An outlier outside the realm of normal expectation, as the past does not point to its possibility. It is characterized by its rarity, extreme impact, and retrospective predictability. That is, explanations after the fact make it seem like it was predictable.
The frequency of 'black swan' events has and will continue to increase over time due to an environment with an increasing number of feedback loops. That is, rapid information flow and compounding effects can lead to more unpredictable extremes.
This involves naively using the past trend to predict the future. From the point of view of the farmer's pig, he just keeps eating and growing and thinks this will continue.
Instead, the pig goes to the butcher shop and becomes ham and bacon. From the point of view of the pig this is an extreme event, but from the point of view of the farmer it is well within the realm of expectations. Extreme events are relative to point of view.
Often normal expectations are applied to situations where non-linear events prevail. With non-linear factors we have to adjust our expectations outward to include the possibility of rare extreme events.