A glint of spume in the moonlight, the hiss and rumble of breakers on rocky shoals, the ominous steepening of the open ocean rollers as they begin to feel the bottom: a lee shore – the sailor’s perennial nemesis.
What the sailor desires above all else in a gale (excepting, perhaps, a snug berth in a sheltered haven) is sea room. Leagues and leagues of sea room. Mile upon mile of empty water off his lee. For every ship makes leeway. It is the universal toll levied by the Aeolian ferryman for passage across the deep. And out in the deeps of the deep it is no great price. If the ship makes two points of leeway, then the navigator simply figures it into his calculations. But a coasting vessel may have precious little room to fudge her course. If another point of leeway means drifting another 500 yards to leeward before she rounds the next point, then she needs 500 extra yards of sea room, or she piles into the rocks. 600 yards would be better. 600 miles would be better yet.
A lee shore, put simply, is a nerve-wracking presence, and a lee shore at night, in a gale, on a poorly charted coast, is a very bad thing indeed. And so a sailor learns to keep off them as a matter of course.
Such generalities are liberating. They allow us to simplify our lives in a way that our world, filled with billions upon billions of unique objects, very much demands. No one can afford to analyze the idiosyncrasies of every little situation before making a decision. Perhaps this particular bit of coast has a quiet little harbor to slip into. Perhaps, in this particular case, there is no need to beat out of the bay and two hundred more miles to seaward at the first hint of dirty weather. There might be a beautiful little estuary two points down the coast. But then again, perhaps there is no good anchorage for 300 miles in any direction. With constantly updated charts of every coast we could know such things, and make a more informed decision, but coasts change, bars drift across harbor mouths, and the world is short of cartographers. Best to be on the safe side and get some sea room. A lee shore is not a beast to be trifled with.
But we may wish we had an updated chart, or better yet, a pilot, to let us know for sure that we aren’t wasting three days running away from a perfectly good anchorage, just because we want to be on the safe side, and this impulse is all to the good, because it recognizes in the gratifying freedom of generalization a (nevertheless troubling) lack of optimization.
As often as we may be forced to rely on a general solution that is “good enough,” we always recognize the possibility for an individualized solution that is better.
I propose then, that a crucial step in our problem solving methods be a determination of just what level of optimization we can afford. Mapmaking is, after all, expensive and time-consuming, but in some cases, it may be worth the cost. In our home port, for example, we may find it profitable to pay someone to notify us if any stone or bar bigger than a fishing dory moves out of place, that we may move about with the utmost confidence and efficiency in a place where we spend so much of our time. Just as reasonable would be to spend no time or money charting the shores of some desolate, rocky upthrust on the far side of the world, whose shores we hope never to see at all.
So too with social policy. At every turn, we must not fail to ask, “How much research can we afford to do (or not do) before we attempt a solution? And given the information we have, what level of generality must we accept in order to find a solution that works?” Far too often in our political discourse I see discussions happening at a level of generality (or specificity) that is quite obviously unsuited to the resources and demands of the situation, and I am irked.
One example: climate change. This is an area in which, as with most long-term forecasters, we are operating with little or no hard data. Fortunately, enormous amounts of data are being produced monthly as we seek to better understand the problems we face, but for the mean time we must base our predictions on computer models operating at a far lower level of resolution than we would like. This is much better than nothing, but it is far short of reliable. Furthermore this data collection is costly. Currently the scientific community judges the expenditure to be a worthy cause, and understandably so, the Earth being something like humanity’s home port. The political class, however, seems incapable of understanding the level of generality at which climate science is currently operating, and insists on either absurdly specific solutions (wind power – this amounts to attempting to tow a supertanker out of harbor with a rowboat) or impossibly general ones (cutting worldwide carbon consumption – which might be compared avoiding all coasts forever for fear of shipwrecks). While both ideas have a great deal of merit when tailored to a level of specificity that makes them meaningful, neither makes any sense as a “solution” to global climate change.
If any “solution” to undesirable human interaction with Earth’s climate is to be found, it will be found by asking, specifically, “In which areas of human interaction do we have sufficient data to make a reliable predictions about how changes in human behavior will affect the climate? What can be done in those areas and at what cost? What data do we need to increase our confidence levels? What costs can we accurately predict at the level of knowledge we have?”
In the case of climate change, many of the proposed costs only introduce greater uncertainty. Most especially in determining social and economic costs, we trade one non-linear, dependent set of uncertainties for another. We must recognize then, that our level of uncertainty at the global level is extremely high. Is the climate changing? Yes. How? It’s absorbing more heat? Why? For many reasons: partly we believe, because of anthropogenic CO2 emissions. Can we stop humans from producing more and more CO2 as they industrialize? We don’t know. Data so far suggests not quickly. Do we know what long term effects higher CO2 concentrations will have? We know some. Can we calculate the costs of increased CO2 emissions? Some of them. We expect them to be high. Can we calculate the costs of CO2 reduction? Not if we don’t know how to make people reduce their consumption.
In short, there is insufficient data to create an optimized solution at the global level. For the nautical equivalent, we are sailing uncharted seas, but we fear we may be running up on some rocks. We cannot change course with confidence because we suspect those courses too to be troubled by unknown shoals. We should very much like to stay where we are, as we have pretty well established the location of all the rocks in sight, and we are not currently aground. All of which leaves us little better than the equivalent of avoiding lee shores as a matter of course whenever possible. The best solution seems to be to reduce sail, and then wait to see what looms up ahead of us. In other words, try to reduce CO2 emissions whenever doing so seems to have an acceptable cost, knowing that with the data we have, the decision will have to be left up to individuals and small groups until more data is available, because no coherent cost-benefit analysis can be done at a global level at present.
This does not, however, free us from the burden of seeking optimized solutions in specific areas where we have more data. There is no doubt that the smog cost of auto-emissions in large cities is high. It is far less incalculable because we can restrict our data to information pertaining to a single population in a single, highly populated area. Not only must we consider fewer variables, but we can also collect data at a much lower cost and at a much higher resolution.
It is possible that a city, then, is a small enough, data-rich enough entity that it can consider a more optimized emissions solution focused on ozone, and carbon, sulfur, and nitrogen oxides. It has tax data through which to estimate economic costs of enforcement. It has hospitals to help calculate the social costs of non-enforcement. It has traffic cameras and counters to measure the effectiveness of enforcement. It might then embark on an emissions reduction plan of its own accord and achieve good results. It would do this not believing itself to be part of a global “solution” to climate change, which belief would be utterly unfounded, given the great uncertainty about the behavior of other cities, but believing it had the data to optimize a local solution.
Ultimately this all comes down to a single point. We are wasting our resources if we fail to come up with specific solutions when we have data, and we are risking unnecessary danger if we fail to heed generalized advice when data is lacking. Too often, we seem to be committing both errors at once.