I'm using this fable to try to clarify my thoughts.
Suppose, just suppose, that Nortel shares had sticky prices. Rather than adjusting almost instantly to ensure market-clearing equilibrium, the price of Nortel shares adjusted slowly over time in response to excess demand or supply of Nortel shares.
That would mean that uninformed traders, who knew only the past history of Nortel share prices, could forecast the price of Nortel shares. Not perfectly, but it wouldn't be an unforecastable random walk. If the price of Nortel shares increased today, it would probably increase tomorrow too. If the price of Nortel shares decreased today, it would probably decrease tomorrow too. The sort of naive extrapolative expectations that create bubbles would then become rational.
But it would be difficult to profit from that ability to forecast the price of Nortel shares. If you saw the price rising, you would expect it to keep on rising, and you would want to buy Nortel shares. But who would take the other side of the trade? Only if you were lucky, and were at the front of the queue of buyers, and found someone willing to sell for idiosyncratic reasons (his car just broke down, so he needed to sell his Nortel shares to buy a new car) would you be able to profit.
Now suppose, just suppose, that in addition to having sticky prices, Nortel shares were used as the medium of exchange. So everyone was always buying and selling Nortel shares for all sorts of idiosyncratic reasons. Everyone is a noise trader too. When you sell your old car, you buy Nortel shares in return. When you buy a new car the next day, you sell Nortel shares in return.
So if you saw the price of Nortel shares had been rising, and you knew it would probably keep on rising, you would want to hold more Nortel shares. And everyone else would want to do the same. You might get lucky, and be at the front of the queue of people wanting to buy more Nortel shares, and find someone to take the other side of the trade. But if Nortel shares were the medium of exchange, there's a sure-fire way for any individual to hold more Nortel shares: simply sell fewer Nortel shares by buying fewer other goods.
As the price of Nortel shares rose, trading volume would fall, because people would want to sell fewer Nortel shares, even as they wanted to buy more, so fewer Nortel shares would actually be bought and sold. And people would notice the declining volume of trade, and the excess demand for Nortel shares, and the excess supplies of other goods, and call it a recession.
Now suppose, just suppose, that in addition to having sticky prices, and being used as a medium of exchange, Nortel shares had very safe earnings that were recession-proof, while earnings on other assets tended to fall in recessions. So if you saw the price of Nortel shares rising, you would expect a recession, and you would expect the earnings on other assets to fall, which would increase the fundamental value of Nortel shares as a safe haven. Any bubble in Nortel shares would increase the fundamental value of Nortel shares.
Now let's go back to the real world.
Government bonds are substitutes for government money. As the price of bonds rises, and the rate of interest on bonds falls towards 0%, they become closer and closer substitutes for money. So a big enough increase in the demand for bonds becomes an increased demand for money+bonds. And the price of money+bonds, in terms of other goods, is sticky, because it's just the reciprocal of the general price level, which is sticky. And money is the medium of exchange. And money+bonds is a safe haven in a recession, when the return on other assets falls.
A bond bubble, leading to a money+bonds bubble, is much more rational, and much more damaging to the economy, than a bubble in real-world Nortel shares.
Money is different from other assets: money doesn't have a price of its own; everyone is a noise trader in money. Update: and when money's trading volume falls, everything's trading volume falls too.