Wednesday, 8 April 2026

Systemic Functional Linguistics Under Constraint — 3 Realisation Is Not Bidirectional

With context restored and register stripped of its variable-function, one move remains available to those who want to preserve a familiar explanatory structure:

realisation is treated as bidirectional.

Sometimes this is explicit:

  • context shapes semantics
  • semantics realises context

Sometimes it is softened:

  • mutual influence
  • co-determination
  • feedback between strata

Sometimes it is implied in practice:

  • analysts move freely from context to semantics and back again
  • explanations invoke both directions as if they were interchangeable

In each case, the claim is the same:

relation between strata runs both ways.


This appears reasonable.

If semantics realises context, then surely context must in some sense shape semantics. Otherwise, how could the relation hold?

This intuition is precisely what must be refused.


Because the moment realisation becomes bidirectional, three consequences follow immediately:

  • relation becomes transitive
  • constraint becomes mechanism
  • and stratification collapses into process

The structure is simple.

If context acts on semantics, then something must pass:

  • from context
  • to semantics

Whether described as:

  • influence
  • shaping
  • constraint
  • probabilistic bias

the same requirement holds:

there must be a pathway along which this effect is transmitted.


But this is exactly what has been excluded.

No transfer.
No mechanism.
No shared substrate.

So bidirectionality is not an innocent extension.

It is a reintroduction of transitivity.


At this point, the account faces a familiar choice.

If realisation is bidirectional, then either:

(1) Context determines semantics

In which case:

  • semantics is responsive
  • meaning is shaped by something outside it
  • and the semiotic is no longer internally sufficient

Or:

(2) Semantics determines context

In which case:

  • context is reduced to a projection of semantic organisation
  • and the distinction between strata collapses

The attempt to hold both directions simultaneously does not resolve this.

It intensifies it.

Because now:

  • each stratum must act on the other
  • each must transmit something across the boundary
  • and each becomes partially dependent on the other for its operation

This is no longer stratification.

It is interaction.


And interaction is precisely what has been excluded.


The appeal to “feedback” makes this collapse appear sophisticated.

But feedback requires:

  • a loop
  • a medium
  • a process unfolding over time

In other words:

  • a system within which signals circulate

This reinstates:

  • a shared substrate
  • a space of operation
  • and a mechanism of transmission

All of which violate the constraints already established.


So the problem is not that bidirectionality is poorly formulated.

It is that it is structurally incompatible with the model it is meant to extend.


The alternative is not to weaken realisation.

It is to specify it correctly.


Realisation is not:

  • causal
  • interactive
  • bidirectional

It is:

a stratified relation in which a lower stratum realises a higher.


This must be read precisely.

  • semantics realises context
  • lexicogrammar realises semantics
  • phonology realises lexicogrammar

At no point does the higher stratum act back on the lower.

Not because it is passive.

But because “acting back” would require:

  • transmission
  • influence
  • or mechanism

Which are not available.


This does not mean that context is irrelevant to semantics.

It means that relevance cannot be expressed as influence.


There is no need to say that context shapes meaning.

Because:

  • meaning is the realisation of context
  • not a response to it

This removes the apparent asymmetry that motivates bidirectionality.

We do not need:

  • context → semantics
  • semantics → context

We need only:

  • semantics realises context

And this relation holds without:

  • causation
  • influence
  • or feedback

At this point, the familiar objection returns:

If context does not act on semantics, how do we account for variation?


The answer is now already available.

Variation is not produced by influence.

It is:

the differentiation of semiotic potential under constraint.


Different contextual configurations are not inputs.

They are:

  • distinct positions in the stratified organisation of the semiotic
  • realised in different semantic actualisations
  • and reproducible as patterns of constraint

No transmission.
No feedback.
No bidirectionality.


This restores the integrity of stratification.

Each stratum:

  • is distinct
  • is internally sufficient at its level
  • and enters into relation only through realisation

And crucially:

realisation is not coupling.


Within the semiotic:

  • relation is stratified (realisation)

Between distinct organisations:

  • relation is non-stratified (coupling)

To confuse these is to collapse the entire framework:

  • coupling becomes interaction
  • realisation becomes causation

So the insistence on bidirectionality is not a minor theoretical preference.

It is the final attempt to preserve a mechanistic model under a different name.


Closing formulation

Realisation is not bidirectional.

It does not transmit, influence, or feed back across strata.

It is the relation by which a lower stratum actualises a higher—
without causation, without mechanism, and without reversal.

The moment it is treated as bidirectional,
stratification collapses into interaction,
and the semiotic loses its internal sufficiency.


With this, the last technical refuge is removed.

There is no longer:

  • environment
  • variable
  • or bidirectional process

through which grounding can be quietly reintroduced.


What remains is exposed.

And under these conditions, most extensions will not need to be argued against.

They will fail on contact.

No comments:

Post a Comment