Hall II · Psychological Phenomena, Tasks & Thought Experiments
Müller-Lyer illusion
Try it
Thread: Perception & awareness
Thread: Expectancy & belief (secondary)
Station · Try it
Both lines below are drawn by this page in your browser. One is fixed; you lengthen or shorten the other until the two shafts look the same to you, and then the page shows you what the true lengths were. There is no right answer to get and nothing to score. Whatever you set is held in this tab only and is never stored or sent anywhere.
The interactive station needs JavaScript, which is off — nothing is hidden from you because of that. The full explanation below covers everything the station demonstrates.
Figure 1 of 2
Drag the slider until the two shafts look the same length to you. Ignore the fins — they are not part of the line.
Fixed line
Your line
The experience
Two lines with little fins on their ends, drawn above by this page. Stretch one until it
looks the same length as the other, then let the page tell you what you actually did. Run
both trials — the second reverses the fins, and the pair is far more interesting than
either half.
The two shafts can be exactly the same length
and still refuse to look it. A shaft whose fins splay outward, so that the whole figure
is wider than the shaft itself, looks longer than an identical shaft whose fins fold back
over it. Franz Carl Müller-Lyer published the effect in 1889, observing that a line
terminating in an obtuse angle appears longer than one terminating in an acute angle,
that the size of the angle matters, and that curved fins produce it as well as straight
ones.Convergent / replicated Everything on this page follows from that one stubborn picture.
What is happening
Explaining it has proved much harder than
producing it. Müller-Lyer's own account had the fins flowing into the length judgement;
later candidates include misapplied size constancy, in which the figure that looks longer
is the one processed as further away, conflicting-cues and confusion accounts that
average over the whole figure or over the distance between fin tips, low-pass spatial
filtering that discards fine detail, and eye-movement accounts descending from Judd. A
century on there is no consensus winner, and it is increasingly suspected that the
illusion reflects several co-existing mechanisms, or several parallel routes to the same
appearance.Mixed / contested
That is a more interesting situation than a settled answer. The figure is trivial to
draw, the effect is trivial to reproduce, and the explanations keep multiplying because
each of them accounts for part of the evidence.
What this shows
The illusion does not dissolve when you know better.
The same person who has just watched the two shafts being measured, and agreed that they
are equal, still sees one as longer — and that stubbornness carried real theoretical
weight. Fodor made it a leading argument that vision is informationally encapsulated,
writing that how the world looks can be "peculiarly unaffected by how one knows it to
be", and Pylyshyn used the same figure to argue for the cognitive impenetrability of
visual processing.Supported, with limits
You can test that for yourself in a way the philosophers could not: the page will tell
you the true lengths, and you can then look again. Most people find the figure entirely
unrepentant.
What this does not show
A gap between your setting and the true length is not a reading of your visual system.
How big this illusion comes out
depends on the figure and the procedure as much as on the observer. Müller-Lyer himself
found apparent length changing with the angle of the fins, and the historical
cross-cultural record is largely a set of comparisons across different instruments: when
Amir and Firestone re-analysed Rivers's turn-of-the-century data with modern statistics,
only five of fifteen group comparisons reached significance, and most of those five
compared groups tested under two different versions of the protocol.Supported, with limits Our page uses
one fin angle, one fin length and one screen, and reports one session. It is a way of
meeting the phenomenon, not of measuring you.
Research and limitations
The most-cited thing about this figure is not the figure.
It became the emblem of a strong claim about
culture and perception. Segall, Campbell and Herskovits presented data from fifteen
societies showing substantial differences in susceptibility to geometric illusions, and
read them as evidence of different habits of perceptual inference tied to the visual
environment — the "carpentered world" idea. Henrich, Heine and Norenzayan then put those
data at the head of their argument that Western undergraduates are a strange sample to
generalise from, reporting that American undergraduates needed one segment to be about a
fifth longer than the other before calling them equal, that they differed significantly
from every other society studied, and that for the San foragers of the Kalahari it was
not an illusion at all.Mixed / contested
That reading is now under sustained attack.
Amir and Firestone argue the cultural account is almost certainly false, assembling
evidence that the illusion arises in birds, fish, reptiles, insects and other mammals,
that the statistics of natural scenes suffice to produce it, that it survives curved fins
and figures with no lines at all, that it appears in touch as well as vision, and that
nine children with dense congenital cataracts showed it on every stimulus variant within
hours of first seeing clearly. They also read the historical record as inconsistent:
Rivers found jungle-dwelling Urali and Sholaga more susceptible than the English, and
Segall and colleagues published their own section headed "Inconsistencies in our Own
Data", worried in print that interpreters may not have conveyed the instructions,
recorded an administrator's impulse to press respondents who gave the unexpected answer,
and excluded more than half of some samples.Mixed / contested Neither side is a footnote here, and
this entry does not pick a winner.
Research drawer — claims, sources and limits
Every highlighted claim above is recorded here with its evidence grade and sources. Grades are per claim; this site never computes an overall truth score.
↩ The two shafts can be exactly the same length and still refuse to look it. A shaft whose fins splay outward, so that the whole figure is wider than the shaft itself, looks longer than an identical shaft whose fins fold back over it. Franz Carl Müller-Lyer published the effect in 1889, observing that a line terminating in an obtuse angle appears longer than one terminating in an acute angle, that the size of the angle matters, and that curved fins produce it as well as straight ones.
Amir, D., & Firestone, C. (2025). Is visual perception WEIRD? The Müller-Lyer illusion and the Cultural Byproduct Hypothesis. Psychological Review (advance online publication). Preprint dated 24 January 2025, doi:10.31234/osf.io/y7mtf https://doi.org/10.31234/osf.io/y7mtf · accessed 2026-08-24 Support note: Section 2.1: "In 1889, the German sociologist Franz Carl Müller-Lyer noted a curious phenomenon: a line terminating in a simple obtuse angle appeared longer than one terminating in an acute angle"; "Müller-Lyer also noted that apparent line length was related to the angle size of the arrowheads, and found similar effects with line segments capped by curves rather than straight lines". The direction of the effect in geometric terms is fixed by the conflicting-cues description in section 2.2.1: the figure that is overall longer is the one whose central line comes to look longer.
Centre for the Study of Perceptual Experience, University of Glasgow. "Müller-Lyer Illusion." Illusions Index (illusionsindex.org). https://www.illusionsindex.org/i/mueller-lyer · accessed 2026-08-24 Support note: The Centre for the Study of Perceptual Experience entry states that the illusion presents two horizontal lines of identical length differing only in the direction of their terminating fins, one of which appears longer, and dates the first publication to Müller-Lyer 1889 in Archiv für Anatomie und Physiologie — an opened tertiary source used to corroborate the phenomenon description and the publication venue.
↩ Explaining it has proved much harder than producing it. Müller-Lyer's own account had the fins flowing into the length judgement; later candidates include misapplied size constancy, in which the figure that looks longer is the one processed as further away, conflicting-cues and confusion accounts that average over the whole figure or over the distance between fin tips, low-pass spatial filtering that discards fine detail, and eye-movement accounts descending from Judd. A century on there is no consensus winner, and it is increasingly suspected that the illusion reflects several co-existing mechanisms, or several parallel routes to the same appearance.
Amir, D., & Firestone, C. (2025). Is visual perception WEIRD? The Müller-Lyer illusion and the Cultural Byproduct Hypothesis. Psychological Review (advance online publication). Preprint dated 24 January 2025, doi:10.31234/osf.io/y7mtf https://doi.org/10.31234/osf.io/y7mtf · accessed 2026-08-24 Support note: Sections 2.1 and 2.2.1 enumerate the accounts named here: Müller-Lyer's "confluxion" theory, Gregory's size scaling theory, Day's conflicting-cues theory, the Sekuler and Erlebacher confusion theory keyed to intertip distance, Morgan et al. on misperceived endpoint locations, Ginsburg's low-pass spatial filtering, and efference theories tracing to Judd (1905). The section concludes that "such an account even eludes the field at large" and that "the Müller-Lyer illusion may actually reflect multiple co-existing mechanisms, or even multiple parallel paths to producing the same perceptual effects".
Centre for the Study of Perceptual Experience, University of Glasgow. "Müller-Lyer Illusion." Illusions Index (illusionsindex.org). https://www.illusionsindex.org/i/mueller-lyer · accessed 2026-08-24 Support note: Lists three competing explanations — misapplied size constancy scaling ("the outward-pointing version ... is processed as being further away"), conflicting cues, and confusion theory keyed to the distance between arrowhead tips — corroborating that the mechanism is presented as unsettled in a current teaching reference, and supplying the depth-scaling direction quoted in the claim.
↩ The illusion does not dissolve when you know better. The same person who has just watched the two shafts being measured, and agreed that they are equal, still sees one as longer — and that stubbornness carried real theoretical weight. Fodor made it a leading argument that vision is informationally encapsulated, writing that how the world looks can be "peculiarly unaffected by how one knows it to be", and Pylyshyn used the same figure to argue for the cognitive impenetrability of visual processing.
Amir, D., & Firestone, C. (2025). Is visual perception WEIRD? The Müller-Lyer illusion and the Cultural Byproduct Hypothesis. Psychological Review (advance online publication). Preprint dated 24 January 2025, doi:10.31234/osf.io/y7mtf https://doi.org/10.31234/osf.io/y7mtf · accessed 2026-08-24 Support note: Section 2.2.1 quotes Fodor (1983, p. 66) on "the widely noted persistence of many perceptual illusions ... even in defiance of the subject's explicit knowledge that the percept is illusory. The very same subject who can tell you that the Müller-Lyre [sic] arrows are identical in length, who indeed has seen them measured, still finds one looking longer than the other", and Fodor (1984, p. 34) on how "how the world looks can be peculiarly unaffected by how one knows it to be"; it then quotes Pylyshyn (2003) making the same move for cognitive impenetrability. Graded supported_with_limits because the source reports these as the premises of influential arguments and does not itself measure the persistence, and because the encapsulation conclusion drawn from it is philosophically contested.
↩ How big this illusion comes out depends on the figure and the procedure as much as on the observer. Müller-Lyer himself found apparent length changing with the angle of the fins, and the historical cross-cultural record is largely a set of comparisons across different instruments: when Amir and Firestone re-analysed Rivers's turn-of-the-century data with modern statistics, only five of fifteen group comparisons reached significance, and most of those five compared groups tested under two different versions of the protocol.
Amir, D., & Firestone, C. (2025). Is visual perception WEIRD? The Müller-Lyer illusion and the Cultural Byproduct Hypothesis. Psychological Review (advance online publication). Preprint dated 24 January 2025, doi:10.31234/osf.io/y7mtf https://doi.org/10.31234/osf.io/y7mtf · accessed 2026-08-24 Support note: Section 2.1 records that Müller-Lyer "noted that apparent line length was related to the angle size of the arrowheads". Section 5.1 reports the reanalysis of the Rivers data: a one-way ANOVA giving F(5, 212) = 7.15, p < 0.001, then "Of the fifteen possible comparisons, only five are significantly different, and of these, most involve comparisons between two different versions of the protocol". Graded supported_with_limits: this is one reanalysis of one historical dataset, presented in an argumentative review.
↩ It became the emblem of a strong claim about culture and perception. Segall, Campbell and Herskovits presented data from fifteen societies showing substantial differences in susceptibility to geometric illusions, and read them as evidence of different habits of perceptual inference tied to the visual environment — the "carpentered world" idea. Henrich, Heine and Norenzayan then put those data at the head of their argument that Western undergraduates are a strange sample to generalise from, reporting that American undergraduates needed one segment to be about a fifth longer than the other before calling them equal, that they differed significantly from every other society studied, and that for the San foragers of the Kalahari it was not an illusion at all.
Segall, M. H., Campbell, D. T., & Herskovits, M. J. (1963). Cultural differences in the perception of geometric illusions. Science, 139(3556), 769-771. https://doi.org/10.1126/science.139.3556.769 · accessed 2026-08-24 Support note: The full published summary reads: "Data from 15 societies are presented showing substantial intersocietal differences of two types in susceptibility to geometric optical illusions. The pattern of response differences suggests the existence of different habits of perceptual inference which relate to cultural and ecological factors in the visual environment." The claim's phrasing follows it closely and does not attribute to the 1963 paper the "carpentered world" label or per-society figures, which come from the 1966 book and from Henrich et al.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2-3), 61-83, with open peer commentary and authors' response. https://www2.psych.ubc.ca/~henrich/pdfs/WeirdPeople.pdf · accessed 2026-08-24 Support note: Section 3.1 (Visual perception) of the target article: "the undergraduates required that line 'a' be about a fifth longer than line 'b' before the two segments were perceived as equal. At the other end, the San foragers of the Kalahari were unaffected by the so-called illusion (it is not an illusion for them). While the San's PSE value cannot be distinguished from zero, the American undergraduates' PSE value is significantly different from all the other societies studied." The figure shows results from 16 societies (14 small-scale) from Segall et al. (1966); the claim cites the 1963 Science report's own count of 15 for the earlier publication.
↩ That reading is now under sustained attack. Amir and Firestone argue the cultural account is almost certainly false, assembling evidence that the illusion arises in birds, fish, reptiles, insects and other mammals, that the statistics of natural scenes suffice to produce it, that it survives curved fins and figures with no lines at all, that it appears in touch as well as vision, and that nine children with dense congenital cataracts showed it on every stimulus variant within hours of first seeing clearly. They also read the historical record as inconsistent: Rivers found jungle-dwelling Urali and Sholaga more susceptible than the English, and Segall and colleagues published their own section headed "Inconsistencies in our Own Data", worried in print that interpreters may not have conveyed the instructions, recorded an administrator's impulse to press respondents who gave the unexpected answer, and excluded more than half of some samples.
Amir, D., & Firestone, C. (2025). Is visual perception WEIRD? The Müller-Lyer illusion and the Cultural Byproduct Hypothesis. Psychological Review (advance online publication). Preprint dated 24 January 2025, doi:10.31234/osf.io/y7mtf https://doi.org/10.31234/osf.io/y7mtf · accessed 2026-08-24 Support note: Abstract lists the five lines of evidence in this order and concludes the Cultural Byproduct Hypothesis "is almost certainly false". Section 4.1 names "birds, fish, reptiles, insects, and other mammals". Section 4.5 describes Gandhi et al. (2015): nine congenitally blind children aged 8-16 with dense bilateral cataracts, tested hours after lens replacement, where "100% of the children, on 100% of the stimulus variants, showed evidence of the illusion". Section 5.1 reports the Rivers (1905) reversal — the Urali and Sholaga showing the greatest effect of all, "even considering the English". Section 5.2 names the "Inconsistencies in our Own Data" section of Segall et al. (1966) and quotes the interpreter concern (p. 178), the administrator's "impulse to correct the respondent" (p. 178), and exclusions "in some cases more than half".
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2-3), 61-83, with open peer commentary and authors' response. https://www2.psych.ubc.ca/~henrich/pdfs/WeirdPeople.pdf · accessed 2026-08-24 Support note: Included as the counterweight the claim is contesting, and to show the dispute is live rather than closed: the target article's own Response section rebuts a commentary (Khemlani et al.) that the illusion is "but one single phenomenon in visual perception, hardly representative of all visual perceptual processes" by pointing to parallel population-level variation in the Sander Parallelogram and two Horizontal-Vertical illusions — evidence that was already being argued over in 2010.