The Paradox of False Belief Understanding

The Role of Cognitive and Situational Factors for the Development of Social Cognition

Julia Wolf

• Publisher: De Gruyter
• Year: 2021
• ISBN:
  9783110758658
• Edition: 1
• Topic: Philosophy
• Subtopic: Epistemology in Philosophy

Our ability to understand others is one of the most central parts of human life, but explaining how this ability develops remains a controversial issue, exercising psychologists and philosophers alike. Within this literature the Paradox of False Belief Understanding remains one of the main open challenges. Based on an up to date overview of the empirical and theoretical literature, this book highlights the significance of this paradox for our understanding of the development of social cognition and provides a new explanation of it in the form of the Situational Mental File Account. Central features of the account are, firstly, identitfying three distinct stages in the development of belief understanding and, secondly, elaborating the role of both cognitive and situational factors as well as their interaction in the development of belief understanding. This account is also applied to the related phenomenon of pretend play, demonstrating the potential for a wider application of the account. This account generates both new empirical predications and a framework for further theoretical work, thereby providing a fruitful ground for further interdisciplinary research in this area.

1 The Significance of the False Belief Task

The main aim of this book is to develop an empirically adequate account of the paradox of false belief understanding. This paradox refers to the finding that children fail explicit versions of the false belief task (FBT) till age 4, but pass implicit versions of the same task much earlier. The aim in this chapter is to set the groundwork for the discussion of the paradox of false belief understanding from Chapter 2 onwards by showing that the paradigm giving rise to the paradox remains an important tool for testing the development of social cognition. In other words, in this first chapter, I will be concerned with the question why we should be interested in the findings from the FBT, thereby providing a background motivation for the discussion of the paradox of false belief understanding to follow in the next chapters.

The FBT has held a central position in the literature on the development of social cognition for more than 30 years. It has become the staple test of whether children have an understanding of the mental states of others or not. In recent years, however, there has been mounting criticism against the FBT, arguing that the FBT is an overly artificial measure that is unable to tell us anything about how social cognition really works and develops. In this chapter, I will argue that although there are reasons to be apprehensive of the dominance of the FBT in the study of the development of social cognition in children, the FBT remains a relevant and useful task that continues to give rise to interesting new insights concerning the development of cognition.

In this chapter, I will first review some of the history of the FBT and clarify what exactly the FBT is supposed to test. This will give me the opportunity to introduce some of the central terms in this debate: mindreading and social cognition, and how these relate to each other. Based on this, I will then address some of the criticisms against the FBT, the main criticism being that the FBT is an inadequate test of social cognition that does not reflect the realities of everyday social cognition. Concerning this I will argue for the importance of mindreading, and that the FBT, as a measure of mindreading, does test an important aspect of social cognition. Having thus defended that FBT paradigm, I will then turn to focus on the paradox of false belief understanding from Chapter 2 onwards.

1.1 Social Cognition, Mindreading and the False Belief Task

Broadly understood, social cognition refers to our ability to understand other people. It is the basis of our ability to interact with others – both in a cooperative and a competitive manner, and our ability to predict and explain the behaviour of others. Traditionally, it was thought that social cognition is fundamentally based on mindreading, i. e. the ability to attribute mental states such as thoughts, beliefs, desires, or emotions to others. It is only in recent years that the idea that social cognition might not be based solely on mindreading has gained in popularity. The debate surrounding mindreading can be traced back to the seminal Premack and Woodruff (1978) paper, “Does the Chimpanzee have a Theory of Mind?” Premack and Woodruff were interested in the question whether animals (in this case, chimpanzees) were able to attribute mental states to others and make use of this to predict the behaviour of others. They famously coined the term ‘Theory of Mind’ (ToM) for this ability of an individual to “imput[e] mental states to himself and others” (Premack and Woodruff, 1978, 515). In the literature, Theory of Mind is used interchangeably with the terms ‘mindreading’ or ‘mentalizing’. I will predominantly use ‘mindreading’ to refer to this ability to attribute mental states to oneself and others,¹ in keeping with my focus on mindreading as an ability rather than a matter of concept possession (see Introduction).

It is critical to note that the term ‘mindreading’ has two meanings in the debate. Understood widely, mindreading refers to the ability to attribute any mental states to another person, including mental states such as thoughts, desires, emotions, and beliefs. This is the way I defined it above. Understood narrowly, mindreading refers to the ability to attribute beliefs to another person. The attribution of beliefs (narrow mindreading) is a subtype of the attribution of mental states (wide mindreading). Someone who is not able to attribute beliefs might still view others as minded beings and be able to attribute mental states other than beliefs. Someone who lacks the ability of mindreading is not able to attribute beliefs. For the sake of clarity, I will reserve the term ‘mindreading’ for mindreading widely understood unless otherwise specified. When talking about mindreading narrowly understood, I will refer specifically to belief attribution.

What is the role of the FBT? In order to understand the significance of the FBT it is important to understand how and why it came about. In the previously mentioned paper, Premack and Woodruff (1978) carried out an experiment to test whether chimpanzees were able to infer the intention of another individual and predict their behaviour accordingly. Their specific methodology is not important for our current purposes. What matters is that in response to this study the FBT was independently suggested by Dennett (1978), Bennett (1978), and Harman (1978) as a means of avoiding some of the criticisms that had been raised against Premack and Woodruff’s experiment, and thereby provide conclusive evidence of the ability to attribute mental states to others. The crucial element of the FBT is that it exploits the fact that beliefs can be false. Therefore, it allows for introducing a distinction between the other’s belief and reality (or one’s own knowledge of reality). Therefore, for example, the chimpanzee sees Bob place his banana in the crate, but the banana is then removed without Bob’s knowledge so that Bob believes the banana is in the crate – which the chimpanzee knows to be false. If the chimpanzee is predicting behaviour on the basis of beliefs, the chimpanzee should expect Bob to go looking for his banana in the crate (even though the banana is not actually in the crate).

It should be noted that in exploiting this distinction between belief and reality, the FBT was intended to provide a particularly stringent test of mindreading by means of testing belief attribution. Failure to pass the FBT is insufficient to show that someone is able to engage in mindreading widely construed.² If a child fails the FBT they might nonetheless be able to attribute other mental states such as desires to another person and therefore be capable of at least some forms of mindreading widely construed, even if they cannot engage in mindreading narrowly construed (i. e. belief attribution). Moreover, failing the FBT alone is also insufficient to show that someone is incapable of belief attribution (i. e. mindreading narrowly construed). The FBT is a demanding task and there might be a range of reasons why someone capable of mindreading could fail. So, while passing the FBT provides evidence of both wide and narrow mindreading (in virtue of belief attribution being a subtype of mindreading), failing the FBT is not sufficient to show an inability to mindread understood either widely or narrowly.

While initially Dennett (1978) made clear that the FBT was to be understood as a sufficient test of mindreading and not a necessary one, the prevalent view that resulted from the finding that children before the age of 4 fail (at least explicit versions) of the FBT is that children are actually incapable of attributing beliefs to others prior to this point (Wellman et al., 2001).³ This view that failing the FBT provided evidence of being incapable of attributing beliefs is something that developed in the literature following a range of variations and modifications of the task (Perner, 1991). This seemed to show that this age 4 boundary was a robust result, persisting across variations. It was this robust finding, which led to the view that there is a genuine change in children’s ability to attribute beliefs to others at age 4.

As noted above, the FBT is specifically concerned with mindreading narrowly construed. That is to say, it aims to test the ability to attribute beliefs to others. Mindreading, broadly construed, concerns a wide range of mental states beyond beliefs. Although the FBT may also require an understanding of some other mental states (especially desires), the focus is very much on attributing beliefs as the necessary ability for passing the FBT. While this exclusive focus on beliefs has been subject to criticism (see, for example, Apperly, 2011), it is worth noting that there are at least two reasons for this fixation on beliefs. Firstly, as noted above, by making use of beliefs it was possible to make use of a distinction between beliefs and reality, making the FBT such a powerful test of mindreading broadly construed. The idea was that if someone passed the FBT we could be sure that they were attributing mental states to someone. Tests making use of other mental states, such as in the original experiment by Premack and Woodruff (1978) would not provide as conclusive evidence as the FBT and might be explainable in non-mentalistic terms.⁴ Secondly, beliefs are often thought to be particularly important mental states, which are crucial for predicting and explaining the behaviour of others (Doherty, 2008). This is inspired by the influential view of belief-desire pairs as both the cause and explanation of action. While the origins of this view can be traced back to the ideas of Hume (1978), in current discussions this idea is most famously associated with the work of Davidson (1963) and his highly influential arguments that belief and desire pairs (so called pro-attitudes) cause our behaviour. Therefore, for example, it is the desire for a cold glass of water, paired with my belief that there is cold water in the fridge that causes me to go to the fridge to get some water. It should be noted that the question of how actions are actually caused and how we predict and explain the actions of others are at best loosely related. If our actions are indeed caused by beliefs and desires, then this could provide reason to think that a social cognition based on the attribution of beliefs and desires to others could be effective, but not necessarily so. Similarly, even if actions are not caused by belief-desire pairs we might still make sense of others by attributing beliefs to them. As my aim is to look at the development of understanding of others’ actions, not the actual cause of actions, I will not discuss the belief-desire framework further here. Although this framework has substantially fed the discussion surrounding the FBT, the rationale of the FBT does not depend on it and I will neither defend nor assume this framework for the purpose of this discussion.

The basic idea behind the FBT can be summarised as follows: the phenomenon we are ultimately interested in is social cognition, which refers broadly to our ability to understand and interact with other people. This ability is thought to be fundamentally based on mindreading, i. e. the ability to attribute mental states to others. Belief attribution is a critical subcomponent of mindreading. The best way to test mindreading is through the FBT, which tests mindreading via testing belief attribution. Motivating the FBT, we therefore have the following picture (Figure 1.1):

Figure 1.1: The relation of social cognition, mindreading and the FBT. Social cognition is based on (or occurs via) mindreading. Belief attribution is a sub-type of mindreading. The FBT is the central test of mindreading in virtue of being a stringent test of belief attribution.

It was in light of this, with the FBT as the ‘acid test’ of mindreading that the FBT rose to prominence and has remained the staple test of mindreading in the developmental literature, in particular following the initial and robust findings of the shift in performance at age 4 (Baron-Cohen et al., 1985; Wimmer and Perner, 1983): not only was the FBT a neat means of testing social cognition, but it also seemed to be picking up an important developmental step.

It should be noted that this simplified picture can be seen as the target of many of the critics of mindreading and the FBT. This is not to say that this view is actually held by adherents of the mindreading account. These views are usually more nuanced and do not claim that mindreading narrowly understood is the sole basis of our social cognition (Apperly, 2011). Rather the claim is that belief attribution is a particularly central social ability (e. g. Doherty, 2008) or that it represents a significant organisational change to the structure of our social reasoning, for example a change to an understanding of representational mental states (Wellman et al., 2001).

1.2 Criticisms of the False Belief Task

Based on the above picture, there are two places where one might object to the FBT (see Figure 1.2). Firstly, it might be objected that the FBT is a poor test of mindreading, either widely or narrowly construed. Secondly, it can be objected that social cognition is not actually based on mindreading. Insofar as the FBT does test mindreading, then it does not test any interesting or important aspect of social cognition...

Table of contents

• Title Page
• Copyright
• Contents
• Acknowledgements
• List of Figures and Tables
• List of Acronyms
• Introduction
• 1 The Significance of the False Belief Task
• 2 The False Belief Task – A Developmental Sketch of the Paradox of False Belief Understanding
• 3 Theories of Mindreading – The Role of Cognitive and Situational Factors in Belief Understanding
• 4 Situational Mental File Account
• 5 Finding a Point of View: Vicarious Mental Files and Perspective Taking
• 6 Pretend Play and Its Importance for Mindreading
• Conclusions and Outlook
• Subject Index
• Author Index