This review article presents a hypothesis that infected but asymptomatic individuals can transmit COVID-19 by generating infectious respiratory particles during normal speech. The author describes two distinct particle types produced while speaking — large droplets and small aerosols — and links these to different transmission modes: droplet infection by direct collision and airborne infection via aerosol accumulation. The work is presented as a hypothesis and review and was published in the Japanese Journal of Infectious Diseases (Jpn J Infect Dis). Publication metadata in the source: Sakae Inouye; Jpn J Infect Dis. 2026 Sep 17;79(5):218–221. PMID 41320299; DOI 10.7883/yoken.JJID.2025.169.
According to the hypothesis, one class of infectious particles arises from saliva when pronouncing plosive consonants, specifically /p/ and /t/. Those speech sounds produce relatively heavy droplets that, because of their larger size and mass, are postulated to travel a short distance and can physically collide with the faces of nearby conversation partners. The author frames this route as a conventional droplet transmission mechanism: expelled droplets contact mucous membranes or contaminate surfaces proximal to the speaker, enabling infection in close conversational settings.
The source emphasizes the role of saliva as the origin for these larger droplets and links plosive articulation to their generation. The account treats this process as a plausible explanation for close-range transmission events attributable to speaking, particularly in face-to-face interactions.
The second class of particles the author proposes consists of much smaller aerosols formed from laryngeal secretions during the voicing of vowels. Vocal cord vibration that produces voiced vowel sounds is hypothesized to aerosolize laryngeal fluids into lighter particles that remain suspended in air. Because these aerosol particles are lighter, they can float and persist for longer periods and therefore may accumulate in enclosed, poorly ventilated spaces during prolonged conversation.
The hypothesis suggests that such accumulation of aerosols inside a non-ventilated room can create an environment conducive to airborne infection in other occupants of that same room, even when the infected person shows no symptoms. The distinction drawn in the paper is functional: heavy droplets cause direct-contact droplet infection at close range, while aerosols enable longer-range or cumulative airborne exposure in enclosed spaces.
The author discusses global epidemiologic patterns of COVID-19 in light of the speaking-derived particle hypothesis. One observation noted in the source is that COVID-19 mortality appeared higher in some high-income, hygienic countries — a pattern the author highlights as inconsistent with typical communicable disease distributions that favor low-income settings.
The paper proposes that factors such as climate and housing structure could influence aerosol accumulation and the likelihood of airborne transmission. For example, differences in indoor ventilation, room size, and behaviors related to heating or air conditioning are presented as possible contributors to between-country variation in transmission dynamics, according to the hypothesis.
Beyond physical factors, the author raises cultural behaviors and linguistic differences as potential contributors to variable transmission. Cultural norms governing close conversation, social spacing, frequency and duration of indoor gatherings, and typical speech volume could all affect the production and spread of speech-generated particles.
Additionally, the author posits that differences in languages — specifically, the phonetic elements prevalent in particular languages (for example, frequency of plosive consonants versus vowels and voicing characteristics) — might influence the relative generation of droplets versus aerosols during typical speech. The source frames this as a hypothesis to explain some heterogeneity in COVID-19 epidemiology across regions.
The article is explicitly framed as a hypothesis and review; the PubMed abstract does not report quantitative experimental data, measurements, or confirmatory studies testing the detailed mechanisms. The source identifies plausible mechanisms by which speaking could produce both droplets and aerosols from different anatomic secretions (saliva versus laryngeal secretions) and links these mechanisms to droplet and airborne transmission pathways, but detailed empirical evidence and numerical parameters were not reported in the abstract.
Readers interested in experimental methods, measurements of particle size distributions, viral load in speech-generated particles, or quantitative assessments of ventilation effects should consult the full text of the article. The PubMed abstract includes the keywords aerosols, asymptomatic infection, droplets, and speaking and lists related MeSH terms such as Respiratory Aerosols and Droplets, COVID-19/transmission, and Speech.
Publication details and identifiers available in the source: Sakae Inouye. Jpn J Infect Dis. 2026;79(5):218–221. PMID 41320299. DOI 10.7883/yoken.JJID.2025.169.