基于成本最小化信息的社会性意图识别：来自脑电和行为的证据

与已有研究着重考察如何识别客体导向性意图(动作以物理对象为目标, 而不涉及其他人)不同, 本研究对人们如何识别社会性意图(动作以指向社会主体为目标以影响对方的交互行为)进行了探讨。基于两交互主体在整体层面应遵循效用最大化的分析, 提出当A协助B达成目标状态所需要的成本小于B单独实现该目标状态所需要的成本时(简称为成本最小化信息), 其可被识别为具有社会性意图。通过在B面前设置栅栏的方法操纵成本最小化信息, 以指示不同意图类型的脑电μ抑制程度、对不同变化的敏感性(辨别力)为指标, 对该假设进行了检验。结果显示, 相比客体导向性意图的控制条件(即A将目标物苹果放置在石头前), 当A将目标物苹果放置在被栅栏挡住的B前, 其动作可减少B单独获取该苹果的动作成本, 即符合成本最小化条件时, μ的抑制程度更高(实验1), 且对结构改变(某两个动画中充当相同角色的智能体互换)的辨别力更强, 但对角色交换(某个动画中两个智能体的角色交换)的辨别力更弱(实验3a); 而当栅栏不存在时, 虽然A的运动路径与实验1相同, 但A将苹果放置在B前的成本大于B自身获取苹果的成本, 即不符合成本最小化条件, 条件间μ抑制的差异消失(实验2), 且对不同动作模式中变化的辨别相当(实验3b)。鉴于已有研究表明社会性意图所诱发的μ抑制强于客体导向性意图, 且人们对存在社会性意图的两个智能体间的结构改变更容易辨别, 而对角色交换不敏感, 故上述结果揭示, 两个个体的行为是否满足成本最小化影响人们对动作意图的识别, 支持成本最小化信息是社会性意图识别线索的观点。     

Infants’ perception of goal-directed actions: A multi-lab replication reveals that infants anticipate paths and not goals

Influential developmental theories claim that infants rely on goals when visually anticipating actions. A widely noticed study suggested that 11-month-olds anticipate that a hand continues to grasp the same object even when it swapped position with another object (Cannon, E., & Woodward, A. L. (2012). Infants generate goal-based action predictions. Developmental Science, 15, 292–298.). Yet, other studies found such flexible goal-directed anticipations only from later ages on. Given the theoretical relevance of this phenomenon and given these contradicting findings, the current work investigated in two different studies and labs, whether infants indeed flexibly anticipate an action goal. Study 1 (N = 144) investigated by means of five experiments, under which circumstances (e.g., animated agent, human agent) 12-month-olds show flexible goal anticipation abilities. Study 2 (N = 104) presented 11-, 32-month-olds and adults both a human grasping action as well as a non-human action. In none of the experiments did infants flexibly anticipate the action based on the goal, but rather on the movement path, irrespective of the type of agent. Although one experiment contained a direct replication of Cannon and Woodward (2012), we were not able to replicate their findings. Overall our work challenges the view that infants are able to flexibly anticipate action goals from early on, but rather rely on movement patterns when processing other’s actions.     

Will any doll do? 12-month-olds' reasoning about goal objects

Infants as young as 5 months of age view familiar actions such as reaching as goal-directed (Woodward, 1998), but how do they construe the goal of an actor's reach? Six experiments investigated whether 12-month-old infants represent reaching actions as directed to a particular individual object, to a narrowly defined object category (e.g., an orange dump truck), or to a more broadly defined object category (e.g., any truck, vehicle, artifact, or inanimate object). The experiments provide evidence that infants are predisposed to represent reaching actions as directed to categories of objects at least as broad as the basic level, both when the objects represent artifacts (trucks) and when they represent people (dolls). Infants do not use either narrower category information or spatiotemporal information to specify goal objects. Because spatiotemporal information is central to infants' representations of inanimate object motions and interactions, the findings are discussed in relation to the development of object knowledge and action representations.     

Goal saliency boosts infants’ action prediction for human manual actions,but not for mechanical claws

Previous research indicates that infants’ prediction of the goals of observed actions is influenced by own experience with the type of agent performing the action (i.e., human hand vs. non-human agent) as well as by action-relevant features of goal objects (e.g., object size). The present study investigated the combined effects of these factors on 12-month-olds’ action prediction. Infants’ (N = 49) goal-directed gaze shifts were recorded as they observed 14 trials in which either a human hand or a mechanical claw reached for a small goal area (low-saliency goal) or a large goal area (high-saliency goal). Only infants who had observed the human hand reaching for a high-saliency goal fixated the goal object ahead of time, and they rapidly learned to predict the action goal across trials. By contrast, infants in all other conditions did not track the observed action in a predictive manner, and their gaze shifts to the action goal did not change systematically across trials. Thus, high-saliency goals seem to boost infants’ predictive gaze shifts during the observation of human manual actions, but not of actions performed by a mechanical device. This supports the assumption that infants’ action predictions are based on interactive effects of action-relevant object features (e.g., size) and own action experience.     

Eighteen-month-olds integrate verbal cues into their action processing: Evidence from ERPs and mu power

Behavioral research has shown that infants use both behavioral cues and verbal cues when processing the goals of others’ actions. For instance, 18-month-olds selectively imitate an observed goal-directed action depending on its (in)congruence with a model’s previous verbal announcement of a desired action goal. This EEG-study analyzed the electrophysiological underpinnings of these behavioral findings on the two functional levels of conceptual action processing and motor activation. Mid-latency mean negative ERP amplitude and mu-frequency band power were analyzed while 18-month-olds (N = 38) watched videos of an adult who performed one out of two potential actions on a novel object. In a within-subjects design, the action demonstration was preceded by either a congruent or an incongruent verbally announced action goal (e.g., “up” or “down” and upward movement). Overall, ERP negativity did not differ between conditions, but a closer inspection revealed that in two subgroups, about half of the infants showed a broadly distributed increased mid-latency ERP negativity (indicating enhanced conceptual action processing) for either the congruent or the incongruent stimuli, respectively. As expected, mu power at sensorimotor sites was reduced (indicating enhanced motor activation) for congruent relative to incongruent stimuli in the entire sample. Both EEG correlates were related to infants’ language skills. Hence, 18-month-olds integrate action-goal-related verbal cues into their processing of others’ actions, at the functional levels of both conceptual processing and motor activation. Further, cue integration when inferring others’ action goals is related to infants’ language proficiency.     

Emergent Goal-Anticipatory Gaze in Infants via Event-Predictive Learning and Inference

From about 7 months of age onward, infants start to reliably fixate the goal of an observed action, such as a grasp, before the action is complete. The available research has identified a variety of factors that influence such goal-anticipatory gaze shifts, including the experience with the shown action events and familiarity with the observed agents. However, the underlying cognitive processes are still heavily debated. We propose that our minds (i) tend to structure sensorimotor dynamics into probabilistic, generative event-predictive, and event boundary predictive models, and, meanwhile, (ii) choose actions with the objective to minimize predicted uncertainty. We implement this proposition by means of event-predictive learning and active inference. The implemented learning mechanism induces an inductive, event-predictive bias, thus developing schematic encodings of experienced events and event boundaries. The implemented active inference principle chooses actions by aiming at minimizing expected future uncertainty. We train our system on multiple object-manipulation events. As a result, the generation of goal-anticipatory gaze shifts emerges while learning about object manipulations: the model starts fixating the inferred goal already at the start of an observed event after having sampled some experience with possible events and when a familiar agent (i.e., a hand) is involved. Meanwhile, the model keeps reactively tracking an unfamiliar agent (i.e., a mechanical claw) that is performing the same movement. We qualitatively compare these modeling results to behavioral data of infants and conclude that event-predictive learning combined with active inference may be critical for eliciting goal-anticipatory gaze behavior in infants.     

Can unpredicted outcomes be intended? The role of outcome-beliefs in children's judgments of intention

An adult-like concept of intention includes a deliberate action to achieve a goal and a belief that one's action (if successful) will cause the desired outcome. For example, good outcomes caused by accident or by chance are not believed to be caused intentionally. In two experiments, we asked whether children understand this connection between intentions and outcomes. Children played two games in which actions could produce unintended outcomes (i.e., causes were unplanned). Children sometimes received a desirable reward independent of intention. In Experiment 1, 4- and 5-year-olds mistakenly claimed they had intended the desirable outcome even when it was unexpected. Four-year-olds judged that they had not intended a deliberate action if it did not yield a rewarding outcome. Experiment 2 demonstrates that 6-year-olds seldom make these errors. The results suggest that 4- and 5-year-old children have not yet attained an adult-like concept of intention. Their inaccurate judgments regarding their intentions, given a rewarding yet unexpected outcome, can be explained by a positivity bias.     

Action outcome probability influences the size of the head-fake effect in basketball

Both kinematic and contextual information (e.g., action outcome probability) play a significant role in action anticipation. However, few researchers have examined the reciprocal influence of the two types of information and fewer still have investigated this issue for deceptive actions in sports. In the present study, we investigate the impact of action outcome probability on the processing of deceptive kinematic cues for the head fake in basketball. We manipulated the probability of the action outcome to either pass the ball to the left or to the right side (i.e., 75%, 50%, 25%) and examined how this contextual information affected the influence of head orientation on pass direction judgments. Outcome probability information was either provided explicitly (Experiment 1) or implicitly (Experiment 2). Both experiments indicated an increased head-fake effect with increasing outcome probability. Moreover, the bias to respond in line with the player’s head direction increased linearly with outcome probability. Also, discriminability between deceptive and genuine actions was poorer for high outcome probability (75%) associated with head orientation than for the 25% and 50% values. Last, a stronger response bias toward the higher probability side for deceptive trials than for genuine trials was only significant when the outcome probability information was processed implicitly in Experiment 2. The results of this study fit well with recent literature on contextual information in action prediction and are discussed in light of confirmation bias and signal detection theory.     

Infants' understanding of object-directed action

When and in what ways do infants recognize humans as intentional actors? An important aspect of this larger question concerns when infants recognize specific human actions (e.g. a reach) as object-directed (i.e. as acting toward goal-objects). In two studies using a visual habituation technique, 12-month-old infants were tested to assess their recognition that an adult's reach is directed toward its target object. Infants in the experimental condition were habituated to a display in which an actor reached over a wall-like barrier with an arcing arm movement, to pick up a ball. After habituation infants saw two test displays, for which the barrier was removed. In the direct test event the actor reached directly for the ball, the arm tracing a visually new path, but the action consistent with attempting to reach for the object as directly as possible. In the indirect test event the actor traced the old path, reaching over in an arc, even though the wall was no longer present. This arm movement was identical to that in habituation but no longer displayed a reach going directly to its object. In a control condition infants saw the same movements but in a situation with no goal-object. In the experimental conditions, with a goal object present, infants looked longer at the indirect test event in comparison to the direct test event. In the control conditions infants looked equally at both indirect and direct test events. We conclude that sensitivity to human object-directed action is established by 12-month-olds and compare these results to recent findings by [Gergely, G., Nadasdy, Z., Csibra, G., & Biro S. (1995). Taking the intentional stance at 12 months of age. Cognition, 56, 165-193] and [Woodward, A. (1998). Infants selectively encode the goal object of an actor's reach. Cognition, 69, 1-34].     

Executive functioning by 18‐24‐month‐old children: effects of inhibition,working memory demands and narrative in a novel detour‐reaching task

Infants can inhibit a prepotent but wrong action towards a goal in order to perform a causal means‐action. It is not clear, however, whether infants can perform an arbitrary means‐action while inhibiting a prepotent response. In four experiments, we explore this executive functioning in 18–24‐month‐old children. The working memory and inhibition demands in a novel means‐end problem were systematically varied in terms of the type and combination of means‐action(s) (causal or arbitrary) contained within the task, the number of means‐actions (1 or 2), the goal visual availability and whether the task was accompanied by a narrative. Experiments 1 and 2 showed that children performed tasks that contained causal as opposed to arbitrary information more accurately; accuracy was also higher in tasks containing only one step. Experiment 2 also demonstrated that performance in the arbitrary task improved significantly when all sources of prepotency were removed. In Experiment 3, task performance improved when the two means‐actions were intelligibly linked to the task goal. Experiment 4 demonstrated that the use of a narrative that provided a meaningful (non‐causal) link between the two means‐actions also improved children's performance by assisting their working memory in the generation of a rationale. Findings provide an initial account of executive functioning in the months that bring the end of infancy. Copyright © 2006 John Wiley & Sons, Ltd.     

动作理解因境而异:动作加工中情境信息的自动整合

针对动作理解的机制,模拟论主张大脑自发模拟他人的动作,就相同的动作其理解也相同,而理论论则认为人们基于合理性原则对他人动作进行推理,相同的动作发生在不同的情境时会有不同的理解。但以往研究所采用动作材料的运动学特性和发生情境存在共变,其难以区分动作理解是支持模拟论还是理论论。通过两项实验,采用动画制作技术来产生有无约束情境下的追逐动作,以指示动作加工过程的脑电μ抑制为指标,对前述两种观点进行了检验。其中,在约束情境中存在障碍物,追逐者需改变运动方向以绕过障碍物,从后方逐渐趋近目标;而无约束情境中不存在障碍物,但追逐者依然保持与存在约束情景下相同的运动模式。结果发现,当追逐动作发生在存在约束的情境时,其基于合理性原则推测可获得清晰的动作目标,该条件下的μ抑制程度高于不存在约束情境的条件(实验1);而当仅追逐者运动,即趋近的目标不确定时,虽然有约束和无约束情境间的物理差异与实验1相同,但条件间μ抑制的差异消失(实验2);且上述μ抑制并非与注意相关的枕叶α活动的泛化。该结果提示,动作发生的情境信息影响人们对动作的理解,即基于推理过程理解动作,支持理论论观点。     

General processes, rather than "goals," explain imitation errors

The goal-directed theory of imitation (GOADI) states that copying of action outcomes (e.g., turning a light switch) takes priority over imitation of the means by which those outcomes are achieved (e.g., choice of effector or grip). The object < effector < grip error pattern in the pen-and-cups task provides strong support for GOADI. Experiment 1 replicated this effect using video stimuli. Experiment 2 showed that shifting the color cue from objects to effectors makes imitation of effector selection more accurate than imitation of object and grip selection. Experiment 3 replicated this result when participants were required to describe actions. Experiment 4 indicated that, when participants are imitating and describing actions, enhancing grip discriminability makes grip selection the most accurately executed component of the task. Consistent with theories that hypothesize that imitation relies on task-general mechanisms (e.g., the associative sequence learning model, ideomotor theory), these findings suggest that imitation is no more or less goal directed than other tasks involving action observation.     

What makes an event: Temporal integration of stimuli or actions?

In this article, we ask what serves as the “glue” that temporarily links information to form an event in an active observer. We examined whether forming a single action event in an active observer is contingent on the temporal presentation of the stimuli (hence, on the temporal availability of the action information associated with these stimuli), or on the learned temporal execution of the actions associated with the stimuli, or on both. A partial-repetition paradigm was used to assess the boundaries of an event for which the temporal properties of the stimuli (i.e., presented either simultaneously or temporally separate) and the intended execution of the actions associated with these stimuli (i.e., executed as one, temporally integrated, response or as two temporally separate responses) were manipulated. The results showed that the temporal features of action execution determined whether one or more events were constructed; the temporal presentation of the stimuli (and hence the availability of their associated actions) did not. This suggests that the action representation, or “task goal,” served as the “glue” in forming an event in an active observer. These findings emphasize the importance of action planning in event construction in an active observer.     

Preverbal infants identify emotional reactions that are incongruent with goal outcomes

Identifying the goal of another agent’s action allows an observer to make inferences not only about the outcomes the agent will pursue in the future and the means to be deployed in a given context, but also about the emotional consequences of goal-related outcomes. While numerous studies have characterized the former abilities in infancy, expectations about emotions have gone relatively unexplored. Using a violation of expectation paradigm, we present infants with an agent who attains or fails to attain a demonstrated goal, and reacts with positive or negative affect. Across several studies, we find that infants’ attention to a given emotional display differs depending on whether that reaction is congruent with the preceding goal outcome. Specifically, infants look longer at a negative emotional display when it follows a completed goal compared to when it follows a failed goal. The present results suggest that infants’ goal representations support expectations not only about future actions but also about emotional reactions, and that infants in the first year of life can relate different emotional reactions to conditions that elicit them.     

Ten‐month‐old infants use prior information to identify an actor's goal

For adults, prior information about an individual's likely goals, preferences or dispositions plays a powerful role in interpreting ambiguous behavior and predicting and interpreting behavior in novel contexts. Across two studies, we investigated whether 10‐month‐old infants’ ability to identify the goal of an ambiguous action sequence was facilitated by seeing prior instances in which the actor directly pursued and obtained her goal, and whether infants could use this prior information to understand the actor's behavior in a new context. Experiment 1 demonstrated that the goal preview impacted infants’ subsequent action understanding, but only if the preview was delivered in the same room as the subsequent action sequence. Experiment 2 demonstrated that infants’ failure to transfer prior goal information across situations arose from a change in the room per se and not other features of the task. Our results suggest that infants may use their understanding of simple actions as a leverage point for understanding novel or ambiguous actions, but that their ability to do so is limited to certain types of contextual changes.     

Sensitivity to communicative relevance tells young children what to imitate

How do children decide which elements of an action demonstration are important to reproduce in the context of an imitation game? We tested whether selective imitation of a demonstrator’s actions may be based on the same search for relevance that drives adult interpretation of ostensive communication. Three groups of 18‐month‐old infants were shown a toy animal either hopping or sliding (action style) into a toy house (action outcome), but the communicative relevance of the action style differed depending on the group. For the no prior information group, all the information in the demonstration was new and so equally relevant. However, for infants in the ostensive prior information group, the potential action outcome was already communicated to the infant prior to the main demonstration, rendering the action style more relevant. Infants in the ostensive prior information group imitated the action style significantly more than infants in the no prior information group, suggesting that the relevance manipulation modulated their interpretation of the action demonstration. A further condition (non‐ostensive prior information) confirmed that this sensitivity to new information is only present when the ‘old’ information had been communicated, and not when infants discovered this information for themselves. These results indicate that, like adults, human infants expect communication to contain relevant content, and imitate action elements that, relative to their current knowledge state or to the common ground with the demonstrator, is identified as most relevant.     

Evaluations of pro-environmental behaviors by 7-month-old infants

Environmental morality is the foundation of a sustainable future, yet its ontogenetic origin remains unknown. In the present study, we asked whether 7-month-olds have a sense of ‘environmental morality’. Infants’ evaluations of two pro-environmental actions were assessed in both visual and reaching preferential tasks. In Experiment 1, the overt behavior of protecting (i.e., collecting artificial objects spread on a lawn) was compared with the action of harming the environment (i.e., by disregarding the objects). In Experiment 2, the covert behavior of protecting the environment (i.e., maintaining artificial objects inside a container) was compared with the action of harming the environment (i.e., littering the artificial objects on a lawn). The results showed infants’ reaching preference for the agent who performed overt pro-environmental actions (Experiment 1), and no preference for the agent who performed covert pro-environmental actions (Experiment 2). These findings reveal a rudimentary ecological sense and suggest that infants require different abilities to evaluate overt impact-oriented and covert intend-oriented pro-environmental behaviors.     

You Can't Always Get What You Want: Infants Understand Failed Goal-Directed Actions

ABSTRACT— At what age do infants understand that goals exist independently of the actions that result from them? Exploring infants' understanding of failed intentional actions—when the goal of the action is unfulfilled and thus not apparent in the actor's movements—is a critical step in answering this question. Using a visual habituation paradigm, we assessed when infants understand that a failed intentional action is goal directed and whether an understanding of successful intentional actions (actions that do overtly attain their goals) precedes an understanding of failed intentional actions. Results demonstrated that 10- and 12-month-olds recognized the goal directedness of both successful and failed reaching actions. Eight-month-olds also recognized the goal directedness of successful actions, but not of unsuccessful attempts. Thus, by the end of the 1st year of life, infants possess an impressive understanding of intentional action, and an understanding of failed intentional actions follows an earlier understanding of successful ones.     

Implementation intentions do not enhance all types of goals: The moderating role of goal difficulty

Three studies tested whether Gollwitzer and Brandst?tter’s (1997) failure to find an implementation effect for easy goals was due to a ceiling effect, to the moderating effect of previously formed habits, or to a moderating effect of earlier implementation intentions. The studies strongly indicated that easy goals did benefit from forming implementation intentions (i.e., specifying where or when one would perform the action). This suggests that Gollwitzer and Brandst?tter’s failure to find significant implementation effects for easy goals was due to a ceiling effect and not to other moderating effects. However, in the three experiments, we found no positive effect of implementation intentions for the enactment of goal-related behavior corresponding to a certain type of difficult goal. More specifically, when the focus was on the outcome of goal-directed action rather than on the goal-directed actions themselves, implementation intentions specifying when or in what conditions the relevant actions were to be performed did not enhance enactment. When the focus was on the goal-directed actions, we replicated the positive effect of forming implementation intentions. We argue that specifying when or where a goal-directed action should be enacted does not enhance enactment when the actor is not aware of the actions that are required to reach the goal. Possibly, implementation intentions specifying what one should do (rather than where or when) might be more helpful to enhance enactment rates of this type of goal.