Deep Dive: Perpendicularity Proofs and Coordinate Geometry

#CODEX Deep-Dive: Perpendicularity Proofs And Coordinate Geometry (WMA11, spec 2.1, 2.2)

Provenance (2026-09-29). Built from a direct read of research/veridian/WMA11-verified-facts.md (grepped for "perpendicular", "parallel", "gradient", "2.1", "2.2" — §1 spec table of contents, §4.2 Coordinate geometry, §6–§7 topic/flag notes, and the full "Lesson-audit log" entry for this exact lesson, including its 2026-09-14 mark-scheme-bullet coverage audit); the full text of the existing lesson content/veridian/curriculum/wma11/perpendicularity-proofs-and-coordinate-geometry.ts; content/veridian/technique-hub/wma11/signal-topic-brief-perpendicularity-proofs-and-coordinate-geometry.md; content/veridian/technique-hub/wma11/codex-formula-and-method-bank.md §1.2 and §2.2 (checked for overlap before writing a line of this document); content/veridian/question-bank/wma11.ts (grepped for "perpendicular"/"gradient"/"rectangle" — no matches, no item exists for this spec point yet); and the verbatim spec text for 2.1/2.2 already transcribed in the facts bank. No new primary-source PDF was opened for this pass — the real Jan 2023 Q2 coordinates used below are the ones the facts bank already re-extracted and verified (§4.2's "UPDATE" entry, cross-checked against the lesson-audit log). Not affiliated with or endorsed by Pearson Edexcel.

#0. What this document adds, and what it deliberately does not repeat

The existing lesson already derives the perpendicular-gradient condition from a 90° rotation, and already works the confirmed Jan 2023 Q2 proof-rigour finding, the confirmed length-squaring trap, and the confirmed "fourth vertex" method-reliability finding — all on its own VERIDIAN-original point set, by exactly two of the mark scheme's three credited routes (gradients; lengths via Pythagoras). The signal brief already gives the frequency count (2 of 5 series), the mark tariffs, and the prove/construct split. The formula bank already gives the one-line method recipes. None of that is re-derived here.

This document is additive in five specific ways none of the above already covers:

1. A second, algebraic derivation of $m_1 m_2 = -1$ (direction-vector dot product), complementary to the lesson's geometric rotation argument, and the edge case neither existing document states in worked form: what happens when one line is vertical, where the gradient formula itself breaks down.
2. The mark scheme's third and fourth credited routes to a perpendicularity proof, actually worked — the vector-dot-product route and the cosine-rule route are both named as fully credited in the real Jan 2023 Q2 mark scheme (verified-facts.md §4.2's Lesson-audit log, "Way 2... Pythagoras or cosine rule" and "Way 3... vectors"), but neither is worked anywhere in this course's existing WMA11 content — the lesson's own method-comparison block covers only gradients and Pythagoras.
3. Those two missing routes worked on the real question, not a VERIDIAN-original substitute — using the actual verified Jan 2023 Q2 coordinates, $P(-3,7)$, $Q(9,11)$, $R(12,2)$, re-extracted and confirmed in the facts bank. This does not touch, or attempt to resolve, the flagged cross-file inconsistency between the lesson and simultaneous-equations-by-substitution.ts (both of which use their own different placeholder numbers, per §4.2's explicit note that a coordinated fix across both files is needed, not a unilateral one) — this is a separate, new document, not an edit to either flagged file.
4. One real mark-scheme condition not yet taught anywhere in this course's WMA11 material: the lengths/Pythagoras route requires exact values, and a decimal-only final comparison is confirmed, verbatim, to score one mark fewer than the full route.
5. The perpendicular-bisector-plus-fixed-distance construction (Oct 2020 Q6's mechanism), generalised to the case where the given segment is not horizontal or vertical — both the lesson's worked chain and its chain-drill use an axis-aligned segment specifically so the perpendicular bisector's direction is trivial (purely vertical or purely horizontal); the general oblique case needs one further tool (a unit direction vector) that neither the lesson nor the formula bank state.

Everything below is either a direct, independently checkable algebraic derivation, or a claim traceable to WMA11-verified-facts.md. Where the real material runs out, each section says so.

#1. Spec scope, exact text

Spec 2.1, as transcribed and verified in the facts bank §1: "Equation of a straight line, including the forms $y - y_1 = m(x - x_1)$ and $ax + by + c = 0$ — to include (i) the equation of a line through two given points, (ii) the equation of a line parallel or perpendicular to a given line through a given point." Spec 2.2: "Conditions for two straight lines to be parallel or perpendicular to each other." Both are flagged [ASSUMED] in the formula bank — universal geometric facts, not given in the P1 formula booklet and not on the spec's own short "must be memorised" list (booklet contents confirmed directly against the primary PDF: sphere/cone mensuration and the cosine rule are the P1 booklet's only two entries — the cosine rule's presence there matters directly for §3.2 below).

#2. The perpendicularity condition itself: two equivalent mechanisms, and where one of them breaks

#2.1 The gradient condition — cross-referenced, not re-derived

The lesson already proves $m_1 m_2 = -1$ by rotating a gradient triangle $90°$ about the origin and reading off the rotated point's coordinates. That derivation is correct and is not repeated here.

#2.2 A second derivation: direction vectors and the dot product

Every non-vertical line with gradient $m$ has a direction vector $\begin{pmatrix}1 \\ m\end{pmatrix}$ — moving one unit across moves $m$ units up. Two vectors are perpendicular exactly when their dot product is zero (the dot-product-zero test is itself what "Way 3" of the real Jan 2023 Q2 mark scheme uses directly, worked on the real question in §3.3 below). Apply that test to the two direction vectors:

$$\begin{pmatrix}1 \\ m_1\end{pmatrix} \cdot \begin{pmatrix}1 \\ m_2\end{pmatrix} = 1 + m_1 m_2 = 0 \iff m_1 m_2 = -1$$

This is the same fact as §2.1, reached by a genuinely different route — one geometric (a rotation), one algebraic (a dot product) — and the agreement between them is why the gradient test (Way 1) and the vector test (Way 3) in the real mark scheme are interchangeable: they are not two different facts that happen to give the same answer on this question, they are the same single fact, stated in two equivalent languages.

One detail worth stating precisely, because it is easy to get backwards under exam pressure: the sign of the direction vector doesn't matter for this test. Reversing either vector (using $\overrightarrow{QP}$ instead of $\overrightarrow{PQ}$, say) negates the dot product, and a dot product of exactly zero stays exactly zero under negation ($-0 = 0$). This is why the real mark scheme can write $\overrightarrow{PQ} \cdot \overrightarrow{QR} = 0$ and conclude "angle $PQR = 90°$" without worrying about which of the four possible vector directions along those two lines was chosen — perpendicularity is a property of the two lines, not of which way along them a vector happens to point.

#2.3 The edge case the gradient formula cannot express: a vertical line

$m_1 m_2 = -1$ requires dividing by $m_1$ to reach $m_2 = -1/m_1$. If $m_1 = 0$ (a horizontal line), this gives $m_2$ undefined — correctly, because the line perpendicular to a horizontal line is vertical, and a vertical line's gradient is not a number at all (it has no well-defined "rise over run," since the run is always zero). The formula doesn't fail here; it correctly refuses to produce an answer, because the true answer isn't a gradient.

The vector test in §2.2 handles this case without any special pleading. A vertical line has direction vector $\begin{pmatrix}0 \\ 1\end{pmatrix}$; a horizontal line has direction vector $\begin{pmatrix}1 \\ 0\end{pmatrix}$. Their dot product is $0 \times 1 + 1 \times 0 = 0$ — perpendicular, confirmed, with no division by zero anywhere in the calculation. The practical rule this gives: the moment either line in a perpendicularity question is vertical (an equation of the form $x = k$, with no $y$-term), stop reaching for $m_1 m_2 = -1$ — state directly that the perpendicular line is horizontal ($y = k$ for some constant), or use direction vectors if a proof is required. Neither the lesson nor the formula bank states this edge case in worked form; it is included here because it is a genuine gap in what "the perpendicularity condition" means once a line's gradient stops being a number.

#3. The mark scheme's four routes, worked on the real question

The real Jan 2023 Q2 mark scheme (re-extracted and verified in the facts bank's Lesson-audit log) credits three interchangeable routes to the 3-mark "prove $PQR = 90°$" part, and a length route that itself splits into two distinct tools (Pythagoras' converse, or the cosine rule) — four calculations in total, all scoring the identical M1 A1 A1. The lesson works two of them (gradients; Pythagoras) on its own point set. Below are all four, worked on the real points: $P(-3, 7)$, $Q(9, 11)$, $R(12, 2)$.

#3.1 Route 1: gradients (confirming the lesson's method, on the real points)

$$m_{PQ} = \frac{11 - 7}{9 - (-3)} = \frac{4}{12} = \frac{1}{3}, \qquad m_{QR} = \frac{2 - 11}{12 - 9} = \frac{-9}{3} = -3$$

$\frac{1}{3} \times (-3) = -1$, so $PQ \perp QR$ — angle $PQR = 90°$.

#3.2 Route 2a: lengths, via the converse of Pythagoras' theorem

$$PQ^2 = 12^2 + 4^2 = 160, \qquad QR^2 = 3^2 + (-9)^2 = 90, \qquad PR^2 = 15^2 + (-5)^2 = 250$$

$PQ^2 + QR^2 = 160 + 90 = 250 = PR^2$, so by the converse of Pythagoras' theorem, angle $PQR = 90°$.

One further observation, not needed for the proof but worth naming because it is presumably why the setter chose these numbers: $160 : 90 : 250$ reduces to $16 : 9 : 25$, i.e. $PQ : QR : PR = 4 : 3 : 5$ (scaled by $\sqrt{10}$) — the real Jan 2023 Q2 triangle is a scaled 3-4-5 right triangle.

#3.3 Route 2b: lengths, via the cosine rule — the credited alternative neither existing document works

The real mark scheme explicitly credits the cosine rule as an alternative within the same "Way 2" (verified-facts.md §4.2 Lesson-audit log: "A1: via Pythagoras or cosine rule e.g. $PQ^2 + QR^2 = PR^2$ so $PQR = 90$"). Unlike the converse-of-Pythagoras check — which only ever answers "is this angle exactly $90°$?" — the cosine rule computes the angle directly, and would work equally well on a triangle that turned out not to be right-angled:

$$\cos(PQR) = \frac{PQ^2 + QR^2 - PR^2}{2 \times PQ \times QR} = \frac{160 + 90 - 250}{2\sqrt{160}\sqrt{90}} = \frac{0}{2\sqrt{14400}} = \frac{0}{240} = 0 \implies PQR = 90°$$

The numerator is exactly the Pythagoras identity from §3.2, restated as a subtraction rather than an equality — the cosine rule and the converse of Pythagoras are not two independent facts here, they are the same right-angle test wrapped in a formula that also happens to be listed in the P1 booklet (§1 above), which is worth knowing if a "prove or find the angle" question ever gives a triangle where the angle in question is not $90°$ and Pythagoras' converse genuinely cannot be used.

#3.4 Route 3: vectors — worked in full, on the real question, for the first time in this course's WMA11 content

$$\overrightarrow{PQ} = \binom{9-(-3)}{11-7} = \binom{12}{4}, \qquad \overrightarrow{QR} = \binom{12-9}{2-11} = \binom{3}{-9}$$ $$\overrightarrow{PQ} \cdot \overrightarrow{QR} = (12)(3) + (4)(-9) = 36 - 36 = 0 \implies \overrightarrow{PQ} \perp \overrightarrow{QR} \implies \text{angle } PQR = 90°$$

Genuinely the fastest of the four routes on this question — two subtractions to build each vector, one multiplication-and-addition to test them, no squaring and no fractions anywhere — which is consistent with the examiner-report finding already documented elsewhere in this course for the different part of this same question (finding the fourth vertex): candidates using vector methods were independently reported as among the most confident on this question type (facts bank §4.2, part (b) finding; the vector route is not separately commented on for part (a) in the report itself, so this is the mark scheme's own credited structure, not a second examiner-report citation).

#3.5 A real, sourced trap none of this course's existing WMA11 content states yet: exact values only

The verified mark scheme for the length route carries a condition that changes which answers actually earn the accuracy mark: "Values must appear as exact for this mark" (verified-facts.md §4.2, Lesson-audit log, Way 2). A candidate who computes $PQ = \sqrt{160} \approx 12.65$, $QR = \sqrt{90} \approx 9.49$, $PR = \sqrt{250} \approx 15.81$, and checks $12.65^2 + 9.49^2 \approx 15.81^2$ using rounded decimals has not earned the same mark as one who keeps $PQ^2 = 160$, $QR^2 = 90$, $PR^2 = 250$ as exact integers throughout — the mark scheme scores a decimal-only version of this comparison as M1 A1 A0, one mark short. This is a direct, practical consequence of a habit this course already teaches for a different reason elsewhere (WMA11's own differentiation content confirms "resorting immediately to a decimal answer" as a real, separately-documented mark loss, per the formula bank §2.3.1) — here it applies specifically to the intermediate squared lengths in a perpendicularity proof, not to a final answer, and it is not yet stated anywhere in this course's coordinate-geometry content.

#4. Part (b) of the same real question, briefly

The real question's part (b) — given $PQRS$ is a rectangle, find $S$ — is already fully worked in the facts bank itself (§4.2's "UPDATE" entry) and in the lesson's own two method-comparison routes (on the lesson's own point set, not the real one). For completeness on the real points: the mark scheme's main route is $S = P + \overrightarrow{QR} = (-3, 7) + (3, -9) = (0, -2)$. The diagonal-midpoint route cross-checks it independently: midpoint of $PR = \left(\frac{-3+12}{2}, \frac{7+2}{2}\right) = (4.5, 4.5)$, which must also be the midpoint of $QS$; solving $\left(\frac{9+x}{2}, \frac{11+y}{2}\right) = (4.5, 4.5)$ gives $x = 0$, $y = -2$ — the same point, confirmed by a second, independent method. Nothing here is new; it is included only so this document's worked treatment of the real question is self-contained rather than half on the real points and half deferred elsewhere.

#5. The "construct" sub-case, generalised beyond the axis-aligned special case

Oct 2020 Q6's mechanism — a point constrained to the perpendicular bisector of a segment $AB$, a fixed distance from $A$, right angle guaranteed at $M$ (the midpoint of $AB$), Pythagoras applied to triangle $AMC$ — is fully worked in the lesson's own worked-chain and chain-drill. Both of the lesson's examples deliberately choose $AB$ horizontal or vertical, so "move along the perpendicular bisector" becomes "change one coordinate by a fixed amount" — genuinely simpler than the general case, and correctly presented as such. Spec 2.1/2.2 place no such restriction on the segment's orientation; here is the same mechanism carried through when $AB$ is neither horizontal nor vertical, a case neither the lesson nor the formula bank works. This is an original worked extension of the verified mechanism, not a transcription of any past paper.

Worked example. $A(1, 1)$ and $B(5, 5)$. $C$ lies on the perpendicular bisector of $AB$ with $AC = 6$. Find both possible positions of $C$.

• Midpoint and half-length. $M = \left(\frac{1+5}{2}, \frac{1+5}{2}\right) = (3, 3)$. $AB = \sqrt{4^2 + 4^2} = \sqrt{32} = 4\sqrt{2}$, so $AM = 2\sqrt{2}$.
• Pythagoras from the right vertex, $M$. $MC^2 = AC^2 - AM^2 = 36 - 8 = 28$, so $MC = 2\sqrt{7}$.
• The step the axis-aligned examples skip: find a unit vector along the perpendicular bisector. Gradient of $AB$ is $\frac{5-1}{5-1} = 1$, so the perpendicular bisector's gradient is $-1$ — direction vector $\binom{1}{-1}$, length $\sqrt{2}$, unit vector $\frac{1}{\sqrt{2}}\binom{1}{-1}$. (When $AB$ is horizontal or vertical, this unit vector is trivially $\binom{1}{0}$ or $\binom{0}{1}$ — exactly why the lesson's own examples never needed this step spelled out.)
• Move from $M$ by $MC$ along that unit vector, both ways. $C = M \pm MC \cdot \frac{1}{\sqrt{2}}\binom{1}{-1} = (3,3) \pm 2\sqrt{7} \cdot \frac{1}{\sqrt{2}}\binom{1}{-1} = (3,3) \pm \sqrt{14}\binom{1}{-1}$, giving $C = (3+\sqrt{14},\, 3-\sqrt{14})$ or $C = (3-\sqrt{14},\, 3+\sqrt{14})$.
• Check. $AC^2 = (2+\sqrt{14})^2 + (2-\sqrt{14})^2 = (4+4\sqrt{14}+14) + (4-4\sqrt{14}+14) = 36$, so $AC = 6$ ✓, and $\overrightarrow{MA} \cdot \overrightarrow{MC} = (-2,-2)\cdot(\sqrt{14},-\sqrt{14}) = -2\sqrt{14}+2\sqrt{14} = 0$ ✓, confirming the right angle at $M$ the whole construction depends on.

The mechanism is identical to the lesson's own worked chain in every respect except one: finding the direction to move in requires a unit vector once the segment is oblique, where an axis-aligned segment lets that step disappear into "up/down" or "left/right." Exact surd-form answers here are consistent with this course's own exact-answer discipline elsewhere (spec 1.2 surd manipulation) — this is exactly the kind of answer a genuinely non-axis-aligned version of this question would produce.

#6. Real exam-angle synthesis

The confirmed record is exactly what the signal brief already states and is not re-derived here: 2 of 5 fully-mined WMA11 series contain a confirmed perpendicularity/coordinate-geometry question (Jan 2023 Q2, Oct 2020 Q6), against 5 marks each time, split as a 3+2 "prove, then construct" pair in one series and embedded inside a longer Pythagoras-based question in the other. The two genuinely different angles this actually gets tested from, on the confirmed evidence: (i) a proof, where the arithmetic is short and the mark lost is the concluding sentence, not a calculation; and (ii) a construction, where the arithmetic is short and the mark lost is misidentifying which point in a diagram is actually guaranteed a right angle. Neither confirmed appearance makes this a paper-dominating topic — it is, by the facts bank's own honest count, the thinnest-tested strand in the whole WMA11 corpus (spec strand 2 has only two numbered sub-points, and no other WMA11 strand scores worse than 4 of 5 series in the same tally).

One further, honest observation this document adds: spec 2.2 names both conditions — parallel and perpendicular — but every confirmed real question found in this research pass tests perpendicularity specifically; the facts bank contains no confirmed "prove two lines are parallel" question (grepped directly for "parallel" across the full facts bank — the only other hits are the spec's own text and an unrelated Jan 2019 region-inequality question using "parallel to the y-axis" as a plain geometric description, not a proof target). By direct mechanical analogy to the confirmed perpendicular mark scheme, a "prove $AB$ is parallel to $CD$" question would presumably be scored M1 (gradient method) A1 (both gradients correct) A1 (dependent concluding statement, $m_1 = m_2$ so parallel) — but this is stated here only as an inferred structural parallel, not a verified fact, and should not be taught as if a real mark scheme confirms it.

#7. Honest limits

This document's theory sections (§2, §5) are original, independently checkable mathematics, correctly scoped to spec 2.1/2.2, but they are not exam-frequency claims — nothing in §2 or §5 asserts that a vertical-line perpendicularity question or an oblique perpendicular-bisector construction has ever appeared on a real WMA11 paper, because the facts bank's research pass did not establish that either way. §3–§4's worked routes are grounded directly in the one real, fully-recovered question this topic has (Jan 2023 Q2); no second real question with full coordinates exists in the facts bank to cross-check the exact-values trap (§3.5) or the cosine-rule route (§3.3) against a second series. §6's "2 of 5" count, and everything about which series were and weren't fully mined, is inherited from the facts bank and signal brief, not independently re-verified in this pass.

#8. Sources

• research/veridian/WMA11-verified-facts.md — §1 (spec 2.1/2.2 text, verified against Specification Issue 3); §4.2 (Jan 2023 Q2's real coordinates, Oct 2020 Q6); the "Lesson-audit log" entry for perpendicularity-proofs-and-coordinate-geometry.ts (the full, verbatim, re-extracted mark scheme for Jan 2023 Q2 parts (a) and (b), including the "exact values" condition used in §3.5).
• content/veridian/curriculum/wma11/perpendicularity-proofs-and-coordinate-geometry.ts — the existing lesson (90° rotation derivation, gradient and Pythagoras routes, the axis-aligned perpendicular-bisector worked chain and chain-drill, the full trap taxonomy) — read in full and not duplicated.
• content/veridian/technique-hub/wma11/signal-topic-brief-perpendicularity-proofs-and-coordinate-geometry.md — frequency count, mark tariffs, prove/construct framing — cross-referenced, not restated.
• content/veridian/technique-hub/wma11/codex-formula-and-method-bank.md §0.3, §1.2.1–1.2.2, §2.2.1–2.2.2 — booklet-status flags, one-line method recipes, checked for overlap.
• content/veridian/question-bank/wma11.ts — grepped; no item tagged to spec 2.1/2.2.
• §2.2, §2.3, §3.2–3.4, §5 are original derivation/computation, independently checkable by hand, not sourced to any external document.