The Effect of Hyperarticulation on Early Language Development (HELD)
The project examines how hyperarticulation (clear articulation) in child-directed speech affects children’s linguistic development. Child-directed speech is a speech-style adults often use when talking to young children. The features of child-directed speech have proven to have a positive impact on various aspects of language development, but very little is known about the specific role of hyperarticulation. Hyperarticulation occurs both in child directed speech, and for example when adults are talking to other adults in noisy environments. Hyperarticulation has the purpose of facilitating or improving the conditions for effective communication, but it is unclear if it has any effect on language development in the long term.
Within the project child-directed speech will be analyzed to assess the degree of hyperarticulation of vowels in speech directed to children at different ages. The degree of hyperarticulation in parents’ speech will also be related to the infants’ development of speech sound categories. A series of experimental studies will be used to determine how the extent, amount and relevance of hyperarticulation affect children’s word learning.
Investigating the impact of hyperarticulation on children’s language development provide an insight into the fundamental language ability and processes involved. It may also contribute to applications in language teaching and the development of intervention of children with delayed language development.
Final report
Purpose and development
The purpose of the project was to investigate the effect of hyperarticulation (exaggerated articulation) on early speech- and language development. Speakers spontaneously adapt their articulation to the situation, for example, speaking more clearly in noisy environments. Speakers also adapt articulation to their conversational partner, for example, when talking to language-learners such as infants.
The project has been implemented according to plan in general terms. Hyperarticulation was studied descriptively in speech directed to infants, as well as relation to early language skills. Two additions were made to the project: the development of a new measure of hyperarticulation, and a comparison between two experimental word-learning paradigms.
Implementation
The first study of the project was based on previously collected data, that is, audio-video recordings of parent-infant interaction. In the recordings, parent speech was tagged and analyzed in terms of the acoustic properties of vowels, specifically fundamental frequency and vowel hyperarticulation. In further studies, infant vocalizations in the same recordings were transcribed, and their phonetic complexity was calculated and related to parents’ vowel hyperarticulation. In the final study, new data of about 100 children was collected to study word-learning, using two different experimental paradigms.
The three most important findings
A new measure for hyperarticulation was developed within the project. Contrary to traditional measures, the novel measure can be used to estimate hyperarticulation of single vowel tokens and normalizes across vowel types and speakers (Marklund & Gustavsson, 2020). This enables studying of dynamics in vowel articulation during a conversation. An evaluation of the measure was performed, comparing parents’ articulation when speaking to their 12-month-old compared to when talking to an adult. Both the new measure and traditional measures show hyperarticulation in infant-directed speech relative to adult-directed speech (Marklund & Gustavsson, 2020).
Furthermore, the project demonstrated a relationship between parent vowel articulation and infant vocalization at 12 months. More hyperarticulated parent vowels were related to more phonetically advanced infant vocalizations (Marklund, E., Marklund, U., & Gustavsson, 2021). On the level of individual utterances, greater hyperarticulation in parent utterances was followed by less advanced child vocalizations (Marklund, U., Marklund, E., & Gustavsson, 2021). A possible explanation for these findings can be that adult hyperarticulation momentarily requires higher processing capacity in the child, which leads to less advanced vocalizations as immediate response, but impacts child speech production positively in the long run.
Since infant-directed speech is typically hyperarticulated relative to adult-directed speech, the impact of infant-directed speech on word-learning has been studied, comparing two different experimental word-learning. The findings show that 14-month-olds are able to learn novel words from infant-directed speech in a simple experiment paradigm, but are unable to learn novel words when the experiment paradigm is cognitively more demanding (Gustavsson, Marklund, Ericsson & Hjerpe, in preparation).
In conclusion, the project demonstrated that adults hyperarticulate their speech to infants, and that there is a relationship between parent vowel hyperarticulation and infant vocalization complexity. Together, these findings suggest that parents’ spontaneous adaptations in interaction with infants, specifically in terms of exaggerating the articulation, may have a positive impact on early language development.
New research questions
The project has generated novel research questions, in particular regarding the more dynamic aspects of hyperarticulation in speech to infants. In speech between adults, the degree of spontaneous hyperarticulation can vary within conversations, or even within single utterances. This raises questions regarding the specific dynamics of hyperarticulation in speech to infants who are still learning language. For example, it would be of interest to study whether the degree of hyperarticulation differs between words that the parent believes that the infant understands and words that the parent does not believe that the infant understands. This would provide additional information about the type of spontaneous communicative adaptations humans make when in conversation with others, as well as how these adaptations may be relevant to early language development.
It would also be interesting to study the role of hyperarticulation in other aspects of early language learning in order to pinpoint specific stages in the developmental process at which it may have an effect, as well as to investigate the relationship between hyperarticulation and other types of communicative adaptations (e.g., prosodic entrainment).
Lastly, the findings of the project have highlighted the question of why infant-directed speech is hyperarticulated relative to adult-directed speech, and proposed a general phonetic explanation. It would be of interest to empirically study how this explanation relates to other hypotheses proposed in the literature.
Dissemination of results and third-stream activities
The findings from the project’s studies has been and will be published in open-access academic journals (Marklund & Gustavsson, 2020; Marklund E., Marklund U. & Gustavsson, 2021; Marklund U., Marklund, E. & Gustavsson, 2021; Gustavsson, Marklund, Ericsson & Hjerpe, in preparation). The findings have also been presented in talks at CLINTEC at the Section of Speech and Language Pathology at Karolinska Institutet, at the Department of Linguistics at Pontifical Catholic University, São Paulo, Brazil, and at an academic conference (Gustavsson & Marklund, 2018). An additional three conference presentations were planned during 2020, but were canceled due to the Covid-19 pandemic. The project and its studies have been presented at internal seminars at the Department of Linguistics, Stockholm University multiple times during the course of the project.
A popular science article about the project’s findings was published in Språktidningen (Hypertydligt tal ger vassare joller, nr 1, 2023), and popular science summaries of the study findings have been continuously published on the project website (su.se/ling/HELD).
The purpose of the project was to investigate the effect of hyperarticulation (exaggerated articulation) on early speech- and language development. Speakers spontaneously adapt their articulation to the situation, for example, speaking more clearly in noisy environments. Speakers also adapt articulation to their conversational partner, for example, when talking to language-learners such as infants.
The project has been implemented according to plan in general terms. Hyperarticulation was studied descriptively in speech directed to infants, as well as relation to early language skills. Two additions were made to the project: the development of a new measure of hyperarticulation, and a comparison between two experimental word-learning paradigms.
Implementation
The first study of the project was based on previously collected data, that is, audio-video recordings of parent-infant interaction. In the recordings, parent speech was tagged and analyzed in terms of the acoustic properties of vowels, specifically fundamental frequency and vowel hyperarticulation. In further studies, infant vocalizations in the same recordings were transcribed, and their phonetic complexity was calculated and related to parents’ vowel hyperarticulation. In the final study, new data of about 100 children was collected to study word-learning, using two different experimental paradigms.
The three most important findings
A new measure for hyperarticulation was developed within the project. Contrary to traditional measures, the novel measure can be used to estimate hyperarticulation of single vowel tokens and normalizes across vowel types and speakers (Marklund & Gustavsson, 2020). This enables studying of dynamics in vowel articulation during a conversation. An evaluation of the measure was performed, comparing parents’ articulation when speaking to their 12-month-old compared to when talking to an adult. Both the new measure and traditional measures show hyperarticulation in infant-directed speech relative to adult-directed speech (Marklund & Gustavsson, 2020).
Furthermore, the project demonstrated a relationship between parent vowel articulation and infant vocalization at 12 months. More hyperarticulated parent vowels were related to more phonetically advanced infant vocalizations (Marklund, E., Marklund, U., & Gustavsson, 2021). On the level of individual utterances, greater hyperarticulation in parent utterances was followed by less advanced child vocalizations (Marklund, U., Marklund, E., & Gustavsson, 2021). A possible explanation for these findings can be that adult hyperarticulation momentarily requires higher processing capacity in the child, which leads to less advanced vocalizations as immediate response, but impacts child speech production positively in the long run.
Since infant-directed speech is typically hyperarticulated relative to adult-directed speech, the impact of infant-directed speech on word-learning has been studied, comparing two different experimental word-learning. The findings show that 14-month-olds are able to learn novel words from infant-directed speech in a simple experiment paradigm, but are unable to learn novel words when the experiment paradigm is cognitively more demanding (Gustavsson, Marklund, Ericsson & Hjerpe, in preparation).
In conclusion, the project demonstrated that adults hyperarticulate their speech to infants, and that there is a relationship between parent vowel hyperarticulation and infant vocalization complexity. Together, these findings suggest that parents’ spontaneous adaptations in interaction with infants, specifically in terms of exaggerating the articulation, may have a positive impact on early language development.
New research questions
The project has generated novel research questions, in particular regarding the more dynamic aspects of hyperarticulation in speech to infants. In speech between adults, the degree of spontaneous hyperarticulation can vary within conversations, or even within single utterances. This raises questions regarding the specific dynamics of hyperarticulation in speech to infants who are still learning language. For example, it would be of interest to study whether the degree of hyperarticulation differs between words that the parent believes that the infant understands and words that the parent does not believe that the infant understands. This would provide additional information about the type of spontaneous communicative adaptations humans make when in conversation with others, as well as how these adaptations may be relevant to early language development.
It would also be interesting to study the role of hyperarticulation in other aspects of early language learning in order to pinpoint specific stages in the developmental process at which it may have an effect, as well as to investigate the relationship between hyperarticulation and other types of communicative adaptations (e.g., prosodic entrainment).
Lastly, the findings of the project have highlighted the question of why infant-directed speech is hyperarticulated relative to adult-directed speech, and proposed a general phonetic explanation. It would be of interest to empirically study how this explanation relates to other hypotheses proposed in the literature.
Dissemination of results and third-stream activities
The findings from the project’s studies has been and will be published in open-access academic journals (Marklund & Gustavsson, 2020; Marklund E., Marklund U. & Gustavsson, 2021; Marklund U., Marklund, E. & Gustavsson, 2021; Gustavsson, Marklund, Ericsson & Hjerpe, in preparation). The findings have also been presented in talks at CLINTEC at the Section of Speech and Language Pathology at Karolinska Institutet, at the Department of Linguistics at Pontifical Catholic University, São Paulo, Brazil, and at an academic conference (Gustavsson & Marklund, 2018). An additional three conference presentations were planned during 2020, but were canceled due to the Covid-19 pandemic. The project and its studies have been presented at internal seminars at the Department of Linguistics, Stockholm University multiple times during the course of the project.
A popular science article about the project’s findings was published in Språktidningen (Hypertydligt tal ger vassare joller, nr 1, 2023), and popular science summaries of the study findings have been continuously published on the project website (su.se/ling/HELD).