diff --git a/spec/API_specification/dataframe_api/column_object.py b/spec/API_specification/dataframe_api/column_object.py
index c11da438..c9253fec 100644
--- a/spec/API_specification/dataframe_api/column_object.py
+++ b/spec/API_specification/dataframe_api/column_object.py
@@ -17,6 +17,95 @@ class Column:
     constructor functions or an already-created dataframe object retrieved via
 
     """
+    def to_array_obj(self, *, null_handling: str | None = None) -> object:
+        """
+        Obtain an object that can be used as input to ``asarray`` or ``from_dlpack``
+
+        The returned object must only be used for one thing: calling the ``asarray``
+        or ``from_dlpack`` functions of a library implementing the array API
+        standard, or equivalent ``asarray``/``from_dlpack`` functions to the
+        one in that standard. In practice that means that the returned object
+        may either already *be* an array type of a specific array library, or
+        it may implement one or more of the following methods or behaviors:
+
+        - ``__dlpack__``
+        - the Python buffer protocol
+        - ``__array__``
+        - ``__array_interface__``
+        - ``__cuda_array_interface__``
+        - the Python ``Sequence`` interface
+        - any other method that is known to work with an ``asarray`` function
+          in a library of interest
+
+        Importantly, the returned object must only implement/expose those
+        methods that work. An ``asarray`` function may try to use any of the
+        above methods, and the order in which it does so is not guaranteed.
+        Hence it is expected that if a method is present, it works correctly
+        and does not raise an exception (e.g., because of an unsupported dtype
+        or device).
+
+        .. admonition:: Tip
+
+           One way to expose methods only if they will work for the dtype,
+           device, and other characteristics of the column is to hide access to
+           the methods dynamically, so ``hasattr`` does the right thing. For
+           example::
+
+              def __dir__(self):
+                  methods = dir(self.__class__)
+                  attrs = list(self.__dict__.keys())
+                  keys = methods + attrs
+                  if not self.dtype in _dlpack_supported_dtypes:
+                      keys.remove("__dlpack__")
+
+                  return keys
+
+              def __dlpack__(self):
+                  ...
+
+        Parameters
+        ----------
+        null_handling : str or None
+            Determine how to treat ``null`` values that may be present in the
+            column. Valid options are:
+
+            - ``None`` (default): no special handling. This assumes that either
+              no missing values are present, or there is an array type with
+              native support for missing values that is/can be converted to.
+              *Note: there is currently no such library that is in wide use;
+              NumPy's masked arrays are non-recommended, and other array
+              libraries do not support missing values at all.*
+            - ``raise``: always raise a ``ValueError`` if nulls are present.
+            - ``to-nan``: for floating-point dtypes, convert any nulls to ``nan``.
+              For other dtypes, do the same as ``None``.
+
+            Note that if it is desired to convert nulls to a dtype-specific
+            sentinel value, the user should do this before calling
+            ``is_array_obj`` with `.isnull()` and replacing the values
+            directly.
+
+        Returns
+        -------
+        array_obj : object
+            Either a custom object or an instance of the  array type of a
+            specific array library. In the latter case, the array must have the
+            closest matching dtype of the array library (e.g., a column with a
+            floating-point dtype must produce the corresponding floating-point
+            dtype of the same precision that the array library offers).
+
+        Raises
+        ------
+        TypeError
+            In case it is not possible to convert the column to any (known) array
+            library type, or use any of the possible interchange methods.
+            This can be due to the dtype (e.g., no array library supports datetime
+            dtypes with a time zone), device, or other reasons.
+        ValueError
+            If the column contains ``null`` values which prevent returning an
+            array object.
+
+        """
+
     @classmethod
     def from_sequence(cls, sequence: Sequence[object], dtype: dtype) -> Column:
         """