HistoricalChristian.Faith

Infant baptism is an ancient practice: dating it from the earliest evidence

Type: supporting argument (off-timeline) · Assessment: plausible — on the dating evidence infant baptism is an ancient practice. The firm, explicit record opens about c. 200, and it opens with a critic: the earliest explicit writer, Tertullian (De Baptismo 18, c. 203), opposes infant baptism — and an argument for delay presupposes a custom already in place and common enough to answer, which sets the practice securely before its first mention and, more likely than not, back into the second century. Within about fifteen years the Apostolic Tradition prescribes it as routine rubric (c. 215), and by c. 248 Origen, writing in a different region, reports it as a custom long inherited. That the earliest datum is hostile yet takes the practice for granted, and that no source anywhere preserves a memory of its arrival, makes an ancient, received custom the more probable reading and a late-third-century invention hard to credit. What the dating does not settle is the stronger claim of strict apostolic institution: no explicit witness survives before c. 200, and the pre-200 baptismal descriptions we do have (the Didache, Justin Martyr) portray instructed adults who “are persuaded and believe” — so whether the ancient custom reaches all the way to the apostles or crystallized within the sub-apostolic second century, the dating alone cannot decide. That stronger, breadth-based inference to apostolic origin is weighed separately, and comes out contested, on the companion antiquity & universality page. The verdict here is bounded: the practice is demonstrably ancient — established by c. 200 and probably older — while its strict apostolicity stays unproven · the early-dating reading advanced by Origen and, in modern form, Joachim Jeremias; the later-development reading by Kurt Aland and Everett Ferguson

Related arguments: the historical case — see also antiquity & universality · the delay of baptism · Irenaeus’s “reborn to God” · Polycarp’s eighty-six years

The claim

This is the chronological wing of the historical case — the question of the practice’s age. Not “is the practice universal?” (that breadth question is weighed on the antiquity & universality page) but “how old is infant baptism — how far back does the dating of the evidence let us trace it?” The paedobaptist answer — that it is an ancient, inherited custom, not a late graft — runs in two steps.

1. The first explicit mention already treats infant baptism as an existing custom. The earliest Christian writer to discuss the baptism of little children is Tertullian, and he writes to discourage it: weighing “the circumstances and disposition, and even age” of the candidate, he judges delay “preferable; principally, however, in the case of little children.” An author arguing that a practice should be slowed is an author for whom the practice already exists and is common enough to need answering — so the earliest explicit datum, though hostile, witnesses a custom already established in North Africa by c. 200, pushing its origin back into the second century.¹

2. The explicit witnesses then arrive fast, and quickly call the custom apostolic. Within about fifteen years the church order traditionally ascribed to Hippolytus prescribes the rite in rubric — “baptize the little children first… if they cannot [answer for themselves], let their parents or other relatives speak for them” (c. 215) — and by the 240s Origen, commenting on Romans, states flatly that “the Church received from the apostles the tradition of giving baptism even to little children.” The men who stand nearest the sub-apostolic period, on this reading, report infant baptism not as a novelty they remember arriving but as an old, inherited custom.² If the reading of Irenaeus that counts infants among those “born again to God” (c. 180) holds, the first — implicit — witness is earlier still, a generation before Tertullian; but that reading is itself contested.

The modern scholarly form of the argument is Joachim Jeremias’s: reading the dated evidence forward from the apostolic mission — the New Testament household baptisms, the analogy of Jewish proselyte baptism (which took in the children of a converting household), a Polycarp who at his martyrdom counted “eighty and six years” in Christ’s service, and the earliest funerary inscriptions of baptized infants — he argued for substantial continuity of infant baptism from the first generation.³

Who advances it

The apostolic-dating claim is as old as Origen, who first fixes the custom’s origin in the apostles, and it is pressed by Augustine and, at the Reformation, by Calvin (“there is no writer, however ancient, who does not trace its origin to the days of the apostles”). The same “silence proves nothing” line is pressed by Philip Schaff, who held that the apostolic fathers’ and the Didache’s failure to mention infant baptism “cannot be fairly used as an argument against” it and that “no time can be fixed at which it was first introduced,” and by Francis Schaeffer, who challenges that anyone denying an original infant-baptist practice “should be able to show in Church History when it started… there is no such break recorded.” Its modern champion is Joachim Jeremias (Infant Baptism in the First Four Centuries, 1958; ET 1960; and the reply The Origins of Infant Baptism, 1962; ET 1963). The rival later-development dating is the credobaptist and much of the critical-historical reading: Kurt Aland (Did the Early Church Baptize Infants?, 1961; ET 1963) answered Jeremias point for point, and Everett Ferguson’s comprehensive survey reaches a similar terminus — firm evidence only from c. 200, general infant baptism a third-century-and-later development. That later-dating reading has its own long pedigree: Adolf Harnack judged the custom’s adoption to lie in complete historical obscurity; the nineteenth-century critical exegetes — Meyer, De Wette, Olshausen — found “no conclusive proof-text” for infant baptism in the apostolic household narratives, Meyer inferring from 1 Corinthians 7:14 that “the baptism of children was not in use at that time”; the Baptist Dan Taylor read Justin’s professing-adult language as “clear and full evidence” that infant baptism was not yet practiced; and Thomas Schreiner reads the Apology of Aristides and the Cappadocians’ preference for candidates old enough to grasp the faith to the same effect.⁴

The evidence for

As a dating inference the paedobaptist case has real footholds.

• The earliest explicit voice is a critic — and a critic is a witness that the thing exists. This is the argument’s strongest single move. Tertullian is not reporting a curiosity; he is trying to curb a going practice, supplying reasons (the risk to the sponsors, the innocence of childhood) why it should be deferred. One does not write to slow down a custom that has not yet started. So even the hostile earliest datum places infant baptism securely in the church of c. 200 and implies an origin earlier than the first surviving complaint about it.⁵
• The explicit chain is short and it runs the right way. Discussion c. 203, rubric c. 215, apostolic-tradition claim c. 248, unanimous conciliar assumption c. 253 (Cyprian’s council debates only the day, never the subject). Across two generations no writer records the practice as new; each treats it as received. The nearer a source stands to the apostolic age, the more it speaks of infant baptism as old.
• Indirect data may push the line back before Tertullian. Jeremias’s reading of the New Testament household formulae, of proselyte baptism (which included a convert’s minor children), and of the Polycarp “eighty-six years” datum — if any of it holds — would locate children within the baptized company from the mission period, filling the apparent pre-200 gap with evidence that is real but not explicit.

Adversarial assessment

What the dating establishes. The core observation is sound and is not answered by dismissing Tertullian as an eccentric: the earliest explicit mention of infant baptism opposes it, and an argument for delay does presuppose a practice already under way. Whatever else is uncertain, infant baptism is a real, established custom in North Africa by c. 200, not a third-century invention out of nothing — and, since a hostile writer treats it as an entrenched thing to be curbed rather than a novelty to be named, the line most probably runs back into the second century. The debate is about how much earlier still it runs, not whether it reaches c. 200. So the practice is demonstrably ancient, and the strong credobaptist claim — that infant baptism is a late graft with no roots before the third century — does not survive the very first datum. That much the dating carries on its own.

What the dating will not settle. What it cannot reach is the further step from ancient to apostolic — that the practice was instituted by the apostles and kept from the first generation. Three things stop that stronger inference short of proof (they limit the reach back to the apostles, not the c. 200–and–probably–earlier antiquity just granted).

• No explicit evidence survives before c. 200, and the pre-200 descriptions we have portray believing adults. The earliest baptismal instructions — the Didache — presuppose a candidate who has been taught the “Two Ways” and who fasts before the rite, and Justin Martyr describes those baptized as people who “are persuaded and believe that what we teach… is true” and who then “are regenerated.” Neither text mentions infants; both describe an instructed, professing candidate. This is the datum Aland pressed: the direct record before Tertullian is not silent about baptism — it is fairly full — and what it describes is adult, faith-professing baptism, so Tertullian’s protest reads as naturally as resistance to a spreading practice as it does as an eccentric objection to an apostolic one.⁶
• Jeremias’s indirect evidence is each re-readable. The household formulae name no infant and repeatedly specify hearing, believing members (weighed on the household-baptisms page); the proselyte-baptism analogy imports a Jewish practice whose bearing on Christian baptism is disputed; Polycarp’s “eighty-six years” need mean only a lifelong Christian, not a baptized infant; and the datable infant tomb-inscriptions Jeremias cited cluster in the later period and often record emergency baptism of a dying child, which is precisely the setting Aland and Ferguson give for the practice’s spread. Indirect evidence that can be read either way cannot fix the date.
• The fourth-century pattern actively resists an early-universal-norm reading. Well after Tertullian, a striking number of children born to devout Christian parents were not baptized in infancy: Basil, Gregory of Nazianzus, Ambrose, Rufinus, Jerome, and Augustine himself were all baptized as adults despite Christian upbringing, and the deferral of baptism — even to the deathbed — became a fourth-century vogue before it receded (the Cappadocian Basil and his brother Gregory of Nyssa, and the Jerusalem catechist Cyril, are all witnesses to that adult catechumenate). Jeremias read this as a temporary decline from an earlier infant-baptism norm; Aland read it as showing that infant baptism of the children of Christian families had not yet become the norm. That two such opposite datings both fit the same fourth-century evidence is exactly why the chronology underdetermines the origin.⁷

The paedobaptist rejoinder keeps the question open rather than closing it the other way. It replies that an argument from the pre-200 silence is weak where the surviving texts are baptismal orders and apologetic sketches, written to describe the conversion of adult pagans, not to legislate the treatment of believers’ newborns — so their focus on professing converts is what the genre predicts and settles nothing about infants; that the fourth-century delayers are a special, theologically motivated cohort (fear of post-baptismal sin), not a survival of an original credobaptist norm; and that Tertullian’s protest, once granted, does the paedobaptist’s work, since it dates an established custom to a generation for which the credobaptist has only inference. On this rejoinder the development thesis must itself posit an early, traceless shift from adult-only to infant baptism — the very kind of silent, undated change it charges the apostolic-origin thesis with.⁸

Assessment

Plausible. Scoped to the question it can actually answer — how old is the practice? — the dating evidence yields a real and defensible result: infant baptism is ancient. The first explicit mention is already an argument against an established custom, so the practice is firmly in place by c. 200; and because that first witness is a hostile one treating the custom as entrenched, with no source anywhere recording its introduction, the more probable reading is that it reaches back into the second century as an inherited rite rather than a late graft. That is more than merely coherent — it is the likelier account of the dates, and it puts the strong credobaptist claim of a third-century novelty out of court. What the dating cannot do is carry the further, stronger step to strict apostolic institution. No explicit witness survives before c. 200; the pre-200 baptismal descriptions we do have portray instructed, professing adults; and Jeremias’s indirect evidence for an unbroken line to the first generation — the household formulae, proselyte baptism, Polycarp, the inscriptions — is at every point re-readable, as Aland and Ferguson showed. So the dates leave two extrapolations open behind the c. 200 terminus (unbroken apostolic continuity, or a development crystallizing within the second century around emergency baptism and original-sin theology), and cannot choose between them. That is why the verdict is bounded rather than a flat standoff: the antiquity of the practice is plausible, indeed probable, on the dating alone; the leap from ancient to apostolic is not, and that stronger inference — which rests less on the bare dates than on the practice’s later universality and the absence of any founding-controversy — is the properly contested question weighed on the antiquity & universality page. One caution for how the argument is deployed follows directly: the datum that Tertullian “argues against an existing custom” establishes antiquity, not apostolic origin — a proponent who treats “the earliest mention presupposes the practice” as equivalent to “the apostles instituted it” has slid a bounded inference into an unbounded one. The reading of Tertullian’s protest itself is set out on the Tertullian detail page.

Sources