In metaphysics, phenomenalism is the view that physical objects cannot justifiably be said to exist in themselves, but only as perceptual phenomena or sensory stimuli (e.g. redness, hardness, softness, sweetness, etc.) situated in time and in space. In particular, some forms of phenomenalism reduce all talk about physical objects in the external world to talk about bundles of sense data.
Phenomenalism is a radical form of empiricism. Its roots as an ontological view of the nature of existence can be traced back to George Berkeley and his subjective idealism, upon which David Hume further elaborated. John Stuart Mill had a theory of perception which is commonly referred to as classical phenomenalism. This differs from Berkeley's idealism in its account of how objects continue to exist when no one is perceiving them. Berkeley claimed that an omniscient God perceived all objects and that this was what kept them in existence, whereas Mill claimed that permanent possibilities of experience were sufficient for an object's existence. These permanent possibilities could be analysed into counterfactual conditionals, such as "if I were to have y-type sensations, then I would also have x-type sensations".
As an epistemological theory about the possibility of knowledge of objects in the external world, however, the most accessible formulation of phenomenalism is perhaps to be found in the transcendental idealism of Immanuel Kant. According to Kant, space and time, which are the a priori forms and preconditions of all sensory experience, "refer to objects only to the extent that these are considered as phenomena, but do not represent the things in themselves". While Kant insisted that knowledge is limited to phenomena, he never denied or excluded the existence of objects which were not knowable by way of experience, the things-in-themselves or noumena, though he never proved them.
Kant's "epistemological phenomenalism", as it has been called, is therefore quite distinct from Berkeley's earlier ontological version.