Something happens.
It leaves a trace.
But a trace is not yet history.
A footprint in sand records a passage, but unless the footprint changes what happens next, its history goes no further.
Something more interesting happens when a past event alters the organisation of a system.
A wound changes a body.
A repeated experience changes a nervous system.
A mutation changes a lineage.
Training changes an LLM.
In each case, the past produces a difference that persists.
But persistence alone is not enough.
The difference must enter into the system's subsequent activity.
The wound changes how the body responds.
The learned pattern changes what the nervous system can do.
The inherited variation changes which organisms are likely to persist.
The trained parameters change which continuations the model can produce.
The trace has become a disposition.
And this is perhaps the crucial transition.
A trace says:
something happened.
A disposition says:
because something happened, something else can now happen differently.
History has therefore become organisational when a past difference acquires consequences for future possibilities.
This does not require a representation of the past.
Indeed, the past may be completely absent from the system's awareness.
What matters is that the system has been changed.
The present now contains a difference that was produced by something earlier.
And because the difference affects what can happen next, the future contains a consequence of the past.
We can therefore extend our earlier sequence:
event → trace → persistence → disposition → altered possibility.
The final step is decisive.
A history that changes nothing is merely a history.
A history that changes what can happen next has become part of organisation.
This may be why memory is such a misleadingly narrow word for the phenomenon.
A genome does not remember in the way a person remembers.
A metabolic cycle does not remember.
A scar does not remember.
An LLM's parameters do not remember a training example in the ordinary sense.
Yet all can carry history forward.
They do so by making the future different from what it would otherwise have been.
Perhaps, then, the deepest form of historical persistence is not the preservation of the past.
It is the alteration of possibility by the past.
No comments:
Post a Comment