While some authors present the Zermelo-Fraenkel axioms as including the axiom of the empty set, for the sake of simplicity, the standard axiomatization of set theory includes no such axiom. In fact, the existence of the empty set is a theorem, which is a consequence of the axiom of infinity, and the axiom schema of collection. Also, I think that discussing the role of the axiom of choice here is important, since it allows us to say more about the continuum hypothesis. Without the axiom of choice, the continuum hypothesis is the statement that there exists no cardinal κ such that Aleph(0) < κ < 2^Aleph(0). Here, Aleph(0) is the cardinal of the set of natural numbers N, and 2^Aleph(0) is the cardinal of the power set of N. However, with the axiom of choice, in conjunction with the Zermelo-Fraenkel axioms, one can prove the well-ordering theorem. This means that for all sets X, there exists some =
Hello Angel! I find your comment important and relevant. During the realization of the video we decided to omit certain points for the sake of simplicity and accessibility to everyone. Some of these points are precisely the ones you are pointing out, so hopefully people will get to read this. In particular the clarification of the syntactic/semantic issue that became confusing at the end. Also, thanks for taking the time to answer a lot of the questions in the comments!
Good comment, mate! Even with a non-math degree I could follow it fine, gives interesting and important context and corrections to the video. Thank you!
The incompleteness of consistency has 6 more orders of complexity, essentially not validly available in some sovereignty regimes, one of which is canada. Beyond the criminals, satanists, and anti-christs, their laws are irrelevant, and not cured by the modeling of civility. The idea is therefore alien to conformal existence, and, patently unlawful, to be even so cheaply discussed on a controlled bulletin board.
This was a very helpful comment. Thank you. People used to tell me that the CH is neither true nor false (it can be either you want). I always thought that sounded ridiculous.
great video as usual! small correction: at 1:47, "CH is not inconsistent" is the same as "is consistent", and thus not disprovable. I think you meant to say that the *negation* of CH is consistent w/ ZFC
Great video. Seriously, huge props for including Luciano in the project and citing him. I'm a grad student in set theory, and math youtubers have, in recent years, constantly made videos about cardinality and CH with demonstrable errors. It's really frustrating seeing my subject consistently misunderstood. So, in all seriousness, taking that extra step to refer to somebody more knowledgeable is more than appreciated.
Great video. I just want to add that "undecidable" should not be thought of as a "truth value" in the way "true" and "false" are. Rather, there is a class of truth values which are not "true" or "false", all of which are called undecidable. All true statements are equivalent, all false statements are equivalent, but not all undecidable statements are equivalent.
@MadAlly And then there is Tarski's undefinability theorem, which states, that there is no L-formula True(n). (That is, you can't prove arithmetically, that Arithmetics is true.)
@@Kleithap Real numbers are fundamentally untenable in the physical world. We are incapable of thinking of uncountability (represent a state of thought by a quantized distribution of energy. There are countable states as it is quantized). It's purely a set theoretic concept.
@@IsomerSoma A representation of a continuum is imaginable. You can not simultaneously imagine uncountable numbers of things (for that matter, aleph nought in untenable)
I am an engineering student and I have always seen math as a chore , something I have to learn to make everything else I study possible I started using Brilliant to make it less of a chore because several "big" youtubers were sponsored by it (im not promoting it and I have so far only used the limited free version) And after years of using derivatives as simple formulas of functions ... I finally understand them ... on a trully intuitive level, I finally get what it means this is what has ignited a bit of a spark in me , trying to learn and UNDERSTAND math ... and thats how i got this video & ur channel in my recommended All those "crazy" math teachers in school always talking about "the beauty of math" , which I never saw... I think I finally got a glimpse of what they see
How do you use the limited version? I've tried it once and it seemed to me like after one or two courses (or even exercises) I'd have to cough up the money to keep going
@@Myrskylintu Yeah , by now i have hit the limit in 5 different courses that I wanna complete , so I will get premium in the next few days In the mean time I am using the free platfrom Khan Academy Ideally I wanna use both to complement one another , since they are built differently
This is an excellent introduction. Perhaps a more advanced version of this video might explore the concept of "Forcing" in greater detail. Paul Cohen's proof is not easy to follow. Thank you.
It's not difficult at all, the only thing that makes it difficult is set theory language, which makes it a pain in the ass to explain how it works. It works like this--- a truly uncountable set is a branching structure, there are at least two paths you can take with any finite amount of information you are given. An example is the real numbers--- you imagine you know some digits and then you can always ask Vanna White to uncover another digit, and there are at least two choices for that digit. This makes the real numbers into a "tree". Now given any structure of ZFC, like some enormous aleph, you can map it to a collection of partially-known real numbers, think of a bunch of partially specified points in the tree, only finitely many digits are known of all these numbers, but you keep asking Vanna White to uncover more digits, one by one. As Vanna uncovers more digits, you prove more and more statements, for example, if one number called "x" has "3.141...." showing and nothing else, and Vanna uncovers the next digit, and it's "7", you immediately will prove "x is not equal to pi" from this information. But if you imagine going on forever, even when Vanna uncovers EVERY digit of every number, there will still be a ton of things you didn't prove either way, meaning the union of all statements you prove from partial data is never going to be complete. So you define a different notion of 'true' and 'false' which is designed to end up complete in the limit, this is Cohen's notion of "forcing". A statement is "forced true" when you can prove it is true from the digits you already know, like "x is not equal to pi" was forced true in the example earlier. A statement is forced false if NO FURTHER EXTENSION is ever going to force it true, meaning, from this point on, no matter what Vanna uncovers, you are never going to prove it true (notice that isn't saying you will prove it false, only that you will never prove it true). You extend this definition in the obvious way by logical deduction, so if you can deduce something from forced true and forced false statements, they are also forced appropriately. If you think about it a second, given this new notion, EVERY statement is either going to be forced true or forced false in the limit of revealing every digit. That's by definition, we designed it to work that way. So now you know that every statement is going to be decided in the sense of forcing, no matter what path you pick for your numbers. now pick a path through the set that is "generic". In the case of real numbers, you're picking digits, and making the real number generic just means that you avoid every single one of the countably many ways to name a property in set theory so specific that it restricts the real number digits to something more precise than any interval. For example "Less than 18 digits of my real number are not equal to 3" is too specific, because once I know 18 digits that aren't 3, the rest are all 3. A "too specific property" is formalized by the notion of a meager set. One way to define a meager set is that if you are playing a digit game, where your adversary gets to define a bunch of digits of a real number, then you get to name one digit, then your adversary gets a bunch more digits, then you get to name one digit, and so on, and your goal in naming digits is to avoid landing in the set eventually, then a set is meager iff you can always win this game, no matter what your opponent does. This means Vanna can always uncover new digits to avoid any given meager set, again, by definition. Because Vanna can avoid any meager set, you can make Vanna avoid ALL the meager sets inside a given countable model of set theory, because the union of countably many meager sets is still meager. These numbers obtained by Vanna are now 'generic' relative to your model of set theory, they have absolutely no specific property which you can write down in set theory which can specify a number in some way significantly better than saying "it's somewhere in that interval over there". In other words, you have freely specified a bunch of real numbers with no specific logically writable-down properties beyond the list of digits that you are uncovering. With these new numbers added to your countable model, you can easily prove they are all forced to be different (meaning you're never going to prove x is equal to y no matter how many equal digits are uncovered, so 'x=y' is forced false), so if you have one such number for each of the countable many elements of some enormous aleph (in a countable model of ZFC), like aleph omega, then the continuum will now become bigger than aleph omega. That's the end of the continuum hypothesis proof. There's one issue here, which is that you have to prove that adding these numbers doesn't "collapse cardinals", this means, you have to prove that aleph omega is still aleph omega at the end of the process of adjoining. This is proved using the 'countable chain condition', which states that any collection of disjoint already for sure unequal partially specified digit-sequences is countable (in any model of set theory, just countable). Then at no stage are you ever going to produce a map between two different cardinalities. The best book on this is probably Nik Weaver's recent book, but Cohen's original book is definitely second best.
Paul Cohen's proof is a very bad place to start with learning forcing. There are books like Kunen's introduction to independent proofs, or Bell's boolean valued models. One can try Jech's set theory 3rd millenium edition, but this book is harder. Forcing is also explained in Kanamori's higher infinite. Nice exposition can be also found in handbook of mathematical logic. But in order to understand forcing well, i would suggest to get basic background in mathematical logic before (formulas, proofs, models, filters, boolean algebras).
Thank you to both Anna Clara Fenyo, and Elizabeth Harper, for your insights and recommendations. Your remarks may contain the germ of a script for the more advanced tutorial that I believe this subject deserves. Some of the references offered (including one addition of my own) follow: "Forcing for Mathematicians" by Nik Weaver, World Scientific Publishing, January 2014 "Set Theory and the Continuum Hypothesis", Paul J. Cohen (W.A. Benjamin, 1966), Dover Books on Mathematics, Illustrated Edition: Reprinted December 2008 "Set Theory An Introduction To Independence Proofs (Studies in Logic and the Foundations of Mathematics, Vol. 102)" by Kenneth Kunen (Elsevier, 1980), North Holland, Reprinted December 1983 "What is Mathematical Logic?" by C. J. Ash, J. N. Crossley, C. J. Brickhill, J. C. Stillwell, and N. H. Williams (Oxford University Press, 1972) Dover Books on Mathematics, Reprinted October 2010 I found the chapter on Set Theory in Crossley, 1972, to offer a clear though incomplete overview. I have been looking for something more thorough, but which does not make quite the same demands as the original proof by Cohen, 1966.
@@christopherhume1631 If i were to choose only one, i would suggest Kunen. It provides not only forcing, but also rudiments of philosophy of set theory and it introduces everything basically from scratch
66 was a strange year where much happened with everything in a one +volume CRC. Thanks, Luciano for your hard work and thanks to your HONEST cohort. Refreshing to see this amongst the brainiacs. Good thing we're leading the way.
Thanks Luciano! And good luck with your studies. Greetings from a colleague from Germany. (But with focus not in set theory, though. I wanted to keep my sanity, lol.)
Good work Aleph 0 and Luciano Salvetti! I just started watching a 12 hour philosophy of mathematics series by Joel David Hamkins and really appreciate your short video's approachability in discussing important concepts and background in the foundations of mathematics. Thanks!
Well, there is a catch here: The accepted standard axiom system of modern math is not just ZFC but actually ZFC + large cardinal axioms. ZFC has an axiom of infinity. But this axiom only gives us the first level of infinity. Higher notions of infinity motivated by a principle known as "reflection" are called "large cardinal axioms". They are considered true by most set theorists. Such axioms are added to ZFC as needed on a case by case basis in mathematical practice. The philosophy behind this is the idea that the "true mathematical universe" that we try to model is undefinably large. A prime example are regularity properies of the real line. By assuming the existance of say a supercompact cardinal one can derive the Lebesgue measurability of all definable sets of reals. In contrast Gödels model L is incompatible with even moderately large cardinals. It's a model of ZFC + V=L. But V=L is considered false in the "true universe" by most contemporary logicians because it implies the nonexistence of e.g. measurable cardinals. V=L implies the contiuum hypothesis (CH), but it also implies many weird things, like the existence of a definable (projective) Banach-Tarksi decomposition of the unit ball. In fact V=L is wrong and such a Banach-Tarksi decomposition does not exist, but large cardinals are needed to show this. So, Gödel's V=L (together with a forcing construction by Robert Solovay) would give the wrong impression, that we could "choose" wether non measurable projective sets of reals exist or not. Both statements are compatible with ZFC (technically an incaccessible cardinal is needed here). But we know now (using large cardinals) that actually V!=L and all projective sets of reals are in fact measurable! Note that Luzin believed the latter question to be unsolvable in 1925. But what about CH? Starting from any ground model one can easily force not-CH as well as CH and both forcing notions will (very different from Gödel's V=L) not destroy any known large cardinals. So CH is even independent of ZFC+large cardinals! But does that mean that we can coose? The answer is we don't know. Wether CH is in fact true or false is an open problem as of today but not necessarily an unsolvable one. All we know is that the statement is independent from todays accepted axioms of math. But there is interesting research going on and one day we might solve this question.
In fact, I was impatiently waiting for a new video from you, especially since there are topics that you did not complete explaining in previous videos, especially the last Fermat theorem. I jumped out of joy when I received a notification from your channel with a new video, but I was waiting for more from you. Especially the explanation using mathematical equations, and not just impromptu speech. Because most of the viewers of this channel are specialists in this field. In any case, I thank you and your colleague who helped you (Luciano). I am waiting for a new video from you soon. Good luck to you and your friend.
"a model of ZFC is a pair (M, epsilon) where M is a set and epsilon is the "belongs to" relation" That's clear. And a set is... A point in a model of ZFC ? That is, to my knowledge, the biggest offense in mathematical logic. In order to work with "clean" set theory, ie study ZFC and its properties in terms of first order logic, you have to work with naive set theory. There are nightmarish things hidden in there as well, like the ordinals and the cardinals actually being in "bijection" (as proper classes), or Lowenheim-Skolem's theorem implying that there exists a countable model of ZFC. Funnily enough, such a model still thinks that what it calls "real numbers" are bigger than what it calls the "natural numbers". Of course, a bigger model containing it knows that it isn't true. That's genuinely terrifying to me. You can have a tower of models of ZFC, each one thinking it is a "genuine" model of ZFC and has "actual" cardinals but actually having a prescribed cardinality itself in the bigger model. The fact that there exists no way to fix (and an actual theorem telling you : "there ISN'T a way to fix this and there will never be one (as far as 1st order logic is concerned)" is mind-numbing.
You don't need naive set theory to assert the existence of first order models. Every model of a first order theory is a "set" in the sense that it's a bunch of objects which satisfy the axioms of that theory, but you don't need to assert that, for example, the power set of a model is a set, or the union of two models, etc. It is a theorem of first order logic that any consistent theory has a model.
Dear Aleph 0! Your videos are all superb! They are packed with so many wonderful and complex ideas in mathematics and physics and their unifying inter-relationship. I have also enjoyed your other presentations on Galois Theory, Stokes Theorem and DeRahm's Theorem. Thank you for creating and and uploading them. Kudos to Luciano also for this Continuum Hypothesis material. I would like to suggest if you can give names of accessible books to further explore these topics in detail if viewers choose to do so. It would greatly help in the dissemination of scientific knowledge among the general public. Cheers!
Brilliant!!!! Thank you Aleph 0 and Luciano!!!! Was looking for a great explanation to this very interesting topic. I remember this lesson during my time as math major (22 years ago). Unfortunately, I changed majors to physics. But I still have tons of love for math.
I'd love to hear more about Forcing. I haven't seen a lot of resources online that discuss the topic. Also, I was wondering how other set theories like NBG and MK play into this question.
NBG is equiconsistent with ZFC, since it is a conservative extension of ZFC. MK is not a conservative extension of ZFC, and is stronger than NBG and ZFC, but the continuum hypothesis is undecidable in MK.
MK uses a stronger form of choice if I recall correctly and is not equivalent to ZFC. Forcing is a hard topic. If you want to learn more about it, I’d suggest starting light with some classical set theory. The standard is Ken Kunen’s book, but I think that is like reading Rudin on a first run through real analysis. Start with Halmos, then find any book on introductory set theory. It should go up to at least the equivalence of AC with Well Ordering and Zorn as well as discuss the reflection principle. You’ll need to know some model theory and Gödel’s theorems, which themselves require some decent background. Once you get that, Kunen should be tolerable. Forcing is essentially “outer model theory” and so you need to understand those things first to get it.
@@seanspartan2023Sure thing. Don’t skip the sections on cardinal arithmetic and infinite combinatorics. They are critical for understanding everything that comes later.
Believe or not, TIL that Cantor lived into the 1900s and there are photographs of him! For some reason I've always thought he was more like Euler's contemporary.
For those who wish to get a flavor for Cohen forcing, this is roughly how one proceeds: Think of an object. If there is no obvious objection to its existence, then it *does* exist in the universe of sets, for that universe is vast. Use the notion of a *poset* to create increasingly consistent approximations to this object, and we obtain a forcing axiom for that object. In other words, think of it this way: The Continuum Hypothesis is like asking whether unicorns exist according to the laws of physics (unicorns being a set of cardinality between the natural numbers and the real numbers). Cantor claims these unicorns *do not exist.* Gödel answers the question using his constructible universes, saying that "yes, Cantor claiming that unicorns do not exist is consistent with the laws of physics." Cohen, however, answered it by saying "Well, we do not observe these unicorns in our (constructible) universe, however, there is no good reason why unicorns should not be allowed to exist, provided we can model our universe in a way that does not violate reasonable laws of physics (such as gravity, quantum mechanics, etc.), but may take liberties with the planets or molecular makeup of these unicorns. Who is to say these unicorns don't feed off silicon, or methane gases, etc.? So, let us create some reasonable guesses, as to the details of these unicorns: They look and run like horses, but they have horns, and they may run through walls (these are not violations of the laws of quantum mechanics, for example). Thus, we can approximate the biological nature and habitats of these unicorns to arbitrarily closer degrees of exactness, by running our arguments to their increasingly *filtered* conclusions, and eventually we obtain a perfectly reasonable model of a universe that works just like ours, except unicorns *do exist!* Thus, the statement "unicorns exist" is in fact independent of our current knowledge of the laws of physics, because consistent arguments can be made for and against such an existence. However, if we extend those laws of physics, (such as adding the law of *constructibility*), we can indeed demonstrate universes in which unicorns are not only plausible, but inevitable.
Thanks so much for uploading this, just when I needed it. This topic was one of the topics I suggested you upload next. And you did it. I am writing an important paper on the foundations of math and the future of math and physics, and this is something I am addressing. Topos theory and Grothendiecks Algebraic Geometry would be an awesome next topic. Or possibly abc conjecture and the purported proof by Mochizuki (although this is pushing it)...
Thanks to you and Luciano. I now feel like I understand the place of the Continuum hypothesis in mathematics better. I also understand a little bit more about zermelo Frankel axiom's
1:05 Saying "larger" and "size" for infinite sets is extremely misleading as there's two mutually exclusive definitions-cardinality and containment. The integers contains the even numbers, but they're the same cardinality (can be mapped onto one another).
One of my favorite statements that depends on the continuum hypothesis involves sets of “strong measure zero”: for any positive sequence (ε_n) you can cover your set by a sequence (I_n) of intervals where the length of I_n is less than ε_n. Whether or not all strong measure zero sets are countable depends on the continuum hypothesis.
My favorite statement independent of ZFC is the question of the existence of inaccessible cardinals, because it is by far one of the easiest of such topics to formulate and comprehend.
Dumb non-mathematician talking. What is an inaccessible cardinal for this purpose? Don’t inaccessible cardinals exist by definition in ZFC because of the axiom of infinity? You can’t get there by finite sequences of power sets on finite sets or performing arithmetic operations on integers finitely many times.
@@headlibrarian1996 What you are describing is an infinite cardinal, not an inaccessible cardinal. A cardinal κ is (weakly) inaccesible if and only if all of the conditions below hold: •κ is uncountably infinite •κ is not the sum of κ smaller cardinals •for all λ, if λ < κ, then λ+ < κ, where the successor of λ is λ+ There is also (strongly) inaccesible cardinals. The only distinction is that the last condition is replaced by: •for all λ, if λ < κ, then 2^λ < κ If the generalized continuum hypothesis is true, then all weakly inaccesible cardinals are strongly inaccesible cardinals, and in that case, they are simply called inaccesible cardinals.
@@angelmendez-rivera351 This formulation is rather difficult to comprehend imo. The way I understand them, is that they are exactly the uncountable regular limit cardinals, limit meaning they are not x+ for some cardinal x (or 2^x in the case of strongly inaccessibles), and regular meaning that the smallest unbounded sequence within them is as large as the cardinal itself.
@@Tsskyx My "formulation" is not more difficult to comprehend than yours. My "formulation" is actually _identical_ to yours, in that my formulation simply provides the definitions for the words "limit" and "regular" without actually using the labels themselves. Nothing about the stated conditions in the definition is actually different. Obviously, if you know how each word is defined individually already, then merely using the words in question is simpler, but I have no reason to assume the person I was replying to is not familiar with the words, and if it does turn out that they are not familiar with the words, I am just doing them a disservice. Either way, the definition is difficult to understand, either because the terminology is completely foreign, or because the definition, without relying on labels, is very dry to get through without prior intuition of it.
@@headlibrarian1996 Inaccesable cardinals are just the ones that don't result from ordinals defined via addition and union. For example, the finite cardinals result from the finite ordinals which you can think of as integers. The first infinite cardinal is the union of all the integers, that's how we're able to access infinite ordinals and cardinals. (A cardinal is just the first ordinal of its size)
Wow, I had no idea there were actual constructions where the continuum hypothesis was false. If it isn't too hard, it would be cool if you could talk about how to create a forcing extension
The details can be quite tricky, but roughly it can be thought of as a generalization of Cantor diagonalization. Or at least adding a generic real can be thought of this way. The main components of a forcing construction are a poset consisting of elements called forcing conditions and a class of special sets called names. These conditions are ordered in such a way that “stronger” conditions provide more information about the intended generic object you want to construct. For example, Cohen forcing uses partial functions with finite domains as conditions. Laver and Sacks forcing uses special types of trees as conditions. Grigorieff forcing generalizes lots of forcings by specifying certain types of Boolean ideals which can be used as domains for partial functions. The structure of the forcing typically is derived from the kind of question one wants to answer. The names, on the other hand, are a way of talking about objects that exist in the forcing extension, but that may not exist yet in the ground model. (The ground model is where you build the forcing poset and the names are kind of like a computer simulation of the generic extension.) Names can be quite tricky to handle due to their recursive construction. I like to conceptualize them as “cloudy” or “probabilistic” sets. We may not be able to say exactly which sets they are (Really I mean which sets they will become in the forcing extension.) given certain forcing conditions, but strong enough forcing conditions can give partial information about the structure of a name. Think of this as “resolving” some of the cloudiness. For example, I might have in my hand a name that contains the correct information to be the Prufer group or a Cohen-generic real, but will get only one or the other if certain conditions are met. If I travel to the left I get the group, if I travel to the right I get the real. The last critical thing to understand is density. Density in forcing roughly corresponds to the idea that a logical statement can be true “generically often”. A better way to think about this is that a property is generic if no matter how much information a condition gives you, you can always add a little bit more information to make the property true. This is the connection to diagonalization. For example, suppose you had a list of real numbers. Then you can generically extend any finite domain partial function to be different from the next real in your list and thus can obtain a “new real” generically by just making the right next digit choice at every step.
Cool topic. Just one thing about the parallel postulate: you can prove from the axioms, that there exists a parallel line, but not if its unique. Parallel postulate is exactly that: the parallel line is unique. Spherical geometry is a poor example for this, because it doesnt satisfy all the euclidian axioms in the first place and doesnt have parallel lines at all. So it doesnt give meaningful input on the PP. Hyperbolic geometry is probably what you were looking for; here the euclidian axioms are satisfied, so you have parallel lines. But they are not unique, you may have a multitude of parallel lines through a point. That proves, that the PP is independent from the euclidian axioms.
Excellent opener. I think people really need to understand more that math (and all other kinds of philosophy such as the sciences) are necessarily self referential and cannot make claim to absolute truth (with the exception of the truthfulness of the statement "something exists"). At the end of the days it's all model building based on our intuitions and practical desires!
I went to Hebrew school in Boro Park, a neighborhood of Brooklyn, NYC, where Paul Cohen also attended. Along with a third pre- bar mitvah boy, we sometimes skipped the class and chatted together. I don't mean to convey that Paul and I were friends, just acquaintances, but we, the other boy and I, knew that he was studying higher maths. I understood that he was close to his grandfather. I don't know if his grandfather was at all mathematical. We met again at Brooklyn College in 1950, where I struggled with physics. I believe that the maths chairman, a Prof Borofsky - a very likable guy - recognized Paul's potential and arranged for him to go to the U of Chicago to do graduate work in mathematics. I envied Paul's escape from Brooklyn College's rigid curriculum.
Be careful when you speak of “true” vs. “provable”. For example, the undecidable proposition at the heart of Goedel’s famous theorem (which translated into English states “this proposition cannot be proven”) indeed cannot be proven nor disproven, as both options would lead to a contradiction - but reasoning outside the formal system, one realizes it’s actually true.
I love this video! Amazing work. One correction though: at 6:20 you state "...or you can choose." I think you mean to say is undecidable, which while one might consider the surface semantics, there is a difference between arbitrary choice and undecidability. See: Maddy, Penelope (2019), What Do We Want a Foundation to Do? In Deniz Sarikaya, Deborah Kant & Stefania Centrone (eds.), Reflections on the Foundations of Mathematics. Springer Verlag., and Kurt Godels Completeness Theorem.
Showing Luciano some love in the comment section. ❤
Love you, Luciano Salvetti
While some authors present the Zermelo-Fraenkel axioms as including the axiom of the empty set, for the sake of simplicity, the standard axiomatization of set theory includes no such axiom. In fact, the existence of the empty set is a theorem, which is a consequence of the axiom of infinity, and the axiom schema of collection.
Also, I think that discussing the role of the axiom of choice here is important, since it allows us to say more about the continuum hypothesis. Without the axiom of choice, the continuum hypothesis is the statement that there exists no cardinal κ such that Aleph(0) < κ < 2^Aleph(0). Here, Aleph(0) is the cardinal of the set of natural numbers N, and 2^Aleph(0) is the cardinal of the power set of N. However, with the axiom of choice, in conjunction with the Zermelo-Fraenkel axioms, one can prove the well-ordering theorem. This means that for all sets X, there exists some =
Hello Angel! I find your comment important and relevant. During the realization of the video we decided to omit certain points for the sake of simplicity and accessibility to everyone. Some of these points are precisely the ones you are pointing out, so hopefully people will get to read this. In particular the clarification of the syntactic/semantic issue that became confusing at the end.
Also, thanks for taking the time to answer a lot of the questions in the comments!
Good comment, mate! Even with a non-math degree I could follow it fine, gives interesting and important context and corrections to the video. Thank you!
The incompleteness of consistency has 6 more orders of complexity, essentially not validly available in some sovereignty regimes, one of which is canada.
Beyond the criminals, satanists, and anti-christs, their laws are irrelevant, and not cured by the modeling of civility.
The idea is therefore alien to conformal existence, and, patently unlawful, to be even so cheaply discussed on a controlled bulletin board.
This was a very helpful comment. Thank you. People used to tell me that the CH is neither true nor false (it can be either you want). I always thought that sounded ridiculous.
Yip, it is what it is. Not an easy subject. Good that infinity just exists in math 😅.
This video felt more like an intro than an actual video. I hope there will be a dedicated series to this. This topic deserves it.
great video as usual!
small correction: at 1:47, "CH is not inconsistent" is the same as "is consistent", and thus not disprovable. I think you meant to say that the *negation* of CH is consistent w/ ZFC
You won’t get many likes with such a perfectionist comment, but I think Aleph0 would appreciate it more than anyone.
Thank you for clarifying this. The same correction occurred to me but I thought I was missing something.
@@charlesbrowne9590 Is there a problem with the idea of mathematical perfection?
@@charlesbrowne9590 He'll get likes from me. I was confused by this.
not ch is consistent
Great video. Seriously, huge props for including Luciano in the project and citing him. I'm a grad student in set theory, and math youtubers have, in recent years, constantly made videos about cardinality and CH with demonstrable errors. It's really frustrating seeing my subject consistently misunderstood. So, in all seriousness, taking that extra step to refer to somebody more knowledgeable is more than appreciated.
Yes! When the world needed him most! He returned!
exactly my reaction!
He never really explains any math in his videos, though. It’s just symbols being drawn, and vague concepts talked about.
@@ophello Someone once said we should take the numbers out of math
@@chaotickreg7024 I agree arithmetic is not mathematics.
@@Alan-zf2tt Arithmetic is what happens after your math is done
Great video. I just want to add that "undecidable" should not be thought of as a "truth value" in the way "true" and "false" are. Rather, there is a class of truth values which are not "true" or "false", all of which are called undecidable. All true statements are equivalent, all false statements are equivalent, but not all undecidable statements are equivalent.
@@selfieelfie Wow, didn't expect to see you here.
If they are either 'true' of 'false' then they are equivalent to one or the other.
You might be right - I can't decide.
@MadAlly And then there is Tarski's undefinability theorem, which states, that there is no L-formula True(n). (That is, you can't prove arithmetically, that Arithmetics is true.)
so one might say "undecidable" acts similarly to the "undefined" one gets when dividing by zero
I vote for more about "forcing" as well. And, please, don't be shy about using abstruse mathematical notation! I'm sure your subscribers will love it!
Glad to see you back at it, friend.
The continuum hypothesis is my absolute favourite maths mystery. It boggles the mind.
It's not a mystery, we know the answer.
@@firefox7857 I find it hard to imagine what the universe in which it is false looks like. That is a mystery to me.
@@Kleithap Real numbers are fundamentally untenable in the physical world. We are incapable of thinking of uncountability (represent a state of thought by a quantized distribution of energy. There are countable states as it is quantized). It's purely a set theoretic concept.
@@ffc1a28c7 Nah a continuum is very much imaginable.
@@IsomerSoma A representation of a continuum is imaginable. You can not simultaneously imagine uncountable numbers of things (for that matter, aleph nought in untenable)
I am an engineering student and I have always seen math as a chore , something I have to learn to make everything else I study possible
I started using Brilliant to make it less of a chore because several "big" youtubers were sponsored by it (im not promoting it and I have so far only used the limited free version)
And after years of using derivatives as simple formulas of functions ... I finally understand them ... on a trully intuitive level, I finally get what it means
this is what has ignited a bit of a spark in me , trying to learn and UNDERSTAND math ... and thats how i got this video & ur channel in my recommended
All those "crazy" math teachers in school always talking about "the beauty of math" , which I never saw... I think I finally got a glimpse of what they see
How do you use the limited version? I've tried it once and it seemed to me like after one or two courses (or even exercises) I'd have to cough up the money to keep going
@@Myrskylintu Yeah , by now i have hit the limit in 5 different courses that I wanna complete , so I will get premium in the next few days
In the mean time I am using the free platfrom Khan Academy
Ideally I wanna use both to complement one another , since they are built differently
i unlocked brilliant completely for free by pretending to be an educator
Got me learning now thanks..
Welcome Back!!!!!!!
I think you are one of the channels that I wait the videos of the most eagerly. Great job !!!!
Glad you're back. Made my day.
This is an excellent introduction. Perhaps a more advanced version of this video might explore the concept of "Forcing" in greater detail. Paul Cohen's proof is not easy to follow. Thank you.
This is a difficult topic (at least to me)
It's not difficult at all, the only thing that makes it difficult is set theory language, which makes it a pain in the ass to explain how it works. It works like this--- a truly uncountable set is a branching structure, there are at least two paths you can take with any finite amount of information you are given. An example is the real numbers--- you imagine you know some digits and then you can always ask Vanna White to uncover another digit, and there are at least two choices for that digit. This makes the real numbers into a "tree".
Now given any structure of ZFC, like some enormous aleph, you can map it to a collection of partially-known real numbers, think of a bunch of partially specified points in the tree, only finitely many digits are known of all these numbers, but you keep asking Vanna White to uncover more digits, one by one. As Vanna uncovers more digits, you prove more and more statements, for example, if one number called "x" has "3.141...." showing and nothing else, and Vanna uncovers the next digit, and it's "7", you immediately will prove "x is not equal to pi" from this information. But if you imagine going on forever, even when Vanna uncovers EVERY digit of every number, there will still be a ton of things you didn't prove either way, meaning the union of all statements you prove from partial data is never going to be complete.
So you define a different notion of 'true' and 'false' which is designed to end up complete in the limit, this is Cohen's notion of "forcing". A statement is "forced true" when you can prove it is true from the digits you already know, like "x is not equal to pi" was forced true in the example earlier. A statement is forced false if NO FURTHER EXTENSION is ever going to force it true, meaning, from this point on, no matter what Vanna uncovers, you are never going to prove it true (notice that isn't saying you will prove it false, only that you will never prove it true). You extend this definition in the obvious way by logical deduction, so if you can deduce something from forced true and forced false statements, they are also forced appropriately. If you think about it a second, given this new notion, EVERY statement is either going to be forced true or forced false in the limit of revealing every digit. That's by definition, we designed it to work that way.
So now you know that every statement is going to be decided in the sense of forcing, no matter what path you pick for your numbers. now pick a path through the set that is "generic". In the case of real numbers, you're picking digits, and making the real number generic just means that you avoid every single one of the countably many ways to name a property in set theory so specific that it restricts the real number digits to something more precise than any interval. For example "Less than 18 digits of my real number are not equal to 3" is too specific, because once I know 18 digits that aren't 3, the rest are all 3. A "too specific property" is formalized by the notion of a meager set. One way to define a meager set is that if you are playing a digit game, where your adversary gets to define a bunch of digits of a real number, then you get to name one digit, then your adversary gets a bunch more digits, then you get to name one digit, and so on, and your goal in naming digits is to avoid landing in the set eventually, then a set is meager iff you can always win this game, no matter what your opponent does. This means Vanna can always uncover new digits to avoid any given meager set, again, by definition.
Because Vanna can avoid any meager set, you can make Vanna avoid ALL the meager sets inside a given countable model of set theory, because the union of countably many meager sets is still meager. These numbers obtained by Vanna are now 'generic' relative to your model of set theory, they have absolutely no specific property which you can write down in set theory which can specify a number in some way significantly better than saying "it's somewhere in that interval over there". In other words, you have freely specified a bunch of real numbers with no specific logically writable-down properties beyond the list of digits that you are uncovering.
With these new numbers added to your countable model, you can easily prove they are all forced to be different (meaning you're never going to prove x is equal to y no matter how many equal digits are uncovered, so 'x=y' is forced false), so if you have one such number for each of the countable many elements of some enormous aleph (in a countable model of ZFC), like aleph omega, then the continuum will now become bigger than aleph omega. That's the end of the continuum hypothesis proof.
There's one issue here, which is that you have to prove that adding these numbers doesn't "collapse cardinals", this means, you have to prove that aleph omega is still aleph omega at the end of the process of adjoining. This is proved using the 'countable chain condition', which states that any collection of disjoint already for sure unequal partially specified digit-sequences is countable (in any model of set theory, just countable). Then at no stage are you ever going to produce a map between two different cardinalities.
The best book on this is probably Nik Weaver's recent book, but Cohen's original book is definitely second best.
Paul Cohen's proof is a very bad place to start with learning forcing.
There are books like Kunen's introduction to independent proofs, or Bell's boolean valued models. One can try Jech's set theory 3rd millenium edition, but this book is harder.
Forcing is also explained in Kanamori's higher infinite.
Nice exposition can be also found in handbook of mathematical logic.
But in order to understand forcing well, i would suggest to get basic background in mathematical logic before (formulas, proofs, models, filters, boolean algebras).
Thank you to both Anna Clara Fenyo, and Elizabeth Harper, for your insights and recommendations. Your remarks may contain the germ of a script for the more advanced tutorial that I believe this subject deserves.
Some of the references offered (including one addition of my own) follow:
"Forcing for Mathematicians" by Nik Weaver, World Scientific Publishing, January 2014
"Set Theory and the Continuum Hypothesis", Paul J. Cohen (W.A. Benjamin, 1966), Dover Books on Mathematics, Illustrated Edition: Reprinted December 2008
"Set Theory An Introduction To Independence Proofs (Studies in Logic and the Foundations of Mathematics, Vol. 102)" by Kenneth Kunen (Elsevier, 1980), North Holland, Reprinted December 1983
"What is Mathematical Logic?" by C. J. Ash, J. N. Crossley, C. J. Brickhill, J. C. Stillwell, and N. H. Williams (Oxford University Press, 1972) Dover Books on Mathematics, Reprinted October 2010
I found the chapter on Set Theory in Crossley, 1972, to offer a clear though incomplete overview. I have been looking for something more thorough, but which does not make quite the same demands as the original proof by Cohen, 1966.
@@christopherhume1631
If i were to choose only one, i would suggest Kunen. It provides not only forcing, but also rudiments of philosophy of set theory and it introduces everything basically from scratch
Outstanding, and many thanks to Luciano for joining forces with you.
Did you end this with the Axiom of Choice?
“Life exists on the fulcrum.” - JG.
This is so awesome. Thanks Luciano for the collaboration!!
After 10 months he returned with 7 minute video. Absolute legend.
Thnak you Aleph 0 and Luciano Salvetti!
Amazing video!
Thank you for making interesting concepts easily accessable to the world.
Good thing I have notifications on!
66 was a strange year where much happened with everything in a one +volume CRC. Thanks, Luciano for your hard work and thanks to your HONEST cohort. Refreshing to see this amongst the brainiacs. Good thing we're leading the way.
Great presentation! 'Love' to Luciano.
Welcome back! Keep up the great work!
Thanks Luciano! And good luck with your studies.
Greetings from a colleague from Germany. (But with focus not in set theory, though. I wanted to keep my sanity, lol.)
What is your focus on?
@@bengal_tiger1984 Differential geometry, cohomology, and scattering theory (in the periphery of the "can you hear the shape of a drum" thing)
Good work Aleph 0 and Luciano Salvetti! I just started watching a 12 hour philosophy of mathematics series by Joel David Hamkins and really appreciate your short video's approachability in discussing important concepts and background in the foundations of mathematics. Thanks!
I'm always looking forward to seeing your videos! Thanks for highlighting this topic! Also, thanks to Luciano!
Great video, thanks for returning
1:37 Who is Kurt Geddel?
Thanks Luciano Salvetti!
It's nice to have you back
Well, there is a catch here: The accepted standard axiom system of modern math is not just ZFC but actually ZFC + large cardinal axioms. ZFC has an axiom of infinity. But this axiom only gives us the first level of infinity. Higher notions of infinity motivated by a principle known as "reflection" are called "large cardinal axioms". They are considered true by most set theorists. Such axioms are added to ZFC as needed on a case by case basis in mathematical practice. The philosophy behind this is the idea that the "true mathematical universe" that we try to model is undefinably large.
A prime example are regularity properies of the real line. By assuming the existance of say a supercompact cardinal one can derive the Lebesgue measurability of all definable sets of reals. In contrast Gödels model L is incompatible with even moderately large cardinals. It's a model of ZFC + V=L. But V=L is considered false in the "true universe" by most contemporary logicians because it implies the nonexistence of e.g. measurable cardinals. V=L implies the contiuum hypothesis (CH), but it also implies many weird things, like the existence of a definable (projective) Banach-Tarksi decomposition of the unit ball. In fact V=L is wrong and such a Banach-Tarksi decomposition does not exist, but large cardinals are needed to show this.
So, Gödel's V=L (together with a forcing construction by Robert Solovay) would give the wrong impression, that we could "choose" wether non measurable projective sets of reals exist or not. Both statements are compatible with ZFC (technically an incaccessible cardinal is needed here). But we know now (using large cardinals) that actually V!=L and all projective sets of reals are in fact measurable! Note that Luzin believed the latter question to be unsolvable in 1925.
But what about CH? Starting from any ground model one can easily force not-CH as well as CH and both forcing notions will (very different from Gödel's V=L) not destroy any known large cardinals. So CH is even independent of ZFC+large cardinals! But does that mean that we can coose? The answer is we don't know. Wether CH is in fact true or false is an open problem as of today but not necessarily an unsolvable one. All we know is that the statement is independent from todays accepted axioms of math. But there is interesting research going on and one day we might solve this question.
Luciano - take some love. ❤
Finally a comprehensive explanation of continuum hypothesis I been trying to find for years. Thanks
In fact, I was impatiently waiting for a new video from you, especially since there are topics that you did not complete explaining in previous videos, especially the last Fermat theorem. I jumped out of joy when I received a notification from your channel with a new video, but I was waiting for more from you.
Especially the explanation using mathematical equations, and not just impromptu speech.
Because most of the viewers of this channel are specialists in this field. In any case, I thank you and your colleague who helped you (Luciano). I am waiting for a new video from you soon. Good luck to you and your friend.
Happy to see that you haven't abbandoned the channel!
Thank you Luciano Salvetti !👍👏
"a model of ZFC is a pair (M, epsilon) where M is a set and epsilon is the "belongs to" relation"
That's clear. And a set is... A point in a model of ZFC ?
That is, to my knowledge, the biggest offense in mathematical logic. In order to work with "clean" set theory, ie study ZFC and its properties in terms of first order logic, you have to work with naive set theory.
There are nightmarish things hidden in there as well, like the ordinals and the cardinals actually being in "bijection" (as proper classes), or Lowenheim-Skolem's theorem implying that there exists a countable model of ZFC.
Funnily enough, such a model still thinks that what it calls "real numbers" are bigger than what it calls the "natural numbers". Of course, a bigger model containing it knows that it isn't true.
That's genuinely terrifying to me. You can have a tower of models of ZFC, each one thinking it is a "genuine" model of ZFC and has "actual" cardinals but actually having a prescribed cardinality itself in the bigger model.
The fact that there exists no way to fix (and an actual theorem telling you : "there ISN'T a way to fix this and there will never be one (as far as 1st order logic is concerned)" is mind-numbing.
You don't need naive set theory to assert the existence of first order models. Every model of a first order theory is a "set" in the sense that it's a bunch of objects which satisfy the axioms of that theory, but you don't need to assert that, for example, the power set of a model is a set, or the union of two models, etc. It is a theorem of first order logic that any consistent theory has a model.
Thanks Luciano to participate on the video.
YOURE BACK :DDDD I remember reading an article about this stuff before, it’s super interesting! Cant wait to take the set theory
course at my Uni
Good to see you again
This is interesting. Thanks Aleph 0 and Luciano!
Fantastic video! Loved it
Luciano, we love you.
Outstanding 🤯 thanks Luciano! And the video itself it’s amazing and so easy to understand and learn.
Woot!! He's back! Thanks for another AMAZING video! Love it! Please keep them coming! Shout out to Luciano Salvetti for his insight! Brilliant stuff
As always, very informative video. You sir, are making an impact on the world. !!
dude i just got interested in pure maths and ran into your channel. please keep posting, love your content!
Dear Aleph 0! Your videos are all superb! They are packed with so many wonderful and complex ideas in mathematics and physics and their unifying inter-relationship. I have also enjoyed your other presentations on Galois Theory, Stokes Theorem and DeRahm's Theorem. Thank you for creating and and uploading them. Kudos to Luciano also for this Continuum Hypothesis material. I would like to suggest if you can give names of accessible books to further explore these topics in detail if viewers choose to do so. It would greatly help in the dissemination of scientific knowledge among the general public. Cheers!
Brilliant!!!! Thank you Aleph 0 and Luciano!!!! Was looking for a great explanation to this very interesting topic. I remember this lesson during my time as math major (22 years ago). Unfortunately, I changed majors to physics. But I still have tons of love for math.
I think people will be surprised by a lot of proof theory stuff, like intuitionist mathematics, constructive proofs, martin lof type theory and etc.
go post constantly man i'm addicted
You’re back!
Thank you very much Luciano!
Wow that was awesome! Thanks to both of you.
Great video! Thanks to Luciano.
You always teach me something new Luciano! You're a legend!
I'd love to hear more about Forcing. I haven't seen a lot of resources online that discuss the topic.
Also, I was wondering how other set theories like NBG and MK play into this question.
NBG is equiconsistent with ZFC, since it is a conservative extension of ZFC. MK is not a conservative extension of ZFC, and is stronger than NBG and ZFC, but the continuum hypothesis is undecidable in MK.
@@angelmendez-rivera351 Thank you
MK uses a stronger form of choice if I recall correctly and is not equivalent to ZFC.
Forcing is a hard topic. If you want to learn more about it, I’d suggest starting light with some classical set theory. The standard is Ken Kunen’s book, but I think that is like reading Rudin on a first run through real analysis. Start with Halmos, then find any book on introductory set theory. It should go up to at least the equivalence of AC with Well Ordering and Zorn as well as discuss the reflection principle. You’ll need to know some model theory and Gödel’s theorems, which themselves require some decent background.
Once you get that, Kunen should be tolerable. Forcing is essentially “outer model theory” and so you need to understand those things first to get it.
@@HPTopoG I have Kunen's book but I was never able to get deeply into it. I'll take your other suggestions. Thank you.
@@seanspartan2023Sure thing. Don’t skip the sections on cardinal arithmetic and infinite combinatorics. They are critical for understanding everything that comes later.
I know this is a science video but that song is incredible. Taking Flight. Holy moly bro. That song is way to good.
thank you for the great intuition on undecidability. had real trouble getting my head around it.
Another great video, I hope you keep making them.
Believe or not, TIL that Cantor lived into the 1900s and there are photographs of him! For some reason I've always thought he was more like Euler's contemporary.
We need more of this videos with same quality
Great, this channel is perfect
6:46 No no, I see YOU in the next video.
For those who wish to get a flavor for Cohen forcing, this is roughly how one proceeds: Think of an object. If there is no obvious objection to its existence, then it *does* exist in the universe of sets, for that universe is vast. Use the notion of a *poset* to create increasingly consistent approximations to this object, and we obtain a forcing axiom for that object.
In other words, think of it this way: The Continuum Hypothesis is like asking whether unicorns exist according to the laws of physics (unicorns being a set of cardinality between the natural numbers and the real numbers). Cantor claims these unicorns *do not exist.* Gödel answers the question using his constructible universes, saying that "yes, Cantor claiming that unicorns do not exist is consistent with the laws of physics." Cohen, however, answered it by saying "Well, we do not observe these unicorns in our (constructible) universe, however, there is no good reason why unicorns should not be allowed to exist, provided we can model our universe in a way that does not violate reasonable laws of physics (such as gravity, quantum mechanics, etc.), but may take liberties with the planets or molecular makeup of these unicorns. Who is to say these unicorns don't feed off silicon, or methane gases, etc.? So, let us create some reasonable guesses, as to the details of these unicorns: They look and run like horses, but they have horns, and they may run through walls (these are not violations of the laws of quantum mechanics, for example). Thus, we can approximate the biological nature and habitats of these unicorns to arbitrarily closer degrees of exactness, by running our arguments to their increasingly *filtered* conclusions, and eventually we obtain a perfectly reasonable model of a universe that works just like ours, except unicorns *do exist!*
Thus, the statement "unicorns exist" is in fact independent of our current knowledge of the laws of physics, because consistent arguments can be made for and against such an existence. However, if we extend those laws of physics, (such as adding the law of *constructibility*), we can indeed demonstrate universes in which unicorns are not only plausible, but inevitable.
The king is back!
Thanks so much for uploading this, just when I needed it. This topic was one of the topics I suggested you upload next. And you did it. I am writing an important paper on the foundations of math and the future of math and physics, and this is something I am addressing. Topos theory and Grothendiecks Algebraic Geometry would be an awesome next topic. Or possibly abc conjecture and the purported proof by Mochizuki (although this is pushing it)...
great video thank you. Please do one about the riemann hypothesis. the biggest problem in mathematics
Great video - thanks to Luciano
welcome back buddy
Yesss the best math channel returns!
Can’t wait for the next video
Love your videos. Keep it up buddy.
"In some sense" is a good way of describing it😌
Thanks to you and Luciano. I now feel like I understand the place of the Continuum hypothesis in mathematics better. I also understand a little bit more about zermelo Frankel axiom's
Thanks Luciano
1:05 Saying "larger" and "size" for infinite sets is extremely misleading as there's two mutually exclusive definitions-cardinality and containment. The integers contains the even numbers, but they're the same cardinality (can be mapped onto one another).
One of my favorite statements that depends on the continuum hypothesis involves sets of “strong measure zero”: for any positive sequence (ε_n) you can cover your set by a sequence (I_n) of intervals where the length of I_n is less than ε_n. Whether or not all strong measure zero sets are countable depends on the continuum hypothesis.
Borel’s Conjecture! This was the invention of Laver Forcing! The original paper is a hell of a read.
My favorite statement independent of ZFC is the question of the existence of inaccessible cardinals, because it is by far one of the easiest of such topics to formulate and comprehend.
Dumb non-mathematician talking. What is an inaccessible cardinal for this purpose? Don’t inaccessible cardinals exist by definition in ZFC because of the axiom of infinity? You can’t get there by finite sequences of power sets on finite sets or performing arithmetic operations on integers finitely many times.
@@headlibrarian1996 What you are describing is an infinite cardinal, not an inaccessible cardinal. A cardinal κ is (weakly) inaccesible if and only if all of the conditions below hold:
•κ is uncountably infinite
•κ is not the sum of κ smaller cardinals
•for all λ, if λ < κ, then λ+ < κ, where the successor of λ is λ+
There is also (strongly) inaccesible cardinals. The only distinction is that the last condition is replaced by:
•for all λ, if λ < κ, then 2^λ < κ
If the generalized continuum hypothesis is true, then all weakly inaccesible cardinals are strongly inaccesible cardinals, and in that case, they are simply called inaccesible cardinals.
@@angelmendez-rivera351 This formulation is rather difficult to comprehend imo. The way I understand them, is that they are exactly the uncountable regular limit cardinals, limit meaning they are not x+ for some cardinal x (or 2^x in the case of strongly inaccessibles), and regular meaning that the smallest unbounded sequence within them is as large as the cardinal itself.
@@Tsskyx My "formulation" is not more difficult to comprehend than yours. My "formulation" is actually _identical_ to yours, in that my formulation simply provides the definitions for the words "limit" and "regular" without actually using the labels themselves. Nothing about the stated conditions in the definition is actually different. Obviously, if you know how each word is defined individually already, then merely using the words in question is simpler, but I have no reason to assume the person I was replying to is not familiar with the words, and if it does turn out that they are not familiar with the words, I am just doing them a disservice. Either way, the definition is difficult to understand, either because the terminology is completely foreign, or because the definition, without relying on labels, is very dry to get through without prior intuition of it.
@@headlibrarian1996 Inaccesable cardinals are just the ones that don't result from ordinals defined via addition and union. For example, the finite cardinals result from the finite ordinals which you can think of as integers. The first infinite cardinal is the union of all the integers, that's how we're able to access infinite ordinals and cardinals. (A cardinal is just the first ordinal of its size)
Wow, I had no idea there were actual constructions where the continuum hypothesis was false. If it isn't too hard, it would be cool if you could talk about how to create a forcing extension
The details can be quite tricky, but roughly it can be thought of as a generalization of Cantor diagonalization. Or at least adding a generic real can be thought of this way.
The main components of a forcing construction are a poset consisting of elements called forcing conditions and a class of special sets called names. These conditions are ordered in such a way that “stronger” conditions provide more information about the intended generic object you want to construct. For example, Cohen forcing uses partial functions with finite domains as conditions. Laver and Sacks forcing uses special types of trees as conditions. Grigorieff forcing generalizes lots of forcings by specifying certain types of Boolean ideals which can be used as domains for partial functions. The structure of the forcing typically is derived from the kind of question one wants to answer.
The names, on the other hand, are a way of talking about objects that exist in the forcing extension, but that may not exist yet in the ground model. (The ground model is where you build the forcing poset and the names are kind of like a computer simulation of the generic extension.) Names can be quite tricky to handle due to their recursive construction. I like to conceptualize them as “cloudy” or “probabilistic” sets. We may not be able to say exactly which sets they are (Really I mean which sets they will become in the forcing extension.) given certain forcing conditions, but strong enough forcing conditions can give partial information about the structure of a name. Think of this as “resolving” some of the cloudiness. For example, I might have in my hand a name that contains the correct information to be the Prufer group or a Cohen-generic real, but will get only one or the other if certain conditions are met. If I travel to the left I get the group, if I travel to the right I get the real.
The last critical thing to understand is density. Density in forcing roughly corresponds to the idea that a logical statement can be true “generically often”. A better way to think about this is that a property is generic if no matter how much information a condition gives you, you can always add a little bit more information to make the property true. This is the connection to diagonalization. For example, suppose you had a list of real numbers. Then you can generically extend any finite domain partial function to be different from the next real in your list and thus can obtain a “new real” generically by just making the right next digit choice at every step.
A beginner's guide to forcing: timothychow.net/forcing.pdf
Awesome work!
Cool topic. Just one thing about the parallel postulate: you can prove from the axioms, that there exists a parallel line, but not if its unique. Parallel postulate is exactly that: the parallel line is unique.
Spherical geometry is a poor example for this, because it doesnt satisfy all the euclidian axioms in the first place and doesnt have parallel lines at all. So it doesnt give meaningful input on the PP.
Hyperbolic geometry is probably what you were looking for; here the euclidian axioms are satisfied, so you have parallel lines. But they are not unique, you may have a multitude of parallel lines through a point. That proves, that the PP is independent from the euclidian axioms.
I love that you continue to make videos. What future video plans do you have?
Thank you Luciano!
welcome back!!!
Ugh thank you. Everything else I saw on this topic was so unnecessarily thick. The bit about “True, False, or Choose” makes a lot of sense
Excellent opener. I think people really need to understand more that math (and all other kinds of philosophy such as the sciences) are necessarily self referential and cannot make claim to absolute truth (with the exception of the truthfulness of the statement "something exists"). At the end of the days it's all model building based on our intuitions and practical desires!
Vai Luciano, gran bel lavoro!!!
I went to Hebrew school in Boro Park, a neighborhood of Brooklyn, NYC, where Paul Cohen also attended. Along with a third pre- bar mitvah boy, we sometimes skipped the class and chatted together. I don't mean to convey that Paul and I were friends, just acquaintances, but we, the other boy and I, knew that he was studying higher maths. I understood that he was close to his grandfather. I don't know if his grandfather was at all mathematical. We met again at Brooklyn College in 1950, where I struggled with physics. I believe that the maths chairman, a Prof Borofsky - a very likable guy - recognized Paul's potential and arranged for him to go to the U of Chicago to do graduate work in mathematics. I envied Paul's escape from Brooklyn College's rigid curriculum.
Thank you Luciano.
Be careful when you speak of “true” vs. “provable”. For example, the undecidable proposition at the heart of Goedel’s famous theorem (which translated into English states “this proposition cannot be proven”) indeed cannot be proven nor disproven, as both options would lead to a contradiction - but reasoning outside the formal system, one realizes it’s actually true.
Awesome video! I kind of want to learn more about model theory now
So whenever some statement of that sort exists, you can _choose_ to use it as an axiom or not? Makes sense.
I love this video! Amazing work. One correction though: at 6:20 you state "...or you can choose." I think you mean to say is undecidable, which while one might consider the surface semantics, there is a difference between arbitrary choice and undecidability. See: Maddy, Penelope (2019), What Do We Want a Foundation to Do? In Deniz Sarikaya, Deborah Kant & Stefania Centrone (eds.), Reflections on the Foundations of Mathematics. Springer Verlag., and Kurt Godels Completeness Theorem.
Highly underrated comment
These are wonderful videos! Thank you so much!
Great video to both you and Luciano!
And, do you choose CH to hold?
Love it!!!!!!!!!!!!!!!
Thanku u Luciano selvetti for delivering such immense knowledge to us💓