BU CAS LX 400 - Week 7a. Cognitive/ processing approaches

Unformatted text preview:

1Week 7a. Cognitive/processing approachesCAS LX 400Second Language Acquisition“Cognitive approaches”• So far:– The state of knowledge (what they (can) know)– Is the form of an L2er’s knowledge like a L1er’s?• Cognitive approaches:– Less concerned with what knowledge is acquired,more with how that knowledge is acquired.– Tend to view language learning as being learning ofa complex skill, using general learning mechanisms(not specific to language).– More plausible for L2A (less so for L1A) butgenerally contrary to a “UG in L2A” view…Today’s mixed bag of topics• Information processing (McLaughlin)– Controlled vs. automatic processing– Fossilization, intake and input processing.– Pedagogical predictions?• Competition model (MacWhinney & Bates)• Learning strategies (O’Malley & Chamot)• Connectionism (Sokolik & Smith)• Operating Principles (Slobin, Anderson)Information processing• Central assumptions:– The mind is a general-purposesymbol-processing system.– The mind is a limited capacity processor.• The mind as a slow computer with limitedRAM.• (This version primarily McLaughlin’s)Information processing• Refining the model further:– Conscious tasks require attention.– Attention is limited(by memory, processing power).– Well-practiced tasks become “automatic”.– Automatic processes no longer requireresources, leaving the resources available forother conscious tasks.Information processing• Intuitively illustrated by learning to drive a standardtransmission car.• Initially, much conscious attention.– Clutch in, ease up on the accelerator, move gearshift to thenext position in the reverse-N pattern, clutch out graduallywhile returning pressure to the accelerator.• Eventually, an almost unconscious process; attentionfreed up to focus on things like finding the turnoff orfinding a place to merge into traffic.2Information processing• The information processing approachextends this intuition about what happenswhen you learn to drive a standardtransmission car to what happens as youlearn a second language.– You learn the rules (explicitly?).– You practice using them over and over.– Eventually, they become habit, “automatic”.Controlled processing• First step: controlled processing of information,conscious attention to processing.• Limitations on attention: only one process can becontrolled at once (otherwise, interference).• Controlled processes are flexible, adaptable.– Can be altered in the face of contradicting evidence– Can be applied to novel situations.• But, slow and resource-intensive.Automatic processing• Learning: transfer of controlled processes tolong-term memory: automatic processes.• In long-term memory, automatic processes:– Rapid– Do not tax working memory, attention(allowing these resources to be used elsewhere)• But difficult to suppress or alter.Fossilization• An explanation for the (otherwise mysterious)phenomenon of fossilization in L2A.•(Recall: Fossilization is reaching a plateau ofgrammatical performance where systematicerrors persist regardless of further training.Idea: A process which has become automaticbefore it had been learned properly.)Restructuring• Controlled⇒Automatic, requires/prefers storingin a way that fits, leading to restructuring of theexisting automatic processes.• Example: U-shaped learning curve:– (automatic) unanalyzed forms (walked, went)– (automatic) analyzed forms (walk-ed)– Prompts reorganization, generalizing a regular rule foradding past tense ending -ed).– Then: over-regularization (go-ed).Restructuring vs. fossilization?• Automatic knowledge is frequently restructuredduring L2A, so why is it that systematic errorsfossilize, again?• Guess: automatic knowledge has developed to apoint where restructuring is a) difficult(magnitude) or b) not required by new additionsto automatic knowledge.3Restructuring• To date: The accounts of what is involved in restructuringautomatic knowledge are hopelessly vague.– Intuition– Qualitative shifts observed.• But a vital area of the model for useful predictions about theend state.– Is this the stage where UG exerts its influence over L2A?– Only language systems of a certain “shape” can “fit” in the brain?Restructuring of automatic knowledge to maintain the shape of theknowledge?Input ≠ intake• Inuktitut—input:• Qasuiirsarvigssarsingitluinarnarpuq• ‘Someone did not find a completely suitableresting place.’tired cause.be suitable not someoneQasu-iir-sar-vig-ssar-si-ngit-luinar-nar-puq not place.for find completely 3sgInput ≠ intake• Krashen: Learner must get comprehensible input(mixture of structures acquired and structures notyet acquired) to advance.• Intake: Input that is used in grammar-building.• VanPatten on input ⇒ intake– Intake (detection) of grammatical form requiresattention to form.– Limited resources, competes with attention tocontent.Comprehensible input• We have already seen some evidence that L2erstend to proceed through similar stages as they gainproficiency.• Coupled with the idea of comprehensible inputbeing required to advance, suggests there is anorder in which structural information can be used.• If true—can we tailor the input provided to L2ersto maximize its effectiveness by specificallyprovide input relevant to the next stage?VanPatten’sInput Processing Model• Learners process input for meaning before theyprocess it for form– Content words first– Lexical (vs. grammatical/functional) items first– More meaningful morphology (vs. less) first– Less meaningful morphology (e.g., agreement) is atthe bottom of the totem pole. It gets attention only ifthere is any left to spare, if the communicativemeaning is clear with little or no cost to attention.VanPatten’sInput Processing Model• Learners possess a default strategy that assignsthe role of agent to the first noun (phrase) theyencounter in a sentence (word order).– Can be overridden by lexical semantics and eventprobabilities– Learners adopt other processing strategies forgrammatical role assignment only after theirdeveloping system has incorporated other cues(case, stress)4Implications so far for instruction• McLaughlin’s controlled ⇒ automatic knowledgemodel suggests that frequent use/repetition ofgrammatical constructions will speed the move toautomatic processing and allow for new(controlled) processes to be


View Full Document
Download Week 7a. Cognitive/ processing approaches
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Week 7a. Cognitive/ processing approaches and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Week 7a. Cognitive/ processing approaches 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?