Languages as Ontologies III – Deep Dives: 3 Typologies of Valency: How Languages Theorise Participation

If Posts 1 and 2 examined how languages carve events and relations, this post turns to a third foundational architecture:

How do languages construe participation?

Typologists call this “valency”: the number and type of participant roles a verb “takes”.

In a relational ontology, this misses the point entirely.

Valency is not a property of verbs.

It is a systemic theory of what it means for something to matter to an event.

Different languages make different ontological commitments about:

• which participants are intrinsic,

• which are optional,

• which are construed as configurational additions,

• and what counts as “involvement” in the first place.

The aim of this post is to show that valency is not argument-counting.

It is the language’s metaphenomenal architecture of how participation becomes meaningful.

---

1. The English Theory: Participation as Verb-Internal Structure

English offers a familiar schema:

• intransitive (1 participant),

• transitive (2 participants),

• ditransitive (3 participants),

• optionally plus obliques/circumstantials.

This schema presumes:

1. events have intrinsic slots,

2. those slots belong to the verb,

3. participants fill pre-existing functional roles,

4. anything outside the verb’s argument structure is peripheral.

This is one possible theory of participation:

• the verb is central,

• participants are dependents of that centre,

• significance is encoded structurally: “core” vs. “peripheral”.

But many languages do not divide the world this way.

---

2. Languages That Construct Participation as Configuration

In many systems — across Australia, the Pacific, the Amazon, the Caucasus — valency is not a verb-internal property at all.

Participation is relational, distributed, and configurational.

Examples of configurational participation

• participant roles are marked on nouns, not verbs;

• involvement is graded (affectedness, control, animacy) rather than categorical;

• “core” vs “peripheral” is not a binary but a continuum;

• multiple participants can be foregrounded simultaneously without being “arguments”;

• or the language allows differential argument indexing that shifts with discourse needs.

In such systems, a clause is not an event-centred structure with participants attached.

It is a configuration of participants, each contributing differently to the unfolding phenomenon.

This realises an ontology where:

events are emergent from configurations, not containers for them.

---

3. Languages That Dynamically Reassign Valency

Some languages allow valency to shift fluidly depending on viewpoint, affectedness, volitionality, or discourse structure.

For instance:

• a participant can be promoted to high salience via morphological marking (voice, alignment, indexation) even if not “core” in English terms;

• valency can be increased or decreased without the English-style alternations of “causativisation” or “applicativisation”;

• or languages may treat roles like “recipient”, “beneficiary”, or “instrument” as equally central when highlighted.

In such systems, there is no fixed intrinsic participant structure.

Participation is not structural.

It is perspectival — and the grammar encodes that perspectival shift.

This perfectly matches the relational ontology:

• participation is a cut, not a property;

• what counts as “involved” is a construal, not a metaphysical fact.

---

4. Languages That Encode Valency Through Interaction Rather Than Structure

In many languages, participation is distributed across:

• discourse markers,

• evidentials,

• switch-reference systems,

• participant-tracking clitics,

• and interactional factors like shared knowledge or attention footing.

This yields an ontology in which:

• participation is as much epistemic as experiential,

• the system embeds “who is relevant” into the intersubjective organisation of the clause,

• involvement is not defined by the verb but by the social configuration of speakers, hearers, and referents.

English cannot do this.

Its valency architecture treats participation as internal to the event, not co-individuated in the interaction.

---

5. Languages With No Clear Core/Peripheral Boundary

Some languages treat all participant-like elements as equally structural.

Others effectively have no valency distinctions at all, relying on:

• flexible constituent ordering,

• predicate-nominal structures,

• or clause-chaining strategies where “arguments” bleed across clauses.

In such systems, valency as a category collapses.

Not because the languages are “underspecified”, but because the system has encoded a non-modular ontology of involvement — one in which:

• phenomena are not composed of discrete roles,

• and involvement is temporally emergent, not structurally assigned.

This aligns elegantly with the relational-ontological claim:

events do not come with participant slots;

participants and events co-actualise each other.

---

6. Reframing Valency Within a Relational Ontology

Across languages, valency reveals the underlying metaphenomenal commitments of a system:

• Is participation intrinsic, configurational, dynamic, perspectival?

• Do events anchor participants, or do participants co-constitute events?

• Is involvement a structural property, an interactional calculus, or a discourse-driven emergence?

The crucial point:

Valency is not about counting arguments.

It is about a system’s theory of what involvement is.

Different languages instantiate different architectures of:

• significance,

• centrality,

• relational relevance,

• and event–participant co-constitution.

Typology, once again, becomes comparative ontology:

the study of how languages carve the conditions under which something matters.

---

Next Post

Post 4 will analyse Typologies of Transitivity, not in the impoverished sense of “high vs low transitivity”, but as contrasting theories of impact, change, dynamism, and commitment — the deep grammar of how languages construe the intensity and texture of experience.