Saturday, 1 July 2017

Spelling and Statistics

The end of another term and various completed pieces of work come home.  I scan through the science tests and assignments. Something has changed since I last looked at my Year 8 child's work.  I am somewhat confused.  It is her spelling that has improved!  I encounter the words 'because', 'materials', 'electricity', 'measure' and all forms of double consonant words spelled correctly.  But how can this be?  I haven't done any explicit spelling work with her for nearly a year!
Professor Diane McGuinness in 'Early Reading Instruction' makes the statement that 'the statistical probabilities of sound-spelling patterns in words is what accounts for why spelling improves with age'.  How?  According to Mark Seidenberg ('Language at the Speed of Sight') it comes down to the brain's implicit (behind the scenes) tracking of statistical regularities (frequencies) and co-occurrences of the elements from which words are composed - sounds and spellings. The brain is always quietly taking a tally of which sounds and spellings (and other elements) occur in the vicinity of each other and which ones don't. And every time a child or young person reads, these statistical patterns are fine tuned.  Amazing.
And of course there are patterns!  The sounds used in our language means that there are some combinations that are possible, whilst others are not.  For toddlers, language acquisition is the first step and development of the statistical data structure is  driven by exposure to a massive amount of data (people speaking).   Through exposure to speech, sound (and other elements) pattern structures are formed.  Reading is the 'add on' to language and brings in data on regularities in how letters combine and how they relate to the sounds in words (spelling).  According to Mark Seidenberg, how we become more accurate spellers involves the absorption of a lot of data (primarily through reading) which is why the amount and variety of texts that children read is important.  For a young child, new words are a unique pattern without any relationship to an overall pattern.  However, major patterns emerge as the child reads and writes more words. But pattern development is not just limited to individual spellings.  As longer words are encountered, the brain starts to tally up statistical probabilities of syllable combinations.  So much is going on in this amazing neural data 'hunting and gathering' symphony.  We begin with speech, introduce explicit instruction and with ongoing development and practise in reading (and writing), the statistical data structure continues to refine itself without conscious effort.  Which is why spelling can improve with age.
Over the years I have kept two things alive - reading aloud to my child and having her read books that her level of code knowledge at the time could accurately decode.  In reading aloud I continually exposed her brain to more sophisticated language.  In making her read aloud I kept her brain engaged in the process of mapping the patterns of sounds and spellings, syllables, words, phrases, etc.  And at intermittent points on that road, I provided explicit instruction.  When she entered high school she had the skills to read (decode) even if her fluency was not well developed. But she had enough to engage with curriculum texts, which she has been increasingly required to do as time progresses.  This year I have reinstated 20 minutes of nightly silent reading of quality literature.  Over the years, at times tough parenting on my part, has ensured that her brain has been exposed to the massive amount of data needed to build its statistical data base.  What I am seeing now is the improvement in her spelling (in the absence of additional explicit instruction) resulting from refinement of that structure by her brain without conscious awareness or effort.  Actively and persistently feeding the statistical data structure development pays off.  We just have to invest in it properly and give it time.  This is probably why random spelling lists in school classrooms have very little impact to spelling accuracy. These lists don't 'feed' the brain, they work against it.
Children must be taught in the proper way.  If I had left my daughter exposed solely to the whole language instructional techniques that were the modus operandi in her early years' school, her disengagement with reading (and probably school) would have been complete. It wasn't until Year 4 that she was exposed to synthetic phonics at another school and I worked intermittently with linguistic phonics at home. Through a change in school, the support of her new teachers, and my input, she got enough for her brain to pick up the foundations of the statistical data structure.  When she refused to read, I read to her.  When she was ready, we read together. And when her data structure was more developed, she read independently.  Through reading, I kept her brain engaged with the 'data gathering' it innately sought.
Because of the rich history of the English language and our writing system, sounds can be represented by more than one spelling. For everyone, there will always be words that require specific explicit spelling instruction that is in addition to the implicit statistical learning that goes on behind the scenes if we want their accurate spelling in long term memory. More importantly, there is also much more that needs to be done in schools to incorporate the innate functions of the brain into reading instruction and spelling development. But in the meantime I can marvel at the magnificence of the human brain and at the work it can do without conscious effort, if we just keep it exposed to data through reading.  But if a young person or adult is never taught how our alphabet code works to represent spoken language in written form, then the brain's amazing 'reading pattern hunting and gathering' function, remains untapped and children's futures are derailed. When we know what we now know, this is unacceptable.

1 comment:

  1. thank you... This is shining light on some questions I have had floating around for a long time. Ever since I have been hearing the explanation that 'exposure improves spelling', I've wanted to know, "How?!" This article has given me some new links/connections to see how this might take place in conjunction with, and as a result of; explicit instruction.