Home > Incompatible Food Triad

The problem is this: Find three foods, such that any two of them go together, but all three do not.

We understand "go together" in any reasonable sense of the expression, as it is ordinarily applied to foods.

One way of seeing the problem is this: given three foods that don't go together, it's usually because two of them don't go together. For example, Richard Feynman's famous example of accidentally requesting milk and lemon in his tea is not a solution: (1) Milk, tea, and lemon do not go together. (2)(a) Tea and lemon do go together, (b) Tea and Milk do go together, but (c) Milk and lemon do not go together. For the solution to work milk and lemon would have to go together as well. Most attempted solutions (so far, according to Hart) tend to overlook one of the three pairs.

The problem can also be formulated thus: Find a counter example to either of the following alleged theorems (where R(x,y,...) means "x, y, ... all go together") :

(1) Given any three foods A, B, and C, if [R(A,B), R(A,C) and R(B,C)] then R(A,B,C)

(2) Given any three foods A, B, and C, if ~R(A,B,C) then [~R(A,B) or ~R(A,C) or ~R(B,C)].

Two decent solutions have in fact been proposed on George Hart's page:

(1) Salted cucumbers, sugar, yogurt.

(2) Orange juice, gin, tonic.

It seems highly likely that whatever one person doesn't like, there exists someone else out there in the world who does like it. If this is accepted as a means of eliminating a solution, then it becomes futile to discover solutions: the whole world cannot be polled. Therefore it would be good to have a clarification of the puzzle such that if it holds for one person, it has been solved.

It also seems appropriate that since any sub-solution must taste good, the combination of all three elements tastes bad out of excess more than anything else.