diff --git a/oxlib/normalize.py b/oxlib/normalize.py
index 0aeb30d..7be2ca5 100644
--- a/oxlib/normalize.py
+++ b/oxlib/normalize.py
@@ -1,6 +1,6 @@
-# -*- coding: utf-8 -*-
+# -*- coding: utf-8 -*-
 # vi:si:et:sw=4:sts=4:ts=4
-# GPL 2008
+# GPL 2008
 import re
 
 _articles = ('the', 'la', 'a', 'die', 'der', 'le', 'el',
@@ -106,6 +106,8 @@ def canonicalName(name):
     #      - single surname, composed name: 186604
     #        (2: 178315, 3: 6573, 4: 1219, 5: 352)
     # Don't convert names already in the canonical format.
+    if name in ('Unknown Director', ):
+        return name
     if name.find(', ') != -1: return name
     sname = name.split(' ')
     snl = len(sname)