Thnaks to research-lover Eric Herman, I’ve been reading Lightbrown and Spada’s How Languages Are Learned (4th ed.) from Oxford University Press. This indispensable reading summarises a load of research on everything to do with second-language acquisition. Two phrases jumped out at me: “Comprehension of meaningful language is the foundation of language acquisition” and “Considerable research and experience challenge the hypothesis that comprehensible input is enough.”
Krashen pretty much nailed 80% of language learning: we learn via comprehensible input in safe and enjoyable settings. If that C.I. can be made compelling and repeatable enough, we learn more, or faster, or both. HOWEVER…there is more to learning a language than understanding it. So…here are some Lightbrown and Spada-mentioned insights from research.
a) people need a minimum of 16 meaningful exposures to a vocab item to remember it. TPRS will serve us well here.
b) fossilised errors— recurring grammatically wrong patterns of speech or interpretation– will continue unless there are deliberate teacher interventions. C.I. is not enough, at least not in a classroom setting. With first language acquisition, or immersion, the sheer volume of input will correct most– but not all– errors.
The classic example is use of the Spanish phrase me gustan los tacos (“I like tacos”). Learners will often figure out that “gusta–” means “like”, but will fail to acquire the -n that you need for plurals. Why specific feedback is necessary is open for debate, but it is necessary. TPRS pop-ups useful here.
The question of why this is necessary is up in the air. Some think the quantity of input in a classroom is low; others see this problem as stemming from various interlanguage processes.
c) Specific training in output– specifically in writing– is necessary. While the effects of writing training (drawing attention to specific parts of grammar or vocab or punctuation, etc, or asking for error correction, and other strategies) are not huge in comparison to the effect of quality C.I., they are significant. For example, in the New Brunswick E.S.L. study, the focused instruction kids did outperform the C.I. kids in writing. However, what was astonishing was, the C.I. kids did 90% as well as the focused-instruction kids with no teacher guidance or feedback. There is speculation on why this is, and it has something to do with the idea that understanding and output, while related, are not exactly the same brain system. So…if you wanted to do only ONE THING for second-language acquisition and writing output, it should be comprehensible input. If you want kids to be noticeably better, give them some specific meaningful instruction in, and feedback on, writing…but only in upper levels. Until a massive amount of language has been heard and read, feedback has limited effects at best.
d) Some specific training in recognising “weird” (i.e. non-native) grammatical forms is necessary for acquisition. For example, English and Spanish pronoun orders are different: in English we say “I ate it,” while in Spanish it’s “Yo lo comi” (“I it ate”). Research suggests that unless we provide a LOT of focused C.I. that (a) uses this structure and (b) draws attention to it, we will get delayed, incorrect or no acquisition of the rule. TPRS again will serve us well here– do those comprehension checks!
e) There is zero evidence to suggest that “smarter” people are “better” at picking up languages. Smarter– i.e. academically proficient symbol-manipulators, rule-followers, etc (you all know who I am talking about here!) are better at learning about languages…but in terms of acquisition, there is little difference between “types” of students.
f) Readers must know 90-95% of the vocab in the text to be able to read independently. This argues for MUCH more use of “easy” readers in a 2nd language classroom and much less “hard” and non-teacher-supported reading.
g) The best thing a teacher can do for S.L.A. is to allow kids to experience authentic success. This means kids should (a) understand everything, which feels good, (b) find it interesting (ay,..there’s the rub) and (c) feel safe and comfortable in class. The links between motivation and acquisition, despite “common sense” thinking, are unclear.
h) Free voluntary reading matters…but FVR with teacher interaction etc is much more effective.
i) For oral error correction, friendly comprehensible recasts (restatements) work best. Grammar explanations do little or nothing most of the time. However, recasts don’t do much, and there is disagreement about why they do sometimes work.
j) Sociocultural competence matters…but not that much. Yes, people need to be taught target-language stuff having to do with how that culture works. E.g. in Spanish, people need to learn when/where to use the usted (“You sir/madame”) form. Knowing this helps…but it’s far less important than people getting loads of C.I.
The upshot? Krashen, 30 years ago, got it 85% right. The other 15% is examined in Lightbrown & Spada. Go read it.