Executive summary
Legal texts, public decisions and authoritative institutional material
Uses a procedural chronology to code legal norms, public practice evidence and individual allegations separately.
Research questions
- How should law and observable practice be distinguished?
- Which stages most affect evidence reliability and defence rights?
- How can external courts avoid undifferentiated system-wide conclusions?
Principal findings
Legal basis of detention, family notice, timing of counsel and interrogation records are central case-level checkpoints.
RSDL, national-security exceptions and secrecy restrictions may change practical access to safeguards.
Assess independence, publicity, counsel, silence, confession evidence and remedy separately.
Law and standards
Legal analysis begins with the applicable threshold rather than a predetermined conclusion. International obligations, domestic law, institutional rules and professional standards are treated separately. China’s signature but non-ratification of the ICCPR is distinguished from its obligations under the ratified Convention against Torture; soft-law standards are not presented as equivalent to treaty duties.
Research design and source evaluation
This report uses reproducible desk research. Priority is given to treaties, judgments, legislation and original official records; then UN, court, parliamentary, government and regulatory material; then specialist research with a disclosed method; and finally reputable reporting for chronology. Each material proposition records source, publication and coverage dates, jurisdiction, procedural status, independent support, contrary material and limitations. Repetition of one underlying item is not counted as independent corroboration.
Treaties, legislation, judgments and official records.
UN, court, parliamentary, government and regulatory documents.
Independent sources are traced to their underlying evidence.
Material inconsistent with the working hypothesis is recorded.
Legal thresholds and unit of analysis
Fair trial, arbitrary detention, torture and political prosecution may overlap but have distinct thresholds. An opinion should state whether material proves background conditions, increased risk, an individual fact or a legal threshold preventing transfer or use of evidence.
The most reliable unit is the case chronology, not a single system-wide label. Initial control, family notice, detention, arrest approval, first counsel access, interrogation, prosecution, hearing, judgment and remedy should be fixed individually.
Counsel, secrecy and national-security exceptions
The presence of a lawyer does not establish effective defence. Choice, timing, confidentiality, disclosure, investigation, exclusion motions, family contact and professional pressure all require examination. Replacement of family counsel by appointed counsel requires scrutiny of genuine choice and effectiveness.
State-secret, investigation, national-security and closed-hearing rules may restrict safeguards. External courts should identify the decision-maker, reasons, duration, review and actual impact rather than cite exceptions abstractly.
Confession acquisition, ill-treatment and reliability
Confession analysis starts with acquisition conditions. Location, sleep, food, restraint, threats, injury, shifts, recordings, lawyer notes and medical requests may corroborate. Later repetition or televised admission does not automatically break earlier coercion.
Exclusion requires more than a statutory rule: could the motion be made, were recordings and medical records accessible, did the court investigate, give reasons and provide effective appellate correction?
Forensic use and expression of conclusions
External proceedings should separate Chinese law, practice material and individual evidence. Missing documents should be classified as nonexistent, unpublished, refused, deleted or not sought. Information gaps neither automatically defeat a claim nor prove the gravest allegation.
Conclusions should state which safeguards exist and operated, which were unavailable in practice, which were displaced by exceptions, what remains unconfirmed and whether defects could be cured by independent, timely and effective remedy.
Professional use
The report is written for courts, law firms, asylum and immigration bodies, legislatures, media and specialist organisations. Use in proceedings requires the primary sources to be supplied and updated to the hearing date, with an explicit bridge between general material and individual facts. It is not case-specific legal advice and does not replace qualified legal, country or medical expertise.
It is not a universal conclusion on all Chinese criminal cases and does not replace Chinese-law or forum-law expertise.
Limitations and conclusion
Public research is limited by language, censorship, closed proceedings, uneven regional coverage, safety and incomplete files. CRM does not treat an information gap as proof of safety or a contextually plausible allegation as established fact. Material corrections increase the version number; typographical or citation corrections are logged as minor updates.
CRM concludes that fair-trial risk indicators in china’s criminal process must be assessed through a transparent chain from source to fact, from fact to legal issue and from general pattern to the individual record. Conclusions should be no broader than the evidence permits.