Why Learning to Write Comes Before Learning to Read
It may sound strange, but the act of learning to handwrite with paper and pencil precedes learning to read.
You’ll be able to read a Mandarin or Cantonese newspaper once you know about 3,000 characters, according to the BBC. And the best way to learn them may be to practice writing them first — with old-fashioned paper and pen.
Today, smartphone language-quiz apps abound and schools spend less and less time teaching printing and cursive, yet according to UCSF researcher Paul Gimenez and team, recent neuroimaging studies “have concluded that while free-form handwriting practice clearly supports reading acquisition, typing (Longcamp et al., 2005) and even tracing (James and Engelhardt, 2012) do not”. Writing by hand may be an important part of an adult’s foray into learning a new alphabet or symbol-based subjects like math, physics, and music. And for children, handwriting may support the ability to recognize individual letters, which studies have found to be the most important predictor of reading ability.
An fMRI study demonstrated that handwriting recruits the same brain regions that underlie reading acquisition in children. In the study, they learned letters by writing, tracing, or typing them. When they saw the letters later during an fMRI scan, only the handwritten ones elicited brain activity in a well-known “reading circuit,” as well as in Broca’s area, a brain area key to speech production. The traced and typed letters didn’t elicit the same response.
Another study tested character recognition in adults after participants learned made-up symbols by writing them or typing them (which required them to find and push the symbol on a keyboard). When tested, the adults recognized the correct orientation of the hand-written characters more accurately and for a longer time, while their memories of typed characters dropped after three weeks. What’s more, the faster the participants had been able to write symbols during training, the better their recognition was later.
So there’s something about the physical movement of writing that helps us learn characters. Neuroimaging studies support this, showing that when we write, we recruit an integrated neural network of motor systems that is also recruited when recognize individual letters. Researchers suggest that this overlap is a direct result of the sensorimotor aspect of writing, and that the way we read may really depend on how we write.
Even though tracing letters does take the hand through the necessary motions to produce a letter, it doesn’t have the same effect as free-form handwriting does, because their outputs are quite different: tracing produces a set of near-perfect copies of a letter, while free-form printing creates noisy, imperfect ones.
Scientists theorize that this is why handwriting supports letter recognition: it’s better at teaching us what makes an “A” an “A,” or a “B” a “B.” A letter’s little features are important because the orientation, overlap, angles, and location of certain strokes can make or break its identity. We can’t rely on global shape to tell us what it is — otherwise, mirror-image “b” and “d” would mean the same thing.
At the same time, there are other small features that are less important. These can change without confusing us about what letter we’re looking at. Think about the differences between fonts, serif versus sans serif, italics versus bold type. We (adults and children alike) can still recognize them and read them, even though sizes, slants, and small additional lines vary.
Researchers argue that free-from handwriting trains this ability to distinguish which features are important. Imagine this: a child tries to write the letter “G” and their rendition is less than perfect, because they’ve just begun learning or don’t have the necessary fine motor skills yet. But presumably, they know they intended to draw a “G,” so producing an imperfect copy actually broadens their understanding of how a “G” is formed and what it can look like. Writing free-form letters over and over creates a noisy dataset that broadens letter categories, making later recognition and reading faster and more accurate — much in the same way an image recognition software program “trains” itself to recognize tulips by scanning thousands of unique flower photos.
Of course, you can’t get an imperfect letter dataset when you’re tracing or typing them, because they come out near-perfect every time. So the next time you need to learn new characters, drop the training wheels, start with a blank piece of paper, and get your hands in motion.
The best part?
It’s good to mess up.
If you enjoyed this article, please share on social links below.