Why the same page is easy for one reader and impossible for another

Here is a real document in ordinary English, and you can open the page it comes from in about a minute.

On the evening of 8 June 1875, Sir John Hay rose in the House of Commons to ask the Government to open talks with the other countries bound by the existing steering and sailing rules, with a view to amending them, so as to reduce, in the words of his motion, "if possible, the lamentable loss of life now arising from collisions at sea". Then this happened. What follows is the entire remaining record of the Commons that evening.

The hon. Baronet was proceeding to address the House, when—

Notice taken, that 40 Members were not present; House counted, and 40 Members not being present,

House adjourned at ten minutes after Nine o'clock.¹

So what happened?

If you know how the Commons works, you have just read a small story with a beginning and an end, and it's not the story the words appear to tell.

If you do not, you have read something that looks like routine housekeeping. A man started speaking, there were not many people in, everybody went home early. Every sentence of that is true. It also misses what happened, which is the more interesting way to be wrong.

Now notice what did not cause the difficulty. There's no unfamiliar word in that passage. Present, counted, adjourned, Members, House. Two items need a gloss and neither is hard: hon. is short for honourable, and a Baronet is a title.

Two phrases are doing more work than they look. "Notice taken" and "House counted" are the formal names of a procedure rather than descriptions of somebody noticing and somebody counting. But you cannot tell that from the words, and nothing on the page flags them, which is the point rather than an exception to it: these are ordinary words being used as a term of art, and the only thing that would tell you so is knowing the institution.

What you were missing, then, is not a word you could look up. It's what each of those plain facts implies, and none of it is on the page.

• Forty is the quorum. Below forty the House cannot transact business at all. So the sentence is not a remark about attendance. It is the meeting ceasing to exist.
• "Notice taken" is somebody's act rather than a state of affairs. Any member present may stand and point out that the House is inquorate. Most often that's simply a member who can see the benches are empty and says so. Sometimes it is a member protecting the House from a vote taken on a thin attendance. And sometimes it is a member who wants this particular debate stopped, which the Victorian Commons complained about often enough that the complaint is part of the record too.
• A count-out ends the whole sitting rather than the item. Whatever else stood on the order paper went with it, and because the record reports only what was reached, you can't see from this page what that was.
• Ten past nine is early. This House regularly sat past midnight. In June 1875 thirteen of its sittings did, one of them until a quarter past three in the morning.¹
• The House had suspended at seven for dinner and resumed at nine. Ten minutes later it was counted out, which is the ordinary setting for a count: a thin House that has not come back. The page you are reading does not say this. The page before it does.¹
• "Was proceeding to address the House" means Hay had not finished his opening speech. The motion was never debated and never voted on.
• A private member had to win a night for a motion in the first place. Hay never brought this one back in the 1875 session.¹

Put those together and the story is this. A proposal to reduce deaths at sea was stopped dead by a member pointing out that fewer than forty men were in the room, before it was debated or voted on. Whether stopping it was the point, on that particular night, the record does not say, and the record not saying is the whole of what this lesson is about.

The record doesn't report a defeat, and the reason is simpler than indifference: there was no defeat to report. Nothing was voted down. The motion lapsed with the sitting, and what the record reports is the thing that actually happened, which is a room too thin to do business. A reader who followed the House knew what that meant, so nothing more had to be written.

Difficulty is a relation, not a property

This lesson is about one idea, and it changes what you do when a book defeats you.

Difficulty is not a property of a text, the way length is. It is a relation between a text and a particular reader. Those few lines of Hansard are trivial for one reader and opaque to another, and the lines didn't change.

That sounds obvious when it is put that way, and almost nobody acts as though it were true. The usual response to a page that will not open is to conclude something about yourself, and the sentence people reach for is "I am a bad reader". Sometimes the honest version of that sentence is "I lack the background this text assumes", and those two diagnoses point at completely different actions.

The mechanism: every text leaves most of itself out

Here is why background matters as much as it does.

No writer writes everything down. It would be impossible, and the result would be unreadable. So every text states some things and leaves the rest to be supplied by the reader: the relations between the things named, the reason one sentence follows another, the significance of a number.

The Hansard page never says that forty is the quorum, that taking notice is a choice rather than an observation, that a count-out kills the whole sitting, that nine o'clock is early, or that the motion won't come back. It assumes all of it. A reader who has it fills those gaps without noticing they were gaps, which is why that reader would tell you the passage was perfectly clear.

So understanding is not the passive reception of what is on the page. It is the page plus what you bring, and when what you bring is missing, you are doing a different and much harder job than the fluent reader beside you, on the same words.

The study everybody quotes, and the reason to hold it loosely

The famous demonstration is Recht and Leslie, 1988.² Sixty-four junior-high students were sorted two ways at once: by measured reading ability, and by how much they knew about baseball. That crossing is the whole design, because it produces a group who read well and know nothing about baseball, and a group who read poorly and know the game.

All of them read a passage describing half an inning, and then reconstructed what had happened using wooden pieces on a model field, which is one of several ways the study measured what they had taken in.

The result is the one the study is famous for: the poor readers who knew baseball did as well as the good readers who did not. Knowledge of the subject compensated for measured reading ability.

That is a striking finding and it gets quoted constantly, usually at second hand. Mine is second hand too: I have read its abstract and the published accounts of it rather than the paper itself, and that is worth your knowing before I tell you what is wrong with it. Here is the part that usually gets left off.

Reynolds, Hattan and Markham reviewed nineteen baseball studies published between 1978 and 2018, and found that thirteen of them used the same two measures of baseball knowledge.³

Worth knowing why the review was written, because the lesson would be quoting a referee when it's quoting a player. Its abstract says these studies keep turning up in teacher journals, books and podcasts, "especially within science of reading conversations", and it closes by asking the field to build on non-baseball work instead.³ That is a live argument about how reading is taught, and the reviewers are in it. It does not make the audit wrong. It does mean their opening concession counts for more rather than less, because they grant the knowledge link to an audience that had reason to want it qualified.

Their argument has two halves, and they are about different things. The first is about the knowledge tests: those tests "focused heavily on vocabulary and baseball trivia".³ Whether that is the same thing as knowing how the game works is the question the whole review turns on, and the reviewers take it that it's not. The second is about the reading passage, and it is the one that gets left out when this critique is summarised: "the most common baseball comprehension text was deceptively complex".³

That second half matters here more than the first, and it cuts against me as well as against them. My whole opening move is that the words are easy and the knowledge is what is missing. If the passage at the centre of these studies was not in fact easy, then the design wasn't cleanly separating what a reader knew from how hard the text was, and a difference the studies read as knowledge could have been partly a difference in handling a difficult text.

Read that carefully, because it is easy to over-read in either direction. The review is not saying that prior knowledge doesn't matter. It is saying that this family of studies may be measuring familiarity with baseball jargon rather than the deeper thing it claims to measure, and that a literature with thirteen studies sharing two instruments is narrower than its citation count suggests.

So where does that leave the claim?

The reviewers answer that themselves, in passing. Their opening clause takes it as given that links between knowledge and reading comprehension "have been widely documented for decades", and their conclusion calls for "reliance on non-baseball studies" to build the field on.³ That is not a case against the claim. It is a case for resting the claim on something other than its famous example, made by the people who have just audited that example.

I should say plainly what that's worth as evidence, because this lesson is about not taking a citation on trust. It is a background clause in an abstract, not a finding, and it points at a literature this course has not read. What it tells you is that the people closest to the critique don't think the critique touches the wider claim. What it does not tell you is how strong the wider claim is, and a course that has read one review and one abstract is not in a position to tell you that either.

So the claim is untouched by this critique, and the flagship demonstration of it is weaker than its fame. Both of those are true at once, and holding them together is a skill this course wants you to have, because you will meet the pattern constantly: a famous study, a real finding, and a gap between the two that nobody mentions when they cite it.

This is also why the opening example of this lesson is four lines of parliamentary record rather than the four lines about cricket that were the first thing I wrote. A cricket paragraph leans on jargon in the obvious way, on words like off-spinner that a newcomer plainly doesn't know, and a lesson that teaches this review's objection and then rests on an example committing it is not teaching the objection.

The Hansard lines are not free of the problem. They make the harder version of it. Two of their phrases are terms of art, as the sixth paragraph of this lesson says, but both are built from words you know, so nothing warns you and the definition test comes back clean. That's the case worth learning, and it is the case the baseball studies did not test. If you want the same shape in a domain you have no history with at all, the checkpoint at the end of the next section is the cleanest version.

Three obstacles, and how to tell them apart

When a passage defeats you, one of three things is usually happening. They feel identical from the inside and they need different responses, so the diagnosis is the skill.

Vocabulary. There are words you do not know, and they are load-bearing. The signature is that you can point at them, and the test is whether a definition fixes it. If you look the word up and find yourself exactly where you were, then what you were missing was not the word. The fix, when it really is vocabulary, is a dictionary or a glossary, and it's quick.

Missing background. You know every word and cannot say what the passage claims, or worse, you can say something confident and wrong. The signature is the Hansard page: fluent reading, and a plausible reading that misses the event. The fix is not in this text. It's another text, easier, on the same subject.

Syntax. One particular sentence will not resolve, and you can feel yourself rereading it and losing your place. The signature is that you can restate the paragraph around it but not that sentence. The fix is structural: find the main verb, find its subject, and put the qualifications to one side.

The three, side by side, with the question that separates them.

The question that separates the first two is the one the top block turns on: would a definition fix it? If looking the word up leaves you where you were, you are in the second block and no amount of dictionary work will move you.

Here is a sentence built to be the third case, in plain words:

The clause that the tenant, having given the notice that the agreement, unless the landlord had objected within the period the schedule sets out, required, believed sufficient was not.

Every word there is one you know, and two of them, clause and schedule, are legal usages you can read past without loss. The difficulty is entirely in the nesting, and the fix is mechanical, so let us do it.

Find the main clause first. Strip out everything that interrupts and what is left is the clause ... was not, which is a sentence with a hole in it. The missing word is not further on. It's behind you, in believed sufficient, and English lets you drop a repeated complement, so was not means was not sufficient. That is the whole claim. Everything else is there to tell you which clause.

Now peel the interruptions, outermost first. The outermost is that the tenant ... believed sufficient, which says whose judgement is at issue. The tenant believed the clause sufficient.

That is the distinction worth carrying: with missing background, more time on this page will not help, because what you need isn't on this page. With syntax, more time on this page is exactly what helps.

What to do when the diagnosis is "missing background"

This is the case I see handled worst, because the honest response feels like defeat.

The response is to find an easier text on the same subject, survey it with the six steps from lesson 2, and read enough of it to get the relations the hard text assumes. Then go back.

Three things make that less painful than it sounds.

It is often much less reading than it looks like from here. What you are missing is usually a small number of relations rather than a subject. The Hansard page needed six lines of explanation, not a course in constitutional history.

Surveying tells you whether the easier book has what you need, in about half an hour, before you commit to it. That's what lesson 2's procedure is for, and this is its most common use.

And you have already met the rule that covers the other half of this, in lesson 2: on a first pass through a hard book, read it through without stopping to look up or ponder what you do not understand.⁴ Now you know why that rule works. The first pass tells you which gaps are load-bearing, so the trip to the easier book is aimed.

What people get wrong

"I am a bad reader." Sometimes true, and in my experience much more often a misdiagnosis of missing background. The test is the page you started this lesson with: if you can read every word fluently and still come away with the wrong event, the problem is not the reading.

"Difficulty is a property of the book." It's a relation. This is also the mistake behind reading-ease scores, which count syllables and sentence lengths and cannot see the thing that actually made a passage hard for you. Writing Clearly lesson 1 carries the evidence for that, from a study of 518 readers in which reader background shaped what made a text difficult.

"Reading it again more slowly will fix it." It fixes syntax. It does nothing for missing background, because the second pass has the same gaps as the first. How to Learn Anything has the evidence on why rereading feels more useful than it is, in its first three lessons; this lesson adds the reason rereading cannot help here.

"Looking things up as I go is the responsible approach." On a second pass, yes. On a first pass it's how people spend a month on chapter 1 and abandon the book, which is the behaviour Adler's rule exists to prevent.

"If I need an easier book I should not be reading the hard one." The easier book is how you become the reader the hard one assumes. That is a route, not a retreat, and the alternative on offer is not reading the hard book successfully; it is reading it unsuccessfully.

Practice

Connections

Lesson 2 gave you a procedure for deciding whether to read a book. This lesson gives you the most common reason a survey comes back "not yet", and what to do about it, which is a different book first.

Lesson 4 takes up reading rates, and this lesson is why your own rate is not one number: the same reader moves at different speeds through texts of equal length depending on what those texts assume. Lesson 4 gives you the figure most adults average, and then has you time yourself on three texts of your own, which is this lesson's claim arriving as a measurement.

Lesson 5 is about coming to terms with an author, which is the other route to a missing vocabulary: building it from inside the book, out of the author's own use of them. The two routes are complementary, and the choice between them is mostly about how much the book assumes.

Go deeper

• How to Learn Anything, lessons 1, 2 and 6, on rereading and highlighting and why they feel more productive than they are. This course does not re-teach that evidence and leans on it.
• Recht and Leslie, "Effect of Prior Knowledge on Good and Poor Readers' Memory of Text", Journal of Educational Psychology 80(1), 1988. The baseball study itself, which is short. Paywalled at the publisher, and university libraries carry it.
• Reynolds, Hattan and Markham, "Fair or Foul? Interrogating the Role of Baseball Knowledge in Studies of Knowledge and Comprehension", Reading Research Quarterly 60 (2025). The review that should be read next to it, and worth reading as an example of how a literature gets audited. Also paywalled, but the abstract carries both halves of the argument.

Sources

1. Hansard, "NAVY—RULE OF THE ROAD AT SEA. RESOLUTION", House of Commons, 8 June 1875, vol 224 cc1561-2, historic-hansard. Quoted verbatim from that page, including the em dash marking the interrupted speech. Sir John Hay's motion is quoted from the same page; the transcription there carries a column number inside the sentence and breaks "lamentable" across a line, which is why the motion is given in summary above with only its closing phrase quoted. Those two artefacts are worth noticing for their own sake, because a digitised record is a transcription and not the thing itself. The dinner suspension and the nine o'clock resumption are on the page immediately before this one, which reads "And it being now Seven of the clock, the House suspended its sitting. The House resumed its sitting at Nine of the clock." The June 1875 adjournment times are on the same site and were read off it one sitting at a time: thirteen ran past midnight, the latest on 17 June at a quarter past three in the morning, and the Wednesday sittings ended around six in the evening, which is the convention rather than an early night. Hay's Hansard index for 1875 shows no return of this motion. On rights: the 1875 text was first published in 1875 and is long out of copyright. The digitisation is published by UK Parliament, whose Historic Hansard index states that material on the site remains under Parliamentary Copyright; three lines are quoted here for criticism and review, with attribution, and where Parliament's own Open Parliament Licence v3.0 applies, this lesson contains Parliamentary information licensed under it.
2. Donna R. Recht and Lauren Leslie, "Effect of Prior Knowledge on Good and Poor Readers' Memory of Text", Journal of Educational Psychology 80(1), 1988, doi:10.1037/0022-0663.80.1.16. Sixty-four junior-high students crossed on reading ability and baseball knowledge; a half-inning narrative; reconstruction with wooden pieces on a model field; poor readers with baseball knowledge matching good readers without it. Read at abstract and secondary level, and the study used more than one measure of what the readers had taken in, so this lesson describes the reconstruction as one of them rather than as the measure. See research/SOURCES.md.
3. Dan Reynolds, Courtney Hattan and Marissa Markham, "Fair or Foul? Interrogating the Role of Baseball Knowledge in Studies of Knowledge and Comprehension", Reading Research Quarterly 60 (2025), e575, doi:10.1002/rrq.575. Read at abstract level, and every phrase quoted above is from that abstract. A systematic review of nineteen baseball studies from 1978 to 2018, thirteen of which share two knowledge measures; the finding that those measures "focused heavily on vocabulary and baseball trivia"; the separate finding that "the most common baseball comprehension text was deceptively complex"; the opening concession that links between knowledge and comprehension "have been widely documented for decades"; and the closing call for "reliance on non-baseball studies".
4. Mortimer J. Adler and Charles Van Doren, How to Read a Book, revised edition, 1972, on superficial reading. See lesson 2 and research/SOURCES.md.
5. Marc Brysbaert, "How many words do we read per minute? A review and meta-analysis of reading rate", Journal of Memory and Language 109 (2019), 104047, full text. Cited here only for the note that reading rates in a second language are lower. That is the extent of what this course's sources say about second-language reading, and research/SOURCES.md records the gap as a standing weakness rather than an oversight.