Cross-linguistic variation in word-initial cluster production in adult and child language: evidence from English and Norwegian

Abstract

Young children simplify word initial consonant clusters by omitting or substituting one (or both) of the elements. Vocalic insertion, coalescence and metathesis are said to be used more seldom (McLeod, van Doorn & Reed, 2001). Data from Norwegian children, however, have shown vocalic insertion to be more frequently used (Simonsen, 1990; Simonsen, Garmann & Kristoffersen, 2019). To investigate the extent to which children use this strategy to differing degrees depending on the ambient language, we analysed word initial cluster production acoustically in nine Norwegian and nine English speaking children aged 2;6–6 years, and eight adults, four from each language. The results showed that Norwegian-speaking children produce significantly more instances of vocalic insertions than English-speaking children do. The same pattern is found in Norwegian- versus English-speaking adults. We argue that this cross-linguistic difference is an example of the influence of prosodic-phonetic biases in language-specific developmental paths in the acquisition of speech.

Vocalic Intrusions in Consonant Clusters in Child-Directed vs. Adult-Directed Speech

“…However, if we consider not just the phonological structures but also how they are implemented , we can model actual divergence in inputs, and make different predictions about the children’s early productions. Indeed, in the same study ( Garmann et al, 2021 ), we found that, while vocalic intrusions were also evident in English child speech [as one possible strategy for tackling clusters, and one that is attested to some degree cross-linguistically ( McLeod et al, 2001 )], they were far less prevalent than in Norwegian child speech. This could indicate that in selecting a strategy for tackling clusters infants are influenced by distributional patterns in their ambient speech input.…”

“…Here, we investigate the role that CDS may have in mediating the prosodic-phonetic bias of vocalic intrusion in Norwegian in the child acquisition process. In our earlier study, we found that child speech displays more vocalic intrusion in consonant clusters than does ADS ( Garmann et al, 2021 ), which suggests vocalic intrusion may be more prevalent in CDS than in ADS. This could come about quite incidentally: since vocalic intrusions may be influenced by the slower tempo and exaggerated prosody that characterise CDS, we hypothesise that CDS has longer and more frequent vocalic intrusions than ADS.…”

“…Looking at the duration of vocalic intrusions between CDS overall (median = 29 ms) and CS overall (median = 43 ms), the former was significantly shorter than the latter (W = 811, p < 0.001). As can be seen from Figure 2, there may be a development corresponding to what occurs in children's speech, where Garmann et al (2021) reported a significant reduction in duration with age. However, there is a lot of variation masking possible changes in CDS over time: Investigating the duration of vocalic intrusions in the CDS data divided by the children's age, we found no significant differences between the CDS addressed to 2;6-, 4-, and 6-year-olds.…”

“…This often leads to vocalic intrusions between a sequence of consonants in a cluster in adult speech (e.g., [1bᵊɭoː] for [1bɭoː] blå , ‘blue’). In a recent study, we found vocalic intrusions in 30.6% of instances for (Urban East) Norwegian, while for (Southern British) English, vocalic intrusions were nearly non-existent (affecting only 0.6% of instances) ( Garmann et al, 2021 ). In the same study, vocalic intrusions were found to be more common in ADS when C2 is a liquid, with the greatest incidence occurring when C2 is a rhotic tap or flap, and a particularly high incidence when this was combined with a voiced stop in C1 position.…”

“…Audio was recorded using a Zoom Handy Recorder H2 with built-in microphones. The mothers and their children took part in a larger study comparing phonetic microvariation in the production of consonants in both child and adult speech in English and Norwegian, reported on by Payne et al (2015Payne et al ( , 2017 and Garmann et al (2021). The studies were reviewed and approved by NSD -Norwegian Centre for Research Data under the reference number 36466.…”