The acquisition of complex onsets in Icelandic: the effects of markedness, sonority, and frequency

Abstract

The acquisition of consonant clusters is determined by several underlying factors, and recent studies demonstrated that input frequency, investigated in isolation, cannot account for the order of acquisition (e.g., Jarosz 2017). The study conducted in this thesis was an empirical test of the theories of phonological acquisition. It investigated the simplification patterns occurring in initial clusters and the role of sonority, type frequency, and articulatory complexity of individual segments on the order of their acquisition. To investigate the influence of these factors on accuracy in producing initial consonant clusters, I analyzed data from 68 monolingual children (aged 2;6-4;3) acquiring Icelandic, available as a part of the Icelandic Másdóttir corpus (Másdóttir 2014, Másdóttir et al. 2021), which is a part of the PhonBank corpora (Rose & MacWhinney 2014). The investigation of simplification patterns revealed a discrepancy in processes targeting branching onsets and sC clusters. Additionally, the results showed that Icelandic children produced clusters composed of stop and liquid as two segments most frequently. Fricative-initial clusters were most frequently realized as one consonant. I demonstrated how two sonority-based generalizations, the Minimal Sonority Distance and the Sonority Dispersion Principle, did not account for all the tendencies in the order of acquisition, and could not predict children’s low accuracy on fricative-initial clusters. The type frequencies of initial onset clusters in Icelandic did not correlate with children’s accuracy. The findings revealed that what has previously been attributed to the role of sonority, could be explained on the basis of articulatory complexity of individual cluster members, or possibly perceptual cues.