Proof of the Generalization of the Sawayama-Thébault Theorem

We prove two conjectures posed in 2016 concerning a generalization of the Sawayama-Thébault Theorem and the Sawayama Lemma. We show that this generalized statement can be viewed in Laguerre geometry, which provides a natural framework for resolving the problem.

💡 Research Summary

The paper tackles two conjectures formulated in 2016 that aim to generalize the classic Sawayama‑Thébault theorem and the associated Sawayama Lemma. After a thorough historical overview—tracing the original 1938 problem posed by Victor Thébault, early solutions by K. B. Taylor, H. Streefkerk, automated proofs by Chou, and later elementary proofs—the authors focus on the modern conjectures put forward by Vietnamese mathematician Dao Thanh Oai. The central claim is that both conjectures can be proved elegantly by interpreting the configuration within Laguerre geometry, specifically using an axial circular transformation (a dilation) that respects the oriented tangency of circles and lines.

The first main result is a “Generalized Sawayama Lemma.” In the classical lemma, a triangle ABC, a point D on BC, and a circle ω tangent to BC, AD and the circumcircle of ABC produce a line EF that always passes through the incenter I of ABC. The generalization replaces the vertex A with a circle ωₐ that is tangent to AB and AC; a second circle Ω passes through B and C and is tangent to ωₐ. A third circle ω is tangent to BC at D, to Ω, and to an arbitrary line k that is tangent to ωₐ at E. The theorem asserts that for any choice of k, the line DE always passes through the fixed point I, the incenter of triangle ABC.

The proof proceeds by applying a Laguerre dilation that maps the non‑degenerate circle ωₐ to its center Oₐ (effectively turning ωₐ into a point). Under this transformation, the configuration collapses to a triangle OₐB′C′ where B′ and C′ are the images of the intersections of Ω with BC. The classical Sawayama Lemma can now be invoked on triangle OₐB′C′, guaranteeing that the transformed line D′E′ passes through the incenter I″ of OₐB′C′. The authors then analyze the relationship between the original and transformed points: the vectors DD′ and EE′ are equal, perpendicular to BC, and have length equal to the radius of ωₐ. Consequently, translating D′E′ back by the fixed vector –DD′ yields the original line DE, and the translated incenter I″ becomes a point I′ that must coincide with the original incenter I of ABC. The coincidence is established by showing that A′ (the tangency point of Ω and ωₐ), D, and the midpoint M of the arc BC are collinear, and that I′ lies on the angle bisector of ∠BAC, which uniquely determines the incenter.

A series of auxiliary lemmas support the main argument. Lemma 2.1 presents an alternative configuration of the Sawayama Lemma, involving a circle tangent to the circumcircle at P, to BC at E, and to AD at F, and shows that EF always passes through the C‑excenter. Lemma 2.2 studies the internal/external tangency of two circles ωₐ and Ω and proves that the line joining their tangency point A′ with the incenter I bisects ∠BAC. Lemma 2.4 establishes a concyclicity of five points (F, I′, E, P, A′) using homothety, inversion, and power‑of‑a‑point arguments. These results are essential for handling the degenerate case where ω collapses to a point, ensuring continuity with the classical theorem.

The second main result is the “Generalized Sawayama‑Thébault Theorem.” Building on the generalized lemma, the authors consider the same configuration but now focus on the line joining the two external tangency points of Ω and ω (instead of the internal ones). By employing Lemma 3.1—a statement about involutions on a circle Ω induced by a pair of intersecting chords through a common internal tangency point A—the paper shows that the mapping Y↔Y′ (where Y and Y′ are second intersections of lines AXB and Af(X)B with Ω) is itself an involution. This involution property guarantees that the cross‑ratio relationships required for the Laguerre transformation remain invariant, allowing the same dilation argument to be applied. Consequently, the line joining the two new tangency points also passes through the incenter I, completing the generalization.

Throughout, the authors emphasize that the degenerate cases (when ωₐ becomes a point, or when the radius of ω tends to zero) reduce exactly to the classical Sawayama‑Thébault theorem and Lemma, confirming that their results are true extensions rather than merely analogous statements.

From a methodological standpoint, the paper’s novelty lies in recasting a classical Euclidean tangency problem into Laguerre geometry, where oriented tangency and axial circular transformations provide a natural language for handling both internal and external contacts simultaneously. This approach sidesteps the heavy algebraic machinery (e.g., Gröbner bases) traditionally used for such problems and yields a more synthetic, geometric proof.

However, the manuscript suffers from several presentation issues. The definitions of “oriented tangency” and the precise nature of the Laguerre dilation are only sketched; a formal statement with coordinates or a transformation matrix would strengthen the argument. Some steps—particularly the claim that the vector DD′ is constant—are justified intuitively but lack a rigorous derivation. Lemma 3.1’s involution claim is introduced without a full proof of existence or uniqueness of the mapping f, leaving a small logical gap. Moreover, the figures are inconsistently labeled, and typographical errors (e.g., “Sa way ama” split across lines) hinder readability.

In summary, the paper successfully demonstrates that the two 2016 conjectures are true, offering a clean geometric proof via Laguerre transformations. It enriches the theory of Sawayama‑Thébault type problems by providing a unified framework that accommodates both internal and external tangencies. With minor revisions to clarify definitions, fill the noted logical gaps, and improve the manuscript’s presentation, the work would make a valuable contribution to modern synthetic geometry.