Ok, lets take this back a step or two and I'll see if I can clarify a few things to help:

1) I'd recommend glancing through my review paper I put the link up for as this summarises the philosophical and ethical issues that are typically discussed with respect to such technology.

2) It can't. I think this is where we're getting confused. However, before we get to that, my attitude is that you have to discuss what running even is. That sounds a little odd but what that means is you have to philosophically define the sport itself first. For example, you need to discuss the role of technology in a sport first. For example, cycling tried to minimise and standardise bicycle technology's impact in the hour record for some years. In this case, are trainers part of running or should their impact be minimise ? If the answer is yes, it's open season on their design (to a point). If the answer is no, you then determine its key performance indicators (i.e. what makes running fast with relation to the footwear specifically) and then after that, you define and validate the best tests that provide those answers. As you infer, the tests need to be achievable and specific to the sport. To date, the full battery of required tests have not been conducted on these shoes - hell, we don't even know what these are yet as everyone concedes they don't know exactly how the shoes even work.

3) Yes, you're right. This goes back to my earlier point about defining first what trainers contribute and then defining tests to do that. Mechanical energy return is only one of those tests and not the only test you should run. However, the use of running shoes are so complex, no study yet has attempted to conduct the full range probably needed.

4) The 100% argument was taken from the IPC's sports rules that suggest any form of sports technology (such as prostheses) should be 'passive in nature' only. If you get more than 100% back it's likely not passive and is illegal. This is easy to investigate with prosthetic limbs but problematic with the human leg as the ankle actually generates power so it requires testing in a slightly different way. So far, the published studies typically do this in two ways 1) with steady state running on a treadmill and 2) mechanically loading the shoes (when not being worn). The often quoted line that the shoes are 4% better than anything else is based upon that they use 4% less physiological energetic cost to use when running in them or that they require less oxygen uptake to use when running at the same speed.

5) What I mean by that is that we simply need to investigate these tests using more than one method. I've mentioned two tests above as examples. The 100% comment is meaningful though as the shoes should be tested for their level of mechanical energy return - even if 100% or greater isn't hypothetically possible.

A key point to this is that the ethics should be discussed BEFORE any tests are designed. Any decision should be based on a joint assessment between ethical discussion and then physical tests. A problem is that few researchers are able to handle both extremes - hence why World Athletics had convened a working group made up of a range of individuals.