREEMENT TARGET POSITION and ACTUAL SOURCE AGREEMENT. Then ask whether the target matches the whole phrase, the nearest conjunct, the first conjunct, the last conjunct, or a default/resolved form.
One-Sentence Answer
Translate closest-conjunct agreement by preserving the coordinated participants and source controller pattern while allowing the target language to recompute agreement from its own coordination rules.
Why This Precision Layer Matters
Agreement is normally used as evidence for reference. Coordination complicates that shortcut. A feminine adjective next to mother and father may not mean the whole group is grammatically feminine; it may simply agree with the closest feminine conjunct. A singular verb beside one conjunct may coexist with a semantically plural coordination under a language-specific agreement strategy. Precision requires separating group meaning from agreement controller.
What Coordination Agreement Is
When two or more noun phrases are conjoined, an agreeing element must determine which person, number, gender, noun-class or honorific features to realize. Languages solve this in several different ways.
The main translation risk is assuming plural coordination always yields one universal plural agreement form. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to list each conjunct’s features and the target’s realized features. MASC noun + FEM noun can trigger masculine resolution, feminine CCA, default form or another pattern depending on language. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate the group meaning first and agreement strategy second. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Resolved Agreement
Resolved agreement computes features for the coordination as a whole rather than copying one conjunct.
The main translation risk is treating a resolved masculine or plural form as if it belonged to one noun. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to derive the group-level feature bundle explicitly. Many Romance grammars prescribe masculine plural agreement for mixed-gender human nouns. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, allow the target to use its own resolution rules. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Closest Conjunct Agreement
Closest conjunct agreement matches the conjunct nearest to the agreement target under a language-specific notion of proximity.
The main translation risk is assuming the nearest noun is merely causing an error. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to test whether the pattern is systematic and licensed in the language. A prenominal adjective may agree with the first noun because that noun is closest on the relevant side. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve controller logic without labelling it a mistake. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
First Conjunct Agreement
Some constructions favor agreement with the first conjunct regardless of overall plural semantics.
The main translation risk is interpreting first-conjunct singular agreement as a singular subject. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to keep semantic coordination separate from agreement controller. N1.SG and N2.SG can still denote two participants even if the verb matches N1. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate both participants and use target plural if required. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Last Conjunct Agreement
Other languages or constructions show agreement with the final conjunct.
The main translation risk is using final-conjunct features to characterize the whole group. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to identify last-conjunct agreement independently. Korean honorific agreement with coordinated subjects has been analyzed as sensitive to the last conjunct in specific patterns. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve group membership and honorific relation. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Number Resolution
Two singular conjuncts often yield plural semantic reference, but source agreement may still be singular under some strategies.
The main translation risk is equating singular agreement with singular participant count. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to count conjuncts before reading verbal number. Maria and João are two people even if a source verb matches one singular conjunct. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve plural group reference in target syntax. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Gender Resolution
Mixed-gender coordination can trigger masculine, feminine, neuter, default or closest-conjunct agreement depending on language.
The main translation risk is projecting source agreement gender onto every member of the group. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to track lexical gender per conjunct and agreement gender separately. A feminine adjective under CCA can modify a mixed-gender coordination. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, do not infer that the masculine conjunct disappeared. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Person Resolution
First, second and third person conjuncts can combine under person-hierarchy rules that differ from closest-conjunct gender agreement.
The main translation risk is using gender-controller logic for person agreement. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to write each conjunct’s person and derive the language’s resolution hierarchy. I + you may trigger first-person plural in many systems. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve inclusive participant set and target person resolution. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Noun-Class Resolution
Bantu and other noun-class systems can resolve or copy noun-class features under coordination.
The main translation risk is reducing noun class to biological gender. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to record noun class as a grammatical feature independent of sex. Xhosa coordinated nouns can provide evidence about closest-conjunct agreement in class features. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate lexical referents while using class only as parsing evidence where target lacks it. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Honorific Agreement
Coordination can mix honorific and nonhonorific conjuncts, creating special agreement choices.
The main translation risk is translating honorific agreement as plural number. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to separate HONORIFICITY from NUMBER. Korean verbal honorific morphology can be sensitive to the final conjunct in coordinated subjects. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve respect toward the relevant referent without changing group size. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Prenominal Adjectives
Adjectives before a coordination can favor agreement with the nearest following conjunct in languages that allow CCA.
The main translation risk is treating adjective gender as resolved over the whole coordination. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to note adjective position relative to each noun. Contemporary French data show strong acceptance of closest-conjunct agreement with prenominal adjectives. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve adjective scope over both nouns even if target agreement differs. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Postnominal Adjectives
Adjectives after coordinated nouns may show different preferences from prenominal adjectives because linear and structural distances change.
The main translation risk is assuming the same controller in both adjective positions. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to compare pre- and postnominal minimal pairs. French postnominal patterns can show more competition between resolution and CCA than prenominal patterns. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate adjective scope and let target adjective agreement recompute. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Preverbal Agreement
A verb before coordinated subjects may agree with the first or nearest conjunct in some languages.
The main translation risk is reading singular preverbal agreement as evidence that only one subject participates. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to identify the entire postverbal coordination. Verb.SG N1.SG and N2.SG can still describe two coordinated actors. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate plural semantics where target requires it. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Postverbal Agreement
A verb after coordinated subjects may have access to a different nearest conjunct or favor resolved agreement.
The main translation risk is assuming word order has no effect on source agreement choice. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to compare controller distance before and after reordering. Moving the coordination can alter which conjunct is closest to the agreement target. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, avoid reordering before the agreement analysis is complete. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Linear vs Structural Proximity
Research debates whether CCA reflects linear closeness, hierarchical structure, derivation, or interactions among these factors.
The main translation risk is turning one theoretical explanation into translation meaning. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to treat the observable controller pattern as primary. Slovenian and South Slavic data have motivated both linear and structural analyses. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate agreement facts without requiring a settled theory. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Conjunct Order
Swapping conjunct order can change closest-conjunct agreement while leaving group membership constant.
The main translation risk is interpreting agreement change as a semantic change in the group. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to compare N1+N2 with N2+N1. FEM+MASC versus MASC+FEM can yield different agreement forms under CCA. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve the same two referents and any discourse effect of order. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Semantic Plurality vs Morphological Agreement
A coordinated phrase can be semantically plural even when the agreement target is singular or copies one conjunct.
The main translation risk is letting morphology override obvious coordination semantics. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to count referents independently of verb form. Two singular names joined by and usually denote more than one participant. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, use target plural agreement if target grammar requires it. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Human vs Nonhuman Coordination
Some languages treat human mixed-gender coordination differently from nonhuman coordination.
The main translation risk is using one resolution rule across animacy domains. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to record humanness or animacy where source grammar uses it. French agreement preferences differ depending on human versus nonhuman coordinated nouns in experimental work. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve lexical referents while applying target-specific agreement. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Agreement Attraction vs Grammatical CCA
Agreement attraction is a processing error in which a nearby noun wrongly influences agreement; CCA can be a legitimate grammatical strategy.
The main translation risk is correcting grammatical CCA as if it were a typo. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to look for systematic acceptability and language-specific licensing. Contemporary French CCA should be distinguished from classic attraction errors. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, respect grammatical source patterns. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Syncretic Agreement Forms
An agreement form can be compatible with more than one controller because gender, number or class distinctions are syncretic.
The main translation risk is claiming CCA when morphology does not uniquely identify the controller. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to list all feature bundles compatible with the form. A plural adjective may look identical across genders. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve ambiguity in the analysis. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Default Agreement
Some languages use a default form when coordination features conflict or cannot resolve normally.
The main translation risk is assigning default morphology to a real third-person singular controller. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to test whether the same form appears with expletive or unvalued agreement. A default masculine/neuter/singular form can signal failed or neutral resolution. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate group meaning independently. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Mixed Number Conjuncts
Singular and plural conjuncts can create complex resolution patterns.
The main translation risk is using the closest noun’s number as group cardinality. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to compute semantic union first. one teacher and several students form a plural group regardless of which conjunct controls morphology. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve participant count. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Mixed Person Conjuncts
Coordinating speaker, addressee and third persons can trigger person hierarchies.
The main translation risk is dropping one participant because the resolved pronoun looks first plural. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to expand the participant set. I and she → we in many target systems, but source agreement may follow different rules. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve all conjuncts and target person semantics. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Correlative Coordination
Both…and, either…or, neither…nor and similar structures can have agreement rules different from simple and-coordination.
The main translation risk is applying ordinary CCA universally. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to identify conjunction type. Either the teachers or the principal may trigger proximity agreement in some English styles. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve logical relation and target agreement convention. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Disjunction
Or-coordination often differs from and-coordination in semantic number and agreement.
The main translation risk is assuming all conjoined phrases denote a plural group. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to ask whether the conjunction is additive or alternative. A or B refers to alternatives rather than necessarily a joint group. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate agreement and logic separately. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Coordination of Unlike Cases
Recent research also shows that coordinated expressions need not always share identical case or category features, so agreement must be read in the full construction.
The main translation risk is rejecting a source coordination merely because conjunct morphology differs. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to check whether each conjunct satisfies its own grammatical constraints. Polish and Estonian evidence demonstrates coordination across unlike cases under certain structures. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, translate functional relations rather than imposing uniform source morphology. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Predicative Adjectives
Predicative adjectives can resolve features differently from attributive adjectives because they stand farther from individual conjuncts.
The main translation risk is copying an attributive CCA rule into predicative syntax. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to compare target position and source construction. Predicative agreement often favors resolution where prenominal adjectives favor CCA. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, generate target predication from group meaning. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Participles
Participles can agree with coordinated subjects or objects in gender and number and show CCA or resolution.
The main translation risk is treating participial gender as semantic gender of the whole group. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to identify controller options and voice role. South Slavic participles are central evidence in coordination-agreement research. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve event participants and target participial agreement. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Translation into English
English has limited gender agreement but does resolve number/person in coordinated subjects and sometimes shows proximity effects with disjunction.
The main translation risk is trying to reproduce source gender agreement visibly. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to use source morphology mainly as parsing evidence. A source feminine CCA adjective can become an invariant English adjective. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, let grammatical gender disappear while preserving noun reference. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Translation into a Gendered Target
A target with grammatical gender may require a resolved form different from the source CCA form.
The main translation risk is copying source adjective gender because it feels more faithful. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to recompute agreement over the target coordination. Source nearest-feminine agreement may map to target masculine plural resolution. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve people and modifier scope, not source ending shape. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
AI Nearest-Noun Bias
Models can produce accidental agreement attraction and then mistake their own error for grammatical CCA.
The main translation risk is using proximity without checking source grammar. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to separate licensed CCA from generic next-word prediction. A nearby feminine noun should control agreement only if the source construction allows it. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, validate against language-specific rules. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
AI Resolution Bias
Models trained on dominant standard-language norms can normalize legitimate CCA into resolved agreement.
The main translation risk is erasing dialectal or grammatical source structure. Coordination is deceptive because a verb, adjective, participle or honorific marker may agree with the coordination as a whole, with a resolved feature bundle, or with one particular conjunct. A target can sound smooth while changing which noun controlled agreement.
A practical diagnostic is to recognize systematic source alternatives before editing. Contemporary French CCA can be well accepted despite prescriptive masculine-resolution norms. The agreement pattern should be analysed before normalizing word order, because changing conjunct order can change the closest controller in languages where linear or structural proximity matters.
For recovery, preserve source grammar in metalinguistic or stylistically relevant translation. Preserve participant reference and the source’s coordination structure first, then generate agreement according to target grammar rather than copying source gender or number endings mechanically.
Worked Examples
Worked Example 1: Two Singular Conjuncts, Singular Source Verb
Source problem: A source language allows the verb to agree with the nearest singular conjunct even though two people are coordinated.
Diagnosis: Morphological singular does not reduce semantic group size.
Target strategy: Translate both conjuncts and use target plural agreement if required.
Verification: No participant may disappear because the source verb is singular.
Worked Example 2: Mixed Gender with Prenominal Adjective
Source problem: A feminine adjective precedes feminine N1 + masculine N2 and scopes over both.
Diagnosis: The source permits closest-conjunct agreement rather than requiring masculine resolution.
Target strategy: Preserve adjective scope over both nouns; recompute agreement in the target.
Verification: Feminine source morphology must not be interpreted as excluding the masculine conjunct.
Worked Example 3: Conjunct Order Reversed
Source problem: Reversing N1 and N2 changes the agreement form.
Diagnosis: Agreement is sensitive to proximity/order while group membership remains constant.
Target strategy: Preserve order if discourse matters, but do not create a new participant set.
Verification: The target should represent the same two referents.
Worked Example 4: Honorific Last-Conjunct Agreement
Source problem: A coordinated Korean subject contains one honorific and one nonhonorific conjunct, and verbal honorific marking tracks the final conjunct under the relevant pattern.
Diagnosis: Honorific agreement is sensitive to conjunct order, not plural number.
Target strategy: Preserve respect toward the relevant referent using target honorific or register resources.
Verification: Do not turn honorific morphology into plural reference.
Practical Diagnostic Workflow
Build a COORDINATION AGREEMENT TABLE with CONJUNCT 1 PERSON/NUMBER/GENDER/CLASS/HONORIFICITY, CONJUNCT 2 FEATURES, COORDINATION TYPE, TARGET POSITION, SOURCE AGREEMENT FORM, POSSIBLE CONTROLLERS, RESOLVED FEATURES and WORD ORDER. Mark whether the observed form uniquely diagnoses CCA or is compatible with both CCA and resolution.
Then run the order-reversal test. Swap the conjuncts conceptually while keeping the same referents. If the predicted agreement changes, the language may use first or last or closest-conjunct agreement. If the form stays constant, resolved or default agreement may be more plausible.
Verification Method
From the target, list every coordinated referent and verify their number, person, gender or honorific properties independently of the source agreement ending. Confirm that the modifier or predicate still applies to the same coordination. A target is correct even when its agreement morphology differs radically from the source because target resolution rules are different.
Recovery Method
If a singular source agreement form caused a participant to disappear, rebuild the coordination first. If source gender was copied mechanically into a target with different resolution rules, regenerate agreement from target features. If legitimate CCA was normalized during editing, restore it only when the source grammar or metalinguistic content requires the original agreement strategy to remain visible.
Practice Sequence
Create coordinated pairs with MASC+FEM, FEM+MASC, SG+SG, SG+PL, 1P+3P and honorific+nonhonorific features. Place the agreement target before and after the coordination. Generate four possible strategies: resolved agreement, first-conjunct agreement, last-conjunct agreement and default agreement. Translate each into English and one hypothetical gendered target.
Quality-Assurance Checklist
- List every conjunct before reading agreement morphology.
- Separate semantic plurality from morphological number.
- Distinguish resolved agreement from CCA.
- Distinguish first- from last-conjunct agreement.
- Track conjunct order.
- Track target position before or after coordination.
- Track gender and number separately.
- Track person-resolution hierarchies separately.
- Track noun class where relevant.
- Track honorificity independently of number.
- Distinguish grammatical CCA from agreement attraction.
- Check syncretic forms before claiming a controller.
- Treat disjunction separately from conjunction.
- Audit AI nearest-noun bias.
- Audit AI standard-resolution bias.
Frequently Asked Questions
What is closest conjunct agreement?
An agreement strategy in which an adjective, verb, participle or other target matches the conjunct closest to it rather than a feature bundle resolved from the full coordination.
Is closest conjunct agreement an error?
Not necessarily. In many languages and constructions it is a licensed grammatical pattern and should be distinguished from accidental agreement attraction.
Can two singular nouns trigger singular agreement even though they denote two people?
Yes under some first-, last- or closest-conjunct strategies. Semantic group size and morphological agreement can differ.
What is resolved agreement?
A strategy that computes person, number, gender or class features for the whole coordinated phrase.
Does word order matter?
In many CCA systems, yes. Reversing conjunct order can change which noun is closest to the agreement target.
Can adjectives show different agreement from verbs?
Yes. Attributive position, predicative position and verbal agreement can favor different strategies.
How should English translate source gender CCA?
Usually by preserving the coordinated nouns and modifier scope; English adjectives normally do not need to reproduce source gender morphology.
Why does AI fail?
Models can either mistake proximity-driven prediction errors for grammar or normalize genuine CCA into the standard resolved form they have seen most often.
Internal Routes Through the eduKate Translation Architecture
Route ordinary person, number and gender agreement to Grammatical Gender and Agreement: https://edukatesg.com/2026/09/18/translate-grammatical-gender-agreement-reference-gender-number-person-differently/
Route extraction-related loss of agreement to Anti-Agreement Effects: https://edukatesg.com/2026/09/19/translate-precisely-anti-agreement-effects-subject-extraction-wh-relatives/
Route remote controllers to Long-Distance Agreement: https://edukatesg.com/2026/09/19/translate-precisely-long-distance-agreement-cross-clausal-embedded-object-features/
Route broad clause and phrase order to Sentence Structure: https://edukatesg.com/2026/09/18/translate-sentence-structure-word-order-grammar-meaning/.
Final Principle
Agreement with a coordination is not a direct count of who is present. Preserve the full participant set first, identify whether the source resolves features or copies one conjunct, and let the target recompute agreement under its own grammar.