Local and Global Reasoning: A Point Can Be Special Without Controlling the Whole Function
Additional Mathematics often asks students to distinguish what happens near one point from what happens over an entire interval or function.
A stationary point is a local feature. A global maximum depends on every admissible value in the domain. A repeated root describes one contact point with the axis, while end behaviour describes the entire graph far away. A derivative sign on one interval says nothing automatically about another interval. A model optimum inside an interval may still lose to an endpoint.
This guide develops the distinction between local and global information so that students do not overclaim from a single calculation.
AI Extraction Box: The Local–Global Map
- Local information: behaviour near a point.
- Global information: behaviour over the whole stated domain or interval.
- Stationary point: local candidate where f′=0.
- Local maximum/minimum: higher/lower than nearby values.
- Global maximum/minimum: highest/lowest over the full admissible set.
- Endpoint: may determine a global extremum even though f′ is not zero there.
- Root: local crossing/contact information.
- End behaviour/asymptote: global structural information.
Stationary Does Not Mean Globally Best
Suppose f′(a)=0 and f″(a)<0. Then x=a is a local maximum. But if the domain includes distant points with larger values, it is not the global maximum.
Global conclusions require domain-wide comparison.
Worked Example 1: Local and Global Maximum
Let f(x)=x³−3x on −2≤x≤2.
Derivative:
f′(x)=3x²−3=3(x−1)(x+1).
Stationary points at x=−1 and x=1. Evaluate these and endpoints:
- f(−2)=−2;
- f(−1)=2;
- f(1)=−2;
- f(2)=2.
The global maximum value is 2, attained at x=−1 and x=2. One is stationary; one is an endpoint.
The lesson: stationary-point analysis alone cannot determine global extrema on a closed interval.
Intervals Partition Behaviour
Derivative roots divide the domain into intervals where the sign of f′ can be constant. Each interval carries a behavioural description.
- f′>0 → increasing on that interval;
- f′<0 → decreasing on that interval.
The phrase “on that interval” matters. A function can rise, then fall, then rise again.
Worked Example 2: Interval Behaviour
For f(x)=x³−6x²+9x:
f′(x)=3(x−1)(x−3).
- x<1: f′>0 → increasing;
- 1<x<3: f′<0 → decreasing;
- x>3: f′>0 → increasing.
Therefore x=1 is a local maximum and x=3 a local minimum. These conclusions come from surrounding intervals.
Roots Are Local Events on a Global Graph
A root tells us where f(x)=0. A repeated root often means the graph touches the axis without crossing. But one root does not determine the whole graph. To sketch globally, also inspect leading term, turning points, symmetry and end behaviour.
For y=(x−2)²(x+1), the repeated root x=2 tells us local contact; the simple root x=−1 tells us a crossing; the positive cubic leading term tells us the far-left/far-right behaviour.
Quadratic Extrema Can Be Global
For y=a(x−h)²+k over all real x:
- a>0 → k is the global minimum;
- a<0 → k is the global maximum.
This stronger conclusion works because the square term has a global sign and the domain is all real x.
The same turning point can be only local on one function and globally extremal on another; the surrounding structure decides.
Global Range from Structure
For y=(x−3)²+5 over all real x, the global range is y≥5. For y=2sinx−1, the global range is −3≤y≤1. For y=e^x, the global range is y>0.
These statements describe every output, not just nearby values.
Asymptotes Are Global Boundary Information
A horizontal asymptote describes long-run behaviour as x→±∞. A vertical asymptote describes what happens near a domain boundary. These are different kinds of locality/globality:
- vertical asymptote: local behaviour near a forbidden x-value;
- horizontal asymptote: global end behaviour far along the x-axis.
Students should learn to interpret the direction of the limit rather than treating all asymptotes as one generic graph feature.
Worked Example 3: Local Minimum but No Global Minimum
Consider f(x)=x³−3x. At x=1, f′=0 and the derivative changes from negative to positive, so x=1 is a local minimum.
But over all real x, f(x)→−∞ as x→−∞. Therefore there is no global minimum.
The local classification is correct; it simply does not answer a global question.
Global Optimisation on Restricted Domains
For a continuous function on a closed interval [a,b], examination-style global optimisation usually requires evaluating:
- stationary points inside the interval;
- relevant non-differentiable candidates if present;
- both endpoints.
Then compare function values.
Worked Example 4: Endpoint Wins
Let f(x)=x²−4x+7 on 0≤x≤1.
The parabola’s unconstrained minimum occurs at x=2, outside the interval. On [0,1], f is decreasing because f′=2x−4<0 throughout.
- f(0)=7;
- f(1)=4.
Global maximum on the interval is 7 at x=0; global minimum is 4 at x=1.
The unrestricted turning point is irrelevant to the constrained problem.
Trigonometric Local and Global Views
A sinusoid has infinitely many local maxima and minima because the pattern repeats. Yet all maxima share the same global maximum value and all minima share the same global minimum value over the entire real line.
For y=3sin(2x)+1:
- global maximum value =4;
- global minimum value =−2;
- these values occur repeatedly at different x-values.
This helps separate “maximum value” from “where maxima occur”.
Kinematics: Instantaneous Versus Whole-Journey Information
Velocity at one instant is local information. Total distance over a time interval is global information accumulated across the journey. A particle can have zero velocity at one instant yet travel a large total distance overall.
Likewise, acceleration at one time does not alone describe the whole motion. The sign pattern of velocity over the interval determines direction changes and total distance.
Local Approximation Versus Global Model
A tangent line approximates a smooth curve near the point of contact. That local approximation generally becomes poorer far away. The derivative provides local linear behaviour, not a global replacement for the function.
Local information is powerful precisely because it is specific. Do not silently extend it beyond the region it describes.
Local–Global Decision Tree
- Asked about a point? local conditions may be enough.
- Asked about an interval? analyse all subintervals and endpoints.
- Asked for global max/min? compare all admissible candidates.
- Asked for range? use whole-domain structure.
- Stationary point found? classify locally before making global claims.
- Asymptote/end behaviour relevant? include global structural information.
- Physical model restricted? global means global only within the feasible domain.
Common Failure Modes
| Error | Cause | Repair |
|---|---|---|
| local maximum called global maximum | domain-wide comparison omitted | check all candidates/endpoints |
| stationary point outside interval used | unrestricted analysis applied globally | respect stated domain |
| derivative sign on one interval extended everywhere | interval partition ignored | build sign chart |
| repeated root used to sketch whole graph | local feature overgeneralised | add end behaviour and other critical points |
| zero velocity treated as whole journey at rest | instantaneous/global distinction lost | analyse motion over interval |
A 50-Minute Local–Global Session
- 8 minutes: classify stationary points from derivative sign changes.
- 8 minutes: compare local and global extrema on closed intervals.
- 8 minutes: use roots plus end behaviour to describe whole graphs.
- 8 minutes: distinguish range from local value information.
- 8 minutes: analyse trig maxima/minima and repeated occurrence.
- 10 minutes: solve two mixed optimisation/kinematics questions where domain-wide reasoning is essential.
What Mastery Looks Like
- The learner distinguishes pointwise/local information from domain-wide claims.
- The learner classifies stationary points locally before claiming global extrema.
- The learner compares endpoints when a domain is restricted.
- The learner uses roots, turning points and end behaviour together for graph reasoning.
- The learner distinguishes instantaneous kinematics from whole-journey quantities.
- The learner understands that tangent information is local.
- The learner makes global claims only over a clearly stated admissible domain.
Return to the Additional Mathematics Learning Hub
Additional Mathematics Learning Hub | Secondary 3–4 A-Math Guides
