Keywords: non-native speech; non-native discourse; rhythm; advanced proficiency levels; language proficiency

This research is grounded in integrated rhythmology, a school of thought that regards rhythm as a multifaceted system structured by different language strata and means [Antipova, 1984; Teixeira, 2023]. Rhythm is also viewed from the perspective of non-nativeness. Non-native discourse is a speech performance marked by non-native language features. Within this domain, rhythm is the least explored phenomenon [Gut, 2012].
This study aims to describe the two fundamental units of rhythmic hierarchy, rhythmic structures (RSs) and rhythmic schemes of syntagms (RSSs), and to investigate how these units are redistributed among advanced learners at B2, C1 and C2 language proficiency levels.
The corpus was collected from October through December 2024. It comprises approximately 20 minutes of oral speech. Each participant (n = 13) reads out a descriptive narrative passage taken from English examination preparation materials. Participants were assigned to language groups using a multistage procedure that included an online English proficiency test, two speaking and writing tasks and a placement interview.
The research encompasses several methods. Among them is Zlatoustova’s approach, involving the classification of RSs (the number of syllables and the position of the stressed syllable) and the analysis of RSSs (the number of RSs within a syntagm). For the statistical analysis of the data, Manhattan (L1) and Euclidean (L2) distances are used to find out how the chosen groups differ from one another.
The results of calculating Manhattan and Euclidean distances demonstrate that the smallest distance values are observed between groups C1 and C2 at the level of RSs (L1 = 45; L2 = 12.5), indicating a high degree of similarity in the distribution of these units. However, group B2 shows significant differences from groups C1 (L1 = 98; L2 = 28.5) and C2 (L1 = 89; L2 = 27.3). This suggests that the rhythmic organisation of speech in group B2 has a set of specific rhythmic features that set it apart from groups C1 and C2, while these groups form a more homogeneous cluster.
A similar analysis conducted for RSSs confirms the identified trend. The distances between groups C1 and C2 are minimal (L1 = 29; L2 = 10.4), underscoring a homogeneity in the distribution of RSSs. Group B2 shows a significant distance from C1 (L1 = 72; L2 = 47.7) and C2 (L1 = 69; L2 = 35.8) groups. Hence, both types of analysis, at the level of RSs and RSSs, indicate that groups C1 and C2 form a close cluster regarding speech rhythm patterning, whereas group B2 represents a distinct group with its own rhythmic features.
The current research attains its primary goal of describing RSs and RSSs across advanced proficiency levels. The results suggest that, when analysed using Manhattan and Euclidean distances, speech rhythm can serve as an indicator of language proficiency.

References:
Antipova A. M. The Rhythmic System of English Speech. Moscow, 1984.
Gut U. Rhythm in L2 Speech // Speech and Language Technology. 2012. 14 (15). 83—94.
Teixeira L. A. S. A Multidimensional Analysis of English-L2 Rhythm Development // Ilha do Desterro. 2023. 76 (3). 223—250.