Strata in Systemic Functional Linguistics
Levels, Layers, Planes, Domains
Elissa D. Asp [+–]
Saint Mary’s University
Elissa D. Asp is Professor of English and Linguistics, and Coordinator of Linguistics at Saint Mary’s University, Halifax, NS. She was educated at Glendon College and York University in Toronto, where she specialized in linguistic description of discourse and linguistic theory. She is interested in functional and formal theories of language and developing models that elucidate language processing in context. Ongoing research addresses three main areas: (1) language and discourse correlates of dementias, especially neurodegenerative diseases associated with ageing; (2) MEG studies of neurocognitive networks supporting language and discourse processing; (3) theoretical implications of these studies for models of language.
One of the central tasks of linguists is to characterize and explain relations between form and meaning in language. Linguists differ in how they go about this according to their theoretical orientations and practical goals. However, they do more or less converge on broad levels or domains – semantics, syntax, morphology, phonology, and phonetics – as relevant to any account. Vocabulary may be treated separately or subsumed as part(s) of other levels or domains, and in functional models, discourse and aspects of context may also be posited as a levels or domains. A large part of the architecture of different linguistic theories has to do with how levels and domains and relations between them are described and which ones are viewed as defining language.
Within systemic functional linguistics (SFL), different levels of representation are defined as distinct ‘strata’ and include discourse and context levels. Treatment of relationships between different levels varies but instantiation and realization are central types. The importance attributed to stratification in SFL can be seen in Michael Halliday’s assertions that it is the separation of form from meaning via lexicogrammatical stratification that distinguishes proto languages from adult human languages. The inclusion of context as a stratum foregrounds a conceptualization of language as an evolved open dynamic system – embedded in and shaped by the biological and social contexts in which it develops and is used.
Strata in Systemic Functional Linguisticsoutlines the historical development of the idea of ‘strata’ in relational models grounded in European structuralism and functionalism, and charts the development of stratified models within the family of SFL descriptions. It compares SFL with other functional and formal linguistic approaches to linguistic levels in order to explicate what a stratal architecture does and does not capture about language. It also probes the neurocognitive plausibility of stratified architectures in SFL and their potential to inform processing perspectives.
Table of Contents
Chapter 1
This chapter provides an overview of the concept of linguistic strata and potential relations between them as compared with related concepts such as modules and domains.
Chapter 2
Chapter two outlines the development of the idea of strata beginning with 19th century aphasiological models which introduced the idea of a language faculty with distinct components contributing different functions and following with the work of de Saussure on the distinction between signifier and signified and the relational character of the sign, Hjelmslev’s elaboration of the relational character of language into distinct ‘planes’, Lamb’s adaptation of Helmslev’s model in stratificational linguistics, and Halliday’s adaptation of strata to his systemic model.
Chapter 3
This chapter sketches the various approaches to the stratal organization of language within the SFL family of models. These include the Sydney and Cardiff grammars, but also the work of Gregory, Martin and others on discourse and contextual levels. The goal is a characterization of the common core as well as of the variety of views with respect to strata and relationships between them.
Chapter 4
This chapter explores a range of other approaches to the phenomena handled by strata in SFL models in order to foreground how stratified models actually differ and where there are overlaps. Other approaches discussed include cognitive construction grammar, functional discourse grammar, the parallel architecture model, role and reference grammar, and word grammar among functional grammars and minimalist, cartographic and parametric substantiation perspectives among formal models.
Chapter 5
This chapter considers whether stratal organization matters when it comes to potential and actual applications. Arguments are made that the utility of strata are reflected primarily in textual analysis and socio-cultural critique through discourse analysis that is enabled by linking discourse and context to the grammar on the one hand, and work on text generation and related computational tasks on the other. The relevance for text analysis and socio-cultural critique means that in practice stratified models have had very broad influence in fields such as education and areas where ‘designed discourse’ has roles. The relevance for computational tasks is perhaps more specific to the grammar insofar as, for example, the treatment of morphosyntax or lexicogrammar as realizing semantic selections simplifies morphosyntax. Stratified models may also have potential to be informative about neural injuries that affect language and discourse though this potential is underdeveloped in SFL work in clinical linguistics.
Chapter 6
This chapter reviews evidence that suggests current and on-going relevance of stratal perspectives for understanding neurocognition but also highlights research which suggests revisions to stratal models. Stratified perspectives appear at the heart of many neuroscience models of speech processing and are at least partially supported by evidences from neurostructural damage and neurophysiologyical evidences derived from imaging. However, neurophysiological evidence also suggests that brains are, in some ways, less ‘tidy’ in their activity patterns than the elegant models we make. They allow both feedback and feed forward activity and even formerly ‘unimodal’ sensory areas allow input from networks associated with other modalities. So, for example, motor representations of articulation may support auditory decoding, auditory perception may be influenced by visual stimuli and so on. SFL as a stratified systems model – where selections on one stratum predict or activate selections on adjacent strata – may be particularly valuable in research investigating such ‘bottom up’ and ‘top down’ modulations in hierarchical generative paradigms for speech and discourse production and comprehension.
Chapter 7
Conclusions [+–]
The conclusion summarises the main points. Strata are here to stay. However, there is a need to revisit relations between the various ‘levels’ that get represented as strata in SFL models and to ask whether some strata (such as discourse and context) mightn’t or couldn’t be better described by other constructs. At issue is the need to continue to develop the view of language as a dynamic open system embedded in socio-cultural contexts but also change, where change is warranted especially by changing understanding of socio-cognitive and neurobiological contexts and the intersections of these with socio-cultural contexts.
