Definition by Recursion and a Question about Induction

62 Views Asked by At

I have some questions to ask.

  1. Suppose I want to define some sequence of propositional formulas $\{\varphi_{j}\}_{j\in\mathbb{N}}$. First, I define it this way. Fix an enumeration $p_{1},p_{2},\ldots$ of propositional variables. For any $j\in\mathbb{N}$, define $\varphi_{j}$ recursively as follows. \begin{align*} \varphi_{1}&=p_{1}\\ \varphi_{j+1}&=\varphi_{j}\wedge p_{j+1}. \end{align*} Now, I don't want to specify it recursively. This time I want to say simply this way: $$\mbox{For each}\ j\in\mathbb{N},\ \mbox{define}\ \varphi_{j}=p_{1}\wedge p_{2}\wedge\cdots\wedge p_{j}.$$ The question is: is the latter way of defining $\{\varphi_{j}\}_{j\in\mathbb{N}}$ ambiguous? While one can easily get a sequence $p_{1},p_{1}\wedge p_{2},p_{1}\wedge p_{2}\wedge p_{3},\ldots$ after sticking to the former way of defining, I wonder if it is possible that one may misinterpret the latter way of defining---that it produces something different from what the former does. Which way of defining---recursively, or bluntly---is more acceptable?
  2. From the definition of $\{\varphi_{j}\}_{j\in\mathbb{N}}$ defined recursively above, I want to show that, for each $j\in\mathbb{N}$, $\varphi_{j+1}\models\varphi_{j}$ but $\varphi_{j}\not\models\varphi_{j+1}$. (Here $\models$ means logically implies and $\not\models$ means does not logically imply). By quick inspection, one can see that $\varphi_{j+1}\models\varphi_{j}$ since a conjunction must logically imply one of its constituents. And one can easily arrive at $\varphi_{j}\not\models\varphi_{j+1}$ too since $p_{j+1}$ does not appear in $\varphi_{j}$, and so there is some truth assignment satisfying $\varphi_{j}$ but not $p_{j+1}$. However, the point is: do I have to use proof by induction? I have tried, but the proof is not complicated and it does not even rely on the induction hypothesis. The induction hypothesis is never needed. Does this mean I can prove the claim directly? Or what?)

Thank you in advance.

1

There are 1 best solutions below

1
On BEST ANSWER

1) Both ways are perfectly acceptable. If you don't like to use dots (which could be considered ambiguous but are generally admitted), you can write $$\varphi_j=\bigwedge_{1\leq i\leq j} p_i.$$

2) You don't need the proof by induction, you can just fix any $j\in\mathbb N$, and explain your reasoning at this level. For the first way it is the rule of conjunction elimination, and for the second you describe the truth assignment satisfying $\varphi_{j}$ but not $\varphi_{j+1}$.