In neoclassical theory, knowledge generates increasing returns - and theref
ore growth - because it is a public good that can be costlessly reused once
created. In fact, however, much knowledge in the economy is actually tacit
and not easily transmitted -and thus not an obvious source of increasing r
eturns. Several writers have responded to this alarming circumstances by af
firming hopefully that knowledge today is increasingly codified, general, a
nd abstract - and increasingly less tacit. This paper disputes such a trend
. But all is not lost: for knowledge does not have to be codified to be reu
sed and therefore to generate economic growth.