The Perception of Laryngeal and Length Contrasts in Early Language Acquisition

Abstract

The present thesis investigates (Standard) German and Swiss German infants' perception of a laryngeal and a length contrast in labial stops. German has a laryngeal contrast with voiceless unaspirated stops opposed to voiceless aspirated stops. The primary phonetic correlate is VOT, which, following Mikuteit (2006), is referred to as 'after closure time' (ACT). Instead of a laryngeal contrast, Swiss German has a length contrast between short stops (singletons) and long stops (geminates). The contrast is phonetically realised by closure duration (CD).Regarding early speech perception, it is assumed that infants start life with universal perception skills, enabling them to discriminate nearly all kinds of phonetic contrasts, irrespective of the phones' relevance for the language the infants are about to acquire. At the end of the first year of life, infants' perception adapts to the native phoneme inventory (e.g., Werker & Tees 1984). In a series of experiments, it is examined whether such a developmental pattern is found for German and Swiss German infants listening to a laryngeal contrast (native in German, non-native in Swiss German) and a length contrast (native in Swiss German, non-native in German).The first part of the thesis discusses phonetic and phonological fundamentals as well as the relevant acquisition literature. The second part comprises the presentation of new empirical work and starts with a pilot production study with adults that confirms that ACT is the primary cue to make a contrast between pretonic labial stops in German. A categorisation study also conducted with adults reveals that Germans rely on ACT but fail to use CD to perceptually contrast two stop categories whereas Swiss Germans are able to rely on both ACT and CD to distinguish stops categorically. The perception study serves as a basis for the infant experiments as it shows where the respective native phoneme boundaries are located.A precursory infant experiment with German 6- to 8-month-olds demonstrates that the Switch Procedure, the method used for all infant tests in the present thesis, works well. New ways of analysing the data are introduced, which allow detecting effects that might be overlooked with the common analysis. Experiments with German and Swiss German 6- to 8-, 10- to 12- and 14- to 16-month-olds examined infants' ability to discriminate a laryngeal and a length contrast.The findings support the assumption that ACT contrasts are universally discriminable in the first half year of life. Moreover, the data suggest that consonantal length contrasts are different from most other contrasts in early perception. The CD contrast is not discriminable at the age of 6 to 8 months. Instead, the results suggest that it has to be acquired with linguistic experience, similar to some other contrasts of low acoustic salience. 14- to 16-month-old infants' failure to discriminate a respectively native phoneme contrast is argued to be due to a phase of perceptual uncertainty at the beginning of the second year of life, which might be triggered an increasingly variable linguistic input. Moreover, German 10- to 12- and 14- to 16-month-olds were not expected to discriminate a (presumed non-native) length contrast. An additional production study with German adults provides two potential explanations for this ability. Either infants are more sensitive to CD as a cue for phoneme categorisation than adults or the infants are using CD differences to detect word boundaries. Swiss German infants' ability to discriminate the non-native laryngeal contrast is in line with the adult data and interpreted with reference to typological aspects of the language. Apart from that, perception studies with the infants' parents confirm the appropriateness of the infant stimuli.In sum, the thesis provides new data on adults' production and perception of laryngeal and length contrasts. Insights into German and Swiss German infants' discrimination skills reveal that the two types of stop contrasts follow different developmental paths in early speech perception. Thus, not only by including the hardly-examined length contrast the thesis crucially contributes to the debate on early infant speech perception. Furthermore, it presents new insights on infants' speech perception in the second year of life. Finally, the thesis introduces new analyses for the data collected with the Switch Procedure which help to better assess infants' perception skills.