1.3 Revisiting Irrational Numbers
Assume that $\sqrt2$ is a fraction in lowest terms and follow the logic until it breaks: a proof by contradiction, powered by the Fundamental Theorem of Arithmetic, shows that $\sqrt2$, $\sqrt3$, $\sqrt5$ and numbers such as $3+2\sqrt5$ are irrational.
Draw a square floor tile with sides of exactly $1$ metre. By Pythagoras, its diagonal is $\sqrt{1^2+1^2}=\sqrt2$ metres. A calculator says $\sqrt2=1.414213562\ldots$, and a tiler would happily call it $141.4$ cm. But is there a fraction that is exactly $\sqrt2$? Perhaps a cleverer fraction such as $\tfrac{41}{29}$ or $\tfrac{577}{408}$ does the job? (Squaring them gives $1.99881\ldots$ and $2.0000060\ldots$ โ very close, but not $2$.)
In Class 9 you were told that $\sqrt2$ is irrational: no fraction equals it. Trying fractions one by one can never settle this, because there are infinitely many to try. In this section we prove it, using a powerful style of argument called proof by contradiction and the Fundamental Theorem of Arithmetic from Section 1.1. The same method then shows that $\sqrt3$, $\sqrt5$, $5\sqrt2$, $3+2\sqrt5$ and many other numbers are irrational. These proofs are asked in board examinations every year, so we write each one out step by step.
Rational and irrational numbers, and proof by contradiction
A real number that is not rational is called irrational: it cannot be written as $\dfrac pq$ with integers $p$, $q$ and $q\ne0$.
Every rational number can be written in lowest terms, that is, with $p$ and $q$ co-prime ($\operatorname{HCF}(p,q)=1$): just divide $p$ and $q$ by their HCF. For example $\dfrac{84}{120}=\dfrac{84\div12}{120\div12}=\dfrac{7}{10}$.
Examples of rational numbers: $\tfrac34$, $-5=\tfrac{-5}{1}$, $0=\tfrac01$, $\sqrt{49}=7$, $\sqrt{\tfrac{9}{16}}=\tfrac34$. To show a number is rational, it is enough to write it as a fraction. To show a number is irrational, we must show that no fraction at all can equal it. That is where proof by contradiction comes in.
- Assume the opposite of what you want to prove. (To prove "$\sqrt2$ is irrational", assume "$\sqrt2$ is rational".)
- Write down what the assumption gives you in symbols (for example $\sqrt2=\tfrac ab$ with $a$, $b$ co-prime).
- Reason correctly, one small step at a time, using facts you know are true.
- Reach a statement that is impossible โ it contradicts the assumption or a known fact.
- Conclude: the assumption must be false, so the statement you wanted is true.
An everyday version: "If my friend had come home, her shoes would be at the door. Her shoes are not at the door. So she has not come home." Here the assumption "she has come home" led to something that is not true, so we reject it.
When a prime divides a square
Example 1 is the case $p=2$ of a more general fact. It works for every prime, and it is the key step in all the proofs of this section.
Proof. By the Fundamental Theorem of Arithmetic (Theorem 1.1), write $a$ as a product of primes, $a=p_1\times p_2\times\cdots\times p_n$, where some of the primes may be repeated. (If $a=1$ there are no primes, $a^2=1$, and no prime divides $1$, so there is nothing to prove.) Squaring,
$$a^2=p_1\times p_1\times p_2\times p_2\times\cdots\times p_n\times p_n .$$This is a prime factorisation of $a^2$, and by uniqueness it is the only one. Now $p$ divides $a^2$, so $a^2=p\times m$ for some positive integer $m$; factorising $m$ gives a prime factorisation of $a^2$ that contains $p$. By uniqueness, $p$ must be one of the primes $p_1, p_2, \dots, p_n$. But each of these divides $a$. So $p$ divides $a$. $\blacksquare$
(b) A positive integer $a$ is such that $5$ divides $a^2$. Show that $25$ divides $a^2$.
(c) $4$ divides $36=6^2$. Does $4$ divide $6$? Why does this not contradict Theorem 1.4?
"$p$ divides $a^2$, so $p$ divides $a$" is true when $p$ is prime, but can fail when $p$ is composite: $4$ divides $6^2$ but not $6$; $9$ divides $3^2=9$ but not $3$; $8$ divides $4^2$ but not $4$. In every proof below, check that the number you apply the theorem to is a prime ($2, 3, 5, 7, 11, \dots$).
$\sqrt2$, $\sqrt3$ and $\sqrt5$ are irrational
Now we can prove the famous result. Read the proof slowly; every later proof in this section copies its shape.
Proof.
- Assume the opposite: suppose $\sqrt2$ is rational. Since $\sqrt2\gt0$, we can write $\sqrt2=\dfrac ab$ with positive integers $a$, $b$ that are co-prime (lowest terms).
- Then $\sqrt2\,b=a$. Squaring both sides gives $2b^2=a^2$. So $2$ divides $a^2$.
- $2$ is prime, so by Theorem 1.4, $2$ divides $a$. Write $a=2c$ for some positive integer $c$.
- Substitute in $2b^2=a^2$: $2b^2=4c^2$, so $b^2=2c^2$. So $2$ divides $b^2$, and by Theorem 1.4 again, $2$ divides $b$.
- Contradiction: $2$ divides both $a$ and $b$, so $a$ and $b$ have the common factor $2$. But we chose $a$ and $b$ co-prime.
- The contradiction came from assuming that $\sqrt2$ is rational. So $\sqrt2$ is irrational. $\blacksquare$
Another way to see why step 2 is impossible uses prime "fingerprints". In $a^2$ the prime $2$ appears an even number of times (each copy of $a$ gives the same number of $2$s). In $2b^2$ it appears an even number of times from $b^2$, plus one โ an odd number of times. By uniqueness of prime factorisation, $a^2$ and $2b^2$ can never be the same number. Interactive 1 shows this fingerprint test live, while it hunts for a fraction that hits $\sqrt2$.
Nothing in the proof of Theorem 1.5 used anything special about $2$ except that it is a prime. Replace $2$ by any prime $p$ and every step still works.
1. Forgetting "co-prime". The contradiction is "$a$ and $b$ have a common factor". If you never said $a$ and $b$ are co-prime, there is nothing to contradict.
2. Skipping the reason. Write "$3$ divides $a^2$, so $3$ divides $a$ because $3$ is prime (Theorem 1.4)". Examiners look for that reason.
3. Stopping half-way. Showing that $3$ divides $a$ is not enough; you must substitute $a=3c$ and show that $3$ divides $b$ too.
4. Using a calculator. "$\sqrt3=1.7320508\ldots$ does not look like a fraction" is not a proof. A calculator shows only a few digits, and the fraction $\tfrac{1351}{780}=1.7320512\ldots$ already shares its first six digits with $\sqrt3=1.7320508\ldots$
5. Trying it on a perfect square. The method does not work for $\sqrt4$ or $\sqrt9$: $4$ and $9$ are not prime, and indeed $\sqrt4=2$ and $\sqrt9=3$ are rational.
Numbers built from irrationals: $5\sqrt2$, $3+2\sqrt5$, โฆ
Once we know $\sqrt2$, $\sqrt3$, $\sqrt5$ are irrational, many more numbers follow quickly. The idea: if the bigger number were rational, we could "undo" the arithmetic and make $\sqrt5$ (say) rational, which we know is false.
- $r+s$ and $r-s$ are irrational;
- if $r\ne0$, then $r\times s$ and $\dfrac sr$ are irrational.
Why. Suppose $r+s=t$ were rational. Then $s=t-r$, a difference of two rational numbers, which is rational โ but $s$ is irrational. Contradiction. In the same way, if $r\times s=t$ were rational with $r\ne0$, then $s=\dfrac tr$ would be rational. (The condition $r\ne0$ matters: $0\times\sqrt2=0$ is rational.)
Theorem 1.7 says nothing about two irrational numbers, and for good reason: $\sqrt2+(-\sqrt2)=0$, $\;\sqrt2\times\sqrt8=\sqrt{16}=4$ and $(3+\sqrt5)+(3-\sqrt5)=6$ are all rational. Before you decide, simplify the expression to the form $a+b\sqrt n$.
Exercises
- 1 $\sqrt{81}$
- 2 $\sqrt{12}$
- 3 $\tfrac27+\sqrt3$
- 4 $\sqrt5\times\sqrt{20}$
- 5 $(3-\sqrt2)+(3+\sqrt2)$
- 6 $(3-\sqrt2)(3+\sqrt2)$
- 7 $(2+\sqrt5)^2$
- 8 $\dfrac{\sqrt{72}}{\sqrt2}$
- 9 $\sqrt{\tfrac{9}{16}}$
- 10 $\sqrt2+\sqrt8$
- 11 $(\sqrt3+1)^2-2\sqrt3$
- 1 Prove that $\sqrt5$ is irrational.
- 2 Prove that $\sqrt7$ is irrational.
- 3 Prove that $3\sqrt2$ is irrational.
- 4 Prove that $2+\sqrt3$ is irrational.
- 5 Prove that $4-3\sqrt2$ is irrational.
- 6 Prove that $\dfrac{1}{\sqrt3}$ is irrational.
- 7 Prove that $\dfrac{\sqrt2+5}{3}$ is irrational.
- 8 Prove that $\sqrt6$ is irrational. (Hint: start as for $\sqrt2$ and use the prime $2$.)
- 9 Challenge. Prove that $\sqrt2+\sqrt3$ is irrational. (Hint: if $r=\sqrt2+\sqrt3$ were rational, square $\sqrt3=r-\sqrt2$.)
- 1 A positive integer $a$ is such that $7$ divides $a^2$. Show that $49$ divides $a^2$.
- 2 Find a positive integer $a$ such that $8$ divides $a^2$ but $8$ does not divide $a$. Why does this not contradict Theorem 1.4?
- 3 True or false? "The sum of two irrational numbers is always irrational." Give a reason or a counter-example.
- 4 True or false? "The product of a non-zero rational number and an irrational number is irrational." Give a reason.
- 5 Find two different irrational numbers whose product is rational but whose sum is irrational.
- 6 If $x=2+\sqrt3$, show that $x^2-4x$ is rational, and find its value.
- 1 A square floor tile has side $30$ cm. Find its diagonal exactly and to $2$ decimal places. Is the exact length rational or irrational?
- 2 A square garden plot has area $50\ \text{m}^2$. Find its side exactly and to $2$ decimal places, and say whether it is rational. For which whole-number areas from $40\ \text{m}^2$ to $60\ \text{m}^2$ is the side a rational number of metres?
- 3 Modelling. A sheet of A4 paper measures $210$ mm by $297$ mm. The A-series is designed so that long side $\div$ short side $=\sqrt2$ (then folding a sheet in half gives the same shape). Show that $\tfrac{297}{210}$ is not exactly $\sqrt2$ by comparing $297^2$ with $2\times210^2$, and explain why no rectangle whose sides are whole numbers of millimetres can have ratio exactly $\sqrt2$.
- 4 A right triangle has legs (the two shorter sides) of whole-number lengths. Find the hypotenuse and say whether it is rational, when the legs are (a) $1$ cm and $2$ cm (b) $6$ cm and $8$ cm (c) $2$ cm and $2$ cm (d) $5$ cm and $12$ cm.
How this is examined: almost every board paper has a 2- or 3-mark proof that $\sqrt p$ ($p$ prime) or a number such as $a+b\sqrt p$ is irrational, plus 1-mark questions that test whether you simplify before deciding. In a proof, the marks are for the assumption (with "co-prime" where needed), the squaring step with its reason, and the stated contradiction.
E1. Board ยท 1 mark Which of the following is irrational? (A) $\sqrt{\tfrac{49}{64}}$ (B) $(2-\sqrt3)(2+\sqrt3)$ (C) $\sqrt2\times\sqrt{18}$ (D) $\sqrt3+\sqrt{12}$
E2. Board ยท 1 mark Assertion (A): $\sqrt2\times\sqrt8$ is a rational number. Reason (R): The product of two irrational numbers is always irrational.
(A) Both A and R are true and R is the correct explanation of A. (B) Both A and R are true but R is not the correct explanation of A. (C) A is true but R is false. (D) A is false but R is true.
E3. Board ยท 2 marks Given that $\sqrt3$ is irrational, prove that $5-2\sqrt3$ is irrational.
E4. Board ยท 3 marks Prove that $\sqrt{11}$ is irrational.
E5. Board ยท Case study A square lawn. The gardeners of a school in Kochi mark out a square lawn of area $72\ \text{m}^2$ and plan a straight path along one diagonal.
(i) Find the side of the lawn in simplest form. Is it rational or irrational? (1 mark)
(ii) Find the length of the diagonal path. Is it rational or irrational? (1 mark)
(iii) Fencing is needed round the whole lawn. Given that $\sqrt2$ is irrational, prove that the length of fencing is irrational. (2 marks)
Research question. No fraction equals $\sqrt2$ โ but how close can a fraction get, and how close do real A-size paper sheets come to the ideal ratio $\sqrt2$?
Why it matters. A clay tablet from ancient Babylon (about 1800 BCE, now at Yale University) gives $\sqrt2$ correct to five decimal places, and the Indian ลulba-sลซtras, written for laying out fire altars, give the approximation $1+\tfrac13+\tfrac{1}{3\times4}-\tfrac{1}{3\times4\times34}=\tfrac{577}{408}$. Today the international paper standard ISO 216 (A4, A3, A5โฆ) uses the ratio $\sqrt2$, so that enlarging A4 to A3 on a photocopier needs no trimming. Engineers, printers and computer programmers all work with good rational approximations of irrational numbers, because measurements and computers can store only fractions.
What to do.
- In a spreadsheet, put $p=1$, $q=1$ in the first row. In each new row use the rule $p_{\text{new}}=p+2q$ and $q_{\text{new}}=p+q$ (in cells:
=A2+2*B2and=A2+B2). Make $10$ rows. - In two more columns compute $\tfrac pq$ and $p^2-2q^2$. What do you notice about $p^2-2q^2$? Why can it never be $0$?
- Measure an A4 sheet (and an A5 or A3 sheet if you have one) with a ruler, to the nearest millimetre. Compute long side $\div$ short side and compare with $\sqrt2\approx1.41421$.
- Which fraction from your table is closest to your measured ratio? Is the paper more accurate than your ruler?
- Fold the A4 sheet in half. Measure again and check the new ratio.
Tools: Ruler, A4 sheet, spreadsheet (Google Sheets / Excel) or calculator
Hand in: Your table of $10$ fractions with the column $p^2-2q^2$, your paper measurements and ratios, and a paragraph (4โ5 sentences) explaining why a sheet of paper can be made close to $\sqrt2$ but never exactly equal to it.
Going further. Change the rule to $p_{\text{new}}=p+3q$, $q_{\text{new}}=p+q$ and look at $p^2-3q^2$. Do the fractions approach $\sqrt3$? Try to find a rule for $\sqrt5$.
๐งช Self-check quiz
Q1. Which of these numbers is rational?
Q2. To prove that $\sqrt3$ is irrational, we begin by assuming $\sqrt3=\tfrac ab$, where $a$ and $b$ are
Q3. Let $p$ be a prime and $a$ a positive integer. If $p$ divides $a^2$, then
Q4. Which statement is always true?
Q5. The number $(3+\sqrt5)(3-\sqrt5)$ is