An unobservable (also called impalpable) is an entity whose existence, nature, properties, qualities or relations are not directly observable by humans. In philosophy of science typical examples of "unobservables" are atomic particles, the force of gravity, causation and beliefs or desires. However, some philosophers (ex. George Berkeley) also characterize all objects — trees, tables, other minds, microbiological things and so on to which humans ascribe as the thing causing their perception—as unobservable.
"Unobservables" is a reference similar to Immanuel Kant's distinction between noumena (things-in-themselves, i.e., raw things in their necessarily unknowable state, before they pass through the formalizing apparatus of the senses and the mind in order to become perceived objects) and phenomena (the perceived object). According to Kant humans can never know noumena; all that humans know is the phenomena. Kant's distinction is similar to John Locke's distinction between primary and secondary qualities. Secondary qualities are what humans perceive such as redness, chirping, heat, mustiness or sweetness. Primary qualities would be the actual qualities of the things themselves which give rise to the secondary qualities which humans perceive.
The ontological nature and epistemological issues concerning unobservables is a central topic in philosophy of science. The notion that a given unobservable exists is referred to as scientific realism, in contrast to instrumentalism, the notion that unobservables such as atoms are useful models but don't necessarily exist.
Metcalf[self-published source?] distinguishes three kinds of unobservables. One is the logically unobservable, which involves a contradiction. An example would be a length which is both longer and shorter than a given length. The second is the practically unobservable, that which we can conceive of as observable by the known sense-faculties of man but we are prevented from observing by practical difficulties. The third kind is the physically unobservable, that which can never be observed by any existing sense-faculties of man.
- Stephen Palmquist, "The Radical Unknowability of Kant's 'Thing in Itself'", Cogito 3:2 (March 1985), pp.101-115; reprinted as Appendix V of Kant's System of Perspectives (University Press of America, 1993).
- W. V. Metcalf, 1940, Philosophy of Science, Vol. 7, No. 3 , pp. 337-341
See also 
- Logical positivism
- Hidden variable theory
- Object of the mind
- If a tree falls in a forest