Archive for November, 2012

Is Reading Aloud the Secret to Proofreading?

November 7, 2012 1 comment

Recently I began re-teaching myself the trumpet, after a brief break of twenty years. I fancy no ambitions of being the next Miles Davis, only the ambition of unwinding between grading essays, planning lessons, and responding to the usual student emails about their grandparents—dead and dying.

Trumpet, Mouthpiece, and Bucket Mute.

How quickly old skills come back! After a few days, I was lighting up the house with the brassy tones of Kumbaya and When Johnny Comes Marching Home. As is, the trumpet is a loud instrument. Our house’s hardwood acoustics amplified things to the point where my wife politely insisted I shove a practice mute in the horn.

At the local music shop, I spoke with a clerk whose main qualification was a Jim Morrison haircut. He couldn’t answer my questions about which mute to buy. Instead, he stood baffled by the store’s selection of mutes, contemplating them the way a stoned teenager contemplates the rainbow reflection off a DVD. Annoyed, I purchased a practice mute with a midrange price, and vowed never to return.

A mute works like a car muffler. This mute’s cork seal lodges into the bell of the horn. Air escapes only from two holes drilled into the mute, each the diameter of a BB. Even when I blew hard, only the faintest sound escaped, tinny and muffled. Against the noise from the street and the ambient hum of household electronics, I could barely hear myself play. Gone was the enjoyment.

Gone also was my feedback loop. I couldn’t hear which notes I was flubbing and which I was nailing. And this meant I couldn’t figure out what I needed to work on improving.

A similar feedback principle holds with writing. Typically, the writer’s feedback loop comes from re-reading what they’ve written, or getting the response of peers, tutors, or teachers. But how often are the words of a student’s essay actually read aloud?

I’ve begun to wonder whether we are asking students to do something akin to playing a horn plugged with a mute. What do students miss when their writing remains silent? When we read aloud the words we write, the language becomes not just textual but acoustic. We hear the music of language. And there’s something different, something richer and more moving, about how we process language we hear.

The crucial way that written and spoken language differ:

Most fundamentally, our facilities with spoken language are intuitive and inborn. In fact, the psychologist/linguist Steven Pinker has referred to our facility with spoken language—rightly—as The Language Instinct. We all have been speaking our native tongue from our earliest years, with no explicit instruction required, and we are geniuses at it before we reach ten. (If you don’t believe me, try teaching a dog or a chimp to understand or speak any sentence that comes out of a kindergartner’s mouth.)

Written language, though related to spoken language, is something else entirely. It’s against nature. It’s something that we produce only after many years of rigorous instruction in the classroom. Even then, many people never get very good at it.

Evidence also suggests humans’ facilities with language are linked to our facilities with music. For one, all cultures have spoken language and music, and just as with language, all normal individuals possess extraordinary talent with music. (If you don’t believe me, try teaching a kindergartner to hum a tune, and then try teaching a chimp or a dog to hum the same tune). Even people with severe intellectual disabilities or brain damage often have remarkable abilities with music and spoken language.

The link between language and music appears to run deeper. Neuroscientist Anirudhh P. Patel argues that both spoken language and music share a hierarchical structure and both “overlap in important ways in the brain” (674). Other psycholinguistic research from the Max Planck Institute shows that when listeners heard a musical sequence with a dissonant chord, brain imaging recorded neural activity in some of the same regions where we see neural activity if a person hears an ungrammatical sentence.

Confession: I myself struggle with proofreading.

When I’ve browsed back through prior blog posts, I’ve noticed them blighted by more than a few mechanical errors. (The irony of the writing teacher making grammatical errors on a blog about grammar teaching doesn’t amuse me as much as it should; I’m a perfectionist.) Before those posts went up, I had proofread them obsessively, and then had a colleague proofread them once more for good measure. What happened?

We both proofread silently.

After the incident with the trumpet mute, I switched to proofreading aloud. I sometimes feel like a lunatic talking to myself, but I was catching three times as many errors in a single read-through.

Was I on to something that could help my students?

I had always instructed students on proofreading strategies, but the results too often left me disappointed. I covered the handbook’s advice on proofreading (the sort of conventional wisdom that’s become obligatory for handbook authors) and reminded them throughout the semester to budget ample time in their writing process for proofreading.

In my experience, many teachers take a product-centric approach to proofreading (“These are the errors to look for”), rather than the process-centric approach I favor (“These are the steps to go through, and this is how much time it takes.”) Many students were following my advice. Many forgot to budget the necessary time. Still, too many essays—even ones where students put in the time—read like marvels of shoddy proofreading.

I recently suggested that my developmental students try proofreading aloud. When students turned in their first essays, I asked for a show of hands to see how many followed this advice and how many found it helpful. About half of the class proofread aloud and about a third of the class found it helpful. Several students vocally endorsed proofreading aloud. But the striking evidence emerged from the final product—essays much more mechanically polished than usual for this point in the semester.

Of course, this evidence falls short of scientific, but it suggests a future experiment: measure what happens when two otherwise similar classrooms of students are given different guidance about whether to proofread aloud or in silence.

Proofreading aloud helps a lot, but I don’t expect it to be a panacea, for the following reasons:

1. It obviously won’t work with in-class writing.

2. English words are spelled in ways that correlate only loosely with the phonetic sequences that issue from our vocal tracts, so it won’t necessarily help with certain spelling errors (including homophones and anti-phonetic spellings).

3. Some students’ teachers have told them to put commas wherever a pause seems natural—a gross oversimplification. Many speakers randomly pause, only because they hesitate in thought. Also, many speakers briefly pause after they finish pronouncing the grammatical subject of a sentence, especially a long subject, even though it’s widely accepted as ungrammatical to separate the subject and verb with a single comma.

4. Punctuation errors are tough to catch reading aloud, especially errors where students are confounding commas and ending punctuation. Both of these sound the same—like silence. Apostrophes, likewise, are silent.

5. Some students, when they read aloud, only loosely scan the lines they read, glossing over written errors and pronouncing what they think they see, rather than focusing on the letters on the page. This problem is exacerbated when students proofread at the computer, where the optics encourage this “scanning.”

However, proofreading aloud does especially help in certain situations. It helps students notice sentences where the syntax muddles in circles, the sentence rambles out of control, or the different phrases don’t quite fit together as a complete sentence. It makes choppy transitions more salient. It also works better when students track the word they’re pronouncing with a pencil, in order to prevent “scanning” past errors.

Why do we catch more errors when proofreading aloud?

Jan Madraso suggests that when students proofread aloud, it relieves the burden on their short-term memory and it forces them to proceed more slowly (32 – 33). I would add, as mentioned earlier, that working with spoken language is easier, since it’s more natural, more intuitive, and hard-wired in the human mind.

Then we should consider the role of prosody (that complex pattern of stress, rhythm, and intonation that voice synthesizers struggle to replicate). When we read aloud, we are forced to map onto the words a prosody. A large body of linguistic research shows that the prosodic structure of a sentence depends heavily on its syntactic structure. So if the syntactic structure of a sentence is malformed, sprawling, or difficult to parse, when you try to pronounce it, your tongue gets tied.

Another body of linguistic research shows that prosody depends on a variety of semantic and pragmatic factors about the discourse. For instance, stress patterns change depending on whether words represent items that are new to the discourse or familiar to the discourse. Similarly, we use a special intonation when we contrast two elements with one another, convey sarcasm, read the items in a list, or transition to a new subject. Thus, when a reader can easily map a prosody onto words, it signals that the text is comprehensible.


In the end, I exchanged that practice mute for a “bucket mute” that muffled the sound much less. Practicing the trumpet grew more enjoyable.

The experience also taught me to appreciate the sound of writing. Written words, by comparison, seem more lifeless. I now ask students to read excerpts from texts aloud in class. It appeals to the auditory learners, and brings the classroom alive with the music of language.