[FOM] Frege's error
rgheck at brown.edu
Sun Jul 17 14:06:07 EDT 2005
Slater's diagnosis of what leads to the contradiction in Frege's system
would be more convincing if it were accompanied by an explanation of how
this error manifests itself in the derivation of Russell's paradox.
That, however, would prove difficult, since the remark reported in
Carnap's notes---roughly, that, if there is a function F(x,y), then
there is also a function F(x,x)---has no correlate in Frege's formal
I take it that Tennant was making essentially this point. There is a
sense in which no expression that "denotes a function" ever occurs in
Frege's system, so there can be no formal rule corresponding to this
informal remark. That is why imposing the framework of the lambda
calculus onto Frege's system can only be misleading and, indeed,
So what did Frege mean by the remark Carnap recorded? It is, in effect,
a remark about the syntax of the system: If F(,) is a two-place
function-symbol, then we can form both the expression F(x,y) and the
expression F(x,x). Not too much controversial there.
That is not to say that there might not be a way of massaging Frege's
system into consistency in accord with some of the ideas of the lambda
calculus. At this point, umpteen ways of accomplishing the task are
known, so that wouldn't be at all surprising, and maybe it would even be
interesting. We'd have to see the system to be able to tell. (Maybe some
of Nino Cocchiarella's modifications of comprehension could be
understood in this light?) But it's hard to see why any one of these
ways should have any special claim to reveal "the error" Frege made.
That case would have to be made on independent philosophical grounds.
For those who might be interested: Please note the new email address.
More information about the FOM