Forestry Commission logo

Overview of Research Findings

Coverage and Gaps in Empirical Studies

There has been a huge amount of work, much of it by foresters, on psychophysical preference by non-experts, as well as aesthetic and value judgments by experts. Ribe (1989) has made a useful summary of many of the former, and in particular those that use variations on and developments of the Scenic Beauty Estimation Method (SBE). Zube, in 1982, summarised a large body of work from the mid 1960s onwards and showed that there is a lack of a unifying theory to inform the way humans and landscape interact and how this relates to preference and behaviour outcomes. He underlines the gap between theory and empirical work focused on application and practice.

It would seem that there is little that is new to be learnt from using SBE variants on the psychophysical approach to preference. Similarly, it is possible that the Kaplans have exhaustively explored their framework of mystery, complexity, legibility and coherence, although it appears that there may be more to be learnt by building on this than on the SBE.

Both of the above approaches, however, rely heavily on photographs and are thus not dealing with perceptions from engagement with the real world. It is interesting to read Zube, quoting Ittleson's 1973 list of properties which relate to the real landscape (as opposed to a photograph or video) and to be reminded how much they mirror Humphry Repton's 18th century riposte (set out in his Red Book for Holme Park of 1793) to the picturesque school of Uvedale Price and Richard Payne Knight (Malins, 1966), pointing out the difference between landscape painting and landscape gardening. Ittleson identified the following properties.

  1. landscapes surround - they permit movement and exploration and force the observer to become a participant
  2. landscapes are multimodal - information is received through multiple senses and processed (broadly speaking) simultaneously
  3. landscapes provide peripheral vision as well as central, from behind and in front, in and out of focus
  4. landscapes provide more information than can be used - they can simultaneously provide redundant, inadequate, ambiguous, conflicting and contradictory information
  5. landscape perception always involves action - landscapes cannot be passively observed, they provide opportunity for action, control, manipulation
  6. landscapes always have an ambience - they are mostly encountered as part of a social activity, they have a definite aesthetic quality and have a systemic quality (various components and elements are related)

Repton would add the quality of continuing change, whether measured hourly, daily, or annually, as an essential landscape quality - weather (especially in Britain), sun angle, humidity, etc., are constantly changing influences on any landscape experience, and Foster (1991) has recognised this in underlining the limited temporal applicability of environmental judgments.

Russell and Snodgrass (1987) state that what we plan to do in a landscape, and whether that plan is frustrated, assisted or modified, will affect our perception of the landscape (this is not the same as Gibson's concept of "affordance", although it may be affected by it). We cannot divorce what we see from what we might do. In this context, photographs are only valid as representations of views of landscapes from a fixed seat, or out of a window, but into which we will never go, and even then they are a flawed, because surrogate, representation of reality.

The Kaplans, in work with photographs, (1989) attempt to address some of these issues when exploring "smoothness" and "ease of locomotion" in views, although these constructs are part of an expert judgement and not made directly by the subjects of the landscape preference study. Preference studies based simply on asking for a preferred photograph view can overlook the importance of the way landscapes call forth action and involvement, for example campers, when asked which photograph of a landscape within a forest they preferred, chose an image with screening and ground cover vegetation but, when the landscape close to them offered the opportunity, they chose to camp in a clear place with little ground cover vegetation (Arthur, Daniel and Boster, 1977). Similarly, Hull and Stewart (1995) have shown that a distant view will call forth a different response from that to a close view. In real life engagement with the landscape, the view within 15m of the viewer is the most important and most often focused upon.

The sequence of experience of landscape is important in how we perceive it and will influence preference. A few studies have explored this in different ways (including Hull and Stewart, 1995), but many aspects remain largely unexplored. It is clear that what a person has seen in the landscape immediately before influences the response to what is seen next. It is also evident that familiarity and habituation influence visual preference, so views of the same scene at the start and end of a journey may evoke different responses. The influences of familiarity, habituation and acculturation mean that tourists are likely to exhibit different responses to locals (Wohlwill, 1974, Russell and Lanius, 1984)

Multisensory and multimodal engagement with the landscape has been touched on by only a handful of researchers. Some work has been done on blind/visually impaired perception and sound perception in the landscape (Anderson et al., 1983, Porteous and Mastin, 1985b), but most work in these domains has been carried out using artificial surrogates for the real environment. Porteous (1985a) says that 90% of our perceptual intake is believed to be visual and much of the rest is auditory and tactile. He has looked (via literature) at perceptions of smell. Children before puberty are much more sensitive to smell than after and less negatively affected by smells in general. It is evident that preferences after adolescence are culturally conditioned and habituation is very important - we get used to smells quickly. But memory of odour does not decay over time (unlike visual memory), perhaps because smell is primitively linked directly to the brain. Smell also involves affect strongly and cognition only a little (the opposite of visual stimuli, according to Porteous). This raises some fascinating issues with regard to smell: the importance of what pre-adolescent children experience and the strength of that memory if re-encountered later in life.

The involvement of affect and emotional response in landscape also seems worth exploring further, and again, not just related to vision. Ulrich (1981) suggests that investigations using physiological or medical measures (cf. intuitive/subjective procedures) have been very much more successful in motivating government action and public concern regarding environmental quality. It is possible, for example, to measure arousal (as Ulrich has) using an electrocardiograph and measures of alpha-amplitude in brain electrical activity. This might be a fruitful way forward, especially if there are techniques that can be carried out in the field. It is possible also to use saliva samples to measure stress levels or subjective responses to mood scales such as PANAF (Positive Affect, Negative Affect) or ZIPERS (Zuckerman Inventory of Personal Reactions) (see Ulrich, 1981).

The exploration of 'smellscapes' by Porteous (1985a) also highlighted the desirability of looking at different age responses to landscape. Some work has already shown (e.g. Lyons, 1983) that preference is influenced by age. Pre-adolescents are more extreme in their views, and the variance in extremes of like/dislike declines with old age. The question of consensus vs. individual or categories of group response is important and perhaps preference studies should look more closely at this variation (and that of gender, ethnicity, etc.) rather than simply seeking the predominant or apparently consensual response.

The work of Purcell (see Purcell, Lamb et al., 1994) is important here too, in exploring how we learn about landscapes and store and categorise that information. Purcell's work suggests that experience of landscape is structured around categories or more complex mental representations such as schemata, which are the result of long term exposure to regularities in the environment. He suggests landscape perception is prototypically based and the prototype is stored in the memory more strongly than other category members. He has explored the relationship between typicality of a scene and affective experience. Judgement of landscape typicality is based on relatively abstract attributes of the landscape and the concept 'landscape' may mask a diversity of types of environments and mixtures of types.

Despite Ulrich's urging towards physiological/medical measures for political expediency, phenomenological methods also seem to hold out a useful way forward, dealing as they do with full engagement with the landscape. The challenge in landscape analysis techniques is that they should stand up well under criteria of reliability, sensitivity, validity and utility (Daniel and Vining, 1983). Some phenomenological approaches may produce reliable and sensitive data on an individual basis only (e.g. Brook, 1998) whose interpretation, by its very nature, may not be generaliseable across large groups, and thus of limited utility.

Developments of Personal Construct Theory techniques, some of which have now moved away from the overly rigid repertory grid methods used by Harrison and Sarre (1975) (like the SBE, too reliant on sophisticated statistical manipulation of measurements of dubious numerical comparability) offer ways forward. These explore the meaning perceived in the landscape through bipolar mental constructs which are personal to each individual. Furthermore, through laddering techniques, they tap into core values held by an individual. Such methods can handle multidimensional responses and have been used successfully with pre-adolescent children as well as adults (Aspinall and Ujam, 1992, Ward Thompson, 1995). They might assist a move from over-simplistic consensus-based measures of landscape preference towards information-gathering which is sensitive to age difference, gender, experience and socio-cultural background, for example, and whether experienced in groups or individually.

One aspect of landscape preference that has barely been touched on in the literature is microclimate. The sensual information received through skin temperature and touch is the least well served by perception studies, yet such sensations as warmth, coolness, humidity or windiness are, in practice, a very strong determinant in how people use and interact with the landscape and, therefore, what type of landscapes are preferred. The microclimate of a location in the landscape is an element over which landscape architects and foresters can have a strong and direct effect, yet the landscape preference literature almost completely ignores it. There would seem to be a clear case for further exploration of this aspect and several techniques, both phenomenological and physiological, which might readily assist in empirical studies.

The experience of taste in the landscape, although not immediately promising as a research focus, might also be explored. There is the element of direct taste of elements found in the landscape - e.g. blackberries or spring water - which may not be regularly part of the landscape experience but an important and vivid sub-set of experience. There is also the issue of how food/drink tastes in the landscape, which might well turn out to be an important aspect of landscape preference relating to how people plan to use and interact with their environment to make an everyday experience more pleasurable.

It is clear that passive visual perception alone does not determine our preferences and actions in the landscape. Harvey (1995) has shown that, when people were asked about how well a particular landscape would tolerate change, information about the landscape was a more important influence than what it looked like, in determining people's responses. Other work at Edinburgh College of Art/Heriot-Watt University (Hope, work in progress) has been trying to develop a profile for aspects of place that encompasses affect, cognition and the role of critical assessment, which in turn influences decision-making. As many have said before, decision-making is ultimately the key to perception and preference that is of most vital concern to landscape architects and foresters, since decision-making will determine how people interact with the landscape and the results (both intentional and otherwise) which derive from this interaction.

Finally, a summary of those areas of research which might usefully be pursued, either because they would fill gaps in present coverage in the literature, or because they would take further promising techniques and methods that have not yet fully been explored, is listed below.

Report Contents