Prepared 1 September 2026 · Eric D. Miller · Talk About Martin
Subject of review: a social-media post widely attributed in local discussion to Barbara “Babs” Clowdus.
Bottom line
The post is not a literature review. It poisons the well by declaring every prior counter-argument “propaganda,” “memes,” and “smears”; treats an election as if it settled a physics question; collapses three different acoustic problems into the single word “infrasound”; cites three papers that do not say what the captions say they say; and never mentions the August 2026 finding that no published study has measured the infrasound / low-frequency spectrum of an operational data center. That last fact alone makes “the science is settled” a false statement.
Part I defines the method so the word “propagandist” is a claim about sentences, not a nickname. Part II checks the sources. Part III states what would actually close the file.
Part I — What a propagandist is
The short definition
A propagandist is not “a person I disagree with.” A propagandist is a person who treats evidence as costume for a pre-selected outcome, and treats the other side’s evidence as moral defect. Opinion is allowed. Advocacy is allowed. What is not allowed, if the word “facts” is going to be used, is building the verdict first and fetching the bibliography second.
What the word is not
If “propagandist” only means “the other side,” the word is useless. Everyone in a land-use fight is trying to persuade someone. Residents write flyers. Utilities hire consultants. Newspapers choose ledes. Calling all of that propaganda empties the term and makes the person using it look like the thing they are naming.
The word is also not a diagnosis of character, party, or motive in the soul. Motives are inferred, not weighed. What can be shown on the page is method: how a writer moves a reader from an open question to a closed verdict, and whether the evidence on the page can survive contact with the documents it cites.
Used correctly, “propagandist” is a claim about method. It has to be demonstrated the same way any other claim is demonstrated: by pointing at sentences.
The working test
A communication is propaganda — and its author is acting as a propagandist in that instance — when most of the following are present at once. One technique is ordinary advocacy. A stack of them, pointed at a pre-selected outcome, is the thing being named.
| Test | What it looks like on the page |
|---|---|
| Poison the well | The opponent is defined as dishonest, hysterical, or captured before any evidence is offered. After that frame is locked, a short reading list is presented as “facts.” |
| Election or power as gavel | A political outcome is treated as if it retired a measurement, a statute, or a physics question. “The election is over” is offered as the on-ramp to science. |
| Motte-and-bailey | A modest, defensible claim is parked in front of a sweeping claim the writer actually wants. When challenged, the writer retreats to the modest claim and pretends that was the whole argument. |
| Equivocation | One word is used for three different problems. Winning the easiest version is treated as winning the stack. |
| False equivalence | A heartbeat, a passing train, or city traffic is treated as the same exposure as a 24-hour industrial plant in a quiet rural night. Duty cycle, tonality, ambient, and duration are not named. |
| Appeal to rigor, then misstate the source | “These are studies, not magazine articles.” The papers are real. The captions are not. Abstracts that ask for more research are printed as no-effect verdicts. |
| Strategic omission | The one paper that is actually about the source being defended is left out. The file can only be “closed” if the gap is invisible. |
| Projection | The other side is accused of looping one study and smearing opponents. The post then loops a short list, over-claims it, and opens by smearing opponents. |
| Bait-and-switch subject | “Facts about data centers” turns out to be facts about the easiest acoustic sub-claim. Water, power, setbacks, codes, and governance are dropped on the floor. |
| Outcome-stable posture | Across months the talking points change — envied site, wait for a proposal, studies have spoken — but the preferred land-use outcome does not. |
Advocacy can use one or two of these without becoming propaganda. Consultants omit. Residents simplify. The line is crossed when the techniques are stacked, the omitted fact is the load-bearing fact, and the other side is framed as a moral problem rather than a party with a falsifiable claim.
The test has to travel. A neighbor who over-reads an animal study at 130 dB as proof of disease at a rural lot line is doing the same method in the other direction. Naming the method only when the other side uses it is also propaganda.
Mapping the circulating post onto the test
Poison the well. The first four sentences do no scientific work. They assign motive and moral status: propagandist, meme-factory, smear artist. Only after that frame is locked does the author offer “verifiable facts.”
Election as gavel. “Okay, the election is over. The tactics overturned the council, which was the objective, so now let’s get to some verifiable facts.” Sound pressure does not change because a council changed. A-weighting still discounts 63 Hz by about 26 dB whether or not anyone won a race.
Motte-and-bailey. The motte is that some short-to-medium controlled trials of pure inaudible infrasound in healthy adults have not shown large effects on standard cognitive batteries or questionnaire sleep scores. The bailey is that noise fears about data centers were implanted propaganda and the health question is closed. Those are not the same proposition.
Equivocation and false equivalence. “Infrasound” is used as if it were the entire acoustic complaint. Winning a lab argument about 6 Hz is treated as winning the stack. “Your body already makes infrasound” erases duty cycle, tonality, ambient, and duration.
Appeal to rigor, then misstate the source. “These are studies, not magazine articles.” Baliatsas 2016 found associations with annoyance, sleep problems, concentration difficulties, and headache, and asked for more research. The post captions it as “no health effects.” Dastan 2026 calls infrasound an under-studied environmental stressor. The post parks it under “NO significant effects.” Møller 1984 found annoyance and ear pressure. It is listed as if it were a long-term no-effect study.
Strategic omission. Wheeler et al., 7 August 2026, state that no published study has measured the ILFN spectrum of an operational data center. A post that claims to close “facts about data centers” and does not mention that paper is not a literature post.
Projection. The post accuses the other side of looping “the same study” and of using smears when argument failed. Then it loops a short list, over-claims it, and opens by smearing the other side.
Bait-and-switch. “Verifiable facts about data centers” becomes a discussion of the easiest acoustic sub-claim. Water, power, setbacks, dBA-only codes, and the FPL-as-landowner-utility-taxpayer structure are not retired by a 28-night 6 Hz trial.
Outcome-stable posture. In February 2026 the same author wrote that Silver Fox “will be envied,” was “NOT next to a residential area,” and sat “two miles from any residence in any direction.” In April, after withdrawal, TCPalm quoted her calling the loss “a potential loss” and saying people cannot judge until they see a proposal. After the election the pitch is that the fears were implanted. The talking points rotate. The preferred door does not.
Part II — The post and the papers
The post, as circulated
The text under review opens by describing opponents of data-center siting as authors of “anti-data center propaganda,” then of “slick memes,” “a torrent of publications, which often referred to the same study,” and finally of “character assassinations and smears.” It declares the election over, the council overturned, “which was the objective,” and then offers “verifiable facts about data centers.” The fact-section that follows is almost entirely about inaudible infrasound. It asserts that “long-term studies — these are studies, not magazine articles — have consistently shown NO significant effects of infrasound on human health, including cognitive performance, psychiatric symptoms, and sleep quality.” It equates data-center sound with traffic and with “our own bodies.” It lists three papers.
- Dastan et al., Applied Sciences 16(3):1553 (3 February 2026), “Infrasound and Human Health: Mechanisms, Effects, and Applications.” mdpi.com/2076-3417/16/3/1553
- Baliatsas, van Kamp, van Poll & Yzermans, Science of the Total Environment 557–558:163–169 (2016), “Health effects from low-frequency noise and infrasound in the general population: Is it time to listen?”
- Møller, Journal of Low Frequency Noise, Vibration and Active Control 3(1):1–17 (1984), “Physiological and Psychological Effects of Infrasound on Humans.”
A fourth study is clearly being gestured at in the “6 Hz, 80–90 dB” line even though it is not named: Ascone, Kling, Wieczorek, Koch & Kühn, Scientific Reports 11:3190 (2021), a 28-night bedroom trial of inaudible 6 Hz infrasound in 38 healthy adults.
Paper-by-paper autopsy
This section decides whether the author is summarizing science or recruiting science. Each row is a claim in the post, then the document the claim points to.
| Paper | How the post uses it | What the paper actually does |
|---|---|---|
| Dastan et al. 2026, Appl. Sci. “Infrasound and Human Health.” | Implied to support “NO significant effects on human health.” | A mechanisms review. Cellular work: PIEZO1 / TRPV4, calcium influx, oxidative stress, apoptosis. Animal work: neuroinflammation and memory impairment at prolonged or intense dose. Human work: short-term moderate exposures show minimal physiologic change; context shapes symptoms. Authors’ own close: infrasound is “an important environmental stressor that has received little attention”; “may be harmful at chronic exposure depending on intensity and frequency.” They do not issue a no-effect finding. |
| Baliatsas et al. 2016, Sci. Total Environ. systematic review. | Captioned as: “A systematic review identified no health effects from everyday low-frequency noise, including infrasound, in the general population.” | Seven studies. Associations reported with annoyance, sleep-related problems, concentration difficulties, and headache near LFN sources. Evidence on chronic diagnosed disease was “very limited.” Pooled high annoyance about 10 percent. Authors: everyday LFN “constitutes an issue that requires more research attention, particularly for people living in the vicinity of relevant sources.” The title of the paper is a question — “Is it time to listen?” — not a no. |
| Møller 1984. Sixteen subjects, three hours. | Listed under “these are the findings” in a no-effect frame. | Standout result in the paper itself: high annoyance ratings and a feeling of pressure on the ear, including near the threshold of perception. It is a 42-year-old acute lab study. It is not a long-term health study, and it is not a null result on annoyance. |
| Ascone et al. 2021 (the unnamed “6 Hz, 80–90 dB” trial). | “Long-term studies with inaudible infrasound (6 Hz, 80–90 dB) found no significant effects on cognitive performance or sleep quality.” | Closest item to the caption. N=38 healthy adults; devices in bedrooms 28 nights versus sham. Behavioral batteries mostly null. Exploratory: trend toward more physical weakness. MRI: regional grey-matter decreases in cerebellar VIIIa and left angular gyrus; authors would not call that proven harm because they could not link it to behavior. Twenty-eight nights in screened volunteers is not a decade next to a cooling yard in a quiet rural night. |
The Baliatsas sentence, word for word
This is the cleanest single misrepresentation in the post, because the paper is a systematic review and its abstract is short. The post says the review “identified no health effects from everyday low-frequency noise, including infrasound, in the general population.” The review’s own highlight bullets, as published, say:
“A part of the population reports high annoyance attributed to LFN sources. LFN is associated with self-reported outcomes, mainly neurological. Current evidence is very limited, especially regarding chronic conditions. More epidemiological research on LFN and health effects is needed.”
“No health effects” and “associated with annoyance, sleep problems, concentration difficulties, and headache, with chronic-disease evidence too thin to close the file” are not interchangeable English. The first is a verdict. The second is a research status report. Propaganda prefers verdicts.
Dastan 2026 is not a no-effect paper
A writer who has read Dastan 2026 and still types “NO significant effects” is not confused by a paywall. The abstract and the concluding remarks of that paper are public. They describe dose-dependent biological pathways and an evidence gap in long-term human work. Using that paper as a prop for a no-effect slogan is quote-mining by caption.
What a fair summary of the human trial literature would say
A fair summary, written without a campaign to win, looks like this:
- Pure inaudible infrasound at moderate laboratory levels, over days to a few weeks, has not reliably moved standard cognitive tests or psychiatric scales in healthy adults. Ascone 2021 and Marshall et al. 2023 (72-hour simulated wind-turbine infrasound at about 90 dB peak; traffic noise, not infrasound, worsened wake after sleep onset) are the strongest exhibits for that limited claim.
- Annoyance, ear-pressure sensation, and sleep interference show up repeatedly when the signal includes audible or near-audible low-frequency energy, or when people live near real industrial LFN sources. That is Baliatsas, Møller, Leventhall’s reviews, and the WHO community-noise discussion of prominent low-frequency content.
- High-SPL and animal work show biological pathways that are real at doses most communities will never see. Those papers are not a permit to invent disease at residential lot lines. They are also not a permit to declare the pathways imaginary.
- No team has published the 1–200 Hz spectrum of an operating hyperscale campus paired with health endpoints in the neighboring houses. Until that exists, “data centers are cleared” is an inference, not a finding.
That paragraph is what “verifiable facts” would have looked like. The post did not write that paragraph, because that paragraph does not shut anyone up.
The measurement trick the post never names
Almost every official noise number a resident will be shown in a site-plan package is A-weighted (dBA). A-weighting was designed around the ear’s sensitivity to mid frequencies at modest levels. It applies on the order of 26 dB of attenuation at 63 Hz, more than 39 dB at 31.5 Hz, and more than 50 dB in the infrasound band. A facility can therefore “meet the ordinance” on the meter the ordinance names while the C-weighted or flat spectrum — the part that rattles a window, a sternum, or a night’s sleep — is a different document.
WHO community-noise guidance has said for years that when low-frequency components are prominent, the difference between C-weighted and A-weighted levels is the flag, and C-weighting should inform the assessment. G-weighting exists specifically for infrasound. A post that talks for several hundred words about infrasound and never mentions that the compliance meter throws infrasound away is not educating the reader. It is leaving the reader with the one number the applicant will be happiest to print.
Wheeler, Everett, Williamson & Prybutok, Sustainability 8(4):126 (7 August 2026), “Infrasound and Low-Frequency Noise in Data Center Environments,” make the data-center version of this point the center of their paper. Central sentence: no published study has measured the ILFN spectrum of an operational data center, nor examined health outcomes in workers or nearby residents against that exposure. A “facts” post published after that paper that still pretends the file is closed is choosing a narrative over the literature.
Source: mdpi.com/2571-8797/8/4/126
Equivocation: one word, three problems
The post uses “infrasound” as if it were the entire acoustic complaint. It is not.
| Layer | Band / heard? | What it actually is |
|---|---|---|
| Infrasound | Below 20 Hz. Usually not heard unless very loud. | Contested physiology. Felt as pressure or vibration at high sound-pressure levels. Almost never regulated. Thrown away by dBA (about 50 dB of attenuation). |
| Low-frequency noise | About 20–200 Hz. Often a hum or rumble. | High annoyance. Poor wall attenuation. Sleep disruption. Poorly captured by dBA (about 26 dB down at 63 Hz). |
| Mid / high mechanical | Above 200 Hz. Heard. | Classic noise: sleep, annoyance, WHO cardiovascular pathway. This is what dBA was designed for. |
| Generators / transformers | Mixed, often tonal. Heard. | Startle, night events, tonal penalty. Sometimes regulated, often poorly. |
Collapsing that stack into “inaudible infrasound” lets the author fight the easiest battle and declare the war over. “Traffic makes infrasound, even our own bodies produce infrasound” is the same move. A heartbeat is not a 24-hour industrial duty cycle. A train two miles away is intermittent. A hyperscale cooling yard is a stationary, high-duty-cycle, often tonal floor dropped into a rural residual ambient low enough that a slowing freight is part of the 3 a.m. soundscape. Duration, tonality, intermittency, and background level are the variables that decide whether a source is a curiosity or a chronic exposure. The post names none of them.
What the post drops on the floor
Even if every sentence about 6 Hz were true, the post would still be a bait-and-switch. The public case against hyperscale siting in Indiantown was never only infrasound.
- Water and power. Hyperscale campuses are grid-scale loads. Tesoro Groves is a 5,722-acre PUD framework on FPL land inside which data processing is a permitted use. Load, transmission, and cooling water are site questions, not vibe questions.
- Land use and the rural night. The Village annexed the FPL tract and approved the PUD framework on 30 April 2026 without an explicit data-center prohibition. Unincorporated neighbors sit on the other side of a line that is, in places, a fraction of a mile from Village jurisdiction. “Two miles from any residence,” the Silver Fox talking point from February 2026, is not a transferable fact about every future pad on a 5,722-acre light-industrial map.
- Codes that cannot see the problem they claim to regulate. If the noise ordinance is dBA-only, LFN and infrasound are structurally invisible. That is a drafting failure, not a health clearance.
- Governance. FPL is the dominant taxpayer in the Village, the landowner of the PUD, and the utility that would serve any campus. Residents who asked for prohibition language, infrastructure-pause language, or tighter acoustic conditions were not required to win an infrasound lab trial in order to ask those questions.
Reducing that stack to “people were scared of a frequency their bodies already make” is not analysis. It is a story that makes the neighbors the problem.
Pattern across months, not just this week
On 16 February 2026, Barbara Clowdus posted a “what do we know” thread on the Silver Fox proposal. It told readers the campus would be “envied by the residents of other data-center cities throughout the US,” that it was “NOT next to a residential area as is the case in many other areas, especially in northern Virginia,” and that it would stand “an estimated two miles — more than 10,000 feet — from any residence in any direction.” On 30 April 2026, after Silver Fox 606 withdrew, TCPalm quoted her calling the withdrawal “a potential loss” and saying she “fails to see how people can make a judgment until they see what’s actually proposed.”
| Date | Posture | Function |
|---|---|---|
| 16 Feb 2026 | Silver Fox “will be envied”; “NOT next to a residential area”; “two miles from any residence.” | Sell the site. Exile the Virginia complaints as somebody else’s problem. |
| 30 Apr 2026 | Withdrawal is “a potential loss”; “cannot judge until we see what’s proposed.” | Keep the industrial door open after the first application dies. |
| Post-election 2026 | Opponents were propagandists; studies show NO effects; sit down. | Convert a political outcome into a scientific gavel. |
Three postures. One outcome preference: the industrial door stays open, and the people who will not accept that door are recast as the information disease. That through-line is not curiosity.
Part III — What would close the file
The measurements that would settle it
If the author wants “verifiable facts about data centers,” the list is shorter and harder than a Facebook caption. None of it requires anyone to win a culture war.
- Full-spectrum measurements at the property line and at representative bedrooms under design cooling load: dBA, dBC, dBG, and unweighted third-octave from 1 Hz to 200 Hz, day and night, with tonality noted.
- Ordinance language that uses those numbers — not dBA alone — including a dBC−dBA trigger and a night limit designed for a rural residual ambient.
- Independent review of water, wastewater, and transmission, with figures tied to a named megawatt range rather than to “light industrial, maybe later.”
- Setbacks measured to the nearest existing dwelling, including unincorporated lots, not to a talking-point radius borrowed from a different project.
- Public posting of the raw acoustic files, not a consultant’s dBA contour map.
Until item 1 exists for a campus of the class being enabled on the Tesoro tract, every local claim that “infrasound from data centers has been studied and cleared” is false as a matter of the published record. Wheeler et al. said so in August. The post that calls other people propagandists did not mention it.
How to answer the post in public
Do not return smear for smear. Naming the method licenses a precise public reply. It does not license a pile-on, a nickname campaign, or a theory of the author’s soul. Those are the same well-poisoning move running in the other direction. The effective reply is three sentences and a demand.
You cited Baliatsas 2016 as finding no health effects. The paper found associations with annoyance, sleep problems, concentration difficulties, and headache, and asked for more research. You cited Dastan 2026 as a no-effect study. The paper calls infrasound an under-studied environmental stressor with dose-dependent biological pathways. You did not cite Wheeler 2026, which states that no operational data center has had its ILFN spectrum published. Post the 1–200 Hz measurements for the facility you are defending, or stop calling this settled.
That reply does not need a nickname, a motive lecture, or a theory of the election. It needs the papers and a meter. The reason the circulating post avoids that ground is that on that ground it loses.
Propaganda analysis is a tool. It is not a substitute for a meter. Even a clean win on this post leaves water, transmission, wetlands, tax structure, and ordinance drafting exactly where they were.
One-line verdict
She framed neighbors as the information disease, declared an election to be a scientific finding, cited papers that contradict her captions, equated a 24-hour industrial plant with a heartbeat, and omitted the one review that is actually about data-center sound. That is not a difference of emphasis. That is a constructed verdict wearing a bibliography. A person who does that in a land-use fight, while calling it “verifiable facts,” is acting as a propagandist. Part II is the exhibit.
Sources cited
- Dastan M, Dyminski Parente Ribeiro E, Bellut-Staeck U, Zhou J, Lehmann C. Infrasound and Human Health: Mechanisms, Effects, and Applications. Applied Sciences. 2026;16(3):1553. https://doi.org/10.3390/app16031553
- Baliatsas C, van Kamp I, van Poll R, Yzermans J. Health effects from low-frequency noise and infrasound in the general population: Is it time to listen? Sci Total Environ. 2016;557–558:163–169. https://doi.org/10.1016/j.scitotenv.2016.03.065
- Møller H. Physiological and Psychological Effects of Infrasound on Humans. J Low Freq Noise Vib Act Control. 1984;3(1):1–17.
- Ascone L, Kling C, Wieczorek J, Koch C, Kühn S. A longitudinal, randomized experimental pilot study to investigate the effects of airborne infrasound on human mental health, cognition, and brain structure. Sci Rep. 2021;11:3190.
- Marshall NS et al. The Health Effects of 72 Hours of Simulated Wind Turbine Infrasound. Environ Health Perspect. 2023. PMC10032045.
- Wheeler MR, Everett B, Williamson SM, Prybutok V. Infrasound and Low-Frequency Noise in Data Center Environments. Sustainability. 2026;8(4):126. https://www.mdpi.com/2571-8797/8/4/126
- Burbank K. Developer withdraws proposal for data center in Indiantown. TCPalm. 30 April 2026.
- Burbank K. A data center could now be built on FPL land in this Florida village. TCPalm. 6 May 2026.
- Randall J. Florida lawsuit against FPL’s land use for data centers gets dismissed. TCPalm. 8 July 2026.
- O’Hara T. Florida county commission requests data center moratorium. TCPalm. 26 August 2026.
- Barbara Clowdus, public Facebook posts, 16 February 2026 (“what do we know about the proposed data center”) and related threads characterizing Silver Fox setbacks and “envied by other data-center cities.”
Circulation note. Quote the papers, not the nickname. If a sentence in the circulating post is later corrected against the primary text, the correction should be acknowledged. Accuracy is the point of the exercise.