People absolutely share uncertain things. A lot of papers have discussions of things that need further scrutiny. But closer to what you’re asking is wha happens at conferences, during talks, or in casual meetings and email exchanges. People are very upfront about what is uncertain or unknown, what has been tried and not worked out, etc. when they speak in person.
Should such things be published more online? Probably, but it’s a lot of effort to go through for researchers either already tight schedules.
It’s highly suggestive that you are not familiar with scientific publishing if you think “no one” accounts for, addresses, engages with and values the conditional, the possible, the uncertain.
Your premise is wrong. I looked up a preprint server and chose the most recent submission. The following section is from the discussion and I have emphasised language which is strictly observational, hedges, qualifies, or otherwise couches terminology to imply the potential for uncertainty:
> While the mechanisms by which CKX2 mutation enhances root development remain
incompletely understood, it is well established that auxin is the principal regulator of root architecture and interacts antagonistically with cytokinin. Notably, this antagonism appears
unidirectional: elevated cytokinin does not suppress auxin function, especially at higher
hormone concentrations41. Studies of other CKX family members, such as OsCKX4, have
shown that upregulation promotes robust root systems, increasing both the initiation and
development of crown roots42. In the present study, KAMALA seedlings exhibited elevated
OsCKX4 and other CKX mRNA levels at the seedling stage, when foundational root
development occurs (Fig. 4f and supplementary Fig. 13). Strikingly, biochemical analysis of
KAMALA roots failed to detect cytokinin, while auxin levels closely mirrored those in the
parental Samba Mahsuri line (Supplementary Fig. 20). This intimates that despite increased
cytokinin in inflorescence tissues (driving enhanced reproductive and grain traits), cytokinin is depleted or absent in developing roots, potentially enabling auxin-mediated root growth to proceed unimpeded. Thus, modifications in CKX2 may lead to a spatial rebalancing of phytohormones within the plant, facilitating superior root and shoot development without the negative feedback typically associated with enriched cytokinin levels.
The “certainty” is often introduced not at the stage of publication, nor when retransmitted by decent science communicators, but in the context of scientific controversies (or other forms of elevated public interest, which is probably just saying the same thing) where scientific research is over-summarised for public consumption. This also happens when the textbooks are written, and uncertainties about CKX2 modifications may become certainties for the benefit of the student trying to learn more.
The problem in this sense is not the publication but transmission of scientific knowledge, which accrues certainty as it is simplified and reshared, even though it is properly qualified in the journal article.
How do you filter out the garbage, especially in a world with generative AI, where it is getting easier to create?
Everyone replicating every study before they can trust it probably isn't feasible. On the other hand, if every paper was delivered as part of a reproducibility package, maybe garbage filtering would be a little easier.
Definitions for clarity:
Replication: Redoing an entire study, including raw data collection.
Reproduction: Redoing the analysis and all other artifact generation with the same raw data.
I believe what you are describing is due to the broad ignorance of the history of science. Specifically, very few people are even capable of engaging in epistemological thought. Even Einstein said that science without epistemology is--insofar as it is thinkable, muddled at best. Practitioners in many fields can't differentiate between high-quality evidence and low-quality evidence, and this often fall back on the vestiges of logical positivism. Personally, I believe philosophy of science holds the key to understanding many contentious issues in science that have devolved into ideological wars, when really the differing sides are taking past each other from what Kuhn called incommensurable paradigms. And since philosophy programs are literally dying in academia right now, it's hard to see this getting any better anytime soon.
There is a whole body of research dedicated to the academic system as a form of meta-knowledge. Like whole institutes, thousands of academics globally. It’s Critical Higher Education studies.
The consensus, in private, is no it cannot change as a rupture of status quo requires academics to challenge their means of income, per se. What puts food on the table and what supports their life, family if they have one.
In public, it is extremely difficult or any other similar set of words academics like to describe unlikelihoods.
The key words for your exploration are: neoliberal higher education, marketisation of research, siloing of science, the ‘PhD’ as (un)fit for purpose and narrowing of research horizons (Nat. Science mainly).
Complexity theory and cybernetics as a descriptor of (un)certainty in research.
I myself work in the sciences (physics), so perhaps I can clarify some things regarding how scientific results are presented. There are different types of articles; by this I mean that the main articles are often called "research papers" and are those that present original results.
By original, we mean presenting new knowledge related to a given problem that provides a significant advance over what was already known about the problem in question. Obviously, defined like this, it is necessary that the results be valid and solidly demonstrated (either through calculations and demonstration in the case of theoretical work, or through experimentation in the case of more experimental work; sometimes both are used in a single study).
Advancement of research in a given field is only possible if it is based on previously established results. Nevertheless, it is possible to discuss less precise topics, even to speculate on certain things, but this must be done according to certain rules.
For example, in the case of a research paper, it is important to structure the article in such a way as to clearly distinguish the state of the art (the starting point of the work), the problem (the question) to be solved or of interest, the method used to answer the question(s), the results obtained, and finally their interpretation within the framework of the research field.
It is especially this last part that allows the author to discuss things that may not yet be well understood, to see how the new results could be interpreted in different contexts, and to open the door to new questions and inquiries.
There are also articles of the "perspectives" or "opinions" type. In general, this type of article does not present new results, but rather offers a structured discussion around certain problems, unknowns, and the associated results. In this type of article, authors can discuss less well-understood aspects, formulate questions, propose new hypotheses, and suggest theoretical or experimental scientific approaches that could validate or invalidate some of the proposed hypotheses.
I think this is a valid critique of public-facing media, but would argue that most scientists are far more inclined to nuance and would rarely describe their research in terms of absolute certainty.
As a biology professor at a small liberal arts college, I've noticed a deterioration of language skills that makes communicating that nuance to students difficult and I think reflects a broader tendancy towards confirmation bias, oversimplification, and an expectation of definitive proof of things due to an overemphasis on memorization rather than critical thinking.
I want to believe a culture shift is possible, otherwise I wouldn't be doing what I do, but I do worry a lot about the impact of anti-intellectualism on the next generation of scientists.
I would like to see some more systematic and nuanced way to evaluate the validity of results. "This is how things work" ... what is the range of situations in which this has been tested? Do we already know of situations where things don't actually work that way? I would like to see some sort of function that maps Situation to Reliability. Probably a Bayesian sort of reliability - in situations where no testing has been done, what other theories are plausible?
I have not seen any uniform approach at all. How findings are published varies a lot. It depends on where it gets published. There are different fields in STEM with quite different notions of what a paper should look like. Even just different journals can have quite different cultures.
And then there are different sorts of publications, from research letters, to survey papers, to textbooks, to the popular press.
For a very interesting angle on some related history (the rise of scientific publishing as a highly profitable business), check out the Behind the Bastards podcast on Robert Maxwell. He’s the father of the whole industry and also happens to be the father of Ghislaine Maxwell of Epstein fame (and it’s where she got all her money to become an heiress). It’s quite a story on a lot of different levels.
Science, currently, is a social game of collaborations and citations, not a particular concern for truth for its own sake. This is to be expected given the human condition, and the only thing that keeps things veridical are the controls that are expected to be requisite of every publication. But given that people/scientists are driven primarily to maximize prestige, there's just a lot of noise out there.
SpectralFormFactor | 7 hours ago
People absolutely share uncertain things. A lot of papers have discussions of things that need further scrutiny. But closer to what you’re asking is wha happens at conferences, during talks, or in casual meetings and email exchanges. People are very upfront about what is uncertain or unknown, what has been tried and not worked out, etc. when they speak in person.
Should such things be published more online? Probably, but it’s a lot of effort to go through for researchers either already tight schedules.
knockingatthegate | 4 hours ago
It’s highly suggestive that you are not familiar with scientific publishing if you think “no one” accounts for, addresses, engages with and values the conditional, the possible, the uncertain.
knockingatthegate | 4 hours ago
Give a single example.
histogrammarian | 3 hours ago
Your premise is wrong. I looked up a preprint server and chose the most recent submission. The following section is from the discussion and I have emphasised language which is strictly observational, hedges, qualifies, or otherwise couches terminology to imply the potential for uncertainty:
> While the mechanisms by which CKX2 mutation enhances root development remain incompletely understood, it is well established that auxin is the principal regulator of root architecture and interacts antagonistically with cytokinin. Notably, this antagonism appears unidirectional: elevated cytokinin does not suppress auxin function, especially at higher hormone concentrations41. Studies of other CKX family members, such as OsCKX4, have shown that upregulation promotes robust root systems, increasing both the initiation and development of crown roots42. In the present study, KAMALA seedlings exhibited elevated OsCKX4 and other CKX mRNA levels at the seedling stage, when foundational root development occurs (Fig. 4f and supplementary Fig. 13). Strikingly, biochemical analysis of KAMALA roots failed to detect cytokinin, while auxin levels closely mirrored those in the parental Samba Mahsuri line (Supplementary Fig. 20). This intimates that despite increased cytokinin in inflorescence tissues (driving enhanced reproductive and grain traits), cytokinin is depleted or absent in developing roots, potentially enabling auxin-mediated root growth to proceed unimpeded. Thus, modifications in CKX2 may lead to a spatial rebalancing of phytohormones within the plant, facilitating superior root and shoot development without the negative feedback typically associated with enriched cytokinin levels.
The “certainty” is often introduced not at the stage of publication, nor when retransmitted by decent science communicators, but in the context of scientific controversies (or other forms of elevated public interest, which is probably just saying the same thing) where scientific research is over-summarised for public consumption. This also happens when the textbooks are written, and uncertainties about CKX2 modifications may become certainties for the benefit of the student trying to learn more.
The problem in this sense is not the publication but transmission of scientific knowledge, which accrues certainty as it is simplified and reshared, even though it is properly qualified in the journal article.
Bach4Ants | 7 hours ago
How do you filter out the garbage, especially in a world with generative AI, where it is getting easier to create?
Everyone replicating every study before they can trust it probably isn't feasible. On the other hand, if every paper was delivered as part of a reproducibility package, maybe garbage filtering would be a little easier.
Definitions for clarity:
liccxolydian | 6 hours ago
Physics tried this with string theory, nowadays it's often used as an example in anti-academic and anti-intellectual propaganda.
wiggum_bwaa | 6 hours ago
I believe what you are describing is due to the broad ignorance of the history of science. Specifically, very few people are even capable of engaging in epistemological thought. Even Einstein said that science without epistemology is--insofar as it is thinkable, muddled at best. Practitioners in many fields can't differentiate between high-quality evidence and low-quality evidence, and this often fall back on the vestiges of logical positivism. Personally, I believe philosophy of science holds the key to understanding many contentious issues in science that have devolved into ideological wars, when really the differing sides are taking past each other from what Kuhn called incommensurable paradigms. And since philosophy programs are literally dying in academia right now, it's hard to see this getting any better anytime soon.
First-Beginning-7513 | 5 hours ago
There is a whole body of research dedicated to the academic system as a form of meta-knowledge. Like whole institutes, thousands of academics globally. It’s Critical Higher Education studies.
The consensus, in private, is no it cannot change as a rupture of status quo requires academics to challenge their means of income, per se. What puts food on the table and what supports their life, family if they have one.
In public, it is extremely difficult or any other similar set of words academics like to describe unlikelihoods.
The key words for your exploration are: neoliberal higher education, marketisation of research, siloing of science, the ‘PhD’ as (un)fit for purpose and narrowing of research horizons (Nat. Science mainly).
Complexity theory and cybernetics as a descriptor of (un)certainty in research.
brainquantum | an hour ago
I myself work in the sciences (physics), so perhaps I can clarify some things regarding how scientific results are presented. There are different types of articles; by this I mean that the main articles are often called "research papers" and are those that present original results.
By original, we mean presenting new knowledge related to a given problem that provides a significant advance over what was already known about the problem in question. Obviously, defined like this, it is necessary that the results be valid and solidly demonstrated (either through calculations and demonstration in the case of theoretical work, or through experimentation in the case of more experimental work; sometimes both are used in a single study).
Advancement of research in a given field is only possible if it is based on previously established results. Nevertheless, it is possible to discuss less precise topics, even to speculate on certain things, but this must be done according to certain rules.
For example, in the case of a research paper, it is important to structure the article in such a way as to clearly distinguish the state of the art (the starting point of the work), the problem (the question) to be solved or of interest, the method used to answer the question(s), the results obtained, and finally their interpretation within the framework of the research field.
It is especially this last part that allows the author to discuss things that may not yet be well understood, to see how the new results could be interpreted in different contexts, and to open the door to new questions and inquiries.
There are also articles of the "perspectives" or "opinions" type. In general, this type of article does not present new results, but rather offers a structured discussion around certain problems, unknowns, and the associated results. In this type of article, authors can discuss less well-understood aspects, formulate questions, propose new hypotheses, and suggest theoretical or experimental scientific approaches that could validate or invalidate some of the proposed hypotheses.
FrontAd9873 | an hour ago
Are you a grad student or do you work in academia? I wonder in what field you’re observing this.
anonyblyss | an hour ago
I think this is a valid critique of public-facing media, but would argue that most scientists are far more inclined to nuance and would rarely describe their research in terms of absolute certainty.
As a biology professor at a small liberal arts college, I've noticed a deterioration of language skills that makes communicating that nuance to students difficult and I think reflects a broader tendancy towards confirmation bias, oversimplification, and an expectation of definitive proof of things due to an overemphasis on memorization rather than critical thinking.
I want to believe a culture shift is possible, otherwise I wouldn't be doing what I do, but I do worry a lot about the impact of anti-intellectualism on the next generation of scientists.
kukulaj | 7 hours ago
I would like to see some more systematic and nuanced way to evaluate the validity of results. "This is how things work" ... what is the range of situations in which this has been tested? Do we already know of situations where things don't actually work that way? I would like to see some sort of function that maps Situation to Reliability. Probably a Bayesian sort of reliability - in situations where no testing has been done, what other theories are plausible?
[OP] spider_in_jerusalem | 7 hours ago
Yeah, kind of a meta-awareness layer where you check for things that may have not been considered.
norb_151 | 2 hours ago
What you're asking for is literally the standard way how findings are published in STEM....
kukulaj | 10 minutes ago
I have not seen any uniform approach at all. How findings are published varies a lot. It depends on where it gets published. There are different fields in STEM with quite different notions of what a paper should look like. Even just different journals can have quite different cultures.
And then there are different sorts of publications, from research letters, to survey papers, to textbooks, to the popular press.
1E4rth | 4 hours ago
For a very interesting angle on some related history (the rise of scientific publishing as a highly profitable business), check out the Behind the Bastards podcast on Robert Maxwell. He’s the father of the whole industry and also happens to be the father of Ghislaine Maxwell of Epstein fame (and it’s where she got all her money to become an heiress). It’s quite a story on a lot of different levels.
seldomtimely | 3 hours ago
Science, currently, is a social game of collaborations and citations, not a particular concern for truth for its own sake. This is to be expected given the human condition, and the only thing that keeps things veridical are the controls that are expected to be requisite of every publication. But given that people/scientists are driven primarily to maximize prestige, there's just a lot of noise out there.