This statistic, the percentage of job losses that are permanent, is a useful way to distinguish Keynesian recessions from Recalculations. In a Keynesian recession, you are temporarily laid off because of excess inventories and deficient aggregate demand. You wait to be recalled by your firm. This was true of recessions from the end of the second World War through the 1980 recession. Even the 1975 recession, which was a "supply shock" (higher oil prices, requiring some permanent readjustments), had a relatively low share of permanent job losses.
In a Recalculation, you permanently lose your job and you have to find something else. The Recalculation model increasingly holds as we move away from an economy dominated by manufacturing. Even though the 1990 and 2000 recessions were relatively mild, a large share of the job losses were permanent.
As commenters have been pointing out, recalculation is always taking place in the form of creative destruction. See Lectures on Macroeconomics, especially number 3. The problem now is that the economy cannot handle the amount of adjustment that is required to deal with the sudden change in the housing and financial markets.
What does a "jobless recovery" mean? I do not like the term. Around 2003, when the term was first coined, I instead described it as a "productivity-cushioned recession." Either term describes the same phenomenon--GDP growing significantly faster than employment.
During the prior jobless recovery (or productivity-cushioned recession), inflation was relatively low. That would suggest that my inflation bet may be wrong. On the other hand, during the prior episode the government was not issuing trillions of dollars of new debt. Ultimately, I think that all this debt will produce a lot of inflation, with very little net gain in employment relative to what would have taken place without it.