Hacker News Reader

Comment by WhompingWindows | Hacker News Reader

Where is the line between objective science and subjective science? Is the work truly science if it is funded and heavily motivated by industry to increase the industry's profits? Or is it advertising and marketing masquerading as science, cloaking itself in claims of objectivity, when only we educated folk know to follow the money and the affiliations?

If science is a set of methods and tools to discover truth, it does make sense that there are self-interested "parasites" which feed off this reputation for their own ends. Sadly, it does seem these parasites are bad for us: smoking, sugar, hydroxychloroquine, ivermectin, low-fat diet lies, not to mention cottage industries of fake internet reviews, bots, trolls, paid shills, who go on forums and social media to sell their snake oil.

View on HN

I sincerely believe the fact that English lacks a distinction between what you mentioned as 'objective' and 'subjective' science is part of the reason why we are in this mess. Science is not a monolith and while we have concepts like validity we lack the distinction between science which is likely to change (diet/nutrition, [xyz] reduces risk of cancer, most psychology research, etc.) and science which is basically law, like kinematics, chemistry, meteorology, etc.

I'm actually struggling to articulate the concept simply because I lack the ability to cleanly categorize these things. To put it another way, a study published in an Evolutionary Psychology Journal is not as 'true' as Evolution itself. Despite this, the study benefits from all the clout established by the broader field.

English does not lack this distinction.

"Hard Sciences" refers to scientific inquiry that is empirical in nature and has results that can be reproduced and confirmed independently. eg: most physics and chemistry

"Soft Science" refers to the rest. eg: psychology

These categories exist, but parent's main example was diet/nutrition, which should fit into the hard science bucket because it's chemistry/biology, but currently involves a lot of soft science-y type studies because it's got so many moving parts.

I'm not sure making a distinction like that is particularly useful, in any case. I think perhaps that people who have studied a lot of science can already make the distinction fairly easily, and having phrases like hard and soft science just serves to create assumptions where they needn't exist.

Some would group biology into a soft science because the margins for error must be relaxed or softened; one cannot eliminate potentially confounding factors from a biological experiment because each organism is unique and fractally complex.

Another categorization is the "natural sciences" and the "social sciences". Natural science is often split into "life science" and "physical science", again because biology is difficult.

I also wonder if we can take it a step further and apply a sort of "instance" versus "principle" science. 'Instance Science' is composed of studies that are trying to observe or experiment with something that is heavily influenced by uncontrollable variables and is highly likely to change. In some sense, the results are more like a snapshot in time than a durably reproducible phenomenon. What we typically call "soft science" and all those studies facing a reproducibility crisis fall into this bucket. Instance Science maps poorly onto "the real world".

Contrast that with 'Principle Science' in which studies are not affected by nearly as many uncontrollable variables and is more closely related to demonstrable cause-and-effect phenomena. The best examples are chemistry and physics. Biology is tricky to categorize in this because I see elements of both in it. For instance, a study investigating whether or not taking an increased dose of Vitamin B helps energy would most certainly belong in Instance Science, but the underlying mechanism of how Vitamin B is involved in the Krebs Cycle is Principle Science.

This idea is still in it's infancy and I'm curious to know people's thoughts on this distinction I'm trying to elaborate on.

Nutrition isn't a "hard science" because people digest food differently, and that varies over time. We adapt. Your first week of a bean diet will be harsh but after the 3rd year you're probably ok. or dead. Some people wouldn't ever adapt to it.

There's "hard science" there but to throw a rope around the whole field is more of an exercise in faith, that there is One True Diet for All People.

Not at all. Adequate science would tease out all the suitable variables for each individual’s diet at any time and situation for their stage of life. Which will include the details of current internal biome, current infections, medical history of their digestive and other organ systems, metabolism cycles, and many more things. And for which outcome where outcomes compete: cancer likelihood, bone health, sperm count, mental dexterity, fat content, etc.

The fact that there are too many variables and that it’s overly challenging to adequately measure them, coupled with challenges in studying people (ethics, self-assessment blind spots, laws against various options) makes nutrituon a squishy science, neither soft (people stuff like economics or psych) nor hard.

Like weather forecasting but worse. Nicely put. "oobleck science" on this continuum perhaps.

Does the fact that people digest food differently make a difference? I feel like the way food is digested is "knowable" in a way that physics is knowable, we just don't have adequate tools to measure all the complexity yet. As opposed to lots of things about sociology being "unknowable," like Asimov's Foundation being fantasy (probably).

In any case, I was attempting to make the point that hard science and soft science are anything but settled categories, which seems borne out by the responses.

I am not a big fan of these distinctions. In fact, they're completely backwards when you look at it from the perspective of difficulty of the sciences. The so-called hard sciences, like physics and chemistry, are far easier than biology, ecology, psychology, etc. These difficulties should not mean that they are lesser as implied by the normal usage of "hard" vs "soft" sciences.

> empirical in nature and has results that can be reproduced and confirmed independently

All sciences conform to this. It's just that reproducing results in physics and chemistry is much, much easier and feasible than in the other sciences.

"hard" doesn't refer to the difficulty level of "hard sciences", but the solidity of the evidence. Basically the same as "hard" evidence in a court case (video footage of a murder, procured from a trusted source, with witnesses and murder weapon intact, as opposed to mere circumstancial evidence).

I know they don’t and didn’t say they did. However, there are implications (if you’ve ever worked with physicists or even engineers you’ll feel the results of these implications), so I suggested that if you do look at it from a different perspective, then you get what I said.

I just don’t think they’re useful terms. In many ways, physics and chemistry are the low hanging fruit of science.

Why do you think it's about the subject, and not a reference to our ability to understand the non-linear outputs from a huge number of inputs?

Personally, I believe that psychology could be a hard science if it weren't for the limitations of our tools.

For now, psychology is too hard (meaning difficult) to be a hard (meaning solid) science.

Exactly! It's about the limitations of our tools, which means as the tools evolve, subjects shift more towards hard science as we understand them better.

It's not a category that is immovable.

> I'm actually struggling to articulate the concept simply because I lack the ability to cleanly categorize these things.

not sure if that categorization is possible.

> Despite this, the study benefits from all the clout established by the broader field.

here an example of AI and tech mumo-jumbo co-opted by psychologists and sold as "fact" to an audience of psychologists (I stumbled over this today while browsing the Psychology section of Springer):

"Chatbots to Support Mental Wellbeing of People Living in Rural Areas" https://link.springer.com/article/10.1007/s41347-021-00222-6

>> Gamification could be used within the chatbot to increase user engagement and retention. Content within the chatbot could include validated mental health scales and appropriate response triggers, such as signposting to external resources should the user disclose potentially harmful information or suicidal intent. Overall, the workshop participants identified user needs which can be transformed into chatbot requirements.

>> In addition to supporting those with mental ill health, digital technologies are also considered to have potential for preventing mental health problems and for improving the overall mental health of the population

>> Further research is necessary to try to equip chatbots with an understanding of emotion-based conversation and appropriate empathic responses, to adjust their personality and mimic emotions

a classic example because it even ignores that there is little to no research within psychology about the effects of when we condition vulnerable groups to pour emotions into these robotic "empathy sinks".

You can't embed this in language because it's fundamentally a contingent property, not intrinsic to the statements themselves. You end up with inadequate words like "settled" science.

I hate to beg the question, but what's stopping us from creating new terminology ourselves? English is malleable and I've witnessed new categories and distinctions spring up in my lifetime. For instance, 'cis' for gender identity matching biological sex. Growing up I never had a word for that concept and I don't think academia adopted the term until relatively recently.

The difficult part is deciding on the definition of the concept, not picking the name of the concept. There are lots of different terms and concepts you could be using to categorize types of science, but most of them probably don't match what you care about.

"cis-" as a prefix for the opposite of "trans-" dates back to Latin ("cisalpine Gaul"); it was first used in relation to gender identity in German in the early half of the 20th century, but all the pioneering research by German doctors was destroyed by Hitler.

Well, lines are fuzzy, at least on one side I would say: if the science has matured enough to be used in engineering it is objective. That doesn't mean evolution is not.

Ultimate truth discovery is the ideal, but in reality all science is biased or follows trends to some extent. Scientists aren't less susceptible to faults than the rest of us.

The reproduction crisis goes to show that objective science is much more myth than reality. Most people need the money, therefore they need the grants.

So much more objective science and thus progress needs tenure and randomized first grants?

> If science is a set of methods and tools to discover truth

Science, as practiced, is a process that consumes dollars and PhDs and outputs peer reviewed papers.

If that's the line, then there can be no objective science, as all funding organizations have motives beyond curiosity.

Isn't it the perfect solution fallacy?

Take a government funding basic research, with some hope there will be profit from this. Take an industry which damages public health and funds lies to cover this up.

Both have motives beyond curiosity, yet I think that there is at least a chance that the government-funded research is objective.

If you take it too far, sure. But I'd suggest that the notion of "objective" is itself the idea taken too far.

In practice, we all have material interests in the world. I think it's much more useful to think of "objective" not as a binary or a destination that one arrives at. But more a direction, like "up", that one moves in.

I think David Deutsch has suggested that to be objective is to be sincere in identifying and correcting errors in one’s analysis.

I'm for that, but it's a very subjective definition of objectivity.

Well yeah, but industry has funded basic research with some hope there will be profit, and governments have damaged public health and lied to cover it up, so they're basically the same unless you happen to know the motive in a given case.

Reply on news.ycombinator.com