Classical Categories vs

Classical View of Categories and Family Resemblance

• Classical view of categories is based on the premise that categories are defined by the presence or absence of specific properties.
• Categories are defined by necessary and sufficient features, with no in-between cases and all members being equal.
• The main claims of the classical view are that concepts are mentally represented as definitions, providing characteristics necessary and sufficient for category membership.
• Membership of categories such as squares and even numbers is defined by specific properties, which are necessary and sufficient for classification.
• Category membership isn't binary, as demonstrated by participants rating membership on a continuum from 0 to 100% and changing their inclusion criteria over time.
• Some category members are considered more typical or representative than others, as shown by high inter-rater reliability in typicality ratings.
• Category membership decisions and categorization decision times are influenced by typicality, with typical members being confirmed or generated more frequently.
• Even classically definable categories show typicality effects, with certain members being more typical than others.
• An alternative approach to the classical view is the concept of family resemblance, where categories are defined by overlapping distributions of features.
• Family resemblance is a weighted sum of featural overlap among category members, with a strong correlation between typicality and family resemblance.
• Category boundaries are fuzzy and category membership doesn't imply equality, with some members being more typical or 'better' category members than others.
• The features generated by participants tend to apply to a small subset of category members, and none of the features are necessary and sufficient.