Two models of Kitchen Cleanliness

and much else

You pour out some cashews for a snack, spill some of them on the kitchen floor, which has not been washed today but appears clean. Do you pick them up and eat them or throw them out and get more out of the bag? The answer depends on which of two mental models you are running.

One model is chemistry, taking those precautions and only those necessary to protect you from eating something that will injure you. The amount of anything the cashews could have picked up from contact with an apparently clean floor is tiny, perhaps a microgram. There is nothing in your kitchen toxic at that small a dose¹ so you eat the cashews.

The other model is contagion. The kitchen floor is unclean, the cashews touched it so they are unclean. You should not eat things that are unclean so you throw the cashews in the trash and pour some more out of the bag.

In that situation chemistry is a more appropriate model, but not always. Any quantity of a bacterium or virus is a potential source of contagion since it can, in a suitable environment, reproduce. The appropriate precaution is to avoid letting anything touch your food that has been polluted by touching something that might hold the disease organism; if something is polluted, purify it. Carried further, make sure that nothing that touches anything that touches your food has touched anything that might. Carried farther still …

Interestingly enough, that looks very much like the attitude to pollution in cultures such as Romani or Orthodox Judaism.² Certain things are unclean, contact with them makes other things unclean. Pollution is contagious.

Explaining Pollution

For most of the time humans have existed they had no easy way of distinguishing effects of disease from effects of poisoning. Under those circumstances intuiting uncleanness as pollution made a lot of sense. Some features of ritual pollution fit that pattern; in Rabbinic law, something that a corpse had touched was polluted. In Romaniya, the Romany legal system, the lower half of the human body is unclean. Maimonides, a doctor as well as a famous scholar of Jewish law, generalized the idea from religion to medicine, adopted a policy of hand washing more than six hundred years before Semmelweis.³

Unfortunately the practice did not catch on.

A model of cleanliness that reduced disease mortality might well, over a period of hundreds of thousands of years, get hardcoded into the human brain by evolution as one way of thinking about risks. It is then available to apply to things that can be classified as clean/dirty, good/bad, pure/polluted, sometimes appropriately, sometimes not.

Beyond the Kitchen: Cancel Culture and Reds

Smith holds unacceptable political opinions, is a Trumpist/alt Right/Nazi. Jones had Smith on his podcast, so Jones is a Nazi too. Johnson mentioned Jones favorably on his blog, so Johnson too is a Nazi. Cancel all of them.

The pattern is older than cancel culture —it was also, on the opposite side of the political spectrum, Senator McCarthy’s approach to identifying reds.

Patterns?

Is there a pattern to which people use which model? Is there a pattern to what things which model is used for? If patterns exist can we explain them?

When I discussed the idea of this post with other people, mostly to check that the idea of the two models made sense to people other than me, two of them separately reported that, by their observation, women tended to apply the contagion model, men chemistry. One of them suggested a poll to test the conjecture — the cashew question would work — but I have not actually done so.

One reason women might be more inclined to the contagion model is that women do most of the job of taking care of babies and small children; prior to modern times, a lot of both died of contagious diseases.

Binary Categories

The contagion model uses binary categories: clean/unclean, sterile/infected. Some issues fit that better than others. With no way of distinguishing between an object with lots of viruses or bacteria and an object with only a few, the binary category makes sense. It also makes sense because, in many contexts, the distinction does not matter; few can become many in an hour or a day. The chemistry model, in contrast, treats variables as continuous; dose matters.

If the contagion model is built into our mental machinery, the fact that it requires binary categories may help to explain the prevalence of binary categories in our thinking and our institutions — even when they are not appropriate.

An example in our legal system is guilt/innocence. In a tort case, at least in legal theory, if the jury believes that the probability of guilt is 51% the defendant owes the full amount of the damage, if 49% he owes nothing. That pattern is echoed in accounting rules where an expense such as a tort judgement deemed more likely than not to occur counts as a liability at its full value, one deemed less likely than not to occur as a liability of zero.

Arguments over abortion show the problems with imposing a binary category on a continuous variable. Few of those who believe abortion should be legal also support legalized infanticide. Few, but not none, of those who believe abortion should be illegal are opposed to birth control, even birth control that prevents a fertilized ovum from implanting, and nobody I am aware of expects a pregnant woman to take the same level of precautions against accidental miscarriage that she would take to protect a living infant. Yet the transition from fertilized ovum to infant is a continuous one.

Tribalism — red/blue, conservative/liberal, Republican/Democrat, believer/heathen — is arguably another example. There is no good reason why someone opposed to legal abortion should also be against gun control or someone in favor of legal abortion in favor of it, but if political or ideological categories are binary …

On The Other Hand

One great advantage of binary categories, hence of the pollution model, is simplicity. If all a jury has to decide is whether the defendant is probably guilty, exact calculations of probability are only necessary for close cases. Returning a probabilistic verdict in order to scale the damage judgement to the probability of guilt would require it in every case. Deciding whether something you dropped on the ground has picked up enough dirt that you shouldn’t eat it, in harder cases than dry nuts dropped on a clean kitchen floor, requires more thought than a judgement of “unclean.” Judging a neighbor as friend or foe in whatever political war is currently running is easier if his opinion on one disputed issue predicts his views on most others.

The case for simplicity over precision is particularly strong for internal discipline, developing and enforcing habits. It is easier for me to commit myself to spending two hours a day, seven days a week, on writing projects, as I did a good many years ago, or not permitting myself ice cream if my morning weight is over 160, than it would have been to commit myself to more flexible, hence more nearly optimal, policies.

---

Past posts, sorted by topic

A search bar for past posts and much of my other writing

1. Unless there happens to be some Botulism toxin lying around.

2. See Chapters 2 and 4 of Legal Systems Very Different from Ours.

3. “The little-known history of cleanliness and the forgotten pioneers of handwashing” by Peter Poczai and László Z. Karvalics.