Children Use Temporal Cues to Learn Causal Directionality

The ability to learn the direction of causal relations is critical for understanding and acting in the world. We investigated how children learn causal directionality in situations in which the states of variables are temporally dependent (i.e., autocorrelated). In Experiment 1, children learned about causal direction by comparing the states of one variable before versus after an intervention on another variable. In Experiment 2, children reliably inferred causal directionality merely from observing how two variables change over time; they interpreted Y changing without a change in X as evidence that Y does not influence X. Both of these strategies make sense if one believes the variables to be temporally dependent. We discuss the implications of these results for interpreting previous findings. More broadly, given that many real‐world environments are characterized by temporal dependency, these results suggest strategies that children may use to learn the causal structure of their environments.     

Cause or effect: What matters? How 12‐month‐old infants learn to categorize artifacts

This paper investigates the role of cause and effect relations for infants' learning about artifacts. Two experiments tested whether 12‐month‐olds categorized a given set of unfamiliar artifacts according to overall similarity and/or according to part similarity, depending on what kind of video demonstration was presented before the start of the categorization task. In both experiments, the actions performed with objects were accompanied by interesting effects but the causal relation between object‐structure and effects was teased apart. In one video demonstration (Expt 1), the experimenter used the object part to produce some kind of effect in a causally plausible way. In another video demonstration (Expt 2), the experimenter performed similar actions with the same objects as in Expt 1, followed by the same effects as before. Importantly, however, no plausible cause–effect relation was provided this time. Only infants participating in Expt 1 categorized the set of unfamiliar objects according to part similarity. This finding suggests that 12‐month‐olds attend to the causal relation between specific object parts and their functional use when categorizing artifacts, rather than merely associating form‐characteristics with an interesting effect.     

Two‐year‐olds are vigilant of others’ non‐verbal cues to credibility

Data from three experiments provide the first evidence that children, at least as young as age two, are vigilant of others’ non‐verbal cues to credibility, and flexibly use these cues to facilitate learning. Experiment 1 revealed that 2‐ and 3‐year‐olds prefer to learn about objects from someone who appears, through non‐verbal cues, to be confident in performing actions on those objects than from someone who appears uncertain when performing actions on those objects. Experiment 2 revealed that when 2‐year‐olds observe only one model perform a single action, either confidently or unconfidently, they do not use the model’s level of confidence in this single instance to influence their learning. Experiment 3 revealed that 2‐year‐olds will use a single model’s level of confidence to guide their learning if they have observed that the model has a history of being either consistently confident or consistently uncertain. These findings reveal that young children selectively alter their learning based on others’ non‐verbal cues of credibility, and underscore the importance of an early sensitivity to socio‐cognitive cues for human learning and development.     

The importance of decision making in causal learning from interventions

Recent research has focused on how interventions benefit causal learning. This research suggests that the main benefit of interventions is in the temporal and conditional probability information that interventions provide a learner. But when one generates interventions, one must also decide what interventions to generate. In three experiments, we investigated the importance of these decision demands to causal learning. Experiment 1 demonstrated that learners were better at learning causal models when they observed intervention data that they had generated, as opposed to observing data generated by another learner. Experiment 2 demonstrated the same effect between self-generated interventions and interventions learners were forced to make. Experiment 3 demonstrated that when learners observed a sequence of interventions such that the decision-making process that generated those interventions was more readily available, learning was less impaired. These data suggest that decision making may be an important part of causal learning from interventions.     

Parallels Between Action‐Object Mapping and Word‐Object Mapping in Young Children

Across a series of four experiments with 3‐ to 4‐year‐olds we demonstrate how cognitive mechanisms supporting noun learning extend to the mapping of actions to objects. In Experiment 1 (n = 61) the demonstration of a novel action led children to select a novel, rather than a familiar object. In Experiment 2 (n = 78) children exhibited long‐term retention of novel action‐object mappings and extended these actions to other category members. In Experiment 3 (n = 60) we showed that children formed an accurate sensorimotor record of the novel action. In Experiment 4 (n = 54) we demonstrate limits on the types of actions mapped to novel objects. Overall these data suggest that certain aspects of noun mapping share common processing with action mapping and support a domain‐general account of word learning.     

The Invisible Hand: Toddlers Connect Probabilistic Events With Agentive Causes

Children posit unobserved causes when events appear to occur spontaneously (e.g., Gelman & Gottfried, 1996). What about when events appear to occur probabilistically? Here toddlers (M = 20.1 months) saw arbitrary causal relationships (Cause A generated Effect A; Cause B generated Effect B) in a fixed, alternating order. The relationships were then changed in one of two ways. In the Deterministic condition, the event order changed (Event B preceded Event A); in the Probabilistic condition, the causal relationships changed (Cause A generated Effect B; Cause B generated Effect A). As intended, toddlers looked equally long at both changes (Experiment 1). We then introduced a previously unseen candidate cause. Toddlers looked longer at the appearance of a hand (Experiment 2) and novel agent (Experiment 3) in the Deterministic than the Probabilistic conditions, but looked equally long at novel non‐agents (Experiment 4), suggesting that by 2 years of age, toddlers connect probabilistic events with unobserved agents.     

Making Probabilistic Relational Categories Learnable

Theories of relational concept acquisition (e.g., schema induction) based on structured intersection discovery predict that relational concepts with a probabilistic (i.e., family resemblance) structure ought to be extremely difficult to learn. We report four experiments testing this prediction by investigating conditions hypothesized to facilitate the learning of such categories. Experiment 1 showed that changing the task from a category‐learning task to choosing the “winning” object in each stimulus greatly facilitated participants' ability to learn probabilistic relational categories. Experiments 2 and 3 further investigated the mechanisms underlying this “who's winning” effect. Experiment 4 replicated and generalized the “who's winning” effect with more natural stimuli. Together, our findings suggest that people learn relational concepts by a process of intersection discovery akin to schema induction, and that any task that encourages people to discover a higher order relation that remains invariant over members of a category will facilitate the learning of putatively probabilistic relational concepts.     

Information from multiple modalities helps 5‐month‐olds learn abstract rules

By 7 months of age, infants are able to learn rules based on the abstract relationships between stimuli ( Marcus et al ., 1999 ), but they are better able to do so when exposed to speech than to some other classes of stimuli. In the current experiments we ask whether multimodal stimulus information will aid younger infants in identifying abstract rules. We habituated 5‐month‐olds to simple abstract patterns (ABA or ABB) instantiated in coordinated looming visual shapes and speech sounds (Experiment 1), shapes alone (Experiment 2), and speech sounds accompanied by uninformative but coordinated shapes (Experiment 3). Infants showed evidence of rule learning only in the presence of the informative multimodal cues. We hypothesize that the additional evidence present in these multimodal displays was responsible for the success of younger infants in learning rules, congruent with both a Bayesian account and with the Intersensory Redundancy Hypothesis.     

Inferences about unobserved causes in human contingency learning

Estimates of the causal efficacy of an event need to take into account the possible presence and influence of other unobserved causes that might have contributed to the occurrence of the effect. Current theoretical approaches deal differently with this problem. Associative theories assume that at least one unobserved cause is always present. In contrast, causal Bayes net theories (including Power PC theory) hypothesize that unobserved causes may be present or absent. These theories generally assume independence of different causes of the same event, which greatly simplifies modelling learning and inference. In two experiments participants were requested to learn about the causal relation between a single cause and an effect by observing their co-occurrence (Experiment 1) or by actively intervening in the cause (Experiment 2). Participants' assumptions about the presence of an unobserved cause were assessed either after each learning trial or at the end of the learning phase. The results show an interesting dissociation. Whereas there was a tendency to assume interdependence of the causes in the online judgements during learning, the final judgements tended to be more in the direction of an independence assumption. Possible explanations and implications of these findings are discussed.     

Semantic detail in the developing verb lexicon: An extension of Naigles and Kako (1993)

Verbs are often uttered before the events they describe. By 2 years of age, toddlers can learn from such an encounter. Hearing a novel verb in transitive sentences (e.g. The boy lorped the cat), even with no visual referent present, they later map it to a causative meaning (e.g. feed) (e.g. Yuan & Fisher, 2009 ). How much semantic detail does their verb representation include on this first, underinformative, encounter? Is the representation sparse, including only information for which they have evidence, or do toddlers make more specific guesses about the verb's meaning? In two experiments (N = 76, mean age 27 months), we address this using an event type studied by Naigles and Kako ( 1993 ); they found that when toddlers hear a novel transitive verb while simultaneously viewing a non‐causative referent—a contact event such as patting—they map the verb to the contact event. In Experiment 1 we replicated this basic result. Further, toddlers’ representations persisted over a 5‐minute delay, manifesting again during a retest. In Experiment 2, toddlers heard the verbs while watching two actors converse instead of while seeing contact events. At test, they showed no evidence of mapping the verbs to contact events, either initially or after a 5‐minute delay, despite that in prior work they mapped verbs to causative events under identical circumstances. We infer that on hearing a novel verb in a transitive frame, absent a relevant visual scene, toddlers posit a more specific representation than the evidence requires—one that incorporates causative semantics. A video abstract of this article can be viewed at: https://youtu.be/aRCqSTbr6Bw     

The Role of Probability and Intentionality in Preschoolers' Causal Generalizations

Three experiments examined whether preschoolers recognize that the causal properties of objects generalize to new members of the same set given either deterministic or probabilistic data. Experiment 1 found that 3- and 4-year-olds were able to make such a generalization given deterministic data but were at chance when they observed probabilistic information. Five-year-olds reliably generalized in both situations. Experiment 2 found that 4-year-olds could make some probabilistic inferences, particularly when comparing sets that had no efficacy with sets in which some members had efficacy. Children had some difficulty discriminating between completely effective sets and stochastic ones. Experiment 3 examined whether 3- and 4-year-olds could reason about probabilistic data when provided with information about the experimenter's beliefs about causal outcomes. Children who were more successful on standard false-belief measures were more likely to respond as if the data were deterministic. These data suggest that children's probabilistic inferences develop into early elementary school, but preschoolers might have some understanding of probability when reasoning about causal generalization.     

Causal structure learning over time: observations and interventions

Seven studies examined how people learn causal relationships in scenarios when the variables are temporally dependent - the states of variables are stable over time. When people intervene on X, and Y subsequently changes state compared to before the intervention, people infer that X influences Y. This strategy allows people to learn causal structures quickly and reliably when variables are temporally stable (Experiments 1 and 2). People use this strategy even when the cover story suggests that the trials are independent (Experiment 3). When observing variables over time, people believe that when a cause changes state, its effects likely change state, but an effect may change state due to an exogenous influence in which case its observed cause may not change state at the same time. People used this strategy to learn the direction of causal relations and a wide variety of causal structures (Experiments 4-6). Finally, considering exogenous influences responsible for the observed changes facilitates learning causal directionality (Experiment 7). Temporal reasoning may be the norm rather than the exception for causal learning and may reflect the way most events are experienced naturalistically.     

Conditional probability versus spatial contiguity in causal learning: Preschoolers use new contingency evidence to overcome prior spatial assumptions

This study examines preschoolers' causal assumptions about spatial contiguity and how these assumptions interact with new evidence in the form of conditional probabilities. Preschoolers saw a toy that activated in the presence of certain objects. Children were shown evidence for the toy's activation rule in the form of patterns of probability: The toy was more likely to activate either when objects made contact with its surface (on condition) or when objects were several inches above its surface (over condition). In Experiment 1, 61 three-year-olds saw a deterministic activation rule. In Experiments 2 and 3, 48 four-year-olds saw an activation rule that was probabilistic. In Experiment 4, 30 four-year-olds saw a screening-off pattern of activation. In all 4 experiments, children used new evidence in the form of patterns of probability to make accurate causal inferences, even in the face of conflicting prior beliefs about spatial contiguity. However, children were more likely to make correct inferences when causes were spatially contiguous, particularly when faced with ambiguous evidence.     

The role of causal models in multiple judgments under uncertainty

Two studies examined a novel prediction of the causal Bayes net approach to judgments under uncertainty, namely that causal knowledge affects the interpretation of statistical evidence obtained over multiple observations. Participants estimated the conditional probability of an uncertain event (breast cancer) given information about the base rate, hit rate (probability of a positive mammogram given cancer) and false positive rate (probability of a positive mammogram in the absence of cancer). Conditional probability estimates were made after observing one or two positive mammograms. Participants exhibited a causal stability effect: there was a smaller increase in estimates of the probability of cancer over multiple positive mammograms when a causal explanation of false positives was provided. This was the case when the judgments were made by different participants (Experiment 1) or by the same participants (Experiment 2). These results show that identical patterns of observed events can lead to different estimates of event probability depending on beliefs about the generative causes of the observations.     

Fourteen‐month‐old infants track the language comprehension of communicative partners

Infants employ sophisticated mechanisms to acquire their first language, including some that rely on taking the perspective of adults as speakers or listeners. When do infants first show awareness of what other people understand? We tested 14‐month‐old infants in two experiments measuring event‐related potentials. In Experiment 1, we established that infants produce the N400 effect, a brain signature of semantic violations, in a live object naming paradigm in the presence of an adult observer. In Experiment 2, we induced false beliefs about the labeled objects in the adult observer to test whether infants keep track of the other person's comprehension. The results revealed that infants reacted to the semantic incongruity heard by the other as if they encountered it themselves: they exhibited an N400‐like response, even though labels were congruous from their perspective. This finding demonstrates that infants track the linguistic understanding of social partners. A video abstract of this article can be viewed at https://youtu.be/pQUv8yFhnbk .     

Gradations of emulation learning in infants' imitation of actions on objects

This study explored different gradations of emulation in the imitation of actions on objects by 17-month-olds. Experiment 1 established levels of behavioral reproduction following prerecorded video demonstrations similar to those levels following live demonstrations. In Experiment 2, two digitally modified videos, where object movements or body movements critical to producing the target action were highlighted in isolation, were developed. Infants produced the target action equally frequently by observing the object movement video and observing the unmodified video. In contrast, their performance was much less successful based on the body movement video. In Experiment 3, the performance obtained following the object movement video was similar to that following a further video that emphasized the object movements produced in unsuccessful attempts to produce the target action. These findings suggest that emulation in the form of object movement reenactment or affordance learning plays a role in the social learning of actions on objects during infancy.     

Four‐month‐old infants individuate and track simple tools following functional demonstrations

Two experiments examined whether 4‐month‐olds (n = 120) who were induced to assign two objects to different categories would then be able to take advantage of these contrastive categorical encodings to individuate and track the objects. In each experiment, infants first watched functional demonstrations of two tools, a masher and tongs (Experiment 1) or a marker and a knife (Experiment 2). Next, half the infants saw the two tools brought out alternately from behind a screen, which was then lowered to reveal only one of the tools (different‐objects condition); the other infants saw similar events except that the same tool was shown on either side of the screen (same‐object condition). In both experiments, infants in the different‐objects condition looked reliably longer than those in the same‐object condition, and this effect was eliminated if the demonstrations involved similar but non‐functional actions. Together, these results indicate that infants (a) were led by the functional demonstrations they observed to assign the two tools to distinct categories, (b) recruited these categorical encodings to individuate and track the tools, and hence (c) detected a violation in the different‐objects condition when the screen was lowered to reveal only one tool. Categorical information thus plays a privileged role in individuation and identity tracking from a very young age.     

Lexical‐processing efficiency leverages novel word learning in infants and toddlers

Children who rapidly recognize and interpret familiar words typically have accelerated lexical growth, providing indirect evidence that lexical processing efficiency (LPE) is related to word‐learning ability. Here we directly tested whether children with better LPE are better able to learn novel words. In Experiment 1, 17‐ and 30‐month‐olds were tested on an LPE task and on a simple word‐learning task. The 17‐month‐olds’ LPE scores predicted word learning in a regression model, and only those with relatively good LPE showed evidence of learning. The 30‐month‐olds learned novel words quite well regardless of LPE, but in a more difficult word‐learning task (Experiment 2), their LPE predicted word‐learning ability. These findings suggest that LPE supports word‐learning processes, especially when learning is difficult.     

Inferring the size of a goal object from an actor's grasping movement in 6‐ and 9‐month‐old infants

The present study applied a preferential looking paradigm to test whether 6‐ and 9‐month old infants are able to infer the size of a goal object from an actor's grasping movement. The target object was a cup with the handle rotated either towards or away from the actor. In two experiments, infants saw the video of an actor's grasping movement towards an occluded target object. The aperture size of the actor's hand was varied as between‐subjects factor. Subsequently, two final states of the grasping movement were presented simultaneously with the occluder being removed. In Experiment 1, the expected final state showed the actor's hand holding a cup in a way that would be expected after the performed grasping movement. In the unexpected final state, the actor's hand held the cup at the side which would be unexpected after the performed grasping movement. Results show that 6‐ as well as 9‐month‐olds looked longer at the unexpected than at the expected final state. Experiment 2 excluded an alternative explanation of these findings, namely that the discrimination of the final states was due to geometrical familiarity or novelty of the final states. These findings provide evidence that infants are able to infer the size of a goal object from the aperture size of the actor's hand during the grasp.     

Is Probabilistic Evidence a Source of Knowledge?

We report a series of experiments examining whether people ascribe knowledge for true beliefs based on probabilistic evidence. Participants were less likely to ascribe knowledge for beliefs based on probabilistic evidence than for beliefs based on perceptual evidence (Experiments 1 and 2A) or testimony providing causal information (Experiment 2B). Denial of knowledge for beliefs based on probabilistic evidence did not arise because participants viewed such beliefs as unjustified, nor because such beliefs leave open the possibility of error. These findings rule out traditional philosophical accounts for why probabilistic evidence does not produce knowledge. The experiments instead suggest that people deny knowledge because they distrust drawing conclusions about an individual based on reasoning about the population to which it belongs, a tendency previously identified by “judgment and decision making” researchers. Consistent with this, participants were more willing to ascribe knowledge for beliefs based on probabilistic evidence that is specific to a particular case (Experiments 3A and 3B).