Our problem is that we have been stalked by an article. For 36 years it has been following us around, appearing in citations, in introductory textbooks, in musty journals. It appears in various disguises, now the mantle of short term memory, then the guise of absolute identification, then the cover of recoding. Our commentary gives us a chance to probe the depths of this mystery.
Miller's article discussed perceptual and memorial limitations, and he excites the reader's curiosity by pointing out numerical commonalities holding across these domains. The numbers mentioned in the title of the article are of course the least significant part of the message (do not expect us to review 36 years of research and report new and better estimates, such as 7.12 plus 1.94 minus 3.44). What are the real reasons for the magic and lasting fame of this article? We try to unravel them here in an oblique and roundabout fashion, by detouring first through the locales of absolute identification and short term memory, the two topics that dominate Miller's classic article.