Skip to main content

Compute $\sin(\sin(...\sin(x)))$

Problem

Find the limit $\sin(\sin(...\sin(x)))$ for any real number $x$.

Solution

The problem can be precisely formulated as follows: Find the limit $L$ such that $$ L \equiv\lim_{n\rightarrow\infty}{f_n(x)}, $$ where $x\in\mathbb{R}$, $n\in\mathbb{N}$, and $$\begin{align} f_1(x) &= \sin(x)\\ f_n(x) &= \sin(f_{n-1}(x)). \end{align}$$ We observe that, if the limit $L$ exists, $$\begin{align} \lim_{n\rightarrow\infty} f_{n-1}(x) &= L\\ \lim_{n\rightarrow\infty} f_n(x) &= L, \end{align}$$ and thus, $$ L = \sin{L}, $$ which we show to be $L=0$ in the following.

First, we show that the limit exists. We accept the Axiom of Continuity of the Real Line, which says that any bounded monotonic sequence converges. Thus, to show that $L$ exists, we need only show that (1) $f_n(x)$ is a bounded sequence of $n$, and that (2) $f_n(x)$ is a monotonic sequence of $n$.

  1. $f_n(x)$ is clearly bounded because $f_n(x) = \sin{y}$, where $y\equiv f_{n-1}(x) \in \mathbb{R}$ and from trigonometry, $-1 \leq \sin{y}\leq 1$ for $\forall y\in\mathbb{R}$. Consequently, $f_n(x)$ is also bounded; in particular, $-1 \leq f_n(x) \leq 1.$
  2. We now show that if $f_1(x)\in[0,1]$, then $f_n(x)\in[0,1]$ and if $f_1(x)\in[-1,0]$, then $f_n(x)\in[-1,0]$. If $f_1(x)\in[0,1]$, then $\sin(f_2(x))\in[0,1]$ because $\sin{x}$ is an odd function and because of the preceding section. Likewise, for all $k\in\mathbb{N}$, $f_k(x)\in[0,1]$ implies $f_{k+1}(x)\in[0,1]$, which, by mathematical induction, means that $\forall n\in\mathbb{N}$, $f_n(x)\in[0,1]$. The same proof can be employed to show the case for $f_1(x)\in[-1,0]$.
  3. In the interval $[0,\frac{\pi}{2})$, $f_n(x)$ is monotonic decreasing because $f_n(x) = \sin(f_{n-1}(x))$ and $\sin(f_{n-1}(x)) \leq f_{n-1}(x)$, which follows from the geometry of the definition of $\sin{x}$ as shown in the figure, which clearly shows that $\sin{\theta} \leq h $ because $h$ is the hypotenuse of the right triangle whose one leg is $\sin{\theta}$, and $h \leq \theta$ because $h$ measures a line segment while $\theta$ measures an arc both of which have the same endpoints. Thus, $ \sin{\theta} \leq \theta$.
  4. In the interval $(-\frac{\pi}{2},0]$, $f_n(x)$ is monotonic increasing because if we let $y \equiv -f_n(x) $, then $y \in [0,\frac{\pi}{2})$, and the preceding section applies so that we can conclude that $$\begin{align} \sin y &\leq y\\ \sin (-f_{n-1}(x)) &\leq -f_{n-1}(x)\\ -\sin (f_{n-1}(x)) &\leq -f_{n-1}(x)\\ \sin (f_{n-1}(x)) &\geq f_{n-1}(x). \end{align}$$
  5. The last two sections imply that $f_n(x)$ is a bounded and monotonic sequence in each of the intervals $(-\frac{\pi}{2},0]$ and $[0, \frac{\pi}{2})$. This means that the limit $L$ exists in those intervals. Furthermore, the monotonicity impies that $L = \sin L$ is a one-to-one function in those intervals, which means that $L = 0$ is the only solution to $L = \sin{L}$. ■

References

Labels:

Comments

Post a Comment

Popular posts from this blog

2020 MEXT Japanese Government Scholarship Undergraduate Students Natural Sciences Qualifying Examination Mathematics (B): Problem 1(4)

This problem appears at the Qualifying Examinations for Applicants for Japanese Government (MEXT) Scholarships 2020 . There are two mathematics exams: one for biology-related natural sciences (Mathematics A), and another for physics- and engineering-related natural sciences (Mathematics B). This problem is from the 2020 Mathematics (B) questionnaire . The official answer key is here . Problem 1(4) The division of a polynomial function $f(x)$ by $(x-1)^2$ gives the remainder of $x+1$, and that by $x^2$ gives the remainder $2x+3$. Thus, the remainder of the division of $f(x)$ by $x^2(x-1)$ is $$ \fbox{ A }x^2 + \fbox{ B }x + \fbox{ C } . $$ Solution We need to find the remainder when $f(x)$ is divided by $x^2(x-1)$. Because $x^2(x-1)$ is of order $n=3$, the remainder will be of at most the order $n=2$, which means that it is of the form $Ax^2 + Bx + C$. The problem is now to find the coefficients $A,B$ and $C$ such that $$ f(x) = Q(x)\
Post a Comment
Read more

Reasonable Majority Rule: Wise Decision-making in a Democracy

Let the people decide what is good for them. On the one hand, this principle is at the core of democracy and manifests in the phrase "majority rule." On the other hand, the protection of the voice of the minority is another principle a functioning democracy must enforce. Democracy requires both. The fundamental assumption is that, after free and informed debate, the majority will be reasonable enough to judge ideas based on their merits (and not on their emotional relationship with the proponents). Because in a democracy, the Government's wisdom is an extension of the people's wisdom, wise decision-making is the duty of every citizen. The "goodness" of a decision or an idea does not necessarily depend on the number of its proponents but on the independent practical assessment of available information. The difficulty in balancing the interests of opposing sides is highlighted on occasions when the minority is unwilling to accept the decision of the
Post a Comment
Read more

2020 MEXT Japanese Government Scholarship Undergraduate Students Natural Sciences Qualifying Examination Mathematics (B): Problem 1(1)

This problem appears at the Qualifying Examinations for Applicants for Japanese Government (MEXT) Scholarships 2020 . There are two mathematics exams: one for biology-related natural sciences (Mathematics A), and another for physics- and engineering-related natural sciences (Mathematics B). This problem is from the 2020 Mathematics (B) questionnaire . The official answer key is here . Problem 1(1) The largest one among natural numbers that are less than $$ \log_2{3}\cdot\log_3{4}\cdot\log_4{5}\cdots\cdots\log_{2019}{2020} $$ is $ \fbox { A } $. Solution The key idea is the following relationship that holds for positive numbers $p, q,$ and $b$. $$ \log_p{q} = \frac{\log_b{q}}{\log_b{p}}, $$ where $p\neq1$ and $b\neq1$. For convenience, let $$ L \equiv \log_2{3}\cdot\log_3{4}\cdot\log_4{5}\cdots\cdots\log_{2019}{2020}. $$ Selecting $b=e$, we obtain $$\begin{align} L &=\frac{\cancel{\log{3}}}{\log{2}}\cdot \frac{\cancel{\
Post a Comment
Read more