This site
The original document
Everything on this site comes from one file. This page says what that file is, what was in it, what happened to each part of it, and what is deliberately not published.
What the source is
- Title
- My case for Conditional Immortality/Annihilation (CI) instead of Eternal Conscious Torment (ECT)
- Author
- Phil Welch
- Started
- Format and length
- A word processor document of about fifty-two pages
The file was written for people who asked the author what he now believed about hell and why. It circulated as a shared document, gathered comments from readers, and grew over the following years into a long argument with its own table of contents.
It is the source for every argument on this site. It is not the authority for any claim: where the document asserts something its evidence does not support, this site states the narrower claim and records the change in the changelog.
How the web edition reorganises and verifies it
The document was read out of its own file format rather than copied from the rendered page, so that anything present only in the underlying markup, such as a comment, a hyperlink relationship or an image, could not be quietly dropped. Every paragraph, table cell, image and link was given an identifier and an entry in a migration ledger.
The material was then reorganised into thirty-seven pages, each covering one argument, with a preface and two appendices alongside them, and each carrying the page numbers of the original it was drawn from. Scripture references were normalised and checked against a verified public-domain corpus. Sources that appeared as bare links in the document were given records with an access date, a link status and a rights status. Historical and language claims were checked against primary sources where those exist, and narrowed or withdrawn where they could not be supported.
The build refuses to complete if a substantive element from the source has no destination, if a Scripture reference cannot be resolved, or if the author’s personal contact details appear anywhere in the output. That is why the counts below can be published rather than estimated.
The source inventory
The checksum below identifies the exact file this site was built from. If the source is ever revised, the checksum changes and this page changes with it.
Source file SHA-256d3e567fd5a945efb2280d058787e808c86109d5ae848744beffafd45cf09187a
| Measure | Count | What it means |
|---|---|---|
| Paragraphs | 997 | Paragraph elements in the file, of which 904 carry text. |
| Tables | 3 | Rebuilt as semantic tables with real header cells. |
| Hyperlinks | 43 | Each given a record in the source library with its status. |
| Images | 8 | None is republished as an image. See the media table below. |
| Editorial comments | 30 | Comments and replies, grouped into 17 threads below. |
| Ledger entries | 1034 | One entry for every element considered, with its treatment. |
| Entries with a destination | 943 | Substantive elements that reached a published page. |
| Kept in the private record only | 2 | Contact details and personal reflections withheld for the reasons below. |
| Formatting only | 89 | Empty spacing paragraphs and interface debris with no content to carry. |
| Unmapped | 0 | Substantive elements with no destination. The build fails if this is not zero. |
Ledger generated July 29, 2026.
How the original wording is preserved
The migration ledger holds the source text of each element alongside the treatment it received: preserved, lightly edited, rewritten for clarity, summarised, relocated, qualified, moved to an appendix, kept in the private record, or omitted as formatting. Where a claim was narrowed or withdrawn, the ledger still holds what the document said.
That is what makes the changelog checkable rather than merely reassuring. A reader who wants to know exactly what was changed can compare the published wording with the recorded original, and the entries that did this are listed with their reasons in the changelog.
What happened to the editorial comments
The document carried 30 comments and replies from the author and from readers, grouped here into 17 threads. They are private editorial correspondence, so no commenter is named anywhere on this site and no comment is reproduced. What is published is the disposition: what was done about each thread, and why.
| Disposition | Threads |
|---|---|
| Incorporated into revised copy | 12 |
| Addressed in a caveat | 1 |
| Added as an open question | 2 |
| Added to internal research backlog | 1 |
| Preserved only in the private source archive | 1 |
| Total threads | 17 |
Where a comment led to a change in the argument, that change appears in the changelog on its own terms, described by what was wrong and what was done rather than by who said it.
What happened to the images
The file contained 8 embedded images. None is republished as an image. Where an image carried part of the argument, that argument was rebuilt as text, as a semantic table or as an accessible graphic, because a screenshot cannot be read by assistive technology, searched, translated or checked.
| File | What it showed | Treatment | Rights |
|---|---|---|---|
| image1.png | Table of four two-outcome games with their win and loss amounts, used in Appendix 1 to show how outcome size shapes assumed odds. | Rebuilt as semantic HTML See where it wentRebuilt as a semantic table with a caption and explicit column headers, with the zero-sum assumption stated in prose. The original conveyed comparisons partly through relative bar size, which carried no text alternative. | Cleared |
| image6.png | Chart of the presumed odds for each of the four games under a zero-sum assumption. | Rebuilt as an accessible SVG See where it wentRebuilt as an accessible SVG with text labels inside the graphic and an equivalent data table immediately beneath it. Nothing is distinguished by colour alone. | Cleared |
| image2.png | Screenshot comparing the Greek of Revelation 4:8 with Revelation 14:11. | Rebuilt as semantic HTML See where it wentReplaced with real text. Both verses are quoted in full from the public-domain corpus and the verbal parallel is described in words. Because the detailed comparison could not be reconstructed from the image with confidence, the page states the parallel only at the level the text supports and carries a specialist-review-pending status. | Paraphrased |
| image3.png | Second screenshot in the Revelation 4:8 and 14:11 Greek comparison. | Rebuilt as semantic HTML See where it wentHandled together with image2.png. An argument that exists only inside a screenshot cannot be checked by a reader, so it is not carried forward in that form. | Paraphrased |
| image8.jpg | Thumbnail for the 28-minute video overview. | Used as a poster image See where it wentThe author owns the video. A restrained text-and-type poster is used instead of the fire-themed thumbnail so that the site’s visual identity is not built around flame imagery. The original title is preserved in the video metadata and on the watch page. | Cleared |
| image5.png | QR code encoding the shortened URL for the shared working document. | Regenerated See where it wentA QR code is of no use on a page the reader is already viewing. A newly generated code encoding the canonical site URL appears only on the printable handout and print cover sheet. The original short link is retired. | No rights question |
| image4.png | Raster YouTube play button placed over the video thumbnail. | Replaced by the icon system See where it wentReplaced with an original inline SVG play control that scales cleanly, carries an accessible name, and is fully keyboard operable. | No rights question |
| image7.png | Small screenshot of the word processor’s document-outline toolbar icon, used in a note telling desktop readers how to open the outline pane. | Omitted as a formatting artefactInterface debris from the authoring tool rather than content. The underlying need, moving quickly between sections, is met by the chapter navigation, the on-this-page list and the case map. | No rights question |
What is deliberately not published
Two categories of material from the document are withheld, and both are recorded in the ledger so that the omission is visible rather than silent. Together they account for 2 ledger entries.
- Personal contact details. The document opens with the author’s email address and phone number and invites readers to comment in the file itself. Republishing either on a public site would expose them to automated harvesting for the rest of their working life, and it would place the burden of handling correspondence on one private inbox. A check on every build fails if either ever appears in the output. Corrections come through the form instead.
- Private editorial comments. The comment threads were written by named people in a private document, in the expectation that they were talking to the author and not to the public. Their substance is accounted for above; their wording and their authors are not published.
A small number of personal reflections in the document were also kept out of the public pages because they were speculation the author himself labelled as such. Those decisions are in the changelog with the reasoning, not hidden.
Why the file itself is not published
The original document is not available for download. It is not being withheld because its contents are embarrassing: the substance is on this site, and the places where it was corrected are listed publicly.
It is withheld because the file carries personal contact details, private comment threads with named authors, and long quotations from copyrighted Bible translations, and because a word processor file also carries revision metadata that is not visible when you read it. Publishing it would mean publishing all of that at once.
The file will be considered for release only once a rights and privacy audit has been completed and the material above has been dealt with. Until then the ledger, the changelog and the pages themselves are the public record. Printable versions of the case are on the downloads page.
If something here is wrong
If a page misrepresents the source document, drops something substantive from it, or narrows a claim further than the evidence requires, please say so through the corrections form. Point at the page and quote the wording you are questioning. Accepted corrections are published in the changelog with the issue and the decision.