Learning Language in Chunks
“... In recent decades, the study of language acquisition and instruction has increasingly focused on ‘chunking’: how children learn language not so much on a word-by-word basis but in larger ‘lexical chunks’ or meaningful strings of words that are committed to memory. Chunks may consist of fixed idioms or conventional speech routines, but they can also simply be combinations of words that appear together frequently, in patterns that are known as ‘collocations.’ In the 1960s, the linguist Michael Halliday pointed out that we tend to talk of ‘strong tea’ instead of ‘powerful tea,’ even though the phrases make equal sense. Rain, on the other hand, is much more likely to be described as ‘heavy’ than ‘strong’... ”
It's funny, I've been thinking about that a lot for the last few months... Especially how there are phrases in English that don't really make sense or sound fluent when you try to break it down into the individual words.
For example: "In the mean time". Everyone knows that that means something like 'While we're waiting for that' or 'while that is taking care of itself' or something, but who has ever used the word 'mean' for that definition? It wouldn't work in any other phrase.
I suspect other languages are the same way, but I don't yet know them well enough to be sure.
I think it's also worth noting that the conventions can be broken for effect.
"This is strong tea."
"This is powerful tea."
The second sentence is more emphatic because it breaks convention and makes you think about it. Of course, you have to know that convention in the first place to get the emphasis.
I'm in agreement with Linguists that 'chunking' is useful, but I really feel that considering the child to be the ultimate language learner is the wrong approach (perhaps more easy to measure though). Unfortunately, great language learners tend to have very fixed processes with a massive amount of variety and time frames. It's near impossible to measure, but, I think, the current approach doesn't reflect that an adult is an adult and a child a child. Still chunking is useful.
Last edited by Cranks (2011 May 21, 7:18 am)
I think a difference is that children are often able to memorize and reuse those chunks easily, sometimes even without being able to break them down into individual words (and occasionally they make funny mistakes that way).
But for adults, I think it's very hard to memorize long chunks of language without first learning to recognize the parts individually. It's just a blur a unknown words that we don't process.
But that doesn't mean that we won't learn those chunks eventually with enough listening and reading practice.
Dood I was just thinking about this while doing RTK. Even though I'm fluent in English, I know more phrases than I know the actual vocabulary. Trips me out bro, crazy I know.
I thought it was a nice bit of synchronicity, finding this piece with its reference to Halliday, as I'm big on the conception of and focus on ‘lexicogrammar’ in Halliday's systemic-functional grammar. (Synchronicity because I had recently read something about NYT's On Language column being defunct and randomly checked out that article.)
Related: Heating up or cooling up the brain? MEG evidence that phrasal verbs are lexical units
Excerpt from the above's concluding remarks:
“It is now generally accepted by cognitive linguists that complex pairings can become cognitive routines or units if they are used with sufficient frequency. Accordingly, even some fully transparent combinations, which in principle could be built compositionally by the combinatorial language system, are thought to be stored and accessed as wholes rather than produced ‘from scratch’. Similarly, in (Shtyrov et al., 2005) and (Goldberg, 2006) formulation of a constructionist outlook on grammar, any frequently occurring combination of words is assumed to be stored, whether that combination is idiosyncratic or not. Frequency of use and cognitive entrenchment are also fundamental in usage-based approaches to language (learning): multiple, concrete instances of language use are claimed to lie at the basis of speakers’ mental grammar: a grammar rule is not ‘innate’ but is extracted from its regular instantiations – it gradually emerges from learned exemplars (e.g. [Barlow and Kemmer, 2000], [Bybee, 2006] and [Tomasello, 2003]). In a neurobiological framework, syntactic rule formation has been related to an emergent property of the learning and storage of word strings and the substitutions between string members normally encountered (Pulvermüller & Knoblauch, 2009).
There is a growing body of experimental evidence that redundant storage of regular forms does indeed occur. With respect to word inflection, for example, several researchers have proved psycholinguistically that full inflectional forms with a sufficiently high rate of actual occurrence are stored as units even if they can also be derived by rule (e.g. [Alegre and Gordon, 1999], [Baayen et al., 1997], [Baayen et al., 2002] and [Sereno and Jongman, 1997]). Even strong proponents of the dual route theory, according to which “irregular and regular inflection, and words and rules more generally, depend on different systems in the brain” (Pinker, 1999, p. 255), recognize that medium- and high-frequent regular forms can also have stored representations (Pinker & Ullman, 2002, p. 458). Our second finding is in consonance with these studies, in that we have provided evidence that common phrasal verbs appear to be lexically listed, even in cases where they could just as well have been put together from their parts by the grammar system. Moreover, the brain responses indicate that lexical access processes have priority over rule-based decomposition in the case of common verb–particle combinations.”
Bonus, just some stuff I've been reading lately:
(First, an accessible sort of primer:
Semantics Returns from the Grave
Words Are More Human than Syntax)
The role of semantics and grammatical class in the neural representation of words
Nouns and verbs in the brain: A review of behavioural, electrophysiological, neuropsychological and imaging studies
Semantic or lexico-syntactic factors: what determines word-class specific activity in the human brain
Nouns and verbs in the brain: Implications of linguistic typology for cognitive neuroscience (They reference Evans & Levinson's paper here, and I really like their suggestion in that paper for Bickel's ‘multivariate typology’ [what they refer to as low-level feature metalanguage]... Bickel links here: http://dx.doi.org/10.1515/LINGTY.2007.018 and http://www.uni-leipzig.de/~bickel/resea … lenary.pdf)
Event-related potentials to event-related words: Grammatical class and semantic attributes in the representation of knowledge
Oh, and this is also interesting (re: amodality and multimodality): Perception, action, and word meanings in the human brain: the case from action verbs
Oops, forgot this one: Cortical representation of the constituent structure of sentences - Somewhat related to not only meaningful integration of semantics/syntax but Gary Marcus' thoughts on subtrees/treelets rather than trees in the brain. So in effect, I think it's best to think of it as small, overlapping lexicogrammatical nests (or perhaps reticulations/nets) rather than syntax trees.
Last edited by nest0r (2011 May 22, 9:29 am)