1

Suppose $x$ is proportional to $y$ and $x$ is proportional to $z$. Then we can write, \begin{align*} x &= k_{1}y \end{align*}

The constant $k_{1}$ is going to depend on $z$. So, $$k_{1} = f(z)$$ Since $x$ is proportional to $z$, $$k_{1} = f(z) = cz$$, for some constant $c$. So we have, $$x = czy $$ $$x \propto zy $$

But, since $x$ is proportional to $z$ we can write, $$x = k_{2}z$$ So multiplying the two equations we have,

$$x^{2} = k_{1}k_{2}yz$$ $$\frac{x^{2}}{yz} = k_{1}k_{2}$$ $$x^{2} \propto yz$$

These are two conflicting results. So what am I missing here? I have looked at this thread: How does one combine proportionality? and while it does give some insight, doesn't fully answer my question.

Edit: Ok, I think I understand it now. In the last part we had $x^2 = k_1k_2yz$, since we are considering $y$ and $z$ as things that can change, we must also consider the constants. We know that if $z$ changes, $k_1$ changes, and if $y$ changes, $k_2$ changes. That is, $k_1 = c_1z$ and $k_2 = c_2y$. So we have \begin{align} x^2 &= c_1c_2y^2z^2 \\ x^2 &\propto (yz)^2 \\ \end{align}

From this can I write $$x \propto (yz)$$ ?

2 Answers2

2

The two results are both correct under different interpretations of what "$x$ is proportional to $y$ and $x$ is proportional to $z$" means.

One degree of freedom

One interpretation is that $x$, $y$, and $z$ are all related in a way that knowing any one of them determines the other two. Geometrically, the set of all points $(x,y,z)$ permitted by the relationship between them is a curve in space.

In this case, "$x$ is proportional to $y$" gives us a relationship $x = k_1 y$ that is absolute: if I tell you what $y$ is, you can immediately use $x = k_1y$ to find $x$. The same thing applies to the statement "$x$ is proportional to $z$": we have $x = k_2 z$, with some different constant.

In this case, $x^2 = k_1 k_2 yz$ is true but slightly misleading: it might make you think that you can choose any value of $y$ and $z$ and use the formula to find what $x^2$ is. In fact, from $x = k_1y$ and $x = k_2z$ we can determine that $z = \frac{k_1}{k_2} y$: if we choose a value of $y$, that determines both $x$ and $z$.

Still, it's true that $x^2$ is proportional to $yz$, in the sense that the ratio between $x^2$ and $yz$ is always the same.

Two degrees of freedom

Another interpretation is that $y$ and $z$ are separate inputs, independently free to take on any value we like. Geometrically, the set of points $(x,y,z)$ permitted by the relationship between them is a surface in space.

In this case, "$x$ is proportional to $y$" should really be said as "$x$ is proportional to $y$ when $z$ is held constant". We have $x = k_1 y$, but $k_1$ is a value that depends on $z$.

If $x$ is proportional to $y$ when $z$ is held constant, and $x$ is proportional to $z$ when $y$ is held constant, then:

  • Because $\frac{x}{y} = k_1$, where $k_1$ does not depend on $y$, the quantity $\frac{x}{yz} = \frac{k_1}{z}$ also does not depend on $y$.
  • Because $\frac{x}{z} = k_2$, where $k_2$ does not depend on $z$, the quantity $\frac{x}{yz} = \frac{k_2}{y}$ also does not depend on $z$.

We conclude that $\frac{x}{yz}$ does not depend on either $y$ or $z$: it is a true constant. Therefore $x$ is proportional to the product $yz$.

Misha Lavrov
  • 142,276
  • Thank you. So the main source of confusion here was the use in English? Also, I think there's a mistake in the first part: shouldn't it be "$x^2=k_1k_2yz$ is true but slightly misleading. Also at the end of the second part, to avoid confusion, shouldn't it be said like this: $x$ is proportional to the product $yz$, when either $y$ or $z$, but not both, is constant. – Orlin Aurum Jan 07 '23 at 04:01
  • I've corrected the first mistake. The second is not a mistake. In that situation, $x$ is proportional to the product $yz$ no matter what is going on with $y$ and $z$: from knowing what happens when $y$ is held constant, and knowing what happens when $z$ is held constant, we deduce what happens in general. – Misha Lavrov Jan 07 '23 at 04:10
  • This is a great explanation! My answer attacks the one degree of freedom perspective, and I didn't know there was another interpretation for proportionality. Thanks for the information! – Zo-Bro-23 Jan 09 '23 at 06:32
0

Explanation

Suppose x is proportional to y and x is proportional to z.

Mathematically, this means that $x=k_1y$ and $x=k_2z$, where $k_1$ and $k_2$ are constants. Multiply these two equations together: $x^2=k_1k_2yz$. Let $k_3=k_1k_2$. $x^2=k_3yz$, and since $k_3$ is a constant ($k_2$ and $k_1$ are constants), this means that $x^2\propto yz$.

Where you're going wrong

Like @David_Quinn mentioned, you are going wrong when you say that the constant $k_1$ is going to depend on $z$, or when you say that "we know that if $z$ changes, $k_1$ changes, and if $y$ changes, $k_2$ changes". Constants are constants; they don't change. The fact that $x$ is proportional to $z$ means that for any $x$, $z$ is equal to that $x$ divided by a constant $k_1$, and that constant has to be the same regardless of $x$ and $z$.

Testing with numbers

We can also arrive at the same conclusion by testing out numbers, although this is hardly rigorous. Let $x=1$, $y=2$, and $z=3$. This means that $k_1=2$ and $k_2=3$. If $x$ were to change to $2$, then, $z$ would change to $6$ and $y$ would change to $4$. $yz$ is $6$ when $x$ is $1$ and $24$ when $x$ is $2$. $x^2$ is $1$ when $x$ is $1$ and $4$ when $x$ is $2$. Since $\dfrac{24}{6}=4$, we can say that $yz$ might be proportional to $x^2$, but cannot be proportional to $x$. This is the same as what we get from our analysis above.

Edit

The other answer is talking about a case where $z\propto k_1$, and not $z\propto x$. This might be why you got confused and wrote $k_1=f(z)$. In your particular case, however, this is not true, so $A$ is not proportional to $BC$.

Zo-Bro-23
  • 216
  • well I don't think I can argue with your testing with numbers part, but what about Aang's answer in that thread I mentioned: https://math.stackexchange.com/a/433767/765852 – Orlin Aurum Jan 06 '23 at 16:14
  • Zarif's answer seems to be a better explanation. They are not talking about $A\propto B$ and $A\propto C$ like you are. Instead, they're talking about $A\propto B$, $A=kB$, and $k\propto C$. In this case, $A\propto BC$.

    Aang's answer uses a similar method, as can be noted by the line "$k=f(C)$".

    – Zo-Bro-23 Jan 07 '23 at 03:05
  • I think you are over-complicating things unnecessarily. You used the tag "algebra-precalculus", and the other thread uses the tag "linear algebra", so that might be something that you do not need to worry about right now. As long as you understand my reasoning, you should be fine! Let me know if you have any questions. – Zo-Bro-23 Jan 07 '23 at 03:06