Prophecies are a staple of fantasy fiction, and for good reason: they are a convenient way of providing the heroes with information to get the plot moving while also imparting an aura of ominous mystique. How do you write a good prophecy for your story or game? Let’s start by looking at how prophecies worked in historical cultures.

Nearly every people in history has believed in some way of communicating with supernatural forces in order to gain special knowledge or insight, but the methods, purposes, and results of that communication could vary widely from culture to culture. By “prophecy” we usually mean something more particular: statements about specific future events which impart the necessary knowledge for the recipient to avert, influence, or at least cope with the effects of those events. Numerous cultures in history believed in some way of gaining these kinds of insights.

The problem that historical oracles faced, of course, was that predicting the future doesn’t actually work. The priestesses at Delphi or the authors of the Sibylline Books at Rome had no special insight into the future any more the authors of modern horoscopes and fortune cookies do. Nevertheless, many people believed in the power prophecy. The Histories by Herodotus, a work which makes frequent references to oracles, gives a useful view of the ways in which people coped with the unreliability of prophecy.

Reasonable guesswork. Prophets may not have special knowledge of the future, but they can make reasonable guesses about what is likely to happen, the same as anyone else. When the small Greek city of Miletus led a rebellion against the powerful Persian Empire, it didn’t take much special foreknowledge to predict that things were going to go badly for Miletus. The Delphic oracle produced this prophecy: “Miletus, you who scheme at evil deeds, will be a feast and splendid gifts for many. Your wives will wash the feet of long-haired men. Strangers will tend my shrine at Didyma.” (Herodotus, Histories 6.19, my own translations) This prophecy turned out to be true, but plenty of other Greeks claiming no connection to the gods also knew that things were going to go badly for Miletus, and so refused to join the revolt.

Vagueness. The standard dodge for prophets from Delphi to Nostradamus is to give an answer vague and cryptic enough that it will seem to suit whatever actually happens. The most famous example is perhaps the Delphic oracle’s response to the Lydian king Croesus, who asked whether he should invade Persia. The oracle replied that by doing so, Croesus would destroy a great empire, neglecting to mention which empire would be destroyed. As it happened, Croesus’ attack on Persia led to the Persian conquest of Lydia, but if things had gone the other way, the oracle would still have been right. (Herodotus 1.53)

Unspecificity. Some prophecies, like the one given to Croesus about his war with Persia, gave vague information about a specific event; others gave detailed information without specifying what event it related to. For example, a little-known Athenian seer named Lysistratus predicted that “The women of Colias will cook with oars,” which came true when wreckage from the naval battle of Salamis washed up on Cape Colias and was used as firewood by the locals. (Herodotus 8.96) This prophecy is unambiguous about what will happen, but says nothing about when or why. Colias was downstream of an important harbor and shipping channel; it was not hard to predict that wreckage from some significant event would wash up there and be salvaged sooner or later.

Selection bias. People tend to remember things that confirm their beliefs and forget things that don’t. People who believed in the power of oracles accordingly tended to remember prophecies that turned out to be true or could be interpreted to be true. Almost all the historical prophecies we have recorded were written down only after they had apparently come true. A number of recorded prophecies from the Delphic oracle begin with the word “But,” suggesting that some preceding part of the oracle has been left out, possibly because it turned out to be wrong or not relevant, such as in another Delphic reply to Croesus: “But when a mule becomes the king of the Medes, then flee, soft-footed Lydian, by the pebbly Hermus, and do not be ashamed to be a coward.” (Herodotus 1.55) This part of the prophecy was interpreted after the fact to refer to the Persian king Cyrus, whose ancestry was both Persian and Median, analogous to a mule, which is the progeny of a horse and a donkey.

Intrigue. Sometimes prophecies were manipulated in order to achieve the results some party wanted. It was an open secret that the priests at Delphi could be bribed to give particular answers. Other oracles and seers were no doubt similarly open to influence. The Alcmaeonid family of Athens were known to have bribed the Delphic priests to encourage the Spartans to help them against their rivals in Athens. (Herodotus 5.63) Another kind of manipulation is exemplified by Onomacritus, a collector of oracles who tried to encourage the Persian king Xerxes to invade Greece by sharing only those prophecies in his collection that seemed positive for him and hiding any that seemed negative. (Herodotus 7.6)

Now, as an author with full control over the world of your imagination, you don’t have to resort to any of these dodges. If you want your ancient prophecies to come true, then they will. The problem with prophecies in fiction, though, is they risk undermining the agency of the main characters. If prophecies predict the threat or its resolution too reliably or in too much detail, opportunities for drama are lost. If your work is for a game or some other setting where other people will have input to the plot, you can bet your dice that as soon as you hand them a prophecy they will try to exploit, invalidate, or weasel out of it in some way.

Uncertainty is a source of drama. When your audience already knows how everything is going to end, it’s harder to keep them interested in the story. Prophecies risk diminishing drama by introducing too much certainty. How do you keep the uncertainty in a story when there’s a prophecy involved? The techniques mentioned above are a good place to start because they serve the same function for a different reason: historical prophets had to keep uncertainty in their predictions because they didn’t actually know what was going to happen. You can use the same ideas in order to avoid tipping your hand too much to your audience or players.

Reasonable guesswork. If an in-story prophecy confirms something your heroes already suspect or adds useful detail to a picture that was already becoming clear, it can add impulse to the plot without dominating it. Conversely, a prophecy that doesn’t provide answers but spurs your heroes to ask important questions can be a good way to get things moving.

Vagueness and unspecificity. Both these techniques are good ways of keeping a prophecy from overwhelming the agency of your characters. If the prophecy refers to a specific event but doesn’t give clear details about it, or gives a clear prediction without specifying when, why, or how it will come about, there’s more room for your characters to work around it.

Selection bias. Lean in to the fact that prophecies can be wrong. If your characters (or their players) are aware that prophecies are unreliable or only seem true after the fact, their doubts about the truth or usefulness of the prophecy they’ve received can be a good source of drama.

Intrigue. There’s even more drama to be mined out of the fact that a prophecy might have been tampered with or invented, or that an authentic prophecy might have been delivered to your characters in such a way as to influence their understanding of it. Such puzzles open up interesting possibilities for side plots and interactions with antagonists.

As an author, the future is in your hands, a power that historical prophets never had. Still, you can learn from their examples how to make your prophecies sufficiently portentous without overwhelming your characters and plot.

History for Writers looks at how history can be a fiction writer’s most useful tool. From worldbuilding to dialogue, history helps you write.