I am broadly interested in the human ability to reason about others, learn from others, and inform others in communicative contexts. How do we construct rich, abstract theories about how the world works from our everyday experiences that often involve other people, and how do we communicate what we know to others? My research brings together various approaches -- primarily developmental, computational, and neuroimaging methods -- aiming to provide a unified description of the cognitive and neural mechanisms that underlie the representations and inferential processes that allow us to learn about the world, and to communicate what we know.
Symbolic Systems Program
Ph.D., MIT, Cognitive Science (2012)
- Independent Studies (4)
Prior Year Courses
- Cognitive Development
PSYCH 141 (Aut)
- Social Cognition and Learning in Early Childhood
PSYCH 175 (Win)
- Cognitive Development
PSYCH 141 (Spr)
- Graduate Seminar on Theory of Mind
PSYCH 285 (Aut)
- Social Cognition and Learning in Early Childhood
PSYCH 175 (Win)
- Cognitive Development
Integrating Incomplete Information With Imperfect Advice.
Topics in cognitive science
When our own knowledge is limited, we often turn to others for information. However, social learning does not guarantee accurate learning or better decisions: Other people's knowledge can be as limited as our own, and their advice is not always helpful. This study examines how human learners put two "imperfect" heads together to make utility-maximizing decisions. Participants played a card game where they chose to "stay" with a card of known value or "switch" to an unknown card, given an advisor's advice to stay or switch. Participants used advice strategically based on which cards the advisor could see (Experiment 1), how helpful the advisor was (Experiment 2), and what strategy the advisor used to select advice (Experiment 3). Overall, participants benefited even from imperfect advice based on incomplete information. Participants' responses were consistent with a Bayesian model that jointly infers how the advisor selects advice and the value of the advisor's card, compared to an alternative model that weights advice based on the advisor's accuracy. By reasoning about others' minds, human learners can make the best of even noisy, impoverished social information.
View details for DOI 10.1111/tops.12388
View details for PubMedID 30414253
Development of Children's Sensitivity to Overinformativeness in Learning and Teaching
2018; 54 (11): 2113–25
Effective communication requires knowing the "right" amount of information to provide; what is necessary for a naïve learner to arrive at a target hypothesis may be superfluous and inefficient for a knowledgeable learner. The current study examines 4- to 7-year-olds' developing sensitivity to overinformative communication and their ability to decide how much information is appropriate depending on the learner's prior knowledge. In Experiment 1 (N = 184, age = 4.09-7.98 years), 5- to 7-year-old children preferred teachers who gave costly, exhaustive demonstrations when learners were naïve, but preferred teachers who gave efficient, selective demonstrations when learners were already knowledgeable given their prior experience (i.e., common ground). However, 4-year-olds did not show a clear preference. In Experiment 2 (N = 80, age = 4.05-6.99 years), we asked whether children flexibly modulated their own teaching based on learners' knowledge. Five and 6-year-olds, but not 4-year-olds, were more likely to provide exhaustive demonstrations to naïve learners than to knowledgeable learners. These results suggest that by 5 years of age, children are sensitive to overinformativeness and understand the trade-off between informativeness and efficiency; they reason about what others know based on the presence or absence of common ground and flexibly decide how much information is appropriate both as learners and as teachers. (PsycINFO Database Record (c) 2018 APA, all rights reserved).
View details for DOI 10.1037/dev0000580
View details for Web of Science ID 000448187100010
View details for PubMedID 30265027
- Means-Inference as a Source of Variability in Early Helping FRONTIERS IN PSYCHOLOGY 2018; 9
From Exploration to Instruction: Children Learn From Exploration and Tailor Their Demonstrations to Observers' Goals and Competence.
This study investigated whether children learn from exploration and act as effective informants by providing informative demonstrations tailored to observers' goals and competence. Children (4.0-6.9years, N=98) explored a causally ambiguous toy to discover its causal structure and then demonstrated the toy to a naive observer. Children provided more costly and informative evidence when the observer wanted to learn about the toy than observe its effects (Experiment 1) and when the observer was ordinary than exceptionally intelligent (Experiment 2). Relative to the evidence they generated during exploration, children produced fewer, less costly actions when the observer wanted or needed less evidence. Children understand the difference between acting-to-learn and acting-to-inform; after learning from exploration, they consider others' goals and competence to provide "uninstructed instruction".
View details for DOI 10.1111/cdev.13059
View details for PubMedID 29635785
'To the victor go the spoils': Infants expect resources to align with dominance structures.
2017; 164: 8-21
Previous research has found that within the first year of life infants possess rich knowledge about social structures (i.e., that some individuals are dominant over other individuals) as well as expectations about resource distributions (i.e., that resources are typically distributed equally to recipients). We investigated whether infants' expectations about resource distribution can be modulated by information about the dominance structure between the recipients. We first replicated the finding that infants attribute a stable dominance hierarchy to a pair of individuals when their goals conflicted and one individual yielded to the other (Expt. 1), and that this sensitivity is not driven by lower-level perceptual factors (Expt. 2). In Experiments 3-5, we tested our main hypothesis that infants' attention to equal and unequal distributions varies as a function of prior social dominance information. We first replicated and extended prior work by establishing that infants looked significantly longer to unequal than equal resource distributions when no prior information about dominance was provided about recipients (Expt. 3). Critically, following social dominance information, infants looked significantly longer to an equal distribution of resources than a distribution that favored the dominant individual (Expt. 4), and looked significantly longer when the submissive individual received more resources compared to when the dominant individual received more resources (Expts. 4 and 5). Together, these findings suggest that infants expect resources to align with social dominance structures.
View details for DOI 10.1016/j.cognition.2017.03.008
View details for PubMedID 28346870
Order Matters: Children's Evaluation of Underinformative Teachers Depends on Context.
The ability to evaluate "sins of omission"-true but pragmatically misleading, underinformative pedagogy-is critical for learning. This study reveals a developmental change in children's evaluation of underinformative teachers and investigates the nature of their limitations. Participants rated a fully informative teacher and an underinformative teacher in two different orders. Six- and 7-year-olds (N = 28) successfully distinguished the teachers regardless of the order (Experiment 1), whereas 4- and 5-year-olds (N = 82) succeeded only when the fully informative teacher came first (Experiments 2 and 3). After seeing both teachers, 4-year-olds (N = 32) successfully preferred the fully informative teacher (Experiment 4). These results are discussed in light of developmental work in pragmatic implicature, suggesting that young children might struggle with spontaneously generating relevant alternatives for evaluating underinformative pedagogy.
View details for DOI 10.1111/cdev.12825
View details for PubMedID 28542838
Learning the Structure of Social Influence
2017; 41: 545-575
We routinely observe others' choices and use them to guide our own. Whose choices influence us more, and why? Prior work has focused on the effect of perceived similarity between two individuals (self and others), such as the degree of overlap in past choices or explicitly recognizable group affiliations. In the real world, however, any dyadic relationship is part of a more complex social structure involving multiple social groups that are not directly observable. Here we suggest that human learners go beyond dyadic similarities in choice behaviors or explicit group memberships; they infer the structure of social influence by grouping individuals (including themselves) based on choices, and they use these groups to decide whose choices to follow. We propose a computational model that formalizes this idea, and we test the model predictions in a series of behavioral experiments. In Experiment 1, we reproduce a well-established finding that people's choices are more likely to be influenced by someone whose past choices are more similar to their own past choices, as predicted by our model as well as dyadic similarity models. In Experiments 2-5, we test a set of unique predictions of our model by looking at cases where the degree of choice overlap between individuals is equated, but their choices indicate a latent group structure. We then apply our model to prior empirical results on infants' understanding of others' preferences, presenting an alternative account of developmental changes. Finally, we discuss how our model relates to classical findings in the social influence literature and the theoretical implications of our model. Taken together, our findings demonstrate that structure learning is a powerful framework for explaining the influence of social information on decision making in a variety of contexts.
View details for DOI 10.1111/cogs.12480
View details for Web of Science ID 000399736400008
View details for PubMedID 28294384
The Naïve Utility Calculus: Computational Principles Underlying Commonsense Psychology.
Trends in cognitive sciences
2016; 20 (8): 589-604
We propose that human social cognition is structured around a basic understanding of ourselves and others as intuitive utility maximizers: from a young age, humans implicitly assume that agents choose goals and actions to maximize the rewards they expect to obtain relative to the costs they expect to incur. This 'naïve utility calculus' allows both children and adults observe the behavior of others and infer their beliefs and desires, their longer-term knowledge and preferences, and even their character: who is knowledgeable or competent, who is praiseworthy or blameworthy, who is friendly, indifferent, or an enemy. We review studies providing support for the naïve utility calculus, and we show how it captures much of the rich social reasoning humans engage in from infancy.
View details for DOI 10.1016/j.tics.2016.05.011
View details for PubMedID 27388875
Learning From Others and Spontaneous Exploration: A Cross-Cultural Investigation
2016; 87 (3): 723-735
How does early social experience affect children's inferences and exploration? Following prior work on children's reasoning in pedagogical contexts, this study examined U.S. children with less experience in formal schooling and Yucatec Mayan children whose early social input is predominantly observational. In Experiment 1, U.S. 2-year-olds (n = 77) showed more restricted exploration of a toy following a pedagogical demonstration than an interrupted, accidental, or no demonstration (baseline). In Experiment 2, Yucatec Mayan and U.S. 2-year-olds (n = 66) showed more restricted exploration following a pedagogical than an observational demonstration, while only Mayan children showed more restriction with age. These results suggest that although schooling is not a necessary precursor for sensitivity to pedagogy, early social experience may influence children's inferences and exploration in pedagogical contexts.
View details for DOI 10.1111/cdev.12502
View details for Web of Science ID 000379913500009
View details for PubMedID 27189400
Children's understanding of the costs and rewards underlying rational action
2015; 140: 14-23
Humans explain and predict other agents' behavior using mental state concepts, such as beliefs and desires. Computational and developmental evidence suggest that such inferences are enabled by a principle of rational action: the expectation that agents act efficiently, within situational constraints, to achieve their goals. Here we propose that the expectation of rational action is instantiated by a naïve utility calculus sensitive to both agent-constant and agent-specific aspects of costs and rewards associated with actions. In four experiments, we show that, given an agent's choices, children (range: 5-6 year olds; N=96) can infer unobservable aspects of costs (differences in agents' competence) from information about subjective differences in rewards (differences in agents' preferences) and vice versa. Moreover, children can design informative experiments on both objects and agents to infer unobservable constraints on agents' actions.
View details for DOI 10.1016/j.cognition.2015.03.006
View details for Web of Science ID 000355042100002
View details for PubMedID 25867996
Sins of omission: Children selectively explore when teachers are under-informative
2014; 132 (3): 335-341
Do children know when people tell the truth but not the whole truth? Here we show that children accurately evaluate informants who omit information and adjust their exploratory behavior to compensate for under-informative pedagogy. Experiment 1 shows that given identical demonstrations of a toy, children (6- and 7-year-olds) rate an informant lower if the toy also had non-demonstrated functions. Experiment 2 shows that given identical demonstrations, six-year-olds explore a toy more broadly if the informant previously committed a sin of omission. These results suggest that children consider both accuracy and informativeness in evaluating others' credibility and adjust their exploratory behavior to compensate for under-informative testimony when an informant's credibility is in doubt.
View details for DOI 10.1016/j.cognition.2014.04.013
View details for Web of Science ID 000340013900009
View details for PubMedID 24873737
Differences in the right inferior longitudinal fasciculus but no general disruption of white matter tracts in children with autism spectrum disorder
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
2014; 111 (5): 1981-1986
One of the most widely cited features of the neural phenotype of autism is reduced "integrity" of long-range white matter tracts, a claim based primarily on diffusion imaging studies. However, many prior studies have small sample sizes and/or fail to address differences in data quality between those with autism spectrum disorder (ASD) and typical participants, and there is little consensus on which tracts are affected. To overcome these problems, we scanned a large sample of children with autism (n = 52) and typically developing children (n = 73). Data quality was variable, and worse in the ASD group, with some scans unusable because of head motion artifacts. When we follow standard data analysis practices (i.e., without matching head motion between groups), we replicate the finding of lower fractional anisotropy (FA) in multiple white matter tracts. However, when we carefully match data quality between groups, all these effects disappear except in one tract, the right inferior longitudinal fasciculus (ILF). Additional analyses showed the expected developmental increases in the FA of fiber tracts within ASD and typical groups individually, demonstrating that we had sufficient statistical power to detect known group differences. Our data challenge the widely claimed general disruption of white matter tracts in autism, instead implicating only one tract, the right ILF, in the ASD phenotype.
View details for DOI 10.1073/pnas.1324037111
View details for Web of Science ID 000330587600075
View details for PubMedID 24449864
Theory of Mind Performance in Children Correlates With Functional Specialization of a Brain Region for Thinking About Thoughts
2012; 83 (6): 1853-1868
Thinking about other people's thoughts recruits a specific group of brain regions, including the temporo-parietal junctions (TPJ), precuneus (PC), and medial prefrontal cortex (MPFC). The same brain regions were recruited when children (N=20, 5-11 years) and adults (N=8) listened to descriptions of characters' mental states, compared to descriptions of physical events. Between ages 5 and 11 years, responses in the bilateral TPJ became increasingly specific to stories describing mental states as opposed to people's appearance and social relationships. Functional activity in the right TPJ was related to children's performance on a high level theory of mind task. These findings provide insights into the origin of neural mechanisms of theory of mind, and how behavioral and neural changes can be related in development.
View details for DOI 10.1111/j.1467-8624.2012.01829.x
View details for Web of Science ID 000314111900002
View details for PubMedID 22849953
The double-edged sword of pedagogy: Instruction limits spontaneous exploration and discovery
2011; 120 (3): 322-330
Motivated by computational analyses, we look at how teaching affects exploration and discovery. In Experiment 1, we investigated children's exploratory play after an adult pedagogically demonstrated a function of a toy, after an interrupted pedagogical demonstration, after a naïve adult demonstrated the function, and at baseline. Preschoolers in the pedagogical condition focused almost exclusively on the target function; by contrast, children in the other conditions explored broadly. In Experiment 2, we show that children restrict their exploration both after direct instruction to themselves and after overhearing direct instruction given to another child; they do not show this constraint after observing direct instruction given to an adult or after observing a non-pedagogical intentional action. We discuss these findings as the result of rational inductive biases. In pedagogical contexts, a teacher's failure to provide evidence for additional functions provides evidence for their absence; such contexts generalize from child to child (because children are likely to have comparable states of knowledge) but not from adult to child. Thus, pedagogy promotes efficient learning but at a cost: children are less likely to perform potentially irrelevant actions but also less likely to discover novel information.
View details for DOI 10.1016/j.cognition.2010.10.001
View details for Web of Science ID 000293312400003
View details for PubMedID 21216395
16-Month-Olds Rationally Infer Causes of Failed Actions
2011; 332 (6037): 1524-1524
Sixteen-month-old infants (N = 83) rationally used sparse data about the distribution of outcomes among agents and objects to solve a fundamental inference problem: deciding whether event outcomes are due to themselves or the world. When infants experienced failed outcomes, their causal attributions affected whether they sought help or explored.
View details for DOI 10.1126/science.1204493
View details for Web of Science ID 000291990000035
View details for PubMedID 21700866
Infants consider both the sample and the sampling process in inductive generalization
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
2010; 107 (20): 9066-9071
The ability to make inductive inferences from sparse data is a critical aspect of human learning. However, the properties observed in a sample of evidence depend not only on the true extension of those properties but also on the process by which evidence is sampled. Because neither the property extension nor the sampling process is directly observable, the learner's ability to make accurate generalizations depends on what is known or can be inferred about both variables. In particular, different inferences are licensed if samples are drawn randomly from the whole population (weak sampling) than if they are drawn only from the property's extension (strong sampling). Given a few positive examples of a concept, only strong sampling supports flexible inferences about how far to generalize as a function of the size and composition of the sample. Here we present a Bayesian model of the joint dependence between observed evidence, the sampling process, and the property extension and test the model behaviorally with human infants (mean age: 15 months). Across five experiments, we show that in the absence of behavioral cues to the sampling process, infants make inferences consistent with the use of strong sampling; given explicit cues to weak or strong sampling, they constrain their inferences accordingly. Finally, consistent with quantitative predictions of the model, we provide suggestive evidence that infants' inferences are graded with respect to the strength of the evidence they observe.
View details for DOI 10.1073/pnas.1003095107
View details for Web of Science ID 000277822600011
View details for PubMedID 20435914