A Poppe –Laka os Epis emological C i ique o he CDC’s 2025
Wo ding on Vaccines and Au ism
Dimas Tadeu Co as
Uni e si y o S˜ao Paulo
Abs ac
In No embe 2025, he Cen e s o Disease Con ol and P e en ion (CDC) modi ied a sec ion o
i s accine sa e y webpage by sugges ing ha he s a emen “ accines do no cause au ism” is no
an e idence-based claim. This a icle o e s an epis emological c i ique o such wo ding h ough he
amewo ks o Ka l Poppe ’s alsi ica ionism and Im e Laka os’s me hodology o scien i ic esea ch
p og ammes. I a gue ha he CDC’s o mula ion is logically lawed because i misunde s ands he
s uc u e o scien i ic in e ence, he epis emic s a us o nega i e causal indings, and he dema ca-
ion be ween p og essi e and degene a i e esea ch p og ammes. The conclusion is ha he claim
“ accines do no cause au ism” is p ecisely he kind o s a emen ha is suppo ed by ex ensi e
empi ical alsi ica ion and i s squa ely wi hin he me hodological no ms a icula ed by Poppe and
Laka os.
Keywo ds
Philosophy o Science; Poppe ; Laka os; Falsi ica ionism; Resea ch P og ammes; Scien i ic Me hod;
Epidemiology; Causa ion; Au ism; Vaccine Sa e y; Public Heal h Communica ion; E idence-Based
Medicine.
1 In oduc ion
Since he la e 1990s, he hypo hesis ha childhood accines cause au ism spec um diso de (ASD)
has been epea edly examined in epidemiological, clinical, and mechanis ic s udies. La ge-scale coho
analyses in ol ing millions o child en ac oss di e en coun ies ha e consis en ly ailed o demons a e
any causal associa ion. These indings ha e led mos scien i ic and egula o y bodies o conclude ha
accines do no cause au ism.
The CDC’s No embe 2025 websi e e ision in oduced signi ican epis emic ambigui y by s a ing
ha his conclusion is “no an e idence-based claim.” This pape sc u inizes his asse ion using p in-
ciples om he philosophy o science, pa icula ly he epis emological amewo ks de eloped by Ka l
Poppe and Im e Laka os. I aim o show ha he CDC’s language con adic s es ablished scien i ic
me hodology and unde mines he logical basis o causal in e ence in epidemiology. P e ious au ho s ha e
in oked Laka os’s no ion o degene a i e esea ch p og ammes o cha ac e ize accine denialism and he
shi ing accine–au ism hypo hesis. Pagnini (2023) ames he ecu ing econ igu a ion o an i- accine
a gumen s— om hime osal o MMR, aluminum, and “immune o e load”—as a pa adigma ic case o a
Laka osian degene a i e p og amme. Simila ly, Callaghan (2021) no es ha he accine–au ism claim
unc ions as a con empo a y illus a ion in discussions o Laka os’s model o scien i ic eg ess. O he
philosophy o science and STS analyses ha e employed Poppe ian and Laka osian ools o unde s and
pseudoscience and misin o ma ion su ounding accines. Howe e , o my knowledge, he e has been no
sys ema ic Poppe –Laka os analysis o he CDC’s own 2025 wo ding change on au ism and accines as a
case o epis emically eg essi e ins i u ional communica ion. This a icle aims o ill ha gap by applying
o mal c i e ia om alsi ica ionism and esea ch p og amme heo y di ec ly o a public heal h agency’s
he o ical aming o causa ion and e idence.
1
2 Poppe ’s Falsi ica ionism and he Logic o Nega i e Causal
Claims
2.1 The s uc u e o scien i ic es ing
Poppe (1959) a gues ha scien i ic hypo heses mus be alsi iable and ha science p og esses h ough he
ejec ion o hypo heses ha ail empi ical es s. Impo an ly: Scien i ic hypo heses, acco ding o Poppe
(1959), a e no con i med h ough he accumula ion o posi i e ins ances bu a e es ed by exposing hem
o po en ial e u a ion. Science does no aim o p o e uni e sal nega i es, no does i es ablish ce ain y;
ins ead, i examines posi i e causal hypo heses by de i ing es able p edic ions and subjec ing hem o
igo ous empi ical sc u iny. When hese p edic ions epea edly ail, he associa ed hypo hesis is ejec ed.
In his amewo k, he ele an ques ion is no whe he a claim can be disp o en wi h absolu e ce ain y,
bu whe he he a ailable e idence consis en ly alsi ies he causal asse ion unde conside a ion.
Thus, he ele an hypo hesis in his case is:
H1: Vaccines cause au ism.
I empi ical s udies sys ema ically ail o co obo a e H1, hen Poppe ian me hodology equi es e-
jec ion o he causal claim.
2.2 Nega i e indings as e idence
In epidemiology, a la ge body o nega i e e idence— esul s showing no associa ion—coun s as e idence
o absence when:
1. The s udies a e su icien ly powe ed,
2. The designs minimize bias,
3. Mul iple popula ions a e examined,
4. Replica ion is achie ed,
5. Con ounding is con olled.
These c i e ia ha e been me in accine-au ism esea ch (Taylo e al., 2014; H iid e al., 2019; Jain
e al., 2015). Thus, in Poppe ian e ms, he hypo hesis has been ho oughly es ed and alsi ied.
2.3 Why he CDC wo ding misapplies Poppe ian logic
By sugges ing ha he s a emen “ accines do no cause au ism” lacks e idence, he CDC implici ly
equi es p oo o a uni e sal nega i e—some hing Poppe explici ly s a ed is impossible and unnecessa y.
Ins ead, he co ec epis emic o mula ion is:
The causal hypo hesis “ accines cause au ism” has been es ed and ejec ed.
3 Laka os: Resea ch P og ammes and he Dema ca ion Be-
ween P og essi e and Degene a i e Science
3.1 The Laka osian amewo k
Laka os (1978) concep ualizes science as a se ies o esea ch p og ammes, each s uc u ed a ound a ha d
co e o undamen al assump ions ha a e no eadily abandoned, su ounded by a p o ec i e bel o
auxilia y hypo heses ha can be adjus ed in esponse o empi ical p essu e. The o e all ajec o y o a
esea ch p og amme is shaped by i s me hodological ules and can de elop in one o wo di ec ions: i
may be p og essi e, gene a ing no el p edic ions and yielding inc easing explana o y powe , o i may
become degene a i e, elying on ad hoc modi ica ions ha do no lead o new co obo a ed indings.
2
3.2 The accine–au ism hypo hesis as a degene a i e p og amme
Using Laka os’s c i e ia, he o iginal MMR–au ism hypo hesis (Wake ield e al., 1998) was ne e epli-
ca ed, and subsequen e o mula ions o he claim shi ed i egula ly om hime osal, o accine sched-
ules, o aluminum, and hen o no ions o “immune o e load,” each appea ing only a e p io e sions had
been alsi ied. None o hese successi e i e a ions p oduced no el p edic ions o independen co obo a-
ion. Taken oge he , he accine–au ism line o inqui y exempli ies a degene a i e esea ch p og amme.
3.3 Au ism esea ch as a p og essi e p og amme
In con as , mode n au ism esea ch—spanning gene ics, neu ode elopmen al biology, and en i onmen-
al ac o s un ela ed o accines—has ad anced as a clea ly p og essi e p og amme. I has gene a ed
obus and inc easingly p edic i e models o ASD, iden i ied mo e han one hund ed gene ic loci as-
socia ed wi h neu ode elopmen al pa hways, and p oduced mechanis ic explana ions ha cohe e wi h
con empo a y neu obiological e idence. Recen la ge-scale genomic and de elopmen al s udies, including
wo k iden i ying mo e han se en y high-con idence isk genes and de ailed mechanis ic pa hways, u -
he illus a e he expanding explana o y powe and empi ical e inemen cha ac e is ic o con empo a y
au ism esea ch (Lo d e al., 2020; Fu e al., 2022).
3.4 Why he CDC wo ding is inconsis en wi h Laka os
By implying ha ejec ing he accine–au ism causal claim lacks e idence, he CDC ails o acknowl-
edge he degene a i e s a us o he accine–au ism esea ch p og amme. As no ed in he p eceding
subsec ions, each successi e e o mula ion o he hypo hesis eme ged only a e p io e sions had been
empi ically e u ed, and none has p oduced no el p edic ions, independen co obo a ion, o mechanis ic
suppo . F om a Laka osian pe spec i e, his pa e n exempli ies a degene a ing esea ch p og amme
whose p o ec i e bel expands only h ough ad hoc modi ica ions ha do no inc ease explana o y o
p edic i e powe .
This makes he CDC’s wo ding especially p oblema ic: i o e looks he degene a i e cha ac e o he
accine–au ism hypo hesis and obscu es he clea me hodological g ounds on which he en i e line o
inqui y has al eady been decisi ely abandoned.
4 A gumen om Igno ance and he Bu den o P oo
The CDC s a emen also commi s a classical allacy: he a gumen om igno ance (a gumen um ad
igno an iam). This allacy a ises when a p oposi ion is ea ed as plausible me ely because i has no
been disp o en wi h absolu e ce ain y. In scien i ic easoning, howe e , unce ain y is he de aul , and
he absence o comple e disp oo does no cons i u e posi i e e idence in a o o a claim.
A s a emen such as “we canno ule ou Xwi h 100% ce ain y” does no imply ha Xis he e o e
plausible. We e his p inciple adop ed, an unlimi ed ange o un ounded causal hypo heses would emain
pe manen ly “open ques ions.” Fo ins ance, i is impossible o p o e wi h absolu e ce ain y ha
sma phones do no cause au ism, ha wea ing ed shi s inc eases he isk o measles, o ha sola
eclipses igge in luenza ou b eaks. I is equally impossible o elimina e he logical possibili y ha
bananas, shoelaces, o sea bel s con ibu e o au ism. Ye none o hese claims wa an s scien i ic
conside a ion, because none mee s e en he minimal e iden ia y h esholds equi ed o causal in e ence.
These examples unde sco e he undamen al poin : scien i ic easoning assigns he bu den o p oo
o he pa y making a posi i e causal claim. The ele an ques ion is no whe he an e ec can be
conclusi ely uled ou , bu whe he he e exis s su icien e idence o jus i y belie in he causal hypo hesis
in he i s place.
In epidemiology, a causal asse ion (“Xcauses Y”) mus be suppo ed by consis en associa ions,
empo al sequencing, biological plausibili y, dose– esponse ela ionships, eplica ion ac oss popula ions,
and con e gence o expe imen al, clinical, and mechanis ic e idence. The accine–au ism hypo hesis
sa is ies none o hese c i e ia. I has been epea edly es ed and epea edly alsi ied. To ea i as “s ill
possible” because uni e sal nega i es canno be p o en is o in e he logic o causa ion.
Thus, he CDC’s wo ding does no me ely so en a scien i ic conclusion; i misplaces he bu den o
p oo and g an s unwa an ed epis emic s anding o a hypo hesis ha has al eady been sys ema ically
ejec ed by he a ailable e idence.
3
5 Public Scien i ic Communica ion and Epis emic Responsibil-
i y
Scien i ic ins i u ions ca y a esponsibili y no only o e alua e e idence, bu also o communica e i
wi h cla i y, consis ency, and epis emic in eg i y. In ields such as accine sa e y—whe e public unde -
s anding di ec ly a ec s popula ion-le el heal h ou comes— he p ecision o o icial language is i sel a
o m o public heal h in e en ion. When au ho i a i e bodies in oduce concep ual ambigui y in o well-
es ablished scien i ic conclusions, hey isk gene a ing misin e p e a ions ha p opaga e a beyond he
echnical con ex in which such s a emen s we e c a ed.
The CDC’s 2025 wo ding on au ism and accines illus a es his dange . By sugges ing ha he claim
“ accines do no cause au ism” is “no e idence-based,” he agency inad e en ly legi imizes a hypo hesis
ha has been epea edly alsi ied in high-quali y esea ch. Such ph asing no only mis ep esen s he
e iden ia y landscape bu also unde mines decades o wo k aimed a coun e ing misin o ma ion. In he
con ex o public heal h, epis emic ambigui y unc ions as a ec o : i lowe s us , shi s indi idual
isk pe cep ion, and p o ides he o ical suppo o accine hesi ancy mo emen s ha ely on p ecisely
his ype o ins i u ional unce ain y. This dynamic is well documen ed in he science-communica ion
li e a u e: e en sub le ambigui ies in ins i u ional messaging ha e been shown o ampli y misin o ma ion,
e ode public us , and educe adhe ence o e idence-based heal h ecommenda ions (Lewandowsky e
al., 2017; Scheu ele and K ause, 2019).
F om he pe spec i e o science communica ion, he p oblem is no me ely seman ic. I cons i u es
a b eakdown o wha may be called an epis emic du y o ca e— he obliga ion o scien i ic agencies o
p e en o eseeable ha m a ising om unclea o misleading s a emen s. Because public heal h decisions
ope a e unde condi ions o impe ec in o ma ion and equi e high le els o collec i e us , e en a small
depa u e om es ablished e iden ia y consensus can ha e disp opo iona e e ec s, including educed
accina ion a es, esu gence o p e en able diseases, and ampli ica ion o pseudoscien i ic na a i es.
In his sense, he CDC’s ph asing ep esen s no only a concep ual e o bu an ins i u ional epis emic
ailu e, one ha con lic s wi h he co e p inciples o scien i ic communica ion in democ a ic socie ies.
E ec i e public heal h policy equi es language ha is no only echnically co ec bu also aligned wi h
es ablished e idence hie a chies and sensi i e o he dynamics o misin o ma ion. When ins i u ions
de ia e om his s anda d, hey inad e en ly pa icipa e in he e osion o he epis emic ounda ions
upon which public heal h elies.
6 Conclusion
Using Poppe ’s alsi ica ionism and Laka os’s esea ch p og ammes, we can conclude:
1. The s a emen “ accines do no cause au ism” is a legi ima e e idence-based s a emen g ounded
in epea ed empi ical alsi ica ion.
2. The accine–au ism hypo hesis is a degene a i e esea ch p og amme.
3. The CDC’s 2025 wo ding miscommunica es he epis emic s a us o exis ing e idence.
4. A mo e accu a e scien i ic s a emen would be:
“The hypo hesis ha accines cause au ism has been ex ensi ely es ed and
consis en ly ejec ed by high-quali y e idence.”
Acknowledgmen s
The au ho used Cha GPT (OpenAI, GPT-5.1) o assis wi h he d a ing and linguis ic e inemen o
po ions o his manusc ip . All in ellec ual con en , a gumen a ion, e isions, and conclusions a e he
sole esponsibili y o he au ho , who has e i ied he accu acy o all in o ma ion p esen ed.
Re e ences
Poppe K. The Logic o Scien i ic Disco e y. Hu chinson; 1959.
Laka os I. The Me hodology o Scien i ic Resea ch P og ammes. Camb idge Uni e si y P ess; 1978.
4
Feye abend P. Agains Me hod. Ve so; 1975.
Taylo LE, Swe d ege AL, Eslick GD. Vaccines a e no associa ed wi h au ism. Vaccine. 2014;32(29):3623–
3629.
H iid A, Hansen JV, F isch M, Melbye M. Measles, mumps, ubella accina ion and au ism. Ann In e n
Med. 2019;170(8):513–520.
Jain A, e al. Au ism occu ence by MMR accine s a us among US child en. JAMA. 2015;313(15):1534–
1540.
Wake ield AJ, e al. Ileal-lymphoid-nodula hype plasia, non-speci ic coli is, and pe asi e de elopmen-
al diso de in child en. Lance . 1998;351(9103):637–641.
Nyhan B, Rei le J. The oles o in o ma ion de ici s and iden i y h ea in science denial. PNAS.
2015;112(52):15837–15842.
Ins i u e o Medicine. Ad e se E ec s o Vaccines: E idence and Causali y. Na ional Academies P ess;
2012.
Lo d C, B ugha TS, Cha man T, e al. Au ism spec um diso de . Na Re Dis P ime s. 2020;6(1):5.
doi:10.1038/s41572-019-0138-4.
G o e J, Ripke S, Als TD, e al. Iden i ica ion o common gene ic isk a ian s o au ism. Na Gene .
2019;51(3):431–444. doi:10.1038/s41588-019-0344-8.
Fu JM, Sa e s om FK, Peng M, e al. Ra e coding a ia ion p o ides insigh in o he gene ic a chi ec-
u e and pheno ypic con ex o au ism. Na Gene . 2022;54:1283–1293. doi:10.1038/s41588-022-01146-2.
Lewandowsky S, Cook J, Fay N, Gignac GE. Ra ional denial o scien i ic e idence. Psychol Sci. 2017;28(6):691–699.
Scheu ele DA, K ause NM. Science audiences, misin o ma ion, and e idence-based communica ion. P oc
Na l Acad Sci USA. 2019;116(16):7662–7669.
O eskes N, Conway EM. Me chan s o Doub . Bloomsbu y P ess; 2010.
Cen e s o Disease Con ol and P e en ion. Au ism and Vaccines — Vaccine Sa e y. Upda ed No embe
2025. U.S. Depa men o Heal h and Human Se ices; 2025. A ailable om: h ps://www.cdc.go /
accinesa e y/au ism.
Pagnini C. Vaccine Hesi ancy and Degene a i e Resea ch P og ammes: A Laka osian Analysis. Phil-
Pha m Blog; 2023. A ailable om: h ps://philpha mblog.wo dp ess.com.
Callaghan M. Laka os Re isi ed: Inno a ion and ‘No el Fac s’. EconS o ; 2021. A ailable om:
h ps://www.econs o .eu.
Au ho (s) unknown. Poppe , Pseudoscience, and he Logic o Dema ca ion in Vaccine Deba es. PhilSci
A chi e; 2020. A ailable om: h ps://philsci-a chi e.pi .edu.
5
