This question should be asked in the Mathematics department.
Usually, plotting against x is a plot of function f(x) against a horizontal value of x:
f(x)
Therefore, plotting y against x
y = f(x)
which is a mapping of y values against a range of x values related thro the function f(x).
OTOH, when mathematically necessary, we would also plot x against y,
x = f(y)
The convention is that x would occupy the horizontal axis, while y occupies the vertical axis, regardless if x is plotted against y, or y against x. Visually, which often would appear mutually indiscriminatable for 1-1 mapping plots. For higher order graphs, it would be rather obvious what is being plotted against which. For example, quadratic functions and open-curve conics,
v <-> u
such that there are more than one value of u for every v, but only one value v for every u, it is quite obvious we should be conveniently plotting v against u, regardless of the orientation of their respective axes.
In the case of closed-conics: circles and ellipses, there is no difference in plotting vert against horz or horz against vert because there are always two values of v for each value of u, and similarly two values of u for each value of v.
However, if neither dimensions are specified in terms of x or y, for example, ROI against Investment, we would usually make ROI the vertical axis and Investment the horizontal axis. But that is not an exclusive practice.
This answer should be answered by a Math professor for 1st year Math students. This is not a question concerning the English language, because you would face the same dilemma whether you approach in Russian, Hindi, Swedish or Tagalog.