Monoidal globular categories as a natural environment for the theory of weak \(n\)-categories (Q1269468): Difference between revisions

Category theory developed from the 1950s onwards. Fairly soon after its start four significant developments occurred: (i) To compare categories you needed functors, to compare functors you needed natural transformations, so 2-categories were introduced to study the overall setting. (ii) Similarities between phenomena in categories with an `internal hom set' (closed categories) and categories of vector spaces, lead to the study of `categories with multiplication' subsequently called `monoidal categories'. (iii) To handle situations where the 2-categorical picture was inadequate lax limits and bicategories were introduced (and the latter were shown to include monoidal categories as a special case). (iv) The final development was that of methods of handling universal algebra in a coherent categorical way using ideas such as monads, algebraic theories and operads. The interplay with homotopy theory in all of this was significant, but MacLane's coherence theorem of 1963 was also equivalent to cut elimination in logic. By 1975 the buildings block were in place; 2-categories, bi-categories, operads and coherence. Bi-categories are weak 2-categories. What about 3-categories, 4-categories, \(n\)-categories, etc. and also their weak cousins? The problem was that straight generalisation might lead to sterility. There were no immediate applications for the concepts therefore no benchmarks to say if any new attempts answered the problem. From 1984 onwards the picture changed. Connections with knot theory and physics and the development of linear logic brought monoidal categories to the fore again and Grothendieck's pursuit of stacks suggested that \(n\)-categories and \(n\)-groupoids might play a significant role in interpreting nonabelian cohomology and higher order invariants across a wide range of areas of algebra and geometry. The notion of tri-category was formulated and the coherence problem was solved in that case. This article describes one of several different ways of approaching the general definition and theory of weak \(n\)-categories. The approach is based on the theory of operads. Alternative approaches are due to \textit{J. C. Baez}, \textit{J. Dolan} and \textit{M. Neuchl} [see \textit{J. C. Baez} and \textit{M. Neuchl}, Adv. Math. 121, No. 2, 196-244 (1996; Zbl 0855.18008)], \textit{J. C. Baez} and \textit{J. Dolan}, J. Math. Phys. 36, No. 11, 6073-6105 (1995; Zbl 0863.18004), and preprints] coming from the direction of topological quantum field theory; \textit{M. Makkai}, \textit{C. Hermida} and \textit{J. Power} (preprint 1997), with a viewpoint close to logic and theoretical computer science, and \textit{Z. Tamsamani} (Ph. D. thesis, Toulouse 1995 available from QA archives) with a simplicial approach using ideas from homotopy theory. (Are they equivalent? Some progress has been made in this by \textit{Leinster} (Preprints, Cambridge, 1998).) This is an area that should grow rapidly. It seems to interact with many different parts of mathematics, physics and theoretical computer science. It should provide insight into the higher order combinatorics and higher dimensional algebra of many disparate situations. The paper itself is very well written and is a very important contribution to the foundations of this exciting new area.