Periodic Reporting for period 1 - Hadza communication (Communication Development in Infants: the Case of the Hadza Hunter-Gatherers of Tanzania)
Berichtszeitraum: 2015-11-01 bis 2017-10-31
The ways children learn to communicate verbally and non-verbally depend to a great extent on the culture in which they are raised. Much is known about the social environment of children living in Western societies, and some is known about those living in many non-Western societies, but much less is known about how children are socialized in hunter-gatherer societies. The objective of the proposed project is to gain insights into the development of communication in hunter-gatherer children from the Hadza living in Tanzania. The project’s aims are
• to develop a minimally intrusive data collection method that gives access to the learning process by including physiological measures.
• to analyze physiological responses to infants’ verbal and non-verbal interactions with their social environment focusing on the Hadza, as a particularly interesting cultural community.
• extract patterns in data from diverse sources (audio, video, observation, electrophysiological measures) with automated data mining techniques.
The approach is interdisciplinary, integrating methods from anthropology and ethology with modern physiological assessment methods, innovative wearable technology and data mining techniques. It is important to know more about the development of communication in Hadza infants as communication in hunter-gatherers has not been comprehensively studied and the Hadza are a community living in conditions that resemble our ancestors’ living conditions. This helps to shed light on the evolution of language in our species. As hunter-gatherer societies are increasingly threatened due to ecological, economic and political changes in their environment, it is crucial to conduct this research as soon as possible.
Using the Multimodal Interaction Recorder for Children (MIRC), a relatively unobtrusive method for collecting daylong interactions between children and their social environment was designed that can record not only speech, but also the relative location of relevant others and heart-rate. The recordings with the MIRC can, indeed, be used to analyse physiological responses to interactional data, but unfortunately no appropriate data mining techniques are currently available to carry out these analyses. Efforts are continued to develop appropriate techniques, but this was not feasible within the scope of the project.
However, additional observations using –more traditional– video recordings allowed us to discover various patterns concerning the way children are being addressed verbally and non-verbally. In particular, it was found that the Hadza society is changing rapidly as the result of increased tourism. These changes appear to be related to the way children are being addressed verbally and non-verbally, and which in turn could impact their language development.
• to develop a minimally intrusive data collection method that gives access to the learning process by including physiological measures.
• to analyze physiological responses to infants’ verbal and non-verbal interactions with their social environment focusing on the Hadza, as a particularly interesting cultural community.
• extract patterns in data from diverse sources (audio, video, observation, electrophysiological measures) with automated data mining techniques.
The approach is interdisciplinary, integrating methods from anthropology and ethology with modern physiological assessment methods, innovative wearable technology and data mining techniques. It is important to know more about the development of communication in Hadza infants as communication in hunter-gatherers has not been comprehensively studied and the Hadza are a community living in conditions that resemble our ancestors’ living conditions. This helps to shed light on the evolution of language in our species. As hunter-gatherer societies are increasingly threatened due to ecological, economic and political changes in their environment, it is crucial to conduct this research as soon as possible.
Using the Multimodal Interaction Recorder for Children (MIRC), a relatively unobtrusive method for collecting daylong interactions between children and their social environment was designed that can record not only speech, but also the relative location of relevant others and heart-rate. The recordings with the MIRC can, indeed, be used to analyse physiological responses to interactional data, but unfortunately no appropriate data mining techniques are currently available to carry out these analyses. Efforts are continued to develop appropriate techniques, but this was not feasible within the scope of the project.
However, additional observations using –more traditional– video recordings allowed us to discover various patterns concerning the way children are being addressed verbally and non-verbally. In particular, it was found that the Hadza society is changing rapidly as the result of increased tourism. These changes appear to be related to the way children are being addressed verbally and non-verbally, and which in turn could impact their language development.
In cooperation with the Max Planck Institute for Ornithology a miniaturized device for data collection, the Multi-modal Interaction Recorder for Children (MIRC), was developed. This wearable device gives access not only to the audio an infant is exposed to but also to the distance of major caregivers and a physiological measure (electric heart activity/heart rate), which can potentially be related to each other as they are fully synchronized. The initial idea to have a “minimally intrusive” device that can be worn on the arm or leg was rejected in favour of the quality of the heart rate data, which can be achieved with chest electrodes.
During the field study, data on more than 20 Hadza infants could be collected in Tanzania, including, observational data, video-recordings and data generated by the MIRC.
The data mining approach envisioned to analyse the data from the MIRC could not be realised, as AI techniques to analyse children’s speech data is not yet sufficiently well advanced to carry out the analyses we were aiming for. However, efforts to improve these techniques are carried out at several institutes worldwide, including at Tilburg University. As soon as the automatic coding of the audio files can be realised, data mining techniques will be applied to the data and interesting results are expected.
Analysis of the additional observations revealed unanticipated, interesting results for the question of cultural changes, as differences in livelihood in different Hadza camps seem to be related to interactional patterns and probably language development in infants. Traditional ways of raising a child thus seem to be changing in response to changes in lifestyle.
As part of the training and networking, a secondment was carried out with Camilla Power from the University of East London, who provided the crucial anthropological input. Further training was achieved regarding co-organisation of workshops, theoretical background in evolution of language, and regarding data mining techniques. The results of these activities will help further the career of Dr. Abels.
The results of the project were presented at several scientific conferences and a workshop, and a public workshop for mothers and professionals (midwife, child care worker) was conducted as an outreach activity.
During the field study, data on more than 20 Hadza infants could be collected in Tanzania, including, observational data, video-recordings and data generated by the MIRC.
The data mining approach envisioned to analyse the data from the MIRC could not be realised, as AI techniques to analyse children’s speech data is not yet sufficiently well advanced to carry out the analyses we were aiming for. However, efforts to improve these techniques are carried out at several institutes worldwide, including at Tilburg University. As soon as the automatic coding of the audio files can be realised, data mining techniques will be applied to the data and interesting results are expected.
Analysis of the additional observations revealed unanticipated, interesting results for the question of cultural changes, as differences in livelihood in different Hadza camps seem to be related to interactional patterns and probably language development in infants. Traditional ways of raising a child thus seem to be changing in response to changes in lifestyle.
As part of the training and networking, a secondment was carried out with Camilla Power from the University of East London, who provided the crucial anthropological input. Further training was achieved regarding co-organisation of workshops, theoretical background in evolution of language, and regarding data mining techniques. The results of these activities will help further the career of Dr. Abels.
The results of the project were presented at several scientific conferences and a workshop, and a public workshop for mothers and professionals (midwife, child care worker) was conducted as an outreach activity.
The MIRC, which was developed to benefit this project, is the first wearable device that can record audio (speech etc), location of several other persons (e.g. caregivers) wearing a matching device, and heart-rate in a synchronised manner and that can be worn by children. This device has received much interest from the scientific community and has the potential to be exploited further to benefit future studies on language development throughout the world, including the EU.
Although the data mining approach envisioned to analyse the data from the MIRC could not be realised within the duration of this project, we expect these to become sufficiently applicable within the next year or two. Once this is realized, all MIRC data from this project will be analysed, allowing us to gain better insights into how interactional patterns surrounding Hadza children relate to their physiological reactions. Given the observed differences in livelihood among different camps, we expect to see these differences being reflected in the results.
Regarding the findings from the observations obtained so far, some interesting conclusions may be drawn. Results indicate that changes in lifestyle are related to changes in interaction patterns. If this is, indeed, true, it can be assumed that similar changes would also occur in the interactions of families with migration experiences. Given that immigration is a major issue in contemporary European society, results that will still come from the data collected in this project can have an impact on the way immigration influences children’s (language) development.
Although the data mining approach envisioned to analyse the data from the MIRC could not be realised within the duration of this project, we expect these to become sufficiently applicable within the next year or two. Once this is realized, all MIRC data from this project will be analysed, allowing us to gain better insights into how interactional patterns surrounding Hadza children relate to their physiological reactions. Given the observed differences in livelihood among different camps, we expect to see these differences being reflected in the results.
Regarding the findings from the observations obtained so far, some interesting conclusions may be drawn. Results indicate that changes in lifestyle are related to changes in interaction patterns. If this is, indeed, true, it can be assumed that similar changes would also occur in the interactions of families with migration experiences. Given that immigration is a major issue in contemporary European society, results that will still come from the data collected in this project can have an impact on the way immigration influences children’s (language) development.