"Is there a good reason why many runners think a soft surface is gentler on their feet and limbs? Or is this another example of a frequent error we all make, trusting what seems like common sense and never asking if the conventional wisdom is correct?"
There is, Kolata writes, "no evidence that softer surfaces prevent injuries," and "no reason to run on softer ground unless you like to."
Kolata is right—the conventional wisdom is wrong. But you should still run on trails.
Perhaps more to the point, there's nothing in the scientific literature on surface hardness and injury to suggest that soft-surface running is safer. Plus, Kolata writes, you're more likely to trip over a root or twist your ankle when you run on trails than you are when you run on pavement. Kolata takes both points as evidence against trails.
But she doesn't address a perfectly reasonably theory in support of trails, one that has nothing to do with impact forces: The main benefit of off-road running is variability.
Trails are never perfectly flat. Repetitive-stress injuries, the kind of injuries that most often affect distance runners, are caused when the same tendons, muscles, and ligaments are stressed in the same ways. Even slightly varying those stresses—say, by running on uneven trails—probably reduces the likelihood of injury.
Of course, there's little peer-reviewed evidence to support this view, either. But in the absence of evidence, the "variability of surfaces" theory seems more plausible than Kolata's belief that you'll trip over your feet—that simply isn't how most runners get hurt.