Complexity

Complexity is both a construct describing the structural elaboration and variety of L2 output and one of three vertices of the CAF triad (accuracy, fluency, complexity) used to characterise second-language performance and proficiency. Housen and Kuiken (2009: 461–463) frame CAF as the dominant multi-componential answer to the question "what makes an L2 user more or less proficient", and complexity is the vertex most contested in its definition and measurement.

The CAF framework

The fluency–accuracy contrast entered ELT through Brumfit's (1984) distinction between fluency-oriented activities, aimed at spontaneous oral production, and accuracy-oriented activities focused on controlled, grammatical production. Skehan (1989, and more fully in his 1998 A Cognitive Approach to Language Learning) added complexity as a third dimension, giving the triangle that now anchors task-based performance research. Housen and Kuiken's (2009) Applied Linguistics special issue is the canonical reference point: CAF as performance descriptors for assessment, as proxies for underlying L2 proficiency, and as indicators of developmental change.

Standard readings treat fluency as the real-time mobilisation of the L2 system (speed, hesitation, repair), accuracy as conformity to target-language norms (error-free clauses, error rates), and complexity as the extent to which learners produce elaborated, varied, and non-default language. Complexity is the outlier: unlike accuracy, it has no external correctness criterion, and unlike fluency, its measurement is not tied to a clear behavioural signal like pause or rate.

Skehan's (1998) Limited Attentional Capacity model, also known as the Trade-off Hypothesis, argues that attentional resources during real-time production are finite, so gains on one CAF dimension are often bought at the cost of another. Skehan (2009: 510–532) develops this into a model linking task-based performance to Levelt's speech-production architecture, arguing specifically that joint gains in accuracy and complexity reflect separable task and task-condition factors rather than a single "push upward" from task difficulty.

Linguistic complexity as a construct

Bulté and Housen (2012, in Housen, Kuiken and Vedder's Dimensions of L2 Performance and Proficiency) decompose L2 complexity into a taxonomy with two top-level distinctions. Absolute (or objective) complexity concerns structural properties of the language itself (how many elements, how deep the embedding, how varied the inventory). Relative (or subjective) complexity concerns processing cost for a given learner. Within absolute complexity, they separate grammatical complexity (further split into morphological and syntactic), lexical complexity (density, diversity, sophistication), and, in some accounts, phonological and propositional complexity.

Operationalisation typically relies on ratio measures. Syntactic complexity is indexed by mean length of T-unit (MLTU), mean length of clause, or clauses-per-T-unit subordination ratios; lexical complexity by type-token ratio variants such as VOCD/D and MTLD, by sophistication measures pegged to frequency bands, and by density ratios over function versus content words. Norris and Ortega (2009: 555–578) audit these measures and show that many studies use a single subordination index and treat it as a proxy for "complexity" tout court, ignoring that phrasal, clausal, and subordination-level growth follow different developmental trajectories. They call for multidimensional, developmentally sensitive measurement built organically from data rather than imported wholesale from L1 child-language research.

Task complexity

Robinson (2001, 2011) uses "complexity" in a second sense: task complexity, the cognitive demands a task imposes on the learner. His Triadic Componential Framework distinguishes task complexity (cognitive factors built into task design), task conditions (interactional and participant variables), and task difficulty (learner-internal ability and affective variables). Task complexity itself splits into resource-directing dimensions (here-and-now vs there-and-then, few vs many elements, no reasoning vs causal reasoning) and resource-dispersing dimensions (planning time, prior knowledge, single vs dual task).

The Cognition Hypothesis predicts that increasing task complexity along resource-directing dimensions pushes learners to deploy more complex and more accurate language simultaneously, because those demands recruit the form-function mappings that development requires. Increasing complexity along resource-dispersing dimensions, by contrast, simply strains processing and degrades performance across the board. This is the direct point of divergence with Skehan: Robinson predicts joint rises in complexity and accuracy under the right task conditions, while Skehan predicts trade-offs. The empirical picture is mixed, and Skehan (2009) reinterprets apparent Cognition-Hypothesis wins as products of separate, additive task design and condition effects.

Critiques and disputes

Norris and Ortega (2009) argue that most CAF complexity research uses measures that cannot support the interpretations placed on them: single indices stand in for multidimensional constructs, ceiling effects distort comparisons across proficiency bands, and phrasal complexity, a key marker of advanced writing, is systematically under-measured. Pallotti (2009) pushes further on the conceptual side. He reserves "complexity" for structural intricacy and insists that structurally complex language is not automatically developmentally advanced or communicatively adequate. A learner who produces long, subordination-heavy sentences is not therefore closer to a target; a late-acquired feature is "difficult", not "complex". Pallotti's (2015) "simple view of linguistic complexity" extends this by arguing for complexity defined strictly in structural terms and measured accordingly, with development, difficulty, and adequacy treated as separate constructs.

Practical implications

For task-based language teaching, the two notions of complexity motivate different design moves. Sequencing by task complexity in Robinson's sense, through the SSARC model, pushes learners along resource-directing dimensions to elicit more elaborated language, while resource-dispersing demands are reduced at first and increased later to build automatised processing. For writing assessment and research, the Norris-Ortega and Pallotti critiques argue against treating any single syntactic ratio as a general proficiency index: reports should specify which sub-dimension of complexity is being measured (phrasal, clausal, subordination, lexical diversity, lexical sophistication) and pair that with evidence of developmental relevance for the learner group and L2 in question. In classroom terms, activities labelled as "complex" should be audited for whether they ask for more elaborated form-meaning mappings or merely impose more cognitive load without a linguistic payoff.

References

• Brumfit, C. (1984). Communicative Methodology in Language Teaching. Cambridge University Press.
• Bulté, B. and Housen, A. (2012). Defining and operationalising L2 complexity. In Housen, A., Kuiken, F. and Vedder, I. (Eds.), Dimensions of L2 Performance and Proficiency: Complexity, Accuracy and Fluency in SLA (pp. 21–46). John Benjamins. doi:10.1075/lllt.32.02bul
• Housen, A. and Kuiken, F. (2009). Complexity, accuracy, and fluency in second language acquisition. Applied Linguistics, 30(4), 461–473. doi:10.1093/applin/amp048
• Norris, J. M. and Ortega, L. (2009). Towards an organic approach to investigating CAF in instructed SLA: The case of complexity. Applied Linguistics, 30(4), 555–578. doi:10.1093/applin/amp044
• Pallotti, G. (2009). CAF: Defining, refining and differentiating constructs. Applied Linguistics, 30(4), 590–601. doi:10.1093/applin/amp045
• Pallotti, G. (2015). A simple view of linguistic complexity. Second Language Research, 31(1), 117–134. doi:10.1177/0267658314536435
• Robinson, P. (2001). Task complexity, cognitive resources, and syllabus design: A triadic framework for examining task influences on SLA. In Robinson, P. (Ed.), Cognition and Second Language Instruction (pp. 287–318). Cambridge University Press.
• Robinson, P. (2011). Second language task complexity, the Cognition Hypothesis, language learning, and performance. In Robinson, P. (Ed.), Second Language Task Complexity: Researching the Cognition Hypothesis of Language Learning and Performance (pp. 3–37). John Benjamins.
• Skehan, P. (1989). Individual Differences in Second Language Learning. Edward Arnold.
• Skehan, P. (1998). A Cognitive Approach to Language Learning. Oxford University Press.
• Skehan, P. (2009). Modelling second language performance: Integrating complexity, accuracy, fluency, and lexis. Applied Linguistics, 30(4), 510–532. doi:10.1093/applin/amp047