Associated Topics || Dr. Math Home || Search Dr. Math

### Mathematics As Analogy

```Date: 12/16/2003 at 17:35:01
From: Mark
Subject: Theoretical Question about Circumference and Diameter

Everyone knows that to date, the number representing pi is infinite,
so, for the sake of convenience, we often round it off to a decimal
place appropriate for whatever purpose we are calculating it.  But can
somebody explain why it's infinite, or so hard to calculate to the
end?

I mean, if pi is the actual relation of the circle's diameter to its
circumference, how can this ratio be infinite when the diameter has a
finite length, and if you were to disconnect the circle and measure
it in a straight line, it too has a finite length.  Is there a formula
to explain this?
```

```
Date: 12/17/2003 at 12:57:12
From: Doctor Ian
Subject: Re: Theoretical Question about Circumference and Diameter

Hi Mark,

A lot of people get confused by thinking that pi is something that is
determined by actually _measuring_ physical circles.  It's not.

Pi is the ratio of circumference to diameter for any _mathematical_
circle.  But there are no mathematical circles in the real world, so
we can't ever find pi by measuring something:

Why Pi?
http://mathforum.org/library/drmath/view/61017.html

Does this help?

- Doctor Ian, The Math Forum
http://mathforum.org/dr.math/
```

```
Date: 12/17/2003 at 14:08:21
From: Mark
Subject: Theoretical Question about Circumference and Diameter

Thank you for the answer.  That clears up a lot of questions, but it
makes me wonder about something else.  I hope you don't mind if I get
a little philosophical here.

Am I to assume then, that if pi as an irrational number highlights the
distinction between the ideal circle and the real world circle, we
can also view other geometric shapes in the same light?  For example,
parallel lines must never cross to be considered parallel, but I
imagine then, this only occurs in the ideal because limitations of
our measuring instruments would preclude two lines from being placed
beside each other perfectly straight, and these imperfections would
become apparent the longer the lines ran.

Second, I saw that it can be mathematically demonstrated that pi is
irrational, but I was wondering if we can still use the idea of
physically measuring a perfect circle to illustrate the irrationality
of pi in more layman's terms.

That is, if we can imagine the idealized circle which doesn't exist in
reality, does that mean the perfect curvature of this circle confuses
the actual meeting points of a hypothetically perfect diameter line
with the perimeter of the circle?  I could understand how pi has no
end if we thought of it in this way.  Does this make any sense?
```

```
Date: 12/17/2003 at 22:16:20
From: Doctor Ian
Subject: Re: Theoretical Question about Circumference and diameter

Hi Mark,

>Am I to assume then, that if pi as an irrational number highlights
>the distinction between the ideal circle and the real world circle,
>we can also view other geometric shapes in the same light?

Yes.  It's not emphasized enough that the way math works is that we
create idealized concepts, and then look for situations where we can
set up _analogies_ between these concepts and things we observe in the
world.

For example, we might have a situation in the world where we have a
collection of objects with unique identities (e.g., humans, or dates,
as opposed to hydrogen atoms).  Numbers also have unique identities,
so we can use a number to represent each object.

Going further, if the objects fill some space of possible objects, as
with dates, then we assign consecutive numbers to consecutive objects.

If they don't, as with people or accounts, then we can be more
arbitrary in assigning numbers (e.g., social security or credit-card
numbers).  But when we do this, we give up the ability to calculate
things like "distances" between pairs of objects.

But as with any analogies, if you push them too hard, they break.

>For example,
>parallel lines must never cross to be considered parallel,

In a flat space, anyway.  In a spherical space, they can cross (e.g.,
the meridians of longitude on a globe intersect at the poles--each
one is a "line" within the surface of the globe):

Why is Pi a Constant?
http://mathforum.org/library/drmath/view/57828.html

>but I
>imagine then, this only occurs in the ideal because limitations of
>our measuring instruments would preclude two lines from being placed
>beside each other perfectly straight, and these imperfections would
>become apparent the longer the lines ran.

Even before you get to that point, you have to deal with the fact that
mathematical lines are infinite in extent!  Normally, we model things
with line segments instead of lines, and not crossing isn't sufficient
to determine whether two line segments are parallel or not.

But the main thing is that we can't make observations of physical
objects, and then use those observations to draw conclusions about
mathematical concepts.

Suggestions, yes!  But not conclusions.  And this is a subtle but
important point.  Looking at objects in the world, we get ideas for
mathematical idealizations of those objects; but then we define the
idealizations within axiomatic systems that aren't in any way based on
the world.  And these axioms are the only valid sources of conclusions
about the mathematical concepts they define.

People often think that the point of math is to "describe the world",
but this is a misunderstanding:

What is Mathematics?
http://mathforum.org/library/drmath/view/52350.html

>Second, I saw that it can be mathematically demonstrated that pi is
>irrational, but I was wondering if we can still use the idea of
>physically measuring a perfect circle to illustrate the irrationality
>of pi in more layman's terms:

Sure, although I would drop the word "physically".  One way to do that
is by constructing an infinite series of approximations, each of which
roughly corresponds to a "measurement".

For example, suppose we construct a circle, and then construct the
largest square that will fit inside it.  The perimeter of the square
is an approximation to the circumference of the circle, although not a
very good one.  :^D

And thus the ratio of a diagonal of the square to the perimeter of the
square is an approximation to the value of pi.  If the diameter of the
circle is 1, that's also the diagonal of the square; and each side of
the square will have length sqrt(2)/2.  So the perimeter would be 4
times sqrt(2)/2, or 2 times sqrt(2).  Given a diameter of 1, this is
also an approximation of pi:

2*sqrt(2) = 2.82

It's a little on the small side, which is what we'd expect, since the
circumference is completely outside the perimeter except at four
points of contact.

Now, suppose instead of a square, we use a hexagon.  (We always want
to use a polygon with an even number of angles, so that a diagonal of
the polygon is the same as the diameter of the circle.)  We can do the
same kind of calculation, and we'll arrive at a larger ratio of
perimeter to diameter.

The more angles we use, the closer the approximation will be.  But to
get all the way to pi, we'd have to use an infinite number of sides!
The only way to get a value whose decimal expansion terminates or
repeats would be to use a finite number of sides.  But then we'd be
computing something other than pi, wouldn't we?

>If we can imagine the idealized circle which doesn't exist in
>reality, does that mean the perfect curvature of this circle confuses
>the actual meeting points of a hypothetically perfect diameter line
>with the perimeter of the circle? I could understand how pi has no
>end if we thought of it in this way. Does this make any sense?

I'm not quite sure what you're getting at here.  Can you say more

- Doctor Ian, The Math Forum
http://mathforum.org/dr.math/
```

```
Date: 12/18/2003 at 01:13:24
From: Mark
Subject: Theoretical Question about Circumference and Diameter

This is very abstract for me so I'll do my best to explain:

I understand that in a perfect circle (a circle which does not exist
in reality) pi is irrational or infinite.  We can only demonstrate
this mathematically.  We could never demonstrate this with a real
world circle using physical measuring devices.

But let's imagine someone measuring the diameter of this perfect
circle with a perfectly precise measuring device that could give
infinite measurement if required.  (This also assumes the line segment
representing the diameter of the circle is perfectly straight and
passes through the real centre of the circle.)

Since pi is infinite, we could never be able to visualize this perfect
line touching the true edge at either poles of the circle.

And this is my question:

If we were trying to visualize the infiniteness of pi based on the
whole technique of using diameter to find circumference, is the
infiniteness of pi explained by the fact that we could never get the
diameter to intersect exactly with the circle because of its perfectly
symmetric curvature?

I mean, if you could imagine drawing a perfectly straight line through
the centre of a perfect square, it would not be hard to demonstrate
that this line has a definite beginning and end, and runs exactly
perpendicular to the two sides of the square, because they are flat.
So I'm sure you could work out a formula to demonstrate the ratio of
this line to the perimeter of this perfect square and show that this
ratio is a finite number.

Unless there is an error in my reasoning, it seems like the reason pi
is irrational is that the perfectly circular shape of the circle makes
it impossible to reach a conclusive measurement of its true  diameter,
and therefore the relation thereto to its circumference.

And I suppose since it can be proven that pi is infinite, the
hypothetical diameter of a perfect circle must also be infinite.  It's
amazing to think that an infinite line could exist within the closed
space of a perfect circle.  I bet the secret of the universe is

This may sound confusing, but I assure you I'm not crazy.
```

```
Date: 12/18/2003 at 09:58:41
From: Doctor Ian
Subject: Re: Theoretical Question about Circumference and diameter

Hi Mark,

>I understand that in a perfect circle (a circle which does not exist
>in reality) pi is irrational or infinite.  We can only demonstrate
>this mathematically. We could never demonstrate this with a real
>world circle using physical measuring devices.

Right.

>But let's imagine someone measuring the diameter of this perfect
>circle with a perfectly precise measuring device

Are you talking about a _physical_ measuring device?  If so, how can
you use a physical measuring device to measure a non-physical object?

To measure a mathematical object, we can only use other mathematical
objects.  That was the point of using polygons to make successively
closer approximations to the circumference of a circle.

>that could give
>infinite measurement if required. (This also assumes the line segment
>representing the diameter of the circle is perfectly straight and
>passes through the real centre of the circle.)

The only thing that can behave this way is an equation, or a geometric
construction, which amounts to the same thing.

>Since pi is infinite, we could never be able to visualize this
>perfect line touching the true edge at either poles of the circle.
>
>And this is my question:
>
>If we were trying to visualize the infiniteness of pi based on the
>whole technique of using diameter to find circumference, is the
>infiniteness of pi explained by the fact that we could never get the
>diameter to intersect exactly with the circle because of its
>perfectly symmetric curvature?

No, because we're not really using diameter to find circumference.
We're using various methods to find the circumference and the diameter
separately, and then finding pi by taking the ratio of circumference
to diameter.

Only after we're confident that the ratio of circumference to diameter
is constant for all circles can we go back and start using that ratio
to find circumference in terms of diameter, or vice versa.

Note that if we create a perfect mathematical circle, and create a
perfect mathematical line passing through its center, the line _will_
intersect the circle at two points.  Now, suppose we define the circle
in such a way that its circumference is a rational number, like 1.
The diameter will be

diameter = circumference / pi

a rational divided by an irrational... so as with pi, it won't be
possible to specify the diameter with a terminating or repeating
decimal expansion.  On the other hand, suppose we define the circle in
such a way that the diameter is a rational number, like 1.  The
circumference will be

circumference = pi * diameter

a rational times an irrational... so as with pi, it won't be possible
to specify the circumference with a terminating or repeating decimal
expansion.

It has less to do with the perfect curvature of the circle, or the
perfect straightness of the line, than with the properties of rational
and irrational numbers.

>I mean, if you could imagine drawing a perfectly straight line
>through the centre of a perfect square, it would not be hard to
>demonstrate that this line has a definite beginning and end, and runs
>exactly perpendicular to the two sides of the square, because they
>are flat.

Actually, you'd have the same problem.  To see why, imagine that we
have a square, and that we define the largest circle that will fit
inside it.  This circle touches the square at four points, i.e., the
midpoint of each side of the square.  A vertical or horizontal line
segment passing through the center of the square will also be a
diameter of the circle.  If one has a definite beginning and end, then
so does the other.

To make this more definite, let's say that the corners of the square
are at (1,1), (-1,1), (-1,-1), and (1,-1).  The circle is centered at
the origin and has radius 1.  The line segment from (0,-1) to (0,1) is
the kind of segment you're describing; and it's also a diameter of the
circle.  The length is very definite, and very rational: 2.  Similarly
for the line segment from (-1,0) to (1,0).

>So I'm sure you could work out a formula to demonstrate the ratio of
>this line to the perimeter of this perfect square and show that this
>ratio is a finite number.

The length of this line segment would be the same as the length of a
side of the square, right?  So what if I define the square so that the
length of each side is sqrt(2), or pi?  Then it's an irrational number.

I wonder if some of the problem that you're having with this comes
from your use of the word "infinite" to describe what are really
finite, irrational numbers.  Because the decimal expansion of a number
doesn't terminate or repeat, that doesn't mean that the number is
infinite.  The square root of two is a very finite number--it's the
length of the hypotenuse of a right triangle whose legs have length 1.
And we can specify it exactly by using square root or exponential
notation:

__    1/2
\| 2 = 2

We just can't write out the decimal expansion using a finite number of
digits.  But that's an issue of notation, not size.

An "infinite" number, on the other hand, would be something like the
size of the set of integers.  But we really use the term "transfinite"

>Unless there is an error in my reasoning, it seems like the reason pi
>is irrational is that the perfectly circular shape of the circle
>makes it impossible to reach a conclusive measurement of its true
>diameter, and therefore the relation thereto to its circumference.

Again, I'm not quite clear on what you mean by this.  There are
certainly interpretations of this statement that would make the
statement correct, but I don't know if those are consistent with your
understanding of what you're saying.

I think the main difficulty is probably that you're trying to import
the concept of direct measurement, which is a physical concept, into
mathematics.  We can't really measure things in mathematics, at least
not the way we do in the world, by putting things next to each other
and comparing the lengths.  The most we can do is define the
dimensions of some objects, and then use deduction to draw conclusions

>And I suppose since it can be proven that pi is infinite, the
>hypothetical diameter of a perfect circle must also be infinite.

Not at all.  When we create a circle, we get to _define_ its diameter,
or its circumference.  We can specify that the circumference of a
circle is 1 unit, or 5 units, or 12 times pi units, or whatever we
want.  Similarly, we can specify that the diameter is 1 unit, or 11
units, or sqrt(29) units.

However, having specified either the circumference or the diameter, we
know that the other is constrained by the _rule_ that the ratio of
circumference to diameter has to be pi.  That rule is what we use

But we can't just say "Let a circle have diameter 1 and circumference
5", any more than we can say "Let a square have a side length of 1 and
an area of 5".  Or rather, we _can_ say it, but only if we're willing
to be flexible about the shape of the space we're operating in!  But
in flat, Euclidean space, as soon as we specify that a particular
shape is a "circle" or a "square", we're saying that it has a
particular definition, and obeys particular rules.

>It's
>amazing to think that an infinite line could exist within the closed
>space of a perfect circle.

See above regarding this use of the word "infinite"... but in any
case, is it more amazing than thinking that an infinite number of
points can be contained in a finite line segment?

>I bet the secret of the universe is

I'll bet it's not.  :^D

The reason I'd make that bet is that what happens in mathematics has
no necessary connection to what happens in the world.  In mathematics,
we make up definitions, and then play with them to see what
conclusions we can derive from them.  These definitions are not
constrained by reality in any way.

For example, in mathematics, we can define sets of infinite size, like
the whole numbers: 0, 1, 2, 3, ...  If the universe is finite (which
it might be), then a set like this doesn't correspond to anything in
the universe.  But that doesn't mean we can't define the set, and play
with it.

Similarly, in mathematics, we can define continuous sets, like the
real numbers between 0 and 1.  Between any two numbers in this set, we
can always find another number.  If the structure of the universe is
discrete (which it might be), then this kind of continuity doesn't
correspond to anything in the universe.  But that doesn't mean we
can't define the concept, and play with it.

And so on, and so on.

As I said earlier, mathematics is really good for setting up analogies
to real world objects and processes.  For example, we might look at
something like a container of water, and decide that for all practical
purposes, the volume of the water in the container can be modeled
using a real number.  But if we start taking out half the water, then
half the remaining water, and so on, at some point we get down to a
single water molecule, right?  But that never happens with the
numbers.  We can just keep dividing numbers in half forever, always
getting more numbers.

Does this mean that there is something wrong with the mathematics, or
with our description of the world?  Not at all.  It just means that if
you push _any_ analogy hard enough, it will eventually break.  Where
people get confused is in forgetting that when we use math to describe
the world, we're _just_ making analogies.

>This may sound confusing, but I assure you I'm not crazy.

I don't think you're crazy.  I think that like most people, you were
taught to think that mathematics is somehow the "language of the
universe".  I was taught that, too.  But it's not true.  It's a
language used by people who are trying to describe the universe.

So I'd say that what you _are_ is ready to make a conceptual
breakthrough that never occurs to most people about the nature of
mathematics, and its relationship to the world.  And when you make
that breakthrough, I think you're going to find that mathematics is a
lot more interesting, and a lot more fun, than you've ever imagined.

- Doctor Ian, The Math Forum
http://mathforum.org/dr.math/
```

```
Date: 12/18/2003 at 12:24:03
From: Mark
Subject: Thank you

Well, that definitely clears up my question and I've had a conceptual
breakthrough.  Thank you for the insight.  I will never look at the
physical world the same way.  I now realize that at a certain level,
math becomes more art than science.
```
Associated Topics: