The King Chicken Theorems

Stephen B. Maurer

Mathematics Magazine · 1980 · 65 citations · 9 references

DOIFull text

Open access

Concepts

Abstract

Fact: In any barnyard of chickens, in every pair of chickens, one is dominant over the other.The dominant chicken asserts its dominance by pecking the other chicken on the head and neck, whence the phrase "pecking order."However, it rarely happens that this pecking order is linear.That is, it is rare that there is a first chicken who pecks all the others, a second chicken who pecks all but the first, and so on.Question: Given that the pecking order may not be linear, is there still a reasonable way to designate a most dominant chicken, i.e., a king?In this paper, we analyze at length one promising king definition, and look briefly at several others.This material illustrates nicely the delights and pitfalls of applied mathematics.One delight is that a model set up to describe one situation may well describe several others as well.A pitfall is that the mathematics done with the model, though good mathematics, may not really be of much use in the situation being modeled.It is not merely a matter of irrelevant theorems.It may be a matter of the wrong mathematical assumptions or even the wrong definitions.This last is a very subtle matter because it may be far from obvious that the definitions are wrong.Specifically, we wish to define a king chicken in such a way that the definition implies a strong sense of dominance and also implies that every flock of chickens has a king.We give such a definition and such a theorem-a definition and a theorem due to the mathematical sociologist H. G. Landau in 1953 [5].But a king should also be unique, or at least rare.It is necessary to ask if our definition assures this.We are led to discover and prove many theorems which say that kings are all too common.In short, the theorems have negative implications about the usefulness of the definition.But we would not even have discovered this if we had not done some mathematics with the definition, and done it with this sort of testing in mind.This lesson, that a model needs to be judged by the theorems it keeps, is perhaps the main point of this paper.I must confess, however, that for me the theorems themselves are also a main point; the pure mathematician in me finds them interesting in their own right.Also, their proofs illustrate some interesting techniques, for instance, duality and induction from n to n + 2. In any event, good modeling does not stop with wrong definitions, no matter how pretty the theorems they give.One must look for better definitions and/or better mathematical assumptions, and then start proving theorems again.We say a bit about this second round towards the end of the paper.This paper is written primarily for undergraduates, both in the style of proofs and in the amount of interpretation of theorems.There are problems stated later which could lead to good undergraduate projects.Indeed, the paper developed from work begun with undergraduates-my 1977 Graph Theory class at Princeton.Some of the results appear to be new.

References

9