Search Results

This chapter offers a selection of theories of causal learning. Some of the theories come out of psychology, while others come out of rational analyses of causal learning. All tend to focus on how people use correlations — information about which events go together — to figure out what is causing what. A number of other, supporting pieces of information about what causes what — a number of other cues to causal structure — that people are responsive to are described. The costs and benefits of allowing learners to actively intervene on the system they are learning are discussed.

According to human perception, the world is full of causal systems composed of autonomous mechanisms that generate events as effects of other events. This chapter attempts to make this idea more precise and therefore more clear by making it more formal. It introduces the causal model framework as an abstract language for representing causal systems. It discusses the three parts of a causal model: the causal system in the world (i.e., the system being represented); the probability distribution that describes how likely events are to happen and how likely they are to occur with other events — how certain we can be about each event and combination of events; and a graph that depicts the causal relations in the system.

This chapter discusses what causes are. It argues that causal relations are not the only kind of invariance useful for representing the world. There are various kinds of mathematical representations, as well as logical and probabilistic representations. But noncausal forms of invariance are less useful than causality for describing relations among events because they don't naturally describe the processes that generate those events and because, therefore, they fail to support key forms of counterfactual inference as directly as causal models do. In short, only causal models represent the invariance that tells us what the effects of our and others' actions would be. As a result, people seem to be particularly adept at representing and reasoning with causal structure.

This chapter focuses on how we induce properties of the world. Some psychologists have been coming back to the view that induction is mediated by causal models — causal models that are often generated online through the application of causal principles, abstract causal relations that have general applicability. Such causal models help to explain how people make inductive inferences when the inference can be conceived as a causal effect, as in the ‘bananas have it; therefore, monkeys have it’ example. In other cases, inference involves analogy: a predicate is applied to one category because it is known to apply to an analogous category, as in the ‘tigers do it; therefore, hawks do it’ example. In such cases, the analogy seems to be between causal structures. Finally, causal models give a psychologically plausible way to think about why people sometimes show sensitivity to statistical information. Instead of assuming that people calculate statistics like measures of the variability of a property, the requisite information can be interpreted as a property's centrality in a causal model.

This chapter focuses on why cause matters. It argues that causal relations hold across space, time, and individuals; therefore, the logic of causality is the best guide to prediction, explanation, and action. Not only is it the best guide around; it is the guide that people use. People are designed to learn and to reason with causal models.

Takes the objections that were discussed in the previous chapter as a starting‐point for arguing in favour of the dynamic view of the world that is defended in the book.

It argues that satisfactory truth conditions for counterfactuals can only be given if the world is a dynamic one. Further, an argument from causation can be put forward in favour of the dynamic view: The concept of a causal relation is indirectly defined by constraints on causal laws. But there can be causal laws only if the past and the present are real, while the future is not.

What is it to represent a relation as causal? The question has a wide interdisciplinary significance. The question itself has been at the centre of philosophical discussion from Hume onward; it is one that is pivotal in human developmental psychology; disputes in ethology about animal tool use cannot be resolved without addressing the question. Three issues about the representation of causality whose significance crosses the usual disciplinary boundaries are: What would be good evidence that some creature is representing a relation as causal? What would make it good evidence? In particular, how is its status as good evidence grounded in a constitutive account of what it is to represent a relation as causal? What is the relation between the constitutive account of what it is to represent a relation as causal, and what makes a relation one of causality? What is the right model of the relation between the content of the representation and what is represented? This is evidently an area in which issues of intentional content, the conditions for attribution of mental states, and the nature of the content's reference bear upon one another. The issues develop very quickly into more general ones about the relations between intentional content and metaphysics. That is part of their interest. But rather than start with such grander themes, this chapter works its way up to them from a starting point often adopted in the ethological literature. It begins with the question of the relation between tool use and the representation of causation.

This chapter returns to the metaphysics of the causal relation proper. It examines generalist theories of causation, theories that seek to reduce that relation between states of affairs tokens, to some law-based relation between states of affairs types. Particular attention is paid to the one such generalist theory that has had a large influence in legal theory. This is the nomic sufficiency theory of John Stuart Mill and his intellectual descendents, Hart, Honore, Mackie, and Wright. This is the view that reduces singular causal relations to instantiated scientific laws, and that reduces scientific laws to sets of minimally sufficient conditions (where each such condition is necessary to the sufficiency of the set). The thesis of the chapter is that generalist theories founder on the same seven arguments that doom the counterfactual theory as well.

This chapter argues that infants' experience of their own actions and the consequences that these actions have on the world play an important role in their developing understanding of causal relations. Recent philosophical theories of causation take an interventionist perspective on causality: If manipulations on one factor (interventions) are associated with a change in a second factor, then the first causes the second. In addition, empirical evidence suggests that both adults and young children readily learn causal structure from enacting and observing interventions. This chapter presents evidence that infants' developing ability to act on the world is intimately linked to their causal understanding. Infants' interventions may enable them to evaluate causal hypotheses and to detect the causal structure of various events in the world.

This chapter offers a quick tour of book 1, part 3 of the Treatise, a portion of the text where Hume sets forth what he takes to be some of his most important discoveries in his pursuit of a science of human nature. These include an examination of necessary connections, causal relations, and probability. The underlying theme of this part of the Treatise is to locate the fundamental source of the fixation of belief in the natural operations of the imagination, not in our intellectual faculties. This is the explicit role of what amounts to a skeptical argument concerning induction. It is also the reason why he treats vivacity as the determinant of belief. Here Hume's naturalism is in full flower.

This concluding chapter sums up the key findings of this study about intentional content. The findings suggests that there are not intentional norms in nature because there are no biological norms in nature and that claims about function and content are claims about objective causal relations between natural items. The results also indicate that determinacy objection against a teleological approach to intentional explanation can be answered using the proposed account of the theory of functions.

This chapter begins with a discussion of the goals of epidemiologic research, which is the quantification of the causal relation between exposure and disease. It then discusses the measurement of causal relations between exposure and disease, inferences from epidemiologic research, descriptive goals and causal inference, efficacy of breast cancer screening, and contribution of epidemiology to policy decisions.

This chapter attempts to show that claims that chimpanzees (and other animals) reason about the world in human-like ways are unfounded. In particular, it argues that only humans rely upon ‘higher order’ causal relations such as force, weight, or gravity, and logical inferential processes such as diagnostic causal reasoning when making sense out of the world. It begins by analysing one of the most basic problems in this field: an ape wielding a stick to obtain an out-of-reach banana.

The two conflicting traditions of causality in psychology are both represented in the book: the arbitrary tradition of Hume in which causality is learned, and the natural tradition of Michotte, in which causality is directly perceived. Delay evidently affects arbitrary and natural causality differently. Although not affecting judgement in the natural case, in an arbitrary case when humans pressed a lever to produce a light, delay weakened both instrumental responding and judgement of causal relationship. While the judgement of natural causal relations remains untouched, delay weakens the judgements of arbitrary causal relations.

Rae Langton's book Kantian Humility offers a provocative interpretation of one of Kant's central theses and one of the arguments for it. The thesis is that we have no knowledge of things in themselves, which Langton interprets as meaning we have no knowledge of the intrinsic properties of things. We have knowledge of mind-independent things, not just of appearances, but what we know about these things is only how they are related to other things and to us, not what they are like in nonrelational respects or ‘in themselves'. This is the doctrine Langton calls Humility. Kant's argument for the doctrine is based on the premise that human sensory knowledge is Receptive — that we have knowledge only of those features of things that have some causal impact on our faculty of sensibility. But how does Receptivity imply Humility? Strawson complained that the link between the two is ‘a fundamental and unargued complex premise of the Critique'. Langton undertakes to identify the Kantian assumption that must be added to Receptivity to get Humility. However, her own reconstruction of Kant's argument still leaves a significant gap. This chapter tries to do for Langton's Kant what Langton did for Strawson's Kant: to identify and assess an unstated premise that is needed to make the argument go through. In this reconstruction, the missing premise is that causal relations hold (when they hold at all) of necessity. When contemporary philosophers maintain this anti-Humean thesis, they generally do so as a consequence of a view about property individuation sometimes known as causal structuralism. The chapter examines the credentials of causal structuralism.

This chapter continues the analysis of divine persons begun in the previous chapter in some small Trinitarian treatises. Two new main elements are added to the description of a divine person: persons are enumerable entities, unlike their nature which cannot be counted; persons can be distinguished by relations of origin or causal relations: Gregory distinguishes the person who is “the cause” of the other two (i.e., the Father) from the person who is “from the cause” (i.e., the Son), and from that who is “by that which is directly from the first” (i.e., the Holy Spirit);

This chapter shows that the perception of others' actions and production of self-action are mapped onto commensurate representations starting from birth. This allows infants not only to learn interventions through their own manipulations but also to multiply greatly their learning opportunities by observing the manipulations of others and profiting from them. Infants imitate but do not blindly copy everything they see. First, they make creative errors. Second, they skip over the literal behavior they see and choose to duplicate inferred interventions: what the adult meant to do, not what the adult did do. Third, when causal relations are difficult, as in the rake case for younger infants, observation alone does not seem to guarantee success; older infants glean more from the modeling than do younger ones.