5

I had recently asked a question about a conjecture in a triangle, and it led me to another question.

In the non-equilateral triangle below, $CD, AD$, and $BD$ concur at point $D$. The angles $(a, b, c, d, e,$ and $f)$ in degrees are also placed as given in the figure below.

enter image description here

Define a solvable triangle such that: For a triangle like the one in the figure above, given two pairs of adjacent angles (say $a,b$ and $c,d$), if the remaining two angles (say $e,f$) can be found by a synthetic solution using elementary geometry, it is a solvable triangle. (Just to be clear, I have added an example at the end.)

My questions are:

  • Are all triangles containing integer angled sextuplets solvable triangles?

  • What is the criterion for being a solvable triangle?


Such a sextuplet must satisfy the Trigonometric Ceva's Theorem:

$$\frac{\sin(a\cdot\frac{\pi}{180})}{\sin(b\cdot\frac{\pi}{180})}\cdot\frac{\sin(c\cdot\frac{\pi}{180})}{\sin(d\cdot\frac{\pi}{180})}\cdot\frac{\sin(e\cdot\frac{\pi}{180})}{\sin(f\cdot\frac{\pi}{180})}= 1$$

We also know that,

$$a+b+c+d+e+f=180°$$

In the comments of their accepted answer to my previous question, @user provided a sextuplet $\frac\pi{180}\{1,30,87,2,29,31\}$ with different integers to work on -I am also sure that there more such sextuplets which consist of different integers.

Here is an example problem using this sextuplet (there are more such triangles that can be formed with this sextuplet):

Let $\measuredangle ACD=31°$, $\measuredangle BCD=30°$, $\measuredangle CAD=1°$, $\measuredangle BAD=2°$ in $\bigtriangleup ABC$. Find $\color{green}{\text {$\measuredangle ABD$}}$.

Example Synthetic Solution:

Since the angles are extremely small, I have reproduced the diagram by rescaling the angles (blue line segments represent the original problem and red line segments represent my constructions).

Let’s select a point $E$ on $AD$ such that $AE=CE$ and connect $E$ and $C$, so $\measuredangle ECA=1°$.

Let’s extend $BC$ and select a point $F$ on $BC$ such that $CE=CF$, constructing the equilateral triangle $\bigtriangleup CEF$.

Angle chasing leads us to: $$\measuredangle DEC=2° \; \text{and} \; \measuredangle DEF=58°$$

Let’s connect points $F$ and $D$. Since $CD$ is the angle bisector of the equilateral triangle $\bigtriangleup CEF$, using congruency: $$\measuredangle DFB=2° \; \text{and} \; \measuredangle DFE=58°$$

Since $\bigtriangleup AEF$ is isosceles, $\measuredangle AFE=29°$.

Since $\measuredangle DFB=\measuredangle BAD=2°$, $AFBD$ is a cyclic quadrilateral, which allows us to conclude that:

$$\color{green}{\text {$\measuredangle AFD=\measuredangle ABD=87°$}}$$

I think the $30°$ angle is the critical angle that allowed this solution.

$\\$

enter image description here

The synthetic solution seems to be difficult when $a, b, c, d, e,$ and $f$ are all different integers.

I am not very sure what the appropriate tags are for this question.

Thanks in advance.

krazy-8
  • 2,294
  • So here's a puzzler for you: obviously, given $a$, $b$, $c$, and $d$, what does it mean to construct $e$ and $f$? Obviously we can construct them - just give yourself $AC$, draw $ABC$ and $ADC$ given the angles (as we can transfer angles using compass and straightedge) and connect $BD$, and, behold, there they are. Of course, given this, we actually also get trigonometry: if we have an angle, we can generate its trigonometric ratios simply by constructing a right triangle with it. – Dan Uznanski Apr 11 '21 at 15:34
  • Does "construct any type of polygon on the figure" mean that given a segment $AC$ you can construct a regular $n$-gon for any $n$ with side $AC$? If so, then you can construct any angle measured by a rational number of degrees. Obviously this would be a definition of "constructible" that is different from the usual compass-and-straightedge definition. – David K Apr 11 '21 at 15:37
  • It's possible the OP had in mind something like "prove that the other two angles have integer value in degrees, with a purely geometric reasoning". – Intelligenti pauca Apr 11 '21 at 15:57
  • @Intelligentipauca Thank you for your comments and answer. I am aware of the constructibility rules, but I am not sure if my problem is directly related to it. I didn’t word this question properly. What I’m trying to ask is: If we are only given angles $a, b, c, d$ can we always find the angle "$e$" using using elementary geometry provided that all angles are integers in this triangle. Can we always produce a synthetic solution to find this angle? Think of it as a "find angle $e$" question rather than a construction problem. Is it more clear now, or am I missing something simple? – krazy-8 Apr 11 '21 at 16:01
  • OK: you want to find the value (in degrees) of angle $e$ by purely geometric means. But this can always be done: just construct an angle of $1°$ (or of $3°$ if all given angles are multiples of it) and then add it repeatedly to itself until you obtain an angle equal to $e$. – Intelligenti pauca Apr 11 '21 at 16:42
  • You can use the trigonometric Ceva's theorem to determine the two angles, provided a tool for proving that the angles satisfy the theorem. – user Apr 11 '21 at 18:38

1 Answers1

3

Only angles whose integer value in degrees is a multiple of 3° can be constructed, in general. But if we are given an angle with integer value in degrees, which is NOT a multiple of 3°, then we can construct all integer-valued angles.

From the above it follows that triangles containing integer angled sextuplets are NOT "solvable" only if the four angles we are given are multiples of 3°, while the other two angles are not.

EDIT.

As noted in a comment below, such a triangle cannot exist, hence all triangles containing integer angled sextuplet are indeed constructible.

It's possible that the OP had another meaning in mind for "constructible": in that case it should be advisable to explain that.

  • If as in this case we are given integer angles that aren't multiples of 3 degrees then we are generally able to construct other ones in that list, because that's enough to get a 1 degree angle. – Dan Uznanski Apr 11 '21 at 15:13
  • @DanUznanski That's exactly what I wrote above. – Intelligenti pauca Apr 11 '21 at 15:20
  • I can read really. Also why can I not undownvote.

    Oh! We do know that such a triangle does not exist: if it did, then- ugh. This is part of why I'm not sure I like this question - we can obviously construct the angles e and f, the construction is obvious given a, b, c, and d. So if a, b, c, and d are actually constructible angles, then e and f must be as well.

    – Dan Uznanski Apr 11 '21 at 15:24
  • @DanUznanski You are right: I'll correct my answer. – Intelligenti pauca Apr 11 '21 at 15:32