Home Injective stability for odd-dimensional unitary K1
Article Publicly Available

Injective stability for odd-dimensional unitary K1

  • Weibo Yu EMAIL logo
Published/Copyright: November 20, 2019

Abstract

In this paper, under the usual stable range condition, a decomposition theorem for the elementary subgroup is obtained, and the injective stability theorem for odd-dimensional unitary K1 is proved.

1 Introduction

Stability results for classical K1, beginning with Bass [5] and Vaserstein [10] for linear K1, Bak and Tang [2] for Hermitian K1, Bak, Petrov and Tang [3] and Tang [9] for quadratic K1, and the author [11] for odd unitary K1, have been available in the literature for a long time and play an important role in the theory of classical groups. The stability theorems in [3, 2, 9, 11] are all proved under the Λ-stable range condition, which is weaker than its predecessors. The difficulty in the proof of stability results is injective stability, which is always concerned with a decomposition of the elementary subgroup. Does injective stability hold under the usual stable range condition which is weaker than the Λ-stable range condition? In [8], Sinchuk proved injective stability for unitary K1 under the usual stable range condition. In [1], Ambily proved injective stability for K1 of Roy’s group under the usual stable range condition over a commutative ring.

Recently, Bak and Preusser [4] introduced odd-dimensional unitary groups, which are isomorphic to Petrov’s odd hyperbolic unitary groups [7] such that V0=R and contain as special cases the groups GL2n+1(R), where R is any ring, the classical groups O2n+1(R) and Sp2n+1(R), where R is any commutative ring, and further all even-dimensional unitary groups U2n(R,Λ), where (R,Λ) is any form ring (see [4, Example 15]). In this paper, using the usual stable range condition, we prove the injective stability theorem for odd-dimensional unitary K1, which generalizes similar results given in previous papers.

Let R be an associative ring with identity and (R,Δ) a Hermitian form ring. Denote by U2n+1(R,Δ) the odd-dimensional unitary group and by EU2n+1(R,Δ) its elementary subgroup. Under certain assumptions on R, EU2n+1(R,Δ) is normal in U2n+1(R,Δ) (see [4, Theorem 39], [7, Theorem 4], [11, Corollary 3.3], [12, Theorem 4.4]). Set

KU1,n(R,Δ):=U2n+1(R,Δ)/EU2n+1(R,Δ).

In this paper, we consider this object as a coset space. Under the embedding

ϕ2n+12n+3:U2n+1(R,Δ)U2n+3(R,Δ),σ(1000σ0001),

we have the stabilization map KU1,n(R,Δ)KU1,n+1(R,Δ). Then we can get the following injective stability theorem.

Theorem 1.1.

Let nsr(R)+2. Then the map

KU1,n-1(R,Δ)KU1,n(R,Δ)

is injective.

The rest of the paper is organized as follows. In Section 2, we recall the necessary notation. In Section 3, we recall in detail the definitions of the odd-dimensional unitary group U2n+1(R,Δ), elementary subgroup EU2n+1(R,Δ), and the usual stable range condition. In Section 4, we establish a decomposition theorem for EU2n+1(R,Δ) and prove our main result Theorem 1.1.

2 Notation

In this paper, denote by the set of all positive integers. If m,n, the set of all m×n matrices with entries in the ring R is denoted by Mm×n(R). Let Aij denote the entry at position (i,j) if AMm×n(R). Denote by At the transpose of A, that is, (At)ij=Aji. Denote the i-th row of A by Ai* and the j-th column of A by A*j. Set Mn(R):=Mn×n(R). The identity matrix in Mn(R) is denoted by e or En. If AMn(R) is invertible, the entry of A-1 at position (i,j) is denoted by Aij. Further, we denote by Rn the set of all columns (a1,,an)t with entries in R.

Let G be a group and H a subset of G. Denote by H the subgroup of G generated by H. If x,yG, let xy:=y-1xy and [x,y]:=xyx-1y-1. Set

Hx:={hxhH}.

3 Odd-dimensional unitary groups

3.1 Hermitian form ring

Our concept closely follows [4]. Let R be an associative ring with identity 1. Recall that an anti-isomorphism ¯:RR, aa¯ on R is a bijective map satisfying a+b¯=a¯+b¯, ab¯=b¯a¯ for all a,bR and 1¯=1. Let λR with the property that a¯¯=λaλ¯ for all aR. It is easy to see that λ¯=λ-1. The element λ is called a symmetry for ¯ and the triple (R,¯,λ) a ring with involution and symmetry. A Hermitian ring is a quadruple (R,¯,λ,μ), where (R,¯,λ) is a ring with involution and symmetry and μR such that μ=μ¯λ.

Remark 3.1.

(1) Define the map ¯:RR, aa¯:=λ¯a¯λ. It is the inverse map of ¯. One checks easily that (R,¯,λ¯,μ¯) is a Hermitian ring.

(2) For any n, (Mn(R),,*λe,μe) is a Hermitian ring, where

A*:=A¯t=((A¯ij)ij)t=(A¯ji)ijfor anyAMn(R).

Let R be the underlying set of the ring equipped with the multiplication of the ring, but not the addition of the ring. A (right) R-module is a not necessarily abelian group (G,) equipped with a map :G×RG, (x,a)xa, satisfying

  1. x0=0 for all xG,

  2. x1=x for all xG,

  3. (xa)b=x(ab) for all xG and a,bR,

  4. (xy)a=(xa)(ya) for all x,yG and aR.

We treat as an operator with higher priority than . An R-module is canonically an R-module, but not conversely. Let G and G be R-modules. By a homomorphism of R-modules, one means a group homomorphism f:GG such that f(xa)=f(x)a for all xG and aR. By an R-submoduleN, we mean a subgroup N of G which is -stable, that is, xaN for all xN and aR.

Let (R,¯,λ,μ) be a Hermitian ring. Consider the Heisenberg group𝔜, where 𝔜=R×R as a set, with a composition operation “”, satisfying

(a1,b1)(a2,b2)=(a1+a2,b1+b2-a¯1μa2).

It is easy to see that the law is associative with identity (0,0), and the inverse is given by

-˙(a,b)=(-a,-b-a¯μa)for any(a,b)𝔜.

Furthermore, supply 𝔜 with a right action of R through (a,b)c=(ac,c¯bc). 𝔜 becomes an R-module.

Let (R,+) have the R-module structure defined by ab=b¯ab. Define the trace maptr:𝔜R by tr((a,b))=a¯μa+b+b¯λ. It is easy to see that tr is a homomorphism of R-modules.

Set

Δmin:={(0,a-a¯λ)aR}andΔmax:=ker(tr).

An odd form parameter of (R,¯,λ,μ) is an R-submodule Δ of 𝔜 such that ΔminΔΔmax. The pair ((R,¯,λ,μ),Δ) is called a Hermitian form ring and is abbreviated by (R,Δ).

3.2 Odd-dimensional unitary groups

Let (R,Δ) be a Hermitian form ring and n. Consider the right R-module M:=R2n+1 with the fixed basis (e1,,en,e0,e-n,,e-1), where ei is the column vector whose i-th entry (1i-1) is 1 and whose other entries are 0. If uM, we call (u1,,un,u-n,,u-1)tR2n the hyperbolic part of u and denote it by uhb. Define the maps

h:M×MR,(u,v)u*(00p0μ0pλ00)v=i=1nu¯iv-i+u¯0μv0+i=-n-1u¯iλv-i,

and

q:M𝔜,u(q1(u),q2(u)):=(u0,uhb*(0p00)uhb)=(u0,i=1nu¯iu-i),

where u*=u¯t, uhb*=u¯hbt and pMn(R) is the matrix with 1 on the skew diagonal and 0 elsewhere. One checks easily the following lemma.

Lemma 3.2.

The following statements hold.

  1. h is a λ -Hermitian form, that is,

    1. h is biadditive,

    2. h(ua,vb)=a¯h(u,v)b for all u,vM and a,bR,

    3. h(u,v)=h(v,u)¯λ for all u,vM.

  2. q has the following properties:

    1. q(ua)=q(u)a for all uM and aR,

    2. q(u+v)q(u)q(v)(0,h(u,v))modΔmin for all u,vM,

    3. tr(q(u))=h(u,u) for all uM.

The odd-dimensional unitary groupU2n+1(R,Δ) is the group consisting of all elements σGL2n+1(R) satisfying

h(σu,σv)=h(u,v),
q(σu)q(u)modΔ

for all u,vM. These groups are isomorphic to Petrov’s odd hyperbolic unitary groups such that V0=R and contain as special cases the groups GL2n+1(R), where R is any ring, the classical groups O2n+1(R) and Sp2n+1(R), where R is any commutative ring, and further all even-dimensional unitary groups U2n(R,Λ), where (R,Λ) is any form ring.

Set

Θ:={1,,n,0,-n,,-1},Θhb:=Θ{0}.

Let ϵ(i) be 1 for i=1,,n and -1 for i=-n,,-1.

Lemma 3.3.

Let σGL2n+1(R). Then σU2n+1(R,Δ) if and only if conditions (1) and (2) below hold.

  1. We have

    σij=λ-(ϵ(i)+1)/2σ¯-j,-iλ(ϵ(j)+1)/2for alli,jΘhb,
    μσ0j=σ¯-j,0λ(ϵ(j)+1)/2for alljΘhb,
    σi0=λ-(ϵ(i)+1)/2σ¯0,-iμfor alliΘhb,
    μσ00=σ¯00μ.

  2. q(σ*j)(δ0j,0)modΔ for all jΘ.

Proof.

See [4, Lemma 17]. ∎

Define the embedding Φ:GLn(R)U2n+1(R,Δ) by

Φ(σ)=(σ0001000p(σ*)-1p).

3.3 The elementary subgroup

Let (R,¯,λ,μ) be a Hermitian ring. We defined an R-module structure on 𝔜. Let (R,¯,λ¯,μ¯) be the Hermitian ring defined in Remark 3.1 (1), and let 𝔜¯ be the Heisenberg group corresponding to (R,¯,λ¯,μ¯). Since the underlying set of both (R,¯,λ,μ) and (R,¯,λ¯,μ¯) is R×R, we get another R-module structure on R×R if we replace the Hermitian ring (R,¯,λ,μ) by (R,¯,λ¯,μ¯). We denote the group operation (resp. scalar multiplication) defined by (R,¯,λ,μ) and (R,¯,λ¯,μ¯) on R×R by 1 and -1 (resp. 1 and -1), respectively. Set

Δ1:=ΔandΔ-1:={(a,b)R×R(a,b¯)Δ}.

It is easy to see that ((R,¯,λ¯,μ¯),Δ-1) is a Hermitian form ring.

If i,jΘ, let eij denote the matrix in M2n+1(R) with 1 in the (i,j)-th position and 0 elsewhere. If i,jΘhb, i±j and aR, the element

Tij(a):=e+aeij-λ(ϵ(j)-1)/2a¯λ(1-ϵ(i))/2e-j,-iofU2n+1(R,Δ)

is called an (elementary) short root matrix. If iΘhb and (a,b)Δ-ϵ(i), the element

Ti(a,b):=e+ae0,-i-λ-(1+ϵ(i))/2a¯μei0+bei,-iofU2n+1(R,Δ)

is called an (elementary) extra short root matrix. The extra short root matrices of the kind Ti(0,b)=e+bei,-i are called (elementary) long root matrices. An element of U2n+1(R,Δ) is called elementary matrix if it is a short or extra short root matrix. The elementary subgroupEU2n+1(R,Δ) is the subgroup of U2n+1(R,Δ) generated by all elementary matrices. The coset space

KU1,n(R,Δ):=U2n+1(R,Δ)/EU2n+1(R,Δ)

is called the odd-dimensional unitary K1-functor. Under the embedding

ϕ2n+12n+3:U2n+1(R,Δ)U2n+3(R,Δ),σ(1000σ0001),

it is easy to see that ϕ2n+12n+3(EU2n+1(R,Δ))EU2n+3(R,Δ). Therefore, we have the stabilization map KU1,n(R,Δ)KU1,n+1(R,Δ).

The elementary matrices satisfy the following properties:

(S1)Tij(a)=T-j,-i(-λ(ϵ(j)-1)/2a¯λ(1-ϵ(i))/2),
(S2)Tij(a)Tij(b)=Tij(a+b),
(S3)[Tij(a),Tkl(b)]=eifk-i,jandli,-j,
(S4)[Tij(a),Tjk(b)]=Tik(ab)ifk±i,
(S5)[Tij(a),Tj,-i(b)]=Ti(0,ab-λ-(1+ϵ(i))/2b¯a¯λ(1-ϵ(i))/2),
(E1)Ti(a1,b1)Ti(a2,b2)=Ti((a1,b1)-ϵ(i)(a2,b2)),
(E2)[Ti(a1,b1),Tj(a2,b2)]=Ti,-j(-λ-(1+ϵ(i))/2a¯1μa2)ifj±i,
(E3)[Ti(a1,b1),Ti(a2,b2)]=Ti(0,-λ-(1+ϵ(i))/2(a¯1μa2-a¯2μa1)),
(SE1)[Tij(a),Tk(b,c)]=eifkj,-i,
[Tij(a),Tj(b,c)]=Tj,-i(cλ(ϵ(j)-1)/2a¯λ(1-ϵ(i))/2)
  Ti(bλ(ϵ(j)-1)/2a¯λ(1-ϵ(i))/2,
(SE2)    acλ(ϵ(j)-1)/2a¯λ(1-ϵ(i))/2).

Let i,jΘhb such that i±j. In the following, the usual ordering of the numbers -n,,-1,0,1,,n is used. Denote by Tij the subgroup consisting of all Tij(a) with aR and by Ti the subgroup consisting of all Ti(a,b) with (a,b)Δ-ϵ(i). Let n3. We define the following subgroups of EU2n+1(R,Δ) which will be used to establish a decomposition theorem in next section.

Un=Tij,Tij<0<i,
Un-=Tij,Tii<0<j,
Ln=Tiji,j>0,
Fn=Tij,Tii-1,j1,
Y+=Tn,Tn-1,Tn-1,-nUn,
Y-=T-n,T-(n-1),T-(n-1),nUn-,
V+=Ti,n,Ti,n-1i=1,,n-2Ln,
V-=Tn,i,Tn-1,ii=1,,n-2Ln.

Denote by the semidirect product. Set

U+=UnV+,U-=Un-V-,Gn=UnLn.

3.4 Stable range condition

Let R be an associative ring with 1. A vector (a1,,an)t with aiR is called (left) unimodular if there exist elements b1,,bnR such that

b1a1++bnan=1.

The stable range conditionSn says that, for every unimodular vector

(a1,,an+1)tRn+1,

there exist elements b1,,bnR such that

(a1+b1an+1,,an+bnan+1)tRn

is unimodular. The stable rank of R, denoted by sr(R), is the smallest number n such that Sn holds.

Lemma 3.4 ([8, Lemma 9]).

Let m<n and sr(R)n-m. Let v be a unimodular column of height n. Then there exists a matrix BM(n-m)×m(R) such that (En-m,B)vRn-m is unimodular.

4 Injective stability

In this section, we get a decomposition theorem for EU2n+1(R,Δ) and prove an injective stability theorem for odd-dimensional unitary K1 under stable range condition.

Lemma 4.1.

Let n3. Then EU2n+1(R,Δ) is generated by Gn and Y-.

Proof.

Denote by H the subgroup of EU2n+1(R,Δ) generated by Gn and Y-. We only need to prove that Un-H. Let i=1,,n-1, j=1,,n-2, aR and (b,c)Δ. It follows from (S1), (S4) and (SE2) that

T-n,j(a)=[T-n,n-1(a),Tn-1,j(1)]H,
T-j,i(a)=[T-j,-n(1),T-n,i(a)]H,
T-j(b,c)=T-n,j-1(c)[Tnj(-1),T-n(b,c)]H.

Hence H contains all the generators of EU2n+1(R,Δ). ∎

Let H1,,Hs be arbitrary subsets of a group G. Define the Minkowski product H1Hs:={h1hshiHi,i=1,,s}.

Lemma 4.2.

Let n3. We have the following inclusions:

  1. Y-U+U+Y-Y+,

  2. Y+U-U-Y+Y-,

  3. Y-U+U-U+U-Y+Y-.

Proof.

(i) Let A=TnTn-1Tn-1,-n and B be the union of all Tij and Ti, where i{1,,n-2} and either j<0 or j{n-1,n}. Clearly, we have U+=AB and Y+=A. If aA and bB, then ab=[a,b]ba.

Case 1: aTn. If bTij, it follows from (SE1) that [a,b]=e. If bTi, it follows from (E2) that [a,b]Ti,-n.

Case 2: aTn-1. If bTij, it follows from (SE1) that [a,b]=e. If bTi, it follows from (E2) that [a,b]Ti,-(n-1).

Case 3: aTn-1,-n. Assume that bTij. If j<0, it follows from (S3) that [a,b]=e. If j=n-1, it follows from (S4) that [a,b]Ti,-n. If j=n, it follows from (S4) that [a,b]Ti,-(n-1). Assume that bTi. It follows from (SE1) that [a,b]=e. Hence ab=ba for some bB and U+=BA=BY+.

Let α=βγU+, where βB and γY+. Since σβσ-1 lies in U+ for any σY-, we have σα=σβσ-1σγU+Y-Y+.

(ii) Similar proof as (i).

(iii) Follows from (i) and (ii). ∎

Set S:={αLnαn-1,1=αn1=0}.

Lemma 4.3.

Let nsr(R)+2. Then, for every αLn, there exist βαV+ and γαV- such that γαβααS.

Proof.

Let αLn, and let v be the first column of α. It follows from Lemma 3.4 that there exists a matrix BM(n-2)×2(R) such that the first n-2 coordinates of

v=Φ((En-2B0E2))v

form a unimodular column. Now, there exists another matrix CM2×(n-2)(R) such that both (n-1)-th and n-th coordinates of

Φ((En-20CE2))v

are 0. Let

βα=Φ((En-2B0E2))V+,γα=Φ((En-20CE2))V-.

Then we have γαβααS. ∎

Corollary 4.4.

Let nsr(R)+2. Then UnUn-LnU+U-S.

Proof.

Let αLn. Since βα normalizes Un-, it follows from Lemma 4.3 that

UnUn-α=(Unβα-1)(βαUn-βα-1γα-1)(γαβαα)U+U-S.

Lemma 4.5.

Let nsr(R)+2. Then Y-U+U-SU+U-LnFn.

Proof.

Let αS and β(Y-)α. Since (Y-)αUn-, we have

β=(En00A10CBEn)

for some AM1×n(R), BMn×1(R) and CMn×n(R). By the definition of S, the first column of α remains unchanged if we multiply α on the left by an element of Y-. Hence the first column of β coincides with that of the identity matrix. Then A11=0 and C*1=0. It follows from Lemma 3.3 that Cn*=0 and Bn1=0. Therefore, we have βUn-FnFn. Since (Y+)αFn, by Lemma 4.2, we have

Y-U+U-αU+U-Y+Y-α=U+U-α(Y+)α(Y-)αU+U-LnFn.

Lemma 4.6.

Let σU2n+1(R,Δ) have the form

(En00A10CBEn),

where AM1×n(R), BMn×1(R), CMn×n(R). Then σEU2n+1(R,Δ).

Proof.

σ-1 has the form

(En00-A10BA-C-BEn).

Let i<0<j. It follows from Lemma 3.3 that

Bn-j+1,1=-A¯1jμ,Bn+i+1,1A1j-Cn+i+1,j=C¯n-j+1,-iλ

and (A1j,Cn-j+1,j)(0,0)modΔ. Then (A1j,Cn-j+1,j)Δ and

σ=j=1nTj(A1j,Cn-j+1,j)j-i<n+1Tij(Cn-j+1,-i).

Hence σEU2n+1(R,Δ). ∎

Now we have the following decomposition theorem.

Theorem 4.7 (Decomposition theorem).

Let nsr(R)+2. Then every element of EU2n+1(R,Δ) has a GnUn-Fn-decomposition, that is,

EU2n+1(R,Δ)=GnUn-Fn.

Proof.

It suffices to prove that GnUn-Fn is stable under left multiplications by the generators of EU2n+1(R,Δ). By Lemma 4.1, it is enough to show that

Y-GnUn-FnGnUn-Fn.

Since Ln normalizes Un and Un-, we have

GnUn-Fn=UnLnUn-Fn=UnUn-LnFn.

It follows from Corollary 4.4 and Lemma 4.5 that

Y-GnUn-Fn=Y-UnUn-LnFnY-U+U-SFnU+U-LnFnGnUn-Fn.

Proof of Theorem 1.1.

Let

σU2(n-1)+1(R,Δ)EU2n+1(R,Δ).

By Theorem 4.7, we have σ=αβγ, where

α=(A11A12A1301A2300A33)Gn,β=(En00B2110B31B32En)Un-,γFn.

Since the first columns of both σ and γ coincide with that of the identity matrix, we get A33(B31)*1=0 and B1121+A23(B31)*1=0. Since A33 is invertible, we have (B31)*1=0 and B1121=0. Hence βFn. Thus we can assume that σ=αγ. Since the first columns of both σ and γ coincide with that of the identity matrix, we get that (A11)*1 equals the first column of the identity matrix and (A33)n* equals the last row of the identity matrix. Hence we can express α as α=α1α2, where α1U2(n-1)+1(R,Δ) and α2Fn. Thus we can assume that σ=αγ with α,γU2(n-1)+1(R,Δ).

Since

α=(EnA12A13(A33)-101A23(A33)-100En)(A110001000A33),

where the second matrix lies in Ln, we get that A11En(R). Let A11=(100A*11), we have A*11En(R)GLn-1(R). Since nsr(R)+2, it follows from the stability of the K1-functor under the stable range condition Sn [6, Theorem 5.4.2] that A*11En-1(R). Then αEU2(n-1)+1(R,Δ).

Since γFn, we can express γ as

γ=(1000C0001)(1C12C130E2(n-1)+1C23001),

where CEU2(n-1)+1(R,Δ). It follows from γU2(n-1)+1(R,Δ) that

C12=0,C13=0andC23=0.

Then γEU2(n-1)+1(R,Δ). Hence σEU2(n-1)+1(R,Δ). ∎


Communicated by John S. Wilson


Award Identifier / Grant number: 11501047

Award Identifier / Grant number: 300102129106

Funding statement: This work is supported by National Science Foundation of China (No. 11501047) and the Fundamental Research Funds for the Central Universities, CHD (No. 300102129106).

Acknowledgements

The author is grateful to the referee for constructive suggestions for improvements. The author would like to thank Prof. Guoping Tang for helpful discussions.

References

[1] A. A. Ambily, Normality and K1-stability of Roy’s elementary orthogonal group, J. Algebra 424 (2015), 522–539. 10.1016/j.jalgebra.2014.07.036Search in Google Scholar

[2] A. Bak and G. Tang, Stability for Hermitian K1, J. Pure Appl. Algebra 150 (2000), no. 2, 109–121. 10.1016/S0022-4049(99)00035-3Search in Google Scholar

[3] A. Bak, V. Petrov and G. Tang, Stability for quadratic K1, J. K-Theory 30 (2003), 1–11. 10.1023/B:KTHE.0000015340.00470.a9Search in Google Scholar

[4] A. Bak and R. Preusser, The E-normal structure of odd dimensional unitary groups, J. Pure Appl. Algebra 222 (2018), no. 9, 2823–2880. 10.1016/j.jpaa.2017.11.002Search in Google Scholar

[5] H. Bass, K-theory and stable algebra, Inst. Hautes Études Sci. Publ. Math. (1964), no. 22, 5–60. 10.1007/BF02684689Search in Google Scholar

[6] H. Bass, Algebraic K-theory, W. A. Benjamin, New York, 1968. Search in Google Scholar

[7] V. A. Petrov, Odd unitary groups, J. Math. Sci. 130 (2005), 4752–4766. 10.1007/s10958-005-0372-zSearch in Google Scholar

[8] S. Sinchuk, Injective stability for unitary K1, revisited, J. K-Theory 11 (2013), no. 2, 233–242. 10.1017/is013001028jkt211Search in Google Scholar

[9] G. P. Tang, Unitary K1-groups under Λ-stable range condition (in Chinese), Chinese Ann. Math. Ser. A 25 (2004), no. 2, 171–178; translation in Chinese J. Contemp. Math. 25 (2004), 143–152. Search in Google Scholar

[10] L. N. Vaseršteĭn, On the stabilization of the general linear group over a ring, Math. USSR-Sb. 8 (1969), 383–400. 10.1070/SM1969v008n03ABEH001279Search in Google Scholar

[11] W. Yu, Stability for odd unitary K1 under the Λ-stable range condition, J. Pure Appl. Algebra 217 (2013), no. 5, 886–891. 10.1016/j.jpaa.2012.09.003Search in Google Scholar

[12] W. Yu and G. Tang, Nilpotency of odd unitary K1-functor, Comm. Algebra 44 (2016), no. 8, 3422–3453. 10.1080/00927872.2015.1085543Search in Google Scholar

Received: 2019-07-04
Revised: 2019-10-13
Published Online: 2019-11-20
Published in Print: 2020-03-01

© 2020 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 25.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/jgth-2019-0099/html
Scroll to top button