3

THIS IS NOT A HOMEWORK QUESTION

Context
I am trying to simplify the expression $\sin{\frac{\pi}{14}}\times\sin{\frac{3\pi}{14}}\times\sin{\frac{5\pi}{14}}$.

I know from numerical approximations that it is equal to $\frac{1}{8}$ and I can prove this using complex numbers and Euler's formula.

However, when trying to use the product-to-sum identities, things get very messy (working shown below). Can anyone see where I have missed something important or where this can be re-directed?

Approach using product-to-sum identities
$\sin{\frac{\pi}{14}}\times\sin{\frac{3\pi}{14}}\times{\frac{5\pi}{14}}=\frac{1}{2}[\cos{\frac{2\pi}{14}}-\cos{\frac{4\pi}{14}}]\times\sin{\frac{5\pi}{14}}$

$=\frac{1}{2}[\sin{\frac{7\pi}{14}}+\sin{\frac{3\pi}{14}}-\sin{\frac{9\pi}{14}}-\sin{\frac{\pi}{14}}]$

Which seems to be going nowhere.

Grouping the terms and multiplying in a different order seems to lead to the same result.

Because I can see that the product has an exact value of $\frac{1}{8}$, and can be shown using Euler's formula, I would assume that it was possible to show this using product-to-sum and sum-to-product identities, so I am guessing that I have missed an obvious step.

If anyone can identify that obvious step that I have missed it is greatly appreciated.

Red Five
  • 3,164
  • 1
    Not sure about how to most convincingly turn it into a use of sum/product formulas. I would not shy away from Euler's formula. Using it we can show that the roots $a_1,a_2,a_3$ of the cubic $x^3+x^2-2x-1$ are $a_1=2\cos(2\pi/7)$, $a_2=2\cos(4\pi/7$) and $a_3=2\cos(8\pi/7)$. Combining that with $$x^3+x^2-2x-1=(x-a_1)(x-a_2)(x-a_3)$$ should give you enough identities to work with. – Jyrki Lahtonen May 06 '24 at 10:53
  • You can also try and start from here. Check out the threads linked to that one as well. – Jyrki Lahtonen May 06 '24 at 10:55
  • Approach0 is the way to find past occurences of a given formula. – Jyrki Lahtonen May 06 '24 at 10:58
  • 2
    Multiply by $\cos\frac{\pi}{14}$ as a catalyst. You get three double-angle formulas in turn. Each gives a factor $1/2$, and you are left with your original $\cos\frac{\pi}{14}$ – Empy2 May 06 '24 at 12:43
  • Thanks - will try that. – Red Five May 06 '24 at 20:13
  • It might be easier to write your product in terms of cosines; it's equal to $\cos\frac{3\pi}{7}\times \cos\frac{2\pi}{7}\times \cos\frac{\pi}{7}$. I'm not sure sum-product formulae are going to get you away from trig functions of multiples of $\frac{\pi}{14}$, though. – Chris Lewis May 08 '24 at 10:58
  • Thanks @ChrisLewis. This was one of my early ideas, including changing one, two or all three of the terms to cosines. It didn't seem to get very far. As you say, every time a nice angle was created, a not-so-nice angle remained. – Red Five May 08 '24 at 11:02

3 Answers3

5

Since $\sin A=\cos(\frac{\pi}{2}-A)$, we have $$\begin{align}&\sin{\frac{\pi}{14}}\sin{\frac{3\pi}{14}}\sin{\frac{5\pi}{14}} \\\\&=\cos\bigg(\frac{\pi}{2}-\frac{\pi}{14}\bigg)\cos\bigg(\frac{\pi}{2}-\frac{3\pi}{14}\bigg)\cos\bigg(\frac{\pi}{2}-\frac{5\pi}{14}\bigg) \\\\&=\cos\frac{3\pi}{7}\cos\frac{2\pi}{7}\cos\frac{\pi}{7}\tag1\end{align}$$

Multiplying $(1)$ by $\dfrac{8\sin(\pi/7)}{8\sin(\pi/7)}(=1)$ and using $2\sin A\cos A=\sin(2A)$ twice, we have

$$\begin{align}(1)&=\frac{8\sin\frac{\pi}{7}\cos\frac{\pi}{7}\cos\frac{2\pi}{7}\cos\frac{3\pi}{7}}{8\sin\frac{\pi}{7}} \\\\&=\frac{4\sin\frac{2\pi}{7}\cos\frac{2\pi}{7}\cos\frac{3\pi}{7}}{8\sin\frac{\pi}{7}} \\\\&=\frac{2\sin\frac{4\pi}{7}\cos\frac{3\pi}{7}}{8\sin\frac{\pi}{7}}\tag2\end{align}$$

Since $2\sin A\cos B=\sin(A+B)+\sin(A-B)$, we finally get $$(2)=\frac{\sin\pi+\sin\frac{\pi}{7}}{8\sin\frac{\pi}{7}}=\color{red}{\frac 18}$$

mathlove
  • 151,597
  • If $7x=\pi, 8x=x+\pi$ and $$\cos3x\cos2x\cos x=-\dfrac{\sin8x}{\sin x}=+\dfrac18$$ – lab bhattacharjee May 08 '24 at 17:06
  • 1
    Wonderful solution. Thanks. While searching for a way to simplify the working (if possible) I stumbled upon this gem: https://www.cambridge.org/core/journals/mathematical-gazette/article/abs/8862-grandmas-identity/AF41FE95D91E607A1FD4BF313E2664C3 which makes the result drop out once you turn the sines into cosines. – Red Five May 08 '24 at 22:00
1

The main ideas here are product to sum trig identities and telescoping sums.

Define for brevity $\,s_n := \sin(n\frac{\pi}{14}),\, c_n := \cos(n\frac{\pi}{14})\,$ and $\,S := s_1 \, s_3 \, s_5.\,$

Use product to sum trig identities and $\,s_7 = 1,\, s_{-n} = -s_n,\, s_{14-n} = s_n \,$ to get

$$ S = \frac14(s_7 + s_3 - s_9 - s_1) = \frac14 - \frac14(s_{-3} + s_1 + s_5). \tag1 $$

Use product to difference trig identities and $\,s_{n-7} = -c_n\,$ to get

$$ 2\,s_2\,s_n = -c_{n+2} + c_{n-2} = s_{n-5} - s_{n-9}. \tag2 $$

Apply this identity to equation $(1)$ and telescope the sum to get

$$ S = \frac14 - \frac14\frac{s_0-s_{-12}}{2\,s_2} = \frac14 - \frac14\frac{s_2}{2\,s_2} = \frac18. \tag3 $$

Somos
  • 37,457
  • 3
  • 35
  • 85
1

Trying to derive how the $\dfrac\pi{14}$ was chosen.

If $\sin8x=\sin x,\sin x(8\cos x\cos2x\cos4x-1)=0$

and $\sin x\ne0\implies x\ne n\pi \ \ \ \ (1)$ where $n$ is any integer

$8\cos x\cos2x\cos4x=1$

Case$\#1:$, $8x=2m\pi+x\implies x=\dfrac{2m\pi}7,$ by $(1), m\ne0, m\in[\pm1,\pm2,\pm3]$

$m=\pm1\implies\cos x\cos2x\cos4x=\cos\dfrac{2\pi}7\cos\dfrac{4\pi}7\cos\dfrac{8\pi}7=\cos\dfrac{2\pi}7\left(-\cos\dfrac{3\pi}7\right)\left(-\cos\dfrac\pi7\right)$

$m=\pm2\implies\cos x\cos2x\cos4x=\cdots=\left(-\cos\dfrac{3\pi}7\right)\left(-\cos\dfrac\pi7\right)\left(\cos\dfrac{2\pi}7\right)$

$m=\pm3\implies\cos x\cos2x\cos4x=\cdots=\left(-\cos\dfrac\pi7\right)\left(\cos\dfrac{2\pi}7\right)\left(-\cos\dfrac{3\pi}7\right)$

So, all the non-zero values of $m$ give us the same identity $8\cos\dfrac\pi7\cos\dfrac{2\pi}7\cos\dfrac{3\pi}7=1$

Case$\#2:$, $8x=(2m+1)\pi-x\implies x=\dfrac{(2m+1)\pi}9, 0\le m\le8$

By $(1), 9\nmid(2m+1)$

If $2m+1=9q\iff2(m-4)=9(q-1)\iff9\mid(m-4), m-4=9t$(say) for any integer $t$

So in $[0,8],m\ne4$

Observe that if $\dfrac{(2m_1+1)\pi}9+\dfrac{(2m_2+1)\pi}9=2r\pi\iff m_1+m_2=9r-1$ where $r$ is any integer, $\cos\dfrac{(2m_1+1)\pi}9=\cos\dfrac{(2m_2+1)\pi}9$

Setting $r=1, m_1+m_2=8$

$\implies m=0,8\implies\cos x\cos2x\cos4x=\cdots=\cos\dfrac{\pi}9\cos\dfrac{2\pi}9\cos\dfrac{4\pi}9$

$\implies m=3,5\implies\cos x\cos2x\cos4x=\cdots=\left(-\cos\dfrac{2\pi}9\right)\left(-\cos\dfrac{4\pi}9\right)\cos\dfrac{\pi}9$

$\implies m=2,6\implies\cos x\cos2x\cos4x=\cdots=\left(-\cos\dfrac{4\pi}9\right)\left(-\cos\dfrac{\pi}9\right)\cos\dfrac{2\pi}9$

If $2m+1=3p\iff2(m-1)=3(p-1)\implies 3$ divides $m-1, m-1=3p$(say)

$\cos\dfrac{3p\pi}9=\cos\dfrac{p\pi}3$

More trivially if $p=1\iff m=1,\cos\dfrac{(2m+1)\pi}9=?$

For $m=1,\cos x\cos2x\cos4x=?$

So, $m=0,8,2,6,3,5$ all lead to the Morrie's law

  • This is a wonderful answer - thanks. I've read about Morrie's law a few times now, but not sure I completely understand the implications of which this is one. The people who wrote these exam questions (over 100 years ago) certainly had some good insights. – Red Five May 12 '24 at 04:48