<p>By uniting the anti-mentalism of behavioral and contextual analyses with the social observations of the social sciences, a more realistic idea of “beliefs” is generated, which is amenable to observations. Beliefs are language behaviors shaped by a person’s societal, group, and interpersonal social exchanges, but such social contexts are rarely observed concurrently with belief behaviors. These ideas are demonstrated by analyzing political beliefs from the left and right, showing that such beliefs act like a common pool that can be adapted for any conversation. This is evidenced by people knowing both the left and the right beliefs (so which is said or written must arise from the consequences in their social contexts), by the changes over time in what has been labeled left or right, by the overlap in using left and right beliefs when this is useful, by both views containing less prominent beliefs which can answer typical criticisms, and by people’s use of beliefs without really “understanding” the issues. For functional conversations, what are called left-leaning views appear as being fair in social relationships but are weaker on how this will be resourced, and what are called right-leaning views appear as supporting society by creating a strong resource-based society (formerly through family values and hereditary but now through economics) but are weaker on how doing this affects people’s social relationships. Both those who might label themselves as left or as right draw on this common pool of functional verbal behaviors as suits their contexts, so functionally, anyone needing to look sympathetic and caring about social relationships can adapt left-type views to what they are saying, while those needing to look strong and decisive can adapt right-type views. The strong practical implication of the paper is that concurrent research observations of social relationship exchanges are necessary to understand any use of political beliefs, and reported beliefs alone tell us little without social context.</p>