Recently, the National Academy of Sciences issued a massive report analyzing how to cut U.S. gasoline use in half by 2030, and beyond that, how to reduce transportation emissions 80 percent by 2050.

If you don’t want to read all 395 pages, Brad Plumer has an excellent write-up. The take-home message is that no one technology or policy — efficiency, feebates, plug-in electrics, fuel cell vehicles, natural gas vehicles, what have you — can get us to those goals. To make the numbers add up, you have to do them all together.

I don’t have a bone to pick with any of the report’s numbers, but I do want to use it to make a point.

In a previous post I made a distinction between widgets, which are the discrete elements of a system, and systems themselves. It is possible to model changes in widgets with some precision. You establish a baseline scenario (“business as usual”) in which status quo policies, behaviors, and trends are locked into place and then model changes in a single variable, or small set of variables. This makes things manageable. If you set out to model changes in a large number of variables and how those changes interact with one another — i.e., to model systemic changes — things quickly become computationally intractable.

This is why modelers are always at pains to say that they are not making predictions, especially regarding scenarios stretching out to 2030 or 2050. Most such scenarios amount to, “what would happen if everything else stayed the same but this particular thing(s) changed?”

In the real world of human societies, of course, everything is changing at once. There are multiple overlapping complex systems, all of them displaying emergent properties and nonlinear dynamics. It makes predictions difficult — especially, as they say, about the future.

You can’t predict long-term changes in human cultures with models, though models can inform such predictions. Envisioning long-term change scenarios requires an understanding of history and political economy, serious engagement with culture and technology, a good grasp of systems dynamics, and a substantial dollop of intuition. Models can bring clarity to such analysis, but they cannot prove or disprove any scenario. Only time can do that.

Anyway, why am I blathering on about this? Because I fear that people take reports like the NAS one the wrong way, as a kind of delineation of what’s possible. It is not that. It merely demonstrates the limits of widget-based analysis.

Plumer notes that the report passes almost entirely over changes in land use and public transit, but I think that’s just one aspect of a larger phenomenon, which is that the report confines itself entirely to changes in vehicles. (To be fair, that’s what lots of people seem to do when imagining changes in transportation.)

But vehicles are only widgets in larger systems. Other parts of the system can change, and are changing. To take one example, NAS says “limited range and long recharge time are likely to limit the use of all-electric vehicles mainly to local driving,” which will restrict their growth. The unspoken presumption is that consumers will continue to demand all-in-one vehicles, cars and trucks that can serve both short-distance and long-distance needs.

But consumer demand is shaped by infrastructure, markets, and policies. What if even more driving was shifted to local urban areas and long-distance travel was shifted to either public transit or rented/shared vehicles (like Zipcar or the like)? What if land use and transit could be shifted so that consumers only (or mainly) need cars for local transportation? That would remove the constraint on the growth of electric vehicles, no?

Another thing the report passes over is the possibility of self-driving vehicles, which I happen to think will be game-changers (yeah, I said “game-changers”). What if local transportation needs in urban areas could be satisfied with a fleet of self-driving electric vehicles? That would mean fewer, lighter cars and less CO2.

Or consider how changes in the economics of renewable energy, particularly distributed energy, might change the market dynamics of electric vehicles. (Remember, electric car batteries could soon be an intrinsic part of grid management.) Or how demographic changes — young people driving less, older people returning to cities — might shift patterns of demand for transportation.

Anyway, the point here is not to explore, or even list, all the ways that transportation system dynamics might change. It is just to say that widget-level changes are only the tip of the iceberg.

I don’t know any better than anyone else how these things will turn out, though I have my hunches and predictions like anybody. (For instance, I think changes in power generation and distribution will make the logic of electrification inevitable, enough so to overwhelm the technological limitations of electric vehicles. That’s why they will win out over natural gas or hydrogen vehicles.) My main point is just that our aspirations, ambitions, and expectations with regard to transportation — and power, agriculture, land use, etc. — should go beyond widget-level change, and should not be constrained by analysis showing the limitations of widget-level change. We’ll only find out how much is possible by pushing on all the levers at once and remaining open and opportunistic, ready to exploit those nonlinear phase shifts when they arise.

Positive system change is impossible to predict. Those who believe in it are too often beaten down by reports and forecasts like the one from NAS, told that their aspirations are impossible, dismissed as naive. But it is only through such naivete that the world is redeemed. “Hope is definitely not the same thing as optimism,” said famed dissident Vaclav Havel. “It is not the conviction that something will turn out well, but the certainty that something makes sense, regardless of how it turns out.”

Sustainability — living on this planet with more compassion and awareness — makes sense. That’s what matters.