Skip to content
Back to the question

Evidence record

Does the language you speak change what you see?

Every sentence on the exploration belongs to at least one claim below. Each says what kind of statement it is, what it rests on, what the evidence does not establish, and how far it was checked. Nothing here is a score, and none of it is a verdict on a belief or a theory.

Last checked 24 September 2026.

  • Source checked: 22
  • Partly checked: 6
  • A judgement, not a fact: 11

If something here is wrong, misquoted, out of date or unfair, write to support@unknownoceans.com and give the claim number, which does not change.

What the question takes for granted

An empirical question in psychology with a famous overstatement attached. The page separates the strong claim, that language determines thought, from a narrow one about speed at a colour boundary, and reports that the narrow one is itself disputed.

  • Languages divide colour differently

    Established

    Russian obliges a speaker to choose between two words across the range English calls blue. That is a fact about the languages, not about what their speakers see.

  • Seeing can be measured by how fast someone tells two colours apart

    Contested

    The 2007 study uses reaction time on a matching task as its measure. The 2020 replication argues that reaction times to blues of different lightness are uneven for reasons that have nothing to do with language, which is a dispute about the measure as much as the result.

How the question is labelled

Claim: Two words for blue

Kind
About a word. What a word in the original language means, or how it is translated.
Review
Source checked. The cited passage was read, and it says this.

Claim 1 · checked 24 September 2026 · AI-assisted audit

Claim: Language

Kind
Framing. Signposting. It asserts nothing a source could check.

The question

Claim: Russian has no single word for blue.

Kind
About a word. What a word in the original language means, or how it is translated.
Review
Source checked. The cited passage was read, and it says this.

Meaning no single basic term covering the whole range English calls blue; Russian has two, one for lighter and one for darker blues.

  • Supports

    Russian blues reveal effects of language on color discriminationIntroduction, first paragraph

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: There is no single generic word for ... in Russian that can be used to describe all of the colors in Fig. 1

Claim 3 · checked 24 September 2026 · AI-assisted audit

Claim: In a 2007 study, Russian speakers were faster to tell two blues apart when the two had different names, and lost the edge when their language was busy.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationabstract

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: blue stimuli that spanned the siniy/goluboy border ... Russian speakers were faster to discriminate two colors when they fell into different linguistic categories in Russian ... than when they were from the same linguistic category ... Moreover, this category advantage was eliminated by a verbal, but not a spatial, dual task.

  • Cuts against

    Russian blues reveal the limits of language influencing colour discriminationabstract

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: Contrary to previous research, our results show that in Russian, stimulus discrimination at the ... blue BCC boundary is not reflected in processing speed.

What this does not establish

One study of twenty-one Russian and twenty-one English speakers in the analysis. The advantage is a comparison within the Russian group; overall, the Russian speakers were slower than the English speakers.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “Russian speakers are measurably faster at telling two blues apart, and lose the advantage when their language is occupied.” The old wording read as Russian speakers being faster than English speakers, and as a settled general fact. The study found a within-group advantage for pairs across the Russian boundary, and Russian speakers were slower overall (1,085 against 938 msec).

Claim 4 · checked 24 September 2026 · AI-assisted audit

Claim: A 2020 attempt to repeat it found no edge.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Not an exact replication: the authors could reproduce fifteen of the original twenty colours, and their first experiment repeated only the no-interference condition.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “” A published replication with the same tasks did not find the effect the subtitle describes; the page did not mention it.

Claim 5 · checked 24 September 2026 · AI-assisted audit

Claim: The strong claim that language determines thought is not supported.

Kind
What the field holds. The settled view of a field, as a review, a consensus statement or a reference work gives it. Not one study.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Linguistic relativityabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Review of the field

    The source reads: While we do not find support for the idea that language determines the basic categories of thought or that it overwrites preexisting conceptual distinctions

What this does not establish

Rests on the abstract of one review; the body of the review was not read. The same abstract reports growing support for weaker effects of language on thought.

Later work looked for

Europe PMC, 2011 on, "linguistic relativity" reviews: nothing found that revives the determinist version.

Claim 6 · checked 24 September 2026 · AI-assisted audit

Claim: A narrower one has been reported: the categories a language marks speed up the discriminations that cross them, and the effect shrinks or disappears when the language system is loaded during the task.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationDiscussion

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: In all four studies (the present work and refs. 21–23), a category advantage was observed in simple perceptual tasks and the category advantage was selectively eliminated or reduced by verbal, but not spatial, interference.

What this does not establish

The four studies named are the 2007 paper and three related ones it cites, two of them about the English blue and green boundary. The 2007 paper read here is the only one of the four this audit read.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “A narrower claim is: the categories a language marks speed up the discriminations that cross them, and the effect can be switched off by loading the language system during the task.” The old wording set the narrow claim beside the unsupported strong one as supported. It has been reported, and its best-known case has since failed a replication (claim 8).

24 September 2026. Previously: “A narrower one has been reported: the categories a language marks speed up the discriminations that cross them, and the effect switches off when the language system is loaded during the task.” The four studies the 2007 paper summarises found the advantage eliminated or reduced by verbal interference, not always switched off.

Claim 7 · checked 24 September 2026 · AI-assisted audit

Claim: The Russian blues demonstration did not reappear when another group repeated it.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

One replication, by one group, with differences of stimulus and sample that its authors list. It does not show the original result was wrong, only that it did not appear again.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “” The summary did not mention the 2020 replication.

Claim 8 · checked 24 September 2026 · AI-assisted audit

Claim: The idea that speakers of different languages inhabit different worlds has been embraced, dismissed and embraced again.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Language, thought, and color: Whorf was half rightabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Review of the field

    The source reads: Over the years, consensus has oscillated between embrace and dismissal of this hypothesis.

What this does not establish

Rests on the review’s abstract; its history of the debate was not read.

Corrected

24 September 2026. Previously: “The idea that speakers of different languages inhabit different worlds has been asserted, refuted, and asserted again for a century, usually with anecdotes about snow.” No cited source dates the debate to a century or says it was usually argued from snow. The review cited says consensus has oscillated between embrace and dismissal, and the sentence now says that.

Claim 9 · checked 24 September 2026 · AI-assisted audit

Claim: The version that can be tested is smaller, and was tested with a colour chip and a stopwatch.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The chip and the stopwatch stand for on-screen colour squares and reaction times, which is what the 2007 study used.

Corrected

24 September 2026. Previously: “The version that survived is smaller, testable, and was tested with a colour chip and a stopwatch.” "Survived" overstates it: the narrow version has itself failed a replication (claim 34).

Claim 10 · checked 24 September 2026 · AI-assisted audit

What was measured

Claim: What was measured

Kind
Framing. Signposting. It asserts nothing a source could check.

The experiment

Claim: The experiment

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: Russian divides the blue range into two basic colour terms, lighter and darker, with no word covering both.

Kind
About a word. What a word in the original language means, or how it is translated.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

The 2020 replication found the lighter and darker boundary less stable than a hue boundary: it shifted with how often each shade was shown.

Claim 13 · checked 24 September 2026 · AI-assisted audit

Claim: English has one.

Kind
About a word. What a word in the original language means, or how it is translated.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationIntroduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: For English speakers, all of the colors in all trials fell into the same basic linguistic category, namely, blue.

Claim 14 · checked 24 September 2026 · AI-assisted audit

Claim: In a 2007 study, speakers of each language saw three blue squares and indicated which of the lower two matched the one above.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationIntroduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: Subjects were simultaneously shown three color squares arranged in a triad (see Fig. 1) and were asked to say which of the bottom two color squares was perceptually identical to the square on top.

Later work looked for

How the task works; not a finding later work revises.

Claim 15 · checked 24 September 2026 · AI-assisted audit

Claim: Russian speakers were faster when the two candidates fell on opposite sides of their category boundary than when both fell inside one category.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Only on the harder comparisons, where the two blues were two steps apart; the page says so in claim 31. The 2020 replication did not find it.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Claim 16 · checked 24 September 2026 · AI-assisted audit

Claim: English speakers showed no such effect.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Claim 17 · checked 24 September 2026 · AI-assisted audit

Claim: The advantage disappeared when participants performed a verbal task at the same time, and survived when the interfering task was spatial.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

The verbal task was silently rehearsing an eight-digit number; the spatial task was holding a grid pattern in memory.

  • Supports

    Russian blues reveal effects of language on color discriminationResults, Detailed Analyses

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: there was a category advantage under both the ... interference and the spatial-interference conditions, but not under the verbal interference condition

What this does not establish

The 2020 replication did not test the interference conditions: it found no advantage to remove.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Claim 18 · checked 24 September 2026 · AI-assisted audit

Claim: The spatial task is the control: it shows the loss came from occupying language, not from doing two things at once.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationIntroduction

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: The spatial-interference control condition was used to examine whether any differences between the baseline condition and verbal-interference condition were specific to language, or whether they were due to nonspecific effects of any dual task.

  • Supports

    Russian blues reveal effects of language on color discriminationMethods, The Interference Conditions

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: The spatial- and verbal-interference tasks were pretested for difficulty in the absence of a primary task and found to result in equal accuracy

Later work looked for

How the design works; not a finding later work revises.

Corrected

24 September 2026. Previously: “That is the control that makes this about language rather than about culture or training.” The spatial task controls for the general cost of a second task, which is what the authors designed it for. It does nothing to separate language from culture or training.

Claim 19 · checked 24 September 2026 · AI-assisted audit

Claim: What the stopwatch showed

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: Russian speakers discriminated across their category boundary faster than within a category.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

The 2007 sample, on near-colour comparisons. Not found in 2020.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “Russian speakers discriminate across their category boundary faster than within a category.” Present tense stated one study’s result as a general fact about Russian speakers, which the 2020 replication does not bear out.

Claim 21 · checked 24 September 2026 · AI-assisted audit

Claim: A distinction the language marks is available to perception under time pressure.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The authors’ reading of claim 21, which the page adopts.

Other readings

The 2020 replication reads the same reaction times as a lightness effect: darker blues take longer to tell apart, whatever the language.

Claim 22 · checked 24 September 2026 · AI-assisted audit

Claim: English speakers showed no boundary effect on the same squares.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationabstract; Methods

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: English speakers tested on the identical stimuli did not show a category advantage in any of the conditions. ... Subjects were simultaneously shown three color squares arranged in a triad

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “English speakers show no boundary effect on the same squares.” Tense, to match the other rows as one study’s result.

Claim 23 · checked 24 September 2026 · AI-assisted audit

Claim: The effect follows the vocabulary rather than the eyes.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

Follows from claims 21 and 23: the same squares, a different result in each language group.

  • Context

    Russian blues reveal effects of language on color discriminationDiscussion

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: color discrimination performance differs across language groups as a function of what perceptual distinctions are habitually made in a particular language

Other readings

Martinovic, Paramei and MacInnes argue the difference may follow the darker blues and the slower responders instead.

Corrected

24 September 2026. Previously: “The effect follows the vocabulary rather than the eyes. Colour vision is the same.” The 2007 study did not test its participants’ colour vision, so "Colour vision is the same" had no source.

Claim 24 · checked 24 September 2026 · AI-assisted audit

Claim: The advantage vanished under a verbal load, not under a spatial one.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationResults, Summary of Results

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: The category advantage found for Russian speakers was disrupted by verbal, but not spatial, interference.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “The advantage vanishes under a verbal load, not under a spatial one.” Tense, to report one study’s result rather than a general fact.

Claim 25 · checked 24 September 2026 · AI-assisted audit

Claim: The language system is participating in the task, live, rather than having shaped perception permanently.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The authors’ own inference from claim 25, which the page adopts.

  • Context

    Russian blues reveal effects of language on color discriminationDiscussion

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: If the language-specific distortions in perceptual discrimination had been a product of a permanent change in perceptual processors, temporarily disabling access to linguistic representations with verbal interference should not have changed the pattern in perceptual performance.

What this does not establish

The authors say the exact nature of the interaction cannot be determined from their data. It stands or falls with the result it interprets.

Claim 26 · checked 24 September 2026 · AI-assisted audit

What this does not support

Claim: What this does not support

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: It is not evidence that a speaker cannot perceive a distinction their language lacks.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The authors draw the same limit; claim 29 gives the data.

  • Context

    Russian blues reveal effects of language on color discriminationDiscussion

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: The critical difference in this case is not that English speakers cannot distinguish between light and dark blues, but rather that Russian speakers cannot avoid distinguishing them

Claim 28 · checked 24 September 2026 · AI-assisted audit

Claim: English speakers matched the blues as accurately as Russian speakers did, and put a light and dark boundary in almost the same place.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationResults, Accuracy

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: Because the stimuli were present on the screen until subjects responded, accuracy was high (96.5 ... and 95.7 ... for English and Russian speakers, respectively)

    96.5 per cent for English speakers and 95.7 per cent for Russian speakers.

  • Supports

    Russian blues reveal effects of language on color discriminationResults, Boundaries

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: dark blue boundary (English speakers) were nearly identical

    Stimulus 8.7 for the Russian speakers and 8.6 for the English speakers, on a scale of twenty.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “English speakers see the difference between the two blues perfectly well and take slightly longer to act on it.” English speakers were not slower: they were faster overall than the Russian speakers (938 against 1,085 msec). What they lacked was a within-group advantage at the boundary. "Perfectly well" is replaced by what the study measured, accuracy and boundary placement.

Claim 29 · checked 24 September 2026 · AI-assisted audit

Claim: What they lacked was the speed-up for crossing it.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “” Replaces "take slightly longer to act on it", which misstated the result (claim 29).

Claim 30 · checked 24 September 2026 · AI-assisted audit

Claim: The difference was a little over a tenth of a second, and only when the two blues were close; on easier pairs there was none.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal effects of language on color discriminationResults, Detailed Analyses

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: This effect was completely due to the near-color condition (Fig. 3)

  • Supports

    Russian blues reveal effects of language on color discriminationResults, Detailed Analyses

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: In near-color trials, Russian speakers showed a category advantage without interference [1,164 vs. 1,288 msec ... the category advantage was significantly larger in no interference blocks than in verbal interference blocks [124

    A category advantage of 124 msec without interference and 109 msec under spatial interference, on near-colour pairs: 0.11 to 0.12 of a second, a little over a tenth.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “The difference is in milliseconds, in a task designed to be hard.” The task mixed easy and hard comparisons, and the effect appeared only on the hard ones, at about 124 msec. The new wording gives the size and the condition.

24 September 2026. Previously: “The difference was about a tenth of a second, and only when the two blues were close; on easier pairs there was none.” The paper's category advantage on near-colour pairs is 124 msec without interference and 109 under spatial interference; 124 is a quarter more than a tenth of a second, outside what "about" allows.

Claim 31 · checked 24 September 2026 · AI-assisted audit

Claim: That is the honest size of the effect.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page’s judgement on claim 31.

Corrected

24 September 2026. Previously: “That is the honest size of the effect, and it is a real one.” "It is a real one" is not settled: the 2020 replication did not find it (claim 34).

Claim 32 · checked 24 September 2026 · AI-assisted audit

Claim: Whether it holds is now argued.

Kind
Disputed. A contested point, where the page reports a side or the disagreement.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

Introduces claims 34 and 35.

Other readings

Winawer and colleagues tested whether overall speed explained their result and found it did not (claim 35). The replication’s authors argue it does.

Corrected

24 September 2026. Previously: “” Added with the report of the 2020 replication.

Claim 33 · checked 24 September 2026 · AI-assisted audit

Claim: In 2020 Martinovic, Paramei and MacInnes repeated the task in two experiments, with seventeen and then twenty Russian speakers, and found no speed advantage at the boundary.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Source checked. The cited passage was read, and it says this.

What this does not establish

Not an exact copy of the 2007 study: fifteen of the twenty colours, no interference conditions in the first experiment, no English group in the second, and Russian speakers who were early bilinguals, mostly from the Baltic states.

Later work looked for

Europe PMC and Crossref, 2020 on, "Russian blues" and "goluboy siniy": Martinovic, Paramei and MacInnes (2020) found no speed advantage at the boundary; the page now reports it. A 2026 corrigendum to that paper exists and could not be read here.

Corrected

24 September 2026. Previously: “” A published replication with the same tasks was missing from the page.

Claim 34 · checked 24 September 2026 · AI-assisted audit

Claim: They argue the original result may be an artefact: darker blues are slower to tell apart, and the Russian group was slower overall.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

  • Supports

    Russian blues reveal the limits of language influencing colour discriminationDiscussion

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: the emergence of an observable RT advantage for cross-category colours in Winawer et al.’s study might have been an artifact of the non-uniform RT distributions driven by discriminations of darker vs. lighter colour pairs, further impacted by considerably longer overall RTs in the Russian group

  • Qualifies

    Russian blues reveal effects of language on color discriminationResults, Detailed Analyses

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Experiment

    The source reads: This analysis confirms that differences in overall speed between the two language groups were not responsible for the ... linguistic differences of interest

    The 2007 authors tested overall speed as a covariate. They did not test the lightness of the individual pairs, which is the replication’s main argument.

What this does not establish

An argument from the replication’s data, not a demonstration that the 2007 result was an artefact.

Corrected

24 September 2026. Previously: “” Reports the replication’s explanation, attributed to its authors.

Claim 35 · checked 24 September 2026 · AI-assisted audit

How far it generalises

Claim: How far it generalises

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: The effect lives in the moment of the task, not in the eye.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

A restatement of claim 26, the 2007 authors’ reading of their interference result.

What this does not establish

Only as secure as the effect itself, which the 2020 replication did not find.

Claim 37 · checked 24 September 2026 · AI-assisted audit

How far beyond colour it goes

Claim: How far beyond colour it goes

Kind
Framing. Signposting. It asserts nothing a source could check.

Claim: A review written two years later put the position as Whorf being half right, in two ways: language influences colour perception “primarily in half the visual field”, and colour naming across languages is shaped by both universal and language-specific forces.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Language, thought, and color: Whorf was half rightabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Review of the field

    The source reads: the Whorf hypothesis is half right, in two different ways: (1) language influences color perception primarily in half the visual field, and (2) color naming across languages is shaped by both universal and language-specific forces

What this does not establish

Rests on the review’s abstract; the review itself was not read.

Claim 39 · checked 24 September 2026 · AI-assisted audit

Claim: That is a specific, checkable version of a claim that spent decades being neither. It also limits it. A perceptual difference that appears in one half of the visual field and not the other is not a different world.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page’s reading of claim 39.

Claim 40 · checked 24 September 2026 · AI-assisted audit

Claim: And the half-field result has not held up in later tests: Witzel and Gegenfurtner ran ten versions of the original experiments with 230 observers and found the category effect equally strong in both halves.

Kind
Experimental result. What an experiment or trial found, under its own conditions. A result in animals, cells or a simulation is not a result in people.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Is there a lateralized category effect for color?abstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Experiment

    The source reads: We ran ten different versions of the two original experiments with overall 230 naive observers. ... However, these effects were not lateralized: They appeared to the same extent in both visual fields.

  • Supports

    Russian blues reveal the limits of language influencing colour discriminationIntroduction, list of findings

    Cited · Link answered, 19 October 2026 · Read in full, 24 September 2026 · Research in people · Review of the field

    The source reads: lateralisation of the CP effect to the right visual field, which was taken to be the landmark evidence for the influence of language lateralised in the left hemisphere

    The same passage lists four studies that could not replicate it; only Witzel and Gegenfurtner’s abstract was read.

What this does not establish

Witzel and Gegenfurtner was read as an abstract. The other failed replications are known here only from the 2020 paper’s list.

Later work looked for

Europe PMC, 2011 on, "lateralized category effect": a 2017 study reports a late right-lateralised effect alongside an early left one; nothing restores the half-field account as stated in 2009.

Corrected

24 September 2026. Previously: “” The page presented the half-field result as the checkable core of the question without saying that large replications did not find it.

Claim 41 · checked 24 September 2026 · AI-assisted audit

Claim: Comparable effects have been reported for spatial description, for how much blame people assign when an accident is described with or without an agent, and for how time is laid out.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

  • Supports

    Linguistic relativityabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Review of the field

    The source reads: across a wide range of domains, including motion, color, spatial relations, number, and false belief understanding

  • Supports

    Subtle linguistic cues influence perceived blame and financial liabilityabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Experiment

    The source reads: Agentive descriptions led participants to attribute more blame and request higher financial penalties than did nonagentive descriptions.

  • Supports

    Does language shape thought? Mandarin and English speakers’ conceptions of timeabstract

    Cited · Link answered, 19 October 2026 · Abstract only, 24 September 2026 · Research in people · Experiment

    The source reads: Mandarin speakers tended to think about time vertically even when they were thinking for English

What this does not establish

All three read as abstracts. The blame study compares two ways of describing an event in English, not two languages.

Later work looked for

Europe PMC, 2001 on, Mandarin time and vertical: the time result is disputed (claim 43); nothing found that revises the other two reports.

Corrected

24 September 2026. Previously: “Comparable effects have been reported for spatial description, for how events are attributed to people or to circumstances, and for how time is laid out, some of them strong and some contested.” The cited study measured blame and financial penalties under agentive and non-agentive descriptions, and the page now says so. "Some contested" was replaced by naming the contested result and who holds each side (claim 43).

Claim 42 · checked 24 September 2026 · AI-assisted audit

Claim: The time result is contested: January and Kako reported six failed attempts to repeat it, and Boroditsky and colleagues reported new evidence for it in 2011.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Partly checked. Only part of the source could be read. What was not read is listed.

What this does not establish

Both read as abstracts. The 2011 paper uses a different task from the 2001 one, so it supports the claim that the two groups think about time differently rather than repeating the original experiment.

Corrected

24 September 2026. Previously: “” The page cited the 2001 time study without its failed replications or the later defence; both sides are now named and cited.

Claim 43 · checked 24 September 2026 · AI-assisted audit

Claim: Colour is the cleanest case because the stimulus can be specified physically,

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The 2007 study gives its colours as physical coordinates, which a study of blame or time cannot do for its stimuli.

Claim 44 · checked 24 September 2026 · AI-assisted audit

Claim: and even there the replication turned on whether two pairs of blues were really equally hard to tell apart.

Kind
What scholars argue. A report of what specialists argue, or where they disagree.
Review
Source checked. The cited passage was read, and it says this.

Corrected

24 September 2026. Previously: “” Added so the passage does not present colour as free of the dispute it reports above.

Claim 45 · checked 24 September 2026 · AI-assisted audit

Claim: Most of the interesting claims are about domains where nothing can be specified that precisely, which is why they stay argued.

Kind
This page’s reading. A judgement the page makes from the evidence. Others read it differently.
Review
A judgement, not a fact. Reviewed for whether it is fairly put and follows from the checked claims. No source can confirm it.

The page’s judgement, following claims 42 to 45.

Corrected

24 September 2026. Previously: “Most of the interesting claims are about domains where it cannot, which is why they stay argued.” Reworded to follow the new clause before it.

Claim 46 · checked 24 September 2026 · AI-assisted audit

What was read