게슈탈트 '부분이 아닌 전체로서의 현상과 상태'

[문형철]

게슈탈트 심리학의 시각 지각에 대한 세기:

(Gestalt Psychology in Visual Perception: I. Perceptual Grouping and Figure-Ground Organization)

논문의 전체 내용을 포괄적으로 다루며,

역사적 배경, 핵심 원리, 메커니즘, 신경 기반, 비교 및 도전 과제 등을 포함.

이 논문은 게슈탈트 심리학의 시각 지각 측면을 중점으로 하며,

게슈탈트 상담(치료)과 직접 연결되지는 않으나,

게슈탈트 원리가 상담에서 활용되는 전체성(holism)과 지각 조직의 기초를 이해하는 데 유용.

게슈탈트 상담은

이 심리학적 기반에서 유래한 치료 접근으로,

현재 경험의 전체성을 강조.

인식 perception ---> 지각 recognition

있는 그대로 ---> 인간의 지각대로 

지각의 왜곡

에고의 착각

있는 그대로 Self realization

1. 서론: 게슈탈트 심리학의 탄생과 발전

게슈탈트 심리학은

1912년 Max Wertheimer의 phi-motion 실험으로 시작되었습니다

이 실험은 두 개의 정지된 자극을 번갈아 제시할 때

순수한 운동 지각이 발생한다는 것을 보여주며,

지각이 개별 요소의 합이 아닌 전체적인 'Gestalt'(형태)로 구성된다는 핵심 아이디어를 제시했습니다.

Wertheimer, Wolfgang Köhler, Kurt Koffka 등의 초기 연구자들은

지각을 holistic(전체적)으로 접근하였으며,

이는 구조주의 원자론적 분석과 대비됩니다.

참고) 문치연 재해석

낮은 의식의 장(200이하)

perception vs cognition

pain                suffering

갈증                배고픔

사랑욕구         가려움, 질병의 고통

자극                stress response

욕구                미해결 과제

높은 의식의 장(200~499)

perception vs cognition

pain                치료대상, 쉬어라 신호

갈증                물마심, 만성탈수의 이해

사랑욕구         연인, 사랑의 체험

자극                반응

욕구                해결과제, 삶의 체험

높은 의식의 장(500이상)

perception vs cognition

pain                가부좌 명상, 승화

갈증                생명 반응, 물에 대한 사랑

사랑욕구         무조건적 사랑, 인류애

자극                자극과 Self 체험

욕구                사랑, 평화, 깨달음

1920년대에 Köhler는 Physical Gestalten에서 물리적 현상에도 Gestalt 원리를 적용하여 필드 이론을 제안했습니다. 이는 뇌의 전기적 필드가 지각을 형성한다는 가설이었으나, 후에 실험적으로 반박되었습니다. Wertheimer의 1923년 논문 Laws of Organization in Perceptual Forms는 지각 조직의 법칙(proximity, similarity, good continuation, closure, Prägnanz)을 체계화했습니다. Koffka의 1935년 Principles of Gestalt Psychology는 이러한 원리를 확장하여, 지각이 동적 상호작용에서 emergence(출현)하는 전체성임을 강조했습니다.

1940-1950년대에 게슈탈트 심리학은 쇠퇴기를 맞았습니다. Karl Lashley(1951)와 Roger Sperry(1955)의 연구가 뇌의 필드 모델을 반박하였고, David Hubel과 Torsten Wiesel(1968)의 시각 피질 연구가 모듈러 접근을 주류로 만들었습니다. 그러나 1980년대부터 revival(부흥)이 시작되었습니다. 예를 들어, von der Heydt(1984)의 환상 윤곽 연구와 Navon(1977)의 global precedence 효과가 게슈탈트 원리를 현대적으로 재해석했습니다.

이 역사적 맥락은

게슈탈트가 단순한 현상학적 관찰이 아닌,

지각의 본질적 메커니즘을 탐구하는 패러다임으로 이해해야 합니다.

게슈탈트 상담에서 이는 클라이언트의 경험을 '전체'로 보는 접근의 기초가 됩니다.

예를 들어,

Perls의 게슈탈트 치료는

이 원리를 바탕으로 '미완성 Gestalt'를 완성하는 과정을 강조합니다.

2. 지각 그룹화의 핵심 원리

게슈탈트 원리는 시각 자극을 그룹화하여 의미 있는 형태로 조직합니다.

고전적 원리와 현대 확장된 원리를 아래 테이블로 요약합니다.

원리                                     정의                                                  강도 (엔트로피 감소율)     발달적 출현            신경 상관

근접성 (Proximity)	공간적으로 가까운 요소가 그룹화됨	~75% (가장 강력)	신생아	V1/V2 맥락적 조절
유사성 (Similarity)	색상, 방향 등 동일한 특징이 그룹화	중간	6-7개월	V2 클러스터링
좋은 연속성 (Good Continuation)	부드럽고 공선적인 경로가 통합	높음 (방향/곡률)	3-4개월	V2 윤곽 검출기
폐쇄성 (Closure)	폐쇄된 형태 완성	약함 (1D → 2D)	3개월 미만 없음; 청소년기 발달	V2 환상 윤곽 뉴런
대칭성/평행성 (Symmetry/Parallelism)	전체적 형태 규칙성	대칭 평행에서 빠른 처리	초기	V4 대칭 검출기
볼록성 (Convexity)	곡선 경계 선호	~90% 다중 영역 디스플레이	선천적 편향	V2 곡률 코딩
새로운 원리	동시성, 공통 영역, 요소 연결성, 극단적 에지, 윤곽 엔트로피	맥락 의존적	학습 의존적	재귀 네트워크 (피드백 루프)

이 원리들은 Wertheimer(1923)의 고전적 법칙에서 유래하며, 현대 연구(Elder & Goldberg, 2002)에서 엔트로피 감소로 정량화됩니다. 예를 들어, 근접성은 무작위 배열에서 그룹화를 75%까지 예측합니다. 발달적으로, 일부 원리는 선천적(근접성)이고 다른 것은 학습적(폐쇄성)입니다. 신경적으로, V1의 저수준 필터에서 시작하여 V2의 고수준 통합으로 이어집니다.

게슈탈트 상담 관점에서,

이러한 원리는 클라이언트의 지각 왜곡을 이해하는 데 유용합니다.

예를 들어,

트라우마 경험에서 '좋은 연속성'이 깨지면

전체적 자아 인식이 왜곡될 수 있습니다.

3. 윤곽 통합과 완성 메커니즘

3.1 윤곽 통합 (Contour Integration)

이산적 요소(점, 에지)를

연속적 윤곽으로 결합합니다.

게슈탈트 큐를 통해 엔트로피를 줄입니다.

생태학적 기반으로,

uncorrelated 큐가 팩토리얼 가능성을 높입니다(Elder & Goldberg, 2002).

컴퓨테이션 모델(Jacobs, 1996; Lowe, 1985)은

컴퓨터 비전 알고리즘으로 활용됩니다.

통합은 확률적 과정으로, 베이지안 추론과 유사합니다.

자연 장면에서 큐의 독립성은 그룹화의 효율성을 설명합니다.

3.2 윤곽 완성 (Contour Completion)

• Amodal 완성: 가려진 부분, 감각적 윤곽 없음.
• Modal 완성: 환상 윤곽, 감각적 품질 있음. Relatability 기준: 교차 연장, <90° 회전 각도(Kellman, 1991). 형태 제약: 곡률 극성, 표면 기하학(Singh, 2007).

상담 적용: 클라이언트의 불완전한 기억을 '완성'하는 과정에서 amodal 완성을 비유적으로 사용.

3.3 큐 조합

고전적 관점: 경쟁적(단순성 원리). 현대 관점: 시너지적 베이지안 통합(Claessens, 2008). 자연 장면에서 큐가 협력합니다.

4. 배경 조직

시각 장면을 фигура(전경)와 배경으로 분리합니다.

주요 큐를 테이블로 요약:

큐                                            역할                                                 임계값/                                      강도예시

볼록성	фигура 영역 선호	~90% 다중 영역	볼록 vs. 오목 완성
대칭성	전체적 형태 편향	~75% 고정 근처	대칭 фигура 팝아웃
극단적 에지	에지 선명화	볼록에 높은 임계	입체적 깊이 지각
에지-영역 그룹화	윤곽이 영역에 결합	맥락 민감	Kanizsa 도형
관절 운동	동적 분리	전주의적	Ternus 디스플레이
전진 영역	운동에서 фигура 출현	주의 조절	겉보기 운동 패러다임

발달: 선천적(근접성) vs. 학습적(폐쇄성). 신경 기반: V1의 фигура 강화(Lamme, 1995); V2의 경계 소유권(Zhou, 2000).

상담 맥락: 클라이언트가 '배경'으로 밀려난 감정을 'фигура'로 가져오는 치료 과정과 유사.

5. 신경 메커니즘

영역                                      기능                                                            증거

V1	저수준 필터, 맥락적 조절	Carandini et al. (2005)
V2	환상 윤곽, 경계 소유권	von der Heydt (1984), Zhou (2000)
재귀 네트워크	맥락 통합, 동역학	Grossberg (1985), Lamme (1995)
피드백 루프	탑다운 주의, 경계 할당	Qiu & von der Heydt (2007)

전주의적 vs. 주의 조절: 경계 소유권은 짧은 지연(<30ms)으로 피드백 관여.

대학원 분석: 재귀 네트워크는 게슈탈트의 동적 필드를 현대적으로 재현합니다.

6. 비교, 대조, 보완성

차원게슈탈트 (원래)주류 비전 과학보완성

전체성 vs. 모듈성	상호의존적 부분/전체	독립 모듈 (V1 채널)	베이지안 모델이 단순성과 가능성 통합
큐 강도	선천적, 내재 법칙	경험적, 학습 (유아 연구)	발달 데이터가 선천 큐 정제
신경 모델	필드 동역학 (반박됨)	계층적, 피드포워드	재귀 네트워크가 게슈탈트 동역학과 정렬
진실성	내부 단순성 법칙	장면 통계 규칙	내부 vs. 외부 제약 조화

게슈탈트는 전체성을 강조하나, 주류 과학은 분석적입니다. 보완으로 베이지안 접근이 제안됩니다.

7. 제한과 도전 과제

도전설명제안 해결

체계적 분석	그룹화/фигура-배경의 공통 vs. 특이 요인	Gestalts의 등급 연속체 (약함 → 강함)
작업 패러다임	자극 유발 vs. 지시 작업	다중 안정 격자와 타겟 검출 연결
이론 통합	그룹화 → фигура-배경 연결	표면 폐쇄 곡선 보완
컴퓨테이션 갭	프로세스 모델 부족	정량적, 심리물리학 적합 모델 개발
신경컴퓨테이션 브리징	생리학 & 심리물리학 단절	다학제 팀 (심리물리학자, 모델러, 신경생리학자)
진실성	내부 단순성 vs. 외부 통계	베이지안 통합 내부 선행과 장면 가능성

8. 결론과 미래 방향

논문은 게슈탈트 원리가 현상학적 법칙에서 정량적, 생태학적, 신경 기반 모델로 진화했다고 결론짓습니다. 그러나 통합 이론 부족, 연구 파편화, 컴퓨테이션 프로세스 모델 필요 등이 남아 있습니다. '게슈탈트 부흥'을 위해 전체성(게슈탈트)을 분석적, 확률적, 신경과학적 접근과 합성합니다.

미래 단계:

1. 심리물리학 통합 중간 수준 비전 모델 개발.
2. 실험 패러다임 연결 (그룹화 → фигура-배경).
3. 동적 장면에서 내부 단순성 법칙 해결.
4. 다학제 노력으로 게슈탈트 원리를 주류 비전 과학에 임베드.

9. 게슈탈트 상담과의 연계 (확장 해석)

이 논문은 지각 중심이지만, 게슈탈트 상담(Fritz Perls, 1940s)에서 핵심입니다. 상담에서 'Gestalt'는 미완성 경험을 전체화합니다. 예: 클라이언트의 지각 그룹화 왜곡을 수정. 대학원 수준에서, 이 원리를 인지치료와 비교(예: Beck의 인지 왜곡)하여 보완성을 탐구할 수 있습니다.

Psychol Bull

. Author manuscript; available in PMC: 2012 Nov 1.

Published in final edited form as: Psychol Bull. 2012 Jul 30;138(6):1172–1217. doi: 10.1037/a0029333

A Century of Gestalt Psychology in Visual Perception I. Perceptual Grouping and Figure-Ground Organization

Johan Wagemans 1, James H Elder 2, Michael Kubovy 3, Stephen E Palmer 4, Mary A Peterson 5, Manish Singh 6, Rüdiger von der Heydt 7

• Author information
• Article notes
• Copyright and License information

PMCID: PMC3482144  NIHMSID: NIHMS390005  PMID: 22845751

The publisher's version of this article is available at Psychol Bull

Abstract

In 1912, Max Wertheimer published his paper on phi motion, widely recognized as the start of Gestalt psychology. Because of its continued relevance in modern psychology, this centennial anniversary is an excellent opportunity to take stock of what Gestalt psychology has offered and how it has changed since its inception. We first introduce the key findings and ideas in the Berlin school of Gestalt psychology, and then briefly sketch its development, rise, and fall. Next, we discuss its empirical and conceptual problems, and indicate how they are addressed in contemporary research on perceptual grouping and figure-ground organization. In particular, we review the principles of grouping, both classical (e.g., proximity, similarity, common fate, good continuation, closure, symmetry, parallelism) and new (e.g., synchrony, common region, element and uniform connectedness), and their role in contour integration and completion. We then review classic and new image-based principles of figure-ground organization, how it is influenced by past experience and attention, and how it relates to shape and depth perception. After an integrated review of the neural mechanisms involved in contour grouping, border-ownership, and figure-ground perception, we conclude by eval‎uating what modern vision science has offered compared to traditional Gestalt psychology, whether we can speak of a Gestalt revival, and where the remaining limitations and challenges lie. A better integration of this research tradition with the rest of vision science requires further progress regarding the conceptual and theoretical foundations of the Gestalt approach, which will be the focus of a second review paper.

Keywords: Gestalt, grouping principles, figure-ground organization, neural mechanisms, vision science

1 Introduction

Exactly 100 years ago Wertheimer (1912) published his paper on phi motion—perception of pure motion, without object motion—which many consider to be the beginning of Gestalt psychology as an important school of thought. The present status of Gestalt psychology is ambiguous. On the one hand, many psychologists believe that the Gestalt school died with its founding fathers in the 1940s or after some devastating empirical findings regarding electrical field theory in the 1950s, or that it declined because of fundamental limitations that blocked further progress, while stronger theoretical and experimental frameworks arose in the 1960s and 1970s that have dominated the field ever since (e.g., cognitive science, neuroscience). On the other hand, almost all psychology textbooks still contain a Gestalt-like chapter on perceptual organization (although often poorly connected to the other chapters), and new empirical papers on Gestalt phenomena appear with increasing frequency.

We are convinced that Gestalt psychology is still relevant to current psychology in several ways. First, questions regarding the emergence of structure in perceptual experience and the subjective nature of phenomenal awareness (e.g., visual illusions, perceptual switching, context effects) continue to inspire contemporary scientific research, using methods and tools that were not at the Gestaltists’ disposal. Second, the revolutionary ideas of the Gestalt movement continue to challenge some of the fundamental assumptions of mainstream vision science and cognitive neuroscience (e.g., elementary building blocks, channels, modules, information-processing stages). Much progress has been made in the field of nonlinear dynamical systems, both theoretically and empirically (e.g., techniques to measure and analyze cortical dynamics), progress that allows modern vision scientists to surpass some of the limitations in old-school Gestalt psychology as well as those in mainstream vision research.

The centennial anniversary of Gestalt psychology is therefore an excellent opportunity to take stock of what we have discovered about core Gestalt phenomena of perceptual organization and how our understanding of the underlying mechanisms has evolved since Wertheimer’s seminal contribution. Due to this review’s scope, we divide it in two parts: This paper deals with perceptual grouping and figure-ground organization, whereas the second covers modern developments regarding the general conceptual and theoretical frameworks that underlie Gestalt ideas (e.g., holism, emergence, dynamics, simplicity). In Table 1, we provide an overview of the topics covered in the first review paper, together with the section headings, the questions or issues being raised, and some of the answers provided. One of the aims of our review is to remove the many misunderstandings surrounding Gestalt psychology, which are listed separately in Table 2, along with a more balanced view on the actual state of affairs.

Table 1.

Overview of the paper with section numbers and headings, questions and issues raised, and answers provided

SectionNo.Section TitleQuestions/Issues/Answers

1.	General Introduction	we motivate why an extensive review of 100 year of research on perceptual organization is valuable
2.	A Brief History of Gestalt Psychology	we address four questions regarding Gestalt psychology:
2.1.	The Emergence of Gestalt Psychology	(1) how did it start?
2.2.	Essentials of Gestalt Theory	(2) what does it stand for?
2.3.	Further Development, Rise, and Fall of Gestalt Psychology	(3) how did it evolve?
2.4.	The Current Status of Gestalt Psychology	(4) where does it stand now?
3.	Perceptual Grouping
3.1.	Introduction	- we distinguish grouping and figure-ground organization - we enumerate the classic grouping principles: proximity, similarity, common fate, symmetry, parallelism, continuity, closure - we review progress in our understanding of perceptual grouping since the early days of Gestalt psychology; specifically:
3.2.	New Principles of Grouping	(1) we discuss a number of additional principles that have been discovered since the initial set were described: generalized common fate, synchrony, common region, element connectedness, uniform connectedness
3.3.	Grouping Principles in Discrete Static Patterns	(2) we demonstrate how at least some grouping principles can be measured experimentally and expressed in quantitative laws: (a) when several orientations can be perceived based on grouping by proximity in a particular dot lattice, the outcome is determined by the relative distance alone, not by the angle between the competing organizations (affecting the global symmetry of the lattice and how it looks) (b) when grouping by proximity and grouping by similarity are concurrently applied to the same pattern, the two principles are combined additively
3.4.	Grouping Principles in Discrete Dynamic Patterns	(3) we review a century of research on grouping in dynamic patterns, incl. Korte’s laws, element and group motion in Ternus displays, space-time coupling versus space-time trade-off
3.5.	At What Level Does Grouping Happen?	(4) we demonstrate that grouping principles operate at multiple levels: provisional grouping takes place at each stage of processing, possibly with feedback from higher levels to lower ones, until a final, conscious experience arises of a grouping that is consistent with the perceived structure of the 3-D environment
3.6.	Conclusion
4.	Contour Integration and Completion
4.1.	Introduction	we distinguish contour grouping (integration) and contour completion
4.2.	Grouping Principles for Contour Integration	we discuss the grouping principles that play a role in contour integration: proximity, good continuation, similarity, closure, symmetry, parallelism, convexity
4.3.	Contour Completion	we review several issues regarding contour completion; specifically:
4.3.1.	Modal and amodal completion	we distinguish modal and amodal completion
4.3.2.	Grouping and shape problem	we distinguish the grouping problem and the shape problem
4.3.3.	Contour interpolation and extrapolation	- we distinguish contour interpolation and extrapolation - we address 2 questions: (a) what geometric properties of the visible contours are used by human vision? (b) how are these variables combined to define the shape of the contour?
4.3.4.	Surface geometry and layout	we discuss the role of surface geometry and layout in contour completion
4.4.	Some General Issues Regarding Perceptual Grouping and Contour Integration	we address the following general questions regarding perceptual grouping and contour integration:
4.4.1.	Development	(1) to what extent are the Gestalt laws innate or learned?
4.4.2.	Cue combination	(2) how are they combined?
4.4.3.	Computational models	(3) how can they be jointly represented in accurate computational models and useful algorithms?
4.5.	Conclusion
5.	Figure-Ground Organization
5.1.	Introduction	- we distinguish the structuralist and Gestalt positions - we discuss Wertheimer’s criteria to demonstrate that past experience affects initial figure-ground organization
5.2.	Classic Image-Based Configural Principles of Figure-Ground Organization	we discuss the classic configural principles of figure-ground organization: convexity, symmetry, small region, surroundedness
5.3.	New Image-Based Principles of Figure-Ground Organization	we discuss new image-based principles of figure-ground organization: lower region, top-bottom polarity, extremal edges and gradient cuts, edge-region grouping, articulating motion, advancing region motion, contour entropy as a ground cue (+ part salience, axiality)
5.4.	Nonimage-Based Influences on Figure-Ground Perception	we discuss the evidence for nonimage-based influences on figure-ground organization: past experience, attention and perceptual set
5.5.	Figure-Ground Organization in Relation to Shape and Depth Perception	we discuss how figure-ground organization relates to shape and depth perception
5.6.	Conclusion
6.	Neural Mechanisms in Contour Grouping, Figure-Ground Organization, and Border-Ownership Assignment
6.1.	Introduction	- we review the neurophysiological studies investigating the neural mechanisms in contour grouping, figure-ground organization, and border-ownership assignment in an integrated way - in doing so, we demonstrate how contemporary neuroscience has embraced Gestalt ideas, while doing justice to Hubel and Wiesel’s heritage in the following three ways:
6.2.	Context Integration in Illusory Contours	(1) we demonstrate how the responses of cortical neurons can depend on the parameters of the stimulus in its receptive field as well as on the properties of the overall configuration in the visual field
6.3.	Figure-Ground Organization and Border-Ownership Assignment	(2) we substantiate the Gestalt postulate of autonomous organization processes that form primary units of perception
6.4.	Involuntary Organization and Volitional Attention	(3) we refine our understanding about the role of attention in these processes of perceptual organization
6.5.	Conclusion
7.	General Discussion and Conclusion
7.1.	The Swinging Pendulum of Gestalt History
7.2.	Gestalt Research Anno 2012
7.3.	Limitations and Challenges to Contemporary Research on Perceptual Organization
7.4.	Conclusion

Open in a new tab

Table 2.

Common misunderstandings about Gestalt psychology

A. GENERALCommon assumptionActual state of affairs

Gestalt psychology is completely dead and buried because its limitations have never been overcome.	Interesting work in the Gestalt tradition is still being carried out and many limitations and shortcomings have been overcome or addressed (see Table 4 for an overview).
Gestalt psychology was a radical, simple-minded theory which has been rejected.	Many of the ideas of Gestalt psychology are still very much alive. A century of research has allowed several more synthetic positions, integrating some of the original Gestalt positions with alternative positions (see Table 5 for an overview).
All fundamental issues pertaining to perceptual grouping and figure-ground organization are solved.	Important problems regarding perceptual grouping and figure-ground organizations are still unsolved. Some of these are mentioned in the course of the discussion in this paper. There are still some controversial issues and open questions that continue to stimulate contemporary research. A number of challenges are listed separately in the final section of this paper (see Table 6 for an overview).

B. SPECIFICCommon assumptionActual state of affairs

Grouping principles are mere textbook curiosities only distantly related to normal perception.	Grouping principles pervade virtually all perceptual experiences because they determine the objects and parts we perceive in the environment.
Gestalt psychology has claimed that all Gestalt laws are innate and that learning or past experience can never play a role.	Gestalt psychology has emphasized the autonomy of the Gestalt laws but it has not claimed that all Gestalt laws are innate and that learning or past experience can never play a role.
The Gestalt theory about brain function is rejected by the empirical evidence.	Köhler’s specific conjecture about electromagnetic brain fields appears to be rejected by experiments by Lashley and Sperry, but advances in neurophysiology have confirmed the existence of pre-attentive mechanisms of visual organization postulated by Gestalt theory. The more abstract notion of the brain as a physical Gestalt can also be implemented as recurrent networks with closed feedback loops, which can be proven to converge to an equilibrium state of minimum energy.
Vague Gestalt notions about whole-processes in the brain are now completely replaced by precise single-cell recordings demonstrating that neurons operate like primitive detectors.	Neurophysiology has come a long way since Hubel and Wiesel’s atomistic approach to orientation-selectivity of single cells in cat and monkey cortex, taken as prototypical feature detectors. The current literature emphasizes the role of context-sensitive, autonomous processes within recurrent networks.

Open in a new tab

To put these two reviews in perspective, we first introduce the key findings and ideas of the founders of Gestalt psychology, along with a brief sketch of its further development, rise, and fall (for an extensive treatment, see Ash, 1995). The historical section ends with a discussion of Gestalt psychology’s empirical and conceptual problems and an indication of how these limitations are being addressed in current research. We then review current research on perceptual grouping and figure-ground organization in more detail in the remaining sections. We focus on these two topics because they were the most important ones in the Gestalt tradition of perceptual organization and still are today, even for vision in general.

2 A Brief History of Gestalt Psychology

This section addresses four questions regarding Gestalt psychology: (1) How did it start? (2) What does it stand for? (3) How did it evolve? (4) Where does it stand now?

2.1 The Emergence of Gestalt Psychology

What Max Wertheimer discovered in 1912 was called phi motion, a special case of apparent motion. (For an excellent discussion of its historical importance, see Sekuler, 1996; for a demonstration of the phenomenon and for a review of its misrepresentation in later sources, see Steinman, Pizlo, & Pizlo, 2000.) According to the conventional view of apparent motion, we see an object at several successive positions and motion is then “added” subjectively. If this were correct, then an object would have to be seen as moving, and at least two positions—the starting and end points—would be required to produce seen motion. Neither of these conditions held in the case of phi motion. In the key experiment, a white strip was placed on a dark background in each of two slits in the wheel of a tachistoscope, and the rotation speed was adjusted to vary the time required for the light to pass from one slit to the next (i.e., the interval between the two). Above a certain threshold value (~200 ms), observers saw the two lines in succession. With much shorter intervals (~30 ms), the two lines appeared to flash simultaneously. At the optimal stage (~60 ms), observers perceived a motion that could not be distinguished from real motion. When the interval was decreased slightly below 60 ms, after repeated exposures, observers perceived motion without a moving object—that is, pure phenomenal or phi motion. Although only three observers were tested, “the characteristic phenomena appeared in every case unequivocally, spontaneously, and compellingly” (Wertheimer 1912/1961, p. 1042). In the same paper, Wertheimer proposed a physiological model described in terms of a short circuit and a flooding back of the current flow (“transverse functions of a special kind;” Wertheimer, 1912/1961, p. 1085), which produced what he called “a unitary continuous whole-process” (Wertheimer, 1912/1961, p. 1087). He then extended this theory to the psychology of pure simultaneity (for the perception of form or shape) and of pure succession (for the perception of rhythm or melody). These extensions were decisive for the emergence of Gestalt theory.

2.2 Essentials of Gestalt Theory

The phi phenomenon was the perception of a pure process, a transition that could not be composed from more primitive percepts of a single object at two locations. In other words, perceived motion was not added subjectively after the sensory registration of two spatiotemporal events but had its own phenomenological characteristics and ontological status. From this phenomenon, Wertheimer concluded that structured wholes or Gestalten, rather than sensations, are the primary units of mental life. This was the key idea of the new and revolutionary Gestalt theory, developed by Wertheimer and his colleagues in Berlin. An overview of how the Berlin school of Gestalt psychology distinguished itself from the dominant view of structuralism and empiricism, as well as of related Gestalt schools is given in Table 3.

Table 3.

Key claims by the Berlin school of Gestalt psychology in opposition to other schools

   Berlin school of Gestalt psychology     (Wertheimer, Köhler, Koffka)Opposing schools

	Structuralism/associationism/empiricism (von Helmholtz, Wundt)
structured wholes or Gestalten are the primary units of mental life	sensations are the primary units of mental life
experimental phenomenology: perceptual experience must be described in terms of the units people naturally perceive	introspection: perceptual experience must be analyzed as combinations of elementary sensations of physical stimuli as their building blocks
percepts arise on the basis of continuous whole-processes in the brain; percepts organize themselves by mutual interactions in the brain	percepts are associated combinations of elementary excitations
perceptual organization is based on innate, intrinsic, autonomous laws	perceptual organization is based on perceptual learning, past experience, intentions
simplicity or minimum principle	likelihood principle
	Graz school of Gestalt psychology (Meinong; von Ehrenfels, Benussi)
Gestalten (structured experiences, wholes) are different from the sum of the parts	Gestalt qualities are more than the sum of the constituent primary sensations
two-sided or reciprocal dependency between parts and wholes - there are specifiable functional relations that decide what will appear or function as a whole and what as parts - often the whole is grasped even before the individual parts enter consciousness	one-sided dependency between parts and wholes (the wholes depend on the parts, but the parts do not depend on the whole)
perception “emerges” through self-organization; perception arises non-mechanistically through an autonomous process in the brain	perception is “produced” on the basis of sensations
	Leipzig school of Gestalt psychology (Krüger, Sander)
no analysis into stages, but functional relations in the emergence of Gestalts can be specified by Gestalt laws of perceptual organization	stage theory: “Aktualgenese”, microgenesis
holism integrated with natural science (physical Gestalten, isomorphism, minimum principle)	mystic holism, segregated from natural science

Open in a new tab

The notion of “Gestalt” had already been introduced into psychology by Christian von Ehrenfels in his essay “On Gestalt qualities” (1890). Based on the observation that humans can recognize two melodies as identical even when no two corresponding notes in them have the same frequency, von Ehrenfels argued that these forms must possess a “Gestalt quality”—a characteristic that is immediately given, along with the elementary sensations that serve as its foundation, a characteristic that is dependent on its constituent objects but rises above them. For von Ehrenfels, Gestalt qualities rest uni-directionally on sense data: Wholes are more than the sums of their parts, but the parts are the foundation (“Grundlage”) of the whole. In contrast, Wertheimer claimed that functional relations determine what will appear as the whole and what will appear as parts (i.e., reciprocal dependency). Often the whole is grasped even before the individual parts enter consciousness. The contents of our awareness are by and large not additive but possess a characteristic coherence. They are structures that are segregated from the background, often with an inner center, to which the other parts are related hierarchically. Such structures or “Gestalten” are different from the sum of the parts. They arise from continuous global processes in the brain, rather than combinations of elementary excitations.

With this step, Wertheimer separated himself from the Graz school of Gestalt psychology, represented by Alexius Meinong, Christian von Ehrenfels, and Vittorio Benussi. They maintained a distinction between sensation and perception, the latter produced on the basis of the former. The Berlin school, represented by Max Wertheimer, Kurt Koffka, and Wolfgang Köhler, considered a Gestalt as a whole in itself, not founded on any more elementary objects. In their view, perception was not the product of sensations but it arose through dynamic physical processes in the brain. As a result, the Berlin school also rejected stage theories of perception proposed by the Leipzig school, represented by Felix Krüger and Friedrich Sander, in which the gradual emergence of Gestalten (“Aktualgenese” or “microgenesis”) played a central role. Although the Berlin theorists adhered to a nonmechanistic theory of causation and did not analyze the processes into stages, they did believe that the functional relations in the emergence of Gestalts could be specified by laws of perceptual organization.

2.3 Further Development, Rise, and Fall of Gestalt Psychology

Two major developments are generally considered as highlights in the history of Gestalt psychology: Köhler’s discussion of “physical Gestalten” (1920) and Wertheimer’s proposal of “Gestalt laws of perceptual organization” (1923). Köhler (1920) extended the Gestalt concept from perception and behavior to the physical world, thus attempting to unify holism (i.e., the doctrine stressing the importance of the whole) and natural science. He proposed to treat the neurophysiological processes underlying Gestalt phenomena in terms of the physics of field continua rather than that of particles or point-masses. In such continuous field systems, which he called strong Gestalten, the mutual dependence among the parts is so great that no displacement or change of state can occur without influencing all the other parts of the system. Köhler showed that stationary electric currents, heat currents, and all phenomena of flow are strong Gestalten in this sense. These he distinguished from what he called weak Gestalten, which do not show this mutual interdependence.

In addition, Köhler tried to construct a specific testable theory of brain processes that could account for perceived Gestalten in vision. He thought of visual Gestalten as the result of an integrated process in what he referred to as “the entire optical sector,” including retina, optical tract, and cortical areas, as well as transverse functional connections among conducting nerve fibers (i.e., a recurrent neural network in modern terms). He proposed an electrical field theory, in which “the lines of flow are free to follow different paths within the homogeneous conducting system, and the place where a given line of flow will end in the central field is determined in every case by the conditions in the system as a whole” (Köhler, 1920/1938, p. 50). In modern terms, Köhler had described the visual system as a self-organizing physical system.

These ideas led Köhler to postulate a psychophysical isomorphism between the psychological reality and the brain events underlying it: “actual consciousness resembles in each case the real structural properties of the corresponding psycho-physiological process” (Köhler, 1920/1938, p. 38). By this he meant functional instead of geometrical similarity indicating that brain processes do not take the form of the perceived objects themselves. In addition, he insisted that such a view does not prescribe complete homogeneity of the cortex but is perfectly compatible with functional articulation. Experiments to establish the postulated connections between experienced and physical Gestalten in the brain were at the time nearly impossible to conduct, but decades later, Köhler attempted to do so (see below).

Around the same time, Max Wertheimer further developed his Gestalt epistemology and outlined the research practice of experimental phenomenology that was based on it. He first stated the principles publically in a manifesto published in Volume 1 of Psychologische Forschung in 1922. Wertheimer called for descriptions of conscious experience in terms of the units people naturally perceive, rather than the artificial ones imposed by standard scientific methods. By assuming that conscious experience is composed of units analogous to physical point-masses or chemical elements, psychologists constrain themselves to a piecemeal inquiry into the contents of consciousness, building up higher entities from constituent elements, using associative connections. In fact, such and-summations (“Und-Summe”), as Wertheimer called them, appear “only rarely, only under certain characteristic conditions, only within very narrow limits, and perhaps never more than approximately” (Wertheimer 1922/1938, p. 13). Rather, what is given in experience “is itself in varying degrees ‘structured’ (‘gestaltet’), it consists of more or less definitely structured wholes and whole-processes with their whole-properties and laws, characteristic whole-tendencies and whole-determinations of parts” (Wertheimer 1922/1938, p. 14). The perceptual field does not appear to us as a collection of disjointed sensations, but possesses a particular organization of spontaneously combined and segregated objects.

In 1923, Wertheimer published a follow-up paper, which was an attempt to elucidate the fundamental principles of that organization. The most general principle was the so-called law of Prägnanz, stating, in its most general sense, that the perceptual field and objects within it will take on the simplest and most encompassing (“ausgezeichnet”) structure permitted by the given conditions. For Köhler (1920), this tendency towards the Prägnanz of the Gestalt was just another example that phenomenal Gestalten were like physical Gestalten: As shown by Maxwell and Planck, all processes in physical systems, left to themselves, show a tendency to achieve the maximal level of stability (homogeneity, simplicity, symmetry) with the minimum expenditure of energy allowed by the prevailing conditions. More specific principles that determine perceptual organization according to Wertheimer were proximity, similarity, uniform density, common fate, direction, good continuation and “whole properties” (or “Ganzeigenschaften”) such as closure, equilibrium, and symmetry.

Empirical work on these principles existed before Wertheimer’s landmark paper (for a recent review, see Vezzani, Marino, & Giora, 2012), but now the general claim that perceptual experience is organized was turned into a complex open-ended research program aimed at the discovery of the laws or principles governing perceptual organization in both its static and dynamic aspects. It is this research program that Wertheimer, Koffka and Köhler started to work on with their students, once they had acquired professorships at major universities in Germany in the 1920s and 1930s. We cannot cover this flourishing period of the Berlin school of Gestalt psychology extensively here, but a few highlights that deserve mentioning in passing are studies by Kurt Gottschaldt on embedded figures (1926), Joseph Ternus on phenomenal identity (1926), Karl Duncker on induced motion (1929), Wolfgang Metzger on a homogeneous Ganzfeld (1930) and motion in depth (1934). In the meantime, Gestalt thinking also affected research on other sense modalities (e.g., binaural hearing by Erich von Hornbostel), on learning and memory (e.g., Otto von Lauenstein and Hedwig von Restorff), and on thought (e.g., Karl Duncker). Later, Gestalt theory was also applied to action and emotion (by Kurt Lewin), to neuropathology and the organism as a whole (by Adhemar Gelb and Kurt Goldstein), and to film theory and aesthetics (by Rudolf Arnheim). This period marked the high point but not the end of Gestalt psychology’s theoretical development, its research productivity, and its impact on German science and culture.

Around this time, Gestalt theory also started to have some impact on research in the U.S., mainly owing to Wolfgang Köhler and Kurt Koffka (see King & Wertheimer, 2005, Chapter 10). For instance, Koffka’s (1935) notion of vector fields inspired some interesting empirical work published in the American Journal of Psychology (Brown & Voth, 1937; Orbison, 1939). Reviews of Gestalt psychology appeared in Psychological Review on a regular basis (e.g., Helson, 1933; Hsiao, 1928), a comprehensive book on state-of-the-art Gestalt psychology was published as early as 1935 (Hartmann, 1935), and three years later Ellis’s (1938) influential collection of translated excerpts of core Gestalt readings made some of the original sources accessible to a non-German-speaking audience. Already in 1922, at Robert Ogden’s invitation, Koffka had published a full account of the Gestalt view on perception in Psychological Bulletin.

At first sight, Gestalt theory seemed to develop rather consistently, from studying the fundamental laws of psychology first under the simplest conditions, in elementary problems of perception, before including complex sets of conditions, and turning to other domains such as memory, thinking, emotion, aesthetics, and so forth. At the same time, however, the findings obtained did not always fit the original theories, which posed serious challenges to the Gestalt framework. Even more devastating to the development of Gestalt psychology was the emergence of the Nazi regime in Germany from 1933 to World War II. In this period, many of the psychology professors at German universities lost their posts because of the discrimination and prosecution of Jews, so they emigrated to the U.S. to take on new positions there. The works by German psychologists who stayed, for instance, Edwin Rausch’s monograph on “summative” and “nonsummative” concepts (1937) and Wolfgang Metzger’s (1941) psychology textbook, were largely ignored outside Germany. Metzger’s synoptic account of research on the Gestalt theory of perception entitled “Gesetze des Sehens” (“Laws of seeing”), first published in 1936 and later reissued and vastly expanded three times, was only translated into English in 2006.

After emigrating to the U.S., the founding fathers of Gestalt psychology did not perform many new experiments. Instead, they mainly wrote books in which they outlined their views (e.g., Koffka, 1935; Köhler, 1940; Wertheimer, 1945). The major exception was Köhler who had taken up physiological psychology using EEG recording and other methods in an attempt to directly verify his isomorphism postulate. Initially, his work with Hans Wallach on figural aftereffects appeared to support his interpretation in terms of satiation of cortical currents (Köhler & Wallach, 1944). Afterwards, he was able to directly measure cortical currents—as EEG responses picked up from electrodes at the scalp—whose flow direction corresponded to the direction of movement of objects in the visual field (Köhler & Held, 1949).

Soon after that breakthrough, however, Lashley and colleagues (Lashley, Chow, & Semmes, 1951) performed a more critical test of Köhler’s electric field theory and its underlying postulate of isomorphism. If the flows of current picked up from the scalp in Köhler and Held’s experiments indeed reflected the organized pattern of perception and not merely the applied stimulation, and if that pattern of perception would result from a global figure-field across the whole cortex, a marked alteration of the currents should distort the perception of these visual figures. By inserting metallic strips and metal pins in large regions of the visual cortex of rhesus monkeys, Lashley et al. could short-circuit the cortical currents. Surprisingly, the monkeys could still perform the learned shape discriminations, demonstrating that global cortical currents were not a necessary condition for pattern perception. In subsequent experiments, Sperry and colleagues (Sperry, Miner, & Myers, 1955) performed extensive slicing and dense impregnation with metallic wires across the entire visual cortex of cats, and showed that these animals too could still perform rather difficult shape discriminations (e.g., between a prototypical triangle and distorted variants). Together, these two studies effectively ruled out electrical field theory as an explanation of cortical integration and undermined the empirical basis of any isomorphism between cortical flows of current and organized patterns of perception. Köhler (1965) naturally reacted to these developments but his counterarguments and suggestions for further experiments were largely ignored, and to most scientists at the time, the matter was closed. Electrical field theory, which had been one of the pillars of Gestalt psychology’s scientific basis, was considered dead and buried.

While Gestalt psychology declined in the English-speaking world after World-War II, Italy remained a stronghold of Gestalt psychology. For instance, Metzger dedicated the third edition of his “Gesetze des Sehens” to his “Italian and Japanese friends.” Among his friends were Musatti, Metelli, and Kanizsa—three major figures in Italian psychology. In spite of being Benussi’s student and successor (from the Graz school), Cesare Musatti was responsible for introducing the Berlin school of Gestalt psychology in Italy and training important students in this tradition—most notably Metelli and Kanizsa, whose contributions continue to be felt today. Fabio Metelli is best known for his work on the perception of transparency (e.g., Metelli, 1974). Gaetano Kanizsa’s most famous studies were performed in the 1950s with papers on subjective contours (e.g., the so-called Kanizsa triangle), modes of color appearance, and phenomenal transparency (Kanizsa, 1954, 1955a, 1955b), although their impact came much later, when he started to publish in English (Kanizsa, 1976, 1979).

In addition to Italy, Gestalt psychology was also strong in Belgium and in Japan. Albert Michotte became famous for his work on the perception of causality (1946/1963), arguing strongly against an inferential, associationist, empiricist account of it, like other Gestalt psychologists had done for other aspects of perception. For him, causality is perceived directly, not derived from more primitive sensations through some cognitive operation, and this percept could be shown to be tightly coupled to specific higher-order attributes in the spatiotemporal events presented to observers. He also introduced the notions of modal and amodal completion (Michotte, Thinès, & Crabbé, 1964), and studied several configural influences on these processes. (For a further discussion of Michotte’s heritage, see Wagemans, van Lier, & Scholl, 2006.) Building on earlier collaborations of Japanese students with major German Gestalt psychologists (e.g., Sakuma with Lewin, Morinaga with Metzger), Gestalt psychology continued to develop in Japan after World-War II. For instance, Oyama did significant work on figural aftereffects (e.g., Sagara & Oyama, 1957) and perceptual grouping (e.g., Oyama, 1961).

2.4 The Current Status of Gestalt Psychology

Despite signs of well-deserved respect in the U.S. and in Germany (e.g., Köhler’s APA presidency in 1957; Wertheimer’s posthumous Wilhelm Wundt Medal in 1983), the ideas of the Gestaltists were received with ambivalence. On the one hand, they were recognized for raising central issues and provoking important debates in psychology, theoretical biology, and other fields, but on the other hand, their mode of thinking and research style did not sit comfortably in the intellectual and social climate of the postwar world, and they were confronted with vehement criticism. Two sets of explanations have been given for this outcome (Ash, 1995). The first emphasizes institutional, political, and biographical contingencies. Koffka, Köhler and Wertheimer all left for the U.S. and obtained positions where they could do excellent research but could not train PhDs. The Gestalt school’s further expansion was also handicapped by the early deaths of Max Wertheimer in 1943 and Kurt Koffka in 1941, as well as many other Gestalt psychologists of the first and second generation (e.g., Duncker, Gelb, Lauenstein, Lewin, von Restorff). In Germany, Metzger, Rausch, and Gottschaldt did have a large number of PhD students, but few of them carried on in the Gestalt tradition. A notable exception is Lothar Spillmann, who obtained his D. Phil. with Metzger in Münster in 1964 and who pioneered the impact of Gestalt ideas in modern neurophysiology ever since (e.g., Spillmann, 1999, 2009).

The second set of explanations concerns scientific issues of a methodological and conceptual nature (summarized in the left column of Table 4). Compared to the rigor of psychophysics and behaviorism, Gestalt psychology was severely criticized for offering mere demonstrations, using either very simple or confounded stimuli, formulating laws with little precision, and adding new “laws” for every factor shown to have an influence on perceptual organization. In the 1950s and 1960s, its critics increasingly insisted on causal explanations, by which they meant cognitive operations in the mind that could be modeled as computer algorithms or neural mechanisms that could be attributed to the properties of single cells that were discovered by Hubel and Wiesel in that period. In addition, serious conceptual limitations appeared when Gestalt thinking was extended to other areas such as personality and social psychology (e.g., Richard Crutchfield, Solomon Asch, Fritz Heider, David Krech). The further the metaphors were stretched, the harder it became to connect them to Köhler’s concept of a self-organizing brain and his speculations about electromagnetic brain fields.

Table 4.

Problems in old-school Gestalt psychology and how they are solved in contemporary research

ProblemsSolutions

Mere demonstrations based on direct (subjective) reports	Real experiments (1) - also indirect methods (matching, priming, cueing) and performance measures (accuracy, reaction time) - also psychophysical techniques (thresholds in detection/discrimination tasks) - also neuropsychological studies with brain-damaged patients
Either very simple or confounded stimuli	Carefully constructed stimuli, sometimes also richer stimuli (1) - allowing research of everyday tasks - allowing research of ecological foundations
Grouping principles and laws of perceptual organization studied in isolation	Also studying relationships with other processes (1), e.g., - perceptual grouping in relation to depth perception, lightness perception - visual contour completion in relation to surface geometry and layout - figure-ground organization in relation to shape and depth perception
Laws formulated with little precision	Quantification, which allows measurement (1, 2)
Proliferation of laws	Unification into stronger, better developed theoretical frameworks (2)
No mechanistic understanding	Computational models (1, 2)
Poor understanding of neural basis	Somewhat better understanding of neural basis (1, 2)

Open in a new tab

Note. 1 = Paper 1 and 2 = Paper 2.

Despite these criticisms, Gestalt thinking did not disappear from the stage completely. In the slipstream of Shannon’s information theory, a few researchers tried to provide a quantitative underpinning to the central Gestalt notion of simplicity (e.g., Attneave, 1954; Attneave & Arnoult, 1956; Hochberg & McAlister, 1953; Leeuwenberg, 1969, 1971; for a review, see Hatfield & Epstein, 1985). A number of independent, original scientists working on perception and information processing kept some Gestalt issues on the research agenda (e.g., Fred Attneave, Wendell Garner, Julian Hochberg, Irvin Rock). These became more prominent again with the discovery of true Gestalt phenomena such as global precedence in hierarchical letters (e.g., Navon, 1977), configural superiority effects based on emergent features (e.g., Pomerantz, Sager, & Stoever, 1977), and the importance of hierarchical structure in perceptual representations (e.g., Palmer, 1977). The experimental paradigms were derived from standard methods in cognitive psychology, and the results were incorporated into mainstream information-processing accounts (e.g., Beck, 1982; Kubovy & Pomerantz, 1981). In the major alternative approaches to visual perception—the ecological (e.g., Gibson, 1971) and computational (e.g., Marr, 1982) approaches—the influence of Gestalt thinking has also been acknowledged explicitly. In the last two or three decades, perceptual grouping and figure-ground organization—the most central topics of Berlin school research—have returned to center stage (e.g., Kimchi, Behrmann, & Olson, 2003), although the relationship to the original Gestalt theory (e.g., two-sided dependency between wholes and parts, minimum principle) is not always clear.

In the remainder of this paper, as well as in a second more theoretically oriented paper (Wagemans et al., 2012), we review the later developments in more detail (summarized in the right column of Table 4). We start with research on perceptual grouping in simple displays (Section 3) and extend this to contour grouping, integration, and completion in more complex shapes and real-world images (Section 4). In the next section, we cover research on figure-ground perception, where many of the factors affecting grouping, in addition to unique factors, exert an influence (Section 5). Although links to neural mechanisms are mentioned throughout, we also provide a more integrated account of the literature on the neural mechanisms of contour grouping and figure-ground organization in a separate section (Section 6). This review demonstrates that research from the last two or three decades has addressed (and partially solved) some of the major methodological and conceptual shortcomings in old-school Gestalt psychology.

3 Perceptual Grouping3.1 Introduction

Historically, the visual phenomenon most closely associated with perceptual organization is grouping: the fact that observers perceive some elements of the visual field as “going together” more strongly than others. Indeed, perceptual grouping and perceptual organization are sometimes presented as though they were synonymous. They are not. Grouping is one particular kind of organizational phenomenon, albeit a very important one. Another is figure-ground organization. In general, grouping determines what the qualitative elements of perception are, and figure-ground determines the interpretation of those elements in terms of their shapes and relative locations in the layout of surfaces in the 3-D world.

Max Wertheimer first posed the problem of perceptual grouping in his ground-breaking 1923 paper by asking what stimulus factors influence the perceived grouping of discrete elements. He first demonstrated that equally spaced dots do not group together into larger perceptual units, except as a uniform line (Figure 1A) and then noted that when he altered the spacing between adjacent dots so that some dots were closer than others, the closer ones grouped together strongly into pairs (Figure 1B). This factor of relative distance, which Wertheimer called proximity, was the first of his famous laws or (more accurately) principles of grouping.

Figure 1.

Open in a new tab

Illustration of several grouping principles (adapted from Palmer, 2002a).

Wertheimer went on to illustrate other grouping principles, several of which are portrayed in Figure 1. Parts C, D, and E demonstrate different versions of the general principle of similarity: all else being equal, the most similar elements (in color, size, and orientation for these examples) tend to be grouped together. Another powerful grouping factor is common fate: all else being equal, elements that move in the same way tend to be grouped together. Notice that both common fate and proximity can actually be considered special cases of similarity grouping, with velocity and position as the relevant properties, respectively. Further factors influencing perceptual grouping of more complex elements, such as lines and curves, include symmetry (Figure 1G), parallelism (Figure 1H), and continuity or good continuation (Figure 1I). Continuity is important in Figure 1I because observers perceive it as containing two continuous intersecting lines rather than as two angles whose vertices meet at a point. Figure 1J illustrates the effect of closure: all else being equal, elements that form a closed figure tend to be grouped together. This display also shows that closure can dominate continuity, since the very same elements that were organized as two intersecting lines in Figure 1I are now organized as two angles meeting at a point in part Figure 1J.

One might think that such grouping principles are mere textbook curiosities only distantly related to normal perception. On the contrary, they pervade virtually all perceptual experiences because they determine the objects and parts that people perceive in the environment. (Hence, they also affect other sensory modalities; for a thorough discussion of grouping principles in audition, see Bregman, 1990; for a recent review of Gestalt principles in tactile perception, see Galace & Spence, 2011.) A practical application of the Gestalt principles is camouflage, which results when the same grouping processes that would normally make an organism stand out from its environment as a separate object, cause it to be grouped together with its surroundings instead. For instance, the same leopard that is clearly visible when it is seen in a tree against the uniform sky is difficult to see against a mottled, leafy backdrop—until it moves. Even perfect static camouflage is undone by the principle of common fate. In sum, camouflage and camouflage breaking provide an ecological rationale for the principles of grouping.

Since the early days of Gestalt psychology, considerable progress has been made, including (1) the discovery of additional principles, (2) the experimental measurement of the strength of grouping factors and the development of quantitative laws, as well as (3) new insights into the level of processing at which perceptual grouping happens. These new developments will be described in the next section, where we will also discuss their possible structural and ecological basis.

3.2 New Principles of Grouping

3.2.1 Generalized common fate

One of the most powerful of the classic grouping principles is common fate—the tendency for elements that “move together” to be perceived as a unitary entity (Wertheimer, 1923). The possibility that Wertheimer may have had a much broader range of phenomena in mind, however, is suggested by a passage of his seminal article that is not widely known because it was not included in Ellis’s translation (Wertheimer, 1923/1938): “Also this principle [of common fate] is valid in a wide range of conditions; how wide is not yet investigated here” (Wertheimer, 1923, p. 316; our own translation). In this vein, Sekuler and Bennett (2001) presented an extension of common fate to grouping by common luminance changes. They found that when elements of a visual scene become brighter or darker simultaneously, even if they have different luminances throughout, observers have a powerful tendency to group those elements perceptually. It is as though the principle of common fate operates not only for the common motion of elements through 3-D physical space, but through luminance space as well. The structural rationale for generalized common fate is clear: It is another example of similarity grouping, but based on similarity of changes in feature values, such as luminance or position, rather than on the similarity of the feature values themselves. An ecological rationale for grouping by common luminance changes might lie in the simultaneous brightening or darkening that occurs across a spatial area when the level of illumination changes (e.g., with the appearance of sunlight or shadows; see also van den Berg, Kubovy, & Schirillo, 2011).

3.2.2 Synchrony

Synchrony is the tendency for elements that change simultaneously to be grouped together (Alais, Blake & Lee, 1998; Lee & Blake, 1999). The changes do not have to be in the same direction, however, as they do in generalized common fate. A random field of black and white dots whose luminances change in polarity randomly over time against a gray background, for example, will segregate into two distinct regions if the dots in one area change synchronously rather than randomly. Grouping by synchrony can be considered as an even more general form of common fate in which the simultaneous changes do not have to involve either motion, as in classic common fate, or common direction of change, as in generalized common fate. The structural basis for grouping by synchrony is clear: the simultaneous occurrence of visible changes of the elements that are grouped. Such grouping makes sense because it reflects a strong temporal regularity in the stimulus event.

The ecological rationale behind grouping by synchrony is far less clear, however. Objects in the natural environment seldom change their properties in different directions or along different dimensions in temporal synchrony. Indeed, it is difficult even to devise plausible examples of ecological situations that would exhibit this kind of temporal regularity without some form of extended common fate being involved. Nevertheless, synchrony grouping may arise from some very general nonaccidentalness detection mechanism, possibly connected to the perception of causality (e.g., Michotte, 1946/1963). The argument is that the temporal coincidence of multiple changes is unlikely to be due to chance alone, and so it must have some common underlying cause related to an ecological event that relates the synchronously changing elements.

A radically different and quite controversial, rationale is that temporal synchrony of changes drives grouping because synchrony of neural firings is the physiological mechanism by which the brain codes all forms of grouping (e.g., Milner, 1974; von der Malsburg, 1981). The argument is that if the environment drives the neural substrate to produce synchronous firing by virtue of synchronous changes, the changing elements will automatically be grouped because of the synchronous firing. Some researchers report evidence that seems to support this claim (e.g., Gray & Singer, 1989; Singer & Gray, 1995), but others disagree (e.g., Shadlen & Movshon, 1999). This issue is discussed further in the second paper (Wagemans et al., 2012; Section 4). Further controversy surrounds synchrony grouping because it has been claimed that such grouping effects are actually produced by stimulus artifacts that can be detected by the early visual system (Farid, 2002; Farid & Adelson, 2001). These challenges are complex, but the bottom line is that both the existence of grouping synchrony and the mechanism by which it occurs are currently unclear. In general, these controversies show quite clearly that the interest in perceptual grouping principles remains strong in contemporary research.

3.2.3 Common region

Common region is the tendency for elements that lie within the same bounded area (or region) to be grouped together (Palmer, 1992). An illustration is provided in Figure 1K, where the black dots that lie within the same ovals are likely to be grouped into pairs. The structural basis for grouping by common region appears to be that all the elements within a given region share the topological property of being “inside of” or “contained by” some larger surrounding contour. If it is viewed as similarity of containment, it can be related to several other grouping principles based on similarity (e.g., color, orientation, and size). Common region also appears to have an ecological rationale arising from textures and hierarchically embedded parts. When a bounded region encloses a number of image elements, they are likely to be elements on the surface of a single object, such as a leopard’s spots or the features of a face, rather than independent objects that just happen accidentally to lie within the same bounding contour.

Experimental evidence for the existence of common region as a grouping factor comes from studies using the Repetition Discrimination Time or RDT method (Beck & Palmer, 2002; Palmer & Beck, 2007). In a speeded discrimination task, observers were able to report the shape of a repeated element more quickly in a line of otherwise alternating shapes (e.g., squares and circles) when the repeated shapes were located within the same surrounding region than when they were located in two separate regions.

3.2.4 Element connectedness

Element connectedness is the tendency for distinct elements that share a common border to be grouped together. The important structural basis for this form of grouping is the topological property of connectedness (Palmer & Rock, 1994). Connectedness can be considered as the limiting case of the classic factor of proximity, but Palmer and Rock argued that framing it this way puts the cart before the horse, in the sense that one needs distinct units to speak meaningfully about their distance in the first place. The compelling rationale for element connectedness is ecological: Pieces of matter that are physically connected to each other in 3-D space are the primary candidates for being parts of the same object, largely because they tend to behave as a single unit. The bristles, metal band, and handle of a paint brush, for example, constitute a single object in large part because of their connectedness, as demonstrated by the fact that when you push one part, the other parts move rigidly along with it.

The effectiveness of element connectedness was demonstrated in a behavioral task using the RDT method (Palmer & Beck, 2007). As was the case for grouping by common region, displays with elements that were connected to each other produced reliably faster responses than displays with unconnected elements. Another behavioral result that provides striking support for the importance of element connectedness comes from a neuropsychological study by Humphreys and Riddoch (1993). Their patient, who was afflicted with Balint’s syndrome—a condition resulting from bilateral damage to parietal cortex that results in a deficit in perceiving more than a single “object” at any given time—was unable to discriminate between arrays containing many circles of just one color (either all red or all green) and arrays in which half of the circles were red and the other half green. However, if pairs consisting of one red circle and one green circle were connected by lines, the same patient was able to make the discrimination between one-color and two-color arrays. Unifying a pair of circles through element connectedness thus appears to enable these patients to perceive them as a single perceptual object so that they could see two circles at once, a feat that was impossible for them in the absence of the connecting line.

3.2.5 Uniform connectedness

This principle represents something of a departure from standard Gestalt ideas about perceptual organization because it addresses the question of how the initial organization into elements might occur. In his classic article on grouping, Wertheimer (1923) never actually mentioned where the to-be-grouped elements came from. Presumably, he believed that they were somehow derived from the grouping principles he articulated, but Palmer and Rock (1994) argued that they arise from the earlier organizational process of uniform connectedness (UC), which is the principle by which the visual system initially partitions an image into a set of mutually exclusive connected regions having uniform (or smoothly changing) properties, such as luminance, color, texture, motion, and depth. The UC elements thus created form the entry level units into a part-whole hierarchy that is created by grouping together different UC regions and, if necessary, by parsing them into lower-level elements at deep concavities (e.g., Hoffman & Richards, 1984).

Palmer and Rock’s claims regarding the foundational status of UC have not been uniformly accepted. Peterson (1994), for instance, argued that UC is one of many properties relevant to partitioning the visual field, and that UC units are not entry-level units. Kimchi (2000) examined the role of UC in experiments designed to reveal the gradual emergence, or microgenesis, of organizational processes using a primed matching task with displays containing connected or disconnected elements. The complex results she obtained were not consistent with UC being the sole determinant of entry-level units in a part-whole hierarchy, as Palmer and Rock (1994) proposed. Rather, they showed that collinearity and closure were at least as important, if not more so, in the initial organization that can be tapped by such methods. Nevertheless, the theoretical rationale for some organizational process like UC to contribute to creating a set of potential perceptual units on which further grouping and parsing can operate seems sound. Indeed, something like it is a standard assumption in most theories of computational vision (e.g., Marr, 1982).

Whereas the research on grouping principles, as described above, provides predictions about what elements in a display are likely to be grouped together, it does not reveal how strong each of the grouping principles is. This is the focus of the next sections, covering studies applying static and dynamic stimuli, respectively.

3.3 Grouping Principles in Discrete Static Patterns

3.3.1 Conceptual background

Starting with Wertheimer (1923), researchers have often used dot lattices to quantify grouping. A dot lattice is a collection of dots in the plane which is invariant under at least two translations, a (with length |a|) and b (whose length is |b|≥ |a|). These two lengths, and the angle between the vectors, γ (constrained by 60° ≤ γ ≤ 90°), define the basic structure of the lattice, by defining the parallelogram between each quartet of dots in the lattice (Kubovy, 1994, extending the work of Bravais, 1850/1949; see Figure 2A). The diagonals of this parallelogram are denoted c and d (where |c|≤|d|). In its canonical orientation, a is horizontal. More generally, the orientation of the lattice can be defined by the angle θ (measured counterclockwise) and |a| is called the scale of the lattice. Since scale is irrelevant to the invariant properties of the lattice and unimportant for grouping over a reasonable range, the relevant parameters are |b|/|a| and γ (Figure 2B). With these two parameters, six different types of lattices can be defined, each characterized by their symmetry properties.

Figure 2.

Open in a new tab

(A) Defining features of a dot lattice stimulus. (B) Two-dimensional space and nomenclature of dot lattices. Adapted from Kubovy (1994), with permission.

When grouping by proximity is pitted against grouping by similarity, displays consist of at least two kinds of elements (called motifs), separated by one of the translation components, resulting in dimotif lattices (Figure 3; see Grünbaum & Shephard, 1987). In order to characterize the relation between two grouping principles, one must construct grouping indifference curves (Figure 4), similar to the indifference curves used in micro-economics (Krantz, Luce, Suppes, & Tversky, 1971): Imagine a consumer who would be equally satisfied with a market basket consisting of 1 kg of meat and 4 kg of potatoes and another consisting of 2 kg of meat and 1 kg of potatoes. In such a case, the (meat, potato) pairs (1, 4) and (2, 1) both lie on an indifference curve.

Figure 3.

(a) 왼쪽 이미지: 근접성이 유사성에 의해 강화되는 경우

• 요소들의 **가장 짧은 거리 방향(shortest distance)**은 동일하지만, 두 번째로 짧은 거리(second shortest distance) 방향에서는 요소가 다릅니다.
• 결과: 근접성에 의한 그룹화(가까운 점들끼리 묶임)가 유사성(같은 모양/색상)에 의해 **강화(reinforced)**됩니다.
• 지각적으로: 수평 또는 특정 방향의 행/열로 강하게 그룹화되어 보입니다. 유사한 요소가 가까이 있으면 근접 효과가 더 강해집니다.

(b) 오른쪽 이미지: 근접성이 유사성에 의해 반대되는 경우

• 가장 짧은 거리 방향에서는 요소가 다르고, 두 번째로 짧은 거리 방향에서는 요소가 동일합니다.
• 결과: 근접성에 의한 그룹화가 유사성에 의해 **반대(opposed)**됩니다.
• 지각적으로: 근접성만으로는 한 방향으로 그룹화되려 하지만, 유사성이 이를 방해하여 다른 방향(예: 같은 모양끼리)으로 그룹화되거나 경쟁이 일어납니다.

Figure - PMC

Two dimotif rectangular dot lattices with |b||a|=1.2.

Figure 4.

Open in a new tab

Two grouping indifference curves. The abscissa, δa, represents the difference in luminance between adjacent elements of a. The ordinate, |b|, represents the distance between the dots of b (assuming |a| = 1). Only the equilibrium grouping indifference curve is achievable without independently measuring the strength of grouping by proximity. The methods to be described later allow us to plot a family of indifference curves. (The θ values are different for each of the four dot lattices.)

With these tools in hand, two important questions regarding perceptual grouping can be formulated. First, when several orientations can be perceived based on grouping by proximity in a particular dot lattice, what determines the preferred grouping? Is the outcome determined by the relative distance alone, or also by the angle between the competing organizations (an aspect that affects the global symmetry of the lattice and the way the overall configuration looks)? Second, when grouping by proximity and grouping by similarity are concurrently applied to the same pattern, what rule governs their joint application? Are these two principles combined additively or not? We briefly review the most important attempts to address these two questions. The initial studies always pitted grouping by proximity against grouping by similarity.

3.3.2 Initial attempts to quantify grouping by proximity by pitting it against similarity

The first to systematically study grouping by proximity in interaction with similarity was Rush (1937). In her experiment she showed observers sequences of dot lattices in which the distance between dots in one orientation was held constant, and the distance between dots in another orientation was reduced from trial to trial. She assumed—incorrectly as argued below—that one could measure the strength of the two principles by finding their point of equilibrium, and concluded that “…it may be said that Similarity equals about 1.5 cm of Proximity” (p. 90).

Roughly two decades later, Hochberg and Silverstein (1956, unaware of Rush’s work, as a footnote in Hochberg & Hardy, 1960, attests) also set out to solve the problem of measuring the strength of grouping by similarity by pitting it against grouping by proximity. In manipulating luminance differences or distances between dots, they produced grouping indifference curves (Figure 4). Reanalysis showed that an additive combination of proximity and similarity described their results best. Unfortunately, the logic employed by Hochberg and his colleagues suffered from the same flaw as Rush. Their method produced only one grouping indifference curve—the one for which both groupings are in equilibrium. Their method cannot produce grouping indifference curves for which one principle is 2× or 3× as strong as the other, which are needed to measure the relative strengths of the two principles.

Quinlan and Wilton (1998) studied the relations between grouping by proximity and two forms of grouping by similarity (by color and by shape). Their stimuli consisted of strips of seven elements, with the center element as the target. They manipulated proximity by slightly shifting the left or right set of three elements, and they also manipulated color and shape similarity. Observers were asked to rate the degree to which the target grouped with the elements on the left or on the right. Although the conception of the experiment is elegant, its reach was curtailed because each grouping principle was either present or absent. Had Quinlan and Wilton used a design in which each type of grouping was a multilevel factor, they could have addressed the additivity question (i.e., the second question introduced above), but they did not.

Oyama, Simizu, and Tozawa (1999) presented rectangular dimotif lattices for 3 s and asked observers to indicate continuously with a joystick whether they saw horizontal or vertical grouping. The horizontal separation was increased by 15’ after a “horizontal” response, and decreased by that amount after a “vertical” response. Using a double-staircase method, the ratio of vertical to horizontal distances was determined that matched a particular dissimilarity. This method produced more complete equilibrium grouping indifference curves than were obtained by Hochberg and his colleagues.

3.3.3 The pure–distance law and the additivity of grouping principles

Oyama (1961) was the first to show that one can measure the strength of grouping by proximity without pitting it against another grouping principle (e.g., grouping by similarity). Using rectangular dot lattices at a fixed orientation, he recorded the amount of time subjects reported seeing the competing horizontal and vertical groupings. The ratio of the time they saw the horizontal and vertical organizations was found to be a power function of the ratio of the horizontal and vertical distances th/tv = (dh/dv)−α, with α ≈ 2.89.

Using dot lattices at near-equilibrium, Kubovy and Wagemans (1995) and Kubovy, Holcombe, and Wagemans (1998) demonstrated that grouping by proximity can be understood as the outcome of a probabilistic competition among potential perceptual organizations. The basic idea is simple. If the distances between dots in two orientations of the lattice are equal, the chances of seeing one orientation or the other are equal too. If one distance becomes larger than the other, the relative chance of seeing that orientation decreases. If the ratio of the longer to the shorter vector is larger than about 1.5, grouping along that orientation is almost never seen. Kubovy and colleagues presented different kinds of dot lattices for 300 ms each and asked observers to indicate the perceived orientation. They could then use the frequencies of the perceived orientations over a large number of trials as estimates of the probabilities, and plot the relative frequencies as a function of relative distance. Their results (shown schematically in Figure 5) were remarkable. All the values of the log-odds fell on the same line, called the attraction function. Its slope is a person-dependent measure of sensitivity to proximity. Although the (relative) strength of grouping decays as an exponential function of (relative) distance, the attraction function in log-space is linear. The fact that all data points—obtained with all pairs of distances and all relative orientations (i.e., all points in the 2-D lattice space of Figure 2B)—could be fitted well by a single straight line indicates that grouping by proximity depends only on the relative distance between dots in competing organizations, not on the overall configuration in which the competition occurs (i.e., the lattice type, each with its own symmetry properties). Hence, this result, which was called the Pure Distance law provided a satisfactory answer to the first question raised above.

Figure 5.

Open in a new tab

The pure distance law (adapted from Kubovy et al., 1998, with permission).

Once it has been established how grouping varies as a function of relative distance, the effect of conjoined grouping principles can be determined by measuring a family of grouping indifference curves. Kubovy and van den Berg (2008) presented participants with rectangular lattices of dots of different contrasts. Dots with the same contrast were either arranged along the shorter axis of each rectangle of dots within the lattice (similarity and proximity in concert) or arranged along the longer axis (similarity and proximity working against each other). Dot lattices varied across two dimensions: the ratio between the short and long axis of each rectangle of dots within the lattice and the contrast difference between the different arrays of dots. As in the previous studies, each lattice was presented for 300 ms, and participants were asked to indicate which of the four orientations best matched the perceived arrangement of the dots in the lattice. By plotting the log likelihood of reporting the direction of the long axis versus the short axis as a function of the ratio of the length of the long and short axis for different values of the contrast difference between dots (shown schematically in Figure 6A), a family of grouping indifference curves was then obtained (depicted in Figure 6B). Because these indifference curves are parallel in log-odds space, the conjoined effects of proximity and similarity are called additive. Using lattices in which dots were replaced by Gabor elements, Claessens and Wagemans (2005) came to similar conclusions regarding proximity and collinearity. These results, therefore, provide a clear answer to the second question raised above.

Figure 6.

Open in a new tab

The conjoined effects of proximity and similarity are additive. The dashed lines in (A) turn into grouping indifference curves in (B). Adapted from Kubovy and van den Berg (2008), with permission.

3.4 Grouping Principles in Discrete Dynamic Patterns

Apparent motion is perceived when an object is presented in two or more successive frames at different spatial locations with proper durations and intervals. As discussed before, Wertheimer (1912) showed that under certain conditions it is possible to perceive pure motion, where motion is perceived without perceiving the moving object itself. The optimal timing and spacing between successively presented object presentations was investigated in more detail by Korte (1915), who found a direct relationship between the optimal temporal and the optimal spatial interval for perceived apparent motion. Later studies, however, have shown that the relationship between the optimal temporal and spatial interval depends on the stimuli used. For example, using horizontal arrays of dots that were displaced on successive frames, Burt and Sperling (1981) found that the spacing of the dots in the array strongly influenced the apparent motion percept in addition to the effects of the temporal interval between frames and the spatial displacement of the entire array.

The influence of spatial and temporal factors in apparent motion was further investigated using the Ternus display (Kramer & Yantis, 1997; Pantle & Picciano, 1976; Ternus, 1926; Wallace & Scott-Samuel, 2007), in which an array of three dots is presented across two frames at different spatial locations. When the two frames are presented in rapid succession (i.e., with a short inter-stimulus interval), it appears that the outmost dot is displaced while the center two dots appear to be stationary: element motion occurs. When the temporal interval between the successive frames is longer, the entire array of dots appears to jump: group motion is perceived. The two different types of perceived apparent motion represent two different solutions to the correspondence problem (Ullman, 1979), referring to the task of matching the objects in the first frame to the (possibly displaced) objects in the second frame. Whether element or group motion was perceived was found to depend on the properties of the individual stimuli in both frames such as their features (Dawson, Nevin-Meadows, & Wright, 1994), their size or the sharpness of their edges (Casco, 1990), as well as on the presence of contextual elements affecting how they are grouped (Kramer & Yantis, 1997) or how they are perceived in 3-D space (He & Ooi, 1999).

The interaction between spatial and temporal aspects was further investigated by Gepshtein and Kubovy (2000), who were able to determine the relationship between spatial grouping (determining which elements in each frame belong together) and temporal grouping (determining which elements across frames belong together), by using successive presentations of dot lattices— motion lattices—which allowed them to independently manipulate the strength of spatial and temporal groupings. A motion lattice (Figure 7) is composed of two identical dot lattices, D1 and D2, displayed in alternation. Two ratios determine the perceived motion: (1) the motion ratio rm = m2/m1, where m1 and m2 are the shortest and the next shortest spatial distances across which the apparent motion could occur between the frames; (2) the baseline ratio rb = b/m1, where b is the shortest spatial distance between the dots within D1 and D2 to which the apparent motion could apply. The orientation of a virtual line drawn through these dots is called the baseline orientation.

Figure 7.

Open in a new tab

A motion lattice. 

 and 

 are dot lattices presented in alternation.

As in the classic Ternus display, two classes of motion can be perceived. First, element motion is now apparent motion from each dot in D1 to a corresponding dot in D2 (and vice versa as the dot lattices alternate). The log-odds of seeing m2 rather than m1 as a function of the ratio of the distances is called an affinity function, by analogy with the concept of an attraction function for static dot lattices (Figure 8A). Second, group motion is now apparent motion orthogonal to the baseline orientation (Figure 7). Sequential models predict that if the spatial configuration of a stimulus remains constant, the likelihood of seeing group motion—an indicator of spatial grouping—cannot be affected by manipulations of the temporal configuration of the stimulus. However, the pattern of interaction in Figure 8B between rm, the temporal configuration of the stimulus, and rb, the relative density of the dots along the baseline, clearly refutes the sequential model.

Figure 8.

Open in a new tab

(A) The affinity function. (B) The objecthood functions.

How spatial and temporal distances interact to determine the strength of apparent motion has been controversial. Some studies report space-time coupling: If the spatial or temporal distance between successive stimuli is increased, the other distance between them must also be increased to maintain a constant strength of apparent motion (i.e., Korte’s third law of motion). Other studies report space-time trade-off: If one of the distances is increased, the other must be decreased to maintain a constant strength of apparent motion. To establish what determines whether coupling or trade-off occurs, Gepshtein and Kubovy (2007) generalized the motion lattice of Figure 7, as illustrated in Figure 9, showing a temporal component of m3, T3, of twice the magnitude of the temporal component of m1, T1. By manipulating the spatial components of these motions, S3 and S1 from S3 ≫ S1 to S3 ≪ S1, an equilibrium point between the extremes was found at r31 = S3/S1, for which the probability of seeing the two motions was the same. If r31 > 1 then space-time coupling holds; if r31 < 1 then space-time trade-off holds. This suggests that previous findings on apparent motion were special cases and that the allegedly inconsistent results can be embraced by a simple law in which a smooth transition from trade-off to coupling occurs as a function of speed: Trade-off holds at low speeds of motion (below ≈ 12°/s), whereas coupling (Korte’s law) holds at high speeds. The deeper theoretical implications of these results for the visual system’s economy principles are discussed in the second review paper (Wagemans et al., 2012; Section 4).

Figure 9.

Open in a new tab

A six-stroke motion lattice. (A) The successive frames are superimposed in space. Gray levels indicate time. b is the baseline distance. (B) The time course of the display. The three most likely motions along m1, m2, and m3 can occur because dots in frame fi can match dots in either frame fi+1 or frame fi+2. (C–D) Conditions in which different motion paths dominate: m1 in Panel C and m3 in Panel D. (The stimuli were designed so that m2 would never dominate.) Adapted from Gepshtein and Kubovy (2007), with permission.

The above research on perceptual grouping in static and dynamic discrete patterns spans a complete century, from Schumann (1900), Wertheimer (1912, 1923) and Korte (1915) up until today. It was mainly using well-controlled, parametrically varied stimuli in order to isolate one factor or another, and trying to quantify its strength. In addition, it sparked a renewed interest in understanding the level at which perceptual grouping operates, which was addressed in studies that use somewhat richer stimuli with additional variations, more typical for naturally occurring stimulation.

3.5 At What Level Does Grouping Happen?

As described above, Wertheimer (1923) demonstrated powerful grouping effects due to a large number of stimulus variables (e.g., proximity, similarity, good continuation) using flat 2-D displays on the printed page (see Figure 1). Subsequent researchers have investigated where in the visual system these effects occur (i.e., before or after the construction of a 3-D representation of the image), by using various kinds of 3-D displays with depth cues, shadows, transparency, and other higher-level factors.

Rock and Brosgole (1964) conducted a classic experiment on this topic to examine whether grouping by proximity operated on retinal 2-D distances or perceived 3-D distances. Observers in a dark room saw a 2-D array of luminous beads either in the frontal plane (perpendicular to the line of sight) or slanted in depth so that the horizontal dimension of the array was foreshortened. The beads were actually closer together vertically than horizontally, so that when they were viewed in the frontal plane, observers always reported seeing them grouped into vertical columns rather than horizontal rows. The critical question was whether or not the beads would be grouped in the same way when the same lattice was viewed slanted in depth such that the beads were retinally closer together in the horizontal direction. When this array was viewed monocularly, so that the beads appeared to be in a frontal plane perpendicular to the line of sight (even though they were actually slanted in depth), observers perceived the grouping to change to a set of rows rather than columns, as one would expect based on retinal distances. However, when viewed binocularly, so that stereoscopic depth information enabled observers to see the beads slanted in depth, they reported grouping them into vertical columns, as predicted by postconstancy grouping based on a 3-D representation of perceived distances in the phenomenal environment (because the beads appeared to be closer in the vertical direction, as was actually the case in the physical world). Rock and Brosgole’s results therefore support the hypothesis that the final, conscious result of grouping occurs after binocular depth perception. Several phenomenological demonstrations supporting the same conclusion are provided by Palmer (2002b; Palmer, Brooks & Nelson, 2003).

Rock, Nijhawan, Palmer and Tudor (1992) later investigated whether grouping based on lightness similarity happened before or after lightness constancy. Using displays that employed cast shadows and translucent overlays, they also found evidence that the final conscious result of grouping depended on a postconstancy representation that reflected the perceived reflectance of surfaces rather than the luminance of retinal regions. Analogous evidence that the final conscious organization resulted from a grouping process that operates on relatively late, postconstancy representations was reported by Palmer, Neff, and Beck (1996) for amodal completion and by Palmer and Nelson (2000) for illusory contours. Further results of Schulz and Sanocki (2003) support the view that prior to achieving the conscious result of perceptual grouping based on a 3-D postconstancy representation, some nonconscious grouping processes operate on a 2-D preconstancy representation. They used the same lightness displays as didRock et al. (1992), but included a brief, masked presentation condition in which they found that observers reported seeing an organization that is based on the retinal luminance of 2-D regions (see also van den Berg et al., 2011). Further evidence that grouping operations occur before constancy has been achieved is based on grouping effects that actually influence the achievement of constancy (see Palmer, 2003).

Perhaps the most parsimonious view consistent with the known facts is that grouping principles operate at multiple levels. It seems most likely that provisional grouping takes place at each stage of processing, possibly with feedback from higher levels to lower ones, until a final, conscious experience arises of a grouping that is consistent with the perceived structure of the 3-D environment. Whereas the above findings provide valuable information about the stages at which grouping operates, these studies have mainly employed relatively artificial stimuli. The next section is dedicated to the role of grouping in contour integration and completion, in ways that are closer to the processing of natural stimuli.