# Descartes’ Dream: Algebra IS Geometry

Charles M . Patton, Principal Scientist Center for Technology in Learning SRI International

Overview
• History, in brief • The Constructions, in Context • Number and Operation Duality - Build-it-yourself Numberline • Variables and Equations - What’s Unknown? • Constructing the Cartesian Plane - Finally • Summary and Commentary

René Descartes (1596 - 1650)
• “I think, therefore I am” • Cartesian Plane, Cartesian Coordinates • Geometry => Algebra => Geometry

=>

x2+y2=4 y=x-1

=>

- 7 + 1- 7 - 1 , ) ( 2 2

(

7+1 7-1 , ) 2 2

But his “secret notebook” suggests that it was actually Algebra => Geometry => Algebra that he was after.
(see, e.g., Amir D. Aczel, “Descartes’ Secret Notebook”)

Constructive Geometry
• Interpreting algebraic operations as geometric constructions

• But issues such as “trisecting an angle” prevented success

The metrical notions, such as length and angle, that all since Euclid assumed must be intrinsic to geometry, were actually a barrier to success of the program.

Axiomatic Geometry
• Following Descartes, there was 200 years of experiments with alternatives to Euclidean geometry: Can you leave out the parallel postulate? Do “straight lines” need to be “straight”? Can the notion of “length” vary from place to place? • As part of his “program” David Hilbert (1862-1943) organized variants as “axiom schemes” arguing, famously, as to the meaning of point, line, and plane, “At any time you have to be ready to substitute the words beer stein, bench, and table.” As long as the relations are preserved, everything else will follow.

Hilbert’s Fundamental Theorem
• Arguably the simplest of the axiom schemes was that for projective geometry, as it is primarily about incidence. • The fundamental theorem (paraphrased) holds that these incidence properties determine the underlying algebra of the projective geometry, providing geometric constructions for all the algebraic operations: Descartes was right! • John von Neuman (1903-1957) realized that the “yes-no” experiments of quantum mechanics fit the projective geometry scheme and used this interpretation to determine that complex, rather than real, numbers underlie quantum mechanics.

Wu’s Method
• Hilbert’s constructions were designed for generality. As a consequence, they are not all that “user friendly”. • In 1977 Chinese mathematician Wu Wen-Tsien devised an automated theorem proving method in geometry that bypassed both the usual axiomatic and coordinate approaches and went straight to the algebra-as-construction correspondence - with great success. • In the process, he set out the basic set-up and constructions in a much more user-friendly way. And it is his constructions that we will use today.

Our Set-up - Informally
• Any two distinct points determine a unique line to which they are both incident. • Any two distinct lines determine a unique point to which they are both incident • There are lots of points and lines • There’s a fixed line we have selected and call the “line at infinity.” Lines are called parallel when their joint point of incidence is incident to the line at infinity. • Desargues Axiom (next) • Pappas Axiom (later) • Order & continuity (later)

Desargues Axiom
ing edge-pairs incident to a Correspond line

If and only if

Corresponding vertexpairs incident to a point

More familiar form: Law of Perspective

Especially useful version

conclude

assume

The Gist of the Constructions

A distant light source casts a shadow from one line to another, taking each point on one line to a point on the other.

A second distant light source casts a shadow back on to the first line.

The combination takes any point on the first line to another point on the first line, transforming the line to itself.

Two Classes of Such Constructions
Choose a (finite) line and (finite) point on it. Call the point ‘0'.
0

Consider the class of ‘shadow’ transformations where the second line is parallel to the first:

And a second class of ‘shadow’ transformations where the second line is incident with the first at 0:
Multiplicative / Scaling 0

Independence of Choice
In each class, a transformation is completely determined by its action on any (generic) point the particular choice of second line and ‘lights’ is immaterial beyond this.

If these agree

0

Then all these will agree as well
(see ‘independence’)

With the choice of a ‘fiduciary’ point (0) there is an association of an (additive) transformation with every (finite) point on the line. To a point A, associate the transformation, “+A” that takes 0 to A. For any other point, B, B+A is the result of applying the transformation +A to B.

0

A

B

“addition by A takes 0 to A”

With the choice of a ‘fiduciary’ point (0) there is an association of an (additive) transformation with every (finite) point on the line. To a point A, associate the transformation, “+A” that takes 0 to A. For any other point, B, B+A is the result of applying the transformation +A to B.

0

A

B

B+A

“and takes B to B+A”

Subtraction on the Line
With the choice of a ‘fiduciary’ point (0) there is an association of the (additive) inverse transformation with every (finite) point on the line. To a point A, associate the transformation, “-A” that takes A to 0. For any other point, B, B-A is the result of applying the transformation -A to B.

0

A

B-A

B

“subtraction by A takes A to 0 and takes B to B-A”

Is this definition of addition commutative?

+B

+B

? +A +A?

?

?

desargues

desargues

desargues

Multiplication on the Line
With the choice of an additional ‘fiduciary’ point (1) there is an association of an (multiplicative) transformation with every (finite) point on the line. To a point A, associate the transformation, “*A” that takes 1 to A. For any other point, B, B*A is the result of applying the transformation *A to B.

0

A

1

B

“multiplication by A takes 1 to A”

Division on the Line
With the choice of an additional ‘fiduciary’ point (1) there is an association of an (multiplicative) inverse transformation with every (non-zero, finite) point on the line. To a point A, associate the transformation, “/A” that takes A to 1. For any other point, B, B/A is the result of applying the transformation /A to B.

0

A

1

B

B/A

“division by A takes A to 1 and takes B to B/A”

Building the Number Line, Equations, etc.