Earlier quoted context omitted.
> (1) there is NO single cause and (2) nothing can be quantified Fortunately causal analysis supports 1 and contradicts 2. With good data, models and algorithms you can _quantify_ the strengths of various causal factors, and confirm or discredit models. Time scales don't matter, complexity can be overcome, and even unobserved causes can be modeled. > This is why subjects like history differ from physics. We should no…
I am afraid your citations don't support your claim. Those works are about using natural experiments and large sample sizes to estimate causes and effects. But that isn't possible for single, large events like "the industrial revolution" because there is only one such event. This is why no one can say "industrial revolutions are 8% caused by geography and 12% caused by social structures" with any actual basis for tha…
> We rationalize these findings using a dynamic spatial model, where slavery investment raises the return to capital accumulation, expanding production in capital-intensive sectors. To establish causality, we use arguably exogenous variation in slave mortality on the passage from Africa to the Indies, driven by weather shocks.
The second paper focuses on labor-capital substitution. In this view, there are again many industrialization locations, but also each tech is a micro-revolution. You can even see the proposed causal DAG in Figure 4.
> To measure technological change, we collect new, granular data on the diffusion of technologies (both labor-saving and not) in industrializing England. To measure the geographical patterns of labor scarcity, we exploit the massive shock to labor supply that resulted from military recruitment during the French Revolutionary and Napoleonic Wars (1793-1815).
Neither paper predicts a boolean "revolution", because thats really an academic construct to describe a shift in practices over time with fuzzy boundaries and definitions. Instead they go after capital formation or adoption of labor saving devices, which seem like reasonable proxies.
[1]: https://www.nber.org/system/files/working_papers/w30451/w30451.pdf
[2]: https://www.gla.ac.uk/media/Media_900972_smxx.pdf