The entanglement of radicals

Chi Wa Chan, Antigona Pajaziti, Flavio Perissinotto and Antonella Perucca · 2025

Read The entanglement of radicals by Chi Wa Chan, Antigona Pajaziti, Flavio Perissinotto and Antonella Perucca (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-19211.txt.

Introduction

Chi Wa Chan, Antigona Pajaziti, Flavio Perissinotto, and Antonella Perucca settle a question on the entanglement of radicals. Their paper is from two thousand twenty-five. Radicals are roots of numbers or of field items, such as square or cube roots. The question is how many plain sum links can hold among a set of them.

They show that, over any field, such links are very few. Their result wraps up a classic theorem from the nineteen seventies on when radicals stand free of each other in sums. It also solves a problem raised years later by Lenstra. The work gives a full picture of when roots can be tied by a sum.

Catalog record

CREATORChi Wa Chan, Antigona Pajaziti, Flavio Perissinotto and Antonella Perucca
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-19211

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The entanglement of radicals Chi Wa Chan, Antigona Pajaziti, Flavio Perissinotto and Antonella Perucca Abstract . In this work we achieve a full understanding of the so-called entanglement of radicals, showing that — over any field — there are extremely few additive relations among radicals. Our results complete a famous theorem by Kneser from 1975 on the linear independence of radicals and solve a problem discussed by Lenstra in 2006. 2000 Mathematics Subject Classification Primary: 11R20; Secondary: 11R18. 1. Introduction Let K be a field (for which we fix an algebraic closure \overline{K} ) and consider a multiplicative group G of radicals of K , that is a group generated by K^{\times} and by elements in \overline{K} that have some power in K^{\times} . Clearly, the multiplicative relations among the radicals in G are encoded in the group structure. We are interested in the additive relations among the radicals in G (also called entanglement ) that become relevant when we consider the field K(G) . To study the additive relations among radicals we may suppose without loss of generality that the index |G:K^{\times}| is finite. Then we can “measure” the additive relations by comparing this index and the degree of the extension K(G)/K . Indeed, for radicals that are dependent (in the sense that they give rise to additive relations) the degree [K(G):K] is smaller than the index |G:K^{\times}| . Roots of unity are radicals, and K -linear relations among them constitute one first type of entanglement, which we call cyclotomic entanglement . The basic relations are the following: if \zeta_{n} is a root of unity of order n , then we have 1+\zeta_{n}+\zeta_{n}^{2}+\cdots+\zeta_{n}^{n-1}=0\,. Over \mathbb{Q} the above relations (and those generated by them) are all the additive relations among roots of unity, but there are more relations involving \zeta_{n} for a field K such that the degree of the cyclotomic extension K(\zeta_{n})/K is less than \varphi(n) . For example, if K=\mathbb{Q}(\sqrt{5})\subset\mathbb{C} and \zeta_{5}=e^{2\pi i/5} , then we have the K -linear relation \sqrt{5}\cdot 1-\zeta_{5}+\zeta_{5}^{2}+\zeta_{5}^{3}-\zeta_{5}^{4}=0\,. We remark that for a number field K there are only finitely many K -linear relations among roots of unity which generate all additive relations: this is because there is a constant c_{K} such that the intersection of K with \mathbb{Q}(\zeta_{\infty}) (the largest cyclotomic extension of \mathbb{Q} ) is contained in \mathbb{Q}(\zeta_{c_{K}}) . Thus the cyclotomic entanglement is rather harmless and measuring it amounts to determining the degree of K\cap\mathbb{Q}(\zeta_{\infty}) . There can be further entangled radicals , for example the above relation \sqrt{5}=\zeta_{5}-\zeta_{5}^{2}-\zeta_{5}^{3}+\zeta_{5}^{4} is a \mathbb{Q} -linear relation for the radical \sqrt{5} involving roots of unity. More generally, all square-roots of rational numbers are contained in a cyclotomic extension of \mathbb{Q} . To generate the corresponding entanglement we may take \sqrt{2}=\zeta_{8}+\zeta_{8}^{7} and for any odd prime number p express \sqrt{p} as a \mathbb{Q} -linear combination of 4p -th roots of unity with a Gauss sum that – with appropriate root choices – can be written as follows: \sqrt{p}=(-1)^{(p-1)/4}\cdot\sum_{i=1}^{p-1}\left(\frac{i}{p}\right)\zeta_{p}^{i}\,. Over a field K different than \mathbb{Q} there could be more entanglement of this type, which we call Kummer entanglement , because there can be further radicals that are contained in a cyclotomic extension of K . For example, over K=\mathbb{Q}(\sqrt{5}) the square root of -\frac{5+\sqrt{5}}{8} is contained in \mathbb{Q}(\zeta_{5}) . For the Kummer entanglement, the entangled radicals generate abelian radical extensions of K and we can invoke Schinzel’s Theorem on abelian radical extensions (Theorem 6 ): this entanglement is due to Kummer extensions of K that are contained in cyclotomic extensions and hence it is well-understood. Over \mathbb{Q} , an entanglement that is neither cyclotomic nor Kummer is given by the following \mathbb{Q} -linear relation (with the appropriate root choices): \sqrt[4]{-4}=1+\zeta_{4}\,. This entanglement relation is due to the decomposition in ( 8 ), which in turn stems from the non-cyclicity of the extension \mathbb{Q}(\zeta_{8})/\mathbb{Q} . In fact, the relations that we presented completely describe the entanglement over \mathbb{Q} (this is also a special case of our results below). In this work we solve the problem of describing the entanglement over any field. As explained by Lenstra in [ 8 ] , beyond the theoretical interest, the understanding of entanglement is crucial for a designer of a computer algebra system who wishes to do computations with radicals e.g. over number fields or function fields (namely, it is the main mathematical obstacle that such a designer needs to overcome). In this work, we achieve a full understanding of the entangled radicals by building on a classical result by Kneser (Theorem 4 ). Rather surprisingly, the entanglement is as limited as possible. In a nutshell, for a general field there are no surprises with respect to \mathbb{Q} , and there may just be one element (of the form 1+\zeta_{2^{w}} ) that plays the role that \sqrt[4]{-4} plays for \mathbb{Q} . Our very general results are presented in the next section. 2. The main results We denote as customary the roots of unity, and we let \mathrm{char}(K) be the characteristic of K . We fix a prime number \ell\neq\mathrm{char}(K) and denote by \sqrt[\ell^{n}]{K^{\times}} the subgroup of \overline{K}^{\times} consisting of the elements whose \ell^{n} -th power is in K^{\times} (and define \sqrt[\ell^{\infty}]{K^{\times}} as the union of \sqrt[\ell^{n}]{K^{\times}} for n\geqslant 0 ). Similarly, we write \zeta_{\ell^{\infty}} to mean all roots of unity whose order is a power of \ell . Moreover, we call K(\zeta_{2\mathcal{P}}) the extension of K that is generated by the roots of unity whose order is 4 or an odd prime number and it is not divisible by \mathrm{char}(K) . Our results show the remarkable fact that the additive relations of radicals are extremely few. This is because by a famous result by Kneser (Theorem 4 ) there are no more additive relations over the field K(\zeta_{2\mathcal{P}}) . Theorem 1 . Suppose that \ell is odd or that \zeta_{4}\in K . Let t\geqslant 1 be the largest integer such that \zeta_{\ell^{t}}\in K(\zeta_{2\mathcal{P}}) , or set t=\infty if no such largest integer exists. If \zeta_{\ell}\notin K , then we have (1) K(\zeta_{2\mathcal{P}})\cap\sqrt[\ell^{\infty}]{K^{\times}}=\langle\zeta_{\ell^{t}},K^{\times}\rangle\qquad\text{and}\quad\qquad K(\zeta_{2\mathcal{P}})\cap K(\!\sqrt[\ell^{\infty}]{K^{\times}})=K(\zeta_{\ell^{t}})\,. If there is some largest integer w>0 such that \zeta_{\ell^{w}}\in K , we have (2) K(\zeta_{2\mathcal{P}})\cap K(\sqrt[\ell^{\infty}]{K^{\times}})=K(\zeta_{\ell^{t}},K(\zeta_{2\mathcal{P}})\cap\sqrt[\ell^{w}]{K^{\times}}) and (3) K(\zeta_{2\mathcal{P}})\cap\sqrt[\ell^{\infty}]{K^{\times}}=\langle\zeta_{\ell^{t}},K(\zeta_{2\mathcal{P}})\cap\sqrt[\ell^{w}]{K^{\times}}\rangle\,. If t\leqslant 2w , then the field K(\zeta_{2\mathcal{P}})\cap K(\!\sqrt[\ell^{\infty}]{K^{\times}})=K(\zeta_{2\mathcal{P}})\cap K(\!\sqrt[\ell^{w}]{K^{\times}}) is the largest subextension of K(\zeta_{2\mathcal{P}})/K that is Kummer and with exponent a power of \ell . Notice that the case \langle\zeta_{\ell^{\infty}}\rangle\subseteq K is not dealt with because setting w=\infty in the above formulas would result in a trivial assertion. We denote by \sqrt{K^{\times}} the subgroup of \overline{K}^{\times} consisting of the elements whose square is in K^{\times} , noticing that K(\sqrt{K^{\times}})/K is a Kummer extension. Theorem 2 . Suppose that \ell=2 and \zeta_{4}\notin K . If \langle\zeta_{2^{\infty}}\rangle\subseteq K(\zeta_{2\mathcal{P}}) , then we have (4) K(\zeta_{2\mathcal{P}})\cap\sqrt[2^{\infty}]{K^{\times}}=\langle\zeta_{2^{\infty}},K(\zeta_{2\mathcal{P}})\cap\sqrt{K^{\times}}\rangle and (5) K(\zeta_{2\mathcal{P}})\cap K(\sqrt[2^{\infty}]{K^{\times}})=K(\zeta_{2^{\infty}},K(\zeta_{2\mathcal{P}})\cap\sqrt{K^{\times}})\,. Else, call w\geqslant 2 the largest integer such that \zeta_{2^{w}}+\zeta_{2^{w}}^{-1}\in K and let t\geqslant w be the largest integer such that \zeta_{2^{t}}\in K(\zeta_{2\mathcal{P}}) . Then we have (6) K(\zeta_{2\mathcal{P}})\cap K(\sqrt[2^{\infty}]{K^{\times}})=K(\zeta_{2^{t}},K(\zeta_{2\mathcal{P}})\cap\sqrt{K^{\times}}) and (7) K(\zeta_{2\mathcal{P}})\cap\sqrt[2^{\infty}]{K^{\times}}=\langle\zeta_{2^{t}},1+\zeta_{2^{w}},K(\zeta_{2\mathcal{P}})\cap\sqrt{K^{\times}}\rangle\,. If t>w , we may omit 1+\zeta_{2^{w}} from the list of generators. The further results of this paper are described in Section 4 . We don’t consider general radicals in \sqrt[\infty]{K} but only radicals in \sqrt[\ell^{\infty}]{K} where \ell is prime. This is sufficient for

Read the complete text