Task Complexity and L2 Narrative Oral Production

This study explores the impact of manipulating Task Complexity simultaneously along both pre-task planning time and the degree of displaced, past time reference on L2 narrative oral production. By drawing on consolidated psycholinguistic models of language production, the study first explores the similarities and differences between L1 and L2 oral production and examines how language processing for comprehension, production, and learning are mediated by the cognitive processes of attention and memory. Secondly, a definition of task is provided, and the concept of Task Complexity is presented and developed as born from a need to design and sequence tasks in a syllabus in a principled way. Thirdly, a review of options in syllabus design as a consequence of different conceptions about language teaching and learning is provided. The main characteristics of a number of synthetic and analytic syllabi are considered, and this is followed by a review of research into task features. In the fourth place, the specific literature related to planning time and the degree of displaced, past time reference is reviewed, and the questions and four hypotheses of the experiment advanced. The first hypothesis states that providing pre-task planning time will positively affect learners fluency and structural complexity with no impact on lexical complexity or accuracy. The second hypothesis predicts that increasing Task Complexity along the +/- Here-and-Now variable will have positive consequences for learners complexity and accuracy but will have detrimental effects on fluency. The third hypothesis foresees a greater impact of planning time in more complex There-and-Then versions of tasks. Finally, the fourth hypothesis predicts that the effects of increasing Task Complexity along the +/- Here-and-Now variable will be enhanced under planned conditions. From a methodological point of view, a repeated measures design is used to obtain production data from 48 subjects with a lower-intermediate level of English, and four levels of Task Complexity which combine the two variables are established. Each learner narrates each of 4 comic strips under a different condition, the sequence of which is controlled by means of a Latin square design. Measures include unpruned Speech Rate A and pruned Speech Rate B for fluency; the percentage of lexical words, the ratio of lexical to function words, and the Guirauds Index of lexical richness for lexical complexity; the number of S-nodes per T-unit for structural complexity; and the percentage of error-free T-units, the target-like use of articles, the percentage of self-initiated self-repairs, and the ratio of repaired to unrepaired errors for accuracy. Finally, learners perception of Task Complexity is also analyzed by means of an affective variables questionnaire, in which they are asked about their perception of difficulty, stress, confidence, interest, and motivation. The questionnaire is validated by means of protocol analysis. Results of repeated measures ANOVAs show that planning time generates higher fluency and lexical complexity with no effects for structural complexity or accuracy. Increasing complexity along the +/- Here-and-Now variable has positive effects on accuracy with no significant differences for lexical or structural complexity, and with detrimental effects for fluency. No significant differences are found between the impact of planning time of simple Here-and-Now and complex There-and-Then versions of tasks. The effects of increasing complexity along the There-and-Then variable are enhanced by planning time without reaching statistical significance. Further results show that more complex tasks are perceived as more difficult, stressful, and generate lower levels of confidence without detrimental effects on the perception of interest or motivation. Results are discussed in the light of L1 and L2 production theories as well as competing theories of attention allocation policies during task performance. It is argued that the dimensions of accuracy and complexity can be attended to by learners as long as tasks are kept simple along resource-dispersing dimensions (i.e. planning time) and complex along resource-directing dimensions (i.e. +/- here-and-now). It is finally argued that Task Complexity is an important construct for task design and task sequencing which can be manipulated to obtain specific effects on L2 production and, potentially, learning. KEYWORDS: Cognitive complexity, Tasks, Syllabus design, Oral production, Narrative, SLA RESUM: Aquest treball investiga limpacte de la manipulacio simultania del temps de planificacio i el grau de desplacament en el temps i lespai sobre la produccio narrativa oral en lL2. En primer lloc, i basant-se en models psicolinguistics consolidats de produccio oral, el treball examina les similituds i les diferencies entre la produccio oral en lL1 i en lL2, i analitza com els processos datencio i memoria condicionen el processament per a la comprensio, la produccio i laprenentatge del llenguatge. En segon lloc, es defineix el concepte de tasca i es desenvolupa el concepte de complexitat cognitiva, el qual sorgeix de la necessitat de dissenyar tasques i organitzar-les en un programa de manera raonada. En tercer lloc, es fa una revisio de les opcions de disseny de programes densenyament de llengua que emanen de diferents concepcions sobre lensenyament i laprenentatge. Es consideren les principals caracteristiques duna serie de programes sintetics i analitics, i es fa una revisio de les investigacions referides a diferents trets de les tasques. En quart lloc, es presenten el estudis que es refereixen especificament a les variables de planificacio i +/- aqui-i-ara, i savancen les preguntes i les quatre hipotesis daquest estudi. La primera hipotesi afirma que proporcionar temps de planificacio afectara positivament la fluidesa i la complexitat estructural de la produccio dels parlants, sense tenir cap impacte significatiu sobre la seva complexitat lexica ni la seva correccio. La segona hipotesi planteja que augmentar la complexitat cognitiva de la tasca forcant la narracio en lalla-i-llavors tindra consequencies positives per a la complexitat i la correccio, pero negatives per a la fluidesa. La tercera hipotesi preveu un impacte mes gran del temps de planificacio sobre la produccio en les tasques cognitivament mes complexes. La quarta hipotesi suggereix que limpacte de laugment de la complexitat cognitiva sobre la produccio sera mes gran quan hi ha hagut temps de planificacio. Metodologicament, sutilitza un disseny de mesures repetides per a lobtencio de les narratives orals de 48 subjectes dun nivell dangles intermig-baix, i sestableixen 4 nivells de complexitat cognitiva que combinen les dues variables. Cada estudiant narra cada historia sota cadascuna de les condicions, les quals es presenten en sequencies diferents mitjancant un disseny de quadres llatins. Les mesures inclouen: per a la fluidesa, el nombre de sillabes per minut de la transcripcio exacta i el nombre de sil.labes per minut de la transcripcio depurada; per a la complexitat lexica, el percentatge de mots lexics, la ratio de mots lexics respecte dels funcionals i lindex de Guiraud de riquesa lexica; per a la complexitat estructural, el nombre de nodes verbals per cada unitat T; i per a la correccio, el percentatge dunitats T sense errors, lus a nivell nadiu dels articles, el percentatge dautocorreccions i la ratio derrors corregits respecte dels no corregits. La percepcio subjectiva dels estudiants respecte de la complexitat cognitiva de les tasques es mesura amb un questionari de variables afectives, el qual demana als estudiants per la seva percepcio de la dificultat, lestres, la confianca, linteres i la motivacio respecte de cada tasca. El questionari es valida mitjancant una analisi de protocol que investiga la interpretacio de les variables afectives per part dels estudiants. Els resultats dels tests ANOVA de mesures repetides demostren que proporcionar temps de planificacio te un impacte positiu sobre la fluidesa i la complexitat lexica de la produccio dels parlants de lL2. Laugment de la complexitat cognitiva respecte de la variable +/- aqui-i-ara te efectes positius sobre la correccio linguistica sense produir efectes significatius per a la complexitat estructural ni lexica, i amb efectes negatius per a la fluidesa. No es troben diferencies significatives de lefecte de la planificacio entre tasques simple en laqui-i-ara, i les complexes en lalla-i-llavors. Els efectes de laugment de la complexitat cognitiva amb la variable +/- aqui-i-ara sincrementen sota condicions de planificacio sense arribar a nivells significatius. Resultats addicionals demostren que les tasques mes complexes cognitivament es perceben com a mes dificils, estressants, i generen nivells mes baixos de confianca sense afectar la percepcio del seu interes o la seva motivacio. Els resultats sexpliquen a la llum de teories de produccio del llenguatge en lL1 i lL2, aixi com en funcio de teories que difereixen en la concepcio de com latencio funciona durant la produccio. Sargumenta que les dimensions de complexitat linguistica i correccio poden ser ateses simultaniament sempre i quan les tasques siguin simples pel que fa a la planificacio i complexes pel que fa a complexitat cognitiva (es a dir, en lalla-i-llavors). Finalment, es defensa que la complexitat cognitiva de les tasques es un constructe empiricament testable i util per al disseny i la sequenciacio de les tasques en un programa densenyament. La complexitat cognitiva de les tasques pot ser manipulada per tal dobtenir efectes especifics per a la produccio i, potencialment, per a laprenentatge.