diff --git a/changelogs/fragments/isidentifier.yml b/changelogs/fragments/isidentifier.yml
new file mode 100644
index 00000000000..3b5b09241e0
--- /dev/null
+++ b/changelogs/fragments/isidentifier.yml
@@ -0,0 +1,3 @@
+---
+minor_changes:
+  - isidentifier - remove unwanted Python 2 specific code.
diff --git a/lib/ansible/utils/vars.py b/lib/ansible/utils/vars.py
index 9e0510a766b..50ddf42e7f8 100644
--- a/lib/ansible/utils/vars.py
+++ b/lib/ansible/utils/vars.py
@@ -32,8 +32,6 @@ from ansible.module_utils.common.text.converters import to_native, to_text
 from ansible.parsing.splitter import parse_kv
 
 
-ADDITIONAL_PY2_KEYWORDS = frozenset(("True", "False", "None"))
-
 _MAXSIZE = 2 ** 32
 cur_id = 0
 node_mac = ("%012x" % uuid.getnode())[:12]
@@ -237,7 +235,22 @@ def load_options_vars(version):
     return load_options_vars.options_vars
 
 
-def _isidentifier_PY3(ident):
+def isidentifier(ident):
+    """Determine if string is valid identifier.
+
+    The purpose of this function is to be used to validate any variables created in
+    a play to be valid Python identifiers and to not conflict with Python keywords
+    to prevent unexpected behavior. Since Python 2 and Python 3 differ in what
+    a valid identifier is, this function unifies the validation so playbooks are
+    portable between the two. The following changes were made:
+
+        * disallow non-ascii characters (Python 3 allows for them as opposed to Python 2)
+
+    :arg ident: A text string of identifier to check. Note: It is callers
+        responsibility to convert ident to text if it is not already.
+
+    Originally posted at https://stackoverflow.com/a/29586366
+    """
     if not isinstance(ident, string_types):
         return False
 
@@ -251,25 +264,3 @@ def _isidentifier_PY3(ident):
         return False
 
     return True
-
-
-isidentifier = _isidentifier_PY3
-
-
-isidentifier.__doc__ = """Determine if string is valid identifier.
-
-The purpose of this function is to be used to validate any variables created in
-a play to be valid Python identifiers and to not conflict with Python keywords
-to prevent unexpected behavior. Since Python 2 and Python 3 differ in what
-a valid identifier is, this function unifies the validation so playbooks are
-portable between the two. The following changes were made:
-
-    * disallow non-ascii characters (Python 3 allows for them as opposed to Python 2)
-    * True, False and None are reserved keywords (these are reserved keywords
-      on Python 3 as opposed to Python 2)
-
-:arg ident: A text string of identifier to check. Note: It is callers
-    responsibility to convert ident to text if it is not already.
-
-Originally posted at http://stackoverflow.com/a/29586366
-"""
diff --git a/test/units/utils/test_isidentifier.py b/test/units/utils/test_isidentifier.py
index e4b2a40b277..1a88f5e3c4c 100644
--- a/test/units/utils/test_isidentifier.py
+++ b/test/units/utils/test_isidentifier.py
@@ -24,21 +24,13 @@ def test_valid_identifier(identifier):
 @pytest.mark.parametrize(
     "identifier", [
         "pass", "foo ", " foo", "1234", "1234abc", "", "   ", "foo bar", "no-dashed-names-for-you",
+        "True", "False", "None"
     ]
 )
 def test_invalid_identifier(identifier):
     assert not isidentifier(identifier)
 
 
-def test_keywords_not_in_PY2():
-    """In Python 2 ("True", "False", "None") are not keywords. The isidentifier
-    method ensures that those are treated as keywords on both Python 2 and 3.
-    """
-    assert not isidentifier("True")
-    assert not isidentifier("False")
-    assert not isidentifier("None")
-
-
 def test_non_ascii():
     """In Python 3 non-ascii characters are allowed as opposed to Python 2. The
     isidentifier method ensures that those are treated as keywords on both