Nate, my understanding is that setting system boundaries is one of the issues the proposed ASPO committee on Net Energy experts will consider and define.  All energy production systems have "externalities".  Nuclear cooling water can warm up river water, U mining has a range of environmental hazards and of course there is the waste disposal issues.  We need a way of adding all that up.

Dual cycle nat gas power plants are often promoted on the basis of their super heat conversion efficiency.  But the utilities may conveniently forget to add in the energy costs of gas discovery, liquification, transportation, regasification before it reaches the boilers.

Bio fuels as you know have a wide range of potential externalities - water use, soil depletion, transportation, monocultures etc.  That's one reason why I liked the talk by Milton Maciel on Brazilian sugar cane at ASPO so much - I still think we should try and get a post on that - because it captured a technlogy that delivered high ERoEI but in an ecologically sustainable way.

Many of the externalities are "hidden" costs to society and they are often environemntal in nature.  Given our dual concerns for energy depletion and environmental sustainability, I believe that a "whole system" method of accounting EroEI + external costs is vital to seeing the best path forward.

"What if we design a great energy technology with high EROI of 20:1, yet it has triple the greenhouse gas emissions and depletes the soil."

That, by the very definition of "great", would not only not be a great technology but simply unsustainable.

In the real world one always has to deal with multiple problems at the same time, yet, it is not only technically correct to seperate concerns but also necessary. The reason why you are not allowed to dispose of your motor oil in the sewage is that the water treatment plant can't deal with it.

For very much the same reasons nuclear power plants are not allowed to ship nuclear waste in carboard boxes with the mail.

Right now global awareness is growing that disposing of CO2 in the atmosphere is just as bad as spilling oil or releasing radioactivity once you get to the Gton/year quantities. The mechanisms to deal with that are the same as with the motor oil and the nuclear waste: laws and regulations.  

Just because you can make up a worse scenario in your mind than exists in reality, real problem solvers will not have to change their strategies. And one of them is seperation of concerns, which starts at the analysis level and then continues on the political one.