When I decide to study a subject, I get a bit obsessive. Part of the reason for that is because I once was a Jesus myth theorist, and I fell for many of the conspiracies that surrounded that. I was burned before by subscribing to an idea that I was ignorant about. I don’t want that to be the case again. I don’t think Matthew Britt and Jaaron Wingo, the authors of “Christ Before Jesus,” take this same view.
This is the final part of my critique of Britt’s and Wingo’s methodology, and I think that after all of this, it should be clear that they should have spent more time researching stylometry. So let’s jump in and see how this plays out with some of their final concerns.
So now we finally get to the final concern they raise. “Could the stylometric analysis be picking up theology, genre, or theme instead of authorship?”
This is a question we’ve already discussed a bit, with the overwhelming evidence saying that yes, all of those things are real concerns. So how do Britt and Wingo defend against such a concern? They largely just brush it off.
When talking about theology or theme, they just claim that “the computer does not see this beyond the word choice of the authors.” They do admit that genre comes into play, but they push that off to just the broader levels. Even with them acknowledging that genre does affect the word choices we make, they still claim that there are some underlying fingerprints that are unique for each individual.
They bolster this claim by saying that the program they are using, Stylo, looks at the most frequent words, and theological words generally don’t come into play here. To be clear, Stylo is not unique in this manner. Looking at the most frequent words is a standard practice in stylometry, and yet, as shown in the studies in the previous articles, genre still makes a huge impact here.
Ignoring the fact we’ve already covered in regards to genre, how do theological words impact the text? Bruce Metzger has a great little book called “Lexical Aids for Students of New Testament Greek.” In it, he breaks down Greek words by how many times they are used. So for instance, the first category is words used more than 500 times. By breaking it down in this manner, those learning Greek can focus first on learning the words that are used the most.
Within that first section, two very clear theological terms appear: theo (God), and kurios (Lord). As a note, Metzger isn’t just looking at one specific form of the word theo. In Greek, there are 5 different cases (or 8, depending on how you separate them), and Metzger combines all of these cases for each word. So, for God, it’s not just theo, but the other cases as well, such as theos. Stylo treats all of these cases as separate words, and as per my email exchange with Britt and Wingo, they don’t do anything to compensate for this either.
With all of that in mind, we can actually observe how this works on a chapter level and see how the results can be skewed.
To start off, we can run an analysis with Stylo on the Pauline Epistles. For this I broke up each epistle by chapter. I included Hebrews along with the other letters attributed to Paul, as it was also at one time thought to have been by Paul. To be fair to Britt and Wingo, I used the default setting on Stylo, as they did, and only changed the distance measure to Eder’s Simple. The last portion doesn’t really factor in here, as we will just be looking at word frequency, but I wanted to keep it with what they had done.
What Stylo will do here, besides creating a graph, is that it will also provide a word list, as well as a frequency table. It’s that frequency table that we will be looking at.
When looking at the table, the first chapter that is listed is First Corinthians, chapter 1, so that’s what we will look at. And let’s just look at the term theou (the genitive form of theos) simply because it’s the predominant form in this chapter.
So how does this all stack up? Well, theou is tied for the fifth most used word in the chapter, at 12 uses. It is tied with the word de, which is a conjunction and often translated as either but or and. And to put this into perspective, de is the fifth most used word in the New Testament. So we can see that theou is well represented here.
In fact, based on that frequency table, we know that it is used 2.4 times for every 100 words in this chapter. There are only two other chapters in which the word is used at a higher frequency, and those are First Corinthians chapters 2 and 3. In chapter 2, it used 9 times, with a frequency of 3.1 times per 100 words, and is tied for fourth place with the word to, which happens to be the most common word in the New Testament, as it translates to the. In chapter 3, it is also used 9 times, with a frequency of 2.6 times per 100 words, and is tied for second place with the word gar, which means for, and is the 18th most used word in the New Testament.
However, there are also at least four chapters within Paul that the term is never used. In general, the amount of use is less than half of what it is in First Corinthians 1. To be clear though, theos, and it’s various cases, are the 14th most used word in the New Testament, so it’s common in general. But it is highly skewed here.
All of this should show that theological words can have an impact here. And this stands in direct contrast with what Britt and Wingo argue. They cite the Oxford English Corpus and highlight that the 10 most common words in English are: the, be, to, of, and, a ,in ,that, have, I. They then follow up that the 100 most common words in English account for some 50% of the words we generally use on a day-to-day basis, with theological words coming way down on that list. They argue that this is also true for Greek, and every other language.
Yet, in these three instances, we see that’s not true. And if we just look at the word God in the New Testament, it comes in at the 14th most used word, meaning that it’s very high up there.
But that’s just one word. What about other theological terms? Going back to 1 Corinthians chapter 1, the next two most used words, tying for 7th place with 10 uses each, are Iesou and Christou. In context here, both are theological words. So in this chapter alone, 3 of the most used terms are theological. If we look at chapters 2 and 3, theou is the only theological term that breaks the top 10. But still, these all show that theology plays a factor here.
Looking at the top 20, in First Corinthians chapter 1 we have kuriou (Lord) ranking at 16, with 6 uses, and then theos (God), ranking at 18, with 5 uses. We can also pick out a handful of other theological words that make the cut. At number 27, we have sophian (wisdom). 29 is adelphoi, which means brother, but here is used in a theological manner by saying they are brothers in the Lord. At 30 we have ebaptisa, which means baptized, but of interest here, it is only used 4 times in the New Testament, with three of those being in this chapter. And then we also have the related term, ebaptisthete, coming in at number 48, which means like, were you baptized. It’s a word that only occurs again, 4 times in the New Testament, and 2 of those are in this chapter.
Those are just the most clear uses of theological language within the top 50 words here. We could also mention Christo, which is number 40, but it’s a just a different case of Christou. Both mean Christ here. And if we go up to the top 100, we see more of this, with different cases of words we’ve already seen. And we could also do this with First Corinthians chapter 2, which within just the top 25, they have multiple forms of the word spirit and wisdom, as well as a second form of God. The same holds true with First Corinthians chapter 3 as well, with terms like Lord popping up.
It should be obvious that theology here, and theological words, have a massive impact on at least some of the chapters. But a bigger problem is exposed while running these chapter analyses, and then looking at both the word list and the frequency table.
First, as already mentioned, Greek has five cases. So you can have various words in different forms, but they mean the same thing. So you can find various forms of the word God, but they all mean God. To try to make this simple, let’s use the term theos as an example. Theos is the nominative case, which means that God would be the subject of the sentence. Theou would be the genitive case, which relates to possession, so like God’s love. The dative case would be theo, and is rather versatile, but is often used to express relationships between nouns. So for instance, in First Corinthians 1:4, it’s used for the expression, my God, or literally, the God of me. Accusative case would be theon, which would place God as the direct object, so like saying, your love of God, in which God is the direct object of that love. And finally, we have the vocative case, which would be thee, and is used mostly for directly addressing someone. In the New Testament, it’s only used twice, in Matthew 27:46, when Jesus cries out, My God (thee), My God (thee).
With it all laid out, it can seem a bit confusing, but we can see it at play in First Corinthians 1:21 really well, as we have three forms right there.
“For since, in the wisdom of God (tou Theou) the world did not know God (ton Theon) through wisdom, God (ho Theos) decided, through the foolishness of our proclamation, to save those who believe.”
As we can see, the term God doesn’t actually change, it’s just how it’s interacting with the terms around it. I also included a second word in Greek in each case, tou ton, ho. These are articles, like the English word the. Specifically, these are definite articles, and they are used to specify a noun. In the verse above, it specifies what god, the Judeo-Christian God.
When looking at Greek articles, they must agree with the noun in case, gender, and number. In our case here, it just means we will be using the singular, masculine version, and it’s just the case which will change which specific word we use, but in general, there are a handful of words that serve as the definite article (and then we have indefinite articles as well, which generally translate to the word a).
What this really means is that the same term is filling up multiple spots on these frequency lists. Going back to First Corinthians chapter 1, the most used word is tou. It’s the singular, genitive, neuter form of the definite article. The fourth most used word is ho, the singular, nominative, masculine form of the article. Number 13, 14, and 15 are ta, ti, and to, all forms of the article. Then at 19, we have tin, which along with ti, are singular, accusative, feminine forms of the article. Number 20 is ton, the plural, genitive, feminine form. And then, just filling out the top 100 words, we also have tus, ton, and tis. And as we saw in the verse above, we don’t even always translate these terms. They just tell us something about the noun. And yes, I included the word ton here twice, because it was listed twice. It’s due to accent marks, which will make sense in just a little while.
One could make a good case that we should combine those terms. But either way, they skew things a good deal. If we count up all of the uses of the definite article, and we will even round down, we get 70 occurrences. With chapter 1 only having 500 words, that’s more than 10% of the text.
Things then get more complicated because Greek also uses accent marks. These don’t change the meaning of the word, but instead indicate pitch. There are three different types of accents: the acute (’), the circumflex (~), and the grave(`). The problem is that Stylo sees them as different words.
For instance, in chapter 1, the 22nd most used word is gar, which means for. The 46th most used word is also gar. The only difference is the accent. If we combine both cases into one, then we get a total of 6 uses, which means it would be tied for 16th place. That pushes it up 6 spots. We see the same thing in chapter 3, with gar being 2 and 27, and if we combine them this time, it places them in the top spot.
Going through the frequency table, there are many instances of these issues. And when dealing with such small numbers, such as the word count in a single chapter, this has a great deal of impact.
But there is a second problem here as well that we find when looking at the frequency table. Looking at First Corinthians chapter 1 again, there are 246 unique words (or semi-unique as we saw above). But only 70 of those are used more than once. And only 40 are used more than 3 times. Only the top 8 words see double-digit use. That means 72 percent of the words in this chapter are unique; words that are used only once. A third of the words within the top 100 are unique. For chapter 2, only 47 words are used more than once, with just three hitting double digits, and chapter 3, there are 54 words used more than once. That means, for chapter 2, 100 are unique, while in chapter 3, there are 147 unique words. That’s nearing an average of 70 percent of the words being unique.
To try to build a profile of a style based on unique words simply doesn’t work. But there’s another problem that presents itself here. While Britt and Wingo are technically running the 100 MFW, in practice, that isn’t happening. If only 47 words are being used more than once, that means everything after that, number 48 through to the end, are tied for last place. They can’t be considered a frequent word, as they are unique.
And sure, some chapters do better, a few even hit the 100 mark. The longest chapter in the Pauline Epistles, 1 Corinthians 15, gets up to 125 words used for than one. But then again, words 73 to 125 all tie for that place, as they all are used just twice. When we take into account that there are many of those words are repeats, as in they are just different forms of the same word (such as a handful being definite articles), or the accent marks are just different, that number of frequent words dwindles.
Taking all of this into consideration, and applying it to the example that Britt and Wingo cite, the passage of the Parable of the Sower that we can find in all of the Synoptic Gospels, we find some serious problems.
Parable of the Sower
The Parable of the Sower is found in Mark 4:13-20, Luke 8:11-15, and Matthew 13:18-23, at least according to Britt and Wingo. We will see why this is a problem later on. Before even jumping into the text though, a question already arises. Why did the authors decide to quote the verses from the American Standard Version?
If the ASV had been the typical translation they had used throughout, there would be no problem. But, at the beginning of the book, they state that the New International Version would be the standard. That means, there was a purposeful shift here.
In introducing the passages, they say the ASV “provides a more direct word-for-word translation,” but there are some problems here. First, if we compare the ASV to the Greek, which Britt and Wingo had access to, we can see that each text in the ASV is around 30 words longer. So the passage in Matthew, in Greek is 128, in the ASV it is 160. In Mark, the Greek is 145, the ASV 178. For Luke, Greek is 109, the ASV is 144. To be fully inclusive here, the NIV contains 159 words for Matthew, 164 for Mark, and 145 for Luke.
Now, word count isn’t everything, but it does show that the ASV isn’t a direct word-for-word translation. There is still some interpretation that has to be done, which is fine. All translations have to navigate that a bit because of the nature of Greek.
There are other problems with the ASV though. Foremost, it’s outdated. It was published in 1901. Many of our best manuscripts used for translating the Bible date after that. Our understanding of Greek in general has greatly improved since then. It uses archaic words, and the English language in general has evolved since then.
And it’s not like one can’t find modern formal equivalent translations, the more word-for-word type of translations. Such as the updated version of the ASV, the New American Standard Bible. Now, while I personally like the New Revised Standard Version, which is a dynamic equivalence translation, which is more thought-for-thought, the NASB is at least decent compared to the ASV. And if you want that more word-for-word feel, it’s a much better choice.
So why did they pick the ASV? Who knows, but it is just one of those things that raises questions, because all that it does is add a bit more confusion and less reliability, because it’s not a good modern translation. It’s just one of those things that makes one think that either they are using the ASV in this case because it betters their argument, or it’s because they simply don’t know any better as they don’t have the relevant background.
Whatever the case may be, it does negatively impact their surface-level look. They start off by claiming that by just looking at the texts, we can see that the authors have different styles, for instance, Mark starts almost every other sentence with the word “and.” This is true in the ASV. But it’s not true in the NIV, their primary translation for their book. In the NIV, for this passage, no sentence begins with “and.” To be fair, both the NASB and NRSV do have around half of the sentences beginning with the word “and,” and this really follows the Greek here.
But does it make Mark stand out? Not at all. If we just look at the passage in Luke, every sentence, besides the first one, begins with the word “and.” With Matthew, it’s half of the sentences that begin with the word and. Mark doesn’t stand out here even on the surface level.
What if we dive a bit deeper though. As explained above, running Stylo on these passages just doesn’t work. All of the passages are less than two hundred words, so it’s far from a suitable length, even according to Britt and Wingo (remember, they stated Philemon was too short, yet it has more than a hundred words than any of these passages). But we can still get a frequency table and make some observations.
First, the idea that you can get the 100 most frequent words from any of these passages is simply false, regardless of what Britt and Wingo claim. For Mark, only 20 words are used more than once; Luke has 27 words used more than once; and Matthew comes in at 23 words used more than once. In addition, none of these passages even have 100 different words. Mark has 93 unique words, Luke has 75, and Matthew has 78. This means that even in the best case, with Luke, almost 2/3rds of the words are only used once, meaning they are all tied for last place; they aren’t frequent words in any sense.
Second, we can see that Mark does use the word “and” slightly more. Mark uses it 17 times, Luke uses it 14, and Matthew 12. This makes sense as Mark is the longest passage here. Statistically, both Mark and Luke use the word the same amount though; it makes up 10% of their word choice. Matthew uses it slightly less statistically, coming in at just 8% of their word choice.
However, for Luke, it’s his most-used term, with 14 uses. The second most used term by Luke is “the”, coming in at 11 uses. For both Matthew and Mark, this is flipped, with Mark using “the” 18 times and “and” 17 times, while Matthew uses “the” 17 times and “and” 12 times.
The term “and” here should not stick out for Mark. But what about the second word that Britt and Wingo highlight: “he?” They claim that Mark uses it once, Luke doesn’t, and Matthew uses it 10 times. When we look at Matthew, it is his third most word in this passage. It’s also true that Mark only uses it once, meaning it’s not a frequent word.
What about Luke though? Yes, in the verses that Britt and Wingo copy, the term “he” is not used. The problem is that they start at verse 11. This cuts off the introduction to the parable. An introduction that Mark has included. For Luke, that introduction starts off in verse 10, which states, “And he said, Unto you it is given to know the mysteries of the kingdom of God: but to the rest in parables; that seeing they may not see, and hearing they may not understand.”
For Luke, that is where the quote begins. In verse 9, the disciples ask Jesus what the parable means, and in verse 10, Jesus begins speaking, thus the phrase, “and he said.” This is how Mark also has Jesus begin his explanation.
Now, to be fair, Mark has the passage really begin in verse 11, with: “And he said unto them, Unto you is given the mystery of the kingdom of God: but unto them that are without, all things are done in parables: 12 that seeing they may see, and not perceive; and hearing they may hear, and not understand; lest haply they should turn again, and it should be forgiven them.”
So this would add a second “he” to Mark’s passage, but as we can see, the passages are very similar. In fact, if we compare the two passages in general, they are exceptionally similar. Especially when you include more in the selection, such as, the entire quote from Jesus that is being discussed here.
This raises a serious question though. Why does Matthew seem so different? The answer is simple, it’s a different passage. For both Mark and Luke, the passage quoted isn’t the parable of the sower, but the explanation of that parable. Both passages explain the parable. For Mark, the parable is quoted in verses 3-9. Following this, the disciples are alone with Jesus, and they ask him what they meant. We then get the passage that Britt and Wingo quote. The same is true with Luke. The actual parable is in verses 5-8, which is then followed with the disciples asking Jesus what it meant. We then get the passage from Britt and Wingo, which is missing verse 10.
For Matthew, we are getting the actual parable. It’s not until verse 36 that we get the disciples and Jesus leaving, and the disciples asking for an explanation. That explanation follows in verses 37-42:
“And he answered and said, He that soweth the good seed is the Son of man; 38 and the field is the world; and the good seed, these are the sons of the kingdom; and the tares are the sons of the evil one; 39 and the enemy that sowed them is the devil: and the harvest is the end of the world; and the reapers are angels. 40 As therefore the tares are gathered up and burned with fire; so shall it be in the end of the world. 41 The Son of man shall send forth his angels, and they shall gather out of his kingdom all things that cause stumbling, and them that do iniquity, 42 and shall cast them into the furnace of fire: there shall be the weeping and the gnashing of teeth. 43 Then shall the righteous shine forth as the sun in the kingdom of their Father. He that hath ears, let him hear.”
Now, this doesn’t change the stats too much. In this passage, there are 161 words compared to 160. “And” is still the second most used word, still coming in at 8% of total word use, even though “and” is used 13 times, instead of 12. The big thing here though is that instead of “he” being used 10 times, it drops to just 3 uses.
We also run into a separate issue here. To anyone reading the Matthew passage here, it is strikingly different from either Luke or Mark. Mark and Luke follow the same pattern in explaining each portion of the parable. Both state that what is being sown is the word. Luke clarifies that it’s the word of God, and makes it more direct by saying that the seed, which is being sown in the parable, is the word.
They both speak of the seed sown by the way side being those who, once they hear the word, the devil comes and takes it away from them so they aren’t saved. They move onto those sown in the rocky places, who have no root, and although they receive it with joy, once they face temptation (Luke) or tribulation (Mark), they fall away.
Next are the thorns, those who hear the word but because of the cares and riches of this world, they are unfruitful. And finally, those in the good ground, they heard the word, they accepted it, and they bear fruit.
The manner in which these flow is the same. There is some difference in word choice, but not a great deal. Most of this is composed of words used just once by both authors, yet they are the same words being used by both authors.
When we compare it to Matthew, it gets quite different. Right off the bat, we are told that the person sowing the seed is the Son of Man. The phrase, Son of Man, is not found in either Mark or Luke, nor do they name the sower. Matthew also states that the field is the world, another detail missing from the others.
Matthew also changes what the seed is. Mark and Luke say it’s the word of God. Matthew says the good seed are the sons of the kingdom, while the tares (a word we don’t see in either Mark or Luke, but refer to a type of weed) are the sons of the evil one. And while both Mark and Luke mention the devil, Matthew goes a step beyond and states that the enemy who sowed the tares is the devil personally.
Matthew then moves on to transforming this parable into an apocalyptic message. He has the end of the world, reapers being the angels, the tares being burned up in fire, again, at the end of the world. He has the Son of Man sending forth angels to sweep up all those things that cause stumbling, and this is then thrown into the furnace of fire, where there will be weeping and gnashing of teeth. After all of this, the righteous will then shine forth in the kingdom of their Father.
If we are to believe Britt and Wingo, all of this that we’ve discussed confirms what scholars agree on, that Matthew follows Mark more closely than Luke does. I think we can agree that that is a load of bunk. In this case, Luke clearly follows Mark more closely, and Matthew is way out in left field playing soccer.
With all of this though, their argument is simply invalid. Even when we ignore whatever reason they have for using the ASV here, virtually every argument they had here was off. Mark is not special with his use of the word “and” here. Luke used it to start his sentences at a higher rate. In the passage they quoted from Matthew, we also see half the sentences beginning with the word “and.” Now, in the actual passage they should have focused on, only 1 out of the 4 sentences started with “and,” but “and” is used nearly as much as in Luke and Mark. Matthew just has longer sentences.
The biggest thing that invalidates the arguments though is that they cut out part of Luke, and had the wrong passage in Matthew. Any result they got by running Stylo is irrelevant as the data is wrong. The best way to highlight this is to point out that when we look at the proper texts, Luke does use the word “he” 1 time, and Matthew only uses it 3 times, not 10. This is a massive oversight on their part.
Corpus Selection
Britt and Wingo finish off chapter 7 by discussing the actual corpus selection. They briefly detail the fact that they broke up each text into the chapter level. I’ve already discussed this and why it’s flawed from the get-go.
Moving on, they speak about how they selected what texts they would view. One thing they mention is that they didn’t usually use anything from before the first century BCE. This is okay, but the reason they give is that they claim they have no connection to Christian texts. However, Christian texts are based on Jewish texts, many of which are prior to the first century BCE.
Without the Hebrew Bible, as well as apocryphal Jewish texts, there would be no Christian texts. It’s why New Testament scholars have to have some background in the Old Testament. New Testament texts quote, or allude to the Old Testament somewhere around 900 times, depending on what source you refer to. That’s not including quotes or allusions to other Jewish texts.
To say there is no connection doesn’t make any sense. Now, maybe they don’t make the right type of connection or something along those lines. But if that’s the case, the issue is that this isn’t the first time they have made vague statements that could be interpreted in a number of ways, which essentially make their statements useless.
Quite possibly the biggest oversight though is just the fact that we aren’t told which version of the Greek texts that are being used. It was actually this question that first led me to email Britt and Wingo. The answer was, we used a variety of texts, but they had a preference for the Society of Biblical Literature Greek New Testament, and the Wescott-Hort, for ease of access.
Specifically, they recommended the WH version found at Tufts University Perseus Library. Personally, I prefer the Nestle-Aland, and I use both the 26th and 27th editions. It’s the ones I learned on, that were recommended while I was learning Greek. But the WH version is fine. The issue is that this should have been mentioned. If one is going to continually encourage others to test their results, this is basic information that should be given out.
As a whole, this chapter should give one a lot of reasons to have doubts about the whole process that Britt and Wingo propose. The available research, including that by the authors of Stylo, all show major flaws in the methodology that Britt and Wingo claim to have created after much study and experimentation. That methodology seems to be basically running Stylo on the default settings.
We’ve also seen some major mistakes, such as quoting the incorrect verse in Matthew when relating the Parable of the Sower, and failing to include the first verse in the Lukan version. Such a major misstep completely invalidated their argument.
There have also been a few examples that should raise suspicion about their understanding of Stylo, or stylometry in general. Not to mention a lot of vagueness in the whole ordeal.
Now, that doesn’t mean we can just disregard their findings, and from here, we will be diving into their results, as shown in chapter 8. It just means there are many reasons to question everything, as the foundation they are working on is shaky at best.
Up to this point, I had still wanted to continue giving them the benefit of the doubt though. I thought they were sincere, as it appeared that they were open to answering questions. But it did seem like they were way over their heads. All of this changed when I actually had a live discussion with Wingo, which changed my view of at least him for the negative. But we will get into that later.
Now, before we move on to their results, if we may just go back to the topic of peer review to bring this all full circle. Going back to the YouTube discussion with the Godless Engineer John Gleason, the idea of their work being peer reviewed was brought up. Britt and Wingo entertained the thought, and suggested that maybe in the future, that’s something they would look at. In its current form though, that simply won’t happen.
The why should be rather clear by now. “Christ Before Jesus” is not a proper research text. As already discussed, the vagueness and lack of discernible methodology are quite telling. There are few citations, end notes are scant, support for their claims is largely missing, and more often than not, what bolsters their argument isn’t facts or data but simple confidence.
Because of the lack of transparency, and the removal of key data, replicating and checking their work is nearly impossible. And with this portion of their work on such shaky ground, many of their other claims also begin tumbling down. As without the support of their stylometric analysis, many of their claims simply are baseless. And as we will see in their results, the evidence simply doesn’t support their conclusions.