---
title: Inverted Habiro Series Analysis
url: https://www.emergentmind.com/topics/inverted-habiro-series
type: topic
---

# Inverted Habiro Series Analysis

Inverted Habiro series denotes a family of constructions that reverse, extend, or analytically dualize the usual Habiro cyclotomic expansion. In the most literal usage, the usual finite products in Habiro’s basis are replaced by reciprocal products after extending the cyclotomic coefficients to negative indices, producing a series naturally expanded near \(x=0\) or \(x=\infty\) and adapted to the two-variable knot-complement invariant \(F_K(x,q)\) and to homological blocks [2106.03942]. In adjacent usages, the phrase refers to a coefficient-enlarged Habiro ring in which \(\Delta_K(t)\) is inverted [2603.01619], to coefficient-recovery mechanisms at roots of unity that reconstruct ADO data from Habiro coefficients [2009.13285], and to resurgence-theoretic reconstructions of analytic structure from formal series extracted from Habiro elements via strange identities [2304.07001]. This suggests that the term designates a thematic direction rather than a single universally fixed definition.

## 1. Classical Habiro framework and the sense of “inversion”

The classical Habiro ring is
\[
\mathcal H=\varprojlim_n \mathbb Z[q]/((q)_n), \qquad (q)_n=(q;q)_n=\prod_{k=1}^n(1-q^k),
\]
and its elements admit expansions
\[
F(q)=\sum_{n=0}^\infty A_n(q)\,(q)_n, \qquad A_n(q)\in \mathbb Z[q].
\]
For knots, Habiro’s cyclotomic expansion expresses the colored Jones polynomial in a basis built from cyclotomic factors. One convenient form is
\[
J_K(x,q)=\sum_{n\ge 0} a_n(K;q)\,\Omega_n(x,q), \qquad 
\Omega_n(x,q)=\prod_{i=1}^n\bigl(x+x^{-1}-q^i-q^{-i}\bigr),
\]
with the specialization \(x=q^N\) recovering the \(N\)-colored Jones polynomial [2009.13285].

The inversion idea first appears as a change of expansion regime. In the standard cyclotomic picture, the series is completed at a finite place, centered at \(x+x^{-1}=2\). The inverted picture replaces the factors \(\Omega_n(x,q)\) by their reciprocals and treats the resulting object as a power series near \(x+x^{-1}=\infty\), equivalently as a series in \(x\) or \(x^{-1}\). In the classical limit \(q\to 1\), the usual Habiro series lives in \(\mathbb Z[[x+x^{-1}-2]]\), while the inverted one lives in \(\mathbb Q[[1/(x+x^{-1}-2)]]\) [2106.03942].

A second, algebraically distinct sense of inversion enlarges the coefficient ring rather than replacing factors by reciprocals. For a knot with Alexander polynomial \(\Delta_K(t)\), one passes from \(\mathbb Z[t^{\pm1}]\) to
\[
R=\mathbb Z[t^{\pm1},\Delta_K(t)^{-1}],
\]
and then forms a relative Habiro completion over this étale algebra. In that setting, “inverted” refers to the fact that \(\Delta_K(t)\) is inverted in the coefficient ring, not to reciprocal cyclotomic factors [2603.01619].

## 2. Reciprocal cyclotomic expansions and knot-complement series

The most direct definition of an inverted Habiro series is the reciprocal cyclotomic expansion proposed for the knot-complement invariant \(F_K(x,q)\). In the formulation based on negative Habiro coefficients,
\[
F_K(x,q)=-(x^{\frac12}-x^{-\frac12}) \sum_{m=1}^{\infty}\frac{a_{-m}(K)}{\prod_{j=0}^{m-1}(x+x^{-1}-q^j-q^{-j})},
\]
and the coefficients \(a_{-m}(K;q)\) are related explicitly to the ordinary \(x\)-power-series coefficients of \(F_K\) by \(q\)-binomial transforms [2106.03942]. In Park’s later normalization, the same object is written as
\[
F_K(x,q)= \sum_{k=1}^{\infty} \frac{a_{-k}(K;q)}{\prod_{j=0}^{k-1}(x+x^{-1}-q^j-q^{-j})},
\]
or equivalently
\[
F_K(x,q) = \sum_{k=0}^{\infty} a_{-k-1}(K;q)\,
\frac{(-1)^k q^{\frac12 k(k+1)}}{(x;q)_{k+1}(x^{-1};q)_{k+1}},
\]
which is the form used to identify \(F_K\) with a balanced homological block [2603.05236].

This inversion principle is closely tied to the Gukov–Manolescu conjecture: the Melvin–Morton–Rozansky expansion should resum to a two-variable \(q\)-series \(F_K(x,q)\). An inverted version of the \(R\)-matrix state sum proves this conjecture for a large class of links, including all homogeneous braid links and all fibered knots up to 10 crossings. In that construction, inversion on the state-sum side parallels inversion on the Habiro side: negative summation indices are introduced through lowest-weight as well as highest-weight Verma modules, and the resulting series is annihilated by the quantum \(A\)-polynomial [2106.03942].

For the figure-eight knot \(4_1\), the inverted form is especially simple:
\[
F_{4_1}(x,q) = \sum_{m=0}^{\infty} \frac{1}{\prod_{j=0}^{m}(x+x^{-1}-q^j-q^{-j})}.
\]
The same example exhibits a structural feature emphasized in the 3d–3d literature: the classical annihilator contains an abelian factor \(y-1\), and the contour selecting poles associated with the abelian branch produces the homological block. By contrast, a different contour choice in the same half-index integral yields the colored Jones polynomial after the specialization \(x=q^n\). Thus the same integral expression encodes both the inverted Habiro series and the usual Jones data, with the distinction implemented by the pole prescription [2603.05236].

The trefoil examples show the same pattern. For the left-handed and right-handed trefoils, the papers write explicit inverted Habiro series, explicit half-index integrals, and explicit contour choices separating the homological block, the non-abelian branch, and the colored Jones polynomial. The abelian branch is present in the classical annihilator, but it is not recovered from the twisted superpotential alone; it is instead isolated by the contour relevant to the half-index [2603.05236].

## 3. Coefficients, truncation, multiplication, and residues

A central algebraic feature of inverted Habiro theory is that the negative-index coefficients are not merely auxiliary. For the Gukov–Manolescu series
\[
F_K(x,q)=\sum_{n\ge 0} f_n(q)x^n,
\]
the associated inverted Habiro series is
\[
P_K(x,q)= (x^{-1}-1)\sum_{k\ge 0}\frac{a_{-k-1}(q)}{\prod_{i=0}^{k}(x+x^{-1}-q^i-q^{-i})},
\]
and the Taylor expansion of \(P_K(x,q)\) at \(x=0\) is required to agree with the GM series. This gives a bijection between the systems \(\{a_{-k-1}\}\) and \(\{f_i\}\) [2509.22610].

The relation is explicit:
\[
f_i = \sum_{k=0}^i \qbinom{k+i}{2k}\, a_{-k-1}, \qquad
a_{-k-1} = \sum_{i=0}^k (-1)^{k+i} \qbinom{2k}{k-i}\frac{[2i+1]}{[k+i+1]} f_i.
\]
Assuming the lower bound condition
\[
\delta(a_n)\ge -\frac{n(n+3)}{2}+C \qquad (n<0),
\]
the inverted series admits a closed formula in terms of the Jacobi theta function and truncated theta functions:
\[
P_K(x,q)=\frac{1}{\theta(x,q)}\sum_{i\ge 0} f_i(q)\,\theta_i(x,q).
\]
In this form, the inverted Habiro series acts as a regularization of the GM series: the truncated theta factors compensate for the negative \(q\)-degrees that may occur in the coefficients \(f_i(q)\) [2509.22610].

The negative-index basis also carries a multiplication law analogous to Habiro’s original multiplication formula. Using the basis elements \(\sigma_n\), the product is
\[
\sigma_m^0 \sigma_n^0 = \sum_{i=0}^\infty \gamma_{m,n}^i\, \sigma_{m+n-i}^0,
\]
which leads to a \(\mathbb Z((q))\)-algebra \(\Omega\) of inverted Habiro series satisfying the lower bound condition. This supplies a genuine ring structure, not just an ad hoc family of formal expansions [2509.22610].

At roots of unity, a different inversion mechanism appears in the ADO/WRT/CGP comparison for double twist knots. There, Habiro’s series
\[
J_K(x,q)=\sum_{n\ge 0} a_n(K;q)\,\Omega_n(x,q)
\]
is evaluated at \(q=\varepsilon_p\), and the ADO invariant is recovered from the truncated sum
\[
\operatorname{ADO}_K(x,\varepsilon_p)=\sum_{n=0}^{p-1} a_n(K;\varepsilon_p)\,\Omega_n(x,\varepsilon_p).
\]
The periodicity relation
\[
a_{n+kp}(K;\varepsilon_p)=a_n(K;\varepsilon_p)\,a_k(K;1)
\]
is the coefficient-recovery mechanism behind this truncation formula. For 0-surgeries on double twist knots, the difference between the CGP and WRT invariants is controlled exactly by the \((p-1)\)-st Habiro coefficient:
\[
\operatorname{CGP}(M,\lambda)=\{p\lambda\}^2\,\operatorname{WRT}(M)+p\,a_{p-1}(K(l,m);\varepsilon_p).
\]
The same paper notes that the clean periodicity property fails for torus knots, so this coefficient-extraction picture does not extend in the same simple form there [2009.13285].

Residue theory provides a further structural layer. For \(n=-k-1<0\), the basis function \(\sigma_n(x)\,dx\) has simple poles at \(x=q^j\) for \(j=0,\pm1,\dots,\pm k\), and the residues are explicit. For a full inverted Habiro series \(P(x)=\sum_{n<0} a_n \sigma_n\), the residues \(r_j\) yield a pole expansion
\[
P_K(x)=\sum_{j=-\infty}^{\infty}\frac{r_j}{x+x^{-1}-q^j-q^{-j}},
\]
and the GM Dehn surgery formula can be rewritten in terms of these residues. In this way, surgery expressions become residue computations in the inverted Habiro framework [2509.22610].

## 4. Relative and coefficient-enlarged Habiro rings

A separate but closely related development constructs a new Habiro-type ring adapted to the colored Jones polynomial as a two-variable object. For
\[
R=\mathbb{Z}[t^{\pm1},\Delta_K(t)^{-1}],
\]
the Habiro ring of the étale map \(\mathbb Z[t^{\pm1}]\to R\) is
\[
\widehat R_{\,/[t^{\pm1}]} = \varprojlim_n R[q]/(q;q)_n\,R[q],
\]
equivalently described by compatible expansions around all roots of unity with Frobenius twist built into the gluing condition. The coefficients now live in a localized Alexander ring, and the Frobenius lifts act by \(t\mapsto t^p\) for all primes \(p\) [2603.01619].

In this setting, the colored Jones polynomial is lifted from a sequence \(J_{K,n}(q)\in\mathbb Z[q^{\pm1}]\) to a single universal element
\[
J_K(t,q)\in \widehat{\mathbb{Z}[t^{\pm1},\Delta_K(t)^{-1}]/[t^{\pm1}]}
\]
satisfying
\[
J_K(q^{n-1},q)=J_{K,n}(q)\qquad (n>0).
\]
The construction begins from a \(q\)-hypergeometric state sum \(f_{A,B,\nu}(t,q)\) deforming the ordinary state sum for \(J_{K,n}(q)\). The \(q\to1\) limit is a rational function,
\[
f_{A,B,\nu}(t,1)=\frac{1}{\Delta_K(t)},
\]
which explains why \(\Delta_K(t)\) must be inverted in the coefficient ring [2603.01619].

This coefficient enlargement yields a loop expansion
\[
J^{\rm loop}_K(t,q)\in \mathbb{Z}[t^{\pm1},\Delta_K(t)^{-1}][[q-1]], \qquad
J^{\rm loop}_K(t,1)=\frac1{\Delta_K(t)},
\]
together with analogous expansions around every root of unity \(\zeta_m\). The local expansions glue by Frobenius, confirming Habiro’s conjecture that the loop expansion extends coherently to neighborhoods of all roots of unity. The universal lift is uniquely determined either by its \(q-1\) expansion or by its values at all roots of unity [2603.01619].

In this version of the subject, “inverted Habiro series” does not mean reciprocal cyclotomic factors. It means a relative or coefficient-enlarged cyclotomic completion in which the Alexander polynomial is inverted and the resulting object simultaneously encodes all color specializations and all root-of-unity expansions.

## 5. Resurgence, strange identities, and analytic reconstruction

A further dual viewpoint arises from Habiro elements associated with strange identities for partial theta series. Starting from a periodic function \(f\), one considers a partial theta series \(\theta_{a,b,f}(q)\) and a Habiro element
\[
\Phi_f(q)=\sum_{n=0}^\infty A_{n,f}(q)\,(q)_n
\]
that agrees with the partial theta series “to infinite order at roots of unity.” From the asymptotics of \(\theta_{a,b,f}\) as \(q\to1\), one extracts coefficients
\[
C_{n,f}=(-1)^nL(-2n-1,f)
\]
and forms the formal Gevrey-1 series
\[
F_f(x)=\sum_{n=0}^\infty C_{n,f}\,x^{-n}.
\]
The paper then studies the Borel transform, median resummation, and radial limits of this formal \(1/x\)-series [2304.07001].

The Borel transform is written as
\[
\mathcal B[F_f](p)=C_M+G_f(p),
\]
with
\[
G_f(p)=g_{1,f}\odot g_{2,f}.
\]
The Hadamard-product decomposition makes the singularity structure explicit: \(g_{1,f}\) has poles at a discrete set, \(g_{2,f}\) has a square-root singularity at \(p=\tfrac14\), and contour deformation yields the representation
\[
G_f(p)=\frac{M^2 b}{3\pi c}\sum_{\ell=1}^\infty \frac{f(\ell)}{\bigl(\frac{M^2}{4\ell^2}-p\bigr)^{5/2}},
\]
with the constants and branch conventions of equation (2.9). Hence \(G_f\) is endlessly analytically continuable on \(\mathbb C\setminus N\), where \(N\) is a discrete singular set on the positive real axis [2304.07001].

The resummation theorem states that the left and right Borel sums
\[
\mathcal S^\pm[F](x)=\int_{\gamma_\pm} e^{-px}\,\mathcal B[F](p)\,dp
\]
exist on sectors and define a median sum
\[
\mathcal S^{\mathrm{med}}[F](x)=\tfrac12\bigl(\mathcal S^+[F](x)+\mathcal S^-[F](x)\bigr).
\]
For \(F_f\), the median sum is analytic on \(\Re(x)>0\), has radial limits at rational points of its natural boundary, and its discontinuity across the positive real axis is expressed in terms of the original partial theta series. In particular, for \(a\in\mathbb Q\setminus\{0\}\), the value
\[
\mathcal S^{\mathrm{med}}[F_f]\!\left(-\frac{1}{2\pi i a}\right)
\]
is written explicitly via \(\theta_{0,4M^2,f}(e^{2\pi i a})\), up to a phase factor depending on the sign of \(a\). For special periodic functions \(\chi_{2st}^{(n,m)}\), this becomes the precise recovery statement of Theorem 1.4. The paper applies the framework to the torus-knot families \(T(2,2u+1)\) and \(T(3,2k)\), proving the conjectural Costin–Garoufalidis picture in those cases [2304.07001].

This is not an algebraic inversion of Habiro coefficients. It is an analytic inversion in which one passes from a Habiro element to a Gevrey-1 asymptotic series, then to its Borel plane, and finally back to radial-limit values. A related but different transformation theory studies coefficients of \(F_f(1-q)\), \(F_f\!\left(\frac1{1+q}\right)\), and \(F_f\!\left(\frac{1-q}{1+q}\right)\), proving asymptotics and sign patterns for the \(1-q\) expansion and formulating positivity conjectures for the transformed series [2204.02628].

## 6. Motivic, arithmetic, and matrix-valued generalizations

The Habiro ring also admits generalizations in which “inversion” means changing the target geometry or the arithmetic context of evaluation. In the \(F_\zeta\)-geometric framework, the Habiro ring is interpreted as a ring of analytic functions on the set of roots of unity through the evaluation maps
\[
e_\zeta:\mathbb{Z}[q]\to \mathbb{Z}[\zeta]
\]
and the injective total evaluation map into the product over all roots of unity. Habiro series are then treated as counting functions of ind-varieties built from auxiliary varieties \(X_m\) and products of punctured affine spaces. The same formalism is lifted to a motivic Habiro-Grothendieck ring by replacing \(q\) with the Lefschetz motive \(\mathbb L\) and completing with respect to the ideals generated by
\[
(\mathbb L-1)(\mathbb L^2-1)\cdots(\mathbb L^n-1).
\]
The paper further introduces formal roots \(\mathbb L^{1/n}\) and Tate motives \(Q(1/n)\), which give a fractional-power extension of the Habiro/Tate framework [1310.2261].

In this motivic setting, the closest analogues of inversion are dualization by \(\mathbb L\mapsto-\mathbb L\), interpolation at negative integers, and the passage to fractional Tate twists. These constructions do not define an inverted Habiro series in the knot-theoretic sense, but they do reverse the usual integer-power or \(q\to1\) perspective and make roots of unity intrinsic to the geometry itself [1310.2261].

A different arithmetic generalization organizes Habiro-like expansions into matrices indexed by boundary parabolic \({\rm SL}_2(\mathbb C)\)-representations. For a hyperbolic knot, the invariant
\[
\boldsymbol{\Phi}_\alpha(h)=\bigl(\Phi_\alpha^{(\sigma,\sigma')}(h)\bigr)
\]
is defined for \(\alpha\in\mathbb Q/\mathbb Z\). Its \((\sigma_0,\sigma_1)\) entry is the power-series expansion of the Kashaev invariant around the root of unity \(e^{2\pi i\alpha}\) as an element of the Habiro ring, while the remaining entries belong to generalized Habiro rings of number fields. The first column is given by the perturbative series of Dimofte–Garoufalidis, the columns form fundamental solutions of a linear \(q\)-difference equation, and the matrix defines an \({\rm SL}_2(\mathbb Z)\)-cocycle \(W_\gamma\) conjecturally extending to smooth and even holomorphic functions on suitable domains [2111.06645].

Here the “inverted” aspect is modular and sectorial rather than reciprocal: one no longer has a single Habiro series at \(q=1\), but a family of Habiro-like expansions around all cusps and across all boundary-parabolic sectors. This broadens the notion of inversion from a formal manipulation of cyclotomic factors to a reorganization of asymptotic, arithmetic, and modular data.

The literature therefore supports several rigorously distinct meanings of inverted Habiro series. The narrowest meaning is the reciprocal cyclotomic expansion for \(F_K(x,q)\) and homological blocks. Broader meanings include coefficient localization by \(\Delta_K(t)^{-1}\), coefficient recovery and truncation at roots of unity, resurgence-theoretic reconstruction from radial-limit asymptotics, and generalized Habiro expansions over motives and number fields. Across these settings, the unifying pattern is a reversal of the standard Habiro viewpoint: information usually extracted from a cyclotomic completion at finite place is reconstructed, extended, or reindexed from the opposite side.

Source: https://www.emergentmind.com/topics/inverted-habiro-series