We often treat a failed description as a temporary inconvenience.
Sometimes that is true.
But sometimes something more interesting occurs.
It is not that we lack a description of what is happening. Our way of describing what is happening has ceased to work.
The distinction matters.
A map can fail because a road has been omitted. That is a problem of information. But a map can also fail because the landscape no longer fits the distinctions by which the map was made.
The difficulty is then not that we have failed to describe one more thing.
It is that the things we know how to describe no longer organise the situation adequately.
This can happen when a relation we had taken for granted stops holding. A category that seemed stable begins to blur. A boundary no longer separates what it once separated. A distinction that once made consequences visible no longer does so.
The world has not disappeared.
Something has happened to the relation between the world and our construal of it.
This is why a failed description can be revealing.
It can show us what the description had depended upon.
When a coordinate system breaks down, we discover that location had depended upon a particular structure of space. When a physical theory reaches a domain in which its equations cease to provide meaningful predictions, we discover that the theory had conditions of applicability. When an everyday distinction becomes impossible to maintain, we may discover that what seemed like an obvious feature of the world was also a feature of the way we had made the world intelligible.
The failure therefore points backwards.
It reveals the conditions that had previously remained invisible because the construal worked so well.
And it points forwards.
For once the old description can no longer carry us, we face a peculiar kind of openness. We are not simply waiting for more facts. We may need a different way of distinguishing what matters.
The horizon of a construal appears precisely here.
Not where reality ends.
But where a way of making reality intelligible reaches the edge of its competence.
The interesting question is therefore no longer simply:
What is beyond our description?
It is:
What has our description required in order to work at all?