April 2007 Archives

...is here. 

 [Thanks go to whoever nominated my post.]

For those of you interested in truth and realism, check out Adam Kovach's (Marymount University) review of Barry Taylor's new book Models, Truth, and Realism (OUP, 2006), here. 

As a colleague of mine, Kamper Floyd, and I work on a paper for the upcoming MPA, I've been thinking a lot about Putnam-esque thought experiments.  But what inspired this post is the paper that generated the idea for the MPA paper, a paper by Sandy Goldberg (University of Kentucky).  In that paper he articulates a challenge for anyone who would combine a robust form of realism with content externalism.  One of Goldberg's upshots is that when the two views are combined, that is when content externalism (CE) and metaphysical realism (MR) are combined, it becomes in principle possible for one to have a thought with a determinate content, a thought that represents the world as being a certain way, but for it also to be possible for the thinker having that thought NOT to know that the thought has the content that it has.  It becomes possible for someone not to know what they are thinking.  In a less confused way, Goldberg (forthcoming/under review) writes, "how we represent the world's joints in thought could be more fine-grained than anything we humans could discern."  Goldberg thinks this is unacceptible.

For this post I am not focusing on the arguments leading to this conclusion, instead I want to try to motivate it using three thought experiments.  I do so only in the attempt to motivate the conclusion in the best way possible and, maybe, by so doing make any points of contention clearer.  It will help if the reader is familiar with Putnam's "The Meaning of "Meaning"" (in Gunderson, ed., Language, Mind and Knowledge, 1975; reprinted in Davidson, M., ed., On Sense and Direct Reference, McGraw Hill, 2007), though I will not cite that paper here.

Scenario 1

In a region of the universe, there is a planet identical to Earth as it is now, call it Fool's Earth , except for the following difference (and its consequences): Fool's Earth's best sciences cannot tell the difference between gold and fool's gold, that is between the element Au and the compound FeS2 (iron pyrite).  Also different are the respective histories of Earth and Fool's Earth.  The history of earth is as we know it, while the history of Fool's Earth is similar except for the following historical linguistic difference.  On Fool's Earth, there was a time when a human being first encountered Au and called it 'Gold' (let's suppose that it was an English speaker, for simplicity's sake).  Let's also suppose that this discoverer of Au had never encountered pyrite of any stripe and in any form, and also that no one else on Earth had ever encountered pyrite.  Now, slowly but surely the word 'gold' spreads as the precious metal spreads.  Then one day, lo and behold, someone discovers pyrite and calls it 'gold' (of course, to them it is not a discovery).  They think they've found more Au.  They sell it, work it, and it gets passed around all the while being mistaken for gold.  Lastly, ex hypothesi, no one can tell the difference between Au and pyrite.  No one will ever be able to tell the difference.

Scenario 2

Similar to scenario 1 except that the discoverer of Au and those that learned the word 'Gold' from her never interacted with the "discoverer" of pyrite and those that learned the word from her, where I understand 'those that learned the word from her' to be something like 'is a link in a causal chain of Putnam's sort leading back to the initial baptism of the substance' (the exact details of the causal story I don't think are important, yet).  Thus we can think of scenario 2 as including two different communities both with a word type 'Gold', except for one community the causal chain traces back to an initial baptism of Au and for the other community the chain traces back to a baptism of pyrite, and nowhere at no time do the chains intersect.  In this scenario as well no one can tell the difference between Au and pyrite.  No one will ever be able to tell the difference.

Scenario 3

Similar to scenarios 1 and 2 except that the discoverer of Au also discovered pyrite at the same time and dubbed what she thought were identical substances 'Gold'.  The purportedly identical substances are peddled around, are all called 'gold', and are all believed to be the same substance.  In this scenario no one can tell the difference between Au and pyrite.  No one will ever be able to tell the difference.

So, enough stage setting, here are my claims about the residents of Fool's Earth as they figure in the respective scenarios:

Claim about Scenario 1: The predicate 'gold' as uttered by anyone has in its extension all and only Au.  All claims of the form 'This (substance) is gold' are false when the speaker is demonstrating pyrite.

Claim about Scenario 2: The predicate 'gold' as uttered by a human link in the chain leading back to Au has in its extension all and only Au; while the predicate 'gold' as uttered by a human link in the chain leading back to pyrite has in its extension all and only pyrite.  We have one word type, but two words.  The truth or falsity of 'This (substance) is gold' depends in part on which chain the speaker is a link in.

Claim about Scenario 3: The predicate 'gold' as uttered by anyone has in its extension all and only both Au and pyrite.  All claims of the form 'This (substance) is gold' are true whether the speaker demonstrates Au or pyrite.

Here's the skeptical rub: How do speakers on Fool's Earth, in any scenario, know what they are thinking when they have gold-thoughts?  WE think we know that we are thinking that this is gold when we have a thought 'This is gold' perhaps because we have an extra-linguistic means of distinguishing pyrite and Au.  But content externalism says that such a means is irrelevant (here is where I'm moving away from Putnam and his division of linguistic labor).  But I think these examples show that it is also not required that there in principle be such a means (and so agreeing with Goldberg's conclusion).  The experts of the community don't determine what the content of anyone's thoughts is.  Why?  Well, consider again scenario 3.  What would happen if at some future time scientists discovered a significant molecular or atomic difference among some of the things typically called 'gold'.  They have a choice: either alter the meaning, by stipulation, of 'gold' so that it has in its extension only pyrite or Au and not both, introducing a new term to cover the remainder; or they can introduce two new terms, say 'gold-Au' and 'gold-pyrite', all the while retaining 'gold' as a term denoting both Au and pyrite.  No choice is forced upon them.  Moreover, such a decision doesn't affect the meaning of 'gold' instantaneously.  A new chain begins with the deciding scientists and moves outward, so that we will get many encounters such as what would happen if in Scenario 2 one member of one chain met a member of the other chain. 

So, what I think these examples motivate is a content externalism which already includes Goldberg's conclusion.  We did not need to add to it any form of metaphysical realism.  If the scenarios are intelligible and correctly analyzed, then it may well be possible to find ourselves in a situation like one of the Fool's Earthers in one of the scenarios.  But this is what we'd expect from externalism.  If meanings ain't in the head, then looking in the head ain't gonna get us any closer to them.  Thinking so may just be a remnant of our Cartesianism.