subagent-factory

Review the applied use of corpus tools and textual analysis — keyword, concordance, and collocation analysis, description of existing translations, and tracking of textual devices (theme dynamics, clausal and iconic linkage, deixis) — set against the commission's purpose and risk level. Use when concrete tooling or textual-device tracking is under review; corpus design choices go to corpus-design-and-methodology.

Review Chinese target prose for Europeanized distortion — defending concision, flexible syntax, and rhythmic force, matching connectors to the actual relation, and preferring concrete subjects over abstract nominalizations. Use when a Chinese target text's Europeanization is under review; translationese used as a quality proxy goes to translation-universals-and-the-third-code.

Review how cognitive and contrastive-pragmatic evidence is used — admitting process research only where it clarifies comprehension or transfer, treating verbal reports and neuroimaging as limited evidence, and triangulating pragmatic claims across data types. Use when a cognition or contrastive-pragmatics claim underlies a quality judgement; the descriptive-norm reconstruction it informs goes to descriptive-studies-and-translational-norms.

Review whether a corpus-based translation study is built to answer its question — corpus type (parallel / monolingual comparable / multilingual comparable), comparability, metadata, alignment, and frequency normalisation — before any difference is attributed to translation status. Use when a corpus design or corpus statistic is under review; universals interpretation goes to translation-universals-and-the-third-code, applied tooling to applied-corpus-tools-and-textual-devices.

Review quality where culture, ideology, English-as-a-lingua-franca, and globalization bear on the choices, grounding any power or ideology critique in detailed linguistic form and function first. Use when a cultural-filter, ideology, or ELF/globalization claim is under review; the overt/covert filtering judgement goes to overt-covert-translation-and-equivalence.

Review whether translated text is studied descriptively and target-orientedly — a reproducible method, norms reconstructed from recurrent corpus regularities, and an individual pattern attributed to a translator only after author style, source-language, and sociolect are ruled out. Use when a descriptive method or a translational-norm claim needs review; universals go to translation-universals-and-the-third-code.

Review the discipline of the evaluation itself — overt errors classified separately from covert dimensional mismatches, impressionistic criteria rejected unless made operationally checkable, and evidence strength preserved. Use when evaluation criteria, an error typology, or a confidence claim need checking; the source-profile basis they weight against goes to overt-covert-translation-and-equivalence.

Review quality where genre and audience accessibility govern the choices — connecting a text to its genre class and checking children's and stage texts against the function they must serve, with accessibility claims backed by quantifiable textual features. Use when a genre-fit or accessibility claim needs review; register detail goes to register-field-tenor-mode-analysis.

Review a House-model quality assessment — is the source-text profile and function specified, is the case overt (source kept intact) or covert (cultural filter for target pragmatics), and has unjustified filtering produced a covert version. Use when an overt/covert classification or a functional-equivalence claim needs situating; register detail goes to register-field-tenor-mode-analysis, error typing to error-analysis-and-evaluation-discipline.

Review register fidelity through the Field/Tenor/Mode split, source-target comparative, with attention to mission statements and comparably exhortative institutional texts. Use when Field/Tenor/Mode assignment or a register shift needs review; the overt/covert judgement it feeds goes to overt-covert-translation-and-equivalence.

Review corpus-based translation and interpreting work in the Russian field, where corpus scarcity and fragmentation are hard design constraints — pressing for multifactorial modelling, significance testing with effect sizes, metadata-rich corpora, and inclusion of the Russian research tradition. Use when a Russian-field corpus or interpreting study is under review; general corpus design goes to corpus-design-and-methodology.

Review claims about translation universals (explicitation, simplification, normalisation) and the third code — checking they are treated as measurable hypotheses, not presumed features, and kept distinct from translationese used as a quality proxy. Use when a universals or third-code claim needs method and faithfulness review; corpus design for such a study goes to corpus-design-and-methodology.

Reviews a translation, translation choice, or translation-studies analysis for its grounding in culture, ideology, power, and institutional rewriting — applying refraction, patronage, feminist, postcolonial, and Bourdieusian lenses. Owns the institutional/agent, reception, and ideology judgment; the fluency-illusion and domesticating/foreignizing-axis judgment routes to the same-package `domestication-foreignization-and-visibility`. It critiques the translation or analysis; it does not produce the finished translation or decide publication.

Reviews a literary translation, translation choice, or translation-studies analysis for Berman's deforming tendencies, subtractive and paratextual losses, over-clarification, and normalization; this skill critiques the translation and does not produce the finished translation or make the publication decision.

Reviews a translation, translation corpus, or translation-studies analysis for whether it applies an explicit, systematic, target-oriented descriptive method and reconstructs translational norms from patterned behaviour rather than asserting or reading them off directly. This skill critiques the method and the norm claims; it does not produce the finished translation or make the publication decision.

Reviews where a translation sits on the domestication-foreignization axis and whether its fluency or readability masks the translator's presence. Owns the fluency-illusion and domesticating/foreignizing-axis judgment; the institutional/agent, reception, and ideology judgment routes to the same-package `culture-ideology-power-and-rewriting`. On the shared free-versus-literal axis, the modern ideological domestication and fluency concern stays here; the classical rhetoric-based literal/free and metaphrase-paraphrase strategy history routes to the same-package `literal-free-strategy-history-and-retranslation`. Review-only — does not produce the finished translation or make the publication decision.

Reviews a translation's formal-versus-dynamic equivalence orientation and its equivalent-effect or "same response" claims for fit to purpose, text type, and audience. Reviews that turn on the linguistic equivalence mechanism itself (word, collocation, idiom, grammar, or information-structure equivalence) route to the sibling `translation-equivalence-advisor`; this skill stays on the formal-versus-dynamic equivalence orientation and equivalent-effect claims. Within this package, orientation/effect fit stays here; the theory-of-meaning premise routes to `meaning-signification-and-equivalence-critique` and procedure/shift naming to `translation-procedures-and-shifts`.

Reviews a translation's interpretive stance, its treatment of untranslatability, or a claim of full determinacy or perfect equivalence against hermeneutic and philosophical grounding (Steiner's fourfold motion, Quine's indeterminacy, Gutt's relevance theory); this skill reviews a translation or translation-studies analysis, it does not produce the finished translation. Tie-breaker — the word- and sign-level signification premise (naive cross-code identity, dictionary-equivalence trust) routes to the sibling `meaning-signification-and-equivalence-critique`; this skill stays on the interpretive and philosophical grounding, indeterminacy of reference, and untranslatability.

Reviews a translation's placement on the metaphrase-paraphrase-imitation and literal-free axes, its verse prosody and historical-register choices, its scholarly apparatus, and its reading against retranslation and a work's afterlife; this skill reviews an existing translation, translation choice, or translation-studies analysis and does not produce the finished translation. On the shared free-versus-literal axis, the classical rhetoric-based literal/free and metaphrase-paraphrase strategy history stays here; a modern ideological domestication or fluency concern routes to the same-package `domestication-foreignization-and-visibility`.

Reviews the theory of meaning and signification underlying an equivalence claim — the critique of naive cross-code identity, dictionary-equivalence trust, and the semiotic nature of meaning across word, grammar, and pragmatic levels — as distinct from producing a translation. Tie-breaker — a review that turns on the practical linguistic equivalence mechanism itself (choosing a word, collocation, idiom, grammatical, or information-structure equivalent) routes to the sibling `translation-equivalence-advisor`; this skill stays on the theory of meaning, not the mechanism of producing an equivalent. Within this package, the theory-of-meaning premise stays here; formal-versus-dynamic orientation/effect fit routes to `equivalence-orientations-and-effect` and procedure/shift naming to `translation-procedures-and-shifts`. The indeterminacy-of-reference and hermeneutic-motion critique (Quine's inscrutability, Steiner's fourfold motion) routes to `hermeneutics-and-the-limits-of-translatability`; this skill stays on the

Reviews a translation's register and discourse choices on the Hallidayan field/tenor/mode model, its cohesion and pragmatic linkage, politeness and face, transitivity and modality shifts, House's overt/covert axes, its audiovisual (AVT) signifying-code analysis, and its subtitling-specific shortening constraints (it does not carry dubbing isochrony/lip-sync or voice-over criteria); it reviews the analysis, it does not produce the finished translation, subtitle file, or publication decision.

Reviews a translation, translation choice, or translation-studies analysis against an explicit brief and the source's predominant text-type function, and against the skopos hierarchy (purpose, then intratextual coherence, then source fidelity); this skill reviews the reasoning behind a translation, it does not produce the finished translation or make the publication decision.

Reviews micro-level translation decisions using the precise metalanguage of translation procedures and shifts — the Vinay & Darbelnet ladder (direct — borrowing, calque, literal; oblique — transposition, modulation, equivalence, adaptation), Catford's level and category shifts, and the servitude-versus-option distinction between obligatory and optional shifts. When a review's core question is which linguistic equivalent to choose (word, collocation, idiom, grammar, or information-structure equivalence), route it instead to the sibling translation-equivalence-advisor; this skill stays on naming and diagnosing the procedure or shift itself, not on selecting among candidate equivalents. Within this package, procedure/shift naming stays here; formal-versus-dynamic orientation/effect fit routes to `equivalence-orientations-and-effect` and the theory-of-meaning premise to `meaning-signification-and-equivalence-critique`.

Reviews translation-quality judgments and applied or empirical translation-studies work for evaluation-method soundness, discipline-map placement, and research-design rigor. Requests to run corpus-based quality metrics or QA scoring and return the scores route to the sibling `translation-quality-reviewer`; this skill reviews only whether the evaluation method itself is sound. It critiques and reports findings and does not itself finalize a quality score, a translation, or a publication decision.

Diagnoses and reviews the surface cohesive network — reference, substitution, ellipsis, conjunction, and lexical cohesion — reworked to the target language's own preferences; owns the surface ties that bind a text, not the reader's coherence judgement, which pragmatic-equivalence-coherence-and-implicature owns. Use when a caller says the pronouns or connectors feel off, the references don't track, or the text reads choppy.

Diagnoses collocation, idiom, and fixed-expression non-equivalence and reviews a chosen rendering — typicality, markedness, the two idiom traps, and register-specific collocability — owning multi-word patterning, not the single-word lexical gap that word-level-nonequivalence-and-strategies owns. Use when a caller asks whether an idiom or phrase sounds natural in the target, or how to render a saying or fixed expression.

Reviews the choice between formal and dynamic equivalence for a brief and audience — receptor response, cultural adjustment, audience decoding capacity, and adequacy evaluation; owns the equivalence-orientation decision, not the register and literary form that register-style-and-literary-form owns. Use when a caller asks whether a rendering should be literal or free, or formal or dynamic for this brief.

Diagnoses and reviews rendering across grammatical systems — obligatory categories, voice, gender, number, tense and aspect, modality, and pronouns of address — by function rather than form; owns grammar-driven shifts, not information structure, which thematic-and-information-structure owns. Use when a caller asks how to handle a tense, aspect, gender, number, voice/passive, or politeness form that the target forces.

Diagnoses and reviews meaning in use — the reader's coherence judgement, implicature, speech-act force, background knowledge, and politeness — checking a rendering does not create wrong implicatures; owns inferred and interactional meaning, not the surface ties that cohesion-and-texture owns. Use when a caller asks whether readers will misread the implication, or whether the rendering lands the intended meaning or politeness.

Diagnoses and reviews register, style, and form-bound material — field/tenor/mode, marked language, poetry, song, and sound effects — weighting form by its communicative function; owns style and the form/content trade-off, not the formal-vs-dynamic orientation that dynamic-and-formal-equivalence owns. Use when a caller asks whether the tone or formality is right, or how to translate a poem, song, or pun.

Diagnoses and reviews the whole-text approach and the standard applied — systematic reflection, guarding against interference, and treating equivalence as relative and partial; owns the overarching stance and evaluation frame, not any single linguistic level. Use when a caller asks whether a translation is "right," "literal enough," or "faithful" overall without naming a single equivalence level.

Diagnoses and reviews information flow — theme and rheme, given and new, markedness, and functional sentence perspective — matching each element's status by the target's own devices; owns clause-level information dynamics, not the surface cohesive ties that cohesion-and-texture owns. Use when a caller asks what should come first, or says the emphasis or word order feels wrong.

Diagnoses word-level non-equivalence — a culture-specific, unlexicalized, or semantically complex item with no direct target equivalent — and reviews the chosen rendering strategy against context and purpose; owns lexical meaning and the single-word/short-phrase gap, not collocation and idiom patterning, which collocation-idiom-and-fixed-expression owns. Use when a caller says there is no target word for a term and asks whether to keep, gloss, or substitute it.

Establish who will read the target text, how they will use and receive it, and the purpose (Skopos) the commission sets, from a translation brief and a document profile — before any wording decision.

Translate recurring, semantically identical information with one uniform target-language construction (iconic linkage), and scale that consistency through controlled language, style guides and translation memory where it fits.

Hold documentation to legal, safety, accuracy and style-guide quality; treat warnings, brand names and EU-market translation as high-stakes; and revise, notify the client, and conduct professional practice with care.

Treat layout, typography, white space, structure, graphics/screenshots, the table of contents and space constraints as communicative usability factors, and build modular, navigable, task-centred guides.

Shape the target text around how readers perceive, remember, attend, learn and solve problems: minimise processing effort and cognitive load so the text is easy to read, understand and act on.

Handle the technical precision layer — terminology, Latin nomenclature, SI units and formulae, regulatory specification terms, mandated/regional naming, acronyms and abbreviations — with the right resources.

Recognise a text's type and genre — technical vs scientific, data sheet, manual, paper, abstract, presentation, case study, reference or hybrid — and translate to its conventions and communicative purpose.