In the defense of 19th century scientists, the best approximate solutions to the Navier-Stokes equations really did predict that mechanical aircraft should not be able to fly, curve balls should not curve in the air, and so on.

It was not until 1904 that Ludwig Prandtl published Über Flüssigkeitsbewegung bei sehr kleiner Reibung (On the Motion of Fluids in Very Little Friction) which first discovered the importance of boundary layers. The breakdown of approximations in those boundary layers allows for all sorts of behaviors that came as a surprise. In time science caught up up with practical engineering advances to finally understand how airplanes can fly, pitchers can throw curve balls, and so on. And even so, those old approximations are still used because they are mostly right!

So they were wrong, but it wasn't gibberish either.

For a similar example of a mostly correct scientific theory producing wrong results, until near the end of the 20th century the linear wave model predicted that rogue waves were impossible. Today we can look back at shipwreck records and laugh at their stupidity. But in fact you can spend a week looking at every wave that passes a point and probably won't find even a single wave that doesn't fit the theory.

Scientific overconfidence in well-tested theories is a systemic error that we are likely to always be prone to. Most of the time it is well justified. But we do nobody a favor by dismissing past examples of this as "gibberish".