Semiotic Bridges and Toxic Transductions


This photo essay tracks toxicity through multiple scales, from the molecular to the global, and the multiple technoscientific investments necessary to visualize its workings.  The entry point is what are called "pattern recognition receptors" (PRR), molecular structures on the surface of cells that are a key mechanism in the complex signal transduction pathways through which toxins exert their (not always deterministic) effects. These "semiotic bridges" (a term used by biosemioticians, biologist/philosophers who think through Bateson [with his interest in "patterns that connect"], C.S. Peirce, Thomas Sebeok, and others) mediate between an outside and an inside--transduce, we can also say, an environmental toxin (such as particulate matter in air pollution) into some range of cellular responses that, cumulatively, become emergent health effects in "a body" and "a population"-- useful but misleading abstractions (i.e. yet another pharmakon in these tangled chains of pharmakons) for a sexed, gendered, classed, raced, geocoded, et cetera-ed person.

The  images in this essay represent different kinds of visualizations that scientists routinely produce or engage with as part of their work, the often publicly-funded work of "basic research" into toxicological mechanisms and effects.  The essay reads images like these ethnographically for what they can signal to us about the more local worlds that produced them, and the more global worlds in which they try to productively intervene.

View essay


Molecular imaging of a toll-like receptor (TLR). Every lung cell is studded with tens of thousands of receptors that form what biosemioticians call "semiotic bridges" -- molecular assemblages that transduce signals from an exterior environment (top) across the cell membrane (the thick and mostly colorless middle) to a semi-fluid cytoplasmic interior (bottom) crowded with diverse complex molecules that get mobilized into "signalling pathways." (These are not pictured here but some show up in the next image in the photo essay.) One common end result of such complex signal transduction is inflammation, or the constriction of lung passageways that are a symptom of respiratory conditions such as asthma. 

At this scale of analysis toxicity acts via molecular mechanism such as these.  Painting with a broad brush: an inhaled molecule binds, beacause of its particular pattern or shape, to the upper portion of the receptor with a complementary patterned shape.  The shape of the receptor thus changes, and this changed shape is sensed at the other end, inside the cell membrane, and triggers a complex cascade of cellular reactions that result in some kind of harm. 

The Toll-like Receptor or TLR is one of my favorite "matters of concern." These particular "pattern recognition molecules" (PRMs) are a relatively recent biomedical discovery, becoming visualizable and knowable only since the 1980s, a basic part of what we now call the innate immune system. On an evolutionary time scale, this is a very old set of molecular structures that we humans share with numerous organisms: fruitflies and fugu, mice and many other mammals.  TLRs have differentiated and multiplied over this evolutionary history; scientists are also interested in how they also differ slightly within species. Enormous investments of time, money, resources, energy, and affect are mobilized -- largely in the Euro-American and East Asian societies that can afford such investments -- to understand these receptors and their complex signalling pathways in the more detailed way they demand.

Some scholars find such "molecularization" of life, health, and toxicity to be reductive, inappropriately mechanistic, or otherwise deserving only critique or dismissal; I am looking to activate a different set of semiotic pathways in my viewers and readers.  I would like this image to transduce the intricate beauty of molecular structures and how that can capture the attention and interests of scientists; the commitments (vocations, for you Weberians) that those scientists embody in working out how variations in the molecule here may be associated with variations in different people's responses to inhaled pollutants; the importance of public investment in such "basic science" that will take years or maybe decades to "pay off," if it ever does; the collective effort to understand toxicity in its most minute enactments, and to stockpile and share data in public repositories like the Protein Database from which this image is taken; and the drives of curiousity, wonder, and for old-fashioned enlightenment that infrastructures a microscopic entity like this Toll-like receptor.

Design statementHyperrealistic visualizations of nanomolecular structures like cell surface receptors are a sign of both collective technical accomplishment, and invitation to sublime wonders. Mobilizing data from multiple expert communities accrued over years in expansive public databases, scientists work (perhaps to obsessive and excessive degrees) to understand the implications of difference at the molecular scale. Viewers are also interpellated into a scientific imaging tradition dating back to Robert Hooke’s Micrographia of 1665: asked to wonder, in amazement and curiosity, at the world contained within the world, the vital beautiful fragile structures of flesh.

TLRs and cars

Air pollution, TLRs, diabetes

This image from a recent article in the journal Diabetes represents what the authors call a "hypothetical framework" by which TLRs (my favorite semiotic bridge and pattern-recognizer) transduce air pollution into chronic disease conditions like diabetes and heart disease. 

The scientific persona of the "modest witness" that requires the authors to designate such an image as "hypothesized" is somehwat at odds with the genre of the scientific illustration and the air of definitiveness and literalness in its iconic mechanicity.  The figure in Figure 2 is clearly male, illustrating an ongoing problem in both journal illustrations and biomedical research itself, both of which normalize and naturalize maleness. Our dude here is seen eating a hot dog, sub, or cheese steak sandwich--these may not be differences that make a difference-- and the accompanying text highlights the synergistic role that "overnutrition" plays (hypothetically, I remind you) in exacerbating the connections between the air-pollution-pattern of an "environment" and the inflamed-diseased-organ-pattern inside our bodies.

I can't say for sure (I'm a modest witness, too) but I'll go out on a limb and say you would not have found an image like this, bringing cars and factories together with cytokines and livers within a single frame, in a major biomedical journal even just five years ago.  The isolated TLR in the previous slide is now almost lost in a complex translational shuffle. I chose this image -- not a useful visualization in the laboratory, but more a mechanism of visual communication -- because it depicts these new patterns of complexity, on multiple scales.  They are the collective result of new inter- and trans-disciplinary research efforts among immunologists, cardiologists, diabetes researchers, geneticists, biochemists, toxicologists, epidemiologists, and others.

A relevant part of the text reads:

Fig. 2 provides a hypothetical framework for these interactions and illustrates how inhalational stimuli may interact with overnutrition to entrain a state of chronic oxidative stress and inflammation...Teleologically, it is thought that pattern recognition receptors were meant to represent a crude but critical early-warning system to rapidly sense changes in lung microenvironment but also, equally importantly, to dissipate early to prevent unfettered inflammation. Thus, the notion that continual activation of these receptors may occur in a feed-forward manner and in concert with other stimuli without dissipation may be somewhat simplistic. However, it is also true that as humans, we did not evolve to be continually exposed to dietary and inhalational stimuli over the years, and such chronic exposure in vivo may have very different effects that we insufficiently understand. 

Design statement: This image conveys how the community of practice I study (environmental health scientists) themselves understand and visualize (“illustrate” might be better) their “object of concern” -- here, how “cardiometabloic diseases” is a multi-scale, emergent product of complex ecologies and systems.  It also directs attention to changing patterns in how objects like toxicity and their associated disease states are understood: toxic causes (factories, cars, eating habits) multiply into new patterns, imbricated with new patterns of in-body objects (organs, tissues, receptors, molecules).  Those interconnected patterns point in turn to another: scientists representing an increasing number of disciplines (pulmonologists, immunologists, geneticists, epidemiologists), each with their own technologies, styles, interests, and research traditions, coming into new patterns of collaboration, shaped by changing patterns of (largely) public research support.


Graphing PM2.5

This is an early diagram of "smog," produced from the air over Los Angeles in September 1969, the start of the contemporary era of air quality research, a time of increased data collection and new data visualizations. I want to emphasize the collective work and scientific attention required to turn imprecise, hazy "smog" -- a recognizable and rapidly worsening civic problem in places like Los Angeles in 1969 -- into something with measurable properties  that can be known in fine-grained detail (literally), and thus acted on -- as in, say, the 1970 Clean Air Act.

This graph comes from one of many scientific articles produced by a collaborative group of white men in white shirts and skinny black ties, some of whom worked mostly in Minnesota where they had developed the Minnesota Aerosol Analyzing System, developed for use in occupational health contexts such as granaries and bakeries.   At the center of this group is Kenneth Whitby, a guy you probably never heard of unless you've been awarded, or know someone who's been awarded, the Kenneth T. Whitby Award from the American Association for Aerosol Research. Whitby for me is an icon of all the undistinguished scientists and engineers who, in the last 50 years, have worked in the largely unrecognized labor of improving and inventing new scientific instrumentation for collecting air quality data, analyzing air's components in the specificities of time and place, and visualizing the data in new ways to more precisely characterize the slew of particles and toxins smogged there.

Design statement: This exemplifies what Hans-Joerg Rheinberger calls (in terms taken up from Derrida) the “graphematic space” that scientists work in, making (writing) “epistemic objects” such as PM2.5. An indistinct toxic “smog” is graphed/written as an epistemic object --something that is knowable and “graspable” as it comes to have fairly precisely specifiable characteristics (size, surface area, etc.), that get defined in response to the technical parameters and capacities of an experimental systems (instrumentation). definable properties.

Death rates from PM2.5 v exposure

This visualization of air pollution data is more recent, and shifts us to a different scale and kind of analysis.  Where the previous image focused on a particular air sample gathered at a particular place (Los Angeles) and a particular time (September 3, 1969), analyzing its components according to particle size and surface area, here the underlying data sets are more global, charting national death rates from PM2.5 (the y-axis) against mean annual PM2.5 exposure (the x-axis). The nations are also color coded by continent, and coded again according to GDP (the size of the circle).  The interactive visualization allows you to foreground different patterns in the data.

This particular data visualization signifies the potentials held by large public health data sets -- so dependent on data limitations, modeling parameters and assumptions, and other factors that simultaneously power and limit analysis -- less for finding answers, and more for their ability to generate new questions and prompt imagination and insight.  Clicking around on different parts of the graph and key shows some of these potentials: why are all the South American countries (in green) so tightly grouped? Why are Asian countries so splayed all over the graph? Is there a better explanation than "Money for cleanup!" for all those nations clustered down toward the graph's zero points?

And we can also see - or at least, see that there is a question to be posed - that high PM2.5 exposures do not necessarily correlate directly to increased death rates.  The many Gulf states that appear as large (wealthy) reddish circles out toward the right end of graph, representing the highest PM2.5 exposures, also have comparatively low death rates -- no worse than Iran, really, with its infamous air pollution, and much, much better than Iraq. (And Afghanistan is, literally, almost off the chart.)  Is it because PM2.5 in Saudi Arbaia, Kuwait, or Qatar is mostly cleaner fine sand particles rather than hydrocarbon-laden particles from cars and industrial facilities?  Or because they are wealthy enough to be managing the situation somehow, through better health care or otherwise?

Design statement: Data visualizations like this are used to produce and explore patterns in “Big Data.” They can (and should) be critiqued as limited, reductive, and otherwise subject to the vicissitudes of measurement, but this can also prevent anthropologists from reading for their productive potentials--not least as generators of new questions. Working with Gregory Bateson’s understanding that information is about “difference that makes a difference,” we can see how operationalizing the carefully characterized and organized differences that constitute large data sets--here, annual exposures to PM2.5 in different nations, their differential death rates, and their differential wealth -- can be used to produce new comparisons, hypotheses, and questions.

Death rates from PM2.5 2016 v 1990

Another interactive graph from the "Our World In Data" site. This one uses data from the Institute for Health Metrics and Evaluation (IHME) in a fairly simple plot: PM2.5-attributed death rates per 100,000 individuals in 2009 (the x-axis) versus those same death in 2016 (the y-axis), with nation-states again color coded by continent and sized according to GDP.  The bisecting line going up through the middle represents no change -- indeed, no progress: the same number of people dying from air pollution in 2016 as seven years previously.  It's thus easy to see, by virtue of being below that line, that the vast majority of nations have indeed made at least some progress. The further down from that mid-line a nation is, the more deaths have decreased there, and the more progress that has been made -- relative to this one metric, anyway.

In significant swaths of (medical) anthropology, the hegemonic attitude towards "data" and especially "Big Data" largely coheres around indifference, skepticism, or flat-out oppositional critique.  The very idea of "metrics," like those produced and analyzed by organizations like the Institute for Health Metrics and Evaluation (more or less equivalent, for some, to saying "the Gates Foundation"), seems to elicit strong reactive statements about the superior qualities of qualitative data and interpretive analysis.  Constructions of death rates like the ones depicted here, or of "DALYs" (Disability Adjusted Life-Years) are troped as (I will now exaggerate and italicize) wonky instruments of colonial control that eclipse or erase the subtleties and nuances so crucial to the quotidian lives that anthropologists alone can access and authorize.

This kind of critique -- and it's not unjustified -- turns on a cluster of notions, however, having to do with data, analysis, and science as tied almost exclusively (I will exaggerate again) to truthful representation -- a logocentrism, if you will, that is hegemonic in the sciences but shared as well, even if unacknowledged or cut somewhat by the apparent alterity of a "humanism," in anthropology.

But data and quantitative, computational analysis have other uses and modes, and the visualization here points to some of them -- indeed, pointing itself is one such valuable function.  Data visualizations like this one don't offer solid universal truths so much as re-direct the attention of scientists (including us), offer patterns to explore and ponder, and spark creative questioning.  Here, for example, this simple graph asks us to ask: how can we understand the differences between China and India? Two nations with similar (enough) GDPs, similar (enough) states of industrialization, similar (enough) headline-making "airpocalypses" in recent years, and similar (enough) death rates in 2009, as easily evidenced by their positions near the "150" marker on the x-axis -- yet something happened in China that made its data-blob move much further down the y-axis, below that midline?  China pushed -- somehow, on something -- and over the course of seven years lowered pretty dramatically -- more dramatically than India, at any rate -- the number of its citizens dying from lung-choking, heart-stopping, brain-eating air pollution.  The data and its visualizations may not tell us a truth, but they do tell us that something real is happening that makes a difference and one way to name that real difference falls under the rubric of governance...

Design statement: Data visualizations like this are used to produce and explore patterns in “Big Data.” They can (and should) be critiqued as limited, reductive, and otherwise subject to the vicissitudes of measurement, but this can also prevent anthropologists from reading for their productive potentials--not least as generators of new questions. Working with Gregory Bateson’s understanding that information is about “difference that makes a difference,” we can see how operationalizing the carefully characterized and organized differences that constitute large data sets--here, changes in death rates due to PM2.5 exposure in different nations, with their differential wealth, at two different points in time -- can be used to produce new comparisons, hypotheses, and questions. It also conveys the difference that programmatic social action -- in this case, air quality regulation and remediation, informed by data and irs knowledges -- has made (or not) in different national governance systems.


Creative Commons Licence


Created date

November 13, 2018