Teach Yourself Logic: A Study Guide [pdf] • Hacker News Board

Re: Teach Yourself Logic: A Study Guide [pdf]

#41

Post by tzjmetron » Wed, Dec 26, 2018, 12:13 PM UTC

nurettin wrote: post #14

edoo wrote: post #13

Studying logic taught me one of the more important distinctions in life is identifying when things are always true vs sometimes true. We are taught to often group things that are mostly true into the always true category. This is effectively lossy compression that can introduce serious error into your line of reasoning.

That doesn't make sense to me as in real life most of our reasoning comes from syllogisms for practical reasons. And if something is consistently demonstrated to be true, chances are it will continue to be so, so we take our chances with that. It isn't lossy anything, it is life.

Agreed.

Re: Teach Yourself Logic: A Study Guide [pdf]

#42

Post by BucketSort » Wed, Dec 26, 2018, 1:01 PM UTC

BucketSort wrote: post #31

_emacsomancer_ wrote: post #28

Earlier quoted context omitted.

> many ideas are more accessible to me in type theory as opposed to logic But how do you understand type theory without some basic understanding of logic?

It's possible, but I guess I should have specified "logic outside the standard curriculum of computer science studies." For example, I think understanding bits of modal logic through how monads behave is much more fruitful for a CS person. Also, I want to further add here that the correspondence between proofs and programs is known as the Curry–Howard correspondence. This is the more general term under which these re…

Also, "Curry–Howard–Lambek correspondence" is the real name for "computational trinitarianism." I like the latter better because I don't like naming mathematical things after people, which have no bearing on the actual meaning.

Re: Teach Yourself Logic: A Study Guide [pdf]

#43

Post by chestervonwinch » Wed, Dec 26, 2018, 2:44 PM UTC

XuMiao wrote: post #19

nurettin wrote: post #14

Earlier quoted context omitted.

That doesn't make sense to me as in real life most of our reasoning comes from syllogisms for practical reasons. And if something is consistently demonstrated to be true, chances are it will continue to be so, so we take our chances with that. It isn't lossy anything, it is life.

There is the mathematical logic that is deduced from axioms which is considered universally true unless the axioms are false. Our everyday "folk" logic is always contextual and ambiguous. The deductions are probabilistic instead of being deterministic. They are sometimes true.

There’s a framework for that: https://en.m.wikipedia.org/wiki/Fuzzy_logic

Re: Teach Yourself Logic: A Study Guide [pdf]

#44

Post by _emacsomancer_ » Wed, Dec 26, 2018, 8:28 PM UTC

BucketSort wrote: post #31

_emacsomancer_ wrote: post #28

Earlier quoted context omitted.

> many ideas are more accessible to me in type theory as opposed to logic But how do you understand type theory without some basic understanding of logic?

It's possible, but I guess I should have specified "logic outside the standard curriculum of computer science studies." For example, I think understanding bits of modal logic through how monads behave is much more fruitful for a CS person. Also, I want to further add here that the correspondence between proofs and programs is known as the Curry–Howard correspondence. This is the more general term under which these re…

Ok, but students need to understand at least ¬, ∧, ∨, →, and there are other useful equivalencies as between sets and (characteristic) functions etc.

Re: Teach Yourself Logic: A Study Guide [pdf]

#45

Post by bostonvaulter2 » Thu, Dec 27, 2018, 6:36 AM UTC

Is there an ePub version of the guide available anywhere?

Re: Teach Yourself Logic: A Study Guide [pdf]

#46

Post by 323454 » Sat, Dec 29, 2018, 4:23 PM UTC

claudiawerner wrote: post #25

elefanten wrote: post #20

Earlier quoted context omitted.

Is this phenomenon uniquely pronounced among hackers/computer types in your experience? In a sense, it kind of sounds like your comment could be summarized as "people often try to apply already-available concepts and reasoning to new problems areas". But it's uniquely funny when hackers do it. On the flip side, how many philosophy, sociology and critical theory types reach for concepts from biology and physics? And h…

I think that philosophy offers much more general mental models and abstractions than computer science, which isn't to say they're better or worse, but applicable differently. To that end, I also feel as though many of the concepts hackers use to describe the world are actually more concrete instances of the ones philosophers might use. There's also a difference between analogy (what I have seen philosophers use more…

The phenomenon of people pulling in and trying to explain problems from outside their specialty occurs in every discipline and is a part of normal scientific work. Usually the overwhelming majority of these attempts fail, but they are useful nonetheless because in failing they help to define the limits of the specialty. Ironically, you're complaint that they do this witout considering "if the concept is valid for the domain" is unanswerable until an attempt is made and its claims critiqued.

That aside, you raised good questions. I think would be useful if you turned your list of critical questions into a template for criticising instances of analogous reasoning anywhere. It would help greatly in evaluating these attempts.

Re: Teach Yourself Logic: A Study Guide [pdf]

#47

Post by hnuser1234 » Sun, Dec 30, 2018, 5:01 AM UTC

claudiawerner wrote: post #15

nurettin wrote: post #14

Earlier quoted context omitted.

That doesn't make sense to me as in real life most of our reasoning comes from syllogisms for practical reasons. And if something is consistently demonstrated to be true, chances are it will continue to be so, so we take our chances with that. It isn't lossy anything, it is life.

As a side note, it's always kind of funny to see hackers/computer types try and frame discussions in terms of things which are familiar to them, and these comments are often highly upvoted because of the explanation. In this instance, "lossy compression". But few people ever seem to question whether the domain of such categorisations is at all applicable to real human situations, and even if it's possible to make suc…

Why is it funny that people understand things in terms of their experience? The original comment about "lossy compression" essentially stated that humans don't have perfect thinking/memory/processing of probabilities. Do you disagree? Does it matter whether or not the wording of this idea uses computer science terms?

Sure, technical people might try to apply methods from technical disciplines to social ones. Sometimes it works, sometimes it doesn't. As a rule, however, I don't think anyone assumes mathematics will completely replace social studies. We might use statistics to better understand social studies, but the social context is still distinctive from pure math, and always will be.

Some of the best hackers are the ones who DO borrow ideas from other disciplines. And all other disciplines can often benefit from applying increasingly complex statistical methods, aided by computer and data scientists.