Relational Cuts: On Images — 2. Relation Before Object

We do not begin with objects.

This is not a preference. It is a restriction imposed by the previous step.

For what was shown there is that “something given to be seen” cannot yet be assumed to be an object. The notion of an object already presupposes a level of stability, a boundary, and a separation from what it is not.

None of this is available in advance.

So the question shifts.

If we cannot begin with objects, what remains?

One answer suggests itself too quickly: relations.

But this answer must also be suspended.

Because “relation” is usually understood as something that connects already-existing things. A bridge between points. A linkage between terms. A structure imposed upon entities that are already there.

If that is what relation means, then it cannot come first.

It would require what it is supposed to explain.

So we cannot yet say “relation” in that sense.

And yet, something is clearly structured.

Something is not undifferentiated.

There is distinction without objecthood.

There is contrast without things that contrast.

There is separation without separable units.

This is the difficulty.

We are confronted with a situation in which the most familiar explanatory units—object, property, relation—are not yet licensed.

And yet the condition for seeing has already been acknowledged.

So we are forced into a more careful formulation:

whatever is given to be seen must involve differentiation that is not yet attributable to objects.

This means:

difference is prior to object.

But even this is too fast.

Because “prior” already introduces an ordering, and ordering already risks reintroducing structure in the form of sequence.

So we should restate it more cautiously:

objects are not required for difference to be operative.

This is enough for now.

What follows from this is not a new ontology of objects, but a reorientation of explanatory priority.

Instead of:

objects → relations between objects → structure of what is seen

we are compelled to consider:

differentiation → stabilisation → the emergence of what can be treated as objects

But even this must not be read as a temporal story.

It is not that first there is raw difference, and then objects are constructed.

That would be another way of sneaking in a substrate.

What is at issue is not sequence but dependency.

Objects, if they are to be spoken of at all, depend on conditions of differentiation that do not already require them.

So the object is not primary.

It is not even secondary.

It is an effect of stabilised differentiation.

At this point, “relation” can be reintroduced, but only under pressure.

Not relation as connector of things.

But relation as the condition under which difference is maintained without collapse into sameness.

We can tentatively say:

relation, in this sense, is not between objects.

it is what makes objecthood possible without assuming it in advance.

But even this formulation must remain provisional.

Because it risks reifying “relation” as a new foundational entity.

That would simply replace one starting point with another.

So we hold it more lightly:

whatever is happening prior to objecthood is not object-like, but it is not nothing.

It is not a collection.

It is not a substrate.

It is not a field in the ordinary sense.

It is a constraint on how differentiation can persist.

Only after such constraints are in place can anything stabilise into what we might later call “an object.”

So when we say “relation before object,” we do not mean:

there are relations first, then objects.

We mean:

the conditions under which anything can appear as an object cannot themselves presuppose objecthood.

This is the minimum constraint.

And it is enough to unsettle the default assumption that we begin with things.

We do not begin with things.

We begin with the conditions under which “thingness” can occur at all.

What those conditions are is still unknown.

But we now know where not to look.