I've been reading, more or less for fun, Stuart Shieber's recent edited book The Turing Test. I was struck by two language-related pseudo-arguments, both made in chapter 7, "Can automatic calculating machines be said to think?", pp. 117-132. This chapter is a transcript of a 1952 radio interview among M. H. A. Newman, Alan M. Turing, Sir Geoffrey Jefferson, and R. B. Braithwaite.
The first is a weak etymological argument made more or less in passing by Sir Geoffrey, on the first page of the transcript (p. 117 of the book, emphasis added):
I don't think that we need waste too much time on a definition of thinking since it will be hard to get beyond phrases in common usage, such as having ideas in the mind, cogitating, meditating, deliberating, solving problems or imagining. Philologists say that the word "Man" is derived from a Sanskrit word that means "to think", probably in the sense of judging between one idea and another. I agree that we could no longer use the word "thinking" in a sense that restricted it to man. No one would deny that many animals think, though in a very limited way. They lack insight. [...]
The OED sort of agrees with the spirit of this, but not the letter:
This word and Sanskrit manu have been together referred by some to the Indo-European base of MIND n.1, on the basis that thought is a distinctive characteristic of human beings.
And furthermore:
A more recent theory suggests a derivation (with loss of an initial obstruent) from the Indo-European base of Lithuanian žmonės people and Old Prussian smunents man, which is a variant (with a different ablaut grade) of the Indo-European base of classical Latin homō man, Old English guma and its Germanic cognates (see GOME n.1), and Old Lithuanian žmuo; but these Indo-European words are usually referred to the Indo-European base of classical Latin humus (see HUMUS n.) and ancient Greek χθώυ (see CHTHONIC n.) meaning 'earth'.
The second is a potentially stronger claim about the connection between words and ideas, made by Newman (pp. 122-123, emphasis again added):
[Consider the] different kinds of number. There are the integers. 0, 1, -2, and so on; there are the real numbers used in comparing lengths, for example the circumference of a circle and its diameter; and the complex numbers involving [the square root of -1]; and so on. It is not at all obvious that these are instances of one thing, "number". The Greek mathematicians used entirely different words for the integers and the real numbers, and had no single idea to cover both. It is really only recently that the general notion of kinds of number has been abstracted from these instances and accurately defined. [...]
What I'm unclear about in this case is the link between the conjoined verb phrases in the bolded part of the quotation: is it and furthermore, or and therefore? If the latter, I'm skeptical about the conclusion that the Greeks lacked the general idea simply because they lacked the general word -- though either way, I'm probably just exposing my ignorance of Ancient Greek mathematics and mathematicians.
[ Comments? ]
Posted by Eric Bakovic at September 28, 2005 05:13 PM