# When Is a Straight Line Not the Shortest Distance between Two Points?

The hypotenuse of a right triangle is not always the shortest distance between the two points that define it

In the Pythagorean theorem episode of NBC Learn's "The Science of NFL Football," you see that a defender in the middle of the field must take the proper angle of pursuit to catch a ball-carrier making a dash down the sideline for the end zone.

In chasing down the ball-carrier, the defender basically runs along the diagonal of a right triangle, in which the sum of the squares of the sides equals the square of that diagonal. You might know this relation, discovered by the 5th-century B.C. Greek mathematician Pythagoras, as a² + b² = c².

The "c " is the hypotenuse, and although it represents the longest side of a right triangle, it is the shortest path between the two points on either end. If the points on the triangle were places to visit in a city, you probably wouldn't bother walking along a and b if you could directly take c.

But the hypotenuse isn't always the shortest route. In fact, it is only the shortest one on football fields and other flat surfaces. On spheres and other shapes, it may not be.

You can see this distinction if you draw a right triangle on a globe. First, let's pick a city on the equator--for simplicity, say it's Quito, Ecuador, on the Pacific coast of South America. From Quito, trace a longitude line to the north pole; then make a 90-degree turn to the right and head straight back down. At the equator, you'll notice a city nearby called Libreville, the capital of the country of Gabon in Africa.

Now draw a line along the globe's surface starting at Quito and going toward Libreville. You probably went eastward, passing over Brazil and the Atlantic Ocean. Indeed, this hypotenuse, traversing one quarter of the globe, marks the shortest distance. But that is not the only hypotenuse.

Mathematically speaking, you would still have a right triangle if you went westward from Quito, circumnavigating Earth along the equator to get to Libreville. The hypotenuse in this case is three fourths of the circumference. It would have been shorter to travel from Quito to the North Pole and then down to Libreville.

The Pythagorean theorem only works on two-dimensional surfaces like football fields; mathematicians refer to such surfaces as Euclidean geometry (named for Euclid, the 3rd-century B.C. Greek mathematician). The theorem fails for non-Euclidean geometries, such as spheres and more complex geometries like saddles. Indeed, all the rules you learned in school, like parallel lines staying parallel, only refer to Euclidean geometry. In the non-Euclidean universe, parallel lines may actually diverge or converge.

Although non-Euclidean geometry may seem exotic and unfamiliar, it is actually common in many fields of science--perhaps most notably, in Einstein's general relativity theory, in which gravity can bend the shape of space and time.

View
1. 1. Cardigan 12:23 PM 11/8/10

Jolly good. Here's another conundrum on akin to zeno's. If I indent the hypotenuse of the 3-4-5 triangle with a set of small steps each parallel and in proportion to the 3-4 sides, the sum of their horizontal and vertical parts is still 7 units. As I make the steps smaller their sum is still 7. From a long distance they merge into the hypotenuse. At which point does the sum of the digital/quantum steps become 5?

2. 2. reflectogenesis in reply to Cardigan 02:26 PM 11/8/10

Is this true of triangles on spheres. What relationship holds there - and on saddles. Does it imply something for space time?

3. 3. reflectogenesis in reply to Cardigan 02:45 PM 11/8/10

Could something like this happen where the magnetic and electrical field of a particle were normal and the direction of travel of the particle was travelling within a dimension which didn't take the fields with them proportionally?

4. 4. sparcboy 03:53 PM 11/8/10

From a long distance they merge into the hypotenuse. That is a visual effect. The actual measurements don't change...do they...or did I miss your point?

5. 5. solspot 03:57 PM 11/8/10

Many predators (eg. wolves, bears catching salmon, your dog catching a frisbie) inherit or learn the behavior of mental estimation which results in successful pursuit. Perhaps someone has (or will) discover the "Pythagorean gene" that makes this mental math so intuitive?

6. 6. Karl Johanson 05:21 PM 11/8/10

You can't draw a triangle on the surface of a sphere. A triangle is composed of straight segments. Connecting three points along the surface of a sphere gives you 3 dots connected by curved segments. Even if there is a right angle at one of the points, there is no hypotenuse, as there is no triangle (so let's call it a "pseudo-hypotenuse"). To connect 3 points with straight lines to form a triangle, on the surface of a sphere, you need the lines to go through the sphere, not stay on the surface. You will find that the shortest distance between the points is... a straight line.

The point about the 'pseudo-hypotenuse' on the surface of a sphere not necessarily being the shortest route, is irrelevant to the title of the article, because it is NOT a straight line. It’s a curved line.

You’ve come up with a complicated way of saying “a straight line isn’t always the shortest distance between two points, if the line in question isn’t actually straight.”

7. 7. tharter in reply to Karl Johanson 06:49 PM 11/8/10

This is only true in "R3", a three dimensional space. A curved 2d space (R2) OTOH is what the author was discussing. In this case the SPACE is curved, but the lines are straight, a straight line being defined as the shortest distance between two points in that space. R3 and higher dimensional spaces can also have curvature, which is analogous to those seen in R2. A positively curved space is like a sphere, finite in extent but unbounded. Negatively curved spaces are often said to be saddle shaped in R2 and are either infinite or bounded.

You can easily reconstruct geometry in these kinds of spaces too and propose axioms analogous to those of the familiar Euclidean (flat) geometry. These geometries are just as internally consistent and valid as Euclidean geometry. For instance geometry on the surface of the Earth confirms that in a positively curved space triangles angles sum to less than 180 degrees.

8. 8. Dr. Strangelove 08:55 PM 11/8/10

The shortest distance between two points is a geodesic. In two-dimensional space, a geodesic is a straight line. Euclidean geometry is geometry in two-dimensional space. It is surprising that non-euclidean geometry was developed only in the 19th century by Gauss and Riemann bec. Ptolemy in the 2nd century already knew spherical trigonometry which is actually 3-dimensional (non-euclidean) geometry.

9. 9. rajnish 02:40 AM 11/9/10

What is straight line anyway? One can easily make a triangle on a (FLAT?) surface with curved lines whose sum is 180 degrees. And what is flat? One can also have a coordinate system in which a single dimension is is represented by a curved line. On surface of earth if one digs a tunnel straight line will still be the shortest distance. In space straight line is still the shortest distance but general relativity does not allow us to travel that path and no one has as yet proved the general relativity to break. We therefore accept the notion that light travels the shortest distance in free space. That is how we define the shortest distance and give physics precedence over mathematics(Euclid). Curvature of space time is a concept and then one should also consider space and time are relative and differ from observer to observer i.e. every you and me has his own space time.

10. 10. Dr. Strangelove in reply to rajnish 04:09 AM 11/9/10

In Euclidean geometry, a line is always straight. If it is not straight, it is not a line. It is a curve. This definition is consistent with analytic geometry - a line is represented by a linear equation. It is always straight. A curve is represented by non-linear equations such as quadratic, cubic, quartic and polynomials. Well, a curve is not a line. A straight line is redundant like a spherical ball.

There is no straight line in space bec. space itself is curved. Even light follows the curvature of space. Light travels the shortest distance, a 3-dimensional geodesic, but it's not a (straight) line.

11. 11. Karl Johanson 01:25 PM 11/9/10

@ tharter, who said: "...a straight line being defined as the shortest distance between two points in that space."

Interesting, but with that definition, the title of the article is still irrelevent.

12. 12. jtdwyer 02:08 PM 11/9/10

And I suppose the surface of a football field is curved because it follows the surface of the spherical Earth? More likely it was leveled using the shortest distance between two points.

Fortunately, a football player can ignore all the superfluous constraints and, unlike the illustration, begin his pursuit of the receiver several yards away from the line of scrimmage or following an initial collision. Representative generalizations seldom occur in reality.

13. 13. jtdwyer in reply to jtdwyer 02:23 PM 11/9/10

Btw, perhaps this inadequate academic idealization of real world problem analysis is the reason why so many pursuers end up tackling the spacetime behind Chris Johnson...

14. 14. Dr. Strangelove in reply to rajnish 08:38 PM 11/9/10

A line is the shortest distance between 2 points in a flat manifold of any n dimensional space. In a curved manifold, the shortest distance is of course a curve. In curved manifolds, a triangle does not have 180 deg. Flat means zero curvature.

In principle, we cannot draw a straigth line or a triangle with 180 deg. bec. our universe is a curved manifold. But the universe is so big compared to earth that all our measurements cannot detect the deviation.

We can do Euclidean geometry on a piece of paper bec. the paper is so small compared to the earth's surface. The error is undetectable. But if your paper is as large as North America, you can detect the error. Your triangle does not give 180 deg. It's non-euclidean geometry.

Same is true with earth and the universe. The latter has a radius of curvature >70 billion light-yrs. Most astronomers think the universe is a 3-sphere (4-dimensional sphere). The famous expanding balloon model is a 3-sphere. The volume of a 3-sphere with 70 gly radius is 6,770,580 billion cubic light-yrs. This is well beyond the visible universe. So the universe is not infinite but it's REALLY BIG.

15. 15. rajnish 11:59 PM 11/9/10

In curved space time angles of triangle may not always be 180 degrees but it is possible to have triangle with 180 degrees and that will also follow cosine formula. Now the question is how can one define straight or non-straight(curved) in presence of force like gravity. Further what is a dimension anyway? One can represent a curved dimension with a single variable. Forget Euclidean and Gaussian what is their in name, if you call an owl a cat it makes no difference as far as their existence, their physical property and every thing else is concerned.

16. 16. Dr. Strangelove in reply to rajnish 01:58 AM 11/10/10

In spherical trigonometry, triangles follow the spherical law of cosines but they never equal 180 deg.

Globally, spacetime is curved. This curvature may be too small to be detectable locally. Spacetime may look perfectly flat locally. The presence of mass distorts spacetime. This is what we call gravitational field. The distortion is a pronounced curvature of local spacetime.

Dimensions are length, width and height in 3-dimensional geometry. A curve, just like a line, has length. But a curve is not a line. They have different properties. A curve has variable slope. A line has constant slope. They are represented by different types of equation.

We have to study Euclid and Gauss bec. geometry has already been invented. It's easy to get lost trying to reinvent it.

17. 17. rajnish 04:51 AM 11/10/10

Suppose I am near a black hole and I shoot photons or light beam parallel to the horizon. Now the beam will return to me after fraction of second. Now without moving an inch does that mean I am travelling in time. Similarly if I could reach a place faster than light with same initial position in space time for both me and light does that mean I am travelling in time.

18. 18. rajnish 04:58 AM 11/10/10

Anyway I will have to use a force to balance the effect of gravity and I suppose my spaceship uses thermonuclear energy to do that.

19. 19. rajnish 05:06 AM 11/10/10

Regarding a slope of geodesic path in isolated space on which I am moving I can never determine that slope I will always move straight.

20. 20. rajnish 08:10 AM 11/10/10

And Dr.Strangelove why are you obsessed with spheres. Geodesic surfaces can be curved, zig-zag, concave-convex with each other in space. Had Mr. Einstein been alive I would have surely discussed this with him.

21. 21. rajnish 08:18 AM 11/10/10

Oh one more thing I was talking of cosine formula with Pythagorean theorem as special case and not direction cosines.

22. 22. Dr. Strangelove in reply to rajnish 07:41 PM 11/10/10

If you travel in space, you also travel in time bec. spacetime is inseparable. That's special relativity.

You can calculate the slope of geodesics without leaving the manifold if you know the radius of curvature and shape of the manifold. That's what cosmologists do. They make calculations about the size and shape of the universe without leaving the universe.

A generalized Pythagorean theorem is applicable to flat manifolds. It defines the length of a line in flat manifolds of any n dimensional space.

I used spherical triangles bec. they are the curved manifold easiest to visualize. Higher dimensional curved manifolds are impossible to visualize but their properties are computable.

I'm sure Mr. Einstein would love to talk to you.

23. 23. HILBILSTILCHEM101 05:46 PM 11/11/10

Only in slow moving things are there 2 points.
As soon as the things move faster, 2 thing are
1 thing and visa versa.

Remember that there is a major fundamental difference between SPACE (distance) and TIME.

When 2 points cool to theoretical absolute zero,
this does no mean there is no TIME.

Be sure what quantity (mixed) you are measuring TIME or SPACE (distance) before there is any type of math and geometrical area between the 2 so called points.

Conjecture, at ZERO TIME there is no SPACE between
2 points.

24. 24. HILBILSTILCHEM101 05:58 PM 11/11/10

Look at this simple problem. Quantity 1 TIME and Quantity 2 SPACE or distance.

At absolute zero degrees Kelvin, there is still TIME.

What space is between 2 points at absolute zero degree Kelvin is there?

The space between 2 points is obviously relative to the heat or energy between 2 points.

Classical mechanics turns thru themodynamics into quantum mechanics where at where at 1 point there
is only probablasitic SPACE TIME between 2 points and
elswhere, there is no SPACE (ecludian distance between 2 points) yet TIME is still existence and has not changed, to longer or shorter.

They teach us this is Physics 101. What is the mystery?

25. 25. jtdwyer in reply to HILBILSTILCHEM101 07:31 AM 11/12/10

I skipped Physics 101, so you'd have to explain to me in more detail how space disappears at absolute zero. I cannot accept that assertion at face value. In a continuously expanding universe, its temperature would eventually reach absolute zero. So at that moment space would disappear?

26. 26. Dr. Strangelove 10:56 PM 11/12/10

The article is misleading and inaccurate. In non-euclidean geometry, a straight line is defined as a geodesic, the shortest distance bet. 2 points in any n dimensional space. So by definition, a straight line is necessarily the shortest distance. Otherwise, it is a curve.

In the example given, the hypotenuse of the right triangle is not the shortest distance bec. it is a curve, not a line. We have to define what are lines and curves in spherical trigonometry, which is non-euclidean geometry.

On earth, longitudes are geodesics (lines) and latitudes are non-geodesics (curves). This may not be obvious but if you pull a string bet. 2 points on earth's surface to get the shortest distance, the string will trace a longitude and deviate from a latitude. Btw, longitudes are non-parallel lines that pass through the poles wherever they are located on the sphere.

To see why geodesics are lines and non-geodesics are curves, we have to use projective geometry. If you do perpendicular projection longitudes on a plane surface, longitudes trace a straight line on a plane surface and latitudes trace a curve.

In short, if you draw a right triangle on a sphere, its sides are not the shortest distances bet. the 3 points bec. the sides are not lines but curves. They may look straight but they are not. To get a triangle whose sides are lines, you have to draw a spherical triangle whose angles are >180 deg., which is not a right triangle whose angles are =180 deg.

The fundamental issue here is not whether a line is the shortest distance bet. 2 points but what appears to be a line in curved surfaces may actually be a curve that only looks straight. The shortest distance is a geodesic that corresponds to a straight line when projected to a plane geometry.

27. 27. TRPartee 02:09 AM 9/16/12

The shortest distance between two points is 0.
Why put anything like a line or curve in the way or 2 points?
Your not debating the "Shortest" distance , you're debating the most direct route.

28. 28. TRPartee 02:22 AM 9/16/12

On second thought the most direct route would also be 0. But you would have to bend the plane between to merge the 2 points, to remove the distance between the 2 points. ;-)

You must sign in or register as a ScientificAmerican.com member to submit a comment.
Click one of the buttons below to register using an existing Social Account.

## More from Scientific American

• Plugged In | 9 hours ago

### What unconventional fuels tell us about the global energy system

• News | 11 hours ago

### Shooting the Wheeze: Whooping Cough Vaccine Falls Short of Previous Shot’s Protection

• Guest Blog | 11 hours ago

### Dissecting the controversy about early psychological response to disasters and trauma

• Ask the Experts | 16 hours ago | 13

### What Role Does Climate Change Play in Tornadoes?

• Reuters | 16 hours ago | 1

More »

## Latest from SA Blog Network

• ### Fat Tuesday: Hungry for love

MIND
Illusion Chasers | 8 hours ago
• ### What unconventional fuels tell us about the global energy system

Plugged In | 9 hours ago
• ### Dissecting the controversy about early psychological response to disasters and trauma

Guest Blog | 11 hours ago
• ### Why Are Barns Red?

Image of the Week | 12 hours ago
• ### The SciArt Buzz: ScienceArt On Exhibit In May/June 2013

Symbiartic | 13 hours ago

## Science Jobs of the Week

When Is a Straight Line Not the Shortest Distance between Two Points?

X

### Subscribe Today

Save 66% off the cover price and get a free gift!

X

X

###### Welcome, . Do you have an existing ScientificAmerican.com account?

No, I would like to create a new account with my profile information.

X

Are you sure?

X