Five Counts, Three Findings: The Hartig Neutral’s Report and What It Actually Says About Judicial Discipline

The Procedural History That Got Here

Hartig was elected to the 52-4 District Court in 2010 and took the bench in January 2011. She was re-elected in 2016 and again in 2022. The first Request for Investigation landed in July 2020, filed jointly by then-Chief Judge Joseph Fabrizio and SCAO Regional Administrator Jennifer Phillips after Oakland County Human Resources concluded that Hartig’s treatment of court administrator Dana O’Neal had created a hostile work environment. A second RFI followed in 2022, filed by the Oakland County Prosecutor’s Office.

The JTC filed a four-count formal complaint on June 4, 2025. Pretrial proceedings produced significant procedural friction: Disciplinary Counsel moved to add a count and compel a further medical evaluation, the Neutral denied both requests, the Michigan Supreme Court reversed, and the case was rescheduled. The Amended Complaint filed on January 9, 2026 carried five counts. Trial began February 2, 2026.

The Neutral’s Report is the product of that process. It is a fact-finding document, not a final disposition. The Michigan Supreme Court receives it and decides what happens next.

Who Is the Neutral

The Michigan Supreme Court appointed retired 25th Circuit Court Judge Jennifer A. Mazzuchi (P51418) to serve as Neutral in this matter. She has no documented connection to Oakland County, the 52nd District Court, the Oakland County Prosecutor’s Office, or any of the named parties in the proceeding.

Mazzuchi spent her career in the Upper Peninsula. She earned her undergraduate degree from the University of Michigan and her law degree from Wayne State University, then ran a civil practice in Detroit until 1997. She joined the Marquette County Prosecutor’s Office that year and stayed for over a decade before being elected to the 25th Circuit Court bench in 2008, where she handled criminal and family cases. She served as Chief Circuit Court Judge from 2016 to 2023 and retired from the bench November 1, 2023, becoming the first female judge to have her portrait hung in the Marquette County Circuit Courtroom.

Her background is prosecution and criminal bench work, in a jurisdiction with no institutional overlap with the parties before her. The Supreme Court’s appointment of a retired UP circuit judge to preside over an Oakland County disciplinary matter is a standard insulation mechanism. In this case, it produced a Neutral whose geographic and professional remove from the parties is clean.

That context matters when evaluating the Report. Mazzuchi’s findings on Count Five, specifically her rejection of the disability theory and her observation that Chief Judge Reeds appeared to harbor a negative opinion of Hartig independent of any cognitive concern, are not the product of institutional loyalty to Oakland County’s court structure. She called the behavioral evidence as she saw it from a distance.

Count One: Withholding the APN Report

Count One is the cleanest count in the complaint. There is not much factual dispute. In October 2023, the JTC requested that Hartig undergo a mental health evaluation through the State Bar’s Lawyers and Judges Assistance Program. She did so. The LJAP report, which both parties received, suggested she undergo further evaluation at All Points North, a Colorado treatment facility. The JTC ordered that evaluation in April 2024. Hartig complied in May 2024 and received the APN report in early June.

APN sent the report only to Hartig. The JTC requested a copy. Through counsel, Hartig objected. After multiple rounds of communication, the JTC gave her a production deadline of July 25, 2024. She continued to object and did not produce the report until December 5, 2024, five months after the deadline.

The Neutral found the reason for the withholding apparent. The APN report had diagnosed a personality disorder and deemed Hartig “unsafe to practice.” The Neutral acknowledged that the APN methodology raised legitimate questions: despite no substance abuse history, the evaluation included drug and alcohol testing; the report repeatedly referred to Hartig as a physician rather than a judge; and the choice of an out-of-state facility for a mental fitness evaluation warranted scrutiny. Those objections go to the report’s validity. They do not answer whether she was required to produce it.

The Neutral also dispatched Hartig’s legal argument that MCR 9.220(D)’s use of “a” evaluation limits the JTC to ordering only one. The argument has a certain grammatical cleverness to it. The Neutral rejected it on grounds of unreasonable results: a construction that would force the JTC to choose between evaluating substance abuse or mental health issues, rather than both, does not serve the rule’s stated purpose of preserving judicial integrity and protecting the public. The singular article does not mean “one and only one” in this context.

Hartig also argued that because MCR 9.220(D) incorporates MCR 2.311(C), which provides a process for court-ordered enforcement of report production, the JTC’s failure to pursue that enforcement was dispositive. The Neutral found this unpersuasive. The JTC did indicate in correspondence that it would seek enforcement under 2.311, but ultimately did not. The Neutral’s response: the question is what Hartig was required to do, not whether there was a secondary mechanism the JTC chose not to deploy.

What the APN Report Actually Says and What the Neutral Did Not Reckon With

The Neutral acknowledged legitimate concerns about APN’s methodology. Hartig had no history of substance abuse, yet the evaluation included drug and alcohol testing. The report repeatedly referred to her as a physician. The choice of a Colorado facility for a Michigan judicial fitness evaluation raised questions. These concerns, the Neutral found, pertain to the report’s validity, not to whether she was required to produce it. That is the right analytical call on Count One.

What the Neutral did not do is examine what the APN report actually concluded and what those conclusions mean for a sitting judge. APN diagnosed a personality disorder and deemed Hartig “unsafe to practice.” The Neutral references these findings only to explain why Hartig withheld the report and to credit her for completing APN’s recommended treatment. That is where the analysis stops. It does not engage the substance of either finding. It does not ask what a personality disorder diagnosis means for judicial fitness, whether the “unsafe to practice” language carries clinical or legal weight under Michigan’s framework, or whether the disability allegation in Count Five might have been better served by engaging the personality disorder theory rather than relying solely on the mild cognitive impairment and frontotemporal dementia theories the experts contested.

That gap matters. Personality disorder is a distinct clinical category from mild cognitive impairment or emerging dementia. The experts in Count Five testified about the latter two. The personality disorder finding from APN entered the record and then largely disappeared from the analysis. The Neutral used it to establish motive for withholding and moved on. The result is a Report that credits Hartig for following clinical advice from a facility whose diagnosis it declined to examine. You cannot have it both ways. If APN’s methodology was sufficiently compromised that the diagnosis deserves no analytical weight, then the treatment compliance credit is also hollow. If the treatment completion matters, then the diagnosis that generated the treatment recommendation deserves more than a footnote.

The Neutral’s specific validation of Hartig’s objection on the substance abuse testing ground is where this gets thinner still. The Report states that it is reasonable to question why drug and alcohol testing was administered given that Hartig had no history of substance abuse. I find that reasoning difficult to sustain. The absence of a documented prior history is not a clinical or investigative reason to omit substance screening from a mental fitness evaluation ordered for a person in a position of public trust. The public record on that point is not abstract.

Federal Judge Thomas Ludington was arrested in Emmet County in October 2025 with a blood alcohol content of 0.27, more than three times the legal limit, after crashing his car into two traffic signs in northern Michigan. No prior documented history of substance abuse. At sentencing, his defense lawyer cited an evaluation by a court-approved behavioral health therapist who determined Ludington did not meet the criteria for an alcohol abuse disorder. He was then placed on probation with twice-weekly alcohol and drug testing requirements, which he allegedly failed to complete in his first week. Emmet County sits in the same northern Michigan judicial geography as Mazzuchi’s former jurisdiction. The pattern of judges presenting with no documented history and then exhibiting behavior that triggers mandatory testing and monitoring is not hypothetical. It is in the recent record of the state whose disciplinary framework she was applying.

The Neutral is right that Hartig’s concerns about APN’s methodology are not groundless. An out-of-state facility that calls a judge “Dr.” throughout its evaluation and cannot get the subject’s professional category correct has a credibility problem. But “the methodology is questionable” and “substance screening was unreasonable because she had no prior history” are two different arguments. The first has traction. The second does not hold up against what Michigan judges have actually done with no prior history on record.

None of this resolves Count Five in Hartig’s favor or against her. The personality disorder finding from a methodologically compromised evaluation is not automatically dispositive. But the Report’s silence on what that diagnosis means, on whether it should have been the theory of Count Five rather than or alongside mild cognitive impairment, and on why no documented prior history insulates a judicial officer from substance screening in a mental fitness context, is a gap that the Michigan Supreme Court and any competent appellate reader of this Record will notice.

Count Two: The False Statements That Were Not “False” Enough

Count Two is the most technically interesting count, and the Neutral’s analysis here has implications well beyond this case.

On December 5, 2024, Hartig’s response to the JTC’s 28-day letter contained two statements that were, as a factual matter, inaccurate. First, she attributed the original Request for Investigation to Dana O’Neal, saying O’Neal had been “subsequently fired for incompetence” by Fabrizio. In fact, the RFI was filed by Fabrizio and Phillips. Hartig had documentation confirming this. Second, she stated that none of her staff had been interviewed as part of the investigation. In fact, staff member Ann Costigan had been interviewed, and Hartig had attached Costigan’s own affidavit describing the interview in a prior filing.

The Neutral found both statements clearly inaccurate. The analysis, however, did not end there. The Michigan Supreme Court, in In re Gorcyca, established that “false” or “misleading” statements in the judicial discipline context require a showing of wrongful intent. The common definitions require that a misrepresentation “generally include an actual intent to deceive.” The audience matters. If the audience cannot be misled or misdirected by a statement, the intent element fails.

The Neutral found both statements careless but not wrongful. Attributing the RFI to O’Neal was inaccurate, but O’Neal was the primary source of the complaints to OCHR that prompted the filing. In context, Hartig was attacking the credibility of the investigation’s origin, not attempting to deceive the JTC about something it did not know. The staff interview statement was characterized as advocacy against the quality of the investigation, not a calculated deception. The Neutral’s framing: arguing against the strength of evidence in your own defense is not “wrongful” intent, even when you do it inaccurately.

The O’Neal record also deserves more precision than Hartig’s letter gave it. O’Neal was not fired for incompetence. According to her federal lawsuit filed in the Eastern District of Michigan (2:22-cv-10470), she was terminated in December 2021 after keeping documented records of Hartig’s conduct at the recommendation of SCAO, while the JTC investigation remained open. The lawsuit alleged Whistleblower Protection Act violations and First Amendment retaliation. Fabrizio, the named defendant, settled the case in 2023. The hostile work environment finding that preceded all of this was not contested at trial: Oakland County Human Resources investigated, interviewed O’Neal for three hours, and concluded a hostile work environment existed. That finding prompted Fabrizio to report to the JTC in the first place. It is in the press record, in the lawsuit, and in the Neutral’s Report itself. Hartig’s 28-day letter called O’Neal fired for incompetence. The documented record says something different.

This is worth sitting with. A judge made two provably inaccurate statements in a formal submission to the body investigating her conduct. The Neutral concluded that this does not constitute sanctionable misconduct because the recipient already knew better. The rule, as interpreted here, is calibrated to deception that functions, not inaccuracy for its own sake.

Hartig and her counsel would not be the first attorneys to submit provably inaccurate statements in a formal proceeding. Clutch finds it in the record daily. What makes this particular instance worth examining is not the act itself but what it reveals about institutional culture: when inaccurate characterizations of documented facts travel this far into formal submissions without correction, it stops being a proofreading problem and starts being a pattern. The Neutral declined to call it misconduct. The record still calls it inaccurate.

Count Three: Fifteen Years of Documented Conduct

Count Three is the largest count in the complaint and consumed the majority of trial time. It covers two broad categories: Hartig’s persistent refusal to defer to the administrative chain of command, and her discourteous treatment of colleagues and court employees across a period stretching back to 2011.

The hostile work environment finding at the center of this count was not a contested characterization invented by the JTC. Oakland County Human Resources investigated, conducted a multi-hour interview with O’Neal, and formally concluded a hostile work environment existed. That finding is documented in O’Neal’s federal whistleblower lawsuit (E.D. Mich. 2:22-cv-10470), in Detroit News coverage from March 2022 and May 2025, and in the Neutral’s Report itself. Press coverage of the proceedings documented the pattern extensively across multiple outlets before the formal complaint was ever filed. The JTC did not create this record. The record created the JTC complaint.

The Neutral’s analysis required understanding the structure of the 52nd District Court. It has four divisions, ten judges, one chief judge for the entire court, and separate court administrators for each division. Individual judges supervise their own chambers staff. All other courthouse departments fall under the court administrator and chief judge. Hartig spent fifteen years resisting that structure.

The record documented a persistent pattern. In 2013, she wrote to SCAO complaining about leadership failures and her inability to effect change. In 2015, she criticized the chief judge court rule at a statewide judicial conference in a manner her colleagues recalled as targeting her own chief judge. In 2019, she wrote directly to State Court Administrator Milt Mack and Chief Justice Bridget McCormack to complain about Chief Judge Fabrizio’s performance. Her benchmate, Judge Maureen McGinnis, described receiving a “gigantic pile” of emails and described Hartig’s conduct as so severe she questioned her own ability to continue working there.

On the specific incident allegations in paragraph 31(a) through (n), the Neutral addressed each individually. She found sufficient proof on most and insufficient on a few. The ones she proved included: Hartig issuing a unilateral “Communication of Absence Policy” demanding O’Neal check in daily “unless she was unconscious”; berating probation supervisor Patti Bates and O’Neal on the record during a drug court session in front of litigants; criticizing O’Neal for not responding to three emails sent while O’Neal was attending a funeral; telling a public defender that his argument made her “want to put a hatchet in your neck”; directing pregnant probation officer Sidorella Arapi to stand up on a Zoom call and display her pregnant belly to the team; and responding to staff concerns about two probation officers by declaring them “persona non grata.”

The Neutral found insufficient proof on three: the allegation that Hartig “insisted” Palulian hire her friend Ann Costigan (the evidence supported a strong recommendation, not a demand); the allegation that she “disrupted” a bench meeting over staff coverage (it was McGinnis who terminated the meeting); and the Arapi belly incident (the Neutral found it misguided but described it as a sincere, if deeply inappropriate, attempt to celebrate good news).

The Neutral’s most analytically important observation on Count Three addresses the relationship between Counts Three and Five directly. Disciplinary Counsel argued that the behavioral conduct in Count Three was evidence of cognitive impairment. The Neutral rejected this framing: if the behaviors stem from cognitive impairment, they are symptoms of an illness, not intentional misconduct. In finding Count Three proved, the Neutral implicitly concluded the conduct was volitional, not pathological.

Count Four: Good Faith in the Face of a Wrong Answer

Count Four addresses Hartig’s dismissals with prejudice in People v. Price and the Armstrong cases, all arising from the same January 2022 state law day docket. The context matters: Hartig had a difficult relationship with the Oakland County Prosecutor’s Office, specifically over the long-standing practice of scheduling criminal matters on Tuesdays. OCPO treated non-Tuesday scheduling as an imposition requiring accommodation that Hartig viewed as assertions of control over her docket.

In Price, a preliminary examination was repeatedly adjourned. On January 11, 2022, the APA appeared but did not produce witnesses, citing a Covid-positive witness. Hartig denied the adjournment request, and when the defense moved for dismissal, she sua sponte raised the possibility of dismissal with prejudice. She directed briefing, and ultimately issued the order. On appeal, the Circuit Court reversed: once Hartig had issued the initial non-specific dismissal order, jurisdiction was gone, and the subsequent dismissal with prejudice was improper.

In Armstrong, three incarcerated defendants appeared on January 31 for a preliminary examination that OCPO skipped, because the date was a Monday rather than a Tuesday. OCPO sent no representative and filed no appearance. Defense attorneys described OCPO’s conduct as “outrageous” and “contemptuous” and requested dismissal with prejudice. Hartig granted it. The Circuit Court reversed there as well, distinguishing the case Hartig had relied upon.

The Neutral found the dismissals erroneous. The legal analysis under MCR 9.211(B) then turns on whether the errors were committed in good faith.

The Neutral’s conclusion: Hartig genuinely believed she had the authority to enter the dismissals with prejudice. She sought briefing in Price. In Armstrong, defense counsel affirmatively requested the remedy and the prosecution did not appear. Her reliance on People v. Borowka was wrong, because Borowka involved a felony at trial stage, not a District Court preliminary examination. But the factual parallels, a prosecutor’s unilateral refusal to proceed after a denied adjournment, made the reliance understandable if mistaken. After the reversals, she complied. There was no pattern of continuing improper dismissals.

Count Five: The Disability Allegation and Why It Failed

Count Five is the most consequential count in the complaint, and the Neutral’s handling of it is the section most likely to define how this case is remembered.

The allegation was specific: that the results of Hartig’s neuropsychological evaluations, coupled with other evidence of her mental status, establish a mental disability preventing performance of her judicial duties, including reading and analyzing testimony, applying law to facts, making credibility determinations, and writing coherent opinions and orders.

By the time of trial, Hartig had undergone multiple evaluations: a 2024 LJAP evaluation, the 2024 APN evaluation (twice, before and after treatment), a September 2024 evaluation at Bright Pines, and a December 2025 evaluation by JTC expert Dr. Michael Lawrence. Respondent’s expert, Dr. Christian Schutte, evaluated her in January 2026. Reviewing expert Dr. Michael Wolff reviewed the materials without conducting his own examination.

The three experts agreed on the expected behavioral manifestations of mild cognitive decline and frontotemporal dementia: perseveration (fixating or getting stuck), disinhibition (failure to control behavior appropriate to circumstances), difficulty encoding new information, and deficits in executive functioning. They also agreed such symptoms would progress over time. Their opinions on whether Hartig had these conditions diverged sharply.

Wolff concluded Hartig suffers from mild cognitive impairment constituting a disability. He relied substantially on the behavioral evidence in Count Three and on Chief Judge Reeds’ written statement, which he found most significant. Lawrence found no generalized neurocognitive decline but identified deficits in language tasks and executive functioning suggesting an emerging frontotemporal dementia, and recommended a PET scan. The PET scan results were normal. Schutte found no neurocognitive disorder, criticized cross-test comparisons Wolff used as methodologically unsound, and argued the behavioral evidence was “encapsulated” in a way inconsistent with cognitive impairment.

The Neutral found Schutte’s testimony persuasive on a specific analytical point: if the behavioral manifestations of cognitive disability would be consistently observed, a pattern that appears selectively directed at certain people rather than broadly exhibited does not fit the expected profile. Hartig’s chambers staff testified favorably. Drug court team members described a thoughtful, engaged colleague. Friends and colleagues saw no cognitive decline. The negative conduct was concentrated among specific court employees, specifically administrators, probation staff, and those she perceived as subordinates who resisted her authority.

The Neutral also flagged that some evidence Wolff weighted heavily did not hold up on closer review. One instance Reeds characterized as Hartig forgetting testimony turned out, on transcript review, to be the prosecutor who had forgotten it. An email chain Reeds described as showing confusion turned out on closer reading not to support that characterization. Some of the “errors” cited as pathological were things the Neutral described as common among cognitively healthy people: attaching the wrong document, forgetting a name.

The Neutral also raised a credibility concern about Reeds as a behavioral witness. He had described Hartig’s conduct as treating staff like “a scullery maid, the lowest of the low,” and testified that when he ran for judicial office in 2014, he had practiced in front of Hartig and “I wouldn’t have wanted her endorsement even if she offered.” He attributed her landlord-tenant procedural approach to “laziness and lack of work ethic.” The Neutral found this degree of animus made it difficult to treat his behavioral observations as supportive of a disability theory rather than simply a low opinion of her conduct.

And then there was the testimony itself. Hartig testified for six full days, three and a half of which were under cross-examination by Disciplinary Counsel. The Neutral observed her throughout. She corrected attorneys when they misspoke, named numerous individuals referenced in the record, demonstrated familiarity with voluminous discovery materials, and showed clear recall throughout. She became emotional once, appropriately, when discussing the formal complaint being filed on the eve of her son’s wedding.

Likely Outcomes: What Prior JTC Cases Tell Us

The Michigan Supreme Court does not issue discipline. It reviews the JTC’s Decision and Recommendation after the Commission processes the Neutral’s Report, and then imposes whatever sanction it determines appropriate. The Court reviews the record de novo and may accept, reject, or modify the Commission’s recommendation entirely. In re Brown, 461 Mich 1291 (2000). It has done all three in recent history.

When the Court calibrates sanctions, it applies the Brown factors, a framework that weighs the nature of the misconduct, whether it was part of a pattern, whether it occurred on the bench, whether it harmed the administration of justice, the judge’s prior disciplinary history, and evidence of remediation. Misconduct that is part of a pattern or practice is treated as more serious than an isolated incident. Misconduct that involves dishonesty to the tribunal is treated as especially grave.

The precedent spectrum runs as follows. In In re Gorcyca, 500 Mich 588 (2017), the Court imposed public censure for a single incident of courtroom misconduct involving demeaning treatment of children in a custody hearing. The JTC had recommended a 30-day suspension without pay. The Court reduced it to censure, noting the absence of prior disciplinary history. In In re Simpson, 500 Mich 533 (2017), the Court imposed a nine-month suspension without pay for misconduct involving interference with a police investigation and misrepresentations to the JTC, finding that lying during disciplinary proceedings is particularly serious. In In re Morrow, the Court imposed a six-month suspension for a pattern of sexually graphic comments toward prosecutors, with prior lesser sanctions on the record. And in In re Brennan, the Court removed a judge from office for a pattern including false statements under oath, evidence tampering, and refusal to recuse. Removal sits at one end of the spectrum; public censure at the other.

Hartig’s proved record does not map cleanly onto any single precedent, which is how most cases work. What the Brown factors produce for her profile looks something like this. The Count Three conduct spans fifteen years and involves multiple chief judges, multiple court administrators, and multiple staff members across documented incidents. That is a pattern, not an incident. The Court treats patterns more seriously. On the other hand, three counts failed, including the disability theory and the false statements count. The Neutral specifically found Hartig capable of controlling and directing her behavior, and found documented evidence of remediation efforts in recent years. That cuts against the most severe sanctions.

Count One adds a compliance violation on top of the pattern: a direct non-compliance with a JTC order for five months. That is a different category from temperament misconduct. It is documented defiance of an investigative demand. The Court will weigh it separately.

What Hartig does not have on the proved record is dishonesty to the tribunal. Count Two failed. That is significant. The cases where the Court has reached for suspension rather than censure tend to involve either dishonesty during proceedings or conduct that directly harmed litigants. The Count Three conduct harmed court employees and made their work harder. The Neutral found the court still performed well. That finding, that Hartig failed to facilitate rather than obstructed, is the kind of distinction that matters in sanction calibration.

A reasonable read of the precedent suggests the range is public censure on the low end to a suspension of several months on the high end. Removal is not supported by the proved record. The false statements count that could have elevated the exposure failed. The disability count that could have supported retirement also failed. What remains is a fifteen-year pattern of interpersonal misconduct and a five-month compliance violation from a judge who is still sitting, was still handling a substantial caseload at the time of the Report, and has no prior formal public discipline on record.

The Court is also not bound by the Neutral’s framing. It reviews the record de novo. If the Commission, when it issues its own Decision and Recommendation, reaches different conclusions than the Neutral on any count, or recommends a harsher sanction than the record here would suggest, the Court will assess that too. The Neutral’s Report is the beginning of the sanctions phase, not the end.

What the Neutral’s Report Actually Settles and What It Does Not

The Report of Neutral is a findings document. It determines what was proved to the fact-finder at trial. It does not impose discipline. The Michigan Supreme Court receives the report and proceeds under its own authority.

On the two proved counts, the conduct is significant in duration and pattern. Count Three spans conduct from 2011 through 2023 and involves multiple chief judges, multiple court administrators, probation staff, an attorney, and a pregnant subordinate. The Neutral’s framing that Hartig “failed to facilitate” rather than “obstructed” administrative functions is a meaningful distinction for the sanctions analysis, but does not diminish the documented record.

Count One, the withholding of the APN report, is narrower in time but involves deliberate non-compliance with a direct JTC order over five months. The Neutral’s finding that the eventual production did not cure the violation matters: this is not characterized as a mistake or delay, but as misconduct.

The three counts that failed are each analytically significant in different ways. Count Two’s failure tells practitioners something about how the Gorcyca intent standard operates: provably false statements made to a recipient who knows the truth are not sanctionable misrepresentations. Count Four’s failure on good faith means erroneous legal conclusions, even significant ones, do not cross into misconduct absent evidence the judge knew the law and chose to defy it. Count Five’s failure is the most consequential: a five-count complaint built in significant part around a disability theory could not survive competing expert testimony and a behavioral record that the Neutral read as selective rather than global.

What the record does not settle is what the Michigan Supreme Court will make of it. The Court is not bound by the Neutral’s findings and may view the pattern in Count Three, combined with the Count One violation, as warranting substantial sanction. Or it may not. The proceedings are ongoing.