Call the Feds!

We’ve Got Nested Radicals!

Alan Craig

F. Lane Hardy Seminar

11-3-08
 What are Nested Radicals?

2 2
Examples:

2 2 2

2 2 2 2
 We could keep this up
forever!

2  2  2  2  2  2 
 If we did,
what would we get?

=?
2  2  2  2  2  2 
 Let’s work up to it.
What are the values of these expressions?
2 ?

2 2 ?

2 2 2 ?

2 2 2 2 ?

2 2 2 2 2 ?
 Let’s work up to it.
What are the values of these expressions?
2  1.414

2 2 ?

2 2 2 ?

2 2 2 2 ?

2 2 2 2 2 ?
What are the Values?
2  1.414

2  2  1.848

2 2 2 ?

2 2 2 2 ?

2 2 2 2 2 ?
What are the Values?
2  1.414

2  2  1.848

2  2  2  1.962

2 2 2 2 ?

2 2 2 2 2 ?
What are the Values?
2  1.414

2  2  1.848

2  2  2  1.962

2  2  2  2  1.990

2 2 2 2 2 ?
What are the Values?
2  1.414

2  2  1.848

2  2  2  1.962

2  2  2  2  1.990

2  2  2  2  2  1.998
What value is this sequence
of numbers approaching?
1.414
1.848
1.962
1.990
1.998

 Now what do you think the value of
this infinite nested radical is?
2  1.414

2  2  1.848

2  2  2  1.962

2  2  2  2  1.990

2  2  2  2  2  1.998

2  2  2  2  2  2   ?
 You’re Right!

2  2  2  2  2  2   2
 Trigonometry

Let’s see an example of where an

infinite nested radical could arise.

Warning: Brief Excursion into

Trigonometry!
 Half-Angle Formula

• We will use the half-angle formula for cosine

to take another look at this sequence and its
limit.

  1  cos 
cos   
2 2
  
Let’s use the formula to find cos  .
8

 
1  cos 
   4
cos  
8 2

  2
Now, cos   , so
4 2
  
Let’s use the formula to find cos  .
8

  2
1  cos  1
   4 2
cos   
8 2 2

Let’s rationalize the last expression by

multiplying numerator and denominator by 2.
  
Let’s use the formula to find cos  .
8

  2
1  cos  1
  4
  2
cos  
8 2 2

 2
1  2
 2 
  
22

2 2

4
  
Let’s use the formula to find cos  .
8

  2
1  cos  1
   4 2
cos   
8 2 2

2 2 2 2
 
4 2
  
Let’s use the formula to find cos  .
8

  2
1  cos  1
   4 2
cos   
8 2 2

2 2 2 2
 
4 2

Now multiply both sides by 2.

  
Let’s use the formula to find cos  .
8

  2
1  cos  1
   4 2
cos   
8 2 2

2 2 2 2
  
4 2

 
2 cos   2  2
8
 Repeatedly using the ½ angle formula:

 
2  2  2 cos 
8

 
2  2  2  2 cos 
 16 

 
2  2  2  2  2 cos 
 32 
 Repeatedly using the ½ angle formula:

 
2  2  2 cos  As the angle  gets
8 smaller and smaller
approaching 0, what
value is the cos()
  approaching?
2  2  2  2 cos 
 16 

 
2  2  2  2  2 cos 
 32 
 Repeatedly using the ½ angle formula:

  Recall cos(0) = 1, so
2  2  2 cos 
8 2 cos() is approaching 2
as  approaches 0.

 
2  2  2  2 cos 
 16 

 
2  2  2  2  2 cos 
 32 
 Repeatedly using the ½ angle formula:

That is,

cos   1 as   0


2  2  2  2    2 cos 0   2 1  2
That’s all the trigonometry
for this session.
 To Recap:
We have shown in two different ways

2  2  2  2  2  2   2

that the equation ‘ought’ to be true

 Now let’s ‘prove’ it.

2  2  2  2  2  2   2
 Set x equal to the expression.

x  2  2  2  2  2  2 
 Square both sides.

x2  2  2  2  2  2  2 
 Subtract the original equation
from the squared equation.

x  2  2  2  2  2  2 
2

x  2  2  2  2  2  2 
 Subtract the original equation
from the squared equation.

x  2  2  2  2  2  2 
2

x  2  2  2  2  2  2 

x x2
2
Now solve the equation.
x x2
2
Solve the equation.
x x2
2

 x x20
2
Solve the equation.
x x2
2

 x x20
2

 ( x  2)( x  1)  0
Solve the equation.
x x2
2

 x x2 0
2

 ( x  2)( x  1)  0
x2

Why did we not use x = -1?

 So

2  2  2  2  2  2   2
 What about?

3  3  3  3  3  3 
 Does

3  3  3  3  3  3 

= 3 ???
 Using the same process as
before, we get

x  3  3  3  3  3  3 
2

x  3  3  3  3  3  3 

x  x  3  x  x 3  0
2 2
 Recall the Quadratic Formula

 b  b  4ac
2
ax  bx  c  0 
2
x
2a

• We have x  x  3  0
2

• So a = 1, b = -1, and c = -3 and

 (1)  (1) 2  4(1)(3)

 x
2(1)
1  13
 x
2
 So, No, we do not get 3

1  13
x  x 3  0  x 
2
 x  2.3, so
2

3  3  3  3  3  3    2.3
 Let’s ask a slightly
different question.
• Is there a positive integer a, such that if we
replace 3 under the nested radical with a, the
nested radical will equal 3?
 Let’s ask a slightly different question.

• That is, is there an a that makes the equation

below true?

a  a  a  a  a  a   3 ?
 Let’s ask a slightly different question.

• That is, is there an a that makes the equation

below true?

a  a  a  a  a  a   3 ?

• Yes! And we are going to find it.

 Subtract the original equation
from the squared equation.

x  a  a  a  a  a  a 
2

x  a  a  a  a  a  a 

x xa
2
 Finding a

1  1  4a
x xa 0  x 
2

(Using the quadratic formula)

 Finding a

1  1  4a
x
2
We want x = 3, so

1  1  4a
3
2
 Finding a
1  1  4a
 3  1  1  4a  6 
2
 Finding a
1  1  4a
 3  1  1  4a  6 
2

1  4a  5  1  4a  25 
 Finding a
1  1  4a
 3  1  1  4a  6 
2

1  4a  5  1  4a  25 

a6
 So we have shown that

6  6  6  6  6  6   3
 Now let’s generalize our result.

• ‘Prove’ that for any integer k > 1, there is a

unique positive integer a, such that

a  a  a  a  a  a   k

Note: The following is not a true mathematical proof of this theorem (which would
use limits of bounded, monotonically increasing sequences) but does suggest the
core reasoning and result of such a proof.
 Finding a
1  1  4a
k  1  4a  2k  1
2
 Finding a
1  1  4a
k  1  4a  2k  1 
2

1  4a  4 k 2  4k  1
 Finding a
1  1  4a
k  1  4a  2k  1 
2

1  4a  4 k 2  4k  1  a  k 2  k 
 Finding a
1  1  4a
k  1  4a  2k  1 
2

1  4a  4k  4k  1  a  k  k 
2 2

a  k (k  1)
 We have shown that
For any integer k > 1, there is exactly one integer
a = k (k - 1), such that

a  a  a  a  a  a   k
 We have shown that
For any integer k > 1, there is exactly one integer
a = k (k - 1), such that

a  a  a  a  a  a   k

That is, every integer can be represented as an

infinite nested radical!
 Example: k = 4

a  k (k  1)  4  3  12

12  12  12  12  12  12    4
 Example: k = 5

a  k (k  1)  5  4  20

20  20  20  20  20  20    5
 Another Way
Alternatively, we might have noticed that
we need to solve
x xa 0
2

in such a way that we get two numbers

that multiply to make a and subtract to
make 1. Further, one of the numbers must
be k. (Why?) Thus, the other number
must be k - 1 and a must be k (k - 1).
 That is
Let k be one of the numbers and h the other :

k  h  a and k  h  1, 

a a
h   k  1 k  a  k
2

k k

 a  k 2  k  k (k  1)
The END?
 The END?

No!

This is way too much fun!

 Let’s Kick it Up a Notch!

a  b a  b a  b a  b a  b a 
 Let’s Kick it Up a Notch!

a  b a  b a  b a  b a  b a 

Note that what we did before

was a special case of this
expression with b = 1.
 Let’s Kick it Up a Notch!

a  b a  b a  b a  b a  b a   k

For each integer k > 1, there are

exactly k - 1 pairs of integers a and
b, 0 < b < k, that satisfy this equation.
Further, a  k (k  b).
 As before, square the equation.

x  a  b a  b a  b a  b a  b a 
2

But before we subtract the original equation

from the squared equation, we must isolate the
radical (so that it will subtract away).

x2  a
 a  b a  b a  b a  b a  b a 
b
 Now subtract.
x2  a
 a  b a  b a  b a  b a  b a 
b

x  a  b a  b a  b a  b a  b a 

x2  a
x0
b
 Now subtract.
x2  a
 a  b a  b a  b a  b a  b a 
b

x  a  b a  b a  b a  b a  b a 

x2  a We will solve this by

 x  0  x 2  bx  a  0 factoring now but keep
b it in mind for later.
 Factor
For integer solutions of

x  bx  a  0
2

we need two integers that multiply to

make a and have a difference of b.
One of the numbers must be k, so the
other is k - b. Thus,

a  k ( k  b)
 (k – 1) Pairs
There are exactly k – 1 such pairs
a and b:
a  k ( k  b) b (difference)

k (k  1) 1
< k k (k  2) 2
b
t 0 < k (k  3) 3
t ha
a ll   
c
Re k  2 k 2
k 1 k 1
 Example: k =4
If k = 4, the k – 1 = 3 pairs a and b
are:
a b
4  3  12 1
42  8 2
4 1  4 3
 Example: k =4

12  12  12  12  12  12    4

8  2 8  2 8  2 8  2 8  2 8   4

4  3 4  3 4  3 4  3 4  3 4   4
 One Last Thought

Consider this continued fraction:

a
b
a
b
a
b
a
b
b 
Suppose it converges to x, then

a
x b
a
b
a
b
a
b
b 
Notice the shaded area is also x

a
x b
a
b
a
b
a
b
b 
 Rewriting the continued fraction

a a
x b  x b
a x
b
a
b
a
b
b 
 See what we get!

a a
x b  x b
a x
b
a
b
a
b
b 

 x  bx  a  0
2
Does this look familiar?
 Yes, these are equal!!!
a
b
a
b
a
b
a
b
b 

 a  b a  b a  b a  b a 
 In particular, set a = b = 1.
1
1
1
1
1
1
1
1
1

 1  1 1  1 1  1 1  1 1    ???
 The Golden Ratio 
1
1
1
1
1
1
1
1
1

1 5
 11 11 11 11 1 
2

(But that’s another F. Lane Hardy talk.)

End  End  End  End  End  

?
Reference
Zimmerman, S., & Ho, C. (2008). On infinitely nested
radicals. Mathematics Magazine, 81(1), 3-15.