COMPLEXITY

Lauren Klein on Data Feminism (Part 1): Surfacing Invisible Labor

Episode Notes

When British scientist and novelist C.P. Snow described the sciences and humanities as “two cultures” in 1959, it wasn’t a statement of what could or should be, but a lament over the sorry state of western society’s fractured intellectual life. Over sixty years later the costs of this fragmentation are even more pronounced and dangerous. But advances in computing now make it possible for historians and engineers to speak in one another’s languages, catalyzing novel insights in each other’s home domains. And doing so, the academics working at these intersections have illuminated hidden veins in history: the unsung influence and cultural significance of those who didn’t write the victors’ stories. Their lives and work come into focus when we view them with the aid of analytic tools, which change our understanding of the stories we’ve inherited and the shape of power in our institutions. One strain of the digital humanities called data feminism helps bring much-needed rigor to textual study at the same time it reintroduces something crucial to a deeper reconciliation of the disciplines: a human “who” and “how” to complement the “what” we have inherited as fact.

Welcome to COMPLEXITY, the official podcast of the Santa Fe Institute. I’m your host, Michael Garfield, and every other week we’ll bring you with us for far-ranging conversations with our worldwide network of rigorous researchers developing new frameworks to explain the deepest mysteries of the universe.

This week we talk to Emory University researcher Lauren Klein, co-author (with Catherine D'Ignazio) of the MIT Press volume Data Feminism. In Part 1 of a two-part conversation, we discuss how her work leverages the new toolkit of quantitative literary studies and transforms our understanding of historical dynamics — not just in the past, but those in action as we speak…

For Part 2 in two weeks, subscribe to Complexity wherever you listen to podcasts — and if you if you value our research and communication efforts, please rate and review us at Apple Podcasts and/or consider making a donation at santafe.edu/give.

You can find numerous other ways to engage with us — including job openings and open online courses — at santafe.edu/engage.

Thank you for listening!

Join our Facebook discussion group to meet like minds and talk about each episode.

Podcast theme music by Mitch Mignano.

Follow us on social media:
Twitter • YouTube • Facebook • Instagram • LinkedIn

 

Related Reading & Listening:

Data Feminism by Catherine D'Ignazio & Lauren Klein

“Dimensions of Scale: Invisible Labor, Editorial Work, and the Future of Quantitative Literary Studies” by Lauren Klein

Our Twitter thread on Lauren’s SFI Seminar (with video link)

Cognition all the way down by Michael Levin & Daniel Dennett

Complexity 34 - Better Scientific Modeling for Ecological & Social Justice

Complexity 42 - Carl Bergstrom and Jevin West on Calling Bullshit: The Art of Skepticism in a Data-Driven World

Complexity 45 - David Wolpert on the No Free Lunch Theorems and Why They Undermine The Scientific Method

Complexity 64 - Reconstructing Ancient Superhighways with Stefani Crabtree & Devin White

 

Mentions Include:

Ruha Benjamin, Joy Buolamwini, Julia Lefkowitz, Ted Underwood, Derrick Spires, David Wolpert, Farita Tasnim, Stefani Crabtree, Devin White, Donna Haraway, Carl Bergstrom, Joe Bak-Coleman, Michael Levin, Dan Dennett

Episode Transcription

Lightly-edited machine transcription by podscribe.ai + Aaron Leventman

 

Lauren Klein (0s): There are other ways to think about abstraction or what can be perceived at scale, then just the zooming out and viewing from a distance. We have all sorts of concepts that are macro or that describe macro phenomenon. So we have the idea of like the collective versus the individual. We have ideas about invisible abstract or collective labor, as opposed to the like hard, tangible labor that is rewarded by deposit in your baking account kind of thing. 


 

If we think more capaciously about what our methods can do, if we sort of think beyond this metaphor of distance scale, as we know, operates on multiple dimensions in the same way, our critical lenses can also take these multi-dimensional approaches. Then we might think of other things to do with these quantitative methods, again, that compliments that extend beyond just sort of zooming out and saying, okay, here's the aggregate view of what these books are saying 


 

Michael Garfield (1m 24s): When British scientists and novelist C.P. Snow describe the sciences and humanities as two cultures in 1959, it wasn't a statement of what could or should be, but a lament over the sorry state of Western societies fractured intellectual life. Over 60 years later, the costs of this fragmentation are even more pronounced and dangerous, but advances in computing now make it possible for historians and engineers to speak in one another's languages, catalyzing novel insights in each other's home domains. 


 

And in so doing the academics working at these intersections, have illuminated hidden veins in history, the unsung influence and cultural significance of those who didn't write the victor's stories. Their lives and work come into focus when we view them with the aid of analytic tools, which change our understanding of the stories we've inherited and the shape of power in our institutions. One strain of the digital humanities called Data Feminism helps bring much needed rigor to textual study. At the same time, it reintroduces something crucial to a deeper reconciliation of the disciplines, a human who, and how to compliment what we have inherited as fact. Welcome to Complexity, the official podcast of the Santa Fe Institute. 


 

I'm your host, Michael Garfield, and every other week, we'll bring you with us for far ranging conversations with our worldwide network of rigorous researchers, developing new frameworks to explain the deepest mysteries of the universe. This week, we talked to Emory University researcher, Lauren Klein, coauthor of the MIT press volume Data Feminism. In part one of our two-part conversation, we discuss how her work leverages the new toolkit of quantitative literary studies and transforms our understanding of historical dynamics, not just in the past, but those inaction as we speak. Subscribe to Complexity wherever you listen to podcasts for part two in two weeks. If you value our research and communication efforts, leave us a rating and review and or consider making a donation @ santafe.edu/give


 

You can find numerous other ways to engage with us, including job openings and open courses @ santafe.edu/engage. Lastly, we are about to release an epic work of the SFI Press gathering essays, reflections, panels, and podcast transcripts about the COVID 19 pandemic through the lens of complex systems science. Visit sfipress.org and sign up for their newsletter or follow us on social media for announcements when the complex alternative is published in a few weeks. 


 

Thank you for listening. Well, Lauren Klein, it's a pleasure to have you on complexity podcast, 


 

Lauren Klein (4m 21s): So happy to be here. 


 

Michael Garfield (4m 24s): So I like to start these conversations by inviting you to talk a little bit about your own background, your biography, what animates you in your research, how you came to the passions that you explore as a scholar and how you got mixed up in this, I guess. 


 

Lauren Klein (4m 44s): Yeah, that's a good question. So I like to say, and because I think my research brings together computational and critical approaches to data to data science and to data's history. And I really pride myself on a syncretic approach, one that isn't sort of an either or, but really is trying to bring things together and to be able to speak to scholars and honestly just nerds on the internet in both of these areas in ways that are allegeable to them and that are meaningful to them. 


 

And, you know, this came about rather kind of through a long path. I went to college and the rise of the first dot com era. So the internet was new email was new. Anyone can make a website and I just was captivated and in college taught myself a lot of stuff. I took a lot of CS classes when I graduated. I worked at a series of dot coms that each sort of imploded and kind of spectacular ways. 


 

And then I sort of had a reactionary response and it was like computers can't trust them. You can trust books. And so I went to grad school in English and actually pursued a pretty traditional English PhD. My area of focus was early American literature. My dissertation was quite literary, quite historicist. And just as I was finishing grad school, I heard about this emerging field called Digital Humanities, which we know what that is now, but I didn't know what it was then. 


 

And I was like, “Hey, this is the thing that I've been looking for.” I always thought I had to choose between technical work and literary analysis and here's this field which lets me do both. And I just threw myself into it. I should say I was also really lucky. I mean, it was also a matter of timing. Right now the job market is a lot different, a lot harder. And someone like me who was just feeling their way through at the time that I was beginning my professional career, like I wouldn't be hireable at this moment, but I was just really lucky. 


 

And so I got, you know, this first job at Georgia Tech, which again was a place where I think that's cretic approach was valued. It's an engineering school, but I was teaching in the humanities. But in order to teach humanities courses to engineers, you need to figure out a way to bridge those two disciplines. And I was ready and eager to do that. And now I'm at Emory and in some ways I'm doing sort of the inverse, trying to figure out how these technical and quantitative methods can be brought back to the humanities for students who either are interested in them or might not know yet that they're interested in them. 


 

So that's sort of been the shape of my career to date. Yeah. I don't know what else do you want to know? 


 

Michael Garfield (7m 36s): No, I mean, that's great. I was really pleased to see your work because even though at SFI, we take this real hard line about our particular strain of transdisciplinarity which is focused on rigorous quantitative research. There's really not a lot of qualitative stuff going on here, but nonetheless, as we'll get into in this call, it's rather tricky in many cases to coax out of a quantitative approach, certain, very obvious and relevant things about our lives as human beings. 


 

And so seeing you square the circle, I found really interesting and inspiring. And so I'm glad that we get to talk about this. Maybe the first place to dive in here is actually the book that you authored with Catherine D’Ignacio. You have an MIT press book, Data Feminismin these strong ideas series. And I'd love to hear you just lay out for folks unfamiliar with your work, what we're talking about when we talk about data feminism. 


 

And I think that that gives us a springboard into a little bit more of a granular examination of what you've been doing with this. 


 

Lauren Klein (8m 57s): Data feminism put really plainly, it's an approach to thinking about data and data science that's informed by feminism and by intersectional feminism in particular. And what we take from that is both this sort of fundamental belief that there should be equality in the world and that that equality has not yet been realized. And so part of the commitment of feminism is not just a sort of imagine the world you want to live in, but also to take steps to make that a reality. 


 

But I would say on a more sort of theoretical level, the book really epitomizes, like I said, the kinds of things that I love to do, which is take theories from one field, or it doesn't have to be serious, but in this case, it is, and show how they have really broad and relevant applicability to a field that might not have seen that yet. And so what we do is we take a lot of conceptual approaches, theoretical structures, examples of actions from feminism really broadly conceived and show how actually these ideas and principles have very direct relevance on data science and in particular to these problems of bias systems, discriminatory algorithms, missing data, all of these ideas, which actually when we started writing the book, I think not that many people were talking about except the people who they were impacting directly. And now actually thanks to the work of not just us, but a lot of amazing scholars in this field, Joy Buolamwini, Margaret Mitchell, Ruha Benjamin. And you're like, I could go on and on, but there's this real, I think cohort of scholar activists, who are trying to redress these harms. 


 

And so that's what we try to do in the book. 


 

Michael Garfield (10m 40s): So to that point, and I think it might, it seems like there's a shift that people need to make here, especially maybe if you're an American. I was just talking with Visiting Scholar, Julia Lefkovitz, yesterday on campus about her work on the difference between American and French journalism and how American journalistic integrity cleaves to this performance of objectivity. Whereas in France, a lot of really excellent, solid journalistic work is very open about the position of the author in a way that would make it look like a mirror opinion piece, quote unquote, to an American audience, and yet by a clued cluing, the author and the stance of the author and the contextual embedding of who's actually doing the research, be it journalistic or academic, you're actually leaving out really, really important information, readdressing the who and the how, and rather than just like pretending, like we're just offering these objective whats when we present claims about the world, seems like an important piece of this. 


 

Lauren Klein (11m 60s): That seems that's such a smart observation. And I appreciate the analogy to journalists. And then actually I should say my co-author Katherine of Data Feminism has a background in journalism. She was teaching data visualization in a journalism department for a lot of years before she went to MIT. But I think that really is what I would say is a fundamental tenant of feminism. And actually there's another helpful phrase. I'll get back to that in a minute. But you know, Ruha Benjamin, who I was just mentioning has this great book called Race After Technology and in it, she talks about imagined objectivity, which is the ideal version of objectivity that we all wish could exist. 


 


 

Of course as people who pride themselves on rigorous research, valid research, producing results that we can bear weight. We all wish that we could achieve some sort of objective stance or scope on any particular problem. But the reality is that no one can achieve that. There is no such thing as just like pure idealized objectivity. And the response to that, and this is where the feminism come in, the response to that is not to just like throw up your hands and say, we can't trust science, or we shouldn't do research or our results are invalid or everything is relative.

Like that's what the people coming to this argument in bad faith might say, but rather we need to embrace all of this contextual information, as you were just saying the who and the what and the why that surrounds our work so that it can inform our work. And the end result is this leads to better science because we're more aware of the position from which we launch your inquiries, whatever, the perspective that we bring to that particular project, what we choose to focus on, what we decide is not especially important or salient at that particular moment, the national political social position from which we launch our inquiries which may lead us to focus on certain issues and minimize other issues. 

All of these things inform our work. They don't invalidate our work, but actually if we can make them tractable or even just articulate them, they let us place our research in a broader frame. And ultimately this leads to a better sense of how far we can to what ends we can apply our results, how far they'll hold to what populations in what contexts. And then it also limits us from either over-applying them, misapplying them injecting potential harms that we didn't foresee, all of this, again, this is what folks like Sandra Harding, Donna Haraway, these sort of early feminists theorists of science and technology would say, this is feminist objectivity. 

This is objectivity plus because it's everything we know from research and it's everything that we bring in from the outside. 

Michael Garfield (14m 56s): So you led directly where I wanted to go with this, which is in some discussion about a piece that you wrote about dimensions of scale, invisible labor editorial work in the future of quantitative literary studies. So this piece really touches on something that I find comes up on the show a lot. I actually came up on the last show that we just released with Stefani Crabtree and Devin White talking about using agent-based modeling and GIS to try and reconstruct likely migration pathways into the peopling of what is now like the Australian continent. 

And they were talking about how, if you try to model a minimal effort, migratory route, then you're missing important details at the ground level of how people actually see and navigate their environment and like the way that they identify specific landmarks and natural features as sacred and like this whole human dimension, which as you quote in your piece “As Donna Haraway has observed the technology of distance, often obscures non-dominant perspectives.” 

And this seems like something that people get stuck on a lot, because I think complexity science in particular has this legacy or this reputation of being about trying to find a single unifying theory of everything like a totalizing framework in which we can make all these universal claims, but it strikes me that what it's really more about, or how it has matured is about being able to fluidly navigate through scales to be as comfortable in the quote unquote micro and to allow it to exist on its own terms, with respect to the macro, rather than seeing like higher levels of organization emerging out of lower levels. 

And like all of these aren't really effective. So, I'm just gonna rant on this a little bit more and then toss it back to you and allow you to unpack this some with your work. But like you give this beautiful example of how Ted Underwood in Distant Horizons and describing his interest in exploring the sweep of long timelines. Analogizes the insights prompted by this perspective to how the curve of the horizon only becomes visible some distance above the earth. But when you look at work like that, work that zooms out as far as possible and finds these regularities that there are always exceptions, like Jevin West's work on the scaling laws and how human beings are an outlier in this like clean graph that we make of like all of the mammals and how much energy we require. 

And like, meanwhile, humans are actually using like orders of magnitude more energy than you would expect for a mammal of our size. So it's like, there's something in escapable as far as like our human existence is concerned that we not simply kinda like dissociate to this orbital view. And so you make a really solid case in this for moving beyond, not just the access that I described, not just this close and distant reading, but like bringing them together, remixing, adding additional axes. 
 

I would love to have you go into some detail about how exactly you're doing this in this particular piece of research on Mary Shadd and Lydia Marie Child and their work as editors. And then from there, I think we can kind of open it up into the discussion of what the study of invisible labor implies or how that changes the way that we're thinking about things more broadly. Okay. That was a mouth full. 

Lauren Klein (18m 48s): No, no, no. I appreciate that framing beause I think that once again, it's kind of the point, like it's always a “both and.” We can learn a lot about culture about society writ large when we do take that sort of zooming back view and really try to grasp the whole as we can see it in its abstract form. But I think that there's a little bit of an inside baseball argument that's embedded in that essay that I don't know how legible it will be to non-literary scholars or scholars who have spent a long time in this field of digital humanities. 

So maybe I'll just frame it a little bit. So the most famous proponent of these computational approaches to literature, which is where my research practice, a lot of where my research practice sits, this Franco Moretti and he coined this phrase distant reading. And it's very evocative. You hear it. And you almost immediately understand as opposed to close reading. What I will do is instead of reading the book, I will zoom out and try to aggregate or abstract so that I can see large patterns in this literature when it's treated as data. 

And the fruits have been born out in work by Ted Underwood and works by Richard Soe, there have been really amazing and interesting claims that have been made and findings that have been uncovered that tell us new things about literature, but the sort of the inside baseball argument that I make there, there are other ways to think about abstraction or what can be perceived at scale than just the zooming out and viewing from a distance. 

We have all sorts of concepts that are macro or that describe macro phenomenon. So we have the idea of like the collective versus the individual. We have the idea of what is when I talk about in the paper ideas about invisible abstract or collective labor, as opposed to the like hard, tangible labor that is rewarded by deposit in your banking account kind of thing. Conceptually, what I was trying to do in that paper, aside from the individual example was say, if we think more capacious about what our methods can do, if we sort of think beyond this metaphor of distance, and we think more in terms of scale, as we know, operates on multiple dimensions in the same way, our critical lenses can also take these multidimensional approaches. 

Then we might think of other things to do with these quantitative methods that again, that compliments that extend beyond just sort of zooming out and saying, okay, here's the aggregate view of what these books are saying? And I believe that with all of my core, I think that, and I will also say like, I don't have the best ideas about what to do next. I think that we have thought fairly narrowly about what computational approaches to texts can do in large part because of this conceptual framing. And I'm really eager to see what the next generation of scholars think of when they allow themselves to bring in additional conceptual models, additional ideas from other disciplines. 
 

I think we'll just start to just learn so much more. Do you want me to get into a little bit about the details of the paper at this point? Where do you want to go? 

Michael Garfield (22m 9s): Yes, definitely. So specifically like surfacing invisible editorial work through topic modeling.

Lauren Klein (22m 18s): So the paper I said a while ago, So my area of expertise is American literature. I focus in particular on the early Republic, but really up until the civil war is like in my wheelhouse. And I'm really an increasingly interested in the relationships among activists within the abolitionist movement. And I'm interested in abolition and folks listening will start to very soon make the connections to the present, but this was this historical, immensely important social movement.


It led to the abolition of slavery in the United States. Arguably it's the most important social movement maybe ever, but it was accomplished by a multiracial coalition of people, including black activists and white activists. It was multi-gender. There were leading women abolitionists at a time when most professional spaces were dominated by white men. So it's this really interesting mix of perspectives coming together. And it's also super interesting to me because, and this is true almost of any social movement ever, everyone agreed about the end goal.

Everyone agreed that slavery is reprehensible. It needs to be abolished, but within the movement and some people even disagree that it could be called a single movement there was intense disagreement about how to accomplish that goal. And so you can think even today about the various proponents within the U.S. government is right, do we try to become more moderate so that we can bring more broad support to some core set of concerns or, since the issue of human freedom is on the line, is there no room for a moderation.

And those are both very valid perspectives that people can have on any issue. Do we just burn it to the ground because this is a flawed democracy or do we try to rebuild it? These are debates that we hear with respect to black lives matter right now, with respect to conversations about defunding the police, even about the infrastructure bill. I don't know when this will air, but all of these ideas about how political change happens, you know, these are still, these are open issues, right? No one really knows there we're still working them out. 

And so that's just sort of a long way of saying that I'm really interested in how these dynamics play it out historically conveniently for me, many of these dynamics, not all. And we can talk about that later, but many of them are documented in these abolitionists newspapers, which were in effect the social media of their day. They were published often weekly, sometimes monthly, but mostly weekly in all sorts of regions, they were circulated and shared. They were editors read each other's work. They often reprinted editorials, or even news items that they agreed with as a sign of endorsement. 

And so it's this really networks textual record that describes a lot of what this movement was up to. And this is ineffective. It's big textual data and that this is exactly the kind of archive as I would call it that I think can open itself up to really interesting new findings when we approach it at scale. And so one of the interesting things that I was looking at in the paper had to do with the editorial work that was invested by different editors.

Because you can read these newspapers, you can read them now, you can go to an archive. You could, they actually, some of them not actually a very small percentage of them have actually been preserved, but you can read them if you want. Many of them have been digitized. You can recreate the process of what it was like to read them at their own time. You can also read countless scholarly accounts of what happened both broadly with respect to the abolitionist movement and in particular, with respect to these newspapers. And yet there are also things that we do not know or are difficult to perceive about these newspapers using extent, critical skills. 

And one of them that really struck me in part because of my own work as an editor, was this idea of what were the editors doing? We know that every week they sat down with these clippings from other newspapers, with their ideas in their own head, with their pen and paper, with their type, because often at this very, very small teams of people who were doing everything from writing, editing, selling, compositing, printing, lthese are small operations. I mean, this was so much work that we were doing and we can see the end result in these newspapers, but there's so much that's missing there about the intellectual labor and other forms of labor, I'll get there in a minute, that had to be invested in order to produce this weekly output. And I was just wondering like, what else could we say about that? How else could we honor that work? Is there a way to not recover the thinking because you can't really do that.  But I was just sort of wondering what else could we do to make this editorial work tractable and legible and sort of make it matter in new ways. And part of this, I would say, is inspired by, I did a lot of archival work at the New York Public Library for this project, reading letters, by these editors where they complained about how hard it was. 

They're like I spent so much time fixing the typos and these other people's newspapers. And then this other newspaper wants me to shove in this anecdote. And like, I wish I could be writing editorials, but I don't have the time to do it. Lydia Maria Child, one of the authors who I, the editors rather who I write about has this letter where she says, “I have three editorials written in my mind, but I haven't had time to write them down.” And so there's like the editor saying, I was thinking of these things. I just couldn't do them. They're not on the page. And so I was just really interested, like where could quantitative approaches come in here? 

And where this took me was to topic modeling, combined with some measures from information theory, but trying to say, okay, like topic modeling is really good at summarizing content. It gives us this like slightly larger, larger view. You're not at the level of the individual article or the level of the sentence, or even the page. It tells us what were the themes that were being discussed in these papers. And you can aggregate them. You could rank them, you can sort them. And what I did in the paper was try to say, okay, in addition to topic, there are some measures from information theory that allow us to identify not only the most dominant topic, like what topic constituted the majority of this newspaper, which P.S., by the way, in all of them is like abolition.

There's a topic that has like slaved freedom, servitude, liberty. Like you look at that and you're like, okay, this is a topic about slavery. And that of course is the number one topic and all of them, but what is the most uniquely significant topic for each newspaper and for each newspaper editor. So if we had to try to identify a theme that was sort of the mark or the hand of this editors intervention, what would that topic be? And in some cases, in my analysis, we find things that confirm what we know. 

So Lydia Maria Child, who I've already mentioned, she was a really prolific author writer, editor, abolitionist. She was brought on as editor in order to attract more women to the cause. She had previously published a cookbook. And so we can see some of the topics that are distinctive about her editorial tenure do have to do with clothing, with dress, with food, with recipes, sort of these domestic issues. And again, we have on the record, William Lloyd Garrison and famous abolitionists saying, I want her because I think she could bring in the women. 

And later on, she goes on about how she tried to do that. And she actually got pushback from some people who thought she was being too moderate. So that confirms known findings. That's great. It's proof that the model holds weight. But then one of the really interesting things that turned up was that I was looking at this newspaper called the Provincial Freeman. This was a newspaper edited by a black woman, Mary Ann Shadd. She married later, became Mary Ann Shadd Cary. And at the time she also wrote about her travails and editing this newspaper, felt that as a woman and as a black woman, she was unduly criticized for her activist positions. 

She felt like, and not felt like she actually did. There was going on the record, people saying maybe you should tone it down a little bit, try to get more readers before you push them away with your radical politics. And she was like, I'm not here for this. Again, human freedom is on the line and this is what it takes. And so interestingly, when you look at the topics that characterize her newspaper, you don't just get these domestic topics, the same as the white woman editor, although that's there. But you also get topics that have to do with like travel with the European politics, with the natural landscape. 

It's just this really capacious view of what freedom would entail once achieved. And this, again, you know, this was really interesting to see, because it tracks some recent scholarship. I'm thinking of folks like Derrek Spires and this amazing work of the Colored Conventions Project who have done a lot of work to say, look, you know, when white people envision emancipation, they envision legal emancipation. For them slavery, the opposite of slavery is, I declare you free, but for the black people who were enslaved, it's not just legal emancipation, it's ability to thrive, it's ability to access to opportunity. 
 

You know, all of the things that living means. And, sure, this comes with legal emancipation, but it also means economic justice, educational justice, social justice, ability to practice your religion, all of these things. And so I saw this in this newspaper and that felt really meaningful to me in large part because Shadd herself had been so vilified really, you know, people were not coming to her for her radical politics for her attempt at radical world making or remaking. 

And then here it was in these topics that was long and rambly. But anyway, I think that's like, that's the heart of this. 

Michael Garfield (32m 18s): Well, first of all, long and rambly is how we do things on the show. You're right at home. But a couple of things come up for me in reading this and in listening to you talk about it and you know, it's my onus to link out to other strains of complex systems research. And so just riffing on how I see rhymes in your work and rhymes and the work of other people. One is this issue of changes and information architecture, and how it's widely analogized. 

Like you just did, that social media and pamphlets have this relationship. And when we're looking at major changes in power structures in society, the work comes in from the margins, which is related to that a lot of really important quote unquote sense-making is being done in these largely uncredentialed forms of communication because the society just hasn't caught up yet to finding ways to encode expertise in these new media. 

And this is related to work that Albert Kao and Mirta Galesic and former External Professor, Carl Bergstrom, and a bunch of other people did with Joe Bak-Coleman at the University of Washington recently on the paper about stewardship of global collective behavior and how they're really arguing that quantitative social science about the impact of social media and other new digital forms is a crisis discipline now. It's so important that we figure this out and then you didn't actually come out and say this per se, but like the term that you alluded to, as an information theory concept is appoint wise mutual information quantifying the degree association between a specific feature and a category. 

This reminds me a lot about using genomics to trace horizontal info transfers through networks. So like just like the relatively recent scientific discovery that rabbits in the laboratory are actually absorbing genes through their gut bacteria and into themselves through the food that they're eating. 

Speaker 1 (34m 34s): I believe that. That's so nutty. 

Michael Garfield (34m 37s): Hypothetically, we're doing this too. And so this kind of thinking spills into the sense of the self as nested networks embedded within networks. And I see my strong rhyme between the work that you're doing in this particular inquiry and this sort of broader paradigmatic shift that is brought about by complex systems research where intersectional feminism and the non-human, or post-human turn. Donna Haraway, you already mentioned, Katherine Hayles and these other people are involved in. 

So it's like in a way there's something about like seeing the human being as a holobiont, you know, in which your gut bacteria are playing an important part in the regulation of your brain activity and this kind of thing seems very much akin to the patterns that you're describing and society. And the way that all of this historically invisible labor is actually contributing to these major transitions in the way that we relate to one another such that the story is I, the self-authoring modern agents, this bound self and making a decision, and now we've moved beyond that. 

And we have say, actually, I am deeply vulnerable to and influenced by all of these other agents that until very recently in ecology and evolutionary biology are getting no credit whatsoever. So you've got like, you know, Michael Levin and Dan Dennett talking about cognition going all the way down. And in both cases we're seeing quote unquote the tail wagging the dog. 

Lauren Klein (36m 16s): I appreciate that observation. I think you're right to identify this belief that I think is consistent across all of my research that ideas happen, change happens, culture forms from the outside in, and that we learn more about a culture by trying to identify and examine and explore the work on the periphery, or that's been placed on the periphery by these dominant power structures rather. It wasn't actually on the periphery at the time at all, like was actually the dominant culture, but that my approach in general, and this is true in my eyes are purely humanistic work, as opposed to, in addition to some of my more technical work that I think that's where the interesting ideas are, that's where the complexity lies. 

Like that's what makes people people, and there's a lot of different residences with different approaches to feminism feminists on, I mean, you mentioned a little bit earlier black feminism, and you could easily trace this or thinking like this to Patricia Hill Collins and this idea of the four different domains she had described. She's formulating this in relationship to oppression. She has this idea called the matrix of domination, but she sort of atomizes the LT Syrian power structure and a little bit more specificity and places, a lot of emphasis on the hegemonic domain, which is system and of culture, as well as the interpersonal domain, which is how in her case oppression affects people. 

But this is true in general about how cultural forces impact people and shape a culture. She also has really interesting ideas about what she calls subjugated knowledge on this idea that we have cultural authorities and outlets that authorize certain types of people to tell certain types of stories and certain registers and anyone who doesn't fit into that particular place gets subjugated. They're forced out of the newspaper of record. 

They're forced out of the formal mode of speaking, and therefore it becomes this vicious cycle where their knowledge becomes subjugated or deprecated, but obviously it's not knowledge. In fact, this is where most of the thinking takes place and it's sort of our work. And I believe this especially as someone who looks back on the past and I would say also like as a white scholar, who is trying to account for the ways in which white scholars have written history and have silenced or minimized the contributions of nonwhite people to all of these social movements that are so central to the country's founding, like it's our job to amplify those and to sort of bring those into the center in the ways that we know how to do it, whether it's through the particular techniques that I apply or really anyone else's work. 

Michael Garfield (39m 8s): So because your child has interrupted this call, and I think that like before we started, we were talking about the effects of the pandemic on exposing the invisible labor of parenting. And again, a link out to Aaron Clauset at SFI and his co-author Alison Morgan at CU Boulder and the work that they've done on just like how antagonistic the academic system is to parents, the lack of understanding, support funding, the incentive structure of this publish or perish thing, where if somebody has a kid, they fall behind, and so this is an enormous problem when we consider that the majority of scholars are parents above the age of 40. 

I would love to hear you talk a little bit about this in terms of like the consequences of making labor visible and of coming to a more kind of complete and human understanding. And then maybe not to put you on the spot with this, but it does seem again to rhyme with work that David Wolpert has been doing on thermodynamics speed limits. So total curve ball here, but like his idea in stochastic thermodynamics, there's this notion of the Landauer Bound, which is sort of like the minimum amount of energy required to process information. 

And we just sort of assume that that is that's the floor. That's the speed limit that we can reach with computer systems, but we'll, it's been decomposing systems into their subsystems. And one of his coauthors for Farita Tasnim at MIT just came and spoke at SFI about this only to it in the show notes about how, when you actually look at the individual contributions of those subsystems in a circuit that the speed limit is actually much higher because the whole thing is only as fast as the slowest components. 

And so, in a way, the computer science conversation seems to be reaching kind of similar conclusions to the data feminism conversation about the invisible contributions of a circuit within a larger circuit, and how actually we have these ideas about what the economy is capable of. But it's based on this nonsense about like we've assumed a spherical cow, we've assumed academics that don't have families, we've assumed a warehouse workers that don't need to take bathroom breaks. 

And so I would love to just invite you to speculate a little bit about what happens when we make the shift and in what ways that sort of like intention with our need for speed as a society and like the efficiencies of economies of scale and so on. 

Lauren Klein (42m 10s): There's a lot there. First of all, I'm glad to hear that you think that computer science has heated this cautionary tale, because I feel like from my perspective, they're still modeling cows as spheres and compartmentalizing and reducing the complexity of our lives, but maybe that's the beacon of hope. I mean, I think, again, this comes back to what I was saying before about the value of multiple perspectives and the value of not just human perspectives, but the value of approaching things from different timeframes, from different scales. 

I think no one would argue that for example, scholarship that takes more time is often more thought through, a little bit deeper, more nuanced in its language, more precise than us argumentation. We can't always take forever on everything to your point about academic time skills and their rigidity, but we learn, there is such a thing as like just-in-time research, especially like I'm thinking, you know, of Joe Gulity, for example, who I know is a SFI Affiliate, who's done really interesting work arguing for the need for really fast scholarship and response to climate change because we are living in a climate emergency. 

We don't have the time to wordsmith and perfect everything. And we just like, get it out there and make change now. But there's also additional, again, not competitive with, but like complimentary scholarship that happens over a long time span when you have time to synthesize, to sit with, to sort of meditate on what the larger impact could be. And we need an academy which makes space for all of this. Another place where this comes up a lot is discussions of participatory design and just community oriented research where academic timescales are such that you often don't have time to build a really good relationship with the community partner before like your grant runs out or you need to get tenure or your RA graduates, or, whatever the constraints make the courses over, whatever the constraints may be. 

And yet we know that human relationships and meaningful relationships take time to build, and there's not really space in academic incentive or reward structure is too, I would say an academic reward structures to incentivize this type of slow relationship building work. And so the scholars who I think make meaningful relationships with communities and therefore do the more impactful work, usually that comes at the expense of merit increase at the end of the year or their promotion or whatever. So that's another example of how now we really do need more flexibility, but again, I may seem that sound like a broken record, but I think this is true, just because certain research fast and certain or research is slow, or certain processes are fast and certain processes are slow it doesn't mean that you can't have people doing both kinds of things, or even the same person doing both kinds of things at the same time. And we absolutely do need an academy that recognizes that there's not only one way to have impact. There's not only one way to produce transformative research. There's all these different ways that we can do it. And that like all are necessary, right. And all are valuable. 

Michael Garfield (45m 28s): So to that point, it seems like a good opportunity to pivot into this other piece that you coauthored with Sandeep Soni and Jacob Eisenstein, abolitionists networks, modeling language change in 19th century activist newspapers. 

 

Thank you for listening. Complexities produced by the Santa Fe Institute, a nonprofit hub for complex systems science located in the high desert of New Mexico. For more information, including transcripts research links and educational resources, or to support our science and communication efforts, visit Santafe.edu/podcast.