SOCI 325: Sociology of Science

Agenda

Public trust, participation,
& implicit values

  1. Technological determinism
    vs. anti-essentialism
  2. Public consumption of
    science

Technological determinism

vs.

anti-
essentialism

Anti-essentialism

Interpretive flexibility

  • A major theme from Herzig’s article on the use of X-rays for hair removal was interpretive flexibility
  • In this view, technologies do not have any essential features
  • The meaning of a technology is determined by its use in society

Animation of Chief Wiggum from The Simpsons sitting at the a dining room table with his son. He is taking walnuts from a bowl and using the butt of his pistol to crack them open. Caption: 'Son, whether you want to win a girl or crack a nut, the key is persistence.'

“No technology – and in fact no object – has only one potential use. Even something as apparently purposeful as a watch can be simultaneously constructed to tell time, to be attractive, to make profits, to refer to a well- known style of clock, to make a statement about its wearer, etc. Even the apparently simple goal of telling time might be seen a multitude of different goals: within a day one might use a watch to keep on schedule, to find out how long a bicycle ride took, to regulate the cooking of a pastry, to notice when the sun set, and so on. Given this diversity, there is no essence to a watch. And if the watch has no essence, then we can say that it has systematic effects only within a specific human environment.”

Sismondo (2009, p. 98)

The case for determinism

Technological determinism

  • In “Do artifacts have politics?” (1980), Langdon winner makes the case that technologies have specific and inherent influences on systems of power

Weak version:

  • Technologies are employed by humans and institutions to resolve socio-political disputes

Strong version:

  • Technologies embody socio-political arrangements as essential features

The case for determinism

Technical arrangements as forms of order

  • The “weak” form of technological determinism in Winner (1980)
  • Use of technology in a deliberate way to intervene in a social or political system
  • E.g. Long Island overpasses
  • Clear examples, but not a direct argument for essentialism

photograph of a cars on a highway going over a very low overpass
photograph of a emergency workers examining a damaged bus at night. The top of the bus appears to have been scraped partially off.

The case for determinism

Inherently political technologies

“Strongly compatible”

  • Certain technologies are “strongly compatible” with particular sociopolitical arrangements.
  • E.g. solar and wind power are inherently decentralizing, and therefore more conducive to democratic or consensus control
photograph of many houses, all of which have solar panels on top their roofs

photograph of many houses, all of which have solar panels on top their roofs

Inherently political

  • Other technologies are fully deterministic in their political
    implications
  • E.g. The inherent dangers and long-term effects of nuclear power necessitate centralized regulation and enforcement

Example: Do LLMs have politics?

Large language models’ role in sociopolitical crises:

  • Political misinformation
  • Environment and climate change
  • Effects on artists and writers
  • Deepfakes and non-concentual images
  • Worker exploitation,
  • Dangerous health advice

Are LLMs:

  • A. Neutral technology,
    not conducive to any particular sociopolitical arrangements
  • B. Strongly compatible
    with certain sociopolitical arrangemets
  • C. Inherently political,
    necessitating specific sociopolitical arrangements

1. Consider

  • Different forms of LLMs
    E.g. chatbots (ChatGPT, Grok); assistants (Apple Intelligence); image generators (Midjourney)’ automatic summaries (Google, Amazon); state–actor tools (military, disinformation); automatic news articles and web pages
  • Aspects of LLMs
    Development
    Production and training
    Access and deployment
    Consumption and meaning

2. Make a claim

  • Write down which of the three options (A, B, or C) you think best describes LLMs
Screenshot of a facebook bot named 'FungiFriend' and described as an 'AI by Northeast Mustroom Identification ...' giving advice on how to cook a poisonous mushroom named Sarcosphaera coronaria

Technological frames

Most STS Scholars:

(why not both meme) Photo of a young child shrugging. Top text: 'Technological determinism or anti-essentialism'. Bottom text: 'Why not both?'

Technological frames

Technological artifacts exist in a web of social, cultural, and material conditions

Technologies may have inherent properties that facilitate certain uses and meanings, but interpretive flexibility can alter that usage in virtually every case

The political and social ramifications of technology are constrained by its essential features and by its established uses and meaning

Sociotechnical ensembles (Wiebe E. Bijker)

A crowd of people sitting in spaced chairs on the astroturf of an indoor stadium, waiting after receiving a vaccine.

Public consumption of science

Public consumption of science

Shelves and shelves of vitamins and supplements

Non-scientists and technoscience

Public (“lay people”) do not just naively use technoscientific output, but have their own understanding of it

E.g. diet and nutrition, algorithmic interactions, public health

What is the relationship between professional scientists’ and non-scientists’ understanding of scientific knowledge?

Two common “models” aim to explain asymmetries:
(both critiqued in STS and in class readings)

  • Dominant model
  • Deficit model

Dominant model

Overview of the ‘dominant model

Science produces reliable, true knowledge when scientists communicate with one another

Science communicators (e.g. journalists) work to translate this knowledge to make it understandable by non-scientists in the public

This translation is viewed as having a necessarily distorting effect on findings

Implicit in the dominant model is a assumption that scientists do not consume popular accounts of scientific output.

The dominant model is argued to be widely believed among professional scientists

Dominant model

Some critiques of the dominant model

  • Scientists directly read popular discussions of research (e.g.
    COVID, tech news, politics)
    ⦙ Scientific research is guided by the questions that people care about (e.g. through funding)
photograph of many houses, all of which have solar panels on top their roofs

photograph of many houses, all of which have solar panels on top their roofs

All scientific communication is translation

  • Clean disconnect between ‘translated’ and ‘untranslated’ knowledge is not realistic
  • E.g. fixation of evidence, actor–network theory

Deficit mondel

Overview of the ‘deficit model

Focus on the degree to which actors have the technical skill/training to understand scientific knowledge

Under this model, widespread misunderstandings (e.g. “vaccines cause autism”) are explained by a lack of scientific literacy

Therefore, important individual and political decision are made with a lack of understanding (e.g. political discussions of reproductive health)

The solution to misunderstandings is better education, better communication

The deficit model is not contradictory to dominant model, but has a different focus

Public consumption of science

Some critiques of the deficit model

Empirical studies often find that publics have important
knowledge that scientists do not

Much scientific knowledge has built-in assumptions
about the way the social world works, and public can
have good grounds for disagreeing with scientists on that
front

Assumption that scientific knowledge does not depend
on historical, social, or political context

Two panels with stills from The Simpsons. Top panel, Homer Simpson poining sternly at a crude line drawing of a futuristic looking car. Bottom panel, Homer Simpson sitting and grinning in the realized version of that car, which looks absurd.

Image credit

Animation of Chief Wiggum from The Simpsons sitting at the a dining room table with his son. He is taking walnuts from a bowl and using the butt of his pistol to crack them open. Caption: 'Son, whether you want to win a girl or crack a nut, the key is persistence.'

Animation from The Simpsons via Giphy

photograph of a cars on a highway going over a very low overpass >

Image from Google Street View via The Washington Post

photograph of a emergency workers examining a damaged bus at night. The top of the bus appears to have been scraped partially off.

Still from video by Amber Ferguson via The Washington Post

photograph of many houses, all of which have solar panels on top their roofs >

Photo from World Economic Forum

photograph of many houses, all of which have solar panels on top their roofs >

Photo by Dan Meyers on Unsplash

Screenshot of a facebook bot named 'FungiFriend' and described as an 'AI by Northeast Mustroom Identification ...' giving advice on how to cook a poisonous mushroom named Sarcosphaera coronaria

Screenshot via 404 Media

(why not both meme) Photo of a young child shrugging. Top text: 'Technological determinism or anti-essentialism'. Bottom text: 'Why not both?'

From imgflip

A crowd of people sitting in spaced chairs on the astroturf of an indoor stadium, waiting after receiving a vaccine.

Video still via News 5 Cleveland

Two panels with stills from The Simpsons. Top panel, Homer Simpson poining sternly at a crude line drawing of a futuristic looking car. Bottom panel, Homer Simpson sitting and grinning in the realized version of that car, which looks absurd.

Stills from The Simpsons

<video autoplay loop> <source src="img/wiggum_nuts.mp4" type='video/mp4'> </video>

Robert Moses, 1920s - 1970s 2018: bus full of over 40 high school students returning from the airport hit one of these Gov. Cuomo said he would put sensors in to prevent it in the future Also: interstates in chicago Also: automation and unions

Note: strongly compatible with X does not mean incompatible with Y

AI bot added to Mushroom ID facebook group by Meta

“VEE-buh BAIK-er” https://forvo.com/word/wiebe_bijker/

implicit critiques over the next couple of weeks - TallBear (2013), Genomic articulations of indigeneity - Allen (2018), Strongly Participatory Science and Knowledge Justice in an Environmentally Contested Region - Harambam and Aupers (2015), Contesting epistemic authority: Conspiracy theories on the boundaries of science

- Scientists may use popular discourse as an alternative route - E.g. Dawkins' _Selfish Gene_, which steered popular and scientific understanding of evolution and genetics

E.g. Todd Akin in 2012 senate race in Missouri: "From what I understand from doctors, [pregancy resulting from rape is] really rare. If it’s a legitimate rape, the female body has ways to try to shut that whole thing down."

e.g. cumbrian sheep farmers, or ACT UP - Assumption that scienctific knowledge is 'context free' - Same for dominant model - tie back to discussions of objectivity