How Much Is Enough?

During the great heat waves of 1995 in Chicago and 2003 in Paris, many people, especially but not only elderly people living alone, died, and the news media in both cities were filled with discussions of what had happened, how and why it had happened, and especially whose fault it all was. The question was raised repeatedly, and continues to be raised: could more have been done to prevent these deaths, by taking greater pains, by being better prepared, by having systems in place to monitor the conditions of life for older people living alone? Most simply and practically, couldn’t air conditioning units have been distributed to those who needed them? In Sweden, as French television viewers learned, social workers visit elderly people living alone several times each day, checking on their well-being. Each social worker has something like four people to care for in this way. Why hadn’t city governments in Chicago and Paris undertaken such simple protective measures?

In another direction, why had the news media covered the story as they did, laying the blame on some city departments which failed to meet the challenges of old age in an urban setting? Later social science analyses asked why the media so quickly blamed, for instance, the culture of various ethnic groups in Chicago. It had been said, for example, that black families did not look out for old people the way Mexicans did. (The Chicago event is described in Klinenberg 2002.)

Heat waves and their attendant pathologies exemplify a larger problem: how much is enough? How much should people, organizations, cities, and nations prepare to deal with the physical, social, and human consequences of recurring (and therefore expectable) natural troubles? Thinking about a wide variety of these situations, we can use the tools of comparative analysis to study their dynamics. Their variations add to our knowledge of underlying dimensions and processes.

Consider snow storms. Large cities in the appropriate weather zones can anticipate periodically being buried in snow, which then remains on the ground, possibly disrupting ordinary life for a long time. How prepared should a large city be for a major snowfall? Every city has some stuff on hand — snowplows, people trained to use them, emergency plans — to deal with an expectable level of trouble. That is, there’s an amount of snow that has happened often enough in the past to be considered “normal” (“normal trouble,” in other words, a concept that pops up everywhere in the analysis of disasters of all kinds). Beyond that is an “abnormal” amount that occurs only once every five, ten, fifty, one hundred years (in other words, this is a variable). Of course, the preparations in place for the normal snowfall can’t handle the abnormal amount. That’s what happened to poor Michael Bilandic (who had succeeded the legendary Boss Richard Daley as mayor of Chicago) during the monumental storms of January, 1979, for which the city was totally unprepared, didn’t know what to do, didn’t have the equipment that might have made something doable — all of which led to the city, its people and its commerce being tied up for most of six weeks, and to Jane Byrne replacing Michael Bilandic in the next election for mayor of the city. (This event is described in Granger and Granger 1980, pp. 209-16, and other such meteorological and political problems are recounted in Mergen 1997, pp. 69-79.)

The normal amount of snow varies from place to place, from city to city. When I lived in Kansas City, where it snowed three or four times during the winter — but, as people always said, “it never sticks”—a three inch snowfall that stayed on the ground for three days was a major disaster. The city’s substantial hills got slippery. Drivers unused to driving in snow easily lost control of their cars and had accidents. It wasn’t very cold, so people didn’t freeze to death as they might have in New York or Chicago. But a three inch snowfall would not have been a major event in Chicago. After all, that’s what happens in winter. It might take 12 or 18 inches of snow in a few hours before Chicago has serious problems, doesn’t have enough snowplows and trucks, and has its traffic and daily routines disrupted. I was once in Montréal when it started to snow, heavily, and (Chicagoan that I was) I feared I would be stuck there for days before the airport could handle traffic again. Not at all. As soon as the snow began, fleets of plows appeared on the streets, taking the snow away as fast as it fell, so that the eighteen inches that might have paralyzed Chicago was handled as a routine event. Of course, larger snowfalls than that occur every once in a while, and they paralyze Montréal.

The same is true of earthquake preparations. San Francisco is prepared for some level of seismic activity — call it X — but not for some larger level — 2X or 3X. A television program some years ago, called “The City That Is Waiting To Die,” compared San Francisco’s earthquake readiness to that of Tokyo, where quakes, small and large, are much more common. The program concluded that if the city didn’t improve its readiness for a large tremor there would eventually be a major disaster. The 1989 quake, some years later, caused a few highway overpasses to collapse, and several people were killed, but it was just a large quake, not the catastrophic Big One often predicted. As a result of that earthquake, one of the co-owners of the San Francisco apartment building we live in insisted that our building needed $30,000 worth of additional earthquake proofing, although it is built on bedrock which did not move in the much larger 1906 tremor. Even though a much larger quake can be expected some day, possibly soon, knowledgeable architects told us that we would be wasting money. Any seismic event that could damage our building would destroy the entire city, so why save ours?

Disasters of human origin, such as the nuclear disasters at Chernobyl and Three-Mile Island, and similar events around the world, present the same problem. Diane Vaughan’s classic examination of the Challenger space flight disaster (Vaughan 1996), and her description of her later involvement with the very similar Columbia accident (Vaughan 2004), add another dimension to the analysis. Here the disaster is possible but not inevitable. The responsible parties, aware of the possibility, could have taken steps to avoid it, but didn’t. Vaughan explains this as the result of, among other things, the “normalization of deviance.” The engineers working on the Challenger space craft knew that the O-rings were not reliable in cold temperatures but, since the rings had not failed in a number of situations of cold weather, the responsible officials decided that that was an acceptable risk, even though doing so violated their own rules and procedures. The same thing occurred in the Columbia accident, where persistent problems with insulation disintegrating and damaging the craft had not made problems and so, contrary to the team’s own rules, they routinely allowed flights despite such damage. They normalized their own rule breaking and so created the conditions for the expectable accident. (Vaughan 1996, pp. 119-52)

You can’t avoid a heat wave or snow storm, though you can prepare for it. But the trouble making events in nuclear matters and in space flight are avoidable. You can follow the rules and have a better chance of avoiding an accident. Yet the two situations are only superficially different. Since the occurrence of natural disasters is foreseeable, even though we can’t say for sure when, there will eventually be a natural event — earthquake, heat wave, snow storm — that will produce terrible damage and loss of life. Knowing this, we could prepare for these foreseeable troubles, just as the flight engineers and administrators could have followed the rules that would have prevented the accident.

To generalize: For every potential disaster, there’s an expected range of dangerous outcomes, and there’s another, less-expected range (statistically minded people might call it the second standard deviation) of much greater and much rarer troubles. It’s in that context that cities and organizations have to decide what magnitude of disaster to prepare for. How big a heat wave? How high a temperature? Going down how far at night? Lasting for how many days? Do we need a special warning system for people, especially older people, who live alone? How much snow? What magnitude of tremor? How serious an accident?

But the extra preparations aren’t free. Spend the money to prepare for this disaster and you don’t have it to spend on preparations for some other kind of disaster, or for the city’s educational or health care needs. So it’s a familiar problem of tradeoffs, somewhat disguised when we say “More could have been done.” That is always true. But it doesn’t solve the political problem: what would we give up for that extra cushion of safety in the heat wave? Similarly, engineers always know how to make their machine better and safer but also know that this could lead them to never certify it as “ready to go.” (Compare Tracy Kidder’s (1981) description of a computer that survived this process and “got out the door” with Latour’s (1996) description of an advanced system of personal transportation designed for the Paris Metro that didn’t.)

We can think of this as an equation which models the relation between expenditure (X) and desired outcome (Y). We want this level of Y. It costs X. We don’t have enough, doing things just as we do now, for all values of Y. So we adjust the equation by lowering Y, what we want, or raising X, the amount we make available to get to this level of Y. The inevitable political arguments are about how to do this.

Arguments over what to do very often treat the level of Y, the desired outcome, as so obviously desirable as to be beyond argument. And, indeed, who will argue that we must let some number of older people die in order to spend more on education or some other worthy project? When the argument is put this way, nothing is negotiable. But, in fact, both X and Y in the equation are variable quantities, the final result of complicated calculations. Here are some of the complications.

Money is not the only scarce resource. Time, energy, interest, and enthusiasm are all scarce, as students of social movements have long emphasized. People and organizations have to be persuaded to care about the heat wave, the snow storm, the earthquake, or the rocket’s safety and to spend their time and energy on it, giving it priority over competing demands.

If we think of resources as finite, the situation looks familiar: a zero-sum game. Putting rocket safety first may hinder making the launches that will persuade the government to continue paying for the project. This is misleading. Resources are finite, but not inflexibly so. People make extraordinary efforts in extraordinary circumstances. They find and use resources they didn’t know they had or were even possible. Time is certainly finite; a day only has twenty-four hours. But the amount of available time is more flexible than it looks. Aubert and White (1965) identified sleep as a source of extra time:

If socially instituted sleep periods are longer than physiological needs require, it provides a constant supply of surplus time. The night contains a time-reservoir which can be drawn on both for emergency actions, ceremonies, and the completion of unfulfilled tasks. (p. 193)

Many people, deciding that the 24 hour day is inefficient, have tried to gain some extra time with a 25 or 26 hour day. But you can’t ordinarily convince others to adopt this convention, so it is usually “impractical” and unworkable (Gleick 1986 pp. 1-3)

Similar reservoirs exist for other resources. Families create different “pots of money,” some set aside for emergencies or anticipated expenses associated with death (Zelizer 1994, pp. 21-5, 109-11, 135-41; Lave 1988, pp. 131-41). Organizations of all sizes routinely set aside “contingency funds” for similar reasons. University administrators have “discretionary funds” they can allocate for expenditures that might otherwise not be allowed. So inflexible financial limits are actually elastic. Every seemingly irreducibly necessary quantity of a resource is socially defined, especially how much of it is untouchable except in extreme situations. The amount available for X, the expenditure in our equation, is more variable than it often seems.

The amount Y, the desirable level of preparation, also varies. An amount of preparation that seemed adequate for the improbable future disaster appears, when that rare but not impossible event finally occurs, nowhere near enough. The accuracy of predictions of disaster-producing natural processes depends on the time scale you use. Patterns of seismic activity invisible over a span of fifty years are obvious when you look at the geological record over several hundred years. Extrapolating from recent experience misses larger patterns which show up in disasters far greater than anything we expect when we calculate how much preparation would be enough. (See the studies of earthquakes in Southern California summarized in Davis 1998, 14-39.)

Short-term political considerations weigh heavily in many such calculations. Thus, American cities allow building in areas which repeatedly flood, or burn in devastating fires, because of the pressure exerted by real estate developers, builders, and others who get rich by ignoring these dangers. This often is decisive in setting the level of preparations considered “realistic.” (See, on this process in Los Angeles, Davis 1998, chapters 2 and 3.)

Calculations of how much preparation is enough, then, are not dictated in any simple way by physical realities, because there is no logical way to choose between methods of calculating Y that produce different results. Our answer is given, instead, by our conventionally accepted ideas of what’s likely, and we don’t prepare for anything beyond that.

If this first category of problem can be called “disasters and accidents,” a superficially very different one has a much less dramatic name: “collections.” Museums and libraries collect stuff — books, art works, anthropological and biological specimens, things considered valuable as part of a cultural patrimony or a data base for research. Which things are worth collecting and saving? How many things are enough to make the collection useful? How long must we keep them? Because — it is one of the lessons of growing up — we can’t keep all our favorite toys. If you never throw anything away you soon run out of space.

If you can afford it (if X in your equation is very high), you don’t have to throw your toys away so soon. You can buy more and keep it all. Starting as a child, and continuing through her adult life, Margaret Woodbury Strong (daughter of a rich man and married to a rich man) collected an enormous number of dolls and toys, but also all kinds of household machinery (kitchen stoves, washing machines, refrigerators, irons) — things most people would not have thought worth saving. And she didn’t just buy one of each: she once bought all the bathtubs in a hotel about to be torn down, and used them as planters around the perimeter of her home. She had no aesthetic or theoretical rationale for her choices. She just bought what she wanted. But she could afford to endow, build, and staff the Strong Museum in Rochester, New York, which contains some 500,000 such objects, some on public display and many more stored less publicly. (Of course, time, money, and energy limited what even she could do, so she just represents how much farther one might go than is conventional.)

More typically, organizations which throw nothing away eventually run up against the equation governing expenditure and outcome: spend more for exhibition and storage space (instead of on further acquisitions), or get rid of things. Nicholson Baker (2001) complained that many libraries, faced with this choice, had discarded irreplaceable collections of newspapers because they no longer had room for them and thought they could, anyway, be replaced by microfilm. Museums of natural history partially solve the problem by displaying a few items publicly and keeping the rest in a much more space-saving way, but that is only a temporary expedient. Art museums, usually quite discreetly, “de-accession” (less euphemistically, sell) items they think no longer worth saving.

It’s hard to find a simple principle governing the value of things that will tell us what’s worth saving and what can safely be thrown away, another complication in evaluating the equation. Few people thought Mrs. Strong was collecting anything worth saving. But what seems easy to discard now is often not so regarded later on. If we keep our collections to a “reasonable” size by getting rid of things not worth saving now, we will inevitably make what later generations will see as terrible mistakes. Some of what we discard now will be just what they prize and need. That’s why librarians collect and store material no one has any use for now. They suppose — experience supports their assumption — that people (especially scholars) will later find that stuff invaluable. Archeologists make great finds in the garbage dumps of the societies they study, the places where people discarded broken, useless things. Now that garbage is just what our scholars use to describe and analyze ways of life long gone. So scholars will use Mrs. Strong’s collections of refrigerators and washing machines to study changes in the nature of household work in the 19th century, and the museum’s extensive holdings of dolls and toys to study changes in the nature of childhood. Future scholars will surely find still other problems we haven’t thought of for which these collections will provide the ideal data.

The equation for disasters and accidents is relatively (only relatively) simple: how much is enough to prevent some foreseeable, if unlikely, calamity? The collection is more complicated because we don’t know the future value of something (what later generations will find it useful for) when we save or discard it (will it be rejected as not having been worth saving after all?). That makes it impossible to calculate the tradeoffs involved. We don’t know what will be gained or lost if we don’t make this investment in conservation.

Scholars and academics experience this problem personally: how many books are enough? Most scholars believe they need every book they own, without exception, and keep books against the barest possibility of a future utility. But how many books do they really need to get their work done? My own working library for a specific project is usually about 25-30 books, and I routinely sell surplus books. But I sometimes sell a book only to find that I need it after all. So I found, writing this paper, that I needed Durkheim’s Suicide (1951), sold long ago when our move to a smaller apartment decreased the available space for books. Of course, I could and did borrow it from the local library. How do I calculate the tradeoff here? If I buy it — is that such a terrible cost? Compared to the peace of mind that came from cleaning out so many books I hadn’t looked at in thirty years?

These calculations are complicated too. I have easy access, in the United States, to free public libraries. But many European scholars assembled their collections exactly because, while libraries existed, and you could use them, you could never just browse. You had to ask the librarian for specific books and then wait for them to be delivered. Having your own library avoided that. But then you couldn’t make the serendipitous finds browsing in open library stacks gives American scholars who, though better off, are never as perfectly supplied as they would like. The unknowable future need creates the urge for a library of one’s own.

These examples take the utility of what is to be saved or bought as given. Of course we want to save lives, prevent the disruption of our cities and the crashes of our space craft, have access to books needed for important work. But the desire to maximize possessions varies among social groups, as became clear when I raised the topic at a dinner party in Paris. The women present (themselves scholars) immediately suggested a similar question based in their gender-specific experience: how many pairs of shoes are enough? And the hostess went around the table asking everyone, male and female, how many pairs they had. Men found the question puzzling, and didn’t always know the answer, but the women understood immediately and knew the precise answer.

The women knew, as the men didn’t, that a more basic question was involved: what do you need different pairs of shoes for? Why isn’t one pair (or two at the most) enough, as is often true for men? The two groups made the calculation differently.

Clothing and appearance mean different things for men and women. You not only must not be naked, and protected against the weather, you must be clothed appropriately for specific situations. Situations are more differentiated for women in this respect than for men. Men in Western and Westernized countries have perhaps three registers of clothing: formal (suit and tie), informal (no tie, perhaps jeans), and completely informal (or sloppy: sweats, shorts, etc.). There are meaningful variations of price and style within categories and some men, perhaps some categories of men, care more about this than others and monitor the subtle differences between brands and styles more carefully, just as some boys take the difference between brands of sports shoes seriously. But the variety of situations men must consider in choosing what they wear is less.

Women in contemporary Western societies, however, find themselves in a great variety of situations which require different kinds of clothing, and the accompanying shoes, which in turn means you must have a larger supply of shoes to meet the situational expectations for varying degrees of formality, sportiness, and being in style. Likewise, clothes and shoes can signify and hint at more things about age, cultural dispositions, sexuality, and attractiveness for women than for men. And that means that how many shoes are enough will be more for women than for men. (“Required” here means that some others will expect you to respect these conventions and that you are unwilling to be the kind of person who does not honor them. I suppose this is why women so often feel they must wear a new dress to someone else’s wedding.)

So we add this to our equation: what we want (Y in the equation), while variable, is also socially defined and socially constrained and thus is more variable taken in the asocial abstract than in the concrete situations of everyday life.

Despite their obvious differences, catastrophes and collections have much in common. Heat waves (it has been demonstrated and much remarked on) do not kill people randomly, but rather select on race, class, and ethnicity. You can think of these disasters, very generally, as part of the selective Darwinian process whose survivors will constitute the world’s collection of human beings, as the survivors of a similar selection constitute the population of books in a library. So disasters make collections. This models a general process found in many disparate areas. I will describe just a few. Each provokes more complications in the rudimentary model I’ve been casually constructing (whose ramifications I’ll leave as an exercise for my listeners and readers, at least for now).

Biologists warn of the terrible consequences of reducing the planet’s biological diversity. In the United States, major construction and harvesting projects have been halted by the discovery that the project would kill off a rare species. But which species to save? All the millions of kinds of beetles? A rare and beautiful species of deer on the verge of extinction? Animals generally considered “cute” or beautiful (but by who? using what criteria?) surely are more likely to be saved than beetles. Because there are so many more kinds of beetles but also because who, other than an entomologist, likes bugs that much?

Further, the enormous number of interconnections between species make it impossible to know what else will be affected by our decision about this one species. And, further still, these consequences take one form in the next ten years, and quite different forms fifty, a hundred, a thousand years from now.

We can’t decide which species to save because, given these complexities, we can’t construct a logically defensible common measure for the many disparate things we want to maximize. (I have taken these examples and arguments from Bowker 2004.) Every way of settling the questions I’ve been discussing runs into a similar impasse, which is why we solve them only by accepting some conventionalized solution, accepted not because it’s “right,” but because (in a way that is perfectly circular) it’s what everyone will accept.

Similar questions arise over the preservation of languages. There are many thousands of languages, but many of them disappear every year. The cost of keeping a language alive is high: you must preserve a population that speaks it and passes it on to the children via educational institutions, a method of writing it, a literature written in it, and a political basis for all that, because the political implications of keeping a language alive or destroying it are important. (See the discussions in Maffi 2001.)

The preservation of historic buildings is a similar process. Every building is historic but, if they are all to be saved for their historical importance, then how will a city grow? Americans really don’t have much of a problem with this, though they think they do, because their cities are so new. In contrast, a two thousand year old city like Rome can only exist as a palimpsest, in which buildings contain the remnants of other buildings, in which every attempt to build something new uncovers something older, sometimes very much older, which nowadays has a prima facie claim to continued existence. (I’m influenced here by the many varieties of city described in Calvino 1974.)

Another version of the biological problem is more familiar and more intimate. How many children are enough for two people to have? The “correct answer” to this question has changed historically, and the criteria for arriving at the answer have changed as well. In earlier times, a couple had enough children to do all the work the family farm or business required for economic viability, taking into account the inevitable deaths along the way. And parents wanted enough children to survive to support them in their old age. More recently, the appropriate number of children has been set in a more complicated way, involving questions of how many can be supported in the style that is thought appropriate. Or, as Gilberto Velho suggested (1976, pp. 272-4), how many children are required to achieve a “family project” of social mobility or preservation of class status?

A public version of the question asks how many children the society or nation needs or should have. So governments espouse family policies suggesting, or encouraging, or requiring so and so many children. We need more children to make more soldiers (France after 1870, Germany under Hitler). We need fewer children, to reduce the amount of poverty, pave the way for economic growth (China’s campaign to reduce the number of children to one per couple), or keep the earth a livable planet.

Underlying all these specific problems is the basic question Durkheim (1951, pp. 246-58) raised. How do we know when we have enough — of anything? He argued that anomic suicide, the special kind which appeared in boom times as well as depressions, resulted from the feeling of insatiability people experienced when the norms governing ordinary consumption were no longer convincing and enforceable. Our wants and desires are ordinarily constrained by norms, sometimes even by rules, governing what people like us should be allowed to possess. Sumptuary laws once limited the clothing people of various social ranks could wear: nobility could wear suits made of the most expensive cloth, but tradesmen had to be satisfied with something less ostentatious. Our clothing tastes and desires are similarly limited by expense. We can’t wear what we can’t pay for. When people no longer feel or accept these constraints, as happens in a boom, they don’t know when to stop (it’s the “Emelda Marcos syndrome,” accounting for her five hundred pairs of shoes).

Sociologists will not be surprised by Durkheim’s general answer to all these questions about how much is enough: it’s enough — preparations for disasters, books in a library, species in the biosphere, shoes in the closet — when everyone who has a voice in the matter agrees to accept it as enough. Since the answer results from compromise, it won’t be logical and will be defensible only as what was possible, at the time, in the circumstances. It will work, like everything in social life, because it is what, at the moment, “everyone” accepts as the way things are done.

Durkheim might have added, had he considered all these permutations of the question, that it’s no easy thing to get all the interested parties to agree — from the politicians and citizens trying to allocate revenues among competing worthy causes to the members of a couple trying to agree on how many children to have or, more difficult yet, on how many books the house can hold.


Aubert, Vilhelm. 1982. The Hidden Society. New Brunswick [N.J.]: Transaction Books.

Aubert, Vilhelm, and Harrison C. White. 1965. “Sleep: a sociological interpretation.” Pp. 168-200 in The Hidden Society, edited by Vilhelm Aubert. Totowa, New Jersey: Bedminster Press.

Baker, Nicholson. 2001. Doublefold :Libraries and the Assault on Paper. New York: Random House.

Bowker, Geoffrey C. 2004. “Time, Money and Biodiversity.” in Global Assemblages: Technology, Politics, and Ethics as Anthropological Problems, edited by Aihwa On and Stephen J. Collier. Oxford: Blackwell.

Calvino, Italo. 1974. Invisible Cities. New York: Harcourt Brace.

Durkheim, Emile. 1951. Suicide, a study in sociology. Glencoe, Ill.,: Free Press.

Gleick, James. 1987. Chaos: Making a New Science. New York, N.Y.: Viking.

Granger, Bill, and Lori Granger. 1980. Fighting Jane: Mayor Jane Byrne and the Chicago Machine. New York: Dial Press.

Kidder, Tracy. 1981. The Soul of a New Machine. Boston: Little Brown and Company.

Klinenberg, Eric. 2002. Heat Wave: A Social Autopsy of Disaster in Chicago. Chicago: University of Chicago Press.

Latour, Bruno. 1996. Aramis or the Love of Technology. Cambridge, MA: Harvard University Press.

Lave, Jean. 1988. Cognition in Practice. Cambridge: Cambridge University Press.

Maffi, Luisa (Ed.). 2001. On Biocultural Diversity: Linking Language, Knowledge, and the Environment. Washington. D.C.: Smithsonian Institution Press.

Mergen, Bernard. 1997. Snow in America. Washington and London: Smithsonian Institution.

Star, Susan Leigh, and James R. Griesemer. 1989. “Institutional Ecology, “Translations,” and Boundary Objects in Berkeley’s Museum of Vertebrate Zoology, 1907-39.” Social Studies of Science 19:387-420.

Vaughan, Diane. 1996. The Challenger Launch Decision :Risky Technology, Culture, and Deviance at NASA. Chicago: University of Chicago Press.

Velho, Gilberto. 1976. “Accusations, Family Mobility and Deviant Behavior.” Social Problems 23:268-75.

Zelizer, Viviana. 1989. “The Social Meaning of Money: “Special Monies.” American Journal of Sociology 95:342-77.

Zelizer, Viviana A. 1994. The Social Meaning of Money. New York: Basic Books.