In building off of this question Can anyone show an example of going through Liouville's differential algebra theorem?
I'm slowly starting to understand more components of Liouville's theorem after organizing what I do and don't know, writing notes and looking back through the theorem again.
But, just when I thought I'd finally figured something out about it, I hit a roadblock.
Let's say I start with the field of rational functions $C(x)$ and I want to work with this to build my starting-field $\mathbb{F}.$
Now, I'm given an element $a \in \mathbb{F}$ where $a = e^{\sqrt{\ln(x)-1}}$ and I suppose $\int{a}$ belongs to some elementary field extension $\mathbb{G}$ of $\mathbb{F}.$
My confusion and question is, what field extensions were necessary to construct $a$ so that I can begin applying Liouville's theorem and adding extensions to check $\int a$?
The reason I'm confused is, rational functions are always the singular go-to example for this theorem. So, I figure that if $a$ belongs to a differential field, it must be $C(x,\ln(x),e^x,\sqrt{x})$ but I'm worried this is incorrect! The reason being, these are functions rational in the listed variables, $\frac{P(x,\ln(x),e^x,\sqrt{x})}{Q(x,\ln(x),e^x,\sqrt{x})}$. This says nothing about composites of these functions in of itself! So if I'm given a function like this, what field extension can I work with other than that of $a$ itself?
And, if I have to add finite field extensions, how could I ever see ahead and know whether it's possible or impossible to add such extensions? Supposedly this is what Liouville's theorem addresses but I'm not sure.
If I need to manually guess at what field extensions to add, how could Liouville's theorem ever tell me ahead of time whether an antiderivative is elementary or not?
– StackQuest Jan 03 '22 at 03:03