Why Did We Not Automate Manual Labor Until the Late 1700s?

We've had the mechanical skill to build clockwork automata—not to mention literal clockwork—for hundreds of years (thousands, if Antikythera is representative).

Why did we not use it to automate manual labor until the late 1700s? https://t.co/X94Azna7iF

— Jason Crawford (@jasoncrawford) March 8, 2023

Your answer should be something that actually was required to begin automating manual labor.

Steam power, for instance, was not. The earliest cases of automation used water and animal power. The reaper was pulled by horses until the 20th century (steam engines were too heavy).

— Jason Crawford (@jasoncrawford) March 10, 2023

Your answer should not “explain too much.”

It can't be that no one cared about saving labor at all, because plenty of labor *was* automated or otherwise reduced. Even in antiquity there were grain mills.

— Jason Crawford (@jasoncrawford) March 10, 2023

Your answer should also not explain too *little*.

This phenomenon is broad, across pretty much all industries. The answer can't be something specific to textiles, or to agriculture, etc.

— Jason Crawford (@jasoncrawford) March 10, 2023

The stylized facts to be explained are something like:

1. We have been inventing labor-saving devices for thousands of years.

2. Automation in particular has existed in some form since at least the earliest grain mills c. 1st c. AD (any earlier examples?)

(continued)

— Jason Crawford (@jasoncrawford) March 10, 2023

3. Where we couldn't automate, we developed hand-operated tools, of increasing sophistication over time (e.g., the spinning wheel).

4. Sophisticated machinery, such as clockwork, has existed since the late medieval / early modern period.

(continued)

— Jason Crawford (@jasoncrawford) March 10, 2023

5. Still, by the early 18th c., very little was automated. Most labor was still done by hand.

6. Starting in the late 18th c., there was a rapid pace of automation until by the mid-20th c. (?) very little was still done by hand.

— Jason Crawford (@jasoncrawford) March 10, 2023

Are any of 1–6 above wrong/misleading? And have I missed anything important?

If not, what hypotheses are consistent with all of them?

— Jason Crawford (@jasoncrawford) March 10, 2023

My top hypothesis is around precision manufacturing.

The kinds of labor that were automated by wind and water mills were pretty crude motions: grinding, sawing, pounding. The kind of automation that waited until the IR involved subtle, delicate motions like spinning thread.

— Jason Crawford (@jasoncrawford) March 10, 2023

But this *seems* to be contradicted by (4). If we had the ability to make clockwork, was making a cotton-spinning machine that much harder?

— Jason Crawford (@jasoncrawford) March 10, 2023

Another hypothesis is about the ratio of the cost of capital to cost of labor. If that ratio is too high then it doesn't make sense to invest in machinery that saves labor. The ROI isn't there.

I think there is something to this—

— Jason Crawford (@jasoncrawford) March 10, 2023

But to make this consistent with all the hypotheses, you need to show that, e.g., a grain mill had good ROI in the 1st c., and a fulling mill had good ROI in the 11th c., but cotton spinning didn't have good ROI until the 18th c.

— Jason Crawford (@jasoncrawford) March 10, 2023

Robert Allen has calculations that show that in the late 18th c., the ROI of a spinning jenny was 38% in England vs. 2.5% in France, which explains why it wasn't adopted in France until the early 19th c. (when the ROI improved).

— Jason Crawford (@jasoncrawford) March 10, 2023

Would similar calculations show the the ROI of a spinning jenny would have been low in 16th- or 17th-c. England? Does that explain why it wasn't invented until the 18th c.? I don't know.

— Jason Crawford (@jasoncrawford) March 10, 2023

Perhaps another hypothesis is that the inventions that happened later weren't *impossible* or undesirable earlier—rather, the overall pace of invention was just slow (because of lack of inventors, lack of capital, small markets, etc.) and so the inventions waited a long time.

— Jason Crawford (@jasoncrawford) March 10, 2023

That would help explain why we got silk-throwing machines in the 15th c., and the stocking frame in the 16th c., but the spinning jenny and water frame waited until the 18th c.

— Jason Crawford (@jasoncrawford) March 10, 2023