diff --git a/tests/test_chinese.py b/tests/test_chinese.py
index 3038fab..5cce82d 100644
--- a/tests/test_chinese.py
+++ b/tests/test_chinese.py
@@ -34,7 +34,7 @@ def test_tokens():
         ]
     )
 
-    # Jieba's original tokenizer knows a lot of names, it sees.
+    # Jieba's original tokenizer knows a lot of names, it seems.
     eq_(
         tokenize(hobart, 'zh', external_wordlist=True),
         ['加勒特', '霍巴特']