The role of task complexity, modality and aptitude in narrative task performance Judit Kormos and Anna Trebits Abstract The study reported in this paper investigated the relationship between components of aptitude and the fluency, accuracy, syntactic complexity and lexical variety of performance in two types of written and spoken narrative tasks. We also addressed the question of how narrative performance varies in tasks of different cognitive complexity in the written and spoken modes. Our findings indicate a complex interaction between aptitude components and task performance under different conditions. The components of aptitude that seemed to be most strongly related to the accuracy and complexity of production were deductive ability and grammatical sensitivity. The results also show that in writing the participants used more varied vocabulary than in speech, but their performance was similar in terms of syntactic complexity. 1 INTRODUCTION Research in the field of second language acquisition (SLA) has long been concerned with the question of why students show great variation in their language learning success. Studies in this area have concluded that individual differences (IDs) are the most important predictors of achievement in a second language (L2) (Dörnyei, 2005). Therefore, it is widely acknowledged that IDs have to be taken into consideration both in theoretical accounts of SLA and in practical pedagogical decision-making. Researchers often point out the necessity of making further advances into uncovering how certain IDs affect and underlie important language learning processes (Dörnyei, 2005; Kormos & Sáfár, 2008; Robinson, 2007a). Learner variables are usually divided into cognitive, affective and personality-related IDs (Dörnyei, 2005). One of the most important cognitive variables influencing the success of L2 learning is foreign language aptitude, which is assumed to be a conglomerate of cognitive factors that might potentially affect the success of language learning (Carroll, 1981). For a long time researchers were interested in the link between foreign language aptitude and global language learning outcomes (for reviews see Ehrman & Oxford, 1995; Grigorenko, Sternberg & Ehrman, 2000). In recent conceptualizations of foreign language aptitude, however, it is argued that different cognitive abilities might be useful in different phases and processes of language learning (Skehan, 2002) and that learners with different cognitive ability profiles might benefit from different types of learning tasks and instructional conditions (Robinson, 2005a). The novelty of our study is that we investigated how various components of language aptitude are related to performance in narrative tasks that differed in their cognitive complexity in both oral and written modes. For this purpose we administered the Hungarian version of the language aptitude test and two oral and written tasks to 44 upper-intermediate learners of English in a Hungarian secondary school. Most studies examining the subtle effects of task characteristics on L2 output conclude that in addition to general measures of L2 production, task-specific measures of production reveal more precise information about how tasks can direct learners’ attention to certain 2 linguistic forms and how IDs may differentiate the ways in which learners can benefit from the manipulation of certain task features (e.g., Kormos & Trebits, in press; Norris & Ortega, 2009; Robinson 2007b). Therefore, in our study we used both general and specific measures of performance in analyzing students’ output. The present study examines foreign language aptitude in relation to the cognitive demands of performing oral and written narrative tasks of different cognitive complexity. Therefore, we first discuss conceptualizations of foreign language aptitude and then review research on communicative tasks and task features with an emphasis on the construct of task complexity. Next, we give a brief summary of research on modality differences in task performance. Finally, we present the results of previous studies which have investigated the influence of IDs on language produced in different tasks. Language learning aptitude Foreign language (L2) aptitude is considered to be one of the best predictors of language learning success (Abrahamsson & Hyltenstam, 2008; Dörnyei, 2005; Nikolov & Ottó, 2006; Sawyer & Ranta, 2001). Carroll (1981) identified four components of language aptitude: i) phonetic coding ability, that is, the “ability to identify distinct sounds, to form associations between those sounds and symbols representing them, and to retain these associations”; ii) grammatical sensitivity, that is, the ability “to recognize the grammatical functions of words (or other linguistic entities) in sentence structures”; iii) rote learning ability, defined as “the ability to learn associations between sounds and meanings rapidly and efficiently, and to retain these associations”; and iv) deductive learning ability, which is “the ability to infer or induce the rules governing a set of language materials, given sample language materials that permit such inferences” (p. 105). Instruments developed to measure language aptitude, such as the Modern Language Aptitude Test (MLAT) (Carroll & Sapon, 1959) and Pimsleur’s Language Aptitude Battery (PLAB) (Pimsleur, 1966), test language learners on the 3 above mentioned four components (for a review see Abrahamsson & Hyltenstam (2008) and Robinson, (2007a)). The Carollian construct of language aptitude was developed in the period when audiolingualism was the prevailing language teaching methodology and when language competence was mainly conceived of as grammatical knowledge. New language teaching methodologies (in particular communicative and task-based approaches) aim to foster the acquisition of communicative competence, which encompasses a wider range of knowledge and abilities than linguistic competence alone. These changes in language teaching methodology and in the conceptualization of the goals in language learning have also necessitated a revision and extension of the concept of aptitude. In response to this, Robinson (2005a) proposed a dynamic aptitude construct, in which cognitive resources and abilities are combined into aptitude complexes. In Robinson’s (2005b) model, primary abilities include pattern recognition, speed of processing in phonological working memory and grammatical sensitivity. These general cognitive abilities, which with the exception of phonological memory, are based on the traditional construct of aptitude, help the so-called second-order abilities. The second-order abilities are specific to language learning and include noticing the gap, memory for contingent speech, deep semantic processing, memory for contingent text, and metalinguistic rule rehearsal. Robinson (2005b) also argued that explicit and implicit learning conditions might require different combinations of cognitive abilities. In his study, which investigated the role of aptitude in different learning conditions, he found a weak link between aptitude and outcomes in implicit learning conditions and a strong relationship between aptitude and learning under explicit conditions. De Graaff’s (1997) research, however, indicated that grammatical sensitivity and the ability to infer the meanings of words from a text were positively related to attainment in an artificial grammar learning experiment under both explicit and implicit conditions. Skehan (2002) also suggested that certain components of the traditional construct of aptitude such as grammatical 4 sensitivity and deductive ability might assist L2 learning in naturalistic contexts, where learners have few opportunities to acquire L2 linguistic rules through explicit explanation. From this brief review of recent research on aptitude it is apparent that several of the components of the original construct of aptitude as defined by Carroll might be relevant underlying cognitive abilities that promote language learning success even in today’s foreign language classrooms. Deductive learning ability and grammatical sensitivity might help learners recognize linguistic patterns in the communicative input, whereas rote learning ability might be one of the significant predictors of the success of vocabulary acquisition. Phonetic coding ability might also play an important role in the acquisition of the phonological system of the L2 and in L2 reading, in which one of the key abilities is phonological awareness (for a recent review see Grabe, 2009). It seems to be important to examine, however, not only how aptitude contributes to the overall success of language learning, but also to analyze how students with different cognitive abilities perform in communicative tasks that they commonly encounter in the process of L2 learning. Task complexity Cognitive abilities do not only assist in the acquisition of L2 knowledge, but they might also influence how learners utilize their acquired knowledge in performing different types of tasks. Consequently, it is important to investigate the relationships between the combination of abilities underlying L2 aptitude and the cognitive demands of pedagogic tasks (Robinson, 2005a). Such research aims to contribute to an understanding of how best to match learners with strengths in certain cognitive abilities to particular types of learning tasks by examining the interplay between the processing demands of pedagogic tasks and the components of aptitude. As cognitive processes may be closely linked to various features of task design (e.g., the availability of planning time, Ellis, 2005; Foster & Skehan, 1996; Yuan & Ellis, 2003, or the number of elements included in the task, Kuiken & Vedder, 2007), the findings of research on the interaction of individual difference factors 5 and task complexity may be used to inform pedagogic decisions in materials development and syllabus design. One of the most important issues in research on language learning tasks concerns the influence of task complexity on task performance. In Robinson’s (2001b) definition, “task complexity is the result of attentional, memory, and other information processing demands imposed by the structure of the task on the language learner” (p. 29). Robinson lists a number of task characteristics such as the number of elements, availability of planning time and prior knowledge, which influence the complexity of the task. In the case of speaking tasks, however, it would also be important to relate the complexity demands of tasks to the different stages of speech production. In Levelt’s (1989) model, speech production has four important components, which follow each other in this order: (1) conceptualization, that is, planning what one wants to say; (2) formulation, which includes the grammatical, lexical and phonological encoding of the message; (3) articulation, in other words, the production of speech sounds, and (4) self-monitoring, which involves checking the correctness and appropriateness of the produced output. In first language (L1) production conceptualizing the message requires attention, whereas formulation and articulation are automatic, and hence processing mechanisms can work in parallel, which makes L1 speech generally smooth and fast. In the case of non-balanced bilinguals and less proficient L2 speakers, however, formulation and articulation are often not sufficiently automatic and require conscious attention, which frequently hinders parallel processing. Task complexity is generally considered to derive from the cognitive demands a task makes in the conceptualization stage. This view is based on the assumption that complex concepts require the use of complex syntactic structures, and therefore cognitively complex tasks are complex both in terms of conceptualization and linguistic formulation (see e.g. Robinson’s (2001b, 2003, 2005b) Cognition Hypothesis). It is, however, possible that tasks make separate and independent complexity demands on the conceptualization and formulation stage. If we consider two speaking tasks such as the ones used in the current study ─ a cartoon description task, in which the storyline 6 is given, and a picture narration task, in which students have to narrate their own story ─ we can see that different aspects of cognitive complexity make different demands on the learners. In a cartoon description task, students do not need to conceptualize the content of the story, which eases the demands in the phase of conceptualization, that is, in selecting and ordering the relevant concepts of their message (Levelt, 1989; Skehan, 2009). Nevertheless, learners have to express the content prescribed by the task in whatever linguistic resources they have available in the L2, which increases the processing load in the linguistic encoding phase of speech production. In a story narration task, however, learners need to design their own story, but they can tailor it to match their linguistic resources, which results in increased conceptualization effort and a potentially reduced load in linguistic encoding. This example illustrates that in certain cases it might be difficult to order tasks in terms of cognitive complexity because tasks might make different and non- comparable demands on different phases of language production (see Pallotti (2009) and Skehan, (2009) for a similar line of argument). Consequently, it might increase our understanding of task complexity, if the cognitive demands of tasks were also considered separately for the conceptualization and linguistic encoding stages of speech production. A key issue in task-based language learning is that in performing a task students need to coordinate the allocation of their attention in order to successfully meet the linguistic demands of the task. This raises questions concerning how attentional resources can be used, coordinated and directed to different aspects of language production during task completion. There are two influential models of task complexity in this field, which have motivated a great number of studies that examine the effects of manipulating the different dimensions of task complexity on L2 output and interaction, and ultimately on L2 learning: Skehan and Foster’s (2001) Limited Attentional Capacity Model and Robinson’s (2001b, 2003, 2005b) Cognition Hypothesis. These models make contrasting predictions as to the effect of increasing task complexity along various dimensions on L2 performance. Skehan and Foster’s (2001) Limited Attentional Capacity Model views attention and memory as limited in capacity; therefore, they suggest that increasing task complexity reduces 7 the pool of available attention and memory resources. As a result, some aspects of performance will be attended to while others will not. Skehan and Foster also claim that cognitively more demanding tasks draw learners’ attention away from linguistic forms so that enough attention can be paid to the content of the message (for a recent account of the model see Skehan, 2009). Skehan and Foster’s (2001) model, however, does not consider that attention is selective and voluntary, in other words that one can have volitional control over choosing relevant stimuli and ignoring irrelevant ones (Allport, 1987; Wickens, 2007). Robinson’s (2001b, 2003, 2005b) Cognition Hypothesis differs from the Limited Attentional Capacity Model in that it assumes that attention is subject to voluntary regulation. In the Cognition Hypothesis, two sets of dimensions of cognitive task complexity are distinguished: resource-directing and resource-dispersing dimensions. The resource-directing dimensions of task performance call learners’ attention to the linguistic features which are needed to meet task demands (e.g., reference to events happening here and now vs. to events taking place then and there), whereas the resource-dispersing dimensions of the task act as attentional limitations in determining what aspect of the task can be heeded (e.g., reducing the pre-task planning time). The Cognition Hypothesis states that sequencing tasks from cognitively simple to complex allows students to progress towards successfully performing real-world target tasks. Robinson proposes that increasing task complexity along resource-directing dimensions can lead to greater accuracy and grammatical complexity of L2 output because such demands can direct learners’ attention to how the concepts and functions required by the task have to be grammaticized using specific linguistic forms. Increasing complexity along resource-dispersing dimensions, however, depletes learners’ attention without having the beneficial effect of directing it to any specific linguistic aspect of L2 production. The hypotheses put forward by the Limited Attentional Capacity model and the Cognition Hypothesis have received mixed support, which is probably due to the fact that in some tasks, certain characteristics make resource-dispersing attentional demands on learners, whereas others simultaneously draw their attention to certain linguistic aspects of performance. Few studies have 8 examined systematically the combined effects of resource-directing and resource-dispersing variables on performance (but see Gilabert, 2005; Iwashita et al., 2001; Révész, 2009) although in real-world communicative tasks these two dimensions simultaneously affect performance (for a recent discussion of this issue see Pallotti, (2009)). In our research we did not intend to test the predictions of the two conflicting models of task complexity, but we were interested in comparing two types of tasks which make different conceptualization demands on L2 speakers: a narrative task, in which the learners need to design the plot of the story, and another type of narrative task, in which they need to linguistically encode a pre-determined storyline. Our aim in using tasks with different attentional demands at different stages of speech production was to extend Robinson’s Cognition Hypothesis (2001b, 2003, 2005b) with respect to complexity demands concerning conceptualization and linguistic encoding. Written vs. spoken task performance In addition to complexity, another key factor to consider with regard to tasks is the mode of performance. Task-based language learning research has traditionally focussed on speaking tasks, and considerably fewer studies have investigated how tasks affect second language writing processes and the product of writing. There are a number of important differences between speaking and writing (Biber, 1988; Chafe, 1982), but from the point of view of psycholinguistic processing, the most crucial ones are that writing is usually not as constrained by time and that it is a recursive process, in which writers plan, linguistically encode their plans and revise them cyclically (Grabe & Kaplan, 1996). Although the availability of time is not unlimited in writing, writers are under somewhat less pressure than speakers to divide their attention between conceptualizing their message and linguistically encoding it, which allows for extensive on-line planning, that is, planning the content of the output while giving it a linguistic form (Yuan & Ellis, 2003). In writing, the time spent on planning the message (i.e. pre-task planning) is also integrated in the writing 9 process because writers can devote considerable time to planning the content before starting to write. Writers also have more attentional resources available for monitoring their output while they are encoding it than speakers. Only a few studies have addressed how the mode of performance influences the output students produce on specific language learning tasks. In a study conducted with learners of French, Granfeldt (2008) found that mode did not have an effect on syntactic complexity, but students used a higher variety of words in writing and were more accurate in speaking than in writing. In Kuiken and Vedder’s (2009) research, however, Dutch learners of Italian produced syntactically more complex language in writing than in speech, but the lexical variety of their spoken and written output did not differ significantly. The contrasting patterns of the findings might be related to the research design of these studies (i.e., different types of tasks used) and to the fact that different L2s and learners with different levels of proficiency were investigated. In our study, we intended to investigate the effect of mode on performance in two types of tasks which pose different conceptualization demands. In our analyses, we focused on differences in lexical variety and in global as well as task-specific measures of syntactic complexity and accuracy. Furthermore, we were also interested in how learners with different cognitive ability profiles perform in these two modes and how various aptitude components are related to the linguistic quality of the output in speech and writing. Individual differences and task performance Individual differences can exert both direct and indirect influence on task performance. Individual difference factors such as anxiety, working memory capacity and aptitude might have a direct effect on students’ decisions concerning the allocation of their attentional resources and on students’ ability to handle their attentional limitations. Individual differences, especially those in cognitive abilities, might also influence how successfully students acquire particular aspects of linguistic 10