Cited by 500Social Robot Interaction and HRIVirtual Reality Applications and ImpactsFace Recognition and Perception

The DOI matches its Crossref record; 73 of the 73 references it cites were found in a registry or an index; none of the 73 sources its claims cite is checked for support yet.

A review of empirical evidence on different uncanny valley hypotheses: support for perceptual mismatch as one road to the valley of eeriness

Open access
Kätsyri, Jari, Förger, Klaus, Mäkäräinen, Meeri, Takala, Tapio·2015·Frontiers in Psychology

Abstract

The uncanny valley hypothesis, proposed already in the 1970s, suggests that almost but not fully humanlike artificial characters will trigger a profound sense of unease. This hypothesis has become widely acknowledged both in the popular media and scientific research. Surprisingly, empirical evidence for the hypothesis has remained inconsistent. In the present article, we reinterpret the original uncanny valley hypothesis and review empirical evidence for different theoretically motivated uncanny valley hypotheses. The uncanny valley could be understood as the naïve claim that any kind of human-likeness manipulation will lead to experienced negative affinity at close-to-realistic levels. More recent hypotheses have suggested that the uncanny valley would be caused by artificial-human categorization difficulty or by a perceptual mismatch between artificial and human features. Original formulation also suggested that movement would modulate the uncanny valley. The reviewed empirical literature failed to provide consistent support for the naïve uncanny valley hypothesis or the modulatory effects of movement. Results on the categorization difficulty hypothesis were still too scarce to allow drawing firm conclusions. In contrast, good support was found for the perceptual mismatch hypothesis. Taken together, the present review findings suggest that the uncanny valley exists only under specific conditions. More research is still needed to pinpoint the exact conditions under which the uncanny valley phenomenon manifests itself.

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This systematic review reinterprets Mori’s uncanny valley hypothesis as several testable hypotheses and reviews empirical evidence for each. It finds little support for the naïve claim that any human-likeness manipulation produces the uncanny valley, but good support for perceptual mismatch explanations, especially inconsistent realism and atypical features on humanlike characters.

Key findings

  • The naïve uncanny valley proper (H1a) received almost no support: 1 of 8 studies supported it, while the general linear human-likeness prediction (H1b) was supported in 7 of 9 studies.
  • Movement hypotheses were not supported: H2a had 0 of 2 studies in favor, H2b had 2 of 2, and no studies addressing movement modulation (H2c) survived selection.
  • Categorization ambiguity evidence was mixed: categorical perception of morphed human-likeness continua was supported (H3a 4/4; H3b 3/3), but negative affinity at the category boundary was not (H3c 0/2; H3d 0/3; H3e 0/1).
  • Perceptual mismatch hypotheses received good support: inconsistent human-likeness (H4a) was supported in 4 of 4 studies, and atypicality (H4b) in 3 of 4 studies.

Methods

  • Systematic review of empirical uncanny valley studies. Searches in Scopus (N = 273), PubMed (N = 23), Science Direct (N = 134), and Web of Science (N = 114), augmented by 6 other articles, yielded an initial list of N = 550; after screening N = 125, the final selected list was N = 17 articles.
  • Studies were evaluated against inclusion criteria and validity threats adapted from Shadish et al. (2002), covering statistical conclusion, internal, construct, and external validity.
  • Because studies were heterogeneous and most did not report effect sizes, the authors used a “box score” count of significant and non-significant findings rather than quantitative meta-analysis, with a 30% threshold for considering positive findings supportive.

Limitations

  • The authors state that their conceptual analysis relied on the English translation and secondary sources rather than Mori’s original Japanese article.
  • The review included only studies using self-report affinity measures, omitting physiological and behavioral studies such as gaze tracking and haemodynamic response measurements; the authors also note that heterogeneous self-report items reduced comparability across studies.
  • The authors state that their evaluation criteria are open to subjective interpretation, do not cover all aspects of experimental validity, and should not be taken as evidence for the validity of the evaluated studies themselves; they also note that only categorization ambiguity and perceptual mismatch explanations were reviewed, leaving other explanations aside.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3389/fpsyg.2015.00390

Cites

73 of 73 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 of 73 cited sources checked · 0 supported · 0 flagged

Show all 81 claims

  • Probably not a substantial claim. 45
  • It may state two separate claims as one. 17
55 of 55
  • The authors state this as an assumption.

    The uncanny valley proper coincides with the category boundary between 'artificial' and 'human' categories.

    “assuming that the uncanny valley proper is thought of as coinciding with the category boundary”
  • The authors state this as an assumption.

    Identification uncertainty at the category boundary leads to negative affinity.

    “The most straightforward assumption would be that identification uncertainty at the category boundary (“categorization ambiguity”) leads to negative affinity.”
  • The authors state this as an assumption.

    Perceptual discrimination is more difficult for characters in the uncanny valley.

    “Assuming that perceptual discrimination would be more difficult for characters in the uncanny valley, this difficulty should also be associated with negative affinity.”
  • The authors state this as an assumption.

    Morphing artifacts are a realistic concern.

    “Assuming that morphing artifacts were a realistic concern”
  • The paper relies on this without support.

    The experimental stimuli clearly represented phenomena that would be likely to exist also in the real world.

    “given that the experimental stimuli clearly represented phenomena that would be likely to exist also in the real world”
  • The paper relies on this without support.

    This article is the first systematic review of the empirical evidence for the uncanny valley.

    “to our knowledge, this article is the first systematic review of the empirical evidence for the uncanny valley.”
  • The paper relies on this without support.

    No prior distinction has been made between different perceptual mismatch hypotheses.

    “we are not aware that a further distinction would have been made between different perceptual mismatch hypotheses.”
  • The paper relies on this without support.

    The original uncanny valley formulation proposed that movement would amplify the characteristic uncanny curve.

    “The original uncanny valley formulation proposed that movement would amplify the characteristic uncanny curve.”
8 of 8

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 6

  1. Brenton, 2005, The uncanny valley: does it exist?
  2. Hanson, 2005, Expanding the aesthetic possibilities for humanoid robots,
  3. Hanson, 2006, Exploring the aesthetic range for humanoid robots,, 16
  4. Harnad, 1987, Introduction: psychophysical and cognitive aspects of categorical perception: a critical overview,, 1
  5. Ishiguro, 2006, Android science—toward a new cross-interdisciplinary framework, Interact. Stud, 7, 297, 10.1075/is.7.03mac · DOI
  6. Ramey, 2005, The uncanny valley of similarities concerning abortion, baldness, heaps of sand, and humanlike robot,

Cited by

Sorted by citations

See all 500

Metadata

Authors
Kätsyri, J., Förger, K., Mäkäräinen, M., & Takala, T.
Affiliations
Aalto University (FI)
Venue
Frontiers in Psychology
Published
2015-04-10
Volume
6
Pages
390-390
ISSN
1664-1078 (ISSN-L)
OpenAlex
W2094482505
Cited by
500
References
73
Citations by year
2026: 71 · 2025: 67 · 2024: 76 · 2023: 36 · 2022: 65 · 2021: 47 · 2020: 27 · 2019: 35 · 2018: 34 · 2017: 21 · 2016: 16 · 2015: 3
FWCI
3.85
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Frontiers in Psychology (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Aaltodoc (Aalto University) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Topics
  • Social Sciences › Psychology › Social Psychology › Social Robot Interaction and HRI
  • Physical Sciences › Computer Science › Human-Computer Interaction › Virtual Reality Applications and Impacts
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
Keywords
uncanny valley, eeriness, anthropomorphism

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-09, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-09 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review