docs/research/history-hong-kong-gap-2026-10-23.md +156 −0
@@ -0,0 +1,156 @@
+# Hong Kong: the gap after 1978
+
+Research date: 23 October 2026. Before this pass, Hong Kong had four history
+entries, with no event between the 1978 definition amendment and the 2023 CBD
+listing. This batch adds nine events, revises three existing descriptions and
+creates four articles. It expands the profile's sentencing account, corrects the
+description of the first CBD imprisonment and converts revised blocks to inline
+citations. Current scores and status classifications are preserved. All twelve profile text
+blocks now use inline citations.
+
+## Primary sources read
+
+- **Legislative Council, 3 June 1987**, printed pp. 1696–1697: written answer
+ announcing more cannabis/mandrax publicity and wider secondary-school talks.
+ The 600 talks and 200,000 pupils concern the whole first cycle of school talks,
+ not a cannabis-only programme. The event records an announcement, category
+ `debate`, not a new criminal rule or completed rollout of every proposal.
+- **Legislative Council, 26 February 1992**, relevant drug-amendment introduction
+ at PDF pp. 55–57: rights rulings, removal/narrowing of presumptions and proposed
+ possession penalties.
+- **Legislative Council, 24 June 1992**, relevant second-reading discussion at
+ PDF pp. 80–84, committee amendments at pp. 111–112 and third-reading passage
+ at p. 130. Read the government's original proposal, Professor Felice Lieh Mak's
+ committee account, the adopted changes to clauses 2 and 4 and final passage.
+ The summary-case fine rose to **HK$100,000**, rather than the government's
+ initial plan to retain HK$10,000. The indictment maximum became seven years
+ and HK$1 million; summary custody remained three years. Removing the separate
+ trafficking-intent possession section did not legalise possession for supply:
+ that conduct was moved into the trafficking definition.
+- **Legislative Council, 20 June 2001**, bilingual printed p. 199: an official
+ commencement table gives **26 June 1992** for the amendment. The table and
+ the penalty discussion at PDF p. 82 of the 1992 sitting were rendered and
+ visually checked. The original gazetted 1992 amendment was not recovered;
+ the entry combines final parliamentary proceedings with the government's
+ later commencement record. The 1994 sitting was downloaded as corroboration
+ but was not needed as a public source.
+- **R v Tuen Shui Ming and another**, 9 June 1995, full judgment read, especially
+ paragraphs 1–4 and 10–20. More than **51 kg**, not 5 kg: the judgment itself
+ discusses the original judge's mistaken wording. The two sentences were raised
+ from two years to four. Weight was the initial basis, with limited potency
+ adjustment; the lowest bands became one band below 2 kg. The court retained
+ different treatment for herbal cannabis, resin and oil and refused a wholesale
+ upward recasting of the tariff.
+- **HKSAR v Phung Hai**, 17 February 2000, full short judgment read: personal
+ possession of 1.18 g; two earlier drug convictions; six months reduced to two;
+ prosecution did not defend the original term. A single outcome does not
+ establish the typical sentence for all repeat offenders.
+- **HKSAR v Chor Lui**, 3 August 2001, full judgment read: two cigarettes,
+ 0.24 g, HK$30 undercover purchase, previous convictions, commercial context,
+ mitigation, plea reduction and dismissed appeal. The four-month herbal
+ benchmark was an explanation of the then-current approach, not a statutory
+ maximum or a guarantee of leniency.
+- **HKSAR v Nguyen Thang Loi and Dang Hung Ngoc**, 20 January 2023: read case
+ facts and sentencing reasoning at paragraphs 1–29 and 94–109, with relevant
+ historical/submission passages. The full judgment was recovered, not merely
+ the official three-page summary. The Department of Justice summary was also
+ read in full and its cultivation table visually checked. The guideline bands
+ concern starting points and individual culpability; they are not mandatory
+ sentences, possession allowances or plant limits. Estimates for growing
+ operations concern annual production, not just the cannabis present on the day.
+- **Security Bureau legislative brief, 20 October 2022**, paragraphs 6–11:
+ CBD market, the government's contamination/conversion rationale, prescription
+ rules since **13 December 2019**, and absence of registered CBD pharmaceutical
+ products at the date of the brief. Chemical/medical statements are attributed
+ to the government's explanation, not adopted as independent efficacy findings.
+- **Government announcements, 30 November 2021 and 31 January 2023**, read in
+ full: THC prohibited regardless of concentration before CBD itself was listed;
+ 1 February 2023 commencement; about 77,400 surrendered items by 29 January.
+- **Information Services Department, 7 July 2023**, read in full: first CBD-related
+ imprisonment, **two months overall for CBD, ketamine and equipment counts**.
+ The report supplies no separate CBD-only sentence. March 29 is arrival/arrest,
+ not the sentencing date.
+- **Hong Kong Police laws/penalties page**, **HKU CLIC possession sentencing
+ page**, and **17 March 2026 government statistics release**, read in full for
+ the revised profile blocks. The 4,108 arrests and 86% conviction rate cover
+ drugs generally, not cannabis alone. Registry reports are not a population
+ prevalence survey.
+
+## Retrieval and provenance
+
+HKLII judgment URLs initially return a JavaScript shell to a simple reader.
+The site's own application bundle documents its public `getjudgment` API;
+that returned the original judgment body and date for all four cases. Both raw
+JSON and extracted text are retained locally. Public citations point to the
+normal HKLII case pages. A vLex reproduction of the 1995 judgment was retrieved
+as a lead, then superseded by the HKLII original. Direct Judiciary retrieval
+timed out; it was not mistaken for a read source.
+
+LegCo downloads failed ordinary local certificate validation, then succeeded
+for these public documents using an unverified SSL context. Their domain,
+document titles, contents and relevant page images were checked. Browser
+search snippets were used as leads, not as substitutes for reading the PDFs.
+
+The same public API supplied the current ordinance consolidation dated 12 June
+2026. Read sections 4, 8, 9, 22, 24, 25 and 54A and the relevant First Schedule
+Part I entries. Section 22(6) excludes paragraphs 8–11 from ordinary professional
+and pharmacy authorities; paragraph 11 expressly lists cannabis and resin. The
+profile therefore replaces a broad professional-exemption description with the
+actual exclusion and licence requirement. The CBD pharmaceutical text is explicitly
+dated to the 2022 report; no operating 2026 patient programme is inferred from
+general controlled-medicine powers. The original current status flag is preserved
+pending specific patient-product/availability evidence. Legacy commentary and
+VOA references were replaced by read statutory and government sources.
+
+The API also exposes earlier ordinance versions, a useful lead for reconstructing
+treatment-report rules and the older score chain in a later pass. Only the sections
+and versions actually inspected in this batch are treated as read.
+
+Research files live under `tmp/history-hong-kong-gap-2026-10-23/`. No complete
+judgment or parliamentary scan was copied into public article assets.
+
+## Classification and score decisions
+
+- **1992: `other`**, because higher maxima and weaker prosecution presumptions
+ are a mixed reform. The explicit unscored reason preserves the criminal
+ possession/personal-enforcement segment levels; the sources do not establish
+ a new typical personal-use disposal deserving a whole-point movement.
+- **1995: `other`**, because the review combined full-weight sentencing, limited
+ low-potency adjustment and wider discretion in the lowest band. It doubled
+ these defendants' sentences but did not simply increase every cannabis tariff.
+- **2000 and 2001: `enforcement`**, individual sentencing outcomes under existing
+ rules; no general decriminalisation or new possession offence is claimed.
+- **2019: `medical`**, a CBD pharmaceutical classification, with an explicit
+ unscored reason. The brief reports no registered CBD medicine in October 2022
+ and establishes no operating THC-cannabis supply route. Current `medical: none`
+ and access zero are preserved; this entry does not imply a functioning patient
+ programme or certify availability in 2026.
+- **2021: `debate`**, a government warning about the existing THC rule while CBD
+ remained outside dangerous-drug controls; no CBD ban is backdated to it.
+- **January 2023: `tightening`**, ending the lower herbal trafficking scale and
+ setting cultivation bands. Sharing and cultivation are already zero; the
+ ordinary possession segment is not changed by these tariffs.
+- **July 2023: `enforcement`**, a combined-offence sentence, not a further legal
+ tightening or proof that every CBD possession case receives custody.
+
+All pre-existing history score vectors and current profile scores are unchanged.
+The old 1928 first step remains a coarse historical baseline requiring further
+penalty reconstruction; this pass does not claim to reconstruct every score
+between 1928 and today. The 1968 and 1978 dates remain verified assent and
+legislative passage respectively, rather than invented commencement dates.
+
+## Validation and isolation
+
+Article, source, category, tag, internal-link and score-preservation checks run
+before the production build. Rendered checks cover citation IDs/targets, source
+links, exports, tag pages and the Hong Kong API. Temporary build output is
+deleted after verification. A private Git index commits only this batch's
+history objects, profile, four articles and this memo; the unrelated staged and
+unstaged `source-archive.json` changes are excluded.
+
+Final verification on 24 October: production build passed with 3,936 pages;
+96 rendered citation links, tag targets, clean JSON/CSV exports and the Hong Kong
+API passed. Exact score remains 6.5 (displayed rounded as 7). All four articles
+passed at 357–416 words with four sections each. The HK profile HEAD matched the
+saved pre-edit baseline, confirming its changes are isolated. Concurrent Zambia
+research and other actors' committed entries are excluded from this batch.