You are viewing the historical archive of The Philosophy Forum.
For current discussions, visit the live forum.
Go to live forum

Karl Popper's Black Ravens

TheMadFool January 21, 2020 at 10:40 7025 views 19 comments
Raven Paradox or Hempel's Paradox:

Claim A = All ravens are black = If it's a raven then it's black

Contrapositive of A = B = If it's not black then it's not a raven = All non-black things are non-ravens

The paradox occurs because anything that proves B e.g. a green apple (non-black but also not a raven) also serves as evidence for A since A and B are logically equivalent. Basically, one can gain information on ravens by looking at non-ravens and this is the paradox.

Karl Popper famously stated that science is about falsifiability i.e. for a hypothesis to be scientific then it must be possible to look for and find counter-evidence.

Consider now the statement A and its equivalent, the contrapositive B.

B = If it's not black then it's not a raven. For B to be false the antecedent has to be true and the consequent needs to be false i.e. B is false if and only if we find a non-black object and it's a raven.

What would amount to as falsifying A = all ravens are black? Of course a raven that is non-black which is exactly what science is supposed to do if A is to be considered scientific according to Karl Popper.

So, [s]falsifying[/s] confirming the claim A can be done by only considering non-black things that need not necessarily be ravens. That is we may look at a green apple or white clouds and be secure that the claim A hasn't been falsified. So long as we don't encounter a non-black raven, we can, according to Karl Popper, invest our trust in the, as yet unfalsified, claim A = all raven are black.

In other words, the Raven paradox is not a paradox in a scientific sense for an inability to falsify a claim counts as support for whatever the claim is and when we see non-raven objects like green apples, yellow bananas, red flags, etc., it implies that the claim A = all ravens are black hasn't been falsified and so we may believe it given that there's also positive evidence (black ravens) to back the claim.


Comments (19)

bongo fury January 21, 2020 at 13:39 #374026
Quoting TheMadFool
So, [s]falsifying[/s] confirming the claim A can be done by only considering non-black things that need not necessarily be ravens. That is we may look at a green apple or white clouds and be secure that the claim A [not only] hasn't been falsified [but has also been positively confirmed].


Which is a potential embarrassment for confirmation theory (induction), but not for falsification theory (hypothetico-deduction), which doesn't pretend to compare and rate equally unfalsified hypotheses according to their confirming evidence.

Quoting TheMadFool
So long as we don't encounter a non-black raven, we can, according to Karl Popper, invest our trust in the, as yet unfalsified, claim A = all ravens are black.


But no more so (according to Popper, as far as I know) than for the equally unfalsified claim that not all ravens are black.

Quoting TheMadFool
In other words, the Raven paradox is not a paradox in a scientific sense


Sure it is. But it's a problem for induction: for deciding between equally as-yet-unfalsified hypotheses.

Quoting TheMadFool
an inability to falsify a claim counts as support for whatever the claim is


You mean confirming evidence counts as support? But how to measure confirmation?

Quoting TheMadFool
we may believe it, given that there's also positive evidence (black ravens) to back the claim.


But what about the equally positive (but intuitively less compelling) evidence of non-black non-ravens? There's the puzzle. (For induction.)
Pantagruel January 21, 2020 at 16:44 #374056
Quoting bongo fury
an inability to falsify a claim counts as support for whatever the claim is
— TheMadFool

You mean confirming evidence counts as support? But how to measure confirmation?


This would be letting confirmation back in through the same door that Popper just tossed it out.
bongo fury January 21, 2020 at 17:05 #374059
Quoting Pantagruel
This would be letting confirmation back in through the same door that Popper just tossed it out.


Agreed. Hempel induce. Popper deduce.

Hempel confirm. Popper falsify.
TheMadFool January 21, 2020 at 17:44 #374061
Reply to bongo fury Thanks for noticing the error in my post. I made the necessary corrections.

Quoting bongo fury
Sure it is. But it's a problem for induction: for deciding between equally as-yet-unfalsified hypotheses.


How is it a paradox when you agree that falsifactionism requires those who make hypotheses to look for counter-evidence by searching outside the domain of the subjects of hypotheses? The statement, all cats are animals is falsifiable precisely by looking for and finding a non-animal that's [s]not[/s] a cat.

TheMadFool January 21, 2020 at 17:49 #374062
Quoting Pantagruel
This would be letting confirmation back in through the same door that Popper just tossed it out.


I guess Popper considers the absence of negative evidence i.e. disconfirming observations as better than positive evidence (confirming evidence). As bongo fury remarked, this type of reasoning, giving greater weightage to disconfirmation, is because of the problem of induction.
Pfhorrest January 21, 2020 at 18:01 #374067
Quoting bongo fury
falsification theory (hypothetico-deduction


Hypothetico-deduction is confirmationist, not falsificationist.
Pfhorrest January 21, 2020 at 18:05 #374071
Quoting TheMadFool
The statement, all cats are animals is falsifiable precisely by looking for and finding a non-animal that's not a cat.


Nope, it's falsifiable by finding a non-animal that is a cat.

"All cats are animals" = "if cat then animal" = "not (cat and non-animal)"

To falsify it, you have to find the negation of that, which would then be "cat and non-animal".
Pantagruel January 21, 2020 at 18:10 #374075
Quoting TheMadFool
I guess Popper considers the absence of negative evidence i.e. disconfirming observations as better than positive evidence (confirming evidence).

Well, it's his main thesis, to be scientific, an hypothesis must be falsifiable, so disconfirming evidence must be at least possible. Whereas induction fails to ever rise to the level of certainty, which he establishes in a variety of ways. I find the logical niceties tortuous at times (like this paradox - what could the status of a non-raven entity ever add to the knowledge of ravens?). However the overall thrust of scientific realism, that objectivity is not what we see, but what has been subjected to critical thought, that I very much like.
Pfhorrest January 21, 2020 at 18:34 #374077
Quoting Pantagruel
I find the logical niceties tortuous at times (like this paradox - what could the status of a non-raven entity ever add to the knowledge of ravens?).


It couldn't, which is the point. Confirmationism implies that it could, which is absurd, hence disproving confirmationism.
Pantagruel January 21, 2020 at 18:43 #374079
TheMadFool January 21, 2020 at 19:02 #374081
Quoting Pfhorrest
Nope, it's falsifiable by finding a non-animal that is a cat.

"All cats are animals" = "if cat then animal" = "not (cat and non-animal)"

To falsify it, you have to find the negation of that, which would then be "cat and non-animal".


:up: So many categories and their complements. I corrected my error.
TheMadFool January 21, 2020 at 19:33 #374089
Quoting Pantagruel
Well, it's his main thesis, to be scientific, an hypothesis must be falsifiable, so disconfirming evidence must be at least possible. Whereas induction fails to ever rise to the level of certainty, which he establishes in a variety of ways. I find the logical niceties tortuous at times (like this paradox - what could the status of a non-raven entity ever add to the knowledge of ravens?). However the overall thrust of scientific realism, that objectivity is not what we see, but what has been subjected to critical thought, that I very much like.


Thanks a lot.
bongo fury January 21, 2020 at 23:46 #374171
Quoting TheMadFool
?bongo fury Thanks for noticing the error in my post. I made the necessary corrections.

Sure it is. But it's a problem for induction: for deciding between equally as-yet-unfalsified hypotheses.
— bongo fury

How is it a paradox when you agree that falsificationism requires those who make hypotheses to look for counter-evidence by searching outside the domain of the subjects of hypotheses? The statement, all cats are animals is falsifiable precisely by looking for and finding a non-animal that's not a cat.


Same confusion here. Corrected:

It's a paradox and potential embarrassment for confirmation theory because it appears to entitle those who make hypotheses to look for confirming evidence by searching outside the domain of the subjects of hypotheses. The statement, all cats are animals is apparently confirmable precisely by looking for and finding a non-animal that's not a cat.


Take rumours of the death of induction with a pinch of salt. (A good habit.) I.e. the embarrassment isn't fatal.

Try to stop confusing the two, though.
TheMadFool January 22, 2020 at 03:44 #374238
Quoting bongo fury
I.e. the embarrassment isn't fatal.


:grin: Thanks for the advice and the clarification.

It seems that what I actually meant was that confirming B = all non-black are non-ravens by observing green apples or red tomatos etc. i.e. non-black non-ravens doesn't amount to a falsification of A = all ravens are black. Ergo, by Popper's account of what a scientific claim is, statement A is not disproved and given that there are some ravens that are black, statement A acquires the status of a scientific theory - to be taken as true for all intents & purposes.

bongo fury January 22, 2020 at 16:05 #374378
Quoting TheMadFool
Ergo, by Popper's account of what a scientific claim is, statement A is not disproved and given [[s]that there are some ravens that are black[/s] that statement A is falsifiable], statement A acquires the status of a scientific theory - to be taken as [[s]true for all intents & purposes[/s] a theory as yet unfalsified and worth testing].


If you want more, fine, there's always induction :smile:
Banno January 22, 2020 at 23:14 #374467
There's work being done on Wikipedia's article on Falsifiability at present.
Pfhorrest January 23, 2020 at 00:06 #374483
Reply to Banno I've been noticing that (it's on my watchlist) and wondering if it was just a coincidence that that was happening there right as these conversations were happening here.

It would be kinda neat to see some kind of cross-collaboration between a forum like this and Wikipedia. Use Wiki as a reference to answer questions here (since WP:NOTFORUM), and then use the conversations here to inspire development of articles there.
TheMadFool January 23, 2020 at 07:15 #374604
Quoting bongo fury
If you want more, fine, there's always induction


Thanks.



Banno January 24, 2020 at 01:06 #374829
Reply to Pfhorrest I don't see it working. To much of a difference in temperament.

But it is an interesting, and falsifiable, hypothesis.

Thinking....


Edit: https://thephilosophyforum.com/discussion/7515/an-hypothesis-is-falsifiable-if-some-observation-might-show-it-to-be-false


A trial. See how it goes.