diff --git a/doc/source/release.rst b/doc/source/release.rst
index ed22348e45c9f..fa541baa4e058 100644
--- a/doc/source/release.rst
+++ b/doc/source/release.rst
@@ -156,8 +156,8 @@ API Changes
 - ``to_excel`` now converts ``np.inf`` into a string representation,
   customizable by the ``inf_rep`` keyword argument (Excel has no native inf
   representation) (:issue:`6782`)
-- Arithmetic ops are now disallowed when passed two bool dtype Series or
-  DataFrames (:issue:`6762`).
+- Arithmetic ops on bool dtype arrays/scalars now give a warning indicating
+  that they are evaluated in Python space (:issue:`6762`, :issue:`7210`).
 - Added ``nunique`` and ``value_counts`` functions to ``Index`` for counting unique elements. (:issue:`6734`)
 
 - ``DataFrame.plot`` and ``Series.plot`` now support a ``table`` keyword for plotting ``matplotlib.Table``. The ``table`` keyword can receive the following values.
diff --git a/doc/source/v0.14.0.txt b/doc/source/v0.14.0.txt
index 56ea0a361e741..ba67e2cd4d4c8 100644
--- a/doc/source/v0.14.0.txt
+++ b/doc/source/v0.14.0.txt
@@ -186,17 +186,18 @@ API changes
 - Added ``factorize`` functions to ``Index`` and ``Series`` to get indexer and unique values (:issue:`7090`)
 - ``describe`` on a DataFrame with a mix of Timestamp and string like objects returns a different Index (:issue:`7088`).
   Previously the index was unintentionally sorted.
-- arithmetic operations with **only** ``bool`` dtypes now raise an error
-  (:issue:`7011`, :issue:`6762`, :issue:`7015`)
+- arithmetic operations with **only** ``bool`` dtypes warn for ``+``, ``-``,
+  and ``*`` operations and raise for all others (:issue:`7011`, :issue:`6762`,
+  :issue:`7015`, :issue:`7210`)
 
   .. code-block:: python
 
      x = pd.Series(np.random.rand(10) > 0.5)
      y = True
-     x * y
+     x + y  # warning generated: should do x | y instead
+     x / y  # this raises because it doesn't make sense
 
-     # this now raises for arith ops like ``+``, ``*``, etc.
-     NotImplementedError: operator '*' not implemented for bool dtypes
+     NotImplementedError: operator '/' not implemented for bool dtypes
 
 
 .. _whatsnew_0140.display:
diff --git a/pandas/computation/expressions.py b/pandas/computation/expressions.py
index 4aff00e3a97d9..47d3fce618f89 100644
--- a/pandas/computation/expressions.py
+++ b/pandas/computation/expressions.py
@@ -6,6 +6,7 @@
 
 """
 
+import warnings
 import numpy as np
 from pandas.core.common import _values_from_object
 from distutils.version import LooseVersion
@@ -170,11 +171,23 @@ def _has_bool_dtype(x):
             return isinstance(x, (bool, np.bool_))
 
 
-def _bool_arith_check(op_str, a, b, not_allowed=frozenset(('+', '*', '-', '/',
-                                                           '//', '**'))):
-    if op_str in not_allowed and _has_bool_dtype(a) and _has_bool_dtype(b):
-        raise NotImplementedError("operator %r not implemented for bool "
-                                  "dtypes" % op_str)
+def _bool_arith_check(op_str, a, b, not_allowed=frozenset(('/', '//', '**')),
+                      unsupported=None):
+    if unsupported is None:
+        unsupported = {'+': '|', '*': '&', '-': '^'}
+
+    if _has_bool_dtype(a) and _has_bool_dtype(b):
+        if op_str in unsupported:
+            warnings.warn("evaluating in Python space because the %r operator"
+                          " is not supported by numexpr for the bool "
+                          "dtype, use %r instead" % (op_str,
+                                                     unsupported[op_str]))
+            return False
+
+        if op_str in not_allowed:
+            raise NotImplementedError("operator %r not implemented for bool "
+                                      "dtypes" % op_str)
+    return True
 
 
 def evaluate(op, op_str, a, b, raise_on_error=False, use_numexpr=True,
@@ -193,7 +206,7 @@ def evaluate(op, op_str, a, b, raise_on_error=False, use_numexpr=True,
                          return the results
         use_numexpr : whether to try to use numexpr (default True)
         """
-    _bool_arith_check(op_str, a, b)
+    use_numexpr = use_numexpr and _bool_arith_check(op_str, a, b)
     if use_numexpr:
         return _evaluate(op, op_str, a, b, raise_on_error=raise_on_error,
                          **eval_kwargs)
diff --git a/pandas/tests/test_expressions.py b/pandas/tests/test_expressions.py
index 777acdf30f1a0..8d012b871d8ca 100644
--- a/pandas/tests/test_expressions.py
+++ b/pandas/tests/test_expressions.py
@@ -343,8 +343,8 @@ def testit():
     def test_bool_ops_raise_on_arithmetic(self):
         df = DataFrame({'a': np.random.rand(10) > 0.5,
                         'b': np.random.rand(10) > 0.5})
-        names = 'add', 'mul', 'sub', 'div', 'truediv', 'floordiv', 'pow'
-        ops = '+', '*', '-', '/', '/', '//', '**'
+        names = 'div', 'truediv', 'floordiv', 'pow'
+        ops = '/', '/', '//', '**'
         msg = 'operator %r not implemented for bool dtypes'
         for op, name in zip(ops, names):
             if not compat.PY3 or name != 'div':
@@ -369,6 +369,49 @@ def test_bool_ops_raise_on_arithmetic(self):
                 with tm.assertRaisesRegexp(TypeError, err_msg):
                     f(df, True)
 
+    def test_bool_ops_warn_on_arithmetic(self):
+        n = 10
+        df = DataFrame({'a': np.random.rand(n) > 0.5,
+                        'b': np.random.rand(n) > 0.5})
+        names = 'add', 'mul', 'sub'
+        ops = '+', '*', '-'
+        subs = {'+': '|', '*': '&', '-': '^'}
+        sub_funcs = {'|': 'or_', '&': 'and_', '^': 'xor'}
+        for op, name in zip(ops, names):
+            f = getattr(operator, name)
+            fe = getattr(operator, sub_funcs[subs[op]])
+
+            with tm.use_numexpr(True, min_elements=5):
+                with tm.assert_produces_warning():
+                    r = f(df, df)
+                    e = fe(df, df)
+                    tm.assert_frame_equal(r, e)
+
+                with tm.assert_produces_warning():
+                    r = f(df.a, df.b)
+                    e = fe(df.a, df.b)
+                    tm.assert_series_equal(r, e)
+
+                with tm.assert_produces_warning():
+                    r = f(df.a, True)
+                    e = fe(df.a, True)
+                    tm.assert_series_equal(r, e)
+
+                with tm.assert_produces_warning():
+                    r = f(False, df.a)
+                    e = fe(False, df.a)
+                    tm.assert_series_equal(r, e)
+
+                with tm.assert_produces_warning():
+                    r = f(False, df)
+                    e = fe(False, df)
+                    tm.assert_frame_equal(r, e)
+
+                with tm.assert_produces_warning():
+                    r = f(df, True)
+                    e = fe(df, True)
+                    tm.assert_frame_equal(r, e)
+
 
 if __name__ == '__main__':
     import nose
diff --git a/pandas/util/testing.py b/pandas/util/testing.py
index 1235aa3cc89d9..e74cf487e75ac 100644
--- a/pandas/util/testing.py
+++ b/pandas/util/testing.py
@@ -35,6 +35,8 @@
     raise_with_traceback, httplib
 )
 
+from pandas.computation import expressions as expr
+
 from pandas import bdate_range
 from pandas.tseries.index import DatetimeIndex
 from pandas.tseries.period import PeriodIndex
@@ -1576,3 +1578,14 @@ def __enter__(self):
     def __exit__(self, exc_type, exc_value, traceback):
 
         np.random.set_state(self.start_state)
+
+
+@contextmanager
+def use_numexpr(use, min_elements=expr._MIN_ELEMENTS):
+    olduse = expr._USE_NUMEXPR
+    oldmin = expr._MIN_ELEMENTS
+    expr.set_use_numexpr(use)
+    expr._MIN_ELEMENTS = min_elements
+    yield
+    expr._MIN_ELEMENTS = oldmin
+    expr.set_use_numexpr(olduse)