A behind-the-scenes look at how this site works — the way I read, argue, check, and decide what to trust — shown on one strange passage in 1 Kings 13.
The workshop, not the showroom. Instead of a finished argument about a passage, this shows how one gets built and tested here — so you can see why what we publish can be trusted. It follows one chapter from a cold read through the lexicon, the commentaries, a study Bible's archaeology, and an argument with an AI that included a blinded skeptic, to the decision that the passage deserved its own Study. It is an illustration of the process — a true-to-life worked example, told lightly, not a verbatim transcript.
Written from an orthodox Christian confession; I read Scripture as inerrant in its Truth rather than in every detail of its composition. I came to this chapter already sure one verse read strangely — and a man who is sure he hears something will find a reason it is there. Catching that is what the later steps are for.
A finished Bible study shows you a verdict and the evidence for it, cleaned up. This is the other thing — the bench it was built on. I lay my own raw observations bare here, the kind that normally never leave the notebook, for one reason: so a reader can see that there is a method, and that the method — not my say-so, and not an AI's fluency — is why the content can be trusted. There is a process; here is an example of it; that is the whole argument of this page.
The workflow has a fixed shape. Read the chapter naively and record the itch. Take it to Strong's and BDB. Read the commentaries — separately, cited, and examined rather than assumed. Pull in the background a study Bible offers, history and archaeology, without mistaking background for the text. Bring the conviction to Claude and argue with it instead of taking the first agreeable answer. Send the surviving claim to a cold, blinded reader told to find what's wrong. Then, only if the question earns it, go to the textual apparatus. Each stage is held loosely, on purpose, so a later step can overturn an earlier one — and several did.
The boxes in this purple are the fourth wall — me stepping out to name the move I'm making and why. Everything else is the example happening. And since this is an illustration: the moves are real and so is the use of Claude, but the specifics are a worked example told lightly, not a court transcript.
I had read 1 and 2 Samuel several times and moved into 1 Kings, and I came to this chapter cold. A man of God comes from Judah "by the word of the LORD" to Bethel, where King Jeroboam stands at the altar. He cries not against the king but against the altar, foretelling that a son of David named Josiah will one day burn human bones on it; he gives a sign, the king's hand withers and is healed at his prayer, the altar splits, and he refuses the king's table on a direct command not to eat, drink, or return the way he came (vv. 1–10). Then an old prophet of Bethel rides after him, lies that an angel has lifted the command, and brings him back; at the table the true word of the LORD comes through the deceiver, sentencing the man for disobedience; a lion kills him on the road but stands over the body without eating it or touching the donkey; the old prophet buries him and asks to share his grave (vv. 11–32). The word lands three centuries on, when Josiah burns bones on that altar (2 Kgs 23:15–18).
Before I trust any observation, I check my retelling against the text, because a paraphrase is where errors sneak in. One I caught on myself: there is no second visit — the oracle, the sign, the withered hand, and the split altar are one continuous scene, not two. Getting the plain sequence right is step one, not a footnote.
These are observations, not conclusions — what the text does to a reader, recorded before any framework is laid over them. The man of God addresses the altar, not the king standing right there. Both central figures are introduced with the indefinite article and never named. The old prophet lies, and the text gives no reason for it. The true word then arrives through that same liar. The lion kills but neither feeds nor mauls the donkey, and the two animals stand together by the body. The deceiver wants to be buried beside the man he deceived to death. My questions out of that: Why did the prophet lie? Why was the man of God judged when he was deceived? What is the lion-and-donkey tableau for? And — the one that nagged most — what to make of the sign given "the same day."
One thread I chased and dropped, kept here on purpose. The shared grave reminded me of Saul's bones (1 Sam 31; 2 Sam 21). I checked: no literary link — different actors, different vocabulary, no allusion. A curiosity tested and discarded is part of the method, not a failure of it. Most of what a study touches ends up on the floor; showing the floor is the point of a piece like this.
Before commentary, the words. I use Strong's and Blue Letter Bible to locate a term, then ground its actual range in a real lexicon (Brown-Driver-Briggs), because a concordance number tells you where a word is, not what it means. The chapter calls the Judahite a "man of God" throughout and the Bethelite a "prophet" — a deliberate-looking pattern worth confirming by direct count. The "sign" is a môpēt, a portent or wonder, more than a bare token. And "bones" is the chapter's hinge: bones threatened on the altar in verse 2, bones spared at the grave in 2 Kings 23:18.
I read commentaries the way you'd canvass witnesses: separately, attributed, and without granting any of them the last word. The older voices I keep — Matthew Henry, Fausset & Brown, Trapp, Clarke, Barnes — and the modern ones I reach for through open scholarship and Enduring Word. The useful thing is that they don't converge into a single answer. They split on the old prophet's motive (Clarke is charitable, Henry and Fausset harsh) and on whether the chapter is one crafted narrative or a layered composition; one is openly troubled by the unfairness, another lands a brisk moral. None of that is binding. They are evidence to be weighed, and weighing them — not assuming them — is the discipline. Kept separate and cited, the canvass looks like this:
"If this were a true prophet, he was a bad man." Reads the lion standing guard over the untouched body and donkey as a deliberate attesting sign.
The man cries to the altar "out of detestation," having no hope of the king; Jeroboam's heart was "as hard as his hand withered."
Unusually kind to the old prophet — curiosity, not malice — and openly troubled about why the man of God was punished at all. The commentator who lets you feel the fairness problem rather than smoothing it.
Raises whether "Josiah by name" could be a later gloss (the 2 Kings parallel never mentions the name), and frames the sign as an early instance of a near token guaranteeing a remote prophecy.
The anonymous man of God is still significantly used; his failure was one of discernment — accepting a new claim that contradicted a clear prior word; the deception "is never explained and should not be harmonized away"; God judges his own more strictly (1 Pet 4:17); "the lion seems more obedient than the man of God."
A compositionally complex unit under Deuteronomistic redaction that preserves older prophetic material; the chapter tests whether a direct divine word can be overridden by a later prophetic claim; the lion as calibrated "judgment-without-contempt." (Exact wording owed from the volume.)
Reads the chapter as built around "the word of the LORD" (debar YHWH), recurring roughly ten times — nearly once every three verses; the liar ends as the genuine witness who honors the dead man's vocation. (Wording owed; the count corroborated by direct count.)
My physical study Bibles add a layer the commentaries mostly don't — the physical world. The NIV's notes point to Jeroboam's calf-shrine at Dan, to how such altars were built, to Canaanite human sacrifice; that last detail quietly sharpens verse 2, where the threat to burn human bones on the altar is the worst defilement it could suffer. This is genuinely useful, interesting background, and worth having. But background is not the text, and it gets read just as critically as everything else: the Dan material describes the twin shrine Jeroboam built (1 Kgs 12:29), not the Bethel altar this chapter is aimed at. The spade at Tel Dan — a stone "high place" of Jeroboam's own century, with a horned-altar fragment — lights up the world of the story, not the event in it.1 A study Bible can hand you something valuable; the discipline is to read even it, and even the commentaries, critically rather than assume them.
Some of the modern commentary first reached me as text an AI had generated in an earlier chat and handed back as raw material. My standing rule then fired: naming a source is a citation, and a citation needs real evidence, quoted correctly — the bar is the same whether the claim comes from a book or from a model. So the unverified summaries stayed flagged until checked against the actual sources. The point of the rule isn't bookkeeping; it's that every claim on a published page traces back to something real.
Now the verse that nagged from the start — the clearest example on this page of friction running in both directions on one text. Read in English, verse 3 lands oddly: "He gave a sign the same day." To my ear that read like later, that same day, as if a beat of time had passed. I wasn't claiming a second visit; I just kept insisting the half-verse felt marked.
Claude pushed back first, and rightly. "On that day" (bayyôm hahûʾ) is a common idiom for "the day already in view" — it carries the prophets' future "in that day" and, in narrative, "that same day" (Gen 15:18) — so it does not, on its own, mean "later." What is actually marked is the verb, not the time-word, and that is a discourse choice, not a clock reading; the commentators, read on this verse, all take it as a same-day sign for a far-off prophecy, with no gap. My grammatical hunch was overreaching, and Claude said so.
I did not take the dismissal. I pushed back two or three times — that the half-verse still read strange to me, that "one day" needn't mean "immediately," and I asked where else the phrase occurs — and then, suspicious that Claude was starting to simply agree with me, I asked it to hand the whole question to a cold, blinded reader told to find what was wrong.
This is the friction the page is really about, and on this verse it ran both directions. Claude pushed back on me; I pushed back on that; and rather than let the exchange settle into agreement either way, I sent the question to a cold, blinded reader — given the Hebrew, not the conversation, and told to find what's wrong. Its verdict on the narrow point went against me: the time-gap reading isn't defensible from the grammar. But it took the back-and-forth to get there — and the back-and-forth is what turned up what came next.
Because the question had been pushed that hard, the skeptic surfaced something neither of us had: the marked verb sits exactly where the Greek text of this chapter diverges from the Hebrew — verse 3 reads differently in the Septuagint, in both the verb and the noun. That is a question about how the text was assembled, not about narrated time, and it sent me to the apparatus, which confirmed the divergence lands right on the half-verse that had bothered me. My specific reading was wrong; the instinct that something was off was not.
My ear was right that something is off, and wrong about what. The verse does not narrate a later moment — the grammar rules that out. But the feeling wasn't a phantom; it tracked a real feature, a marked verb on the single most unstable half-verse in the chapter. The reading wasn't discarded; it was relocated — from chronology to composition. That relocation — reached only because the verse was pushed from both sides and then handed to a skeptic — is the whole value of the procedure.
A method should produce something, so here is the honest ledger of what this one yielded. Answered: the man-of-God-versus-prophet wording is a real pattern; the "same day" oddity is genuine but compositional, not chronological; the Dan/Bethel question is clarified; the lion and donkey are a controlled sign, not an accident. Left open: whether the man's death was fair, which the text names as disobedience but never calls just; why the old prophet lied, which the text simply does not say; and how a word certain three centuries out fits with agents the text holds genuinely responsible.
Then the decision a method exists to serve: is this interesting enough to warrant its own Study? My judgment was yes — the live question isn't the grammar but the providence and the discernment, a man judged for a sin he was deceived into, and a purpose that runs through the failure without excusing it. That is contestable and rich enough to argue properly. So it became a Study, where the reading is made carefully and — unlike this companion — graded.
The True Word in a False Mouth — providence and discernment in 1 Kings 13: the man of God, the deceiver, and the word that ran true through a false mouth. The argued reading this companion only points toward.
If a reader takes one thing from this, let it be the shape of the thing, not the verse. A conviction was brought and held loosely; it was checked against the lexicon, weighed against commentaries examined rather than assumed, enriched by background read critically, and argued with an AI deliberately turned into a skeptic. The friction ran both ways, and that is the part that matters. Twice the rebuttal won against what I first believed: my reading of the verse-3 grammar did not survive the blinded skeptic, and Claude reined me in when I pushed a historical claim further than the evidence allowed. But the traffic was not one-way — I moved Claude too. Its first instinct was to file the chapter under a tidy lesson about discernment, and I pushed back, more than once, that this was too small for what the passage actually asks; it conceded, and that harder question — not the grammar — became the spine of the Study this produced. Neither of us simply deferred to the other. That is the point of showing the bench: not that any one reading is certain, but that a published page here has been pulled at from both sides, and that the discipline, not the fluency, is what earns the trust.
Marginalia like this is not graded; the Studies it produces are. A finished Study gets a blind reasoning review — scored for the strength of its claim, its evidence, its honesty about what it can't show — and, on contestable points, a check by a different AI model or a human, precisely so a warm, agreeable read can't pass unchecked. That review is the last reason the content can be trusted, and it is the natural end of the process this page has been illustrating.
The bias, named once more where it lands: I am drawn to passages that resist resolution, and a reader who wants Scripture to yield a clean rule each time would press harder than I did for the tidy lesson. The discipline cuts both ways — the aim was never to manufacture mystery, only to refuse a certainty I hadn't earned. Where the evidence went against my first reading, I have tried to say so plainly, which is the only thing that makes the rest worth trusting.
Sources
The Holy Bible, New Revised Standard Version Updated Edition (the site default), with the Hebrew consulted where a word is discussed. 1 Kings 12:28–31; 13:1–34; Genesis 15:18; 2 Kings 23:15–18.
Primary source.
1 Kings 13 commentary: Matthew Henry; Jamieson, Fausset & Brown; John Trapp; Adam Clarke; Barnes (public domain); David Guzik, Enduring Word (verified directly).
Position: Devotional to mildly critical / pastoral-expository. Canvassed as witnesses; none adopted as authority.
NIV (Cultural Backgrounds) study notes; Tel Dan excavation reports (A. Biran), Madain Project, J. C. H. Laughlin (BAR 1981); Septuagint (Brenton, 3 Kingdoms 13).
Position: Study apparatus, archaeology, and textual witness — background read critically, cited for the world of the text, not the historicity of any narrated event.
The True Word in a False Mouth — the argued Study on 1 Kings 13 that this process produced.
The old prophet "was not really a godly man"; God judges his own more strictly than the wicked — "those have not learned self-denial who cannot forbear one forbidden meal."