[1]
This is a simplified view. A closer look into Gopnik & Meltzoff (1997) shows that their version of TT accounts for interaction as part of the development of an understanding of action and agency (Chap. 5). But interaction is not accounted for in the further dimensions of understanding others. From a bird’s eye view this characterization is not inadequate, although it needs qualification. As the reader will see, my person model theory integrates this initial understanding of action and agency as elements of forming implicit person models that at the beginning may not be rich and abstract enough to warrant being called a theory (see n. 6 below).
[2]
It is important to note that I leave it open whether we have to rely on a package of knowledge we are warranted in calling a theory, since I only discuss the strategy of information processing, not the organization of prior knowledge in experts.
[3]
Although our basic perception mainly relies on certain central cues—e.g., wide-open eyes for fear—the fearful face is not recognized only in one central feature of the face. It requires the integration of several facial features, and not static ones alone. The perceiver also benefits from noticing dynamic visual features like gaze direction: If the gaze is directed away from the perceiver instead of towards her, then this makes the recognition of fear occur faster (see Adams & Kleck 2003; Sander et al. 2006). Together with colleagues I have argued elsewhere that emotion recognition is essentially a process of pattern recognition (Newen et al. forthcoming). This is true for these basic perceptions of emotions. The face is integrated with body posture, since facial expressions are categorized as expressing a specific emotion most rapidly when they are paired with emotionally congruent body postures (Meeren et al. 2005; van den Stock et al. 2007).
[4]
This view was worked out in parallel by Anika Fiebich in her PhD thesis, under my supervision. She applied the thesis in discussing the case of autism (defended January 2013).
[5]
Gopnik and Meltzoff insist that the basic registration of objects—e.g., their being sensitive to object permanence, as well as the basic registration of agents rooted in their being able to distinguish inanimate objects and living beings—which babies develop very early on, shows that they already have an initial theory of objects and agents. They argue that the already innate “structures are rich enough and abstract enough to merit the name of theories themselves” (Gopnik & Meltzoff 1997, p. 82). But it is questionable whether the notion of theory really has any fruitful role here, because, for example, explanations and predictions of the behaviour of a baby when seeing an object are extremely constrained. The developmental story told by Gopnik and Meltzoff is of course very plausible and at some point turns into a theory, because the transformation of the representation in the context of new cognitive abilities comes with a rich and systematic package of explanations and predictions.
[6]
This includes, e.g., the ST, which mainly offers a claim about how we use our knowledge to understand others, and that the main source of this knowledge—in addition to situational input—is one’s own experience. But a representative of ST can easily grant that relevant prior knowledge is organized in a folk-psychological theory. She only insists that the strategy of application of this knowledge in a situation is a simulation process.
[7]
An important question which I cannot discuss in this paper is the question of the development of person model and the limits of application. Some very sketchy remarks may be of help here for urgent questions: Concerning the development I suggest that person model enfolds gradually from an early model of living agents which is based on sensitivity for clusters of features indicating animacy and agency. This “agent models” enfold into person models which are systematically enriched by the features I describe as belonging to person schemata and person images. Furthermore, a creation of a person model (which is a unity of information clustered together) does not presuppose a concept of a person. Person models are developed in fact if some typical features of adult healthy human beings are clustered to model an individual or a group of entities which are relevantly similar to adult healthy human beings. Typical core features are e.g., 1. being an agent, 2. being a sentient being, 3. having some minimal control of action. We use person models to understand babies and pets since we usually perceive them as having a minimal amount of core features.
[8]
I am only presupposing a minimal consensus on using the distinction of implicit versus explicit. It indicates a (gradual) difference in epistemic access such that paradigmatic cases of explicit contents are easily accessible (by the subject’s experience, memory, thinking, imagining etc.) while paradigmatic cases of implicit contents are very difficult to access by the subject while they nevertheless influence the subject’s cognition and behaviour. Intuitively, explicit content are correlated with our intuitive understanding of conscious accessibility, but since the latter is scientifically pretty unclear, I do not want to ground the implicit/explicit distinction on the difference between being or not being consciously accessible.
[9]
Mental phenomena have different ontological types: states, events, processes, and dispositions. So not only are stable mental phenomena included but so are situational experiences (like tokens of perceptions, emotions, attitudes, etc.).
[10]
The distinction between person schema and person image is based on Shaun Gallagher’s distinction between body schema and body image. Establishing a person schema of my own body amounts to Gallagher’s body schema, while a person image of my own body is similar to what he introduces as body image (2005, p. 24).
[11]
In a more detailed explication of the theory, it would indeed be useful to distinguish short-term person models (only stored in working memory) and long-term person models (stored in a long-term memory). In addition, other established distinctions in memory can be used to characterize the content of person models, such as procedural and declarative contents as well as episodic and semantic contents. I will, however, ignore these distinctions in this paper.
[12]
The time course can be observed in ERP studies. These studies all support claims about the early information processing of faces, although there is an ongoing debate about how best to interpret the results. The main observations are enhanced responsiveness to faces relative to a variety of other objects with peaks at approximately 100 milliseconds (Herrmann et al. 2005; Liu et al. 2002; Pegna et al. 2004), 170 milliseconds (Bentin et al. 1996; Eimer & McCarthy 1999; Itier & Taylor 2004), and 250 milliseconds (Bentin & Deouell 2000; Schweinberger et al. 2004) after stimulus onset. (For review see Macrae & Quadflieg 2010). Whole bodies (without faces) are evaluated with a delay of 20 milliseconds compared to the evaluation of faces (Gliga & Dehaene-Lambertz 2005). Concerning faces with emotional expressions, the following rather stable result is reported: there is a frontocentral positivity as early as 120 milliseconds after stimulus onset and a later more broadly distributed positivity beyond 250 milliseconds; both are modulated by emotional facial expressions (Eimer & Holmes 2002; Holmes et al. 2003; Vuilleumier & Pourtois 2007; Williams et al. 2006).
[13]
We leave the question open as to what extent person schemata are constituted by innate or by learned dispositions. The examples mentioned above indicate that they involve properties of both kinds.
[14]
Interestingly, Germans could perceive the friendliness of students from the US and UAE partially (as well as the other way around), while students from UAE and USA could not read the level of friendliness from the other culture at all (Bente et al. 2010).
[15]
This is the aspect of the narrative approach to understanding other minds, mentioned above (e.g., Hutto 2008). But narratives are only one method of establishing a person model. Representatives of a pure narrative approach underestimate the importance of other sources, such as perceptions, feelings, interactions, etc., which often do not involve narratives.
[16]
There is a long and not fully understood process of development from implicit false belief sensitivity to explicit false belief understanding (de Bruin & Newen 2012a; 2012b). Person images actually presuppose an explicit representation of false beliefs.
[17]
The involvement of identity already at the level of implicit schemata is supported by Haxby’s model of face perception according to which we have to distinguish a core cognitive system involving the recognition of face identity and an extended cognitive system which is enabling the recognition of facial expression (Haxby et al. 2000).
[18]
In the literature there are discussed one-factor accounts to explain mental disorders, e.g., in the case of schizophrenia (Gallagher 2004): a top-down approach argues that disturbances of higher-order cognition is the only source for thought insertion (Stephens & Graham 2000) while a bottom-up approach argues that thought insertation is a product of disturbances of neural or basic cognitive processes (like perception). Most of the recent accounts are hybrid account which we call two-factor theories.
[19]
The fact that person identity as a component of person schema formation is not only based on visual but also on an emotional evaluation is supported by the case of prosopagnosia, i.e., the inability to recognize the face of the person one is seeing, even though one is able to see and perceive the rest of the person adequately. Despite the fact that a person suffering from prosopagnosia is not able to see the familiarity of the face, we can measure increased skin conductance for familiar but not unfamiliar faces, thereby demonstrating intact (albeit covert) emotional recognition of known others (de Haan et al. 1992; Tranel & Damasio 1985).
[20]
For a discussion of delusional phenomena, see Coltheart et al. (2007) and Hirstein (2005).
[21]
These types of cases are considered in Gallagher & Hutto (2008), in the section “Pragmatic Intersubjectivity”. Their view is close to a multiplicity view. A minor criticism is that we have to account for such cases independently from being in interaction with someone. They may also involve only observing the other.
[22]
There is already one dynamic model of person construal available in the literature that also supports my dynamic theory of understanding others with person models, i.e., the model of Freeman & Ambady (2011). Despite its merits in describing social perception in more detail as regards the interrelation of bottom-up and top-down processes, the authors neither account for the claim that our rich prior information is mainly organized on the level of persons (not faces or subpersonal features), nor do they account for the interaction between person models and situation models.