1

Let $U_1, U_2$ be subspaces of the real vector space $V$. Prove that $U_1 \cup U_2$ is a subspace of $V$ iff $U_1 \subset U_2$ or $U_2 \subset U_1$

Is my proof valid?

  1. Firstly proving $U_1 \cup U_2 \space \text{is a subspace} \implies U_1 \subset U_2$ or $U_2 \subset U_1$

$\begin{align*}U_1 \cup U_2 \space \text{is a subspace} & \implies \alpha u_1+\beta u_2 \in U_1 \cup U_2 \space\space\forall u_1 \in U_1, \forall u_2 \in U_2, \forall\alpha,\beta \in \mathbb{R} \\ & \implies \alpha_0 u_1+\beta_0 u_2 \in U_1 \space\text{or} \space \alpha_0 u_1+\beta_0 u_2 \in U_2 \space \text{for some $\alpha_0, \beta_0 \neq 0$} \end{align*}$

$\text{WLOG assume}\space \alpha_0 u_1+\beta_0 u_2 \in U_1$

$\begin{align*} & \alpha_0 u_1+\beta_0 u_2 &&\in U_1 \\ \implies & \alpha_0 u_1+\beta_0 u_2 + (-\alpha_0)u_1 &&\in U_1 \\ \implies & \beta_0 u_2 &&\in U_1 \\ \implies & \frac{1}{\beta_0}\cdot\beta_0 u_2 &&\in U_1 \\ \implies & u_2 &&\in U_1 \text{(All by the properties defined on a vector subspace.)} \end{align*}$

Hence $U_2 \subset U_1 \space \text{and similarly} \space U_1 \subset U_2.$ $\text{Thus} \space U_1 \cup U_2 \space \text{is a subspace} \implies U_1 \subset U_2$ or $U_2 \subset U_1$

  1. Secondly proving $ U_1 \subset U_2$ or $U_2 \subset U_1 \implies U_1 \cup U_2 \space \text{is a subspace}$

$\text{WLOG assume $U_1 \subset U_2$}$

$\begin{align*} \text{Hence} \\ &u_1 &&\in U_2 \space \forall u_1 \in U_1 \\ &\alpha u_1+\beta u_2 &&\in U_2 \space \forall u_2 \in U_2, \space \forall \alpha, \beta \in \mathbb{R} \\ &\alpha u_1+\beta u_2 &&\in U_1 \cup U_2 \end{align*}$

Hence $U_1 \subset U_2$ or $U_2 \subset U_1 \implies U_1 \cup U_2 \space \text{is a subspace}$

Please let me know if this is correct.

  • It would be clearer if you write your proofs in English. In your proof that $U_1\subset U_2$ or $U_2\subset U_1$ it is unclear what the $u_1$ and $u_2$ are on the second line. In the first line they are arbitrary elements of $U_1$ and $U_2$, now on the second line are they still arbitrary elements? Do the coefficients $\alpha_0$ and $\beta_0$ depend on the $u_i$? – Joshua Tilley Nov 07 '23 at 22:02
  • I see my mistake now. My assumption they're still arbitrary elements in the second line is wrong. Thanks, and I'd appreciate if you could clarify what exactly you found unclear about the way I wrote my proof. Whether it was the formatting that made it difficult, or whether I didn't define what I was doing properly, so I can work on these things in the future. – thecountofmontecristo Nov 07 '23 at 22:36
  • 1
    Symbols like $\implies$ and $\forall$ can be good short-hands when you are sketching out possible proofs on paper (because you know what you mea, but not good for presentation of a proof (unless you really are working in a formal logic, where the meaning is well-defined). It is clear from the algebra in the first part that your are trying to show that if a linear combination of $u_1 \in U_1$ and $u_2 \in U_2$ belongs to $U_1$ then $u_2 \in U_2$, but this isn't quite strong enough to give what you are looking for: I think you need to use contradiction as I did in my answer. Hope that helps. – Rob Arthan Nov 07 '23 at 22:52
  • 1
  • This is most certainly a duplicate. 2. This is not a "please check my proof" site. If you have specific doubts about your argument, then you should point out exactly which steps you are unsure about and why you are unsure about them. Otherwise, the question is off-topic as being "too broad". You are essentially asking one question for each step in your argument, "Is this step right?" "Is this step right?" "Is this step right?" That's multiple questions in a single post. Voting to close.
  • – Arturo Magidin Nov 07 '23 at 23:42
  • I'd appreciate if you could post a link to the duplicate you're most certain exists if you found one it so I could learn from the discussions there. I couldn't find one myself that had anything similar to my proof attempt and could help me understand if I was going wrong with it. I've noted your other point and I'll be careful to be specific about my doubts in the future. Cheers. – thecountofmontecristo Nov 08 '23 at 12:27
  • Union of two vector subspaces not a subspace?; 99 votes; 7 answers with positive score, top answer has 124 votes, second answer, with 87 votes has a proof, as does the third top answer, with 30 votes, and the fourth, with 14 votes. – Arturo Magidin Nov 08 '23 at 17:57