Theoretical framework

Media Ecology & Technology Trust

How trust in technology changed as machines disappeared into infrastructure, read through media ecology and American Studies.

Technological authority grows more pervasive even as its mechanisms grow harder to observe. I read the history of technology trust through two traditions. Media Ecology asks how the form of a medium organizes what people perceive, say, and question. American Studies asks how the nation has turned technology into a source of shared meaning, and what happens to trust when that meaning thins.

The media ecology lens

The Toronto School tradition treats media as environments rather than neutral channels. Three thinkers supply the core of my framework.

Harold Innis

Monopolies of knowledge

Every medium carries a bias toward time or space, and control over the means of communication concentrates power. A monopoly of knowledge grows until

“equilibrium is disturbed”

Innis, 1951/2008, p. 4

I read artificial intelligence as a contemporary monopoly of knowledge: citizens are asked to trust systems they have no means to inspect.

Marshall McLuhan

Form over content

McLuhan warned against dividing form from content, insisting that

“the medium is the message”

McLuhan, 1964, p. 7

His rear-view mirror explains why societies judge new technologies through older ones, the basis of my contrast between technostalgic and techno-skeptical memes.

Neil Postman

Technopoly

Postman described a culture that

“seeks its authorization in technology, finds its satisfactions in technology, and takes its orders from technology”

Postman, 1992, p. 71

In a technopoly the cultural immune system that once questioned new tools has broken down.

“Societies have always been shaped more by the nature of the media by which humans communicate than by the content of the communication”

McLuhan, as quoted in Jarvis, 2023, p. 138
A short history of technology trust

My genealogy traces how technological authority lost the visible, narratable form earlier citizens engaged directly.

  1. 1904 to 1905 · Max Weber

    The iron cage

    Rationalization replaced traditional and charismatic authority with calculable procedure that outlasted its original purpose.

    “fate decreed that the cloak should become an iron cage”

    Weber, 1904-1905/1958, p. 181
  2. 1934 · Lewis Mumford

    The clock as key machine

    Technology reorganizes consciousness itself, abstracting time from bodily and seasonal rhythm.

    “The clock, not the steam-engine, is the key-machine of the modern industrial age”

    Mumford, 1934, p. 14
  3. 1985 · Neil Postman

    Huxley, not Orwell

    Algorithmic governance resembles Huxley's dystopia more than Orwell's: convenience and personalization, not coercion.

    “Orwell feared that what we hate will ruin us. Huxley feared that what we love will ruin us”

    Postman, 1985, p. viii
  4. 1991 · Mark Weiser

    Disappearance into infrastructure

    Ubiquitous computing names the endpoint: technology that governs everyday life while dropping out of view.

    “the most profound technologies are those that disappear”

    Weiser, 1991, p. 94
  5. 2015 to 2024 · Pasquale, Zuboff, Schaake

    Algorithmic governance

    Black-box decision systems, surveillance capitalism, and what Schaake calls

Literature review

Memes & Technology Trust

Why a meme is a medium worth testing: genre, rhetoric, cool media, and hidden transcripts.

Internet memes are more than ephemeral humor. As cool, coded, participatory media, they make otherwise invisible technological systems available for public judgment. This section summarizes the scholarship that let me treat a meme as a medium worth testing.

From replicator to genre
Richard Dawkins, 1976

The original meme

Dawkins coined the word for

“a unit of cultural transmission, or a unit of imitation”

Dawkins, 1976, p. 192

His meme is a faithful copy; its fidelity to an original is what matters.

Limor Shifman, 2013

The internet meme as genre

Shifman redefined internet memes as

“groups of content items that were created with awareness of each other and share common characteristics”

Shifman, 2013, p. 367

A meme is a family of variations, read through its history of reuse.

Participatory culture scholars complete the shift. Jenkins, Ford, and Green replace contagion with spreadability: people choose to share, alter, and recontextualize. Knobel and Lankshear show that making and reading memes is a literacy learned informally, and Nissenbaum and Shifman show that this literacy is unevenly distributed, a marker of belonging.

How memes argue

Wiggins reads the meme as an enthymeme, an argument that leaves a premise for its audience to supply. Finnegan's account of visual argument, which Wiggins builds on, describes a form that

“assumes an audience of judges capable of filling in the blanks”

Finnegan, 2001, p. 143

Image and caption anchor one another, template history shapes every new instance, and irony means the same template carries opposite stances. Reading a meme correctly depends on recognizing which stance is in play.

Memes as a cool medium

McLuhan's distinction between hot and cool media gives the formal argument. A hot medium delivers a complete message in high definition; a cool medium withholds and asks the audience to complete it.

“A cartoon is ‘low definition,’ simply because very little visual information is provided”

McLuhan, 1964, p. 22

At every level a meme withholds completion: the image needs the caption, the caption needs the template's history, and the template needs the community's conventions. Read against the Gutenberg Parenthesis and Ong's secondary orality, memes look less like a novelty than an older, communal, variable mode of communication returning after five centuries of print. A cool medium suits an opaque object: a meme gestures toward a system it never fully describes, and the viewer supplies the rest.

Hidden transcripts in public view

James C. Scott distinguished the public transcript, the open interaction between subordinates and those who dominate, from the hidden transcript:

“discourse that takes place ‘offstage,’ beyond direct observation by powerholders”

Scott, 1990, p. 4

Scott names three conditions for hidden transcripts: a marked asymmetry of power, pervasive observation, and real costs to open dissent. Each has a counterpart on contemporary platforms. Yet a meme criticizing a technology company circulates on infrastructure that company operates, so its hiddenness is interpretive rather than spatial. I therefore describe technology memes as publicly visible but culturally coded hidden transcripts: critique carried by irony, intertextuality, and template convention that institutions find hard to recognize, aggregate, or act on. For S

PhD Dissertation Research Project

Hidden Transcripts in the Algorithmic Age

01 · The problem and the plan

Trusting systems we never see

Digital technologies increasingly shape what people see, know, and decide, yet the systems producing those effects remain largely invisible. I asked how citizens evaluate technologies they never directly observe, and whether internet memes shape public trust in those technologies.

Do internet memes measurably shape reported trust in digital technology, and through which visual, rhetorical, and interpretive processes?

Research questions and hypotheses

RQ1 · H1 · confirmatory

Critical memes versus neutral material

Does exposure to technology-critical memes change reported trust relative to neutral material?

H1: technology-critical memes lower reported trust relative to neutral images.

RQ2 · H2 · confirmatory

Technostalgic versus techno-skeptical

Do technostalgic memes, which invoke favorable comparisons with earlier technologies, differ from directly techno-skeptical memes in their trust effects?

H2: technostalgic memes lower trust more than techno-skeptical memes.

RQ3 · H3 · exploratory

Meme literacy

Does meme literacy moderate the relationship between exposure and trust change?

H3: higher meme literacy strengthens the effect.

H4 · exploratory

Interpretive closure

Interpretive closure, the viewer's sense of having understood the meme's point, carries the effect.

H4 has no separate research question; it tests one process named in the overarching question.

The research plan

I used a four-phase sequential explanatory mixed-methods design. The outcome throughout is reported trust in digital technology, measured on four subscales: reliability, benevolence, integrity, and predictability. Each phase discovered something that decided what the next phase did.

  1. Phase 1, collect and code. Build a corpus of technology memes and code each one on seven dimensions.
  2. Phase 2, experiment. Turn the coded trust frames into experimental arms and test them on reported trust.
  3. Phase 3, focused survey. Return to selected participants with the memes that behaved unexpectedly.
  4. Phase 4, interviews. Talk with readers whose interpretations diverged.
  5. Framework. Turn the distinctions the phases produced into the Grammar of Trust.

Back to top ↑

02 · Theoretical frameworks

Four pillars, two instruments

I treated memes as everyday commentary on systems people use daily yet seldom see. Four pillars, fixed before any data collection, shaped what I measured and how I read the results.

Pillar 1

Hidden transcripts

James C. Scott's distinction between the public transcript and the critique spoken offstage. I read memes as publicly visible but culturally coded commentary, critique in full view that irony keeps deniable.

Pillar 2

Media ecology

The Toronto School tradition of Innis, McLuhan, and Postman. Form shapes meaning: memes are cool, participatory media, and McLuhan's rear-view mirror explains why new technologies get judged through older ones.

Pillar 3

Trust theory

Mayer, Davis, and Schoorman's ability, benevolence, and integrity model, translated into four digital subscales with predictability added from research on trust in automation.

Pillar 4

Civil religion

Bellah's civil religion and Nye's technological sublime: technology as a source of American legitimation, and what happens to trust when that faith thins.

From theory to instruments

The pillars became two instruments. The first is a seven-dimension coding schema, with each dimension derived from theory before I coded any meme.

DimensionCodesTheory source
Trust frameTN, TS, AMMedia ecology: McLuhan's rear-view mirror; technostalgia
Content focus10 domains, e.g. privacy, AITrust research: Mayer, Davis, and Schoorman
Format7 templates, e.g. image macroThe structure of the meme
Tone7 tones, e.g. satirical, ironicHidden transcripts: Scott
Intertextual reference6 domains, e.g. meme historyParticipatory meaning-making: McLuhan, Wiggins, Shifman
Nostalgic resonancehigh, medium, lowMedia ecology; technostalgia
Literacy thresholdhigh, medium, lowParticipatory meaning-making

The second is the Digital Technology Trust Scale: twelve items on a seven-point agreement scale, three for each subscale. Reliability and predictability items refer to digital technologies; benevolence and integrity items refer to technology companies and developers. Internal consistency was high before and after exposure (Cronbach's α = .918 and .925, N = 276).

Reliability

Does digital technology work as intended?

Benevolence

Do technology companies and developers act in good faith?

Integrity

Do they act with sound ethics?

Predictability

Is the behavior of digital technology predictable?

Back to top ↑

03 · Methodology

Four phases, each deciding the next

A sequential explanatory design runs the quantitative work first and then explains it with qualitative work. I built mine as a chain: each phase discovered something that decided what the next phase did.

  1. 1
    Corpus content analysis

    Phase 1 · Collect and code

    I collected 121 candidate technology memes from Reddit, X, Instagram, and Facebook, kept 54 that met the inclusion criteria, and coded each on the seven dimensions. I recoded 24 of them 14 days later to check my own consistency.

    Trust-frame codes built the three experimental arms.

  2. 2
    Randomized pre and post design, N = 276

    Phase 2 · Experiment

    U.S. adults recruited through Prolific completed the trust scale, then saw 8 memes drawn at random from one arm's pool of 12: technostalgia, techno-skepticism, or neutral control images. They rated each meme, completed a short filler task, and took the trust scale again.

    The four memes furthest from their pool's average, plus trust-change thirds, selected Phase 3's memes and respondents.

  3. 3
    93 invited, 65 complete

    Phase 3 · Focused survey

    Respondents described six memes in their own words, rated their confidence in each reading and their likelihood of sharing it, and named an audience for the four outliers. A second coder, blind to my codes and to participant age, classified every reading.

    Sincere and ironic readers of one meme selected the interviewees.

  4. 4
    15 of 16 completed

    Phase 4 · Recorded interviews

    I moderated recorded interviews with 8 sincere and 7 ironic readers, using mirrored interview tracks, and analyzed the transcripts through directed coding plus reflexive thematic analysis.

    The distinctions the interviews drew built the Grammar of Trust.

Phase 2 in detail

ArmStimuliPrimary n
Technostalgia12 technostalgic memes91
Techno-skepticism12 of 31 techno-skeptical memes, 5 carrying nostalgia92
Neutral control12 images normed from 18 candidates93

For every meme viewed, participants rated its trustworthiness, humor, likelihood of sharing, and whether they understood its point. The trustworthiness rating tested my coding at the level of the single meme. After the post-test, a condition check asked which description best matched the images they had seen.

Analysis. ANCOVA with pre-test trust as covariate and planned contrasts for H1 and H2; a multiple-membership model that treats the memes as a sample; a within-arm test of the skeptical memes carrying nostalgia; arm by literacy interactions for H3; and bootstrapped mediation through closure for H4.

Ethics: approved IRB protocol; consent through Prolific and on the recording; recordings deleted after transcription.

Back to top ↑

04 · Memes

What the corpus looked like

The final corpus held 54 technology memes. Critique of the present dominated, and the corpus leaned toward satire and indirection rather than direct statement, which fits the hidden-transcript reading of memes as coded critique.

31techno-skeptical memes
12technostalgic memes
11ambivalent memes, not fielded
7 in 10+memes carrying satire

Four memes that carried the argument

I name each meme by type first, with its number in parentheses. These four did the most to show where the coding held and where it did not.

A skeptical meme with inconsistent tone (M42)

M42, the Expanding Brain meme

Read a book, then an audiobook, then an AI summary, then an AI summary of the summary, beside ever brighter brains.

Techno-skeptical · AI and machine learning · multi-panel · satirical, ironic · internet and meme history · nostalgia low · literacy medium

Viewers rated it the most trust-suggesting of all 36 stimuli (4.30). In Phase 3, 30 of 62 read it as sincere praise of AI and 21 as ironic criticism. My own two coding passes had disagreed on its tone.

View the meme on Imgflip

A clear skeptical meme (M26)

M26, the “Always Has Been” astronaut meme

Two astronauts above Earth: it's all about surveillance; always has been.

Techno-skeptical · privacy and surveillance · object labeling · satirical, ironic · internet and meme history · nostalgia low · literacy medium

A typical member of the skeptical pool, which viewers rated lowest (2.65) and recognized most reliably (89.4%). About 57% to 60% of Phase 3 responses named its surveillance message: the coding held.

View the meme on Imgflip

A nostalgic meme turned critical (M25)

M25, the Buff Doge vs. Cheems meme

A paper road map and remembered competence set against phone navigation.

Technostalgic · labor · image macro · nostalgic, satirical · internet and meme history · nostalgia high · literacy medium

Rated 2.57, well below its pool's 3.38, and heard in both interview groups as frustration with dependence on phones rather than warmth for the past.

View the meme on Imgflip

A nostalgic meme received favorably (M22)

M22, the Drake meme

Music on a CD versus streaming.

Technostalgic · obsolete technology · image macro · nostalgic, satirical · internet and meme history · nostalgia medium · literacy low

Rated 4.20 in Phase 2. Respondents tended to name a younger audience for it, whatever their own reading: a favorable reception and a nostalgic frame did not line up neatly.

View the meme on Imgflip

Codes are the final Phase 1 codes (Appendix E of the dissertation). Meme images link to their Imgflip sources (Appendix N).

Back to top ↑

05 · Findings, part 1

Quantitative results

The findings come in two parts. Phases 1 and 2 tested whether memes moved reported trust; Phases 3 and 4, in the next section, explained the results the numbers alone left open.

Supported

H1: critical memes lowered trust

Adjusted difference −0.144, d = −0.355, p = .006, for the 36 memes fielded.

Reversed

H2: skepticism outweighed nostalgia

Technostalgic memes lowered trust less than skeptical memes: d = 0.341, p = .022, opposite to the prediction.

Not supported

H3: meme literacy

Arm by literacy interactions p = .480 and .605. The corpus held no high-threshold memes, so the test covers a limited range.

Not supported

H4: interpretive closure

Indirect effect −0.0003, 95% CI [−0.014, 0.013]. Closure records that a viewer reached a reading, not which reading.

H1 critical vs. controlH2 technostalgia vs. skepticism
−0.4−0.200.20.4Adjusted difference in post-test trust, 7-point scalePrimary ANCOVA, N = 276H1 d = −0.355, p = .006 · H2 d = 0.341, p = .022H1, Primary ANCOVA, N = 276: -0.144 [-0.247, -0.042]H2, Primary ANCOVA, N = 276: +0.139 [+0.020, +0.257]Condition-verified, N = 147H1 d = −0.053, p = .791 · H2 d = 0.393, p = .079H1, Condition-verified, N = 147: -0.022 [-0.188, +0.143]H2, Condition-verified, N = 147: +0.166 [-0.020, +0.352]Memes treated as a sampleboth intervals include 0H1, Memes treated as a sample: -0.146 [-0.383, +0.091]H2, Memes treated as a sample: +0.140 [-0.134, +0.413]Bayesian, memes as a samplePr(direction) = .90 and .87
05 · Findings, part 2

Qualitative results

Phase 3 asked 62 finished respondents what the outlier memes meant. The skeptical meme with inconsistent tone split its readers almost in half.

Response 1: sincereResponse 2: sincereResponse 3: sincereResponse 4: sincereResponse 5: sincereResponse 6: sincereResponse 7: sincereResponse 8: sincereResponse 9: sincereResponse 10: sincereResponse 11: sincereResponse 12: sincereResponse 13: sincereResponse 14: sincereResponse 15: sincereResponse 16: sincereResponse 17: sincereResponse 18: sincereResponse 19: sincereResponse 20: sincereResponse 21: sincereResponse 22: sincereResponse 23: sincereResponse 24: sincereResponse 25: sincereResponse 26: sincereResponse 27: sincereResponse 28: sincereResponse 29: sincereResponse 30: sincereResponse 31: ironicResponse 32: ironicResponse 33: ironicResponse 34: ironicResponse 35: ironicResponse 36: ironicResponse 37: ironicResponse 38: ironicResponse 39: ironicResponse 40: ironicResponse 41: ironicResponse 42: ironicResponse 43: ironicResponse 44: ironicResponse 45: ironicResponse 46: ironicResponse 47: ironicResponse 48: ironicResponse 49: ironicResponse 50: ironicResponse 51: ironicResponse 52: confusedResponse 53: confusedResponse 54: confusedResponse 55: confusedResponse 56: confusedResponse 57: confusedResponse 58: confusedResponse 59: confusedResponse 60: meta-awareResponse 61: meta-awareResponse 62: meta-awareSincere 30praise of AI capability
06 · Framework proposal

The Grammar of Trust

The findings did not fit the model I proposed before fielding. The trust effect was real for the memes I fielded, but the closure measure missed how it worked, and one meme showed why: viewers decoded it in opposite ways and took different stances toward the same reading. I therefore propose the Grammar of Trust, which keeps my seven-dimension coding schema whole as its first moment and replaces interpretive closure with separate decoding and stance-taking.

measuredobserved in Phases 3 and 4proposed

The arrows organize a research agenda; they do not certify a completed causal chain. The terms encoding and decoding come from Stuart Hall (1980); the Grammar of Trust applies them to vernacular memes and ties them to reported trust.

How the model evolved

Original modelChangeGrammar of Trust
Artifact levelKeptEncoding: the seven coded dimensions
DecodingSplitDecoding: propositional and affective
Interpretive closureReplacedStance-taking: endorse, reject, or distance
Trust judgmentRefinedTrust revision by subscale, plus a candidate extension

One meme through the four moments

MomentThe Expanding Brain meme (M42)
EncodingI coded it techno-skeptical, with satirical and ironic tone. My own two passes disagreed on its tone.
DecodingThirty respondents took it to praise AI efficiency; 21 took it to criticize outsourcing thought; 8 gave no committed reading; 3 named the ambiguity.
Stance-takingInterviewees described the opposite reading without adopting it, and 3 of 15 revised their reading under different prompts.
Trust revisionViewers rated it the most trust-suggesting of the 36 stimuli (4.30).

Coded public discourse

The findings also suggest a category between Scott's public and hidden transcripts. Coded public discourse circulates in full view, including the view of the party it criticizes, while permitting its audience to extract the criticism, to decline to extract it, or to extract it without endorsing it. The interviews show the separation of recognizing a message from endorsing it; they do not show resistance, sanction, or institutional response.

Future research

I recommend the Grammar of Trust, with the coding schema as its encoding layer, as the framework for future research on how people interpret technology memes and how those readings bear on trust. Five studies would test it.

Study 1

Nostalgia by framing

A preregistered factorial design crossing critical framing with nostalgic affect, using memes normed and matched in advance and a model of participants and memes together.

Settles: Whether nostalgia softens criticism, or the H2 reversal came from the particular memes fielded.

Study 2

The four moments, measured separately

Separate items for what a meme says, confidence in the reading, endorsement, feeling, and trust, across several inconsistent-tone memes.

Settles: Whether decoding and stance separate beyond one meme; a replacement for the closure item.

Study 3

Literacy at full range

High-threshold memes from imageboard culture and a broader, validated literacy index.

Settles: A fair test of H3.

Study 4

Age and lived experience

Two targeted age groups, or a much larger age-stratified sample, with nostalgic memes about technologies from a clear era.

Settles: Whether nostalgia works through lived experience, and whether younger viewers lean toward ironic readings.

Study 5

Time and context

A simulated feed, repeated exposure, and later follow-up; validation of the trust scale; new institutional-accountability items.

Settles: Whether the effect outlasts one instructed viewing, and whether accountability is a fifth domain of trust.

The workflow transfers: code a documented corpus on the seven dimensions, assign memes within arms by trust frame while sampling from pools, measure trust by subscale before and after exposure, and return to viewers' own readings where the categories fall short.

Back to top ↑

07 · References

Selected references

Selected references from the dissertation, in APA 7 format. The full list appears in the dissertation.

Bellah, R. N. (1967). Civil religion in America. Daedalus, 96(1), 1-21.

Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101. https://doi.org/10.1191/1478088706qp063oa

Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health, 11(4), 589-597. https://doi.org/10.1080/2159676X.2019.1628806

Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research (3rd ed.). Sage.

Hall, S. (1980). Encoding/decoding. In S. Hall, D. Hobson, A. Lowe, & P. Willis (Eds.), Culture, media, language: Working papers in cultural studies, 1972-79 (pp. 128-138). Hutchinson.

Hoff, K. A., & Bashir, M. (2015). Trust in automation: Integrating empirical evidence on factors that influence trust. Human Factors, 57(3), 407-434. https://doi.org/10.1177/0018720814547570

Innis, H. A. (1950). Empire and communications. Clarendon Press.

Knobel, M., & Lankshear, C. (2007). Online memes, affinities, and cultural production. In C. Lankshear & M. Knobel (Eds.), A new literacies sampler (pp. 199-227). Peter Lang.

Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159-174. https://doi.org/10.2307/2529310

Lee, J. D., & See, K. A. (2004). Trust in automation: Designing for appropriate reliance. Human Factors, 46(1), 50-80. https://doi.org/10.1518/hfes.46.1.50.30392

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709-734. https://doi.org/10.5465/amr.1995.9508080335

McLuhan, M. (1964). Understanding media: The extensions of man. McGraw-Hill.

Milner, R. M. (2016). The world made meme: Public conversations and participatory media. MIT Press.

Nye, D. E. (1994). American technological sublime. MIT Press.

Postman, N. (1992). Technopoly: The surrender of culture to technology. Alfred A. Knopf.

Scott, J. C. (1990). Domination and the arts of resistance: Hidden transcripts. Yale University Press.

Shifman, L. (2014). Memes in digital culture. MIT Press.

van der Heijden, T. (2015). Technostalgia of the present: From technologies of memory to a memory of technologies. NECSUS: European Journal of Media Studies, 4(2), 103-121. https://doi.org/10.5117/NECSUS2015.2.HEIJ

Wiggins, B. E. (2019). The discursive power of memes in digital culture: Ideology, semiotics, and intertextuality. Routledge.

Back to top ↑


Contact