Revisiting Hugo Volger’s paper Über die Existenz der freien Algebren

Matías Menni and Walter Tholen · 2024

Read Revisiting Hugo Volger’s paper Über die Existenz der freien Algebren by Matías Menni and Walter Tholen (2024) 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-2412-20286.txt.

Description

Abstract We give a modern account of Volger’s 1967 paper which, motivated by the construction of free algebras for a Lawvere-Linton theory, gives a very constructive proof that the left Kan extension of a product-preserving 𝖲𝖾𝗍 \mathsf{Set} -valued functor is product-preserving. We also analyze how it anticipates, and in part even exceeds, subsequent work of the 1970s.

Catalog record

CREATORMatías Menni and Walter Tholen
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-20286

Browse the full library · About the Tower of Babel

Full text

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

Revisiting Hugo Volger’s paper Über die Existenz der freien Algebren Matías Menni and Walter Tholen Abstract We give a modern account of Volger’s 1967 paper which, motivated by the construction of free algebras for a Lawvere-Linton theory, gives a very constructive proof that the left Kan extension of a product-preserving 𝖲𝖾𝗍 \mathsf{Set} -valued functor is product-preserving. We also analyze how it anticipates, and in part even exceeds, subsequent work of the 1970s. Keywords: Algebraic theory, free algebra, Kan extension, final functor, weakly cofiltered category. Mathematics Classification: 18C10, 08B20, 18A40, 18D20. 1 Introduction In the reprint of his celebrated thesis [ 11 ] (published in part in [ 12 , 13 ] ), as part of the Author’s Comments , Lawvere writes in reference to Chapter 4 on Algebraic Functors (see [ 14 , Pages 19/20] ): The calculus of algebraic functors and their adjoints is at least as important in practice as the algebraic categories themselves. Thus it is unfortunate that there was no indication of the fundamental fact that these functors preserve reflexive coequalizers (of course their preservation of filtered colimits has always been implicit). As remarked above, “the algebra engendered by a prealgebra” is a special case of an adjoint to a (generalized) algebraic functor. However, contrary to what might be suggested by the treatment in this chapter, the use of that reflection is not a necessary supplement to the use of Kan extensions in proving the general existence of algebraic adjoints: as remarked only later by Michel André, Jean Bénabou, Hugo Volger, and others, the special exactness properties of the background category of sets imply that the left Kan extension, along any morphism of algebraic theories, of any algebra in sets, is already itself again product-preserving. Clearly, the same sort of thing holds, for example, with any topos as background. Especially the mentioning of Volger’s name made us curious. Although the paper is not listed in the References of [ 14 ] , we suppose that Lawvere is alluding to Volger’s free-algebra construction in the paper [ 28 ] , which was submitted to Mathematische Zeitschrift in May of 1967; its German title translates to About the existence of free algebras . Rather than proving their mere existence (as may easily be done by Adjoint Functor Theorem methods [ 16 ] ), Volger gives a rather concrete construction of free algebras, for theories that allow infinitary operations up to a regular cardinal. Nevertheless, the paper is rarely mentioned anywhere in the subsequent literature pertaining to Lawvere’s algebraic theories, including the books or monographs that treat Lawvere theories in a signifcant manner, such as [ 20 , 21 ] or [ 22 , 1 ] . Schubert’s book [ 24 ] and its German precursor [ 23 ] list Volger’s 1967 Diplomarbeit (Master’s thesis) at the University of Freiburg, titled Kategorien von Algebren über algebraischen Theorien ( Categories of algebras over algebraic theories ), on which we presume Volger’s paper to be based (we have no access to his thesis). It is however not apparent to which extent Schubert used Volger’s work in his proofs. We found only two articles which directly expand on Volger’s paper, namely [ 25 ] and [ 8 ] , but like Volger’s paper, they hardly get cited in the subsequent literature. (In fact, they did not turn up for us with the search function of Zentralblatt for papers referencing Volger’s paper!) We believe, however, that both, Volger’s paper and the Howlett-Schumacher paper [ 8 ] contain important steps toward Borceux’s later work with Day [ 3 , 5 , 6 ] , which is duly acknowledged in the literature as a pillar of the enriched treatment of algebraic theories and their algebras; see, for example, the often cited paper [ 10 ] , or the more recent [ 9 , 18 ] . In addition, unlike most later works, Volger’s construction covers algebras for theories with infinitary operations. The language barrier and the sometimes cumbersome notation in Volger’s paper may have contributed to its limited resonance in the literature and the likely small number of its actual readers. The goal of this article is to make the achievements of Volger’s paper easily accessible to a wider audience and to demonstrate their importance. In Section 2 we prove a generalized version of a key technical lemma in Volger’s paper. This will help us present his result in a simpler form. In Section 3 we present a rather “liberal” and shortened translation of Volger’s paper in which, while following the original structure of the paper, we use in some instances more modern terminology and simplify some of the original lengthy proofs, thus using methods that have become common place only after the appearance of Volger’s paper, especially with the publication of Mac Lane’s book [ 16 ] . In order to preserve the flow of the original text, these deviations from it and our improvements are indicated and explained only in a sequence of footnotes listed at the end of the paper. Then, in Section 4 , we briefly compare Volger’s result with the presentation of free algebras in Lawvere’s thesis, and analyze on one hand to which extent it is, in conjunction with the Howlett-Schumacher paper [ 8 ] , a true forerunner to Borceux’s work with Day, and on the other hand, when applied to \mathsf{Set} -based algebras, to which extent it goes beyond these papers since his result includes the consideration of infinitary theories (in the sense of Linton [ 15 ] ). The common underlying problem considered in the papers cited above may be formulated as follows: Given functors F:\mathcal{B}\to\mathcal{C} and X:\mathcal{B}\to\mathcal{E} of locally small categories, for \mathcal{B} small with finite products and X preserving them, and for \mathcal{E} (small-)cocomplete, which additional conditions are needed to guarantee that then also the left Kan extension \mathrm{Lan}_{F}X:\mathcal{C}\to\mathcal{E} of X along F preserves (any existing) finite products in \mathcal{C} ? The starting point for the work by Borceux and Day is their elegant short argument that, when \mathcal{E} is Cartesian closed, no additional condition is needed for an affirmaitive answer to the problem, while Howlett and Schumacher require that F should preserve finite products. Their argumentation adopts parts of Volger’s auxiliary proofs which produce an affirmative answer in case \mathcal{E} is the category \mathsf{Set} and \mathcal{B} has also pullbacks, in addition to finite products. But in our rendering of Volger’s proofs we indicate how the use of the finite-product preservation by F may easily be avoided. Furthermore, we show that the pullback assumption in Volger’s paper, which is relevant only when one moves from finite products to products of size below an infinite regular cardinal, may be weakened considerably without losing an affirmative answer of the stated problem in case \mathcal{E}=\mathsf{Set} , but at the expense of some additional restriction on the functor X ; for details, see Theorem 4.2 . 2 A generalization of a lemma due to Volger In this section we prove a generalization of the key [ 28 , Lemma 5] ; a result that, today, may be seen as about the calculation of colimits for certain special functors whose categories of elements are weakly cofiltered. 2.1 Colimits We

Read the complete text