distinct. Decrements in a scuba diver'sperformance which result from thesedistortions, however, may be consider-ably lessened by adaptation to the un-derwater environment. Our involvementstems from a need to improve the visualperformance of divers. But in the courseof this work we have become increas-ingly aware of another great oppor-tunity that the underwater world pro-vides: It is a unique laboratory for theinvestigation of countless perceptualproblems which bear on the mostfundamental theories of perception.

References and Notes

1. New York Times 1969, 1 (12 Jan. 1969).2. W. E. K. Middleton, Vision through the

Atmosphere (Univ. of Toronto Press, Toronto,1952), p. 12.

3. S. Q. Duntley, J. Opt. Soc. Amer. 53, 246(1963).

4. H. R. Blackwell, ibid. 36, 624 (1946).5. L. E. Kinsler, Amer. J. Phys. 13, 255 (1945).6. The maximum amount of magnification is

1.33. The exact amount obtained depends onthe distance of the eye from the interface;with a typical face mask, the magnificationis approximately 27 percent.

7. P. R. Kent and S. Weissman, Nav. Submarine

Med. Center Rep. No. 476 (May 1966); P. R.Kent, Amer. J. Optom. 43, 553 (1966).

8. C. H. Graham, in Handbook of ExperimentalPsychology, S. S. Stevens, Ed. (Wiley, NewYork, 1952), p. 885.

9. S. M. Luria and J. A. S. Kinney, Amer. J.Psychol. 81, 359 (1968).

10. H. Ross, in Underwater Association Report1966-67, J. N. Lythgoe and J. D. Woods,Eds. (T. G. Williams Industrial and ResearchPromotions Ltd., London, 1967), p. 61.

11. A. Lit, Optom. Weekly 59, 42 (14 Nov. 1968).12. L. L. Avant, Psychol. Bull. 64, 246 (1965).13. T. C. D. Whiteside, Roy. Air Force Inst.

Aviat. Med. Flying Personnel Res. Comm.Rep. No. 910 (Nov. 1954).

14. T. Eames, Amer. J. Ophthalmol. 43, 279(1957); J. Educ. Res. 19, 469 (1936).

15. S. M. Luria, Science 164, 452 (1969).16. J. Goldstein, A. Clahane, S. Sanfilippo, Amer.

J. Ophthalmol. 62, 702 (1966).17. C. Phinizy, Skin Diver 18, No. 5, 41 (1969).18. J. A. S. Kinney and S. M. Luria, paper pre-

sented at a meeting of the Eastern Psycho-logical Association, April 1969.

19. , D. 0. Weitzman, Percept. Mot. Skills28, 331 (1969).

20. S. M. Luria, J. A. S. Kinney, S. Weissman,Amer. J. Psychol. 80, 282 (1967).

21. H. E. Ross, Quart. J. Exp. Psychol. 17, 329(1965).

22. G. Fry, C. Bridgman, V. Ellerbrock, Amer.J. Optom. 26, 9 (1949); H. E. Ross, Brit. J.Psychol. 58, 301 (1967).

23. S. M. Luria and J. A. S. Kinney, Percept.Mot. Skills 26, 1019 (1968).

24. J. E. Tyler, Limnol. Oceanogr. 60, 102 (1959).25. J. A. S. Kinney, S. M. Luria, D. 0. Weitz-

man, J. Opt. Soc. Amer. 57, 802 (1967).

The Experience of Living in CitiesAdaptations to urban overload create characteristic

qualities of city life that can be measured.

Stanley Milgram

"When I first came to New York it seemed like a nightmare. As soon as Igot ofi the train at Grand Central I was caught up in pushing, shoving crowdson 42nd Street. Sometimes people bumped into me without apology; what reallyfrightened me was to see two people literally engaged in combat for possessionof a cab. Why were they so rushed? Even drunks on the street were bypassedwithout a glance. People didn't seem to care about each other at all."

This statement represents a commonreaction to a great city, but it does nottell the whole story. Obviously citieshave great appeal because of their vari-ety, eventfulness, possibility of choice,and the stimulation of an intenseatmosphere that many individuals finda desirable background to their lives.Where face-to-face contacts are impor-tant, the city offers unparalleled possi-bilities. It has been calculated by the

13 MARCH 1970

Regional Plan Association (1) that inNassau County, a suburb of New YorkCity, an individual can meet 11,000others within a 10-minute radius of hisoffice by foot or car. In Newark, amoderate-sized city, he can meet morethan 20,000 persons within this radius.But in midtown Manhattan he can meetfully 220,000. So there is an order-of-magnitude increment in the communica-tion possibilities offered by a great city.

26. , ibid. 59, 640 (1969).27. W. Epstein, Varieties of Perceptual Learning

(McGraw-Hill, New York, 1967); J. F. Wohl-will, Annut. Rev. Psychol. 17, 201 (1966).

28. G. Stratton, Psychol. Rev. 4, 341 (1897);ibid., p. 463.

29. See, for example, S. J. Freedman, The Neuro-psychology of Spatially Oriented Behavior(Dorsey, Homewood, Ill., 1968).

30. I. Kohler, Acta Psychol. 11, 176 (1955);Psychol. Issues 3, No. 4 (1964).

31. J. A. S. Kinney, S. M. Luria, D. 0. Weitz-man, Nav. Submarine Med. Center Rep. No.541 (July 1968).

32. 1. Rock, The Nature of Perceptual Adaptation(Basic Books, New York, 1966); C. S. Harris,Psychol. Rev. 72, 419 (1965); J. G. Taylor,The Behavioral Basis of Perception (YaleUniv. Press, New Haven, 1962); K. U. Smithand W. K. Smith, Perception and Motion(Saunders, Philadelphia, 1962).

33. R. Held and S. J. Freedman, Science 142,455 (1963); R. Held, Sci. Amer. 213, 84(Nov. 1965).

34. I. P. Howard and W. B. Templeton, HumanSpatial Orientation (Wiley, New York, 1966),chap. 15.

35. S. Weinstein, E. A. Sersen, L. Fisher, M.Weisinger, Percept. Mot. Skills 18, 641(1964).

36. This work was carried out as part of theBureau of Medicine and Surgery, Navy De-partment, Research Work Unit M4306.03-2050D. The opinions or assertions containedherein are the private ones of the authors andare not to be construed as official or reflect-ing the views of the Navy Department orthe Naval Service at large.

That is one of the bases of its appealand, indeed, of its functional necessity.The city provides options that no othersocial arrangement permits. But thereis a negative side also, as we shall see.

Granted that cities are indispensablein complex society, we may still askwhat contribution psychology can maketo understanding the experience of liv-ing in them. What theories are relevant?How can we extend our knowledge ofthe psychological aspects of life in citiesthrough empirical inquiry? If empiricalinquiry is possible, along what linesshould it proceed? In short, where dowe start in constructing urban theoryand in laying out lines of research?

Observation is the indispensable start-ing point. Any observer in the streetsof midtown Manhattan will see (i)large numbers of people, (ii) a highpopulation density, and (iii) heteroge-neity of population. These three factorsneed to be at the root of any sociopsy-chological theory of city life, for theycondition all aspects of our experiencein the metropolis. Louis Wirth (2), ifnot the first to point to these factors, isnonetheless the sociologist who reliedmost heavily on them in his analysis ofthe city. Yet, for a psychologist, there

The author is professor of psychology at theGraduate Center of The City University of NewYork, New York 10036. This article is based onan address given on 2 September 1969 at the 77thannual meeting of the American PsychologicalAssociation, in Washington, D.C.

1461

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

is something unsatisfactory aboutWirth's theoretical variables. Numbers,density, and heterogeneity are demo-graphic facts but they are not yet psy-chological facts. They are external tothe individual. Psychology needs an ideathat links the individual's experience tothe demographic circumstances of urbanlife.One link is provided by the concept

of overload. This term, drawn from sys-tems analysis, refers to a system's in-ability to process inputs from the en-vironment because there are too manyinputs for the system to cope with, orbecause successive inputs come so fastthat input A cannot be processed wheninput B is presented. When overload ispresent, adaptations occur. The systemmust set priorities and make choices. Amay be processed first while B is keptin abeyance, or one input may be sacri-ficed altogether. City life, as we experi-ence it, constitutes a continuous set ofencounters with overload, and of result-ant adaptations. Overload characteris-tically deforms daily life on severallevels, impinging on role performance,the evolution of social norms, cognitivefunctioning, and the use of facilities.The concept has been implicit in

several theories of urban experience. In1903 George Simmel (3) pointed outthat, since urban dwellers come intocontact with vast numbers of peopleeach day, they conserve psychic energyby becoming acquainted with a farsmaller proportion of people than theirrural counterparts do, and by maintain-ing more superficial relationships evenwith these acquaintances. Wirth (2)points specifically to "the superficiality,the anonymity, and the transitory char-acter of urban social relations."One adaptive response to overload,

therefore, is the allocation of less timeto each input. A second adaptive mech-anism is disregard of low-priority inputs.Principles of selectivity are formulatedsuch that investment of time and energyare reserved for carefully defined inputs(the urbanite disregards the drunk sickon the street as he purposefully navi-gates through the crowd). Third, bound-aries are redrawn in certain social trans-actions so that the overloaded systemcan shift the burden to the other partyin the exchange; thus, harried NewYork bus drivers once made change forcustomers, but now this responsibilityhas been shifted to the client, who musthave the exact fare ready. Fourth, recep-tion is blocked off prior to entrance intoa system; city dwellers increasingly useunlisted telephone numbers to prevent

1462

individuals from calling them, and asmall but growing number resort tokeeping the telephone off the hook toprevent incoming calls. More subtly, acity dweller blocks inputs by assumingan unfriendly countenance, which dis-courages others from initiating contact.Additionally, social screening devicesare interposed between the individualand environmental inputs (in a townof 5000 anyone can drop in to chatwith the mayor, but in the metropolisorganizational screening devices deflectinputs to other destinations). Fifth, theintensity of inputs is diminished by filter-ing devices, so that only weak and rela-tively superficial forms of involvementwith others are allowed. Sixth, special-ized institutions are created to absorbinputs that would otherwise swamp theindividual (welfare departments handlethe financial needs of a million individ-uals in New York City, who wouldotherwise create an army of mendicantscontinuously importuning the pedes-trian). The interposition of institutionsbetween the individual and the socialworld, a characteristic of all modernsociety, and most notably of the largemetropolis, has its negative side. Itdeprives the individual of a sense ofdirect contact and spontaneous integra-tion in the life around him. It simulta-neously protects and estranges the in-dividual from his social environment.Many of these adaptive mechanisms

apply not only to individuals but to in-stitutional systems as well, as Meier (4)has so brilliantly shown in connectionwith the library and the stock exchange.

In sum, the observed behavior of theurbanite in a wide range of situationsappears to be determined largely by avariety of adaptations to overload. I nowdeal with several specific consequencesof responses to overload, which makefor differences in the tone of city andtown.

Social Responsibility

The principal point of interest for asocial psychology of the city is thatmoral and social involvement with in-dividuals is necessarily restricted. Thisis a direct and necessary function ofexcess of input over capacity to process.Such restriction of involvement runs abroad spectrum from refusal to becomeinvolved in the needs of another person,even when the person desperately needsassistance, through refusal to do favors,to the simple withdrawal of courtesies(such as offering a lady a seat, or saying

"sorry" when a pedestrian collision oc-curs). In any transaction more andmore details need to be dropped as thetotal number of units to be processedincreases and assaults an instrument oflimited processing capacity.The ultimate adaptation to an over-

loaded social environment is to totallydisregard the needs, interests, and de-mands of those whom one does notdefine as relevant to the satisfaction ofpersonal needs, and to develop highlyefficient perceptual means of determi-ning whether an individual falls into thecategory of friend or stranger. Thedisparity in the treatment of friendsand strangers ought to be greater incities than in towns; the time allotmentand willingness to become involved withthose who have no personal claim onone's time is likely to be less in citiesthan in towns.

Bystander intervention in crises. Themost striking deficiencies in social re-sponsibility in cities occur in crisis situa-tions, such as the Genovese murder inQueens. In 1964, Catherine Genovese,coming home from a night job in theearly hours of an April morning, wasstabbed repeatedly, over an extendedperiod of time. Thirty-eight residents ofa respectable New York City neighbor-hood admit to having witnessed at leasta part of the attack, but none went toher aid or called the police until aftershe was dead. Milgram and Hollander,writing in The Nation (5), analyzed theevent in these terms:Urban friendships and associations arenot primarily formed on the basis ofphysical proximity. A person with numer-ous close friends in different parts of thecity may not know the occupant of anadjacent apartment. This does not meanthat a city dweller has fewer friends thandoes a villager, or knows fewer personswho will come to his aid; however, itdoes mean that his allies are not con-stantly at hand. Miss Genovese requiredimmediate aid from those physically pres-ent. There is no evidence that the cityhad deprived Miss Genovese of humanassociations, but the friends who mighthave rushed to her side were miles fromthe scene of her tragedy.

Further, it is known that her cries forhelp were not directed to a specific person;they were general. But only individualscan act, and as the cries were not spe-cifically directed, no particular person felta special responsibility. The crime and thefailure of community response seem ab-surd to us. At the time, it may well haveseemed equally absurd to the Kew Gardensresidents that not one of the neighborswould have called the police. A collectiveparalysis may have developed from thebelief of each of the witnesses that some-one else must surely have taken thatobvious step.

SCIENCE, VOL. 167

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

Latane and Darley (6) have reportedlaboratory approaches to the study ofbystander intervention and have es-tablished experimentally the followingprinciple: the larger the number of by-standers, the less the likelihood that anyone of them will intervene in an emer-gency. Gaertner and Bickman (7) ofThe City University of New York haveextended the bystander studies to anexamination of help across ethnic lines.Blacks and whites, with clearly identifi-able accents, called strangers (throughwhat the caller represented as an errorin telephone dialing), gave them a plau-sible story of being stranded on an out-lying highway without more dimes, andasked the stranger to call a garage. Theexperimenters found that the whitecallers had a significantly better chanceof obtaining assistance than the blackcallers. This suggests that ethnic alle-giance may well be another means ofcoping with overload: the city dwellercan reduce excessive demands andscreen out urban heterogeneity by re-sponding along ethnic lines; overload ismade more manageable by limiting the"span of sympathy."

In any quantitative characterizationof the social texture of city life, anecessary first step is the application ofsuch experimental methods as these tofield situations in large cities and smalltowns. Theorists argue that the indiffer-ence shown in the Genovese case wouldnot be found in a small town, but in theabsence of solid experimental evidencethe question remains an open one.More than just callousness prevents

bystanders from participating in alterca-tions between people. A rule of urbanlife is respect for other people's emo-tional and social privacy, perhaps be-cause physical privacy is so hard toachieve. And in situations for which thestandards are heterogeneous, it is muchharder to know whether taking an activerole is unwarranted meddling or an ap-propriate response to a critical situation.If a husband and wife are quarreling inpublic, at what point should a bystanderstep in? On the one hand, the hetero-geneity of the city produces substantiallygreater tolerance about behavior, dress,and codes of ethics than is generallyfound in the small town, but this diver-sity also encourages people to withholdaid for fear of antagonizing the partici-pants or crossing an inappropriate anddifficult-to-define line.

Moreover, the frequency of demandspresent in the city gives rise to norms ofnoninvolvement. There are practicallimitations to the Samaritan impulse ina major city. If a citizen attended to13 MARCH 1970

Table 1. Percentage of entries achieved byinvestigators for city and town dwellings (seetext).

Entries achieved (%)Experimenter

City* Small townt

MaleNo. 1 16 40No. 2 12 60

FemaleNo. 3 40 87No. 4 40 100

* Number of requests for entry, 100. t Number ofrequests for entry, 60.

every needy person, if he were sensitiveto and acted on every altruistic impulsethat was evoked in the city, he couldscarely keep his own affairs in order.

Willingness to trust and assist stran-gers. We now move away from crisissituations to less urgent examples ofsocial responsibility. For it is not onlyin situations of dramatic need but inthe ordinary, everyday willingness tolend a hand that the city dweller is saidto be deficient relative to his small-towncousin. The comparative method mustbe used in any empirical examinationof this question. A commonplace socialsituation is staged in an urban settingand in a small town-a situation towhich a subject can respond by eitherextending help or withholding it. Theresponses in town and city are com-pared.One factor in the purported unwill-

ingness of urbanites to be helpful tostrangers may well be their heightenedsense of physical (and emotional) vul-nerability-a feeling that is supportedby urban crime statistics. A key test fordistinguishing between city and townbehavior, therefore, is determining howcity dwellers compare with town dwellersin offering aid that increases their per-sonal vulnerability and requires sometrust of strangers. Altman, Levine, Na-dien, and Villena (8) of The City Uni-versity of New York devised a study tocompare the behaviors of city and towndwellers in this respect. The criterionused in this study was the willingness ofhouseholders to allow strangers to entertheir home to use the telephone. Thestudent investigators individually rangdoorbells, explained that they had mis-placed the address of a friend nearby,and asked to use the phone. The in-vestigators (two males and two females)made 100 requests for entry into homesin the city and 60 requests in the smalltowns. The results for middle-incomehousing developments in Manhattanwere compared with data for severalsmall towns (Stony Point, Spring Valley,Ramapo, Nyack, New City, and West

Clarkstown) in Rockland County, out-side of New York City. As Table 1shows, in all cases there was a sharp in-crease in the proportion of entriesachieved by an experimenter when hemoved from the city to a small town. Inthe most extreme case the experimenterwas five times as likely to gain admissionto homes in a small town as to homes inManhattan. Although the female experi-menters had notably greater successboth in cities and in towns than themale experimenters had, each of the fourstudents did at least twice as well intowns as in cities. This suggests that thecity-town distinction overrides even thepredictably greater fear of male stran-gers than of female ones.The lower level of helpfulness by city

dwellers seems due in part to recogni-tion of the dangers of living in Man-hattan, rather than to mere indifferenceor coldness. It is significant that 75 per-cent of all the city respondents receivedand answered messages by shoutingthrough closed doors and by peering outthrough peepholes; in the towns, bycontrast, about 75 percent of the re-spondents opened the door.

Supporting the experimenters' quanti-tative results was their general observa-tion that the town dwellers were notice-ably more friendly and less suspiciousthan the city dwellers. In seeking to ex-plain the reasons for the greater senseof psychological vulnerability city dwell-ers feel, above and beyond the differ-ences in crime statistics, Villena (8)points out that, if a crime is committedin a village, a resident of a neighboringvillage may not perceive the crime aspersonally relevant, though the geo-graphic distance may be small, whereasa criminal act committed anywhere inthe city, though miles from the city-dweller's home is still verbally locatedwithin the city; thus, Villena says, "theinhabitant of the city possesses a largervulnerable space."

Civilities. Even at the most superficiallevel of involvement-the exercise ofeveryday civilities-urbanites are reput-edly deficient. People bump into eachother and often do not apologize. Theyknock over another person's packagesand, as often as not, proceed on theirway with a grumpy exclamation insteadof an offer of assistance. Such behavior,which many visitors to great cities finddistasteful, is less common, we aretold, in smaller communities, wheretraditional courtesies are more likely tobe observed.

In some instances it is not simplythat, in the city, traditional courtesiesare violated; rather, the cities develop

1463

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

new norms of noninvolvement. Theseare so well defined and so deeply a partof city life that they constitute thenorms people are reluctant to violate.Men are actually embarrassed to giveup a seat on the subway to an oldwoman; they mumble "I was gettingoff anyway," instead of making thegesture in a straightforward and graciousway. These norms develop becauseeveryone realizes that, in situations ofhigh population density, people cannotimplicate themselves in each others' af-fairs, for to do so would create condi-tions of continual distraction whichwould frustrate purposeful action.

In discussing the effects of overloadI do not imply that at every instant thecity dweller is bombarded with an un-manageable number of inputs, and thathis responses are determined by theexcess of input at any given instant.Rather, adaptation occurs in the formof gradual evolution of norms of be-havior. Norms are evolved in responseto frequent discrete experiences of over-load; they persist and become general-ized modes of responding.

Overload on cognitive capacities: ano-nymity. That we respond differentlytoward those whom we know and thosewho are strangers to us is a truism. Aneager patron aggressively cuts in frontof someone in a long movie line to savetime only to confront a friend; he thenbehaves sheepishly. A man is involvedin an automobile accident caused byanother driver, emerges from his carshouting in rage, then moderates hisbehavior on discovering a friend drivingthe other car. The city dweller, whenwalking through the midtown streets,is in a state of continual anonymity vis-a-vis the other pedestrians.Anonymity is part of a continuous

spectrum ranging from total anonymityto full acquaintance, and it may well bethat measurement of the precise degreesof anonymity in cities and towns wouldhelp to explain important distinctionsbetween the quality of life in each.Conditions of full acquaintance, for ex-ample, offer security and-familiarity, butthey may also be stifling, because theindividual is caught in a web of estab-lished relationships. Conditions of com-plete anonymity, by contrast, providefreedom from routinized social ties, butthey may also create feelings of aliena-tion and detachment.

Empirically one could investigate theproportion of activities in which thecity dweller or the town dweller isknown by others at given times in hisdaily life, and the proportion of activi-

1464

ties in the course of which he interactswith individuals who know him. At hisjob, for instance, the city dweller maybe known to as many people as his ruralcounterpart. However, when he is notfulfilling his occupational role-say,when merely traveling about the city-the urbanite is doubtless more anony-mous than his rural counterpart.

Limited empirical work on anonymityhas begun. Zimbardo (9) has testedwhether the social anonymity and im-personality of the big city encouragegreater vandalism than do small towns.Zimbardo arranged for one automobileto be left for 64 hours near the Bronxcampus of New York University andfor a counterpart to be left for the samenumber of hours near Stanford Univer-sity in Palo Alto. The license plates onthe two cars were removed and thehoods were opened, to provide "releasercues" for potential vandals. The NewYork car was stripped of all movableparts within the first 24 hours, and bythe end of 3 days was only a hunk ofmetal rubble. Unexpectedly, however,most of the destruction occurred duringdaylight hours, usually under the scru-tiny of observers, and the leaders in thevandalism were well-dressed, whiteadults. The Palo Alto car was left un-touched.Zimbardo attributes the difference in

the treatment accorded the two cars tothe "acquired feelings of social anonym-ity provided by life in a city like NewYork," and he supports his conclusionswith several other anecdotes illustratingcasual, wanton vandalism in the city. Inany comparative study of the effects ofanonymity in city and town, however,there must be satisfactory control forother confounding factors: the largenumber of drug addicts in a city likeNew York; the higher proportion ofslum-dwellers in the city; and so on.

Another direction for empirical studyis investigation of the beneficial effectsof anonymity. The impersonality of citylife breeds its own tolerance for the pri-vate lives of the inhabitants. Individu-ality and even eccentricity, we may as-sume, can flourish more readily in themetropolis than in the small town.Stigmatized persons may find it easierto lead comfortable lives in the city,free of the constant scrutiny of neigh-bors. To what degree can this assumeddifference between city and town beshown empirically? Judith Waters (10),at The City University of New York,hypothesized that avowed homosexualswould be more likely to be accepted astenants in a large city than in small

towns, and she dispatched letters fromhomosexuals and from normal individ-uals to real estate agents in cities andtowns across the country. The resultsof her study were inconclusive. But thegeneral idea of examining the protectivebenefits of city life to the stigmatizedought to be pursued.

Role behavior in cities and towns.Another product of urban overload isthe adjustment in roles made by urban-ites in daily interactions. As Wirth hassaid (2): "Urbanites meet one anotherin highly segmental roles.... They areless dependent upon particular persons,and their dependence upon others isconfined to a highly fractionalized aspectof the other's round of activity." Thistendency is particularly noticeable intransactions between customers and in-dividuals offering professional or salesservices. The owner of a country storehas time to become well acquainted withhis dozen-or-so daily customers, but thegirl at the checkout counter of a busyA & P, serving hundreds of customersa day, barely has time to toss the greenstamps into one customer's shoppingbag before the next customer con-fronts her with his pile of groceries.

Meier, in his stimulating analysis ofthe city (4), discusses several adaptationsa system may make when confronted byinputs that exceed its capacity to proc-ess them. Meier argues that, accordingto the principle of competition forscarce resources, the scope and time ofthe transaction shrink as customervolume and daily turnover rise. This,in fact, is what is meant by the"brusque" quality of city life. Newstandards have developed in cities con-cerning what levels of services are ap-propriate in business transactions (seeFig. 1).McKenna and Morgenthau (11), in a

seminar at The City University of NewYork, devised a study (i) to comparethe willingness of city dwellers andsmall-town dwellers to do favors forstrangers that entailed expenditure of asmall amount of time and slight incon-venience but no personal vulnerability,and (ii) to determine whether the morecompartmentalized, transitory relation-ships of the city would make urbansalesgirls less likely than small-townsalesgirls to carry out, for strangers,tasks not related to their customaryroles.To test for differences between city

dwellers and small-town dwellers, asimple experiment was devised in whichpersons from both settings were asked(by telephone) to perform increasingly

SCIENCE, VOL. 167

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

onerous favors for anonymous strangers.Within the cities (Chicago, New

York, and Philadelphia), half the callswere to housewives and the other halfto salesgirls in women's apparel shops;the division was the same for the 37small towns of the study, which were inthe same states as the cities. Each ex-perimenter represented herself as a long-distance caller who had, through error,been connected with the respondent bythe operator. The experimenter beganby asking for simple information aboutthe weather for purposes of travel. Nextthe experimenter excused herself onsome pretext (asking the respondent to"please hold on"), put the phone downfor almost a full minute, and thenpicked it up again and asked the re-spondent to provide the phone numberof a hotel or motel in her vicinity atwhich the experimenter might stayduring a forthcoming visit. Scores wereassigned the subjects on the basis ofhow helpful they had been. McKennasummarizes her results in this manner:

People in the city, whether they are en-gaged in a specific job or not, are less help-ful and informative than people in smalltowns; . . . People at home, regardless ofwhere they live, are less helpful and in-formative than people working in shops.

However, the absolute level of coop-erativeness for urban subjects was foundto be quite high, and does not accordwith the stereotype of the urbanite asaloof, self-centered, and unwilling tohelp strangers. The quantitative differ-ences obtained by McKenna andMorgenthau are less great than onemight have expected. This again pointsup the need for extensive empirical re-search in rural-urban differences, re-search that goes far beyond that providedin the few illustrative pilot studies pre-sented here. At this point we have verylimited objective evidence on differencesin the quality of social encounters incity and small town.

But the research needs to be guidedby unifying theoretical concepts. As Ihave tried to demonstrate, the conceptof overload helps to explain a widevariety of contrasts between city behav-ior and town behavior: (i) the differ-ences in role enactment (the tendencyof urban dwellers to deal with one an-other in highly segmented, functionalterms, and of urban sales personnel todevote limited time and attention totheir customers); (ii) the evolution ofurban norms quite different from tradi-tional town values (such as the accept-ance of noninvolvement, impersonality,

13 MARCH 1970

t

c0

co>1

U-

New norm, A

II ,I II I

//T\Old norm

I m i

I / "I -

I ua ieepne e rnato

Fig. 1. Changes in the demand for timefor a given task when the overall trans-action frequency increases in a social sys-tem. [Reprinted with permission fromR. L. Meier, A Communications Theoryof Urban Growth, 1962. Copyrighted byM.I.T. Press, 1962]

and aloofness in urban life); (iii) theadaptation of the urban dweller's cogni-tive processes (his inability to identifymost of the people he sees daily, hisscreening of sensory stimuli, his develop-ment of blase attitudes toward deviantor bizarre behavior, and his selectivityin responding to human demands); and(iv) the competition for scarce facilitiesin the city (the subway rush; the fight fortaxis; traffic jams; standing in line toawait services). I suggest that contrastsbetween city and rural behavior prob-ably reflect the responses of similarpeople to very different situations, ratherthan intrinsic differences in the person-alities of rural and city dwellers. Thecity is a situation to which individualsrespond adaptively.

Further Aspects of Urban Experience

Some features of urban experience donot fit neatly into the system of analysispresented thus far. They are no lessimportant for that reason. The issuesraised next are difficult to treat inquantitative fashion. Yet I prefer dis-cussing them in a loose way to excludingthem because appropriate language anddata have not yet been developed. Myaim is to suggest how phenomena suchas "urban atmosphere" can be pinneddown through techniques of measure-ment.

The "atmosphere" of great cities. Thecontrast in the behavior of city andtown dwellers has been a natural startingpoint for urban social scientists. Buteven among great cities there are markeddifferences in "atmosphere." The tone,pacing, and texture of social encountersare different in London and New York,

and many persons willingly make finan-cial sacrifices for the privilege of livingwithin a specific urban atmosphere whichthey find pleasing or stimulating. Asecond perspective in the study of cities,therefore, is to define exactly what ismeant by the atmosphere of a city andto pinpoint the factors that give rise toit. It may seem that urban atmosphereis too evanescent a quality to be reducedto a set of measurable variables, but Ido not believe the matter can be judgedbefore substantial effort has been madein this direction. It is obvious that anysuch approach must be comparative. Itmakes no sense at all to say that NewYork is "vibrant" and "frenetic" unlessone has some specific city in mind as abasis of comparison.

In an undergraduate tutorial that Iconducted at Harvard University someyears ago, New York, London, andParis were selected as referencepoints for attempts to measure urbanatmosphere. We began with a simplequestion: Does any consensus existabout the qualities that typify givencities? To answer this question one couldundertake a content analysis of travel-book, literary, and journalistic accountsof cities. A second approach, which weadopted, is to ask people to character-ize (with descriptive terms and accountsof typical experiences) cities they havelived in or visited. In advertisementsplaced in the New York Times and theHarvard Crimson we asked people togive us accounts of specific incidents inLondon, Paris, or New York that bestilluminated the character of that par-ticular city. Questionnaires were thendeveloped, and administered to personswho were familiar with at least two ofthe three cities.Some distinctive patterns emerged

(12). The distinguishing themes con-cerning New York, for example, dealtwith its diversity, its great size, its paceand level of activity, its cultural andentertainment opportunities, and theheterogeneity and segmentation ("ghet-toization") of its population. New Yorkelicited more descriptions in terms ofphysical qualities, pace, and emotionalimpact than Paris or London did, a factwhich suggests that these are particularlyimportant aspects of New York's ambi-ance.A contrasting profile emerges for

London; in this case respondents placedfar greater emphasis on their interac-tions with the inhabitants than on physi-cal surroundings. There was near una-nimity on certain themes: those dealing

1465

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

with the tolerance and courtesy ofLondon's inhabitanits. One respondentsaid:

When I was 12, my grandfather took meto the British Museum . . . one day bytube and recited the Aeneid in Latin formy benefit.... He is rather deaf, speaksvery loudly and it embarrassed the hell outof me, until I realized that nobody waspaying any attention. Londoners are ex-tremely worldly and tolerant.

In contrast, respondents who describedNew Yorkers as aloof, cold, and rudereferred to such incidents as the fol-lowing:

I saw a boy of 19 passing out anti-warleaflets to passersby. When he stopped ata corner, a man dressed in a business suitwalked by him at a brisk pace, hit theboy's arm, and scattered the leaflets allover the street. The man kept walking atthe same pace down the block.

We need to obtain many more suchdescriptions of incidents, using carefulmethods of sampling. By the applicationof factor-analytic techniques, relevantdimensions for each city can be dis-cerned.The responses for Paris were about

equally divided between responses con-cerning its inhabitants and those regard-ing its physical and sensory attributes.Cafes and parks were often mentionedas contributing to the sense that Parisis a city of amenities, but many respond-ents complained that Parisians were in-hospitable, nasty, and cold.We cannot be certain, of course, to

what degree these statements reflect ac-tual characteristics of the cities in ques-tion and to what degree they simplytap the respondents' knowledge of widelyheld preconceptions. Indeed, one maypoint to three factors, apart from theactual atmospheres of the cities, thatdetermine the subjects' responses.

1) A person's impression of a givencity depends on his implicit standard ofcomparison. A New Yorker who visitsParis may well describe that city as"leisurely," whereas a compatriot fromRichmond, Virginia, may consider Paristoo "hectic." Obtaining reciprocal judg-ment, in which New Yorkers judgeLondoners, and Londoners judge NewYorkers, seems a useful way to take intoaccount not only the city being judgedbut also the home city that serves as thevisitor's base line.

2) Perceptions of a city are also af-fected by whether the observer is atourist, a newcomer, or a longer-termresident. First, a tourist will be exposedto features of the city different from

1466

those familiar to a long-time resident.Second, a prerequisite for adapting tocontinuing life in a given city seems tobe the filtering out of many observationsabout the city that the newcomer ortourist finds particularly arresting; thisselective process seems to be part ofthe long-term resident's mechanism forcoping with overload. In the interest ofpsychic economy, the resident simplylearns to tune out many aspects ofdaily life. One method for studyingthe specific impact of adaptation onperception of the city is to ask severalpairs of newcomers and old-timers (onenewcomer and one old-timer to a pair)to walk down certain city blocks andthen report separately what each hasobserved.

Additionally, many persons havenoted that when travelers return to NewYork from an extended sojourn abroadthey often feel themselves confrontedwith "brutal ugliness" (13) and a distinc-tive, frenetic atmosphere whose contri-buting details are, for a few hours ordays, remarkably sharp and clear. Thisperiod of fresh perception should receivespecial attention in the study of cityatmosphere. For, in a few days, detailswhich are initially arresting become lesseasy to specify. They are assimilatedinto an increasingly familiar backgroundatmosphere which, though important insetting the tone of things, is difficult toanalyze. There is no better point atwhich to begin the study of city atmo-sphere than at the moment when a trav-eler returns from abroad.

3) The popular myths and expecta-tions each visitor brings to the city willalso affect the way in which he perceivesit (see 14). Sometimes a person's pre-conceptions about a city are relativelyaccurate distillations of its character,but preconceptions may also reinforcemyths by filtering the visitor's percep-tions to conform with his expectations.Preconceptions affect not only a per-son's perceptions of a city but what hereports about it.The influence of a person's urban

base line on his perceptions of a givencity, the differences between the observa-tions of the long-time inhibitant andthose of the newcomer, and the filteringeffect of personal expectations andstereotypes raise serious questions aboutthe validity of travelers' reports. More-over, no social psychologist wants torely exclusively on verbal accounts if heis attempting to obtain an accurate andobjective description of the cities' socialtexture, pace, and general atmosphere.

What he needs to do is to devise meansof embedding objective experimentalmeasures in the daily flux of city life,measures that can accurately index thequalities of a given urban atmosphere.

Experimental Comparisons of Behavior

Roy Feldman (15) incorporated theseprinciples in a comparative study of be-havior toward compatriots and foreign-ers in Paris, Athens, and Boston. Feld-man wanted to see (i) whether absolutelevels and patterns of helpfulness variedsignificantly from city to city, and (ii)whether inhabitants in each city tendedto treat compatriots differently fromforeigners. He examined five concretebehavioral episodes, each carried outby a team of native experimenters anda team of American experimenters inthe three cities. The episodes involved(i) asking natives of the city for streetdirections; (ii) asking natives to maila letter for the experimenter; (iii) askingnatives if they had just dropped a dollarbill (or the Greek or French equivalent)when the money actually belonged tothe experimenter himself; (iv) deliber-ately overpaying for goods in a storeto see if the cashier would correct themistake and return the excess money;and (v) determining whether taxicabdrivers overcharged strangers andwhether they took the most direct routeavailable.

Feldman's results suggest some in-teresting contrasts in the profiles of thethree cities. In Paris, for instance, certainstereotypes were borne out. Parisian cabdrivers overcharged foreigners signifi-cantly more often than they overchargedcompatriots. But other aspects of theParisians' behavior were not in accordwith American preconceptions: in mail-ing a letter for a stranger, Parisianstreated foreigners significantly betterthan Athenians or Bostonians did, and,when asked to mail letters that werealready stamped, Parisians actuallytreated foreigners better than theytreated compatriots. Similarly, Parisianswere significantly more honest thanAthenians or Bostonians in resisting thetemptation to claim money that was nottheirs, and Parisians were the only citi-zens who were more honest with for-eigners than with compatriots in thisexperiment.

Feldman's studies not only begin toquantify some of the variables that givea city its distinctive texture but theyalso provide a methodological model for

SCIENCE, VOL. 167

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

other comparative research. His mostimportant contribution is his successfulapplication of objective, experimentalmeasures to everyday situations, a modeof study which provides conclusionsabout urban life that are more pertinentthan those achieved through laboratoryexperiments.

Tempo and Pace

-Another important component of acity's atmosphere is its tempo or pace,an attribute frequently remarked on butless often studied. Does a city have afrenetic, hectic quality, or is it easygoingand leisurely? In any empirical treat-ment of this question, it is best to startin a very simple way. Walking speedsof pedestrians in different cities and incities and towns should be measuredand compared. William Berkowitz (16)of Lafayette College has undertaken anextensive series of studies of walkingspeeds in Philadelphia, New York, andBoston, as well as in small and moder-ate-sized towns. Berkowitz writes that"there does appear to be a significantlinear relation between walking speedand size of municipality, but the abso-lute size of the difference varies by lessthan ten percent."

Perhaps the feeling of rapid tempo isdue not so much to absolute pedestrianspeeds as to the constant need to dodgeothers in a large city to avoid collisionswith other pedestrians. (One basis forcomputing the adjustments needed toavoid collisions is to hypothesize a setof mechanical manikins sent walkingalong a city street and to calculate thenumber of collisions when no adjust-ments are made. Clearly, the higher thedensity of manikins the greater thenumber of collisions per unit of time,or, conversely, the greater the frequencyof adjustments needed in higher popula-tion densities to avoid collisions.)

Patterns of automobile traffic contri-bute to a city's tempo. Driving an auto-mobile provides a direct means of trans-lating feelings about tempo intomeasurable acceleration, and a city'space should be particularly evident invehicular velocities, patterns of accelera-tion, and latency of response to trafficsignals. The inexorable tempo of NewYork is expressed, further, in the man-ner in which pedestrians stand at busyintersections, impatiently awaiting achange in traffic light, making tentativeexcursions into the intersection, andfrequently surging into the street evenbefore the green light appears.13 MARCH 1970

Visual Components

Hall has remarked (17) that thephysical layout of the city also affectsits atmosphere. A gridiron pattern ofstreets gives the visitor a feeling of ra-tionality, orderliness, and predictabilitybut is sometimes monotonous. Windinglanes or streets branching off at strangeangles, with many forks (as in Paris orGreenwich Village), create feelings ofsurprise and esthetic pleasure, whileforcing greater decision-making in plot-ting one's course. Some would arguethat the visual component is all-impor-tant-that the "look" of Paris or NewYork can almost be equated with itsatmosphere. To investigate this hypoth-esis, we might conduct studies inwhich only blind, or at least blindfolded,respondents were used. We would nodoubt discover that each city has a dis-tinctive texture even when the visualcomponent is eliminated.

Sources of Ambiance

Thus far we have tried to pinpointand measure some of the factors thatcontribute to the distinctive atmosphereof a great city. But we may also ask,Why do differences in urban atmosphereexist? How did they come about, andare they in any way related to thefactors of density, large numbers, andheterogeneity discussed above?

First, there is the obvious factor that,even among great cities, populations anddensities differ. The metropolitan areasof New York, London, and Paris, forexample, contain 15 million, 12 million,and 8 million persons, respectively.London has average densities of 43 per-sons per acre, while Paris is morecongested, with average densities of 114persons per acre (18). Whatever charac-teristics are specifically attributable todensity are more likely to be pronouncedin Paris than in London.A second factor affecting the atmo-

sphere of cities is the source from whichthe populations are drawn (19). It is acharacteristic of great cities that they donot reproduce their own populations,but that their numbers are constantlymaintained and augmented by the influxof residents from other parts of thecountry. This can have a determiningeffect on the city's atmosphere. For ex-ample, Oslo is a city in which almostall of the residents are only one or twogenerations removed from a purelyrural existence, and this contribuLtes toits almost agricultural norms.

A third source of atmosphere is thegeneral national culture. Paris combinesadaptations to the demography of citiesand certain values specific to Frenchculture. New York is an admixture ofAmerican values and values that ariseas a result of extraordinarily high den-sity and large population.

Finally, one could speculate that theatmosphere of a great city is traceableto the specific historical conditions underwhich adaptations to urban overloadoccurred. For example, a city whichacquired its mass and density duringa period of commercial expansion willrespond to new demographic conditionsby adaptations designed to serve purelycommercial needs. Thus, Chicago,which grew and became a great cityunder a purely commercial stimulus,adapted in a manner that emphasizesbusiness needs. European capitals, onthe other hand, incorporate many ofthe adaptations which were appropriateto the period of their increasing num-bers and density. Because aristocraticvalues were prevalent at the time ofthe growth of these cities, the mecha-nisms developed for coping with over-load were based on considerations otherthan pure efficiency. Thus, the manners,norms, and facilities of Paris and Viennacontinue to reflect esthetic values andthe idealization of leisure.

Cognitive Maps of Cities

When we speak of "behavioral com-parisons" among cities, we must specifywhich parts of the city are most relevantfor sampling purposes. In a sampling of"New Yorkers," should we include resi-dents of Bay Ridge or Flatbush as wellas inhabitants of Manhattan? And, ifso, how should we weight our sampledistribution? One approach to definingrelevant boundaries in sampling is todetermine which areas form the psy-chological or cognitive core of the city.We weight our samples most heavily inthe areas considered by most people torepresent the "essence" of the city.The psychologist is less interested in

the geographic layout of a city or inits political boundaries than in the cog-nitive representation of the city. HansBlumenfeld (20) points out that theperceptual structure of a modem citycan be expressed by the "silhouette" ofthe group of skyscrapers at its centerand that of smaller groups of officebuildings at its "subcenters" but thaturban areas can no longer, because oftheir vast extent, be experienced as fully

1467

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

articulated sets of streets, squares, andspace.

In The Image of the City (21), KevinLynch created a cognitive map of Bos-ton by interviewing Bostonians. Perhapshis most significant finding was that,while certain landmarks, such as PaulRevere's house and the Boston Com-mon, as well as the paths linking them,are known to almost all Bostonians, vastareas of the city are simply unknown toits inhabitants.

Using Lynch's technique, DonaldHooper (22) created a psychologicalmap of New York from the answers tothe study questionnaire on Paris, Lon-don, and New York. Hooper's resultswere similar to those of Lynch: NewYork appears to have a dense core ofwell-known landmarks in midtownManhattan, surrounded by the vastunknown reaches of Queens, Brooklyn,and the Bronx. Times Square, Rocke-feller Center, and the Fifth Avenuedepartment stores alone comprise halfthe places specifically cited by respond-ents as the haunts in which they spentmost of their time. However, outside the

* 0

g 2. o a a

to/idntf th loato ofec don.T

O

/* * *-9 */**/*.*

J * ./** */ * * *

. . ...

toidntf th oaino ec on.T

each point a numerical index is assignedindicating the proportion of persons ableto identify its location.

1468

midtown area, only scattered landmarkswere recognized. Another interestingpattern is evident: even the best-knownsymbols of New York are relativelyself-contained, and the pathways joiningthem appear to be insignificant on themap.The psychological map can be used

for more than just sampling techniques.Lynch (21) argues, for instance, that agood city is highly "imageable," havingmany known symbols joined by widelyknown pathways, whereas dull cities aregray and nondescript. We might test therelative "imagibility" of several citiesby determining the proportion of resi-dents who recognize sampled geographicpoints and their accompanying pathways.

If we wanted to be even more precisewe could construct a cognitive map thatwould not only show the symbols ofthe city but would measure the precisedegree of cognitive significance of anygiven point in the city relative to anyother. By applying a pattern of pointsto a map of New York City, for ex-ample, and taking photographs fromeach point, we could determine whatproportion of a sample of the city's in-habitants could identify the localespecified by each point (see Fig. 2).We might even take the subjects blind-folded to a point represented on themap, then remove the blindfold and askthem to identify their location from theview around them.One might also use psychological

maps to gain insight into the differingperceptions of a given city that are heldby members of its cultural subgroups,and into the manner in which theirperceptions may change. In the earlierstages of life, whites and Negroes alikeprobably have only a limited view of thecity, centering on the immediate neigh-borhood in which they are raised. Inadolescence, however, the field of knowl-edge of the white teen-ager probablyundergoes rapid enlargement; he learnsof opportunities in midtown and out-lying sections and comes to see himselfas functioning in a larger urban field.But the process of ghettoization, towhich the black teen-ager is subjected,may well hamper the expansion of hissense of the city. These ,are speculativenotions, but they are readily subject toprecise test.

Conclusion

I have tried to indicate some orga-nizing theory that starts with the basic

facts of city life: large numbers, density,and heterogeneity. These are externalto the individual. He experiences thesefactors as overloads at the level of roles,norms, cognitive functions, and facilities.These overloads lead to adaptive mecha-nisms which create the distinctive toneand behaviors of city life. These notions,of course, need to be examined by ob-jective comparative studies of cities andtowns.A second perspective concerns the

differing atmospheres of great cities,such as Paris, London, and New York.Each has a distinctive flavor, offeringa differentiable quality of experience.More precise knowledge of urban at-mosphere seems attainable through ap-plication of the tools of experimentalinquiry.

References and Notes

1. New York Times (15 June 1969).2. L. Wirth, Amer. J. Soc. 44, 1 (1938). Wirth's

ideas have come under heavy criticism by con-temporary city planners, who point out that thecity is broken down into neighborhoods, whichfulfill many of the functions of small towns. See,for example, H. J. Gans, People and Plans:Essays on Urban Problems and Solutions (BasicBooks, New York, 1968); J. Jacobs, The Deathand Life of Great American Cities (RandomHouse, New York, 1961); G. D. Suttles, TheSocial Order of the Slum (Univ. of ChicagoPress, Chicago, 1968).

3. G. Simmel, The Sociology of Georg Simmel,K. H. Wolff, Ed. (Macmillan, New York, 1950)[English translation of G. Simmel, Die Gross-stadte und das Goistesleben Die Grossstadt(Jansch, Dresden, 1903) ].

4. R. L. Meier, A Communications Theory of Ur-ban Growth (M.I.T. Press, Cambridge, Mass.,1962).

5. S. Milgram and P. Hollander, Nation 25, 602(1964).

6. B. Latan6 and J. Darley, Amer. Sci. 57, 244(1969).

7. S. Gaertner and L. Bickman (Graduate Center,The City University of New York), unpublishedresearch.

8. D. Altman, M. Levine, M. Nadien, J. Villena(Graduate Center, The City University of NewYork), unpublished research.

9. P. G. Zimbardo, paper presented at the Ne-braska Symposium on Motivation (1969).

10. J. Waters (Graduate Center, The City Univer-sity of New York), unpublished research.

11. W. McKenna and S. Morgenthau (GraduateCenter, The City University of New York),unpublished research.

12. N. Abuza (Harvard University), "The Paris-London-New York Questionnaires," unpub-lished.

13. P. Abelson, Science 165, 853 (1969).14. A. L. Strauss, Ed., The American City: A

Sourcebook of Urban Itnageri' (Aldine, Chi-cago, 1968).

15. R. E. Feldman, J. Persontaliti' Soc. Psi chol. 10,202 (1968).

16. W. Berkowitz, personal communication.17. E. T. Hall, The Hidden Dimension (Doubleday,

New York, 1966).18. P. Hall, The World Cities (McGraw-Hill, New

York, 1966).19. R. E. Park, E. W. Burgess, R. D. McKenzie,

The City (Univ. of Chicago Press, Chicago,1967), pp. 1-45.

20. H. Blumenfeld, in The Quality of Urban Life(Sage, Beverly Hills, Calif., 1969).

21. K. Lynch, The Image of the City (M.I.T. andHarvard Univ. Press, Cambridge, Mass., 1960).

22. D. Hooper (Harvard University), unpublished.23. Barbara Bengen worked closely with me in

preparing the present version of this article. Ithank Dr. Gary Winkel, editor of Environmentanzd Behavior, for useful suggestions and advice.

SCIENCE, VOL. 167

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from

The Experience of Living in CitiesStanley Milgram

DOI: 10.1126/science.167.3924.1461 (3924), 1461-1468.167Science 

ARTICLE TOOLS http://science.sciencemag.org/content/167/3924/1461

REFERENCES

http://science.sciencemag.org/content/167/3924/1461#BIBLThis article cites 5 articles, 1 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceGovernment Works. The title reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S.Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights

(print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for theScience

on June 11, 2019

http://science.sciencemag.org/

Dow

nloaded from