« Ideas of Imperfection | Main | Mothers of Invention »
April 28, 2005
Google Definitions
Warning: causal theory of reference geekery ahead.
At a 1962 conference in Helsinki (the same conference at which Kripke presented "Semantic Considerations on Modal Logic"), Ruth Barcan Marcus said the following:
[T]o discover that we have alternative proper names for the same object we turn to a lexicon, or, in the case of a formal language, to the meaning postulates, ...[o]ne doesn't investigate the planets, but the accompanying lexicon.
(Aside: this Barcan-Marcus quote is taken from John P. Burgess' "Quinus ab Omni Naevo Vindicatus", a paper which John usually refers to excitedly and mysteriously as "the paper with the Latin title." The title, he explains in the paper, echoes Saccheri's Euclidus ab Omni Naevo Vindicatus or Euclid freed freed every blemish, and the paper defends (successfully in my opinion, and I have an extremely fractious relationship with Quine's writings) Quine's argument that de re modality cannot be reduced to de dicto modality.)
Of course, Barcan-Marcus' claim here looks crazy. One cannot always tell that a single object has been given two names just by looking up the meanings of the names in the dictionary. The discovery that 'Hesperus' names the same object as 'Phosphorus' for example (they are both names for Venus), required substantial empirical research in astronomy, not just the consultation of a dictionary. I used to think this quote was an example of someone saying something crazy because an attractive theory seems to imply it. And I thought the train of thought probably went something like this: names are just tags and their meanings are just the objects tagged - like names in modal logic. How does one find out about the meaning of a linguistic expression? One looks it up in the dictionary (right?) So if two names have the same meaning (tag the same object) then we'll be able to tell from the dictionary entries.
And you might think that Barcan Marcus' comment contained some important and radical semantic ideas but wasn't yet very clear on one of the epistemic possibilities that could go along with those ideas (namely that two names, e.g. "Hesperus" and "Phosphorus" could have the same meaning without it being possible to tell, on the basis of one's semantic competance alone, that a sentence expressing the identity of the object(s) referred to is true, e.g., without being able to tell that "Hesperus is Phosphorus" is true. (Forgivably of course, no-one else got there until Kripke's "Naming and Necessity" lectures at Princeton 8 years later.)
However I have just discovered Google Definitions. If one feeds Google the expression "define: " followed by the term one wants defined (it's a little too late for corner quotes,) it will return definitions from all over the web. So naturally I had to feed it all the old philosophical examples, and behold:
Hesperus
evening star: a planet (usually Venus) seen at sunset in the western sky
www.cogsci.princeton.edu/cgi-bin/webwn
Phosphorus
Phosphorus means Venus when it is seen in the morning (the morning star).
en.wikipedia.org/wiki/Phosphorus_(morning_star)
There you go, Ruth Barcan-Marcus was right and all that hard astronomy was for nothing.
Except, er, not. My new electronic super-lexicon is surprisingly quiet on the identity of the referent of '2' with the referent of either '{0,{0}}' (a la van Neumann) or with '{{0}}' (a la Zermelo), and it didn't have anything to say about the identity of Plumwood with Routley. (Though it did tell me that "Londres" was a name for London, and that "Tully" is a name for Cicero.)
Posted by logican at April 28, 2005 09:54 AM
Trackback Pings
The trackback address for this entry is:
http://www.logicandlanguage.net/trakbak.cgi/31
Comments
I'm very intrigued to learn that "two" means "United States" (among other things). And that "2" only has meaning in French. But I suppose by inspection of all those definitions, we see that Benacerraf was right - the number two isn't a set of any sort at all.
Posted by: Kenny Easwaran at April 28, 2005 04:44 PM
That's 1962, not 1972.
Posted by: Pedantic Bore at April 28, 2005 04:54 PM
Hi Pedantic,
You are right, my mistake. I've fixed it now.
Posted by: Gillian Russell at April 28, 2005 04:59 PM
And it's not so good for natural kinds, either.
http://www.google.com/search?hl=en&q=define%3A+water
http://www.google.com/search?hl=en&lr=&q=define%3A+aluminum
Posted by: Franklin Scott at May 3, 2005 11:18 PM
There have ben dictionary entries of proper names ling before 1962 and thereafter. Burgess bullies us in Latin.
Posted by: ruth b marcus at June 20, 2006 02:13 PM