So, first, let's lay out what we mean by gravy. What Americans mean by it can be quite different to what I, in Ireland, mean by it, and different again from what a French cook would understand, if they'd even accept the term rather than jus. However, in all these cases, your baseline for gravy is the juices from the roasting pan, with flour or some other starch added as a thickener to varying degrees. I confess I'm not quite sure what happens from there to get to the off-white stuff the Americans eat with their biscuits, but having established the baseline, any time I mention it hereafter, you can assume it's the British Isles variety I'm talking about.

I'll trace it backward a bit through various cookery books. Mrs Beeton, in 1861, is more or less the origin of what we think of as British cookery. She says:

By the addition of various store sauces, thickening and flavouring, the stocks here referred to may be converted into very good gravies. It should be borne in mind, however, that the goodness and strength of spices, wines, flavourings, &c., evaporate, and that they lose a great deal of their fragrance, if added to the gravy a long time before they are wanted. [...] The shank-bones of mutton, previously well soaked, will be found a great assistance in enriching gravies; a kidney or melt, beef skirt, trimmings of meat, &c. &c., answer very well when only a small quantity is wanted [...] the fragrance of gravies should not be overpowered by too much spice, or any strong essences

So at that stage, we're clearly looking at something very similar to the modern gravy, albeit with plenty of variation. Isabella Beeton was publishing a guide on how to be middle class as much as a cookbook, which is why there's so much detail here - she didn't want to assume that anyone knew any of this.

Back another 60 years, then, to John Mollard's The Art of Cookery Made Easy and Refined.

TAKE slices of ham, veal, celery, carrots, turnips, onions, leeks, a small bunch of sweet herbs, some allspice, black pepper, mace, a piece of lemon-peel, and two bay leaves; put them into a pan with a quart of water, and draw them down till of a light brown colour, but be careful not to let it burn; then discharge it with beef stock. When it boils, skim it very clean from fat, and thicken it with flour and water, or flour and butter passed. Let it boil gently three quarters of an hour; season it to the palate with cayenne pepper, lemon juice, and salt; strain it through a tamis cloth or sieve, and add a little liquid of colour, which may be made as in the following receipt.

This looks more complicated, but since the ham, veal, and everything up as far as the flour is basically just a stock, it's pretty much the same as the later recipe. It's notable that flour and butter are an option for the thickening here - that's a roux, the basis of many French sauces, the New Orleans gumbo, and the English white sauce. Mollard also uses the word cullis to describe this gravy, which might be the root of coulis, used now to describe a simple fruit sauce.

Let's jump a bit further back, then. The Form of Curye is from 1390. This was before spelling was in any way standardised, so you see 'gravy', 'gravey', 'grauy', and a variety of other spellings. Gravy at this point doesn't appear on its own; it's there as part of dishes like 'Rabbit in Gravy', 'Oysters in Gravy', and so forth. For an example:

Take Connynges smyte hem to pecys. Parboile hem and drawe hem with a gode broth with almandes blanched and brayed. Do þerinne sugur and powdour gynger and boyle it and the flessh þerwith. Flour it with sugur and with powdour gynger an serue forth.

or:

Schyl Oysters and seeþ hem in wyne and in hare own broth. Cole the broth thurgh a cloth. Take almandes blaunched, grynde hem and drawe hem up with the self broth. & alye it wiþ flour of Rys. And do the oysters þerinne, cast in powdour of gyngur, sugur, macys. Seeþ it not to stondyng and serue forth.

The general gist here is "boil up whatever you're having in its own broth, add ground almonds and maybe some other stuff, including sugar and ginger, and serve it". So it is essentially the same thing, albeit now with ground almond as the thickener rather than flour (although the second recipe also uses rice flour). It would taste quite different, too, with the almonds, sugar and ginger, so it might not be recognisable as gravy to us. As it happens, I've cooked the chicken version of this, and the taste of ginger is far and away the strongest element in it. But it's clearly the root of the same dish.

There are probably older antecedents - something mentioned in three or four recipes here hasn't just sprung into being - but there are very few cookery books older than this, and I haven't found any mentions of gravy in them. Middle Eastern cookery doesn't, as far as I can see from Annals of the Caliphs' Kitchens, a modern translation of a 10th century Baghdadi cookbook, have anything terribly similar; most of the sauces there are vinegar-based.

So to summarise: we can trace it back to the 14th century, at least, although it's getting hard to recognise by taste then.