You are in luck; we actually do know one three thousand years old European language, maybe one and a half. We also do know another (non-European) language spoken three thousand years ago which is quite conceivable that the ancient supernatural being might know.
Homeric Greek.
There are millions of people who can read Homer's and Hesiod's works in the original. We are even pretty certain that we know how pronounce them in a way which is not all that far removed from the original. Homer's Iliad and Odyssey, and Hesiod's Theogony and Works and Days were originally composed in the 9th, most likely 8th or maybe 7th century before the common era, so 2700 to 2900 years ago, most likely around 2800 years ago.
At that time the Greeks were busy using their ships to trade all over the Mediterranean, so their language was already beginning to be known internationally. Not so much as option number 3, but still plausible.
(The major advantage of Homeric Greek is that there are many people in the modern world who know it well enough to communicate. This is how the stranded Natucketers begin to communicate with their Tartessian frenemies in S. M. Stirling's celebrated Nantucket series beginning with Island in the Sea of Time.)
For completeness, we should also mention the remote possibility that the ancient supernatural being knows some sort of very archaic Latin. It is not at all likely, but hey, maybe his daughter dated a particulary exotic adventurer from Alba Longa, which, at that time, was the most important Latin city. (Rome did not exist yet.) (Yes, there was a people called Latins.) The main problem is that at that time Latin was a small language, spoken by an insignificant small nation in central Italy, and one would have to explain how come the ancient supernatural being knows it.
The great language of trade widely known at that time around the shores of Europe was Phoenician.
The Phoenicians actually traded as far north as Britain in the right time-frame, and they were the dominant long-distance traders in Europe at that time.
Phoenician was actually written at that time, so it is plausible to find inscriptions.
Phoenician is quite similar to the oldest layers of Biblical Hebrew, so that it is quite easily believable that a modern scholar would quickly make sense of it. (There are very many people who study Biblical Hebrew.)
So basically, that's it: if they are to communicate orally without spending time actually learning each other's language, it would be in Homeric Greek or in Phoenician.
If there are inscriptions to be found and quickly deciphered, they would most likely be in Phoenician.
But! It is perfectly conceivable (and, it my opinion, it would make a great episode in the story) that the inscriptions are in Proto-Germanic written with a local adaptation of the Phoenician alphabet. At that time, Proto-Germanic was spoken in southern Scandinavia and the Jutland peninsula; and you are speaking exactly about the time when the Phoenician invention of the alphabet was spreading like wildfire west and east -- it was such a simple idea, it makes writing so easy, why on Earth didn't we think about it! If it spread to the west, and it spread to the east, why not also to the north?
(We know Proto-Germanic a lot better than Proto-Celtic, mainly because the Germanic languages are much more conservative than Celtic languages and thus the reconstruction is much easier, because Germanic languages are attested at an earlier stage, and because Proto-Germanic has so many more living descendants than Proto-Celtic.)
Note about the idea of some sort of Proto-Basque: While we are certain that the ancestors of the Basques spoke a language which is the ancestor of Basque, we unfortunately don't know anything much about that language. Yes, it must have existed. No, we don't know it. Not that very few people study it; no: we simply have very very little data (basically, a handful or names recorded by the Greeks and the Romans) so that, at present, we don't even pretend to have the foggiest idea about it.
