From object manipulation to tool use: infants, robots, animals (2013) – ABSTRACTS

From reaching to grasping a tool-shaped object in the 1st year of life

This presentation will focus on preliminary longitudinal data tracking developmental changes in a few infants followed from the age of 2.5 months (prior to reach onset) up to the end of the first year while looking at and reaching for an object shaped like a drumstick.    The analyses will focus on exploratory looking behavior preceding the act of reaching as captured via eye-tracking, and on the reaching behavior per se.  The data will attempt to address (1) how much visual attention and reaching behaviors are matching each other in developmental time, and (2) identify to which part of the object visual attention and reaching are directed primarily over time; the handle or the drum part of the object.

  • Jeff Lockman, Dept of Psychology, University of Minnesota, USA

When does tool use become distinctively human?:  The development of percussive tool use

Jeffrey J. Lockman and B. J. Kahrs, Tulane University

Percussive tool use—striking one object against another—is a common form of tool use in humans and some non-human primates.  We address at a motor level how this form of tool use begins to manifest qualities that are uniquely human.  We report a series of studies using motion tracking technology from infancy to the third year to show how banging transitions into  hammering and how hammering only begins to evidence relatively greater involvement of the wrist near the third year.  Further, this distinctively human form of hammering is lateralized, appearing only in the preferred hand.  Our findings suggest that achieving distal control of a percussive tool is a gradual process, spanning the infancy and early childhood years and builds on earlier action patterns and forms of lateralization.

  • Alex Kacelnik, Behavioural Ecology Research Group, University of Oxford, UK:

Dedicated and general cognition involved in avian tool use.

I will address tool use in crows, parrots and finches, discussing the recruitment of specialised and general problem solving processes.

  • Elizabeth Price, Center for Behaviour and Evolution, University of Newcastle, UK

Processes underlying tool modification in children and chimpanzees

Although many animals use tools there is to date little compelling evidence of the racheting effect common to human technology. One prominent account for this divide is that racheting of complexity in other animals is limited by reliance on less faithful forms of social learning (e.g. emulation) than found in humans (e.g. imitation). In two experiments we investigated whether preschool children and chimpanzees would learn from a model to modify tools, and further, how much information they required to do so. Key similarities and differences in learning processes are discussed, with reference to the evolution of cumulative culture.

  • Luciano Fadiga, Robotics, Brain and Cognitive Science, Italian Institute of Technology

The Neurophysiology of Object Grasping
In recent years a lot of attention has been dedicated to the study of the cortical circuits sub serving the motor control of the grasping hand, the visuomotor transformation of objects into hand poses and the interrelation between pragmatic and semantic representations of objects and actions. In my presentation I will show and discuss the most recent ‘state of the art’ on this topic integrating the perspective with some very recent empirical results.

Curiosity-Driven Development: From the Discovery of Object Affordances to the Discovery of Communication

Abstract: Based on robotics models, I will discuss how mechanisms of intrinsically motivated learning, also called curiosity-driven exploration, can self-organize developmental trajectories, starting from discovery of the body, then object affordances, then vocal babbling and vocal interactions with others. In such a vision, a limited set of meta-cognitive structures allow a learner to order its own learning experiences, creating its own curriculum where skills, including the manipulation of external objects, get naturally sequenced towards increasing complexity. In particular, I will show that the onset of language spontaneously forms out of such sensorimotor development, where vocalizations are discovered by the learner to be special forms of “tools” that allow to manipulate a special kind of external entities, i.e. others.


Gottlieb, J., Oudeyer, P-Y., Lopes, M., Baranes, A. (in press) Information seeking, curiosity and attention: computational and neural mechanisms, Trends in Cognitive Science.

Moulin-Frier, C., Nguyen, M., Oudeyer, P-Y. (submitted) Self-Organization of Early Vocal Development in Infants and Machines: The Role of Intrinsic Motivation,

Oudeyer, P-Y., Kaplan, F. (2006) Discovering communication, Connection Science, 18(2), pp. 189–206.

  • Philippe Gaussier, Braud, A. Pitti, ETIS (Image and Signal processing Lab), UMR CNRS 8051, ENSEA, University of Cergy-Pontoise

Modeling the sensory-motor development: from the emergence of imitation capabilities to the discovery of tool-use

Learning how to use a tool for an infant is clearly a complex task involving several cognitive capabilities: object recognition, reaching and manipulation using visual and tactile information, planning… Following a developmental perspective, we propose that these cognitive capabilities emerge first as a side-effect of sensory-motor development. In previous works, we have shown how low-level imitation capabilities can emerge in a simple sensory-motor architecture (homeostatic system) thanks to the perceptual ambiguity between vision and proprioception. In front of a person, the robot, being unable to differentiate its own hand from a human hand, tries to reduce the error between its proprioception and the visual information inducing the mimicking of the gesture (delayed imitation being obtained by a simple inhibition of the final motor command). Following the same approach, we have shown how a robot can learn to recognize facial expressions if the human partner mimics the robot’s facial expression. Next, the emotional state can be associated to an arbitrary object allowing the robot to develop reaching capabilities toward “positive” objects and avoidance of “negative” objects. In current researches, we investigate how the body schema can be adapted or tuned fastly enough to allow the use of a tool in order to manipulate an object and how the interest of using a tool could be discovered from either social interactions or exploratory behaviors.

  • Frank Guerin, Dept of Computing Science, University of Aberdeen, UK

How could we get robots up to infant-level competence in manipulation activities?

  • Rana Esseily, LPP, Université Paris Descartes

Observational learning of tool use

  • Eszter Somogyi, LPP, Université Paris Descartes

Seeing Is Better Than Doing: Visual Familiarization With The Function Of A Tool Helps Learning More Than Manipulation Alone. A longitudinal study.