diff --git a/content/posts/spoiler-element.gmi b/content/posts/spoiler-element.gmi
index 4abea82..8975dc1 100644
--- a/content/posts/spoiler-element.gmi
+++ b/content/posts/spoiler-element.gmi
@@ -110,7 +110,7 @@ Some semantic configurability should be possible with HTML attributes:
## Further work
-We may need to create additional corresponding ARIA roles for spoilers. I can't really on an ideal one.
+We may need to create additional corresponding ARIA roles for spoilers. I can't decide on an ideal existing one.
We may also need to define a vocabulary for well-known spoilers, and encourage authors to use that vocabulary when relevant. This could improve how well auto-revealing spoilers works. Of course, I'm not proposing that any content matching that vocabulary should receive a spoiler; I'm only proposing to choose that vocabulary when writing spoiler-text, if it's relevant. I don't think we should be prescriptive about what to use spoilers on across the entire Web.
diff --git a/content/posts/spoiler-element.md b/content/posts/spoiler-element.md
index 5158918..1743b73 100644
--- a/content/posts/spoiler-element.md
+++ b/content/posts/spoiler-element.md
@@ -184,7 +184,7 @@ Some semantic configurability should be possible with HTML attributes:
### Further work
-We may need to create additional corresponding ARIA roles for spoilers. I can't really on an ideal one.
+We may need to create additional corresponding ARIA roles for spoilers. I can't decide on an ideal existing one.
We may also need to define a vocabulary for well-known spoilers, and encourage authors to use that vocabulary when relevant. This could improve how well auto-revealing spoilers works. Of course, I'm not proposing that any content matching that vocabulary should receive a spoiler; I'm only proposing to choose that vocabulary when writing spoiler-text, if it's relevant. I don't think we should be prescriptive about what to use spoilers on across the entire Web.