Reasoning refers to the process of forming conclusions, judgments or inferences from facts or premises (Random House 1991). As such, it is central to human understanding and decision making.
As with so many other terms, reason has a variety of meanings. In the philosophical sense, any justification is a reason, even if it is a wild, off-the-wall, unbelievable premise. The crucial questions are the validity of the argument (logical concerns) and the truth of the premises (epistemological and methodological concerns).

Aside from these, it is worth considering some other points regarding reason (including the note on Janian logic below).
Aristotle noted two categories: theoretical and practical reasoning. The former is about the way things are, the latter, about the way things ought to be. Theoretical reasoning - which centers on description, causation, explanation, and prediction - is the category applied within science. Practical reasoning is prescriptive and leads to action.
Regarding either category, to form a complete line of reasoning, there are three requirements (Figure 1):

1. There can be no infinite regression - all statements or premises must be supported by another reason - 'just because' is not sufficient.
2. There can be no circular reasoning - the reasons cannot rely on each other.
3. There can be no appeal to non-cognitive justifications - all the reasons must make sense within the context of their application. 'Because green is a nice colour' may be a valid reason for explaining why someone painted their house, but not for explaining the colour of leaves.

Difficulty arises because it is never possible to meet all three requirements simultaneously, so perfect reasoning is not possible. This leads to three types of reasoning based on which requirement is given up.

1. Foundationalism relinquishes the first requirement, asserting a particular starting point and building an argument from there.
2. Coherentism relies on making a big enough circle, assuming that the reasoning can be trusted if the argument is consistent; if all the reasons fit and are coherent.
3. Pragmatism relinquishes the third requirement, accepting a line of reasoning because it works, or because it makes intuitive sense.

As all other modes of inquiry, science relies on reason, and is generally taken to rely on foundationalism in particular. I believe, however, that it actually relies on a combination of all three. Foundational reasoning, and only when the assertions are based on observation, is the only reasoning clearly articulated, although the others are used.

(Based on notes taken during Philosophy 417 (1992),
taught by Paul Wood at the University of British Columbia.)

But just to prevent anyone from being too rigidly locked into any particular perspective, I include another possibility...