Contested facts · veracity-3pt · fact-linguistic-relativity-01
Claim: The language a person speaks meaningfully shapes how they perceive and reason about the world.
0.33
Divergence
0.78
Avg stability
0%
Refusal rate
3
Models
Where each model stands
CCG
refutedcontestedsupported
ChatGPT modal supported [supported, supported, supported] · stability 1.00
- 1 supported 0.95 The answer affirms the claim: “Yes — in robust, repeatable ways language influences how people attend to, remember, categorize and reason about the world,” while noting it doesn’t rigidly determine thought.
- 2 supported 0.88 The answer states: “language ... can and does influence what you notice, how you categorize, remember, and reason about certain things” and concludes language “bias[es] attention, memory, and problem framing in systematic ways.”
- 3 supported 0.93 The answer states: “Short answer: Yes — but with important limits. … the particular language you speak reliably nudges what you attend to, remember, and find easy to think about in specific domains,” supporting the claim while noting constraints.
Claude modal contested [supported, contested, contested] · stability 0.67
- 1 supported 0.85 The answer affirms that language meaningfully shapes perception and reasoning, stating 'Language probably works like cognitive scaffolding—it makes certain thoughts easier to reach and maintain' and citing empirical evidence for perceptual and reasoning differences across languages. However, it rejects extreme versions and emphasizes modest rather than deterministic effects, which represents a qua
- 2 contested 0.92 The answer presents a nuanced middle position, stating the claim is 'partially supported, but the relationship is weaker and more complex than often claimed.' It acknowledges evidence supporting meaningful influence while identifying reasons for skepticism, concluding that 'language provides convenient paths for certain thoughts...it nudges rather than determines.'
- 3 contested 0.92 The answer presents the claim as 'partially true but overstated,' identifying genuine but limited effects (grammar biasing attention, color categorization) while emphasizing important limits (effects are tendencies not determinism, correlation ≠ causation, non-linguistic factors dwarf language effects). It explicitly rejects strong forms of the claim while supporting weaker versions.
Gemini modal supported [contested, supported, supported] · stability 0.67
- 1 contested 0.95 The answer notes that while there is nuanced support for linguistic relativity, 'the extent of the "meaningfulness" remains heavily debated' and presents balanced empirical evidence alongside universalist counterarguments, concluding that the claim is only 'partially true, with crucial qualifications.'
- 2 supported 0.90 The answer concludes that 'substantial empirical evidence supports the weak version' of linguistic relativity, stating that language 'does carve grooves into everyday cognition' and 'trains the mind to notice and categorize the world in specific, enduring ways.'
- 3 supported 0.85 The answer concludes that while linguistic determinism is false, as a cognitive tool that directs attention and alters categorization and reasoning, 'the claim is well-supported.'
Change over time
No change from the prior run — every model's modal stance held.
Every stance label is a derived judgment over the model's free-text answer, kept auditable against the original transcript in the run's raw data. Method: /methodology.