A Deeper Readability

Beyond a surface-level approach to readable writing

This article caps a series on crafting highly readable content based on cognitive science, an approach that goes beyond the most common, surface-level approaches. These techniques can reduce a reader’s cognitive load and lead to faster reading with greater understanding, recall, and enjoyment.

Image by FelixMittermeier from Pixabay

The approach includes two techniques based on how people read, seven techniques based on how people learn, and a grab bag of further suggestions to try. If you want to know more about the science and other background, read the other articles of this series, including A Cognitive Model of Reading. The entire series is listed at the bottom.

What I’m calling “surface-level” readability here is basically the use of simple words and sentences, which is the most common suggestion we hear for increasing readability. It’s popular because it’s easy to measure and it works. But we can do a lot more. My concept of “deeper” readability here is the sum of things like clear, visible structure, up-front context and reading goals, definitions, clear units in small groups with logical sequence, connections, examples, stories, and uninterrupted reading flow. These significantly help the reader to naturally, easily convert the internal word stream into schema — cognitive psychology’s term for structured memory.

Studies show that what makes experts different than beginners is not just more knowledge, it’s knowledge that’s better organized.¹ Creating that for readers lowers their effort and strengthens what they learn. But it asks more of the writer than just shorter words and sentences.

Helping Readers Read

How do letters on a page become a voice in our heads? This section explains why short words and sentences improve readability.

A Quick Look

The visual, auditory, & language areas of the brain try two strategies to read a word: a fast visual-recognition path for short, familiar words, and a slow, phonetic path for long or less-common words. The decoding process is not linear or black-and-white; it streams word possibilities through multiple parts of the brain, all influencing each other. One consequence is that context has a surprisingly powerful influence on how we interpret text and even what letters we see. The result is an internal audio stream, like a voice in the reader’s head.

What It Tells Us

• Prefer short, familiar words. It’s easier for all readers, and especially for dyslexics, to recognize a word visually than to “sound it out”. This saves time and mental energy. Because of the brain’s holistic approach, even using words that are easier to spell or faster to pronounce has been shown to improve reading speed. Readers pay for your word choices, so don’t use a $5 word when a cheap one will do. Save their brain power for thinking about your awesome ideas! If intentionally preferring a short, simple word seems crazy, take comfort from Ernest Hemingway, who won the Nobel and Pulitzer prizes but wrote some of his most famous novels (The Old Man and the Sea and The Sun Also Rises) at the 5th-grade reading level. When you use a new or strange term, immediately provide a simple definition; don’t make readers write their own.
• Provide context: Context changes a reader’s comprehension. Give them a mental framework for what they’ll read — for instance with clear titles, headings, and introductions. Your table of contents can help tell the story. Down at the word level, provide enough info to disambiguate a word (i.e., choose the right meaning), either before the word appears, or in the next 2–3 words while the ambiguous word is still in the reading flow. Context can lead or mislead, so provide topic, context, or other reading goals at the start of the document, section, and sometimes paragraph. You can even place context at the start of a sentence. This often leads to subject-verb-object (SVO) sentence structure, which is simple, active, and easier to read than complex or passive structures.

Helping Readers Learn

How does that voice become understanding and knowledge? This section opens the door to a more structural approach to readability.

A Quick Look

Working memory: Sensory inputs, such as that internal reading voice and the visual image of the page, are processed in small buffers (temporary memory areas) that include an audio buffer, an episodic (story) buffer, and a visual-spatial sketchpad (the mind’s eye). Without much conscious effort we can track about 3 seconds of audio, a general storyline, and an image with about 4–7 chunks (items or groups). For our purposes, you can think of these natural cognitive limits as “magic numbers”. Working memory is managed by a central executive process with similar limits. Writing that exceeds these natural limits makes readers stumble, break focus, and re-read. That reduces comprehension and recall, and creates frustration that some readers internalize.

Schemas: Knowledge is stored in long-term memory as a schema or structure of connected units. You can visualize a schema as an outline, a tree structure, or a large molecular model. We basically remember just a schema, not every sentence in a book. Because schema is built on connections and structure, it’s hard to learn random or disorganized information. Good writers provide structure, clear connections, and clues to pre-schematize content for the reader. Once a thing is in long-term memory, a little challenge or redundancy makes the new memory stronger. And over the next few weeks, sleep consolidates and further connects the new memory.

What It Tells Us

• Create units: The central executive looks for identifiable units of meaning, such as a name, phrase, clause, sentence, paragraph, or section. If a sentence or phrase is too long (such as a thing with a very long name), or contains too many chunks to easily juggle (more than 4–7), you’ll exceed the limits of working memory and the reader will have to re-read the sentence. Now, that last sentence pushes the limits a little — it’s over 50 words or 8 seconds long, with 6 main phrases and a long final stretch. The parentheses probably make it worse. If it made you stumble or stop in the middle, learn from my error. Signal the end of each distinct unit with punctuation, a conjunction (such as and, or, although, but…), or a paragraph break. In non-fiction, use section headings to show structure and set context, and limit the number of paragraphs per section to a magic number. This section has 7 bullets, a good maximum; how do you like the length? People I’ve interviewed prefer short paragraphs, and so do I. Every paragraph is a bite that the mind eventually has to swallow.
• Chunk & sequence everything: If you have a long list of, say, 10 units, you’ve exceeded the magic number that working memory can easily wrestle into a schema. Break it into multiple shorter lists, like an outline. At each level, each sub-list should have less than your chosen magic number. (A good max is 7, but 5 or less is better.) Also make sure that every list at every level has a logical sequence. Studies show that people can easily memorize long lists that are logically grouped and sequenced. That’s what the mind likes, the stuff of schema. So use it for your overall structure and for sections, paragraphs, bullet lists, instructions, and even diagrams.
• Connect units with structure and language: Make large and small structures visible with headings, bullets, numbers, and so on. Use phrases like “the following 3 parameters:” that reinforce quantity and category. Make schemas stronger with connecting language that shows logical relationships between units, such as: first, also, or, but, although, still, because, therefore, always, as a result of, for instance. I’ll give you 3 reasons: It helps you write structurally, it strengthens the reader’s schema, and it gives your writing a sense of forward motion.
• Create structure at multiple levels: In the end, knowledge is structure and information is structure — from a 0 or 1, to the sound of the shortest word, to the sequence of words and chapters in a book, to the totality of the world’s libraries and data clouds. Like molecules gathering to form a crystal, thoughts and facts arrange themselves naturally on a well-designed structure, connecting in powerful ways that make them easy to read and remember. If readers can’t identify the structure of your book, chapter, or article, you’ve missed a chance to create deep understanding. Trying to fix a rambling, unclear structure with a coat of short, readable sentences creates a loose collection of fading impressions that readers struggle to connect for themselves.
• Don’t break the flow! Good structure plus readable, connective language can lift readers off the page and into the idea space. Avoid breaking that flow with large tables, irrelevant information, or other distractions. Would you have enjoyed Harry Potter more if every paragraph had a hyperlink in it? When you include a large diagram or table, provide a logical break so readers know when to stop and change mode. Avoid making readers look back and forth between text and a diagram. Consider moving large tables or non-essential side content to an appendix, and links to a resource list.
• Use story: Humans have learned through story since at least the first campfire. Both the working memory and long-term memory have episodic memory that integrates information into experiences. (Some say episodic memory is only about lived events, but I’m using it here as a stand-in for general story-based learning.) So engage it by using goals, problems, conflicts, themes, etc. — the essence of story — as context. For this year’s model, we needed to either increase performance while decreasing cost, or get out of the market. Here’s what we did. This turns a dry list of details into a cohesive solution in which everything connects and gleams with meaning. A list of unconnected items must be memorized, but a contextualized story is easily remembered.
• Use examples: They’re like stories, and they provide a good kind of redundancy that confirms and expands the reader’s understanding without boring repetition. Challenging and using new knowledge with questions and examples is a more powerful learning technique than simple repetition, probably because it forces the reader to strengthen connections in their schema. And examples can prove to readers that they really are learning something useful.

What Comes Next

Those nine tips are the main point of this article, but I’ll mention a few other possible suggestions to sketch a fuller picture of readability for the future.

Complexity

We can do more to limit the complexity of words, sentences, and overall structure. For instance, data suggests that we read a root word like name differently than a complex, built-up word like nominalization that’s constructed of roots plus prefixes and suffixes. Those elemental root words make writing easier to read while creating a strong, direct voice.

Then there’s sentence complexity. My simple suggestion was just to limit sentence length and item count, but Steven Pinker, in his 2014 book The Sense of Style: A Thinking Person’s Guide to Writing in the 21st Century, rolls up his sleeves and analyzes sentence diagrams. He shows how some sentence structures cause exponentially more work for the reader — which a committed writer could learn to avoid.

And then there’s the overall structure or information architecture of a piece, which has its own principles worth exploring separately.

Pace

There must be a maximum rate at which humans can absorb information. One way to measure this is lexical density, which is the ratio of content words like kite or crash to more functional words like to, of, or that. Text with low lexical density seems to go nowhere, but text that’s too dense is hard to keep up with. There are other ways to measure this, such as the concepts of entropy², work, or velocity — how much information or surprise is transmitted, or how quickly a text gets to the point. These are ways to look at conciseness. Writers should distinguish between real information (you won the lottery!) and the other stuff that’s repetition (yes, the actual lottery!) or fluff (congratulations!). The ideal readable text is in a Goldilocks zone that’s not too fast or too slow. So be concise, but use a pleasing mix of sentence lengths and create little cognitive breaks with phrasing, paragraph breaks, and so on.

Getting Personal

We know a lot about how memories are initially formed, then strengthened by rehearsal and challenge, and gradually consolidated over many nights of sleep. And we know that engaging more parts of the mind can make a message more understandable and memorable. So it’s smart to use the reader’s existing concepts in definitions, and use sensory and emotional input. One example is the SUCCESs model⁴ of stickiness used by two psychologists from Duke and Stanford, which tells us to use information that is: simple, unexpected, concrete, credible, emotional, and/or a story. Future readability tools could look for this. Are emotion and sensation really part of “readability”? Well, if the reader can’t recall your message the next day, what’s the point?

Most readability measures don’t check this, but the 1948 paper³ that told us to use short words and sentences also told us to use personal words. This meant names, personal pronouns, quoted speech, and direct communications to the reader. I would add that, when possible, it’s better to talk directly to your reader (hit Enter) than to use elaborate third-party phrasing (the user should hit Enter).

Research at the Alda Center for Communicating Science at Stony Brook University⁵ shows that when we make a connection with the audience we measurably improve communication — again suggesting both personal language and simple, direct talk. Along these lines, studies show that a reader’s sense of value or interest in information, along with their confidence that they can handle it, are key predictors of successful learning.⁶ To help with this, our writing might provide meta-information (this one-hour section covers 3 points…) and offer encouragement (most readers find this part challenging, but taking a break in the middle helps). Again, we might even phrase this directly to readers (you’re almost done!).

This brings me to cultural specificity. Who are we speaking to? If readable best practices include short, familiar words and short, familiar sentence structures, what about language and ideas that are familiar to one group but unfamiliar to others? If we try to include everyone, will we wind up in a dull world of gray jumpsuits and rounded corners where everyone speaks blandly and colorlessly in the world’s shortest words? That’s a lot of questions. I explore some possible answers in Readability Equals Translatability. One possibility is to let your ideas create the excitement, and don’t rely too much on difficult or culturally-specific language.

The Silicon Valley marketing writer and consultant Anne Janzer, in her 2018 book Writing to Be Understood⁷, suggests we should be empathetic — commit to knowing our audience, understanding their situation, and showing a little humanity, humility, and respect. Here again is the suggestion to think about and connect with readers, not only to target their needs and speak their language, but to earn their trust and keep them reading.

And finally, we’ve mentioned enjoyment a few times but not really dug into its effect on readability. Seems like a lot to ask: be informative… be easy to read… and be entertaining? But remember, humans can enjoy things like useful knowledge, a sense of surprise and forward motion, a well-made argument, and the occasional glimpse of a person behind the curtain.

Final Thoughts

Readability is a collective quality that makes text easy to read, as measured by tests of speed, comprehension, and recall. A growing number of remarkable tools can analyze text for these qualities, but they only take a surface-level approach. To really explore the potential of readability — that maybe it’s not the subject that’s murky, but the writing — we need to see writing as more than words and sentences lying flat on a page. And we need to challenge people’s low expectation of just how transformative good writing can be.

Reading researchers come at it from the other side. They’re trying to help people read the world’s text — not make the text more readable. It’s up to us as writers to do that. Readers will get better at reading us, and writing tools will get better at advising us. But we writers should go to meet them with our own improved craft. Some of that can come from an understanding of reading, minds, and learning.

Resources

Articles in This Series

• A Cognitive Model of Reading - The science of how people read
• Surface Readability - The value and limits of simple “short words & sentences” approaches to readability, and some interesting variations
• A Deeper Readability - Techniques that go beyond surface-based approaches, based on cognitive science and other sources
• Writing for Challenged Readers - About ESL & dyslexic readers, and what they said helps them most
• Readability Equals Translatability - How the right approach to readability becomes a scalable approach to fast, consistent translation across multiple languages, how that works in a modular, single-source content management system, and whether language must be “dumbed-down” to achieve readability

Bibliography

The “nine suggestions” are based on the science in A Cognitive Model of Reading, which lists its own sources. These other sources were also helpful to this article.

[1] How People Learn, National Research Council, 1999 — Lots of good stuff in this free PDF book, such as the discussion of expertise and organized knowledge in Chapters 1 & 2.

[2] Introduction to Information Theory, John Pierce, 1980 — In information theory, entropy is a measure of transmitted information or surprise. (This is different than entropy in thermodynamics.) In a chapter on “Psychology and Information Theory”, Pierce explores way to measure the entropy of human language in bits per second. You might also be interested in more recent approaches to this difficult problem, e.g., Entropy Measures of Human Communication Dynamics 2018.

[3] A New Readability Yardstick, Rudolf Flesch 1948, Journal of Applied Psychology, available for free in The Classic Readability Studies, William H. DuBay, editorhttps://www.thegreatcourses.com/courses/the-learning-brain

[4] Made to Stick, Chip & Dan Heath, 2007 — Two psychologists from Duke and Stanford discuss how to make information “sticky” (understandable, memorable, and perhaps viral) by using information that is: simple, unexpected, concrete, credible, emotional, and/or a story. (This roughly spells SUCCESS, they point out, though mnemonics are not on their list.)

[5] If I Understood You, Would I Have This Look on My Face?, Alan Alda, 2017 — This entertaining book describes the Alda Center for Communicating Science at Stony Brook University and what’s been learned there.

[6] The Learning Brain, Thad A. Polk, 2018 — This entertaining short lecture series is a great way to pick up current theory on learning and memory formation and consolidation, including the four key predictors of successful learning (“How Motivation Affects Learning”).

[7] Writing to Be Understood, Anne Janzer, 2018 — Good advice, often from unexpected directions, on how to connect and communicate through business and technical writing.

[8] The Sense of Style: A Thinking Person’s Guide to Writing in the 21st Century, Steven Pinker 2014