Back in 1987 Lindley Darden published “Viewing the History of Science as Compiled Abstraction,” AI Magazine, Volume 8, Number 2, Summer 1987. David Hays and I replied with a letter, William Benzon and David Hays, Reply to Darden, AI Magazine, Volume 8, Number 4, 1987, pp. 7-8.
Summary: Abstract patterns can be recognized in stories, which are likened to one another through Wittgenstein’s ‘family resemblance.’ But it is possible to use language to rationalize such abstract patterns. Thus “charity” can be rationalized as happening when “someone does something nice for someone else without thought of reward.” The notion of reward must itself be abstract, as the variety of physical rewards defies description and abstract rewards (such as fame) are possible. Thus the linguistic means of dealing with abstraction must be recursive. “Metalingual definition,” in recognition of Roman Jakobson, seems an apt term. Note further that science frequently concocts abstract accounts of physical things. Diamonds and lampblack are physical things, and quite different from one another. Yet both consist primarily of carbon, which is abstractly defined through the concepts and experimental technique of modern chemistry. Knowledge must be grounded in phenomenologically naïve commonsense knowledge; but the recursive nature of metalingual definition allows thought to move far away from the sensory base.