The number systems as a chain of containments, and intervals as pieces of R

0:00Five collections of numbers, five capital letters, and a great deal of confusion about how they fit together. The natural numbers: one, two, three, and onward. The integers: those, and zero, and all the negatives. The rationals: everything you can write as one whole number over another. The reals: every point on the line, with no gaps anywhere. And the irrationals: the reals that no quotient can reach. Most people meet these as five separate boxes to memorise.

0:36They are not five boxes. Four of them sit inside one another in a single chain, and the fifth hangs off the side of it. Getting that picture right saves you more work than any definition on the list. Here is the chain. Every natural number is an integer. Every integer is a rational. Every rational is a real. Each arrow is a promise about every single member, with no exceptions hiding anywhere.

1:09And here is what that promise buys you. Suppose you prove something about every real number. You have already proved it about every rational, because each rational is a real. And about every integer, and about every natural. One proof, four collections. That is not a coincidence and not a convenience. It is exactly what containment means, and it is why the chain is worth drawing before anything else in this subject.

1:42The rationals deserve a closer look, because their definition carries a condition that almost everybody skips past. A rational number is p over q, where p and q are integers, and q is not zero. That last clause is not fussiness. One over zero is not a number at all; it does not name a point on the line. Take the condition away and the description stops describing anything. So the rationals are not a list you memorise.

2:15They are a search. Given a number, ask: is there an integer on top, and a non-zero integer underneath, that produce it? If the answer is yes, the number is rational, and if no such pair exists, it is not. Four numbers usually get offered as rationals in disguise: minus five, five sevenths, three and a half, and minus eleven thirds. Look at them honestly. Five sevenths is already one integer over another.

2:50So is minus eleven thirds. Nothing whatsoever is disguised about either of them. The two that genuinely take a step are minus five, which becomes minus five over one, and three and a half, which is three plus a half, and that comes to seven over two. Two of the four, not four. This matters more than it sounds. A student told to hunt for difficulty in four places will invent it in the two where there is none.

3:22Now the irrationals, and they are built differently from everything else here. Every other collection was described by what its members are. The irrationals are described by what they are not: a real number that is not rational. Take the square root of two. Could it be a quotient? Here is the argument, and it is short. Any rational solution of x squared equals two would have to be a whole number that divides two.

3:53So the only candidates in the world are one, two, and their negatives. Square them: one, four, one, four. Not one of them is two. That list was complete, so nothing was missed, and the square root of two is irrational. That argument is finished, and it runs the same way for the square root of five: the candidates are one and five and their negatives, and none of them squares to five.

4:24But be careful about the next example everybody reaches for. Pi is irrational too. That is true, and the proof is nothing like the one you just watched. It is long, it is nineteenth-century, and it is not going to fit here. So when this video tells you the square root of two is irrational, it has shown you why. When it tells you pi is irrational, it is asking you to take that on trust.

4:52Those are two different kinds of statement, and it is worth knowing which one you are being handed. Four containments, then, and one relation that fails. Are the natural numbers inside the irrationals? No. And the disproof costs almost nothing. Take one. One is a natural number. One is also rational, being one over one. So one is not irrational. There is a natural number sitting outside the irrationals, and the containment is dead.

5:28One member did that. Two would have worked just as well, and so would three. That is the whole cost of disproving a containment, and it never gets more expensive than this. Five collections, and you can ask the containment question about any ordered pair of them. That is twenty questions. Twenty, not ten, because the question is not symmetric. The integers sitting inside the rationals tells you nothing about the rationals sitting inside the integers.

6:03Work through all twenty, and seven of them come out yes. Seven. Everything else fails, and every single failure has a member you can point at. Minus one, for the integers not being naturals. A half, for the rationals not being integers. The square root of two, for the reals not being rational. The picture is far emptier than four tidy arrows suggest. Second half of this video, and a change of subject.

6:35An interval is a piece of the real line, and its notation is the most information-dense thing in the topic. Here is what a bracket actually is. An interval is two numbers and two yes-or-no answers. The two numbers say where it starts and where it stops. The two answers say whether each of those endpoints is itself a member. That is the entire content. A round bracket means no, this endpoint is not in.

7:07A square bracket means yes, it is. Not bigger, not smaller, not stronger: one membership verdict about one number. Two answers, two ways each, gives four shapes, and those four are the whole system. Draw all four on the line, with the ends at minus three and five. Neither endpoint in: two hollow circles, round brackets on both sides. Both in: two filled circles, square brackets on both sides. First one in, second one out: filled on the left, hollow on the right.

7:44First out, second in: the other way around. Four pictures, and the only thing that changes is whether a circle is filled. The stretch between the circles is identical in all four. Every number strictly between minus three and five belongs to all four of these intervals. They disagree about exactly two numbers in the entire world, and those two numbers are the endpoints. So how long are they? Second end minus first, every time.

8:18Five minus minus three, which is eight. All four have length eight. Not approximately, not roughly: exactly the same number, four times. Which means length cannot tell these four apart, and a quantity that cannot tell two things apart is not identifying them. Length measures reach. It says how far the interval extends. It says nothing whatsoever about which endpoints came along. If you want to know what an interval holds, you read the brackets.

8:53The length is a different question, with a different answer, and running the two together is the most common error in this material. Intervals can sit inside intervals, and you settle that at the ends as well. Is the open interval from minus three to five inside the closed interval from minus seven to nine? Minus seven is below minus three, and nine is above five, so yes. The second holds the first entirely.

9:25Is the containment proper? That asks whether the second holds something the first does not, and you need one number to show it. Minus seven works: it is in the second and not in the first. But so does minus six. So does minus five. So does five itself, and nine, and every number between minus seven and minus three. That is worth stopping on. When you disprove a containment you need a witness, meaning one member on the wrong side.

9:59You do not need a particular witness. There is no canonical one, no official answer, nothing to circle in red. Minus seven is the usual choice because it is an endpoint and it catches the eye. The checker written for this video works through the ends mechanically and produces minus five instead. Both are completely correct. If you ever catch yourself hunting for the right counterexample rather than a right one, you have misread what a disproof is.

10:33Any member on the wrong side ends the argument the moment you name it. Time to translate, from words into brackets. The reals above minus four and at most six: minus four is out, six is in, so round then square, and the length is ten. Strictly between minus twelve and minus ten: both ends out, round on both sides, length two. From zero up to but not including seven: square then round, length seven.

11:07From three to four inclusive: square on both sides, length one. Ten, two, seven, one. Notice that the brackets and the lengths are answering different questions, and you work them out independently. Nothing about a bracket changes a length, and nothing about a length reveals a bracket. Now the other direction, and one pair that settles the whole argument. Open from minus three to zero: length three. Closed from six to twelve: length six.

11:41Six to twelve with only the right end in: length six as well. And minus twenty-three up to but not including five: length twenty eight. Look hard at the middle two. Both have length six. Both run between the same two numbers. And they are not the same set, because one of them holds six and the other refuses it. One number apart, and identical lengths. If length identified an interval, that pair could not exist.

12:13It does exist, so it does not. One more shape to know. Intervals can run off with no far end at all. The non-negative reals start at zero and never stop. The negative reals run up to zero from below. These get written with the infinity symbol, and here is the rule people get wrong. The bracket at an infinite end is always round. Never square. And the reason is not convention.

12:46A square bracket is a promise that the endpoint is a member, and there is no number down there to be a member. Infinity is not sitting at the end of the line waiting to be included or excluded. There is nothing to make a verdict about, so the notation refuses to pretend that there is. Last idea, and it is the one that repairs the intuition. Take the interval from zero to one thousandth.

13:16Its length is one thousandth. Tiny. How many numbers does it hold? Ask for five hundred of them, and here they come: halve towards the far end, then halve again, five hundred times. Every one of them lands inside. Nothing in that construction cared how long the interval was. Run it on an interval of length one thousand and you get five hundred, exactly the same way. Every interval, however short, holds endlessly many numbers.

13:50Length measures how far, never how many, and those two questions have never had the same answer. So: four collections in a chain, one hanging off the side of it, and seven containments out of the twenty you could ask about. An interval is two ends and two verdicts, and the brackets are where the verdicts are written down. Length is the reach between the ends, and it identifies nothing.

14:19To prove a containment you owe a promise about every member. To disprove one you owe a single number, and any number on the wrong side will do. Those two costs are wildly different. Almost every mistake in this material comes from paying the wrong one.