This post is part of Summary Judgment Motions and Evidence.
In an August 2025 hearing about discovery disputes, the court requested full briefing on my allegation that Harvard failed to preserve relevant evidence. My Motion for Sanctions for Spoliation of Evidence flags some of the evidence known to be lost. Highlights:
• Senior Associate Dean Paul Healy personally wiped his tablet, destroying the only contemporaneous notes of at least one key meeting. He erased the tablet eight months after emailing Dean Nohria about the likelihood of litigation and two months after receiving my increasingly formal claims both in person and by email. (More on timing in the next section.)
• Amy Edmondson, chair of the FRB, was by her own testimony “very disorganized” in email. She failed to retain and produce 68% of messages known to have been sent to or from her, relating to the FRB. In addition, she produced nine relevant emails from her Recoverable Items\Purges folder, indicating that someone not only deleted them but cleared them from Deleted Items after the litigation hold was in place. Indeed, Microsoft documentation indicates that messages in turn end up in Purges if a user manually purges them. To this, HBS offers a declaration from IT leader Christopher Pringle, and argues that HBS’ “background retention expiration policy” (not a manual command by Edmondson or anyone else) probably moved the messages to Purges. But even on Harvard’s telling, someone cleared these messages out of Edmondson’s Deleted Items folder (i.e. by emptying that folder) at some point after Harvard put Edmondson on a litigation hold. Once Edmondson was told to retain documents, she should not have cleared her Deleted Items folder.
The emails found in Purges portray Edmondson in a bad light: She received letter excerpts that were, on her own evaluation, overwhelmingly favorable to me (“these quotes are largely positive”), yet she omitted them from the FRB’s report. In addition, she attempted to turn Standing Committee discussion against me and expressed a personal preference against my candidacy (“hope that our report will alter…”), hoped I would see the “writing on the wall” and leave HBS, and regretted that I had procedural rights that I might use to my benefit (“unfortunately, our FRB policies give him a chance to respond”). (Russcol Aff. Attachments K and L.)
• FRB member Stu Gilson intentionally deleted every responsive email. His production, from five months on the FRB, contained exactly zero emails.
In response, Harvard argues that it retained enough, listing evidence it did retain. But this ignores the critical gaps created by spoliation. My Reply calls out the thin evidence of who said what in the Standing Committee deliberation about the FRB report—a pivotal moment where later notes indicate members were initially skeptical (“Who did we really talk to?”). The only official record is a one-page “addendum” summary claiming Len Schlesinger “provided additional context”—but the addendum directly contradicts Schlesinger’s testimony that he told the Standing Committee nothing beyond what was in the FRB report. With this incomplete and contradictory record, Paul Healy’s destroyed notes would have been the primary evidence to resolve what was actually said. The Standing Committee discussion is particularly important because I allege that committee should not have evaluated the FRB report at all. (See Amended Complaint ¶103-111 and Plaintiff’s Opp to MSJ 21.) Evidence of who said what would confirm the harm from that breach.
Another important gap comes from emails sent to a single recipient. If both participants deleted such an email, the message is permanently lost. For example, Edmondson’s 2015 discussions with Reinhardt largely survived because Reinhardt kept copies, but we have no record of similar discussions with Gilson. Given that Edmondson deleted 68% of her emails and Gilson deleted 100%, any private deliberations between them are now unavailable. To this, Harvard says we have no evidence that there were one-on-one discussions between Edmondson and Gilson. But there were Edmondson-Reinhardt one-on-one emails, and Gilson filled Reinhardt’s seat. And Edmondson and Gilson were the only two Appointments Committee members on the FRB, making certain subjects natural for them to discuss directly. Where Edmondson was so disorganized that less than a third of known messages remain, and Gilson intentionally deleted all messages without exception, I say inferences should be drawn against them about what messages they lost and destroyed.
Yet other messages were nearly lost when no custodian kept a copy, but the messages are nonetheless known because they were copied within other messages. By Harvard’s count (citing Athay Aff.), there are 20 messages like this.
As is typical in a spoliation case, I cannot know exactly what I do not have. However, the evidence recovered from Edmondson’s “Purges” folder—showing her mishandling evidence and lamenting my procedural rights—demonstrates that witnesses deleted significant emails. Indeed, Edmondson’s “Purges” messages provide “smoking gun” evidence of animus. I’m lucky that these messages were, just barely, preserved. I can only imagine what more is, alas, completely gone.
Timing of the duty to preserve evidence
A party’s obligation to preserve evidence arises as soon as litigation is “reasonably foreseeable.” We need not speculate on when Harvard should have foreseen litigation, because Harvard’s own leaders repeatedly memorialized their expectations of a lawsuit in writing:
- March-May 2018 Meetings and Emails. In a series of meetings and emails from March to May, I alerted both Healy and Nohria of my claims. Healy testified that he “assumed [I] might consider litigation” and “didn’t know whether [I] would go so far as … bring a lawsuit” but thought it was possible.”
- May 9, 2018 Healy Email. On May 9, Healy wrote to Nohria that I “indicated that [I am] seriously considering suing the school.” (source.)
- May-June 2018 Provost Correspondence. On May 31 and June 11, I wrote to then-Provost Alan Garber explaining that “rules” had been “violated” and “procedural commitments … weren’t followed.” (source.) My June 19 follow-up quoted and analyzed the P&P and identified “violations” of Harvard’s “commitments.” (transmission email, memo.)
Inexplicably, Harvard did not initiate a litigation hold until August 2021—more than three years later—and even then, only for a small fraction of custodians (including just one FRB member). The other four FRB members (and various other custodians) were not placed on a hold until February 2023, after I filed this lawsuit. (Athay letter, July 16, 2025.)
Harvard’s “internal university review” defense
In its Opposition, Harvard argues that my March to June 2018 communications indicated only that I was complaining to the Provost, but not contemplating litigation. Harvard’s Opposition oddly ignores the 2015 and 2017 documents in which Harvard witnesses themselves discussed—and memorialized in writing—the prospect that I “will sue” and “that there will [be] a legal follow up.”
Even worse, Harvard ignores Healy’s May 9, 2018 email to Nohria in which he says I “indicated that [I am] seriously considering suing the school.” Healy is right that that’s what I said. With that statement, contemporaneously memorialized by Healy to Nohria, how can Harvard argue I was only planning to complain to the Provost and not planning litigation?
Quantifying the harm
The impact of Harvard’s multi-year delay is not speculative; it can be mathematically estimated from Amy Edmondson’s production. The data shows a clear correlation between the passage of time and the loss of evidence: For emails from the 2015 FRB, Edmondson’s still had 27% as of the August 2021 start of litigation hold. For emails from the 2017 FRB, two years more recent, 40%. This analysis supports the unremarkable proposition that emails tend to get lost over time. The best fit for these numbers is a geometric decay of about 20% per year. (Calculations: After four years, at a 20% decay rate, a person would have just (1-0.2)^4=41% of messages left, while after six years, a person would have (1-0.2)^6=26% left. Those two numbers are both within 1% of what we observe from Edmondson’s production.)
If 20% of messages get lost each year, on average, it’s obviously important to preserve as early as possible. It is correspondingly harmful that Harvard inexplicably delayed activating a litigation hold for years, as described above.