Statistical learning of multiple speech streams: A challenge for monolingual infants

Abstract

Language acquisition depends on the ability to detect and track the distributional properties of speech. Successful acquisition also necessitates detecting changes in those properties, which can occur when the learner encounters different speakers, topics, dialects, or languages. When encountering multiple speech streams with different underlying statistics but overlapping features, how do infants keep track of the properties of each speech stream separately? In four experiments, we tested whether 8‐month‐old monolingual infants (N = 144) can track the underlying statistics of two artificial speech streams that share a portion of their syllables. We first presented each stream individually. We then presented the two speech streams in sequence, without contextual cues signaling the different speech streams, and subsequently added pitch and accent cues to help learners track each stream separately. The results reveal that monolingual infants experience difficulty tracking the statistical regularities in two speech streams presented sequentially, even when provided with contextual cues intended to facilitate separation of the speech streams. We discuss the implications of our findings for understanding how infants learn and separate the input when confronted with multiple statistical structures.

One language or two? Navigating cross‐language conflict in statistical word segmentation

“…This was recently replicatedwith infants by Benitez, Bulgarelli, Byers-Heinlein, Saffran, and Weiss (2019), who found that 8-monthold monolinguals were unable to learn two sequentially presented artificial languages with syllable overlap regardless of whether they were tested on the first or second language or whether the languages were differentiated by vocal pitch or speaker accent. This finding aligns with monolingual performance in the dual language task from Antovich and Graf Estes (2018).…”

Dynamics of nonlinguistic statistical learning: From neural entrainment to the emergence of explicit knowledge

“…This finding suggested that statistical learning plays an important role in language acquisition, and spurred a large body of additional work in this area, not only in developmental populations (e.g. Pelucchi et al, 2009;Fló et al, 2019;Benitez et al, 2020), but also in adults (e.g. Saffran et al, 1996b;Cunillera et al, 2009;Batterink and Paller, 2017).…”

Dual language statistical word segmentation in infancy: Simulating a language‐mixing bilingual environment

“…Previous studies have explored how speaker-specific contextual cues (e.g., two individual speakers differing in gender (Weiss et al, 2009), or accent (Benitez et al, 2020), or languages associated with two distinct faces (Mitchel & Weiss, 2010) facilitate word segmentation in two languages. However, in naturalistic bilingual environments, different speakers do not always speak distinct languages.…”

“…Three recent studies have shed light on whether infants can statistically segment words from two language inputs. Benitez et al (2020) reported that 8-month-old monolingual infants failed to segment words in a dual-input segmentation task that lacked contextual cues to differentiate the two languages. In follow-up studies, they investigated whether adding distinguishing contextual cues (e.g., differences in pitch, speakers, and person-based accent) could help infants succeed.…”