Miracles and Historical Method

by Luke Muehlhauser on November 26, 2009 in Reviews

I’m blogging my way through Sense and Goodness Without God, Richard Carrier’s handy worldview-in-a-box for atheists. (See the post index for all sections.) Last time, we discussed the paranormal vs. scientific method. Today we discuss miracles vs. historical method.

Here and elsewhere, we have already discussed several weaknesses of metaphysical speculation about the supernatural. It offers no way to check its methods or facts for accuracy. There’s a massive lack of evidence, and a great many hoaxes and crazies. We also lack plausible mechanisms that would explain the existence and activity of supernatural beings. Finally, there’s the basic argument for naturalism.

As if that wasn’t enough to doubt miracle claims, perhaps the best reason to doubt miracle claims is not philosophical but historical. Every time we look at a miracle claim, we see there are adequate natural explanations available, such that we need not resort to magical ones.

Carrier’s example is The Rain Miracle of Marcus Aurelius.

In 172 C.E. a legion commanded by Marcus Aurelius was surrounded by barbarians somewhere in Eastern Europe. The soldiers, dying of thirst under the hot summer sun, pleaded with the gods for salvation. Clouds quickly filled the previously clear sky. Cool rain fell, and lightning struck down their enemies.

A few years later, the miracle was carved into stone in the Column of Marcus Aurelius, still visible today.

Only eight years after the event, Christian apologist Apollinarius claimed the legion was composed of Christians who prayed to their own god for help, and thus the miracle proved the deity of Christ. Seventeen years later, Christian apologist Tertullian claimed the same thing. But 25 years later, pagan historian Cassius Dio said it was the Egyptian wizard Harnouphis who summoned the god Hermes to rescue the legion.

So what really happened? How could we know?

We have no evidence that Christian prayers or Egyptian magic can call down rain or army-slaughtering precision lightning strikes. In contrast, we have ample evidence that natural causes could product the same result:

After all, rapidly-appearing and violent thunderstorms are a frequent occurrence in nature, especially in the summer months, while the human propensity to exaggerate is both infamous and ubiquitous. So we have much more reason to believe this was a natural event than a natural one.

But this does not mean, of course, that the legion was not made of Christians praying to Christ, or that it did not contain an Egyptian wizard who tried to summon help by casting spells. Is there anything to be said about these claims?

Truly, it would be odd for a pagan emperor who persecuted Christians to allow any of them into his army, let alone fill an entire legion with them. Moreover, all legionaries were required to offer daily prayers to the pagan god Jupiter, protector of the legions. Besides, we just have no evidence there were any Christians in Aurelius’ legions.

Now what about the pagan historian’s tale?

An inscription found in Eastern Europe attests that an Egyptian sorcerer by the very name Harnouphis was traveling with the imperial legions at the time, and the coin minted by Aurelius [a few years after the event] specifically depicts the god Hermes standing in an Egyptian temple…

What we have discovered is one among many examples of Christian authors making up stories in the context of deliberately trying to persuade readers to convert.

But guess what? Centuries later, and throughout the Middle Ages, ignorant Christians celebrated the “Miracle of the Christian Legion.” The Christian apologists won the propaganda war, as they often did – despite the absurdity of their claim and the lack of evidence for it.

The pagan historian’s story about the Egyptian wizard is probably correct, but nobody paid attention to it.

Instead we have a legend springing up just eight years after the fact, when thousands of eye witnesses were surely still alive, when the government was already promoting its alternative account, when all the necessary records were available. And despite these seemingly unfavorable conditions, this legend beat out the truth.

(Those who read contemporary Christian apologists will notice the relevance of all this to the stubbornly obtuse arguments of those who defend the magical resurrection of Jesus.)

So the First Rule of Historical Method is this: Don’t believe everything you read! Humans are notorious liars, eager exaggerators, and happy to believe anything that affirms their pre-existing beliefs.

The Ancient Milieu

It’s also important to understand the ancient culture.

In the 18th century, David Hume observed how suspicious it was “that such prodigious events never happen in our days,” yet supposedly they were routine in ancient times. Historian Edward Gibbon also detailed the “fraud and sophistry in the defense of revelation” seen throughout antiquity.

So what happened? Was it the case that amazing and public miracles were common throughout human history, but suddenly stopped in the modern era? Or is it more likely that people themselves were more credulous and superstitious, but this gradually abated during the Enlightenment?

This was an age when only 20% of the population could read anything at all, and far fewer had access to books, which were very expensive to produce. They knew nothing of science. To them, disease was caused by demons, not microorganisms. The sun went ’round the earth, and the universe was little bigger than the earth and sky. They thought lunar eclipses was the result of monsters trying to devour the moon. When an eclipse occurred, they would bang pots and blow horns to scare away the monsters. The few educated authors of the time complained in their writing of how the racket filled entire cities and countrysides.

Miraculous healers – in flesh or statue – were also common. We have testimonies carved in stone by those who were “cured” of blindness, muteness, and paralysis by pagan gods like Asclepius.

But it wasn’t just superstition and gullibility. There were skilled hucksters, too, such as Alexander of Abonoteichus. Alexander started his own religion: Glycon was the god, and Alexander was his prophet. Alexander tricked so many people into thinking he could perform miracles that they successfully petitioned the government to print coins with the image of Glycon on them – and we still have those coins to this day. Even when the educated skeptic Lucian revealed Alexander’s tricks, his believers only defended Alexander all the more, sometimes violently.1

Historical Method Saves the Day

So if this is what fills the testimony of ancient history, how can we know anything about what happened? How can we separate the truth from the bunk? How do we know it’s not all bunk?

A reasonable belief in historical facts is usually established by two kinds of arguments: the argument to the best explanation, and the argument from evidence.

Both these methods require that the historian figures out where all the evidence comes from, its date of composition or creation, what the claim or object means in its historical context, what sources were used by the author, whether there is evidence of tampering, and so on. If he can’t get his hands on this information, any conclusions he draws are dubious. And this is the reason for Rule Number Two of Historical Method: Always ask for the primary sources of a claim you find incredible. Gathering all the relevant evidence also entails Rule Number Three of Historical Method: Make sure you understand all aspects of the historical context.

Now, let’s look at those two kinds of historical arguments: the argument to the best explanation, and the argument from evidence.

Argument to the Best Explanation

The best explanation of some set of evidence is generally taken to have five virtues:

  1. Explanatory scope. The best explanation explains more facts than its rivals. Consider the rain miracle: the “sudden storm” theory explains all the evidence we have. The Christian Miracle theory fails to explain the original credit given to pagan gods, though the Egyptian wizard theory explains the pagan reports and also the inscription indicating he was with the army in Eastern Europe at the time.
  2. Explanatory power. The best explanation renders the facts more probable than competing explanations do. Consider the Egyptian wizard theory about the rain miracle. If the Egyptian wizard did bring on a storm by magic, this would make the evidence we have very probable. So that theory has strong explanatory power. It even makes the event’s hijacking by Christian apologists probable, since these Christians would not want proof of pagan magic to triumph over the Gospel. And while the Christian theory would make some of the evidence probable, it wouldn’t make the pagan report from Cassius Dio probable, since his report is rather sober and does not praise pagan magic, so it doesn’t appear Cassius Dio is making something up for polemical ends.
  3. Plausibility. The best explanation is more plausible than other theories. It should fit true generalizations about the relevant time, place, and setting. The idea of a Christian legion under Marcus Aurelius is quite implausible. The presence of a pagan wizard in a pagan army is not so implausible. Nor is a sudden storm, the details about which were exaggerated by superstitious minds.
  4. Ad Hocness. The best explanation relies on fewer undemonstrated assumptions than its competitors. Undemonstrated assumptions are called ad hoc features of a theory, things just made up to explain what otherwise doesn’t make sense. The Christian theory requires inventing some reason why Christians would be in a pagan army, while the pagan theory requires a reason why Christians would claim ownership of the miracle so quickly. Since there are no precedents for Christians in a pagan army, the Christian assumption is very ad hoc. In contrast, there is lots of precedence for Christian apologists making shit up, so the pagan assumption is much less ad hoc, or perhaps not ad hoc at all. Likewise, the claim that a sudden storm arrived just when the soldiers were trapped and thirsty is somewhat ad hoc.
  5. Fit to evidence. The best explanation fits with well-established facts. The Christian story does not fit well with known facts, but the pagan story does, and the story of a sudden storm certainly does. Sudden storms happen all the time.

So when we make an argument to the best explanation, we find it most likely that a freak storm occurred, that an Egyptian wizard took credit for causing it, that no Christians were likely involved at all, and that Christian apologists claimed ownership of the “miracle” after the fact.

We also discover we cannot justify believing this was a supernatural miracle. For example, consider the best magical explanation, concerning the Egyptian wizard. This explanation lacks explanatory scope: it does not explain the absence of Harnouphis’ magic throughout the rest of the war, nor why he was apparently successful only once in his life. But the theory of a coincidental storm explains these facts perfectly. There is also the problem that we have no evidence that magic works at all. The Egyptian magic theory is also extremely ad hoc, for it posits that Egyptian magic works but we don’t see it working today. The theory is also implausible, for why would Roman legions begin to meet more frequent defeats after this year, when they had pagan wizards with genuine magical powers in their ranks?

But we must admit the Egyptian magic theory has excellent explanatory power:

If Egyptian magic works, and Hermes really can be persuaded to play with the weather, this would make the occurrence of the storm very probable indeed. But the failure of the theory on every other point, and the fact that the alternative, of a lucky natural coincidence, also makes the event probable enough to be believed, leads us to conclude that no miracle happened that day.

The Argument from Evidence

Another method is the argument from evidence, which precedes the argument to the best explanation – before we can make an argument to the best explanation, we must show we have a believable fact to explain in the first place. This is a question of how much evidence we have, and how much we can trust it.

Carrier distinguishes five types of evidence the historian would like to have:

  1. Physical-historical necessity.
  2. Direct physical evidence.
  3. Unbiased or counterbiased corroboration.
  4. Credible critical accounts by known scholars from the period.
  5. An eyewitness account.

To illustrate these types of evidence, Carrier takes issue with Christian apologist Douglas Geivett’s claim that the evidence for the resurrection of Jesus meets “the highest standards of historical inquiry” and is as well-attested as Julius Caesar’s crossing of the Rubicon in 49 B.C.E.

Consider the evidence for Caesar’s crossing of the Rubicon. It’s physical-historical necessity is very high, since the history of Rome could not have preceded if Caesar had not physically moved an army into Italy. He could not have physically taken Rome without physically moving into Italy. In contrast, all we need to explain the rise of Christianity is a belief. The resurrection of a Jewish profit is not a physical-historical necessity to explain the testimony of superstitious believers or the rise of Christianity.

Second, we have some direct physical evidence for Caesar crossing the Rubicon. We have inscriptions and coins soon commemorating the Republican Civil War related to the crossing. In contrast, we have no physical evidence of any kind for the resurrection: no documents or inscriptions or genuine Shroud of Turin. (And even if we did, it wouldn’t mean he resurrected. It would mean he was vaporized or flash-frozen.)

Third, we have good corroboration for Caesar crossing the Rubicon. An unbiased source is one who would supposedly know if a story was true or not, but has no known reason to be credulous or distort the account. A counterbiased source is even more desirable, for he is one who is biased against the event being reported, so if they admit it happened, there is a good chance it did. And guess what? Many of Caesar’s enemies refer to his crossing the Rubicon, as did many neutral observers. In contrast, we have no hostile or neutral sources for the resurrection of Jesus until over 100 years after the event, and 50 years after Christians had been spreading their stories far and wide.

Fourth, we have credible accounts from critical scholars of the period for Caesar’s crossing the Rubicon, including Seutonius, Appian, Cassius Dio, and Plutarch. In contrast, no scholar mentions the resurrection of Jesus until two or three centuries later, and even they do not quote multiple sources or show skill and accuracy and critical thought like the Roman historians do.

Fifth, there is an eyewitness account for the Rubicon crossing. We have Caesar’s own word for it, in The Civil War. In contrast, we have nothing written by Jesus nor any eyewitnesses.

The Rubicon is well-attested by the evidence. In contrast, the resurrection of Jesus fails miserably with regard to every type of evidence a historian wants.

And when we turn the Historical Method upon other miracle claims, we find much the same thing.

  1. See “Alexander the Quack Prophet” in Selected Satires of Lucian by Lionel Casson. []

Previous post:

Next post:

{ 15 comments… read them below or add one }

Bill Maher November 26, 2009 at 7:28 am

gj defending mine and Dr. Carrier’s discipline. People have a really shitty idea from the History Channel that history is some sort of guessing game when we have very standardized arguments and a rigorous, peer reviewed methodology.

  (Quote)

Edson November 26, 2009 at 9:13 am

I think miracles are much effective at strengthening the faith of the one experiencing a miracle. To this person, she won’t give a damn about whether a miracle had a natural expalnation or happened supernaturally. She prayed to God and God did a miracle. This is what matters to her.

  (Quote)

Mark November 26, 2009 at 10:01 am

So the First Rule of Historical Method is this: Don’t believe everything you read! Humans are notorious liars, eager exaggerators, and happy to believe anything that affirms their pre-existing beliefs.

Fine, then I shall not believe ANYTHING I read here.

Happy Thanksgiving Luke.

Peace and love,

Mark

  (Quote)

Bill Maher November 26, 2009 at 10:57 am

Edson: I think miracles are much effective at strengthening the faith of the one experiencing a miracle. To this person, she won’t give a damn about whether a miracle had a natural expalnation or happened supernaturally. She prayed to God and God did a miracle. This is what matters to her.  

True story: I used to have a friend that was a Jedi and thought could use the force. One day we were sitting down and I was trying to be a smart-ass and asked him to turn off the lights using his powers. He put out his hand and the lights went out. Does this authenticate the existence of the Force and Jedi powers?

  (Quote)

lukeprog November 26, 2009 at 11:31 am

What a lame use for Jedi powers!

  (Quote)

Bill Maher November 26, 2009 at 12:04 pm

lukeprog: What a lame use for Jedi powers!  

I concur, but I am way more of a Star Trek fan.

  (Quote)

Ben November 26, 2009 at 12:14 pm

Those are interesting stories we rarely hear about.

  (Quote)

John D November 26, 2009 at 1:01 pm

Some random thoughts on methodology:

I notice that Carrier is mooting a Bayesian approach to historical methodology these days. He seems to be critical of the use of ‘historicity criteria’, which is what he argues for in Sense and Goodness.

http://www.richardcarrier.info/CarrierDec08.pdf

I’m working my way through Eliot Sober’s Evidence and Evolution at the moment. He argues for a Bayesian approach to assessing the merits of evolutionary theory. It’s quite intense.

I also note that Gregory Dawes rejects Bayesianism in Theism and Explanation (I’m basing that on the interview he did with Luke, not on actually reading it) in favour of a criterial approach.

Any thoughts on the respective merits? Some of the figures put into Bayes theorem when assessing the truth of historical propositions seem distinctly made-up to me. It seems more like a way to quantify your biases, which isn’t necessarily a bad thing.

I must confess: statistics is slightly outside my comfort zone, although I am trying to learn.

  (Quote)

Ben November 26, 2009 at 1:17 pm

John D: I notice that Carrier is mooting a Bayesian approach to historical methodology these days. He seems to be critical of the use of ‘historicity criteria’, which is what he argues for in Sense and Goodness.

These aren’t mutually exclusive. Bayesian reasoning models correct historical reasoning and merely entails more sophisticated criteria. It doesn’t negate the criteria already laid out. At least, that’s what I gather from his most current writings.

Ben

  (Quote)

Silver Bullet November 26, 2009 at 2:22 pm

Luke,

You and others indicate that the gospels were not authored by eyewitnesses. How do you know that?

Best,
SB

  (Quote)

Bill Maher November 26, 2009 at 2:34 pm

Silver Bullet: Luke,You and others indicate that the gospels were not authored by eyewitnesses. How do you know that?
Best,
SB  

Jesus was supposed to die around the year 30 and the historic consensus is that the four canonical gospels (Matthew, Luke, Luke, John) at the earliest were written after the 70′s. Ironically, the Pauline epistles are the earliest records there are of the early church, but these were not written till the 50′s and it is thought be many historians that they are falsely attributed to Saul of Tarsus.

If you want a better answer than that, I can ask my Ancient Near East history professor. His specialty is Biblical text studies and human sacrifice.

  (Quote)

lukeprog November 26, 2009 at 3:48 pm

Silver Bullet,

Maybe I’ll write about that sometime, but for now just check the ‘Composition’ section of the Wikipedia page for each gospel for links to other sources on why this is the scholarly consensus.

  (Quote)

lukeprog November 26, 2009 at 4:00 pm

John D,

I suspect Carrier did not discuss Bayesianism in S&GWG because it would have been waaaaaay to technical. You can’t even talk about Bayesianism without giving a 15-page introduction. I think Bayesianism and “explanationism” are both useful tools, and both are the subject of highly unresolved debates in epistemology and philosophy of science.

  (Quote)

Silver Bullet November 26, 2009 at 7:18 pm

Bill Maher:
Jesus was supposed to die around the year 30 and the historic consensus is that the four canonical gospels (Matthew, Luke, Luke, John) at the earliest were written after the 70’s. Ironically, the Pauline epistles are the earliest records there are of the early church, but these were not written till the 50’s and it is thought be many historians that they are falsely attributed to Saul of Tarsus.
If you want a better answer than that, I can ask my Ancient Near East history professor. His specialty is Biblical text studies and human sacrifice.  

Yes – this is the common response: the gospels were written too long after Jesus to have been written by eyewitnesses. But 30 or 40 years is technically not long enough that all eyewitnesses must have been dead is it? Surely, something could have been written by an eyewitness of the battle of Dieppe in the year 1972 or 1982 no?!

  (Quote)

Kiwi Dave November 27, 2009 at 4:45 am

Silver Bullet: “But 30 or 40 years is technically not long enough that all eyewitnesses must have been dead is it? Surely, something could have been written by an eyewitness of the battle of Dieppe in the year 1972 or 1982 no?!”

True, but given life expectancies of the time, the illiteracy, the rarity of written media and, the immobility of the population, and perhaps the disruption brought about by the uprising against the Romans, we can’t expect too much challenging of false claims.

Nor need we wait a long time for myths to arise. How many days did it take for 9/11 truthers to come up with their alternative accounts, which apparently a sizeable chunk of the population believes, despite readily available debunkings?

Luke, you write: ‘The resurrection of a Jewish profit is not a physical-historical necessity…’ etc. True, but not what you meant to say.

  (Quote)

Leave a Comment