HistoricalChristian.Faith

Baptism as the new-covenant circumcision: does the covenant sign belong to believers’ infants?

Type: supporting argument (off-timeline) · Assessment: contested — the systematic case is genuinely strong: one covenant of grace runs through both Testaments, its sign passes to believers and their seed (Genesis 17), and Paul does draw circumcision and baptism together (Colossians 2:11–12), so on a continuity reading the child of a believer stands within the covenant and its sign follows him. But the whole inference rests on one disputed premise — that new-covenant membership tracks physical descent as the Abrahamic sign did — which the credobaptist denies from the very texts that describe the new covenant (Jeremiah 31:31–34: its members “shall all know” the Lord); and the two proof-texts pressed hardest (Colossians 2:11–12, Acts 2:39) neither state infant baptism nor were read that way by the earliest commentators on them. Strong by systematic theology’s lights, weak by the exegesis of the specific texts — the signature of a genuinely disputed claim, and the fault line between whole confessional traditions · advanced first by Huldrych Zwingli (1525) and given its classic form by John Calvin and the Reformed tradition (scholarly form: John Murray, Pierre-Charles Marcel), answered by Paul K. Jewett and the Reformed-Baptist covenant theologians

Related arguments: the New Testament warrant — see also the household baptisms · children “holy” · suffer the little children

The claim

This is the backbone of the Reformed case for infant baptism, and it is a covenantal argument rather than a proof-text one. It runs in three steps.

1. God gave the covenant sign to believers’ infants. When the Lord covenanted with Abraham, he promised, “I will establish my covenant between me and thee and thy seed after thee in their generations for an everlasting covenant, to be a God unto thee, and to thy seed after thee” (Genesis 17:7), and commanded that the sign of that covenant be placed on the eight-day-old infant: “he that is eight days old shall be circumcised among you, every man child in your generations” (Genesis 17:12). The child received the sign not on his own profession but as a member of the covenant household.

2. Baptism is the new-covenant counterpart of circumcision. Writing to the Colossians, Paul sets the two side by side:

In whom also ye are circumcised with the circumcision made without hands, in putting off the body of the sins of the flesh by the circumcision of Christ: Buried with him in baptism, wherein also ye are risen with him through the faith of the operation of God, who hath raised him from the dead. (Colossians 2:11–12)

The “circumcision made without hands” is identified with what happens in baptism — so baptism is the sign that now does what circumcision once did.

3. The promise still runs to the children. At Pentecost Peter closes his call to “repent, and be baptized” with, “For the promise is unto you, and to your children, and to all that are afar off, even as many as the Lord our God shall call” (Acts 2:38–39). The covenant structure — the sign to the believer and his seed — is unchanged. Therefore the infant children of believers rightly receive baptism, the covenant sign, exactly as they once received circumcision.¹

Who advances it

Its first Reformation proponent was Zwingli, who pressed the circumcision-covenant parallel against the Zürich Anabaptists a decade before the Institutes (baptism “the same as” circumcision, so given to a believer’s infants as circumcision was); but the argument’s classic and most sustained form is Calvin’s. He treats circumcision and baptism as two signs of one covenant, differing only in outward ceremony: “We have, therefore, a spiritual promise given to the fathers in circumcision, similar to that which is given to us in baptism, since it figured to them both the forgiveness of sins and the mortification of the flesh”; the covenant is “common, and the reason for confirming it is common… they had circumcision, instead of which we now have baptism.” From this identity of covenant he concludes that “the baptism of the children of Christian parents” is “as competent as the circumcision of Jewish children.”² His positive ground is the seal: baptism “confirms the promise given to the godly parent, and declares that the Lord will be a God not to him only, but to his seed” (Genesis 17), so a believer’s child already belongs to the covenant the sign confirms and must not be denied it.³ In modern systematics the same case is built out by John Murray (Christian Baptism), Pierre-Charles Marcel (The Biblical Doctrine of Infant Baptism), and Geoffrey Bromiley (Children of Promise), all resting it on the unity of the covenant of grace across the Testaments and the continuity of its sign.⁴

The evidence for

The steelman is a systematic argument, and on its own terms it is coherent and forceful:

• One covenant of grace. Scripture presents the Abrahamic covenant not as a superseded national arrangement but as the covenant in which the gospel was “preached before… unto Abraham” (Galatians 3:8) and into which Gentile believers are grafted (Romans 11). If the covenant of grace is one, and God once appointed its sign for believers and their children together, the burden of proof falls on anyone claiming that the children have now been excluded — a narrowing the New Testament nowhere announces. Continuity is the default; discontinuity must be shown.
• Paul himself links the two signs. Colossians 2:11–12 is not an invention of later theology: Paul really does describe baptism in the language of circumcision (“the circumcision made without hands… buried with him in baptism”). The oldest exegesis of the passage reads it the same way. Already c. 150 Justin Martyr tells Trypho that Christians “have received not carnal, but spiritual circumcision… through baptism” (Dialogue 43); Chrysostom: “No longer… is the circumcision with the knife, but in Christ Himself… When and where? In Baptism”; Cyprian ranks it among the texts showing “the first circumcision of the flesh is made void, and the second circumcision of the spirit is promised instead.” And when a mid-third-century African bishop, Fidus, proposed timing infant baptism by the circumcision calendar — waiting the eighth day — Cyprian’s council of sixty-six answered only that the eighth-day figure had been fulfilled in Christ’s “circumcision of the spirit,” never questioning that baptism was the church’s circumcision: the parallel was so settled that the whole dispute over the day presupposed it. That baptism is the fulfilment of circumcision is thus apostolic and patristic common ground, not a Reformed novelty.⁵
• The sign followed the covenant member, not his profession. Under the old administration the infant received the sign because he belonged to the covenant people, before and without any faith of his own. If the new-covenant sign works the same way — marking covenant membership, not certifying an accomplished regeneration — then a believer’s child is its natural subject, and Acts 2:39 (“the promise is unto you, and to your children”) reads as the expected continuation of “a God… to thee, and to thy seed.”

Its power is that it does not depend on finding an infant named at any apostolic baptism (the weakness of the household-baptisms argument): it argues from the structure of the covenant, so the silence of the New Testament about excluding children counts for it. That is why the argument persuades even some who concede the household texts prove nothing, and why a careful credobaptist critic could title his book-length answer An Appraisal of the Argument That as Infants Were Once Circumcised, So They Should Now Be Baptized — the argument is serious enough to need a book.

Adversarial assessment

What holds. The continuity framework is real and cannot be dismissed. The covenant of grace does run through both Testaments; the sign of that covenant was once given to infants by God’s own command; and Paul does connect baptism to circumcision. If new-covenant membership is constituted the way old-covenant membership was — by birth into a believing household as much as by personal faith — the paedobaptist conclusion follows almost automatically, and the demand that Scripture explicitly re-authorize infants misreads where the burden lies. As a piece of systematic theology the argument is internally consistent and has persuaded a large part of the Reformed world for five centuries.

What is disputed. The whole case turns on a single hinge — whether the new covenant is administered along the same mixed, genealogical lines as the old — and that hinge is exactly what the credobaptist contests, from the texts that describe the new covenant:

• The new covenant is described as regenerate in membership. Jeremiah’s new-covenant oracle — the one the New Testament quotes to interpret what Christ instituted — makes internal knowledge of God constitutive of belonging: “they shall all know me, from the least of them unto the greatest of them… for I will forgive their iniquity” (Jeremiah 31:33–34; cited at Hebrews 8:8–12). Where the Abrahamic sign was given to a mixed body (believing and unbelieving Israel alike), the new-covenant community is said to consist of those who know the Lord. If that is right, the covenant sign belongs to professed believers, not to every child of the household — the continuity the paedobaptist assumes is precisely the discontinuity Jeremiah announces. Leading credobaptists (Jewett; Wellum and Gentry, Kingdom through Covenant) make this the load-bearing point; John MacArthur calls the regenerate membership of the new covenant “the watershed issue” of the whole debate.⁶
• Galatians 3 redefines Abraham’s “seed” by faith, not descent. The paedobaptist leans on the Abrahamic promise “to thee, and to thy seed” (Genesis 17) and on Galatians 3:8 (the gospel “preached before… unto Abraham”) — but the same chapter says who that seed now is: “they which are of faith, the same are the children of Abraham” (Galatians 3:7), and “if ye be Christ’s, then are ye Abraham’s seed, and heirs according to the promise” (3:29). Gavin Ortlund presses this: the church was “defined by physical descent throughout the OT (Genesis 17:9), and defined by spiritual descent throughout the NT (Galatians 3:7),” so the covenant sign that once marked the genealogical seed now marks believers — and he sharpens it into a reductio (“Why Not Grandchildren?”): circumcision ran intergenerationally down the national family, so if baptism simply replaces it on continuity grounds, the paedobaptist should baptize believers’ grandchildren too, which none does. The reading is ancient, not a modern novelty: Jerome, commenting on the verse, holds that Abraham’s true children are those who believe as he believed while yet uncircumcised, and that circumcision was only a temporary national marker — a “sign of prior faith” given to keep Israel distinct among the nations until Christ, now “ceased” and replaced by “the spiritual circumcision of the Lord Jesus Christ.” The paedobaptist answers that Galatians is defining the heirs of justification, not legislating who may receive the covenant sign — the administrative question Acts 2:39 keeps open — but the credobaptist has at least located his discontinuity inside Paul’s own argument, in the very chapter the paedo cites for continuity.⁷
• Colossians 2:11–12 conditions the whole thing on faith. The very verse that links baptism to circumcision ties the burial-and-rising to the recipient’s response: “buried with him in baptism, wherein also ye are risen with him through the faith of the operation of God” (v. 12). Read as a whole, the passage describes what happens to a believer who has put off the body of sins “through faith,” not a rite worked on a passive infant — so the text that supplies the circumcision-baptism link also, on the credo reading, specifies a believing subject. The parallel cuts both ways.
• The specific proof-texts were not read as commanding infant baptism. Neither key verse states the conclusion, and the earliest commentators drew it from neither. On Colossians 2:11–12 the fathers expound spiritual circumcision, the putting-off of sins in baptism, and burial-and-rising — none reasons from it to the baptism of infants. On Acts 2:39 they gloss “and to your children” as descendants / heirs, marking the reach of the promise down the generations and out to the Gentiles (“those afar off”), not as infants present to be baptized: Chrysostom, “the gift is greater, when these are to be heirs of the blessings”; Oecumenius and Theophylact, that the promise is “sufficient not only for them but also for their children” because “a gift is more valuable when the blessings pass also to one’s heirs”; Bede refers it to Joel’s “your sons and your daughters shall prophesy.” The covenantal reading may still be true, but it is a systematic inference laid over the texts, not the plain sense the ancient exegetes found in them.⁸

A popular sharpening of the objection makes the parallel itself the target. The credobaptist apologetic site GotQuestions puts it flatly: the baptism–circumcision equation “is unbiblical… The New Testament nowhere describes baptism as the New Covenant replacement for Old Covenant circumcision” and “nowhere describes baptism as a sign of the New Covenant”; what admits a person to the new covenant’s blessings is faith in Christ, not a covenant sign laid on him in infancy (1 Corinthians 11:25; 2 Corinthians 3:6; Hebrews 9:15). This is an argument from silence, and it overstates the silence: Colossians 2:11–12 is precisely the place where the apostle does draw the two together, tying the “circumcision made without hands” directly to being “buried with him in baptism,” and the second-century church was already reading baptism as the “spiritual circumcision” that replaced the carnal one (Justin, Dialogue 43) — so the equation is ancient, not a Reformation invention, and “nowhere” is too strong. But the concession the silence does register is real: even at Colossians 2 the burial-and-rising is “through the faith of the operation of God,” and no infant is in view; the New Testament nowhere makes the further step from “baptism is the new circumcision” to “therefore administer it to believers’ infants.” That step is supplied by the covenant framework, not by an explicit apostolic command — which is exactly the point in dispute. The exchange does not break the deadlock; it relocates it back onto Colossians 2:11–12, where each side finds what it brought.⁹

A structural pressure point turns the covenant logic back on the paedobaptist. If baptism succeeds circumcision because both are covenant-entry signs laid on the household, the same typology runs on one square further: the Lord’s Supper stands to Passover as baptism stands to circumcision — and Passover, too, was a covenant meal eaten in the household with its children (Exodus 12:26–27). By the identical continuity reasoning the covenant child belongs at the Table as much as at the font — i.e. paedocommunion, the corollary the Eastern churches and a growing Reformed minority in fact draw. Yet Calvin, who pressed the circumcision parallel hardest, barred infants from the Supper, appealing to Paul’s demand that a communicant “examine himself” (1 Corinthians 11:28) — a self-examination no infant can perform. The credobaptist presses the dilemma: either the covenant-sign logic proves too much (carrying the infant to the Table as well as the font), or the very thing that bars him from the Table — personal, examinable faith — is exactly what should bar him from baptism. Calvin’s distinction — that the Supper requires an active, self-examining reception the passive infant cannot give, while baptism does not — is a coherent reply, but it concedes the point in dispute: the two covenant signs are not simply parallel, and once the parallel is qualified at the Table it no longer carries the child straight from circumcision into baptism unqualified.¹⁰

The paedobaptist rejoinder is itself substantial, which is why the question stays open. Paedobaptists answer that the new-covenant community is still visibly mixed — Hebrews’ own warnings against apostates who “counted the blood of the covenant… an unholy thing” (Hebrews 10:29) treat covenant-breakers as members — and distinguish the covenant’s external administration (into which children are born) from its internal saving reality; some read Jeremiah’s “all shall know me” as an already/not-yet promise, or (with a recent exegetical study) argue that Jeremiah’s “from the least to the greatest” is elsewhere a pervasive-but-not-exhaustive quantifier, so it cannot by itself carry the regenerate-only reading. The credobaptist replies that this empties the prophesied “newness” of its content. Neither side has dislodged the other, and even a paedobaptist as careful as Richard Pratt can grant the regenerate reading of Jeremiah 31 while relocating its full realization to the consummation.¹¹

Assessment

Contested. This is the strongest single argument for infant baptism — not a proof-text but a whole covenantal architecture — and it is genuinely strong by systematic-theology lights: one covenant of grace, a sign once given to believers’ children by divine command, and an apostle who links baptism to circumcision. If new-covenant membership is constituted as old-covenant membership was, the conclusion follows. But that “if” is the crux, and it is exactly what the new-covenant texts are invoked on both sides to decide: the credobaptist reads Jeremiah 31 and Hebrews 8 to say the new covenant is regenerate in membership (so its sign is for believers), Colossians 2:12 to condition baptism on faith, and Galatians 3 to redefine Abraham’s seed by faith rather than descent (so the genealogical premise is undercut in the very chapter the paedo cites for continuity); the paedobaptist reads the covenant’s external administration and Hebrews’ apostasy warnings to keep the community mixed (so its sign still runs to the children). A structural pressure point sharpens the impasse without breaking it: the Baptism : Circumcision :: Supper : Passover typology would carry the covenant child to the Table as well as the font, yet Calvin barred infant communion on the ground that the Supper requires a self-examination (1 Corinthians 11:28) baptism does not — a coherent reply that nonetheless concedes the two signs are not simply parallel, so the parallel cannot by itself carry the child from circumcision into baptism. The dispute is not over a detail but over the nature of the new covenant itself — which is why it divides confessions rather than resolving, and why the honest verdict is contested, not sound or weak. Two narrower cautions attach to the proponent’s use of the argument: the covenantal reading of Colossians 2:11–12 and Acts 2:39 is a systematic inference, not the sense the earliest exegetes drew from those verses, so presenting either as plainly commanding infant baptism overreads it (the load-bearing work is done by the continuity framework, not by the texts); and the argument establishes at most that the covenant sign belongs to believers’ children if the covenant is administered genealogically — it does not by itself settle that prior question. Where a proponent treats the circumcision parallel as by itself decisive, that overreach is flagged on Calvin’s page. See also the related covenantal proof from the New Testament household baptisms (assessment: contested).

Sources