# L4 — Construct distinctions (do NOT collapse)

**Job:** `lane-L4-boundary-permeability` · Hub `b73c292a` · MODEL_CHANGE=NO  
**Machine-readable:** `data/lanes/L4/construct_map_v0.json`

## 0. What BOUNDARY_PERMEABILITY is asking

Whether **cross-boundary interaction** can be measured as a construct **distinct from** polarization (preference/ideological distance or affective hostility). Permeability = *traffic across a boundary*; polarization = *positioning / affect around a boundary*.

## 1. Five interaction families (not one construct)

| Family | Layer | What is measured | Not the same as |
|--------|-------|------------------|-----------------|
| **A. Legislative cooperation** | Elite behavioral | Cross-party legislative *acts* (esp. original cosponsorship offer/attract; joint sponsorship; bipartisan passage acts) | Roll-call geometric distance; party-unity stock |
| **B. Bipartisan networks** | Elite relational | Sustained/voluntary network ties across party (caucus co-membership graphs; repeated cosponsor edges; friendship/office networks where measured) | Single-shot cooperation rates; compulsory committee co-service alone |
| **C. Cross-party social relationships** | Mass social | Self-reported friendships / discussion partners / social contact across party | Elite cosponsorship; GSS generalized trust |
| **D. Institutional reciprocity** | Elite institutional | Reciprocal exchange norms inside institutions (offer↔attract dynamics; selective reciprocity in lawmaking) | S2 mass generalized reciprocity / trust |
| **E. Cross-group civic interaction** | Mass civic | Bridging civic participation, cross-cutting organizational membership, deliberate bridging programs | Affective thermometer gaps; elite legislative acts |

### Hard rule — elite legislative reciprocity ≠ S2

| | Elite institutional reciprocity (L4-D / E2-N4 adjacent) | S2 Reciprocity Breakdown |
|--|------------------------------------------------------|---------------------------|
| Actors | Legislators / institutional elites | Mass public |
| Observables | Cosponsorship offer/attract, reciprocal bill support, institutional exchange | Generalized trust, fairness, help-stranger norms (GSS trust = CORROBORATION only) |
| Unit | Member-term / chamber×Congress rates | Survey proportions |
| Failure mode if collapsed | Treating CBS decline as “society lost reciprocity” | Treating GSS trust as elite cooperation |

**Standing SRP rule preserved:** affective polarization ≠ S2; GSS trust ≠ S1 MEASUREMENT; GSS trust ≠ L4 elite permeability.

## 2. Relation to L1 dimension C and E2-N4

- L1 rigidity candidate table already parks **C boundary permeability** on CBS offer (KEEP) / attract (MODIFY, no composite).
- L4 is the **dedicated construct-feasibility lane** for permeability: it must say whether C can be measured **without reproducing polarization**, and which of A–E are KEEP / MODIFY / INSUFFICIENT.
- E2-N4 remains the landed **elite cooperation practice** series — L4 **reuses, does not recompute or composite**.

## 3. Measurable without reproducing polarization?

**Yes, conditionally** — if sensors stay in the **interaction / traffic** family and reject:

1. Between-party NOMINATE center distance (E2-N3 = ELITE_IDEOLOGICAL_SEPARATION) as a permeability proxy  
2. Within-party dispersion / party-unity as permeability  
3. Out-party feeling thermometer / affective gap (S1) as permeability  
4. Silent `1 − cooperation` inversion sold as “polarization” without a named fracture memo  
5. Offer+attract composite index  

Independence test (conceptual): a chamber can show **high ideological separation** and **nonzero cross-party offering** (or the reverse). N3 and N4 already illustrate partial contrary patterns in E2 work — keep them **separate sensors**, not substitutes.

## Explicit non-claims
No L4 score. No S2 score. No E2 composite. No Rigidity Index. No outcome correlations.