Questions of the record
Questions about the MV Barima disaster, answered from the 209 documents in this archive. Anyone can send a question in.
These answers are written by AI, over the whole archive
Everywhere else this site only sets out what published sources said, attributed. Here the question is answered directly, by AI reasoning across all 209 documents. Evidence is named; where the record cannot settle something, that is the answer given. If this page and the record pages disagree, the record pages are right. Question five also draws on international instruments held outside the archive, and says so. Nothing here bears on the guilt of the three men charged on 28 July, who have not been tried. The research and the reasoning are done with Claude Opus 5, Anthropic’s frontier model, over the corpus described on the method page.
This page answers questions about the MV Barima disaster. Anyone can send one in, and the answers are written by AI reading across every document in the archive — currently 209 of them. Each answer opens with the short version; the working sits behind "keep reading" for anyone who wants it.
The chronology, figures and positions pages hold the record these answers are drawn from. Three men were charged with murder on 28 July and have not been tried; this page does not speculate about their guilt, and does not suggest they are being blamed for someone else's decisions. That is for the court and the inquiry.
Send a question. Anyone can. Every question gets published with its answer, and a question the archive cannot answer gets published too, saying so — because what the record cannot answer is itself worth knowing. The same form takes corrections, and documents the archive is missing.
1Who is accountable for the MV Barima tragedy?
Legally, nobody yet. Three Transport & Harbours Department employees — the captain, the chief mate and the goods superintendent — have been charged with murder and remanded. They have not been tried.
Institutionally, three bodies are answerable: the Transport & Harbours Department, which operated the vessel; MARAD, which regulates and certificates it; and the Ministry of Public Works, which owns the fleet and holds the budget. In this archive, formal action of any kind has been taken only at the first.
The reasoned view, in short
The record cannot name who is at fault, but it can locate the layers. What happened in the loading bay and the wheelhouse on 18 July is for the court. Why that vessel was the only one on the route is a capital and procurement question, and it is the layer against which no formal action of any kind appears. That does not clear the operator — the government's own 2025 finding of "longstanding issues of negligence, mismanagement, and accountability" at the department speaks against it.
Keep reading
What has actually happened to people so far.
Table view — all formal action recorded, not only charges
| Level | Charged | Held or detained | Suspended or on leave |
|---|---|---|---|
| Transport & Harbours Department (operator) | 3 | crew from 19 July; a superintendent 22 July | loading and dispatch team, 20 July |
| MARAD (regulator) | 0 | none | none |
| Departmental board | 0 | none | none |
| Ministry of Public Works | 0 | none | none |
The three were charged on 28 July on the advice of the Director of Public Prosecutions. Before that, crew members were held in custody from 19 July and a departmental superintendent was detained on 22 July. The team responsible for loading and dispatch was suspended on 20 July.
Every one of them works for the operator. In the 209 documents this archive holds, there is no charge, no suspension, no leave and no detention involving anyone at the regulator, on the departmental board, or in the ministry.
Who is answerable for what, on paper.
The Transport & Harbours Department operated the vessel. It employed the crew, sold the tickets, loaded the cargo and dispatched the ferry. It has its own board.
MARAD, the Maritime Administration Department, is the regulator. It surveys and certificates ships, licenses them, and — under Guyana's Shipping Act — investigates marine casualties. The body that surveys and certificates vessels is also the body its own law names to investigate their casualties.
The Ministry of Public Works owns the fleet and holds the budget; Juan Edghill is its minister. MARAD's responsible minister is Deodat Indar, at Public Utilities and Aviation. Both have been the subject of demands, and the demands differ: some are for resignation, some only for recusal from the investigation.
What the record does and does not show.
It shows a vessel built in 1939 still running the route because the ferry bought to replace it could not dock at the port, because the wharf it needed was never finished. It shows no certificate of survey, load line or passenger capacity in any of the 209 documents. It shows a regulator whose Director-General, asked on what basis the passenger rating was raised from 150 to 394, said: "I don't have that information." It shows 46 more people aboard than the manifest recorded.
What it does not show is a decision, by a named person, that this archive can point to and call the cause. The inquiry exists to find that, and it has not sat.
Both arguments are on the record. Five opposition parties and the Amerindian Peoples Association have called for Minister Edghill to resign, some also naming Minister Indar. The Transparency Institute of Guyana asked for something narrower — that Edghill, Indar and MARAD's senior leadership recuse themselves from the investigation rather than leave office. The Guyana Human Rights Association has not called for anyone's resignation; its objection is to how the inquiry itself was set up.
Against them stand two things. The government's own position, put by Minister Edghill on 19 July, is that the Transport & Harbours Department is the service provider and MARAD the regulator, and that the disaster stemmed from operational breaches at ground level rather than from regulatory failure — a position he stated before anyone was charged. And the Kaieteur News "Peeping Tom" column argues that the department has its own board, that a minister's duty is "not to supervise the loading of cargo or verify passenger manifests on a daily basis," and that the cause is not yet established. All of it is set out at length on the positions page.
Sources for this answer 9
- MV Barima captain, two crew members charged with murder
- Major discrepancies found on MV Barima manifest
- MV Barima team suspended as probe moves ahead
- 3rd person detained as police probe continues in MV Barima tragedy
- MV Barima Disaster... 73 bodies found, 69 identified
- Opposition unites in call for Edghill to go
- Transparency Institute wants International Maritime Organisation to investigate MV Barima tragedy
- GHRA demands Parliament-led COI into MV Barima sinking
- Should Minister Edghill resign?
Each links to this archive’s page for that document, which carries the publisher, the date and a link to the original.
2By international standards, how should a disaster like this be handled?
There is a recognised sequence, set out in the IMO's Casualty Investigation Code: rescue and log everything; secure the evidence including the wreck; stand up an independent safety investigation separate from any criminal case; look after the families; publish the report and act on it.
Guyana is not in breach of a treaty — none of the major conventions reach a domestic ferry. Measured against the sequence itself, four things stand out: the criminal charges came before the inquiry was constituted, the inquiry still has no gazetted instrument or deadline, no commitment to publish its report appears anywhere in this archive, and it is not established that the wreck will be examined before any decision is taken to leave it.
The reasoned view, in short
No treaty was broken, and that is the weaker half of the answer. The stronger half is that where no external standard reaches, someone still sets the domestic one — so the standard becomes locatable rather than inherited. Whether Guyana's administration could have applied the international standard to a domestic route is not established here in either direction, and those two possibilities call for opposite remedies.
Keep reading
This answer uses sources from outside the archive
Every other answer on this site reasons only over the 209 documents the archive holds. This one reaches outside them, to the international instruments that set the standard. Each is named and linked so you can check it yourself.
First, the legal position, because it is usually got wrong. The big international rules — SOLAS, the Load Lines Convention and the Casualty Investigation Code — apply to ships on international voyages. The MV Barima ran between two Guyanese ports. It sat outside all of them. It was never required to hold a SOLAS passenger safety certificate, and no requirement to hold one was in force.
That gap is well known. In 2022 the IMO adopted Model Regulations on Domestic Ferry Safety expressly because domestic ferries fall through it — the resolution records the IMO as "deeply concerned with the high number of lives lost through domestic ferry accidents around the world." Those model regulations are voluntary.
So the standard below is not a legal obligation Guyana has broken. It is the sequence the same government already applies to its ships on international voyages, because Guyana acceded to SOLAS in 1997 per the IMO's own status of treaties. The question it raises is why a domestic route should get a lower one.
Stage one — the first day: rescue, and one reliable account
What should happen. A maritime rescue coordination centre runs the response and logs it: when the distress call came, who received it, what was tasked and when. The state publishes what it knows, flags what it does not, and does not name a cause. Under the Code a casualty involving even a single death is a "very serious marine casualty," which triggers a safety investigation as of right.
What happened. A distress call at about 23:01 is reported by many outlets, several of which name air traffic control as the recipient. On 25 July the Prime Minister said of a Mayday from the captain, and an instruction to passengers to put on life jackets: "We have no such record." The rescue coordination log has been requested publicly and not produced. The first air asset with a time against it in this archive is a private aircraft, airborne at 01:36; the Guyana Defence Force air corps is reported as mobilised that night with no time given. And on 19 July, before any investigation had reported, the responsible minister named a cause in public — a wave and the tide — and ruled out engine failure, water ingress and overloading. The Kaieteur News editorial of 24 July put the objection: those assurances "were not based on an independent inquiry."
Stage two — the first days: secure the evidence
What should happen. The IMO's guidelines for investigators are explicit that collecting evidence "that can deteriorate or disappear over time will always be the first priority." The wreck is located, protected from interference and examined — by remote vehicle and divers if not raised. Manifests, ticketing data, CCTV, maintenance files and certificates are sequestered before anyone can lose them.
No international instrument requires a wreck to be raised for investigation purposes — separately, the Nairobi Wreck Removal Convention can require removal where a wreck is a hazard, which is a different question. The Estonia was deliberately left in place and surveyed by sonar and remote vehicle; the Sewol was raised after about three years, and the panels that examined it did not settle the cause beyond dispute. What matters is examination, not recovery.
What happened. The wreck was positively located at about 13:31 on 20 July, after information from a fisherman and survey work by private vessels; the state credits civilian and private effort for it. A French dive team reported on 27 July that entering was too dangerous. On 28 July the Prime Minister said the vessel is "likely to remain there as is," and a draft agreement circulated to families three days earlier asks them to accept it as an undisturbed memorial. Whether the hull will be examined before that is settled does not appear in this archive. On the documents: the manifest was published, and the CCTV review's conclusion but not the footage. No certificate, maintenance file, rescue log or tender file has been.
Stage three — the first week: an independent safety investigation
What should happen. This is the heart of the Code, and it has four parts. The safety investigation "should be separate from, and independent of, any other form of investigation." Its purpose is preventing the next disaster, not assigning blame — and investigators "should not refrain from fully reporting on the causal factors because fault or liability may be inferred," so a pending prosecution is not a reason to hold back. The investigating body must be independent of anyone whose interests could conflict. And crew being interviewed must be told of any risk of self-incrimination and of their right to legal advice and to remain silent; the joint IMO/ILO Guidelines on Fair Treatment of Seafarers add that a seafarer must be told which kind of interview it is, because the consequences differ.
What happened. The Prime Minister committed to an inquiry on 22 July, the President announced one on 25 July and named five members on 26 July. All five sit outside Guyana, which answers the independence concern about MARAD investigating itself — though the announcement had promised "local and international experts" and no Guyanese member was named. As of the most recent document here, the establishing instrument is not gazetted, and there is no secretary, no reporting deadline and no published budget. The criminal charges were laid on 28 July, before the inquiry was constituted. Nothing in this archive records whether crew were told which kind of interview they were giving.
Table view
| Day | Date | Step |
|---|---|---|
| 0 | 18 July | The MV Barima sinks |
| 4 | 22 July | Prime Minister commits to a Commission of Inquiry; declines to give a timeline |
| 7 | 25 July | President announces an independent international Commission |
| 8 | 26 July | Five commissioners named; instruments "being formalised" |
| 10 | 28 July | Three operator employees charged with murder; the Commission is not yet constituted |
| — | — | No gazetted instrument, secretary, deadline or budget appears in this archive |
| 1 | 29 Sept 1994 | Separate case, for comparison: Estonia, Finland and Sweden establish a joint commission the day after the MV Estonia sinks |
Stage four — throughout: the families
What should happen — and here the international instruments are silent. The Casualty Investigation Code deals with seafarers, not with bereaved families; the fair-treatment guidelines likewise. So this stage is not measured against a standard, because there is no maritime one. What is on the record instead is what the Guyana Human Rights Association has proposed: one reliable point of contact, published and updated lists, and independent legal advice for bereaved families funded by the state, so that anything they are asked to sign is explained to them by someone who is not the state. That is a proposal by a Guyanese civil-society body, not an international requirement, and it is recorded here as such.
What happened. Information centres were set up and names of the rescued and missing published. Support has been offered — transport, funerals, counselling. On the other side: the Guyana Human Rights Association has asked for state-funded independent legal representation for families and has not got it; the Transparency Institute of Guyana points to the aftermath of the 2023 Mahdia school fire, where families were still in court years later; and families are being asked to sign a declaration about the wreck while no inquiry is sitting.
Stage five — the end: publish, and act
What should happen. Chapter 14 of the Code is mandatory: the final report on a very serious casualty goes to the IMO and "shall be made available to the public." Recommendations are implemented. If new evidence appears, the investigation is reopened.
What happened. Nothing yet, because nothing has started. Guyana's Opposition Leader has said reports of previous commissions were not published and has asked for this one to be livestreamed and released in full. No commitment to publish appears anywhere in this archive.
The comparison, in one table
| Stage | What the standard expects | Where Guyana stands |
|---|---|---|
| First day | Logged rescue timeline; publish facts, not causes | Rescue log requested and not produced; a cause named publicly on day one |
| First days | Wreck located, protected, examined; records sequestered | Wreck located largely by private effort; examination before any memorial decision not established; the manifest was published, but no certificate, rescue log, maintenance file or tender file has been |
| First week | Independent safety investigation, separate from any criminal case | Inquiry announced day 7, members named day 8, not yet constituted; charges laid day 10 |
| Throughout | No maritime standard; the proposal on the record is Guyanese | Information centres and lists provided; funded legal representation asked for by the GHRA, not provided |
| The end | Report published; recommendations acted on | No commitment to publish appears in this archive |
One comparison, and its limits. After the Estonia sank on 28 September 1994, three countries stood up a joint investigation commission on 29 September — the next day. The Estonia was an international-voyage casualty with three states and treaty machinery already engaged, so the comparison is one of speed rather than of like institutions.
Sources for this answer 9
- Huge wave suspected in Guyana ferry disaster as 67 rescued
- MV Barima disaster demands accountability
- Timeline of the rescue effort for the MV Barima ferry sinking?
- Too dangerous to enter sunken MV Barima – French diver
- Govt. floats leaving MV Barima on seabed as memorial site
- International, regional experts appointed to conduct thorough investigation of MV Barima tragedy
- Pres. Ali names 5-member COI to investigate MV Barima disaster
- Transparency Institute wants International Maritime Organisation to investigate MV Barima tragedy
- GHRA demands Parliament-led COI into MV Barima sinking
Each links to this archive’s page for that document, which carries the publisher, the date and a link to the original.
3What conditions was the ferry operating under, and what could have been done differently?
The record does not show what capsized the ferry. It shows an 87-year-old vessel alone on a route because its replacement had nowhere to dock, a count of passengers that was wrong by 46, and no published document saying what it was allowed to carry.
Two of the three were documented before 18 July, in public and largely on the government's own record. The passenger gap was not: it emerged afterwards, from CCTV.
The reasoned view, in short
Of the four measures usually described as things that could have prevented this, two would actually have done so and two would not. And the deeper pattern is not about any of them: the question of whether this vessel should still be carrying passengers was asked repeatedly from outside — in 2010, in 2015, and by the Amerindian Peoples Association over a period of years — and no body inside the operator, the regulator or the ministry is recorded answering it.
Keep reading
It was the only vessel on the route. A replacement, the MV Ma Lisha, was contracted in January 2021 for about US$12.7m and delivered in 2023. It was too big for Port Kaituma, so the wharf had to be rebuilt. That contract went out in December 2022 at G$1.4bn. In February 2024 the work stopped at about ninety per cent complete after cracks appeared and, as reported, the design was found deficient. It was still unfinished when the ferry sank two and a half years later. So the Ma Lisha served other Region One ports and the 87-year-old Barima kept the Port Kaituma run. The Prime Minister said so on 25 July: "The delay has contributed to us not being able to use the MV Ma Lisha to journey from Georgetown to Port Kaituma."
Its condition had been raised in public. In 2015 a master mariner wrote about this named vessel, citing a newspaper inspection report that described its deck, winch and generator area as "a mass of junk and rust." That inspection report was searched for and not found; only the letter citing it is in this archive. In March 2026 — four months before the sinking — G$124.5m of hull work was put out to tender. Whether that work was ever done is answered nowhere in this archive.
The department's own problems were on the government's record. In March 2025 the government's information service reported the Transport & Harbours Department addressing "longstanding issues of negligence, mismanagement, and accountability." That is about governance rather than about this hull, which is why it is recorded separately here.
The record of who was on board was wrong by 46. The manifest said 133. A review of boarding CCTV afterwards showed 179. Only 35 of the first 67 people rescued appeared on the manifest at all. The government has since confirmed that people boarded without being ticketed. None of this was caught by a check before departure; it emerged afterwards, from the CCTV.
Nobody can say what the ferry was allowed to carry. Three different official cargo limits were given in the first week, from 120 to 284 tonnes, and no certificate has been published. See question 4.
What was available and was not in place. Four things:
Finish the wharf — or tranship. Within a week of the sinking the government set up exactly that arrangement: passengers to Kumaka on the Ma Lisha, then smaller boats onward, with a separate free cargo barge alongside it. Whether it works as a permanent substitute is not established here. Part of the answer to why it was not running earlier is that the Kumaka stelling had only just been completed — the Minister said on 22 July it was still awaiting commissioning. That explains some of the delay. It does not explain the years before it.
Count the passengers at the gate. A headcount against the manifest before departure would have shown a count 46 above the manifest. It requires no equipment. (46 is the difference between two totals, not a count of individuals: some people aboard were unrecorded, and some on the manifest did not sail.)
Produce the certificate. If the permitted load is genuinely 120 tonnes and the manifest says 268, that is visible to anyone holding the paperwork.
Retire the hull. Repeated rounds of repairs — 2017, 2021, 2022 and 2024, with more tendered in March 2026 — kept an 87-year-old vessel in passenger service. MARAD's Director-General argues that age is not determinative, because modern thin steel is stronger than the old; a letter-writer makes the same point, that age is not the same as unseaworthiness. Both are on the positions page. A replacement plan did exist: the Minister told Parliament the Ma Lisha would replace the Barima and the Kimbia. What the archive contains no trace of is a date by which the Barima would stop carrying passengers regardless of whether its replacement could dock.
One caveat. None of this proves what sank the ferry. Survivors describe engine trouble before departure and water coming in; the Minister said on day one there was no engine or mechanical problem. Those accounts have never been reconciled, and the inquiry has not tested either.
Sources for this answer 7
- Long delay in completion of Kaituma stelling kept MV Barima on the choppy sea. Who is to blame?
- Over $440M to be spent on major ferry repairs
- T&HD vessels are in disrepair because they haven't been maintained
- Crucial reforms underway within transport, harbours department – Min Edghill
- Major discrepancies found on MV Barima manifest
- MV Barima Disaster... 73 bodies found, 69 identified
- 87-YEAR-OLD MV BARIMA REMAINED IN SERVICE DESPITE REPLACEMENT PLANS
Each links to this archive’s page for that document, which carries the publisher, the date and a link to the original.
4Was the boat overloaded?
Nobody can check. The manifest recorded 268 tonnes of cargo. Three official cargo limits were given in the first week: 284 tonnes, then 126, then 120. The load is inside the first and more than double the other two.
No certificate settling the point appears in any of the 209 documents, and MARAD's Director-General said he did not have the basis for the vessel's rating. MARAD says it was not overloaded. A survivor describes cargo boxes stacked at the rear. Neither account can be checked here.
The reasoned view, in short
Whether the certificate exists or does not, the same thing follows: the system could not state the vessel's permitted load consistently, and the people responsible gave three different figures in six days. Separately, the load-line finding — even taken entirely at face value — answers a question about draught, not the question about stowage and stability that a survivor account actually raises.
Keep reading
Table view
| Stated by | When | Cargo limit | 268 tonnes against it |
|---|---|---|---|
| Minister Juan Edghill | 19 July | 284 tonnes | within the limit |
| MARAD Director-General Thomas | 21–22 July | 126 tonnes | 2.1× the limit |
| MARAD Director-General Thomas | 24–25 July | 120 tonnes | 2.2× the limit |
| Original certification, as described by MARAD | 1938 | 120 tonnes | 2.2× the limit |
Minister Edghill said on 19 July the vessel was licensed for 284 tonnes, which would put the load inside the limit. MARAD's Director-General said on 21–22 July it was 126 tonnes. He revised that on 24–25 July to 120 tonnes. Either of the last two would put the load at more than double the figure MARAD gave.
The vessel's original rating, which the same official dates to 1938, was 120 tonnes and 150 passengers. (The builder's register dates the hull to 1939; the archive flags that conflict on the chronology page.) It sailed licensed for 394 passengers according to MARAD, or 397 according to the Minister — the passenger figure does not agree either.
So which is right? No survey certificate, load line certificate or passenger capacity certificate appears in any of the 209 documents this archive holds. Asked when and on what basis the passenger rating went from 150 to 394, MARAD's Director-General said: "I don't have that information."
What about the "not overloaded" finding? MARAD's Director-General said on 22 July the vessel was not overloaded. He was citing the load line — the mark on the hull showing how deep the vessel is sitting — rather than the tonnage certificate. That is a legitimate thing to cite, and it means his two statements do not contradict each other. But it also means the overloading question was answered by looking at where the water came up the hull, not by checking a document.
What survivors say. One describes cargo boxes stacked at the rear of the vessel — the sole source in this archive for that detail. Another says water entered through the hatch and that the ferry was stuck on a sandbank at the wharf for almost an hour before leaving. These are accounts, not findings, and they are recorded as such on the chronology page.
Sources for this answer 5
- Huge wave suspected in Guyana ferry disaster as 67 rescued
- MV Barima was not overloaded – MARAD Director General
- MV Barima Disaster... 73 bodies found, 69 identified
- Major discrepancies found on MV Barima manifest
- Survivor says may have lost seven relatives in MV Barima capsizing
Each links to this archive’s page for that document, which carries the publisher, the date and a link to the original.
5How many people actually died?
76 people survived. Somewhere around 103 did not. 179 were aboard, 76 were rescued, 73 bodies have been recovered and about 30 remain unaccounted for.
No government statement in this archive gives a death toll. Official releases report bodies recovered and stop there. The 103 has never appeared in one.
The reasoned view, in short
The missing death toll does not need a motive to explain it, and this archive does not supply one. What does need explaining is the combination: two decisions that are each defensible alone — do not publish a toll, leave the vessel where it is — produce roughly thirty families with neither a body nor a death certificate, and no stated process for obtaining either.
Keep reading
Table view
| Measure | People | Basis |
|---|---|---|
| Aboard | 179 | Boarding-area CCTV review; the manifest recorded 133 |
| Rescued | 76 | Settled figure from 21 July |
| Bodies recovered | 73 | Unchanged since 24 July |
| Unaccounted for | about 30 | The Prime Minister's phrasing; never described as dead |
| Recovered plus unaccounted | about 103 | Never published as a figure by any government source |
The arithmetic is all from official figures. The complement of 179 was established by reviewing CCTV of the boarding area rather than from any paper record. 76 were rescued. 73 bodies have been recovered. That leaves about 30 people unaccounted for. 73 plus 30 is 103.
Why it matters. "Unaccounted for" is a reasonable thing to say while a search is running and people might yet be found alive. It is harder to sustain now. On 28 July the Prime Minister said "it's obvious that bodies are trapped in the wreck," and that the vessel is "likely to remain there as is."
Once the state has said both that people are inside and that the wreck is likely to stay where it is, the gap between "unaccounted for" and a death toll is a question the state has not answered.
One thing that does check out. The figures published on 20 July — 27 recovered, 69 rescued, 83 missing — add to 179. The figures published on 24 July — 73 recovered, 76 rescued, 30 unaccounted — also add to 179. The apparent collapse in the number missing, from 83 to 30, is fully explained by bodies being recovered and the survivor count being corrected. It looks alarming and it is not; the same standard applies in both directions.
Sources for this answer 4
- Government tightens ferry safety procedures following MV Barima tragedy
- MV Barima Disaster... 73 bodies found, 69 identified
- Govt. floats leaving MV Barima on seabed as memorial site
- Major discrepancies found on MV Barima manifest
Each links to this archive’s page for that document, which carries the publisher, the date and a link to the original.