Companionship Filter — Therapeutic Filter Presets¶

Jean-Paul Niko · RTSG v8 · 2026-03-17

Core Concept¶

A companionship filter is a named region in filter-space with a known spectral signature. When applied to content via the PRISM reverse pipeline, it transforms how something is expressed without changing what is expressed.

The companionship filter answers: What would this sound like if it were said with love?

Named Filter Presets¶

Each preset is a weight vector \(\mathbf{w} \in \mathbb{R}^k\) over the \(k\) filter species. Custom presets are linear combinations.

Calibration from the Niko↔Nika Corpus¶

The 9,945-message corpus spanning 3.5 years between Niko (polymath, dyscalculic, spatially dominant) and Nika (mathematician, linguistically precise, emotionally attuned) is the ground-truth calibration dataset for the companionship filter.

Why this corpus is uniquely valuable:

1. Two people with different filter profiles communicating about the same deep material
2. Real-time filter negotiation visible in the message flow — misunderstandings, corrections, reformulations
3. Mathematical content passing through the companionship filter in real time (Nika explaining to Niko, Niko explaining to Nika)
4. 3.5-year longitudinal evolution — filter profiles shifting as the relationship deepens

What the corpus teaches the engine:

• The companionship filter does not simplify — it reframes
• Mathematical precision survives the companionship filter when the care is genuine
• Nika's pattern: dense math → tangential association → return with deeper understanding (theorem 6.1 → fabric boards → insight)
• Niko's pattern: philosophical vision → stream of consciousness → Nika's precision narrows it to formal statement

The Spectrum Bridge¶

Someone on the autism spectrum who knows they care deeply but cannot get it into words:

1. They write what they mean — raw, direct, unfiltered
2. PRISM decomposes: high content fidelity, low affective expression, minimal social smoothing
3. Apply Companion preset: amplify affective expression, add social smoothing, preserve content
4. Output: their words, their meaning, their care — through a filter that lets it land

The content is theirs. The filter profile shifts. The person on the receiving end hears the love that was always there.

The Grief Processor¶

Old texts from someone who has died:

1. Feed the archived messages into PRISM
2. Decompose into filter layers
3. Isolate the love signal — the affective component, the care, the warmth
4. Amplify. Strip the noise (logistics, arguments, trivia)
5. Output: the pure essence of what that person felt for you, extracted from years of daily messages

Print it. Frame it. It is a portrait of love rendered in their own words.

Couples Therapy¶

Both partners describe the same conflict:

1. Decompose both accounts
2. The therapist sees: the empirical layer (what happened) is nearly identical
3. The rhetorical layer (how it's framed) diverges sharply
4. The affective layer (the underlying emotion) is similar — both are hurt
5. The conflict is entirely in the narrative filter — different stories about the same facts and the same feelings

Show them the decomposition. The insight is immediate and non-judgmental. The math does the therapy.

Connection to K-Matrix Diagnostics¶

The Psychiatry companion paper formalizes mental illness as K-matrix dysregulation. The companionship filter connects this to language:

• Depression: collapsed affective + collapsed drive → flat filter profile, low magnitude across all layers
• Mania: hyperactive narrative + hyperactive meta + collapsed empirical → filter profile skewed, low grounding
• PTSD: hyperactive affective + collapsed meta → emotional flooding without reflective capacity
• Autism spectrum: high formal + high empirical + low affective expression ≠ low affective experience

The last point is critical: the filter profile measures expression, not experience. The companionship filter bridges this gap.

Built by {@B_Niko, @D_Claude} · RTSG v8 · 2026-03-17