The Johannesburg extradition hearing involving Ndodana Mkhanyisi Tshuma has moved from a police alert into a formal cross-border court process. South African authorities arrested Tshuma in Johannesburg after cooperation involving the South African Police Service’s Interpol National Central Bureau. British prosecutors had authorised three murder charges after the deaths of his wife and two daughters in Bedfordshire. The Crown Prosecution Service has expressly warned that a charging decision is not a finding of guilt, and Tshuma remains entitled to a fair trial.
At the 27 August 2026 hearing, the National Prosecuting Authority said that the documentation supporting the United Kingdom’s request was complete. The extradition matter was postponed to 8 September, while a separate South African firearm and ammunition case was listed for 10 September. That timetable matters because an international arrest does not itself transfer a person across a border. A local judge must still supervise the next procedural step, the defence must be able to test the request, and the requesting state must satisfy the legal conditions governing surrender.
This article uses the Tshuma case as a live test of a question asked whenever an accused person crosses a frontier: what does an arrest actually achieve, and what can still stop a transfer? The French and European comparison is particularly useful. France distinguishes police cooperation, provisional arrest, judicial review and the final act of surrender. Its courts do not retry the foreign case, but they do test legality, double criminality, fundamental rights, documentation, proportionality and the reliability of guarantees. Those controls explain why an Interpol-assisted arrest is the beginning of an extradition case, not its conclusion.
I. What happens after a UK extradition request reaches South Africa?
A. Why an Interpol-assisted arrest still requires a local court hearing
The first legal distinction is between locating a person and surrendering that person. The public record says that South African police traced and arrested Tshuma in Kensington, Johannesburg, after British authorities sought assistance. The official South African account identifies the participation of the Interpol National Central Bureau, Crime Intelligence and the Organised Crime Investigation Unit. The Associated Press reported that South Africa and the United Kingdom have an extradition treaty, while also making clear that a person arrested in South Africa must appear before a South African court before extradition can take place. The government account is recorded by SAnews, and the international sequence was also reported by The Associated Press.
That sequence is familiar across extradition systems. An alert, diffusion, red notice or bilateral police request may justify urgent action under the law of the place where the person is found. It does not replace the formal instrument on which surrender must ultimately rest. The executing state must know who issued the warrant, what conduct is alleged, which offence is relied upon, what sentence may follow, and whether the request complies with the applicable treaty and domestic law. The requested person must also have a meaningful opportunity to challenge detention and surrender.
The legal risk is easy to see. Police systems are designed for speed and information exchange. Extradition courts are designed for attribution, authentication and legal limits. A database hit can be enough to trigger contact between authorities. It cannot, by itself, answer whether the foreign warrant is valid, whether the conduct is an extraditable offence, whether the request is politically motivated, or whether surrender would expose the person to treatment prohibited by human-rights law.
The United Kingdom’s own charging statement illustrates the same separation. On 10 July 2026, the Crown Prosecution Service said it had authorised three murder charges after reviewing evidence provided by Bedfordshire Police. It also stressed that the decision did not determine guilt and that only a court can make findings of fact. The Crown Prosecution Service press release records both the charges and the fair-trial warning. In an extradition file, that distinction remains important: the requesting state must show a lawful basis for prosecution, but the requested state is not being asked to decide the eventual verdict.
The 27 August hearing therefore has two different dimensions. The first is operational: South Africa located a person whom British authorities wanted to question and prosecute. The second is judicial: South Africa must decide whether the request may progress under its treaty and domestic law, whether the documents are sufficient, and how the requested person’s local rights are protected. The postponement to 8 September does not mean that surrender has been refused or granted. It means that the case remains in the supervisory phase.
That is also why headlines using the word “extradited” before a judicial decision can mislead. An arrest is a coercive step. Extradition is a legal outcome. Between the two lie authentication, notice, counsel, evidence thresholds, possible objections and a decision by the competent authority. The more serious the accusation, the stronger the public interest in cooperation; it does not remove the need for a lawful process.
The separate South African firearm case adds another layer. Reports from the 27 August hearing state that Tshuma faces a local allegation concerning unlawful possession of a firearm and ammunition, and that the defence and prosecution were discussing how that matter might proceed. Those allegations must be kept separate from the British murder charges. A local prosecution can affect custody, evidence, scheduling and the timing of any transfer, but it cannot turn an extradition request into a domestic conviction. The contemporary IOL report on the Johannesburg hearing records the NPA’s position that the British documentation was complete and the two matters’ separate timetables.
For international practitioners, the practical lesson is immediate. The first document to obtain is not a press article and not an Interpol screenshot. It is the actual request, warrant, treaty certificate or equivalent instrument transmitted through the proper channel. The second is the custody record showing when the person was arrested, what reasons were given, and when the materials were supplied. The third is the local court order. Without those three layers, it is unsafe to describe the procedural posture with precision.
B. How the Johannesburg hearing separates extradition from local criminal liability
The Tshuma file also demonstrates why a requested person may be involved in two proceedings at once. The United Kingdom seeks access to a person for proceedings concerning alleged conduct in England. South Africa is separately concerned with conduct allegedly discovered on its own territory. The two cases can share witnesses, dates and custody facts, but they have different legal objects. The extradition case asks whether the person may be surrendered. The local case asks whether South Africa can and should prosecute a domestic offence.
In a properly managed case, the court will keep those objects distinct. The foreign request should not become an informal substitute for a South African trial. Conversely, the existence of a local allegation should not automatically destroy the foreign request. The court must identify the sequence required by the governing treaty and statute. It may need to decide whether the requested person remains in custody, whether local proceedings have priority, whether evidence can be shared, and whether any eventual surrender is subject to conditions.
This separation is especially important where the public narrative is intense. The deaths described by the British prosecution are grave, and the allegations have obvious human consequences. Yet the formal question at an extradition hearing is narrower than the question at trial. It normally concerns the existence and legal sufficiency of the request, not whether the foreign prosecution will ultimately prove every element beyond reasonable doubt. A court that treats the extradition hearing as a trial risks exceeding its role; a court that treats it as a rubber stamp risks abandoning the requested person’s rights.
The same tension appears in French law. In ordinary extradition proceedings, Article 696-4 of the French Code of Criminal Procedure lists situations in which extradition is not granted, including a political purpose, extinction of the public action, a punishment contrary to French public order, or a trial before a court that does not provide fundamental procedural and defence guarantees. The provision is not a direction to retry the foreign file. It is a set of external limits on cooperation.
French law also insists on a factual comparison rather than a mechanical comparison of labels. In Cour de cassation, Criminal Division, 8 October 2024, no. 24-81.917, the Court explained that the French courts must assess whether the facts are punishable under French law “indépendamment de la qualification donnée par l’Etat requérant”. The decision concerned the double-criminality requirement. Its significance for a murder allegation is practical: a requesting state may use its own offence name, but the French court examines the conduct described and asks whether it corresponds to an offence meeting the statutory threshold in France.
That assessment is not an evidentiary trial. The same decision states that the court does not have to assess, save for an obvious error, the value of the indices gathered against the requested person. The French judge checks the legal bridge between the alleged facts and the offence, not the complete credibility of every witness. A defence strategy that confuses those two questions may miss the strongest argument. The useful challenge may concern missing factual particulars, an offence that is not criminal in both systems, a defective warrant, or a procedural bar rather than a full merits debate.
Documentation is central to that distinction. Under Article 696-8 of the Code of Criminal Procedure, an ordinary French extradition request is accompanied by a conviction, a procedural act sending the person for trial, or an arrest warrant with a precise indication of the conduct and its date. The text also requires original or certified copies and the foreign legal provisions applicable to the conduct. That is a demanding documentary architecture. It allows the defence and the judicial authority to test what the requesting state is actually asking for.
The wording “l’indication précise du fait pour lequel ils sont délivrés” is not decorative. A request that merely says “murder” without a date, location, victim, conduct and legal basis may be impossible to test. In a file involving several alleged victims, the precision must also permit the requested person to understand whether each count is separate, whether the warrant covers conduct beyond the initial charge, and whether later prosecution would remain within the speciality rule. The South African court will apply its own law and treaty, but the same logic explains why the NPA’s statement that the documents are complete matters procedurally.
In France, consent and opposition also change the route. Article 696-15 of the Code of Criminal Procedure provides that, where the requested person does not consent, the investigating chamber of the competent court of appeal is seized without delay and gives a reasoned opinion. Its assessment concerns the legal conditions and an obvious error in the request. The provision shows how a formal demand becomes an adversarial judicial process. It also shows why a hearing date cannot be treated as a ceremonial step: the defence must decide what objections to raise, what documents to request, and whether detention remains proportionate.
The distinction between local liability and foreign surrender has a further consequence for communication. A public authority may accurately say that it has received a complete extradition request without saying that the requested person will be surrendered. A prosecutor may accurately say that charges have been authorised without saying that the accused is guilty. Counsel, journalists and courts should preserve those distinctions because later judicial decisions will depend on the precise legal posture, not on the intensity of the initial headline.
II. What would French and European safeguards add to the comparison?
A. Which legal objections can stop surrender from France
The first French safeguard is the treaty and statutory framework. Article 696 of the Code of Criminal Procedure provides that, unless an international convention says otherwise, the chapter’s conditions, procedure and effects govern extradition and also fill points that the convention does not regulate. The starting question is therefore not simply whether the foreign police want the person. It is which legal instrument governs the request, which authority transmitted it, and which refusal grounds apply.
For a British request involving France, the answer may differ from classic diplomatic extradition. The EU-United Kingdom Trade and Cooperation Agreement contains a surrender mechanism based on an arrest warrant. Its Article 597 says that cooperation through that warrant must be necessary and proportionate, taking into account the requested person’s rights and the victims’ interests, the seriousness of the conduct, the likely sentence and the availability of less coercive measures, including measures that avoid unnecessarily long pre-trial detention. The official EU-UK Trade and Cooperation Agreement text on EUR-Lex sets out that rule in Title VII.
The French courts have already applied the post-Brexit framework. In Cour de cassation, Criminal Division, 23 November 2022, no. 22-86.162, the Court referred to the TCA’s requirement that cooperation be necessary and proportionate. The case also concerned a condition about serving a sentence in the executing state. Its lesson is that the UK is no longer handled through a simple assumption that every British request is a pre-Brexit European Arrest Warrant. Counsel must identify the TCA mechanism, the relevant article, the issuing judicial authority and the specific ground of opposition.
The second safeguard is the prohibition on surrender where the requested person faces a real and sufficiently established human-rights risk. Article 3 of the European Convention on Human Rights is absolute. In Cour de cassation, Criminal Division, 6 May 2025, no. 24-85.773, the Court held that extradition must be refused when the person demonstrates a real risk of a manifestly disproportionate sentence in the requesting state. It described that assessment as “cette notion devant faire l’objet d’une interprétation stricte” and reserved it for very exceptional cases.
The narrow wording has two consequences. Serious charges do not automatically create a human-rights bar. A defence must identify a concrete risk connected to the requested person and the foreign sentencing framework. At the same time, the fact that a sentence is authorised by foreign law does not end the analysis. The French court must examine the nature of the alleged conduct, the statutory range, the circumstances in which the extreme sentence could be imposed, and any reliable evidence about how the rule operates.
Capital punishment creates a separate example of the guarantee requirement. In Cour de cassation, Criminal Division, 30 January 2024, no. 23-83.549, the Court stated that, where the death penalty is available for the facts, the requesting state must provide an undertaking “dénué de toute ambiguïté” that the penalty will not be applied to the specific person. The United Kingdom does not impose capital punishment, so this rule is not a factual issue in the Tshuma matter. It is nevertheless a useful illustration of French practice: assurances must address the individual case and the actual risk, not repeat a general promise with no operational effect.
The same approach applies to a risk of onward removal or treatment in a third country. In Cour de cassation, Criminal Division, 5 November 2024, no. 24-85.705, the Court examined whether authorities had provided an “engagement ferme et non équivoque” for a person protected against return to a country of origin. An extradition lawyer should therefore ask not only where the person will be sent first, but also what can happen after arrival: transfer, expulsion, detention conditions, access to counsel and the possibility of removal to a place where the original risk reappears.
Political purpose, fundamental defence guarantees, prescription and public order are expressly addressed by Article 696-4. The provision also illustrates a wider principle: extradition cooperation is conditional. The requesting state is not entitled to use a criminal request to pursue a political opponent, to bypass an extinguished prosecution, or to expose a person to a process that does not meet basic procedural standards. The defence must connect the objection to evidence in the request and to the applicable treaty or Convention; a general allegation is not enough.
A final structural control comes from the constitutional place of treaties. Article 55 of the French Constitution states that regularly ratified or approved treaties, once published, have “une autorité supérieure à celle des lois”, subject to reciprocity. In an international criminal case, this means that a lawyer must read the bilateral or multilateral instrument together with the Code of Criminal Procedure. The domestic code may supply the procedure, while the treaty defines a surrender condition, a channel of transmission, a speciality rule or a time limit.
B. What the Tshuma file teaches about proof, delay and guarantees
The most useful French comparison is not a prediction that France would decide the South African matter. France has no jurisdiction to rule on the South African request merely because the facts are reported internationally. The comparison instead identifies the questions that arise whenever a person is arrested in one country for proceedings in another. In the Tshuma file, the public record presently supports a clear chain: alleged deaths in England, British charging decision, flight or travel to South Africa as described by authorities, Interpol-assisted tracing, arrest in Johannesburg, a South African extradition hearing and a separate local firearms matter.
The first question is proof of identity and proof of the request. Under the French documentary model, the warrant must identify the facts and date with precision, and the requesting state must provide the relevant law. In Cour de cassation, Criminal Division, 8 April 2025, no. 25-80.279, the Court addressed the information owed to a person placed under extradition custody. It repeated the requirement that the person be informed “dans le plus court délai” of the reasons for arrest and the accusation. The case concerned the interface between Articles 5 and 6 of the Convention and the extradition procedure.
This is a practical defence point, not a technicality. A person cannot effectively challenge a request without knowing whether the arrest concerns prosecution, sentence enforcement, or a provisional step before a full request. The information must permit counsel to test the warrant, identify the alleged conduct, raise double-criminality issues, and investigate rights-based objections. If the request evolves after arrest, counsel should record when each document was received and whether the later material changes the case’s scope.
The second question is double criminality. Murder allegations will ordinarily pass the basic conceptual comparison between the United Kingdom, South Africa and France, but the legal test still operates count by count. The facts described in the warrant must amount to conduct punishable in the executing state at the relevant level. The label “murder” cannot cure a warrant that fails to describe conduct, dates or participation. Nor can a foreign charge automatically authorise prosecution for additional conduct not covered by the surrender decision.
The third question is speciality. Surrender is permission for the requesting state to proceed within the terms of its request. It is not a blank cheque for unrelated charges. For a file involving three alleged victims and separate firearm allegations, the difference is concrete. The British proceedings and the South African proceedings must remain identifiable. If the requesting state later wants to prosecute a materially different offence or conduct, it may need a further consent or judicial authorisation under the governing instrument.
The fourth question is local detention and delay. A person can remain in custody while authorities exchange documents, arrange hearings and resolve parallel proceedings. That custody has a legal limit. In Cour de cassation, Criminal Division, 17 June 2025, no. 25-82.352, the Court faulted an assessment that failed to examine whether the extradition procedure had been conducted “avec la diligence requise de manière globale” after a person had spent two years and seven months under extradition detention. The point is not that every long case must end on a fixed date. The point is that authorities must assess the whole timetable and their own diligence.
The EU-UK TCA expresses a related proportionality concern. Its Article 597 asks decision-makers to consider less coercive measures, specifically mentioning the avoidance of unnecessarily long pre-trial detention. That principle matters for any future UK-France case. It also gives international readers a way to understand the South African adjournment: a postponement may be justified to allow consultation and complete the defence process, but repeated postponements should be recorded against the reason for the delay, the custody conditions and the availability of alternatives.
The fifth question is the reliability of assurances. If the requesting state promises fair proceedings, medical care, non-application of a death sentence, protection against onward removal or access to counsel, the court must ask who made the promise and whether that authority can deliver it. The French decisions use demanding language. The death-penalty case requires an assurance “dénué de toute ambiguïté”. The refugee-removal case refers to an “engagement ferme et non équivoque”. Those words turn a diplomatic exchange into a legal test. A letter from an official without power over detention or removal may not answer the risk.
The sixth question is the boundary between the extradition judge and the trial judge. In the double-criminality decision, the French Court said that the executing judges need not assess the value of the evidence except in the case of an obvious error. That preserves the requesting state’s role in trying the allegations while ensuring that the request is not legally incoherent. For the Tshuma proceedings, this means that public reports about evidence, post-mortem findings or police theories should not be mistaken for the evidence package that the Johannesburg court must legally evaluate.
The seventh question is what happens if a rights problem is found after an initial decision. The Conseil d’État addressed that issue in Section decision, 22 December 2017, no. 408811. Interpreting Article 46 of the European Convention, it said that full execution of a Strasbourg judgment may require “la disparition de la source de cette violation”. It also held that a later extradition decision based on new elements and new guarantees requires a new decree and a fresh consultation of the investigating chamber. The quoted rule is significant because it rejects the idea that a failed surrender can be repaired by simply reusing the old decision.
That principle has an international equivalent even where the executing state is not France. A refusal based on defective papers, an unmanageable rights risk or an unlawful procedure should identify what must change before a new request can be considered. A requesting state that supplies a new warrant, clarifies the charges or gives enforceable guarantees may be able to try again. A requesting state that merely repeats the same request has not answered the reason for refusal.
The eighth question is the effect of a domestic case in the state of arrest. The South African firearm matter may affect when and how custody is dealt with. A state may have to finish its own prosecution before surrender, may arrange temporary transfer, or may decide which proceeding should take priority under its law and treaty. The answer cannot be inferred from the existence of an Interpol request. It must be found in the South African court record, the extradition treaty and the relevant statutory provisions. The public reporting establishes that two matters are proceeding on separate schedules; it does not establish the final order of surrender.
For a person or family facing a comparable case in France, the immediate checklist is therefore document-based. Obtain the arrest record, the warrant, the request, the treaty provisions, the translation, the custody decisions and every assurance. Compare each count with French law without relying on its foreign label. Check whether the request identifies facts and dates. Record when the accused was informed and when counsel received the papers. Test any claim about detention conditions or onward removal with country-specific evidence. Finally, separate objections that go to legal admissibility from arguments reserved for the foreign trial.
This checklist also explains the commercial and public-interest value of the story. Cross-border arrests produce fast headlines, but the decisive issues are often invisible: the formal warrant may not match the alert, the evidence may be incomplete, a local case may change the timetable, or a guarantee may be too vague to protect the person after surrender. An international audience needs that procedural translation. It is where a French and European lawyer can add value without claiming to decide a foreign case or prejudge an accused person.
Conclusion
The Tshuma hearing is a reminder that international criminal cooperation has two speeds. Police cooperation can locate and arrest a person within hours. Judicial cooperation must then slow down enough to verify the request, protect the defence, separate local allegations from foreign charges, and decide whether surrender is lawful. The Johannesburg court’s 8 September return date is therefore a procedural checkpoint, not a final verdict on extradition and not a finding of guilt.
The French comparison supplies a clear vocabulary for what comes next: precise documents, double criminality, speciality, fair-trial and Article 3 safeguards, reliable assurances, proportionate detention and an independent judicial decision. The French cases cited above show that the court neither ignores the requesting state’s case nor conducts the foreign trial itself. It tests the legal bridge that permits one state to place a person in the hands of another. That bridge is strongest when the warrant is precise, the timetable is disciplined and every rights guarantee is capable of being enforced.