Which Clause Did It Name?

Last week Congress passed a funding act. Buried in it is one sentence about a scheduled federal hemp provision, and if you sell consumable hemp in Texas you have almost certainly been told what that sentence does.

Here is the sentence. It is section 2019 of the enrolled act, quoted in full:

> Until December 11, 2026, the amendments made by section 781 of division B of Public Law 119-37 (7 U.S.C. 1639o note) shall only apply with respect to products described in paragraphs (1)(C)(ii)(I) and (1)(C)(iv)(I) of section 297A of the Agricultural Marketing Act of 1946 (7 U.S.C. 1639o) (as amended by such section 781).

That is the whole provision. Read it twice, because almost everything written about it says something the sentence does not.

First, what it does not say

We pulled the enrolled text from the Government Publishing Office and searched it. Across the entire act:

• the word hemp appears zero times
• the word cannabinoid appears zero times
• the date November 12 appears zero times

So every account describing this as "the hemp ban delay, moving November 12 to December 11" is paraphrasing. That is not a criticism by itself — paraphrase is how anyone reads a statute in under an hour. It matters here because of what the paraphrase compresses.

Notice also what the sentence is shaped like. It does not move a date. It says that until a date, an existing provision shall only apply with respect to two named paragraphs. That is a narrowing of scope for a window, not a postponement. The scheduled provision's own effective date, one year after the law was signed, is untouched.

Second, what those two paragraphs are

This is the part that decides everything, and it takes one more document. Section 781 rewrites the federal definition of hemp and lists what the term does not include. Two of those exclusions appear in nearly identical wording twice — once for intermediate products, at (1)(C)(ii), and once for finished ones, at (1)(C)(iv). Quoting the statute:

> (I) cannabinoids that are not capable of being naturally produced by a Cannabis sativa L. plant;

>

> (II) cannabinoids that — (aa) are capable of being naturally produced by a Cannabis sativa L. plant; and (bb) were synthesized or manufactured outside the plant

Two subclauses. Different tests. Subclause (I) asks whether the plant could make this compound at all. Subclause (II) assumes it could, and asks where this particular batch came from.

Section 2019 names subclause (I). It does not name subclause (II).

It also does not name (iv)(III), which is where the 0.4 milligram per container cap lives.

Third, what the summaries say instead

We read what is circulating. A legal trade publication, dated two days ago:

> Synthetic and converted cannabinoids are excluded from every delay bill discussed above. None of the current delay proposals reach delta-8, delta-10, HHC, THC-O, or THCP made by chemical conversion.

A compliance blog quotes subclause (I)'s language correctly — "not capable of being naturally produced by a Cannabis sativa L. plant" — and then attaches conversion-derived compounds to it in the next breath. Two independently phrased search-answer summaries returned the same construction: converted compounds sit outside the extension.

Conversion — isomerisation, hydrogenation — is the ordinary description of synthesized or manufactured outside the plant. That is the wording of subclause (II), which is the subclause section 2019 does not name.

Not one source we read cites the paragraph numbers. They cite the category, and the category they describe is twice the size of the one in the text.

The thing that makes this different from a bad-sources story

We have written four of these now. Find the author. [Ask it twice.](https://gethempdash.com/learn/ask-it-twice) [As of when.](https://gethempdash.com/learn/both-numbers-are-real) [Under which rule.](https://gethempdash.com/learn/under-which-rule)

Every one of those tests works by finding a disagreement — a claim with no author, a number that moves when you re-ask, a date that has quietly expired, four answers citing three different rules. You catch the problem because something does not line up.

This morning nothing failed to line up. Every source agreed with every other source. They were authored, current, stable on re-asking, and internally consistent. One of them was a legal publication. The claim passed all four tests and it still does not match the sentence.

That is the fifth test, and it is the only one that cannot be run by comparing sources to each other: which clause did it name? You answer it by reading the operative text, or you do not answer it at all. Unanimity is not verification. It is often just one paraphrase that propagated before anyone opened the statute.

What we are not telling you

This is the part that matters most, and we would rather say it too loudly than not loudly enough.

We are not telling you that any product is unaffected until December 11. Not delta-8, not anything else. The finding here is about what a sentence names. It is not a finding about what happens to a compound, and reading it that way would be a worse error than the one we started with.

The reason is a rule we hold ourselves to: an absence is not a finding. That section 2019 omits subclause (II) is not evidence that subclause (II) does not operate on its scheduled date. It is evidence about the drafting of one sentence. We refuse to make that inverse inference for the same reason we refuse the original claim — because neither of them is a citation.

We are also not telling you the coverage is wrong as a matter of law. It does not track the subclause the text names. Whether a given cannabinoid is "capable of being naturally produced by a Cannabis sativa L. plant" is a chemical and legal question, it is genuinely contested, and it is not one a compliance vendor gets to settle in a blog post. If your inventory turns on it, that is a conversation with counsel, and it is worth having this month rather than in November.

And we are not telling you how any of this is being enforced, federally or in Texas. Our record holds enforcement posture open and asserts it in neither direction.

One more, for completeness: the act has passed both chambers and the government's own record still shows it as an enrolled bill with no public law number assigned. Press accounts report it signed. We will publish the number when a government surface carries it.

Our own page has the same gap

We read gethempdash.com/november-12 this morning, after doing the work above. It presents the date as a single threshold with three changes and no scope qualifier — the same shape as the coverage this article is about.

We have not edited it. The reading that would justify an edit is the reading we just told you is open, and quietly rewriting a live rule page on the strength of an unsettled interpretation is the failure mode this whole series is about. It is logged for review with the record it came from, and this paragraph is here so you know before you click.

What you can actually act on this morning

Nothing above changes anything on a Texas shelf today, and the honest version of this piece has to end by saying so.

What has not moved is the Texas rule already in force. The consumable hemp threshold does not bind delta-9 THC on its own. It binds total delta-9 THC, at 0.3% or less on a dry weight basis, inclusive, with an allowance where the accredited laboratory's measurement of uncertainty produces a range including that figure:

> total delta-9 THC = delta-9 THC + (0.877 × THCA)

And 25 TAC §300.301(d)(12) asks for three values on the certificate, not one — delta-9 THC, total delta-9 THC, and total THC per container. A certificate missing any of the three is non-passing on the field set alone, before anyone reaches the arithmetic. That is worked through in [Total Delta-9 THC Is a Computed Value](https://gethempdash.com/learn/total-delta-9-thc-how-it-is-computed) and the wider field set in [The Twelve Fields a Texas COA Must Carry](https://gethempdash.com/learn/texas-coa-requirements-twelve-fields).

So the one useful move this morning is the boring one: pull a certificate and check whether it reports all three values. It earns its place precisely because it does not depend on how the federal scope question resolves. Whichever way that lands, a certificate that cannot show the value the threshold binds is one you cannot reason from — and fixing it is a call to your laboratory you can make today.

The second move costs nothing. When the next summary tells you what a rule does, ask which clause it names. If it can give you a paragraph number, you can go read it. If it can only give you a category, you have learned something useful about how far to trust the category.

Source: H.R. 6500, Continuing Appropriations and Extensions Act, 2027, §2019 (enrolled text, read in full from the Government Publishing Office, 2026-09-06); P.L. 119-37 §781, amending 7 U.S.C. 1639o, read in full from the same source. Our federal records are carried as pending, and none was promoted on the strength of this reading. Texas figures: 25 TAC §300.301(a)(4), (b)(4), §300.301(d)(12); §300.302(b)(1), (c)(3), filed with the Secretary of State 2026-03-02 (TRD-202601066), effective 2026-03-31, verified 2026-08-16 against two independent official publications. Enforcement posture: under review, asserted in neither direction.

The full record set, each status shown rather than summarised, is at [the Texas Rule Center](https://gethempdash.com/texas-rules).