Jun 13 ====== Saffran/etal:1996 ----------------- Niranjan Can we attribute any other strategies of learning apart from statistical learning for any aspects of language acquisition? What are the aspects of experience-independent mechanisms? Mihan How might the duration of exposure affect language development? (According to the study in which the researchers exposed infants to only 2-minute speech.) I think there are also two limitations in this study: one of them is using synthesized speech instead of natural daily dialogues. The other one is that infants were exposed to this synthesized speech without natural linguistic cues such as pauses, intonation, or stress patterns. However, it would be better if the study included these features to check infant´s abilities in segmenting words. Marc The paper mainly deals with experiments concerning word segmentation. Is one type of task enough to investigate the proportions of experience-dependent and experience-independent mechanisms involved during the acquisition of language? Are studies conducted with the help of artificial language insightful concerning the acquisition of natural language? The way children learn an artificial language is for example rather related to their short-term memory, whereas learning a natural language leads to a profound and lasting proficiency in the acquired language. Anina How scalable can we expect this capacity for statistical learning to be? Assuming that real-world input is much sparser and extensive in quantity, this would open a wide space of transitional probabilities to keep track of. How can such information be stored? Nataliia According to the authors, it is unclear whether the statistical learning observed is domain-specific to language or a domain-general cognitive mechanism. What kind of experiments could effectively differentiate between these two possibilities? If statistical learning is found to be domain-specific, would this support the Universal Grammar theory? Ema Observation: After reading the conclusions of the article, I thought to myself that it does not necessarily negate the importance of innate knowledge in the language acquisition process. Instead, it could support the theory that innate mechanisms serve as a blueprint for language learning. This blueprint provides general strategies and constraints, rather than specific knowledge. For example, parameters like "head-first" vs. "head-last" in syntax are set based on linguistic exposure. The fact that 8-month-old infants can extract relevant information and recognize structures after a brief exposure suggests that there must be some innate mechanisms guiding them, such as knowing to look for word boundaries. Question: Don't the results of this study actually support the theory of innate knowledge guiding language acquisition? Given that 8-month-old infants can use short input to extract and apply relevant information, could this indicate that there are innate mechanisms and skills that help them in this process, such as knowing to search for word boundaries? How might these findings be reconciled with the idea that infants have an innate blueprint for language learning that includes general strategies and constraints?