Research Overview
Work in the Cognition and Development lab most broadly is concerned with questions of how children and adults construct causal interpretations of the world around them and how those interpretations compare to other ways of tracking information. This orientation leads naturally to questions of how adults and children cognitively reduce the enormous causal complexity of the world to a more manageable forms and what distortions of that information occur as it is necessarily simplified. Just as image processing software must engage in compressions of information to handle otherwise overwhelming storage and computation requirements, so also must humans construct coarse causal gists of a much more complex reality. We are interested in studying the nature of those gists and what they do and do not capture about real world causal relations. A related question concerns how we deal with gaps in our knowledge. To what extent do we recognize our own gaps and how do we construe and access knowledge in other minds when we need fill in those gaps? These questions are critically informed by taking developmental perspective that explores how young children, and even infants, cognitively grasp the many levels and types of causal structure inherent in the world. We also extensively consider the interplay between domain specific and domain general processes and structures, such as between folkbiology and folkphysics. More broadly all these issues bear on the topic of folkscience, that is, how people naturally construct explanatory accounts of various domains of phenomena.
This general orientation has led to a wide variety of specific research projects, including:
- Levels of Causal Tracking
- Grasping the Terrain of Knowledge
- Deference and Deliberation
- Causal Explanatory Styles
- Evaluating Knowledge and Understanding
- Encoding of Event Structures
- Folkbiology
- Physical and non-physical Causality
- Causal and non-causal
- Overimitation and Causal Interpretation
- Intentionality and Order
There are several different levels at which we can track causal structure of the world, some of which we share with other species, others of which appear to be uniquely human. One of the simplest levels involves causal relevance, simply noting what kinds of properties are likely to be causally important in a domain without knowing their specific effects. This may be as simple as noting that color tends to be more causally central to most natural kinds than it is to most artifacts. A second level involves causal powers, knowing that particular properties have specific causal consequences, such that intentional agents are capable of creating order out of disorder. One may know these powers without having any sense of mechanism. A third level involves knowledge of functions, such as that a computer mouse is meant as a means of moving a cursor on a screen. The most enriched level is having sense of mechanisms analogous to having detailed blueprint like mental models in one's head. We have explored causal thinking at each of these levels in both children and adults. [Back to top]
Grasping the Terrain of Knowledge
Given that we cannot possibly know everything in every domain, the question arises as to how we learn how to navigate the terrain of knowledge, that is, how we learn who knows what in our culture. Several studies have been conducted looking at this question in both children and adults and now in cross cultural contexts as well. [Back to top]
Because much of what learn is gathered through what others tell us, the problem arises as to when we should and should not defer to what explanations and insights offered by others. In other words, what principles do we use to decide whether another person is a legitimate expert on a particular question? How do such principles vary across different domains? We have conducted several studies in this area and have others underway. [Back to top]
Causal explanations can have several different structural forms. For example, one can offer physical mechanistic and teleological explanations of many phenomena in biololgy. A series of studies has examined how children and adults map different explanatory types onto different domains and has also asked under what conditions they seek out these different kinds of explanations. Another line of work asks about the conditions under which we find explanations that refer to vivid physical substrates compelling, such as images of the brain as explanations for psychological effects. [Back to top]
Evaluating Knowledge and Understanding
How well do we understand the limits of our own knowledge and of knowledge in others? We have documented a powerful illusion of explanatory depth in which people think they understand things much better than they do. Further studies have examined how bias develops and ongoing studies are asking how this bias influences other judgments about knowledge in the self and others. Other lines of research in this area ask about the conditions under which one should be cynical of what another person says. Other studies ask about the criteria we use to assess whether another is likely to know what they are talking about. Still others ask about how children and adults are able to evaluate the coherence and circularity of explanations. Another line of work asks about the heuristics we use to evaluate the plausibility or legitimacy of various alleged areas of expertise. Finally, a series of studies has explored biases concerning the perceived relative complexity of judgments in across broad domains such as the natural sciences. [Back to top]
When people observe a causal event, there are several levels at which they might encode that event. We have conducted a series of studies asking about the most natural levels of encoding the functions of tool-based events and how those levels might change with development. This is part of a broader line of work asking about levels of causal analysis. Our findings so far indicate that younger children actually encode at a higher level than older children, contrary to more classic ideas of concrete to abstract shift sin development. [Back to top]
Our lab has had a longstanding interest in folkbiology, having conducted studies on intuitions about species, disease, inheritance of properties, and aspects of physiology. We continue to conduct several lines of work in this broad area as part of an approach that asks how understandings in broad domains such as folkbiology and folkpsychology might differ form each other both in adults and children of all ages. [Back to top]
Physical and non-Physical causality
Not all causality is understood in physical terms. We often talk about psychological causality, such when beliefs cause desires. Even when psychological causality may ultimately supervene on physical substrates, there are reasons to believe that people think about such kinds of causation as non-physical. We are conducting studies examining how such cases of physical and non-physical causation are understood and how they are seen as interacting. [Back to top]
In addition to relational patterns that we encode as causal, there are many other patterns, such as spatial, mathematical relations, or moral relations that we encode as non-causal. We are interested in how relations in each of these domains compare and contrast with respect to intuitions about centrality of features, areas of expertise, coherence, and other factors. [Back to top]
An extensive line of studies asks about how we understand the causal factors that are responsible for departures from randomness in the natural and human created world. We have found that even infants expect that intentional agents are more likely causes of ordered arrays. Several lines of work are pursuing these general ideas in different directions. [Back to top]
Overimitation and Causal Interpretation
Human children, in contrast to other higher primates, will often engage in "overimitation", namely a tendency to imitate unnecessary steps intentionally engaged in by a model to get a goal. A series of studies is showing that this overimitation is far more than simple conformity and instead seems to be a means for inferring causal relevance of properties of systems that are not yet fully understood. [Back to top]