Nine Draft Rule Proposals for Reading the Bible, Drawn from Published Hermeneutics
On this page (11 sections)
Draft proposals only. These are not accepted rules. The author makes every decision. Nothing here is final, and nothing here is a proof.
The short version
The author asked a plain question: can we get rules for reading a word or a phrase in the Bible from the Bible itself, and not only from the say-so of scholars and traditions? Hermeneutics is the study of how to interpret the Bible. The author read the published research on it, then checked what he found against the Bible's own text. Nine draft rule proposals came out. They have been attacked twice on purpose and are now in a third draft. None is accepted, and none has been through the full testing yet. They are proposals only, drawn from known published work on hermeneutics. This post tells where they came from, what each one says, and what is still open.
Why ask this at all
Hermeneutics, as the churches and scholars have taught it, has been used and tested for hundreds of years. Much of it rests on accepted authority. The author does not say it is wrong. This project turns the question around and asks one different thing of each published principle: does the Bible's own use show it?
The author expects most results to line up with the usual methods. That is an expectation, not a result. Some results may differ, and if they do, this series will say so plainly. The author is not a trained Bible scholar, and the project does not claim to be one.
The test
A rule stands only if one of two things is true. Either the Bible's own use shows it, with named examples. Or denying it would make two Bible texts contradict each other. (The Bible is without error, so it cannot contradict itself.) No rule may rest on silence. And every rule must say what one example would make it lose. The author decides each one, but not casually. A testing process still has to be run on each proposal first, and the decision comes after the testing.
The search
First, the project read 27 published sources on how readers say a word or phrase should be read: books, encyclopedias and articles from several traditions. Then it asked four AI assistants (Claude, Meta AI, Gemini and Perplexity) to list the principles they know. A fifth, DeepSeek, was not run.
That made a list of 320 principles: 240 from the sources and 80 from the AI answers. Many said the same thing in different words. With the repeats taken out, 282 different principles were left.
The sources and the AI answers also pointed to Bible passages as examples. Taken together they named 351 separate passages.
The check
Then each of the 282 was checked against the Bible text itself. 350 of those 351 passages were printed from a local copy of the NET Bible, so the words are exact. Everything was then sorted into groups.
Of the 282, 13 could be shown by the Bible's own examples. Another 31 were not about how to read at all. They were advice, preferences or beliefs. And 238 were not shown, because the passages named for them did not show the principle.
Please hear this part clearly. "Not shown" does not mean "wrong." Many of those 238 are well-tested methods. This one test cannot yet vouch for them. That is all it means.
From 13 to nine
The sorted list went to the project's designer, who drew up a draft list. From the 13 came nine draft rule proposals and three small additions to rules the project already had. Then two separate skeptic reviews attacked the draft on purpose. A skeptic here is a reader whose only job is to find holes. The second review came after the first round of repairs, and the draft was reviewed in this way before it became the simple list below.
The first review found 31 problems. Two of the nine failed their own "what would make it lose" test, and they were cut back. The second found 20 more, including problems in the repairs. The reviews were part of building the drafts. The proposals are now in a third draft. They are not final.
Where they stand today
These nine are not in the project's latest published rule list for AI readers, which is Beta 3, Version 3. Each will be tested rigorously first. Only after that will any of them be added to the project's Alpha or Beta lists, and only if the author decides to.
The nine, in plain words
All nine are draft proposals, not accepted rules. Each one below is a short version.
1. What a picture stands for. If the text says what a picture means, that is its meaning in that place. Revelation 1:20 says "seven stars are the angels" of the seven churches. The same picture can mean different things elsewhere. The lion in Genesis 49:9 ("a lion's cub, Judah") is not the lion in 1 Peter 5:8 ("like a roaring lion"). If the text does not say, the rule has no answer.
2. Chapter and verse numbers. They were added long after the Bible was written. A number does not mean the speaker stopped or a new topic began. Ephesians 3:1 starts "For this reason I, Paul" and then breaks off. Verse 14 picks it up again with "For this reason I kneel."
3. Paired lines in poetry. Hebrew poems often say a thing twice in different words, as in Psalm 51:2: "Wash away my wrongdoing. Cleanse me of my sin." The second line may name the same thing as the first. But the pairing alone does not prove it. The author thinks this one is closer to a caution than a rule.
4. Reporting words does not make them true. The Bible reports that the serpent said "you will not die" (Genesis 3:4). The report is accurate. The words are false. Whether reported words are true is settled by the Bible's own voice, such as God's verdict on Job's friends in Job 42:7: "what is right."
5. A question the speaker answers. In Romans 6:1-2 Paul asks whether we should keep on sinning so that grace grows. His own next words are "Absolutely not!" His answer is his position. The question is not.
6. A command or wish is not a report that it happened. "May peoples serve you" (Genesis 27:29) is a wish. The text has to say it happened. If it does not, the rule writes "not shown." That is not the same as "it did not happen."
7. How two covenants stand. A covenant is a binding agreement with God. This rule says to take what the Bible itself says about how two of them relate. It does not say to follow a chart a reader draws. Galatians 3:17 says the later law "does not cancel a covenant" made before it. Hebrews 8:13 says the new "makes the first obsolete." Where two statements seem to differ, the rule writes both and does not pick one.
8. A later use of an earlier text. Sometimes a later Bible writer quotes or retells an earlier passage and teaches from it. That use is the Bible's own reading of the part it uses. First the two texts must be matched by their words. Genesis 21:12 and Romans 9:7 share the words "through Isaac."
9. Passages about Jesus. In Luke 24:44 Jesus says "everything written about me" in the Law, the Prophets and the Psalms must come true. A particular passage counts as about him where the Bible itself says so. John 12:41 does this for Isaiah: "spoke about him." A note that something was "fulfilled" shows it came true. It does not, by itself, show the passage was about him.
What is still open
The author has not decided on any of the nine proposals, and will not until each has been through the testing process. Several questions are waiting.
The biggest is whose words count as the Bible speaking in its own voice. The Bible's narrator counts, and so do God and Jesus. But does a letter writer such as Paul count? Four of the nine have their main examples in letters, and they wait on that answer.
A second question is what shows that a speaker is speaking for God. The testing found that a title alone is not enough. In 1 Kings 13:18 an old prophet says "I too am a prophet" and then tells a lie, and the very same verse says so. In Jeremiah 28:2 the prophet Hananiah speaks in the Lord's name, and Jeremiah later says "did not send you" (28:15). If a title alone made the words the Bible's own voice, the Bible would say opposite things about the same words.
A third question is whether the project should run a third skeptic review before the testing begins.
Honest limits
This is one author's project. AI assistants did much of the reading and the checking, and they did it by following instructions the author gave them. People can be wrong, and so can AI, and either can be wrong when the method is wrong. So the results are the author's responsibility, and the project leans on good discipline and process: printed Bible text instead of memory, written tests, outside attacks on every draft, and testing before any decision. No result is proof, whoever produced it. A proposal that passes these checks has only passed these checks. Nothing here replaces reading the Bible for yourself.
What this means for you
You can try the same habits on any claim about a Bible word or phrase. Read the verses around it. Ask whether the Bible itself says what the word or picture means, and where. Ask whether the speaker is the Bible talking or a person the Bible only reports. And ask what one example would show the claim to be wrong. If the claim cannot answer that, hold it loosely.
Soli Deo Gloria.
Share this post:Share on Facebook

