The Old Guys were not dumb. Although it was not yet known that the amount carbon affected properties, they did make steel in different "tempers", those different "tempers" being different carbon contents. If you look closely at specification for modern "W1" steel, which is a plain carbon tool steel, you will find that terminology is still in use. W1 is made to a range of different carbon contents, referred to as "tempers". I have an Italian double-dog wheel lock (lock only) with very nice springs. Look close and one may see long slag inclusions, but they don't hurt very much. You may recall that many crossbows had tempered steel bows.
Since they did not have thermocouple temperature controlled furnaces, they did not necessarily heat treat springs the same as a modern heat treater might. To get the hardness (also known as the "temper") of the spring right, one would perform what is today known as a "slack quench", that is, a rather slow quench that did not develop full hardness. Today the guys who oil quench their 1095 or 1070 springs are performing a "slack quench", though I don't suppose they are charmed by that term. With a slack quench one may then be pretty far off in the reheat process, which today is called "tempering" or "drawing", and still get a useable spring. Not ideal, but much better than having your spring too hard. In which case they can just break all by themselves, under tension in that pretty lock setting on the table there.