Thanks for watching everyone! Don't forget to submit your questions on my subreddit for a potential 10k Q&A video. www.reddit.com/r/anotherroof/ This concludes my series on defining numbers from the ground up, but there are plenty more videos to come.
@Spydragon Animations Simply use the definition with [(ac + bd, ad + bc)] for the integers. You can show that "negative * negative = positive". In particular, you can show that (- a) * (- b) = a * b if this is what you are unsure about
@Spydragon Animations The partition kind of "drops out" from the relation definition. Two elements are in the same partition if they are related. If you're being formal, you have to prove that (2,2) is related to (1,1) just as much as you have to prove that (-1,-1) is related to (1,1). I think, when he was at that part with the board and all the number pairs laid out in a grid, he was already being less formal - by that point he has shown that the definition is the same as our intuitive understanding of rational numbers, so he relied on our intuition to explain things visually without interrupting the rhythm with unnecessary proofs. I hope that helps, even if I'm a few months late :)
Why didn't you define the real numbers as an integer-subset with a highest element, representing the sum of 2^(each subset-member)? This allows you to avoid the rational numbers altogether. Also, why don't you consider quaternions numbers?
i gennuenly don't know if you are using that as an exemple of "really hard thing to explain" or of "nobody solved this yet". If it is the second option... Someone did.
As a graduate engineer I always thought I had a reasonable understanding of maths, now I see I was just given a box full of tools and shown how to use them, but with no explanation as to why the tools work. The biggest take for me from this video is that there is a fundamental difference between integers and naturals, the difference between the other sets being a bit easier to 'see'.
Yes, engineering curricula are designed that way because you want to go out into the real world and start solving real engineering problems (and earning money) as soon as possible! You don't want to spend 4 years learning all about the fundamentals of math used in engineering, then another 4 years learning about all the physics used in engineering, then another 4 years learning about all the chemistry used in engineering ... it's turtles all the way down, and you'll die before you graduate! Besides, if you want to learn more, you can always do so later, either in your spare time (like I'm doing now), or by going back to school.
In Icelandic, Imaginary numbers are called _"þvertölur"_ or "Across Numbers" (It sounds a lot less awkward in Icelandic) and complex numbers are broken down to _Vertical_ and _Horizontal_ components (Or just length and angle, if that's what you prefer); Because Imaginary numbers are just as real as any other..
I personally treat complex numbers more geometrically than algebraically, so I call them "spherical numbers" due to their intrinsic connection to spherical geometry. The name also hints at the possibility of numbers for other geometries, which do in fact exist, though they lack some of the nice features of the complex numbers.
Polish doubles down on the imaginary and the number "i" is called "jednostka urojona" or "Deluded Unit" as though we've lost our marbles, and gone fucking mad square rooting negatives.
@@Zaniahiononzenbei There are two main cousins of the spherical/complex numbers. One is called the "dual numbers," and the other is, depending on who you ask, either the "split-complex numbers," or the _"hyperbolic_ numbers." Knowing that the spherical unit squares to -1, I'll let you guess what its cousins square to.
I don't know why it's said we can never fully write down pi when it's easy: π Sure, writing it down as a decimal number is hard, but if you pick the right base it's easy! In base π it's just 10. Now, sure, that's an absurd base for everything else but sometimes sacrifices need to be made! More seriously - I am loving your series so far. Thanks for the hard work you've put into it.
We can do the same tricks that we used to jump up a number system, to define pi, though it takes some steps. We need to define powers, using ordered pairs. Then we can define rotations, where powers of a complex number are equivalent if they give you the same complex number but scaled. That leads naturally to exp and ln.
Firstly, pi in base pi is not pi. It is "10". As that's how base systems work. But base pi doesn't really make sense, you end up with an unevenly spaced number line.
My favorite is the completion as a metric space. It requires a lot of conceptual overhead (if compared to dedekind cuts), but it’s so natural and gives a good intuition why we use reals to begin with
This is one of the most geeky, pedantic and "why does anyone care" videos I've found so far; but yet it is so beautiful (and nicely presented). The definition of negative numbers just drops out so easily. You did manage to refrain yourself from the obvious joke: "Z from the German for 'zee integers'" ;)
I'd love to see an appendix video of sorts on transcendentals, though that certainly sounds like a challenge to make. This series has been great to watch, thank you
Transcendentals are a subset of real numbers, (or maybe complex if you want) so they can be expressed in the same way. You just define it as the set of all rational numbers less than it paired with the set of rational numbers more than it.
@@thomaspeck4537 No, the transcendental numbers are a subset of the complex numbers, because the algebraic numbers are a subset of the complex numbers. The algebraic numbers are defined as the algebraic closure of the rational numbers, and this closure necessarily includes i. Therefore, they are not a subset of the real numbers.
@@Cammymoop To explain the surreal numbers would require an entire video series. An appendix certainly would not be sufficient. You would need to start by changing the axioms of set theory, since in reality, if we are being rigorous, the surreal numbers do not exist in the Zermelo-Fraenkel axioms, even if you include the axiom of choice or its negation. You need axioms of set theory that are equipped with the ability to speak about proper classes, in order to define the surreal numbers.
This is such an incredible video massive props to you for taking on such a massive project and such difficult concepts and explaining them in a way that I actually feel like I understand :D
2:15 Illustrating the gap between the whole numbers as being foamy is such a splendidly intuitive way of conveying that only a proportion of that space can be represented with fractions.
oh my goodness when I saw a new proof video was out I got so excited I could hardly wait! Your channel is my comfort channel and I really appreciate the work you do, and the jokes and the math and the bricks and everything!! Thank you so much
30:45 “it’s like upgrading the operating system; it does everything the old one did, and more.” Little did he know that now upgrading your OS will remove features.
Here's a Dedekind cut definition of π: Lower set A = Union over all n in N of the sets: { r in Q : r < 4 * the sum from k = 0 to (2n + 1) of [ (-1)^k / (2k + 1) ] } Upper set B = Union over all n in N of the sets: { r in Q : r > 4 * the sum from k = 0 to (2n) of [ (-1)^k / (2k + 1) ] } You're welcome! :)
@@acrm-sjork That's because the rationals are not complete. Real numbers have the completeness property saying that every converging series has a limit point in the set.
@yanntal954 not sure that I got why infinite sum of rationals MUST or even CAN be irrational? Why not all sums of rationals converge to rationals? It is not contradict your statement
Really hoping to see Dedekind cuts definition of R, but it might be a bit too advanced for RU-vid. Definitely gonna be a banger video though, we all know it
Me too. I don’t have the requisite knowledge to understand the formal definition, so I was hoping there would be a more intuitive way to understand them without getting rid of the complicated details; this channel seems very good at that.
Dedekind‘s construction is relatively simple compared to completion of Q as a metric space. Most of the ingredients are readily available at this point in the series. But the completion thing connects most naturally to what we’re actually doing with reals
@@AnotherRoof You could perhaps take a moment to emphasize on what you wrote for the "R" block at 1:11:31, which might not be obvious from the chalkboard: That the "mother set" in this case is power_set(Q). . This is important because (1) it is a conceptual leap from the previous cases (where the "mother set" is just a cartesian pdt), and (2) R is somewhat different than N Z and Q in "size", in that its construction asks you to start from something much "larger". . (Explaining (2), "size" of infinite sets, etc... = topic for another video maybe?)
The most beautiful thing I saw in this video was that the question "why minus times minus is plus" never arises! It just is baked on the rules of the definition for the integers. Great stuff, marvelous presentation.
At 52:46, when the expression is multiplied by 'q' is important to notice that q > 0, otherwise the ''. But we know that q is positive because we chose the right representative of the rational number.
Loved the series! I don’t know how useful it is for someone who isn’t already familiar with this stuff, but I like the balance between intuition and rigor here
Well, I was familiar with all the building blocks (relations, power sets etc.). The definitions of each class of numbers were completely new to me, though, but the explanations were impeccable and I understood all of them. So definitely very useful for me at least:)
I'm actually proud of myself because I barely understood the information presented here. I'm going to have to binge watch your videos, I want to be sure I understand things of this nature.
I was hoping you might explore how C starts to lose properties compared to R. In a sense R is the top of the "number pyramid". Because you lose ordering in C, and the other way, you lose completeness in Q. If you go to quaternions you lose more (commutativity), and octonions lose even more (associativity).
You lose any natural or intuitive ordering in C, but it would certainly be possible to define an arbitrary ordering. Start by saying that anything that is closer to the origin on the imaginary plane than x+iy is less than x+iy, and then define an arbitrary ordering for all the points that are equidistant from the origin (e.g. saying that x is the highest within that set, and then going clockwise around the circle).
@@stephengray1344 Yes, but no. In the context of fields, when we talk about ordered fields, we talk about fields with an ordering such that the field operations are isotonic (in a standard sense) with respect to the ordering. The real numbers are an order field. The complex numbers are not.
@@stephengray1344 Also, specifying that an ordering does exist is sort of redundant, since by the well-ordering theorem, all sets can be well-ordered. I know the theorem requires the axioms of choice, but if you relax the conditions so that you only need a total order, then you can significantly weaken the axioms needed. Also, most mathematicians do accept the axiom of choice, despite it being controversial.
You have an amazing teaching style. I really appreciate your approach of laying out a few axioms and incrementally building up to every topic covered step by step. One thing that shows up in a number of math explainer videos is the Monster group, but without a background in group theory I've never had a foundation to be able to make sese of any of the writing about it, besides "big number interesting". Given your incremental approach to explaining things a series of videos on group theory that build up to the point of having the context and understanding to make sense of statements like "monster group M is the largest sporadic simple group", would be super nifty.
I've mentioned this in other comments and maybe I'll go into more detail in my 10k Q&A video, but my PhD was in group theory and I would love to do a series on the various families of finite groups. That said, because it a topic so dear to me, I want to gain more experience to make sure I do the topic justice!
@@gracenc vaguely Yorkshire? I'm from Lancashire -- and you're all lucky I'm not one of those proud Lancastrians who actually care about this sort of thing!
Any chance of you covering the Surreal numbers? It seems like a good place to go off of this. I'd also just like to say that I really like that you're using physical blocks for your axioms/theorems and stacking them together as you go. I think it's a great visual metaphor for how math concepts build off of a few foundational concepts.
do you mean the trancendental numbers?, or do you actually mean the surreal number that include: the number greater than every real nubmber and the rnumber less than every real muber but greater than zero?
@@plopgoot5458 He means numbers greater and less than any real number. Unfortunately the surreal numbers' construction require the use of transfinite ordinals and thus cannot be a set as there is no "set of all transfinite ordinals" in ZFC. They are a proper class.
@@crosseyedcat1183 Is that an issue, that they are a proper class as opposed to a set? It is an important caveat, yes, but I fail to see what is so "unfortunate" about it.
@@plopgoot5458 He is talking about the Conway surreal numbers, which extend the real numbers, and they are the "largest" ordered field, in the category theoretic sense.
In a very real respect, you're single-handedly undertaking what PBS Infinite Series was doing. :) I greatly appreciated that channel while they were active, but it turns out there are just not that many channels engagingly tackling topics at the foundations of mathematics outside of "headline-friendly" ones like trying to quickly explain Goedel's theorems ad nauseum. (I'm sure you'll get to those theorems, too, but if it's anything like this presentation, I expect it will be thorough and cohesive!)
It's funny you mention PBS infinite series, I was literally just thinking about that channel because I got a question asking if I'll ever make a video on the axiom of choice and I thought their video was good. I'm glad your enjoying my videos! A slight disclaimer: I'm not planning for every single one of my videos to be a deep dive into foundational topics, but I'll always try to approach ideas with my own style and an emphasis on rigour. Anyway thanks for watching :)
@@AnotherRoof Of course! I'm enjoying it, and it's been long enough since my university coursework that I appreciate the refresher either way. I find this stuff fascinating, even if it has very little bearing on my day to day life, but it's not at all like a bicycle; the gist of it may stick for ages, the details and practice not so much.
Your proof of the denseness of Q made me realise how denseness is related to division. It's so obvious now! Of course if a set of numbers (containing at least two members) is closed under averaging, it is dense! And if it's not closed under division of two, it's not going to be closed under averaging.
@@cadekachelmeier7251 I'm sure you're right, but that seems weird to me. It's just 2^(countable infinity), right? Why doesn't that just return countable infinity again?
@@vanderkarl3927 For reference, it's Cantor's Theorem. Honestly, it's pretty much the same logic as Cantor's diagonal argument. It's just more abstract because instead of reasoning about lists and numbers, you have to reason about any potential set and a set containing an infinite number of other sets. Other than that, it's just constructing an element that can't exist in any potential mapping, same as the diagonal argument. en.wikipedia.org/wiki/Cantor%27s_theorem
I knew you couldn't possibly be defining every number ever, but I could not imagine you would speak so fondly of Conway after leaving his surreal numbers out!
I've always loved educational videos, but this gets me some different levels of joy. I'm actually having popcorns right now while watching. Thanks so much for this series of videos!!
At 59:00, you explain how to construct a Dedekind set, defining it as being all the elements of the rationals which are smaller than a given number (r). You then prove that this set will not have a maximum value, because of the density property. This implies (as you say in the video) that both m and r (in your demonstration) are part of the rationals. However, you then says that "this is how we define the real value r". But wasn't r supposed to be a rational? Otherwise we cannot use the density we used to construct the Dedekind set
Let's say that we have some rational number m that is smaller than r (which is assumed to be irrational). Then we can add a positive rational number n to m to make a new rational number that is larger than m but still smaller than r. And we can make n as small as needed because there is always another rational number between 0 and n. That's as far as I got, but I'm not convinced it's a complete proof since I haven't ruled out the possibility of there being a real number that is "smaller than" all the positive rational numbers but larger than 0
Great for a start. Examples of additional stuff: 1)Quotient systems of numbers, for example Z modulo an integer n. In particular, if n= a prime p, the only interesting case to be honest, then Z modulo p is a field, i.e. it satisfies all axioms satisfied by Q and beyond. It's an example of a finite field; it's incomparable to N, Z, Q, R, C etc. It can be shown that ANY finite field is essentially unique (i.e. unique up to isomorphism) and has q=p^n elements, where p a prime and n any integer from 1 and up. It's symbolised F_q. The prime p is called the Characteristic. Every field has a characteristic which is defined to be the smallest positive number that annihilates (i.e. if multiplied with, it zeroes out) all elements, equivalently the multiplicative unit, if such a positive number exists; otherwise the characteristic is 0. So, characteristic p means that 1+...+1 (p times) is 0, and is the first time we get 0 if we add 1 a non-zero number of times to itself, equivalently all other numbers in the list of 1+...+1 etc. are distinct. Likewise if we add any other element to itself consecutively. Every field has characteristic 0 or a prime number p. In the first case it contains a copy of Z, thus of Q (considering fractions), thus it's an extension of Q. In the latter it contains a copy of F_p. So basically, if we stick to only Q, R, C we miss out on plenty of action. 2) Transcendental extensions: If F is a field and V a set of "variables" (could be finite or infinite) we can form the set of polynomials on the variables of V and call it F[V]. The elements of F[V] are called polynomials in V with coefficients in F. They are (including 0) the finite sums of monomials of F[V], who are in turn finite products of elements of V (could be empty, which gives 1), multiplied with an element of F (could be 1). If we copy the construction that gives Q from Z (it's called the "Field of Fractions" construction), we get the set of rational "functions" in F[V], symbolised as F(V). The idea is that F(V) is the smallest field containing F and the elements of V, which are considered ALGEBRAICALLY INDEPENDENT over F. This means they satisfy no non-obvious algebraic relation. In case V={x_1,...,x_n} this is just the sets of polynomials and rational "functions" in x_1,...,x_n repsectively. If furthermore n=1, then we get V={x} and we write F[x] and F(x) for the set of polynomials in x with coefficients in F and rational "functions" in x with coeffs in F. F(x) is called a simple transcendental extension. Basically, it behaves like π and e in R with rational coefficients, every algebraic relations using x and elements of F are trivial, i.e. they merely follow from the axioms of arithmetic, equivalently: a_nx^n+...a_0=0 => all coefficients are 0. 3) Algebraic Extensions: We can add "algebraic relations" to F(V) above using quotient constructions (an example below), so that the images of V are no longer algebraically independent. Alternatively, if F, L fields such that F is a subfield of L (like Q in R and R in C) then L is algebraic over F if every element a in L satisfies a polynomial equation with coefficients in F. C is an algebraic extension of R since every complex number has a polynomial (more below), R is DEFINITELY NOT an algebraic extension of Q, in fact almost everything is transcendental over Q. So, in the example of F[x] x a variable over F, we can take any polynomial m(x) (a prime/irreducible polynomial for good results) and form F[x] modulo m(x). This isn't a field if m(x) isn't irreducible, in fact if m(x)=p(x)q(x) then p(x) and q(x) modulo m(x) would be non-zero, but their product m(x) is 0 mod m(x) by definition. If m(x) is irreducible, then F[x] modulo m(x) is a field, similar to 1), and is the "smallest" field containing F such that m(x) has a root; clearly x mod m(x) satisfies m(x) itself. If we define x mod m(x)=: ρ, then we call the above set F[ρ]=F(ρ) the simple extension defined by a root of m(x). The reason the above equality holds is that inverses of non-zero elements already exist, due to the following: F[ρ] is the set of all (unique) linear combinations of 1, ρ, ρ^2,..., ρ^(n-1) with coefficients in F. This is being proven by Euclidean division of polynomials. So, any such non-zero combination, corresponds to a polynomial g with a degree smaller than the one of m(x), thus it isn't being divided by m(x), and since m(x) is prime it must be relatively prime, thus there exist polynomials s,t such that sg+tm=1, which modulo m(x) gives sg=1 modulo m(x) (same proof works in 1) ). So we could define "square root of 2", and any other algebraic number in general, without dedekind cuts necessarily (if we aren't interested in ordering numbers on a line) as follows: Start with Q, form Q[x], then consider m(x)=x^2-2 which can be shown to be irreducible, and form Q[x] mod m(x). We could call it Q[j]=Q(j), where j is an "imaginary number" such that j^2=2 (note it could also be minus square root of 2). It contains all elements a+bj where a,b are rationals; any such writing is unique and if we have a non-zero element, we can find the inverse by multiplying and dividing with the conjugate element a-bj. Same construction can give us C from R, just take x^2+1 as m(x). In general if L is an algebraic extension of F, and a an element of L, then the polynomial of F[x] of smallest degree having a as a root is called the MINIMAL POLYNOMIAL, and it also happens to be the unique polynomial which divides all such polynomials. If it is defined to be m(x), then the smallest field containing F and a, F(a) is going to be essentially the same thing as the above construction F[x] modulo m(x). The degree of m(x) is called the degree of a. In particular F(a)=F[a]. Note that if F=L then the minimal polynomial of a is x-a which has degree 1. We can keep adjoining elements to a field F. If a is a root of a (w.l.o.g. irreducible) polynomial p, then we form F[a]=F(a), and then if b is a root of the (irreducible over F[a] ) polynomial q, we can form F[a,b]:=(F[a])[b]. In theory, every algebraic extension can be obtained by simple extensions, if we "take the limit" for infinite-dimensional extensions. So we can have Q[sqrt(2), sqrt(3)] etc. Theorem: Any field extension can be seen as a transcendental extension followed by an algebraic one. So, wlog for more "numbers" we can take, for example, C(x,y), C with two extra algebraically independent elements, identified by a "surface" over C if you do algebraic geometry, then you add the relation, say y^2-x=0 and you get... something. 4) Algebraic closure. If you take the subfield of all complex numbers who are algebraic (over Q) you get the Algebraic Numbers. This is called the algebraic closure of Q, since every polynomial over Q has a root now (but without the additional transcendental elements of R and Q). You can always take the algebraic closure of any field! So, you can take e.g. F_3, the finite field with 3 elements, and form the algebraic closure of it, the smallest field with characteristic 3 s.t. every polynomial in it splits into linear factors. This is endless. You can form F_5[x,y,z], take field of fractions, then algebraic closure. The algebraic closure of C(x) deserves to be called the field of Algebraic Functions over C. 5) p-adics. He mentions them in another video I think? You can combine them with all the other examples. "p-adics" in F(x) will end up being Laurent series in x, another example of field. 6) Power series, Laurent series, Puiseux series, etc. 7) Infinitesimals, that is, Levi-civita fields, Hyperreals, Surreals, etc. I hope I didn't forget anything! PS: Of course I forgot; Quaternions, Octonions and more exotic constructions like that.
I had to pause this several times to revel in the mind blasts this gave me. It's insane how much sense it will makes now. Thank you for the concise explanation!
Honestly one of the best maths channels out there. You actually go into the weeds of how it operates, and it's always a joy following you along. You never know what strange things you'll encounter. This idea that an equivalence induces a partition is just an excellent little nugget of how things operate. I'm trained in physics myself, but have always had a strong fondness for maths, and I can't help but see an analogy between equivalence relations and how they induce a partition, and the event of a uranium235 atom absorbing a neutron and fissioning. The job of physicists over time has been to account for all the bits, to make sure that the physics we are calculating is actually accounting for everything - and it's led us down some strange rabbit holes. It feels likely that you can derive some form of symmetry between a relation having equivalence and some conservation law in physics (like energy conservation - the way it functions is analogous to the parts of an equivalence relation, thermodynamically). Just speculating here, but this is what I love with this channel - you provide tools to facilitate the thinking of these kinds of thoughts!
Complex numbers are just the algebratic closure of the topologic closure of the groupification under multiplication of the groupification under addition of the monoidisation of natural numbers provided with the standart z-module stracture.
Yes! I contemplated using this one in the video but decided that the average of the two rationals is more intuitive but I agree regarding the elegance.
I'm 10 years separated from my compsci BSci, and this video is an entertaining bit of bringing it back to the fundamental axioms of the inductive power of diagonalization. Great stuff.
41:05 I never saw why q != 0. I can see that if we were to include q=0 then this relation is no longer transitive. (p,q) ~ (r,s) : ps=rq => (1,1) ~ (0,0) (0,0) ~ (2,3) But (1,1) !~ (2,3)
There is a slight generalization of this construction in the mathematical subject known as commutative algebra. It is called the localization. However, even in this context, including 0 results in a structure that is not very interesting. en.wikipedia.org/wiki/Field_of_fractions#Localization
Can I just say, this is possibly one of the most societally beneficial channels on all of RU-vid. Access to quality education is slipping in most of the world. You're doing the best work with this!
John Conway is amazing, although I'll admit I mostly know him because of the Game of Life. As I recall, in an interview he expressed an annoyance that that was the thing for which he was known to many people (I believe it was a Numberphile video), but it's still a beautiful piece of maths (it's Turing complete too, as I recall), and exposed many people to cellular automata as a subject. Langton's Ant is another interesting automaton, and I got part way through trying to knit a scarf in the pattern of Rule 135, but never finished it.
*STANDING OVATION* WOW! I am astonished at how clear this breakdown (or rather, build-up) series was! I am also flabbergasted by how incomplete my own understanding of our entire number system was. This all made so much sense and I really feel like I have a much more fundamental grasp on what all of these things, which I thought I had already understood, actually are and mean. Thank you so so much for this!! I can not imagine a world where your channel isn't as much of a cornerstone of internet mathematics as 3b1b or mathologer; you truly deserve it. Please keep up the incredible work, I really appreciate all of the time and effort that went into this as well as the risk you took on in embarking upon this journey. Can not wait to see what you do next!!
For as much as u didn't like the division demo, I liked it. It was cool to see how division creates a reflection along the zero line, literally dividing the integers in two. It reminded me of matrix transformations, and I love anything that involves transformations
Extremely well done as always. My freshman son enjoyed it with me. I think he got a little bit lost on the definition of reals but got through it. Interestingly, he struggled with complex simply because he'd never been introduced to them. As soon as I explained them he looked back at your definition and saw it immediately. Thank you so much for all the hard work!
Wow this was an awesome culmination to an awesome series. I am not a mathematician, but I have taken some advanced courses (Real Analysis, Measure Theory, etc), so I knew more or less what you were talking about with all the "hand waving". I had never seen this conception of numbers, and I loved it! Keep up the great work!
Thank you so much! Really instructive masterpiece! Personally I would have added a remark on operator-preserving relation on top of equivalent relation ("congruent" relation), and on embedding (N in Z by identifying n with [(n,0)], e.g.), but I am sure that you intentionally omitted this. Only one high level critique: Your proof sketch of "completeness" of R appears to me to show only density rather than completeness on the real line (for every a
I feel like you keep bumping right up against group theory and algebra. Hopeful to see some vids on those topics some day from you! Fantastic video as usual.
@@AnotherRoof it's also my favorite area of mathematics, and I am REALLY enjoying your videos and style and just everything, hence my yearning for the topic. Whenever you do get to it, I'm sure it will be fantastic. Thanks again!
I've not looked at group theory at all but I kept feeling like category theory was around the corner. Not that I have a much of an understanding of that either :). I'm sure I'd enjoy your take on either/both.
@@brendanobrien4095 I only have an undergrad degree in mathematics so I haven't studied category theory at all (that's a graduate topic, 100 percent), but my understanding of it is that it is essentially another layer of abstraction on top of group theory/algebra (those terms are not at all interchangable, but whatever; it's fine for our purpose here.) Like, my understanding is that category theory is essentially an algebra of algebras. But again: I don't have any formal training in category theory. So if I'm wrong, that is not unexpected. Also, you should look at group theory. If you know what category theory is, you ought to know what group theory is :p. No judgment here, just a recommendation. It would be a cool video idea for sure though. Not sure how much another roof knows about category theory, but any video he wants to make is fine by me, cause he's killing it. I've rewatched every video at least twice and just loved every second of it.
@@kruksog You should understand that my path to this stuff is atypical. I am a software engineer and I took none of this in college. I came across category theory by way of lambda calculus which I was studying for applications in functional programming paradigms. This is the first I've even heard of group theory and I'm very interested if it relates to these topics.
At 59:35, since we want to use the Dedekind cuts to define the real numbers, if r is not a rational number (1:00:49) and not a member of Q, how can we then prove that there exists another rational number between m and r? Also, how are we certain that in the Dedekind cut, there aren't two or more irrational numbers? (great video btw)
Thanks this was very thorough, and I can see how various ideas take the 'blocks' of number theory that you explain so well, and build a whole set of new theories out of them... i find this actually useful as a presrequisite to the PBS infinity series... Thanks!
I was fairly familiar with the construction of the rationals from integers and complex numbers from rationals, but had never seen/considered the similar construction of integers from naturals! In hindsight, the parallels seem obvious, but I guess I'd never really considered "negativeness" with enough of a critical, "what _are_ you, really?" attitude.
The “A must have no greatest element” thing from the dedekind cut reminds me of Zeno’s dichotomy. To walk out of a room, you first have to walk halfway to the door, then half of the remaining distance, then half of the remaining distance, then half of the remaining distance, etc., etc.
Zeno's paradox, in whichever form it is presented, makes a number of unfounded and erroneous assumptions about both time and space that are irrelevant to the purely conceptual space of the number line.
One of my high school teachers described it as getting half the distance to his girlfriend, then half of that distance, etc., with the conclusion that "Eventually I get as close as I need to," which sort of ties it into the concept of a limit.
Actually you have skipped redefining operations on real numbers. But some of them are not trivial like multiplication or division, so it would be interesting to have it in the video. But I understand that the video is quite long. Great series!
Yeah I just kind of glossed over that and said "it can be done" haha -- probably should have just displayed them at least for a few seconds but yeah, 80 minutes was long enough!
I’ve been eagerly, perhaps a bit impatiently, awaiting this video. I have to say that it was worth the wait. With the length and depth of this video, you certainly didn’t disappoint. Thank you! I can’t wait for the next one.
0:25 Temperature and altitude are bad examples for negative numbers. They have an arbitrary zero value. Imagine it like this: Define 0 °Apples as 10 Apples. A basket with 3 Apples in it contains −7 °Apples. It that sounds like nonsense to you, it’s because it is, but temperature and altitude work exactly like that, it only feels natural because of exposure. There is absolute zero of temperature (0 Kelvin) and altitude (the center of the earth). Negatives only exist in differences; if you ask me “how much taller is your brother than you are” the answer is −2 cm because he’s actually the smaller one, still body height can’t be negative.
I would argue that extended groups like quaternions and such are useful because some problems really have more than 2 independent axes/quantities and the ability to still treat those combined elements as numbers is useful. However they are definitely even more esoteric than complex numbers
We use the theory of Clifford algebras for that. Quaternions are almost never used in practice, because although they were originally invented with the purpose of generalizing complex number algebra to higher dimensions of space, mathematicians later learned that they are not actually the most conceptually useful way of doing that, nor are they the most practical to use either.
43:59 heh, I came back to this video after a while and noticed that you reached right through the naturals and the integers to grab the chalk to partition out the rationals, which is rather pleasing to my sense of irony.
Operations in whole numbers and in rationals are defined in terms of representatives of the classes, so it should be mention that one should prove that the result doesn't depend on the choice of representatives. I think that wasn't emphasized in the video.
I hadn't thought about constructing integers as a cartesian product over the natural numbers combined with the paired addition/subtraction relation. It's a very nice way of defining them.
Although the diagram for defining the set of rational numbers is much messier than the diagram for defining integers, people are far more familiar with the former. "Every one" is already familiar with the concept that "1/2", "3/6" and "5/10" are the same rational number. That's primary school material. But that [(1, 2)] and [(3, 4)] are the same integer will confuse most people.
I was curious about the proof that the real number are continuous, at 1:06:50 you hint at it but it seems to me that all you are proving is that the real numbers are dense, as in there's never an empty gap. But the rationals share that property, you even proved it, yet there were numbers that were missing. How can we be sure that there are no numbers hiding between the real numbers?
Agreed, it seems like he left out the key part of the proof. As I understand it the way to prove the real numbers are continuous is to do the dedekind cut procedure again using the real numbers, and then you can show that the resulting number is just another real and doesn't give you something new.
It's very subtle. In one sense, there are numbers between the real numbers, because you can put more numbers on the line with constructions like Robinson's Hyperreals or the Surreal Numbers, etc. But in another sense, the reals are gapless in a way those things are not: for instance, if you have a nonempty set of reals that doesn't grow arbitrarily big (e.g. maybe they're all less than a million), then there is a "cap" real that is greater than or equal to everything in the set, but smaller than any other candidate. (The cap may or may not lie in the original set. Think about intervals like (0,1) and [0,1], both of which have cap 1.) This "dedekind completeness" works for the reals and stops working if you try to stuff even more numbers inside.
The short answer is, it depends what you mean by a number. There are number systems (see hyperreal numbers or surreal numbers) in which there are other numbers hiding between the real numbers. What is meant in this context is that the real numbers are "complete" (this is a technical term). There are several equivalent definitions of what this means, but perhaps the simplest one to see (relative to this video) is that if we were to repeat the process of taking Dedekind cuts we wouldn't get any new numbers. That is to say, instead of the sets A and B consisting of rational numbers, say they consisted of real numbers. Then the set of Dedekind cuts of this form would just be the reals again and wouldn't contain any new numbers.
I watched this while doing my dishes, and when the number line came up, I thought it was CG. When I looked up and saw you hang the number labels on it, it was almost a jump scare! 😂
I would like to explain the concept of negative numbers with an old joke: A mathematician leaves an empty room. He sees two people entering this room. Then five people leave the room. And the mathematician thinks: Ok. Now three more people have to enter the room, and it is empty again.
This is actually amazing, I don't think any course actually goes through the building of all the number sets. Also, I noticed that if we let all numbers of the form (1,0) in rationals definition (we are never dividing after all), then we have natural notion of infinity (or rather, the point at infinity from projective planes). However, (0,0) still has to be excluded because it breaks transitivity of the equivalence (curse you indeterminacy). (1,0) + (a,b) = (1b + 0 a, 0b) = (b,0) = (1,0) (1,0) - (a,b) = (1b - 0a, 0b) = (b,0) = (1,0) if b is not 0 i.e. (a,b) is not (1,0) (a,b) - (1,0) = (a0 - b1, b0) = (-b,0) = (1,0) if b is not 0 i.e. (a,b) is not (1,0) (1,0) * (a,b) = (a*1, b*0) = (a,0) = (1,0) if a is not 0 i.e. (a,b) is not 0 (1,0) * (0,b) = (1*0, 0b) = (0,0) = 0 (a,b) / (0,c) = (a * c, 0b) = (ac,0) = (1,0) (a,b) / (1,0) = (a*0, b*1) = (0,1) = 0 But the following dont work because indeterminacy (1,0) - (1,0) = (b,0) - (a,0) = (b*0 - 0*a, 0*0) = (0,0) (1,0) * (0,1) = (1*0, 0*1) = (0,0) Of course, addition and subtraction have lost associativity, and (1,0) has no inverse, like (0,0) = (0,1).
you say not to worry about tagging ordered pairs with 1 and 2 because we can use different numbers, but what if we want to use the set of all natural numbers? We’d need new tags to avoid duplicates. And then what if we use the set of all of those tags, and so on. There must be a cleaner solution
Yes, I wish I'd never used the Hausdorff ordered pair but I felt it was more intuitive to non-mathematicians. Anyway, for a set A it is always possible to find a thing which is not contained in A, namely A itself by the axiom of regularity. So if worse comes to worst we can use the set A itself to tag the elements of A. So an ordered pair (x,y) with x in A and y in B could be encoded {{a,A},{b,B}}. Hope this helps!
There are sets not in Naturals for example sets {0,2} and {1,2} you can use this two for tagging. This set are distinct from {},{0},{0,1},{0,1,2} which are the set represented by 0,1,2 and 3. Sets {0,2} and {1,2} are not in Naturals.
@@AnotherRoof this works except that a lot of the cartiesian products you use in this video are of the form A×A. So you probably want something like A and {B} as the two tags (which should work for this video/most applications in math).
Dude I'd be absolutely enamored with this video if I was high. Even while sober, this is amazing. You have a great way of presenting information with a combination of tactile visuals and animation. I especially love when you give a few seconds for the viewer to apply a newly introduced concept. It builds confidence and I'm sure it helps with retention. Once I finish, I'm sending this to my old modern algebra teacher. Him teaching these concepts, where all of our assumptions are stripped away and we re-establish the fundamentals with a deep understanding is why I'm majoring in mathematics. I will be binging your vids.
Do you see how the "equivalence class" is astutely changed to "collection or set of relatives"? There is plenty of time to learn the usual expression. By now it's time to learn and boy we are doing it!
This was amazing. I think the next logical step is to define complex numbers as the pair of pairs (matrix) [[a, -b], [b, a]]. This allows you to use the relations to define more equivalence groups, and create new kinds of numbers, like the split-complex and the dual numbers. It also allows you to define rotation, boost (hyperbolic rotation), and "dual derivatives", which are like "rotation" around a point at infinity. The latter allows you to define calculus with just sets. Dual numbers are really cool. Even though you just need *C* for algebraic closure, dual numbers give you a "back door" to defining calculus without needing limits or any of the usual calculus introductions. You just need ε, which is [[0,1],[0,0]] and the rules you have thus defined (though I guess you need functions first!) Of course, ε ties in with surreal numbers, which is also a logical extension of this video!
A really nice video. I love the sense of building on foundations, with literal examples of your building blocks. The time flew by, although I ended up watching it in a couple of bites, due to watching it during the working day. Couldn't wait until the end of the day! Really interested to see where you go next. And very happy to think that this community can be part of your escape plan. I only lasted a year as a maths teacher, and it took my wife twenty years to escape, so wishing you well in your attempt to make it out.
technical terms for further reading 1.Construction of natural numbers using ZF axioms 2,3.Quotient field of an integral domain (abstract algebra/ ring theory) It's awesome that he break down the whole general construct, picked out the ideas and names that will be used in this specific case (Rational domain from integer ring), explained them using easy logic, and provide graphic while going through the process. Very impressed. I always thought integers as forming a ring from natural numbers, its intuitive to just add all the negatives to fill out the (additional inverse) gap in natural number. Never thought it can be constructed like that using this method, this easily 4.Dedekind cut, this might be the most famous among math major students First thing to learn in Calculus, the real numbers, the completeness and continuity of real numbers The proof (That the set of real number is continuous) is very elegant, the use of logic, the rules of dedekind cut, chasing down every single case into a contradiction. My first textbook skipped this part, stating only the theorem and not the proof. On my second learn through, I got another textbook, and the proof is beautiful, just like you demonstrated in the video. Do read about this, yes it is very technical, but once you go through the underlying logic, everything in the theorem is just about splitting proof into cases This video is AMAZING! You're leaving the hard technical terms out yet this is so rigorous in logic! Bravo!
This is all the numbers necessary to gaplessly add, subtract, multiply, divide, and take roots. I'd argue quaternions and octonions still qualify as numbers because they are still sufficient for all of those properties. If you go above octonions, these fundamental operations start breaking again. Of course, if you insist that numbers must be associative, then things already break with the octonions, and if you insist that they must be commutative, quaternions are already broken. Meanhile, if you are ok with not having gapless division, there are suddenly LOADS of numbers. Just no limit of them. Defining addition and multiplication is usually quite easy. Subtraction and Division each pose problems that, if you don't require those operations to be defined without gaps, suddenly Pandora's Box is opened and you get infinite variations of what's possible. A really simple way to extend numbers arbitrarily is to take your favourite polynomial equation and declare some constant to behave according to it. For instance: 0 = e² + 1 (a + e b)(c + e d) = a c + (a d + b c) e + e² b d but according to the above equation, e² = -1 a c - b d + e (a d + b c) whoops, redefined the complex numbers! how about e² = 0? a c + e (a d + b c) these are the dual numbers! Very useful for calculus, it turns out. Try plugging it into, for instance, a taylor series to see why. However, they don't generally have division: If (c + d e) is the inverse of (a + b e), that means a c + e (a d + b c) = 1 Let's assume a!= 0 so then c = 1 / a and a d + b c = 0 a d + b / a = 0 a² d + b = 0 d = -b/a² so 1/(a + b e) = 1/a - e b/a² Now, let's assume a=0 e (b c) = 1 this has no solution! If a = 0, the number can't be divided. You could go wilder, and define: 0 = u e² + v e + w (where u, v, w are real) in which case you get a multiplication like a c + (a d + b c) e - (v e + w)/u b d a c - w/u b d + (a d + b c - v/u b d) e Or if you want to, you can establish higher order relationships, which nets you extra dimensions: e³ = -1 (a + b e)(c + d e)(f + g e) = a c f + (a d f + b c f + a c g) e + e² (b d f + a d g + b c g) +b d g e³ = a c f - b d g + (a d f + b c f + a c g) e + (b d f + a d g + b c g) e² etc. Every single polynomial thus defines a potential "number system" with its own multiplication rules. Addition and subtraction here always work in the usual way. Division is likely a restricted notion. But all of these will remain commutative and associative. Unless you specially denote it. For instance, you could go: i² = -1 j² = -1 k² = -1 i = jk j = ki k = ij using a bunch of (in this case six) polynomial rules to establish how they work together, and you get their entire behavior. For instance, using these rules, note that: ijk = jki = kij = -1 This one carries over from complex numbers: iiii = jjjj = kkkk = 1 And here is how anticommutativity comes in: i = i jjjj = ij jjj = k jjj = kj j² = - kj Voila, you just got quaternions! This way you can generate as many number systems as you desire.
Love how you say you're not a computer guy but write equality as :=, the same trick we use in computer languages to differenciate an assignation (set the value to the value of the math result [:=]) from a comparison test (is left member equal to right member [==]) !
@@FrostDirt i don't come from math, as many in my situation. But there is a little of me in every high power UPS around the world. And i've been paid "generously", like 75k€ in 3 years... And they don't understand why i won't be back ^^
I like defining the reals as equivalence classes of Cauchy sequences. I think that’s a much more natural definition. It makes the completeness a bit more obvious.
