Hacker News Reader

Comment by kerkeslager | Hacker News Reader

> To justify substantial investments, I am told, an experiment needs a clear goal and at least a promise of breakthrough discoveries.

This is antithetical to science. If you're promising a breakthrough discovery, you're approaching the experiment with bias.

The fact that more new particles have not emerged at energy levels the LHC can produce is a discovery--if I'm understanding the blog post correctly[1], it's the beginnings of a disproof of naturalness in supersymmmetry. It's not as exciting as if they had discovered hundreds of new things to study, but it's equally important.

And that's exactly why I agree with the author: science is about finding what's true not about finding what's exciting. As a taxpayer, I think one of the most valuable things particle physics could do here is to educate people on that bias and lead by example. I get that they fear losing their funding to do science, but if you let that fear push you into pursuing exciting results over the truth, then you're not doing science anyway.

[1] I'm not a particle physicist--my post is about the social problem that physicists are facing, not about the physics.

View on HN

"This is antithetical to science. If you're promising a breakthrough discovery, you're approaching the experiment with bias."

I think all experiments are began with some premonition of what to expect. For example Michaelson and Morley very much expected to measure the speed through which earth would pass through aether ... except they couldn't. An the results pushed physics toward theory of relativity.

I disagree that's it's an entirely bad process to bankroll experiments based on unproven promises. This is exactly how the Manhattan Project happened. The physicists promised that it was very likely they could create a very large explosion, but they did not know if it would bang or fizzle. So, the US government began a huge industrial scale operation to enrich uranium and to assemble the bomb. The first atomic bomb explosion was very much empirical science that was bankrolled by "unsound" promises.

In this sense going beyond LHC would be kinda ground breaking - big budget science with absolutely no clue on what to expect. It's how discoveries are made, yes, but I'm not sure if any large scale scientific project has been funded without at least some clue on what to expect.

> I think all experiments are began with some premonition of what to expect. For example Michaelson and Morley very much expected to measure the speed through which earth would pass through aether ... except they couldn't. An the results pushed physics toward theory of relativity.

> I disagree that's it's an entirely bad process to bankroll experiments based on unproven promises. This is exactly how the Manhattan Project happened. The physicists promised that it was very likely they could create a very large explosion, but they did not know if it would bang or fizzle. So, the US government began a huge industrial scale operation to enrich uranium and to assemble the bomb. The first atomic bomb explosion was very much empirical science that was bankrolled by "unsound" promises.

What you're describing is the "hypothesis" step in the scientific process. Properly done, a hypothesis isn't a promise--it's simply a statement of the possibility you're testing, without any commitment to the possibility being the reality or not.

A good hypothesis results in the same experiment as its negative: "There are more supersymmetric particles at higher energies" is the same hypothesis as "There are not more supersymmetric particles at higher energies" because you test both hypotheses in the same way. Contrast this with a promise: you can't promise something and its opposite.

> In this sense going beyond LHC would be kinda ground breaking - big budget science with absolutely no clue on what to expect. It's how discoveries are made, yes, but I'm not sure if any large scale scientific project has been funded without at least some clue on what to expect.

I don't think we have absolutely no clue what to expect--the article goes into some of the possibilities.

> A good hypothesis results in the same experiment as its negative: "There are more supersymmetric particles at higher energies" is the same hypothesis as "There are not more supersymmetric particles at higher energies" because you test both hypotheses in the same way.

Actually, no, they are not the same. You are excluding the middle, as it is necessary for an experiment to disprove the null hypothesis before any conclusion can be made. Just because you don't prove your hypothesis doesn't mean you prove its negation. Typically, an experiment will find no result at all.

The negative of we "will see X via some experiment." Is not "X does not exist," but rather, "we will not see X via some experiment."

Agreed--thanks for the correction.

I think you're splitting hairs now. Once you have a hypothesis, you can come up with a promise. IF there are (or are not) more supersymmetric particles at higher energies, THEN we can leverage X plus this to move towards possibly curing cancer [or whatever, I have no idea].

The word "promise" doesn't just mean that something will definitely happen, it has a secondary meaning of something being promising, having "the quality of potential excellence".

It's about articulating where this experiment slots into the context, articulating why it's interesting to look at this thing, and not the fifteen other things that won't be funded if your thing does.

I think you maybe can make that sort of conditional promise sometimes, but in context, that's not the kind of promise described in the article. It's specifically saying, "a promise of breakthrough discoveries" (this is the quote from the article). A scientist can't promise breakthrough discoveries. A lot of discoveries are just, "this thing we thought might happen didn't happen, I guess that's a dead end".

That's simply not true. Once you have a hypothesis, you do not come up with a promise. You come up with a test.

Promise and Promising do not have the same meaning.

promise = statement that something will happen

promising = showing signs of future success

You do both. You come up with a test, yes, and you come up with an idea of what it will mean if the test either confirms or rejects the hypothesis.

And no, promise can mean exactly what I said it can mean. https://en.oxforddictionaries.com/definition/promise

From the article:

> To justify substantial investments, I am told, an experiment needs a clear goal and at least a promise of breakthrough discoveries

That sentence is meaningless if "promise" means "A declaration or assurance that one will do something or that a particular thing will happen" -- the "at least" is totally redundant in that interpretation. If it means "the quality of potential excellence", then "at least" makes perfect sense.

If it was "promise" in the sense of "showing promise", then it's a mass noun and they wouldn't have said "a promise". That's like saying "a money" or "a knowledge".

Again, use the word promise here with its other meaning and it makes perfect sense. Both framed questions have the promise of revealing something big. They are not a guarantee of a big result but there is the possibility of a big result.

> What you're describing is the "hypothesis" step in the scientific process. Properly done, a hypothesis isn't a promise--it's simply a statement of the possibility you're testing, without any commitment to the possibility being the reality or not.

Once there is money on the line, the concept of "without any commitment" goes out the window. You are committing money to testing that hypothesis and there is an opportunity cost for other more promising hypotheses you could instead test with that same money.

Saying, "I think the particle collider X will demonstrate the existence of the Higgs boson" (or whatever) is a simple hypothesis.

Saying, "I think you should give me $9 billion to build particle collider X that will demonstrate the existence of the Higgs boson" is a much different statement that requires more sophisticated analysis before smart action can be taken.

I think what really should be said is that we need governments to continue funding pure science - research that doesn't necessarily have immediate benefits, but rather expands our understanding and may provide building blocks for those breakthrough discoveries that clearly move us forward.

That's not to say that there isn't also a place for funding specific research that shows promise for solving specific problems or that would provide specific benefits - the Manhattan Project is certainly an example of this.

Yes, and the LHC cost about $13 billion. Was it worth it? Would another--probably more expensive--collider be worth a likely negative result? Is there a cheaper way to achieve most of the same goals? Are there more promising things to do with that research money?

Top comment suggests it's too early to say that the results from the LHC are entirely negative.

> At the end of 2018, the LHC will have recorded a mere 3% of the intended research program. That means that there is 30x more data to come. I think you'd need to see the results of all of the data before you say that the LHC was a bust. It may be. But your claim is hasty.

>"I disagree that's it's an entirely bad process to bankroll experiments based on unproven promises. This is exactly how the Manhattan Project happened. The physicists promised that it was very likely they could create a very large explosion, but they did not know if it would bang or fizzle."

The difference is that they predicted a large explosion before, now they predict a bump on a graph representing an event (actually events) that nobody otherwise notices ever happened...

> The first atomic bomb explosion was very much empirical science that was bankrolled by "unsound" promises.

War often forces you to take risks that would be imprudent under normal circumstances.

> This is antithetical to science. If you're promising a breakthrough discovery, you're approaching the experiment with bias.

You're mixing two uses of the word promise here.

One is like a guarantee. I promise to deliver a breakthrough.

Another is an expectation. There is the promise of a breakthrough.

https://dictionary.cambridge.org/dictionary/english/promise

The statement here means simply that there could be a breakthrough discovery and that the chances are at least relatively high.

I'm not sure that matters: either use of the word is a form of bias.

It is absolutely not a form of bias in the experiments (it is of course technically a bias in which experiments are funded, but this does not affect the integrity of the results).

Promising a breakthrough means I must either be lucky or force my results towards something that sounds good. That is bias.

Having an experiment where there is promise of a breakthrough simply means my experiment could deliver something huge.

I could fling Fabergé eggs at a wall and it'd be expensive but exceptionally unlikely to reveal anything big. Testing the warmth of fires lit with Rembrandts would be similarly unenlightening but expensive. Firing particles at each other at energies we've never tested before with newly designed detectors has a chance of a breakthrough (however you choose to define a breakthrough). Picking the latter over the former because it can give a breakthrough does not mean the experiments done with it are biased.

The point is though, these criticisms (that LHC might find nothing) had been making the rounds since before the LHC was built, while many promoters claimed we would for sure find evidence of supersymmetry. So while LHC may not have been a mistake, the right response now would arguably be to reassess fundamental theories in light of the new evidence accumulated at the cost of billions of dollars - not to go back and tweak the same old theories to suggest that many more billions and years need to be spent to make really really sure we were wrong.

If late 1800s physicists had spent decades building ever more accurate devices for trying to prove the existence of the eether that held together the universe, perhaps some useful engineering or data analysis work would have come out of it, but it could also be a way for the field to go on an extremely expensive wild goose chase and stall out actual theoretical breakthroughs.

Well, alternate uses of the money is the whole problem.

How many mathematicians can you let loose on long-standing physical problems (qualitative dynamics of the large-N body problem, the freaking turbulence motion of fluids, etc.) -- at some level of "big bet" that frees them from staccato publication pressure -- with the money spent trying to find gluinos or some such ill-developed theoretical construct?

Let's say the total collider budget could fund 100K mathematicians for 10 years, working on whatever they thought was most important.

It's hard to imagine the output of that wouldn't be amazing.

It's actually pretty easy to imagine when you consider the problem of selecting which 100K mathematicians will receive this beneficence. The first few thousand would be straightforward, but they're presumably all the ones who have tenure and can already spend the next 10 years working on whatever they want.

After that, how do you separate the promising mathematicians from the lazy and the crackpots?

This is the same problem that a "Manhattan Project" to cure cancer or what have you always runs into: It's easy to see where to get value from the first dollar, but the 30 billionth dollar likely costs more than a dollar just to figure out how to productively spend it!

"Fortunately", experimental particle physics doesn't have this problem, since you can always use that next dollar to build a bigger collider.

With 100k grants to give, I think you'd want quite a few crackpots, and maybe even some lazy mathematicians. Those could be the types who come up with a game changing result.

100K mathematicians is a lot, but Wikipedia says there are 9267 people with Erdös number of 2, which is a huge mark of distinction: the median Erdös number in Fields medal laureates is 3.

I say, start a program with the Erdös-2 people and as it develops let these hire Erdös-3 folks.

this isn’t sustainable though, as people with smaller Erdös numbers will die out and larger ones will be too common.

so we should probably allow the smaller Erdös numbers to be inherited through primogeniture, to make sure we still have an identifiable class of good mathematicians to give money to.

And that's exactly why I agree with the author: science is about finding what's true not about finding what's exciting.

It's not like the choice about what to spend money on is between "true" and "exciting". The choice is between "true and exciting" and "true and not exciting". We have to use something to choose what to invest in, so why not choose based on how exciting the potential discoveries are?

Because there is a third category, 'exciting but not true', which is at high risk of distorting the decision making process and getting funding that is badly needed elsewhere (see, for example, Scott Kelley's DNA, anything to do with homeopathy, etc).

The article is about funding the Large Hadron Collider. If someone was to suggest spending that amount on researching homeopathy I don't think they'd get very far.

Because, historically, many important discoveries (I would guess the majority) came from things people thought weren't terribly important. Often things that most thought had no value at all.

I think of it like central planning vs free markets. It can sound good to plan things out and direct things towards the outcomes you want, but it's less efficient.

> Because, historically, many important discoveries (I would guess the majority) came from things people thought weren't terribly important. Often things that most thought had no value at all.

Or a massive case of "hmm, that's odd". Like Fleming noticing that bacteria were not growing near certain molds.

Well investing in the exciting thing is what the free market does so...?

It's the difference between central planning and a free market -- thinking your can plan out the overall system vs a more decentralised system -- that I'm focusing on with that analogy, and not making any other particular points about the relationship between science funding and business.

You could also choose to do experiments which teach you nothing at all, and I think that's the worry here.

Most people don't go into science in a completely calm, cool search for raw data. They go into it because they find it exciting and want to solve interesting things or discover "cool" things. Scientists aren't robots. The US went to the moon because it was an exciting challenge, it inspired generations of kids to become scientists. We could discover lots of meaningless facts about logic, but most people want to do something exciting, and the people who discover meaningless facts about logic do it because they think it is interesting and exciting to an extent. Spending money on something that could be exciting is better than paying bean counters to discover meaningless logical facts. Especial when the bean counters don't really find anything, either expected or unexpected.

> Most people don't go into science in a completely calm, cool search for raw data. They go into it because they find it exciting and want to solve interesting things or discover "cool" things.

This is true, but as a great philosopher said, you can't always get what you want. If you search for the truth, a lot of it won't be "cool". Some of it will be cool of course: but if you prioritize coolness over truth, it might prevent you from discovering anything at all. A cool lie isn't a discovery.

That reminded me of two quotes:

> If you look for truth, you may find comfort in the end; if you look for comfort you will not get either comfort or truth only soft soap and wishful thinking to begin, and in the end, despair.

-- C. S. Lewis

> A man may imagine things that are false, but he can only understand things that are true, for if the things be false, the apprehension of them is not understanding.

-- Isaac Newton

> This is antithetical to science. If you're promising a breakthrough discovery, you're approaching the experiment with bias.

That's not true. In fact it's probably the opposite of true!

You do experiments specifically because you have some a priori reason to think that this experiment will tell you something interesting. In fact, one of the major ways scientists are trying to deal with the replication crisis is pre-registering their methodologies and expectations of experiments.

Which isn't to say it's not a good idea to do fundamental research, but it's absolutely valid to try and consider where funding should go based on what we expect to get from experiments.

Of course, I think the current process is pretty bad, since it relies so much on theatrics, as the OP mentioned. But I agree with OP here, at least in what should happen - physicists shouldn't hype or over-promise what an experiment can deliver. I just wish we lived in a world which valued these kinds of fundamental results enough to agree to support them financially!

> You do experiments specifically because you have some a priori reason to think that this experiment will tell you something interesting.

There's a big difference between saying, "This experiment will tell me something interesting" and "This is the interesting thing that this experiment will tell me". The former is what you're describing, the latter is what I'm objecting to.

Then you are objecting to a strawman. "This experiment promises interesting results" is different from "I promise that this experiment will yield interesting results". As in, it's a completely different definition.

Hence the the inclusion of "at least" in the source quote, which wouldn't make sense alongside the other definition. See the dictionary links that have been posted multiple times.

Hm, I can see how the wording of the article is ambiguous.

I don't think that naturalness is something that can be disproved, but it is not necessarily the case that finding out the universe does not work in some particular way is just as useful as finding out that it does work some specific way. That is because there are so many more of the former.

The issue of the article is a resource allocation problem, and there is something unethical about bending scientific prognostications (these cannot be distinguished with the label 'hypotheses') to that end.

"Pursuing exciting results over the truth" doesn't come into it - no-one is accused of falsifying anything here (though it has happened elsewhere.) At worst, the truth will be delayed, though that might be the outcome of pouring resources into a bigger machine, rather than of not doing so.

If you look at the history of Bell Labs, the scientists were mostly given free reign of what to work on, with the idea that it would somehow benefit communications.

What did we get out of it? The transistor, the laser, cellular technology, solar cells, and tons of other things that nobody would have bothered to research -- without the simple curiosity of our scientists.

While I do agree that you should be able to do experiments for experiment's sake, and to just "see what happens" so to speak, projects like the LHC and the fusion reactor projects are all multi-billion projects; that's a lot of money to be spending on something where people don't know what to expect.

But yeah, experiments, especially nowadays, are to prove theories, not to push breakthroughs - in fact, given the higgs boson was already theorized, science could already use it. Not sure what the LHC added to that besides proving it exists.

> But yeah, experiments, especially nowadays, are to prove theories, not to push breakthroughs - in fact, given the higgs boson was already theorized, science could already use it. Not sure what the LHC added to that besides proving it exists.

Theory verification is a very important part of physics. For example lots of people who work in string theory write down lots of crazy theories of how physics looks beyond the standard model. Perhaps they are all wrong, but we currenly have nothing better. So we would really love to have any currently experimentally viable way with which we were able to check these theories. Thanks to LHC we could do this at least for the Higgs boson. Before this experiment the Higgs boson was also "just a crazy theory that was able to resolve a hole in the standard model".

Because building future science upon unverified discoveries is dangerous. Better to prove and measure the higs now rather than have some later and even more monumental experiment to discover a super-higs fail because the underlying theory was off by a few percentage points.

That proof of existence alone is what now holds up decades of work that was done by theorists based on the (then) assumption that the Higgs mechanism is real. The observation is a major puzzle piece in that picture. Without the experiment, all that theory would be worthless.

> While I do agree that you should be able to do experiments for experiment's sake, and to just "see what happens" so to speak, projects like the LHC and the fusion reactor projects are all multi-billion projects; that's a lot of money to be spending on something where people don't know what to expect.

On the contrary, not knowing what to expect is exactly why you should spend money on it. If you know what to expect, there's no reason to spend billions of dollars testing what you already know.

I think the author's argument is that we pretty much know what to expect--a negative result--and all the people arguing otherwise can't give good reasons beyond "I want my funding to continue".

It's also antithetical to scientific principle, though not to common practice, to look for your keys under the streetlight. If there's no good reason to believe that a new accelerator will produce breakthroughs, then there's no reason to fund it. There are plenty of areas of research where large investments of money would have a great chance of advancing science rather than just enabling a small group of scientists to continue doing the work they're accustomed to.

> The fact that more new particles have not emerged at energy levels the LHC can produce is a discovery

You want to find out the distance to the Moon. You build a 100 meters high tower, but you still cannot reach the Moon. So is building a 200 meters high tower now a good idea? Maybe if you build your tower a little higher, you could finally reach the Moon.

Or maybe you should go back to the drawing board and re-think your whole approach.

> This is antithetical to science. If you're promising a breakthrough discovery, you're approaching the experiment with bias.

No, but you have to have some sort of hypothesis to justify the experiment. You don't throw effort and money at the wall either, you make a guess about what you'll find, and use that to drive the decision as to what to investigate.

And new particles, at this point, don't qualify. The LHC was probably "worth it" for the Higgs result alone, but absent a new target (like the Higgs) that we really think will be there, no one sane would build another bigger collider at these budgets.

> This is antithetical to science. If you're promising a breakthrough discovery, you're approaching the experiment with bias.

Whether to build something expensive isn't a scientific question, it's a political one.

To take an example to the extreme, if it were just about science, science might decide to convert the entire mass of the Earth into a particle collider and kill us all in the process.

> Whether to build something expensive isn't a scientific question, it's a political one.

Well, if you're building something expensive with the intent to perform science, one would hope that the political answer to this would be informed by science.

> To take an example to the extreme, if it were just about science, science might decide to convert the entire mass of the Earth into a particle collider and kill us all in the process.

Relevant: http://cowbirdsinlove.com/46

Reply on news.ycombinator.com