Western herbal knowledge ran through a chain of books for a very long time. Dioscorides compiled his account of medicinal substances in the first century AD, and it remained the standard reference in Greek, Arabic and Latin for something like fifteen hundred years, copied, translated and illustrated but rarely challenged. Monastic and vernacular traditions added to it; the Anglo-Saxon leechbooks preserve a distinct northern layer.
Nicholas Culpeper broke a different kind of authority. In 1649 he published an English translation of the London physicians' pharmacopoeia, which had been kept in Latin, and followed it with a herbal that assigned each plant a planetary ruler. The astrology was traditional. The politics were not: putting the material in English put it in the hands of people who could not afford a physician, and the College of Physicians was furious.
Alongside all this ran the doctrine of signatures, the principle that a plant's appearance indicates its use, so that a yellow flower treats jaundice and a walnut the brain. It is a memorable system and it is not a method of discovery, since the resemblances are found after the use is known.
The parting came with isolation. Once morphine was separated from opium at the start of the nineteenth century, and salicylate and quinine followed, it became possible to give a known quantity of a single compound. Pharmacology went that way and herbal practice did not, and the argument between whole plant and isolated constituent is still live.