diff --git a/decisions/0004-one-table-two-places.md b/decisions/0004-one-table-two-places.md
index 40ec30f..a08e757 100644
--- a/decisions/0004-one-table-two-places.md
+++ b/decisions/0004-one-table-two-places.md
@@ -61,7 +61,7 @@ remove, and it is the same failure record 0007 refuses for the design tokens.
 
 Have this build write the profile table directly. The cost is a credential that
 reaches another repository, held by a job here, plus commits landing in that
-tree from a source its own maintainer is not reading.
+tree from a source I am not reading when I edit it by hand.
 
 ## When this is worth revisiting
 
diff --git a/decisions/0006-what-answers-at-the-catalogue-address.md b/decisions/0006-what-answers-at-the-catalogue-address.md
index 0afe747..7661fb3 100644
--- a/decisions/0006-what-answers-at-the-catalogue-address.md
+++ b/decisions/0006-what-answers-at-the-catalogue-address.md
@@ -75,9 +75,9 @@ invisible.
 
 ## What this record does not decide
 
-Whether the domain moves at all is the maintainer's call, as entry 3 of issue #7.
-Nothing here decides it, and what this record states has to hold on whichever
-answer comes back.
+Whether the domain moves at all is my call, as entry 3 of issue #7. Nothing
+here decides it, and what this record states has to hold on whichever answer
+comes back.
 
 ## When this is worth revisiting
 
diff --git a/decisions/0008-the-url-shape.md b/decisions/0008-the-url-shape.md
index b9432d6..7367481 100644
--- a/decisions/0008-the-url-shape.md
+++ b/decisions/0008-the-url-shape.md
@@ -96,9 +96,9 @@ which is the kind this project cannot see and cannot repair.
 
 ## What this record does not decide
 
-Where the addresses are served from is entry 3 of issue #7 and is the maintainer's
-call. The catalogue address is held by record 0006 and is fixed by a commitment
-rather than by this scheme.
+Where the addresses are served from is entry 3 of issue #7 and is my call. The
+catalogue address is held by record 0006 and is fixed by a commitment rather
+than by this scheme.
 
 The scheme above is written for a site published in one language. Entry 2 of issue
 #7 has an option that publishes two, and on that answer every address here gains a