One thing that is interesting from the Lampedusa Island, Italy CO2 concentration data, is that it is very cyclical with a single year period. Obviously, the CO2 spikes in the summer, when it is warm, and drops dramatically in the winter, typically the min-to-max range is 10-15 ppm. obviously, warm temperatures cause the CO2 level to spike, while the colder winter temperatures cause the CO2 levels to dip. This further suggests that it is temperature that is driving CO2 concentrations instead of CO2 concentrations driving temperature. If it were CO2 concentrations driving temperatures, there should not be an oscillation of the CO2 concentrations in such a cyclical fashion.
That's good point and there was a very interesting article posted recently that analyzed Vostok ice core temp and CO2 inferred data in those terms. Specifically he suggested that the correlation between global temperatures and CO2 concentrations was explained by the temperature dependence of CO2 solubility in water.
The chart you're referring to shows, in addition to the yearly variation, a distinct upward trend. If we had the regional temps for that time range, we might be able to test your hypothesis in that other guys terms.
The cyclic change is caused by seasonal variation of plant growth - if you look around the net (GOOGLE "seasonal variations CO2") you will find some nice graphs of concentration shift by latitude and season. We know this, for example, because CO2 with fossil flue carbon isotope markers is relatively stable while the CO2 produced by living organisms varies seasonally.
The oscillations are due to seasonal uptake of CO2 by northern deciduous forests.