Open Mind最新文献

We investigate the roles of linguistic and sensory experience in the early-produced visual, auditory, and abstract words of congenitally-blind toddlers, deaf toddlers, and typically-sighted/hearing peers. We also assess the role of language access by comparing early word production in children learning English or American Sign Language (ASL) from birth, versus at a delay. Using parental report data on child word production from the MacArthur-Bates Communicative Development Inventory, we found evidence that while children produced words referring to imperceptible referents before age 2, such words were less likely to be produced relative to words with perceptible referents. For instance, blind (vs. sighted) children said fewer highly visual words like "blue" or "see"; deaf signing (vs. hearing) children produced fewer auditory signs like hear. Additionally, in spoken English and ASL, children who received delayed language access were less likely to produce words overall. These results demonstrate and begin to quantify how linguistic and sensory access may influence which words young children produce.

Differences in the processing of subject and object relative clauses have been explained by a combination of syntactic, semantic, and pragmatic factors, such as a general subject advantage based on syntactic constraints, effects of animacy, and the discourse status of relative clause internal subjects. In this paper, we will focus on a factor related to verb meaning, the implicit causality of the verb, which biases the principal causer of the event described by the verb. Depending on whether the bias is on the subject or the object, implicit causality can conflict with the foregrounded antecedent of the relative clause, leading to increased difficulty in comprehension. We tested this hypothesis by manipulating implicit causality in subject and object relative clauses. We used both offline (acceptability judgment task) and online (self-paced reading task) methods to observe at which stage of processing implicit causality influences comprehension. Our findings from acceptability judgments showed that object relative clauses with subject-biased verbs were the least acceptable and the least understood. Conversely, object relative clauses with object-biased verbs were as acceptable and easy to understand as subject relative clauses in French. However, results from self-paced reading indicated that subject-biased verbs were more difficult to process regardless of the construction, suggesting that the integration of implicit causality occurs at a later level of processing, such as in acceptability judgments and comprehension questions. Further acceptability judgment tasks suggested that implicit causality influences relative clause acceptability beyond word order and thematic roles. We propose linking the role of implicit causality with the function of a restrictive relative clause and introduce the Aboutness Hypothesis to explain relative clause processing: a relative clause is more acceptable and easier to understand when everything contributes to making the head its optimal aboutness topic.

A core aim of developmental cognitive science is to uncover the basic building blocks of human thought. For instance, work revealing that even young children, adults without formal education, and distant animal species are sensitive to basic Euclidean properties indicates that humans may be endowed with some primitive understanding of Euclidean geometry. But what about other forms of geometry? Here, we explore children's sensitivity to topological spatial forms. We show that children, like adults, spontaneously distinguish and match items in accordance with their topological relations. As well, we show that children's judgments about object similarity are remarkably consistent with adults', indicating stability in object concepts throughout the lifespan. Finally, we compare children's sensitivity to various topological forms with their sensitivity to geometric properties like curvature, perpendicularity, and symmetry, and find that while there is some variability in performance across all the features tested, overall performance for geometric vs. topological is comparable. Collectively, these findings suggest that even young children have an intuitive understanding of topological relations and suggest that topological relations may be among the building blocks of human visuospatial representation.

Questions may be asked with an intent to acquire new information from the recipient (i.e., information-seeking questions) or with the intent to teach (i.e., pedagogical questions). Understanding how the questions' recipients infer the intent of questions is important, because the recipients' inferences have important consequences for reasoning and learning. In the present series of studies, we tested the hypothesis that i) askers use prosodic cues-an ever-present signal-to encode information-seeking and pedagogical intent both in deliberate and spontaneous speech and that ii) adults and children can draw appropriate inferences about the question's intent on the basis of prosody alone. In Experiments 1 and 2, we found that naïve adult listeners and children aged 5 years and above have the capacity to explicitly identify which asker has an intention to teach on the basis of prosody alone. In Experiment 3, we found that parents' spontaneous speech in pedagogical or information-seeking contexts is appropriately recognized by naïve listeners as pedagogical or information-seeking. Thus, the intent of pedagogical and information-seeking questions is acoustically encoded by askers, and it can be appropriately decoded by recipients.

Humans make rich inferences about the geometry of the visual world. While deep neural networks (DNNs) achieve human-level performance on some psychophysical tasks (e.g., rapid classification of object or scene categories), they often fail in tasks requiring inferences about the underlying shape of objects or scenes. Here, we ask whether and how this gap in 3D shape representation between DNNs and humans can be closed. First, we define the problem space: after generating a stimulus set to evaluate 3D shape inferences using a match-to-sample task, we confirm that standard DNNs are unable to reach human performance. Next, we construct a set of candidate 3D-aware DNNs including 3D neural field (Light Field Network), autoencoder, and convolutional architectures. We investigate the role of the learning objective and dataset by training single-view (the model only sees one viewpoint of an object per training trial) and multi-view (the model is trained to associate multiple viewpoints of each object per training trial) versions of each architecture. When the same object categories appear in the model training and match-to-sample test sets, multi-view DNNs approach human-level performance for 3D shape matching, highlighting the importance of a learning objective that enforces a common representation across viewpoints of the same object. Furthermore, the 3D Light Field Network was the model most similar to humans across all tests, suggesting that building in 3D inductive biases increases human-model alignment. Finally, we explore the generalization performance of multi-view DNNs to out-of-distribution object categories not seen during training. Overall, our work shows that multi-view learning objectives for DNNs are necessary but not sufficient to make similar 3D shape inferences as humans and reveals limitations in capturing human-like shape inferences that may be inherent to DNN modeling approaches. We provide a methodology for understanding human 3D shape perception within a deep learning framework and highlight out-of-domain generalization as the next challenge for learning human-like 3D representations with DNNs.

Owners are often blamed when their property causes harm but might not receive corresponding praise when their property does good. This suggests a double standard of ownership, wherein owning property poses risks for moral blame that are not balanced with equal opportunities for credit. We investigated this possibility in three preregistered experiments on 746 US residents. Participants read vignettes where agentic property (e.g., animals, robots) produced bad or good outcomes, and judged whether owners and the property were morally responsible. With bad outcomes, participants assigned owners more blame than property (Experiments 1 and 2) or similar blame (Experiment 3). But with good outcomes, participants consistently assigned owners much less praise relative to their property. The first two experiments also examined if the double standard arises in two other relationships between authorities and subordinates; participants showed the double standard when assessing moral responsibility for parents and children, but not for employers and employees. Together, these findings point to a novel asymmetry in how owners are assigned responsibility.

Pragmatic inferences are based on assumptions about how speakers communicate: speakers are taken to be cooperative and rational; they consider alternatives and make intentional choices to produce maximally informative utterances. In principle, this analysis applies to linguistic but also non-linguistic communicative actions, but this prediction is typically only tested in children and not in more systematic implicature contexts. We test key implications of this view across six online experiments with American English speaking adults (total N = 231). Experiments 1A and 1B showed that participants made pragmatic inferences based on different types of communicative actions, some being non-linguistic. In Experiment 2, pragmatic inferences were found to be conditional on the speaker's epistemic states. Finally, Experiments 3A to 3C showed that pragmatic inferences were more likely to be made when the communicative action was produced intentionally. Taken together, these results strengthen the view that pragmatics includes social inference about cooperative communication over intentional actions, even non-linguistic actions.

Children and adults alike rely on others to learn about the world, but also need to be able to determine the strength of both their own evidence as well as the evidence that other people provide, particularly when different sources of information disagree. For example, if two informants agree on a belief but share the same evidence, their testimony is statistically dependent on each other, and may be weaker evidence for that belief than two informants who draw on different pieces of evidence to support that belief. Across three experiments (total N = 492), we examine how 4- and 5-year-old children evaluate statistical dependency on a task where they must determine which of two jars that toys were drawn from. A majority of informants, whose testimony could draw from the same evidence or different evidence, always endorsed one jar. Then, children were presented with a dissenting informant or their own personal data that was consistent with the other jar. Children showed no sensitivity to statistical dependency, choosing the majority with equal probability regardless of the independence of their testimony, but also systematically overweighted their own personal data, endorsing the jar consistent with their own evidence more often than would be predicted by an optimal Bayesian model. In contrast, children made choices consistent with this model on a similar task in which the data was presented to children without testimony. Our findings suggest that young children treat majorities as broadly informative, but that the challenges of inferring others' experiences may lead them to rely on concrete, visible evidence when it is available.

Differentiation and Combination theories make different predictions about the order in which information is processed during categorization. Differentiation theory posits that holistic processing of a stimulus occurs before individual features are processed. According to Differentiation theory, overall similarity-based categorization is faster and less effortful compared to unidimensional categorization. In contrast, Combination theory posits that individual features are processed first and that information regarding these features must be combined during multidimensional categorization. According to Combination theory, overall similarity-based categorization is more effortful and takes more time compared to unidimensional categorization. In this study, we trained participants to learn artificial categories using classification learning and observation learning procedures. We used participants' categorization descriptions to determine the number of stimuli dimensions used for categorization. Our results from the first three experiments show that participants who used more dimensions took more time to categorize the transfer stimuli, consistent with Combination theory. In Experiment 4, we tested the hypothesis that using more dimensions takes more time solely due to multiple eye fixations and saccades. In our study, we used visual stimuli with features that do not overlap in space. Our results show that while performing a multidimensional task, participants need more time to recall the feature-category associations learned during the experiment, making the task more effortful, as predicted by Combination theory. Further studies are needed to determine whether Combination theory applies to other types of stimuli, particularly those with spatially non-separable features.

Instrumental helping is one of the paradigmatic "prosocial" behaviors featured in developmental research on sociomoral reasoning, but not much is known about how children recognize instances of helping behaviors or understand the term 'help'. Here, we examined whether young children represent helping as a second-order goal and take it to mean increasing the utility of another agent. In Study 1, we tested whether 12-month-old infants would expect an agent who previously helped to perform an action that reduced the Helpee's action cost. We found that while infants expected agents to act individually efficiently (Experiment 1C), they did not expect the agent to choose the action that maximally reduced the Helpee's cost compared to an action that reduced the cost less (Experiment 1A) or not at all (Experiment 1B). In Study 2, we examined whether three-year-old preschoolers (1) maximize a Helpee's cost reduction when prompted to help in a first-person task, and (2) identify in a third-party context which of two agents, performing superficially similar behaviors with varying effects on the Helpee's action options, actually helped. Contrary to our predictions, preschoolers did not help in a way that maximally reduced the Helpee's cost in (1). In (2), however, they indicated that the agent who reduced the Helpee's action cost was the one who helped. Taken together, these results support the proposal that, at least by preschool age, children possess a second-order utility-based concept of helping, but that they may not exhibit efficiency when choosing their own helping actions.