Cotangent Complex: Derived Differentials

• The cotangent complex is the derived module replacing classical Kähler differentials, capturing both first-order data and higher deformation characteristics.
• It is constructed via simplicial resolutions or cofibrant replacements, enabling precise analysis of smooth, singular, and étale cases.
• Its applications span algebraic geometry, operadic algebras, and enriched categories, making it essential for understanding deformation theory and obstruction phenomena.

The cotangent complex is the derived replacement for the module of Kähler differentials attached to a morphism of rings, schemes, operads, enriched categories, or other homotopical algebraic objects. In its classical form it is the total left derived functor of $\Omega^1$; in model-categorical and $\infty$-categorical formulations it is the suspension spectrum of the identity object in the relevant tangent category. Its degree-zero homology recovers ordinary differentials, while its higher homology and cohomology encode derived derivations, square-zero extensions, deformation classes, and obstruction theory (Conde-Lago et al., 2022, Vezzosi, 2010, Harpaz et al., 2016).

1. Classical construction and derived refinement

For a morphism of rings $A \to B$, the cotangent complex $L_{B/A}$ is constructed by resolving $B$ simplicially by polynomial $A$-algebras and applying the two-term construction

$$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$

where $R \to B$ is a surjection with kernel $I$. Equivalently, $L_{B/A}$ is the left derived functor of the module of Kähler differentials. This is the formulation emphasized both in the classical simplicial-algebraic account and in the logarithmic generalization (Conde-Lago et al., 2022).

In derived algebraic geometry, the same construction is performed in a homotopical category of cdga’s or simplicial commutative algebras by choosing a cofibrant resolution and deriving $\infty$0. For a morphism $\infty$1 of derived schemes, the local affine complexes glue to a quasi-coherent complex $\infty$2. This derived cotangent complex specializes correctly in standard cases: if $\infty$3 is smooth, then $\infty$4 in degree $\infty$5; if $\infty$6 is a regular closed immersion, then $\infty$7; and if $\infty$8 is étale, then $\infty$9 (Vezzosi, 2010).

A common shorthand identifies the cotangent complex with “derived differentials,” but the derived enhancement is essential precisely because $A \to B$0 is generally insufficient in singular or non-smooth situations. The point of $A \to B$1 is that $A \to B$2, while the higher $A \to B$3 retain deformation-theoretic information that disappears on passage to ordinary differentials (Conde-Lago et al., 2022, Vezzosi, 2010).

2. Universal property, tangent categories, and deformation theory

In Quillen’s abstract framework, if $A \to B$4 is a left proper combinatorial model category and $A \to B$5, the tangent model category at $A \to B$6 is obtained by stabilizing the slice or over-under category. The cotangent complex is then defined as the suspension spectrum of the identity map,

$A \to B$7

and the relative cotangent complex of $A \to B$8 is the homotopy cofiber of the induced map between these suspension spectra. This realizes the cotangent complex as the universal stable derived invariant at the point $A \to B$9 (Harpaz et al., 2016).

The universal property appears concretely in derivations and square-zero extensions. In the classical algebraic setting one has

$L_{B/A}$0

and the same object classifies square-zero extensions of $L_{B/A}$1 by $L_{B/A}$2. In derived algebraic geometry, Vezzosi’s formulation makes the higher $L_{B/A}$3-groups geometric: for a classical $L_{B/A}$4-point $L_{B/A}$5, the groups $L_{B/A}$6 are represented by maps from the derived $L_{B/A}$7-th infinitesimal disk $L_{B/A}$8 into $L_{B/A}$9 (Rasekh et al., 2020, Vezzosi, 2010).

The cotangent complex also organizes deformation-obstruction sequences. For a composite $B$0 there is a transitivity triangle

$B$1

and in a homotopy-Cartesian square there is a base-change equivalence

$B$2

Kodaira–Spencer and obstruction classes arise from this formalism: the obstruction to lifting across a square-zero extension lies in an $B$3-group of the relevant cotangent complex, while smoothness forces those higher obstructions to vanish (Vezzosi, 2010, Conde-Lago et al., 2022).

3. André–Quillen homology, Quillen cohomology, and rigidity

For a homomorphism of commutative noetherian rings $B$4, locally of finite flat dimension, the cotangent complex $B$5 yields André–Quillen homology by

$B$6

Its zeroth homology is $B$7, and in the surjective case $B$8 one has $B$9 and $A$0 for $A$1 (Briggs et al., 2020).

In the abstract tangent-category setting, Quillen cohomology is obtained by mapping out of the cotangent complex. For $A$2,

$A$3

For operadic algebras in stable contexts this becomes the familiar formula

$A$4

so Quillen cohomology is literally represented by the cotangent complex (Harpaz et al., 2016, Harpaz et al., 2016).

A major structural result is the rigidity theorem of Briggs–Iyengar: if $A$5 is locally of finite flat dimension and

$A$6

then $A$7 for all $A$8 and $A$9 is locally complete intersection. This sharpens Avramov’s earlier asymptotic vanishing criterion and confirms a conjecture of Quillen. The proof uses the universal Atiyah class and an equivariance theorem relating $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$0 to Hochschild cohomology (Briggs et al., 2020).

This rigidity result clarifies a frequent misunderstanding: higher André–Quillen groups are not merely auxiliary invariants. Their vanishing pattern detects whether the map is locally complete intersection, so the amplitude and homology of the cotangent complex carry decisive singularity-theoretic content (Briggs et al., 2020).

4. Operadic algebras and cotangent complexes over operads

For a $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$1-algebra $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$2 in a suitable symmetric-monoidal model category $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$3, Harpaz–Nuiten–Prasma identify the tangent category at $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$4 with operadic $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$5-modules after stabilization. More precisely, under hypotheses such as differentiability, left properness, combinatoriality, and suitable admissibility of $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$6, the stabilized tangent category $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$7 is Quillen equivalent to the stabilization of $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$8, and in the stable strictly pointed case to the ordinary model category of $$\Omega^1_{R/A}\otimes_R B \xleftarrow{\,d\,} I/I^2,$$9-modules $R \to B$0 itself (Harpaz et al., 2016).

In that setting, the cotangent complex of $R \to B$1 is the image of the suspended identity object $R \to B$2 under the equivalence

$R \to B$3

The same paper gives an operadic $R \to B$4-ary approximation: in good cases there is a cofiber sequence

$R \to B$5

whose cofiber is the usual topological André–Quillen complex in the one-colored case (Harpaz et al., 2016).

Millès provides an explicit bar-cobar model for the cotangent complex of a $R \to B$6-algebra: $R \to B$7 In the associative case this identifies the cotangent complex with the classical Koszul complex; in the commutative case it yields the familiar symmetric-exterior resolution; and in the Lie case it recovers the Chevalley–Eilenberg complex. The slogan “the Koszul complex is the cotangent complex” is therefore literal in the operadic framework (Milles, 2010).

At the level of dg operads themselves, the tangent category at a $R \to B$8-cofibrant colored dg operad $R \to B$9 is Quillen equivalent to the category of infinitesimal $I$0-bimodules. Under this equivalence the cotangent complex corresponds to $I$1, where $I$2. Special cases are especially concrete: for a dg $I$3-operad the cotangent complex is represented by the Pirashvili functor $I$4, while for a dg $I$5-operad it is related to operadic Hochschild cohomology through a cofiber sequence and controls first-order deformations via Quillen cohomology (Harpaz et al., 7 Feb 2026).

5. Enriched categories, spectral categories, and $I$6-ring spectra

The cotangent-complex formalism extends from algebras to enriched and spectral categories. For a spectral category $I$7, the absolute cotangent complex is defined as the derived indecomposables of the augmentation ideal,

$I$8

and one has a homotopy-cofiber sequence of $I$9-bimodules

$L_{B/A}$0

Relative cotangent complexes, topological derivation homology, and topological derivation cohomology are defined from the same object, and these invariants are invariant under Dwyer–Kan equivalence, hence descend to small stable $L_{B/A}$1-categories (Campbell, 2015).

For enriched categories, Harpaz–Nuiten–Prasma show that if $L_{B/A}$2 is an enriched category with object set $L_{B/A}$3, then the tangent category at $L_{B/A}$4 is equivalent to the model category of enriched $L_{B/A}$5-bimodules, identified as

$L_{B/A}$6

In the $L_{B/A}$7-categorical limit, the tangent $L_{B/A}$8-category at $L_{B/A}$9 in $\infty$00 is equivalent to

$\infty$01

and the cotangent complex is described as the image of $\infty$02 under this equivalence (Harpaz et al., 2016). A parallel description identifies the cotangent complex of an enriched category as a spectrum-valued functor on the twisted arrow $\infty$03-category; Quillen cohomology becomes the homotopy groups of a total homotopy limit over $\infty$04 (Harpaz et al., 2016).

For $\infty$05-ring spectra, several a priori different constructions of the cotangent complex coincide. The Goodwillie-stabilization definition, the sequential-colimit formula, and Basterra’s model-categorical topological André–Quillen construction all agree in $\infty$06 for a map $\infty$07 of $\infty$08-rings (Rasekh et al., 2020). This setting also admits explicit computations: if $\infty$09 is the Thom $\infty$10-$\infty$11-algebra associated to an $\infty$12-map $\infty$13, then

$\infty$14

Moreover, if $\infty$15 is étale or more generally solid, then $\infty$16, so vanishing of the cotangent complex detects formal étaleness in the $\infty$17-setting (Rasekh et al., 2020).

A further operadic refinement expresses both Quillen and Hochschild theories of an $\infty$18-algebra by spectrum-valued functors on the operadic twisted arrow $\infty$19-category $\infty$20. In this picture the cotangent complex of the operad itself becomes a functor $\infty$21 with

$\infty$22

for an $\infty$23-ary operation $\infty$24, and Quillen cohomology of an $\infty$25-algebra $\infty$26 with coefficients in $\infty$27 is computed as a mapping object from $\infty$28 to the corresponding endomorphism functor $\infty$29 (Hoang, 2023).

6. Logarithmic, valuation-theoretic, and Poisson-enhanced variants

The logarithmic cotangent complex $\infty$30 is defined for a morphism of prelog rings by resolving both the underlying rings and monoids by free cofibrations and deriving logarithmic differentials. Its degree-zero homology is the module of logarithmic differentials

$\infty$31

and it satisfies transitivity and functoriality analogous to the ordinary cotangent complex (Conde-Lago et al., 2022).

Logarithmic geometry modifies the low-degree shape of the theory. Conde-Lago and Majadas construct a logarithmic Lichtenbaum–Schlessinger complex: a $\infty$32-term complex concentrated in degrees $\infty$33 that computes $\infty$34 for $\infty$35. They also construct a logarithmic analogue of Quillen’s fundamental spectral sequence. In contrast with the classical case, the log spectral sequence has only one nonzero row at the $\infty$36-page, together with error terms $\infty$37 measuring failure of flatness of the monoid map; when the log structures are trivial, the construction collapses to the classical cotangent complex (Conde-Lago et al., 2022).

For valued fields, the cotangent complex admits an explicit non-Noetherian description. For a finite or purely transcendental extension of real valued fields $\infty$38, the integral cotangent and log cotangent complexes are constructed from a MacLane–Vaquié chain approximating the induced semi-valuation. The resulting complexes satisfy

$\infty$39

and the higher homology obeys

$\infty$40

with $\infty$41 torsion-free and vanishing precisely when $\infty$42 is separable. The same analysis yields explicit formulas for the different, log different, discrepancy, Kähler norm, and weight norm in terms of augmentation steps of the chain (Maex, 2 Apr 2026).

The cotangent complex also acquires additional algebraic structure in Poisson and singular settings. If $\infty$43 is a Poisson algebra with $\infty$44 a Poisson ideal, then $\infty$45 carries an $\infty$46-algebroid structure induced from a $\infty$47-structure on a semifree resolution. In low degrees the bracket extends the usual Koszul bracket on Kähler differentials, and in hypersurface or suitable complete-intersection cases the higher operations collapse so that one recovers a dg Lie algebroid. This demonstrates that the cotangent complex can serve not only as a derived replacement for $\infty$48 but also as the natural receptacle for higher Poisson or Lie–Rinehart structures on singular spaces (Herbig et al., 2021).

Across these variants, the same pattern persists: the cotangent complex is the universal derived receptacle for differentials, derivations, and square-zero extensions, but its concrete form depends strongly on the ambient geometry. In smooth, étale, or solid situations it collapses to degree $\infty$49 or vanishes; in singular, logarithmic, operadic, enriched, and spectral contexts it retains higher structure that governs deformation theory, exactness properties, and cohomological invariants (Vezzosi, 2010, Rasekh et al., 2020, Conde-Lago et al., 2022).