Trope Particularity

Try it first: the particularity trainer — an interactive drill on naming the nine effects — runs in your browser. The pipeline map sits beside it. This repository is documentation only: the vocabulary, the trainer, the map, and the essays and proposals around them.

A diagnostic vocabulary for property-instance individuation and loss. It names — precisely, with a fixed and deliberately small set of terms — the ways a specific, situated quality gets flattened as it moves through a system, and whether what survives is enough for the use at hand.

Duck typing asks whether it quacks. Particularity asks which quack you just laundered into a boolean.

Table of Contents

• What this is
• What this is not
• The vocabulary
  • Effects — what happens to a property-instance
  • Residue — what is left
  • Verdict — relative to a declared use-model
• Background
• Status
• Where this repo sits
• Machine-readable record
• Repository map
• Where to go next
• Licence

What this is

A workbench, not a language. The artefact is a vocabulary you think with: a board of named tools you pick up to say what just happened to a property-instance.

A property-instance is a quality as borne by this entity, in this context, under these conditions — the hoarseness of this duck, this morning, on this pond. Not the universal "hoarseness"; the particular, situated instance of it.

The vocabulary sorts into three kinds of term, because they are three different kinds of thing:

• Effects — things that happen to a property-instance (nine terms).

• Residue — what is left after an effect (one term).

• Verdict — a judgement on whether the residue was enough, always relative to a declared use-model (three terms: the use-model and the two verdicts).

Figure 1. The twelve terms, by family

See also assets/trope-particularity-map.svg (the pipeline, with the effects split two ways: by loss amount and by veridicality) and assets/particularity-trainer.html (an interactive drill on naming the nine effects).

The shape of the whole thing in one line:

A property-instance undergoes some p-effect; relative to its declared use-model, the surviving residual is judged p-sufficient or p-insufficient.

What this is not

Three boundaries keep the project honest:

1. Not "trope safety." This is not a rival to type safety sold on the same ground. That comparison is a losing one, and not the point.

2. Not "type theory cannot express this." Type theory can encode many such structures, and so can trope theory. The claim is narrower: neither, by itself, supplies the diagnostic vocabulary for property-instance individuation and loss.

3. Not a bag of subtle metadata. The centre stays fixed: individuated property-instances. A term that does not bear on individuation or loss does not belong here.

And a practical boundary for when any of this matters at all:

Particularity matters when loss, detachment, conflation, deception, or portability of property-instances affects interpretation, correctness, accountability, or later use.

The positive counterpart to these boundaries — where the vocabulary sits beside type theory, graded/coeffect types, provenance, and information-flow control — is docs/positioning.adoc.

The vocabulary

Effects — what happens to a property-instance

The nine effects split two independent ways. Loss amount: kept whole (p-preserving) versus lost-or-changed. Veridicality: faithful versus deceptive — the three deceptive duals (p-falsifying, p-misbinding, p-conflating) present something false rather than merely lossy.

Term	Arity	Veridicality	What it is
p-preserving	one	faithful	The baseline: nothing lost. Quality, bearer, context, and an honest record all survive. Every other effect is preserving minus one part (or a lie about one part).
p-projecting	one	faithful	Named parts dropped on purpose. You know exactly which coordinates are gone. A reduced view, not a lie — provided nobody mistakes the projection for the whole.
p-collapsing	one	faithful	Flattened to a predicate or scalar. The quality survives as a test; the individual that bore it is gone. The quack becomes true.
p-detaching	one	faithful	The quality severed from its bearer — the quality–bearer bond broken. A trait floating free of whose it was.
p-attenuating	one	faithful	Fidelity degraded with no clean structure. Unlike projecting, you have lost an unknown amount.
p-falsifying	one	deceptive	A false value: the residual claims faithfulness, but the quality presented is not the one that was there. Not loss — substitution.
p-misbinding	one	deceptive	A false bearer: the bond is asserted intact, but the quality is pinned to the wrong entity. The tie holds — to the wrong one.
p-conflating	two	deceptive	A false singleness: two numerically distinct instances asserted to be the same one, with no label saying so. Downstream reasons about one where there were two, and is silently wrong.
p-fusing	two	faithful	Two instances deliberately blended into one labelled aggregate. Not an error — the plurality survives in the label, so downstream knows not to pull them apart.

The arity column marks a real seam: seven single-instance effects, and two two-instance (p-conflating, p-fusing — the dishonest and honest versions of the same two-into-one move). The veridicality column marks the other seam: the three deceptive duals are false value / false bearer / false singleness. The two splits are independent — loss-amount is about how much survived; veridicality is about whether what survived is true.

Note

The three deceptive duals exist in the vocabulary because the formal semantics needs them, but none is writable in a faithful producer: a value-substitution, a mis-attribution, or an untagged merge is a fault, caught at the checker’s validation boundary, not a move you compose with (see the checker’s prevent profile, R-2026-06-22).

Residue — what is left

Term	What it is
p-residual	What actually survived an effect. The handle you inspect to decide whether the loss mattered.

Verdict — relative to a declared use-model

A verdict is meaningless on its own. The same collapse is sufficient for counting ducks and insufficient for diagnosing one — nothing about the collapse changed, only the use. So a verdict always names the model it is relative to.

Term	What it is
use-model	The declared statement of what a property-instance is for. The thing every verdict is relative to.
p-sufficient	The surviving residual is enough for the declared use. What was lost did not matter for this purpose.
p-insufficient	The surviving residual is not enough for the declared use. Something needed was lost (or falsified), and reasoning from the residue will go wrong.

Background

The vocabulary’s centre — qualities as borne by this thing, in context — is the founding move of trope theory (the metaphysics of abstract particulars; D. C. Williams, Keith Campbell). The full grounding and the compatibility test are in docs/background.adoc. In short: a property-instance is a trope; whose it is is compresence; and p-conflating is exactly the collapse of exact resemblance into numerical identity — two tropes can be exactly alike and still be two.

What trope theory does not supply is the dynamics of loss and deception — collapsing, attenuating, detaching, falsifying, mis-binding, and the sufficiency verdict. That layer is this project’s own.

Status

Early, but no longer alone. This repository is the vocabulary — the named effects, the residual, the verdict, their discriminating tests, and worked examples.

The checker is a separate repository: github.com/hyperpolymath/trope-checker. It is a checker, not a language — it consumes a language-neutral Trope IR and returns a verdict (p-sufficient / p-insufficient) against each consumer’s declared use-model, with a witness edge. Its reference core is in Idris2 (quantitative types track the loss-plumbing). The first front-end that lowers to the IR is the Haec language (github.com/hyperpolymath/haec). This workbench defines the words both of them realise; it depends on neither.

Where this repo sits

Vocabulary track: trope-particularity-workbench → haec → trope-checker (verdict). Store/query track: haec → vocarium → hermeneia → trope-checker (verdict). trope-checker depends on nothing (HC-1); the workbench depends on nothing; vocarium and hermeneia meet at the M2 StoreProvider adapter (vocarium ADR-0002: wire, do not link).

Machine-readable record

The repo deed trope-particularity-workbench_chora.deed at the root is the single machine-readable record. The retired A2ML tree is frozen, read-only, at https://github.com/hyperpolymath/trope-particularity-workbench/tree/b18eb67dcf10e80d3560b0468a7a85c484231f02/.machine_readable.

Repository map

Path	What lives there
README.adoc	This file.
vocabulary/	One entry per effect: definition, the test against its nearest neighbour, one worked example.
examples/	The same situation laundered several ways (the duck, a record, a person), in lockstep with the checker’s conformance fixtures.
assets/	particularity-trainer.html (interactive drill) and trope-particularity-map.svg (the pipeline).
docs/background.adoc	The trope-theory grounding (Williams, Campbell) and the compatibility test, in full.
docs/	Positioning, theory papers, whitepapers, funding proposals, and the research programme.
the checker	A separate repository: github.com/hyperpolymath/trope-checker — annotate, propagate, flag.

Where to go next

• vocabulary/ — the nine effects, sharpened. Start here.

• the trainer (source: assets/particularity-trainer.html) — drill until you can name the move from the test.

• docs/background.adoc — why the vocabulary is compatible with the metaphysics it borrows from.

• State of the trope stack — one table, one row per repo: status, what exists, the obligations it owns, proof/test citations, and the ground-truth command that verifies it.

• docs/status/RESEARCH-PROGRAMME.adoc — the prioritised, trope-stack-wide programme of work (what is worth doing next in the field, and who might fund it).

• docs/theory/computing/trope-theory-and-neural-representation.adoc — a position paper reading learned neural features as tropes (abstract particulars), and the estate’s loss/deception effects as names for superposition, polysemanticity, and the binding problem. Provisional.

• docs/whitepapers/outreach/baffled-couch-potatoes.adoc — an essay on the neglect of trope theory and why it leaves language models baffled.

• docs/whitepapers/academic/funding-landscape.adoc — candidate funders for the programme, mapped to work packages.

Licence

Code and config: Mozilla Public License 2.0. Prose docs and vocabulary: CC-BY-SA-4.0. Per-file SPDX-License-Identifier headers are authoritative.