Visions of Nature and Landscape Type Preferences: an Exploration in the Netherlands W.T

Landscape and Urban Planning 63 (2003) 127–138

Visions of nature and landscape type preferences: an exploration in The Netherlands W.T. de Groot a,∗, R.J.G. van den Born b a Social Environmental Science, Nijmegen University and Centre of Environmental Science, Leiden University, P.O. Box 9104, Nijmegen 6500 HE, The Netherlands b Social Environmental Science, Nijmegen University, Nijmegen, The Netherlands

Abstract A survey was carried out among inhabitants of Gennep, a small-town municipality in the east of The Netherlands, questioning about (1) the types of nature that people distinguish and the levels of naturalness ascribed to these types of nature; (2) the images that people hold of the appropriate relationship between people and nature and the level of adherence to these images, and (3) people’s preference of broadly defined landscape types. Types of nature inferred by means of factor analysis were labeled arcadian, wild and penetrative nature, the last category comprising elements such as mosquitoes and rats in the barn. Factor analysis was used as well to infer images of appropriate relationship, which appeared to hinge around the concepts of mastery over nature, responsibility for nature and participation in nature, respectively. Levels of adherence to the responsibility and participation images were very high, indicative of a ‘new biophilia’ mainstream in Dutch culture. Landscape types were defined, roughly, as (1) landscape made by and for people; (2) park-like, arcadian landscape; (3) wild, interactive landscape and (4) landscape “in which one may experience the greatness and forces of nature”. Strikingly, more than half of the respondents expressed preference for this last (‘deep ecology’) landscape type, with another third preferring the wild, interactive landscape. The highest preference of the ‘greatness and forces’ landscape was found among the respondents with high ascription of naturalness to the penetrative type of nature and respondents adhering to the participation-in-nature image of relationship. Preferences for landscape types as defined here should be distinguished from visual or behavioral landscape preferences. Even if people may not select landscapes of the ‘greatness and forces of nature’ in daily behaviors, these landscapes of wilderness and greatness do connect with human visions and desires. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Landscape; Nature; The Netherlands; Attitudes; Naturalness; Relationship with nature; Wilderness; Deep ecology; NEP

1. Introduction Norway and Sweden, an average of 80% of the respondents acknowledge the intrinsic value of nature, that In Western countries in general and in highly ur- is, nature’s right to exist irrespective of its uses and banized ones such as The Netherlands in particular, functions for humankind (Grendstad and Wollebaek, remarkable levels of nature-friendliness are currently 1998). In surveys in The Netherlands, this percentage found to exist within the general public. In surveys in is usually 90% or higher (Van den Born et al., 2001). This ‘new biophilia’, as it is sometimes called, may be ∗ Corresponding author. Tel.: +31-24-3611581; hypothicated as a new cultural phase of the Western so- fax: +31-24-3615957. cieties, that is now succeeding the previous phases of E-mail address: [email protected] (W.T. de Groot). nature’s conquest and domestication (De Groot, 1999).

Apart from this aspect of the value of nature, Van erences to the two aforementioned elements of Van den Born et al. (2001) have coined the term ‘visions den Born et al.’s (2001) visions of nature. These two of nature’ as an umbrella that comprises two more elements concern relatively abstract categories and elements: in order to arrive at a proper connection with the landscape concept, we have separated the landscape • the ‘images of nature’, defined as the types of nature concept into categories that are likewise broad and that people distinguish, such as, in Western culture, abstract, hence without specific visual details such wild nature or arcadian nature; as openness, foliar density or foreground/background • the ‘images of relationship’, defined as the images distinction, and without specific ecosystem content that people hold of the appropriate relationship be- such as forest, wetland, grassland or sea. The land- tween humans and nature, examples of which are scapes thus defined, of which we have distinguished dominion or guardianship. only four, are here called ‘landscape types’. Images of nature have been subject to empirical Thus, our main research questions were: (1) What research before (see Section 3). The same holds for images of nature do people distinguish? (2) What im- aspects of the people-nature relationship, e.g. be- ages of the appropriate relationship with nature do tween nature and health (Kaplan and Kaplan, 1989; people distinguish? (3) What preference do people Frumkin, 2001), nature and children (Kahn, 1999; have for landscape types? (4) How do these prefer- Nevers et al., 1997) and nature and farmers (Aarts, ences relate to the images of nature and the images 1998; Kaltoft, 1999); Van den Born et al. (2001) give of the appropriate relationship with nature that people an overview. The more ethical/philosophical ‘images adhere to? of relationship’ as defined above, however, are yet Below, the next section introduces the methodology. to be addressed empirically. The present paper will Sections 3–5 then report on the results of the first three present the first results in this area. research questions separately, and Section 6 presents ‘Nature’ in these types of research is understood, the interconnections. Section 7 summarizes these re- roughly, as everything that lives or organizes itself sults and Section 8 gives the overall discussion and outside humans and human decisions. The concept conclusion. of ‘landscape’, on the other hand, usually stands for the sensory (most often visual) aspects of nature, arti- facts and their mixtures, usually taken on a relatively 2. Research approach and method large scale. Human preference for different landscapes has been the object of much literature, e.g. Coeterier With respect to the images of nature (research ques- (1987), Herzog et al. (2000) and Misgav (2000), that tion 1), a quantitative research method has been well connects the visual qualities of landscapes with pref- established (see below), and many data are already erence statements of respondents. available. The component was included, however, in Nature-friendliness may sometimes be glimpsed order to establish the relationships with the images of indirectly from landscape preference research. Ulrich relationship (research question 2) and the landscape (1986) and Purcell and Lamb (1998), for instance, type preferences (research question 3). assert that the perceived degree of naturalness in land- The latter two research questions have never been scapes is a powerful factor in the preference that peo- addressed empirically yet. The obvious advantage of ple have for these landscapes; see Van den Berg (1999, this situation is that in terms of substance, innova- p. 118) for group differences within this general pic- tive results may be expected (see Sections 4 and 5). ture. Respondents in these landscape-oriented surveys Methodologically, however, the issue is how to orga- or interviews are not invited, however, to express their nize the first empirical exploration. One design is to ideas about nature as such or their relationship with first validate the research concepts in qualitative inter- nature. Hence, no relationships between landscape views, and then move to quantitative survey in order preferences and views on nature may be established. to establish the distribution of these categories over The present paper is designed to fill this gap. Our the population. The other research design is to accept strategy has been to connect stated landscape pref- less validated research concepts for the time being and W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 129 focus on distributions first. Based on such a quantita- through the regular mail; eight were returned to sender tive exploration, it becomes much better known which due to wrong address, and a total of N = 172 com- categories are the most relevant for further qualitative pleted questionnaires were returned, representing a validation and enrichment. We have chosen for the response of 35% (which is high for The Netherlands). latter approach, after an informal validity test that In terms of educational level, the sample contained checked the researchers’ interpretations of the ques- 21.6% of respondents with lower education (pri- tionnaire items with a number of non-researchers. mary school and lower vocational training), 47.6% Furthermore, the chosen method of factor analysis with middle-level education (high school or voca- also contains a validity check; typologies used by tional middle-level), and 29.0% with higher education the researchers for the ‘top–down’ formulation of the (higher vocational training or university). These fig- questionnaire items, if not re-produced in the factors, ures match well with The Netherlands as a whole may be rejected as non-valid (compare, for instance, (Maters, 2000). The age distribution matched as Laumann et al., 2001). well. The same holds for the political affiliations, Against this background, a survey was designed. It with social democrats, conservative liberals and focused on the municipality of Gennep, southwest of Christian-democrats, in that sequence, in the lead, as Nijmegen town. Gennep was selected for its normal- they are in The Netherlands as a whole. By and large, ness; it has no special demography or features such as then, there are no reasons to suspect serious sampling a university, a large chemical factory, a special land- errors. scape or ongoing landscape projects that could set The images of nature were elicited in the same it apart from the rest of The Netherlands in terms way as done by Bervaes et al. (1997), Buijs and of nature and landscape visions. The municipality is Filius (1998), Van den Berg (1999) and Van den semi-rural; out of its 16,000 inhabitants, 9000 live Born et al. (2001), based on the work of Rosch in Gennep town. The small-town character of Gen- and Mervis (1975) on cognitive schemata. Existing nep ensures that our data will not be biased towards expert-based listings of types of nature (e.g. Van ‘eco-centric’ visions that are often associated with den Born et al., 2001) were used by the researchers more urbanite respondents (Van den Berg, 1999). The to generate concrete types-of-nature instances. The landscape surrounding Gennep is nice without being listings distinguish between, for instance, ‘domestic special, dominated by dairy farm grasslands but also nature’, ‘wild nature’, ‘functional nature’ etc. and containing the Maas river and floodplain, the smaller the types-of-nature examples then are, for instance, and more idyllic Niers river, a recreation lake and some ‘cats and dogs’, ‘the rainforest’ and ‘meadows’, re- patches of forest and heath land. spectively. These instances were used as items in the The sample for the survey was drawn randomly questionnaire, of which respondents were asked to from the municipality’s telephone directory. This indicate (on a three-point scale) the degree to which generates two biases in the sample. First, not all they “associate these items with real nature”. The re- inhabitants are listed, because they have a secret num- sponses were subjected to a factor analysis (extraction ber or no telephone. Overall in The Netherlands, the method: principal axis factoring; rotation method: total of these groups is <15%, however; we have ac- varimax with Kaiser rotation), a technique that ex- cepted this source of possible bias for an exploratory tracts the items to which respondents tend to assign study. The second bias stems from the fact that within the same degree of being “real nature”, whether high, families, the husband is often the only one listed, or middle or low. These new groups of items (‘factors’) the one filling out the questionnaire anyway. Thus in may then be labeled by the researchers, forming a our sample, it turned out that 60% of the respondents new, empirically grounded typology of images of were male, as compared to 52% in Gennep as a whole nature (see, for instance, Gorsuch, 1974: p. 186 and (Maters, 2000). Although gender differentiation in Kim and Mueller, 1994: p. 113 on the interpretation attitudes towards nature is most often found to be of identified factors). Since our purpose was to con- insignificant (e.g. Van den Berg, 1999), this calls for nect the images of nature with landscape preferences some caution when generalizing for the population as rather than identifying new images, the questionnaire a whole. Five hundred questionnaires were sent out items largely coincide with those used by Van den 130 W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138

Table 1 Results of the factor analysis of images of nature Image of nature Items Factor score Mean level of naturalness S.D.

Arcadian nature Lambs in the meadow 0.751 1.52 0.63 Grain fields 0.740 1.23 0.70 Pollard willows 0.474 1.72 0.50 Birds brooding in the grassland 0.397 1.69 0.58 Polder ditches 0.396 1.42 0.64 Penetrative nature Mosquitoes 0.821 1.38 0.73 Rats in the barn 0.627 0.84 0.78 Weeds in the garden 0.468 1.58 0.60 Elementary nature The North Pole 0.736 1.91 0.36 The sea 0.583 1.95 0.25 The wind 0.505 1.94 0.23 Earthquakes 0.314 1.81 0.52 A swamp A swamp 0.555 1.91 0.32 The items are those of the questionnaire. The images are the ‘factors’ identified by the statistical procedure and named by the researchers. Factor scores are the degrees to which the items fit into the factor (image of nature) as a whole. The mean levels of naturalness ascribed to the items are calculated by assigning 2 to “full”, 1 to “a bit” and 0 to “no” association with real nature. S.D. is the standard deviation of the mean. ‘Elementary nature’ and ‘a swamp’ may be combined to form ‘wild nature’.

Born et al. (2001), only somewhat reduced in number; includes its own spirituality, in which humans may they are listed in Table 1. participate. The same factor analysis method was used for the As is not difficult to imagine, formulating instances elicitation of images of the appropriate relationship that may express these images in a questionnaire is an between people and nature. As a starting point, we as yet tentative job. The first attempt of Van den Born used the typology of ‘relationship images’ shown in et al. (2001) had not been successful. Trying a second Table 2, based on the work of De Groot (1992) who summarized from Dutch and American environmental Table 2 philosophy. These images range from the most an- Images of the relationship between humans and nature thropocentric image of man the technocrat-adventurer, Degree of Images conquering nature and trusting technology to fix all anthropocentrism problems that might arise, up to the most ecocentric Anthropocentrism Man the technocrat-adventurer image of ‘Oneness with nature’, i.e. the New-Age Man the manager-engineer ‘deep ecologist’ re-uniting with nature, the divine Man the steward of nature process of being. In-between, we find the anthro- Non-anthropocentric Man the guardian of nature pocentric but more cautious man the manager of Man and nature as partners Man as participant in nature nature, the well-known image of stewardship, and Oneness with nature (‘Unio mystica’) three non-anthropocentric images, defined by that in these, an intrinsic value of nature has been acknowl- Based on Van den Born et al. (2001). The abstract singular “Man” here comes closer to the Dutch questionnaire term “de mens” than edged. In the guardianship image, humans are seen does the English concrete plural “humans”; in Dutch, “de mens” as lording over an intrinsic yet not fully up-to-level is ungendered. The images are ordered in decreasing order of value. The image of partnership between humans anthropocentricity. The main dividing line is drawn at the point and nature is much more one of horizontal, dynamic of recognition of real intrinsic value in nature. The first three exchanges in which values (e.g. non-dominance and images roughly represent the traditions of the three mainstream currents in Dutch politics (conservative liberals, social democrats intensity) are acknowledged also of these exchanges and Christian-democrats, respectively). The partnership image is themselves. In the image of participation in nature, largely a Dutch product (e.g. De Groot, 1992). ‘Oneness with nature is seen as an all-encompassing whole that nature’ is related to New-Age and deep ecology philosophy. W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 131

Table 3 Result of the factor analysis of images of relationship between humans and nature Factors and items Factor score Level of adherence S.D.

Man the adventurer and exploiter of nature Technology and science can solve environmental problems in the future (TAME) 0.507 0.50 1.01 Economic growth is necessary to solve environmental problems (TAME) 0.446 −0.22 1.12 Humans have the right to use nature (TAME) 0.403 0.56 1.05 Although nature has value of itself, humans stand above it (TAME) 0.582 −0.37 1.28 Mean of ‘Adventurer and exploiter’ 0.14 Man responsible for nature People are part of nature (PR) 0.435 1.46 0.70 It is our duty to conserve nature for future generations (SG) 0.405 1.79 0.56 Nature may blossom if people interact with it in the right manner (PR) 0.458 1.68 0.53 We should not stand above nature, but work together like partners (PR) 0.478 1.32 0.77 We have to take care that plants and animals keep their own places to live (SG) 0.505 1.74 0.47 Mean of ‘Man responsible for nature’ 1.60 Man the participant in nature I would love to once join the wild geese on their journey (PT) 0.740 0.39 1.24 I would like to be alone for a month in the heath lands, feeling at one 0.669 −0.15 1.26 with my surroundings (PT) The spiritual aspect of nature is very important (PT) 0.332 0.45 1.06 Mean of ‘Man the participant in nature’ 0.23 The items are literally those of the questionnaire. After each item, it is marked which theoretical image(s) of relationship (Table 1) were used to generate the item (technocrat-adventurer and manager-engineer: TAME; steward and guardian: SG; partner: PR; participant and ‘oneness’: PT). The images are the factors identiﬁed by the statistical procedure and named by the researchers. They are arranged in increasing order of ecocentricity. Factor scores are the degrees to which the items ﬁt into the factor (image of relationship) as a whole. The levels of adherence are calculated by assigning 2 to “strongly agree”, 1 to “agree”, 0 to “neutral”, −1 to “disagree” and −2 to “strongly disagree”, and averaging over the respondents. S.D. is the standard deviation of the mean.

time for the present questionnaire, we were satisfied Bervaes et al. (1997), Buijs and Filius (1998) and Van by some of our inventions, such as the item of “I would den Born et al. (2001) had indicated already that most love to once join the wild geese on their journey”, respondents would fall into the non-anthropocentric designed to express the participation image. Others, group, only one landscape type was designed for to our feeling, do not seem fully adequate yet and are association with the two most anthropocentric im- open for improvement, e.g. by way of open interviews ages of Table 1: “A well-ordered landscape, made by or stronger links with recent empirical research on the and for people”. The other extreme, “A landscape in spiritual aspects of nature such as Williams and Harvey which one may experience the greatness and forces (2001). The full list of items is in Table 3, where it is of nature”1, was designed for association with the indicated after each item which image of relationship it two ecocentric extremes. The other two, the English was designed to express. We used a five-point scale to park landscape and the rough, natural outback were measure level of agreement, assigning +2 to “strongly designed for association with the stewards/guardians agree”, +1 to “agree”, 0 to “neutral”, −1 to “disagree” of nature and the partnership-with-nature adher- and −2 to “strongly disagree”. ents, respectively. As touched upon already in the The landscape types were put into a single questionnaire question, in which respondents were asked 1 As a linguistic note, it may be remarked here that the Dutch to express their preference (“This landscape I like the original “groots/grootsheid” is somewhat more confined to bigness ... and hence less strongly positive than the English “great/greatness”. best: ”). The landscape types were designed as In Dutch, all four landscape types items were designed to have a most likely preferences of respondents adhering to cer- slightly positive connotation (in keeping with Ulrich, 1993), but tain images of relationship. Since previous research of not one stronger than the other. 132 W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 introduction, these verbal formulations leave the re- The overall level of naturalness ascribed to this image spondent free to think about any concrete landscape of nature is 1.27 on the scale between 0 and 2. while filling out the questionnaire. The landscape of The third image, grouping the North Pole, the sea, ‘greatness and forces of nature’ may be the wide-open the wind and the earthquakes, has been called ‘ele- spaces for one respondent, the sea for an other, or the mentary nature’. As may be seen in the Table, the high mountains for the next. Thus for all landscape swamp is left as a lone item not grouped with any other types, the responses are independent of the tastes of in the factor analysis. Since it shares its wild charac- respondents in terms of concrete landscapes such as ter with elementary nature and the fact that virtually the sea or wetland or forest. the same number of respondents associates the swamp The data were analyzed by means of the SPSS sta- with real nature, we will group elementary nature and tistical software. the swamp together in the rest of our analysis, calling it ‘wild nature’. The overall level of naturalness ascribed to this group is the highest of the three, with a 3. Images of nature mean of 1.90 on the scale between 0 and 2. In order to identify relations with background vari- The scree-plot of the factor analysis indicated that ables such as age and gender, as well as to prepare a distinction between four factors (images of nature) for the analysis of possible association with landscape yields the most significant classification. Table 1 type preferences, respondents were classified into cat- shows the factor scores through which items are egories of those who characteristically assign higher grouped together by the factor analysis, as well as degrees of naturalness to arcadian, penetrative and the average levels of naturalness ascribed to these wild nature, respectively. Two selection strategies may images. The latter levels were calculated by assigning be distinguished in this respect. The first is to simply 2 to “full”, 1 to “a bit” and 0 to “no” association with select all respondents who assign “full” association real nature. with real nature to all items of the three respective We named the first group of items, headed by the groups. This has the disadvantage, however, that many lambs in the meadow (factor score 0.751), ‘arcadian respondents may be excluded rather arbitrarily, and nature’, because the items grouped here suggest the that respondents who assign a high level of naturalness peaceful, small-scale harmony of people and nature to many items would be selected for possibly all three in the traditional arcadia of Western culture (Schama, groups. In order to arrive at a less arbitrary and more 1995; Eisenberg, 1998). The inclusion of the grain mutually exclusive classification, a procedure was fol- field could be seen as slightly contradictory to this. lowed that expresses that the typical ‘arcadian’ respon- The reason of its inclusion could be, however, that dent is one who will not ascribe high naturalness to respondents were not offered more items such as ‘a such very non-arcadian things as mosquitoes or rats soccer field’ that could have formed a group of ‘utility in the barn, while at the same time another category nature’ together with the grain field, as in Van den should be formed by respondents who (rightfully) as- Born et al. (2001). The overall level of naturalness sume that the rat in the barn is in fact a more free, ascribed to arcadian nature is 1.52 on the scale between more natural creature than the very arcadian lambs in 0 and 2. the meadow. Thus, the classification became: The second image of nature, headed by the mosquitoes (factor score 0.821) could be interpreted 1. Characteristic respondents ascribing a high level of as joining negative elements of nature but on the naturalness to arcadian nature are those with a level other hand, the (dangerous) earthquake and swamp of naturalness assigned to arcadian nature that is are excluded here and in the longer list of items in (a) higher than the mean and (b) higher than the Van den Born et al. (2001), the dandelion growing level of naturalness assigned to penetrative nature. in the pavement is included in this group. Following With these criteria, 74 respondents were selected. Van den Born et al. we labeled this image of nature 2. Characteristic respondents ascribing a high level of as ‘penetrative nature’, i.e. the type of nature that naturalness to wild nature are those with a level creeps into places we have designed to be our own. of naturalness assigned to wild nature that is (a) W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 133

higher than the mean and (b) higher than the level Since we had generated our questionnaire items ba- of naturalness assigned to arcadian nature. With sically from four images (technocrat-adventurer and these criteria, 104 respondents were selected. manager-engineer, steward and guardian, partner, and 3. Characteristic respondents ascribing a high level of participant), our preference was for the four-factor so- naturalness to penetrative nature are those with a lution. It turned out, however, that although the two ex- level of naturalness assigned to penetrative nature tremes of this set were readily interpretable, the middle that is (a) higher than the mean and (b) higher than two were not; both were mixtures of steward/guardian the level of naturalness assigned to arcadian nature. and partner items. Thus, we decided to choose the sim- With these criteria, 51 respondents were selected. pler, three-factor solution, given in Table 3. As may be seen in the Table, the factors come out with good statis- This classification yielded the following results tical clarity. Jointly they explain 32% of the variance. (with levels of significance in Pearson’s χ 2-test added As for the substantive interpretation of the three fac- in parentheses). The category of characteristic respon- tors, Table 3 shows that the first and the last factor dents ascribing a high level of naturalness to arcadian coincide with the theoretical images (Table 2) used to nature was composed of relatively many women generate the questionnaire items. In other words, the (0.01), with age and level of education non-significant. strongly anthropocentric theoretical images of man the The category of characteristic respondents ascribing technocrat-adventurer and man the manager-engineer, a high level of naturalness to wild nature was com- as well as the strongly ecocentric theoretical images posed of relatively many men (0.04), people in the of participation in nature and oneness with nature, 35 to 54-age bracket (0.03) and people with higher have been reproduced empirically. Naming these two education (0.03). The category of characteristic re- factors ‘Man the Adventurer and Exploiter of nature’ spondents ascribing a high level of naturalness to and ‘Man the Participant in nature’ respectively, as penetrant nature was composed of relatively many well as placing them first and last on the anthropocen- men (0.01), people in the 45 to 54-age bracket (0.05) tricity/ecocentricity scale, was therefore unambiguous. and people with higher education (0.06). The analysis also showed that the ‘middle images’ of Overall, the analysis shows that respondents distin- Table 1 (man the steward, guardian and partner of na- guish between a consistent set of images of nature. ture), although coming out well as a group, were not As said in the previous section, our purpose here has reproduced separately through this survey. It cannot been mainly to classify our respondents in Van den be known yet if this muddledness is present in reality, Born et al.’s (2001) images-of-nature system, prepar- implying that lay people do not agree with the envi- ing to investigate the association with landscape type ronmental philosophers that the three images denote preferences. As it turned out, the only difference be- relevantly different ethical stances, or that the ques- tween Table 1 and Van den Born et al. is that the latter tionnaire items through which we have tried to elicit have generated two additional images of nature (‘do- the images are as yet not fully adequate. At this point, mesticated nature’ and ‘utility nature’), due to the fact we have given the whole group the heading of ‘Man that they included more items in their questionnaire, Responsible for Nature’, and taken this image as point including ones such as ‘cats and dogs’ and ‘a soccer of departure for further analysis. field’. The levels of naturalness ascribed to the images Table 3 also shows the levels of adherence on of arcadian, penetrative and wild nature in our and the five-point scale between −2 and +2. With a their surveys are basically equal (after transferring the mean level of adherence of 1.60 on this scale, the scales). responsibility-for-nature image is clearly Dutch mainstream culture at present. This contrasts strongly with the assertions of environmental philosophers, even 4. Images of the people-nature relationship Dutch ones such as Zweers (1995), that the Cartesian image of ‘Man the Master and Possessor of nature’ The scree-plot of the factor analysis indicated that (or ‘Man the Despot’) is still dominant in Western a distinction between three or four factors (images of societies. In Table 3, the equivalent ‘Adventurer and relationship) yields the most significant classification. Exploiter’ image has an overall level of adherence of 134 W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 only 0.14, meaning that the negative responses of re- translation, the questionnaire items were: spondents to the items of this image almost outweigh the positive ones; the Cartesain vision has become 1. A well-ordered landscape, made by and for people. a subcurrent within the general public. On the other 2. A varied, park-like landscape. end of the scale, we encounter the strongly romantic, 3. Untamed nature, with which one may have many spiritualizing elements of ‘Man the Participant’. With interactions. its overall level of adherence of 0.23 this is a subcur- 4. A landscape in which one may experience the rent too but jointly with the mainstream Responsiblity greatness and forces of nature. image, it is a strong empirical grounding of Van den Born et al.’s (2001) ‘new biophilia’. Following De Groot (1999), the first of these In order to identify relations with background vari- items was designed to appeal to the strongly an- ables and landscape type preferences, respondents thropocentric (‘Adventurer and Exploiter’) respon- were classified as adherents of the three images. The dents, the second to the arcadian and stewardship classification procedure was that in order for a re- types of respondents, the third to appeal to the spondent to be classified in one of the images, the partnership-with-nature image, and the fourth to the respondent had agreed or strongly agreed with all participant and oneness-with-nature ideas. With that, items making up that image. This yielded 12, 127 we expected the second image to draw most of the and 25 respondents in the three categories of Ad- respondents, and the fourth only a spiritualistic few. venturer/exploiter, Responsible and Participant cate- The result, however, was the reverse. As shown in gories, respectively. The classification was considered Table 4, the preferences were 4% to the man-made satisfactory because these numbers of respondents landscape, 13% to the park-like landscape, 31% to reflect the overall levels of adherence. untamed, interactive nature, while a majority of 51% The analysis of association with the background of the respondents preferred the landscape of ‘great- variable yielded the following results (with levels of ness and forces of nature’. Surprising though this is, it significance of Pearson’s χ 2-test again in parenthe- should be borne in mind that responses such as these ses). The category of respondents adhering to the may be highly dependent on the type of questions and ‘Adventurer and Exploiter’ image was composed of the research setting. This issue will be taken up in the relatively many people over 55 years of age (0.04) and Section 8. many people with a low level of education (0.01), with Relationships were tested between landscape type gender influence insignificant. These levels of signifi- preference and the background variables of age, gen- cance are remarkable because they exist in spite of the der and level of education. We mention levels of low number of adherents of this image; with respect significance in Pearson’s χ 2-test in parentheses. As- to level of education, for instance, 8 of the 12 respon- sociation with age turned out to be non-significant dents in this category had only primary school or lower (0.14). Only very slight tendencies seemed to exist that vocational training. The large mainstream category of people above 55 have more preference for the ordered adherents to the Responsibility-for-nature image was and park-like landscapes, while people below 45 are composed of somewhat more younger people (0.01), less interested in these landscape types. The landscape with influences of gender and level of education in- of greatness and forces of nature is certainly no desire significant. The category of Participants-in-nature was of only the young; the preference for this landscape is composed of somewhat less people with lower levels high in all ages except above 65, and slightly elevated of education (0.05) with influences of age and gender in the 25 to 54-age bracket. Relationship with gender insignificant. was squarely insignificant (0.69). The relationship with level of education was well clustered (0.05), with the lowest educated preferring the man-made and 5. Landscape type preferences park-like landscapes and the highest educated preferring the ‘greatness and forces’ landscape even more In the survey, respondents were asked to indicate than the average. Overall, these findings are congru- one preference between four landscape types. In literal ent with those of Van den Berg (1999) who studied W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 135

Table 4 The association of levels of naturalness ascribed to the three images of nature with preference for the four landscape types N Landscape type preference (%)

Man-made Park-like Untamed, interactive Experience of landscape landscape landscape greatness and forces All respondents 171 3 14 32 52 Arcadian nature is natural, more so 74 1 15 40 43 than penetrative nature Wild nature is natural, more so than 104 0 12 33 55 arcadian nature Penetrative nature is natural, more so 51 0 10 24 66 than arcadian nature The method of categorization of the respondents is given in Section 3. The landscape types labels (‘man-made’ etc.) summarize the questionnaire formulations given in Section 5. The items that make up the images of nature are in Table 2. N is the number of respondents. visual landscape preferences in The Netherlands, and should be interpreted against the background of the theories such as those of Ulrich (1993), which suggest strong overall (‘all respondents’) preference of the that personal security and confidence (factors which ‘wild’ and ‘greatness’ landscape types, discussed in are likely to be more prevalent among the higher the previous section. It then appears that: educated and not-too-old) are associated with higher preference for wilder, less secure landscapes. 1. Respondents ascribing a high level of naturalness to arcadian nature appear to be somewhat less charmed than the average by the landscape of ‘greatness and forces’, but this relationship is not 6. Associations of landscape type preferences significant in the χ 2-test. with visions of nature 2. Respondents ascribing a high level of naturalness to In this section, the associations are described be- wild nature tend to move their preferences towards tween landscape type preferences on the one hand the wilder landscapes, especially that of the ‘great- and visions of nature, i.e. the levels of naturalness that ness and forces’ (significance 0.01 in Pearson’s χ 2 respondents ascribe to elements of their environment, -test). and their ideas of the appropriate relationship between 3. Respondents ascribing a high level of naturalness to people and nature, on the other hand. Based on the the- penetrative nature move even stronger in that direc- oretical notions put forward in the previous sections, tion, with 66% preferring the latter landscape (sig- a number of associations were expected, e.g. (1) asso- nificance is only 0.06 due to their lower number). ciation of the idea that arcadian nature is natural with Overall, this outcome confirms the expectations. preference for the most traditionally arcadian, ‘park- Table 5 shows the association between land- like’ landscape type, (2) association of the idea that in scape type preference and adherence to images of fact penetrative nature is more natural with preference people-nature relationship. Again, the figures should for the more wild types of landscape, (3) and a more be interpreted as variation around the general (‘all or less one-to-one association of adherence to the respondents’) rejection of man-made and preference three images of the people-nature relationship with for wild and great landscapes. So doing, it appears that: preference of the four landscape types: the adventurer/ exploiter with the man-made landscape, the ‘steward/ 1. Adherents to the Adventurer and Exploiter im- guardian/partner’ with the park-like and the interac- age have a higher-than-average preference for the tive landscapes, and the participant-in-nature with the man-made and park-like landscape, while rejecting ‘greatness and forces of nature’ landscape type. the ‘greatness and forces’ landscape. Table 4 shows the result of the cross tabulation 2. Adherents to the Responsibility-for-nature im- of images of nature and landscape types. The figures age have a lower-than-average preference for 136 W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138

Table 5 The association of adherence to images of people-nature relationship with preference for the four landscape types N Landscape type preference (%)

Man-made Park-like Untamed, interactive Experience of greatness landscape landscape landscape and forces All respondents 170 3 14 32 52 Man the adventurer and 12 17 33 42 8 exploiter of nature Man responsible for nature 127 1 8 32 60 Man the participant in nature 25 4 4 12 80 The method of categorization of the respondents is described in Section 4. The landscape types labels (‘man-made’ etc.) summarize the questionnaire formulations given in Section 5. The items that make up the images of people-nature relationship are in Table 3. N is the number of respondents.

the man-made and park-like landscapes, and a responsibility for nature is the broad mainstream, somewhat stronger-than-average preference for the and the spiritual/romantic image of participation in landscape of ‘greatness and forces’. nature is present as an undercurrent that has gained 3. Adherents of the Participant-in-nature image are more strength than the mastership image. These find- almost completely (80%) clustered in preference ings represent the first empirical confirmation of the for the ‘greatness and forces’ landscape. evolutionary theory of De Groot (1999), who conjec- tured that in the rapidly urbanizing settings such as All these relationships were significant in Pearson’s The Netherlands, new protective attitudes and a new χ 2-test at a level of 0.02 or lower. Overall, they con- longings for nature are rising to cultural dominance. firm the expectations. Landscape type preferences were investigated by means of verbal formulations (hence not photographs) 7. Summarizing conclusion that left respondents free in terms of ecosystem types or visual aspects such as foreground/background de- In the foregoing, it has been shown that people tails. Man-made and traditionally arcadian landscapes in Gennep distinguish between a consistent set of (“well-ordered”, “park-like”) were preferred by only images of nature such as arcadian nature and wild a small minority of the respondents, while the two nature. Levels of naturalness ascribed to wild nature wild and experiential landscape types, formulated with elements were 1.90 on a scale running between 0 terms such as “untamed” and “the greatness and forces and 2. On that same scale, elements of ‘penetrative of nature”, were preferred jointly by more than 80% nature’ such as mosquitoes or rats in the barn still of the respondents. This preference was even stronger have an ascribed level of naturalness of 1.27, indi- among the higher educated. cating that respondents have a wider view on nature Associations between landscape type preferences than including only what is nice and green. and the images of nature follow expected patterns. Applying the same methodology but new in terms Respondents who ascribe higher levels of naturalness of substance, we also found that people in Gennep to wild and penetrative nature than to arcadian nature distinguish between three images of the appropri- have an even higher preference than the average for ate relationship of people and nature, and that these the wild and great landscapes, running up to 66% images, speaking in terms of mastership of nature, preference for the ‘greatness and forces’ landscape responsibility for nature or participation in nature, of the ‘penetrant nature’ respondents. As for the im- partly reproduce images articulated in environmental ages of people-nature relationship, adherents of the philosophy. The degree to which respondents ad- ‘Adventurer and Exploiter’ image appear to have a here to these images, however, contrasts strongly to higher-than-average preference for the man-made what philosophers usually assert; the mastership im- and park-like landscape types, while the other two age is only an undercurrent among the respondents, groups lean stronger than average toward the wild W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138 137 and great landscapes, running up to a preference for Another point to note is that the respondents the ‘greatness and forces’ landscape type of 80% of have filled out their questionnaires in the safety and the Participant-in-nature respondents. warmth of their homes, and not while being swept around or threatened in the landscape of “experi- 8. Discussion ences of the greatness and forces of nature” where they so intensely appear to want to be. In this con- Van den Berg (1999: p. 63) studied the association text, it must be borne in mind that what has been of two variables allied to the images of relationship and identified in our study might represent the highest the landscape type preference. Related to the images rung on a ladder that goes down from people’s gen- of relationship, a degree of ecocentricity was mea- eral dreams and desires to people’s specific thoughts sured by means of the widely used NEP scale (Dunlap and behaviors in the varied situations of daily life. and Van Liere, 1978). Related to the landscape type With the eye on landscape preference research, we preferences, Van den Berg measured preferences for may consider this ladder to consist of (1) preferences photographed landscapes. Contrary to our results, for landscape types elicited as in the present paper, no association of the two variables was found. One i.e. in general and verbal terms, (2) preferences for reason may be, as Van den Berg puts it, that “accep- landscapes elicited as in most landscape preference tance of ecocentric ideas has become so widespread studies, i.e. visualized and depicting more concrete that new measures are needed to capture ... people’s landscapes, and (3) preferences for landscapes as ex- cognitions regarding the relationship between humans pressed in concrete behaviors of daily life, such as and nature”. The items shown in Table 3 may be the picnics, play, recreation or recuperation from daily beginning of such a new scale ‘beyond NEP’. stress. The most important methodological caveat that Obviously, preferences do not need to be constant may be brought in against the data presented here is when going up or down the ladder. People may ex- that both their substance and their internal coherence press a preference for the wild open spaces on the may be biased by socially desirable responses. The highest level and yet, on the behavioral level, spread questionnaire items we have analyzed were part of their picnic blanket in a cosy corner of the forest. This a larger questionnaire that also contained items on is not to say that people are “inconsistent”, as is often water and water management, but did not touch on thought when discrepancies between attitudes and be- non-environmental issues such as crime, traffic jams haviors are found. In terms of our landscape types, it or ethnicity. Implicitly, it must also have been clear only means that people neither desire to have ‘great- to virtually all respondents that the survey had been ness and forces’ everywhere continuously, nor a world sent to them by a nature-friendly university depart- composed of picnic sites without end. Or putting it ment. Thus, a bias may have arisen that only the more positively, it may all be true that on the rungs of more nature-friendly respondents may have returned the ladder in people’s cognitive and value schemata, the questionnaire and that within this group, people the great blue whale should swim the ocean even if have tended to be nice to the department and give only for us to dream about, the wilderness should be nature-friendly answers. We have of course been there even if peak experiences of wilderness solitude aware of this and have addressed the issue in all are rare, the recreational landscapes should be there to ways possible, inter alia by choosing Gennep in the admire their visual beauty, the picnic sites should be first place, by using factor analysis instead of direct accessible, cosy and safe, and nature around the block questions and by formulating the items such that they should be our children’s challenging playscape. equally fall in line with people’s overall attitudes. The steps and variants between dreams and behav- Also, the consistency of the results and the distribu- iors also have their relevance for nature and landscape tion of background variables (see Section 2)donot policies. Visual landscape preferences are obviously indicate a strong bias. It does appear to be preferable relevant for landscape protection and design, as are in general, however, to embed this type of questions behavioral preferences for policies with respect to the in surveys that also address non-nature issues, such zoning of national parks, the design of amenities, ur- as the ones mentioned already. ban parks planning and so on. 138 W.T. de Groot, R.J.G. van den Born / Landscape and Urban Planning 63 (2003) 127–138

Analogously, the landscape type preferences as Herzog, T.R., Herbert, E., Kaplan, R., Crooks, C., 2000. Cultural identified in the present study appear to have a policy and developmental comparisons of landscape perceptions and relevance in terms of their own substance and level preferences. Environ. Behav. 32 (3), 323–346. Kahn, P.H., 1999. The Human Relationship with Nature: of scale. In The Netherlands, for instance, the Wad- Development and Culture. The MIT Press, Cambridge. den Sea, the coast, the great lakes and big skies are Kaltoft, P., 1999. Values about nature in organic farming practice our landscapes of ‘greatness and forces of nature’. and knowledge. Sociologia Ruralis 39 (3), 39–53. The river floodplains and the Veluwe forests are our Kaplan, R., Kaplan, S., 1989. The Experience of Nature. potential wildernesses of ‘untamed and interactive A Psychological Perspective. Cambridge University Press, Cambridge. nature’. All other countries in Western Europe have Kim, J., Mueller, C.W., 1994. Factor Analysis: Statistical Methods areas of those qualities and potentials. Protecting and Practical Issues. In: Lewis-Beck, M.S. (Ed.), Factor and developing landscapes such as these, irrespec- Analysis and Related Techniques. Sage Publications, London. tive of visual beauty or recreational-behavioral use, Laumann, K., Gärling, T., Stormark, K.M., 2001. Rating scale is of great value not only for biodiversity or nature measures of restorative components of environments. J. Environ. Psychol. 21 (1), 31–44. as such; it also connects directly with human visions Maters, G., 2000. Water-en natuurbeleving in Gennep. Een and desires as explored in this paper. kwantitatief onderzoek naar waterbeleving, grondhoudingen en natuurbeelden. Doctoraalscriptie. KUN, Nijmegen. Misgav, A., 2000. Visual preference of the public for vegetation Acknowledgements groups in Israel. Landsc. Urban Planning 48, 143–159. Nevers, P., Gebhard, U., Bilmann-Mahecha, E., 1997. Patterns of The authors wish to thank Grietje Maters for the reasoning exhibited by children and adolescents in response to moral dilemmas involving plants, animals and ecosystems. J. data gathering and Otto Hermkens for the support in Moral Educ. 26 (2), 169–186. the data analysis. Purcell, A.T., Lamb, R.J., 1998. Preference and naturalness: an ecological approach. Landsc. Urban Planning 42, 57–66. References Rosch, E., Mervis, C.B., 1975. Family resemblances: studies in the internal structure of categories. Cognitive Psychol. 7, 159–177. Schama, S., 1995. Landscape and Memory. Harper-Collins Aarts, N., 1998. Een kwestie van natuur. studie naar de aard Publishers, London. en het verloop van communicatie over natuur en natuurbeleid. Ulrich, S.R., 1986. Human responses to vegetation and landscapes. Wageningen University, Wageningen. Landsc. Urban Planning 13, 29–44. Bervaes, J., Buijs, A.E., Filius, P., Volker, C., 1997. Draagvlak voor Ulrich, S.R., 1993. Biophilia, Biophobia and Natural Landscapes. natuur? Peiling bij het publiek en maatschappelijke organisaties. In: Kellert, S.R., Wilson, E.O. (Eds.), The Biophilia Hypothesis. DLO-Staringcentrum, Wageningen. Island Press, Washington, DC. Buijs, A.E., Filius, P., 1998. Natuurbeelden in de praktijk. De Van den Berg, A.E., 1999. Individual Differences in the Aesthetic invloed van natuurbeelden en natuurvisies op gedrag en mening Evaluation of Natural Landscapes. Groningen University, over het beleid. DLO-Staringcentrum, Wageningen. Groningen. Coeterier, J.F., 1987. De waarneming en waardering van lands- Van den Born, R.J.G., Lenders, R.H.J., de Groot, W.T., Huijsman, chappen. Resultaten van omgevingspsychologisch onderzoek. E., 2001. The New Biophilia: an exploration of visions of nature Wageningen University, Wageningen. in Western countries. Environ. Conserv. 28 (1), 65–75. De Groot, W.T., 1992. Evironmental Science Theory: Concepts and Williams, K., Harvey, D., 2001. Transcedent experience in forest Methods in a Problem-Oriented, One-World Paradigm. Elsevier, environments. Environ. Psychol. 21, 249–260. Amsterdam. Zweers, W., 1995. Participeren aan de natuur. Jan van Arkel, De Groot, W.T., 1999. Van vriend naar vijand naar verslagene en Utrecht. verder: een evolutionair perspectief op de verhouding tussen mens en natuur. Nijmegen University Press, Nijmegen. Wouter T. de Groot started out as a civil engineer, worked in Dunlap, R.E., Van Liere, K.D., 1978. The ‘New Environmental Kenya and moved into environmental science in the 1970s, teaching Paradigm’: a proposed measuring instrument and preliminary and publishing in the fields of environmental ecology, economics, results. J. Environ. Educ. 9, 10–19. philosophy, methodology, land use theory and social science. He Eisenberg, E., 1998. The Ecology of Eden. Picador, London. now is professor in social environmental science at the Leiden and Frumkin, H., 2001. Beyond toxicity; human health and the natural Nijmegen Universities, The Netherlands. environment. Am. J. Prevent. Med. 20 (3), 234–240. Gorsuch, R.L., 1974. Factor Analysis. Saunders, Philadelphia. Grendstad, G., Wollebaek, D., 1998. Greener still? An empirical Riyan J.G. van den Born has a masters in social environmental examination of Eckersley’s ecocentric approach. Environ. science and now teaches this discipline at Nijmegen University; Behav. 30 (5), 653–675. her work focuses on visions of nature and rural issues.