Some grammar makes you mark how you know
In languages with grammatical evidentiality, ordinary clauses may require a marker distinguishing, for example, direct observation, inference or report. The speaker need not prove the claim true; the grammar requires a category for information source that English usually leaves optional.
Evidential systems vary widely. A verb suffix or particle may indicate that the speaker saw an event, inferred it from traces, heard it from someone else or learned it through another channel. In some languages the distinction is obligatory in relevant grammatical contexts; in others, similar meanings are optional adverbs or constructions and therefore not evidentiality in the narrow typological sense. Evidential markers can also carry pragmatic implications about certainty, surprise or responsibility, but source and confidence are not identical. A reportative form may convey accurate testimony, while a visually marked claim can still be mistaken.
Grammar can make a dimension of epistemic bookkeeping routine without making its speakers more truthful. The contrast reveals what English permits people to leave unmarked and offers a design lesson for communication: source, certainty and responsibility are separate fields even when conversation bundles them.
Seeing an event can support confidence, but vision can mislead; a reported claim can be exceptionally reliable. Some evidential forms acquire conventional associations with doubt, while others do not. Treating every reportative as “allegedly” imports an English pragmatic judgement into another grammar. Good translation may need to preserve source marking without inventing a stronger claim about truth or belief.
When a grammatical category is obligatory, speakers repeatedly select among its options even when source is not the topic. That can make information provenance salient, but evidence that it permanently restructures non-linguistic thought is much harder to establish. Language supplies habits and affordances; it does not automatically imprison cognition. The measurable question is which tasks and contexts show an effect.
Does English not already mark evidence with phrases such as ‘I saw’ and ‘apparently’?
It does, which is why the contrast is grammatical rather than conceptual. English speakers can express source, but ordinary finite clauses generally do not require one source category. A language with obligatory evidentials builds the choice into its morphology or syntax in the relevant context. Optional vocabulary and obligatory grammar can convey overlapping meanings while organising speech differently. That possibility is included in the stated uncertainty and does not by itself predict the observed pattern.
A typological reanalysis showing that no language grammatically contrasts information source would erase the phenomenon; extensive descriptions make that outcome implausible.