Who Verifies the Verifier?

On 14 May 2018, the Hong Kong Companies Registry issued a certificate of incorporation to International Growers Standards Organization Limited — 國際農民標準組織有限公司 — under company number 2694452. The certificate is an ordinary document, and that is precisely what makes it useful. Anyone who wants to establish that IGSO is a registered entity in a named jurisdiction can take that number to the Registry's own public record and get an answer without asking IGSO anything at all.

That is a small fact. It is also the correct place to start, because almost everything IGSO publishes argues that a claim about a product should be checkable rather than believed. An argument shaped like that has an obvious edge, and it is entirely fair to turn it around.

A Question We Left Standing

In an earlier article setting out trust infrastructure as five layers, IGSO wrote that the governance layer must answer who governs the system, who defines the standards, who verifies the verifiers, and who resolves disputes. The question was posed in order to distinguish a governed institution from a software tool. It was not answered. This article answers the part of it that points inward.

The Most Common Answer Is the Weakest One

Ask a verification body who verifies it and the usual reply is accreditation: it is assessed by an accreditation body, which is itself recognised under a wider arrangement. This is real work performed by serious institutions, and nothing here disputes its value. But as an answer to the question it has a structural problem — it moves the question up one level rather than resolving it. Whoever accredits the accreditor is the same question again, one storey higher.

In agricultural trade the recursion has a practical cost. A buyer in Hong Kong holding a carton from a province they have never visited is not going to audit an accreditation body in a third country before deciding whether to accept the shipment. The chain may be sound and still be unreadable at the moment and place where someone actually needs it. That is the same objection IGSO makes about a certificate sitting in a drawer: correctness that cannot be reached at the point of use does limited work.

The second common answer is reputation — years in operation, members served, logos on a page. This one is circular by construction. Reputation is the aggregate of people having trusted an institution before. Offering it as grounds for trusting the institution now assumes the thing being asked about.

Four Things That Make a Verifier Checkable

First, a legal identity that a stranger can look up. A jurisdiction, a date, a number held by a public registry rather than by the organisation itself. Registration is emphatically not a mark of quality — a registry number establishes that an entity exists and says nothing about whether it is any good. What it does is convert the question "who are you" from something answered by a website into something answered by a government record. A verifier that cannot be located in any public register is asking for trust while providing no instrument for testing it.

Second, criteria published before they are applied. IGSO maintains six numbered standards — producer and product authentication, freshness certification, aging certification, location verification, multi-node safety verification, and data space application — published openly and in full, in English, Traditional Chinese and Simplified Chinese, ahead of any product being assessed against them. Criteria kept private can be adjusted after the fact to fit a result. Criteria published in advance can be read by the party being assessed, by the buyer relying on the outcome, and by anyone who wishes to argue that they are wrong. Being open to that argument is not a weakness in a standard. It is most of what makes it one.

Third, a position with no competing interest. IGSO is based in Hong Kong, and the reasoning behind that has been set out at length elsewhere; it does not need rebuilding here. The narrow point that belongs in this article is about the kind of evidence involved. Hong Kong does not grow fruit at commercial scale, so a standards body seated there is not adjudicating between origins while holding one of them. That is a structural fact about a place, and a reader can confirm it independently. It is a different class of assurance from a promise of impartiality, which can only be accepted or refused.

Geography is the weaker half of that answer, though, because it addresses where the body sits rather than what it is allowed to do. The stronger half is constitutional. IGSO is incorporated as a company limited by guarantee, and its Articles of Association carry three provisions that a reader can pull and check: the income and property of the organisation are applied solely to its objects, with no portion paid to members by dividend, bonus or otherwise; no member of the board may be appointed to any salaried office of the organisation; and on winding up, any surplus cannot be distributed among the members at all, but transfers to another institution carrying an equal or stronger version of the same restriction.

The point is not that these clauses make anyone virtuous. It is that they remove a specific pressure. A body that cannot pay its directors a salary, cannot distribute to its members, and cannot recover its assets on dissolution has no financial position that improves when a certificate continues to hold. That is the pressure the next section is about, and it is worth noticing that the answer to it is a matter of public record rather than a matter of trust.

Fourth, an anchor that the party benefiting cannot move. This is the function of the resolver at id.igsoworld.org. When a mark on a product resolves, the answer arrives from an institutional address rather than from a system the seller controls. A trust mark that only its beneficiary can issue and interpret is a brand asset that happens to be shaped like a verification.

“An institution that can only ever return a positive result is not verifying anything. It is publishing.”

A Verifier Is Tested by What It Withdraws

The fourth point carries an obligation that is easy to state and uncomfortable to operate. If the resolver is the authority, it has to be able to return a negative — a status that has lapsed, a certificate suspended, a batch that no longer holds what it held last week. Live Cert is built on the premise that certification is a status rather than a document, and a status that genuinely updates is one that can expire and be revoked.

The awkward part is who that acts against. Withdrawal is exercised against a party that is usually paying, usually cooperative, and usually not at fault in any deliberate way. An institution that finds withdrawal too costly will discover reasons not to do it, and those reasons will sound reasonable each time. The credibility of every positive result a system issues rests entirely on the existence of negative ones.

This is the same reason certification and commercial promotion have to sit apart. A body that both certifies a product and sells the promotion of it has an interest in the certificate holding. Separating those functions is not fastidiousness about org charts; it removes the specific pressure that makes withdrawal hard.

What Is Published Today, and What Is Not

Applying the four tests to IGSO produces a mixed result, and the mixed result is the honest one to publish. Checkable today, without IGSO's assistance: the entity, its jurisdiction and its registration; the six standards, in full, in three languages; the resolver and the institutional address a mark resolves to; and the relationship between the three ecosystem products, each of which states what it covers.

Not yet published to the same standard: the composition of the body that approves a standard before it takes effect; the procedure by which a standard is revised, and what notice a certified party receives when it is; and the route by which a disputed certification decision is escalated, together with who decides it finally. These are ordinary things for a buyer, a regulator or a producer to ask about, and at present the answer is that they are documented internally rather than openly.

A body that publishes only the parts of its governance that flatter it has produced a brochure. Setting out the gap in public is how it becomes something to be measured against, and the intention is to close it in the same place it has been named.

The Arc, in One Line

A standard is only as credible as the institution behind it, and an institution is only as credible as the number of its own claims a stranger can check without its help.

Frequently Asked Questions

Who is IGSO, in verifiable terms? International Growers Standards Organization Limited (國際農民標準組織有限公司) was incorporated in Hong Kong on 14 May 2018 under the Companies Ordinance (Chapter 622 of the Laws of Hong Kong), with Companies Registry number 2694452. That registration is a matter of public record and can be checked independently of anything IGSO publishes about itself.

Is IGSO accredited by an accreditation body? IGSO operates as a standards organization that publishes its criteria openly and anchors certification status to an institutional resolver. It does not claim third-party accreditation, and this article deliberately argues that accreditation alone would not answer the question in any case, because it relocates it rather than resolving it.

Why does publishing standards in advance matter? Criteria that are published before assessment can be read by the party being assessed and by anyone relying on the result, which makes them contestable. Criteria disclosed only after a decision cannot be distinguished from criteria written to fit it.

What does the resolver at id.igsoworld.org actually do? It is the institutional address a product mark resolves to, so that the answer about a product's certification status comes from outside the party selling it — including when that answer is that the status no longer holds.

What parts of IGSO governance are not yet public? The composition of the body approving a standard, the revision procedure and the notice it requires, and the escalation route for a disputed certification decision. This article states that openly rather than omitting it.