It has become an unwritten law of teaching that when reading aloud to students, the teacher must ensure students are reading along in their own copy of the text. This is, I contend, a bad idea. To understand why we need to consider working memory in some detail.

It’s well-known that the capacity of working memory is strictly limited – estimates range from anywhere between 4 to 9 items at any one time – but it’s less well-known that working memory is almost certainly not a single edifice. Baddeley and Hitch‘s widely accepted working memory model contains four distinct components. The central executive* (CE) focuses our attention on the information we want to process. Where there are competing stimuli – listening to the teacher vs. making sculptures out of the contents of a pencil case – it will decide which should be attended to at any given moment. As every teacher knows, even if students are determined to pay attention, they’re liable to switch focus if a bee flies into the classroom or someone farts. The CE uses three dynamic sub-components to process information. These are the phonological loop (PL), which deals with verbal information, the visuo-spatial sketchpad (VSS) which processes visual information and the spatial relationships between objects, and the episodic buffer (EB) which integrates new information from the PL and VSS with information already stored in long-term memory.

Despite the bottleneck of working memory, we are capable of holding information in the different components without too much difficulty. For instance, images processed in the VSS can be used to anchor verbal instructions and explanations, held in the PL. So, if I were to explain what a sepulchre was whilst also showing an image of what one looked like, working memory would not be in danger of becoming overloaded; in fact, the image would support the explanation. The general rule is that while we should avoid over taxing any of the individual components, visual and auditory information can be processed simultaneously without creating additional cognitive load.

So, what about reading? There’s no doubt that text read aloud will be processed by the phonological loop, but is text on the page an image processed by the visuo-spatial sketchpad, or is it sounded out internally and therefore processed by the phonological loop? This is important, because if it’s the latter then asking students to read along at the same time they are being read to runs the risk of over-burdening working memory and leading to cognitive overload.

For most of my career I uncritically accepted the self-evident truth that children must be made to ‘read along’ whenever I read aloud. I often found this problematic. Some children, typically the more able readers, often wanted to read ahead and I had to invest time in preventing them from doing so. (I should be clear here that by ‘more able’ reader I’m referring here to reading fluency and not comprehension. As I explain in this post, there is no link between reading fluency and intelligence.) Other children, often weaker readers, would tend to put down their books and just listen as I read. I developed various techniques to nip these minor acts of defiance in the bud because I knew that was in my students’ best interest.

I now think this was misguided. Silent reading is a relatively recent innovation. St Ambrose, 4th century Bishop of Milan, was considered unusual in that when he read he made no sound:

…his eyes scanned the page and his heart sought out the meaning, but his voice was silent and his tongue was still. Anyone could approach him freely and guests were not commonly announced, so that often, when we came to visit him, we found him reading like this in silence, for he never read aloud. (Augustine, Confessions Book Six, Chapter Three)

So called ‘silent’ reading actually involves turned print code into sounds internally, and thus all language whether read as text or listened to is processed in the phonological loop. The loop represents a system made up of a short-term store in which memory is represented phonologically, and a rehearsal process which preserves and refreshes the information, partly through the process of ‘subvocalisation‘.

More able readers will attempt to block out my voice so that their phonological loop can be used solely to process the text they’re reading silently to themselves. Weaker readers are often distracted by their lack of fluency – their phonological loop is divided between trying to decode phonemes and trying to attend to my voice. To resolve the issue, their central executive will switch focus away from the text so they can better listen to me. Whenever teachers insist students follow along as text is read out loud, they run the risk of over burdening students’ working memories.

For able readers this isn’t too much of a problem – as long as they’re not discovered to have been reading ahead they will have processed the text and transferred much of it to long-term memory. But for weaker readers this practice can be disastrous. Switching between different components of working memory may well induce a ‘task-switching’ penalty. Additionally, so much working memory capacity is tied up with trying to decode that there is little room left to think about meaning, when we add in the need to also block out the distraction the teacher’s voice, the battle is almost always lost.

Although ‘reading along’ is well-intentioned, it’s important to remember that students who have not stored the knowledge of phoneme/grapheme relationships in long-term memory as automated schema will not acquire this vital knowledge by being made to read real texts in classroom situations. Most children in this position require specific, targeted interventions before they can be expected to read fluently. Far, far better to allow them to dedicate all their attention on listening so that they have greater capacity to think about meaning, draw inferences and make hypotheses.

To be absolutely clear: reading aloud to students, especially weaker readers, is a good thing to do. Prosody – the sound and rhythm the words make – really aids comprehension, especially with difficult texts. The problem comes when children are expected to follow along at the same speed. Because they lack the fluency to do this their working memory overloads and derails comprehension.

For more detail on all this it’s worth reading the following:

* The existence of a central executive has been called into doubt by Sweller, Ayres & Kalyuga (2011) who suggest its theorised functions are performed in long-term memory. They argue that “an independent central executive disassociated from knowledge held in long-term memory results in an infinite regress of central executives.” Although this is currently a minority opinion, I find it convincing.