Children's imitation of causal action sequences is influenced by statistical and pedagogical evidence

Children are ubiquitous imitators, but how do they decide which actions to imitate? One possibility is that children rationally combine multiple sources of information about which actions are necessary to cause a particular outcome. For instance, children might learn from contingencies between action sequences and outcomes across repeated demonstrations, and they might also use information about the actor’s knowledge state and pedagogical intentions. We define a Bayesian model that predicts children will decide whether to imitate part or all of an action sequence based on both the pattern of statistical evidence and the demonstrator’s pedagogical stance. To test this prediction, we conducted an experiment in which preschool children watched an experimenter repeatedly perform sequences of varying actions followed by an outcome. Children’s imitation of sequences that produced the outcome increased, in some cases resulting in production of shorter sequences of actions that the children had never seen performed in isolation. A second experiment established that children interpret the same statistical evidence differently when it comes from a knowledgeable teacher versus a naïve demonstrator. In particular, in the pedagogical case children are more likely to “overimitate” by reproducing the entire demonstrated sequence. This behavior is consistent with our model’s predictions, and suggests that children attend to both statistical and pedagogical evidence in deciding which actions to imitate, rather than obligately imitating successful action sequences.     

Inferring hidden causal structure

We used a new method to assess how people can infer unobserved causal structure from patterns of observed events. Participants were taught to draw causal graphs, and then shown a pattern of associations and interventions on a novel causal system. Given minimal training and no feedback, participants in Experiment 1 used causal graph notation to spontaneously draw structures containing one observed cause, one unobserved common cause, and two unobserved independent causes, depending on the pattern of associations and interventions they saw. We replicated these findings with less-informative training (Experiments 2 and 3) and a new apparatus (Experiment 3) to show that the pattern of data leads to hidden causal inferences across a range of prior constraints on causal knowledge.     

Segmenting dynamic human action via statistical structure

Human social, cognitive, and linguistic functioning depends on skills for rapidly processing action. Identifying distinct acts within the dynamic motion flow is one basic component of action processing; for example, skill at segmenting action is foundational to action categorization, verb learning, and comprehension of novel action sequences. Yet little is currently known about mechanisms that may subserve action segmentation. The present research documents that adults can register statistical regularities providing clues to action segmentation. This finding provides new evidence that structural knowledge gained by mechanisms such as statistical learning can play a role in action segmentation, and highlights a striking parallel between processing of action and processing in other domains, such as language.     

Inferring causal networks from observations and interventions

Information about the structure of a causal system can come in the form of observational data—random samples of the system's autonomous behavior—or interventional data—samples conditioned on the particular values of one or more variables that have been experimentally manipulated. Here we study people's ability to infer causal structure from both observation and intervention, and to choose informative interventions on the basis of observational data. In three causal inference tasks, participants were to some degree capable of distinguishing between competing causal hypotheses on the basis of purely observational data. Performance improved substantially when participants were allowed to observe the effects of interventions that they performed on the systems. We develop computational models of how people infer causal structure from data and how they plan intervention experiments, based on the representational framework of causal graphical models and the inferential principles of optimal Bayesian decision‐making and maximizing expected information gain. These analyses suggest that people can make rational causal inferences, subject to psychologically reasonable representational assumptions and computationally reasonable processing constraints.     

Learning the form of causal relationships using hierarchical bayesian models

People learn quickly when reasoning about causal relationships, making inferences from limited data and avoiding spurious inferences. Efficient learning depends on abstract knowledge, which is often domain or context specific, and much of it must be learned. While such knowledge effects are well documented, little is known about exactly how we acquire knowledge that constrains learning. This work focuses on knowledge of the functional form of causal relationships; there are many kinds of relationships that can apply between causes and their effects, and knowledge of the form such a relationship takes is important in order to quickly identify the real causes of an observed effect. We developed a hierarchical Bayesian model of the acquisition of knowledge of the functional form of causal relationships and tested it in five experimental studies, considering disjunctive and conjunctive relationships, failure rates, and cross-domain effects. The Bayesian model accurately predicted human judgments and outperformed several alternative models.     

Identifying expectations about the strength of causal relationships

When we try to identify causal relationships, how strong do we expect that relationship to be? Bayesian models of causal induction rely on assumptions regarding people’s a priori beliefs about causal systems, with recent research focusing on people’s expectations about the strength of causes. These expectations are expressed in terms of prior probability distributions. While proposals about the form of such prior distributions have been made previously, many different distributions are possible, making it difficult to test such proposals exhaustively. In Experiment 1 we used iterated learning—a method in which participants make inferences about data generated based on their own responses in previous trials—to estimate participants’ prior beliefs about the strengths of causes. This method produced estimated prior distributions that were quite different from those previously proposed in the literature. Experiment 2 collected a large set of human judgments on the strength of causal relationships to be used as a benchmark for evaluating different models, using stimuli that cover a wider and more systematic set of contingencies than previous research. Using these judgments, we evaluated the predictions of various Bayesian models. The Bayesian model with priors estimated via iterated learning compared favorably against the others. Experiment 3 estimated participants’ prior beliefs concerning different causal systems, revealing key similarities in their expectations across diverse scenarios.     

Structure and strength in causal induction

We present a framework for the rational analysis of elemental causal induction-learning about the existence of a relationship between a single cause and effect-based upon causal graphical models. This framework makes precise the distinction between causal structure and causal strength: the difference between asking whether a causal relationship exists and asking how strong that causal relationship might be. We show that two leading rational models of elemental causal induction, DeltaP and causal power, both estimate causal strength, and we introduce a new rational model, causal support, that assesses causal structure. Causal support predicts several key phenomena of causal induction that cannot be accounted for by other rational models, which we explore through a series of experiments. These phenomena include the complex interaction between DeltaP and the base-rate probability of the effect in the absence of the cause, sample size effects, inferences from incomplete contingency tables, and causal learning from rates. Causal support also provides a better account of a number of existing datasets than either DeltaP or causal power.     

Independence and dependence in human causal reasoning

Causal graphical models (CGMs) are a popular formalism used to model human causal reasoning and learning. The key property of CGMs is the causal Markov condition, which stipulates patterns of independence and dependence among causally related variables. Five experiments found that while adult’s causal inferences exhibited aspects of veridical causal reasoning, they also exhibited a small but tenacious tendency to violate the Markov condition. They also failed to exhibit robust discounting in which the presence of one cause as an explanation of an effect makes the presence of another less likely. Instead, subjects often reasoned “associatively,” that is, assumed that the presence of one variable implied the presence of other, causally related variables, even those that were (according to the Markov condition) conditionally independent. This tendency was unaffected by manipulations (e.g., response deadlines) known to influence fast and intuitive reasoning processes, suggesting that an associative response to a causal reasoning question is sometimes the product of careful and deliberate thinking. That about 60% of the erroneous associative inferences were made by about a quarter of the subjects suggests the presence of substantial individual differences in this tendency. There was also evidence that inferences were influenced by subjects’ assumptions about factors that disable causal relations and their use of a conjunctive reasoning strategy. Theories that strive to provide high fidelity accounts of human causal reasoning will need to relax the independence constraints imposed by CGMs.     

Error probabilities for inference of causal directions

A main message from the causal modelling literature in the last several decades is that under some plausible assumptions, there can be statistically consistent procedures for inferring (features of) the causal structure of a set of random variables from observational data. But whether we can control the error probabilities with a finite sample size depends on the kind of consistency the procedures can achieve. It has been shown that in general, under the standard causal Markov and Faithfulness assumptions, the procedures can only be pointwise but not uniformly consistent without substantial background knowledge. This implies the impossibility of choosing a finite sample size to control the worst case error probabilities. In this paper, I consider the simpler task of inferring causal directions when the skeleton of the causal structure is known, and establish a similarly negative result concerning the possibility of controlling error probabilities. Although the result is negative in form, it has an interesting positive implication for causal discovery methods.     

Bayes and blickets: effects of knowledge on causal induction in children and adults

People are adept at inferring novel causal relations, even from only a few observations. Prior knowledge about the probability of encountering causal relations of various types and the nature of the mechanisms relating causes and effects plays a crucial role in these inferences. We test a formal account of how this knowledge can be used and acquired, based on analyzing causal induction as Bayesian inference. Five studies explored the predictions of this account with adults and 4-year-olds, using tasks in which participants learned about the causal properties of a set of objects. The studies varied the two factors that our Bayesian approach predicted should be relevant to causal induction: the prior probability with which causal relations exist, and the assumption of a deterministic or a probabilistic relation between cause and effect. Adults' judgments (Experiments 1, 2, and 4) were in close correspondence with the quantitative predictions of the model, and children's judgments (Experiments 3 and 5) agreed qualitatively with this account.     

The mental representation of causal conditional reasoning: mental models or causal models

In this paper, two experiments are reported investigating the nature of the cognitive representations underlying causal conditional reasoning performance. The predictions of causal and logical interpretations of the conditional diverge sharply when inferences involving pairs of conditionals—such as if P₁then Q and if P₂then Q—are considered. From a causal perspective, the causal direction of these conditionals is critical: are the P_icauses of Q; or symptoms caused byQ. The rich variety of inference patterns can naturally be modelled by Bayesian networks. A pair of causal conditionals where Q is an effect corresponds to a “collider” structure where the two causes (P_i) converge on a common effect. In contrast, a pair of causal conditionals where Q is a cause corresponds to a network where two effects (P_i) diverge from a common cause. Very different predictions are made by fully explicit or initial mental models interpretations. These predictions were tested in two experiments, each of which yielded data most consistent with causal model theory, rather than with mental models.     

The origins of probabilistic inference in human infants

Reasoning under uncertainty is the bread and butter of everyday life. Many areas of psychology, from cognitive, developmental, social, to clinical, are interested in how individuals make inferences and decisions with incomplete information. The ability to reason under uncertainty necessarily involves probability computations, be they exact calculations or estimations. What are the developmental origins of probabilistic reasoning? Recent work has begun to examine whether infants and toddlers can compute probabilities; however, previous experiments have confounded quantity and probability—in most cases young human learners could have relied on simple comparisons of absolute quantities, as opposed to proportions, to succeed in these tasks. We present four experiments providing evidence that infants younger than 12 months show sensitivity to probabilities based on proportions. Furthermore, infants use this sensitivity to make predictions and fulfill their own desires, providing the first demonstration that even preverbal learners use probabilistic information to navigate the world. These results provide strong evidence for a rich quantitative and statistical reasoning system in infants.     

Second-order conditioning of human causal learning

This article provides the first demonstration of a reliable second-order conditioning (SOC) effect in human causal learning tasks. It demonstrates the human ability to infer relationships between a cause and an effect that were never paired together during training. Experiments 1a and 1b showed a clear and reliable SOC effect, while Experiments 2a and 2b demonstrated that first-order extinction did not affect SOC. These results were similar to those found in animal and human conditioning and suggested that a similar associative mechanism could explain these effects. However, they can also be used to look into the underlying causal mental model people build and store while they are learning this task. From a cognitive view, overall results suggest that an independent rather than a chain causal mental model is stored after second-order learning in human causal tasks.     

Generality of the summation effect in human causal learning

Considerable research has examined the contrasting predictions of the elemental and configural association theories proposed by Rescorla and Wagner (1972) and Pearce (1987), respectively. One simple method to distinguish between these approaches is the summation test, in which the associative strength attributed to a novel compound of two separately trained cues is examined. Under common assumptions, the configural view predicts that the strength of the compound will approximate to the average strength of its components, whereas the elemental approach predicts that the strength of the compound will be greater than the strength of either component. Different studies have produced mixed outcomes. In studies of human causal learning, Collins and Shanks (2006) suggested that the observation of summation is encouraged by training, in which different stimuli are associated with different submaximal outcomes, and by testing, in which the alternative outcomes can be scaled. The reported experiments further pursued this reasoning. In Experiment 1, summation was more substantial when the participants were trained with outcomes identified as submaximal than when trained with simple categorical (presence/absence) outcomes. Experiments 2 and 3 demonstrated that summation can also be obtained with categorical outcomes during training, if the participants are encouraged by instruction or the character of training to rate the separately trained components with submaximal ratings. The results are interpreted in terms of apparent performance constraints in evaluations of the contrasting theoretical predictions concerning summation.     

When global structure "Explains Away" local grammar: a Bayesian account of rule-induction in tone sequences

While many constraints on learning must be relatively experience-independent, past experience provides a rich source of guidance for subsequent learning. Discovering structure in some domain can inform a learner’s future hypotheses about that domain. If a general property accounts for particular sub-patterns, a rational learner should not stipulate separate explanations for each detail without additional evidence, as the general structure has “explained away” the original evidence. In a grammar-learning experiment using tone sequences, manipulating learners’ prior exposure to a tone environment affects their sensitivity to the grammar-defining feature, in this case consecutive repeated tones. Grammar-learning performance is worse if context melodies are “smooth” — when small intervals occur more than large ones — as Smoothness is a general property accounting for a high rate of repetition. We present an idealized Bayesian model as a “best case” benchmark for learning repetition grammars. When context melodies are Smooth, the model places greater weight on the small-interval constraint, and does not learn the repetition rule as well as when context melodies are not Smooth, paralleling the human learners. These findings support an account of abstract grammar-induction in which learners rationally assess the statistical evidence for underlying structure based on a generative model of the environment.     

The power of causal beliefs and conflicting evidence on causal judgments and decision making

In two experiments, we investigated the relative impact of causal beliefs and empirical evidence on both decision making and causal judgments, and whether this relative impact could be altered by previous experience. Participants had to decide which of two alternatives would attain a higher outcome on the basis of four cues. After completing the decision task, they were asked to estimate to what extent each cue was a reliable cause of the outcome. Participants were provided with instructions that causally related two of the cues to the outcome, whereas they received neutral information about the other two cues. Two of the four cues—a causal and a neutral cue—had high validity and were both generative. The remaining two cues had low validity, and were generative in Experiment 1, but almost not related to the outcome in Experiment 2. Selected groups of participants in both experiments received pre-training with either causal or neutral cues, or no pre-training was provided. Results revealed that the impact of causal beliefs and empirical evidence depends on both the experienced pre-training and cue validity. When all cues were generative and participants received pre-training with causal cues, they mostly relied on their causal beliefs, whereas they relied on empirical evidence when they received pre-training with neutral cues. In contrast, when some of the cues were almost not related to the outcome, participants’ responses were primarily influenced by validity and—to a lesser extent—by causal beliefs. In either case, however, the influence of causal beliefs was higher in causal judgments than in decision making. While current theoretical approaches in causal learning focus either on the effect of causal beliefs or empirical evidence, the present research shows that both factors are required to explain the flexibility involved in human inferences.     

Perception in action: multiple roles of sensory information in action control

Perceptual processes play a central role in the planning and control of human voluntary action. Indeed, planning an action is a sensorimotor process operating on sensorimotor units, a process that is based on anticipations of perceptual action effects. I discuss how the underlying sensorimotor units emerge, and how they can be employed to tailor action plans to the goals at hand. I also discuss how even a single action can induce sensorimotor binding, how intentionally implemented short-term associations between stimuli and responses become autonomous, how feature overlap between stimulus events and actions makes them compatible, and why action plans are necessarily incomplete.

A Bayesian Theory of Sequential Causal Learning and Abstract Transfer

Two key research issues in the field of causal learning are how people acquire causal knowledge when observing data that are presented sequentially, and the level of abstraction at which learning takes place. Does sequential causal learning solely involve the acquisition of specific cause‐effect links, or do learners also acquire knowledge about abstract causal constraints? Recent empirical studies have revealed that experience with one set of causal cues can dramatically alter subsequent learning and performance with entirely different cues, suggesting that learning involves abstract transfer, and such transfer effects involve sequential presentation of distinct sets of causal cues. It has been demonstrated that pre‐training (or even post‐training) can modulate classic causal learning phenomena such as forward and backward blocking. To account for these effects, we propose a Bayesian theory of sequential causal learning. The theory assumes that humans are able to consider and use several alternative causal generative models, each instantiating a different causal integration rule. Model selection is used to decide which integration rule to use in a given learning environment in order to infer causal knowledge from sequential data. Detailed computer simulations demonstrate that humans rely on the abstract characteristics of outcome variables (e.g., binary vs. continuous) to select a causal integration rule, which in turn alters causal learning in a variety of blocking and overshadowing paradigms. When the nature of the outcome variable is ambiguous, humans select the model that yields the best fit with the recent environment, and then apply it to subsequent learning tasks. Based on sequential patterns of cue‐outcome co‐occurrence, the theory can account for a range of phenomena in sequential causal learning, including various blocking effects, primacy effects in some experimental conditions, and apparently abstract transfer of causal knowledge.     

Preschoolers,adolescents, and adults visually anticipate an agent's efficient action; but only after having observed it frequently

The present study examined the contribution of efficiency reasoning and statistical learning on visual action anticipation in preschool children, adolescents, and adults. To this end, Experiment 1 assessed proactive eye movements of 5-year-old children, 15-year-old adolescents, and adults, who observed an agent stating the intent to reach a goal as quickly as possible. Subsequently the agent could four times either take a short, hence efficient, or long, hence inefficient, path to get to the goal. The results showed that in the first trial participants in none of the age groups predicted above chance level that the agent would produce the efficient action. Instead, we observed an age-dependent increase in action predictions in the subsequent repeated presentation of the same action. Experiment 2 ruled out that participants’ nonconsideration of the efficient path was due to a lack of understanding of the agent's action goal. Moreover, it demonstrated that 5-year-old children do predict that the agent will act efficiently when verbally reasoning about his future action. Overall, the study supports the view that rapid learning from frequency information guides visual action anticipations.