Easy, just take it as a given.
The British Antarctic Survey, who really ought to know better, have come up with this piece of nonsense:
It is often said that the temperature ‘leads’ the CO2 during the warming out of a glacial period. On the most recent records, there is a hint that the temperature started to rise slightly (at most a few tenths of a degree) before the CO2, as expected if changes in Earth’s orbit cause an initial small warming. But for most of the 6,000-year long ‘transition’, Antarctic temperature and CO2 rose together, consistent with the role of CO2 as an important amplifier of climate change (see Fig. 4 overleaf). In our modern era, of course, it is human emissions of CO2 that are expected to kick-start the sequence of events. We see no examples in the ice core record of a major increase in CO2 that was not accompanied by an increase in temperature.
If we look at Fig 4: we can also see that the converse could be true, as they start to imply, then rapidly back-pedal from, presumably before they are hauled up in front of the Inquisition for heresy. They then go on to state a fairly conclusive proof why CO2 isn't causing warming, presenting it as proof of the opposite.
From the air in our oldest Antarctic ice core, we can see that CO2 changed in a remarkably similar way to Antarctic climate, with low concentrations during cold times, and high concentrations during warm periods (see Fig. 3 overleaf). This is entirely consistent with the idea that temperature and CO2 are intimately linked, and each acts to amplify changes in the other (what we call a positive feedback).
Well, no, if it was positive feedback, the warming would release more CO2, which would cause more warming, which would release more CO2, which would cause more warming and so on, until the seas boiled away into space. This they haven't done, from which we conclude that a rise in CO2 levels cannot both cause warming and be caused by it. Since no-one seems to dispute the former, then the conclusion about the latter is obvious.