Pure Mathematics 4

Deep Dive: Implicit and Parametric Differentiation Tangents and Normals

Provenance: written 2026-09-29. Read/checked this session: research/veridian/WMA14-verified-facts.md §1 (spec text), §3 (formula booklet contents), §4 (mark-scheme conventions), §5.5 (the topic's own citations), §7–§8 (evidence-strength notes) in full; the existing lesson file content/veridian/curriculum/wma14/implicit-and-parametric-differentiation-tangents-and-normals.ts in full (mechanism blocks, worked chains, method-comparison, trap-taxonomy, and its own header/footer provenance notes — including its self-audited gap, reused below); content/veridian/technique-hub/wma14/codex-formula-and-method-bank.md §2.5 (the existing method skeleton), forge-exam-technique-guide.md (§ on cross-topic meta-traps and the per-question mark table), and signal-frequency-and-boundaries.md §1 (frequency row for this topic); content/veridian/question-bank/wma14.ts (checked — contains no item for this spec point yet, so nothing to cross-reference there). No new primary Pearson document was fetched this session; every claim below traces to the facts bank's own already-verified citations, re-synthesized at greater theoretical depth than the existing method-skeleton or lesson-header treatments give it. Not affiliated with or endorsed by Pearson Edexcel.

#WMA14 · CODEX Deep-Dive — Implicit and Parametric Differentiation, Tangents and Normals

#Why this document exists

The 12 curriculum lessons teach this topic through drill (mechanism blocks, a worked chain, a marked solution, MCQs). The CODEX formula bank's §2.5 gives it a five-line method skeleton. Neither is the right place for the thing this document does instead: laying out the complete real model — why implicit and parametric differentiation are two applications of one underlying idea, exactly which mechanical sub-cases the real exam record actually exercises, and how those sub-cases compose into the handful of genuine exam angles this specific spec point (5.1) has actually been tested from. This is additive to, not a restatement of, the lesson file and the three existing technique-hub documents — it does not re-run the lesson's own worked chains (the $xy^2$ compound-step derivation, the parametric tangent-at- $t=2$ chain, the trap-taxonomy items) and cites them by name rather than reproducing them.

Scope boundary, confirmed against the spec, not assumed: spec point 5.1 requires, in the spec's own words (WMA14 spec, §1, quoted by the facts bank), "differentiation of simple functions defined implicitly or parametrically" — and names, as guidance rather than an optional extra, "the finding of equations of tangents and normals" to such curves. Spec point 5.2 (connected rates of change) is a separate numbered item, already the subject of this course's separable-differential-equations.ts. Nothing about a rate-of-change context (a filling tank, a growing shadow, $dV/dt$ ) appears anywhere below — that boundary is the lesson file's own, and this document holds it too.

#1. The theoretical model: one idea, two applications

A curve relating $x$ and $y$ can be given three ways: explicitly ( $y = f(x)$ , $y$ isolated), implicitly ( $F(x,y) = c$ , $x$ and $y$ tangled together with no isolation performed or possible in elementary terms), or parametrically ( $x = x(t)$ , $y = y(t)$ , both expressed through a third variable). WMA14's spec 5.1 is built entirely on one fact that makes the second and third cases tractable without ever solving for $y$ : whatever its closed form turns out to be (if it has one at all), $y$ is a function of $x$ . That single premise is what licenses treating $y^n$ , $\sin y$ , $e^y$ , and so on as composite functions of $x$ and differentiating them with the chain rule exactly as one would differentiate $(g(x))^n$ — substituting $g(x) = y$ changes nothing about the rule itself. The lesson file's own mechanism block derives this substitution explicitly and is the right place for the full derivation; it is restated here only because every sub-case below is a direct consequence of it, not a new rule per term.

Implicit differentiation applies that premise directly: differentiate $F(x,y) = c$ term by term with respect to $x$ , treating every $y$ -containing term as a composite function of $x$ , then rearrange algebraically to isolate $\frac{dy}{dx}$ . No discrete formula exists for this — the CODEX bank tags it correctly as a technique, not a lookup — because the exact differentiation performed depends entirely on the specific term shapes present in $F(x,y)$ , which is precisely why the sub-case taxonomy in §2 below is where the real difficulty (and the real exam record) actually lives.

Parametric differentiation applies the same premise one layer removed: if $x$ and $y$ are both functions of a third variable $t$ , the chain rule gives $\frac{dy}{dt} = \frac{dy}{dx}\cdot\frac{dx}{dt}$ directly from treating $y$ as a function of $x$ which is itself a function of $t$ . Rearranging for the quantity actually wanted,

\frac{dy}{dx} = \frac{dy/dt}{dx/dt},

which is why the correct order of operations is "differentiate $x(t)$ and $y(t)$ separately first, then divide" and never "eliminate $t$ , then differentiate the resulting Cartesian relation" (the Cartesian relation may not even exist in closed elementary form, and even where it does, this course's own parametric-equations-conversion-and-domain-range.ts documents real candidates reaching for calculus on questions where none was wanted — the mirror-image error, reaching for algebra where calculus was actually required, is what Oct 2022 Q1's "A few candidates thought it appropriate to use calculus and scored no marks" describes for a different spec point, 3.1, but the underlying lesson — match the technique to the representation you're actually given — is the same one 5.1 depends on).

Tangents and normals are the payoff both routes are built for: once $\frac{dy}{dx}$ is known at a specific point (or parameter value), the tangent is the line through that point with that gradient, and the normal is the line through that point with gradient $-1/\left(\frac{dy}{dx}\right)$ — the perpendicular-gradient relation $m_1 m_2 = -1$ , derived on this course in wma11-perpendicularity-proofs-and-coordinate-geometry.ts and used here, not re-derived. Spec 5.1's own guidance clause names this pairing explicitly rather than leaving it implicit, which is confirmed directly by the spec quote above — tangent/normal construction is not an optional extension of the differentiation skill on this paper, it is named as required alongside it.

#2. Implicit differentiation — the real sub-cases, each grounded to a citation

Reading $F(x,y)=c$ term by term, five structurally distinct term-shapes appear across the real record reviewed. Each genuinely needs a different piece of the rule below, and each is grounded to a specific question rather than presented as a general claim.

#2.1 A bare power of $y$ alone: $y^n \to ny^{n-1}\frac{dy}{dx}$

The direct chain-rule case — no product rule needed, because the term has no $x$ -factor to trigger one. This is the base case every other sub-case below is a variation on, and is not separately citation-worthy on its own (it appears inside nearly every real question below as one term among several); it is stated here because the remaining four sub-cases are all defined by what else gets combined with it.

#2.2 A mixed term, $x$ -power times $y$ -power: product rule, then chain rule on the $y$ -factor

Micro-example (fragment, not a full question). Differentiate $3x^2y$ with respect to $x$ : product rule with $u=3x^2$ , $v=y$ gives $\frac{d}{dx}[3x^2y] = 3x^2\frac{dy}{dx} + 6xy$ — the chain rule fires only on the $v=y$ factor (giving it a $\frac{dy}{dx}$ ), never on the $u=3x^2$ factor (which differentiates by the ordinary power rule, no $\frac{dy}{dx}$ attached).

This is confirmed as the single hardest mechanical step in implicit differentiation across the whole archive reviewed — January 2024, Q3(a), a term of the shape $y^2x$ requiring the product rule and chain rule together, where the examiner's report records candidates who "obtained one term as a result of the product rule" (dropping the other factor entirely) or who "obtained an incorrect term" on the other. The lesson file's own worked chain derives this exact compound step in full for $xy^2$ and is the right place for the complete derivation; this document names it as the sub-case's defining difficulty rather than re-deriving it.

#2.3 An $x$ -only term: no $\frac{dy}{dx}$ at all, by design, not by exception

Because $\frac{d}{dx}(x) = 1$ , a term containing $x$ but no $y$ differentiates by the ordinary rules with no chain-rule factor attached anywhere — the "invisible" multiplier is exactly 1. This sounds trivial stated abstractly, but the real record shows it is not trivially applied: October 2021, Q1's examiner report documents candidates "surprisingly" writing $4x^2+8 \to 8x+8\,\frac{dy}{dx}$ — an over-application of " $y$ -terms pick up a $\frac{dy}{dx}$ " pattern-matched onto a term that never contained a $y$ at all. The same report line supplies two further, independently-quotable slips on the $y$ -side of the same question: $3x^2y \to 6x\,\frac{dy}{dx}$ (dropping the product-rule term entirely, collapsing sub-case 2.2 into sub-case 2.1's rule) and $4y^2 \to 8\,\frac{dy}{dx}$ (dropping the $y$ itself, leaving only its coefficient after differentiating). All three are from one examiner-report line about one question (Oct 2021, Q1) — three distinct, real, quotable error patterns, not three independent series confirmations, and this document states that scope explicitly rather than let three items read as three data points.

#2.4 A mixed term where the $x$ -factor is itself composite: both factors need their own chain rule

This is the sub-case the existing lesson file's own header names as a documented gap in its worked material, and it is genuinely distinct from 2.2 rather than a restatement of it: in every product-rule example the lesson works through, the $x$ -side factor is a bare power of $x$ , whose derivative needs no chain rule of its own. The real anchor question for this topic, January 2021 Q6 (curve $4y^2+3x=6ye^{-2x}$ ), contains a term where that is not true. Re-deriving it in full as the micro-example this sub-case needs:

\frac{d}{dx}\!\left[6ye^{-2x}\right] = 6e^{-2x}\frac{dy}{dx} + 6y\cdot(-2e^{-2x}) = 6e^{-2x}\frac{dy}{dx} - 12ye^{-2x}

— product rule across the $y$ -factor and the $e^{-2x}$ -factor, and a second, separate chain-rule application on the $x$ -side factor itself (differentiating $e^{-2x}$ needs its own inner-function multiplier, $-2$ , exactly as it would outside any implicit-differentiation context). This matches the real mark scheme's own A1 criterion for this line, extracted directly from the primary source during this course's own lesson-audit pass: "Differentiates $6ye^{-2x}$ to obtain $-12ye^{-2x}+6e^{-2x}\frac{dy}{dx}$ ." The distinction from 2.2 is precisely this: sub-case 2.2's difficulty is remembering to apply the chain rule to the $y$ -factor at all; sub-case 2.4 additionally requires the ordinary (non-implicit) chain rule on the $x$ -factor, in the same line, and a candidate fluent in 2.2 alone can still stall here.

#2.5 A power of a sum, $(x+y)^n$ : two methods credited on exactly equal footing

Micro-example. For $(x+y)^3$ , Route A applies the chain rule directly: $\frac{d}{dx}[(x+y)^3] = 3(x+y)^2\left(1+\frac{dy}{dx}\right)$ (the inner derivative of $x+y$ with respect to $x$ is $1+\frac{dy}{dx}$ , since $x$ contributes 1 and $y$ contributes $\frac{dy}{dx}$ ). Route B expands first, $(x+y)^3 = x^3+3x^2y+3xy^2+y^3$ , then differentiates term by term using sub-cases 2.1–2.2 above.

Confirmed directly from October 2022 Q11(a), a "show that" question built on exactly this term shape: the mark scheme credits both routes on equal footing — "once one of the above two methods were clearly seen to be applied", full marks for the part became available (paraphrasing the mark scheme's own remaining wording). This is a real, checkable instance of a broader pattern this paper's mark schemes apply consistently (§4 of the facts bank documents the same equal-footing treatment for the three legitimate routes through a 3-term quadratic) — worth naming here because a candidate taught only the "efficient" chain-rule route may not realize the slower expand-and-differentiate route is not a downgrade, it is fully creditable.

#3. Parametric differentiation — the one real tested failure mode

The mechanism (§1 above) reduces to two ordinary derivatives and a division: differentiate $x(t)$ , differentiate $y(t)$ , then form $\frac{dy/dt}{dx/dt}$ . The real record contains exactly one question-level citation for this half of the spec point, but it isolates a specific, quotable mechanical slip rather than a setup error: October 2022 Q6, where the examiner's report records candidates simplifying $\frac{1}{3\sin^2 t}$ incorrectly to $3\cos^2 t$ — inverting the fraction and substituting $\cos^2 t$ for $\sin^2 t$ in the same slip, rather than the correct $\frac{1}{3}\csc^2 t$ or equivalent. This is a simplification error occurring after the correct differentiation and division have already happened, not a failure to know $\frac{dy}{dx}=\frac{dy/dt}{dx/dt}$ itself — worth distinguishing, because it means the real documented risk in this sub-case is algebraic tidying of a compound fraction, not the calculus step that produces it. The lesson file's own worked chain (parametric tangent at $t=2$ ) drills exactly this reciprocal risk in full and is the place for the complete worked derivation; this document names the sub-case's real failure point rather than re-running the chain.

#4. Forming the tangent or normal: the one trap that dominates this topic's exam record

Once $\frac{dy}{dx}$ (implicit or parametric) is known, constructing the tangent or normal is mechanically simple — substitute the specific point's coordinates (or parameter value) to get a numeric gradient, apply $m_1m_2=-1$ for a normal, then use the point-gradient line form. The real record shows the failure is almost never this mechanism; it is which point gets used. January 2021 Q6 (the 9-mark anchor for this whole topic, 5 marks for the implicit differentiation, 4 for the normal) is the source. Its examiner's report records, in its own words, "a surprisingly large number of candidates" who simply took the named point $P$ to be the origin outright and lost every mark in the part; and, a second, distinct depth of the same failure, candidates who found $P$ 's correct coordinates and then evaluated the gradient at $(0,0)$ anyway — losing every mark in the part either way. Two distinct failure depths, both converging on the same behavioural root: substituting a convenient default point instead of the one the question actually specifies. This is the same underlying instinct the FORGE guide's own cross-lesson synthesis names in a different spec point (position vectors, substituting a line's defining point instead of forming the vector the geometry actually requires) — two topics, one shared "reach for the nearby default object instead of the one the question names" pattern, independently confirmed rather than asserted here.

#5. Synthesis — every real exam angle this specific topic has actually been tested from

Restricting strictly to spec 5.1 (excluding the Jan 2024 cone/rates-of-change question, which is spec 5.2 and out of this document's scope), the facts bank's question-level review yields exactly five real, citable angles, across four series:

Series, question	Real angle tested	Marks	What the report/scheme actually documents
Jan 2021, Q6	Implicit differentiation of a mixed equation (incl. a composite $x$ -side factor, §2.4), then the normal at a named non-origin point	9 (5+4)	The wrong-point trap (§4) — the dominant, most consistently cited failure for this topic
Oct 2021, Q1	Implicit differentiation of a general $F(x,y)=c$ , testing multiple term-shapes in one question	— (not separately captured in the facts bank)	Three distinct slip patterns from one report line: dropped product-rule term, dropped chain-rule $y$ -factor, reflexive $\frac{dy}{dx}$ on an $x$ -only term (§2.2, §2.3)
Oct 2022, Q11(a)	Implicit differentiation of $(x+y)^3$ , a "show that" question	— (part-mark only, not separately captured)	Two methods (direct chain rule vs. binomial expansion) credited on exactly equal footing (§2.5)
Jan 2024, Q3(a)	Implicit differentiation of a $y^2x$ -shaped mixed term	— (part-mark only)	Confirmed as the single hardest mechanical step in the topic's whole reviewed record (§2.2)
Oct 2022, Q6	Parametric differentiation to a tangent, then a further part using it	— (part-mark only)	The reciprocal/sign simplification slip on a compound trig fraction (§3)

Per this course's own SIGNAL document, this topic is confirmed seen in 4 of 7 series that received real question-level mining (the honest denominator — see that document's §1 for why 7, not the facts bank's coarser 11-series count, is the correct base) — stated exactly that way, as a count against a small, named denominator, never as a percentage or a predictive likelihood, per this course's own no-fabrication discipline. No grade-boundary figure exists anywhere in this course's WMA14 source review for any topic, this one included; none is stated here.

What this synthesis adds beyond SIGNAL's frequency row and FORGE's per-question mark table: those two documents record that these five angles exist and how many marks the largest is worth. This section is the first place in the technique-hub layer that states which mechanical sub-case (§2.1–2.5, §3, §4) each real angle actually exercises — the connective layer between the theory in §§1–4 and the bare citation list SIGNAL and FORGE already carry.

#6. What is genuinely not evidenced — stated rather than invented

Per this document's own hard discipline: where the real record runs out, it is stated as absent, not filled in. Three plausible-sounding extensions of this topic have zero citation anywhere in the facts bank's WMA14 review, and are not present anywhere below as a result:

Second derivatives ( $\frac{d^2y}{dx^2}$ ) of an implicitly or parametrically defined curve. Spec 5.1's own text names only first-derivative differentiation and the resulting tangent/normal; no examiner report or mark scheme reviewed for this paper cites a second-derivative question for this spec point.
Stationary points or vertical tangents from $\frac{dx}{dt}=0$ on a parametrically-defined curve. Plausible in principle (a horizontal tangent occurs where $\frac{dy}{dt}=0$ with $\frac{dx}{dt}\neq0$ ; a vertical one where $\frac{dx}{dt}=0$ with $\frac{dy}{dt}\neq0$ ), but no real question citing this specific construction was found in the reviewed record for this spec point.
A grade-boundary or likelihood figure for this topic specifically. None exists anywhere in this course's WMA14 source review (confirmed directly in the SIGNAL document's §2) — restated here because a deep-dive document is exactly the kind of place a fabricated "commonly worth N% of the paper" claim would otherwise creep in.

This is the honest shape of the evidence for this topic: real, and sufficient for a genuine sub-case taxonomy and a five-item exam-angle synthesis, but narrower than this course's vectors or proof-by- contradiction material (per the facts bank's own §8 flag, echoed in the lesson file's header) — which is why this document is the length it is rather than longer.

#Cross-references (cited, not duplicated)

Full worked derivations for sub-cases 2.1–2.3 and §3, plus the complete trap-taxonomy and MCQ set: content/veridian/curriculum/wma14/implicit-and-parametric-differentiation-tangents-and-normals.ts.
The five-line method skeleton this document expands on: codex-formula-and-method-bank.md §2.5.
Per-question mark tariffs and the cross-topic "wrong default point/object" meta-pattern (§4 above): forge-exam-technique-guide.md.
Frequency-of-appearance figures (§5 above): signal-frequency-and-boundaries.md §1.
The perpendicular-gradient relation used without re-derivation in §1 and §4: wma11-perpendicularity-proofs-and-coordinate-geometry.ts.

#Sources used

research/veridian/WMA14-verified-facts.md §1 (spec 5.1/5.2 verbatim text), §3 (formula-booklet contents — confirmed no discrete implicit/parametric-differentiation formula exists there), §4 (general mark-scheme conventions), §5.5 (this topic's five citations: Jan 2021 Q6, Oct 2021 Q1, Oct 2022 Q11(a), Jan 2024 Q3(a), Oct 2022 Q6), §7 (topics-with-enough-material note), §8 (evidence-strength flags). The Jan 2021 Q6 mark-scheme fragment for the $6ye^{-2x}$ term (§2.4 above) is quoted from this course's own lesson-audit log for the existing lesson file (same facts-bank document, "Lesson-audit log" section), itself re-extracted directly from the primary Pearson mark-scheme PDF. No new primary Pearson document was fetched this session.