ArticleLiterature Review

Becoming an intentional agent: Introduction to the special issue

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... After their first birthday, infants begin to understand that an actor may consider various plans to pursue a goal, and choose one to intentionally enact based on environmental reality [307]. Eighteen-month-old children are able to both infer and imitate the intended goal of an action even if the action repeatedly fails to achieve the goal [308]. Moreover, infants can imitate actions in a rational, efficient way based on an evaluation of the action's situational constraints instead of merely copying movements, indicating that infants have a deep understanding of relationships among the environment, action, and underlying intent [309]. ...
... Simulation theory claims that the mechanism underlying the attribution of intentions to actions might rely on simulating the observed action and mapping it onto our own experiences and intent representations [328]; and that such simulation processes are at the heart of the development of intentional action interpretation [308]. In order to understand others' intentions, humans subconsciously empathize with the person they are observing and estimate what their own actions and intentions might be in that situation. ...
Preprint
Full-text available
Recent progress in deep learning is essentially based on a "big data for small tasks" paradigm, under which massive amounts of data are used to train a classifier for a single narrow task. In this paper, we call for a shift that flips this paradigm upside down. Specifically, we propose a "small data for big tasks" paradigm, wherein a single artificial intelligence (AI) system is challenged to develop "common sense", enabling it to solve a wide range of tasks with little training data. We illustrate the potential power of this new paradigm by reviewing models of common sense that synthesize recent breakthroughs in both machine and human vision. We identify functionality, physics, intent, causality, and utility (FPICU) as the five core domains of cognitive AI with humanlike common sense. When taken as a unified concept, FPICU is concerned with the questions of "why" and "how", beyond the dominant "what" and "where" framework for understanding vision. They are invisible in terms of pixels but nevertheless drive the creation, maintenance, and development of visual scenes. We therefore coin them the "dark matter" of vision. Just as our universe cannot be understood by merely studying observable matter, we argue that vision cannot be understood without studying FPICU. We demonstrate the power of this perspective to develop cognitive AI systems with humanlike common sense by showing how to observe and apply FPICU with little training data to solve a wide range of challenging tasks, including tool use, planning, utility inference, and social learning. In summary, we argue that the next generation of AI must embrace "dark" humanlike common sense for solving novel tasks.
... After their first birthday, infants begin to understand that an actor may consider various plans to pursue a goal, and choose one to intentionally enact based on environmental reality [305]. Eighteen-month-old children are able to both infer and imitate the intended goal of an action even if the action repeatedly fails to achieve the goal [306]. Moreover, infants can imitate actions in a rational, efficient way based on an evaluation of the action's situational constraints instead of merely copying movements, indicating that infants have a deep understanding of relationships among the environment, action, and underlying intent [307]. ...
... Simulation theory claims that the mechanism underlying the attribution of intentions to actions might rely on simulating the observed action and mapping it onto our own experiences and intent representations [326]; and that such simulation processes are at the heart of the development of intentional action interpretation [306]. In order to understand others' intentions, humans subconsciously empathize with the person they are observing and estimate what their own actions and intentions might be in that situation. ...
Article
Full-text available
Recent progress in deep learning is essentially based on a “big data for small tasks” paradigm, under which massive amounts of data are used to train a classifier for a single narrow task. In this paper, we call for a shift that flips this paradigm upside down. Specifically, we propose a “small data for big tasks” paradigm, wherein a single artificial intelligence (AI) system is challenged to develop “common sense,” enabling it to solve a wide range of tasks with little training data. We illustrate the potential power of this new paradigm by reviewing models of common sense that synthesize recent breakthroughs in both machine and human vision. We identify functionality, physics, intent, causality, and utility (FPICU) as the five core domains of cognitive AI with humanlike common sense. When taken as a unified concept, FPICU is concerned with the questions of “why” and “how,” beyond the dominant “what” and “where” framework for understanding vision. They are invisible in terms of pixels but nevertheless drive the creation, maintenance, and development of visual scenes. We therefore coin them the “dark matter” of vision. Just as our universe cannot be understood by merely studying observable matter, we argue that vision cannot be understood without studying FPICU. We demonstrate the power of this perspective to develop cognitive AI systems with humanlike common sense by showing how to observe and apply FPICU with little training data to solve a wide range of challenging tasks, including tool use, planning, utility inference, and social learning. In summary, we argue that the next generation of AI must embrace “dark” humanlike common sense for solving novel tasks.
... The foundation for understanding preferences involves the appreciation that human behaviors are often goal or intention directed (Johnson, 2000;Sommerville et al., 2005;Biro and Hommel, 2007;Csibra and Gergely, 2007). Research evidence reveals that infants tend to interpret people's actions such as gazes and reaches as reflections of their underlying intentions (Meltzoff, 1995;Woodward, 1998;Woodward and Sommerville, 2000;Woodward and Guajardo, 2002;Woodward, 2003;Behne et al., 2005;Sommerville et al., 2005;Johnson et al., 2007;Woodward et al., 2009;Luo and Baillargeon, 2010). ...
... Taken together, past research indicates that infants appreciate the goal-directed nature of human behaviors and understand that some of these behaviors can reflect people's preferences (Repacholi and Gopnik, 1997;Woodward, 1998;Buresh and Woodward, 2007;Biro and Hommel, 2007;Luo and Johnson, 2009;Luo, 2011). They also understand that one person's preference does not necessarily generalize to another individual but can be generalized to other objects (Buresh and Woodward, 2007;Luo and Beck, 2010;Novack et al., 2013). ...
Article
Full-text available
Human preferences are person specific because different individuals do not necessarily have the same preference. Although existing empirical evidence demonstrates that infants have a basic understanding about people’s preferences, there remained one question as to whether infants appreciate that a person’s preference can be generalized to objects that belong to the same kind. This study addressed this gap with 13-month-old Chinese infants. In Experiment 1, infants were first habituated to an actor preferring a target object over a different shaped distractor object. Next, the objects’ position and colors were changed and infants watched the actor from habituation and a new actor each alternated preference between the two objects. Results revealed that infants looked longer at the event when the old actor preferred the distractor object. Throughout the experiment, infants were also presented with additional tests in which each actor requested them to visually identify his or her preferred object. Results showed that infants identified the different colored target object for the old actor but not for the new actor. Experiment 2 removed the additional tests and replicated the results of Experiment 1. Experiment 3 confirmed that 13-month olds could differentiate the two different colored target objects. Together, the present findings provide the first known evidence that 13-month olds expect object preferences to be generalized across objects of like kind but not necessarily across individuals. Such sensitivity to the rules guiding preference generalization could help infants predict people’ behaviors and facilitate more successful social interactions.
... A new potential class of applications is also emerging, which may be described as Mandated Intentional Agents (MIAs). The concept of an Intentional Agent, acting according to its own beliefs, desires and intentions in the world, is an established concept in cognitive science [12]. An MIA is a machine-based agent which represents the beliefs, desires and intentions of a human client in the real world. ...
Chapter
Full-text available
This paper extends a novel technique for the classification of sentences as Dialogue Acts, based on structural information contained in function words. Initial experiments on classifying questions in the presence of a mix of straightforward and “difficult” non-questions yielded promising results, with classification accuracy approaching 90%. However, this initial dataset does not fully represent the various permutations of natural language in which sentences may occur. Also, a higher Classification Accuracy is desirable for real-world applications. Following an analysis of categorisation of sentences, we present a series of experiments that show improved performance over the initial experiment and promising performance for categorising more complex combinations in the future.
... Subsequently, when the object is instead hidden at location B, infants tend still to reach to location A. This error is attributable at least partially to a deficit in the ability to inhibit previously rewarded action (Diamond, Cruttenden, & Neiderman, 1994; Marcovitch & Zelazo, 2009), as evidenced, for example, by the fact that infants are more likely to look at the correct location than they are to reach to it (Hofstadter & Reznick, 1996). Nevertheless, Rochat and Striano's (2000) description of infants as agents deliberately exploiting instrumental contingencies is not without some foundation: much evidence has been put forward to argue that infants do expect outcomes instrumentally contingent on their actions, and the available evidence is often interpreted as demonstrating that even young infants behave -4 intentionally (e.g. Biro & Hommel, 2007; but see Haith, 1998). Many arguments can be discounted because evidence is interpreted without taking into account the possibility of reinforcement learning leading to habit acquisition (e.g. ...
Article
It is known that young infants can learn to perform an action that elicits a reinforcer, and that they can visually anticipate a predictable stimulus by looking at its location before it begins. Here, in an investigation of the display of these abilities in tandem, I report that 10-month-olds anticipate a reward stimulus that they generate through their own action: .5 sec before pushing a button to start a video reward, they increase their rate of gaze shifts to the reward location; and during periods of extinction, reward location gaze shifts correlate with bouts of button pushing. The results are consistent with the hypothesis that the infants have an expectation of the outcome of their actions: several alternative hypotheses are ruled out by yoked controls. Such an expectation may, however, be procedural, have minimal content, and is not necessarily sufficient to motivate action.
... It is widely accepted that infants' actions are goal-directed from an early age (Biro & Hommel, 2007). Most of the evidence cited in support of such claims demonstrates goaldirectedness as we have defined it here in terms of motor organisation (e.g. ...
Article
Full-text available
The term goal directed conventionally refers to either of 2 separate process types-motor processes organizing action oriented toward physical targets and decision-making processes that select these targets by integrating desire for and knowledge of action outcomes. Even newborns are goal directed in the first sense, but the status of infants as decision makers (the focus here) is unknown. In this study, 24-month-olds learned to retrieve an object from a box by pressing a button, and then the object's value was increased. After the object's subsequent disappearance, these children were more likely to press the button to try to retrieve the object than were control 24-month-olds who had learned to retrieve the object but for whom the object's value was unchanged. Such sensitivity to outcome value when selecting actions is a hallmark of decision making. However, 14- and 19-month-olds showed no such sensitivity. Possible explanations include that they had not learned the specifics of the action outcome; they had not acquired the necessary desire; or they had acquired both but did not integrate them to make a decision.
... Understanding actions in terms of the goals they are designed to achieve is a fundamental human faculty that emerges early in development (Biro and Hommel, 2007;Csibra and Gergely, 2007). How do infants decide whether a certain behavior is an ''action'' worthy of goal attribution? ...
Article
Infants show very early sensitivity to a variety of behavioral cues (such as self-propulsion, equifinal movement, free variability, and situational adjustment of behavior) that can be exploited when identifying, predicting, and interpreting goal-directed actions of intentional agents. We compare and contrast recent alternative models concerning the role that different types of behavioral cues play in human infants' early understanding of animacy, agency, and intentional action. We present new experimental evidence from violation of expectation studies to evaluate these alternative models on the nature of early development of understanding goal-directedness by human infants. Our results support the view that, while infants initially do not restrict goal attribution to behaviors of agents exhibiting self-propelled motion, they quickly develop such expectations.
Thesis
Full-text available
Au regard de la place croissante occupée par l'intelligence artificielle dans nos sociétés, la nécessité d'affiner notre compréhension de la notion d'apprentissage semble plus importante que jamais. S'intégrant dans une telle optique, cette thèse de doctorat cherche à mettre en lumière les mécanismes d'apprentissage permettant au bébé d'acquérir au cours de la première année de vie une utilisation appropriée et différenciée de son corps lui permettant d'interagir de manière efficace avec son environnement physique et social, ce que nous désignons ici par le terme de "savoir-fairecorporel". Le postulat au cœur de ce travail de recherche est le suivant : l'acquisition progressive du savoir-faire corporel au cours de la vie fœtale et des premiers mois de vie post-partum est sous-tendue par deux mécanismes d'apprentissage, l'exploitation de la sensibilité aux contingencessensorimotrices et la motivation intrinsèque. Ce travail de thèse explore la première partie de ce postulat, c'est-à-dire le rôle jouée par la sensibilité aux contingences sensorimotrices dans le développement du savoir-faire corporel. Afin d'investiguer cette hypothèse, ce travail de thèsese concentre dans un premier temps sur l'analyse critique des données expérimentales déjà existantes sur le sujet, à la fois en psychologie du développement et en robotique développementale. Dans un second temps, cette recherche vise à approfondir notre compréhension de l'acquisition du savoir-faire corporel à travers l'expérimentation chez le bébé âgé de moins d'un an.
Chapter
It is likely that in AI, Robotics, Neuroscience and Cognitive Sciences, what we need is an integrated approach putting together concepts and methods from fields so far considered well distinct like non linear dynamics, information, computation and control theory as well as general AI, psychology, cognitive sciences in general, neurosciences and system biology. These disciplines usually share many problems, but have very different languages and experimental methodologies. It is thought that while tackling with many serious ‘hard core’ scientific issues it is imperative, probably a necessary (pre) requisite, that we do serious efforts to clarify and merge the underlying paradigms, the proper methodologies, the metrics and success criteria of this new branch of science. Many of these questions have already been approached by philosophy, but they acquire in this context a scientific nature: e.g.: Is it possible cognition without consciousness? And without ‘sentience’? In the context of AI and neuroscience research various definition of consciousness have been proposed (for example by Tononi, [44], to quote an example liked by the author). How they relate to the previous and contemporary philosophical analysis? Sometimes scientists may look as poor philosophers, and the opposite: philosophers may look as poor scientists, yet, the critical passages of history of science during a paradigm change or the birth of a new discipline have often involved a highly critical conceptual analysis intertwined with scientific and mathematical advancements. The scientific enterprise is now somehow close to unbundle the basic foundation of our consciousness and of our apperception of reality, and, it is clear that there are some circularity issues with the possible ‘explanations’, at least.
Conference Paper
This paper proposes and discusses a modeling framework for embodied anticipatory behavior systems. This conceptual and theoretical framework is quite general and aims to be a, quite preliminary, step towards a general theory of cognitive adaptation to the environment of natural intelligent systems and to provide a possible approach to develop new more autonomous artificial systems. The main purpose of this discussion outline is to identify at least a few of the issues we have to cope with, and some of the possible methods to be used, if we aim to understand from a rigorous standpoint the dynamics of embodied adaptive learning systems both natural and artificial.
Article
Full-text available
This chapter presents and extends a model of human socio-emotional development that we have recently proposed (Gergely & Watson, 1996). Our model gains its uniqueness by the fact that even though it embraces the common assumption that young infants are sensitive to contingency experience, at the same time, it rejects the general view that they are initially perceptually aware of their specific basic emotion states. Indeed, it is our contention that contingency detection is crucially involved in an infant's progressively developing awareness of his or her internal affective states. More specifically, our "social-biofeedback model" holds that the caregiver's contingent reflections of the infant's emotion expressive dis- plays play a central causal role in the development of emotional self-aware- ness and control that is mediated by the contingency detection module. We begin by trying to make very clear what we mean by contingency perception, contingency seeking, and its special limitation to perceivable contin- gencies, because we shall place considerable theoretical weight on these foundational constructs. We then consider the implications of this view of early contingency perception when conjoined with an assumption that an infant begins life with little or no awareness of his or her dispositional states. That leads us to our social-biofeedback model of how the infant progressively becomes aware of his or her emotional dispositions through the process previously identified as social mirroring. Our model includes an assumption about a change in the target magnitude of contingency seeking that appears to occur at about 3 months of age. The possible relevance of this for the understanding of the deviant developmental pat- tern in autism is also briefly considered.
Article
Full-text available
Babies may opt for a simpler way to turn on a light after watching an adult do it. Here we show that if an adult demonstrates a new way to execute a task to a group of infants aged 14 months, the children will use this action to achieve the same goal only if they consider it to be the most rational alternative. Our results indicate that imitation of goal-directed action by preverbal infants is a selective, interpretative process, rather than a simple re-enactment of the means used by a demonstrator, as was previously thought1, 2, 3.
Article
Full-text available
According to the authors' 2-phase model of action control, people first incidentally acquire bidirectional associations between motor patterns and movement-contingent events and then intentionally use these associations for goal-directed action. The authors tested the model in 4 experiments, each comprising an acquisition phase, in which participants experienced co-occurrences between left and right keypresses and low- and high-pitched tones, and a test phase, in which the tones preceded the responses in forced- and free-choice designs. Both reaction time and response frequency in the test phase depended on the learned associations, indicating that presenting a tone activated the associated response. Results are interpreted as evidence for automatic action–outcome integration and automatic response priming through learned action effects. These processes may be basic for the control of voluntary action by the anticipation of action goals.
Article
Full-text available
Three longitudinal studies are reported in which 6-8-month-old infants were tested on means-end problems involving pulling a cloth to retrieve a toy. Production of intentional means-end behavior increased between 6 and 7 months, but although 6-month-olds' behavior was unaffected by the presence or absence of a toy on the cloth, 7-month-olds more often produced means-end sequences when a toy could be retrieved. Infants' performance remained the same when the cloth was either attached to the toy or separate, suggesting that goal-subgoal conflict does not interfere with performance of means-end sequences. By 8 months, infants could appropriately adjust their means-end behavior to the distance of the toy. These results confirm J. Piaget's (1953) original description of a shift from transitional to intentional means-end behavior and suggest that development of means-end behavior involves acquisition of knowledge of appropriate means-end relations.
Article
Full-text available
Here we show that if an adult demonstrates a new way to execute a task to a group of infants aged 14 months, the children will use this action to achieve the same goal only if they consider it to be the most rational alternative. Our results indicate that imitation of goal-directed action by preverbal infants is a selective, interpretative process, rather than a simple re-enactment of the means used by a demonstrator, as was previously thought.
Article
Full-text available
Traditional approaches to human information processing tend to deal with perception and action planning in isolation, so that an adequate account of the perception-action interface is still missing. On the perceptual side, the dominant cognitive view largely underestimates, and thus fails to account for, the impact of action-related processes on both the processing of perceptual information and on perceptual learning. On the action side, most approaches conceive of action planning as a mere continuation of stimulus processing, thus failing to account for the goal-directedness of even the simplest reaction in an experimental task. We propose a new framework for a more adequate theoretical treatment of perception and action planning, in which perceptual contents and action plans are coded in a common representational medium by feature codes with distal reference. Perceived events (perceptions) and to-be-produced events (actions) are equally represented by integrated, task-tuned networks of feature codes--cognitive structures we call event codes. We give an overview of evidence from a wide variety of empirical domains, such as spatial stimulus-response compatibility, sensorimotor synchronization, and ideomotor action, showing that our main assumptions are well supported by the data.
Article
Full-text available
For more than two decades, researchers have argued that young children do not understand mental states such as beliefs. Part of the evidence for this claim comes from preschoolers' failure at verbal tasks that require the understanding that others may hold false beliefs. Here, we used a novel nonverbal task to examine 15-month-old infants' ability to predict an actor's behavior on the basis of her true or false belief about a toy's hiding place. Results were positive, supporting the view that, from a young age, children appeal to mental states--goals, perceptions, and beliefs--to explain the behavior of others.
Article
Full-text available
Some researchers have suggested that infants' ability to reason about goals develops as a result of their experiences with human agents and is then gradually extended to other agents. Other researchers have proposed that goal attribution is rooted in a specialized system of reasoning that is activated whenever infants encounter entities with appropriate features (e.g., self-propulsion). The first view predicts that young infants should attribute goals to human but not other agents; the second view predicts that young infants should attribute goals to both human and nonhuman agents. The present research revealed that 5-month-old infants (the youngest found thus far to attribute goals to human agents) also attribute goals to nonhuman agents. In two experiments, infants interpreted the actions of a self-propelled box as goal-directed. These results provide support for the view that from an early age, infants attribute goals to any entity they identify as an agent.
Article
Full-text available
Humans show a strong and early inclination to interpret observed behaviours of others as goal-directed actions. We identify two main epistemic functions that this 'teleological obsession' serves: on-line prediction and social learning. We show how teleological action interpretations can serve these functions by drawing on two kinds of inference ('action-to-goal' or 'goal-to-action'), and argue that both types of teleological inference constitute inverse problems that can only be solved by further assumptions. We pinpoint the assumptions that the three currently proposed mechanisms of goal attribution (action-effect associations, simulation procedures, and teleological reasoning) imply, and contrast them with the functions they are supposed to fulfil. We argue that while action-effect associations and simulation procedures are generally well suited to serve on-line action monitoring and prediction, social learning of new means actions and artefact functions requires the inferential productivity of teleological reasoning.
Article
Infants between 12 and 21 days of age can imitate both facial and manual gestures; this behavior cannot be explained in terms of either conditioning or innate releasing mechanisms. Such imitation implies that human neonates can equate their own unseen behaviors with gestures they see others perform.
Article
A long-standing puzzle in developmental psychology is how infants imitate gestures they cannot see themselves perform (facial gestures). Two critical issues are: (a) the metric infants use to detect cross-modal equivalences in human acts and (b) the process by which they correct their imitative errors. We address these issues in a detailed model of the mechanisms underlying facial imitation. The model can be extended to encompass other types of imitation. The model capitalizes on three new theoretical concepts. First, organ identification is the means by which infants relate parts of their own bodies to corresponding ones of the adult's. Second, body babbling (infants' movement practice gained through self-generated activity) provides experience mapping movements to the resulting body configurations. Third, organ relations provide the metric by which infant and adult acts are perceived in commensurate terms. In imitating, infants attempt to match the organ relations they see exhibited by the adults with those they feel themselves make. We show how development restructures the meaning and function of early imitation. We argue that important aspects of later social cognition are rooted in the initial cross-modal equivalence between self and other found in newborns.
Article
In this review; a description is offered of the way actions are represented, how these representations are built, and how their content can be accessed by the agent and by other agents. Such a description will appear critical for understanding how an action is attributed to its proper origin, or, in other words, how a subject can make a conscious judgement about who the agent of that action is (an agency judgement). This question is central to the problem of self-consciousness: Action is one of the main channels used for communication between individuals, so that determining the agent of an action contributes to differentiating the self from others.
Article
Based on a series of research studies of reactions to facial orientation and response-contingent stimulation, "the game" hypothesis has been developed and explained. Its name is derived from the game-like interactions which are assumed to generate the infant's 1st awareness of a clear contingency. As a stimulus initially becomes more familiar, it becomes increasingly a candidate for classification as contingent where 1 of 3 things can occur: (a) if contingency analysis through successive exposures of the stimulus indicates that the stimulus is a clearly contingent one, then the smiling and cooing and attention of the infant is expected sharply to increase to an asymtote of vigorous responding; (b) if the continuing contingency analysis indicates that the stimulus is clearly noncontingent, then the modest level of smiling and cooing is expected to decline back to base rate elicited by completely novel stimuli; and (c) if, through continuous exposure, the contingency analysis of the stimulus is continuously indeterminate due either to ambiguous aspects of the contingency itself or to confusing situational characteristics, then the response to the stimulus is expected to become progressively less positive and eventually to begin eliciting negative emotional responses. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Investigated whether children would re-enact what an adult actually did or what the adult intended to do. In Experiment, 1 children were shown an adult who tried, but failed, to perform certain target acts. Completed target acts were thus not observed. Children in comparison groups either saw the full target act or appropriate controls. Results showed that children could infer the adult's intended act by watching the failed attempts. Experiment 2 tested children's understanding of an inanimate object that traced the same movements as the person had followed. Children showed a completely different reaction to the mechanical device than to the person: They did not produce the target acts in this case. Eighteen-mo-olds situate people within a psychological framework that differentiates between the surface behavior of people and a deeper level involving goals and intentions. They have already adopted a fundamental aspect of folk psychology—persons (but not inanimate objects) are understood within a framework involving goals and intentions. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Exp 1 demonstrated that autistic Ss continue to fail a task originally designed as one of strategic deception when there is no opponent present: They perseveratively indicate the target object. The authors argue that this behavior is better explained in terms of failing to disengage from an object than in terms of a theory-of-mind deficit. To ensure that their difficulties were not due to failure to construe the task in a competitive manner, the authors ran a 2nd study, on detour reaching. Compared with control Ss, the autistic Ss had great difficulty with the task. Children's difficulties with these 2 tasks are discussed in light of recent evidence that autism is associated with failing executive tasks, and it is argued that viewing the syndrome as an executive deficit has clear advantages. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Sixty 3- and 4-5-month-old infants were simultaneously presented with 2 on-line video images of their own legs. In 3 experiments, the temporal contingency between the 2 images and the infant's actual movements was maintained constant while their spatial relationships were systematically manipulated. In Experiment 1, both spatial orientation and directionality of movement on the visual display were varied. In Experiment 2, only directionality of movement was varied. In Experiment 3, only spatial orientation was varied. Analyses focused on infants' preferential looking and relative amount of leg activity while looking at either view. Results show that both groups of infants actively compared and explored the 2 views of their legs. They looked significantly longer and generated significantly more leg activity while looking at the view displaying a left-right inversion. These results demonstrate that the perception of self-produced leg movements by young infants is partially determined by spatial information about movement directionality. The results are interpreted as evidence of an early detection of intermodal invariants which specify the body as a situated agent in the environment. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Two-month-olds and newborns were tested in a situation where they had the opportunity to experience different auditory consequences of their own oral activity on a dummy pacifier. Modulation of oral activity was scored and analyzed relative to two types of contingent auditory feedback, either analog or non-analog to the effort exerted by the infant on the pacifier. The dummy pacifier was connected to an air pressure transducer for recording of oral action. In two different experimental conditions, each time the infant sucked above a certain pressure threshold they heard a perfectly contingent sound of varying pitch. In one condition, the pitch variation was analog to the pressure applied by the infant on the pacifier (analog condition). In another, the pitch variation was random (non-analog condition). As rationale, a differential modulation of oral activity in these two conditions was construed as indexing some voluntary control and the sense of a causal link between sucking and its auditory consequences, beyond mere temporal contingency detection and response–stimulus association. Results indicated that 2-month-olds showed clear signs of modulation of their oral activity on the pacifier as a function of analog versus non-analog condition. In contrast, newborns did not show any signs of such modulation either between experimental conditions (analog versus non-analog contingent sounds) or between baseline (no contingent sounds condition) and experimental conditions. These observations are interpreted as evidence of self-exploration and the emergence of a sense of self-agency by 2 months of age.
Book
• This work, a second edition of which has very kindly been requested, was followed by La Construction du réel chez l'enfant and was to have been completed by a study of the genesis of imitation in the child. The latter piece of research, whose publication we have postponed because it is so closely connected with the analysis of play and representational symbolism, appeared in 1945, inserted in a third work, La formation du symbole chez l'enfant. Together these three works form one entity dedicated to the beginnings of intelligence, that is to say, to the various manifestations of sensorimotor intelligence and to the most elementary forms of expression. The theses developed in this volume, which concern in particular the formation of the sensorimotor schemata and the mechanism of mental assimilation, have given rise to much discussion which pleases us and prompts us to thank both our opponents and our sympathizers for their kind interest in our work. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
In this paper we shall argue that mentalistic action explanations, which form an essential component of a mature theory of mind, are conceptually and developmentally derived from an earlier and purely teleological interpretational system present in infancy. First we summarize our evidence demonstrating teleological action explanations in one-year-olds. Then we shall briefly contrast the structure of teleological vs. causal mentalistic action explanations and outline four logical possibilities concerning the nature of the developmental relationship between them. We shall argue for the view that causal mentalistic action explanations are constructed as useful theoretical extensions of the earlier, purely teleological, nonmentalistic interpretational stance.
Article
This study explored infants' ability to discriminate between, and their tendency to reproduce, the accidental and intentional actions of others. Twenty 14- through 18-month-olds watched an adult perform a series of two-step actions on objects that made interesting results occur. Some of the modeled actions were marked vocally as intentional (“There!”), some were marked vocally as accidental (“Woops!”). Following each demonstration, infants were given a chance to make the result occur themselves. Overall, infants imitated almost twice as many of the adult's intentional actions as her accidental ones. Infants before age 18 months thus may understand something about the intentions of other persons. This understanding represents infants' first step toward adult-like social cognition and underlies their acquisition of language and other cultural skills.
Article
Prior studies and have found that infants focus on the goals of an action over other details. The current studies tested whether infants would distinguish between a behavior that seemed to be goal-directed and one that seemed not to be. Infants in one condition saw an actor grasp one of two toys that sat side by side on a stage. Infants in the other condition saw the actor drop her hand onto one of the toys in a manner that looked unintentional. Once infants had been habituated to these events, they were shown test events in which either the path of motion or the object that was touched had changed. Nine-month-olds differentiated between these two actions. When they saw the actor grasp the toy, they looked longer on trials with a change in goal object than on trials with a change in path. When they saw the actor drop her hand onto the toy, they looked equally at the two test events. These findings did not result from infants being more interested in grasping as compared to inert hands. In a second study, 5-month-old infants showed patterns similar to those seen in 9-month-olds. These findings have implications for theories of the development of the concept of intention. They argue against the claim that infants are innately predisposed to interpret any motion of an animate agent as intentional.
Article
Two experiments investigated whether infants represent goal-directed actions of others in a way that allows them to draw inferences to unobserved states of affairs (such as unseen goal states or occluded obstacles). We measured looking times to assess violation of infants’ expectations upon perceiving either a change in the actions of computer-animated figures or in the context of such actions. The first experiment tested whether infants would attribute a goal to an action that they had not seen completed. The second experiment tested whether infants would infer from an observed action the presence of an occluded object that functions as an obstacle. The looking time patterns of 12-month-olds indicated that they were able to make both types of inferences, while 9-month-olds failed in both tasks. These results demonstrate that, by the end of the first year of life, infants use the principle of rational action not only for the interpretation and prediction of goal-directed actions, but also for making productive inferences about unseen aspects of their context. We discuss the underlying mechanisms that may be involved in the developmental change from 9 to 12 months of age in the ability to infer hypothetical (unseen) states of affairs in teleological action representations.
Article
A new framework for the understanding of functional relationships between perception and action is discussed. According to this framework, perceived events and planned actions share a common representational domain (common-coding approach). Supporting evidence from two classes of experimental paradigms is presented: induction paradigms and interference paradigms. Induction paradigms study how certain stimuli induce certain actions by virtue of similarity. Evidence from two types of induction tasks is reviewed: sensorimotor synchronisation and spatial compatibility tasks. Interference paradigms study the mutual interference between the perception of ongoing events and the preparation and control of ongoing action. Again, evidence from two types of such tasks is reviewed, implying interference in either direction. It is concluded that the evidence available supports the common coding principle. A further general principle emerging from these studies is the action effect principle that is, the principle that cognitive representations of action effects play a critical role in the planning and control of these actions.
Article
The effect of conflicting visual, and mechanical proprioceptive information on postural stability was compared in two groups of infants. The older group could sit or stand unsupported, while the younger group could sit but could not yet stand. The results suggest that visual proprioception functions to maintain postural stability before infants have learned to stand.
Article
Infants between 12 and 21 days of age can imitate both facial and manual gestures; this behavior cannot be explained in terms of either conditioning or innate releasing mechanisms. Such imitation implies that human neonates can equate their own unseen behaviors with gestures they see others perform.
Article
The development of an ability to use vision in adjusting the hand and the fingers to the orientation of an object to be grasped was studied in a group of 15 infants. They were 18 weeks at the first session and were seen at 4-week intervals until 34 weeks old. At each session they were presented with horizontal and vertical rods. The orientation of the hand of the infant when reaching for these rods was measured at each 60-msec interval during the last 540 msec of the approach. It was found that even at the youngest age there were signs of adjustment of the hand to the orientation of the object. However, at that age the adjustments were rather incomplete. During the months that followed there was a rapid improvement in the skill studied. The findings were in accordance with the idea that information about object orientation is accessible to the manual system when infants start reaching for objects but that the system has yet to be tuned and calibrated before functioning adequately.
Article
This paper reports a habituation study indicating that 12-month-old infants can take the "intentional stance" in interpreting the goal-directed spatial behavior of a rational agent. First, we examine previous empirical claims suggesting that the ability to attribute intentions to others emerges during the second half of the first year. It is argued that neither the perceptual evidence (concerning the early ability to discriminate agents), nor the behavioral data (indicating the use of communicative gestures for instrumental purposes) are sufficient to support such claims about the early appearance of a theory of mind, as there are alternative explanations for these phenomena in terms of simpler psychological processes. It is then suggested that to show that an infant indeed attributes an intention to interpret the goal-directed behavior of a rational agent, one needs to demonstrate that the baby can generate an expectation about the most rational future means action that the agent will perform in a new situation to achieve its goal. We then describe a visual habituation study that meets this requirement. The results demonstrate that based on the equifinal structure of an agent's spatial behavior, 12-month-old infants can identify the agent's goal and interpret its actions causally in relation to it. Furthermore, our study indicates that infants of this age are able to evaluate the rationality of the agent's goal-directed actions, which is a necessary requirement for applying the intentional stance. In closing, we discuss some of the theoretical and methodological implications of our study.
Article
Research with young children has shown that, like adults, they focus selectively on the aspects of an actor's behavior that are relevant to his or her underlying intentions. The current studies used the visual habituation paradigm to ask whether infants would similarly attend to those aspects of an action that are related to the actor's goals. Infants saw an actor reach for and grasp one of two toys sitting side by side on a curtained stage. After habituation, the positions of the toys were switched and babies saw test events in which there was a change in either the path of motion taken by the actor's arm or the object that was grasped by the actor. In the first study, 9-month-old infants looked longer when the actor grasped a new toy than when she moved through a new path. Nine-month-olds who saw an inanimate object of approximately the same dimensions as the actor's arm touch the toy did not show this pattern in test. In the second study, 5-month-old infants showed similar, though weaker, patterns. A third study provided evidence that the findings for the events involving a person were not due to perceptual changes in the objects caused by occlusion by the hand. A fourth study replicated the 9 month results for a human grasp at 6 months, and revealed that these effects did not emerge when infants saw an inanimate object with digits that moved to grasp the toy. Taken together, these findings indicate that young infants distinguish in their reasoning about human action and object motion, and that by 6 months infants encode the actions of other people in ways that are consistent with more mature understandings of goal-directed action.
Article
Two experiments assessed infants' understanding that actions that occur in sequence may be related to an overarching goal. Experiment 1 tested whether embedding an ambiguous action (touching the lid of a box) in a sequence that culminated with an action infants readily construe as goal-directed (grasping a toy inside the box) would alter infants' construal of the ambiguous action. Having seen the ambiguous action in this context, infants later construed this action in isolation as being directed at the toy within the box. Experiment 2 tested whether infants related the two actions on the basis of the temporal or the causal relation between them. When the causal relation was disrupted but the temporal relation was preserved, infants no longer related the two actions. These findings indicate that 12-month-old infants relate single actions to overarching goals and that they do so by construing goal-directed action in a causal framework.
Article
We contrast two positions concerning the initial domain of actions that infants interpret as goal-directed. The 'narrow scope' view holds that goal-attribution in 6- and 9-month-olds is restricted to highly familiar actions (such as grasping). The cue-based approach of the infant's 'teleological stance', however, predicts that if the cues of equifinal variation of action and a salient action effect are present, young infants can attribute goals to a 'wide scope' of entities including unfamiliar human actions and actions of novel objects lacking human features. It is argued that previous failures to show goal-attribution to unfamiliar actions were due to the absence of these cues. We report a modified replication of Woodward (1999) showing that when a salient action-effect is presented, even young infants can attribute a goal to an unfamiliar manual action. This study together with other recent experiments reviewed support the 'wide scope' approach indicating that if the cues of goal-directedness are present even 6-month-olds attribute goals to unfamiliar actions.
Article
We investigated the neural correlates of the perception of human goal-directed action by 8-month-old infants. Infants viewed video loops of complete and incomplete actions, which they could discriminate according to our pilot study, while we recorded their electrophysiological brain activity. Analysis of bursts of gamma-band oscillations resulting from passive viewing of these stimuli indicated increased gamma-band activity over left frontal regions when viewing incomplete actions as compared with complete actions. These results suggest that by 8 months infants are sensitive to the disruption of perceived goal-directed actions.
Article
This paper reviews studies on infants' imitation of goal-directed actions in the first two years of life. Special emphasis is given to the role of the two observable components of an action, that is, the movement and the action effects, on infants' replication of target actions. The reviewed studies provide evidence that infants benefit most from a full demonstration of both movements and effects. If movements are demonstrated in isolation, infants may encode this information, but they preferentially reproduce actions that lead to salient effects. If action effects are presented in isolation, infants younger than 19 months usually fail to emulate the unseen movements that would be necessary to produce these effects. Infants' ability to predict action effects or to infer unseen movements from incomplete demonstrations improves substantially at the end of the second year of life. It is concluded that the capability to learn relations between movements and action effects by observation, and the knowledge about movement-effect relations acquired so far, may be important factors underlying the developmental changes in infants' imitation of goal-directed actions.
Article
Infant imitation demonstrates that the perception and production of human action are closely linked by a 'supramodal' representation of action. This action representation unites observation and execution into a common framework, and it has far-reaching implications for the development of social cognition. It allows infants to see the behaviors of others as commensurate with their own-as 'like me.' Based on the 'like me' perception of others, social encounters are interpretable and informative. Infants can use themselves as a framework for understanding others and can learn about the possibilities and consequences of their own potential acts by observing the behavior of others. Through social interaction with other intentional agents who are viewed as 'like me,' infants develop a richer social cognition. This paper explores the early manifestations and cascading developmental effects of the 'like me' conception.