How Discourse Uses a Fundamental Operation of the Brain
By: Curtis Kelly
Whenever someone mentions discourse, I think about one particular area of discourse: expository writing. I might be biased since I co-authored a Cambridge textbook series teaching it, but expository writing is the mainstay of school report writing, media writing, and business writing. Why? Because exposition basically just means explaining stuff. It is a kind of writing we teach by focusing on organization, a key component of discourse. Organization refers to a particular set of rules we use to encode information so that the reader can decode it in the same way. Expository[1] organization is important to teach because it is not as natural to us as narrative (time-ordered) organization is, and the rules vary across cultures. My students in Japan had a terrible time writing comprehensible papers, partly as we shall see because of Japanese discourse interference, and that led me to make teaching organization the mainstay of my classes.
[1] Ken Macrorie, author of Uptaught and The I-Search Paper said that “expository” is a terrible academic word; most students think it is something you buy in a drugstore.
In fact, it was another teacher, Robert Kaplan, who taught us how different cultures use different styles of organization, and his discovery is one of my favorite stories:
Kaplan’s flash of light
The sixties were a time when large numbers of foreign students started coming to the US to study in huge numbers. Regular subject teachers all over the country started to complain about the “poor writing skills” of these students, since their reports and essays were so hard to read. Most universities set up special ESL programs for these students to improve their writing through intensive grammar and vocabulary study, but these efforts seemed to have little effect.
Then one ESL teacher, Robert Kaplan, noticed something interesting. It was not weak grammar and vocabulary that led to their “poor” writing; there was something wrong with the organization. In fact, their troubles with organizing seemed to be related to where they came from. So, he and his colleagues divided 598 student papers up into five groups, representing their general cultural backgrounds, and analyzed their rhetorical styles. He found that each cultural group seemed to be using a different system of organization that was being transferred from their own language. He sketched this famous diagram to illustrate how these rhetorical styles looked from an English reader’s perspective (Kaplan, 1966).
Students from the Middle East tend to write in parallels connected by inference, as found in the King James Bible: A host came upon the land. The followers were smote (my own creation for simplicity’s sake; for the real stuff, go here). “Oriental” students from East Asian countries seemed to write around and around the main point of their essay, and not declare it until the end. And so on.[2]
When Kaplan wrote his discovery up, it turned on a light for language teachers everywhere. Before Kaplan, these teachers were beating their heads against the wall trying to get their students to write better by teaching more grammar and vocabulary. It was not until he pointed out that these students were transferring whole discourse structures from their own languages into English that these teachers could understand why their efforts had been so futile. And, with that light, came an understanding that it is not grammar that writing language teachers should be focusing on, but rather, organization.
[2] Kaplan was much criticized as being culturally biased for using a straight line for English and tortuous paths for everything else, but as long as you see these squiggles as representing the perspective of English speakers, it makes sense. A Russian paper on rhetorical style might portray the Slavic style as a straight line.
Organization represents the way information is encoded so that a reader (or listener) can decode it in the same way. In English we chunk information into paragraphs with single topics, usually implied or stated near the top, so that readers, using these same organization rules, can follow what we are writing about. It is this process of encoding and decoding that is the crux of communication, and it involves a fundamental operation of the brain, one every teacher should know about. This fundamental operation is built into all language processing, not just expository writing.
The fundamental operation in everyday situations
Read the mini-stories below, all true, that illustrate this fundamental operation. See if you can figure out what it is.
A few days ago, when my daughter, a non-native speaker, worked on her science fair project, she wrote, suger, syrop, and cristle. It was interesting how logical these misspellings were, based on analogs and phonics.
In an interview, Lisa Feldman Barrett said that we usually complete someone else’s utterance “even before the words come out of her ______.”
I commented that my Kyoto wife’s pasta was a little more tomatoey than usual. To my surprise, she snapped back, “If you don’t like it, you make it yourself next time!” (Oh no! I actually like it tomatoey.)
All the language teachers at a Panasonic speech contest wailed with laughter when a female contestant explained why she liked studying English. She said she wanted to “have intercourse with many foreigners,” not realizing the dictionary definition of the word had changed.
My far-right cousin describes the January 6 rioters in the US as “tourists,” while my far-left cousin describes them as “terrorists.”
I proofed a paper a friend had written for a journal and was disturbed by how many personal opinions and inferences were packed into the literature review. I thought he had violated basic research writing rules, but then began to wonder if the standards had changed since my day.
Can you see what these stories have in common? In each case, meaning is created by matching language to some pre-existing internal model. In some cases, the models are language definitions, as in what “intercourse” is, or whether my cousins elected to see rioters as “tourists” or “terrorists” to fit a larger world view. In others, they are discourse models, as in Kaplan’s diagrams, my understanding of what a literature review should be, or in my wife interpreting my otherwise neutral comment as being a criticism (as is common in Kyoto). And in some cases, the internal models represent systematic processes, as in the way my daughter used phonics and analogs to generate spellings, and how we subconsciously used grammar rules and collocations to complete Feldman Barrett’s sentence with the word “mouth.”
Comparing incoming information to existing mental models is the heart of predictive processing, a fundamental operation of the brain.[3] This is the way we make meaning not only of language, but of all our experiences in the world, and even what is happening in our bodies. I do not call it a “fundamental operation” lightly. Our brains are predicting irrepressibly and unceasingly. Brain scans have found that 70% of the brain’s activity is related to predicting, vs. only 14% for running our bodies (Feldman Barrett, 2020).
[3] For a closer look at predictive processing, see our Think Tank on it.
Why prediction?
At this point you might be wondering why comparing incoming information to models in memory is called “predicting”? How is it different from what we have always said about brain processing: We sense, we identify, we evaluate, and then we act. The difference is that rather than the traditional model, where these actions take place in separate steps, we now see them happening all together, using memory to identify what is happening even before the sensing is complete and with our physical reaction to it built into the memory. We predict. It is a neat little trick our brain uses to save huge amounts of time and energy.[4] As soon as we get just a smidgen of sensory input, we use templates in memory to automatically fill in the rest and mobilize our resources to take appropriate actions.
These templates, or concepts, are cartoon-like amalgamations of all previous experiences we can overlay on incoming data. As Lisa Feldman Barrett put it: “When the brain is faced with sense data, it’s not asking the question, What is this? It’s asking, What is this like?” (2020). In brain terms, memories are really just firing patterns stored in our sensory and motor areas, so that when new incoming sensory or motor firing patterns fit the pattern in memory, the memory itself is activated and the incoming data thereafter is ignored (as long as it fits the prediction[5]). That is what happened when you heard “before the words even come out of her _____.” You had already filled in the word “mouth” even before you got to the end of the sentence, thereby reducing the cognitive load of listening. Lisa left the last word out just to make you aware you had predicted it.
[4] If we had to figure everything out from scratch, we would be frozen in thought all the time. That is exactly what happened to me one Wednesday morning when I waked into class and what I encountered did not fit my subconscious prediction. The classroom was empty. I stopped. I stared. My head whirled while I tried to figure out what was wrong. Wrong room? Wrong day? Then I remembered that there was a general assembly for 3rd year students that day, so classes were cancelled.
[5] Not ignored if the prediction is off, in which case we take notice, direct all our sensory tools towards reconfirming, and release a dollop of dopamine, the neurotransmitter that causes new synapses to form. We experience prediction error as surprise, dismay, or wonder, and it is really at the core of all learning.
Let’s look at another example. I’m sitting at home and I hear a noise coming from the kitchen. As soon as it starts, my brain is going, “Hmm. This auditory neural firing pattern matches the pattern I have in memory for ceramic rattling, and since it is wired to my kitchen memory, I am visualizing someone opening a cookie jar. This cookie-jar model is connected to “Yum, Cookies!” and an action sequence to go get one: Heart, increase blood pressure so that I can get up out of this chair without passing out! Mouth, start salivating!” And all that just comes from hearing a ceramic rattle in my kitchen. Had I heard the same sound in a restaurant, it might have activated memories of waiters stacking dishes instead.
Language processing as prediction
Predicting what comes next is at the center of language processing as well. As soon as you start reading something, your brain activates a slew of top-down models that give meaning to language. It uses phonic models and spelling patterns to predict and identify words. It uses pre-existing concepts to define them. It uses grammar to predict how the words are connected. And it uses discourse models to predict how each part fits into the whole.
Some of the discourse rules for English expository writing are that the main ideas are most likely to be given at the beginning of the paper, a section, or a paragraph.[6] Thereafter, as you read along, these higher-level ideas are backed up with supporting ideas (logic, examples, analogies, etc.). This subordination is the straight line Kapan drew for English. Once in a while, we put the details first and main idea at the end as an inference, but not that often, and when we do, we usually start with a question to show the reader we are using inductive organization.
[6] Fascinating! While editing, Stephen M. Ryan wrote: “This is not how I was taught to write essays (in English) in the UK. Everything in the paper builds to the conclusion which, for the first time, makes explicit the main idea of the paper. I had to get over this when I started to use US-conceived textbooks for my writing courses and they insisted on early revelation of the main point.” Do we need another squiggle?
To understand the importance of discourse rules, think about how important an introductory paragraph is to an expository article, speech, or YouTube video. It usually has three parts: 1) an attention getter (Can’t sleep?), 2) the main topic (Then you are probably using your cell phone too much.), and a guide to the rest (Let’s look at how blue light affects sleep, how it has changed sleeping patterns in today’s youth, and what we can do about it.). This particular feature of English exposition makes following the whole much easier. Why? Because it directly aids prediction as you read, listen, or watch along.
Think about how much you predict about a reading just by knowing what kind it is beforehand. Were I to say that you will read a limerick, Times opinion piece, love letter, results section, or advertisement, you already call up mental models on what the structure and language should be. Or even without me providing the priors, think about how sentences like these activate expectations on what will follow:
It was a cold and stormy night.
Brexit or Regrexit?
She could not get his musty fragrance out of her mind.
There were three hypotheses.
Burdened by debt?
To be or not to be.
What do these ten movie stars look like today?
Roses are red and violets are blue.
A departure from these rules, though sometimes done for artistic purposes, can be jarring:
Roses are red and violets are blue.
Only the second was supported by the results.
As long as a passage follows the expected discourse rules, whether they be overall structure, paragraph organization, or word meaning, we can pretty much get the meaning the writer intended. Not following the rules makes a passage seem disconnected and incomprehensible. As Kaplan discovered, that is usually what happens when non-native speakers use their own L1 discourse rules to write or read English. At Kaplan’s university, the regular course teachers, unfamiliar with the discourse models their international students were using and unable to follow their logic, would scrawl the same three words on their papers, “What’s the point?”
Because discourse rules can be seen as top-down models that aid in predicting meaning, then, it is clear that teaching our students how to use these models is of utmost importance in the writing class.[7] However, it is not unusual for reading and writing teachers to be unaware that we use these rules, and so, to focus on teaching grammar and vocabulary instead.
Now that we know the importance of discourse tools for predictive language processing, let’s avoid that mistake. Let’s make sure that students have the right discourse models available when reading, and use them every time they put pen
to _____.
Curtis Kelly (EdD.) was until recently, a professor at Kansai University. He is the founder of the JALT Mind, Brain, and Education SIG, and producer of the MindBrainEd Think Tanks. He has written over 30 books and given over 500 presentations. His life mission is “to relieve the suffering of the classroom.”