A Review and Clarification of the Terms "holistic," "configural," and "relational" in the Face Perception Literature

Abstract

It is widely agreed that the human face is processed differently from other objects. However there is a lack of consensus on what is meant by a wide array of terms used to describe this "special" face processing (e.g., holistic and configural) and the perceptually relevant information within a face (e.g., relational properties and configuration). This paper will review existing models of holistic/configural processing, discuss how they differ from one another conceptually, and review the wide variety of measures used to tap into these concepts. In general we favor a model where holistic processing of a face includes some or all of the interrelations between features and has separate coding for features. However, some aspects of the model remain unclear. We propose the use of moving faces as a way of clarifying what types of information are included in the holistic representation of a face.

Keywords: composite task; configural; holistic; inversion; moving faces; part-whole task; relational.

Figure 1

An interpretation of a relational/configural model of face perception integrating the Diamond and Carey (1986) and Rhodes (1988) models. The model starts with first-order (isolated) features (that are processed in a part-based manner), moves on to the emergence of second-order (relational) features (that are subject to configural processing), and finishes with increasingly complex higher-order features (attributions) which can involve a combination of first and/or second-order features.

Figure 2

Two different models of holistic processing. (A) A face perception model adapted from Farah (1996). Object and face perception are independent of one another. Faces are perceived as undifferentiated wholes while objects are processed on the basis of their individual parts. (B) A holistic/part-based model of face perception. Holistic/configural and part-based processing work in parallel, making separate contributions to face perception that can be assessed independently of one another.

Figure 3

The original versions of the composite task: (A) the naming version of the task used for familiar faces. The top (or bottom) half of the composite is harder to identify when the halves are aligned (pictured left) as opposed to misaligned (pictured right). In this example the top half belongs to Kevin Rudd and the bottom half belongs to John Howard (both former Prime Ministers of Australia). (B) The same/different version using unfamiliar faces (where the top half is the half-to-match).

Figure 4

Examples of unaltered, feature shape inclusive and feature center point only conceptual models [based on models proposed by McKone and Yovel (2009)] for (A) a neutral face, (B) a turning face, and (C) a yawning face. Black squares represent key points from which configural information is calculated, black lines represent configural information, white circles represent other contour points, and white lines follow the shapes of the features. Center-point only images have been blurred to emphasize that feature shape does not factor into the holistic representation in this model.