Conceptual models used to facilitate the interaction with life's uncertainties are only built up by multi-trials and mini-errors. A student learning to fly needs to acquire the skill of recovering from a stall or a spiral dive well above the hard. A spiral dive at low altitude during instrument training can bring any learning to a very abrupt end. Designated pathways, whether on land, sea or in the air, are likely to be where they are, because being somewhere else happens to be a very bad idea. The error of straying off a pathway will in all probability result in a plunge over a bank into a river, a grounding on a sandbar, or smacking full throttle into the local mountain. Even inadvertently transmitting one's intentions to an enemy or rival can be terminal, since there may well be a response beyond any ability to correct.

All living things remain living whilst they are able to correct for the errors in their biological equilibrium and only make voluntary errors small enough to profit by learning. There is no point in striving to achieve some sort of hypothetical error-free performance. The aim must be to establish error boundaries and strive to keep within them. The stark existential reality then, is that mini-errors and repetition are necessary to integrate and facilitate learning, but fatal errors are terminal by definition. That's the way it works.

In some circumstances, error can be quantified and used to estimate a correction that would ensure the situation remained within predetermined parameters of balance. Auto-pilot systems read transducers which monitor performance and make adjustments which will ensure the aircraft sustains its predetermined behaviour. Industrial processes ensure optimum levels and flow-rates by feeding error signals back to a control mechanism which then determines the action necessary to maintain balance. Even economists experiment with poking various discontinuous disturbances at an economic system, in the occult hope that bulges and deviations might be altered in a direction supposed by them to be optimal.

Actions are not intrinsically safe or erroneous in themselves. Depending on the circumstances, an action which is usually benign in one context could be anything but in another. Closing the throttle before engine shutdown is a prudent move, whereas performing the same action just after take-off from a forestry airstrip is quite the opposite. Breathing whilst competing in the national euphonium finals has quite different consequences to taking a breath under water during the finals of the synchronized swimming.