diff --git a/spacy/en/morph_rules.py b/spacy/en/morph_rules.py
index 129ae799f..2b8aae823 100644
--- a/spacy/en/morph_rules.py
+++ b/spacy/en/morph_rules.py
@@ -68,3 +68,8 @@ MORPH_RULES = {
         "were":         {LEMMA: "be", "VerbForm": "Fin", "Tense": "Past", "Number": "Plur"}
     }
 }
+
+
+for tag, rules in MORPH_RULES.items():
+    for key, attrs in dict(rules).items():
+        rules[key.title()] = attrs
diff --git a/spacy/tests/regression/test_issue686.py b/spacy/tests/regression/test_issue686.py
index 2ca882514..d3807808a 100644
--- a/spacy/tests/regression/test_issue686.py
+++ b/spacy/tests/regression/test_issue686.py
@@ -4,9 +4,8 @@ from __future__ import unicode_literals
 import pytest
 
 
-@pytest.mark.xfail
 @pytest.mark.models
-@pytest.mark.parametrize('text', ["He is the man.", "They are the men."])
+@pytest.mark.parametrize('text', ["He is the man", "he is the man"])
 def test_issue686(EN, text):
     """Test that pronoun lemmas are assigned correctly."""
     tokens = EN(text)