It's a very fuzzy form of ecological accounting, much harder to calculate in any consistent and plausible way than (for example) carbon footprints. But the "N Earths" concept is dubious for reasons beyond simple accounting imprecision. Simply put, it's adding together the wrong things.

Assertions that we'd need three (or five, or ten) Earths to support our now-unsustainable lifestyles may make for nice graphics, but miss a more important story. The key to sustainability isn't just reducing consumption. The key to sustainability is shifting consumption from limited sources to the functionally limitless.

Broadly put, there are three different kinds of resources:

LIMITED-SUBTRACTIVE

These are resources that have a finite limit, and once used, would be difficult or impossible to reuse. The most visible example would be fossil fuels, but most extractive resources would also fit this category. For some resources, the limits may be extended through recycling, but this has limits as well. As a resource dwindles, the resulting high costs may make otherwise expensive extraction methods feasible, but eventually the resource will just be gone. In the language of economics, these are both rivalrous and excludable resources.

The implication for the "N Earths" model: given enough time, we'd never have enough Earths. Oil will run out, whether in a decade or a millennium, as long as someone continues to use it.

LIMITED-RENEWABLE

These are resources that renew over time, but face a limit to total concurrent availability. These are largely (but not exclusively) organic resources: food, fish, topsoil, people. Water arguably could be included here, as well. These resources can be over-used or abused, but absent catastrophe, will eventually recover. Economically, these are considered rivalrous but non-excludable -- that is, they're the "commons."

This is probably the closest fit for the "N Earths" concept, but misses two very important aspects: use management (encompassing conservation, efficiency, and recycling), which can alter the calculus of how much of a given resource may be considered "in use" in a sustainable environment; and substitution, which can cut or eliminate ongoing demand for a given resource (the classic example being guano as fertilizer).

UNLIMITED-RENEWABLE

These are resources that renew over time, but where the limits to availability are so far beyond what we could possibly capture as to make them effectively limitless. These run the gamut from energy (solar and wind) to materials (environmental carbon) to abstract phenomena (ideas). No current or foreseeable mechanisms could fully use the total output of these resources. Economically, they're both non-rivalrous and non-excludable.

Where the limited-subtractive resources make any use non-sustainable, given enough time, with unlimited-renewable resources, all uses are inherently sustainable.

The argument behind the "N Earths" model is that we -- the global we, but especially the West -- need to reduce our consumption to the point where we no longer use more resources than the planet can provide. The argument behind this alternative model -- call it the "Smarter Earth" model -- is that we need to shift our consumption away from limited resources, especially limited-subtractive resources, as much as possible. It's not a question of consuming less (or more, for that matter), but a question of consuming smarter.

The immediate rejoinder to this notion is that "we can't eat ideas or solar energy." That's superficially true; however, plants are embodiments of solar energy, and ideas can allow us to use limited resources more efficiently. It's not possible with current or foreseeable technologies to shift entirely to unlimited-renewable resources, but every step along the way improves our sustainability.

Another response to this model is that it's essentially an argument for a techno-fix. Despite appearances, it's not. What I'm arguing for is more of a design framework, a guide for decision-making. Yes, that may often mean technological design, but it also encompasses community design (as John Robb has engaged in with his "Resilient Communities" work), economic design (do tax and regulation patterns promote a shift from limited-subtractive to unlimited-renewable consumption?), and especially memetic design (how do we construct a coherent narrative of what's happening around us?).

The goal of shifting consumption boils down to this: moving from a "never enough Earths" model for society, to an "all the Earth we need" model.