On the Narrow 2-Class Field Tower of Some Real Quadratic Number Fields: Lengths Heuristics Follow-Up

Elliot Benjamin, Mohamed Mahmoud Chems-Eddin · 2026

Read On the Narrow 2-Class Field Tower of Some Real Quadratic Number Fields: Lengths Heuristics Follow-Up by Elliot Benjamin, Mohamed Mahmoud Chems-Eddin (2026) 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-2601-11773.txt.

Introduction

Elliot Benjamin and Mohamed Mahmoud Chems-Eddin work on a deep question in number theory. They study real quadratic fields, which add a square root to the rationals. For each field, a tower of class fields can be built step by step. They ask how long that tower runs.

They pick fields whose class group is a small group of order four. They also pick fields whose discriminant is not a sum of two squares. The work is a follow-up, checking past guesses on tower length. Such towers test how class groups grow in pure maths.

Catalog record

CREATORElliot Benjamin, Mohamed Mahmoud Chems-Eddin
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE97 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-11773

Browse the full library · About the Tower of Babel

Full text

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

On the Narrow 2-Class Field Tower of Some Real Quadratic Number Fields: Lengths Heuristics Follow-Up Elliot Benjamin, Mohamed Mahmoud Chems-Eddin Abstract . In this article we continue the investigation of the length of the narrow 2 2 -class field tower of real quadratic number fields k \mathrm{k} whose discriminants are not a sum of two squares and for which their 2 2 -class groups are elementary of order 4 4 . Letting G \mathrm{G} equal the Galois group of the second Hilbert narrow 2 2 -class field over k \mathrm{k} , and [ G i ] [\mathrm{G}_{i}] denote the lower central series of G \mathrm{G} , we give heuristic evidence that the length of the narrow 2 2 -class field tower of k \mathrm{k} is equal to 2 2 when G / G 3 \mathrm{G}/\mathrm{G}_{3} is of type 64.150 64.150 (in the tables of Hall and Senior). We also give the formulation of the relevant unit groups of the narrow Hilbert 2 2 -class field for these fields. 2020 Mathematics Subject Classification 11R29, 20D15. 1. Introduction In previous works, for real quadratic number fields \mathrm{k} with 2 -class group elementary of order 4 the first author named and Snyder have studied when the length of the narrow 2 -class field tower of \mathrm{k} is 2 and when it is \geq 3 (cf. [ 4 , 6 ] ). For fields whose discriminant is a sum of two squares the first author named and Snyder have completely determined criteria to distinguish between these two possibilities [ 4 ] . For fields whose discriminant is not a sum of two squares they have determined criteria to distinguish between these two possibilities for all fields k with the exception of one particular type [ 6 ] . We let \mathrm{k}^{1} denote the Hilbert 2-class field of \mathrm{k} , and \mathrm{k}^{1}_{+} denote the narrow Hilbert 2 -class field of \mathrm{k} . Letting \mathrm{k}^{2}_{+} denote the second narrow Hilbert 2 -class field of \mathrm{k} , \mathrm{G}=\mathrm{Gal}(\mathrm{k}^{2}_{+}/\mathrm{k}) , and the symbol \simeq denote ”isomorphic to,“ and making use of the lower central series [\mathrm{G}_{i}] of \mathrm{G} , this particular type of fields \mathrm{k} are fields such that \mathrm{G}/\mathrm{G}_{3}\simeq 64.150 , as listed in Hall and Senior [ 10 ] . More specifically, there are two cases of this type: (c_{3}) and (d_{1}) as listed in [ 5 , 6 ] , and they can be described as \mathrm{k}=\mathbb{Q}(\sqrt{d_{1}d_{2}d_{3}d_{4}}) where d_{j} are distinct negative prime discriminants, each divisible by the prime p_{j} such that if d_{\mathrm{k}}\equiv 4\pmod{8} then d_{4}=-4 , and \left(\dfrac{d_{1}}{p_{2}}\right)=\left(\dfrac{d_{2}}{p_{3}}\right)=\left(\dfrac{d_{3}}{p_{1}}\right)=-1\text{ and }\left(\dfrac{d_{1}}{p_{4}}\right)=\left(\dfrac{d_{2}}{p_{4}}\right)=\left(\dfrac{d_{3}}{p_{4}}\right)=1; type (d_{1}) are these fields \mathrm{k} for which d_{4}=-4 , and type (c_{3}) are these fields \mathrm{k} for which d_{4}\not=-4 . Throughout the paper, let h_{2}(m) (resp. \varepsilon_{m} ) denote the 2 -class number (resp. the fundamental unit) of a real quadratic field \mathbb{Q}(\sqrt{m}) . Moreover, let h(k) (resp. h_{2}(k) , \mathbf{C}l_{2}(k) ) denote the class number (resp. the 2 -class number, the 2 -class group) of a number field k . 2. Narrow 2 -class field tower: Results and Examples Letting \mathrm{k} be a field as described above for which \mathrm{G}/\mathrm{G}_{3}\simeq 64.150 , \mathbf{C}l_{2}(\mathrm{k}^{1}_{+}) denote the narrow 2 -class group of \mathrm{k} , and t denote the length of the narrow 2 -class field tower of \mathrm{k} , we state the following result from [ 6 ] . Lemma 2.1 . If \mathbf{C}l_{2}(\mathrm{k}^{1}_{+}) has 8 - \mathrm{rank}=0 , then \mathrm{k} has narrow 2 -class field tower length t=2 . Furthermore, the authors of [ 6 ] have determined criteria in general to distinguish between t=2 and t\geq 3 for fields \mathrm{k} as above. To describe this criteria we let \mathrm{F}_{1}=\mathbb{Q}(\sqrt{d_{1}d_{3}d_{4}}) , \mathrm{F}_{2}=\mathbb{Q}(\sqrt{d_{1}d_{2}d_{4}}) , \mathrm{F}_{3}=\mathbb{Q}(\sqrt{d_{2}d_{3}d_{4}}) and we note that these discriminants have the following C_{4} -splittings: d(\mathrm{F}_{1})=d_{1}\cdot d_{3}d_{4} , d(\mathrm{F}_{2})=d_{2}\cdot d_{1}d_{4} and d(\mathrm{F}_{3})=d_{3}\cdot d_{2}d_{4} (cf. [ 6 , 12 ] ). From the C_{4} -splittings we obtained unramified cyclic quartic extensions \mathbb{F}_{j} of \mathrm{F}_{j} (wlog j = 1, 2) (cf. [ 6 ] ). If we let \mathrm{L}_{j}=k^{1}_{+}\mathbb{F}_{j} , where \mathrm{k}^{1}_{+} is the narrow Hilbert 2 -class field of \mathrm{k} , then we have two unramified quadratic extensions of \mathrm{k}^{1}_{+} such that the composite \mathrm{L}_{1}\mathrm{L}_{2} is a V_{4} -extension of \mathrm{k}^{1}_{+} and hence contains a third unramified quadratic extension \mathrm{L}_{0} of \mathrm{k}^{1}_{+} . By the C_{4} -splittings of discriminants d(\mathrm{F}_{j}) , there are primitive rational solutions (a_{j},b_{j},c_{j}) satisfying a_{1}^{2}=d_{1}b_{1}^{2}+d_{3}d_{4}c_{1}^{2} , a_{2}^{2}=d_{2}b_{2}^{2}+d_{1}d_{4}c_{1}c_{2}^{2} , and if we let u_{1}=a_{1}+c_{1}\sqrt{d_{3}d_{4}} and u_{2}=a_{2}+c_{1}\sqrt{d_{1}d_{4}} then we see for example that u_{1}u_{1}^{\prime}=a_{1}^{2}-c_{1}^{2}d_{3}d_{4}=b_{1}^{2}d_{1}<0 and wlog we assume that u_{1}>0 and u_{1}^{\prime}<0 , and similarly for u_{2} . Then, \mathbb{F}_{j}=\mathrm{F}_{j}(\sqrt{u_{j}}) are our initial two desired extensions, and \mathrm{L}_{0}=\mathrm{k}_{+}^{1}(\sqrt{u_{1}u_{2}}) (cf. [ 6 , 12 ] ). We are now able to state our general criteria from [ 6 ] to distinguish between t=2 and t\geq 3 for fields as above. Lemma 2.2 . t=2 if and only if h_{2}(\mathrm{L}_{j})/h_{2}(\mathrm{k}_{1}^{+})=\frac{1}{2} , for j=0 , 1 , 2 . The authors of [ 6 ] gave two examples that satisfy Lemma 2.2 and therefore for which t=2 . Using PARI/GP ( https://pari.math.u-bordeaux.fr/ )(assuming GRH), we extend this to the following 12 examples in Table 1 , for which our first two examples are the ones included in [ 6 ] . In our table we list the 2 -class group of \mathrm{k}^{1}_{+} and the 2 -class groups of \mathrm{L}_{j} , j=1 , 2 , 0 denoting our 2 -class groups that are direct sums of cyclic groups of orders 2^{i} , 2^{j} , 2^{s} as (2^{i},2^{j},2^{s}) , for i\geq 1 , j\geq 1 , s\geq 1 , and our fields by the discriminant of the field. Type Field u_{1} u_{2} \mathbf{C}l_{2}(\mathrm{k}^{1}_{+}) \mathbf{C}l_{2}(\mathrm{L}_{1}) \mathbf{C}l_{2}(\mathrm{L}_{2}) \mathbf{C}l_{2}(\mathrm{L}_{0}) (c_{3}) 3\cdot 8\cdot 11\cdot 23 -1+2\sqrt{6} 5+2\sqrt{33} (2,4,4) (4,4) (4,4) (2,2,4) (d_{1}) 7\cdot 23\cdot 31\cdot 4 19+4\sqrt{23} 9+4\sqrt{7} (2,4,8) (2,4,4) (4,8) (4,8) (c_{3}) 7\cdot 3\cdot 19\cdot 131 -25+6\sqrt{21} 23+2\sqrt{133} (2,4,8) (4,8) (2,4,4) (4,8) (d_{1}) 7\cdot 31\cdot 79\cdot 4 391+44\sqrt{79} 111+20\sqrt{31} (2,4,4) (4,4) (4,4) (2,2,4) (c_{3}) 3\cdot 8\cdot 11\cdot 47 -7+4\sqrt{6} 5+\sqrt{33} (2,4,16) (4,16) (4,16) (2,4,8) (d_{1}) 7\cdot 31\cdot 127\cdot 4 45+4\sqrt{127} 423+76\sqrt{31} (2,4,8) (2,4,4) (4,8) (4,8) (c_{3}) 3\cdot 11\cdot 23\cdot 347 -29+6\sqrt{33} 83+10\sqrt{69} (2,4,16) (2,4,8) (4,16) (4,16) (d_{1}) 23\cdot 7\cdot 167\cdot 4 29+12\sqrt{7} 19+4\sqrt{23} (4,8,8) (2,8,8) (2,8,8) (4,4,8) (c_{3}) 3\cdot 8\cdot 11\cdot 71 5+\sqrt{33} 5+4\sqrt{6} (2,4,16) (2,4,16) (4,16) (4,16) (d_{1}) 47\cdot 23\cdot 103\cdot 4 27+4\sqrt{47} 37+8\sqrt{23} (2,8,32) (2,4,32) (8,32) (8,32) (c_{3}) 7\cdot 3\cdot 19\cdot 251 43+10\sqrt{21} 23+2\sqrt{133} (4,16,32) (4,16,16) (2,16,32) (2,16,32) (d_{1}) 7\cdot 103\cdot 127\cdot 4 3+4\sqrt{7} 39+4\sqrt{103} (2,8,8) (2,4,8) (8,8) (8,8) Table 1. Fields \mathrm{k} Satisfying the Criteria of Lemma 2.2 Remark 2.3 . It follows from Lemma 2.2 that for all our fields in Table 1 , we have t=2 . And we note from Table 1 that for our fields k = \mathbb{Q}(\sqrt{3\cdot 8\cdot 11\cdot 23}) and \mathrm{k}=\mathbb{Q}(\sqrt{7\cdot 31\cdot 79\cdot 4}) we have \mathbf{C}l_{2}(\mathrm{k}^{1}_{+})=(2,4,4) . Therefore \mathbf{C}l_{2}(\mathrm{k}^{1}_{+}) has 8 -rank 0 for these two fields and it also follows from Lemma 2.1 that t=2 . Based upon our examples in Table 1 and the fact that all the examples we have found satisfy the criteria of Lemma 2.2 , we make the following conjecture. Conjecture 1 . Let \mathrm{k} be a field as above, satisfying \mathrm{G}/\mathrm{G}_{3}\simeq 64.150 where \mathrm{G}=\mathrm{Gal}(\mathrm{k}^{2}_{+}/\mathrm{k}) . Then \mathrm{k} satisfies the criteria of Lemma

Read the complete text