Over the past forty years, U.S. inflation has exhibited highly persistent movements. Moreover, these shifts in inflation have typically had real consequences, implying a "sacrifice ratio," whereby disinflations are typically associated with recessions and persistent increases in inflation often associated with booms. One hypothesis about the source of the sacrifice ratio is that inflation - and not just the price level - is sticky. Another is that private-sector agents typically must infer changes in inflation objectives indirectly from central bank interest-rate policy. The resulting learning process can lead to a sacrifice ratio trade-off. In this paper, I allow for both sticky inflation and learning in interpreting U.S. macroeconomic developments since 1955. Two key empirical findings are, first, that allowing for learning reduces the evidence for sticky inflation. Second, there is less evidence for sticky inflation in the post-1983 period than earlier. Indeed, in some estimates, there is little evidence of sticky inflation in the period since 1983, although this result is sensitive to the details of the specification. Nonetheless, simulation results suggest that for realistic models, the sacrifice ratio can be accounted for entirely by learning.