How “Common” is Common Sense?
Common sense is usually defined as “what all sensible people know,” but this definition is circular: how do we know someone is sensible other than that they possess common sense? As a result, most people believe that they themselves possess common sense, but have trouble articulating which of their beliefs are commonsense or how common they are.
This project tackles the definitional conundrum of common sense head on via a massive online survey experiment. Participants are asked to rate thousands of statements, spanning a wide range of knowledge domains, both in terms of their own agreement with the statement and their belief about the agreement of others. The team developed novel methods to extract statements from several diverse sources including appearances in mass media, non-fiction books, and political campaign emails, as well as statements elicited from human respondents and generated by AI systems. We have also developed new taxonomies to classify statements by domain and type.
Ultimately, we hope to provide insight into the nature and limits of common sense, thereby aiding research communities (e.g. AI and ML) who wish to explore and simulate this ubiquitous yet frustratingly elusive concept.
ABOVE: Belief calibration by topic
Calibration of one’s own beliefs about what others think is true—one characterization of common sense—is heavily dependent on the topic of a statement. For example, for statements related to religion (represented by the large light green circle to the lower left of the plot), people systematically overestimate others’ beliefs. Statements related to technical topics (represented by the yellow circle in the upper right of the plot) are almost perfectly calibrated on average.
Stevens University Professor & twenty-third Penn Integrates Knowledge Professor
Undergraduate Student Researcher