Judging what you have understood

Lessons 5 and 6 leave you holding a good deal. An outline of the chapter. The author's terms, in his sense. The argument, set out in lines. A margin full of marks, including the places where you wrote an X.

This lesson is what those X marks are for, and the first thing it will tell you is that they aren't disagreements yet.

Rule 9, and why the order isn't politeness

Adler's third stage opens with a rule that reads like manners and isn't.

Don't say you agree, disagree, or suspend judgment, until you can say "I understand."¹

Take that literally, because it is meant literally. Until you can state the author's argument in your own words, you have no verdict available to you. Not disagreement, not agreement, and not the third option either. Suspending judgement counts as a verdict here too, which is worth saying because readers treat it as the absence of one.

What goes wrong when the order is reversed is not that you are rude. It is that you end up arguing with something nobody wrote. A reader who forms a verdict at page thirty has built it out of the first thirty pages plus their own expectations about what the rest will say, and when the book does something else the verdict often does not update, because it was never attached to the book in the first place.

This is also why the effort of lesson 5 is worth the trouble. An outline is the evidence that you have the thing you are about to judge.

Rules 10 and 11

Two more before the grounds themselves.

Rule 10: don't disagree disputatiously.¹ Disagreement is not a contest to be won, and a reader who is reading in order to disagree will find something regardless of what the page says.

Logic and Argument, later in this term, puts a practical test on the same instinct: you should be able to state the author's position in a way the author would accept before you say what is wrong with it. This course leans on that rather than re-teaching it.

Rule 11: present good reasons for any critical judgement.¹ This is the one that separates knowledge from personal opinion, and Adler means the distinction seriously: an author's propositions, he says, "are nothing but expressions of personal opinion unless they are supported by reasons", and he asks the same of the reader, who must "distinguish between genuine knowledge and mere opinion".¹

So "I disagree" is a report about you. "I disagree because the second premise is asserted and never argued for" is a claim about the book, and a claim about the book is a thing somebody else can check, agree with, or refute.

The test is simple and it is worth applying to your own verdicts. Can somebody else disagree with your objection? If not, you have described a feeling.

The four grounds, and nothing else

Here is the part of the method that does the work. When you find an author wrong, you are making one of exactly four claims.¹ Three of them are grounds for saying so. The fourth is not a basis for disagreement at all, and "The wrinkle" below says what it is instead.

Plenty of what a reader wants to say about a book is none of the four. Of the four objections worked below, two are on no ground at all, one lands on a ground, and one is two claims of which only the first can be sorted. That's not the method failing. It is the method telling you what kind of thing you are holding.

The author is uninformed. In Adler's words, "he lacks some piece of knowledge that is relevant to the problem he is trying to solve", and to make the objection stick "you must be able yourself to state the knowledge that the author lacks and show how it is relevant, how it makes a difference to his conclusions".¹ So you owe two things, not one. Notice also that this is not a criticism of anyone's character, and it is usually nobody's fault.

Adler's own example here is the book you worked in lesson 5. Darwin "lacked the knowledge of genetics that the work of Mendel and later experimentalists now provides", and Adler calls that ignorance of the mechanism of inheritance "one of the major defects in The Origin of Species".¹ Notice what that objection is and is not. It is a claim about what Darwin had to work with. It's not a verdict on his conclusion, and a reader who turns a missing mechanism into a verdict on the book has left the ground he named.

The author is misinformed. He "asserts what is not the case".¹ The difference from the first ground is that the first is a silence and this is a positive claim, and Adler is explicit that the cause does not matter: the error "may be owing to lack of knowledge, but the error is more than that. Whatever its cause, it consists in making assertions contrary to fact."¹ The two grounds ask different evidence of you. For the first you produce the knowledge he lacked. For the second you produce the fact his sentence gets wrong.

The author is illogical. He "has committed a fallacy in reasoning", which Adler splits into the conclusion that does not follow and the two statements that cannot both be true, and either way the reader "must be able to show the precise respect in which" it goes wrong.¹ This is the ground lesson 5 prepared you for: an argument set out in three lines shows you the gap that a paragraph hides.

The author's account is incomplete. He has not solved all the problems he set out to solve, or hasn't seen all the ones his subject raises. This ground is different from the other three, and "The wrinkle" below says why.

An objection that cannot be placed on one of those four is one of three things, and the worked examples below give you all three. It may be a point about the author's support rather than about his conclusion, which is rule 11's territory rather than the four grounds'. It may be a judgement of value, which no ground reaches, and the fourth objection below is half of one. Or it is a report of taste, or of a prior belief you brought with you. Naming the ground, or failing to and saying which of those three you have instead, is what makes the difference visible.

Why does naming it help? Because each ground tells you what you would have to produce to be right. If you say the author is misinformed, you have to name the fact. If you say he is illogical, you have to point at the step. The demand for evidence is built into the category, which is why the sorting does more work than it looks like it should.

Three objections to one paragraph

Take the Ruskin paragraph you marked in lesson 6, on looking intensely at words. Here are three objections a reader might make to it, sorted.

Objection 1: "He offers his own certainty in place of a reason." You marked this one in lesson 6, and here is what marking it was for. The parenthesis reads "(I know I am right in this)", and it is doing work in the sentence that no argument in the paragraph does.²

Now sort it honestly, and watch it refuse to go cleanly into a box. The paragraph is not empty of reasons: the British Museum contrast is a reason offered, and it's the only one before the point where lesson 6 stopped. Whether it is a good one is a further question, and lesson 6 has already shown you the catch, which is that it works only in Ruskin's own sense of "illiterate", the one he signals himself with the quotation marks. What the parenthesis adds is confidence, and confidence is not a premise.

So this is not the third ground, because Adler's illogical means a fallacy committed, and no step has gone wrong here. It is not any of the four, and that's not the taxonomy failing. The four grounds are grounds for saying an author is wrong, and this objection does not say he is wrong. It says one sentence of his is unsupported, which is rule 11 turned back on the author who wrote it.

What survives is a narrow objection, and the narrow form is the useful one: the sentence would be exactly as strong without the parenthesis, and a reader who was persuaded by the parenthesis was persuaded by nothing.

Notice what the objection does not claim. It does not claim Ruskin is wrong. It claims that one part of this paragraph does no work, and those are different verdicts with different consequences.

Objection 2: "Accuracy with words is not enough on its own." Ruskin writes that if you read ten pages of a good book "letter by letter,—that is to say, with real accuracy,—you are for evermore in some measure an educated person".²

Deal with his hedge first, because lesson 6 told you it was working. "In some measure" makes the claim small, and a small claim is hard to be wrong about. So the objection is not that accuracy gives a reader nothing. It's that Ruskin offers accuracy as the thing that makes the difference, and for a great many readers it is not the thing.

And notice that this is not the second ground. The hedge is what keeps it off: a claim that ten accurate pages make you an educated person "in some measure" is not one you can produce a contrary fact against, so misinformed has nothing to bite on. What there is, is a piece of knowledge he did not have.

That's uninformed, and this course can name what is missing, because lesson 3 was about it. One of the three things that makes a passage hard for a particular reader is what that reader already knows about its subject, and a reader can be perfectly accurate with every word on the page and still not know what the page claims, because what the page assumes is not on it. What Ruskin would have needed is the finding that the obstacle is often what the reader brings to the page rather than the care they take over it, and lecturing in 1864 he did not have it. That is the ordinary case on this ground, and it is why the ground isn't an accusation.

Which is the shape of most good objections on this ground. You are not catching the author out. You are saying what he would have needed, which is what Adler's rule asks of you.

Objection 3: "Ruskin is insufferable about this." Plenty of readers put that page down thinking it, and it is worth taking seriously rather than being embarrassed about.

And it is not on any of the four grounds, which is why it is worth including. Sort it and it dissolves: it's not that he lacks knowledge, nor that he asserts a falsehood, nor that his conclusion fails to follow, nor that his account is incomplete. It is that his manner irritated you, which is a fact about the reading and not about the argument.

This is the most useful thing the sorting does, and it is why the second exercise below is worth its twenty minutes. My own experience of doing it is that a good deal of what a reader wants to say after a difficult page is objection 3 wearing the clothes of objection 1.

A fourth objection, and what the method cannot do

Lesson 6 stopped quoting Ruskin's paragraph partway and told you what was in the rest. Here is the sentence that matters most for this lesson, and this time you need the whole of it. Ruskin has just said that accent, or turn of expression, "will at once mark a scholar". Then:

And this is so strongly felt, so conclusively admitted, by educated persons, that a false accent or a mistaken syllable is enough, in the parliament of any civilized nation, to assign to a man a certain degree of inferior standing for ever.²

His next words open the paragraph after it: "And this is right; but it is a pity that the accuracy insisted on is not greater, and required to a serious purpose."²

Sort that, and notice first that it's two claims and not one, which is why it had to be quoted whole.

The first is a claim about the world: that in the parliament of any civilized nation a mistaken syllable does in fact fix a man's standing for ever. A ground can get hold of that. Show that it is not so, or not so for ever, or not so in any parliament, and you have found Ruskin misinformed, and what you owe is the fact. This course does not have that fact and will not pretend to. What it can tell you is that the ground is available and the sentence is checkable.

The second claim is "And this is right", and no ground reaches the judgement itself. It is not uninformed, because no missing fact would change it. It's not misinformed, because it asserts nothing about the world at all. It is not incomplete, because it is not a problem left unsolved. And it is not illogical on its face, though a reader who thinks Ruskin has moved from what educated persons feel to what is right has found a step, and should say so, because that would be the third ground.

What is left, once those are used, is a judgement about whether a practice is right, and Adler's four grounds are not built to reach judgements of that kind.

That's a limit of the method, and it is worth knowing rather than working around. A reader who thinks Ruskin is right about this and a reader who finds the sentence repugnant have both made a judgement of value, and neither has found a ground. No amount of sorting will turn either into one.

What the method does give you is the discipline to say which you have. "I think that is wrong about how people should be treated" is a different sentence from "I think that is wrong about how reading works", and running them together is how a reader ends up rejecting a page about words on the strength of the page after it.

This course does not adjudicate Ruskin's view, and it is not equipped to. It notices that the tool it has just handed you reaches the first half of that sentence and not the second, and that the courses built for the second half are Ethics and Political Philosophy, later on the Path.

The wrinkle, and the part nobody quotes

Here is the rule that makes this a method rather than a set of manners, and readers do not like it.

Adler puts it about as hard as it can be put. "You have no freedom of will about this. It is not your sacred privilege to decide whether you are going to agree or disagree. If you have not been able to show that the author is uninformed, misinformed, or illogical on relevant matters, you simply cannot disagree. You must agree."¹ And he names the sentence he is forbidding, which you have said and I have said: "I find nothing wrong with your premises, and no errors in reasoning, but I don't agree with your conclusions."¹

Read that again, because it takes something away from you. The ordinary practice is that a reader is entitled to remain unpersuaded for no stated reason, and this rule says you are not. If you have understood the argument and you cannot name what is wrong with it, the honest verdict is agreement. Adler will let you agree in part, and the next paragraph says on what terms. He will not let you withhold.

Incompleteness is not a way out of that, and Adler is careful about it: the fourth point "is strictly not a basis for disagreement. It is critically adverse only to the extent that it marks the limitations of the author's achievement."¹ What it permits is narrower and stranger. A reader who agrees in part, having found no other ground, "may, nevertheless, suspend judgment on the whole, in the light of this fourth point about the book's incompleteness".¹ So it's a separate finding producing a different verdict: agreement on what was argued, suspended judgement on the whole, and the gap named.

Ruskin cannot be your example here, and it is worth a sentence to say why. You found him uninformed a page ago, and a ground found is a ground used. Adler's suspension is for a reader who agrees in part "because he finds no reason to make any of the other points of adverse criticism",¹ and you have found one. So the verdict on Ruskin stands as disagreement, and that is the discipline working rather than failing.

Notice what separates the two, because it is the whole of what this section teaches. Objection 2 named a piece of knowledge Ruskin lacked and showed what it does to a claim he made, which is a ground for disagreeing with that claim. Incompleteness names nothing the author got wrong. It says a question his subject raises is one his account doesn't answer, and a question left unanswered cannot be answered wrongly. The first is a ground for disagreeing with him. The second is the fourth ground, which is not a ground for disagreeing at all: it marks, in Adler's words, "the limitations of the author's achievement".¹ Only the second earns a suspension, and only where none of the first three has been found.

That gate is narrow on purpose. Suspension is not the verdict you reach for when a book leaves you unsatisfied. It is what is left over when the book has survived everything you could bring against it and still doesn't reach far enough.

That verdict is uncomfortable to write because it concedes and withholds in the same breath. It is also more informative than either "I agree" or "I disagree", and in my experience it is the verdict a good book most often deserves. Here is one.

What people get wrong

"Criticism means finding fault." It means judging, and the judgement can go either way. A reader who has worked through the four grounds and found none of them is doing criticism when they write down that the argument holds, and doing it better than a reader who found a fault because they were looking for one.

"I disagree" is a finished thought. It is the beginning of one. The finished version names a ground and produces what that ground requires.

"I can judge it from the summary." You can judge the summary. Rule 9 is about the book. This is worth saying plainly, because a great deal of what circulates as criticism of a book is criticism of an account of it, and the two are often not the same argument at all.

"Suspending judgement is a cop-out." It is the hardest of the three verdicts to earn, because you have to show both that the argument holds as far as it goes and that it does not go far enough. A reader who suspends judgement without doing that's not suspending judgement. They are declining to read.

"Being unconvinced is a reason." It is a symptom. Sometimes the cause is one of the four grounds and you have not found it yet, in which case keep looking. Sometimes the cause is that you disagreed before you started, in which case that is what to write down.

Practice

Connections

Lesson 5 is the precondition this lesson names. The outline, the terms and the reconstructed argument are what rule 9 means by "I understand", and without them a verdict has nothing to attach to.

Lesson 6's X marks are raw material here rather than conclusions, which is why they are worth making fast and sorting slowly.

Logic and Argument teaches the same discipline from the other direction: how to build an argument that survives this treatment, and what the common failures are called. This lesson names the third ground, illogical, and hands you to that course for the taxonomy under it.

Lesson 10 needs all of this, because reading several authors against each other means judging each of them first, and an unjudged author can't be compared with anything.

Go deeper

• Logic and Argument. Term 1, and the companion to this lesson. Steelmanning, the fallacies, and what a good reason looks like from the writer's side.
• Adler and Van Doren, How to Read a Book, revised edition, 1972. The third stage in full: rules 9 to 15, with the four grounds worked out at length and illustrated. In copyright, so try a library or Open Library.
• John Ruskin, Sesame and Lilies. Free, and the lecture the objections above are aimed at is the first one. Read the rest of it and sort your own objections. There are a good many available.

Sources

1. Mortimer J. Adler and Charles Van Doren, How to Read a Book, revised edition, 1972. The third stage of analytical reading. Every quotation above that carries this note is his; quotation marks elsewhere in the lesson mark words this course has put in a reader's mouth, or a misconception heading, or Ruskin, who carries note 2. Rule 9 and its gloss; rule 10; rule 11 and the two sentences on personal opinion and on distinguishing genuine knowledge from mere opinion; his definitions of the first three grounds, including that a misinformed author "asserts what is not the case" and that the error counts "whatever its cause", which is why this lesson does not make the difference between the first two grounds turn on whether a fact was available; the Darwin illustration of the first ground; the passage forbidding the reader to decline, and the sentence Adler names as the one it forbids; and the incompleteness passage with its suspension clause. The definition of the fourth ground, at the fourth bullet above, is this course's wording rather than a quotation, because what Adler gives there is the ground's status rather than a definition. research/SOURCES.md records all of it verbatim, re-read from the scan on 2026-09-18, and records how that text was consulted and why it is not linked here.
2. John Ruskin, "Of Kings' Treasuries", lecture I of Sesame and Lilies, delivered in Manchester in December 1864 and published in 1865, Project Gutenberg #1293, public domain. Lesson 6 quotes the first half of this paragraph, as far as "consists in this accuracy". This lesson quotes phrases from that half, and in the fourth objection one sentence from past the cut, all checked against the Gutenberg text. Where it reaches past the cut it quotes the sentence whole, including the clause about the parliament and Ruskin's own qualification in the paragraph after, because a lesson about judging an author on what he said cannot do otherwise. Between the two lie the "peerage" of words and words of "modern canaille", which this lesson does not sort and which a reader sorting their own objections should go and read.