I just want to say, Dave, that this article is probably my favorite timeline-related article. Ever. You illustrate ideas that I agree with 100% so eloquently and powerfully, and actually manage to back them up, something I have never been able to do. On top of that you wove a delicious trap for those who tried to argue against it. I am honored that you came to me with this.
A penny from you, LoS, is worth a hundred in the bank.

I'm honoured by your words, and you make me want to write more. I might have to "dual blog" more often.
TML: That's the exact reason I hated the split timeline, and I said it many times on this board. But after TP, I think it's the much more likely answer. I'm more interested in trying to find the theory with the least inconsistancies. That doesn't mean it's true, after all the creators could be making all sorts of huge inconsistancies and not even care, but I think from an objective viewpoint it's the most logical. I just cannot believe that WW preserves its connection to OOT so pristinely and clearly and yet TP looks like a blithering mess to both OOT and WW in a single timeline.
This actually brings up a set of rules that I came up with a long time ago. I actually was inspired by my neural networks class (Everything can be related back to Zelda. Everything.) to come up with this. The principle is called the Curse of Dimensionality. In Neural Networks, it goes like this:
As the number of considered dimensions (variables) within a neural network increases, the more possible an accurate representation of the answer become. However, as the number of considered dimensions within a neural network increases, the more trivial the representation of the answer becomes.
I have typically applied this premise to simply the "Multiple Link Theory" (now known as the as "Single Timeline Theory") as follows: The more Links you have in your theory, the more likely it is that you can make a workable timeline; however, as the number of Links increase, the more trivial and unrealistic your timeline becomes. In short, you could easily solve all timeline problems with thirteen Links (one for each game) and a lot of time to rebuild Hyrule whenever needed. However, such a timeline is inevitably bad because it's trivial; it doesn't truly seek to connect games whatsoever. Thus, the "best" timeline would be a Single Timeline, Single Link timeline, but we all know that trying to crap all 13 games into that construct is impossible, and if it really is possible, completely unrealistic.
My thoughts on the Split Timeline have been similar, except that adding additional timelines beyond the first increase the freedom and triviality much faster than adding additional Links. The exact ratio is something I've never sought to compute, but it has a hard upper bound (and in fact is much less than) 12:1, which is the number of Zelda games minus 1 to 1. (After all, a Split Timeline is more believable than 13 Links.) However, I have tried my best to stay away from the Split Timeline because of that very reason, that so long as I could fathom a Single Timeline candidate that worked within the bounds of my conceivable imagination, I would never convert. I hesitate to call the Split Timeline "cheating," but if it is "cheating," then it "cheats" much more than adding a single Link does.
Twilight does its damage to Single Timeline, but it is an equal opportunity offender; it damages Split Timeline just as well. And if that's the case, well... then I'm content where I'm at.

This is the flaw of your argument - You choose to disregard the rule about only adhering to the strictest level of canon. Basically, this is nothing but personal speculation,
Let's do this. Leerooooooooooy... Jenkins!
and you don't even consider every aspect of the ending, such as Zelda saying she can use it "as a Sage" to return Link to his original time (outright stating that it was due to her own power, not the Ocarina's, that she could send Link back)
OBJECTION! Now, let's break this down and weed out the assumptions that you added to her statement.
It is true that Zelda said the line which you quoted, but what did she
really say. She said that she had the power to send Link back to his original time; most of that is word-for-word, although I don't want to spend the effort to look it up, so I won't put quotes around it for safety's sake. Now, this sentence she uses is a very simple logical statement. Because I am a Sage, I can do this. Zelda is a member of a set of people S that have the ability to manipulate time; we'll call that action X. So, S proves X. Now, let's see what she
didn't say:
She
didn't say that
only Sages could do this. In short, she didn't say that X proves S, or more simply, S
equals X. There's nothing that states that Link doesn't share this property either. Keep this in mind because we'll use it later. Also, she also
didn't say that she has a special version of the ability to manipulate time; in other words, while members of NOT(S) could do something like X, they cannot do X itself. This also has the corollary that the other six sages could conceivably do the very same thing that Zelda herself did.
By strict canon, we must assume that her statements are 100% accurate. Thus, every statement above is assumed true.
Now, let's get back to her "ability". If Zelda's power was all that was needed, so why did she play the Ocarina when sending Link back? Herein lies the trick. Truly we have to define what X is, or rather how X works. Is X simply the ability to manipulate time? If that were possible, why didn't she just screw with time more? Why didn't she send someone else back in time to hide the Ocarina of Time from herself before she ever did it? That someone else very well could have been Link. Because she plays the ocarina (which initiates the effect of the spell, I might add) and the fact that she cannot seem to exhibit this trait without the Ocarina of Time, we must assume by canon that it is the Ocarina that is causing this effect. Thus, X is not a native skill; X, rather, is the
ability to manipulate time
using the Ocarina.
Now, I mentioned that we'd use the the fact that not just Sages could do this as well, yes? Well, so we have it. Link can use the Song of Sun, which manipulates time, spinning forward the clock a full half day (maximum). Thus, Link also is able to do X, and since Zelda's parsed statement does not preclude him from doing such, the ending of the game is self-consistent with itself. End of story.
or that she plays Zelda's Lullaby, a song with 'mysterious powers' which Link never uses to travel in time
OBJECTION! We also know from canon a peculiar part of the ocarina songs. They have zero effect when played... until the moment you actually realise what they do. Go ahead, boot up a new game and try to play the Nocturne of Shadow when you get the Ocarina of Time. Doesn't work. Funny that...
Thus, we must assume that there is something
more to the song than just the melody. And that is where canon saves the day. There is
intent. A magical melody when played without expecting some result will inevitably yield nothing because there is no intent behind the playing of the song. However, if I expect a certain melody to produce some effect (and indeed that melody is capable of doing such),
then and only then is the magical ability unlocked.
Zelda never taught Link that Zelda's Lullaby could be used to manipulate time. Merely, he found out that it had mysterious properties. Thus, Link likely never imagined it could manipulate time, and thus the song therefore did not. All of which canon. All of which self-consistent. (Ironically, this is also self-consistent with the Song of Storms.)
(BTW, I understand that you ignored Majora's Mask, because you are only comparing one game to itself, although it could be argued that Nintendo fixed this error in OoT with the sequel, thereby disqualifying it as an argument to be used this debate).
OBJECTION! I actually have taken that into account over at the 'Blog because of some very simple things that we must consider.
First and foremost, any possible intervention of the Song of Storms from Termina--regardless of source--had to take place
after (with respect to the normal flow of time) Link played the Song of Storms for the Guru-Guru Man as a child. Otherwise, we have the contradiction that the well would already be empty because the windmill already would have went crazy. (In addition, the Guru-Guru Man would already be angry, but again, the well is sufficient.)
We can now strictly narrow down any possible person who could have done the deed. First off, from our experience, no one in Hyrule seems to be aware Termina exists, even Zelda. Second off, there are only two people in Termina that knew of Hyrule (or at least knew of places other than Termina): the Skullkid and the Happy Mask Salesman. The Happy Mask Salesman never was shown to have an ocarina in his possession, so it could not have been him. The Skullkid did have possession of the ocarina, so it theoretically COULD have been him. Strict canon reduces down to that it must have been him if this theory holds true.
HOWEVER, we see from the opening scene when Link is ambushed that, when the Skullkid tries to play the ocarina, he cannot play it well. He can get a few notes out of it, but he doesn't start playing any sort of complicated (or even simple!) tune. Thus, he does not know how to play it. Thus, he couldn't have possibly played the Song of Storms for the Guru-Guru Man (assuming he went back to Hyrule during the three day stretch). Thus, that means he didn't do it.
This is further supported by the fact that Link going back in time to play the Song of Storms
did not change the future at all. Everything seemed to exist exactly as it was before. If someone else
had played the song, Link had just changed the past so that two people would have played the song, and there could have been some changed outcome in the future (unless you believe Future Predestination). Now granted, this is a much harder leap of faith to overcome than the first, but I'm using this only as supporting evidence only, a secondary backup plan. Strict canon shows it wasn't the Skullkid, and canon shows it wasn't anyone else in Termina.
That's not to say canon does not contradict itself within the series. Twilight Princess alone has enough plot holes to 'disprove' the ultimate infallibility of canon, and I don't quite understand why you chose the Song of Storms as the subject of your argument (perhaps to intentionally confuse people?) although I would argue that none of those contradictions could qualify as evidence for anything (more on that below).
My reason for choosing the Song of Storms over
Twilight Princess is actually a very simple reason.
You see, to disprove the possibility of the existence of something that can potentially have many forms, you have to meticulously disprove every such possibility, one after the other. It is why my article had three separate sections, one for each of the major timeline templates. Because it all revolved around a single game, I didn't have to worry about placement of games with regard to other games. I just had to say, "Look here... this game is self-inconsistent, and reordering your timeline won't save you. QED."
Now, the moment I bring into account other games, I have to worry about
ordering. What if, for instance, you decided to place
Twilight before
Ocarina? How about the child timeline? The adult timeline? A single timeline where it takes place after
A Link to the Past, after
Legend of Zelda, or any possible permutation of games therein? Each of those are a theoretically valid possibility as far as timeline goes, and I would have had to disprove each one through proof by exhaustion. And with the number of possible permutations, how exhausting it would be! And THAT doesn't even account for the various timeline templates. The number of variables by accounting for any other Zelda title (with the exception of
maybe Majora's Mask) became much too complicated (which is why I didn't even use
Majora's Mask except as supporting evidence).
Here I would argue that outright contradictions of the main plot within a game should be ignored, even if you consider canon to be the ultimate truth. This is because such plot holes are obviously not part of the intended story, and are due to human error. Thus they should be ignored, because they defeat the very point of the game.
Alright, so let's now be candid about this.
It is very obvious to say for certain that, if some event A exists that produces result X and some event B exists that produces result NOT(X), one of the two must be invalidated. It's paradoxical, I understand that. One of the two, A or B, must get ignored, and to prevent loss of generality, either or may be assumed false. But herein lies the problem. The fact that A or B is false proves that the canon is not
completely true. There's no 100% any more. There's no guaranteed truth to any of it (although we still, for sanity's sake, assume there's truth to most of it).
But let's then say that we've arranged our timeline in some fashionable order that
in our order, we find some event C that produces Y and some event D that produces NOT(Y). However, another timeline may not have this contradiction... but will have yet a different contradiction set (event E and outcome Z, event F and outcome NOT(Z)). Is C false? Is D false? Are C and D both true, thus invalidating that one timeline? Likewise, E and F may have the same problem. Suffice to say, when you get into conditional paradoxes--paradoxes that only exist within specific timelines--we run into problems assigning universal truth values. The first timeline may have no problem assigning E and F as true events, but the second one does not... yet the second one is better than the first one. Which is superior? No one can say.
So much of this becomes subjective that I stop worrying about specific truth values of individual events in the canon and then just start giving everything a specific weight. "This fact isn't a big deal so I can declare it false if need be; this fact is a very big deal, so I should strictly adhere to this." I then find out which canon events my timeline must have set to "not true" and sum them up, and hopefully that number's not too big. So long as I don't completely go hog wild and kill everything, I'm good. After all, it's subjective, and we can't do anything to change that. Once you declare that at least some nonzero percentage of the canon is false,
any percent can be arbitrarily chosen as false.
That said, I still use in-game canon to prove, or disprove a theory, and argue the timeline just like everybody else. This is because the games reflect the creator’s intention, and so long as no new information is given we must assume what we know from them is true. However, if a new game is released that contradicts past games, then the new game reflects the current intention of the creators, and thereby overrides the older games.
And that's the thing I'm really driving at. We really shouldn't be so big and grant ourselves the "right" to hold timeline "debates." The timeline we know is broken at some level, and it's just a matter of time before it becomes truly irreparable. Instead, our purpose now should be "discussing" rather than "debating." We should be comparing ideas and possibly being swayed by the other side, but with a broken canon, the goal of finding te "truth" is a subjective affair rather than an objective one.
At any rate, your argument is interesting and might entertain certain people (while deluding most), but it ultimately does not hold up to any closer scrutiny.
I'd like you to consider that after this post.

Edit: I noticed that you tend to use the "Well, prove that I'm wrong!" argument a lot. However, I could say the same thing, and thus you are not qualified to make any claims about the absoluteness of your idea either. For instance, take the part where you said the OoT might work without the MS in MM because the flow of time is different in Termina. Fair enough. However, can you prove that, especially with Zelda stating to Link that the Goddess of Time (who apparently exists in both worlds) would come to his aid even whenever he played the Song of Time (in Hyrule or otherwise)? No, Zelda says the OoT can control time on its own. Thus falls your entire argument.
As LoS wisely said, this article is a trap.

And you have just fallen in it. Let me explain how:
When I discuss strict canon, my precise definition for it is as follows: "If the canon suggests (shows some evidence for) some 'fact' F but
does not suggest (e.g., there is no evidence to support) any alternative theory G, then F may be assumed to be true
even if it does not prove F is true. This is the very epitome of Occam's Razor, namely that the simplest or most strongly supported explanation is as true as the game mechanics and dialogue itself. That is the assumption I am going with, and my proof attempts to show a contradiction using that philosophy.
I do this by creating what most people would call a "Smoke and Mirrors" proof. Think about it like those amusement park rides where you're going through a haunted house. I construct for you a scene that looks real, but it only looks real when you look at it from the little trolley car path that I laid out. You might even not be scared because you know it's all smoke and mirrors, a presentation that really isn't real but looks real enough. Such is the case here.
However, we all know that once you go behind the scenes or look at things from a different vantage point, the illusion disappears. You realise it's all animatronics and a painted, fiberglass set. When you look from things from that perspective and you realise the whole thing is false, you cry foul. It's just smoke and mirrors; the proof is a fraud, and the scene I created for you--the conclusion--is invalid.
However, I was darn careful in the construction of this article. You see, my goal is not to prove that the scene I constructed is real; rather, my goal was to show you that the
path--the
means, not the end--is invalid. Because I assumed my definition of strict canon, I instantly invalidated anyone leaving the path to investigate things from that alternate perspective. The moment anyone does it,
I declare foul and can ignore you because you either (1) didn't follow my assumption and therefore proved something different or (2) used circular logic and thus didn't prove anything at all.
Can I prove any of this? Can I prove that the Master Sword time travel is the same as the Ocarina of Time time travel? Absolutely not.
But I don't have to.
All I have to do is show that canon suggests it as a possibility... and shows zero evidence for everything else. Canon doesn't have to prove or disprove anything; merely the suggestion and the lack of suggestion is enough.
So, I turn it back to you, as you expected. Can you provide an alternate explanation that canon supports? I'm not asking you to
prove; after all, I didn't prove myself; all you need is suggestive evidence, logical deductions from canon. Until you can do that, my theory holds. The scene remains valid when viewed from the path.... but the path itself is invalid.
Have fun.

Oh and as far as the Goddess of Time goes, well, I can boil that one down to intent as well. Link didn't know the Goddess of Time would come to him in
Ocarina of Time, so the effect of the song was different in
Majora's Mask than
Ocarina. Similarly, I can also say that the Goddess of Time
did help Link whenever Link played the Song of Time and needed help the "Goddess of Time" could provide; you can look at that in terms of the "context sensitive" buttons from
Conker's Bad Fur Day, a song that gives you precisely what you need when you need it. In
Ocarina, it moved blocks when Link needed it; in
Majora, it toyed with time to help Link out. Either of those canon suggestions cause the theory to hold.
Edited by The Missing Link, 02 February 2007 - 04:02 AM.