BUG: Enforce parse_dates as bool when scalar

doc/source/whatsnew/v0.18.1.txt

@@ -134,6 +134,7 @@ API changes
 
 - ``Series.apply`` for category dtype now applies passed function to each ``.categories`` (not ``.codes``), and returns "category" dtype if possible (:issue:`12473`)
 
+- ``read_csv`` will now raise a ``TypeError`` if ``parse_dates`` is neither a boolean, list, or dictionary (:issue:`5636`)
 
 - The default for ``.query()/.eval()`` is now ``engine=None``, which will use ``numexpr`` if it's installed; otherwise it will fallback to the ``python`` engine. This mimics the pre-0.18.1 behavior if ``numexpr`` is installed (and which Previously, if numexpr was not installed, ``.query()/.eval()`` would raise). (:issue:`12749`)
 

pandas/io/parsers.py

@@ -825,6 +825,27 @@ def _validate_usecols_arg(usecols):
     return usecols
 
 
+def _validate_parse_dates_arg(parse_dates):
+    """
+    Check whether or not the 'parse_dates' parameter
+    is a non-boolean scalar. Raises a ValueError if
+    that is the case.
+    """
+    msg = ("Only booleans, lists, and "
+           "dictionaries are accepted "
+           "for the 'parse_dates' parameter")
+
+    if parse_dates is not None:
+        if lib.isscalar(parse_dates):
+            if not lib.is_bool(parse_dates):
+                raise TypeError(msg)
+
+        elif not isinstance(parse_dates, (list, dict)):
+            raise TypeError(msg)
+
+    return parse_dates
+
+
 class ParserBase(object):
 
     def __init__(self, kwds):
@@ -836,7 +857,8 @@ def __init__(self, kwds):
         self.index_names = None
         self.col_names = None
 
-        self.parse_dates = kwds.pop('parse_dates', False)
+        self.parse_dates = _validate_parse_dates_arg(
+            kwds.pop('parse_dates', False))
         self.date_parser = kwds.pop('date_parser', None)
         self.dayfirst = kwds.pop('dayfirst', False)
         self.keep_date_col = kwds.pop('keep_date_col', False)