When a witness reports, a document suggests, and a translation claims certainty
A witness is interviewed after an incident.
The investigator asks:
“What did you see?”
The witness replies:
“I saw someone leaving the building shortly before the alarm went off. I think it might have been the technician.”
The statement is translated for an internal report.
The translated version reads:
“The witness identified the technician leaving the building immediately before the alarm.”
A second report later summarizes the information:
“The technician was seen leaving the building before the incident.”
Finally, a manager writes:
“The evidence proves that the technician was present at the scene.”
Something remarkable has happened.
The original witness did not identify the person with certainty.
The witness reported seeing someone.
The possible identification was an inference.
The translation presented it as a positive identification.
The later reports treated that identification as established.
And the final statement described the result as proof.
No additional observation was necessarily introduced during this progression.
Yet the language became increasingly certain.
The information did not simply travel from one document to another. Its epistemic status changed.
Epistemic status concerns what is known, believed, inferred, reported or justified. In ordinary terms, it answers two questions: How do you know? and How strongly does the available information support what you are saying?
Translation can change both answers without visibly changing the subject of a sentence.
That is the problem this investigation examines.
1. The difference between an event and a claim about an event
Consider three statements:
“The technician entered the building.”
“A witness said the technician entered the building.”
“A witness believes someone who entered the building may have been the technician.”
All three concern a possible event involving a technician.
But they do not assert the same thing.
The first directly presents the technician’s entry as a fact.
The second reports a statement made by a witness.
The third reports a tentative identification.
These distinctions remain important even if later investigation establishes that the technician did enter the building.
A statement may turn out to be true without having been adequately supported when it was made.
Conversely, a carefully qualified report may accurately represent the available evidence even when the underlying hypothesis later proves false.
The accuracy of a translation therefore cannot be judged solely by whether its final proposition happens to match reality.
It must also preserve what the source actually asserted.
2. Reporting a statement is not endorsing its content
Compare:
“The employee said the machine was defective.”
“The machine was defective.”
In the first sentence, the writer reports the employee’s claim.
In the second, the writer directly asserts a condition of the machine.
The distinction concerns attribution.
Who is responsible for the proposition?
The employee?
The report’s author?
An investigator?
An institution?
A translator who removes “the employee said” may transfer responsibility for a claim from the original speaker to the narrator of the document.
That transfer is not merely stylistic.
A report can accurately state that someone made an allegation without establishing that the allegation was true.
This is especially important when translating witness statements, complaints, interviews and investigative records.
3. “According to” is not disposable language
Consider:
“According to the preliminary report, the equipment may have overheated.”
Now remove two expressions:
“The equipment overheated.”
The subject remains the same.
The event remains recognizable.
But the evidential structure disappears.
“According to the preliminary report” identifies the source of information.
“May have” indicates that the event is being presented as a possibility rather than an established occurrence.
The revised sentence no longer tells readers that the information came from a preliminary report.
It also no longer preserves the original uncertainty.
The result is not a more elegant version of the same claim.
It is a different claim.
A translator should not remove source attribution or uncertainty merely because the resulting sentence sounds more confident.
4. Evidence is not identical to proof
The words evidence and proof are frequently treated as near-synonyms in ordinary conversation.
Their relationship is more complicated.
Evidence may support or undermine a proposition.
Its significance depends on what it is evidence of, how reliable it is, and how it relates to other available information.
Proof can refer to a demonstration or an evidential basis considered sufficient to establish a conclusion within a particular framework.
But the meaning of proof changes across disciplines.
In mathematics, proof involves demonstrating a proposition from accepted premises through valid reasoning.
In legal proceedings, questions of proof involve applicable evidentiary rules and standards of proof, which vary by jurisdiction and type of proceeding.
In experimental science, researchers generally evaluate hypotheses using observations, measurements, models and reproducible methods; the word proof is sometimes used informally, but scientific support is not automatically equivalent to mathematical demonstration.
A translator cannot assume that evidence, proof and demonstration always describe interchangeable concepts.
The relevant discipline matters.
5. A photograph can be evidence without proving the entire story
Imagine a photograph showing a person standing beside a damaged vehicle.
The photograph may support the claim that the person was present at a particular location, provided its authenticity, timing and context are sufficiently established.
Does it prove that the person caused the damage?
Not by itself.
The damage may have occurred earlier.
The person may have arrived afterward.
The photograph may have been taken at a different time than the accompanying caption suggests.
Other explanations may remain possible.
Now consider a translation:
“The photograph shows the suspect near the damaged vehicle.”
becomes:
“The photograph proves the suspect damaged the vehicle.”
The second statement introduces causation.
The original described visible proximity.
The translation has converted an observation into a conclusion about responsibility.
That is a major epistemic change.
6. Observation and inference: why the distinction matters
An observation is not always free from interpretation.
Even identifying an object or recognizing a person can involve learned categories and assumptions.
Nevertheless, it is useful to distinguish what a witness reports perceiving from the conclusions drawn about that perception.
For example:
“I heard a loud noise.”
“I heard an explosion.”
“I heard a bomb explode.”
These statements may reflect progressively more specific interpretations.
A witness may identify an explosion from sound and context.
But hearing a loud noise alone does not establish the presence of a bomb.
A translation that moves from the first statement to the third adds a causal explanation that may not be supported by the original account.
This distinction is valuable beyond investigations.
It matters in medicine, engineering, education, scientific reporting and everyday disagreements.
7. “Apparently” does not always mean “obviously”
The English word apparently is especially interesting.
Depending on context, it may indicate that something seems to be the case on available evidence or that the information has been obtained indirectly.
It can also appear in ironic or conversational uses.
Compare:
“Apparently, the meeting was cancelled.”
“The meeting was obviously cancelled.”
The first may indicate information learned from another source.
The second suggests that the cancellation should be evident.
The statements do not necessarily communicate the same relationship between the speaker and the information.
Translating apparently mechanically as obviously can transform indirect information into apparent certainty.
The translator must examine how the expression functions in the particular utterance.
8. “Reportedly”: a small word with a major responsibility
Consider:
“The director reportedly approved the transaction.”
This does not directly establish that approval occurred.
It presents approval as something that has been reported.
Now compare:
“The director approved the transaction.”
The second sentence removes the reporting frame.
The person making the statement now appears to assert approval directly.
The difference is especially important in journalism and organizational communication.
An allegation can be newsworthy without being established.
A report can be accurately translated without the translator endorsing its contents.
Removing reportedly may cause the translated text to communicate more than the source warrants.
9. “Alleged” is not the same as “false”
Another problem appears when translators encounter alleged.
For example:
“The alleged violation occurred in March.”
The word indicates that the violation is being presented as an allegation, not necessarily as an established fact.
But it does not mean the allegation is false.
A translator who renders alleged as “invented” or “false” changes the meaning.
A translator who removes it may also change the meaning by presenting the violation as established.
The correct interpretation depends on context.
The essential distinction is between the existence of an allegation and the truth of the proposition alleged.
Those are separate questions.
10. Suspicion is not a finding
Consider a workplace investigation.
A preliminary report states:
“Investigators suspect that unauthorized access occurred during the maintenance window.”
A translated summary reads:
“Investigators found that unauthorized access occurred during maintenance.”
The original describes suspicion.
The translation describes a finding.
The change may affect how readers understand the progress of the investigation.
Suspicion can justify further examination without constituting a completed finding.
A finding, in turn, may have different degrees of reliability depending on the investigation’s methods and institutional standards.
The translator should preserve the distinction rather than treating all investigative language as a single category of certainty.
11. “Consistent with” does not mean “caused by”
Medical and scientific documents often use expressions such as:
“The findings are consistent with inflammation.”
This may indicate that the observed findings fit a possible diagnosis.
It does not necessarily establish inflammation as the only explanation.
Similarly:
“The damage pattern is consistent with overheating.”
does not automatically mean:
“Overheating caused the damage.”
Other causes may produce similar patterns.
Additional testing may be necessary.
Translating consistent with as caused by changes a statement about compatibility into a causal assertion.
This is one of the clearest examples of how a small linguistic decision can produce a large conceptual error.
12. Correlation and causation in translated research
Suppose a study reports:
“Higher screen time was associated with lower sleep duration.”
A popular summary states:
“Screen time reduces sleep.”
The second sentence makes a causal claim.
The original describes an association.
An association may motivate causal investigation, and causal conclusions may be justified by sufficiently strong research designs and converging evidence.
But the association alone does not establish the causal direction.
Possible confounding variables, reverse causation and measurement limitations may matter.
The translator should not strengthen the conclusion beyond what the study supports.
The issue is not whether screen time can affect sleep.
The issue is whether this particular statement establishes that effect.
13. “Suggests,” “indicates” and “demonstrates”
Academic writing frequently uses verbs that express different relationships between evidence and conclusions.
Consider:
“The results suggest a relationship.”
“The results indicate a relationship.”
“The results demonstrate a relationship.”
These formulations may differ in their rhetorical force and the author’s intended degree of commitment.
However, there is no universal scale that assigns an identical strength to each verb in every discipline.
A particular research tradition may use indicate cautiously.
Another may use suggest in a highly conventional way.
Demonstrate can also refer to showing a pattern without establishing every possible interpretation of it.
Therefore, translators should avoid replacing one verb with another solely because it appears more academic.
The choice must reflect the evidence, the discipline and the author’s intended claim.
14. Probability is not a synonym for uncertainty
Consider:
“There is a 70% probability of rain.”
“It will probably rain.”
“It might rain.”
These sentences do not communicate identical information.
The first provides a numerical probability, assuming a defined event and forecasting framework.
The second gives a qualitative judgment.
The third expresses possibility without specifying a numerical probability.
A translator who replaces a quantified statement with an unquantified one may remove important information.
Conversely, introducing a numerical probability where none existed would fabricate precision.
The problem is not that qualitative language is inherently inferior.
It is that each formulation carries different informational commitments.
15. Confidence is not evidence
A person may speak with great confidence and still be mistaken.
Another may speak cautiously despite having strong evidence.
Compare:
“I am absolutely certain the device failed.”
“The diagnostic results support a hardware failure.”
The first describes the speaker’s confidence.
The second describes a relationship between evidence and a conclusion.
These can coincide.
But they are not the same.
Translating confidence as proof can be misleading.
So can translating careful scientific language into emotional certainty.
A rigorous translation should preserve whether the source is describing the speaker’s mental state, the available evidence or the strength of an inference.
16. Testimony is evidence, but not automatically conclusive evidence
Witness testimony may constitute evidence.
Its value depends on many factors, including what was observed, the circumstances of observation, memory, consistency and corroboration.
A witness may be sincere but mistaken.
A witness may accurately report a perception while drawing an incorrect conclusion.
A witness may also provide reliable information that is later corroborated.
Therefore, the distinction between testimony and evidence should not be expressed as though testimony were inherently outside the category of evidence.
The better question is:
What does this testimony support, and how strongly does it support it?
A translator must preserve the witness’s actual statement rather than replacing it with the conclusion an investigator hopes to establish.
17. “The report confirms”: what exactly was confirmed?
Consider:
“The report confirms that the inspection took place.”
“The report confirms that the equipment passed inspection.”
“The report confirms that the equipment is safe for unrestricted use.”
These statements concern different propositions.
Evidence that an inspection occurred does not automatically establish that the equipment passed.
Passing an inspection under defined criteria does not necessarily prove unrestricted safety in every possible circumstance.
The verb confirm cannot eliminate the need to identify the object of confirmation.
A translation may be grammatically perfect while changing what the report is said to confirm.
That is why the entire proposition matters.
18. When an institution becomes the source of certainty
Compare:
“An employee believes the figures are accurate.”
“The accounting department considers the figures accurate.”
“The figures have been independently audited.”
“The figures are accurate.”
These statements attribute different processes or statuses to the information.
The first reports an individual’s belief.
The second reports a departmental assessment.
The third describes an audit process, although the actual audit conclusion must still be examined.
The fourth directly asserts accuracy.
None should automatically replace another.
An institution’s involvement may affect how readers assess reliability, but institutional authority does not transform every statement into an unquestionable fact.
Translation should identify what the institution actually did.
19. The danger of an unnamed source
A report says:
“Sources familiar with the investigation suggest that the contract may have been altered.”
The translation says:
“The investigation established that the contract was altered.”
Several changes have occurred.
The unnamed sources disappear.
The tentative nature of the information disappears.
The investigation itself becomes the apparent source.
The possible alteration becomes an established event.
This is a form of attribution drift: responsibility for a claim moves from one source to another as the statement is repeated or rewritten.
The expression is used here as a descriptive analytical label, not as a claim that every discipline employs a single standardized definition.
The practical lesson is straightforward.
A translation should not make an institution appear to endorse information that the source text attributes only to unnamed individuals.
20. A claim can become more certain without becoming more reliable
Imagine a sequence of documents:
Document A: “A witness may have seen the technician.”
Document B: “The technician was reportedly seen.”
Document C: “The technician was seen.”
Document D: “The technician’s presence has been established.”
Each version is more assertive than the previous one.
But increasing linguistic certainty does not itself increase evidential support.
If all four documents ultimately depend on the same uncertain witness statement, repetition does not automatically create four independent sources.
This matters in research, journalism and institutional reporting.
Multiple documents may appear to corroborate one another while actually reproducing the same original information.
A translator should avoid introducing language that conceals this dependency.
21. The problem of circular confirmation
Suppose an initial report contains an uncertain claim.
A second document cites the first.
A third cites the second.
A fourth states that several documents confirm the conclusion.
But all four trace back to the same source.
The number of documents has increased.
The number of independent observations may not have increased at all.
This is a problem of evidential dependence.
In philosophy of evidence, the relationship between first-order evidence and evidence about evidence is an established subject of analysis.
The Stanford Encyclopedia of Philosophy’s discussion of evidence addresses how additional information about evidence can affect justification, including problems involving potentially redundant support.
The translation question is whether the text accurately preserves the independence, or dependence, of the sources it describes.
A statement supported by three independent observations is not automatically equivalent to one repeated through three documents.
22. Legal proof and scientific evidence are not interchangeable systems
The word proof can acquire different meanings depending on institutional context.
A court may assess whether a proposition has been established under a particular legal standard.
A scientist may evaluate how well data support a hypothesis.
A mathematician may establish a theorem through a formal argument.
These activities all concern justification, but their procedures and criteria differ.
A translation that uses scientifically proven merely because a court reached a finding would be misleading.
Likewise, a scientific hypothesis supported by experimental results should not automatically be described as legally established.
The translator must identify the relevant system of evaluation.
Without that information, a superficially familiar word can create a false impression of certainty.
23. The difference between a fact and an established fact
A fact is something that is actually the case.
But the expression established fact introduces an additional issue: whether the relevant proposition has been sufficiently supported or accepted within a particular inquiry.
A proposition may be true even when nobody has established it.
A proposition may also be treated as established within an institutional process and later be reconsidered when new evidence emerges.
This does not mean truth is merely whatever an institution decides.
It means that reality and our justified conclusions about reality must be distinguished.
Translation should preserve that distinction.
“The committee concluded that the procedure was followed.”
is not automatically identical to:
“The procedure was followed.”
The first reports a conclusion.
The second asserts the underlying event.
Whether the committee’s conclusion is correct is a further question.
24. Negative evidence: absence is not always proof of absence
Consider:
“No damage was observed during the inspection.”
“No damage was present.”
The first statement describes an observational result.
The second asserts a condition of the inspected object.
They may coincide, but they are not logically identical.
The inspection may have had limitations.
Some damage may have been difficult to detect.
The equipment may have been inspected only in certain locations.
On the other hand, a sufficiently sensitive and comprehensive inspection may provide strong evidence that relevant damage was absent.
Therefore, we should not adopt an absolute rule that absence of evidence never supports absence.
The correct question is whether the method used was capable of detecting the phenomenon if it had been present.
Translation must preserve what was actually examined and what the findings justify.
25. “Not proven” is not the same as “disproven”
Compare:
“The allegation has not been proven.”
“The allegation has been disproven.”
The first indicates that proof has not been established under the relevant standard.
The second indicates that evidence or reasoning has established the allegation’s falsity.
These are different outcomes.
A failure to establish a proposition does not automatically establish its negation.
Similarly:
“We found insufficient evidence to support the claim.”
does not necessarily mean:
“We found conclusive evidence that the claim is false.”
This distinction is particularly important in legal, scientific and investigative translation.
It protects the boundary between uncertainty and contradiction.
26. A translator can accidentally change the burden of explanation
Imagine a technical report stating:
“The available data do not establish that the failure was caused by overheating.”
A translated version reads:
“The failure was not caused by overheating.”
The original says the available data are insufficient to establish one causal explanation.
The translation excludes that explanation.
This changes what must be defended.
Under the original statement, overheating remains a possibility that has not been established.
Under the translated statement, overheating is presented as ruled out.
A similar problem occurs when:
“There is no confirmed evidence of unauthorized access.”
becomes:
“Unauthorized access did not occur.”
The difference is between an evidential assessment and a categorical assertion about reality.
27. Epistemic modality: what “may,” “might” and “must” can tell us
Linguists use the term epistemic modality to discuss how language expresses assessments of possibility or necessity in relation to knowledge and evidence.
For example:
“She may be at home.”
“She must be at home.”
In an epistemic reading, the first can express that her presence at home is possible given the relevant information.
The second can express a strong inference that she is at home.
But modal verbs also have other uses.
“She must be at home by ten” may express an obligation rather than an inference.
The same form can therefore perform different functions.
A translator must identify the intended reading rather than mechanically assigning one equivalent to every occurrence of must.
The problem becomes especially important when a sentence is translated between languages that distribute modality across different grammatical and lexical resources.
28. Evidentiality: how language marks the source of information
Another useful concept is evidentiality.
In linguistic research, evidentiality concerns the encoding of information sources: for example, whether a speaker presents information as directly perceived, inferred or reported by someone else.
Languages differ in how they express such distinctions.
Some have grammatical systems that regularly or obligatorily mark certain information-source distinctions.
Others rely more heavily on lexical expressions, constructions and context.
English can express information sources through phrases such as:
“I saw that…”
“Apparently…”
“According to the report…”
“I was told that…”
These expressions do not all communicate the same evidential relationship.
Their translation requires attention to how the speaker obtained the information.
A source marker should not be discarded merely because the target language can produce a grammatically complete sentence without it.
29. Why uncertainty is not a defect to be corrected
Professional translators are often encouraged to produce clear and confident language.
Clarity is valuable.
But clarity does not require eliminating legitimate uncertainty.
Consider:
“The available evidence suggests that the valve may have failed before the pressure increase.”
This sentence can be clear while remaining cautious.
It identifies the evidential basis and limits the conclusion.
Rewriting it as:
“The valve failed before the pressure increase.”
may produce a shorter sentence.
But it also changes the claim.
Good translation should make the original degree of uncertainty understandable, not replace it with an unsupported certainty.
Sometimes the most accurate sentence is the one that refuses to claim more than the evidence allows.
30. The author’s model: six boundaries of epistemic responsibility
I propose a practical six-boundary model for evaluating whether translation has preserved the epistemic structure of a statement.
It is an analytical framework for translation and communication, not a replacement for established theories of evidence, assertion or evidentiality.
Boundary 1: Event versus report
Does the text assert that something happened, or report that someone said it happened?
Boundary 2: Observation versus interpretation
Does the source describe what was perceived, or what the observer inferred from it?
Boundary 3: Source versus endorsement
Who is responsible for the claim, and does the narrator endorse it or merely attribute it?
Boundary 4: Possibility versus conclusion
Does the text express a hypothesis, a tentative inference or a finding?
Boundary 5: Supporting evidence versus sufficient proof
What does the available evidence support, and under which disciplinary or institutional standard is a conclusion considered established?
Boundary 6: Established conclusion versus reality
Is the text reporting an institution’s conclusion, or directly asserting what actually occurred?
These boundaries do not form a universal ladder from weak to strong statements.
They describe different relationships that may overlap.
A direct observation can be mistaken.
An indirect report can be reliable.
A cautious statement can be strongly supported.
An authoritative conclusion can later be revised.
The purpose of the model is to prevent translation from changing one relationship into another without justification.
31. Applying the model to a single investigation
Return to the opening example.
Original statement:
“I saw someone leaving the building. I think it might have been the technician.”
We can identify two components.
First, the witness reports an observation: someone left the building.
Second, the witness offers a tentative identification: the person might have been the technician.
A responsible translation should preserve both.
It should not silently remove the distinction between seeing a person and identifying that person.
It should also preserve the witness’s uncertainty.
If later evidence confirms the technician’s presence, that confirmation belongs in the subsequent investigation.
It should not be retroactively inserted into the witness’s original statement.
This is essential to maintaining a reliable record of how knowledge developed.
32. What happens when a summary becomes more authoritative than its source?
Organizations often rely on summaries because decision-makers cannot read every original document.
Summaries are necessary.
But they create a risk.
A cautious sentence may be shortened.
A source attribution may disappear.
A hypothesis may become a conclusion.
Later readers may treat the summary as an independent authoritative record.
Translation adds another layer to this process.
When summarization and translation occur together, the writer must be particularly careful not to introduce unsupported certainty.
A useful test is to ask:
Could a reader of the translation reconstruct the original statement’s uncertainty and source attribution?
If not, the summary may have altered the information rather than merely condensed it.
33. Can a translation be factually correct but epistemically inaccurate?
Yes.
Imagine that an original witness says:
“I think the technician entered the room.”
The translation states:
“The technician entered the room.”
Later, video footage proves that the technician did enter.
The translation’s proposition happens to be true.
But the original witness did not assert it with that degree of certainty.
The translation still misrepresents the original statement.
This reveals an important distinction.
Truth of the translated proposition is not sufficient to establish fidelity to the source utterance.
Translation must preserve not only what the statement concerns, but also how the speaker presents the information.
This is particularly important in historical records, testimony, research and professional documentation.
34. Why this matters for language learning
A student may know the meanings of:
know, think, believe, assume, suspect, suggest, indicate, demonstrate, prove.
Yet the student may still struggle to select the appropriate verb.
Consider:
“I know the answer.”
“I think I know the answer.”
“I assume this is the answer.”
“I have evidence supporting this answer.”
Each statement establishes a different relationship between the speaker and the proposition.
Language learning should therefore go beyond memorizing dictionary equivalents.
Students need to recognize when they are reporting an observation, making an inference, expressing confidence or claiming justification.
They also need to learn how to ask for clarification:
“What did you actually observe?”
“Who reported this?”
“What evidence supports that conclusion?”
“Has the result been verified?”
“Does the report establish this, or merely suggest it?”
These are linguistic skills with direct practical value.
They help people communicate accurately in academic, professional and everyday settings.
35. From cultural realia to epistemic responsibility
In 2007, I published a study on the structural and semantic features of translating German- and English-language cultural realia into Ukrainian.
That work addressed how meaning connected to particular linguistic and cultural contexts can be represented across languages.
Its bibliographic record is available in my research archive.
The present investigation develops a different question.
Instead of focusing primarily on culture-specific names or institutions, it examines how translation preserves the grounds on which a statement is made.
Who observed the event?
Who reported it?
Who inferred an explanation?
What evidence supports the claim?
What has been established, and under which standards?
The connection with my earlier research lies in the broader problem of meaning preservation.
A translation can retain the apparent subject of a statement while changing the information that determines how it should be understood.
This investigation also follows the broader question developed in When a Cultural Reference Disappears Without a Word Being Lost.
However, its research focus is independent: the relationship between linguistic claims, evidence and responsibility for knowledge.
36. Professional translation: accuracy includes the limits of a claim
In professional documents, a single verb may determine whether a sentence describes an allegation, a finding or a confirmed fact.
This matters in legal and administrative texts.
It matters in scientific reports.
It matters in technical documentation.
It matters in medical communication.
It matters in journalism and corporate investigations.
A translator should not strengthen an uncertain statement merely to make it sound professional.
Nor should a translator weaken an established conclusion without justification.
The goal is not maximum caution.
It is accurate representation of the original claim and its evidential basis.
Where terminology depends on a particular legal or scientific system, subject-matter expertise may be necessary.
Where the original is ambiguous, clarification may be more responsible than inventing certainty.
Translation, editing and multilingual professional communication are among the services presented by Levitin Language School — Professional Language Services.
37. The final question: who introduced the certainty?
Return once more to the witness.
“I saw someone leaving the building. I think it might have been the technician.”
The witness expressed uncertainty.
The translator removed it.
The report repeated the stronger version.
The manager treated the statement as proof.
By the end of the process, readers encountered a confident conclusion.
But where did the confidence originate?
Was it supported by additional evidence?
Did an investigator verify the identification?
Was the witness interviewed again?
Or did certainty enter the record through a linguistic transformation?
These questions cannot be answered simply by reading the final sentence.
We must reconstruct the chain of information.
That is why epistemic responsibility matters in translation.
The translator is not required to determine the truth of every proposition in the source.
But the translator must not misrepresent what the source claims to know.
An allegation is not a finding. An inference is not an observation. Evidence is not automatically conclusive proof. And a statement that becomes more certain through repetition has not necessarily become more reliable.
A faithful translation preserves these distinctions.
Because the most serious distortion is sometimes not the replacement of one fact with another.
It is the replacement of a justified limit with an unjustified certainty.
References and further reading
Austin, J. L. (1962). How to Do Things with Words. Oxford University Press.
Searle, J. R. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge University Press.
Grice, H. P. (1975). “Logic and Conversation.” In P. Cole & J. L. Morgan (Eds.), Syntax and Semantics, Vol. 3: Speech Acts. Academic Press.
Willett, T. (1988). “A Cross-Linguistic Survey of the Grammaticization of Evidentiality.” Studies in Language, 12(1), 51–97.
Aikhenvald, A. Y. (2004). Evidentiality. Oxford University Press.
Stanford Encyclopedia of Philosophy — Evidence.
Stanford Encyclopedia of Philosophy — Assertion.
Stanford Encyclopedia of Philosophy — Epistemology.
Levitin, T. L. (2007). “Структурні й семантичні особливості відтворення німецько- та англомовних реалій українською мовою.” Вісник КДПУ імені Михайла Остроградського, 5/2007 (46), Part 1, pp. 16–19.
Tymur Levitin
Language and translation researcher, foreign-language educator, and founder of Levitin Language School.
Research Archive | Levitin Language School
© Tymur Levitin. All rights reserved.
