An advanced lesson on overgeneralisation, built on one sentence from your teacher’s own published work — a sentence that was accurate when written and became false on a datable afternoon in September 2026.
| Lesson Basics | |
|---|---|
| Topic | Academic writing · Hedging and scope · Falsifiability · Planetary science as source material |
| Objective | Identify where the scope of a claim lives in an English sentence, recognise when it has been deleted, and repair the sentence so that it can be disproved. |
| Level | Advanced to Proficient (C1–D2). Suitable for researchers writing in English as a second language. |
| Duration | 120–145 min · eight stages · splits cleanly into two sessions after Stage 4 |
| Skills | Reading · Analytical writing · Speaking · Self-editing · Critical thinking |
| Grammar | Identifying vs describing copula · Restrictive relative clauses · Quantifiers · Hedges and boosters · Generic reference |
| Materials | This page (Stage 1 figure projected or printed) · Whiteboard · 150 words of the student’s own draft writing · The two Saturn images |
| Sources | Principia Orthogona Ch. I One Shape, Many Infinities · Vol IV preface Lemniscata & Analema · Book 3 §7.2 · WP-97, WP-100 · Sánchez-Lavega et al., Science Advances 12, eaee4251 (2026) |
Show the figure. Say nothing about which is which.
A and B, drawn to the same scale on the same axes. Both are figure-eights. Both stand upright. One of them is made of dots — forty-one of them, the exposure count of the photograph on the screen.
This is the point, and it is not a metaphor. An analemma is not a curve anyone drew. It is forty-one separate photographs, each taken on a different day, each with its own exposure and its own weather, composited into one frame. The dots are the data. The line your eye draws between them is interpolation, and the interpolation is yours — it is not in the sky. B has uncountably many points and every single one of them satisfies the equation.
Ask (5 min, pairs): one of these is a mathematical object and one is a photograph. Which is which, and how can you tell by looking?
A is the solar analemma — where the Sun stands above Newark, New Jersey, photographed at 12:00 EST on each day of one year. It is an observation, and it is the superposition of two independent effects, orbital eccentricity and axial tilt, at a frequency ratio of 2∶1. Its shape depends on your latitude, on the clock time you pick, on the tilt, and on the fact that the orbit is not a circle. Stand in Quito and it changes. Photograph at 08:00 and it changes. On Mars it is not a figure-eight at all.
B is the Lemniscate of Bernoulli, the quartic (x²+y²)² = 2(x²−y²). It is a definition. It has no latitude, no clock and no planet. It is the same curve in Newark, on Mars, and in a universe with no Sun in it.
Now look again. A’s two lobes are unequal — the summer lobe tall and narrow, the winter lobe short and wide — and the whole figure leans. That asymmetry is the eccentricity of Earth’s orbit, made visible. B’s lobes are identical and it is symmetric about both axes. They are not the same curve. They were never the same curve.
And the student is wrong twice. There is no such thing as the lemniscate. The Lemniscate of Gerono is x⁴ = x² − y², which fits inside the unit square; Bernoulli’s is wider by a factor of √2 and carries an arc-length structure the other one has not got. Two mathematical objects, also figure-eights, also distinct — plus the lunar analemma, a narrower cousin of A traced by the Moon each month. Four curves, four different equations, four different symmetries, four different bounding boxes.
Read the source. This is Chapter I of the series, One Shape, Many Infinities, where the four are separated formally and the separation is machine-checked in Lean 4. The chapter also finds the one thing they genuinely share — a single geometric invariant at the crossing point, which singularity theory classifies exactly. That is the honest version of “they are the same”: not the curves, the node.
The bridge, and write it on the board: six sides at Saturn’s north pole is an analemma, not a lemniscate. It looked like a constant of nature. It was an observation with parameters attached — and when the parameters changed, at the other pole of the same planet, the figure had ten sides.
And the sharper version, once the room has seen the dots: “Six is what stability looks like” draws a smooth curve through a single observation. Saturn had given us one dot. We drew a lemniscate through it.
Teacher writes both on the board and says nothing about either.
Question: which of these could a single photograph prove wrong?
Students vote, then justify in pairs (5 min). Most rooms choose B, and describe A as the “stronger” or “more scientific” sentence. Do not correct anyone yet. Write the vote count on the board and leave it there for the rest of the lesson.
Read the source paragraph aloud. It is from your teacher’s Principia Orthogona, Book 3, §7.2, a section titled The Sixth Mode — Saturn’s Proof:
Task (10 min, written, individually). Complete this sentence in one line: “This paragraph would be false if …”
Collect five or six answers on the board without judging them. Typical answers: if a crystal had five faces; if the hexagon disappeared; if scientists changed the definition of wavenumber. Note which answers describe something a telescope could actually see.
On 2 September 2026, Hubble and ground-based images established a ten-sided wave around Saturn’s south pole, at 63°S — the first regular polygonal jet ever seen in that hemisphere. It has ten vertices. It sits on an eastward jet of 116 metres per second. It was reported in Science Advances.
The same planet. The same rotating fluid. The same constraint. Ten.
Discussion (10 min, whole class). The paragraph in Stage 2 did not contain a mistake when it was written. Every word in it was accurate. So what happened? Push the room past “new data” — the useful answer is that the sentence had claimed something the evidence never supported, and nobody noticed until the world tested it.
Two English sentences that look alike and are not.
| Sentence | Structure | What it claims |
|---|---|---|
| Six is what stability looks like. | identifying (equative) copula — X is Y, the two are the same thing | everything. There is no condition attached, so it is a claim about rotating fluids in general |
| Six is the number of wavelengths that fit around a ring of 83,316 km. | describing copula + restrictive relative clause | one ring, one number. The relative clause carries the entire scope of the claim |
The rule, and it is the lesson:
“Six is the number of wavelengths” still sounds finished. That is exactly the danger. English lets you drop a restrictive clause without leaving a grammatical hole, and writers drop them for rhythm, for confidence, for a stronger closing line. The parameter — here, the size of the ring — disappears, and the number is left standing on its own as though it were a constant of nature.
Practice (15 min, pairs). Each sentence below has a deleted scope. Restore it. You may invent the missing parameter; the point is to notice the hole.
Sample repair for 2: The treatment is effective in patients under sixty who completed the full twelve-week course. Now a trial can disagree with it.
Return to the second sentence in the source paragraph:
The phrase on record looks like a hedge, and students almost always read it as one. It is not doing the work they think.
The rule: a hedge placed on your evidence does not weaken your claim. To weaken a claim you must change the quantifier, or attach a date.
Repair: No other polygonal jet had been reported as of 2026. That sentence can only be falsified by a dated observation — which is precisely what arrived, and the sentence would have survived it, because it says when it was true.
Sorting task (10 min, pairs). Which of these genuinely weaken a claim, and which only move the burden onto the evidence?
to date · as far as we know · generally · in the cases examined · tends to · always · in our sample · arguably
And the opposite move. “Six is not decoration” is a booster: it denies a weaker reading in order to strengthen the claim. Boosters are not wrong in themselves. Attached to a sentence whose scope has already been deleted, they are how confident prose becomes false prose.
A second real pair, from the same corpus, five days apart. First the published version:
And the correction, published on 5 September 2026:
Compare them. The first names nothing anyone could photograph — a trivial field is not a thing a telescope resolves. The second names an object: a thirty-sided figure, which nobody has ever seen on Saturn. The corrected sentence is easier to disprove than the one it replaced.
Your turn (15 min). Take the 150 words of your own writing that you brought. Find one sentence with a deleted scope. Restore the parameter. Read the before and the after aloud to your partner, and have them tell you what would now count as evidence against you.
Point at the Stage 1 vote count, still on the board. Sentence B — the one the room called weaker — is the one that is still true today. Sentence A is the one that had to be withdrawn.
Into the notebook, in the student’s own hand:
| Vocabulary — C1 to D2 | ||
|---|---|---|
| Word | Part of speech | Meaning in this lesson |
| scope | noun | the range of cases a claim actually covers |
| hedge | noun / verb | a word that reduces the strength of a statement (may, tends to, in our sample) |
| booster | noun | a word that increases it (clearly, certainly, not merely) |
| quantifier | noun | a word setting how many cases are covered (all, no, some, most) |
| restrictive relative clause | noun phrase | a that- or which-clause that narrows the noun before it, and cannot be removed without changing the meaning |
| generic reference | noun phrase | using a noun to mean the whole class (the hexagon = all hexagons) |
| overgeneralisation | noun | claiming for every case what was shown for some |
| refutation condition | noun phrase | the state of the world that would make your claim false |
| load-bearing | adjective | (of a claim) holding up other claims, so its failure spreads |
| to supersede | verb | to replace something that is now out of date |
| to withdraw | verb | to formally take back a published claim |
| qualifier | noun | any phrase limiting a statement’s application |
| Timing Summary | ||
|---|---|---|
| Stage 1 | Two curves — the analemma and the lemniscate | 15 min |
| Stage 2 | Warm-up — two sentences, one vote | 15 min |
| Stage 3 | The claim — write the refutation condition | 20 min |
| Stage 4 | The date — the photograph arrives | 15 min |
| Stage 5 | Where the scope lives — the deletable clause | 30 min |
| Stage 6 | The hedge that hedges nothing | 20 min |
| Stage 7 | A good repair is easier to disprove — own writing | 25 min |
| Stage 8 | Close — return to the vote | 5 min |
| Total | 145 min — run over two sessions, or trim Stage 6’s sorting task and Stage 4 to fit 120 | |