Attention has been drawn by Judith to three “features” of the temperature record without any formal assessment of whether the features are real or are just a figment of the mind being drawn to two cherries – one at around 1910 and another at around 1940.
These features are being attributed to some sort of non-specific “regime change” that forces climate into a different state.
But two of the features are accentuated only by two short-term periods of what could be “high frequency ‘noise’” which is apparently explainable by “Hypothesis I” – i.e. a short term period when some of the “natural” cooling or warming influences converge.
If you look at datasets with different coverage (eg. Land only, northern or southern hemisphere only, US only) then these “features” look very different suggesting that they would be sensitive to changes in coverage of the datasets.
None of this, though, takes anything away from ongoing warming driven by ongoing increases in forcing as Hypothesis 3 is only concerned with shifts away from the longer term forced trend. It doesn’t displace “AGW theory”.