28

Why is $2\pi$ radians not replaced by $1$ turn in formulas?

The majority of them would be simpler. If such a replacement was proposed earlier, why was it declined?

Peter Mortensen
  • 674
DSblizzard
  • 923
  • 1
    I like "revolutions" as a measurement because I don't have to stoop to using degrees, but don't confuse people by using radians. (But, for mathematicians known to me, I will use radians, as they have desirable properties and come straight from the usual definitions of an angle) – Milo Brandt Mar 24 '16 at 23:33
  • 1
    "The majority of them would be simpler." The majority of what would be simpler? – fleablood Apr 15 '25 at 07:21

6 Answers6

45

Because if you use $1$ for the turn instead of $2\pi$ (forgetting what a radian is), you get that $\frac{\mathrm d}{\mathrm dx} \sin x = 2\pi\cos x$, which is probably not what you want, and many other problems, such as $\sin x$ not being a solution to $y''+y=0$ etc.

You can discuss whether the full turn is $2\pi$ or $\tau$ (tau), but you can't change the fact that it's equal to $6.2831853071\cdots$.

yo'
  • 4,667
  • 16
  • 30
  • 2
    I think a better link for tau would be www.tauday.com which quite accurately describes tau's superiority to pi. – corsiKa Mar 24 '16 at 19:59
  • Or if you want to link to Wikipedia, https://en.wikipedia.org/wiki/Tau_(mathematical_constant) would probably at least be a better target. – Ilmari Karonen Mar 24 '16 at 20:53
  • @IlmariKaronen Thanks! – yo' Mar 24 '16 at 21:58
  • What exactly is wrong with $\frac{d}{dx}\sin x = 2\pi \cos x$? – Stack Exchange Broke The Law Mar 25 '16 at 03:59
  • 2
    "You can discuss whether the full turn is 2π or τ (tau), but you can't change the fact that it's equal to 6.2831853071⋯." I strongly disagree. The radian is a derived unit. You can define angle to be measured with many arbitrary choices. We defined one full cycle of angle measurement to be $2\pi$. It is a natural choice only in that $\pi$ is an exceedingly pertinent parameter in the context of a circle. – zahbaz Mar 25 '16 at 04:47
  • 5
    @zahbaz Well, you can define the angle as you wish. My favourite definition of sine is that it's the solution to $y''+y=0$ for which $y(0)=0$ and $y'(0)=1$. Or that $\sin x = \frac{e^{ix}-e^{-ix}}{2i}$, where $e$ is such number that the derivative of $e^x$ at $0$ is $1$. Note that I as a mathematician cannot care less about SI and their units; sine does not take an angle as an argument, it takes a number. – yo' Mar 25 '16 at 06:29
  • 2
    So the question is: Do you want to use such a measure of angle that the mathematical sine is the same as the physical one? If yes, you can't but have 6.28 as the full turn. If no, you can do whatever you wish of course, but you're likely not simplifying things once you need any higher mathematics than pure Euclidean geometry involved. – yo' Mar 25 '16 at 06:38
  • Yes, I don't mean to come off that $2\pi$ is a purely arbitrary choice, only that it is essentially a choice, without which we'd have to redefine or scale all the aforementioned. It'd be a mess. Thanks for the response and examples. – zahbaz Mar 25 '16 at 06:59
17

There is a movement to make $2\,\pi=\tau$ a fundamental constant instead of $\pi$ (read The Tau Manifesto.) But not as far as I know to use $1$ as the measure of the whole circunference. One possible reason is that $\pi$ encodes the relation between the radius an the length of a circumference: a circular sector of radius $r$ and angle $\alpha$ radians has a length equal to $\alpha\,r$; it would be $2\,\beta\,\pi\,r$, where $\beta$ is the measure of the angle in the units you propose. Moreover, you cannot avoid $\pi$ in formulas like $$ e^{\pi i}+1=0. $$

Julián Aguirre
  • 77,704
  • 12
    And the final expression could then be $e^{\tau i} = 1$ – Henry Mar 24 '16 at 12:10
  • @Henry Yeah, and that one is missing the important thing: the original includes all 3 basic operations, together with the 5 constants: addition, multiplication, and power. You're missing one constant and one operation ;) – yo' Mar 24 '16 at 13:48
  • 13
    @yo': $e^{\tau i} = 1 + 0$ if you insist on having it. There's no reason to have the "$+ 1 = 0$" rather than "$= -1$" in the original. – Deusovi Mar 24 '16 at 13:53
  • 3
    Or $e^{i\frac{\tau}{2}}+1=0$, which includes 4 basic operations (now with division) and not 5 but 6 of the most important numbers in all mathematics. – N. Virgo Mar 24 '16 at 14:15
  • 2
    Aesthetic pickiness aside, the equation $e^{i\tau} = 1$ is strictly weaker than $e^{i\tau/2} = -1$. The former is consistent with $e^{i\tau/2} = 1$, too. – Ryan Reich Mar 24 '16 at 23:40
  • 2
    @RyanReich And you can say the same about $e^{i\pi}=-1$ as opposed to $e^{i\pi/2}=i$, since it is also consistent with $e^{i\pi/2}=-i$. And so it goes all the way down; you won't be able to pin down the $e^{ix}$ function using single point equations. – Mario Carneiro Mar 25 '16 at 02:43
  • 1
    Yeah, but no one currently cares about $e^{i\pi/2}$, whereas we are talking about replacing an existing, popular equation with a less impressive one. – Ryan Reich Mar 25 '16 at 03:46
  • Ah, and the most important point: $e^{i\pi} = -1$ is surprising both because of its use of imaginary numbers and because it exhibits a negative value of the exponential function. You lose a lot with the $+1$ version. – Ryan Reich Mar 25 '16 at 22:24
  • @RyanReich It's conceptually half a turn (in the complex plane), which would be emphasized by the use of $\tau/2$. – user76284 Apr 18 '19 at 19:21
6

In fact, "radians" are not a unit like meters or second, so you can't rescale them to make $2\pi$ become $1$ (as sometimes you do in physics, rescaling e.g. meters to make $c=1$).

As an aside, degrees are an absolutely artificial concept, and you should try to never use them.

Daniel Robert-Nicoud
  • 30,663
  • 1
    From what I've heard from my physicist colleagues, they frequently assume that all constants are equal to $1$, including $\pi$. – tomasz Mar 24 '16 at 11:12
  • @tomasz Certanily not $\pi$ by itself! However, if you have something like $\cos(2\pi t)$, then you can rescale the time $t$ to have simply $\cos(t)$. – Daniel Robert-Nicoud Mar 24 '16 at 12:54
  • 7
    "From what I've heard from my physicist colleagues, they frequently assume that all constants are equal to 1" is just purely wrong. What physicists do is to re-scale measurement processes so that some measured units end up having a different numerical value with respect to the original units used for the measurement itself. Whoever upvoted that comment should be truly prosecuted. – gented Mar 24 '16 at 13:53
  • 5
    As a matter of fact, physicists do occasionally set pi equal to one (as a very coarse approximation rather than as a definition) - see Fermi estimation. – N. Virgo Mar 24 '16 at 14:18
  • 8
    "you should try to never use them" unless you want to communicate with the other 99% of the population! – Oddthinking Mar 24 '16 at 15:21
  • 2
    @Nathaniel Even as a very coarse approximation, it's probably not wise: $\pi^2\approx 10$, so if you have multiple factors of $\pi$ you're at risk to start losing orders of magnitude. – Semiclassical Mar 24 '16 at 16:31
  • 1
    @Oddthinking Well, I admit that my point of view might be biased by the fact that I'm a mathematician... – Daniel Robert-Nicoud Mar 24 '16 at 17:40
  • 4
    @Semiclassical This would be in a context where a few orders of magnitude lost here or there are not a problem. Like estimating how many stars are in the universe, or the field strength of a graviton (where the real answer is "way too weak to ever detect" and a few orders of magnitude won't change that). – David Z Mar 24 '16 at 17:54
  • 1
    Are degrees truly artificial? I always assumed they were based on an approximation of the number of terrestrial rotations per revolution. Which actually makes them quite natural, albeit very biased towards the terrestrial year. – JAB Mar 24 '16 at 18:36
  • Supposedly Ramsey taught a course for a while which he began by saying "in this course we'll use units where $\hbar = c = 1 = -1 = \pi$," though I think the last step was probably a joke. I do frequently use $\pi = 3 = \sqrt{10}$ – rob Mar 24 '16 at 21:22
  • 2
    Having known about radians for the last 20 years and holding a PhD in math, I still think in degrees first. They exist for a good practical reason. – Ryan Reich Mar 24 '16 at 23:42
  • @JAB defining something in terms of properties of a particular planet, or even universe, is not considered natural in math, and this can also essentially be said for physics, at least with regards to planets. – Matt Samuel Mar 25 '16 at 04:04
4

Radians are the natural, dimensionless choice of unit to measure angle. We could certainly define the full turn as $1$. That's a nice, wholesome unit. It's nicer than pushing our historical comfort with base $60$ with $360^\circ$ or dealing with cumbersome $400 \text{ gon}$. But does $1$ really reflect the system we're trying to describe?

In choosing an appropriate unit, we should ask "What is an angle measure?" One appropriate definition is the measure $\theta$ of the angle subtended by an arc length $s$ of a circle of radius $r$. This definition, though, needs some work to become a practical tool. Given the obvious relationship between an angle and a circle, we can call upon Euclidean geometry to find ourselves a nice parameter. Behold the ratio of circumference to diameter: $C/d=\pi$! Now we observe that the relationship between arc length and circumference goes as

$$s = (\text{some fraction})\times C = \dfrac{\theta \times C}{\theta_{turn}}$$

We're getting somewhere. Let's adopt a convenient parametrization of angle measure, something we can throw numbers at and interpret easily. Perhaps a convenient choice would necessitate a convenient circle. Well, the unit circle has a nice radius and area. Its circumference is $2\pi$. If the distance around the unit circle is $2\pi$, then why don't I adopt this for my unit of angle measure? I'll define the angle measure $\theta_{turn}$ that takes me around the circle to be $2\pi$.

Arc length is now cleanly given by

$$s = r\theta $$

That's the definition of radian in that last line. Why is it a superior choice? Any other choice of units would have left over some other constants. With the choice of $2\pi$, we have a clean, relevant, and dimensionless unit.

zahbaz
  • 10,736
3

Just a complement to zahbaz answer.

With angles in radian, you get the nice formulas:

sin' = cos
cos' = -sin

sin(x) = 1/2 * (exp(ix) - exp(-ix))
cos(x) = 1/2 * (exp(ix) + exp(-ix))

Of course those are mathematical considerations, and that's the reason why in common life we use degrees.

But there are other units. In artillery, the thousandth (millième in french) is used. You have 6400 of then in one turn (why not...) but as an approximation, it gives an elevation of 1m at 1km of distance. If you have this in binoculars and can guess the height of something, you can easily compute a distance - even a military can...

Serge Ballesta
  • 1,888
-2

Histrocially you would use look up tables to approximate functions, and $sin$ and $cos$ where the tables you where given (propably for good reason, due to their nice derivatives). Tables for $x \mapsto \sin(2\pi x)$, $x \mapsto \cos(2\pi x)$ would have been redundant.

Radians are just an unit for a usless intermediate calculation. You probably want to work in turns and evaluate the composite functions $x \mapsto \sin(2\pi x)$, $x \mapsto \cos(2\pi x)$, $x \mapsto e^{2\pi i x}$ in one go.

Tom Huntington
  • 125
  • 1
    Angular velocity in rad/s plays very nicely with tangential velocity. And I wonder, have you ever used trig tables? – David K Apr 15 '25 at 05:09