Chapter 5 — Trigonometric Functions of Real Numbers
5.2 - Unit Circle and Wrapping Function
With the total length of the circle already constructed in §5.1, what gives meaning to “wrapping” the real line onto it is missing: knowing that each partial arc length corresponds to a single point, and that every point corresponds to some length.
§1. Length of a partial arc
Section titled “§1. Length of a partial arc”§5.1 constructed by means of the length of the complete circle. This section repeats the same method, restricted to a portion of that circle —the arc traveled from , counterclockwise, to any other point—.
The inscribed polygons of §5.1 were constructed starting from a square with vertices , , , , and doubling sides by means of the arc midpoint —the point on the ray , with the midpoint of the chord—. That same procedure, applied repeatedly starting only from and —without completing the rest of the square—, generates a sequence of vertices that subdivides, ever more finely, the arc of the first quadrant.
Symmetry of the bisection. Each application of the arc midpoint bisects that arc into two parts of exactly equal length —not only approximately—: the reflection about the line exchanges the two sub-arcs, and since that reflection is an isometry —it preserves all distances, hence it preserves perimeters of inscribed and circumscribed polygons approximating each half—, both halves have the same length.
Consequently, if level of subdivision produces equally spaced vertices along the quadrant —, with —, the arc length from to is, by symmetry, exactly
—a -th part of of the quarter circle, whose total length is by the definition of §5.1—.
With arc length defined at the vertices of the subdivision, it remains to verify that those vertices, and the length that separates them, behave consistently with the order in which they appear along the arc. It is the content of §2.
§2. Monotonicity and the squeezing bound
Section titled “§2. Monotonicity and the squeezing bound”Proposition (monotonicity). For , , and has -coordinate strictly greater than that of .
Proof. The first claim is immediate from the formula of §1: . For the second, it suffices to verify it for consecutive vertices —— and proceed by induction on the levels of bisection. Let be two points of the first quadrant on the circle with —inductive hypothesis—, and the midpoint of the arc , with . Since are both in the first quadrant, —a convex combination of and — has -coordinate strictly between and ; and , the radial projection of onto the circle, preserves that relative position —scaling a vector by a positive factor does not change on which side of a line through the origin it lies, and the lines and seen as cones from the origin separate in the same angular region as and —. Hence . Since the initial square already satisfies , the induction on the levels of bisection gives the monotonicity at each level.
Let now and , for a given and each level , be the consecutive vertices of level such that —they exist and are unique by the monotonicity just proved, and because is a complete partition of into equal subintervals—. By construction, coincides with or advances to the new midpoint inserted between and , never moving back; similarly, never advances. Let , .
Proposition (squeezing bound). is decreasing, is increasing, for all , and in the sense that, for every , there exists with .
Proof. Monotonicity of and : a direct consequence of the previous Proposition and of the fact that does not move back, does not advance. : since precedes in the subdivision, by monotonicity. For the difference: the distance between and —two consecutive vertices of level — is, by Pythagoras, at most the length of the arc that separates them, —a chord never exceeds the arc it subtends, since the chord is, itself, an inscribed approximation of a single segment—; and , the difference of the -coordinates of two points, is at most that same distance. By the Archimedean Lemma —section 4.1, §4—, given there exists with .
With decreasing and bounded and increasing and bounded, squeezing against each other, the supremum axiom produces a single candidate point. It remains to verify that this point has, indeed, arc length . It is the content of §3.
§3. Existence of W(t) in the first quadrant
Section titled “§3. Existence of W(t) in the first quadrant”By the supremum axiom, decreasing and bounded below by has ; by the Lemma of existence of the infimum —or, symmetrically, the axiom itself applied to —, has . By the same approximation argument that closed the definition of in §4.2 §4 and that of in §5.1 §4 —bounding the difference and forcing it below any via the Archimedean Lemma, already verified in the Proposition of §2—, .
Definition. , for —the positive root, corresponding to the first quadrant—.
Proposition. , where is extended to any point of the arc by means of the same supremum/infimum of inscribed/circumscribed polygons of §1, now with final endpoint instead of a vertex of the subdivision.
Proof. For each , the point —with -coordinate equal to , since is the common value of infimum and supremum just proved— is located, by the monotonicity of §2, between and in the subdivision. Extending the inscribed path with the additional chord gives an inscribed approximation of the arc up to of length ; since —same bound as in §2, since is in the same sub-arc as —, it follows that taking supremum over this family of approximations, and bounding from above with the dual argument —chords plus the tangent at —. Since —definition of —, both chains of inequalities, valid for all , force by the Archimedean Lemma, exactly as it forced in the Proposition of §4.2 §4.
With defined on , extending it to the three remaining quadrants and then to all of does not require repeating the construction: it requires, only, reflecting and repeating with periodicity. It is the content of §4.
§4. Extension to the whole circle and to ℝ
Section titled “§4. Extension to the whole circle and to ℝ”Extension to the four quadrants. The points , , , divide the circle into four arcs, each of length —by construction of the initial square in §5.1—. For , define by repeating the construction of §3 on the arc from to , with playing the role of the variable of §3; similarly for and . In each case, the Proposition of §3 —with the coordinates and signs adjusted to the corresponding quadrant— gives .
Extension to by periodicity. For arbitrary , the Archimedean Lemma gives an integer such that —the “remainder” of dividing by —; that integer is unique, by the same uniqueness of Euclidean division. Define
using the construction already made on .
Proposition (periodicity and restricted injectivity). for all , and restricted to is injective.
Proof. The periodicity is immediate from the definition — and leave the same remainder modulo —. The injectivity on follows from the strict monotonicity of within each quadrant —§2— and from the fact that distinguishes different quadrants by the signs of the coordinates, which are not repeated within a single turn.
With completely constructed, there remain to be computed, without any new tool, the four points that divide the circle into exact quadrants. It is the content of §5.
§5. Distinguished points
Section titled “§5. Distinguished points”By direct construction —without additional computation—:
These four values are not a computation: they are, literally, the four vertices of the square with which the construction of §5.1 started, and the reason why that square —and not a triangle or a hexagon— was chosen as the starting point.
§6. Close
Section titled “§6. Close”This section translated, step by step, each piece of machinery already constructed in the book into a new geometric context: the Archimedean Lemma again bounded a difference until making it arbitrarily small, the supremum axiom again produced a value from two sequences that squeeze together, and the extension by periodicity —modulo — used the same Euclidean division with remainder that had been familiar long before this chapter. The only thing genuinely new was geometric, not analytic: the points one works with are no longer numbers on a line, but positions on a curve, and it had to be verified, with care, that the notions of “closer” and “farther” inherited from the real line still make sense when the independent variable travels around a circle and returns, every , over its own steps.
With unit circle completely constructed, the chapter has, at last, everything necessary to define sine and cosine as what they genuinely are: not relations between sides of a triangle, but the two coordinates of a point that moves, at constant speed in arc length, around the unit circle.
§5.3 defines the trigonometric functions of a real number from , and derives from that definition —does not postulate them— the fundamental identities.
Copyright © 2026 Carlos Dagorret — Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)