Posts Tagged ‘learning’
the best kind of sleep
Canto: So Dr Seheult tells us that the most important sleep is in the earliest period. This is called slow-wave sleep or N3 sleep. It doesn’t last long, maybe half an hour…
Jacinta: Why is it called N3 sleep?
Canto: Well, here’s the detail – you have three stages of this early sleep. N1 is when you fall asleep. It lasts no more than ten minutes generally until you’re really there, in sleep. Then there’s N2 of course, which lasts from 30-60 minutes, your muscles relax and you begin to enter this slow-wave, also called delta-wave or delta brain activity sleep. That’s the deepest sleep of the night, at its deepest in the N3 period.
Jacinta: Well that explains the 1-2-3, sort of, but what about the N?
Canto: I haven’t been able to find that out specifically, but these are all phases of NREM (non-rapid eye movement) sleep, which are followed by the REM phase. So I think the N is just short for NREM. Anyway there are two types of sleep for maintaining good health – slow wave sleep and REM sleep towards the end of the night…
Jacinta: But don’t those two make up the whole of the sleep cycle?
Canto: Let me tell the story. Slow wave sleep is when you secrete valuable growth hormone, vital for children, in the time before midnight, according to the good doctor.
Jacinta: But I virtually never sleep before midnight.
Canto: Well you’re not alone there. In cities now, which are growing ever larger, we’re going to bed later and getting up earlier, and so sleeping less…
Jacinta: But generally living longer. So what’s the problem? I’ve heard that Hong Kong, which is about as urban as it gets, has the longest life expectancy on Earth – but that was probably measured before the China crackdown haha.
Canto: Well it’s no joke that China’s thugocracy will jeopardise everything in HK’s future, but good public healthcare and a very low infant mortality rate helps. People today can still live well with diabetes, obesity and slow-developing cancers, but they’d be even better with good sleep habits, if the rat-race allows them. But cities present us with a kind of eternal daylight, at great cost, not only in electric lighting, but in lack of sleep. Not to mention brightly lit screens that we take to bed with us…
Jacinta: Okay so what are the other benefits of slow wave and REM sleep, however delayed?
Canto: Dr Seheult describes a study showing that general sleep deprivation actually reduces the levels of antibodies produced after influenza vaccination. That’s to say, vaccination is less effective for the sleep-deprived. Another study used rhinovirus, a common cold virus. They paid students to be infected and found that those with good sleep efficiency, that’s to say, a high ratio of in-bed time to sleep time – their risk of being infected was reduced five to seven-fold, an extraordinary result. Actually this ‘extraordinary result’ finding comes up again and again in Matthew Walker’s book.
Jacinta: Yes, but it’s surely good to be awake sometimes too. But again, what is it about slow-wave and REM sleep that provides such benefits. What are the mechanisms?
Canto: Well, we’re talking about N3 sleep, the deepest sleep. This sleep phase is particularly important for memory consolidation, the stabilisation of a memory trace once it’s been acquired – meaning presumably the event itself, or its impact. It’s also called sleep-dependent memory processing. Now, how this precisely works is still being researched, but it appears to have much to do with interactions between neurons or neuronal complexes in the neocortex and the hippocampus. So here I should introduce sleep spindles, which are essential to all mammalian species.
Jacinta: They’re brainwaves, aren’t they?
Canto: Neural oscillations, indeed. They’re generated in the thalamic reticular nucleus (TRN), in richest quantities during N2 and N3 sleep. Wikipedia tells me this:
The density of spindles has been shown to increase after extensive learning of declarative memory tasks and the degree of increase in stage 2 spindle activity correlates with memory performance.
This is confirmed in experiments described in Why we sleep, showing that people who slept for a night between being asked to memorise certain data, like putting a name to a face, did a significantly better job than those who tried to remember the data after eight hours without sleep (from morning to evening). During the sleep period, subjects’ brain waves were recorded, and this is Dr Walker’s account:
The memory refreshment was related to lighter, stage 2 NREM sleep, and specifically the short, powerful bursts of electrical activity called sleep spindles… The more sleep spindles an individual obtained during the nap, the greater the restoration of their learning when they woke up. Importantly, sleep spindles did not predict someone’s innate learning aptitude. That would be a less interesting result, as it would imply that inherent learning ability and spindles simply go hand in hand. Instead, it was specifically the change in learning from before relative to after sleep, which is to say the replenishment of learning ability, that spindles predicted.
Jacinta: So they were correlating the number of spindles with their memorising performance, and memory here is being equated with learning. Is that right? I mean, is learning really just memorising?
Canto: Well, no, but it helps. I’m trying to memorise Newton’s inverse square law for gravity, but I know that even if I can reel it off like a favourite poem that doesn’t mean I fully understand it. Let me see G = m1.m2 over r². I’m not sure if that’s right.
Jacinta: Yeah, basically you have to know that the gravitational attraction between two bodies is equal to the product of their masses divided by the square of the distance between their ‘centres of mass’. I think. Though why that happens to be the case I have no idea. Does anyone?
Canto: Because… the universe? I’m beginning to feel sleepy…
References
M. Walker, Why we sleep, 2017
https://www.uofmhealth.org/health-library/hw48331
How to get the best sleep for your immune system | Roger Seheult (video)
https://www.oal.cuhk.edu.hk/cuhkenews_202101_life_expectancy/
adventures in second language acquisition – an intro to the usage-based hypothesis of language learning
don’t you just hate it when slide presentations on grammar contain grammar errors
So now I’m going to describe and reflect on a rather more interesting video by two academics and teachers, Marjo Mitsutomi, a specialist in SLA, and Minna Kirjavainen, who researches first language acquisition. They’re working in the teaching of English in Osaka, and they’re describing the uni course they’ve just set up there. Kirjavainen, the first speaker, describes her research as being on ‘the acquisition of syntax and morphology in monolingual, typically developing children, from about the ages two to six’. So now for some definitions. How does syntax differ from grammar, and what is morphology? A rough answer is that a grammar involves everything about how a language works, which includes syntax, which is essentially about how words are ordered in sentences. Morphology is often described as the corollary of syntax. Grammar can be divided into syntax, the external economy of words (i.e. in sentences), and morphology, the internal economy of words (i.e. from morphemes). For example, ‘robbed’ contains the verb ‘rob’ plus the unit of meaning ‘-ed’, or ‘-bed’, which means ‘in the past’. But presumably ‘went’ is made up of two morphemes, ‘go’ and ‘in the past’, both of which are in a sense hidden in the word?
Kirjavainen describes herself as coming from the ‘usage-based, constructivist view-point’, and says
‘this means that I don’t assume there’s innate syntactic components in the child’s mind like many first language and second language acquisition researchers do. Instead, the usage-based viewpoint assumes that language exposure and general cognitive processes result in language acquisition in children.’
A slide accompanying this anouncement indicated that Michael Tomasello is one of the major developers of this approach. So I’ll need to familiarise myself with Tomasello’s work, especially as I’m currently reading Steven Pinker’s The language instinct, which appears to be an attempt to popularise Chomsky’s universal grammar theories. Chomsky and Pinker argue that there is something innate about grammar, though just what it is is hard to capture.
Kirjavainen is of the view that a child’s general cognitive processes (e.g. pattern finding, analogy-making and categorisation) together with exposure to language, lead to competent language acquisition. She argues that these processes are effective for non-linguistic tasks, so the same mechanisms are sufficient for decoding the language they hear and want to use. She divides her inquiry into first language acquisition into three questions:
- how do children learn to speak their native language?
- what kind of things do children pay attention to when they learn their native language?
- how do caregivers talk to children? What effect does it have on children’s language development?
Firstly, children pay a lot of attention to language input, and apparently research is starting to show that it’s not just lexical items but grammatical structures that children mirror from caregivers. The most frequently used grammatical structures of caregivers are the ones used earliest by children, and they then become the most used by children. So the ambient language heavily influences the child’s language development. The constructivists also argue that syntactic constructions are built on the language that children already know. So they chunk things together and try them out for effective communication with those around them, they absorb responses and corrections and adjust their language accordingly. Examples are ‘Mummy’, ‘I want mummy’, ‘Mummy do it’ ‘I want mummy do it’, ‘I want mummy to do it’.
All of this makes a lot of sense to me prima facie. Mitsotomi, who next takes up the talk, is of Finnish background like Kirjavainen, but with a more pronounced accent, having learned English later in life. She begins by mentioning the critical period hypothesis for SLA, which might be the subject of a future post. Her concern is in how SLA is affected not only by the towering presence of the learner’s L1, but by many other life experiences. So given these influences and possible constraints how do we create a space and an atmosphere conducive to SLA? Also, what does SLA mean to the identity of learners, and how is it that some acquire an L2 more quickly and effectively than others?
Kirjavainen then continues by introducing what might be seen as the obstacles to a collaboration between first and second language theorists. First, some linguists argue that there are inherent differences between first and second language learning. She lists three (out of many) common assumptions about these differences:
- all (typically developing) children learn to speak their L1 natively, whereas most people (with normal cognitive skills) don’t learn their L2 to a native-like standard
- children learn their L1 very quickly whereas it takes L2 learners years to master their target language
- L1 learners make few errors in comparison to L2 learners, i.e. children find it easier to learn the grammatical rules of their language, whereas L2 learners find it difficult to learn these rules
Kirjavainen questions the first assumption, first on the basis of vocabulary – a child’s L1 vocab will depend on her socio-economic background, the level of education, experience and language competence of those she’s learning from and other such factors. These factors also affect syntax, and she described a study of native speakers’ knowledge of and proficiency in the passive construction. The study compared the proficiency of university teaching staff (academics) with non-teaching staff. They were tested on their understanding of active and passive sentences based on pictures, a fairly easy test, and it was found that while both teaching and non-teaching staff had full understanding of the active constructions, only the academics had full understanding of the passive construction. The non-teaching staff were significantly below full understanding. The general point here is that not all native speakers know all the grammatical rules of their L1, and that it depends more on regular usage than is sometimes admitted.
Next Kirjavainen gets stuck into the claim that children learn their native language quickly. She points out that an average 5 year old is quite a competent L1 user, but far from having adult proficiency. She then does a breakdown of how many hours a day children have spent, up to the age of five, exposed to and using the L1. That’s 5 years@6-14 hours a day of exposure, and about 3.5 years@8-14 hours a day in using the language. Conclusion: it takes children years to reach a relatively high level of L1 proficiency.
All of this strikes me as really thought-provoking stuff, and some of the thoughts provoked in me are memories of my callow youth – for example an occasion when as a 15 year-old or so I found myself at a party full of uni students types, all a few years my senior, and was awed by their vocabulary and language proficiency, and fearful that I’d get roped into conversation and be mocked for my verbal incompetence. So, again, I’m finding Kirjavainen’s arguments persuasive here at first blush.
The third assumption is more or less demolished by Kirjavainen as she cites research by herself and others to show that children make lots of errors, especially between the ages of 18 months and 4 years – these include pronoun errors, omission of infinitival to, agreement errors, subject omissions and verb inflection errors. Even at five and upwards there are mistakes with past tense, relative clauses and complement clause constructions. A complement clause? I’ve only just heard of them, but let me explain.
Here are two examples of complement clauses, taken from Kidd et al, 2007.
(1) That Rufus was late angered his boss.
(2) Rufus could see that he had angered his boss.
The complement clauses are underlined. The first here functions as the subject of an ‘argument’ sentence, the second as the object. The second sentence is described as an unmarked case, in which the complementiser that is optional and generally not used in naturalistic speech. There are many other forms of complement clause construction, so I won’t get bogged down by exploring them here.
So this has been an enlightening post for me, and an enlightening view of a new (to me) usage-based constructivist view of language acquisition. Next time I’ll report on the latter part of this talk, which will focus more on the implications for SLA.
just a thought to end with