A computational model to investigate assumptions in the headturn preference procedure

Abstract

In this paper we use a computational model to investigate four assumptions that are tacitly present in interpreting the results of studies on infants' speech processing abilities using the Headturn Preference Procedure (HPP): (1) behavioral differences originate in different processing; (2) processing involves some form of recognition; (3) words are segmented from connected speech; and (4) differences between infants should not affect overall results. In addition, we investigate the impact of two potentially important aspects in the design and execution of the experiments: (a) the specific voices used in the two parts on HPP experiments (familiarization and test) and (b) the experimenter's criterion for what is a sufficient headturn angle. The model is designed to be maximize cognitive plausibility. It takes real speech as input, and it contains a module that converts the output of internal speech processing and recognition into headturns that can yield real-time listening preference measurements. Internal processing is based on distributed episodic representations in combination with a matching procedure based on the assumptions that complex episodes can be decomposed as positive weighted sums of simpler constituents. Model simulations show that the first assumptions hold under two different definitions of recognition. However, explicit segmentation is not necessary to simulate the behaviors observed in infant studies. Differences in attention span between infants can affect the outcomes of an experiment. The same holds for the experimenter's decision criterion. The speakers used in experiments affect outcomes in complex ways that require further investigation. The paper ends with recommendations for future studies using the HPP.

Development of infants' segmentation of words from native speech: a meta‐analytic approach

“…Similarly, and somewhat closer to the skill discussed here, Gerken, Balcomb and Minton (2011) document that infants look longer when they are listening to stimuli from which a rule can be discovered compared to very similar sequences where no such rule is present. Put more generally, in the context of word segmentation from native speech, infants' attention may be better predicted by other factors, such as the characteristics of the stimuli (see also Aslin, 2007;Bergmann, ten Bosch, Fikkert & Boves, 2013).…”

Making Sense of Infant Familiarity and Novelty Responses to Words at Lexical Onset

“…For example, although words are often explored in novelty/familiarity paradigms either in isolation (e.g., Hallé and de Boysson-Bardies, 1994 , 1996 ; Vihman et al, 2004 ; Swingley, 2005 ) or in passages of sentences containing target words (e.g., Jusczyk and Aslin, 1995 ; Jusczyk et al, 1999 ; Bortfeld et al, 2005 ; Singh, 2008 ; Singh et al, 2012 ; DePaolis et al, 2014 ), there are few methodological examinations of familiarity and novelty as they apply specifically to the developing lexicon. One exception is a computational model of factors affecting word segmentation in AHPP experiments ( Bergmann et al, 2013 ). Another is a study ( DePaolis et al, 2013 ) that found that 12-month-old infants’ preference for non-words was linearly related to the number of consonants each infant produced that were featured in the test stimuli; effectively, the infants showed either familiarity or novelty, based upon the extent of their previous practice with the test stimuli (see DePaolis et al, 2011 and Majorano et al, 2014 for similar differences in infant preference based upon their babbling patterns).…”

Top-down versus bottom-up theories of phonological acquisition: A big data approach

“…In the whole database, only 6 studies delay the test phase and/or use exposures in the child's home; and none of them were part of the final analyses given our constraint of including only studies that measured two age groups with the same stimuli. Further, since infants are usually familiarized and tested in close succession, it might be sufficient for them to perform a sort of phonetic pattern matching, which need not employ symbolic representations and/or their long-term memory [24]. We hope future work will help address that this important gap in the current literature.…”