In short, we can not presently seperate the signal of CO2 from
the noise of natural variation.
Warm records are set through the 1930s and 40s, with a long stretch of no records until the 1980s, when the global warming signal firmly emerges from
the noise of natural variation.
My personal belief, from years of doing my own research about anthropogenic global warming, is that there is an effect but it is so small that it is lost in
the noise of natural variation.
Before that it would be harder to detect the signal of its effect against
the noise of natural variations being a trend of less than 0.01 W / m2 / yr.
Rather than waiting for models and observations to improve, or for the anthropogenic signal to emerge from
the noise of natural variation, the scientific community is starting to be able to provide information to policy makers and societies on the timescales that they need it.