29.1A translation and its source
A translation should be treated as a derived text, not as a free-standing question that happens to resemble another. Its claim to accuracy is relative to a particular source version from which it was produced. That qualification does most of the work here: the relevant comparison is never between an abstract question and a rendering, but between one wording on record and a text derived from it. Once the source version is fixed, differences can be sorted into literal wording, pragmatic purpose, locally contested terms, and the translator's or editor's interpretation, instead of being lumped together as "the meaning."
The first invented English example, referred to below as T01, makes the point without pretending to supply a bilingual case. The stipulated source asks, "Is the report available?" with context leaving open whether "available" means that the report appears in the catalog or that the intended reader can inspect it. Two possible reformulations are "Does the catalog contain the report?" and "Can the intended reader inspect the report?" These are stipulated English readings proposed for a future translation-review design; they are not a validated translation between languages and not evidence about any actual language pair. Their value is that each rendering selects one reading that the source wording leaves undecided.
From this, a modest inference follows. If a translator chooses one rendering and the reviewer compares it with a different reading of the source, the apparent disagreement may concern which source sense was selected rather than whether the translation is faithful to the sense actually chosen. Recording the source wording alongside the chosen sense therefore separates two questions that are otherwise easy to conflate: whether the rendering expresses a defensible reading, and whether that reading was the intended one. The limit is equally modest: this bookkeeping cannot establish that the intended sense was correct, only which interpretation the derived text carries.
Because a translated question is a derived text, its adequacy has to be judged against a particular source version. That seemingly simple requirement has practical consequences. A translation faithful to version v1 can differ from version v2 without mistranslating either. Establishing which version it translates must precede an error judgment.
Suppose, as a stipulated case, that an English source question is revised after a translation has already been produced: its literal wording is tightened, while its pragmatic purpose remains the same. Agreement with the revised wording is then expected only if the translation is itself revised or replaced. Conversely, a translation that departs from the current edited text may faithfully render the version from which it was made. Neither fact establishes that one side is defective without identifying the source used.
The application is therefore to translation review rather than to an abstract question of equivalence. Review should first record which source version generated each derived text. Only then can the reviewer separate changes in literal wording from continuities of pragmatic purpose, from locally contested terms, and from interpretations introduced by editors. Without that ordering, an editorial revision and a translation decision become indistinguishable.
Paired review follows naturally: a translator working on a target question and a source editor reviewing the question it derives from can compare versions side by side. Their task is not to certify that the two texts are interchangeable in every respect, but to identify which relation holds—revision, faithful rendering, deliberate narrowing, or interpretation—and whether that relation suits the intended use. This is a proposed procedure for a future translation-review design, not a validation result from multilingual research. The important limit is that locating versions removes one source of false alarms; it does not guarantee agreement about purpose, terms, or the right translation choice.
29.2Ambiguous terms
When the invented English source asks whether a customer was charged, without further context, the word "charge" carries two declared possible senses: a financial fee imposed on the customer, or an accusation levelled against them. The wording alone does not settle which reading is operative. That stipulation has to be fixed before any comparison across languages begins, because the ambiguity here belongs to the source expression as defined for the exercise, not to a demonstrated feature of any particular language.
The next step is to keep separate what easily collapses together. The literal wording is identical in both readings; the pragmatic purpose may differ, since a billing question and a disciplinary question call for different kinds of answer; and editorial interpretation enters only when someone records which sense was taken, and why. To treat any one layer as the question itself is to mistake the words on the page for what a respondent is being asked to do.
The example carries a clear limit. It is invented English, prepared for a later translation-review design, so it supports neither a claim about a verified bilingual pair nor one about how speakers of any language resolve the ambiguity. No completed translation, no multilingual study, and no evidence about readers stands behind it. If further cases are supplied later, the same caution governs them: conditional additions remain conditional.
What follows from the stipulation is modest but consequential for the chapter's question. Before asking what can change when a question crosses languages, one must establish which version of the question is in play and which sense of its terms was intended. Without that diagnostic work, a difference between versions cannot be located, let alone explained; with it, later review has something definite to examine.
Once the source term carries two declared readings — for "charge," a financial fee or an accusation — the translator's first decision is whether the surrounding material has already settled which one applies. Where it has, selecting that sense is simply accurate translation. The difficulty arises when the alternative stays live: here a choice made without supporting context does not preserve the question but narrows it, converting an open inquiry into a question about one thing.
T01 makes the narrowing concrete in English. The invented source asks, "Is the report available?", with "available" left open between catalogued existence and access by the intended reader. Two invented reformulations separate the strands: "Does the catalog contain the report?" and "Can the intended reader inspect the report?" A rendering that silently adopts the first asks whether a record exists; one that adopts the second asks whether the reader can get to the document. The answers can diverge without either reformulation having mistranslated the words it was given.
Where the ambiguity is doing the work of the question, preserving the alternatives can therefore be preferable to an early commitment: the translator carries both readings forward, marks the point of uncertainty, and lets paired review decide whether the intended sense is recoverable from the context the parties actually share. This is a practical preference, not a rule that every undecided word must survive into the target version. When context supports one reading and the other is idle, resolving the ambiguity is the better choice.
The limit of the example is important. T01 supplies stipulated English reformulations for a later translation-review design; they are not verified bilingual pairs, and no completed multilingual study licenses a broader claim about what carrying alternatives does across languages. The defensible claim is correspondingly narrow. Where a source question leaves a term genuinely undecided, and where the choice of sense would change what is being asked, a single-sense rendering silently reduces the scope of the question. The translator should make that reduction visible rather than let it pass as fidelity to the wording.
29.3Politeness and implied requests
The same separation of wording from purpose becomes visible when an apparently informational question also asks for action. Consider an invented English exchange: A says, "My hands are full," and then asks B, "Can you open the door?" Taken by itself, the second utterance can be read as a question about B's ability. The preceding utterance changes what the exchange is doing: it explains why A cannot act, and it thereby supports reading the question as a request that B open the door.
Two kinds of information now have to be kept apart. The literal wording is an interrogative about capability. The pragmatic purpose, supported by the surrounding utterance, is to get B to perform an action. A reviewer who records only the first will miss the request; a reviewer who records only the second will lose sight of the question's actual words. Both descriptions belong in the record, and the inferential step that connects them is the sequence of the two utterances, not anything carried by the interrogative alone.
The limit is equally specific. This is a stipulated English dialogue, written to display the contrast, and the request reading is one that the stated context is designed to support. It is not an observed exchange, a validated account of how requests are normally formed, or evidence that a translation of the question will carry the same force in another language. Any proposed rendering would therefore be hypothetical: a translation team might, for instance, need to add an explicit request, retain the ability question and rely on context, or combine both, but none of these choices has been tested.
What has been established is local. Surrounding utterances can change the pragmatic work a question performs, and review should record that work separately from the literal question being asked.
The request reading of “Can you open the door?” follows from the invented exchange as a whole: A’s preceding “My hands are full” supplies a reason for asking B to act, while the question itself remains literally an inquiry about ability. Politeness and indirectness therefore arise from how the utterances work together, not from some hidden extra wording. The analysis may attribute that force to this exchange, but it may not attribute it to English generally, much less to a language community or culture. Nothing in the stipulated dialogue shows how often English speakers use ability questions as requests, which contexts favor that reading, or whether other languages package comparable purposes differently.
The immediate application is correspondingly limited. A translation review can preserve the distinction by recording the literal question and the locally supported request purpose as separate entries. Reviewers can then ask whether a proposed rendering does justice to both, rather than treating either the literal wording or the inferred purpose as the whole meaning. This does not establish an equivalent construction in another language; it merely supplies a disciplined way to examine a future candidate.
The same division of labor applies to lexical choice, where the two stipulated senses of “available” mark the narrowing a rendering can introduce, just as the request reading marks the pragmatic work added by context.
On both axes, the useful inference stays local: surrounding utterances can make a question act as a request, and a word can support more than one reformulation. Contributor confirmation can settle which reading is intended in a particular version; until then, the range of readings should remain visible rather than being silently collapsed into one interpretation.
29.4Who may confirm an interpretation
When an English question leaves its referent open, the person who put the question forward can often settle which referent was intended. Take the invented source "Is the report available?", which by stipulation allows either catalogued existence or inspectability by the intended reader. The contributor is shown the two English reformulations prepared for this purpose — "Does the catalog contain the report?" and "Can the intended reader inspect the report?" — and selects the one that captures the question as posed. That selection narrows the stipulated ambiguity to a single intended referent.
The inferential step is modest but useful. Once the contributor confirms that the intended referent was catalogued existence, the editor can carry that confirmation into the next stage of paired review and ask whether a proposed rendering preserves it. Confirmation therefore does its work on the question actually asked: it converts an open reading into a stated intention, against which a translation choice can be checked rather than guessed. Nothing more general follows from it.
The limit deserves stating plainly, because the same example can easily be overread. These two sentences are invented English reformulations prepared for a future translation-review design; they are not a validated translation between languages, and no completed translation research stands behind them. Contributor confirmation settles the intended referent at one point in one reformulation. It does not establish that "available" carries a fixed sense across languages, and it does not remove the need to notice where a proposed rendering departs from the confirmed reading. Any particular translation choice remains explicitly hypothetical until paired review examines it against the confirmed referent and the pragmatic purpose of the exchange together.
This narrower scope is precisely what makes contributor confirmation available at all. Because the claim concerns a particular question at a particular moment, the editor need not grant anyone authority over an entire vocabulary in order to proceed. The harder cases are those in which the intended referent remains unsettled, and they need a different record.
The warrant a contributor supplies is narrower than authority over a word. In T01's stipulated question, a contributor who confirms that "available" was meant to ask whether the catalog contains the report selects "Does the catalog contain the report?" over "Can the intended reader inspect the report?" as the rendering that tracks the intention. That selection narrows the stipulated ambiguity for this question and this task. It does not show that "available" generally denotes catalogued existence, and it does not make the reader-access rendering linguistically defective; the invented source explicitly leaves both open.
Scope accordingly belongs to the confirmation itself. A report of one's own intention is the natural evidence about that intention; where wording was produced collectively, or where the contributor cannot reconstruct what was meant, the proper record is uncertainty rather than a compelled choice. Disagreement is recorded the same way. A community interpreter may judge that context favors reader access, while the contributor dissents or remains unsure. Keeping both judgments, and the reasons for each, preserves the material a later reviewer needs without installing any participant as permanent authority on the uses of "available" in general.
"Charge" makes the limit visible from the other side. Because financial fee and accusation are both explicitly allowed readings of the invented source, a contributor who picks one sense narrows that stipulated ambiguity for the purpose at hand; the choice does not retire the other sense, and an interpreter who hesitates is not thereby correcting the contributor. Confirmation therefore answers a local question — who intended what in this reformulation — while the pair of candidate readings survives to be compared under the stated task. Useful translation need not settle every sense before proceeding; it needs the disagreement and the uncertainty attached to the particular wording they concern.
29.5Different answer spaces
When a question moves between candidate renderings, the practical question is not only whether the wording is close, but which answers would satisfy the task. Take the stipulated English source, “Is the report available?” Context explicitly leaves “available” open between catalogued existence and inspectability by the intended reader. The first candidate rendering, “Does the catalog contain the report?”, calls for an answer about catalog membership: whether the report is listed there. The second, “Can the intended reader inspect the report?”, calls for an answer about access: whether that reader can actually view it.
Under the stated translation-review task, these two renderings therefore define different answer spaces. A response that settles catalog membership may leave access unresolved, while a response that establishes access may need information beyond catalog status. Selecting either rendering narrows the stipulated ambiguity, but that selection is an editorial interpretation of what the review needs, not a discovery about the source wording or about language use in general. If the task concerns retrieval from a collection, the catalog reading may fit; if it concerns whether an intended reader can consult the report, the access reading may fit.
This comparison keeps literal wording, pragmatic purpose, and editor interpretation distinct without pretending that one rendering is the uniquely correct translation. It also marks the limit of the example: these are invented English reformulations for a future translation-review design, not a verified bilingual pair or evidence from completed translation research. The useful test is consequently modest and task-relative: ask which answers each candidate would require, then record the narrowing that the choice introduces.
Comparing two candidate renderings identifies when they license different answers under one stated task. The further question is narrower: which grammatical forms keep a stipulated ambiguity open, and which resolve it without saying so. Take an invented English source question, "Was the charge justified?", in which context leaves charge undecided between a financial fee and an accusation. A fluent question such as "Was the fee justified?" invites a verdict on pricing; "Was the accusation justified?" invites a verdict on blame. Each is idiomatic, and each silently closes one branch of the answer space. By contrast, "In what sense of charge is it justified?" leaves both branches available, but at the cost of requesting clarification instead of a verdict. Form can preserve the alternatives; fluency is no guarantee that it does.
That cost marks the relevant limit. Preserving an answer space is not always the right aim. If the stipulated task calls for a decision on the fee alone, the narrowing question is the fitting one, and its narrowing is a feature of the task rather than a defect of the rendering. The criterion is therefore task-relative: ask which answers would satisfy the stated task, then ask whether the rendering still admits exactly those answers. A form that keeps too much open can become unusable for a task that demands a single decision, just as a form that closes too early can make a contested reading invisible.
The same test applies where pragmatic purpose drives the exchange. In an invented English scene, A says "My hands are full" and asks B, "Can you open the door?" The preceding utterance supports a request reading here. A rendering that demands only a yes-or-no answer about B's physical ability would change what counts as a satisfactory response, from compliance to a prediction of capacity. Whether that alteration matters depends on what the task requires the answer to do.
One qualification holds throughout: these are invented English reformulations for a later translation-review design, not verified bilingual pairs and not findings about any language.
29.6A synthetic paired example
The distinction just drawn becomes concrete in one stipulated case. An invented source question asks, "Is the report available?" The wording leaves the predicate's boundary open: "available" can mean that the report exists in the catalog, or that the intended reader can inspect it. Both readings are grammatical in English, both are plausible, and the context assigned to the example settles neither. The ambiguity is in the source question itself, not something introduced by later paraphrase.
Literal wording must here be kept apart from pragmatic purpose. An asker who wants a document located has one purpose; one who wants to know whether access is possible has another. The same sentence can serve either, so selecting a purpose is an interpretation rather than a discovery about the words. Where a term is locally contested in just this way, the analyst who reports one reading as the meaning is supplying an editor's interpretation, and should say so.
Two English reformulations make the alternatives explicit: "Does the catalog contain the report?" and "Can the intended reader inspect the report?" Each resolves the ambiguity stipulated in the source, and therefore each changes the required answer relative to a question that admits both. That consequence is narrow but real: a fluent rendering that resolves the predicate narrows the answer space even as it removes the ambiguity that made the question difficult to route.
Nothing follows yet about which expansion is preferable. Where later context does settle "available," the appropriate reformulation is the one matching that context; the unsettled case is worth keeping only because it exposes what a translator and reviewer must actually decide. The qualification is firm. These are stipulated English readings, offered as instruments for a future translation-review design. They are not a validated bilingual pair, and no completed multilingual study supports them.
The two expansions differ only in what they make explicit. "Does the catalog contain the report?" treats availability as a relation between the report and a collection; "Can the intended reader inspect the report?" treats it as a relation between the report and a person with an intended role. Neither expansion adds a fact about the report: the source question, as stipulated, supplies no catalog status and no inspection access. Each rendering selects one of the two allowed senses, and in doing so converts the original ambiguity into a question that admits a determinate answer once facts are supplied.
That is what makes the pair useful as a design rather than as a result. In a later translation review, the two English expansions can serve as probes: for a candidate translation of the source question, reviewers would ask which expansion the wording supports, whether it commits to one sense or preserves both, and which equivalence judgment it leaves open. The pair thereby specifies the comparison to be performed before any judgment is offered; it does not itself perform one. No target-language formulation has been supplied here.
The same limit governs any extension. If a future review were to add a target-language formulation, it would still have to establish, from that formulation and its context, whether the catalog sense or the access sense is supported; the English pair alone cannot settle this. Nor does the ambiguity of "available" show that any other language has the same two readings or lacks a distinction between them. The example fixes what a reviewer must look for, not what a reviewer will find.
That restraint is the point of keeping the pair within English. A demand for verified equivalence before the review begins would leave no concrete task to undertake; the expansions provide one without claiming validation that has not occurred.
29.7Reporting unresolved equivalence
The stipulated English exchange yields a determinate reading at one level. When A says "My hands are full" and then asks "Can you open the door?", the preceding utterance supports understanding the question as a request that B open the door. That is a local interpretation, supported by the surrounding sentence. It does not establish anything about how requests are formed across a language or culture, and the example does not license such a generalization.
The report must be narrower still for T01. Here the invented source asks "Is the report available?" while leaving open whether available means contained in the catalog or inspectable by the intended reader. The two English reformulations, "Does the catalog contain the report?" and "Can the intended reader inspect the report?", make that ambiguity explicit. Nothing in the stipulated context selects between them. Choosing one sense would narrow the case artificially: selecting catalogued existence over reader access would resolve that stipulated ambiguity by choice rather than by evidence.
So the report should state the supported interpretation and keep the unresolved judgment visible: the request reading is warranted for the hands-and-door exchange, while the sense of available in T01 remains open. The two expansions of T01 are design materials for a later translation-review exercise, not a validated pair of translations, and no multilingual result has been demonstrated. What remains unresolved is thus an equivalence judgment between two candidate construals of one invented source, not a factual question already settled. Paired review can examine such a split without pretending that the source itself fixes the answer.
A provisional translation need not resolve the disputed term before it can be used. In T01, the invented English source asks whether the report is available, and the context leaves open two stipulated readings: that the catalog contains the report, or that the intended reader can inspect it. A rendering that keeps the disputed word and adds a short explanation stating both readings stays readable and stays honest. The explanation does not select between the two English reformulations; it marks the choice as open, so that a later translation review can make it explicitly rather than silently.
What such a rendering supports is conditional reading. A reader who needs catalogued existence may reason from that sense, provided the gloss travels with the term and the selection is reported as a local narrowing. The same discipline applies to the word charge in the other invented example: the financial-fee sense and the accusation sense are both stipulated as allowed, and picking one narrows the ambiguity without discovering anything further about the source. The useful output is therefore not a verdict but a documented pair of live senses, together with the sense actually relied on.
The scope of this application is narrow. The T01 reformulations are English stipulations prepared for a future translation-review design, not verified bilingual pairs and not the outcome of completed translation research. Conditional on such a review being run, reviewers might recover the readings described here and test whether the explanatory note preserves them; until that exists, the provisional rendering licenses reading under a stated assumption rather than any claim of multilingual equivalence. Its value is that unresolved uncertainty constrains what may be asserted while leaving ordinary interpretation available.