The Presence of Background Noise Extends the Competitor Space in Native and Non‐Native Spoken‐Word Recognition: Insights from Computational Modeling

Abstract

Oral communication often takes place in noisy environments, which challenge spoken-word recognition. Previous research has suggested that the presence of background noise extends the number of candidate words competing with the target word for recognition and that this extension affects the time course and accuracy of spoken-word recognition. In this study, we further investigated the temporal dynamics of competition processes in the presence of background noise, and how these vary in listeners with different language proficiency (i.e., native and non-native) using computational modeling. We developed ListenIN (Listen-In-Noise), a neural-network model based on an autoencoder architecture, which learns to map phonological forms onto meanings in two languages and simulates native and non-native spoken-word comprehension. We also examined the model's activation states during online spoken-word recognition. These analyses demonstrated that the presence of background noise increases the number of competitor words, which are engaged in phonological competition and that this happens in similar ways intra and interlinguistically and in native and non-native listening. Taken together, our results support accounts positing a "many-additional-competitors scenario" for the effects of noise on spoken-word recognition.

Noise Modulates Crosslinguistic Effects on Second-Language Auditory Word Recognition

“…The same negative effect of a large PhonNDtrans for cognates compared to non-cognates was found for both accuracy and reaction times. This is consistent with prior work that shows effects of competition on both accuracy and reaction time measures (Karaminis et al, 2022).…”

Recognizing non-native spoken words in background noise increases interference from the native language

“…Using the visual-world paradigm, where participants' eye movements to objects are tracked as they listen to input related to a visual scene (for a review, see Huettig et al, 2011a), we found that non-native target word recognition (looks to the picture of a candle as Dutch listeners heard "candle") was delayed when targets were masked by background noise, reflecting delayed target word recognition (Hintz et al, 2021). Moreover, in a word transcription task, Scharenborg et al (2018) found that the number of Dutch listeners' unique misperceptions, an offline measure reflecting the number of words competing for recognition, on hearing an English target increased as the severity of noise on the target increased (see Karaminis et al, 2022, for a computational model capturing this behavior). The results from these two studies are compatible with a phonetically-based account of word recognition in noise where the presence of noise enhances ambiguity in the nonnative speech signal, which in turn delays non-native target word recognition and increases the number of misperceptions.…”

A network model of referent identification by toddlers in a visual world task

“…The semantic outputs reflect real-world corpora insofar as they are vector representations taken from the GloVe model pre-trained on aggregated global word-word cooccurrence statistics from a 6 billion token corpus (Pennington, Socher, & Manning, 2014). Phonological inputs consist of an unfolding sequence of standard phonological features used to uniquely identify each word in the toddler's vocabulary (Karaminis, 2018).…”

“…Dynamic phonological representations are constructed from encodings of the phones making up the item's label. Each phone is assigned a distributed binary encoding based on 20 articulatory and phonological features (Karaminis, 2018). The dynamic phonological representation for each vocabulary item is a matrix composed of the phonological feature representations of its phones in the order in which they appear as the spoken word unfolds, in which each row corresponds to a time step in the unfolding word.…”