Doug Lemov's field notes

Reflections on teaching, literacy, coaching, and practice.

12.20.16My Favorite Mistake: Knowledge Problems and Skill Problems

Image result for bathing on saturday

Above: How to spend a Saturday


I’m lucky enough to get to learn constantly in my work and that inevitably involves recognizing my own mistaken thinking—things I believed but that I think time and evidence have proven were wrong.

Here’s an example: A few nights ago as I was tucked in, reading aloud from Laura Ingalls Wilder’s On the Banks of Plum Creek to my littlest at bed time, we came across a seemingly innocuous sentence and I paused.  As I reflected on it, I realized it was a perfect demonstration of my personal biggest mistake about learning.

Here’s the passage. It’s about the family deciding to go into town for Christmas services at church (a big deal). The phrase that caught my attention is in italics.

“But why….” asked Laura. “Why are we going to town at night?”

“It’s a surprise,” said Ma. “Now no more questions. We must all take baths and be our very nicest.”

In the middle of the week, Ma brought in the washtub and heated water for Mary’s bath, then again for Laura’s bath and again for Carrie’s.  There had never been such scrubbing and scampering.

In the middle of the week: I put emphasis on the word ‘middle’ as I read the phrase to little Goose and added a tone of surprise.

I paused briefly to ask her why that phrase had been included in the sentence. Why would she make the point that the washtub was being used in the middle of the week? Goose was stumped, and the reason she was stumped is the reason most readers miss most of the inferences they miss when reading.

First, here’s what you’re supposed to understand:  Families like Laura’s, living on the Western frontier in the second half of the 1800s, typically bathed once a week, on Saturday specifically, in anticipation of going to church on Sunday.  The task of fetching water and heating it was onerous and time consuming- probably wasteful too since water was scarce.  Because I’d read the entire Little House series to my older daughter, not to mention a bunch of other frontier narratives, I knew this.  It was clear to me that “In the middle of the week” was used to express surprise. As in, “Can you believe we took baths in the middle of the week? Wow!”

My daughter of course did not realize this. She’s an attentive and intuitive reader- insightful about people’s motivations, but the problem here was that the inference, like most inferences, didn’t require “skill.”  No amount of attentive and intuitive reading and instinct about Laura’s or Ma’s motivations would have helped her realize that Laura’s family would have only bathed on Saturdays.

So I explained the background to her- what it was like on the prairie and why people only bathed once a week.

But in our schools—honestly, even in the classrooms at Uncommon Schools where I work—we perceive problems of understanding to be problems of skill.  We hammer away at lessons teaching kids to ‘make inferences’—telling them to combine what they know with clues in the text for example, and then practicing making inferences from a text they understand.

This clangs with me because, to paraphrase Jung, the faults that annoy us most are the ones that remind us of ourselves.  Honestly, I over-concluded about the importance of “skills” in reading for years. I advised the teachers who worked for me to plan lessons based on things that appeared to me to be skills: inferences foremost among them.  Now I’m pretty sure that was a mistake.  I should note that I’m not big on binaries.  I don’t think it’s true as some advocates of background knowledge state that there are “no such thing as skills” in reading. I just think we vastly over-emphasize them and overlook the ways that knowledge is the larger factor.

There are lots of great explanations of why making inferences is a knowledge problem, not a skills problem, by the way.  My favorite is in Dan Willingham’s in Why Don’t Students Like School?

But I actually think this problem of mistaking knowledge problems for skills problems goes further than just inferences, as Willingham discusses.  Main idea questions are a great example.  For years I thought of this as a skill and I suggested teachers practice it. There’s probably a bit of skill to describing the main idea. If I practice describing a section of text in a single sentence multiple times, I’ll probably get a bit better at spotting and describing what’s most salient.  But a much bigger proportion of the problem when a student can’t explain the main idea is likely to be knowledge based: she lacked the vocabulary (a form of knowledge) to grasp meaning; she lacked the background knowledge to get the inferences that built up to the meaning.  Main idea just isn’t that tricky a skill to execute. It doesn’t need to be broken out into four steps.  You can practice it a few times but then, really, it’s about working to build and reinforce knowledge.

Consider that the weakest reader in your class makes successful inferences all the time about things he knows about: in movies he is quick to spot a villain via a series of deft inferences-his tone of voice or the music that signals his entrance.  Your weak student can easily infer the main idea in situations where he has context too.  The problem at school is much more likely to be an issue of knowledge. The main idea of the passage above is that taking a bath on the weekday signaled to the family that something very special was afoot. The problem if a reader didn’t get it isn’t that they don’t understand how to try to isolate what’s most important. It’s that they don’t understand. So we should probably spend less time trying to skill-i-fy reading gaps and more time investing in knowledge building. No amount of practicing main idea exercises or meta-discussions about ‘what we’re really doing when we find the main idea’ are going to help. The skill is simple; the application is complex because it requires knowledge.

There’s a good analogy here is the idea of context clues.  A lot of teachers believe in context clues as the best way to build vocabulary.  You teach students to infer word meaning from the clues around a hard word in the text, they say, and you have the killer app. Students can then learn any word form any sentence. But as Isabel Beck brilliantly explicates in Bringing Words to Life, context clues are unreliable at best. In some texts they can be used to infer word meaning, but in more sentences the clues are so subtle as to be easily misinterpreted by novices. Or they require knowledge to make the inference and therefor are not accessible to many readers. Or they are neutral—the word is not supported in anyway by the text around it.  “There had never been such scrubbing and scampering,” Wilder writes in the passage above.  There is no indication at all what scampering means.

In other cases context clues are mis-directive.  The more you understand what appear to be hints the more they lead you astray. For example: “The climb up the cliff face was grueling and the boys’ hands were chafed and bloody when, exhilarated, they reached to top.”  The context in that sentence would cause you to think that exhilarated meant something like miserable or exhausted. In actuality the word exhilarated is placed as it is in contrast to what you might expect from the rest of the sentence in order to create a sense of surprise- it’s craft and meaning are deliberately unexpected and therefore the context is mis-directive.

Similarly if you understood that the sentence, ‘There had never been such scrubbing and scampering,’ was about bathing you might assume scampering was something that happened inside the tub, maybe a form of washing, but here it introduces, without clues, a new idea into the sentence: what happened after the washing.

Guessing when you lack knowledge and when cues are often non-directive or mis-directive is therefore a waste of time, Beck asserts. The more you try to guess, the further astray you’ll get. You can’t guess (or infer) your way out of a knowledge problem very easily. Better to give students the definition (or other form of knowledge)  and ask them to apply it in various ways. “Why might the boys be exhilarated at the top of the climb?”

Or “Scampering means running with quick and often irregular steps. An animal trapped inside might scamper. So: What sort of scampering do you think was going on in the Wilder house? What about their living conditions might have encouraged ‘scampering’ after a bath?”

You might do something similar with other applications of background knowledge. Better to give students the information and ask them to apply it.

“People on the frontier generally only bathed once a week- usually on Saturday so they’d be clear for church on Sunday. “What do you think they had to do to get ready for a bath?” Or even: Here is a picture of a child having his Saturday bath around this time in history.  What observations can you draw about the lives of people in this era that are useful to understanding Laura’s surprise at taking a bath mid-week?”  (see the picture at the top of this post) Ideally in this case rather than asking the reader to interpret something she doesn’t know enough to interpret, i am building more background knowledge so she knows more next time around.

For a long time I thought this was a waste of time.  That it was ‘teaching isolated facts instead of skills’; now I am not so sure there is a skill there but I increasingly see how those facts don’t stay isolated for very long.

Anyway, this is the place I’ve evolved my thinking most this year in reflecting on teaching: we over skill it.  We hammer away at questions that presume kids don’t know how to make and inference or spot a main idea when it would be better to invest in giving them the background knowledge and then asking them to apply that knowledge in thinking about the text.

Now when I plan a lesson (or read with my kids) I prime them with background knowledge or ask them to apply background knowledge as often as i ask them to interpret and infer.

I’m going to keep thinking about that and looking for other mistakes in my thinking in 2017.  Happy Holidays.

, , , ,

6 Responses to “My Favorite Mistake: Knowledge Problems and Skill Problems”

  1. James
    December 31, 2016 at 8:47 pm

    Loved this post! It really got me thinking and sent me to re-read The Knowledge Deficit over break. Thanks! I do wonder if there is a general knowledge test for elementary students to assess their levels of knowledge in terms of history, literature, science, phrases and idioms. Preferably, it would be a listening test. It might be nice to have one as it would keep general knowledge top of mind throughout the school year and help close knowledge gaps. Thanks again for this post!

  2. Helen Farrell
    January 8, 2017 at 8:09 am

    But would not the very fact that the author has stated “in the middle of the week” imply that this was not a usual occurrence, even if you were not previously aware of the fact?

    • Robert Columbia
      September 26, 2018 at 1:20 pm

      Ms. Farrell,

      I don’t think that mentioning that this happened in the middle of the week necessarily indicates that it’s unusual. We can distinguish between these passages where there really is a clue and passages that are more “flavor text” – intended to allow the reader to immerse themselves into the author’s vision. Sometimes the entire description is not relevant to the plot or theme per se – certainly you could “get” the themes and plots of Tolkien by skipping all of his poetic descriptions of the landscape of Middle Earth, but would it really be the same experience?

      For example, suppose a character is described as wearing a brown dress to school and it is mentioned that she had two pencils tucked into her school bag. What can we conclude from the passage?

      1) That wearing brown dresses to school was unusual?
      2) That it was unusual for girls to wear dresses at all?
      3) That students carrying their own pencils to school was unusual?
      4) That it was unusual for girls to actually go to school at all?
      5) That school bags were rare?
      6) Perhaps this is one of those post-apocalyptic stories where the birth rate has plummeted and so seeing any child at all is astonishing?

      The answer is, of course, that we don’t know without more context. And that goes back to the original issue – that without the knowledge that mid-week bathing was unusual, how can a reader even hope to determine whether the passage is intended to indicate something special or whether it was just supposed to be flavor text?

  3. Colin Heydt
    January 14, 2017 at 2:52 am

    I wonder about the ‘skill vs. knowledge’ framing of the question (as I did in reading your excellent book on literacy). I’m a philosophy professor, and though I don’t teach much logic or critical thinking, those courses are staples of the undergraduate philosophy curriculum. It could be, of course, that students learn very little in them. But here’s an alternative account: by understanding how inference works, students get better at spotting errors (e.g. standard fallacies). Students don’t learn a “skill” of inference–it’s not like learning how to switch chords deftly on a guitar. Rather, they learn about how they infer (and how, ideally, they should infer). It’s a form of reflection–knowledge about how one comes to believe something to be true.

    So, shouldn’t the ultimate question be: “Are we helping students by giving them knowledge about how they infer while reading?” I agree that most problems of understanding are due to knowledge deficits: vocabulary, historical knowledge, understanding of pragmatic context, etc.. Children benefit more from learning about science, history, the vocabulary of their language than they do from learning ‘how to infer x or y from context clues.’ But there is a lot for students to learn about cognitive bias and about the process of inference itself that could be genuinely helpful. This way of changing the debate–not skill vs. knowledge, but first order knowledge (e.g. history) vs. second order knowledge (e.g. understanding how we infer)–might help our deliberation about what to teach and how to spend our classroom time.

  4. Ned
    July 5, 2018 at 8:02 pm

    I m wondering if there are any series for middle grade that could help syatematically build this kind of knowledge. This summer I’ll be helping a struggling reader who is a recent immigrant to the US and want to basically have him just read a bunch of books that will help him A) spend more time with a book in his hand, and B) just learn about the world. If there are good fiction series kids can get hooked to that help, it would be appreciated!

  5. January 5, 2019 at 6:35 am

    You might want to look to the Magic treehouse series (beginning chapter books) which has recently added non fiction companions to many of their stories. Grade 2-3 reading level, I think.

Leave a Reply