# Can an approximation be “lifted”?

Let's say we make some approximation in some theory (such as there being no friction, when in fact there is).

Now we want to remove this approximation, and thus improve the real-world meaning of the theory.

Is it correct to say this?

This approximation can easily be lifted by (whatever, adding whatnot constant to whatnot formula on whatnot page).

Any alternatives?

-
Assumption or approximation? – JAM Jun 8 '12 at 22:02
@JAM I didn't even notice that at first. Now I'm not sure what's being asked. – Mark Beadles Jun 8 '12 at 22:07
@JAM Sorry, I meant approximation. Fixed. – meh Jun 8 '12 at 22:28
I think "removed" would be much clearer in this context. – Peter Shor Jun 9 '12 at 9:44

Yes, approximations can be lifted. To lift carries the sense of remove or annul, as in lifting an embargo.

The usage is common in technical papers. These are the Google Scholar results which use lift along with approximation.

-
Actually, your Google Scholar search returns only 15 papers, three of which are by the same author. I would say that it demonstrates quite the opposite — that such usage is not common. – 200_success Jun 10 '12 at 6:24

I've never heard of an approximation being lifted. I would say instead that some phenomena (e.g. friction, Coriolis force, relativity) can be corrected for. You can apply a correction factor if the refinement is multiplicative or add a correction term if it's additive.

-

I don't think 'lifted' is the word you want here. Once you remove or negate your assumption, providing instead real data, you can 'refine' or 'improve' your calculations, turning your approximation into an exact answer.

-
But isn't "refine" or "improve" rather weird? To me, it means that the approximation is still in place--just improved or refined. – meh Jun 8 '12 at 22:29
You can say you refine or improve your working theory when you remove the approximation or assumption. It would also make sense if you're laying our an argument for something, then refining that argument by addressing possible objections. – gatkin Jun 8 '12 at 23:02

You may write for example

Without this approximation/assumption, the conclusion still holds if you add this constant to that formula.

-