diff --git a/i18n/README.md b/i18n/README.md
index 3a8dfe6..9707ef4 100644
--- a/i18n/README.md
+++ b/i18n/README.md
@@ -35,7 +35,7 @@ I went to town on this problem because I wasn't sure what the best approach was:
 
 * "naive": Converts the pattern to a regular expression and sequentially scans
   a word list. The implementation is pretty straight-forward.
-* "sets": Coverts the word dictionary into a mapping of
+* "sets": Converts the word dictionary into a mapping of
   (*\<char\>*, *\<index of char\>*, *\<length\>*) => *\<set of words\>*
   so matching becomes a series of set intersections.
 
@@ -47,7 +47,7 @@ I went to town on this problem because I wasn't sure what the best approach was:
   least one intermediate set is small e.g. "i18n" leads to a single
   intersection operation between a length 23 set and a length 11 set.
 
-* "prefix": Coverts the word dictionary into a mapping of
+* "prefix": Converts the word dictionary into a mapping of
   *\<length\>* => *\<prefix tree\>* so matching involves maintaining
   a work queue of prefix trees to traverse in parallel.