Periodic Reporting for period 1 - LETITIA (Language Ecology and Modality in Turn-Taking Acquisition)
Periodo di rendicontazione: 2024-10-01 al 2026-09-30
Turn-taking is one of the fundamental skills children need to learn to become competent conversationalists; learning how and when to contribute to conversation and when your interlocutor will do so is critical to maintaining smooth interactions. Adults across different spoken and signed languages do this effortlessly, they monitor their interlocutors’ turn while planning their own response and carefully time their turn to achieve a swift back and forth of utterances with minimal gaps and overlaps between turns and use minimally interruptive cues to signal understanding or nonunderstandings when they arise. Children’s turn-structure is less complex than that of adults; at around one year, children produce turns involving only one word that gradually grow in complexity alongside their linguistic and
pragmatic skills. Notwithstanding, children under six often show latencies in turn-taking; in caregiver-child interactions, adults may manage slow responses but they might cause non-coordinated turns and misunderstandings when there are more than two conversation partners or in peer-to-peer conversations.8 Nevertheless, insights into the development of conversational turn-taking come exclusively from hearing children acquiring spoken languages. For deaf children, the developmental trajectory of conversational turn-taking is unclear because of three major issues. First, research into sign language acquisition has focused on how deaf children acquire linguistic forms (e.g. phonology, syntax) with very little research on how they develop the key pragmatic underpinnings of conversation. Second, the visual-spatial modality of sign languages provides a fundamentally different ecology for monitoring conversational turn-taking than the auditory-vocal modality of spoken languages. Third, insights into conversational turn-taking from spoken languages may not be directly applicable to signing children as they experience a very different acquisition ecology - 95% of deaf children are born to hearing, non-signing parents and lack rich language models. This makes it challenging to understand what aspects of turn-taking development are linked to (1) language modality, i.e. signed vs. spoken language, and (2) language ecology, i.e. languages where one-to-one vs. one-to-many conversations are the norm. The LETITIA project contributes towards disentangling these threads by studying the development of turntaking
in an extraordinary context: a Balinese village, characterised by high intergenerational deafness and many signers of the local sign language, Kata Kolok. In this setting, communal child-rearing practices lead to most conversations available to children involving many interlocutors. Investigating how deaf children who grow up in such a signing-rich environment develop this fundamental pragmatic skill is unprecedented and affords us the unique opportunity to (1) study sign language acquisition in an ecology that is in many ways similar to that of spoken language acquisition, (2) pioneer in applying methods from social interaction to sign language acquisition data, and (3) begin to disentangle effects of modality and language ecology on language development. To better understand how a signing-rich environment shapes deaf children’s development of turn-taking, this project pursues three research objectives (RO) through two related research questions (RQ):
RQ1: What are the characteristics of turn-taking among adult and child signers?
RO1: document the typological and pragmatic characteristics of turn-taking in Kata Kolok.
RO2: track the developmental trajectory of turn-taking skills in deaf Kata Kolok acquiring children.
RQ2: How do conversational settings with many signing interlocutors modulate the turn-taking skills of signing adults and children?
RO3: explore how the turn-taking skills of adult and child signers are modulated by conversational factors, specifically age of conversation partners (adult vs. child) and number of interlocutors (two vs. multiple).
pragmatic skills. Notwithstanding, children under six often show latencies in turn-taking; in caregiver-child interactions, adults may manage slow responses but they might cause non-coordinated turns and misunderstandings when there are more than two conversation partners or in peer-to-peer conversations.8 Nevertheless, insights into the development of conversational turn-taking come exclusively from hearing children acquiring spoken languages. For deaf children, the developmental trajectory of conversational turn-taking is unclear because of three major issues. First, research into sign language acquisition has focused on how deaf children acquire linguistic forms (e.g. phonology, syntax) with very little research on how they develop the key pragmatic underpinnings of conversation. Second, the visual-spatial modality of sign languages provides a fundamentally different ecology for monitoring conversational turn-taking than the auditory-vocal modality of spoken languages. Third, insights into conversational turn-taking from spoken languages may not be directly applicable to signing children as they experience a very different acquisition ecology - 95% of deaf children are born to hearing, non-signing parents and lack rich language models. This makes it challenging to understand what aspects of turn-taking development are linked to (1) language modality, i.e. signed vs. spoken language, and (2) language ecology, i.e. languages where one-to-one vs. one-to-many conversations are the norm. The LETITIA project contributes towards disentangling these threads by studying the development of turntaking
in an extraordinary context: a Balinese village, characterised by high intergenerational deafness and many signers of the local sign language, Kata Kolok. In this setting, communal child-rearing practices lead to most conversations available to children involving many interlocutors. Investigating how deaf children who grow up in such a signing-rich environment develop this fundamental pragmatic skill is unprecedented and affords us the unique opportunity to (1) study sign language acquisition in an ecology that is in many ways similar to that of spoken language acquisition, (2) pioneer in applying methods from social interaction to sign language acquisition data, and (3) begin to disentangle effects of modality and language ecology on language development. To better understand how a signing-rich environment shapes deaf children’s development of turn-taking, this project pursues three research objectives (RO) through two related research questions (RQ):
RQ1: What are the characteristics of turn-taking among adult and child signers?
RO1: document the typological and pragmatic characteristics of turn-taking in Kata Kolok.
RO2: track the developmental trajectory of turn-taking skills in deaf Kata Kolok acquiring children.
RQ2: How do conversational settings with many signing interlocutors modulate the turn-taking skills of signing adults and children?
RO3: explore how the turn-taking skills of adult and child signers are modulated by conversational factors, specifically age of conversation partners (adult vs. child) and number of interlocutors (two vs. multiple).
In the 12 months of the project, I have created and piloted a detailed coding scheme that can be used for both adult and child data from the Kata Kolok Child Corpus and have used it to annotate part of the spontaneous data related to WP1 and initial annotations for WP2. I have also obtained ethical approval for the project to the Swedish Ethical Authority.
Results at this stage are preliminary as annotation of all data has not been completed. The results suggest that polar questions are highly frequent in adult-adult conversations in Kata Kolok, with a tendency to ask test-like questions even among adults. This is unusual as test questions are usually particularly frequent in adult-child interactions. However, we are dealing with adult participants who have a high degree of shared knowledge and shared context and yet, the number and type of confirmation questions that pretend to be information seeking but are likely to be already known is surprisingly high. Another aspect that might become critical is the amount of overlap between question and answer sequences that appears quite higher than previously reported. This might also be related to the shared context facilitating easier prediction of sequences. How this translates to child-directed and child-produced question-answer sequences is yet to be confirmed as the data annotation is incomplete.