Asymptotic idealizations are used to set non-parameter properties of a model.md

These are cases where we need [[Infinite Idealizations|infinite/asymptotic idealizations]], method wise, as we cannot simply set the parameter to the value we want.

This idea is due to @Strevens2019a

Example: Infinite populations

To ease thinking/calculation about gene pools, sometimes you want to set genetic drift to zero. However, when making a model of genes and such (hmm biology) "genetic drift" isn't a parameter or basic component of this model, there isn't a "genetic drift particle" nor do organisms just have one feature which causes genetic drift. A lot of it is due to e.g. meiosis.

In order to set genetic drift to $0$, you could instead let the population size go to $\infty$, as that will even out any effect that genetic drift has on the gene pool. This is why people do this.

Example: FQHE?

The FQHE shares similarities with this idea: the relevant property we want to have is [[Simply Connected|simply connectedness]] i.e. no holes, but that's not a parameter of the model: we can't just have any random space and call it day by saying: abracadabra no holes possible! We can't tweak it directly, so we have to set the width of the strip to $0$ instead, which isn't possible.

Now, I think this is in super-stark contrast to Strevens' claim that [[Setting a parameter to 0 or infinity means it doesn't carry explanatory power]], as clearly the fact of whether or not the FQHE sample is 2D or not is pretty important, so important in fact that @Shech2019 claims it's enough to justify platonism!