Summary

  • The 3 May 2026 Board resolutions amending Articles 17 and 18 say they were made under Article 25, Section 25.2, and the Empowered Community stage ran from 11 May to 3 July 2026, ending four in support, one abstention, no objection.
  • The threshold that decided it counts bodies, not hands: three of five Decisional Participants in support with no more than one objecting, which makes approval a three-body purchase and refusal a two-body one.

Start with the correction, because the title of the record invites the opposite reading.

The document this piece begins from is a notice filed under a long name: the ASO's notice to the Address Supporting Organization community, on the Approval Action Community Forum for fundamental bylaws amendments to Articles 17 and 18. Read that title quickly and you will assume Articles 17 and 18 are the amendment machinery. They are not, and the correction matters for everything below. Article 17 is the Customer Standing Committee. Article 18 is the IANA Naming Function Review. The machinery is Article 25, Section 25.2, plus Annex D, Article 1. Nothing in the record reviewed for this piece amends either of them.

So what did ICANN do on 3 May 2026? The Board's approved resolutions record three decisions in sequence. Resolution 2026.05.03.11 approves the amendment to Article 18, Section 18.2(b), made under Article 25, Section 25.2. Resolution 2026.05.03.12 approves the amendments to Article 17, on the same basis. Resolution 2026.05.03.13 approves an updated Customer Standing Committee Charter, which the same material says takes effect only if and when the Article 17 amendments are approved through the Empowered Community process. The Board initiated that process on 25 January 2026.

Two gates, then, not one. The Board's side is a three-fourths vote of all Directors, quoted in a GNSO readiness paper of 15 May 2018 as the text of Section 25.2(b). The community's side is a separate count entirely, and it is the count that decides whether the Board's arithmetic is ever brought to bear.

The counting unit is a body

The Empowered Community has five Decisional Participants: the ASO, the ccNSO, the GNSO, the ALAC and the GAC. When the decisive period closed, the record states the test that was applied. The message announcing expiry of the Approval Action Decision Period, dated 3 July 2026, puts it as at least three Decisional Participants in support of the Approval Action, with no more than one Decisional Participant objecting.

Now do the arithmetic yourself, because the shape of it is the story. Five deciders, three needed in support, at most one objection allowed. Three supports carry the action unless the remaining two both object. Two objections defeat three supports. Put plainly: approval costs three of five, refusal costs two of five. That is my subtraction from the printed test, not a finding of ICANN's, and it describes the structure of the fractions rather than anyone's behaviour. A coalition of three bodies is sufficient. A pair is sufficient to stop one. The fifth signature is decorative once the other four have lined up.

What actually happened is recorded, dated, on ICANN's own Empowered Community action page. The action started on 11 May 2026 and terminated on 3 July 2026. The ccNSO Council supported on 19 June, the ASO on 29 June, the ALAC on 30 June, the GNSO on 2 July. The GAC abstained on 2 July. The Empowered Community Administration wrote to the ICANN Secretary approving the amendments on 3 July, and that letter was published on 7 July.

Four supports, one abstention, no objection. The threshold was cleared with room to spare, which means the asymmetry never had to work for a living on this file. Worth saying plainly: no document reviewed here states what difference the abstention made to that arithmetic, and none is inferred. An abstention is not an objection on the face of the printed test, and that is as far as the record goes.

Two clocks in one notice

The other piece of arithmetic sits in the timeline, and it is harder to dismiss because both numbers are printed on the same page.

The ASO notice fixes the forum window twice in a single sentence: it ends at 11:59 pm PDT on the 30th day after 11 May 2026, which the notice prints as 9 June 2026, or during the next scheduled ICANN Public Meeting, ICANN86, taking place 8 to 11 June 2026. The same notice puts the comment deadline at the ASO Secretariat at 9 June 2026.

Then the Decision Period rule. The GNSO's 12 December 2019 guidance on Annex D reproduces Section 1.4(a) as ending the Approval Action Decision Period at 11:59 p.m. on the 21st day after the expiration of the Community Forum Period, and labels forum expiry Day 0 with the Decision Period deadline as Day 21.

Apply that rule to both printed endpoints. Nine June plus twenty-one days is 30 June 2026. Eleven June plus twenty-one days is 2 July 2026. The recorded expiry, in the same 3 July message that states the threshold, is 23:59 PDT on 2 July 2026. Only the second endpoint reconciles with it. This is my subtraction, using a reproduced rule and two printed dates, and it establishes something narrow: the record is consistent with the meeting branch and inconsistent with the calendar branch. It does not tell us which branch was formally selected, and no reviewed document does.

Even ICANN's action page carries the tension, in miniature. Its phase graphic shows a forum phase running 12 May to 12 June, captioned as thirty days and extensible if the forum is held at an ICANN meeting. The page's own correspondence table, which records the operative dates, is the part that lines up with the Decision Period that actually expired.

What the amendments do to the clocks

Article 18 first, because here the change is crisp. The 11 February 2026 redline for Article 18 confines itself to Section 18.2(b) and does one thing: the five-year period stays five years, and its starting event moves. Before, the period ran from the date the previous review team was convened. After, it runs from the date that team submitted its final report to the Board under Section 18.10(c). Section 18.2(a) and the delay mechanism in Section 18.2(c) are unchanged on the text read.

On the record, the review team for the second review was convened on 10 September 2023 and delivered its final report on 4 September 2025. The Board's material then does the honest thing and prints a worked example: without the change the next periodic review must convene by 9 September 2028; under the amendment it would be required to convene by 3 September 2030. Those two dates are two years and six days apart. That subtraction is mine, between two dates on one page.

Article 17 is where the framing gets slippery. The 11 February 2026 redline for Article 17 covers composition, alternates identified by the Registries Stakeholder Group and the ccNSO, alternate liaisons identified by each appointing entity with roles and selection methods left to the Committee, and the effectiveness review. On that review, as retrieved, it moves the interval to every five years measured from delivery of the final report of the prior review, with a shorter interval available at the request of the Committee, the ccNSO, the GNSO, the Board or the PTI Board. The same redline shows the earlier arrangement beside it: a first review two years after the first Committee meeting, then every three years.

Now put that next to the Board rationale, which describes the same change as a move from every two years to every five years. Both formulations sit inside one amendment package. So how much longer between reviews? Three years, if you take the redline's three-year cycle as the baseline. Two years, if you take the rationale's every two years. The answers differ by one printed year, and the record reviewed does not reconcile them. That subtraction is mine, made between two baselines printed in two documents of the same package.

Where the reading stops

Here is the boundary, and it is not a small one.

The consolidated Bylaws page is headed as amended 3 July 2026, and its contents list both Article 25 and Annex D. I requested it twice. Both retrievals returned text that stopped inside Article 1, nowhere near either provision. Two attempts on one instrument class in one sitting is my limit, and I am not going to dress the gap up as a scoop. What it means is simple and it belongs next to the numbers rather than at the end of the piece: the three-of-five test and the twenty-one-day rule are quoted here from a 2019 GNSO reproduction and applied from a 2026 staff message, and the three-fourths Board vote is quoted from a 2018 GNSO paper. None of the three is quoted from the consolidated instrument that carries it. The Secretary's notice of 11 May 2026 is likewise cited only through the action page's correspondence table; the notice itself was not obtained in readable form here.

That is the finding, and it is a finding about documents rather than about people. ICANN changed two of its own review intervals this year, and the consolidated text that would let an outsider check the threshold at the same moment is the one document in the set that would not open to a reader. The directory entry for the notice the whole action began with is here.

So what does a reader plan against? Not the rationale, when the redline prints a different baseline. Not the thirty-day graphic, when the correspondence table records a Decision Period that only one printed endpoint produces. And not a three-of-five headline that treats the fifth body as the interesting one. Two bodies can refuse. That sentence is the one to carry away from this file, and it is printed, twice over, in documents anyone can open.