Talk:Wiles's proof of Fermat's Last Theorem

Higher detail
I hope that we can attain a high level of technical detail. We are not providing a "howto" of the proof (which entire books have been devoted to), but we should be able to catalog many of the previous subjects and the results of other mathematicians that Wiles drew upon. We should be able to provide more detail than say, what the NOVA program "The Proof" provided. I would like to provide a key for the non-mathematicians college graduate so that they have a chance of drilling down through the Wiles proof. I hope to provide a key of notations that Wiles uses without introduction so that the reader can identify the subject area being referred to and read up. This is not a "howto", it is merely a review of the subjects Wiles covers in his discussion so that we can play to Wikipedia's strength: some depth of coverage in topics and the convenience of hyperlinks. Ideally, this will help to enthusiastic college student to gather the background information needed to understand the Wiles paper and not feel at a loss just because the notation, while standard within the specialty, it not obvious to the layman.--Lagelspeil (talk) 03:42, 16 March 2009 (UTC)

Wiles's?
Shouldn't it be "Wiles's" or "Wiles'" or "The Wiles proof"? Staecker (talk) 22:43, 19 March 2009 (UTC)
 * You are correct. I will do the rename as figure out which one to use.--75.25.136.186 (talk) 10:53, 25 March 2009 (UTC)


 * Either Wiles' or Wiles's is acceptable. Manual of Style says "For a singular noun ending in one s, there are two widely accepted forms... Either of those forms may be acceptable in Wikipedia articles, as long as consistency is maintained within a given article." At the moment, the article text uses both forms in different places. Gandalf61 (talk) 12:55, 25 March 2009 (UTC)
 * I agree on the additon of an apostrophe to Wiles. If Bob Smith had solved the problem we wouldn't call it "Smith proof of Fermat's Last Theorem", though it could be "The Smith proof of Fermat's Last Theorem". I'd stick with an apostrophe after the name. Alansohn (talk) 13:06, 26 March 2009 (UTC)
 * I updated the text to match the title. It had been consistently Wiles's . --Lagelspeil (talk) 08:24, 27 March 2009 (UTC)

Paragraph 1 on page 1 of Strunk and White ("The Elements of Style," any edition) is unequivocal: It should be Wiles's. One can't go wrong by following Strunk and White. Jay Janzen, Ph. D. (talk) 15:29, 25 September 2013 (UTC)

Proof of Theorem
Is there actually a site with the complete proof? If so, could you add it to the external links? Math Champion (talk) 01:04, 24 April 2009 (UTC)


 * This appears to be Wiles's paper and this appears to be a more-accessible refinement aimed at being taught as a graduate course in number theory and/or "arithmetical geometry". But I haven't read thru them enough to be sure. 134.114.148.66 (talk) 23:09, 31 March 2017 (UTC)

p vs n
Why is p used as the integer greater than two, instead of n? n seems to be the symbol used in other articles, such as Fermat's Last Theorem, and is usually used for arbitrary integers. --90.149.188.185 (talk) 22:28, 5 July 2009 (UTC)
 * I agree. p is also usually used for prime integers, I believe. I edited it, hope people agree. --Spug (talk) 23:14, 17 July 2009 (UTC)

Full proof
Someone need to write the full proof the of theorem here. 128.189.212.115 (talk) 02:36, 7 December 2012 (UTC)


 * Yes, if only it were helpful. At present, unlike the statement of the theorem, the full proof of the theorem is a large subject. To explain further, as the article does: It seems one or two handfuls of people have learned it fully. Any reader of the proof needs considerable preparation. The published proof itself is 108 pages long. Some simplifications are known. No tremendous simplifications are known. You and I and all of us are the editors of Wikipedia. (Actually, I for one have done nothing yet for this particular article.) We have achieved an outline of the subject, a sketch of its logic, and an indication of how people approach the details. We have links and citations to the published proof. -Minopret (talk) 14:44, 12 January 2014 (UTC)

This is a Wikipedia talk page, not a forum
A Wikipedia talk page is for discussions relating to editing the corresponding Wikipedia article, not for publishing supposed proofs of theorems, for linking to such supposed proofs on other web sites, or anything similar. An editor from Vietnam has recently been using several talk pages of articles to publicise attempts at mathematical proofs. To that editor, please don't. There are plenty of internet forums, blogs, etc where the kind of thing you have been doing would be accepted, but a Wikipedia talk page is not one of them. Continuing in the same way may lead to the range of IP addresses that you use being blocked from editing. That would not be a disaster, since the majority of the other editing from that range is vandalism anyway, but it would be a pity, since there have occasionally been constructive edits from that range, and they would be blocked along with the unconstructive edits. The editor who uses the pseudonym "JamesBWatson" (talk) 14:22, 28 February 2014 (UTC)

September or October
The "Announcement and final proof"-part seems to imply that it was released in October. The site also seems to say so. –St.nerol (talk) 19:38, 27 April 2014 (UTC)

The lowest degree of Fermat's Last Theorem with a Frey curve
Ribet proved the non-modularity of Frey's curve for the n &ge; 5 case of Fermat's Last Theorem. The "Overview" section claims that Frey dealt with every n &gt; 2. I don't have access to Frey's paper so I do not know whether this claim is incorrect, but given Ribet's result it seems so. Can someone source-check it? A similar situation exists in the main article for Fermat's Last Theorem. Nxavar (talk) 09:09, 28 January 2016 (UTC)

R=T?
In the article I read
 * Wiles's goal was to verify that the map $$ R \rightarrow T$$ is an isomorphism and ultimately that $$ R = T$$.

I don't see the difference between having to isomorphic rings R and T and showing R=T? Mathematically I can't tell two isomorphic rings apart. Or does this mean that the isomorphsim is in some sense canonical? --Jobu0101 (talk) 07:45, 19 March 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 6 one external links on Wiles' proof of Fermat's Last Theorem. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive http://web.archive.org/web/20090228225556/http://www.mcs.csuhayward.edu:80/~malek/Mathlinks/Lasttheorem.html to http://www.mcs.csuhayward.edu/~malek/Mathlinks/Lasttheorem.html
 * Added archive https://web.archive.org/20011127181043/http://www.math.harvard.edu:80/~rtaylor/hecke.ps to http://www.math.harvard.edu/~rtaylor/hecke.ps
 * Added archive http://web.archive.org/web/20081210102243/http://cgd.best.vwh.net:80/home/flt/flt08.htm to http://cgd.best.vwh.net/home/flt/flt08.htm
 * Added archive https://web.archive.org/20040803221632/http://cgd.best.vwh.net:80/home/flt/flt01.htm to http://cgd.best.vwh.net/home/flt/flt01.htm
 * Added archive http://web.archive.org/web/20110510130013/http://www.simonsingh.net:80/books/fermats-last-theorem/the-whole-story/ to http://www.simonsingh.net/books/fermats-last-theorem/the-whole-story/
 * Added archive http://web.archive.org/web/20090511230730/http://www.maa.org:80/reviews/fermatwiles.html to http://www.maa.org/reviews/fermatwiles.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.—cyberbot II  Talk to my owner :Online 03:45, 22 March 2016 (UTC)

Bad link or page in reference note
There is a inserted reference (currently auto-numbered 29) after the words Fermat famously in the section titled Popular accessibility that does not seem to have anything to do with any type of statement from Fermat. It links to a book page that is titled WILES' THEOREM AND THE ARITHMETIC OF ELLIPIC CURVES. Without reading the subsequent pages of the chapter it is not possible to see how this reference support the phrase Fermat famously. Perhaps the page number is wrong or perhaps the reference is just completely misplaced in the article.  Nyth 63  12:53, 23 March 2016 (UTC)

Length of proof
The lede states that the proof is more than 150 pages but the section titled Reading and notation guide states that it is more than 100 page. While not mutually exclusive, it does seem contradictory. I realize that document page counts for different printings can change with font, line spacing, and margin sizes, but that is a pretty large swing.  Nyth 63  13:04, 23 March 2016 (UTC)

Summary of Wiles' proof
In the section "Summary of Wiles' proof" are the "Comments" direct quotes? If not they should not use contractions, as per Contractions. You may also want to revise the improper use of the pronouns "we" and "our" per Manual of Style. 5.151.0.121 (talk) 00:09, 21 January 2018 (UTC)

"with all numbers rational"
shouldn't this say "with all numbers integer"? — Preceding unsigned comment added by 77.2.91.57 (talk) 12:10, 13 June 2019 (UTC)


 * Same thing since the equation is homogeneous : if $$(a_1/b_1)^n = (a_2/b_2)^n + (a_3/b_3)^n$$ is a rational solution then $$(a_1b_2b_3)^n = (b_1a_2b_3)^n + (b_1b_2a_3)^n $$ is an integral solution. jraimbau (talk) 11:10, 14 June 2019 (UTC)

Fermat's Last Equation
Thought you'd never ask. Concerning the set of all positive cubic numbers, when two cubic numbers have been added together, has long been proven. And recalling that any set containing zero when multiplying any set higher than the cubes of three equal zero, it seems likely that demostration is what Fermat had in mind when he wrote that his margin was too small to contain it. What do you think? 166.181.249.95 (talk) 08:31, 17 February 2022 (UTC)

Described proof structure is not a proof by contradiction
The article describes the following top-level structure of Wiles' proof of FLT:


 * 1) Assume that there are positive integers a, b, c, and n > 2, such that an + bn = cn.
 * 2) Derive a contradiction (absurd).
 * 3) Conclude that there are no such positive integers (FLT).

Despite the fact that the proof involves a contradiction, this is not what a "proof by contradiction" is. Rather, these kind of arguments are known as "proof of negation":


 * 1) Assume A.
 * 2) Derive a contradiction.
 * 3) Conclude ¬A.

(In the case of FLT, the proposition A is ∃ a, b, c, n ∈ ℕ: n > 2 ∧ an + bn = cn.)

Instead, a "proof by contradiction" has the following shape:


 * 1) Assume ¬A.
 * 2) Derive a contradiction.
 * 3) Conclude A.

The difference might seem negligible, but it actually isn't since, e.g., the law of excluded middle (LEM), can be derived using the principle of proof by contradiction, but not using the principle of proof of negation (of course, assuming a base system in which LEM is not derivable by other means). In other words, "proof of negation" is intuitionistically valid, while "proof by contradiction" isn't, hence the two principles are nonequivalent from a metamathematical point of view.

Importantly, this remark applies to the top-level structure of the proof, and Wiles' proof may very well rely in proof by contradiction in an essential way, but it does not at this point. — Preceding unsigned comment added by 46.10.124.136 (talk) 09:53, 13 June 2022 (UTC)

ZFC and Grothendieck universe
My understanding is that Wiles's proof relied implicitly on the existence of a Grothendieck universe, which is not provable within ZFC. Recently, a work-around has been found that does not rely on the existence of a Grothendieck universe, i.e. the proof now lies within ZFC. I think including a treatment of this issue would improve this article.

I have a couple of references that supposedly support this, but I have not examined them so I'm reluctant to add this material myself.

Mr. Swordfish (talk) 14:48, 17 May 2023 (UTC)

Repetition in summary
The summary assumes FLT is false twice. A semistable elliptic curve arising from the counterexample also seems to be introduced twice, both in 2 and 3. I don't know anything about algebraic number theory so I'm not confident in making these changes myself. --Caliburn &middot; (Talk &middot; Contribs &middot; CentralAuth &middot; Log) 20:15, 1 June 2023 (UTC)