We can now address the second independent question concerning understanding others: How do we organize the information about other people that we already have? This question presupposes that in standard cases of understanding others we are not in a situation in which we are bereft of relevant background knowledge. Quite the contrary: most of the time, we interact with people about whom we have a lot of background knowledge—family members, colleagues, friends, etc. Furthermore, we have background knowledge about the general needs of human beings, the special needs of students, homeless people, etc. It seems clear that we are relying on this type of knowledge in an essential way when we understand others. There may be very short period as a newborn baby when we start from scratch, armed only with certain inborn minimal mechanisms such as neonate imitation. Even the social smile developed with two months is dependent on external stimulation and learning processes, and babies very quickly start to react selectively towards familiar and foreign individuals. They also expect a typical behavioural interactive pattern from the caregiver. If a mother stops reacting intuitively through normal facial expressions and gestures, and instead reacts with a “still face”, then the baby quickly starts to cry (Bertin & Striano 2006; Nagy 2008). The baby is irritated by the unexpected pattern of reaction. How, then, are all these different types of background information about the other organized and used in social understanding?
The question of whether we are organizing our knowledge according to folk-psychological theories has received a number of different answers. According to TT, this is exactly what happens. In understanding others we rely on folk-psychological rules such as: “If she desires an ice-cream and she believes that she can get one with her money at the cafeteria, then she will go to the cafeteria”. No doubt folk-psychological rules, organized according to a belief–desire psychology, are an important instrument for understanding others; but they are by no means the only one. Often it is sufficient to know the conventions in a society to understand what someone is doing and will do next, e.g., if someone is in Japan and he enters a restaurant, he will first take off his shoes, then take a seat, and then will be asked to order. So, seeing someone entering a restaurant who looks like a guest (and not a waiter) allows us to expect a specific conventionally-regulated sequence of behaviour. If one has a liberal notion of folk-psychological theory, then we may add such behavioural conventions into that theory. But even then the question remains whether our understanding of others always relies on knowledge organized as a folk-psychological theory. A counterexample can be proposed by reference to cases of basic intuitive understanding: e.g., the still-face reaction by the caregiver, instead of a typical smiling facial expression and gestural response, makes the baby start to cry (as we saw above). There is thus an intuitive recognition of basic emotions like fear, anger, happiness, or sadness. This may rely on inborn emotion recognition mechanisms, or mechanisms learned very early, which may be evolutionarily anchored, since recognizing such basic emotions is essential for survival (Griffiths 1997; Panksepp 2005). There are two ways in which the counterexample might be blocked: (i) It could be maintained that some folk-psychological theories are inborn (Baron-Cohen 1995) and that intuitive understanding such as face-based recognition of emotion already involves a theoretical package. The problem with this line of reasoning is that the notion of theory, stretched that far, starts to look very implausible. A theory is constituted by a minimal package of systematically interconnected beliefs; and even if a belief is understood in a liberal way such that it does not presuppose linguistic representations, it remains highly questionable whether basic cases of faced-based recognition can be characterized as a systematically interconnected set of beliefs. The standard descriptions of face-based recognition of emotion (e.g., Goldman 2006) on a neural level highlight the relevance of mirror neuron mechanisms and characterize the underlying mechanism as a rather basic and partially independent pattern-recognition process, and thus as not forming a theory. A defect in recognizing disgust does not automatically lead to a defect in recognizing other basic emotions like happiness or sadness (Calder et al. 2000). (ii) A more promising move is to claim that the folk-psychological theory is learned (Gopnik 1993). This view is compatible with some basic processes of understanding which do not yet form a theory, but are developed into one as they are integrated step by step into a systematically-organized body of knowledge. This is a plausible and to some extent empirically grounded view (Gopnik & Meltzoff 1997; Newen & Vogeley 2003).[5] One shortcoming of this view, however, is that its proponents tend to appeal to examples that have a strong focus on general folk-psychological rules, such as: “All humans need to drink, thus if someone picks up a glass in the kitchen, he intends to pour into it some liquid to drink”. This neglects a very important phenomenon, namely that we mostly interact not with complete strangers but with persons we know at least partly and often very well. For example, if Michael observes his son in the kitchen grasping a glass he does not appeal to the folk-psychological rule at all, since he knows that his son—despite his education—still only drinks from a bottle when at home, and that if he takes up a glass it is just because he wants to use it for practising magic tricks. This indicates that all the theories canvassed thus far have a blind spot: so far it seems simply to have been neglected that we rely extensively on knowledge of properties of individuals, which is organized as belonging to one specific individual (the son, the partner etc.) or to a group (students, managers, etc.). The general worry concerning the organization of this knowledge, according to TT, can also be expressed as follows: How are we able to apply a general theory of typically human features in a specific social situation? If we want to integrate our prior background knowledge of persons as individuals or as belonging to a group, e.g., to a profession, then we can characterize the organization of this knowledge as person models. Person models of individuals and groups are by far the most important source of understanding others, I will argue, and since they involve specific knowledge, they are the natural candidate for enabling adequate deployment of more general knowledge of human psychology in concrete everyday situations. It remains to be discussed, then, whether person models have the status of a folk-psychological theory or not. To adumbrate my line of argument: no doubt some elaborate person models are systematically-interconnected sets of beliefs, but not all of them have to be, because some person models only involve very sparse and basic properties that are not highly interconnected.
As we saw earlier, one recent account of understanding others, proposed by Dan Hutto (2008), holds that understanding others mainly relies on telling stories and using this knowledge to understand individuals. The core claim of his NPH (Narrative Practice Hypothesis) is
[…] that direct encounters with stories about persons who act for reasons—those supplied in interactive contexts by responsive caregivers—is the normal route through which children become familiar with both (1) the basic structure of folk psychology and (2) the norm-governed possibilities for wielding it in practice, thus learning both how and when to use it. (Hutto 2008, preface, p. x)
One focus of his theory is not so much how the prior background knowledge of others is organized, but rather how children are able to acquire it. His developmental claim is that the central route for learning relevant background knowledge is listening to stories about persons. I grant that this is an important additional route of epistemic access to relevant knowledge about others; but it is already an advanced method, not normally used before the second year of life. Furthermore, in such cases the focus is not epistemic access to knowledge used to understand the other in the situation (i.e., when listening to the storyteller), but rather to gain new background knowledge with an eye to future understanding of others. In a follow-up paper written together with Gallagher (Gallagher & Hutto 2008), Hutto and Gallagher enrich their views about epistemic access through appeal to direct perception and interaction (see above) in addition to learning by narratives. It is important to note the difference between epistemic access to information that allows me to understand the other in the actual situation (see section 3) and epistemic access to background knowledge relevant for future usage. Thus, by granting that narratives are an additional instrument for learning about important properties of persons, I can enrich my multiplicity claim as characterized above. In integrating this idea, one should also generalize it: we not only learn important background information that helps us to understand others by listening to stories told by a caregiver, but also by reading stories, especially novels.
Let us now briefly discuss the NPH considered as a claim about the organization of our background knowledge. If I have elaborate and explicit knowledge of a person, I may have acquired it by listening to or reading a story, and I may tell a story if someone asks me about this person. But, as the interaction view highlights, sometimes my knowledge may be anchored in the interaction, yet still be non-linguistically represented, and only activated in similar interactive situations. Our rich non-linguistic knowledge about other human beings, which we acquire when directly perceiving them (tone of voice, what they look like) or interacting with them, or when realizing a joint action, etc., are often not linguistically coded and thus not memorized as a linguistic story. If we widen the notion of a story such that it includes any sequence of memorized events, we lose track of any interesting notion of “story”. In fact, we are instead going in the direction that I propose, i.e., that we organize our prior knowledge about others through unifying it in person models. Some such models may include properties of a person that are connected as or with stories, but the core of a person model is a unity of features of a person that are grouped together as belonging to one individual or to a group, where the features may be as primitive as the tone of voice of a person, and have no connection to any story, even in a wide sense.
Although our prior knowledge about others is the main component of our understanding of others in a specific situation, most of the theories canvassed above did not present any clear view on how this knowledge is organized.[6] We found only two suggestions: relevant prior knowledge is organized either as a folk-psychological theory or as a narrative. Neither proposal covers all relevant cases: neither accounts for the innate or very-early-learned (nontheoretical) basic background knowledge that enables us to effect smooth interaction and allows us to rely on a basic intuitive understanding of others. And, furthermore, as I argue in the following, there is an alternative view, the person model theory, which is able to integrate the plausible aspects of these two suggestions, and additionally allows us to explain a variety of phenomena that the alternative views did not or cannot take into account—especially the integration of features of other human beings that allow us to realize an intuitive understanding of them.