I don't think there'd have been any reason to think it would act differently than any other small object passing near a more massive one?
Another key figure was Pierre-Simon Laplace, who, in the first editions of Exposition du Système du Monde (1796), independently proposed the existence of such massive celestial bodies that could trap light. However, he later removed this idea from later editions, possibly due to skepticism over the corpuscular theory of light.
The first real attempt to quantify how much gravity might bend light was made by Henry Cavendish (circa 1783–1784). Though unpublished at the time, his notes (discovered in 1921) contained the first calculations of the bending of light under Newtonian gravity. Later, in 1801, Johann Georg von Soldner formally published a paper calculating the deflection angle of light passing near the Sun under Newtonian mechanics, obtaining a result (~0.85 arcseconds) that was later found to be half of what general relativity predicted.
These early discussions framed gravitational lensing in a purely Newtonian context, treating light as a particle affected by gravitational attraction.
Given that, why shouldn't light also fall at 9.8 m/2^s as it passes the earth? That would be extremely hard to measure, but if you extrapolate the phenomenon, you get gravitational lensing.
18th century scientists wouldn't have known that light was a zero-mass particle. The results of the calculations are ambiguous with a zero-mass particle (a=F/m is a divide-by-zero error). But if you don't know about that, you can just wonder if light particles (or waves, or whatever) also gets deflected at the same acceleration as ordinary matter.
Light in fact doesn't accelerate at a, but at 2a in a gravitational field. The Eddington experiment in 1919 agreed with that figure, disproving the Newtonian version.
That's kind of surprising to me. I assumed in the newtonian picture gravity would not bend light.