Multidimensional analysis (MDA) is a methodological approach used to describe the underlying three-way functional relationship between situational characteristics, linguistic structures, and communicative function. The MDA approach was pioneered by Biber (1988), who used it to describe linguistic variation in spoken and written English in terms of linguistic features co-occurrence. In recent years, the MDA approach has experienced considerable methodological growth. This chapter begins with an overview of five MDA studies that played a pivotal role in the historical development of the MDA approach through the introduction of novel methods, applications, and findings. In this overview, the chapter focuses on the novel aspects of each study and the impact that it had on the field. It then discusses the current methodological state of the art, focusing on recent innovative developments occurring at three crucial stages of an MDA: selection of variables for analysis, selection of lexico-grammatical tagger, and selection of the appropriate statistical technique. The chapter concludes with a case study representing the current state of the art in MDA research.

This chapter examines corpus-based research on intonation and prosodic features, focusing on tune patterns, f0 range, intonational phrasing, and speech rate, alongside tools, annotation systems, and corpus linguistic methods. It highlights multidimensional analysis (MDA; Biber, 1988) and the integration of linguistic and extra-linguistic factors (e.g., gender, speaking style; Gut, 2009; Staples & Puga, 2022; McClellan, 2024), addressing challenges and solutions in these areas. To illustrate research in this area, the case study using Staples and Puga’s (2022) data demonstrates the integration of prosody, fluency, and lexico-grammatical features via MDA, offering insights into functional language use among L1 and L2 English speakers. Data from the LINDSEI and LOCNEC corpora, comprising 259 files from diverse L1 and L2 groups, were analyzed. Results revealed significant differences in fluency, prosodic patterns, and discourse functions. Dimension 1 (D1) showed L1 speakers produced more fluent, informational speech, while L2 groups exhibited simplified, less fluent discourse, highlighting the need to incorporate informational language choices into spoken discourse instruction. Gender also influenced D1 scores: male speakers spoke faster and used fewer high-ending tunes. Findings emphasize distinctions across gender and linguistic groups, advocating for advanced methods that integrate linguistic and social variables to enhance language proficiency assessment and teaching.