ABSTRACT
When asked to estimate their ability to identify common objects (e.g., typical household products) by smell alone, people tend to be very confident about their ability (Cain, 1982). But those of us who have actually tried to name such odors in the absence of visual or other contextual cues have been surprised and vexed by the difficulty of the task; many odors seem very familiar but either fail to evoke a name or elicit only broad categories (e.g., fruit, edible). Although early experiments suggest that people can accurately identify only about 16 odors (Engen & Pfaffman, 1960) even after practice (Jones, 1968), more recent work (Desor & Beauchamp, 1974) indicates that the number is higher, especially for common odors and when feedback on the identification performance is provided. But, even then, performance is still less than perfect. Cain (1979) reported that correct identification of 80 odorants was less than 50% on the first trial, and improved to more than 90% over the course of five additional trials. Experts (e.g., perfumers), moreover, frequently attain impressive levels of identification performance, although this skill acquisition is measured in years rather than trials. Nevertheless, typical free identification of a set of familiar, everyday stimuli by the layperson rarely exceeds 50% correct. Roughly 10% of the incorrectly named stimuli are identified by reasonably close or related names (e.g., orange for lemon), whereas the remaining 40% of incorrectly named stimuli are called by clearly inappropriate names (e.g., turpentine for banana) or by no name at all.
