Ovadia Salama in collaboration with Alexander Tzonis

Ways of reducing crime and increasing security in high-rise housing are receiving appropriate attention. Taking earlier

studies into other use areas will require a framework.

Authors: Ovadia Salama teaches planning methodology at Columbia University, and has written numerous papers on inter- faces between technology, public projects, and social-psycho- logical factors. Alexander Tzonis teaches at Harvard's Graduate School of Design, and is the author of books and articles on architecture, planning and art. Research undertaken at Abt Associates, Inc., in Cambridge, Mass. by both authors forms the basis for this article.

In order to build upon the seminal work of Oscar Newman (Newman, 0., Defensible Space, Macmillan, New York, 1972 and PIA, Oct. 1972, p. 92), the authors have developed a framework for generating crime-reducing design features in a variety of environments. Our approach uses the potential for generalization and for cross-environmental transfer of Newman's findings in the area of public housing. The result of our investigations will enable us to delineate, for each of the four environments of concern-private residences, schools, transportation systems, and commercial areas:

Which environmental functional characteristics are associated with crime occurrence or nonoccurrence;

Which functional characteristics are to be included in, or excluded from, successful designs; and

What specific physical features these functional characteristics suggest as programmatic components in the design of experimental projects.

Our approach is outlined in the light of three general examples drawn from Newman's experience. The general concepts underlying these examples are stated, using his empirical evidence and theoretical approach as guides to our analysis.

This enables us to outline a systematic framework of logically interrelated concepts, describing generic environmental design conditions that impede crime. The framework takes a "tree" structure form including alternative strategies for making the environment more defensible. It will then be possible to show how the three examples from Newman's book fit into this structure and how they represent particular cases of fecund generic conceptual approaches. In the near future, we plan to infer from these generic concepts specific design consequences which can be tested in experimental projects.

In analyzing Newman's suggestions we must first illustrate the distinction introduced above between functional characteristics and design features. For example, if "restrict access to the building through the roof" is the functional characteristic, the associated specific design features include: "eliminate skylights, fortify lock on roof door, install electronically assisted visual and/or auditory surveillance of roof door, post guard on roof, eliminate exterior fire escapes," etc. By functional characteristics, therefore, we mean the functions achieved by the design features as they relate positively or negatively to the criminal activity.

The first example we analyze is the presence of an elevator in the building. Newman points out that this design feature indicates the existence of a circulation channel which offers limited visual access to anyone watching the building and consequently provides good surveillance potential to superintendents. The functional characteristic of interest here is "visual access by supervisor." The presence of an elevator can, however, reflect more general relationships than the specific relationship between the elevator user and the building supervisor. Such relationships (also functional characteristics) pertain to interactions between user and device and can be characterized as direct feelings towards the device. For example, feelings of gratification could result from the user's association of the elevator with technological progress and social improvement: he may feel socially improved since he belongs to the community in which social progress has meaning associated with technological advance. Conversely, feelings of alienation could result if the user associates the mechanical nature of the elevator with his own exclusion from the larger community; it could be seen to emphasize isolation, anomie, and ultimately rejection of the individual from the group. Whether gratification or alienation will result depends on the "many differences in family structure and lifestyles" (Oscar Newman, Defensible Space, p. 189) which lead individuals to interpret design features according to their own unique "codes."

High-rise buildings, our second example from Newman's work, can also be viewed in similar light. Currently these buildings can be recommended or rejected on the grounds of their having a fragmented circulation network and multiple accesses, both of which are difficult to supervise. Newman, however, introduces a distinction which goes beyond this level of analysis and states that some factors which make high-rise buildings workable for middle-income housing make them unworkable for low-income housing. The functional characteristic involved here is related to the difficulty (in a low-income housing development) of distinguishing an intruder from a resident. This functional characteristic suggests the possibility of having the building supervision carried out by the community of users itself, and not by a professional who, generally, is not a regular resident. Newman points out that efficient control by the residents is contingent upon social-psychological variables (op. cit., p 193), such as gratification, belongingness, and respect for order. Implementation of design features which reinforce these social-psychological characteristics can deter intruders – and residents – from crime and vandalism.

Let us consider finally the example of setting "play facilities for infants at each floor level of an apartment building." New- man has established that these are desirable features for crime prevention in public housing (op. cit., p. 206). The factors involved here are 1) the; increased number of actors in the process of supervising circulation and access, 2) their presence on the premises during work hours, arid 3) the in- creased probability of recognizing intruders resulting from a high level of interaction – and therefore of cognisance – among residents. All of these factors, of course, refer to the control of circulation channels. However, in addition to these factors, play facilities often increase the cohesion of the resident group and the individual's sensitivity to the common interests to the group. This reduces the chances of criminal behavior from within the community of residents but, more importantly, it gives the potential intruder an image of cohesion which strongly deters him from selecting this community as a target. At the same time, the residents gain encouragement to act as a group in the face of an intruder.

From the above discussion of some examples drawn from Newman's experience, we can infer that strategies for reducing crime occurrence in public housing projects include one or more of the following functional characteristics: 1) observability of the movements of the individuals, 2) identification of intruders, 3) willingness to act to apprehend the intruder, 4) ability to act to apprehend the intruder, 5) development of an image of community that acts as a deterrent, 6) reduction of provocation, and 7) a reminder of the protective and punitive nature of society.

These functional characteristics can be achieved in various ways by introducing alternative design features. The examples discussed include design features which contribute in varying degrees to the achievement of one or more of these functional characteristics. However, the advantage of introducing the concepts underlying specific design features (as opposed to working directly with the design features them- selves) is that they suggest alternative design options and thus are more easily generalized and transferred between the areas of housing, schools, transportation, and commerce.

A procedure for the systematic generation of all relevant functional characteristics involves a feedback process between the literature on environmental design and crime prevention on one hand, and the progressively refined logical classification of the design features of these environments, in terms of the functions they achieve, on the other hand. For example, two broad categories can first be identified among the multiple strategies used for crime reduction through environmental design. These two categories are 1) those actions which are exogenous to the intruder (such as those involving tighter protection of accesses) and 2) those actions which are internal to the intruder, such as those which tend to modify his behavior. The former are effective immediately after they are implemented but present various shortcomings because they may tend to transfer crime to alternative areas, they are not self-enforceable in every case, and they may have negative consequences on other environmental features, particularly on the social-psychological impact of the environment on its habitual users. In contrast, strategies oriented toward modifying the behavior of the intruder have effects that be- come apparent in the longer term, but these effects are longer lasting, self-enforcing, and have more positive effects on the users' social-psychological climate.

Within the first category (i.e., actions exogenous to the intruder), one can further distinguish controls exerted by professionals who are not residents (or, more generally, users) and those which are exerted by the residents themselves. In the first case, circulation channels, accesses, and egresses are supervised by professionals-guards, superintendents, policemen, doormen, etc.-who are assigned specific locations to control, and who mayor may not be assisted by various technological, visual, and auditory surveillance tools.

These professionals must be on the premises at a given time or all the time. In the second case, control is provided by residents who are present on the premises at random times, stay on these premises for a short duration, and leave. They may or may not be replaced by other persons coming on the premises for the same or different purposes. In general, these purposes are not primarily related to building supervision. Continuity in the presence of some resident is a random process, as opposed to a deterministic process in the first case.

In the "tree" diagram, the concepts which emerged from the above discussion have been numbered as items 1, 2, 1.1, and 1.2. Then, as examples of the way the procedure can be developed, we have indicated additional branches and blocks. The tree can thus be expanded step by step. Each step of this expansion requires, in addition to systematic reasoning, inputs from architects, planners, and designers as well as from a literature review, along the lines we have followed to analyze the examples from Newman's work. The development of this alternative strategies tree (generic approaches for organizing space to reduce crime incidence) is a fundamental step toward the identification of specific de- sign features. The next steps are to generate classifications of crimes according to type, to the circumstances of their occurrence, to the four environments of interest, and to the places in those environments where the crimes can occur. Then, considering each of these environments in turn, one can match the classification of crimes with the various functional characteristics which are the block entries in our "tree" structure. This matching can lead to the identification of a number of desirable alternative design features in order to prevent crime. The most promising of these features are to be included in trial projects to determine their efficiency.

We must note here that a given design feature can contribute to the realization of more than one functional characteristic and therefore can be more effective than a design feature which contributes only to a single strategy. The play areas discussed previously exemplify this fact. Furthermore, by using the "tree" structure, it becomes easier to measure interaction effects of the design features recommended. Thus, if feature A is present and increases the security by its contribution say, to strategy 1.1 and if simultaneously B is present and contributes to strategy 1.2, it is possible to identify the joint effect of A and B by looking at the way strategies 1.1 and 1.2 are interrelated in the tree hierarchy. In other words, the combined effect of A and B is estimated by looking at interactions of functional characteristics.

There are two additional types of effects which the tree allows us to isolate. First, there can be negative interactions. For example, as discussed above, exogenous controls may have a negative impact on the development of some types of internal controls. Secondly, the existence of some functional characteristics is a prerequisite for the effectiveness of other functional characteristics. The clearest example here is the Kitty Genovese incident that occurred some years ago in New York. In this incident both visual and auditory surveillance of the incident occurred, but there was not, tragically, enough sense of community to motivate any of the observers to intervene in any way. Thus, those mechanisms which depend on the casual surveillance of user (i.e., those in branch 1.3 of the tree) depend on the development of certain attitude and behavior sets on the part of users (i.e., those in 2.1).

In order to better understand the tree structure, it is necessary to examine how the examples drawn from Newman's work fit in it, and then to show how the tree can be used to generate testable design features.

From our analysis of the effect of an elevator's presence in a building, we concluded that the elevator's negative effect on security stems from the fact that it creates a channel of circulation which is not visually controlled. The elimination of the elevator falls then under strategy 1.2.1.4. Other effects of the elevator associated with feelings of alienation relate the absence of the elevator to strategy 2.2.2. Thus, eliminating an elevator from the program of the building contributes to these two functional characteristics. The presence of an elevator, however, may also have positive impacts according to strategy 2.1.2. In addition, of course, the elimination of elevators may be highly undesirable in view of other design considerations, primarily building height. In the same way, the effects of building high-rise structures can be seen to fall under 1.2.1.4 (accesses and circulations which are numerous with respect to the number of attendants), under 1.2.1.2 (accesses and circulations which are not visually controlled) and under 1.3.1 (accesses and circulations which require "random walk" as a means of supervision). Finally, the inclusion of play facilities in a building contributes clearly to strategy 1.3.2.2 (setting permanently used activities on premises) to strategy 1.3.1 (use of "random walk" of resident as a means of super- vision) and to strategy 2.1.1 (changing the behavior of residents by increasing their cohesion as a group through the development of communal activities).

Let us note that our tree structure allows us to explore several alternative strategies within the same generic group. For example, 1.3.2.2.2 suggests that it may be more effective to locate the play facilities, not on every floor as recommended by Newman, but close to circulation intersections; strategy 1.3.1.2 used in combination with 1.3.2.2.2 suggests that laundry facilities can be placed close to play facilities, so that mothers can watch their children play, while the fact that the children are present and playing deters the intruder. In addition, the juxtaposition of these two activities addresses functional characteristic 2.1.1.

We have mentioned that the value of the functional characteristics isolated to build our tree structure lies in the fact that they are potentially easy to generalize and transfer to and among environments such as schools, transportation systems, commercial centers, and private residences. We shall illustrate this statement by developing examples of physical features which can be associated with strategy 1.3.2.2.2 which consists of setting permanently or frequently used facilities at intersections or access points.

In the case of a school, let us assume that the crimes to be avoided are burglaries during which photographic equipment or typewriters are stolen. Let us further assume that the most frequently used facility is the cafeteria. Application of strategy 1.3.2.2.2 suggests, then, locating the cafeteria close to the intersection of heavily frequented corridors and preferably as close as possible to egress (or access) points. Note, however, that we may choose, due to social-psychological considerations, to employ a functional characteristic very different from 1.3.2.2.2, such as 1.2.1.2, physically controlling all ac- cess and egress by attendants. If we assume, however, that the crimes to be avoided are assaults, rather than burglaries, then functional characteristic 1.3.2.2.2 suggests that we seek to prevent the school from being deserted during part of the day by calling for activities on its premises which will take place when children are away. Such activities can be community meetings, family movies, continuing education, etc. The presence of such activities can be made possible if the school facilities are adaptable. Hence, a variety of physical features are required which can be specified.

In the case of transportation systems, strategy 1.3.2.2.2 would suggest setting newsstands (or other similar, busy activities operating most of the day) at various locations in the corridors of a subway station. When considering commercial centers, we may find that drugstores are the best facilities for locations close to entrances, with visual access to parking lots and/or storage areas, where most crimes take place. Drugstores indeed remain open longer hours than other shops and usually sustain a high degree of activity. (Of course, anonymous crowds of users may not tend to discourage crime the way groups of users identified with the environment would and thus, by considering functional characteristics in branch 2.1 we may not choose 1.3.2.2.2).

Finally, private residences offer a more complex situation where privacy, status, and comfort requirements conflict with such crime preventing strategy as 1.3.2.2.2. Nevertheless, this functional characteristic has possible applications here, too. It suggests considering the street itself as "the most frequented" place of activity. Using this concept, we find that maximum security might be achieved by removing all landscaping from the street (if it restricts visibility) and by shaping the lots so that they present shorter fronts and more depth than normal. This tends to intensify the use of the street in a sense that the number of doors per linear foot is greater. Visual access is thus heightened, as is the probability of having neighbors coming in and out, using the street, and identifying intruders. (Naturally, this trades off with functional characteristics that help the user identify with his space and set up symbolic barriers; it is, however, an alternative that should at least be considered in the context of its virtues and liabilities.)

In general, since there is a continuum of the concept of privacy from private residences to schools, to commercial areas, to transportation systems, each functional characteristic, such as 1.3.2.2.2, which relates to user surveillance (and, desirably, the possibility of intervention) must be examined closely before transfer is attempted. And in those cases where such functional characteristics are transferable, the design features appropriate for implementing them may be very different for different environments.

Although the above illustrations are not suggested security features, they point out the method's potential. When this procedure is actually put to use by a team of architects, planners, and engineers, it will generate creative solutions in a systematic way more effectively than by trial and error.