Stop Using Scientific Language

[integral]

Quote

 

Here's the working list of the most common obfuscation moves in research papers, roughly ordered by how often you'll encounter them:

1. Nominalization. Turning verbs into nouns. "We measured" → "measurements were obtained." "We decided" → "a determination was made." Strips out the human agent and makes the action sound like a thing that just happened.

2. Passive voice. "It was found that..." "The data were analyzed..." "Subjects were administered..." Hides who did what. Often combined with nominalization for maximum agency removal.

3. Latinate word inflation. Replacing short words with longer Latin-derived ones.

use → utilize

show → demonstrate

start → initiate

end → terminate

enough → sufficient

help → facilitate

about → approximately

before → prior to

after → subsequent to

get → obtain

make → generate

try → attempt

need → require

a lot → a substantial amount

4. Hedge stacking. Piling qualifiers so claims become unfalsifiable. "These results may potentially suggest a possible tendency that could conceivably indicate a relationship that warrants further investigation." Each hedge is defensible alone; together they say nothing.

5. Weasel quantifiers. "Some," "many," "often," "frequently," "a number of," "increasingly" — used without numbers attached. Sounds substantive, commits to nothing. "Many researchers believe..." (how many? which ones?).

6. Throat-clearing phrases. Filler that signals seriousness without adding content.

"It is important to note that..."

"It should be emphasized that..."

"In light of the foregoing..."

"It is worth mentioning..."

"It is well established that..." (often followed by something that isn't well established)

"A growing body of literature suggests..."

"Recent advances have shown..."

7. Methodological inflation. Dressing up simple procedures in technical clothing.

"Conducted a survey" → "employed a cross-sectional quantitative instrument"

"Interviewed people" → "performed semi-structured qualitative inquiry"

"Looked for patterns" → "applied thematic analysis using constant comparative methodology"

"Asked people questions" → "administered a validated self-report instrument"

8. Statistical hedging language. Specific to quantitative papers.

"Trended toward significance" (it wasn't significant)

"Approached significance" (it wasn't significant)

"Marginally significant" (it wasn't significant)

"A robust association" (often a weak correlation)

"Strongly suggests" (correlates with)

"Consistent with the hypothesis that..." (didn't actually prove the hypothesis)

9. Citation chaining. Stacking references to manufacture consensus. "(Smith 2003; Jones 2007; Patel 2011; Lee 2015)" where each paper cites the previous and the original finding was weak. Reader sees four citations, assumes settled science.

10. Theory name-dropping. Invoking a framework to foreclose challenge. "Drawing on a Foucauldian lens..." "From a Bourdieusian perspective..." "Through the framework of intersectionality..." Now disagreement requires arguing with the theorist rather than the claim.

11. Acronym soup. Inventing or using field-specific acronyms that exclude outsiders. By page three the paper is unreadable to anyone outside the subfield, which is sometimes the point.

12. Tautological dressing. Restating the obvious in technical language. "Participants who reported higher levels of subjective wellbeing tended to indicate greater life satisfaction" = "happy people said they were happy."

13. Negative result laundering. Reframing failed findings as discoveries.

"Findings did not reach conventional thresholds of significance, suggesting the need for further investigation with larger samples." (It didn't work.)

"While the primary hypothesis was not supported, exploratory analyses revealed..." (We went fishing after the fact.)

"The intervention showed promise in certain subgroups." (It failed overall, but we found one slice where it looked good.)

14. Limitation burial. Tucking real problems into a "Limitations" section near the end, written in maximally bland language so reviewers and journalists skim past them.

15. Discussion-section drift. The Results say X. The Discussion says X probably means Y. The Abstract says Y. The press release says Y causes Z. Each step adds confidence the data doesn't support, and most readers only read the abstract.

16. The royal scientific "we." Single-author papers using "we" to sound institutional. Multi-author papers using it to diffuse responsibility — if the finding is challenged, no individual owns it.

17. Defensive precision theater. Reporting numbers to absurd decimal places to imply measurement precision the method can't actually deliver. "Participants showed a 23.847% improvement" from a sample of 40 people on a self-report scale.

18. Bridging language between weak claims. "Taken together, these findings suggest..." "Considered as a whole, the evidence indicates..." Used to make a collection of weak results sound like a strong conclusion when none of the individual pieces could carry it alone.

19. Future-tense escape hatch. "Further research is needed to..." "Future studies should examine..." Acknowledges the current work doesn't prove the thing while implying the thing is basically true and just awaits confirmation.

20. Operationalization sleight of hand. Defining a contested concept narrowly to make it measurable, then drawing broad conclusions about the contested concept. "Intelligence" gets operationalized as "score on this 30-minute test," findings are about the test scores, conclusions get stated as if they're about intelligence.

The diagnostic test for any passage is to rewrite it in plain English with specific agents, concrete verbs, and actual numbers. If the claim survives that rewrite, the original complexity was carrying meaning. If the rewrite makes the claim sound trivial or unsupported, the complexity was carrying authority instead of meaning.

 

 

Why would anyone who cares about truth play all these games when writing a research paper?

Obscuritism 🌈

It blows my mind that most scientist are not conscious that they are partaking in obscuritism. 🤯

The first Insight that that should come to someone's mind when they are exposed to Scientific language is that their intentionally fluffing it all up to make them sound as smart as possible. Which is one of the stupidest things you can do, and lacks intellectual integrity

Edited 3 hours ago by integral

[Jacob Morres]

True

[Jirh]

Very nice and important. Thank you!

Who compiled this, by the way?

[Jirh]

2 hours ago, integral said:

Why would anyone who cares about truth play all these games when writing a research paper?

I think it mostly has to do with control and power dynamics. Just like how a social media influencer would speak with confidence to get their point through and make it take effect. They have to demonstrate absolute confidence and certainty and sophistication to gain your trust and establish their authority. Typical human behaviour, and totally understandable too. A humble scientist will not be taken seriously by the layman and will not be respected by peer scientists. So scientists have to play these games, even if mostly subconsciously.

Edited 29 minutes ago by Jirh

[QVx]

Enjoyed that, thanks for sharing.

Also, none of those things are obvious deception to most people, least of all to the persons writing those into their papers.

It takes a lot to stand back and see it, let alone then detect it without heavy attention.

[Natasha Tori Maru]

This sort of thing isn't unique to science.

[zurew]

I think some of that is obscuring things, but there are bad examples in your list because depending on the context it makes sense to use slightly different words and the other thing that you forget is that scientists are not writing researchg papers for layman, they are writing it for their peers who are already competent and familiar in the area.

The implication that you are making is that for every given specific field, everyone should only use natural language and should only use natural language the way layman do, and that it is just a bad requirement to make, because it would make writing papers and understanding papers harder, because of the introduced ambiguity.

Like imagine if you would make the exact same criticism about math and to lets ditch symbols like "/,*, +,-,=" etc, lets ditch algebra and then express all mathematical operations in natural language only. If you thought math was hard, then try to express and understand math only using layman , natural language from now on (if thats even possible).

Or I dare you to express some extremely dense physics equation only using natural language and lets see whats your shortest most understandable way to express and explicate the meaning of said equation (I have no idea wtf is going on here, im just giving it as an example)

Edited 1 hour ago by zurew

[Jirh]

2 minutes ago, zurew said:

I think some of that is obscuring things, but there are bad examples in your list because depending on the context it makes sense to use slightly different words and the other thing that you forget is that scientists are not writing researchg papers for layman, they are writing it for their peers who are already competent and familiar in the area.

The implication that you are making is that for every given specific field, everyone should only use natural language and should only use natural language the way layman do, and that it is just a bad requirement to make, because it would make writing papers and understanding papers harder, because of the introduced ambiguity.

Like imagine if you would make the exact same criticism about math and to lets ditch symbols like "/,*, +,-,=" etc, lets ditch algebra and then express all mathematical operations in natural language only. If you thought math was hard, then try to express and understand math only using layman - natural language from now on (if thats even possible).

Agreed, but the criticism (at least for me) isn't of the use of language to express a perspective, rather it's of the abuse of that feature.

Imagine instead of using + for arithmetic operations, I use it to draw a dog, in the form of ASCII Art. But I don't share it as art. I share it as truth. That's obviously problematic.

The language has its place and time and usage. That still doesn't excuse the abuse.

[Jirh]

8 minutes ago, zurew said:

Or I dare you to express some extremely dense physics equation only using natural language and lets see whats your shortest most understandable way to express and explicatge the meaning of said equation (I have no idea wtf is going on here, im just giving it as an example)

It's actually very possible to express any physics or maths equation in natural language. It's how we originally learn the abstracted language in schools and universities.

For example, E = 1/2 mv^2 is easily translated as Energy equals half the mass multiplied by the velocity squared.

[zurew]

16 minutes ago, Jirh said:

Imagine instead of using + for arithmetic operations, I use it to draw a dog, in the form of ASCII Art. But I don't share it as art. I share it as truth. That's obviously problematic.

Pretty crazy how many things AI and people can express in ASCII.

12 minutes ago, Jirh said:

For example, E = 1/2 mv^2 is easily translated as Energy equals half the mass multiplied by the velocity squared.

Sure but there you are making a big info reduction by assuming shared context. The reason why its enough to say it in natural language the way you did, is because you assume that the reader will interpret your natural language in a physics specific way.

Almost all of those terms can be used outside the context of physics. For instance "energy" is extremely ambigous and is used in a 100 different way outside of physics.

Now do it for the whole equation and express even numbers in natural language.

Edited 1 hour ago by zurew

[Jirh]

58 minutes ago, zurew said:

Pretty crazy how many things AI and people can express in ASCII.

I know, right? It's amazing 

Quote

Sure but there you are making a big info reduction by assuming shared context. The reason why its enough to say it in natural language the way you did, is because you assume that the reader will interpret your natural language in a physics specific way.

Of course, that's because there's always a shared context. We are all humans, sharing some context with each other. If there's no shared context, it wouldn't even make sense to try and communicate anything in the first place. Like for example, the equation you posted. I don't have a damn clue what it means. It would be pointless for a scientist to come and cite it for me to convince me of some idea. They would be crazy to even attempt that.

I think much of the friction is happening because some specialized topics and concepts are leaking to the layman through easy access and social media. But that's not all. Sometimes, the concepts are used on purpose, to confuse, hypnotize, and appear sophisticated. Like imagine someone goes to the doctor, and instead of them explaining the patient's specific problem, they start spouting advanced medical terms at them. Partly, they might not be even aware that they're doing this, because their knowledge might feel like common sense or second nature to them after years of practice. But the problem remains that you don't understand your condition.

"Oh, it looks like you have meningoencephalitis, we're going to have to initiate a targeted anti-inflammatory and antimicrobial intervention protocol with close hemodynamic monitoring and serial lumbar punctures as indicated by your neurological status."

Edited 29 minutes ago by Jirh