In science, we work with hypotheses and theories and only abandon them when they have been proven wrong – by the analysis of empirical and theoretical data – at a sufficient confidence level. There is nothing in between. In particular, a scientist cannot and mustn't abandon a possible explanation that seems just slightly – e.g. 5 times – less likely than another according to some counting; that would be like "elimination at the 80 percent confidence level" which is just way too weak. The "retired idea" may very well be correct and the correct explanations usually are explanations that could be abandoned by a similar sloppy, prejudiced methodology.
The basic process in science is falsification. Painfully enough, John Brockman's website Edge.org doesn't seem to grasp this fundamental feature of the scientific method. The scientists who are answering don't seem to be any better in that respect because (almost?) no one seems to protest against the hidden misconception hiding in the annual 2014 question:
What scientific idea is ready for retirement? (main hyperlink of this blog post; cache is here)You may easily see that by the "retirement", they mean the very same thing as the "hospital for theories" discussed in late 2012. It's a "sort of falsification" which however doesn't require actual, objectively valid (or valuable) arguments that could imply proper falsification, just vague feelings and personal idiosyncratic philosophical preferences. Lisa Randall didn't disappoint me. She agreed with me in this tweet of hers:
Lisa Randall, @lirarandall: Yes @lumidek "What scientific idea I don't like" shouldn't be equated with "What scientific idea is ready for retirement."This disclaimer seems to challenge the annual Edge.org question but it should be a part of common sense: subjective impressions aren't the same thing as objective knowledge or arguments. Science ultimately moves exclusively according to the latter – even though the former may represent intermediate steps that allow individual scientists to achieve the latter.
And indeed, the scientists are offering their (in many cases wild) streams of emotions – personal feelings that they would like to silently promote to objective science with the help of Edge.org.
They would like to "retire" lots of concepts and hypotheses even though they can't falsify them. They would like others to stop the research in research directions that may very well be right, according to the available evidence. And that's too bad. (Some of the concepts they want to "retire" have actually been established to be true and important, but the hardcore science deniers represent a minority – but not tiny minority – of the people who were answering.)
Andrei Linde wants to "retire" the uniqueness and uniformity of the Universe.
However, there's really no evidence that our Universe is non-unique or non-uniform. Linde repeats the anthropic "arguments" in favor of a multiverse but they're just tendentious slogans, not real scientific arguments. It's perfectly acceptable to argue that the would-be regions before the birth of our visible Universe (and perhaps even the spatial regions behind the cosmic horizon "now") are non-existent or unphysical (and therefore physically inconsequential and unusable in valid explanations of anything). It's perfectly sensible to hypothesize that the anthropic explanations suffer from the same need to "realize processes with a small probability" as the non-anthropic ones (because the prior probability of a theory depending on the anthropic selection should be lowered by the same factor as the factor that this anthropic model "explains"). The fundamental equations of string theory admit many solutions ("landscape") but that doesn't mean that all of them are equally real or realized somewhere in the Cosmos ("the multiverse") as ours. And while Linde's contributions to inflation and some other aspects are great, one must still distinguish things that have been established or semi-established by actual scientific arguments (at least the ability to explain some actual patterns in the empirical data) from those that want to be "established" just by Linde's personal authority. It seems to me that at least in this answer, Linde either can't or doesn't want to distinguish these two groups.
Later in the thread, Amanda Gefter of Nude Socialist offers a similar "recommendation to retire" as Linde.
Nina Jablonski wants to deny the existence of races.
This is one of the cases of the straight denial of important concepts. The association of individual humans to races is something that may be done almost perfectly for a vast majority of the mankind. The genetic, anatomic, physiological, and even behavioral differences between individuals in the same race are significantly smaller than the differences between people of different races which is exactly what makes the concept – and, more generally, any concept – legitimate and usually useful. These differences have evolved and increased due to thousands of years of (nearly) separate evolution history of the groups (races). She clearly wants to "retire" an important scientific concept just because it is inconvenient for her tendentious ideology. Hume and Kant could perhaps be considered white supremacists according to our current standards but that doesn't imply that the races don't exist at all.
Martin Rees thinks that our understanding is inevitably bounded.
It's a myth waiting for the "retirement" that we will never hit the boundaries of our understanding. But he can't really prove that such boundaries do exist. We don't have any "really new" yet solid arguments that our ancestors didn't have and that would imply that the hypothetical "boundaries of our understanding" are more real or more guaranteed to exist than they were in the past. At any moment in the past, it would have been profoundly stupid and counterproductive to believe that we were hitting the boundaries of our understanding. The particular realization of the "idea about the boundary" has been falsified in all cases – the idea has been repeatedly falsified by the scientific progress. The idea that "this time is different" isn't substantiated by anything whatsoever. The progress may always look difficult and sometimes it is difficult but that doesn't mean that it is impossible. There are tons of progress occurring also in theoretical physics at the conceptual level, especially within string theory, and if someone chooses to pretend that those things don't exist or even can't exist, he is really fighting against the scientific progress, the recently achieved scientific results, and against science itself.
Julia Clarke wants to retire the Urvogel.
It's the forefather of birds if you don't speak any German. She is surely right that the question "Who was a bird?" depends on the definition of the word – and Nature can't give us any rock-solid universal definition of words that were created to classify currently living species and not all species. However, some definitions are still better than others. Some pairs of species are closer to each other in the tree than other pairs. Archaeopteryx was a blunder (yes, I actually learned about that "discovery" at school!) but I do think that it will be increasingly established that the old ancestors of birds were very close to the animals we started to call "dinosaurs". Some of them were the first ones who could fly. They may very well be called the ("die") Urvögel.
She credits the late TRF reader Michael Crichton with the widespread belief that the birds' ancestors were dinosaurs. I am not defending the idea she opposes because I've interacted with Crichton but because I am convinced that it's the most economic and likely explanation of the known data. In the Jurassic and adjacent periods, the dinosaurs simply covered pretty much all animals that were sufficiently close, and it seems economic to believe that the then-living ancestors of present birds belonged to this class of dinosaurs (I really mean the clade Dinosauria), too.
Fiery Cushman, a psychologist, opposes the idea that big effects have big explanations.
It's a wonderful and correct answer. You must remember that I agree with him as I argued in the context of JFK and other conspiracy theories. The idea that big effects/events must actually have big explanations/theories has been genuinely falsified. It's an idea that can't be proved by correct logic, either. So a psychologist is actually the first pundit who wants to "retire" something that is actually demonstrably flawed!
His arguments and examples are actually rather similar to my arguments against the "two sharply separated roads" in the physicists' debates on naturalness.
There are about 173 contributions like that in total. It's a very mixed bag. I would say that most of the ideas "demanded" to be "retired" are actually either demonstrably correct or "totally plausible". Edge.org is an interesting venue clumping intellectuals and scientists but I do believe that campaigns trying to "retire" ideas without actual robust enough arguments are making a disservice to science because they are designed to reduce the diversity of research, eliminate ideas that may very well be correct, and impose group think rooted in random personal preferences that are only defended by vague and/or emotional words.
Just some briefer comments about some other answers.
I didn't understand the complaints about the "shared genes". Different species surely have different genes, right? A journalism professor says that there's nothing such as "information overload" – only the failure of filters. Well, kind of right. One could always say that there is "information overload" and in all cases, some filter is a solution, so no situation is really different from another. Well, except that some filters are inevitably failing if the incoming data are numerous yet look equally relevant or equally reliable, and that's the case when the term "information overload" looks right and unavoidable to me.
Some cognitive philosophers think that "knowing is a tiny portion of decisions". Bizarre. Well, the percentage is hard to define and context-dependent, anyway. Alan Guth claims that the Universe didn't have to have a low entropy to start with. Well, it surely had a lower entropy because of the second law – regardless of the ensemble of degrees of freedom that we call "the Universe" and its geometric arrangement. Because we may always ask "what was before that" and the only effect that prevents us from further extrapolations to the past is when we hit \(S=0\), we can say that \(S=0\) is unavoidably the property of the \(t=0\) moment. (A maritime guy named Bruce Parker wants to "retire" the concept of entropy itself. Embarrassing.)
Someone opposes Moore's law because other quantities than the number of transistors – namely the number of computers connected to yours – will become more important. It doesn't seem likely to me. There are billions of computers connected to yours through the Internet. Why should it matter? What should it matter for? And even if a new quantity became more important at some moment, one may gradually switch Moore's law from one quantity to another (like when a stock index is readjusted when a new large company is added).
A medical research official thinks that a "larger trial" doesn't necessarily bring more accurate results. Well, the greater size doesn't reduce the systematic errors but it surely reduces the statistical ones! And the statistical errors are often important – and they are often the "excuse" why people are satisfied with weak, 2-sigma-like results (that often turn out to be wrong). Note that the previous sentence is more accurate, more correct, and more informative than his longish diatribe.
Universal grammar is attacked because languages are more diverse than previously thought. Well, great, but the term "universal grammar" is still vague and may be adjusted to agree with the enhanced diversity. Yes, I do realize that I have just been defending Noam Chomsky, kind of. ;-) Some grammar-like abilities are surely hardwired to the brain although most of the detailed laws that Chomsky and others later added may be found incorrect in general.
Not shockingly for this world flooded with anti-quantum zealots, a software engineer and armchair physicist wants to "retire" the uncertainty principle, too. The arguments are complete gibberish – partly based on authorities, partly based on discussions of the right translation of Heisenberg's "Unschärfe Relation". You can't get to any real physics in this way, Mr Krause. At most, you will create feces. The word may sound similar to "physics" but it is not the same thing. Only crackpots doubt the uncertainty principle.
Gordon Kane wants to retire the idea that the world only has 3+1 dimensions. I sympathetic to the extent that many people should get used to the (likely and contradicting no established insights) possibility of extra dimensions. On the other hand, it still seems remotely plausible that the right future (marginally improved or even complete) explanations will only deal with \(d=4\) at some pragmatic level (even if they're rooted in string theory) so I wouldn't "retire" all researchers that assume \(d=4\) in their research even though I think it is very likely that their starting point misses some important points.
I had the temptation to react to every single answer but 173 is too many and many of the answers are too silly.