Recently a student of mine – a middle school teacher – asserted that he did not believe that 0.999… = 1. I tried several methods of convincing him that he was wrong; for example calculating 1 = 3(1/3) = 3(0.333…) = 0.999…, and showing that 0.999 … < 1 led to a contradiction. I browbeat him into submission, but I don’t think I convinced him. After reading the book under review, I see I could have handled things better. I learned that in non-standard analysis, 0.999 … < 1 is true, and more importantly I learned that understanding even the simplest infinite processes involves mastering some tricky metaphors.
This is an important and flawed book. It has generated much commentary, pro and con, and if your curiosity is whetted by my review I suggest you check the reviews of the book in Amazon.com. Like The Number Sense by Stanislaus Dehaene (reviewed below), this book applies cognitive science to an analysis of mathematical thought. But unlike that book, Where Mathematics Comes From goes far beyond an analysis of arithmetic skill, and analyzes some very sophisticated concepts of higher mathematics. Also, compared with Dehaene’s book, this book does not depend very much on laboratory science, but depends almost entirely on theoretical cognitive science.
This is a long book, and a somewhat difficult read, unless you happen to be well versed in the jargon of both mathematics and cognitive science. Even though the authors cover a lot of territory in 450 pages, and the quality of the writing is generally good, I had the feeling that I was reading the same thing over and over. If you want an “executive summary” to get the gist of what the authors believe they have accomplished, I suggest looking at the section “A Portrait of Mathematics” on pages 377 – 379.
In the first four chapters the authors describe the brain’s innate arithmetic, which is quite rudimentary and similar to the innate arithmetic of many other species, and then details how humans have learned to extend these basic concepts to an arithmetic which enables efficient calculation and obeys certain “laws”, such as the commutative property for addition. This section sets the tone for the rest of the book, by introducing basic cognitive mechanisms that the authors believe explain how mathematics is invented and understood. These include grounding metaphors, which yield basic, directly grounded ideas, and linking metaphors, which yield abstract ideas.
Chapters 5 – 7 study the linking metaphors that determine algebra, logic, and sets, and chapters 8 – 11 deal with “The Embodiment of Infinity”. The concept of infinity underlies most of modern mathematics including various number systems (integers, rationals, real numbers, and complex numbers). It appears in many different guises from points at infinity in projective geometry to cardinal and ordinal infinities in Cantor’s theory. The authors introduce a “basic metaphor of infinity” (BMI) that is supposed to account for our understanding of all these concepts. Chapter 11, “Infinitesimals”, is perhaps the most mathematically interesting part of the book. The authors present the hyperreal numbers of Robinson and Keisler, which include infinitesimal and huge quantities and provide be an intuitive and direct way of dealing with calculus. They also introduce a system of their own invention, the granular numbers, which is a subset of the hyperreal numbers that seems to be easier to use.
Chapters 12 – 14 continue the discussion of the infinite by critiquing the program of modern analysis that was pioneered by Dedekind and Weierstrass and continues to this day. While the authors profess the highest admiration for the intellectual achievements of these men and their followers, the subtext here is that their (Dedekind’s and Weierstrass’) purpose in separating analysis from its roots based in an intuitive geometric understanding was to obfuscate the subject and make mathematics the preserve of a specially trained elite. I disagree. My understanding is that the development of technology was beginning to make the traditional conceptions of mathematics inadequate. For example, on page 307 the authors approvingly cite James Pierpont’s (1899) list of “prototypical properties of a curve” including that it is continuous and has a tangent. With this definition, it would be difficult to have a consistent theory of Fourier series necessary for the analysis of radio waves or to solve differential equation with a driving function given by a step function in electrical engineering. The Mandelbrot set and related constructions are rooted in the real world (Mandelbrot’s prototypical example is the coastline of England.) and these “monstrous” sets seem to describe nature better than classical curves and regions.
Chapters 15 and 16 contrast the author’s implications for a philosophy of mathematics that is grounded on the human mind –“embodied mathematics”– with other philosophies that either posit mathematics existing outside the real world (Platonism and “the romance of mathematics”) or see mathematics the manipulation of essentially meaningless strings according to given rules (formalism) or as a cultural construct. I will describe my own view in a separate posting.
The last section of the book is a “case study” in four parts shows that the methods developed in the book can be used successfully to teach mathematics in a way that focuses on meaning. The subject is Euler’s famous equation e^(i*pi) = -1. This is quite good, though I might do a few things differently