Applying a Relational Ontology to SFL: 2 Instantiation Recut: Token–Type Without Temporality

The purpose of this post is to reconstruct instantiation within canonical SFL, remove the accumulated methodological debris, and integrate the relational ontology in a way that clarifies rather than revises Halliday.

Instantiation is a simple relation in Halliday’s architecture.

It is the token–type relation inside a stratum.

And yet it has been the source of decades of confusion, largely because of two unintended distortions:

1. A computational-temporal interpretation (selection from a pre-existing store)

2. A cross-stratal interpretation (texts instantiate context, or semantics “instantiates” field/tenor/mode)

This post repairs both, using the relational ontology to sharpen the conceptual knife.

---

1. The Canonical Definition: Instantiation Within a Stratum

Halliday defines instantiation as:

• the relation between the potential of a semiotic system and

• an instance of that system.

Crucially:

• Instantiation never crosses strata.

• It never describes relations between context and semantics.

• It is not processual, derivational, or generative.

The relation is purely semiotic:

• A situation is an instance of the contextual potential (culture).

• A text is an instance of the semantic potential (language).

Anything else is a category mistake.

This is the first restoration.

---

2. Rejecting Temporal and Computational Drift

In later SFL descriptions, particularly computationally oriented work, instantiation was sometimes described as:

• “selecting features over time”

• “traversing a system network”

• “the unfolding of choices”

These metaphors conflate logogenesis with instantiation, and procedural modelling with metaphysics.

Halliday himself never treated instantiation as:

• A temporal process

• A decision sequence

• A computational pipeline

These are modelling conveniences, not theoretical commitments.

In canonical SFL:

Instantiation is perspectival, not temporal.

This is where the relational ontology proves decisive.

---

3. Recutting Instantiation: The Perspectival Cut

In the relational ontology:

• A system is not prior to its instances.
  It is a theory of possible instances: a structured potential.

• An instance is not produced from the system.
  It is a perspectival actualisation: an event construed as a token of that potential.

The cut between potential and instance is therefore:

• epistemological (a relation of construal)

• perspectival (a way of viewing the event)

• non-temporal (not a sequence or derivation)

Thus:

Instantiation = the perspectival relation in which an event is taken as token of a potential.

This aligns perfectly with Halliday’s architecture and clarifies the metaphysics underlying it.

---

4. The Three Canonical Clines of Instantiation

Once instantiation is restored as a perspectival token–type relation, the canonical clines can be properly articulated:

(a) Contextual instantiation

• culture ↔ situation
  Context is the semiotic potential of culture.
  A situation is an instance of that potential.

(b) Linguistic instantiation

• meaning potential ↔ text
  The semantic system of language is the potential.
  A text is an instance in meaning.

A register is a functional variety of the language system.

A register instance is a text construed as an instance of that subpotential.

Nothing in this architecture allows “register variables” or the cross-stratal drift that has plagued other models.

Each cline is strictly internal to its own stratum.

---

5. System Reconstruction: The Reverse Perspectival Cut

If instantiation is the perspectival movement from potential → instance,

then reconstructing the system is the complementary movement:

• instance → potential

This is not:

• generalisation,

• abstraction, or

• induction.

It is a perspectival shift:

• The text is taken as evidence for the organisation of potential.

• The situation is taken as evidence for contextual potential.

In relational-ontological terms:

• Instance → metaphenomenon → system
  We move from first-order meaning
  to second-order construed meaning
  to theory of potential meaning.

This preserves Halliday’s logic while clarifying its ontological grounding.

---

6. What This Reconstruction Secures

With the relational ontology integrated and the canonical architecture preserved, we now have:

• Instantiation as a non-temporal, perspectival, semiotic relation

• The removal of computational metaphors from the metaphysics of SFL

• Contextual instantiation purified

• Linguistic instantiation properly internal

• Register restored to its rightful place as a semantic subpotential

• A coherent method for reconstructing systems from instances

• A unified ontology of type and token across the entire model

In short:

Instantiation is no longer a process; it is a relation.

No longer a sequence; a perspective.

No longer a drift; a cut.

This is the purified, generalisable metaphysics required for the rest of the series.

---

Next Post

Post 3 will reconstruct context itself—strictly as Halliday meant it—and then reframe it using the relational ontology so that:

• context is restored as semiotic potential,

• field/tenor/mode are systems (not properties of situations), and

• the category mistakes of environmental or sociological interpretations are eliminated.