Blurb:

The promise of the Knowledge Framework is immense, and its potential can be further realized through corpus studies. To obtain compelling and conclusive corpus-based findings, however, it is important to attend to major methodological challenges inherent to applying this framework to the analysis of large collections of text data. One such challenge is posed by qualitative corpus annotation, especially when the corpus is composed of spoken learner language. As a starting point to a scholarly forum on streamlining the methods and laying the groundwork for replicability and reproducibility of Knowledge Framework analysis, this chapter describes the methodology of a corpus-based project investigating the knowledge structures and language functions in spoken discourse produced by non-native speakers of English in two different types of situations—teaching and testing. In parallel with that, it provides rationales for specific procedural steps in view of latent challenges. The methodological rigor pursued in this study resulted in a set of tangible outcomes applicable to future research, including descriptors of functional realizations and linguistic instantiations of the knowledge structures and an annotation protocol. Additionally, these outcomes have important implications for language teaching and assessment, as they are indicative of the functional language ability that non-native English speakers need in order to be effective communicators in instructional settings.