Hacker News Reader

Comment by sdwr | Hacker News Reader

Easier to spot the trick when it's all written in code. "There-is" is a huge overreach from a computational perspective, opens the door for "this sentence is false" type logical knots and unbounded complexity.

I know its the first lesson in discrete math, but "there exists a _" makes no sense from an algorithmic perspective. How do you find it?

View on HN

I think I was confused about the when function... I was at first thinking of when as when in clojure which is essentially just a macro for

(if .. do ..)

(macroexpand '(when 1 2 3 4)) (if 1 (do 2 3 4))

but that is not what it seems to do in PM-Lisp take the example of `=` definition:

(def = (and (when A B) (when B A)))

If this were the same functionality clojure it would be testing the truthyness of the evaluation of each when expression

so If I had (and (when 3 5) (when 5 3)) it would evaluate to (and 5 3) which would evaluate to 3 which is truthy.

So that got me stuck in the article. But it seems that when is more like a test/assertion... (when x z) saying for all values of x ... z is implied (i.e. the same way one mathematical rule implies another).

I'm going to reread this article and try to understand better.

Don't know about clojure (is that a lisp?), here when is therefore in formal logic - "A implies B". Ex, "If it's raining, the streets are wet".

It doesn't "do" anything, just sets a constraint which can be matched against later. I took a class on Prolog in school, which works by back-matching against conditionals like that. I remember just sorting a list being mind-bendingly recursive.

I still think there's some funny business going on in the intersection of computable functions, infinite sets and "there exists a _" style conditionals. Can't put my finger on why "there exists a number of the form 2n + 1" feels fine, but "there exists a function in the set of all possible functions" feels wrong.

The imperative if in programming is different from the logical if. But it isn't disconnected from logic.

In fact, it is closely related to the logical and: it is and with short-circuiting, enabling the left expression to control effects in the right.

In Lisp, the two-form if, namely (if x y), can be replaced by (and x y), because and has the right logic and result value, and also the short-circuiting semantics that y is not evaluated if x is false.

If we want the logical if, we can use the equivalent expression (or (not x) y), which has the same truth table:

     x     y  | (if x y)  |  (or (not x) y)
  ------------+-----------+----------------
   nil   nil  |    nil    |     t
   nil     t  |    nil    |     t
     t   nil  |    nil    |   nil
     t     t  |      t    |     t

The if/and relationship of course shows up in other languages:

   if (pointer != NULL) {
      if (pointer->foo == 42)
        do();
   }

-->

  if (pointer != NULL && pointer->foo)
    do();

There's no overreach whatsoever and you can eliminate "there exists" altogether without any loss of expressivity.

"There exists a number X such that P(X) is true." is nothing more than a short hand for "It is false that for all numbers X, P(X) is false."

In other words, if you accept that it's possible to reason about "for all" then you must accept that it's possible to reason about "there exists" since all "there exists" does is assert that a property is not false for every number.

I don't think OP was saying that "for all" is any more meaningful from a computational perspective than "there exits". The point is that both are higher order functions that return an answer for any function on all natural numbers (or inputs in a more general setting).

In that case I'd argue the objection has nothing to do with "for all" or "there exists" but rather reasoning about infinite sets.

A set doesn't need to be infinite for us to not be able to say something about all it's members. Each busy beaver number is very much finite but you'd run out of universe if you tried calculating BB value for a number as small as 12.

Whether something is practical to compute is a very different matter from what OP stated was their concern, namely logical knots, potential inconsistencies/contradictions of the "this sentence is false" type, and unbounded complexity.

It really isn't. Math proofs need to be not only finite but small enough to fit in the universe to be discovered. This strictly finite nonsense is a left over from a time when we could only do symbol manipulation with a human brain.

That could be true, but it has nothing to do with what was being discussed.

Not an expert but the proof is in fact entirely constructive and witnesses to the existential statement are constructible.

https://math.stackexchange.com/questions/448527/are-the-g%C3...

Reply on news.ycombinator.com