Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How did people determined the values of trigonometric functions before calculators, like e.g. $\sin 37^\circ$ up to five decimal places? Was that possible to find before series were invented?

share|cite|improve this question
They probably just used the Power series expansion of the trigonometric functions. – Mustafa Said Apr 17 '14 at 11:43
You can do a surprising amount with the multiple angle formulas. – Dan Rust Apr 17 '14 at 11:48
In the era before pocket calculators, we used to use a combination of trigonometric tables ( and slide-rule for interpolation in these tables. – Axel Kemper Apr 17 '14 at 11:49
$\sin x = x - \frac13 x^3 + \frac15 x^5 - \frac17x^7 + \cdots$ – wchargin Apr 17 '14 at 17:53
You could use unit circles and a ruler. – PyRulez Apr 18 '14 at 0:13
up vote 27 down vote accepted

Before computers, trigonometric functions were something you looked up in tables. The computation of these tables was a lot of work, and often tables were copied and reused for centuries after they were computed, until eventually more precision was needed and someone had to start it all over from scratch.

The basic tools for constructing trigonometric tables from Hipparchus up to the computer age were formulas that allow you to find the values from smaller angles if you already know them for larger ones. In particular, the half-angle formula $$ \cos \frac{\theta}{2} = \sqrt{\frac{1+\cos\theta}{2}} $$

Then, starting from $\cos 90^\circ=0$ you can successively find the cosines of $45^\circ$, $22^\circ30'$, $11^\circ15'$ and so forth, down to as small an angle as you'd like your precision to be. Then you get the corresponding sines by $\sin\theta = \sqrt{1-\cos^2\theta}$.

From there, you can build up the value for each degree value by using the angle addition formulas -- for example, the values for $67^\circ 30'$ can be gotten by combining the values for $45^\circ$ and $22^\circ30'$, and in this way you can fill in the gaps in your table, to any desired precision, given enough labor.

(The ancients thought in terms of chords rather than sines and cosines, and their algebra was set up slightly differently, but the basic idea is the same).

Doing pure halvings leaves you with your base values being sines of rather clumsy angles. Various tricks to get around this were developed. For example in the 1400s Islamic mathematicians developed a method where you first compute the sine of $3^\circ$ exactly (my source doesn't say how, but possibly by using the angle-difference formulas on the sines and cosines of $72^\circ$ and $60^\circ$, which can be derived by considering pentagrams and equilateral triangles, and then halving twice). From there the sine of $1^\circ$ is found by solving the cubic equation $$ \sin 3^\circ = 3\sin 1^\circ - 4\sin^3 1^\circ$$ numerically, and then you can build up your trig table in whole-degree steps as before.

share|cite|improve this answer
i like this because it highlights how the ancients could pick arbitrary high levels of accuracy without the use of power series. (although eventually the amount of pen and paper work would be maddening) – frogeyedpeas Apr 17 '14 at 16:08
The part "For example around 1400 Islamic mathematicians" should include commas like "For example, around 1400, Islamic mathematicians" to clarify the meaning. Otherwise, it sounds like you're talking about 1400 mathematicians. – ThomasW Apr 18 '14 at 4:33
it's true that just with halvings you end up with clumsy angles, but it's also true that since $\sin(x) ~ x$ for $x$ small, you may interpolate, losing some precision. Luckily you may then sum the interpolated angles and find the error wrt the precise value already computed, so you have a feedback. Since you need just a single angle to start with, it's doable – mau Apr 18 '14 at 8:15

You could also compute it from the series. Not that this is better than the accepted answer, but just for completeness.

share|cite|improve this answer
Note, however, that the power series is less useful if you want to produce tables for angles denominated in degrees, because then you also need a precise value of $\pi$ in order to convert them to radians. – Henning Makholm Apr 17 '14 at 19:34

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.