Praxis / case study
Coordinated Harassment in Hakuouki Fandom
A closed case study (2024–2026) of coordinated harassment, defamation, and platform inaction across Tumblr, AO3, and adjacent communities. AO3 ruled against the appeal; the author has since withdrawn from the platform and removed all works hosted there. The materials below are preserved for accountability, citation, and as a reference for others navigating similar situations.
Overview
Between September 2024 and March 2026, a coordinated group of five Anglophone accounts on Tumblr, AO3, and Google Docs conducted a sustained harassment campaign against the author’s Hakuouki fan writing. The campaign spans 18 months and roughly 5,700 archived posts; attackers deployed a specific cross-linguistic tactic — reaching into the target’s native Chinese for family- and ancestor-targeted curses while writing predominantly in English — to exploit the target’s identity-of-access as a non-native English-speaking fan writer located inside the Chinese Great Firewall.
Platform governance during the active period (Tumblr report backlog, AO3 Policy & Abuse Committee ruling, Bluesky appeal bounce) is documented in the Forensic Brief and Platform Tickets below.
What this case uniquely documents: (a) the first publicly-curated syntactic corpus of coordinated cross-linguistic harassment targeting a fan writer, paired with (b) a syntax-first analyzer that surfaces per-actor rhetorical signatures — bureaucratic- polite DARVO, direct-insult, supporting-cast personae — as quantitative rather than purely interpretive claims. The analyzer is released as a standalone software artifact; quantified signatures against this case’s corpora are visible live at the hate-speech monitor.
The method-level reflection on this case — how records become evidence, why evidence does not automatically become public knowledge, and how the platform’s Terms of Service clause Section II.H pre-loads a target’s documentation as offense — is developed as a separate method note at From Records to Public Knowledge. Concepts named there (witness infrastructure, adverse reading regime, asymmetric evidentiary privilege, bookmark substrate) are accountable to the materials archived on this page; the method note can be read as the theoretical layer of what the case page documents at the material layer.
Content warning & positionality
Exhibits linked below include harassment content: direct insults, body- and identity-targeting slurs, sexual humiliation, pathologizing diagnoses, manipulated or captioned imagery, and cross-linguistic identity-targeted curses in Chinese. The Forensic Brief and Raw Evidence archive reference this material verbatim for documentary purposes.
This case study is authored by the harassment target. It is autoethnographic forensics, not detached journalism. The analysis framework is explicit about this positionality: the author has unique access to attacker intent that a third-party researcher would not, and correspondingly carries analytic bias that third-party annotation would not. Readers should weigh findings accordingly.
Case Files
A navigable catalog of exhibits with permanent URLs for citation.
Full Report
50 chapters · ~33k wordsThe complete first-person case documentation, originally published on AO3 and now removed. The most comprehensive single record of the case.
View →Forensic Brief
PDF · 60 pages · 10.7 MBAnalysis of public Tumblr-homepage statements through July 29, 2025.
Download ↓Platform Tickets
ZIP · 11.8 MBArchive of filed reports and platform responses.
Download ↓Raw Evidence Browser
217 files · 6 foldersSearchable index of unprocessed source materials and screenshots.
View →Medium Articles
ZIP · 4 articles + 1 takedown notice · 67 MBLong-form articles published on Medium in August 2025 documenting the case — forensic timeline, Bluesky appeal, and rebuttal of LadyYomi's framing. Preserved here after the original posts were subject to takedown.
Download ↓
Relationship Network
Force-directed graph of documented actors and their quantified interactions. Edge weights are derived from the hatewatch analyzer corpus data. Methodology ↗
Five accounts form the documented attack cluster, with distinct roles and rhetorical signatures: AYWS (open-hostile, meme-heavy, spun up within 24 hours of a predecessor account’s deactivation), LadyYomi (bureaucratic-polite DARVO dossier; a ~16k-word Google Doc was the primary output), and Vice, SaltyNeo, and LightMelody (supporting personas with intel, emotional-amplification, and validation flows into AYWS). Each actor’s quantitative signature — hedging rate, softener rate, imperative-to-self ratio, all-caps frequency, cross-linguistic attack count — is computed by the syntax-first analyzer and visible at the live monitor. Those signatures provide the empirical basis for the role labels above: they are not characterizations imposed on the data but structural summaries extracted from it.
Timeline
Chronological reconstruction across the major phases of the case.
- Sep 2024preludeprelude
Public creative collaboration with AO3 user Autumn469469 begins. 17 works published in one month. The two had no personal relationship outside fanwork; they discussed Kazama character interpretation continuously for nearly three months.
- Dec 2024preludeprelude
Chinese-language X user @fragmento_iwbyc posts in praise of Yukimura Chizuru — claiming her appeal and narrative role are underrated. This is unrelated to LadyYomi's English-speaking circle, but becomes the seed of a parallel conflict that will later be conflated with the AO3/Tumblr case.
- Jan 2025early signalsearly signals
First anonymous hostile comment received on an ygpgsgl AO3 work. Hostility and insider knowledge suggest the comment was not isolated. ygpgsgl deletes it, assumes it's a one-off.
- Jan 2025early signalsearly signals
The 'Lucy incident' begins: an anonymous user calling themselves 'Lucy' bombards Autumn's AO3 comment section under guest-account names like 'nope.' Lucy claims to be a SaltyNeo fan and accuses Autumn of writing 'Twenty-Three Nights' without SaltyNeo's authorization — which is false (SaltyNeo's own statement confirms consent was given). When provocation fails, Lucy shifts to accusing Autumn of plagiarism and of promoting her Kazama fic. Autumn deletes the comments after a confrontation, which is later reframed as 'destroying evidence of plagiarism.'
- Jan 2025early signalsearly signals
ygpgsgl posts a critique on Lofter responding to @fragmento_iwbyc: 'Chizuru could have been a better character, but the work didn't develop her enough.' Initially she doesn't censor the user ID; she corrects this after feedback. This is a Chinese-language disagreement among Chinese fans about character development — entirely separate from the English fandom feud.
- Feb 2025early signalsearly signals
@fragmento_iwbyc (also active on Chinese platforms) and supporter @TTT_hades come directly into ygpgsgl's Lofter post and attack in the comments, leading a coordinated pile-on across Weibo and Lofter. They simultaneously spread the narrative — 'she insulted the teacher,' 'her ego is too big,' 'she shouts across the air at people.' This user ships Okita Souji × Chizuru and later quits fandom in November 2025. Months of indirect mockery and gossip on Chinese social media follow. AYWS later reframes this incident on English Tumblr as 'ygp cyberbullied an innocent Okita fan' — a cross-platform narrative laundering move that strips the Chinese-language evidence showing this user as the attacker, not the victim.
- Feb 2025early signalsearly signals
Lucy suddenly 'reveals' she is a 14-year-old minor. This retroactively reframes the entire dispute: Autumn sharing the link to 'Twenty-Three Nights' (which contains NSFW content) now becomes 'sending NSFW literature to a minor.' The minor status was never disclosed at the start of the confrontation — it was deployed after the conflict as an identity-manipulation trap. SaltyNeo severs her working relationship with Autumn, citing 'protecting minors.' The plagiarism rumor resurfaces alongside new claims that 'Autumn is an old woman.' Despite Autumn publishing a statement denying all allegations, LadyYomi and her allies continue spreading these claims.
- Mar 2025early signalsearly signals
SaltyNeo posts a plagiarism accusation against Autumn, also claiming Autumn sent NSFW Kazama stories to someone claiming to be 14. LadyYomi publicly supports the accusation in the comments (later admits two now-deleted comments were hers and posts backups). SaltyNeo's original statement (linked) confirms she had given consent for Autumn to use her work as inspiration — contradicting the plagiarism framing.
- May 2025early signalsearly signals
LadyYomi claims to receive a string of insulting AO3 guest comments — allegedly from Autumn — attacking her original LGBTQ characters. The guest comments cannot be verified (anyone can post as 'guest'); LadyYomi treats them as fact and begins citing them as proof of Autumn's character. This becomes the personal grievance she uses to justify everything that follows.
- Jun 4, 2025smear campaignsmear campaign
ygpgsgl publishes 'Psychoanalytic Reflections on My Relationship with Kazama Chikage' on AO3 — anonymously, because the content is intensely personal and critical. The essay is loaded with tags that explicitly call out the harassment patterns by name; anyone searching the relevant tags now sees a counter-narrative they cannot ignore.
- Jun 2025smear campaignsmear campaign
LadyYomi sees the anonymous Psychoanalytic essay and immediately concludes it is 'intentional harassment to evade the block' — she had already blocked Autumn, and since she insists ygpgsgl = Autumn, the very act of posting anonymously is interpreted as 'hiding to keep attacking me.' She and her circle reframe the tag-loaded critique as a 'callout,' manufacturing the accusation that ygpgsgl is the aggressor. The inversion is complete: a defensive critical essay is rewritten as targeted harassment of LadyYomi.
- Jun 2025smear campaignsmear campaign
LadyYomi begins secretly slandering ygpgsgl in the comment sections of her own AO3 works, deliberately conflating ygpgsgl with Autumn as the same person. She quotes anonymous guest comments as evidence and treats unverified accusations as established facts, creating a feedback loop where rumor becomes record. Vulgar language directed at both ygpgsgl and Autumn appears in these threads.
- Jun 2025smear campaignsmear campaign
Without any prior interaction with ygpgsgl, LadyYomi blocks her on AO3 (ygpgsgl had no Tumblr account at this point), then claims ygpgsgl was 'harassing her comment section' and 'stalking' her. The claim is preemptive — designed to position LadyYomi as a victim before ygpgsgl can respond.
- Jun 2025smear campaignsmear campaign
LadyYomi spreads the 'ygpgsgl = Autumn469469' conflation across the comment sections of multiple Hakuouki original-character writers' works. Mass-blocking begins; ygpgsgl's works are tagged as 'unreliable,' 'AI-generated,' 'fake progressive feminist.' Autumn contacts ygpgsgl mid-month: 'Someone is leaving defamatory comments about you under some posts.'
- Jun 30, 2025smear campaignsmear campaign
The 'two-voice bait-and-redirect' operation begins: on AO3, a guest account 'Concerned Reader' (likely hiraethum, based on matching rhetorical structure) posts a fake-neutral question under Autumn's work: 'Just kinda confused tbh… what exactly was copied?' The bait is set — a seemingly innocent question designed to draw out a one-sided 'receipt dump' from a coordinated participant.
- Jul 5, 2025smear campaignsmear campaign
The bait is answered: vice (as AO3 guest 'Hello,' Jul 5 02:28 CST) replies to 'Concerned Reader,' directing them to 'saltyneo's pinned post on tumblr' and adding 'receipts are in our discord and in google docs from the chinese fandom.' The word 'our' is a critical slip — a neutral bystander does not say 'our discord'; only an insider does. 'Concerned Reader' then feeds a second round of bait (Jul 5 06:28 CST): 'Thanks for explaining... Is there any actual screenshot or link?' — keeping the receipt-dump pipeline flowing. Within hours, ygpgsgl notices the entire operation and immediately publishes her Notice document on AO3. Vice panics — switching to a new guest alias 'sorry autumn is reading this' to abort the evidence drop, signaling that the target is watching in real time. The alias name itself is a tell: only the person executing the attack would break character because the target has counter-surveillance. Later the same day or early July 6, hiraethum repeats the exact same bait on Tumblr under her real name ('WHAT'S GOING ON AT HAKUOUKI'S AO3?? 😭'), and hours later LadyYomi begins writing her defamation posts. On July 6, ygpgsgl re-registers a Tumblr account to fight back.
- Jul 5, 2025smear campaignsmear campaign
ygpgsgl posts a formal Notice on AO3 requesting verifiable evidence by July 10. Submissions must include screenshots, links, or timestamped documentation; vague assertions will not be acknowledged. The 50-chapter case documentation 'Anonymous Harassment' begins.
- Jul 6, 2025smear campaignsmear campaign
Hakuouki zine mod msbeastlyeevee announces Autumn's removal from the contributor list in LadyYomi's reply thread — within 24 hours of LadyYomi's defamatory claim. vicecommanderhijikata (Tumblr) — same person as oninofukuchou (AO3); both account names are variations on Hijikata Toshizō — immediately jumps in declaring he '1000% supports LadyYomi' the moment her systematic defamation post goes up. He calls for community-wide blocks against ygpgsgl and becomes LadyYomi's most loyal enforcer. In the linked reblog chain (July 7), vice's language toward ygpgsgl includes 'go fucking touch grass,' 'you're a bully and a coward,' 'you act like a little bitch,' and 'fuck off' — then one month later (Aug 7) he claims 'I've been civil up until now.' He also demands ygpgsgl prove she is not Autumn by sharing Discord screenshots — a demand that would violate Autumn's privacy, creating a deliberate no-win scenario.
- Jul 6, 2025smear campaignsmear campaign
ygpgsgl re-registers a Tumblr account, in order to break out of the AO3 silo and respond publicly. She posts a statement and debates LadyYomi's allies through the day. That same evening, LadyYomi publishes a meme celebrating: ygpgsgl shown alone and isolated, LadyYomi's clique shown united.
- Jul 2025smear campaignsmear campaign
After LadyYomi publishes her isolation meme and her supporters flood ygpgsgl's AO3 comment sections, a wave of anonymous hostile comments arrives — including fabricated accusations recycled from the Lucy incident, identity-based attacks, and coordinated defamation. ygpgsgl archives the full comment threads in two documents (linked). During this period, responding in frustration to a suspected machine-translation sockpuppet commenter, she writes a line borrowing Okita Souji's voice: 'Damn it, do you believe that once I find you, I'll kill you? (in Okita Souji's tone)' — a character-voice vent with an explicit parenthetical annotation marking it as non-literal. This single line is later stripped of all context by LadyYomi and cited 14 times in her 16,000-word Google Doc as evidence of 'death threats against my readers.'
- Jul 7, 2025smear campaignsmear campaign
Autumn posts a one-sentence denial: 'ygpgsgl and I are not the same person.' No explanation, no evidence, no rebuttal. LadyYomi pins a 'Declaration of Victory' on Tumblr stubbornly insisting they are the same person. Followers (including @onino and @lightmelodyva) continue with memes and ridicule.
- Jul 13, 2025escalationescalation
ygpgsgl publishes 'Statement Regarding Authorship and Withdrawal of Collaboration Rights,' formally revoking Autumn's right to translate or host any of her works.
- Jul 17, 2025escalationescalation
SaltyNeo reposts LadyYomi's defamatory post and demands a public apology from ygpgsgl. Around the same time, she leaves an AO3 comment on ygpgsgl's documentation: 'Hello, if you are not associated with Autumn/Saya and have no beef with me, can you remove my name from your tags? It's disturbing.' Three days later (July 20), she sends a private message performing complete ignorance: 'Excuse me, but...Who are you? I really don't get who you are in this mess... May I ask what the problem with me is? As far as I know we never interacted.' The root cause of the entire harassment chain is pretending she doesn't know the person she harmed. ygpgsgl ignores the PM entirely. Nine days later (July 29), the 62-page forensic PDF is published with SaltyNeo named and analyzed — and by August 1, SaltyNeo responds with public humiliation and open hostility. LadyYomi continues to refuse to file formal AO3 reports — because actual reports would require evidence she does not have.
- Jul 2025smear campaignsmear campaign
Throughout July, vicecommanderhijikata and LightMelodyVA aggressively demand that ygpgsgl and Autumn hand over their Discord accounts — ostensibly to 'prove they are not the same person,' but likely motivated by fear that ygpgsgl might infiltrate their private coordination channels and extract more evidence. The demand creates a closed-logic trap: refusing = 'proof you're hiding something,' complying = exposing Autumn's privacy and surrendering counter-intelligence capability.
- Jul 29, 2025escalationescalation
Forensic Brief v1 (62 pages) published — assembled with GPT-4o assistance, mapping the six core figures shaping the narrative. Includes a 'Linguistic Anatomy of Defamation' analyzing AO3 anonymous comments, a 'Stylistic Similarities Between oninofukuchou and Anonymous Comments' section, and a 'Major Finding' on machine-translated 'public opinion.'
- Jul 29, 2025escalationescalation
GPT-4o introduces a single error into the PDF: in one passage analyzing oninofukuchou's rhetoric, it accidentally treats one of its own generated sentences as if it were a quote from onino. ygpgsgl catches and corrects the mistake quickly. But LadyYomi and her circle, reading the PDF in fury, seize on this single attribution error as proof that the entire 62-page document is 'AI slop' and therefore unreliable. From this point on, every subsequent piece of writing or evidence ygpgsgl publishes is labeled 'AI-generated,' 'fake,' 'fabricated by a chatbot,' regardless of content.
- Aug 1, 2025escalationescalation
SaltyNeo begins a public mockery cycle: 'forgets ID,' publicly ridiculing ygpgsgl. The 'AI-using bastard' framing now circulates freely: 'AI can fabricate lies, since she likes using AI, she's fake as a whole.'
- Aug 2025escalationescalation
Aug 7: After ygpgsgl posts a reminder on Tumblr — 'if you're gonna talk shit, do it on Tumblr, not on Bluesky. Don't waste your performance on an empty stage' — vicecommanderhijikata responds within hours: 'I'll talk shit about you wherever the fuck I want. If I want to go to deviant art and talk shit about you, I'll do it there, too. I've been civil up until now... I don't owe you a damn thing.' Two days later (Aug 9), LadyYomi crosses to DeviantArt and posts defamation there — exactly as vice announced. On Aug 13, vice stacks 9 identity markers in a single post ('trans, gay, mentally ill, disabled, unemployed, neurodivergent, fat, poor, atheist, cult survivor, childhood sexual abuse survivor') as a preemptive shield. On Aug 25, vice publishes a longer manifesto: 'Don't pick a fight if you're not willing to stand by what you say... you weren't prepared for the consequences of your actions... you're a shitty, bad person.'
- Aug 9, 2025escalationescalation
LadyYomi crosses to DeviantArt and posts new defamation about ygpgsgl there. The post drives a wave of curious rubberneckers from DeviantArt to ygpgsgl's AO3 documentation work (which had been intentionally left open to public comments for transparency). The rubberneckers begin collectively 'judging' ygpgsgl in the comment section. In anger, ygpgsgl deletes every comment on the documentation work and closes the comment section entirely. The bitter irony: a piece of writing meant to invite scrutiny had to be sealed off because the scrutiny was deployed as a weapon.
- Aug 13, 2025platform suppressionplatform suppression
Coordinated mass-reporting campaign takes down ygpgsgl's Bluesky account (@ygpgsgl.bsky.social) — suspended without notice, escalated to permanent ban within 3 hours, no opportunity to appeal. LadyYomi and vicecommanderhijikata celebrate — vice reblogs LadyYomi's celebration post with a chibi meme of Kondō Isami and Hijikata Toshizō clinking glasses in a toast. hiraethum is visible in the like activity. ygpgsgl writes directly to Bluesky CEO Jay Graber. Account permanently lost — never restored.
- Aug 2025platform suppressionplatform suppression
Emboldened by the Bluesky takedown, AYWS, vicecommanderhijikata, and LightMelodyVA openly discuss strategies for taking down ygpgsgl's Medium account in AYWS's ask/answer threads. Medium moderators initially ignore their reports. Vice devises the tactic of reporting ygpgsgl for 'using AI to write' — weaponizing Medium's AI-content policies as a takedown vector, turning an aesthetic opinion ('AI slop') into a fabricated policy violation. Medium eventually falls. SaltyNeo is the first to celebrate, posting a CG image from the game depicting Kazama Chikage (ygpgsgl's favorite character) being killed by Saitō Hajime (SaltyNeo's favorite) — a deliberate use of in-game violence imagery as symbolic triumph over the target. Account permanently lost — ygpgsgl had no energy left to appeal.
- Aug 18, 2025escalationescalation
LadyYomi's Tumblr account @ladyyomiart is deactivated. Within 24 hours, a new anti-ygpgsgl Tumblr account network appears: @anti-ygpgsgl-widemushroom-show (first post: Aug 19, 19:09 GMT), ygpgsgl_syndrome (stealing the term 'YGP Syndrome' that ygpgsgl herself coined to analyze the harassment pattern, and weaponizing it back against her), and @anti-ygpgsgl-ygppoop-show. The same-day handoff from ladyyomiart's deactivation to AYWS's launch indicates a pre-planned platform migration, not a spontaneous emergence. Content includes 'fuck your fake bot,' 'sex doll kazama wife,' 'maybe she's actually a man,' '你这算哪门子的中国人' ('what kind of Chinese are you'), 'So keep fucking that alibaba Kazama sex doll.' Racialized, sexualized, and ableist attacks running in parallel with LadyYomi's discourse-level smear (now operating via Google Doc instead of Tumblr). The same person also published explicit pornographic AO3 fiction sexualizing ygpgsgl as a character (since removed from AO3). The Hate Speech Monitor (linked) was built specifically to analyze the output of @anti-ygpgsgl-widemushroom-show.
- Aug 19, 2025escalationescalation
SaltyNeo escalates to personal attacks ('cunt,' 'sad, lonely thing'). She publicly uses ygpgsgl's own statistical data to humiliate her — 'Oh wow, thanks for the data, Y-something (I can't remember the random letters after Y, sorry)' — pretending to forget the name while weaponizing the research. SaltyNeo's contradictory statements on the '14 years old' Lucy claim are exposed: she originally invoked the teenager's age to discredit Autumn, then later denied ever knowing about it. This direct contradiction proves she was an active participant, not an ignorant bystander. SaltyNeo is the root cause of the entire harassment chain: without her framing of Autumn for plagiarism and her endorsement of the Lucy narrative, none of the subsequent escalation would have occurred.
- Aug 27, 2025escalationescalation
SaltyNeo posts and pins 'About the future of my translations,' publicly announcing plans to 'kick out people like YGP' from Discord.
- Aug 2025escalationescalation
Tumblr user @lightmelodyva publishes a PSA (public service announcement) framing ygpgsgl as 'AI-dependent, mentally unstable, a structural perpetrator who corners others.' The PSA was written without correctly identifying who ygpgsgl was actually criticizing — ygpgsgl's targets had been the Chinese-speaking X users (@fragmento_iwbyc, @TTT_hades), but English-speaker LadyYomi read the criticism as aimed at herself. The PSA spreads. anti-ygpgsgl-widemushroom-show then deliberately distorts the facts further, completing the merger: the Chinese-fandom dispute and the AO3/Tumblr feud are bundled into a single 'ygpgsgl is a serial perpetrator' narrative, retrofitting her into the moral-bankrupt strawman LadyYomi's circle had been building since June.
- Sep 11, 2025platform suppressionplatform suppression
Tumblr blog deleted without notice — the result of coordinated mass-reporting following ygpgsgl's submission of a documentation package to Idea Factory International and EastAsiaSoft. Zine mod msbeastlyeevee is the first to celebrate the takedown. Full backups preserved; account later restored, but the celebratory response from the harassers confirms the takedowns were not organic moderation decisions but targeted retaliation for reporting to the copyright holder.
- Sep 2025documentationdocumentation
Documentation phase: archive compiled, journalists and NGOs contacted, AI usage clarified through formal public statement. ygpgsgl coins the term 'YGP Syndrome' to describe the structural pattern of being forced to bear the entire burden of proof while being denied recognition as a victim.
- Sep 2025documentationdocumentation
After ygpgsgl publishes her investigation into the Hakuouki Zine's role in the harassment (linked), zine mod msbeastlyeevee panics and attempts to reblog her way out. ygpgsgl blocks her. In retaliation, msbeastlyeevee joins the harassment circle: participates in AYWS's ask sessions (5 asks), flaunts her Hakuouki merchandise collection with stated intent to cause harm ('I want her to feel jealous'), triggering a group merchandise-flaunting chain where vice, SaltyNeo, and others compete to show off their collections as proof of 'real fan' status. AYWS also proposes a 'recommend your favorite Hakuouki fanfics' thread, explicitly excluding ygpgsgl and Autumn as 'trash whose garbage would never be considered.' The zine mod's arc — from neutral institutional gatekeeper to radicalized participant — follows the 'secondary core peripheral figure' pattern identified in ygpgsgl's forensic PDF: appearing neutral, then defecting at a critical moment to add institutional legitimacy to the harassment.
- Sep 11, 2025documentationdocumentation
ygpgsgl submits a full documentation package to Idea Factory International and EastAsiaSoft, detailing copyright violations (pirated CGs, game materials, voice lines distributed by SaltyNeo), brand misuse by the Hakuouki Zine project, and coordinated harassment enabled by zine contributors. The letter includes evidence that zine mod msbeastlyeevee, LightMelodyVA, and vicecommanderhijikata publicly endorsed the anti-ygpgsgl hate accounts while claiming to represent the Hakuouki brand community. One zine mod stated 'I want to make her feel jealous' — documenting stated intent to cause emotional harm. The merchandise-hoarding culture is cited as a gatekeeping mechanism: flaunting collections as proof of 'real fan' status to justify excluding others.
- Sep 13, 2025documentationdocumentation
ygpgsgl writes on AO3, in reference to her harassers: 'they are dead to me now.' The phrase is a literary statement of severance. LadyYomi's circle reframes it as a 'death threat.'
- Sep 23, 2025documentationdocumentation
ygpgsgl publishes a critical analysis of the harassment network on note.com. LadyYomi reports it to note.com, resulting in a warning letter. Within two hours, ygpgsgl publishes a counter-article dismantling the report. Shortly after, @ladyyomi and @oninofukuchou (= vicecommanderhijikata) escalate to openly hostile bookmark comments on ygpgsgl's AO3 works — bookmarks the author cannot delete. Content includes 'a structural perpetrator who frames herself as victim,' 'mental breakdown caused by AI,' and similar personal-attack framings. Likely triggered by ygpgsgl publishing a Japanese-language media brief on September 22 (the day before the bookmark attacks begin) and then tagging the official Hakuouki accounts (@HakuokiShinkai, @OtomateWeb, @IdeaFactoryIntl) plus Japanese public intellectuals on the franchise's 17th anniversary (September 25) — escalating the case from fandom-internal to copyright-holder-visible. The three regulars — LadyYomi, onino, and the anti-ygpgsgl operator — continue this bookmark campaign through October and November.
- Oct 2025documentationdocumentation
LadyYomi weaponizes AO3 bookmarks at scale: repeated 'ygpgsgl = autumn alt' bookmarks on ygpgsgl's works — comments the author cannot delete. She updates a 100+ page Google doc continuously to keep spreading the claim, completely ignoring Autumn's July 7 statement as if it never existed.
- Nov 3, 2025rupturerupture
Autumn finally posts a full public statement under an alt — detailed identity (Arab American), explicit refutation of the conflation. Four months late. LadyYomi falls silent, not in apology but because the lie can no longer be sustained.
- Nov 4, 2025rupturerupture
11 attacking comments appear on Autumn's work within 22 minutes. Open threats: 'other works are my next targets.' The harassers go into revenge overdrive because Autumn's full statement has shattered their core lie.
- Nov 2025rupturerupture
Bookmark harassment continues into November. LadyYomi, onino, and the anti-ygpgsgl operator are still leaving hostile bookmarks on ygpgsgl's works — three months after Autumn's denial, weeks after Autumn's full statement should have ended the conflation.
- Nov 2025rupturerupture
Side note from the parallel Chinese-fandom dispute: @fragmento_iwbyc — the X user whose December 2024 Chizuru post had set off the original Chinese-language conflict — quietly announces she will no longer update any Hakuouki-related works. One of the people who helped seed the strawman walks away from the fandom without addressing any of it.
- Dec 30, 2025aftermathaftermath
Chapter 50 of the documentation work is published — 'Clarity After Sleeping for a Day and a Night.' The summary line: 'LadyYomi is a bitch. So are her gang.' The chapter is unrestrained, profanity included, an exhausted author finally swearing on the page after six months of measured documentation. The harassers fall strangely silent — no bookmark replies, no AO3 reactions in any space ygpgsgl can see. They may have raged elsewhere on DeviantArt or Tumblr, but ygpgsgl is offline by this point, in extended digital rest. The 50-chapter case file effectively ends here.
- Feb 17, 2026aftermathaftermath
Critical essay 'This is Not Love, This is Conceptual Violence' marked as updated with new chapters processing the events.
- Mar 8, 2026ao3 rulingao3 ruling
AO3 Policy & Abuse volunteer 'Oscar' issues a 15-day suspension. Stated reason: copyright violation in a Marx translation. Actual effect: the suspension is bundled with the 50-chapter harassment documentation, removing both works simultaneously — using a plausibly legitimate copyright issue as cover to silence the documentation of abuse.
- Mar 9, 2026ao3 rulingao3 ruling
Public response published: 'The Bait and Switch: How AO3's Oscar Used Copyright as Cover to Silence a Victim.'
- Mar 2026withdrawalwithdrawal
AO3 account deleted. All works hosted on AO3 removed by the author. Case effectively closed. Materials preserved on this site for accountability and reference.
Cite
Suggested citation for this case study as an archival resource:
Primary structured summary: the Forensic Brief PDF linked in § 01. Quantitative analysis of the case’s corpora: the live monitor linked in § 02. The underlying analyzer is released as an independent software artifact; readers interested in the tool itself can locate it via the archive homepage.