1574: "Trouble for Science"

This forum is for the individual discussion thread that goes with each new comic.

Moderators: Moderators General, Prelates, Magistrates

User avatar
Kozmo
Posts: 23
Joined: Mon Mar 09, 2015 8:07 pm UTC
Contact:

1574: "Trouble for Science"

Postby Kozmo » Mon Sep 07, 2015 12:09 pm UTC

Image

title="Careful mathematical analysis demonstrates small-scale irregularities in Gaussian distribution"

It's true, when making a Gaussian distribution histogram I never get it perfectly smooth, although decreasing the bin size does appear to help.

User avatar
orthogon
Posts: 2962
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1574: "Trouble for Science"

Postby orthogon » Mon Sep 07, 2015 12:16 pm UTC

Fortunately the irregularities in the Gaussian Distribution are roughly normally distributed.
xtifr wrote:... and orthogon merely sounds undecided.

User avatar
DavesNotHere
Posts: 21
Joined: Mon Jan 23, 2012 8:17 am UTC

Re: 1574: "Trouble for Science"

Postby DavesNotHere » Mon Sep 07, 2015 12:28 pm UTC

orthogon wrote:Fortunately the irregularities in the Gaussian Distribution are roughly normally distributed.


Obviously the irregularities in the Gaussian Distribution have a Gaussian distribution.

User avatar
Locoluis
Posts: 99
Joined: Mon Dec 11, 2006 7:30 pm UTC
Location: Santiago, Chile
Contact:

Re: 1574: "Trouble for Science"

Postby Locoluis » Mon Sep 07, 2015 12:44 pm UTC

Sueños del Sur - A webcomic about four siblings, their family, friends, adventures and dreams.
http://sds.lgm.cl/

sotanaht
Posts: 212
Joined: Sat Nov 27, 2010 2:14 am UTC

Re: 1574: "Trouble for Science"

Postby sotanaht » Mon Sep 07, 2015 12:44 pm UTC

DavesNotHere wrote:
orthogon wrote:Fortunately the irregularities in the Gaussian Distribution are roughly normally distributed.


Obviously the irregularities in the Gaussian Distribution have a Gaussian distribution.


Which itself has irregularities.

These headlines may either hint at a breakdown of reality itself, or merely at societies rejection of objective reality.

User avatar
Echo244
Posts: 511
Joined: Wed May 20, 2015 9:49 am UTC
Location: Ping! Ping! Ping! Ping!

Re: 1574: "Trouble for Science"

Postby Echo244 » Mon Sep 07, 2015 12:56 pm UTC

I'm wondering if there's any connection to the feedback from the xkcd survey...
Unstoppable force of nature. That means she/her/hers.
Has committed an act of treason.

User avatar
ucim
Posts: 6488
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: 1574: "Trouble for Science"

Postby ucim » Mon Sep 07, 2015 1:30 pm UTC

At first I read the fourth panel as "Republican study fails to reproduce many published results".

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Please help addams if you can. She needs all of us.

belliott4488
Posts: 79
Joined: Tue Mar 25, 2008 8:53 pm UTC

Re: 1574: "Trouble for Science"

Postby belliott4488 » Mon Sep 07, 2015 1:32 pm UTC

Oh, that's just great. The deniers are gonna take this and run with it.

rmsgrey
Posts: 3433
Joined: Wed Nov 16, 2011 6:35 pm UTC

Re: 1574: "Trouble for Science"

Postby rmsgrey » Mon Sep 07, 2015 2:06 pm UTC

What I want to know is where the joke is...

hamjudo
Posts: 99
Joined: Wed Feb 16, 2011 6:56 pm UTC

Re: 1574: "Trouble for Science"

Postby hamjudo » Mon Sep 07, 2015 2:17 pm UTC

Of course Bunsen burners make things colder, it's called the nozzle effect. If you want a Bunsen burner to heat something, you have to take the additional step of lighting it.

I assume the controlled study did not include that step. They must have had good ventilation. It is hard to write up a study after you and your lab are engulfed in a fireball.

This might be like Franklin's lightning and kite experiment. The first experimenter got lucky, but some of those that try to repeat it die.

wayne
Posts: 31
Joined: Fri Apr 19, 2013 3:41 pm UTC

Re: 1574: "Trouble for Science"

Postby wayne » Mon Sep 07, 2015 2:41 pm UTC

Maybe these are all mnemonics?
But for what....?

mschmitt
Posts: 3
Joined: Fri Feb 27, 2015 2:08 pm UTC

Re: 1574: "Trouble for Science"

Postby mschmitt » Mon Sep 07, 2015 4:06 pm UTC

To replicate the Bunsen burner trial, use the burner as the heat source in an absorption refrigerator.

User avatar
orthogon
Posts: 2962
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1574: "Trouble for Science"

Postby orthogon » Mon Sep 07, 2015 4:40 pm UTC

rmsgrey wrote:What I want to know is where the joke is...

There isn't always a joke, but I agree that the message is a bit ambiguous. Some of those issues are the kind of thing that people like Ben Goldacre have been talking about for years: commercial interests and especially publication bias distorting the literature; the need for study findings to be replicated; the 5% probability of a p<0.05 result arising by chance; the need to scrutinise studies for methodological flaws. But perhaps the point of the Bunsen burner example is to point out that vast areas of science are extraordinarily well understood and backed up by countless replications of the key results. Bunsen burners are hot, and if one study shows that they cool things down, that study is up against an awful lot of evidence the other way, and is almost certainly flawed (I like the suggestions of mschmitt and hamjudo as to where those flaws might lie). Likewise vast swathes of medicine, physics, biology etc. are proven way beyond reasonable doubt. I heard recently that evidence for the Higgs Boson is now at something like sixteen sigmas.
xtifr wrote:... and orthogon merely sounds undecided.

Steko
Posts: 5
Joined: Sat Dec 22, 2012 12:15 am UTC

Re: 1574: "Trouble for Science"

Postby Steko » Mon Sep 07, 2015 6:13 pm UTC

I think the joke is that all of these are reported as "trouble for science" but the first four are examples of science working as intended. Only the last example would really challenge science (and even then ...).

User avatar
keithl
Posts: 638
Joined: Mon Aug 01, 2011 3:46 pm UTC

Re: 1574: "Trouble for Science"

Postby keithl » Mon Sep 07, 2015 6:42 pm UTC

While walking on a park trail a few weekends ago, I encountered a man and a dog. The average member of that ensemble was observed to have three legs, one floppy ear, and half a tail. I did not observe millions of ensembles, so I cannot replicate the results to 5 sigma accuracy.

Nothing is measured to 16 sigma accuracy. Indeed, nothing is Gaussian to such extremes - real measurements have "fat tails", "heavy tails", kurtosis, other ways of saying that "there's more at the extremes than you would guess from measurements of the middle." I encountered this when helping design a system for optical enhancment of transistor repeatability on chips - one slow outlier can limit the speed of a whole chip, it matters less how many are "slower". The optimum was to design the optics to reduce the probability of the "very slows" rather than reduce the statistical spread per se - and that was something that could be achieved with the correct enhancement choices for the photomask, while different (and inferior) choices would merely reduce the standard deviation of a few thousand transistors. Not good enough when silicon wafers contain hundreds of trillions of transistors, and a few hundred slow ones can reduce the value of the wafer by half.

Bart Kosko's book "Noise" explains this well for the layman. So does N. N. Taleb's "Fooled by Randomness". We look at reality through "Gaussian colored glasses" - reality is more complex, and most large samples are not Gaussian. Statistical thinking applied properly is a partial antidote for confirmation bias, but it can only disprove, never prove with high confidence. Multiple independent measurement paths are required, and inclusive models; for example, some supposed Sn1a supernovae may actually be quark star collapses, more common in the early universe, making them dubious standard candles.

Statistics-based experiments measure imagination as well as reality; who could imagine that disconnected cables produce superluminal neutrinos?

User avatar
orthogon
Posts: 2962
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1574: "Trouble for Science"

Postby orthogon » Mon Sep 07, 2015 7:07 pm UTC

Steko wrote:I think the joke is that all of these are reported as "trouble for science" but the first four are examples of science working as intended. Only the last example would really challenge science (and even then ...).

In a way, although they're more meta than that, particularly judging by some of the real papers cited over at explainxkcd.com: rather than just tilting evidence against a particular hypothesis, they're suggesting a more systematic, if not systemic problem. The implication is that whole journals may be full of spurious and never replicated results, whole fields using unreliable methods. Of course, that's part of science, too, that we learn how to do science better.

My source for sixteen sigma was none other than Professor Brian Cox, though he was speaking from memory. If I'm not mistaken, the non Gaussian-ness of the data doesn't affect how many sigmas a result has, though of course it does affect the significance calculation.
xtifr wrote:... and orthogon merely sounds undecided.

User avatar
Kalium_Puceon
Posts: 66
Joined: Mon Aug 31, 2015 12:44 pm UTC
Location: Cape Town, South Africa
Contact:

Re: 1574: "Trouble for Science"

Postby Kalium_Puceon » Mon Sep 07, 2015 7:37 pm UTC

As far as the bunsen burner goes, if you're using high-pressure gas and forgetting to light the thing, you could easily freeze water with it. Which would be a novel use of lab gas. And almost definitely dangerous.
"You never get over the desire to do stupid things. You simply have to overrule your stupid urges with an acquired sense of fear."

-Dr. Richard Weisiger

bazza
Posts: 14
Joined: Wed May 25, 2011 6:54 am UTC

The REAL "Trouble for Science"

Postby bazza » Mon Sep 07, 2015 8:18 pm UTC

The real problem for science is the question of how it is run.

Unhelpful Competition
The peer review system is widely seen as being the best way of ensuring that only correct scientific conclusions are published. The trouble is that is assumes that all publication reviewers are utterly impassive, that all correct results are valid, etc.

However we know that reviewers are not perfect, especially if they're about to publish a paper on the same topic as they're reviewing. We know that journals are mostly interested in exciting and sexy results and will not publish negative results. We know that certain groups / researchers achieve positions of being unchallengable, and they exert disproportionate control on the exposure and publication of other's work (especially if it is contrary to their own), whilst often having their own work published with barely a second glance from publications. We know that whilst the system encourages cooperation amongst small sets of scientists, it practically elminates national or global coordination. There's a load of competition between research groups globally, all figthing to be the first to publish. There's a ton of negative results that never get published and no one knows about them, so work gets needlessly repeated with a great waste of money, time, effort and morale.

Why is this? Every one has to earn a living, and with money generally in short supply everyone plays the system to their best advantage. That's the human thing to do. It's especially true in fields where there is potentially a massive cash reward for solving a huge problem (cure for cancer, etc).

It's Not As Good As It's Cracked Up To Be
Why is that a problem? Well, for one thing it means that the really bright people get out of science as quickly as possible. The science system guarantees that the brightest people are not working in science. They're far too clever to stick around in a system that 1) grossly limits their chances of actually making a difference, 2) bring precious little chance of decent renumeration, 3) runs the risk of sinking years of efffort into unknowingly going down a previously explored yet unpublished dead end, 4) makes their lives hell, and 5) basically robs them of a large number of years of decent pension contributions whilst they're in the system. Don't even ask what rights a pregnant female scientist has.

Above Politics? WTF?
Then there's the politics. This is something that massively affects climate research. If there's one thing science should be doing is giving advice as to what to do, not pontificating as to what's wrong. Does mankind affect the climate? Sure, just like everything else does. Without mankind's inteference, would the climate change anyway? Sure, just like it always has done. There's not really any debate to be had.

What would be more useful is if the scientists stopped bleating on about it and started making useful recommendations. What scientists always seem to forget is that there's not a politician on the planet who actually wants to burn fossil fuels. What there actually is is a bunch of politicians who know that without a continual supply of energy their populations are mere days away from riots, anarchy and strife.

Politicians are sick to the hind teeth of hearing about climate change; what they actually need is a single reasoned message about where to get energy from instead of burning fossil fuels. (The answer is probably nuclear fusion, not that the scientific community is unanimous on that either...).

Reform

The scientific world is badly in need of reform, but there's precious little chance of it happening.

Disclosure
No, I've not had a paper rejected, nor am I involved in scientific research, etc. I'm an engineer.

User avatar
da Doctah
Posts: 904
Joined: Fri Feb 03, 2012 6:27 am UTC

Re: 1574: "Trouble for Science"

Postby da Doctah » Mon Sep 07, 2015 9:04 pm UTC

If you need the meta for this strip, it reads Study Proves That All Studies Are Wrong.

Quirkafleeg
Posts: 1
Joined: Mon Sep 07, 2015 8:36 pm UTC

Re: 1574: "Trouble for Science"

Postby Quirkafleeg » Mon Sep 07, 2015 9:10 pm UTC

What would be more useful is if the scientists stopped bleating on about it and started making useful recommendations.


They have. And have been doing so for decades. In journals, conferences, government consultations, and yes, even the media.

So, for example, the IPCC has three panels - one studying the science involved in climate change, one studying the adaptation to climate change and one studying the mitigation of climate change. Just because you've only listened to news reports on one of them, that doesn't mean the other two haven't been hard at work.

User avatar
Eternal Density
Posts: 5547
Joined: Thu Oct 02, 2008 12:37 am UTC
Contact:

Re: 1574: "Trouble for Science"

Postby Eternal Density » Mon Sep 07, 2015 9:20 pm UTC

keithl wrote:While walking on a park trail a few weekends ago, I encountered a man and a dog. The average member of that ensemble was observed to have three legs, one floppy ear, and half a tail. I did not observe millions of ensembles, so I cannot replicate the results to 5 sigma accuracy.

Nothing is measured to 16 sigma accuracy. Indeed, nothing is Gaussian to such extremes - real measurements have "fat tails", "heavy tails", kurtosis, other ways of saying that "there's more at the extremes than you would guess from measurements of the middle."

Eh, I've seen dogs with fat tails and heavy tails, but I've never seen one suffering kurtosis.
Play the game of Time! castle.chirpingmustard.com Hotdog Vending Supplier But what is this?
In the Marvel vs. DC film-making war, we're all winners.

User avatar
orthogon
Posts: 2962
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1574: "Trouble for Science"

Postby orthogon » Mon Sep 07, 2015 9:27 pm UTC

Eternal Density wrote:
keithl wrote:While walking on a park trail a few weekends ago, I encountered a man and a dog. The average member of that ensemble was observed to have three legs, one floppy ear, and half a tail. I did not observe millions of ensembles, so I cannot replicate the results to 5 sigma accuracy.

Nothing is measured to 16 sigma accuracy. Indeed, nothing is Gaussian to such extremes - real measurements have "fat tails", "heavy tails", kurtosis, other ways of saying that "there's more at the extremes than you would guess from measurements of the middle."

Eh, I've seen dogs with fat tails and heavy tails, but I've never seen one suffering kurtosis.

There's a lot of nonsense spoken. Kurtosis is OK; it's excess kurtosis you need to worry about in a dog.
xtifr wrote:... and orthogon merely sounds undecided.

Story
Posts: 78
Joined: Wed Aug 26, 2009 9:03 pm UTC

Re: The REAL "Trouble for Science"

Postby Story » Mon Sep 07, 2015 10:55 pm UTC

bazza wrote:What would be more useful is if the scientists stopped bleating on about it and started making useful recommendations. What scientists always seem to forget is that there's not a politician on the planet who actually wants to burn fossil fuels. What there actually is is a bunch of politicians who know that without a continual supply of energy their populations are mere days away from riots, anarchy and strife.

Politicians are sick to the hind teeth of hearing about climate change; what they actually need is a single reasoned message about where to get energy from instead of burning fossil fuels. (The answer is probably nuclear fusion, not that the scientific community is unanimous on that either...).
[b]


You're projecting your own views on politicians here. I'm sure quite a few want to burn fossil fuels, if not as an end to itself than as a way to stick it to the man/liberal elites, a way to support the local coal production, demonstrate American spirit, etc.

There is no shortage of useful recommendations about what to do and solutions are well known and available. The real problem is that climate science in the US became entangled with politics to the point where people's opinion on what to do about climate change depends on how they identify rather than the actual science. If the "other team" is advocating that we do something, that must mean that it is bad and we should do the opposite.

User avatar
zjxs
Posts: 62
Joined: Tue Sep 29, 2009 6:04 am UTC
Location: The Cloud

Re: 1574: "Trouble for Science"

Postby zjxs » Tue Sep 08, 2015 1:17 am UTC

This cheese bunsen burner is burning me.

User avatar
jc
Posts: 346
Joined: Fri May 04, 2007 5:48 pm UTC
Location: Waltham, Massachusetts, USA, Earth, Solar System, Milky Way Galaxy
Contact:

Re: 1574: "Trouble for Science"

Postby jc » Tue Sep 08, 2015 1:39 am UTC

hamjudo wrote:Of course Bunsen burners make things colder, it's called the nozzle effect. If you want a Bunsen burner to heat something, you have to take the additional step of lighting it. ... This might be like Franklin's lightning and kite experiment. The first experimenter got lucky, but some of those that try to repeat it die.


Actually, we don't know that. Franklin may not have been the first one to try that experiment; he may have merely been the first who was lucky enough to survive it and write a report.

It may also be an example of the observation that discoveries are almost always named for the last person who made them. After all, once a name is attached to a discovery, nobody else can claim to be the discoverer. We might not know about the others before Franklin simply because they didn't publish their results, for whatever reason, so their name wasn't associated with the phenomenon.

Some historians have extended this to another important idea: The "scientific method" (more properly pluralized) wasn't discovered by European scientists; it has been understood and used by smart people in every society. The problem is that groups of such smart people naturally tend to form guilds that keep their information secret, for their own profit. This means that what they learn tends to disappear with time, since each generation can only "remember" what they hold in the minds of their secret groups. By this theory, the actual advance that led to the development of modern Western science was open publication, so that (as Isaac Newton put it), each generation could stand on the shoulders of the giants that preceded them. By this theory, the ongoing fusses over publication, replication, etc. are just the ongoing extension of this ancient situation. The push for trade secrets, the use of patents to block independent research, and the general disinclination to publish "negative" results are all throwbacks to the ancient regime, since their effect is to discard knowledge and make later researchers rediscover things that were actually known earlier. Franklin may have made a name for himself by going against this, and publishing what he managed to discover about the world around him. He probably also didn't know of any previous discoverers of his results. (Or maybe he did, but didn't tell anyone where he read about it. ;-)

ijuin
Posts: 868
Joined: Fri Jan 09, 2009 6:02 pm UTC

Re: 1574: "Trouble for Science"

Postby ijuin » Tue Sep 08, 2015 3:41 am UTC

What Franklin is credited for in the kite experiment is not the simplified image that kids see in educational cartoons, but rather the discovery that weather systems (thunderclouds/rainclouds) contain electric charges and that lightning, rather than being made of fire or some unknown or even supernatural phenomenon, was simply a gigantic electrical discharge.

As for the scientific method itself, the keystone of the method is not the accumulation and/or organization of research or data. The key is the willingness to admit that your preconceptions were wrong when the data disagrees with prior assumptions. It is the freedom from dogma in and of itself that allows for science as we know it--the willingness to consider that no, the heavens do not orbit the Earth, that lighter-than-air substances are not antigravity, and other notions that contradict "the wisdom handed down by The Gods Themselves".

piratejohn
Posts: 15
Joined: Mon Aug 03, 2015 5:26 am UTC

Re: 1574: "Trouble for Science"

Postby piratejohn » Tue Sep 08, 2015 4:15 am UTC

sotanaht wrote:
DavesNotHere wrote:
orthogon wrote:Fortunately the irregularities in the Gaussian Distribution are roughly normally distributed.


Obviously the irregularities in the Gaussian Distribution have a Gaussian distribution.


Which itself has irregularities.

These headlines may either hint at a breakdown of reality itself, or merely at societies rejection of objective reality.


Gaussian Fractals...

piratejohn
Posts: 15
Joined: Mon Aug 03, 2015 5:26 am UTC

Re: 1574: "Trouble for Science"

Postby piratejohn » Tue Sep 08, 2015 4:17 am UTC

I dunno about any of you, but I have never burned a single Bunsen on one of those burners.

User avatar
1337geek
Posts: 551
Joined: Mon Oct 01, 2007 4:21 am UTC

Re: 1574: "Trouble for Science"

Postby 1337geek » Tue Sep 08, 2015 5:31 am UTC

Over half of psychology studies fail reproducibility test

27 August 2015
Largest replication study to date casts doubt on many published positive results.
"Duct tape is like the Force. It has a light side, a dark side, and it holds the universe together." --Carl Zwanzig

User avatar
Copper Bezel
Posts: 2426
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: 1574: "Trouble for Science"

Postby Copper Bezel » Tue Sep 08, 2015 12:28 pm UTC

I think the joke is that many, many stories in popular news media try to pitch a small irregularity or correction as "everything we think we know about X is wrong!", generally in a way that would not at all impress anyone with a basic understanding of science, and these are stories that actually do imply that, but in a way that the people who are targeted by those stories would not recognize. (Until the bunsen burner one, I mean.) The p-value one throws me a little since misunderstanding it and even abusing it offer one of the strategies used to support said actual stories about not-actually-world-breaking things, but it's also making me think they're the thing being spoofed here.

Edit: Like, "If you want to panic on seeing a science news headline, here are some science news headlines you are allowed to panic about."
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her

rmsgrey
Posts: 3433
Joined: Wed Nov 16, 2011 6:35 pm UTC

Re: 1574: "Trouble for Science"

Postby rmsgrey » Tue Sep 08, 2015 3:39 pm UTC

ijuin wrote:As for the scientific method itself, the keystone of the method is not the accumulation and/or organization of research or data. The key is the willingness to admit that your preconceptions were wrong when the data disagrees with prior assumptions. It is the freedom from dogma in and of itself that allows for science as we know it--the willingness to consider that no, the heavens do not orbit the Earth, that lighter-than-air substances are not antigravity, and other notions that contradict "the wisdom handed down by The Gods Themselves".


I'd have said that the heart of the scientific method is not just a willingness to say that you're wrong, but a constant attempt to prove yourself (and others) wrong. You try your best to prove something wrong, and when you fail, you tentatively accept it as probably close enough to true, at least until someone does manage to find something that proves it wrong after all.

It's complicated a little by the fact that your "facts" are themselves only approximations, so something that doesn't quite match your hypothesis can still be close enough to not contradict it.

operagost
Posts: 102
Joined: Tue Dec 21, 2010 6:23 pm UTC

Re: 1574: "Trouble for Science"

Postby operagost » Tue Sep 08, 2015 6:31 pm UTC

ijuin wrote:What Franklin is credited for in the kite experiment is not the simplified image that kids see in educational cartoons, but rather the discovery that weather systems (thunderclouds/rainclouds) contain electric charges and that lightning, rather than being made of fire or some unknown or even supernatural phenomenon, was simply a gigantic electrical discharge.

That's exactly what I was taught in grade school, actually.

I can't speak for other public schools, as my informal research has told me a lot of them sucked and wasted a lot of time telling kids stuff that was useless or just plain wrong.

lgw
Posts: 436
Joined: Mon Apr 12, 2010 10:52 pm UTC

Re: 1574: "Trouble for Science"

Postby lgw » Tue Sep 08, 2015 8:46 pm UTC

hamjudo wrote:This might be like Franklin's lightning and kite experiment. The first experimenter got lucky, but some of those that try to repeat it die.


Franklin wrote up his experiment for the Royal Society. The write up starts with a list of steps for safety, as other experimenters had died. Franklin's lightning and kite experiment was actually the safe one (I don't think he was the first to do the experiment, he's just the one who understood what he was seeing well enough to publish), instructing the experimenter to stay inside and be flying the kite out the window. Also, his experiment didn't involve lightning directly: simply flying the kite up high enough in the right conditions produced the observed effects, from which he deduced that lightning was an electrical discharge.
"In no set of physics laws do you get two cats." - doogly

User avatar
mathmannix
Posts: 1410
Joined: Fri Jul 06, 2012 2:12 pm UTC
Location: Washington, DC

Re: 1574: "Trouble for Science"

Postby mathmannix » Tue Sep 08, 2015 9:17 pm UTC

Yup, about time those so-called "scientists" admitted they were up to a bunch of mumbo-jumbo. Praise the Lord and pass the ammunition!
I hear velociraptor tastes like chicken.

Azaph
Posts: 8
Joined: Wed Sep 09, 2015 12:25 am UTC

Re: 1574: "Trouble for Science"

Postby Azaph » Wed Sep 09, 2015 12:38 am UTC

Locoluis wrote:The Flaw in Gödel’s proof of his Incompleteness theorem :shock:

I'm not sure even the 'Bunsen burners make things colder' paper would be willing to publish that one. 'The world is a disk carried through space on the back of four elephants' is more plausible! And better argued.
Yeah, I made the mistake of clicking the link. It angered me :P.

ijuin
Posts: 868
Joined: Fri Jan 09, 2009 6:02 pm UTC

Re: 1574: "Trouble for Science"

Postby ijuin » Wed Sep 09, 2015 6:10 am UTC

operagost wrote:
ijuin wrote:What Franklin is credited for in the kite experiment is not the simplified image that kids see in educational cartoons, but rather the discovery that weather systems (thunderclouds/rainclouds) contain electric charges and that lightning, rather than being made of fire or some unknown or even supernatural phenomenon, was simply a gigantic electrical discharge.

That's exactly what I was taught in grade school, actually.

I can't speak for other public schools, as my informal research has told me a lot of them sucked and wasted a lot of time telling kids stuff that was useless or just plain wrong.


Yah the "simplified version" that I was speaking of where they show Franklin's kite as actually getting struck by a lightning bolt that gets conducted down to his apparatus (and sometimes to him as well), leading him to be literally "struck by inspiration". In terms of exaggerated "eureka" moments, it's rather like the portrayal of Newton getting hit on the head by apple--in real life he saw an apple fall and wondered if terrestrial gravity is the same force that maintains planetary orbits, but he was never actually struck by the apple.

rmsgrey
Posts: 3433
Joined: Wed Nov 16, 2011 6:35 pm UTC

Re: 1574: "Trouble for Science"

Postby rmsgrey » Wed Sep 09, 2015 2:18 pm UTC

Azaph wrote:
Locoluis wrote:The Flaw in Gödel’s proof of his Incompleteness theorem :shock:

I'm not sure even the 'Bunsen burners make things colder' paper would be willing to publish that one. 'The world is a disk carried through space on the back of four elephants' is more plausible! And better argued.
Yeah, I made the mistake of clicking the link. It angered me :P.


Actually, if you get past the "I won't summarise my argument because it makes it too easy to counter" and drill down into the technical paper, he may have a point - it's not entirely clear because I didn't feel like spending an hour or so translating his argument into a more transparent notation, but his starting point - that when attempting to introduce self-reference into a system, you need to be clear about which level each of your symbols belongs to - is valid (at least in some formulations).

dg61
Posts: 282
Joined: Mon Jan 18, 2010 8:30 am UTC

Re: 1574: "Trouble for Science"

Postby dg61 » Wed Sep 09, 2015 6:57 pm UTC

jc wrote: Some historians have extended this to another important idea: The "scientific method" (more properly pluralized) wasn't discovered by European scientists; it has been understood and used by smart people in every society. The problem is that groups of such smart people naturally tend to form guilds that keep their information secret, for their own profit. This means that what they learn tends to disappear with time, since each generation can only "remember" what they hold in the minds of their secret groups. By this theory, the actual advance that led to the development of modern Western science was open publication, so that (as Isaac Newton put it), each generation could stand on the shoulders of the giants that preceded them.


Do you have a citation for this? I'd be interested in seeing who's arguing for it.

Azaph
Posts: 8
Joined: Wed Sep 09, 2015 12:25 am UTC

Re: 1574: "Trouble for Science"

Postby Azaph » Thu Sep 10, 2015 7:40 am UTC

rmsgrey wrote:
Azaph wrote:
Locoluis wrote:The Flaw in Gödel’s proof of his Incompleteness theorem :shock:

I'm not sure even the 'Bunsen burners make things colder' paper would be willing to publish that one. 'The world is a disk carried through space on the back of four elephants' is more plausible! And better argued.
Yeah, I made the mistake of clicking the link. It angered me :P.


Actually, if you get past the "I won't summarise my argument because it makes it too easy to counter" and drill down into the technical paper, he may have a point - it's not entirely clear because I didn't feel like spending an hour or so translating his argument into a more transparent notation, but his starting point - that when attempting to introduce self-reference into a system, you need to be clear about which level each of your symbols belongs to - is valid (at least in some formulations).


True.
For reverence, I'm coming here from the perspective of someone who did read the paper, and has a decent level of logical competency. The thing is that given how utterly unclear he is, and how he refuses to give a summation in English, I have difficulty imagining anyone who WOULD be competent enough to actually understand the argument, but not sufficiently well versed in logic to realise that it doesn't actually work once they bothered to translate it (among other things, he seems confused about the difference between 'the godel number of x' and 'the godel number of the godel number of x'). With 9/11 or 'the world is flat', you have some superficially plausible(ish) stuff that anyone can understand. At the very least, a graph. Everyone likes graphs! With a bad scientific paper, people can read bits of it, and kinda think they understand it, without spotting the huge flaws. In this case, anyone who can understand the argument, should be able to understand why it doesn't work.
And also that the claim itself isn't really very plausible. Like, if Bunsen burners really do make stuff colder, there is always the faint possibility of a massive conspiracy.
The other thing is... like, with a paper like 'Bunsen burners make things colder, you can say 'these are our results', and it is conceivably possible that you actually got freak results that actually do bear out your hypothesis, and noone can gainsay you. Not likely, but possible. In this case, you know where it's wrong. You can literally point to where the mistakes are. And people have, in reviews of his paper. Repeatedly (and when we move from symbolic notation to 'actual English', the mistakes start seeming a lot clearer).
And on top of that, he doesn't even have the credentials that things like creationism rely on to make absurd claims seem plausible.
Yeah, I've studied logic, so I might be being optimistic. I just genuinely do have an easier time seeing how someone could be a flat-earth advocate. And like I said, this angers me. Hence, rant ^.

User avatar
Copper Bezel
Posts: 2426
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: 1574: "Trouble for Science"

Postby Copper Bezel » Thu Sep 10, 2015 9:14 am UTC

As someone with zero understanding of formal logic, I can see greater value in publishing a paper describing an apparent flaw in the Incompleteness Theorem than a study that found some erroneous behavior with bunsen burners - so long as it's one of the online journals with the handy links to the refutations of the paper up there in the little box at the upper right. In logic, there's a limited set of common dead end roads that someone might follow, where inept handling of lab equipment is a little harder to categorically prevent and a little easier to write off as a study methodology. Maybe that's just me as an outsider, though. That's leaving aside the nonsense rationalization for why writing abstracts is hard and I don't wanna.
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her


Return to “Individual XKCD Comic Threads”

Who is online

Users browsing this forum: No registered users and 25 guests