Thread 2: A corpus-assisted discourse analysis of linguistic transphobia on Mumsnet

[VitoCorleoneOfMNMafia]

At fewer than 20 posts left, time for part two.

Thread 1: https://www.mumsnet.com/talk/womens_rights/5057460-a-corpus-assisted-discourse-analysis-of-linguistic-transphobia-on-mumsnet

Site Stuff thread that tipped off MNHQ: https://www.mumsnet.com/talk/site_stuff/5057903-mumsnet-corpus

There’s something really wrong at Aston if students are being encouraged to present their ethically sensitive work at a public forum before it’s got ethics clearance. Discussing a project at an internal seminar can be a useful part of the process of shaping and nuancing the research questions and methodology and indeed the ethical issues, but I wouldn’t expect it to be at an event that is publicly advertised and bookable on Eventbrite.

If I understand this correctly, it seems, in part, that FoLD obtained it because of the corpus data to which they control access…

I thought I’d have a look for transactivist/trans affirming output from linguistic disciplines, perhaps Eden can pivot towards something positive rather than accusing mothers of spurious crimes.

Found this quite interesting but also extremely dated? It just screams 2020-2021 the way a black Instagram square or a handmade cotton face mask do:

https://www.whysocialscience.com/blog/2021/6/29/because-trans-activism-can-change-how-we-understand-language

Eden Palmer's paper is listed on the upcoming IAFLL 2024 conference programme. Session chaired by Prof Tim Grant. Wed 26 June: 11:30-1200

P#66 Changes in Linguistic Transphobia on Mumsnet over time, 2008-2023. Eden Palmer (Aston University, UK)

https://www.aston.ac.uk/sites/default/files/IAFLL2024_Programme%20%2826.03.24%29_4.pdf

Aston's FoLD really should consider livestreaming that, and other sessions, to MN.

tbf tha pic I shrd wuz frm me linkdin profyl - tha vid yoo got wuz me n me matez out un a transfurbic rampge.

One of the things that makes writing a PhD thesis 'stressful' or at least something that requires effort and concentration is the knowledge that every single thing you write is going to be parsed and analysed and challenged. There is NO scope at all for 'normal colloquialism'. If you're describing getting dressed for field work and you assert that socks should be put on before shoes, you need to define socks and define shoes and either cite an authority or make a very strong tested scientific case for your claim.

So a PhD student who uses the term 'transphobia' would need to have thought long and hard about what that was and how it was defined. A student who uses the term 'hate speech' would need to have a definition of that too. And if they're claiming an overlap between those two things, guess what they'd need to do.

And if they were then going to link those things to thousands of people (who have no right of reply in this, who do not know they are involved) then they really really need to have definitions and justifications for that.

If a student doesn't know that, is not aware of the basic, elementary requirement of writing at that level either for a thesis OR for a public presentation (and the public presentation is arguably more sensitive than the thesis, then yes, I do question whether they are up to working at that level.

I know exactly what it's like.

Including knowing that there is always a moment when the student realises that they are now in a place where they know more than their supervisor does, and needs to be able to demonstrate that and act accordingly. That is a marker of original research.

P#66 Changes in Linguistic Transphobia on Mumsnet over time, 2008-2023. Eden Palmer (Aston University, UK)

So the starting premise is that there is linguistic transphobia on Mumsnet - the PhD research is merely into changes in it over time.

The starting premise is that there is such a thing as linguistic transphobia, that it can be defined and demonstrated. And that every word and phrase used to demonstrate it can be skewered precisely to that definition and not to any other. I think it would also have to be demonstrated that the definition stands the test of time - that the same words would have been linguistic transphobia if they'd been used in 1974, and if not, why not.

I think it is saying that Mumsnet had previously been used by Koschate at Exeter, to successfully distinguish between "feminist" and parent "identity" (I have questions about that but never mind). It's not saying that Liu used Mumsnet at Lancaster.

"ASIA (Automated Social Identity Assessment toolkit) (Koschate et al., 2021), a toolkit which leverages machine learning and natural language processing to automatically assess which identity is situationally salient through sociolinguistic styles, has been proven to be successful in assessing feminist and parent identity in Reddit and Mumsnet online communities."

So, I have had a quick poke on Google Scholar for ASIA Koschate 2021 and found 2 papers
https://link.springer.com/article/10.3758/s13428-020-01511-3
https://ore.exeter.ac.uk/repository/handle/10871/124775

The first one says that MumsNet gave permission as followsL
Proof of concept: Ethics
For our proof-of-concept case, in which we aim to detect parent and feminist identity salience, we chose identities that are widely held and not highly stigmatized. We collected the datasets with permission of the platform owners (Mumsnet UK, Netmums UK) or where permission for research use is granted by the terms and conditions (Reddit). All three platforms explicitly inform users that any content created is in the public domain and rights are owned by the platform rather than the user. Furthermore, forums on all these platforms are clearly signposted as being in the public domain rather than a place for private conversations, and therefore do not fall under the principle of “reasonable expectation of privacy”. For instance, Netmums UK calls their forum “Coffeehouse” to indicate its public nature. All of these platforms allow private messaging between users, thereby highlighting the distinction between public and private channels. No private messages are included in any of our datasets. All five studies presented here received ethical approval from the University of Exeter psychology ethics committee.

Later on they say:
The online forum data for training our model were gathered from the online website Mumsnet UK (www.mumsnet.com/talk), the largest parent online network in the UK, with the kind permission of Mumsnet UK. This site provides different sub-forums in which users can discuss particular topics and themes. We analyzed posts from two sub-forums, “Being a Parent” and “Feminism”. The posts were collected in September 2012 from 2500 threads per sub-forum.

I wonder, did they get permission from Mumsnet to look so specifically at feminism and the Feminism board?

So a key difference is "widely used and not highly stigmatized" presumbly meaning both parenting and feminism. So not throwing round accusations of extremism or transphobia.

Edited to correct quote.

That sounds like quite a challenge to me. Much of what trans activists refer to as "transphobia" rests on the idea of the dogwhistle - namely that the language you use may appear innocent, but in reality is a sinister code signalling bigotry. Phrases such as "biological female", "sex matters", "adult human female" and "safeguarding" are all dogwhistles. I'm not an expert on linguistics, but I imagine this would present an obstacle for the researcher wanting to demonstrate a rise in the use of transphobic language on Mumsnet.

PhDs are challenging. Or should be. They are really not the place for playground name-calling.

Update from Justine on site stuff thread.

Update - Aston Uni have responded and offered a call with their Vice Chancellor to explain the reasons for the research, how they manage ethical approval and protect privacy and data. I'll be taking them up on that and putting some of our own (and your questions). Will report back!

https://www.mumsnet.com/talk/site_stuff/5057903-mumsnet-corpus?reply=134751159&

@AmaryllisNightAndDay Thanks for doing the reading - my post was a bit rushed as I had to go into a meeting. But they did indeed get permission.

Two things though. Unless they have had some significant help from MN - which I doubt - they have also confused username with 'distinct user ID'. I am not sure that it matters for their study, but nonetheless, shows again how they don't understand what they are studying.

This conclusion though...
The overall pattern suggests that a feminist identity (positive coefficients) is expressed through a more intellectual style (e.g., use of long words (sixltrs), articles, semicolons, words related to causality and insights) with more negative connotations (e.g., negating words, negative emotions, swear words) than the parent identity.

We are smart and we are sweary. But I think we knew that already.

Edited because we are sweary not sweaty

It's great that there has been a response, but it sounds like they are expecting to just "explain" what they have done/are doing and then carry on?

Okay thanks 🙏

It's the way that it struck me, tbh.

We shall see. It's MN's data. At least some of it seems to have been collected against MN's terms and conditions.

Let's see if the same way that OU's VC initially adopted one stance in re: the Phoenix tribunal verdict and shortly thereafter adopted another, proves true again.

No courage of their convictions once facing the possibility of having some audience members present whose data has been collected and ‘analysed’. This leaves the data Guinea pigs in the dark about the collection, use and dissemination of their data. Adding yet another layer of unethical behaviour.

will the AI now assume that @VitoCorleoneOfMNMafia and @IwantToRetire are actually the same person

Can humans, especially wombens out smart AI which is inevitably sucking up male patriarchal misogynistic incel thought patterns?

We could have a regular user name swamp system which could be randomly achieved so AI would never know if it was a direct swamp of in fact a swamp generated by a mathematical calculation.

Speak for yourself!

I'm smart sweary and sweaty!

Well, I hope I'm available to see what arrogant, patronising and specious guff they come out with, to explain why they felt justified in doing what they did.

Interesting use of language there.

'Learning the ropes of their discipline' by publicly declaring a specific identifiable group of people to be guilty of a crime, specifically transphobic hate crime, seems pretty confrontational. Ranty even.

Is the suggestion here that posters on Mumsnet should just ignore that accusation?

So they did get the money for their dubious approach to ethics then.

They will also be reluctant to delete the Mumsnet Corpus data for the same reason