diff --git a/CHANGELOG.md b/CHANGELOG.md
index bcaf30802..ec28b4764 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4213,6 +4213,25 @@ documented at release-notes length in the first place, and are still in
 
 ### Documentation and the website
 
+- **`psbt.silent_payments` said its `k` follows an order it does not
+  follow.** The module docstring described the BIP375 prose -- codes
+  sorted lexicographically by spend key -- and named `_ordered_sp_outputs`
+  as that sort, while the function is output index order and its own
+  docstring says so and says why: the BIP's vectors and its reference
+  validator are index order, and only its prose dissents. A reader had
+  the two halves of the file contradicting each other, and the API docs
+  carried the wrong one. The module docstring now describes the code and
+  points at the function for the discrepancy.
+
+  `_ordered_sp_outputs` gains the correction the discrepancy itself
+  needed: the two invalid vectors named after ordering are refused under
+  either rule not because their scripts match no `k` assignment, but
+  because each carries one no ascending rule produces -- two values
+  swapped in one, three permuted in the other. It also names bips PR
+  2207, which proposes settling the question the other way by correcting
+  the vectors and the validator to the prose; if that merges, the test
+  that pins index order is what fails here.
+
 - **`dsa.Sig.parse`'s `strict` says which direction each accident goes**
   (issue #873). The flag stays read for its truth, which is what
   `bool_parameter_test.py` classifies it as and what the line
diff --git a/btclib/psbt/silent_payments.py b/btclib/psbt/silent_payments.py
index ccb27221a..14178b9d8 100644
--- a/btclib/psbt/silent_payments.py
+++ b/btclib/psbt/silent_payments.py
@@ -53,11 +53,12 @@
 keys* of the eligible inputs -- which is why reading those keys is the
 first thing here and not an aside.
 
-**The k of an output** is BIP375's own rule and not BIP352's ordering: the
-codes sharing one scan key are sorted lexicographically by spend key, and
-a subgroup sharing both keys by output index. `_ordered_sp_outputs` is
-that sort, and `tests/psbt/silent_payments_test.py` says which of the
-vectors pins it.
+**The k of an output** is BIP375's own rule and not BIP352's ordering: it
+counts per scan key, over the outputs in index order. That is what the
+BIP's vectors and its reference validator do and *not* what its prose
+says, which asks for the codes to be sorted lexicographically;
+`_ordered_sp_outputs` is where the discrepancy is documented, and
+`tests/psbt/silent_payments_test.py` says which of the vectors pins it.
 """
 
 from __future__ import annotations
@@ -272,10 +273,12 @@ def _ordered_sp_outputs(psbt: Psbt) -> list[tuple[int, PsbtOut]]:
     scan key while walking the outputs in index order -- and only the
     prose dissents. Index order is therefore what interoperates, and the
     two "output scripts" invalid vectors named after ordering are refused
-    under it anyway: their scripts match no assignment at all.
-    `tests/psbt/silent_payments_test.py` pins each of those facts, so a
-    later revision of the BIP that settles it the other way fails here
-    rather than passing quietly.
+    under it anyway: each carries a k assignment no ascending rule
+    produces, the two values swapped in one and three permuted in the
+    other. `tests/psbt/silent_payments_test.py` pins each of those facts,
+    so a revision of the BIP that settles it the other way fails here
+    rather than passing quietly -- which is what bips PR 2207 proposes,
+    correcting the vectors and the validator to match the prose.
     """
     return [(i, o) for i, o in enumerate(psbt.outputs) if o.sp_v0_info]