The ontological status of gold

By Felix Salmon
December 22, 2009

I was pleasantly surprised by the volume of email response I got to a passing reference to Kripkenstein on this blog — clearly quite a lot of you enjoy a bit of analytical philosophy! I went out to lunch today with a couple of philosophically-inclined finance types as a result, and, since I’m still high on Sichuan peppercorns and it seems to be something of a slow news day, I thought I’d put up a poll.

Remember this wonderful graph, from Paul Kedrosky, showing the price of gold in gold? Pay attention, there will be a quiz.

So here’s the question, for those of you who remember the analytic-synthetic distinction:

Or to put it another way: Can an analytic a priori statement be funny?

Update: With 125 votes cast, a clear majority of you (59%) are voting for the first option, analytic a priori. But you’re wrong, as dsquared explained to me in an email this morning — what happens to the graph if the price of gold goes to zero?

The assertion that the price of gold, in gold, is 1, is not analytic because it depends on the truth of at least one other proposition (that gold has a nonzero price) and is not a priori because there are possible worlds in which gold does not have a nonzero price.

Which just goes to prove, if nothing else, that philosophy is probably not best conducted by polling blog readers.

Update 2: Natecha defends the analytic-a-priori crew against dsquared, saying that the statement “x/x=1″ isn’t false when x=0, just undefined. He adds for good measure that “Daniel’s comment flies in the face of philosophical orthodoxy about analyticity and apriority”. Which is a statement I daresay Daniel would agree with.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

I think it depends entirely on whether the subject is considered to be ‘gold’ or ‘price’.

The former is analytic a priori, the latter is synthetic a posteriori.

Posted by Bergamot | Report as abusive

It appears to me that the subject is, fully stated, the price of gold in gold.

Not sure that it requires a complete analytical deconstruction, but maybe. Umm. Price implies something that is being valued, in this case, gold. Price also implies that there is a medium of exchange, typically currency, but it could be a form of barter, in this case, gold.

So, given the nature of “price” as a concept, you can’t really examine the nature of propositions about price without first having a subject, as price itself, is not a subject. Gold itself is not a subject. The price of gold, that is a subject. But it lacks a conceptual component in which to continue in analysis – that is where the principle shift lies.

Price gold in dollars, and the analysis becomes something which must be discussed and derived on the basis of experience, ie data. Price gold in gold, and the analysis reveals an obvious logical determination within the nature of our concept of price. Hence, the analytic a priori.

In this case, for the subject to be synthetic a posteriori, the price of gold would have to be examined in some other denominating currency.


Posted by Ken_H | Report as abusive

the USD/USD does not look much different. in fact the correlation between the gold/gold graph and the USD/USD graph is 1.

Posted by q_is_too_short | Report as abusive

This post makes me want to chase Soros around the parlour with a red hot poker.

If you want some trenchant feedback on this question, you might want to offshore it to, or joint venture it with, the mind-melting folk at

Posted by Uncle_Billy | Report as abusive

Not to be a stick in the mud, but I argue with the premise. What about cons like Cash for Gold? It’s plausible to argue that the USD/Gold for them is not the same as the USD/Gold you can get on a commodity exchange, hence an inequality in the Gold/Gold relationship.

Oooh…maybe that chart would be WAY more informative than the spot price chart, eh?

Posted by jamesagain | Report as abusive

A priori and analytic mean the same thing. The options you possibly wanted to list (given that you mentioned Kripke) are probably as follows:

1. A priori and necessary
2. A priori and contingent
3. A posteriori and necessary
4. A posteriori and contingent

Possibly, (2) would be the answer. Another example of (2) is the standard meter in Paris. It’s defined as being as long as itself, but it would have had been slightly longer or shorter that it actually is if the room it is kept in had a different temperature.

Posted by Jonas82 | Report as abusive

@Jonas82: But this is more like “there is one meter per meter”, which would be true regardless of what’s sitting in Paris. (It would be true even if you replaced that standard meter with a small hunk of gold!)

Posted by absinthe | Report as abusive

Hm. Reading up on this, it seems Kant did distinguish the a priori from the analytic and the a posteriori from the synthetic. However, the examples I was able to google, are all terribly unconvincing and I have never noticed the distinction being made in today’s literature. I’m more of a logician than a philosopher of language though.

Posted by Jonas82 | Report as abusive


Yes, you’re right. Sorry. It’s 2AM here. :)

Posted by Jonas82 | Report as abusive

Every object is self-identical.
Yet propositions of self-identity are not obvious and uninformative.

I’m not sure I accept your distinctions.

However, “Can an analytic a priori statement be funny?”


Therefore: Something which may not exist may still be funny.

Posted by T_S | Report as abusive

“Price” is a truly meaningless concept outside a specific culture and context where exchange/trading occurs. So the assertion is necessarily about a specific cultural context, already making it unlikely to be analytic.
We are further told that this culture considers gold essentially fungible, but there’s no necessary reason for that: and, two idential 1oz gold coins do not have exactly the same atoms, or the identical isotopic composition to 100 decimal places, and they cannot be in the same physical position at once. Yet this particular culture doesn’t care about any of this, it seems.

Undoubtedly contingent, a posteriori.

Posted by axg4 | Report as abusive

Consider: “The price of an ‘X’ slave, in ‘X’ slaves, is 1″ where is X is some reasonably broad description (e.g. “healthy male”). For us, no matter what X is, this is meaningless, but abhhorrent and false to the extent we can even try to make sense of it. At other times and cultures, the question can be meaningful, but individual slaves weren’t considered fungible enough for this ever to be a general truth. At other times perhaps, for suitable X, both meaningful and approximately true.

Challenge: how is this philosophically so different from the gold question?

Posted by axg4 | Report as abusive

Before anyone invests much more time in this, remember that Wittgenstein went insane and locked himself in a cabin on a remote Norwegian fjord after thinking about such things (roughly). He was the victim of some severe inbreeding, but still.

This does make the case for letting others do the heavy lifting, however, a little stronger.

Posted by Uncle_Billy | Report as abusive

The price of gold is the price of gold (always). That is, of course, analytic.

I am surprised that analytic a posteriori is a choice. A sociological remark: I don’t know of anyone who accepts the analytic/synthetic distinction, but holds that analytic truths are not knowable a priori. (This is not to say, as a commenter above erroneously remarks, that “a priori” and “analytic” express the very same concept. They do not.)

– a professional philosopher

Posted by Anonymous | Report as abusive

I’ve been to Wittgenstein’s cabin at the fjord. It was a place of nonzero value. Which is more than I can say for a lot of analytic philosophy. Beautiful and peaceful, although the sunlight was rather limited.

Posted by MartinGreen | Report as abusive

gosh, am I really going to have to establish a Reuters login for this?

Natecha is wrong wrong wrong. When x=0, x/x is undefined, but “x/x=1″ is definitely false – it doesn’t take much work to derive a contradiction from it, think of what the value of 2x/2x would be. “X/X =1″ isn’t a failure of an existential supposition, it’s the assignment of a value (1) to a function that (analytically) can’t take that value.

Posted by dsquared | Report as abusive

note following on from the above that “the USD/USD exchange rate =1″ *is* an analytic and a priori truth, because it is analytic from the definition of “a dollar” that it has a nonzero value in dollars”, whereas this isn’t analytic from the definition of gold.

Posted by dsquared | Report as abusive

The price of gold is the price of gold. Positive, negative or otherwise. The analytic is the logic.

Posted by BarbaraRobbins | Report as abusive