The Reading Mind: A Cognitive Approach to Understanding How the Mind Reads

Introduction

The Chicken Milanese Problem

Agenda for the Introduction

To consider the question how does the mind read? More specifically, to understand why it is a terrible scientific question, and why we pose it anyway.

Picture this commonplace scene. I was on an airplane, reading E. L. Doctorow's Billy Bathgate on my Kindle. The following passage is found near the end of the book, and when I read it, I softly gasped.

Before he got through it I was hearing the distant sound of police sirens, and it was so arduous for him to speak it that he died of the effort: Right, he said. Three three. Left twice. Two seven. Right twice. Three three.¹

My goal in this book is to account for what happened in the few moments it took me to read those 43 words.

The environment held nothing more remarkable than black marks on a white screen, yet somehow I was mentally transported to another world, indeed, to a world quite alien to me: New York City some 30 years before my birth, populated by gangsters. How does the mind create a mental world from black marks? And why would I care enough about Otto Abbadabba Berman—a real‐life gangster portrayed sympathetically in this novel—to gasp when he's murdered?

The approach I'll take to answering these questions is cognitive. I'll describe what the mind is doing as we read, but I'll seldom consider what the brain is doing. That may sound shortsighted (the mind is, after all, what the brain does), but it's a common scientific approach taken over the last fifty years. Computer science offers an analogy. You can describe the steps of a calculation—say, figuring out the date of the next lunar eclipse viewable in Toronto—without describing what's happening in the electronic guts of the computer during this calculation. In the same way, I'm going to describe the steps by which your mind reads without specifying how the brain carries out those steps.

Cognitive psychologists commonly tackle large, daunting questions by breaking them down into smaller, more manageable questions. We do that by thinking through what had to happen in order for some bit of mental work to get done.

What had to happen between my seeing the letters on the screen and my emotional reaction to the events in an imagined world? I had to see the letters and identify them. I had to assemble the letters into words, and then the words into sentences, which I comprehended by applying grammatical rules. My emotional reaction entails not just comprehension, but memory. He died of the effort prompts pity only if you feel like you know Berman. So over the course of the novel I must have built and updated a sort of personality picture of this character. And of course memory is needed to organize the sequence of events into a coherent sense of the plot.

So, will this skeletal outline of what happened as I read Billy Bathgate serve as a starting point for a theory of reading?

How Do You Make Chicken Milanese?

Even my crude analysis shows that what happens when we read? is a bad scientific question, the type of question psychologists usually don't pose. Why? Think of all the millions of activities your mind can direct: you can guess the cost of a paperweight you see in an antique store, ride a child's tricycle for comic effect, make Chicken Milanese, invent a plausible excuse for missing your neighbor's son's middle school play, and so on. For each of these we might pose the question What's happening in the mind when you do that? But scientists don't. The reasons that scientists don't ask how you cook Chicken Milanese inform what I've included and excluded in this book, so it's worth describing these reasons in some detail.

The first reason is that task descriptions are not quite as simple as I've made out. I said let's consider what had to happen as I read that passage from Billy Bathgate, and then I said something like you have to perceive the letters, and understand the words, and so on. The history of psychology shows that it's easy to be fooled when you try to describe a task.

Here's a simple example. When we read it feels as if we move our eyes smoothly—we sweep from the start of a line to the end, and then snap back to the far left of the page for the next line. That impression is easily disconfirmed by watching the eyes of another person as she reads. Her eyes don't move smoothly, but instead jump from one spot to the next, usually a distance of seven to nine letters.² That's so easily observed it's probably been known for centuries. But even that observation—jumping movements, not smooth tracking—is an incomplete description. In fact, your eyes are not always pointing at the same letter when you read.³ About half the time each eye looks at a different letter. They may even be slightly crossed.

The implications of this fact for an understanding of reading are not obvious. I raise the issue to point out that researchers have been working at an account of reading for over a century, and they are still finding ways of improving their description of what's actually happening when someone reads—not how they do it, but what they are doing. That's one reason psychologists usually don't try to explain really complicated behaviors. They figure that they probably shouldn't be confident they can adequately describe what they are explaining.

Suppose we give up on the idea that we'll have a perfect description of what people actually do during a complicated task, and we decide to settle for a provisional description. That's not a bad strategy—as we learn more, our description of the task will improve. One thing we're pretty confident about is that a complex task will require many different cognitive processes. My off‐the‐top‐of‐the‐head analysis of reading called for vision, memory, grammatical analysis, language comprehension, and emotion. Any one of these mental processes is known to be terribly complicated.

Consider seeing letters. One challenge is that letters can take on quite different appearances, varying in size, typeface, and typographical emphasis (bold, italic, etc.) (Figure I.1). How does my visual system treat these very different‐looking objects as equivalent?

Figure I.1. One letter, different fonts. These letters must all be interpreted as equivalent, even though they look different.

© Daniel Willingham

Worse yet, the very same shape might be interpreted as representing different letters, depending on the surrounding context (Figure I.2). So we need to do more than define what makes an A an A; we need to specify the context in which it will be seen as an A.

Figure I.2. Ambiguous letters. Although I'm sure you read this sentence easily, if you look closely you'll notice the es in the word need are the same shape as the c in clean. And the same shape is interpreted as a v and as a u in the word volunteer.

© Daniel Willingham

Finally, note that we've taken for granted that we're looking at black characters on a white background. How could it be otherwise? But what dictates that the black bits define the objects, whereas the white is background? Differentiating objects from their background is so embedded in our visual system that we seldom notice that it's an issue, unless we're looking at one of those clever images where the object and background are ambiguous (Figure I.3).

Figure I.3. Reversible figures. At left, the black area can be seen as an object (vase) and the white area as background, or the white area seen as the object (dog profiles) and the black as background. At right, the mirror is seen as background when we focus on the woman and her reflection, but it may also been seen as the foreground object—a skull.

Vase © Tasha Volkova via Shutterstock; All is Vanity by C. Allan Gilbert. Public Domain http://bit.ly/2a2Nddd

Now suppose the complicated work of identifying letters is complete, and I've assembled the letters into words. My mind is trying to sort out the meaning of what I've read. One problem is that some words have multiple meanings. In the Billy Bathgate passage, Berman starts his brief speech with the word Right. What did I think he meant when I read that word? Morally correct, as in the right thing to do? Or perhaps agreeing with the facts of the matter, as in You got that right. Or appropriate, as in the right tool for the job. A few words later Berman uses the word left and so it becomes clearer that when he said right he was probably signifying a direction. But before I read left, what did I suppose right meant? Did I suspend judgment, hoping for clarifying information later? It doesn't feel like we do that. For example, if you read, Later that afternoon, he went to the bank, you probably don't think to yourself, he might have gone to a financial institution or he might have gone to the edge of a river bed, but I don't know which. You pick one meaning and go with it. But how do you pick? And what about sentences that require you to keep more than one sense of the word in mind simultaneously? For example, in Chapter 1 of The Adventures of Tom Sawyer, Twain says that two fighting boys covered themselves in dust and glory. So a single instance of the word covered is simultaneously literal (for dust) and figurative (for glory). How does that work?

I could go on and describe how reading depends on memory, on your powers of inference and problem solving, even on your ability to time movements with great precision—eye movements must be perfectly synchronized with ongoing reading—but you get the idea. Reading—or cooking Chicken Milanese, or any moderately complex task—calls on so many mental processes that an agenda to provide a cognitive explanation comes perilously close to the goal I'm gonna explain the mind. There's too much.

Let's Do It Anyway

I've discussed two reasons psychologists don't pose questions like how do people cook Chicken Milanese? or how do people read? First, because the task is complex we suspect we'd botch our description of what people are doing as they attempt it, so our theory of the mental events underlying the task would be wrong from the start. Second, the complexity of the task suggests it calls on many mental processes, and a theory of how the mind achieves the task may be too ambitious.

But reading differs from cooking Chicken Milanese in an important way: it matters. Reading matters in our day‐to‐day affairs, in our culture, in our economy, in our civic lives, in the arts, and so on. There are stakes attached to people reading well or poorly (Figure I.4). It's true that psychologists seldom try to account for really complex tasks, but they make an exception for reading, as well as a handful of other consequential tasks, like driving a car.

Figure I.4. Medicine packaging. Literate people likely do not notice how frequently they rely on the ability to read. Here I've blurred the print on medicine packaging to help you imagine being unable to read, and trying to select a medicine for your child's sore throat, perhaps by attempting to match packaging with your memory of television commercials.

Still, the Chicken Milanese problems are real, so we need to deal with them as best we can. What can we do about the task description problem? What if we're trying to account for how people read but we're getting wrong what reading really means, just as people used to get wrong the bit about eye movements during reading? The brute truth is that there's no solution. That's the nature of science, and the best we can do is keep the possibility in mind, and try to be clear‐eyed when we describe the process of reading. Thus, in this book I'll devote a lot of energy to analyzing the task of reading.

How about the complexity problem, the fact that there are so many processes required to read? Here, we must accept that our account of reading will be incomplete. We can't fully describe how people identify letters, how they separate objects from background, and all the rest. So the question is, what will we try to explain, and what will we disregard?

We might say to ourselves well, we don't need to explain how people separate letters from background because that's not really a reading process. It's a process for all of vision, and you happen to use it when you read. So let's set the goal of accounting for the reading stuff, and we'll leave the more general‐purpose thinking processes for someone else to figure out. That won't work, because all of the mental machinery that supports reading is borrowed. Reading is less than 6,000 years old; that's precious little time for any reading‐specific thinking processes to have evolved, and there's not much evidence that any have. The mental processes that contribute to reading evolved for another purpose, and we co‐opt them for the act of reading.

A better principle will be for us to ignore the mental processes that don't vary much from person to person. Separating objects from background is a good example. Yes, it's a complicated, mysterious process, but somehow anyone with typical vision does it. Crucial to our purposes, when people struggle to read it's not due to a failure of this process. And strong readers are no better than average readers in separating objects from background. So although this process is indispensable for reading it's not the first thing we want to explain.

Researchers do know something about the mental differences between strong and weak readers. A strong reader has a broad vocabulary, and would know the meaning of the word arduous in the Billy Bathgate passage. A strong reader would comprehend that Berman is telling Billy the combination to a safe, even though the safe goes unmentioned in the passage. Come to that, we'd guess that a strong reader would be reading in the first place, on a plane, when he could be playing a game on his phone, watching a movie, or sleeping. These factors—broad vocabulary, good comprehension skills, motivation—are quite commonsensical. Accounting for them will get more interesting when we start to engage in task analysis, as I promised we would. What are the differences in personality or attitudes between the people on an airplane who choose to read and those who don't? Is reading from a Kindle different than reading from paper? How can you make sense of a passage that depends on your knowledge of combination safes, but doesn't mention a combination safe?

The Plan of the Book

I've emphasized the complexity of reading. We will therefore take it one step at a time, starting from the ground floor, so to speak: how readers see letters, then moving on to how they see words, then sentences, and so on, with one chapter devoted to each topic. In keeping with the emphasis on close analysis of the task, each chapter frames an aspect of reading as a problem: how is this bit of mental work accomplished?

That emphasis on task analysis will also prompt us to begin not with reading, but with writing (Chapter 1). Considering its purpose will help us better appreciate what readers actually do when they read letters and words (Chapters 2 and 3). From there we can consider word meaning (Chapter 4), and the comprehension of sentences and paragraphs (Chapter 5). Having this understanding of the process of reading will prepare us to consider why people might be motivated to read—or not (Chapter 6). Finally, we'll consider the possibility that digital technologies should prompt us to rethink everything we know about reading, as they have so radically changed other aspects of our lives (Chapter 7).

Before we plunge into this content, let me draw your attention to two limitations of the topics covered. First, this book is offered as a summary of one scientific approach to the study of reading—namely, the cognitive approach. That's not the only scientific perspective on reading. Another scientific literature employs the sociocultural view, which emphasizes the role of the social environment in reading; what you read, how often you read, your interpretation of what you read, and your thoughts and beliefs about reading are all influenced by the people around you and your relationship to them. The cognitive approach is not in opposition to this view; it's just different. As you'll see over the course of the book, it is more concerned with picking apart the mind of the reading individual. It's no accident that I titled the book The Reading Mind, and not The Science of Reading.

The second limitation of this book is that it's not about how people learn to read. I aim to describe how an experienced reader reads, not how a novice learns. That said, a great deal of reading research has been conducted with novice readers, and some it will be relevant to our purpose. I'll flag these studies when I refer to them, to help keep clear in your mind the difference between the mind of the expert reader and the mind of the learner.

Although this book is not offered as a summary of the learning‐to‐read research, some of the conclusions drawn may be applicable to education. However, these implications must be drawn with caution. This book is based on basic science, and basic science seeks to describe the world as it is; in this case, to describe the mind of a reader. Education is not a basic science, but an applied science. Applied sciences do not seek to describe the world as it is, but rather to change the world, to make it more like some ideal vision of what the world ought to be like. In the case of reading education the change is the transformation of people who cannot read into readers.

Applying findings from basic science to that effort is not straightforward.⁴ For example, many of the studies I'll cite were conducted with experienced readers, and their reading may be different than that of those learning to read. In addition, many studies deal with one, isolated aspect of reading—how we know the meaning of a word, for example, or how we read a misspelled word in the middle of a text. But when we consider reading education, we can't think about aspects of reading in isolation. Doing so entails the risk that we'll change instruction to improve one aspect of reading and unwittingly worsen another aspect. To provide an obvious example, long practice sessions studying letter‐sound relationships may help improve decoding, but it may also prompt a decline in reading motivation. With these cautions in mind, I will offer some thoughts at the end of each chapter as to conclusions that scientists can offer that might be useful to practitioners.

But I'm getting ahead of myself. Before we contemplate how the science of reading can be useful to educators, let's review some of the science of reading. True to our commitment to examine carefully what a task really entails, we will begin our analysis not with the reading mind, but with the