You're reading the public-facing archive of the Category Theory Zulip server.
To join the server you need an invite. Anybody can get an invite by contacting Matteo Capucci at name dot surname at gmail dot com.
For all things related to this archive refer to the same person.
Arthur Parzygnat (Jun 11 2020 at 07:09):TLDR: Are we (writers, editors, reviewers) doing enough to make our research papers and resources inclusive?
Original: Please disregard if this has already been brought up in another stream, but I was curious about the following idea and what opinions people had on this with regards to inclusivity on the way that we write, specifically research papers (and some books, though fewer). From the hundreds of (math, physics, CS) papers I have read/tried to read, only a few have stuck out as being very user friendly. This is in part due to a combination of language (using potentially demeaning sounding words such as "obvious" and "trivial"), not supplementing statements with references (for example, when one writes "it is well known that..." but then refuses to state a precise reference), a lack of accessible examples, and so on. I noticed at some point @John Baez mentioned in a stream in a response to a question raise by @Oliver Shetler regarding getting published:
John Baez said:
For example when I publish in TAC I'd never stoop to defining a lax monoidal functor, because in that journal it's assumed everyone knows what that is, and if I did define it people would instantly be suspicious that I was an outsider. I might however define the Heisenberg algebra, since that's a concept in mathematical physics.
This suddenly made me question the inclusivity of journals in general (don't get me wrong, I have actually found that category theory journals have been overwhelmingly welcoming!). But we have to acknowledge that the way we write math papers today was dictated by a very narrow group of individuals, those privileged enough to develop the subject. We learned from them and followed their footsteps. I am sure there are genuinely good aspects to this format, but is it too exclusive?
As we are growing more interdisciplinary in math and the sciences, it is more important to make sure our work is accessible to wider audiences. Although I might not define a lax monoidal functor in TAC, I should point to a reference where it is defined (imagine a non-category theorist who wants to understand a specific result in a category theory paper and wants to understand that result as painlessly as possible). We often find ourselves wanting to learn more about a subject but are shunned by a lack of background and guidance. Authors should provide references to the shortest route they know towards attaining that background. For example, would it hurt to encourage the following actions for journals, writers, reviewers?
(=_=) (Jun 11 2020 at 08:41):I should preface this by saying that all these suggestions are very desirable for improving the quality of scientific publication in general (this could also be a topic for #practice: communication). However, I'd like to discuss some nuances in the status quo that could make these suggestions problematic.
Arthur Parzygnat said:
For example, would it hurt to encourage the following actions for journals, writers, reviewers?
1. Avoid/discourage potentially demeaning language.
2. When certain tools and techniques are used but not explicitly written, provide a detailed reference and location within that reference. For example, don't just reference MacLane's entire book, but rather reference a specific section or whatever, or ideally a place where you feel the exposition is more accessible.
This is what copy-editors used to do, back when journals had these roles. I think a very select few still do, because the language of their published material tends to be more polished than what is normal today. These days, we apparently have no need for copy-editors, and reviewers are loathe to recommend minor changes because it can come across, ironically enough, as being not inclusive.
Point 1., for instance, demands of the author a certain sophistication. Some people (this may include ESL speakers) may not even realise that saying "obvious" or "trivial" is now considered demeaning, for example, or even feel that they're fully justified in saying so because some proposition appears to be transparently true to them. If the journal then insists on those changes, things could escalate to the point where a complaint of the journal not being inclusive enough could, ironically, be justified.
(=_=) (Jun 11 2020 at 08:52):Arthur Parzygnat said:
- If possible, include/encourage simple examples (I didn't realize how important this is until Bart Jacobs explicitly mentioned it) before progressing to more complicated ones. Personally, I don't think many math journals welcome this as they might consider the examples "trivial" or a waste of space.
- Express ideas in multiple ways (algebraically, visually, via examples, via analogies).
These also demand a certain level of sophistication that some people, particularly ESL speakers, feel is superfluous. I've met ESL speakers who don't care for writing more polished prose because they felt it wasn't necessary for them to write "beautiful English" to get their point across -- "beautiful" means, for them, an inferior quality better expressed by the word "florid". Point 5 feels like something that would also be rejected on the same grounds. Both of these requirements also demand more depth in the authors' knowledge, which could feel like a burden to some ESL speakers who're just starting out on their research career.
Oliver Shetler (Jun 11 2020 at 11:01):Arthur Parzygnat said:
TLDR: Are we (writers, editors, reviewers) doing enough to make our research papers and resources inclusive?
Original: Please disregard if this has already been brought up in another stream, but I was curious about the following idea and what opinions people had on this with regards to inclusivity on the way that we write, specifically research papers (and some books, though fewer). From the hundreds of (math, physics, CS) papers I have read/tried to read, only a few have stuck out as being very user friendly. This is in part due to a combination of language (using potentially demeaning sounding words such as "obvious" and "trivial"), not supplementing statements with references (for example, when one writes "it is well known that..." but then refuses to state a precise reference), a lack of accessible examples, and so on. I noticed at some point John Baez mentioned in a stream in a response to a question raise by Oliver Shetler regarding getting published:
John Baez said:For example when I publish in TAC I'd never stoop to defining a lax monoidal functor, because in that journal it's assumed everyone knows what that is, and if I did define it people would instantly be suspicious that I was an outsider. I might however define the Heisenberg algebra, since that's a concept in mathematical physics.
This suddenly made me question the inclusivity of journals in general (don't get me wrong, I have actually found that category theory journals have been overwhelmingly welcoming!). But we have to acknowledge that the way we write math papers today was dictated by a very narrow group of individuals, those privileged enough to develop the subject. We learned from them and followed their footsteps. I am sure there are genuinely good aspects to this format, but is it too exclusive?
As we are growing more interdisciplinary in math and the sciences, it is more important to make sure our work is accessible to wider audiences. Although I might not define a lax monoidal functor in TAC, I should point to a reference where it is defined (imagine a non-category theorist who wants to understand a specific result in a category theory paper and wants to understand that result as painlessly as possible). We often find ourselves wanting to learn more about a subject but are shunned by a lack of background and guidance. Authors should provide references to the shortest route they know towards attaining that background. For example, would it hurt to encourage the following actions for journals, writers, reviewers?
- Avoid/discourage potentially demeaning language.
- When certain tools and techniques are used but not explicitly written, provide a detailed reference and location within that reference. For example, don't just reference MacLane's entire book, but rather reference a specific section or whatever, or ideally a place where you feel the exposition is more accessible.
- Specify notation (I'm proud of the mathematical community for doing this to a reasonable extent, at least from the papers I have read)
- If possible, include/encourage simple examples (I didn't realize how important this is until Bart Jacobs explicitly mentioned it) before progressing to more complicated ones. Personally, I don't think many math journals welcome this as they might consider the examples "trivial" or a waste of space.
- Express ideas in multiple ways (algebraically, visually, via examples, via analogies).
- Prove statements that are used but not proved elsewhere (this also makes it easier for reviewers to verify work).
There are probably many other suggestions to be made, but that's what's on my mind for now. (It's very likely I've made these mistakes by the way, and I'm writing them down to make me aware of this so that I can improve.)
This might be my naïveté speaking, but I would generally hope––in my own work––to write a highly accessible paper first, and then edit the pedagogy out of it for publication, then publish the accessible version next to the condensed version on my personal site / XArchives. I believe that this would make the final draft better in the same way that world building makes a science fiction or fantasy novel better, even though almost none of it makes it into the final draft. It might not even be that much more work because having concrete examples, intuitions and motivations in your mind can make the path from first draft to polished paper shorter. How reasonable or unreasonable does this sound to the ear of someone who has actually published in mathematics journals?
Morgan Rogers (he/him) (Jun 11 2020 at 11:07):Arthur Parzygnat said:
Authors should provide references to the shortest route they know towards attaining that background.
One day a decade from now, the tools will exist to make identifying that shortest route through the literature possible; this is one of the implicit goals of the (for now hypothetical) programme which has been discussed in #practice: communication . But most researchers have taken years to acquire the knowledge relevant to the research they're trying to publish. The skill of condensing that into a shortest entry path is arguably an art form that takes years more work to acquire. That's why it takes so many years for research-level mathematics to get condensed into a form that can filter down to be taught at graduate or undergraduate level.
I'm not saying that we shouldn't try, though! The most actionable version of what you're asking for would be to include these recommendations in publication guidelines (there are people reading this who might be able to do that) so that these issues can be flagged by reviewers in good faith. For it to work in the long term, some training (perhaps a pithy but accessible video tutorial?) about these guidelines which can be disseminated should be made.
(=_=) (Jun 11 2020 at 12:55):Oliver Shetler said:
This might be my naïveté speaking, but I would generally hope––in my own work––to write a highly accessible paper first, and then edit the pedagogy out of it for publication, then publish the accessible version next to the condensed version on my personal site / XArchives. I believe that this would make the final draft better in the same way that world building makes a science fiction or fantasy novel better, even though almost none of it makes it into the final draft. It might not even be that much more work because having concrete examples, intuitions and motivations in your mind can make the path from first draft to polished paper shorter. How reasonable or unreasonable does this sound to the ear of someone who has actually published in mathematics journals?
Using is hard work, so this is "that much more work" because it's painful to cut stuff out that you've already painstakingly typeset, even if you're going to put a "director's cut" later on. Hopefully this would become less of a chore with new tech that automatically generates the for you.
(=_=) (Jun 11 2020 at 13:01):On the other hand, if you've always been conscientious about ing your working notes, it's actually not that painful. So it'd work for someone who's very proficient with and uses it to write up notes every day, and very painful for someone who keeps notes in a handwritten format. This is why I find it important to improve the UX of writing maths.
(=_=) (Jun 11 2020 at 13:06):[Mod] Morgan Rogers said:
But most researchers have taken years to acquire the knowledge relevant to the research they're trying to publish. The skill of condensing that into a shortest entry path is arguably an art form that takes years more work to acquire.
One thing that often gets ignored is the economic implication of the length of time it takes to become good at something. There are reasons why people want that length of time to remain of the order of magnitude of years. For now, it's because we don't have the tech to do it, but once the tech gets close to being mature enough, things will get more "exciting".
John Baez (Jun 11 2020 at 15:39):Oliver wrote:
This might be my naïveté speaking, but I would generally hope––in my own work––to write a highly accessible paper first, and then edit the pedagogy out of it for publication, then publish the accessible version next to the condensed version on my personal site / XArchives. I believe that this would make the final draft better in the same way that world building makes a science fiction or fantasy novel better, even though almost none of it makes it into the final draft. It might not even be that much more work because having concrete examples, intuitions and motivations in your mind can make the path from first draft to polished paper shorter. How reasonable or unreasonable does this sound to the ear of someone who has actually published in mathematics journals?
People on the n-Category Café tend to do something a bit different. We often write blog articles summarizing our papers, and these summaries are often clearer and more fun to read than the original papers. How is this possible? One reason is that these summaries can leave out technicalities that are necessary in a fully rigorous paper. One can say "basically because of the Yoneda lemma" instead of writing out a detailed argument, or "see the paper for details". This lets us focus on the interesting stuff. People who want more details can read the paper.
John Baez (Jun 11 2020 at 15:39):I think this might be easier to do than first writing a big fat expository paper and then trimming it down.
John Baez (Jun 11 2020 at 15:42):Anyway, no matter what you do, if you write stuff that's easier to read than the usual math paper, make it easy for people to find (like on a statistics group blog, advertised on Twitter), and people will seek it out, and if they like it that will prove helpful to your career.
Daniel Geisler (Jun 11 2020 at 19:13):I wrote my first math paper in 1990, but it was horribly written. Thirty years later I am retiring without even publishing to https://arxiv.org/. So in 1999 I built http://tetration.org which has been successful in its own way.
Arthur Parzygnat (Jun 13 2020 at 08:49):Thanks for all the interesting pros and cons listed here! I really enjoyed reading these and seeing your perspectives.
Arthur Parzygnat (Jun 13 2020 at 08:49):@Rongmin Lu, you make an excellent point that some of what I'm suggesting is far more difficult for certain individuals. Speaking only from personal experience, I received excellent grammatical suggestions/corrections to papers I have submitted to category theory journals that (I hope) have improved (and will continue to improve) my writing. If I can say something in a clearer way, I would like to strive towards achieving that. Yes, it's easier to not do that, but I think we should all be heading towards a direction of improvement, not just for others, but also for ourselves, at whatever level we are currently at. On the other side, I have also judiciously corrected drafts, where initially the writer was upset about the enormous number of corrections, but soon realized the importance of those corrections and was able to see how much better their writing flowed. But yes, I do see your point and one should not go overboard (which is why I used the word "encourage" as opposed to "enforce"). It might be helpful to also reassure writers that certain things are suggestions and would improve readability but would not prevent their work from being published (which should itself be based on the content).
Arthur Parzygnat (Jun 13 2020 at 08:50):@Oliver Shetler and @John Baez, these are both great ideas. I have also seen other options, which includes making video summaries of a paper (see for example Tai's video on her thesis or Timothy Nguyen's video on path integrals ... btw, @John van de Wetering might find this second video interesting since he's learning about QFT from a more mathematically rigorous POV), but this seems to take a lot of effort. In either way, I'm worried about what happens many years from now when this information is lost or difficult to find and only the article remains. Though perhaps this may be related to @[Mod] Morgan Rogers, who makes a good point that in the future, it is very likely that locating information will become even more efficient. For example, you can use google scholar to look up papers that cited a specific paper and then search all those papers for specific key words to see if they were used. Imagine how time consuming this would have been just a decade ago.
(=_=) (Jun 13 2020 at 12:20):Arthur Parzygnat said:
Speaking only from personal experience, I received excellent grammatical suggestions/corrections to papers I have submitted to category theory journals that (I hope) have improved (and will continue to improve) my writing.
It's good to hear that you have such a positive experience.
On the other side, I have also judiciously corrected drafts, where initially the writer was upset about the enormous number of corrections, but soon realized the importance of those corrections and was able to see how much better their writing flowed.
Yeah, this is the kind of reaction I've heard about. I think less confrontational or persuasive reviewers may be averse to the prospect of this happening, and thereby letting things slide. I've certainly read my fair share of badly-written publications.
It might be helpful to also reassure writers that certain things are suggestions and would improve readability but would not prevent their work from being published (which should itself be based on the content).
Yeah, that's definitely something worth implementing.
John van de Wetering (Jun 16 2020 at 12:57):A while back I read through the publication guidelines and requirements for the journal Nature. And it explicitly says that quality of writing is not a factor in whether the work will be published. I guess this is to be inclusive towards non-native speakers, but I still find it odd. A little bit more effort on the writing side can result in a lot of time saved on the reading side (since there would hopefully be way more people reading your article then writing it).
David Michael Roberts (Jun 29 2020 at 02:53):And if you can't say in four pages with numerous plots the size of postage stamps, it's not terse enough.
David Michael Roberts (Jun 29 2020 at 02:54):I recently read a paper where the author referenced Johnstone's Elephant for a specific result, with nothing as to where it was in there. Wasn't happy about that one.
(=_=) (Jun 29 2020 at 02:56):I've recently seen several papers where the author refers to several "famous theorems" without citation.
David Michael Roberts (Jun 29 2020 at 02:59):And it was more of a vague specific result, so roughly the halves of two disjoint subsections put together. Nothing so easily localisable as an actual theorem/etc statement.