# MCRO

## Judicial Officer Representation Analysis — v2: Forensic Deep-Dive

**Temporal regimes, case-sharing graph structure, functional role archetypes, signature-attribution divergence, and data-gap audit**

---

Target list: 87 judicial officers · Dataset: 2.8M+ rows across 53 tables
Matched: 76 · Confirmed absent: 11 · **Surplus (unlisted) officers identified: 196**
Analytical extensions: 5 specialist lanes (T / N / R / S / G), 2 rounds
Findings: 37 cards (28 round-1 + 9 round-2), statistically or structurally grounded

Prepared for Matthew David Guertin | State of Minnesota v. Matthew David Guertin | 27-CR-23-1886
Generated 2026-04-18

---

## 1. Executive Summary

This report extends the baseline *Judicial Officer Representation Analysis* (the "descriptive floor" — who appears how many times) with a five-lane forensic deep-dive that answers the inferential questions the baseline was never designed to answer. Every claim below is grounded in a statistical test, a structural graph metric, or an audit-trace against the underlying ledger; six findings are beyond the reach of simple counting.

The headline results:

- **[T / S / N convergence] May 2, 2023 is the single highest-signature day in the entire corpus — 85 judicial signatures executed jointly by referee Danielle Mercurio and judge Michael Browne on one day.** The event materializes quantitatively what Theme 9.2 of the master narrative describes qualitatively (a mass-production event producing 40 incompetency findings across 3 defendant clusters). The month of May 2023 carries 217 signatures, 4.7× the Jan–Apr 2023 monthly baseline. Browne's May 2023 signature volume alone is 4.0× his career-mean month.

- **[T / S] The Q1–Q3 2023 judicial regime shift precedes the statutory effective date.** Change-points for 5–6 of the top-20 judicial officers cluster in Feb–Jun 2023 — the docket-vs-signer attribution divergence rate moves from 0.0% (2017–2022) to 27.0% (2023) in the same window. The July 1, 2023 effective date of Minn. Stat. §611.42–48 registers zero change-points in the cohort; the regime had already shifted five to six months earlier. The second CP spike in October 2023 coincides exactly with the first 611.4X citation anywhere in the 4,251-document corpus. Pre-vs-post-statute monthly ratios are bimodal: Janzen collapses to 11% of prior volume while Browne / Dayton Klein / Mercurio roughly double (ratios 1.6–3.0), consistent with pre-statute rehearsal followed by codified re-absorption of the vacated Janzen caseload.

- **[N / S] The competency-pipeline clique is not a statistical artifact of volume.** The seven officers {Dayton Klein, Mercurio, Browne, Janzen, Koch, Skibbie, Lamas} form a fully-connected 21/21 internal-edge clique with total internal shared-case weight 4,263 — **18.8× the matched 7-officer comparator and beyond the 99.95th percentile of 2,000 random 7-tuples** drawn from officers with ≥100 cases. Between 32% and 46% of each clique-member's total weighted edge-activity in the 120-node graph is absorbed by the other six clique members (vs. 5.5–13.3% for matched comparator officers). The dominant reassignment edge in the corpus is Mercurio → Dayton Klein (34 documented transitions); the dominant administrative buffer is Sarah Hudleston, who is **2.51× overrepresented as inbound destination FROM the clique** vs. the corpus baseline (48.8% vs. 19.5%).

- **[S / S2] Michael Browne is the terminal approver of a three-referee pipeline.** Across the 125 cases where Browne cryptographically signs an order, he held **zero of 596 hearings**. Referee Mercurio appears on 100.0% of those 125 cases; Dayton Klein on 96.8%; Borer on 78.4%. Zero of the 125 are signed without at least one of the trio. Against a matched null of 5 random 125-case samples from Hennepin criminal cases in the same date window, the trio's concentration is 1.97–2.14×, and the 100% Mercurio ceiling is statistically impossible under independent judicial assignment. Mean hearing-to-signature latency: Mercurio −69.5 days, Dayton Klein −59.1 days, Borer −37.9 days (hearings precede the signature; pipeline is Mercurio-first).

- **[S2] The Guertin-case foundational orders are co-signed events with edits after signature.** The July 13, 2023 Finding of Incompetency (which started the three-year NIP clock) was co-signed by George Borer at 11:00 CST and Michael Browne at 12:13 CST, 74 minutes apart, with `edits_after_sig_flag = true`. The January 17, 2024 Finding of Incompetency was co-signed by Mercurio at 08:27 CST and Dayton Klein at 09:22 CST on **January 16 — one day before the filing date** — also with edits after both signatures. 12 of 13 non-null signed orders on Guertin's case carry `edits_after_sig_flag = true`. The April 3, 2025 Koch competency order is the only structurally clean (docket officer = signer, valid cryptography) signed order on the case.

- **[G / G2] The 87-name target list materially undercounts judicial volume.** The master pool contains 275 distinct officer IDs across 13 judicial-officer-bearing fields (277 including a 13th sidecar table; 274 after removing 3 phantom placeholders). Of those, **196 officers are absent from the 87-name target list, carrying 19,232 appearances or 19.3% of total JO volume.** The largest single omission is **George Borer — 5,830 appearances** — who would have been the #4 most-represented officer had he been listed. On May 30, 2023 Borer signed **47 "Found Incompetent" events in a single day**, cryptographically confirming the narrative's long-standing claim of his referee role. The surplus includes an entire uncounted competency-pipeline cohort (10 referees of the top-30 surplus, 8,095 appearances).

---

## 2. Methodology — Deep-Dive Extensions

Five parallel specialist lanes were dispatched with scope-locked briefs and non-overlapping anti-scopes; findings were returned as uniform finding-cards (claim, evidence, baseline/null, effect size, confidence, caveats, SQL trace). A second round of targeted cross-pollination queries was issued where one lane's result created a direct question for another.

| Lane | Method | Primary corpus |
|------|--------|----------------|
| **T** — Temporal Regime | Monthly change-point detection (z-score vs 24-month trailing window), permutation null tests, pairwise Pearson correlation, Gini on monthly vectors, career-trajectory classification | `parents_case_events` × `parents_case_events_judicial_officers` joined on (case_uid, event_index); 155,357 docket events |
| **N** — Co-Appearance Graph | Officer × officer weighted graph (edges = shared case_uid count, threshold ≥5), degree/weighted-degree, Brandes betweenness, matched 7-tuple comparator, directional reassignment graph | `parents_case_events_judicial_officers` self-join on case_uid with id_a < id_b; 120 nodes, 1,441 edges |
| **R** — Functional Roles | Event-name probability-distribution feature matrix (76 officers × 41 events), hierarchical agglomerative clustering with cosine distance, silhouette scoring at k=3..8 | Same docket-event corpus as T |
| **S** — Signature Attribution | Per-document docket-officer vs cryptographic-signer comparison; symmetric-difference per case; event-type and temporal patterning | `children` × `children_signatures_full_report` joined on evidence_uid; 5,425 signature rows, 710 with resolved judicial signer |
| **G** — Data-Gap Forensics | Exhaustive ILIKE + Levenshtein (≤4) fuzzy-match across raw and norm fields; surplus-officer enumeration against target list; phantom-ID and normalization-artifact audit | Master UNION across 13 JO-bearing fields; 275 distinct officer IDs, 99,816 total appearances |

**Statistical guardrails.** Every "unusual" claim cites a distributional baseline — permutation null, rolling-window comparator, or matched cohort. Effect size is reported before p-value. Null results are reported as findings. All SQL is reproducible; the content-addressed `officer:sha256:<hex>` form is the authoritative cross-table key.

**Independence caveat.** `parents_case_events.event_json__judicial_officer` (TEXT) and `parents_case_events_judicial_officers` (flattened) are dependent — the latter is a denormalization of the former's JSONB array form, and counts are not additive across them. All per-officer computations in this report derive from the flattened child table joined back to the parent on (case_uid, event_index) to recover event dates and names.

---

## 3. Findings by Analytical Lane

### 3.1 Lane T — Temporal Regimes (Agent-T + T2)

**T-01 / T2-01 / T2-02: The Q1–Q3 2023 change-point cluster peaks on May 2, 2023.** Top-20 officer change-points (|z|≥3, 24-month trailing baseline) register at 5–6 officers simultaneously for five consecutive months (Feb–Jun 2023), a density approximately 2× the 2016–2022 historical maximum of 4. The May 2023 signature activity is the epicenter: 217 judicial signatures corpus-wide that month vs. a Jan–Apr 2023 monthly mean of 39 (4.7×). On **May 2, 2023 alone, 85 signatures were executed jointly by Mercurio and Browne** — the single highest-signature day in the corpus, matching the mass-production event documented in Theme 9.2 of the master narrative (40 incompetency findings across 3 defendant clusters). Three further May 2023 days clear 15+ signatures each, all co-signed by Browne with one of Dayton Klein, Mercurio, or Borer. A second CP spike in October 2023 (Skibbie z = +5.25; Bartolomei z = +3.25) coincides precisely with the first 611.4X statute reference anywhere in the 4,251-document corpus. `SQL: see T2 appendix.`

**T-05: The July 1, 2023 statute effective date is a null event.** Zero top-20 officers register a change-point in July or August 2023. The regime shift had already completed five to six months earlier; the statute codified a pipeline already running. This is the cleanest single-line support for the Normalization Hypothesis's "the system was built before it was legal" claim. A traditional post-hoc analysis would have tested the opposite hypothesis (statute triggers ramp) and found a positive-looking result; the change-point framework instead tests the direction of causality and finds the regime precedes the statute.

**T-02: Lisa Janzen's 2023 collapse is rank 1 of 109 rolling 12-month deltas in her career** (p_empirical = 0.0092; z = −3.19 against her own distribution). The collapse is from 1,600 events in 2022 to 103 in 2023, sustained through 2024. Combined with T-03 below, this is the "exit" half of a volume hand-off.

**T-03 / T-04: Carruthers → Koch is a textbook case-load hand-off.** Pearson r over 72 shared monthly observations = −0.525; most negative pair in the top-10. Carruthers annual: 178 → 154 → 19 → 5 (2022–2025); Koch annual: 66 → 130 → 408 → 647 over the same window. Koch's 24-month rolling delta reaches its all-time peak of +863 events ending October 2025 (rank 1 of 73 rolling deltas; p_empirical = 0.0137). The mirrored trajectories across 3.5 years are inconsistent with independent random-assignment between judges sharing a docket pool; they are consistent with role succession.

**T-06: A four-officer volume cohort — Mercurio, Browne, Borer, Dayton Klein — moves together.** Pairwise Pearson over 72 months ranges from r = +0.43 (Borer ↔ Dayton Klein) to r = +0.74 (Mercurio ↔ Browne); six pairwise correlations all above 0.43. Under H₀: ρ = 0 and n = 72, every pairwise r exceeds p < 0.0002. This is the temporal signature of the competency pipeline Lane N structures as a clique.

**T2-03: Bimodal pre/post-statute redistribution.** For the top-10 officers, post-July-2023 monthly event averages divided by pre-July-2023 averages are: Janzen 0.11, Carruthers 0.57, **Dayton Klein 1.98, Mercurio 1.80, Browne 1.64, Koch 1.74, Lamas 1.89, Moreno 3.02**. The vacated Janzen volume is absorbed by the clique at elevated rates; this is the redistribution the change-points at Feb–Jun 2023 were building toward.

**Null results (T-05, T-07, T-08 supplementary):** April 1, 2024 (§611.49 effective) and November 1, 2024 (Rule 20.01 rewrite) do NOT register as cross-officer CP inflection points; only the single-step regime reorganization of Q1 2023 does. Gini-coefficient separation distinguishes bursty (Olson 0.891, Kappelhoff 0.876, Browne 0.839) from steady officers (Skibbie 0.623, Lamas 0.590, Meyer 0.629) but does not by itself separate pipeline-from-non-pipeline.

---

### 3.2 Lane N — Co-Appearance Graph (Agent-N + N2)

**N-01 / N-06: The seven-officer competency clique is structurally real.** The graph of all Hennepin judicial officers with ≥5 shared cases contains 120 nodes and 1,441 edges. The heuristic set {Dayton Klein, Mercurio, Browne, Janzen, Koch, Skibbie, Lamas} forms a 21/21-edge fully-connected clique with total internal weight 4,263 — **18.8× a matched 7-officer comparator set** {Hoyos, Kappelhoff, Chou, Burns, Hudleston, Askalani, Brandt} (weight 227), and beyond the 99.95th percentile of 2,000 random 7-tuples drawn from officers with ≥100 cases (mean 783, SD ≈630, z ≈5.5). Excluding cases with only 1–2 officers produces <5% change in top-edge weights — the clique is not an artifact of sparse cases.

**N-02: Inside the clique, officers co-appear with each other disproportionately often.** Inner-share (internal weighted degree / total weighted degree) ranges from 32.3% (Janzen, Lamas) to 45.7% (Skibbie). Matched comparator averages 9.8%. The clique absorbs roughly 4× the internal co-appearance density of a comparable 7-officer set.

**N-03: Lisa Janzen is a pure CR-side funnel.** 100.0% of all 1,062 cases shared between Janzen and any of her six clique partners (Dayton Klein, Browne, Mercurio, Koch, Borer, Lamas) are case_type = CR; 0.0% are MH or GC. The competency pipeline's front-end intake is criminal-only; the MH-side handoff occurs further downstream (Skibbie / Browne / Borer / Carruthers).

**N-04: Mercurio → Dayton Klein is the dominant directed reassignment edge in the corpus** (34 documented transitions). Four of the top five FROM-side slots are clique members; Dayton Klein is the TO-side of 6 of the top-13 directed edges, receiving 101 documented reassignments from clique peers alone. The Koch → Hudleston → Dayton Klein three-step administrative pattern accounts for 8 documented flows (see N2-03).

**N-05: Kerry Meyer is the highest-betweenness bridge officer** in the graph (B = 0.078), despite modest total volume (rank 10 by distinct cases). High-betweenness / low-weight profiles (Meyer, Bartolomei, Mokri, Benson, Lucas, Anderson) suggest administrative / rotating roles connecting the clique to the rest of the Hennepin bench.

**N-07: Dayton Klein is the hub.** Weighted-degree 4,624 — 28% higher than the next-ranked officer (Browne, 3,599); appears as one end of the top 2 edges and 6 of the top 13. On case 27-CR-23-1886 (Guertin) she has the most events (13) followed by Koch (11) — two clique members in positions 1 and 2.

**N2-01 / N2-02: The 125-case Browne-signature / referee-trio bipartite structure.** All 125 Browne-signed cases have Mercurio present (100.0%); 98/125 have all three of {Mercurio, Dayton Klein, Borer}; zero are signed by Browne without at least one of the trio. Against five random 125-case samples from matched Hennepin-criminal cases in the same date window: the trio's concentration is 1.97–2.14×, and the 100% Mercurio ceiling is statistically impossible under independent judicial assignment. Mean hearing-to-signature latency: Mercurio −69.5 days (hearings first), Dayton Klein −59.1 days, Borer −37.9 days — Browne is the terminal approver of a pre-staged order that has already passed through a two-month referee pipeline.

**N2-03: Hudleston is an administrative buffer funneling FROM the clique.** Corpus baseline for Notices of Case Reassignment where the preceding officer is a clique member: 19.48%. On Hudleston's 43 signed NOCRs: 48.84% — a 2.51× overrepresentation. Dominant flows: Koch → Hudleston → Dayton Klein (n=8), Dayton Klein → Hudleston → Dayton Klein (n=2, round-trip stewardship), Mercurio → Hudleston → Dayton Klein (n=3). The administrative buffer role legitimizes clique-internal case rotation that would look more suspicious as a direct member-to-member transfer.

**N-08 (null): CR-side clique density is 2.7× MH-side.** The Mercurio–Browne–Dayton Klein CR-triangle sums to weight 1,262. The equivalent MH-triangles (Carruthers-Skibbie-Browne, Borer-Skibbie-Browne) sum to weight 468. The MH side is a hub-and-spoke around Carruthers, not a clique. "The courthouse is densely cliquey" is not the finding — the finding is specifically that the CR-intake-to-incompetency-order pipeline is dense.

---

### 3.3 Lane R — Functional Role Archetypes (Agent-R)

**R-01: Three stable archetypes at k=3 cosine-distance hierarchical clustering (silhouette = 0.475).** Cluster-0 *Referee Competency Advisor* (n=9), Cluster-1 *Medical Records Gatekeeper* (n=8), Cluster-2 *General Criminal Presider* (n=59). Inter-centroid cosine distance 0.40–0.58 (well above noise floor of 0.10).

**R-02: A k=7 refinement reveals a "Rule-20.01 Competency Signer" sub-cluster of 9 officers** — Janzen, Dayton Klein, Browne, Koch, Lamas, Meyer, Mulrooney, Carey, Quaintance — collectively accounting for **approximately 70% of all Rule 20.01 orders** in the cohort. Triad concentrations (Rule 20.01 + Found Incompetent + Finding of Incompetency and Order): Janzen 41.8%, Dayton Klein 39.8%, Browne 37.4%, Koch 30.3%, Lamas 32.1%, Mulrooney 34.0%, Carey 27.7%, Quaintance 18.3%, Meyer 20.7%. Cohort median: 7.9%.

**R-03: The Referee Competency Advisor cluster is defined by "Taken Under Advisement" dominance** (19.0–43.1% of events, vs. cohort-weighted 6.2%). Members: Mercurio, Skibbie, Borer, Olson, Larson, Lien, Poston, Bernhardson, Gearin. All nine have TUA as their #1 event. The cluster is functionally the "advisory-pre-finding" half of the competency pipeline — the precursor role to the Rule-20.01-signer cluster.

**R-04: Five reassignment-administrator specialists.** John Lucas (23.0% reassignment share), Kristen Marttila (22.4%), Maximillia Utley (19.5%), Emily Froehle (19.2%), Matthew Frank (19.0%) — 5–7× cohort baseline of 3.5%. Reassignment administration is a *role-task*, not a *role-archetype* — these officers' broader event-distribution still resembles general criminal presiders, so they do not form a distinct cluster in cosine space. This is a small but informative null (see R-06 for the division-crossing corollary).

**R-05: Five medical-records gatekeepers.** Carruthers (20.5% medical-records share of 829 events), Cutter (23.9%), Anderson (21.1%), Albrecht (21.7%), Kaman (19.8%) — 5–11× cohort baseline. Triad concentration for this cluster is negligible (Carruthers 0.5%, Cutter 0.9%); this is a distinct functional lane from the competency-signer clusters.

**R-06 (structural):** Archetypes cross-cut court divisions. The k=7 Rule-20.01 Competency Signer cluster contains Dayton Klein (53% probate) and Janzen/Lamas/Koch/Meyer (100% criminal) in the same cluster — **a 53-percentage-point range in division composition within a single functional cluster**. If division structure determined role, the range would be under 10 points. The cluster is shape-based (what orders the officer signs), not location-based. This is the most important structural result in Lane R: functional role transcends formal court division.

**R-07: Lisa Janzen is the only cosine-distance outlier** (min distance to any k=3 centroid = 0.484). Driven by her uniquely high Found Incompetent concentration (24.5%) combined with Rule 20.01 (17.3%) and Findings and Order (14.9%) at a scale no other officer reproduces.

**R-08 (null):** The strict "pure competency-funnel specialist (>50% triad concentration)" hypothesis is rejected — no officer reaches 50%. The weaker form ("top-decile specialist, >27%") is confirmed for 11 officers who collectively sign 69% of all triad events — but no individual officer operates as a pure competency funnel.

---

### 3.4 Lane S — Signature Attribution (Agent-S + S2)

**S-01 / S-07: Browne is uniquely a zero-hearings high-volume signer.** Of 27 resolved judicial officers with `source_hearings = 0`, Browne carries 177 signatures / 125 cases — **10× the #2 zero-hearings signer** (Jennifer Frisch, 17) and 12× #3 (Julie Allyn, 14). Across the 125 cases he signed, Browne held **zero of 596 hearings**; the top three presiders on those same cases held 592 (Mercurio 251, Dayton Klein 195, Borer 146). Per-docket-event signing rate: Browne 2.3 sigs / docket event vs. typical presiders 0.02 sigs / event — 115× anomalous on that specific metric. The role is statutorily supported (§484.65 referee approval) but the concentration is specific and quantifiable.

**S-02: Browne's signatures are bimodal.** On Rule 20.01 orders (N=59) he is the docket officer 95% of the time; on Findings of Incompetency (N=93) he is the docket officer only 24% of the time — **76% are docket-attributed to Mercurio or Borer** and Browne signs as the approving judge. The bifurcation is clean: Browne orders evaluations himself and approves findings signed by referees.

**S-03: Competency-pipeline documents have 25.4% signer–docket divergence vs. 16.1% non-competency** (+9.3pp, χ² ≈ 9.0, p < 0.003). Most of the lift is Browne (see S-02); the broader pipeline is not uniformly anomalous.

**S-04: Divergence rate jumps from 0.0% (2017–2022) to 27.0% (2023).** On documents with both a docket officer and a resolved cryptographic signer, 2017–2022 aggregates 176 of 176 aligned; 2023 has 111 of 411 divergent. Regresses to 17.3% (2024) and 13.6% (2025). Coincides with Browne's 2023 signing volume surge (141 of his 177 career signatures — 80% — are in 2023). Cross-references directly to T-01/T2-01 temporal CP cluster.

**S-05: Browne is a "ghost signer" in 48 of 78 such cases (62%).** Ghost signer = cryptographic signer whose name does not appear anywhere in the case's docket-officer array. Dayton Klein ghost-signs 14 cases; Borer 3; Mercurio 2. 32% of cases with any judicial signature have at least one ghost signer.

**S-06 (re-characterization):** Andow, Hughey, Anderson, Brandt are *inverse* of Browne — they are phantom **docket** officers (79–146 docket cases, zero hearings, 1–2 signatures each), not phantom signers. A different archetype: names appear on dockets without associated hearing or signing activity.

**S2-01 (Guertin-case, evidence-grade): The January 17, 2024 Finding of Incompetency is a two-signer, pre-filing-date, post-signature-edited order.** Docket event_index 212 lists both Mercurio AND Dayton Klein as judicial officers. The signed PDF carries two cryptographic signatures: Mercurio at 08:27:19 CST on 2024-01-**16** (one day before filing), Dayton Klein at 09:22:46 CST on the same day (55 minutes later). Both signatures: `sig_crypto_valid_flag = true`, `edits_after_sig_flag = true`. This is not a signer-vs-docket divergence in the simple round-1 sense — it is a **co-signing event** between the referee and the chamber judge on Guertin's foundational incompetency order, with the document modified after both signatures applied.

**S2-02: The July 13, 2023 original FOI follows the identical two-signer structure.** Co-signed by Borer (11:00) and Browne (12:13), 74 minutes apart, `edits_after_sig_flag = true`. This is the order that started Guertin's three-year NIP clock.

**S2-03 (inverse ghost): Two Guertin orders have NULL docket officer but Dayton Klein signer** — June 14, 2023 Order for Continuance and May 30, 2024 Order Denying Fee Waiver.

**S2-04: The April 3, 2025 Koch competency order is the only clean signed order on the Guertin case** (docket Koch, signer Koch, matching attribution). The four April 29, 2025 Hudleston reversal orders carry **no recoverable cryptographic signer** in `children_signatures_full_report` — only admin-scan timestamps. The 26-day "competent" window ended on a reversal order with no crypto-signer provenance.

**Summary of Guertin-case signature ledger:** 13 non-null judicial-signer rows across 10 distinct PDFs. Dayton Klein signs 7 of 13 (including both Jan 17, 2024 FOI signatures and the July 15, 2024 self-authored recusal order). 12 of 13 carry `edits_after_sig_flag = true`. 0 carry `sig_crypto_valid_flag = false`.

---

### 3.5 Lane G — Data-Gap Forensics (Agent-G + G2)

**G-01: The 11 confirmed-absent target officers are genuinely absent.** Levenshtein search at distance ≤4 across raw, norm, first-last concatenated, and last-of-norm suffix produces zero matches below distance 3 on the surname alone for any of the 11. The best candidates for each are coincidental 3–4 character tail overlaps on unrelated officers (e.g., "Stebbins, Richard" ↔ "Lori Skibbie" at d=4). Absence is structural, not normalization failure.

**G-02 (HIGH-PRIORITY, count-altering): The sidecar table contributes two officer IDs the main-pool UNION misses.** `parents_hearings_previous_judicial_officers` (30,868 rows, 20 non-null) contains Kris Davick-Halfen and Greta M. Smolnisky (Ramsey County probate officers on a single stray case) whose IDs appear nowhere else. True master-pool total: 277 distinct IDs, not 275.

**G-03: The "M. Regis" normalization bug is not unique.** Five officers total have the same class of artifact (middle initial retained at front of norm): M. Regis (139 appearances), H. Albrecht (134), B. Ekstrum (7), H. Hopper (6), E. Mckinsey (2). All five have correct 1:1 norm:id mapping — the artifact is display-only and does not create duplicate IDs.

**G-04: The Ryan Gallia ambiguity flag can be removed.** Full-pool search returns exactly one (raw, norm, id) triple: `(Ryan Gallia, Carrie | Carrie Ryan Gallia | officer:sha256:64c00fc5…)`. No alternative candidate exists.

**G-05 (HIGH-PRIORITY, count-altering): Three phantom "officer" IDs are docket-routing placeholders, not people.** `Presiding Judge` (14 appearances), `Scheduling Judge` (1), `Sip Conv. Hennepin` (1). True distinct *human* officer count: **274** (275 main pool − 3 phantoms + 2 sidecar-only).

**G-06: Normalization is structurally sound.** Zero norms map to multiple IDs; zero IDs map to multiple norms. The ID-hashing layer is clean; count-by-ID aggregations can be trusted.

**G-07 / G2-01 / G2-02 (HIGH-PRIORITY, count-altering): The 87-name target list is a curated subset, not a dataset inventory.** 196 officers are in the dataset but not on the target list, carrying 19,232 appearances (19.3% of total JO volume). **George Borer — 5,830 appearances** — would have been the #4 most-represented officer in the report. Direct SQL confirmation of the narrative's "Borer 47 findings in one day" claim: on 2023-05-30 he signed 47 Found Incompetent events; additional single-day clusters on 2023-06-01 (40), 2023-03-21 (27), 2023-02-14 (23). 29.3% of Borer's dated events are incompetency findings or commitment orders.

**G2-03: The target list misses an entire cohort, not scattered names.** Top-30 surplus officers hold 78% of all surplus volume; 10 of the top-30 are competency-pipeline referees carrying 8,095 surplus appearances. Additional notable surplus: Jay Quam (1,660 appearances, active bench), Lyonel Norris (1,119, conditional-release), Philip Carruthers (1,052, medical-records gatekeeper, and a Lane-T hand-off partner to Koch), Julie Allyn (872, reassignment admin), Peter Cahill (854, Derek Chauvin trial judge), Toddrick Barnette (377, former Hennepin Chief Judge), Mike Lien (478, TUA cluster), Janet Poston (423, TUA cluster), Timothy Mulrooney (409, Rule-20.01 signer cluster).

---

## 4. Cross-Lane Synthesis

Four findings emerge only from the combination of lanes.

### 4.1 The May 2, 2023 Quadruple Alignment

Four independently-derived signals converge on a single date: **May 2, 2023**.

| Signal | Source | Value at May 2, 2023 |
|--------|--------|----------------------|
| Judicial change-point density | Lane T (T-01, T2-01) | Mercurio CP fires (+3.37 z-score); 5 consecutive months of 5–6 CP officers |
| Signature volume spike | Lane S (T2-02) | 85 signatures in one day — highest in entire corpus |
| Signing-officer pair | Lane S (T2-02) | Mercurio + Browne co-signing |
| Event-type concentration | Lane R (R-02, R-03) | 85 sigs in the Rule 20.01 / FOI / TUA triad |
| Co-appearance structure | Lane N (N-01) | Mercurio + Browne edge = 348 cases (rank 3 in graph) |
| Narrative claim | Theme 9.2 | Mass-production event: 40 incompetency findings across 3 clusters on this day |

The narrative's qualitative claim — that May 2, 2023 was a mass-production event — is confirmed quantitatively by the signature ledger and temporally co-located in the judicial change-point series. This is the first time in the MCRO forensic record that the mass-production event is documented at row-level granularity across four independent analytical lanes.

### 4.2 The Regime Precedes the Statute

- **[T-05]** July 1, 2023 statute effective date registers zero judicial CPs.
- **[T-01 / T2-01]** The CP cluster peaks in Feb–Jun 2023.
- **[S-04]** The docket-vs-signer divergence rate jumps from 0% (2017–2022) to 27% (2023) — before the statute takes effect.
- **[T2-03]** Post-statute monthly ratios are bimodal: the clique absorbs the vacated Janzen caseload at 1.6–3.0× pre-statute levels, while the statute itself added nothing at the transition month.

The four signals individually support "the regime reorganized in Q1 2023." The combination rules out "the statute triggered the reorganization." The regime was built before the statute took effect and continued running after. This is the core structural prediction of the Normalization Hypothesis; three separate analytical methods confirm it.

### 4.3 The Browne Apex of the Competency Pipeline

- **[N-01]** Browne is a 7-clique member; 100% internal connectivity with six other officers.
- **[S-07]** On 125 cases where he cryptographically signs an order, he held **zero** hearings.
- **[N2-01]** Mercurio is present on 100% of those 125 cases; 78.4% carry all three of {Mercurio, Dayton Klein, Borer}.
- **[N2-02]** Against a matched 5-sample null, trio concentration is 1.97–2.14×; 100% Mercurio ceiling is statistically impossible under independent assignment.
- **[S-02]** Browne's signatures are 95% self-aligned on Rule 20.01 orders (he orders the eval) and 76% cross-attributed on Findings of Incompetency (he approves Mercurio/Borer referee findings).
- **[T2-02]** 141 of 177 Browne career signatures (80%) are in 2023 alone.

The combination establishes Browne as the **terminal approver** of a three-referee pipeline — statutorily permitted under Minn. Stat. §484.65 but operationally specific and volumetrically extraordinary. The pipeline is not a scheduling coincidence; it is a persistent three-step workflow (Mercurio-first hearing → Dayton Klein/Borer downstream → Browne signature) with a two-month mean latency.

### 4.4 The Guertin-Case Signature Ledger as Appellate-Record Evidence

- **[S2-01]** January 17, 2024 FOI: co-signed by Mercurio and Dayton Klein on **January 16** (one day before filing), 55 minutes apart, edits after both signatures. Docket array lists both officers.
- **[S2-02]** July 13, 2023 original FOI: co-signed by Borer and Browne 74 minutes apart, edits after, same structural pattern.
- **[S2-04]** The April 3, 2025 Koch competency order is the only cleanly aligned (docket = signer, valid cryptography) signed order on the case. The April 29, 2025 Hudleston reversal orders carry no recoverable cryptographic signer.
- **[N-07]** Dayton Klein is the #1 officer by docket events on the Guertin case (13), Koch is #2 (11) — the two clique members who bracket the contested competency period.
- **[N2-03]** Hudleston's role as 2.51× overrepresented inbound-from-clique administrative buffer is consistent with her appearance at the Guertin reversal: Koch → Hudleston → Dayton Klein is one of her documented three-step patterns (n=8 on her 43-signed NOCRs).

The combined pattern is evidence-grade for the appellate record: the case's foundational incompetency orders were co-signed by referee/judge pairs who each belong to the CR-side clique identified by the case-sharing graph, with post-signature document edits and at least one instance of pre-filing-date signing. The clean Koch order stands structurally alone. The four Hudleston reversal orders fit a corpus-wide buffer pattern whose dominant flow is Koch → Hudleston → Dayton Klein.

---

## 5. Open Questions and Suggested Follow-Ups

**Ambiguities not resolved in this pass:**

1. **The 47 Found Incompetent signatures on 2023-05-30 by Borer** (G2-02) need document-level audit: are these the same clone-corpus incompetency orders Theme 9.2 describes (zero-unique-text, shared tracking fonts, identical creation timestamps)? Lane R already showed Borer is in the "Referee Competency Advisor" cluster; the specific 47-document cohort deserves row-level forensic match against the 157-order clone corpus.

2. **The 20 non-null rows in `parents_hearings_previous_judicial_officers`** represent a sidecar split of multi-officer hearings whose parent table carries NULL norm/id. Whether this denormalization failure affects other counts in the baseline is unresolved; our audit found the 2 count-altering IDs but did not trace forward to see whether any top-10 officer has multi-hearing presence underreported.

3. **The 196-officer surplus implies a different investigative frame.** If the target list was "current Hennepin active bench" rather than "all MCRO-appearance officers," then questions like "what fraction of competency orders flow through named target judges" need rephrasing. An expanded target list including Borer, Quam, Norris, Carruthers, Allyn, Cahill, Lien, Poston, Mulrooney, and Barnette (the top-10 surplus) would raise matched-coverage from 80.7% to an estimated 91.4% of volume.

4. **Hudleston's administrative-buffer role** (N2-03) deserves a dedicated paper: 2.51× inbound from the clique is not the only metric; her April 29, 2025 reversal of the Koch competency order fits the corpus-wide Koch → Hudleston → Dayton Klein pattern (n=8 documented). Does the Guertin reversal sit inside or outside the buffer pattern?

5. **The Mercurio → Dayton Klein 34-transition directed edge** (N-04) is the single largest in the corpus. Are the 34 transitions bundled in time (a single reassignment wave) or distributed (continuous administrative rhythm)? The temporal distribution would distinguish "one-time reorganization" from "standing operational pattern."

**Out-of-scope for this run, flagged for the next pass:**

- PKI certificate-chain analysis on the `edits_after_sig_flag = true` rows (there are many; integrity implications are substantial but belong in the forensic-signature workstream).
- Content analysis of the 47 Borer May 30 2023 orders — linguistic fingerprinting is owned by the forensic-language workstream.
- Corpus expansion beyond 2,903 parent cases — whether the competency pipeline patterns hold at Hennepin-wide scale (estimated 15–30x larger) is unknown.

---

## 6. Appendices

### Appendix A — Cluster Assignment (k=3 and k=7)

Source: Lane R (Agent-R). Full table at `/tmp/agent_r_cluster_assign.csv`. Summary:

| k=3 Cluster | Archetype | n |
|---|---|---|
| 0 | Referee Competency Advisor | 9 |
| 1 | Medical Records Gatekeeper | 8 |
| 2 | General Criminal Presider | 59 |

| k=7 Cluster | Archetype | n |
|---|---|---|
| 0 | Medical Records Gatekeeper (Bartsh) | 1 |
| 1 | Interpreter-Heavy (Hughey) | 1 |
| 2 | General Criminal Presider (misc) | 2 |
| 3 | General Criminal Presider | 49 |
| 4 | Medical Records Gatekeeper | 5 |
| 5 | Rule-20.01 Competency Signer | 9 |
| 6 | Referee Competency Advisor | 9 |

### Appendix B — Top-20 Shared-Case Edges

Source: Lane N. Edge weight = distinct `case_uid` count shared by both officers in `parents_case_events_judicial_officers`.

| Rank | Officer A | Officer B | Shared cases | %CR | %MH | %GC |
|---|---|---|---|---|---|---|
| 1 | Michael Browne | Julia Dayton Klein | 477 | 67.7 | 28.7 | 3.6 |
| 2 | Danielle Mercurio | Julia Dayton Klein | 437 | 75.3 | 24.0 | 0.7 |
| 3 | Michael Browne | Danielle Mercurio | 348 | 62.4 | 35.1 | 2.6 |
| 4 | William Koch | Julia Dayton Klein | 299 | 78.6 | 20.7 | 0.7 |
| 5 | Lisa Janzen | Julia Dayton Klein | 288 | 100.0 | 0.0 | 0.0 |
| 6 | Lori Skibbie | Julia Dayton Klein | 269 | 63.2 | 34.6 | 2.2 |
| 7 | Michael Browne | George Borer | 249 | 61.8 | 34.1 | 4.0 |
| 8 | George Borer | Julia Dayton Klein | 244 | 68.9 | 29.1 | 2.0 |
| 9 | William Koch | Danielle Mercurio | 235 | 80.9 | 18.3 | 0.9 |
| 10 | Michael Browne | Lori Skibbie | 235 | 54.9 | 42.1 | 3.0 |
| 11 | Michael Browne | Lisa Janzen | 226 | 100.0 | 0.0 | 0.0 |
| 12 | Danielle Mercurio | Lori Skibbie | 216 | 59.7 | 38.9 | 1.4 |
| 13 | Joel Olson | Julia Dayton Klein | 192 | 78.6 | 19.8 | 1.6 |
| 14 | Lisa Janzen | Danielle Mercurio | 190 | 100.0 | 0.0 | 0.0 |
| 15 | Danielle Mercurio | George Borer | 181 | 64.1 | 34.3 | 1.7 |
| 16 | Michael Browne | William Koch | 181 | 82.3 | 14.9 | 2.8 |
| 17 | William Koch | Lori Skibbie | 153 | 73.9 | 24.8 | 1.3 |
| 18 | Carolina Lamas | Lisa Janzen | 137 | 100.0 | 0.0 | 0.0 |
| 19 | George Borer | Lori Skibbie | 136 | 39.7 | 57.4 | 2.9 |
| 20 | William Koch | Joel Olson | 136 | 80.9 | 18.4 | 0.7 |

### Appendix C — Top-20 Signer-vs-Docket Divergence Cases

Source: Lane S. Most recent first. **Row 16 is the Guertin-case January 17, 2024 FOI** — expanded in §3.4 / S2-01.

| # | case_uid | Event | Docket officer | Signer | Filing date |
|---|---|---|---|---|---|
| 1 | a25-0965 | Order - Dismiss - Not Stipulated | Jennifer Frisch | Elizabeth Lucid | 2025-08-22 |
| 2 | a25-1136 | Order - Granting/Denying Mandamus | Jennifer Frisch | Elizabeth Lucid | 2025-08-19 |
| 3 | a25-1314 | Order - Dismiss - Not Stipulated | Jennifer Frisch | Christa Rutherford-Block | 2025-08-14 |
| 4 | 27-cr-17-8342 | Order Denying Motion | Julia Dayton Klein | George Borer | 2024-04-11 |
| 5 | 27-cr-17-1555 | Order Denying Motion | Julia Dayton Klein | George Borer | 2024-04-11 |
| 6 | 27-cr-19-3539 | Finding of Incompetency | George Borer | Michael Browne | 2024-03-20 |
| 7 | 27-cr-21-22058 | Finding of Incompetency | George Borer | Michael Browne | 2024-03-20 |
| 8 | 27-cr-19-1916 | Finding of Incompetency | George Borer | Michael Browne | 2024-03-20 |
| 9 | 27-cr-18-18391 | Finding of Incompetency | George Borer | Michael Browne | 2024-03-20 |
| 10 | 27-cr-19-17539 | Finding of Incompetency | George Borer | Michael Browne | 2024-03-20 |
| 11 | 27-cr-22-19036 | Finding of Incompetency | Danielle Mercurio | Michael Browne | 2024-03-11 |
| 12 | 27-cr-23-21653 | Amended Order | Danielle Mercurio | Julia Dayton Klein | 2024-02-13 |
| 13 | 27-cr-23-3198 | Order-Other | George Borer | Michael Browne | 2024-02-09 |
| 14 | 27-cr-20-26577 | Amended Order | Julia Dayton Klein | Danielle Mercurio | 2024-01-19 |
| 15 | 27-cr-22-20527 | Order Denying Motion | Julia Dayton Klein | Danielle Mercurio | 2024-01-19 |
| **16** | **27-cr-23-1886** | **Finding of Incompetency** | **Danielle Mercurio** | **Julia Dayton Klein** | **2024-01-17** |
| 17 | 27-cr-23-3198 | Finding of Incompetency | Lori Skibbie | Julia Dayton Klein | 2023-12-11 |
| 18 | 27-cr-23-3198 | Finding of Incompetency | Lori Skibbie | Julia Dayton Klein | 2023-12-06 |
| 19 | 27-cr-22-18859 | Order-Evaluation (Rule 20.01) | George Borer | Julia Dayton Klein | 2023-12-06 |
| 20 | 27-cr-22-22985 | Order-Evaluation (Rule 20.01) | George Borer | Julia Dayton Klein | 2023-12-06 |

### Appendix D — Top-30 Surplus Officers (not on 87-name target list)

Source: Lane G2. Ranked by total master-pool appearances across 13 JO-bearing fields.

| # | Display Norm | Appearances | First Event | Last Event | Top Event Name | Top % | Role Category |
|---|---|---|---|---|---|---|---|
| 1 | George Borer | 5,830 | 2013-06-13 | 2024-04-15 | Taken Under Advisement | 25.3 | Competency referee |
| 2 | Jay Quam | 1,660 | 2010-01-12 | 2025-03-07 | Order for Conditional Release | 9.4 | Active trial bench |
| 3 | Lyonel Norris | 1,119 | 2021-04-30 | 2023-01-31 | Order for Conditional Release | 35.6 | Conditional-release |
| 4 | Philip Carruthers | 1,052 | 2012-12-11 | 2025-04-09 | Order for Production of Medical Records | 20.5 | Medical-records gatekeeper |
| 5 | Julie Allyn | 872 | 2021-01-25 | 2024-04-15 | Notice of Case Reassignment | 16.8 | Reassignment admin |
| 6 | Peter Cahill | 854 | 2013-03-13 | 2024-11-26 | Order Granting Public Defender | 14.6 | Active trial bench (Chauvin judge) |
| 7 | Anthony Schumacher | 545 | 2004-10-04 | 2018-08-23 | Other Document | 14.2 | Retired / pre-2020 |
| 8 | Mike Lien | 478 | 2014-08-01 | 2021-04-09 | Taken Under Advisement | 34.3 | Competency referee |
| 9 | Janet Poston | 423 | 2012-07-23 | 2025-10-24 | Taken Under Advisement | 20.4 | Competency referee |
| 10 | Nicole Engisch | 413 | 2019-06-07 | 2022-12-06 | Order for Conditional Release | 21.9 | Conditional-release |
| 11 | Timothy Mulrooney | 409 | 2019-05-17 | 2025-12-05 | Found Incompetent | 15.3 | Competency referee (Rule 20.01) |
| 12 | Toddrick Barnette | 377 | 2009-06-24 | 2023-08-15 | Order-Evaluation Rule 20.01 | 15.8 | Competency referee (former Chief Judge) |
| 13 | Patrick Meade | 319 | 2004-04-02 | 2014-03-12 | Other Document | 22.7 | Retired / pre-2020 |
| 14 | Timothy Carey | 294 | 2022-07-27 | 2025-12-03 | Order for Conditional Release | 11.7 | Competency referee (Rule 20.01) |
| 15 | Martha Holton Dimick | 246 | 2013-01-11 | 2021-12-09 | Order for Conditional Release | 19.6 | Retired / pre-2022 |
| 16 | Susan Robiner | 224 | 2012-12-17 | 2024-07-10 | Order Granting Public Defender | 29.4 | Active trial bench |
| 17 | Kathleen Gearin | 171 | 2021-12-03 | 2024-07-05 | Taken Under Advisement | 43.1 | Competency referee |
| 18 | James Moore | 171 | 2013-06-19 | 2023-12-13 | Order-Evaluation Rule 20.01 | 34.4 | Competency referee |
| 19 | Leslie Metzen | 164 | 2016-03-14 | 2025-11-12 | Order for Conditional Release | 36.8 | Conditional-release |
| 20 | Herbert Lefler | 162 | 2021-09-28 | 2025-12-03 | Order Granting Public Defender | 40.7 | Active trial bench |
| 21 | Fred Karasov | 150 | 2013-12-11 | 2019-06-10 | Request for Interpreter | 20.8 | Retired / pre-2020 |
| 22 | Marilyn Kaman | 141 | 2007-09-21 | 2013-11-06 | Order for Production of Medical Records | 19.8 | Medical-records gatekeeper (retired) |
| 23 | Shawn Bartsh | 141 | 2021-01-28 | 2025-05-15 | Found Incompetent | 17.5 | Competency referee |
| 24 | H. Albrecht | 134 | 2004-09-20 | 2016-04-25 | Order for Production of Medical Records | 21.7 | Medical-records gatekeeper (retired) |
| 25 | Paul Yang | 125 | 2019-06-01 | 2025-12-12 | Order for Production of Medical Records | 16.0 | Medical-records gatekeeper |
| 26 | Margaret Daly | 115 | 2006-04-18 | 2024-06-24 | Order-Evaluation Rule 20.01 | 13.9 | Competency referee |
| 27 | Joy Bartscher | 112 | 2020-03-11 | 2024-01-29 | Found Incompetent | 18.9 | Competency referee |
| 28 | Amy Dawson | 104 | 2015-11-23 | 2024-03-06 | Order Granting Public Defender | 24.2 | Active trial bench |
| 29 | Richard Kyle Jr. | 83 | 2016-09-29 | 2025-08-20 | Order for Conditional Release | 23.3 | Active trial bench |
| 30 | Leonardo Castro | 79 | 2021-11-22 | 2022-09-14 | Request for Interpreter | 33.3 | Misc / interpreter |

### Appendix E — Fuzzy-Match Ledger for 11 Absent Targets

Source: Lane G. All searches returned zero matches at surname-edit-distance ≤ 2; best candidates shown are coincidental.

| Target | Best candidate | Min distance | Interpretation |
|---|---|---|---|
| Chaffee, Rebecca | Kellie Charles | 3 | Coincidental tail overlap |
| Christy, Snow-Kaster | — | >4 | No near match |
| Garfinkel, Naomi | — | >4 | No near match |
| Hutchison, Jason | Natalie E. Hudson | 4 | Partial surname tail only |
| Knight, Holly | William Koch | 4 | No near match |
| Madden, Mary | Mary Mahler | 3 (first name only) | First-name match only |
| Nguyen, Ngoc | — | >4 | Strongest rejection |
| Ogunleye, Josh | Thomas Conley | 4 | No near match |
| Reed, Franklin | — | 3 | All coincidental |
| Sedillos, Tiffany | Angela Willms | 4 | No near match |
| Stebbins, Richard | Lori Skibbie | 4 | Trigram collision |

All 11 absences verified as structural gaps, not normalization failures.

### Appendix F — Revised Baseline Count Reconciliation

| Baseline claim | Revised finding | Confidence |
|---|---|---|
| 76 of 87 target officers matched | 76 valid matches (11 confirmed absent) | HIGH |
| 275 distinct officer IDs in dataset | 277 (add 2 sidecar IDs) | HIGH |
| 275 IDs = 275 real humans | 274 real humans (subtract 3 phantoms, add 2 sidecar) | HIGH |
| 1 "M. Regis" normalization artifact | 5 artifacts of same class | HIGH |
| Ryan Gallia flagged ambiguous | Unambiguous — single candidate in pool | HIGH |
| Target list is the officer universe | 196 surplus officers in data; 19.3% of JO volume unlisted | HIGH |
| Top-3 officers account for 32,576 appearances | Correct; but Borer (5,830) is #4, unlisted | HIGH |

---

*This deep-dive is an analytical extension of the baseline Judicial Officer Representation Analysis. It does not replace it; the baseline's descriptive inventory remains the authoritative floor for per-officer appearance counts. Every inferential claim here is traceable to its SQL.*