Problem Statement:

Prove that $\sqrt3$ is irrational. Hint: To treat $\sqrt3$, for example, use the fact that every integer is of the form $3n$, $3n+1$ or $3n+2$.

Solution from the "Answers" Chapter:

Since $$ \\ (3n+1)^2 = 9n^2 +6n + 1 = 3(3n^2+2n)+1 \\ (3n+2)^2 = 9n^2 +12n + 4 = 3(3n^2+4n +1)+1 $$ He proceeds then to state that if $k^2$ is divisible by 3, then so is $k$. I have a hard time understanding what $k$ he is talking about in these equations, it hasn't been defined earlier. He then proceeds:

Supose $\sqrt3$ were rational, and let $\sqrt3 = p/q$, where $p$ and $q$ have no common factor. Then $p^2=3q^2$, so $p^2$ is divisible by 3, so $p$ must be. Thus, $p=3p'$ for some natural number $p'$ and consequently $(3p')^2=3q^2$, or $3(p')^2=q^2$. Thus, $q$ is also divisible by 3, a contradiction.

I have a hard time linking the first part where he states that "$k^2$ being divisible by 3 leads to $k$ being divisible by 3" and the second part. Where does the first conclusion come from?

NOTE: Problem statement in the book in the same manner requests the proof of $\sqrt5$ and $\sqrt6$ being irrational. However, in this question I'm interested in the proof for the case of $\sqrt3$.

  • $\begingroup$ To those who don't have a copy of Spivak open at all times (and managed to guess correctly the edition you're using), how about giving the title the content of the actual problem? I mean, is "Understanding the proof that $\sqrt3$ is irrational in Spivak" is a far superior title. Don't you agree? $\endgroup$ – Asaf Karagila May 20 '18 at 23:22
  • $\begingroup$ I've edited the title and the description. $\endgroup$ – Eval May 21 '18 at 11:47
  • 1
    $\begingroup$ That's much better. Thanks. $\endgroup$ – Asaf Karagila May 21 '18 at 11:48

Let $k$ be an integer. It can take one of the following forms: $3n$ or $3n+1$ or $3n+2$ where $n$ is simply the quotient in the euclidean division of $k$ by $3$.


  • if $k=3n$ then $k^2 = 9n^2$
  • if $k=3n+1$, then $k^2 = 3\times ... +1$ (we don't really care of the exact value of $...$), thus $k^2$ isn't multiple of $3$
  • if $k=3n+2$, then $k^2 = 3\times ... +1$, thus $k^2$ isn't multiple of $3$.

Thus, if $k^2$ is multiple of 3, then $k$ can't be of the form $3n+1$ or $3n+2$. Hence, it must be of the form $3n$, that is, $k$ must be multiple of 3.

Hence, if $k^2$ is multiple of $3$, then $k$ is also multiple of $3$.

  • $\begingroup$ Got it. Nice explanation. $\endgroup$ – Eval May 20 '18 at 16:15
  • 1
    $\begingroup$ Yeah. A slightly more complicated way of saying that $k^2 \equiv 0\;(\operatorname{mod} 3)$ if $k \equiv 0\;(\operatorname{mod} 3)$, $1$ if $k \equiv 1\;(\operatorname{mod} 3)$, or 1 if $k \equiv 2\;(\operatorname{mod} 3)$, Good answer. $\endgroup$ – Davislor May 20 '18 at 20:56

The state ment

If $k^2$ is divisible by $3$ then so is $k$.

holds for all integers $k$. He uses this fact below for $k=p$ and for $k=q$.


In the first part, Spivak proves that of a number is not a multiple of $3$ (that is, if it is of the type $3n+1$ or $3n+2$), then its square is not a multiple of $3$. Therefore, if the square is a multiple of $3$, then the number itself must be a multiple of $3$ too.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.