Robotics agents are meant to play an increasingly larger role in our everyday lives. To be successfully integrated in our environment, robots will need to develop and display adaptive, robust, and socially suitable behaviours. To tackle these issues, the robotics research community has invested a considerable amount of efforts in modelling robotic architectures inspired by research on living systems, from ethology to developmental psychology. Following a similar approach, this thesis presents the research results of the modelling and experimental testing of robotic architectures based on affective and attachment bonds between young infants and their primary caregiver. I follow a bottom-up approach to the modelling of such bonds, examining how they can promote the situated development of an autonomous robot. Specifically, the models used and the results from the experiments carried out in laboratory settings and with naive users demonstrate the impact such affective bonds have on the learning outcomes of an autonomous robot and on the perception and behaviour of humans. This research leads to the emphasis on the importance of the interplay between the dynamics of the regulatory behaviours performed by a robot and the responsiveness of the human partner. The coupling of such signals and behaviours in an attachment-like dyad determines the nature of the outcomes for the robot, in terms of learning or the satisfaction of other needs. The experiments carried out also demonstrate of the attachment system can help a robot adapt its own social behaviour to that of the human partners, as infants are thought to do during their development.

Social learning has given us insight into how children learn actions from others across different domains (e.g., actions on objects, pretend play, and tool use). However, little research exists to confirm whether young children can generate their own novel actions. Three different settings were chosen to offer a varied investigation of children’s ability to generate novel actions: generating multiple actions with novel objects; generating iconic gestures in order to communicate; and generating pretend actions using object substitution. Generating multiple actions with novel objects: The Unusual Box test was developed to investigate children’s ability to generate multiple actions with novel objects (Chapter 2). The Unusual Box test involves children playing with a wooden box that contains many different features (e.g., rings, stairs, strings), and five novel objects. The number of different actions performed on the box and with the objects (i.e., fluency) was used as a measure of their individual learning. Positive correlations between the fluency scores of 24 3- and 4-year-olds on the Unusual Box test and two existing measures of divergent thinking were found. Divergent thinking relates to the ability to think of multiple answers based on one premise. Furthermore, a large range of fluency scores indicated individual differences in children’s ability to generate multiple actions with novel objects. In addition, 16 2-year-olds were assessed on the Unusual Box test, twice two weeks apart, to investigate test-retest reliability and the possibility that the Unusual Box test could be used with children younger than 3 years. A strong positive correlation between the scores on the two assessments showed high test-retest reliability, while individual differences in fluency scores and the absence of a floor effect indicated that the Unusual Box test was usable in children from 2 years of age. Generating iconic gestures in order to communicate: Children’s ability to generate iconic gestures in order to communicate was assessed using a game to request stickers from an experimenter (N = 20, Chapter 3). In order to get a sticker children had to communicate to the experimenter which out of two objects they wanted (only one object had a sticker attached to it). Children’s use of speech or pointing was ineffective; therefore only generating an iconic gesture was sufficient to retrieve the sticker. Children generated a correct iconic gesture on 71% of the trials. These findings indicate that children generate their own iconic gestures in order to communicate; and that they understand the representational nature of iconic gestures, and use this in their own generation of iconic gestures. Generating pretend actions using object substitution: In order to determine whether children are able to generate their own object substitution actions and understand the representational nature of these actions, 45 3- and 4-year-olds were familiarized with the goal of a task through modelling actions. Children distinguished between the intentions of an experimenter to pretend, or try and perform a correct action. Children mainly imitated the pretend actions, while correcting the trying actions. Next, children were presented with objects for which they had to generate their own object substitution actions without being shown a model. When children had previously been shown pretend actions, children generated their own object substitution actions. This indicates that children generate their own object substitution actions, and that they understand the representational nature of these actions. An additional study with 34 3-year-olds, revealed no significant correlations between divergent thinking, inhibitory control, or children’s object substitution in a free play setting, and children’s ability to generate object substitution actions in the experimental setting.