Walk past someone composing a difficult paragraph and you’ll hear them say the sentence out loud, sometimes twice, before writing it down, sometimes correcting the word on the second pass. A common reading of this is that the person loses concentration and speaks to fill the silence. What actually happens is closer to the opposite: the two cognitive processes are happening simultaneously, and speaking the sentence out loud is how the second can verify the first.
What the working memory model actually separates
Writing researcher Ronald Kellogg’s compositional model, most recently a 2013 progress report Co-authored with Alison Whiteford, Justin Turner, Erin Cahill, and Piers Mertens, Alan Baddeley’s writing maps onto the architecture of working memory, rather than treating it as a single undifferentiated skill. Planning and review are based on what Baddeley calls the central executive, the system that directs attention and coordinates other resources. The report claims that turning a thought into an actual sentence relies specifically on verbal working memory, the phonological circuit, the same system that stores a phone number long enough to dial it. Constructing a sentence and watching it sound right is not the same operation performed twice. They are two different systems, doing two different things at the same time.
Saying this sharpens rather than distracts
This dual-system account receives direct experimental support from research on self-regulated speech more broadly. a 2012 studypsychologists Gary Lupyan and Daniel Swingley found that when searching for an object visually, people who say its name out loud find it faster than those who search for it silently. Saying the target’s name aloud was not a distraction, but competed with the visual task. He sharpened the visual task by giving the brain an additional verbal cue to search. The same logic naturally applies to drafting: saying a sentence out loud gives the writer an auditory version of the sentence, which is not a drain on attention but a resource, checking the still-formulating version.
What happens when speech is blocked?
The clearest evidence that this verbal channel is actually burdensome rather than a voluntary habit of some writers comes from the experiences that inhibit it. Durham University psychologist Charles Fernyhough, explaining the work of his laboratory a Psychology Today piecereport that articulatory suppression, forcing a person to repeat an irrelevant sound while trying to plan, significantly impairs their planning performance, regardless of whether the blocked speech would otherwise be spoken aloud or silently.
Fernyhough writes: “We make our plans, at least in part, in language, whether overt (public speech) or semi-internal (private speech), or in such a way that only we can hear.” Completely remove access to that channel, in spoken or silent form, and the planning gets even worse. This is a difficult conclusion to combat with the idea that speaking in a sentence is idle noise that competes with real cognitive work. Speaking is in some functional sense part of real cognitive work.
A speech that builds as much as it checks a sentence
To write Psycheresearcher Nana Ariel extends Lev Vygotsky’s theory of individual speech, first to explain why young children narrate their actions to thinking and writing specifically for adults. His idea goes a step further than monitoring: speaking an incomplete sentence out loud doesn’t give the writer a chance to hear whether it’s correct, it actively creates incomplete material there until the words are spoken.
Ariel writes that “speech not only regurgitates pre-existing ideas, but also creates new information in the retrieval process as in the writing process.” His practical advice for writers stuck on a blank page is not to sit still and wait for a sentence to fully form, but to keep talking and writing on the basis that the generative process needs an active verbal channel to create anything.
Two things, one sentence
Together, the research describes something more concrete than some writers just think out loud. It describes a sentence constructed by one process and checked by another, as an interface between speech, sometimes generating the next statement, sometimes catching the last one before it lands.
A writer who mumbles before writing a line is not focused on the page. It drives the page through an additional system that silent layout can’t access and accesses the resource the study actually needs to design under pressure.
Why reading a draft out loud catches what a silent proofreader misses
It’s also the mechanism behind an editing tip that gets repeated without much explanation: read your draft out loud before you consider it finished. The advice works for the same reason when speaking in a sentence, it works when composing it. Silent reading allows the eye to skip a difficult phrase because the brain already knows what the sentence is supposed to say and automatically fills in the blanks. Reading aloud forces the phonological loop to process the sentence in real time like spoken language, where an awkward rhyme or missing word is revealed. Writers who dismiss the read-aloud passage as an unnecessary extra step skip a part of editing that uses a different cognitive channel than the drafting itself.
None of this is to say that every writer has to start talking in every sentence, or that silence is some kind of low mode. Kellogg’s model and Fernyhough’s experiments describe a verbal channel that can operate both covertly, in inner speech, and out loud. What the research rules out is the assumption that voice drafting is a lesser, less disciplined version of “real” writing. It’s a visible example of the process every writer works through, whether or not anyone else in the room hears it happening.
None of this requires the writer to change the way he works. It’s mostly useful as a reason to stop apologizing for it in shared offices and workplaces where mumbling before you’ve even written a sentence to someone unfamiliar with what’s going on can be seen.
