Can a "Done" Increment support empiricism?

I think the answer is YES and is bound to the core of Scrum.

Empiricism, or knowledge frm experience, is the basis of Scrum. The only way of learning true value, is to deliver increments as soon as possible, so they can provide feedback. Empiricism is the reason why Done increments are created in the first place, right?

Given an online application where customers can order stuff, and an increment where customers can track their order progress, the only way to learn the true value of this "feature" is by providing it to the customers and collect insights, data, feedback etc.

Am I thinking too easy, or missing the point here? 

Can a "Done" Increment support empiricism?, If so, how?

Only if it is released.

Only if it is released.

In regards to the product, but I suppose it supports empiricism from the perspective of a Dev Team's capabilities, processes, techniques, etc.?

I believe that it can. Empiricism is a philosophy rooted in the concept that all knowledge is gained by experience (or sense-experience, sensory experience depending on where look).  Combine the description of Definition of Done as a way for everyone involved in the problem being addressed by a Scrum Team to know when an increment is complete and I see empiricism in practice, at least partially.  You still need to add to this whole picture the Sprint Review where stakeholders are given a chance to provide feedback based on the insights they gain from seeing the Done increment in use. Also add the transparency of the work being done, information related to the problem and the continuous communication that occurs in Scrum. 

Add all of that up and I see that a Done Increment can indeed support empiricism.  You can even flip that and say that empiricism supports a Done increment.

It may be appropriate to draw a distinction between empiricism and empirical process control. Empiricism can be applied in a laboratory setting where external factors are deliberately excluded and variability is minimized. Rapid prototyping in a non-production or sandboxed environment with invited users could be an example. Such an experimental approach, however, would not allow empirical process control to be established, which would require the inspection and adaptation of a production system functioning in a live environment.

Good point Ian!

It may be appropriate to draw a distinction between empiricism and empirical process control.

@Ian Mitchell, The point you make, makes sense, however, when I googled empirical process control, it always results in the term empiricism such that I am not able to see the distinction made between the two terms. I also checked the Scrum Glossary but these are not two separate terms.

Would you be able to shed some light? Thanks.

Suppose a team released a change in functionality to all users. They may not like it and the risks are high, but empirical observations can be made. Process control can also be established. If customer satisfaction nosedives, then the company can roll back or replace the change. If inspection and adaptation happens quickly enough then the company might recover.

Now suppose they tested the change first with a small representative cohort of users. Risk is minimized since little control is exercised over the business environment. Yet empirical observations about the sample can still be made, and hypotheses tested and conclusions drawn.