Frequency effects and markedness in phonotactics

Abstract

In this paper, we take up the challenge of exploring the relationship between markedness and frequency in phonotactics. The study is based on word-initial and word-final consonant clusters in Polish and English. The aim of this study is threefold. First, we establish logarithmic frequencies for word-initial and final consonant clusters compiled from two resources, a dictionary (or paradigm) and a written corpus. Second, we examine the preferability status of clusters in three frequency bands (high, mid, low) in terms of two phonotactic principles, i.e. sonority and Net Auditory Distance. Finally, we test the correlations between degrees of markedness and frequency. The present paper extends our previous studies on comparative Polish–English phonotactics, where markedness and frequency constitute the core of the analysis. The study shows that there is no relationship between cluster markedness and its frequency. As to frequencies, Polish and English differ from each other with respect to the distribution of clusters in the dictionary list, while the disproportions are neutralized in usage.

On the Shapes of the Polish Word: Phonotactic Complexity and Diversity

“…/t ͡ ʂʨ-/ in czcić 'to worship', /kt-/ kto 'who' or /rv-/ in rwać 'to tear' (these tautomorphemic examples demonstrate that sonority plateaus and reversals are attested, and although they are statistically rarer, they are not exceptional and can be very frequent in language use). Cluster markedness does not correlate with token frequency (Orzechowska & Zydorowicz 2019). There are numerous clusters which are considered universally dispreferred and are still very common in language use in Polish, e.g., /kt-/ in który 'which' or kto 'who' and /ɡd-/ in gdy 'when'.…”