Social and Statistical Mechanisms of Prelinguistic Vocal Development

语言前声音发展的社会和统计机制

• 批准号: 0844015
• 负责人: Michael Goldstein
• 金额: $ 35.19万
• 依托单位: Cornell University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-04-01 至 2013-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0844015&HistoricalAwards=false
• 关键词: Social; Statistical; Mechanisms; Prelinguistic; Vocal

How do infants learn to produce the sounds of their language? The vocal abilities of infants change dramatically over the first year of life. Beginning with the earliest, immature vocalizations, infants make rapid progress, typically producing their first words by 12 months of age. Along the way, they begin to produce speech-like syllables and to structure sequences of syllables in accordance with the phonological patterns of their language environment. While the vocal achievements of the first year are well-described, not nearly as much is known about the mechanisms of change that drive vocal development. However, many studies of song learning in birds, as well as the investigators' preliminary studies, indicate that the social environment is a source of rich structure and feedback that can guide early vocal learning and development. Can prelinguistic infants use social feedback to create new, more developmentally advanced vocalizations? The investigators hypothesize that two processes enable social influences on vocal learning. During the first 6 months, infants learn that their babbling reliably affects adults' behavior. From this early learning, infants become attuned to social responses to their vocalizing. Over the next 6 months, infants use the contingent reactions of caregivers to their babbling to guide their vocalizations towards the phonological patterns of their language. Acquiring knowledge of underlying patterns that govern the relations among multiple examples is a process known as statistical learning. To investigate these processes of vocal learning, the form and timing of caregivers' reactions to their infants' sounds will be manipulated. First, the process of learning associations between babbling and caregivers' reactions will be investigated in 2- to 5-month-old infants. Infants' information-processing capacities and caregivers' natural responsiveness to babbling will be used to predict individual differences in learning. Second, the statistical properties and contingency of speech to 9-month-old infants will be manipulated to assess their effects on vocal learning. The role of statistical learning in speech perception and production will be directly compared. Finally, the relation between the two learning mechanisms (associative learning at 2 to 5 months and statistical learning at 9 months) will be determined. Findings from these studies will allow the investigators to construct a new theoretical framework for vocal development, integrating data from social learning and speech perception. The experiments will benefit from new applications of digital technologies that allow caregiver behavior to be manipulated in real-time (via instructions given over wireless headphones) during naturalistic interaction. At the same time, changes in infants' vocalizations will be recorded and analyzed. The proposed research has broad and fundamental implications for educating parents in providing optimal learning environments for their infants. An understanding of the role of socially guided learning in speech and language could be used to help caregivers to be more sensitive to infants' behavior in ways that would facilitate development. Investigating social influences on phonological development will contribute to the study of speech-language pathology and of processes underlying both successful and disordered communicative development. This research will provide a framework for understanding the role of early social learning in communication and language deficits such as those seen in autism and Down's syndrome. By elucidating mechanisms by which infants learn from caregivers, these findings will be useful in designing behavioral interventions for infants with developmental disorders.

婴儿是如何学会发出自己语言的声音的？婴儿的发声能力在生命的第一年发生了巨大的变化。 从最早的、不成熟的发声开始，婴儿进步很快，通常在12个月大的时候就能说出第一个单词。 沿着这个过程，他们开始产生类似语音的音节，并根据他们所处语言环境的语音模式来构建音节序列。 虽然第一年的声乐成就被很好地描述，但对推动声乐发展的变化机制知之甚少。然而，许多关于鸟类歌唱学习的研究，以及研究人员的初步研究表明，社会环境是丰富的结构和反馈的来源，可以指导早期的声乐学习和发展。 语言发育前的婴儿能否利用社会反馈来创造新的、更先进的发声？ 研究人员假设，两个过程使社会影响声乐学习。 在最初的6个月里，婴儿知道他们的牙牙学语确实会影响成年人的行为。从这个早期的学习，婴儿成为协调社会反应，他们的发声。 在接下来的6个月里，婴儿利用照顾者对他们牙牙学语的偶然反应来指导他们的发声，使其符合他们语言的语音模式。 获取管理多个示例之间关系的基本模式的知识是一个称为统计学习的过程。 为了研究这些声音学习的过程，将操纵照顾者对婴儿声音的反应的形式和时间。首先，将在2至5个月大的婴儿中调查牙牙学语和照顾者反应之间的学习关联过程。 婴儿的信息处理能力和照顾者对牙牙学语的自然反应将被用来预测学习的个体差异。 第二，对9个月大的婴儿讲话的统计特性和偶然性将被操纵，以评估其对声乐学习的影响。 统计学习在言语感知和产生中的作用将被直接比较。 最后，将确定两种学习机制（2至5个月的联想学习和9个月的统计学习）之间的关系。这些研究的结果将使研究人员能够构建一个新的声乐发展理论框架，整合社会学习和言语感知的数据。这些实验将受益于数字技术的新应用，这些技术允许在自然互动期间实时操纵护理人员的行为（通过无线耳机给出的指令）。 同时，婴儿发声的变化将被记录和分析。这项研究对教育父母为婴儿提供最佳学习环境具有广泛和根本的意义。对社会引导学习在言语和语言中的作用的理解可以用来帮助照顾者以促进发展的方式对婴儿的行为更加敏感。调查语音发展的社会影响将有助于研究语音语言病理学和过程的成功和混乱的交际发展。 这项研究将为理解早期社会学习在沟通和语言缺陷中的作用提供一个框架，如自闭症和唐氏综合症中所见。通过阐明婴儿从照顾者那里学习的机制，这些发现将有助于为发育障碍婴儿设计行为干预措施。