Types That Never Were: 1 The Illusion of Type

In classrooms, conferences, and reviewer reports, we often begin with an apparently innocent gesture:

“This is a report.”

“This is a narrative.”

“This is a research article.”

“This is an explanation.”

The move feels harmless. It feels descriptive.

We are simply identifying what kind of text we are dealing with.

But pause for a moment.

What exactly is being classified?

A text is an instance — an event of meaning.

It is a point on the cline of instantiation.

Yet when we name a “type,” we treat that instance as if it belonged to a pre-existing category — as if somewhere in the system there existed a stable entity called Report, and this text simply realises it.

That move deserves scrutiny.

---

1. System and Instance

Within a Hallidayan ontology, language is a system of potential.

• The system is structured potential.

• An instance is an actualisation of that potential.

• Instantiation is not a binary (system vs instance), but a cline.

Between system and text lies a gradient: subpotentials, intermediate regularities, recurrent patterns of construal.

Crucially:

The system is not a taxonomy of text types.

It is a network of options.

When a text occurs, it does not “belong to” the system the way a specimen belongs to a species.

It actualises selections from the system.

This distinction is small, but decisive.

---

2. Register as Subpotential

Register, in canonical terms, is a functional variety of language:

variation according to field, tenor, and mode.

From the pole of system, register is a region of potential.

It is not a box labelled “narrative” or “report.”

It is a patterned clustering of probabilities within the semantic system.

Register is therefore:

• A tendency within potential.

• A subpotential on the cline of instantiation.

Already, the ontology resists classification.

---

3. How Types Appear

So how do “text types” arise?

From the pole of instance.

When we observe many similar instances, we notice recurrent patterns:

• Similar staging.

• Similar semantic domains.

• Similar lexicogrammatical tendencies.

• Similar contextual configurations.

We group them.

We label the grouping.

We treat the label as if it named something ontologically stable.

But what we have actually done is this:

We have taken regularities of actualisation

and re-construed them as essences.

The “type” is a retrospective abstraction from repeated instances.

It does not pre-exist them.

---

4. The Perspectival Shift

Here is the critical move:

From the system pole → register is subpotential.

From the instance pole → recurrence appears as type.

The difference is not empirical.

It is perspectival.

If we stand at the system end of the cline, we see structured probability.

If we stand at the instance end, we see clusters of resemblance.

Neither perspective is illegitimate.

But only one preserves the ontology of potential.

When we forget the cline, we slide from gradient to category.

And once we slide to category, boundaries begin to harden.

---

5. A Small Unsettling Question

Take a text commonly labelled a “research article.”

Is there a stable entity in the system called Research Article?

Or are there:

• Recurrent contextual configurations,

• Habitual semantic patterns,

• Conventional staging structures,

• And institutional expectations

that constrain actualisation in relatively stable ways?

If the latter, then the “type” is not a thing.

It is the shadow cast by patterned actualisation across the cline.

The difference is subtle.

But it changes what we think we are describing.

---

6. Where This Series Is Heading

In the posts that follow, we will:

• Re-clarify register from the pole of potential.

• Examine how recurrence becomes reification.

• Contrast cline with taxonomy.

• Re-think “text type” as perspectival artefact.

• Reassert instantiation as gradient, not classification.

The goal is not to abolish useful labels.

It is to relocate them.

Types may be pedagogically helpful.

They are not ontological units.

What we call a “type” may never have been there.

Only potential.

And its patterned actualisation.