Believing, Desiring, or Just Thinking About: Toward a Neuroscientific Account of Propositional Attitudes

Abstract

Human minds can relate to a single idea in different ways. I can believe that Starbucks now sells donuts, but I can also hope or fear that Starbucks now sells donuts. Propositions, such as Starbucks now sells donuts are potential states of the world, or ideas thereof, that can be either true or false (McGrath, 2014). Propositional attitudes, on the other hand, are the mental states held by an agent toward a proposition (McKay & Nelson, 2014). The human mind has a remarkable capacity to hold a range of attitudes about an effectively infinite number of potential propositions. But how does our brain appropriately connect an attitude, such as believing or desiring, to a proposition? In the present research we use functional magnetic resonance imaging to identify brain regions that contain information about a limited set of object-location propositions (for example “the dog is in the upper right corner”), as well as regions associated with believing, desiring, or merely thinking about those propositions. We find increased neural activation associated with desiring a proposition in portions of the default mode network (left medial prefrontal cortex, posterior inferior parietal lobule, and middle temporal gyrus) as well as in the left dorsolateral prefrontal cortex. We find increased activation associated with believing a proposition in the right posterior parietal cortex and with thinking about a proposition that is not necessarily believed in the right inferior frontal gyrus. We find reliably decodable location information in the occipital lobe and object information in the right parahippocampal gyrus that is stable across propositional attitude. We also find that that attitude has an effect on proposition representation. When a proposition is being desired, rather than merely being thought about, we find more object decodability in the left putamen, cingulate, cuneus, precuneus, and medial frontal gryus. When a proposition is being believed, rather than merely being thought about, we find more object decodability in the right posterior superior temporal gyrus, right paracentral lobule, and bilateral cuneus. Finally, we take the first steps towards developing a model for how the brain flexibly combines propositional attitudes and propositions by using functional connectivity to identify correlations between attitude-associated neural activation and attitude-specific object representation. We find a small but significant correlation between the activation in the left medial prefrontal cortex and the object representation in the left middle temporal gyrus when the object is a constituent of a desired proposition. We also find a small but significant correlation between activation in the right inferior frontal gyrus and object representation in the right posterior superior temporal gyrus when object is a constituent of a proposition that is being thought about but not necessarily believed.