QnAs with Elizabeth Spelke

If humans have a defensible claim of supremacy in the animal kingdom, it might lie in our superior cognitive abilities, which have endowed our species with unparalleled powers to probe our world. Those fundamental abilities—some shared with other species—help us make sense of space and number, use language, and forge social partnerships, to name a few skills. To gain sophisticated insight into the origin and development of such basic skills, Harvard University cognitive psychologist Elizabeth Spelke has long focused on the roots of the traits, charting their dynamics among infants and toddlers. Set in a halcyon world of cooing babies and prancing preschoolers, Spelke’s decades-long efforts strike at the heart of the human condition, which preoccupied preeminent philosophers of the past and continues to puzzle psychologists today. For her insights into human cognition, Spelke has earned a well-merited reputation among her peers, not to mention a raft of accolades. Most recently, Spelke was named the inaugural recipient of the 2014 National Academy of Sciences prize in psychological and cognitive sciences. In honor of the recognition, PNAS spoke to Spelke about her influential work.

Spelke reviews an infant study session with former graduate student Lindsey Powell. Image credit, Rachel Katz.

PNAS: Your studies on infants are a window into the origins of human cognitive abilities, such as our sense of space and number and our social interactions. Why are infants so well suited to these explorations?

Spelke: Infants have two features that make them suitable for the study of mature human cognition. First, infants learn about space, number, geometry, objects, and other aspects of social cognition from the beginning of life. To do this, they must come into the world equipped with cognitive mechanisms for detecting objects, sets with numerical magnitudes, geometrical relationships between places, and potential social partners. Secondly, when we compare infants’ and adults’ understanding of these elements, we see a huge difference. So we can use infants to ask how our knowledge of the world grows.

PNAS: In a 2014 Cognition paper (1) you reported that six- to eight-year-old children who exercised their approximate number sense (ANS) were better at solving symbolic arithmetic problems than those who did not. What is the ANS?

Spelke: The ANS is an abstract ability that we share with other animals and that has been studied for decades. It is a sense that allows us to tell without counting, performing serial enumeration, or establishing one-to-one correspondence the approximate cardinal value of sets of objects or events. Babies can distinguish on the basis of number between events in which a puppet jumps 8 or 16 times, for example. Newborn infants can also distinguish with approximate accuracy between sequences of 4 or 12 syllables, and relate an auditory sequence of four syllables to a visual display of four objects. This sense is approximate at all ages but, like most of our perceptual abilities, its precision increases with age.

PNAS: What are the practical implications of the finding?

Spelke: For these findings to have practical implications, we need to demonstrate that the effects are lasting, that the effects translate into gains in a math classroom over time, and that the cost-benefit analysis from the perspective of time spent exercising the ANS favors the approach over alternative approaches. For now, we don’t have evidence to support any of those points, but the research is underway. Based on our preliminary work, I can say that kids enjoy the exercise so, at the very least, the approach might turn out to be a way of making math fun for kids; it’s too soon to say whether it will be academically significant.

PNAS: Last fall, you reported in PNAS (2) that four-year-old children use abstract geometry in recognizing spatial symbols and map reading, but how do the results help establish the link between these early spatial abilities with the later development of abstractions of Euclidean geometry?

Spelke: Intuitions about formal geometry develop much later than symbolic arithmetic in children. We previously showed that although some basic intuitions about Euclidean geometry do not depend on formal instruction and are culturally universal among adults—adults in Cambridge, MA, Paris, France, and a remote Amazonian group all have comparable Euclidean geometric intuitions—they develop slowly in children. It is not until about age 12 or 13 that children’s intuitions of Euclidean geometry approximate those of adults.

PNAS: If the developmental course for geometric intuitions is linear, why do some of us lose those abilities as adults; many people are easily fazed by maps and have poor sense of orientation?

Spelke: It is true that adults with developmental or neurological deficits may have impairments in the fundamental spatial abilities of the kind we explore in infants. However, for the garden-variety variability in navigation skills among normal adults, I think what it means to be a poor navigator is that your representations of the geometry of a layout are imprecise. I don’t think we lose these abilities as we get older, but in our technology-driven societies (where GPS and Siri act as guides) even adults with well-developed geometric abilities simply don’t bother to use these skills. Spatial cognition has been a crucial trait since human prehistory, so I am a little concerned that we might be becoming overly dependent on our devices.

PNAS: Previously, you reported in PNAS (3) that by the age of seven months preverbal infants begin to expect members of the same social group to act alike, even before they develop language or can imitate others. This expectation may lay the foundation for social mimicry and stereotyping reported in toddlers and preschoolers, but does it suggest an early-developing “us-versus-them” tendency for social categorization that can be a potential cause of conflict among adults?

Spelke: I hope not. The paper you mention (3) and our previous PNAS paper on infants’ social affinity toward people who share an accent with them (4) together suggest that infants might use such cues to identify with their own group, but I don’t think that they use the cues for active social exclusion. Nor do I think that this early tendency to associate with people who show certain social cues predisposes infants to xenophobia. I say this based on decades of observational evidence, not experimental findings. Besides, infants are born with almost no capacity for effective, independent action; they can’t run away from strangers, for example. So, my guess is that in the early years infants have an incentive for positive social association toward everyone.