A fact — from the Latin factum meaning something made or done — is an event or thing in the real world. Facts should be distinguished from data, theories, and truths, as well as from opinion and things of such ilk. Just as a painting of a pipe is not a pipe, even accurate statements about facts are not the same as the facts themselves.

A fact, again, is an element of the world; something made or done; something that exists or happens. A truth, or true statement, is a linguistic expression or sentence which accurately describes a fact. "I ate salad with lunch" is not a fact — it is a true statement. My eating of the salad, at the time and place it occurred, is the fact that true statement describes.

A truth must describe a fact accurately — how accurately? That, of course, depends on the context in which one utters the purported truth. If a militant vegan asks me, "Did you eat any meat with lunch?" and I respond "I ate salad with lunch," I am describing a fact. However, I still may not be telling the truth, if I ate a chicken salad: in context, I am failing to describe fact with adequate accuracy.

We very frequently cannot judge immediately whether or not a statement is true. For instance, it is not immediately evident whether the statement "Humankind and the gorilla species evolved from a common ancestor" is true, in the sense that it is evident (to me) that "I ate salad with lunch" is true. In order to judge whether non-obvious statements describe facts, we can gather data, formulate hypotheses, and attempt to arrive at theories.

A datum is not a fact. A datum is the result of measuring some attribute of a fact. A sample of carbon extracted from a fossil is a fact; a measurement of the percentage of 14C in that sample is a datum. A datum, furthermore, is not necessarily an accurate datum — if the mass spectrometer is not working correctly, a Carbon-14 reading taken with it will be a completely spurious datum.

A hypothesis, then, is a statement which describes a set of data, and a theory is a hypothesis which describes its set of data adequately accurately. Again, this word "adequate" — adequate for what? The adequacy of a hypothesis is demonstrated by testing it; using it to predict data, and then going out and gathering the data which might confirm or refute the hypothesis.

A demonstrated theory is not the same as a truth. A theory is a statement which describes data, has proven accurate in predicting new data. Within the bounds of the data gathered, and within the bounds of the accuracy of those data, we may reasonably take the theory as true. (To do so is not to exert belief or faith in the claim that it is true, but to plan and act as if it is.) We cannot take a theory as true beyond the bounds of data; to do so would be speculating beyond the data. For instance, we may take Newton's Laws of Motion (which are theories, not "laws") as true for largish objects moving at reasonably slow speeds — but not for quantum-scale objects or relativistic speeds. Newton had no way of measuring data at those extremes, and indeed his theories have been found wildly inaccurate there.

Many have suggested that we can never know for certain if a statement is true, even an apparently evident one such as "I ate a salad with lunch." Many more have said that we can never safely take a theory, such as the theories which describe evolution, as true, since we can (they say) never have adequate data. Various derivatives of Descartes suggest that our everyday senses, or even our brains, could be under the control of forces which deny us access to fact. Hyper-relativism and some forms of postmodernism go so far as to deny that facts even exist; that there is a real world for our statements to describe.

There may be value in what these critics of fact, theory, and truth have to say. However, ironically enough, there is no way their claims can possibly be true — if there are no facts and there is no truth, then certainly the statement "There are no facts and no truth!" itself cannot be true, cannot describe a fact.