The non-Euclidean revolution. With an introduction by H. S. M. Coxeter (Q2469814)

It would be hard, if not impossible, to introduce the Copernican revolution or the Darwinian revolution from the vantage points of the centuries that predated them, to a present-day undergraduate audience on the East Coast of the US (where the author has taught), for these are at the core of the contemporary mindset, the cognitive bias in their favour being too pronounced to allow for a candid turn back to a naive worldview. The situation is altogether different in the case of the non-Euclidean revolution, which has left untouched the naive belief of the mathematically unsophisticated that Euclidean geometry is the only geometry there is. Textbooks on the foundations of geometry commonly introduce a set of axioms, similar to those of Hilbert or of Birkhoff, and then proceed in a rigorous manner to deduce their consequences, pointing out which theorems rely on the Euclidean parallel postulate (EPP), which don't, and what consequences can be derived from the negation of the EPP. There are two kinds of problems with this approach. First, looking from the vantage point of modern axiomatics, the difficulties encountered over so many centuries with the question regarding the (in)dependence of the EPP, which these textbooks refer to in their historical notes, are hard to relate to. Second, students find modern axiomatics abstract and insufficiently motivated for what they perceive to be a subject as plain as geometry. This book introduces the non-Euclidean revolution as experienced by a reader of Euclid's \textit{Elements}, with the author as guide, and where all the resistance to non-Euclidean ideas a student can muster is channeled towards understanding the difficulties history had with both daring to think a non-Euclidean world possible and to accept it once it was exhibited. The level of rigor is kept deliberately below that of the modern Hilbert-style foundational enterprise (``We'll plug the major holes in the \textit{Elements}, root out the intuitive components that are likely to cause trouble when we get to non-Euclidean geometry, and let it go at that. Harmless intuitive references [\(\ldots\)] we will allow to remain; we won't bother to construct the involved logical substructure with which, as Hilbert has shown, they can be supported.'' (p.\ 38)) to make the book first and foremost enjoyable for the intended audience. It starts off with Euclid's postulates and common notions (adding a Common Notion 6 ensuring that geometric magnitudes are positive real numbers), and adds postulates as needed (one corresponding to the circle axiom, one corresponding to the Pasch axiom, the SAS congruence criterion for triangles, one stating that congruent triangles have equal area (where the concept of area is left undefined), and a postulate allowing the choice of random points with certain betweenness properties). It goes on to reprove most of Theorems in the \textit{Elements, Book I}, while carefully mentioning, in a two-column format, why the deductions are allowed. A chapter titled ``The Diamond Theory of Truth'' presents the Kantian view, followed by one on equivalents of the EPP (given neutral geometry with continuity), Saccheri's quadrilateral, and an explanation of the difference between ancient and modern axiom systems. Hyperbolic geometry, based on the Hyperbolic Parallel Postulate as stated by Hilbert, receives the same synthetic treatment the theorems of the \textit{Elements, Book I} were given, with special emphasis on both the various notions of parallelism in hyperbolic geometry and on properties of asymptotically parallel lines. The Poincaré model is used to prove consistency, and there is a section on the Luneburg-Blank theory of binocular vision. The book remains the only one of its kind, and, as long as there will be students at liberal arts colleges pondering the Big Questions rather than being prepared to become professional mathematicians, it will continue to have a captive audience if it finds adequate teachers. This being a reprint, none of the minor infelicities mentioned by J. F. Rigby in his review in [Math. Gazette 73, 68--69 (1989)] could be remedied.