Irwin Silverman and Jean Choi
Considerations of space impact virtually every aspect of the organism's adaptive behavior, including the search for food, water, and shelter; predator avoidance; mating strategies; social structure; and parenting. In this chapter, we focus on the evolutionary approach to two aspects of spatial behavior in humans and nonhuman species. The first is navigation: How do individual travels systematically from place to place? The other is landscape preference: the kinds of places in which the individual chooses to live or visit.
Watch a dog leap over a fence leaving barely enough room to clear, a child throw a ball on a near perfect arc to a target, a honeybee return directly to its hive after a meandering search for food. Ponder the physical laws of space and motion inherent in all these abilities. In no case are the subjects aware of these laws or capable of generalizing from them, yet the actions are performed with ease.
This is illustrative of the prevailing paradigm of evolutionary psychology, which seeks to replace the traditional social science model of mind as primarily consisting of general purpose learning mechanisms. The evolutionary-based perspective is that mind is comprised of functionally independent or semi-independent domain-specific cognitive mechanisms (Cosmides & Tooby, 1992), which evolved in response to specific problems for adaptation and survival in the organism's evolutionary history.
The areas of spatial perception and behavior afford excellent examples of this model. The honeybee's navigational skills enabled it to forage over a relatively large area. The canine's ability to leap over a barrier facilitated both hunting and escape from predators. The human's capacity to accurately throw a projectile also aided in hunting and in agonistic encounters with conspecifics.
Though the concept of domain specificity may ultimately revolutionize our views of mind, contemporary theories of cognitive psychology still remain largely domain general.
Gallistel's (1990) model, frequently cited in the spatial navigation literature, provides a case in point. Gallistel contends that the representational and computational rules presumed to underlie learning and problem solving in adult humans can be applied to all cognitive capacities of all animal species. In regard to spatial navigation, he says:
The fact that dead reckoning computations for unrestricted courses are sufficiently complex to have imposed restrictions on the courses human navigators followed would seem an intuitive argument against the hypothesis that the nervous system of infrahuman animals like the ant routinely and accurately perform such computations. I raise this point to argue specifically against such intuitions, which I believe have been an obstacle to the acceptance of computational-representational theories of brain function. Symbolic manipulations that seem complex, hard to learn and difficult to carry out by human beings often have simple physical realizations. Integrating a variable with respect to time sounds like an impressive operation, yet a bucket receiving a flow of water integrates that flow with respect to time. The filling of a bucket strikes most people as a simple physical operation. A symbolic (mathematical) presentation of the trigonometric and integrative operations involved in dead reckoning computations makes them sound forbidding, but the dead reckoning device on a ship is not complex. The trigonometric, decomposition operations it performs are easily simulated with plausible neural circuits. (1990, pp. 38–39)
What are the essential differences between models?
The domain-specificity approach of evolutionary psychology focuses on ecological requirements that could account for the evolution of a given attribute, navigational skill, or otherwise, in a given species. The goal is to uncover the evolved cognitive mechanisms mediating the behavior, however these are best described. In contrast, Gallistel's (1990) approach has taken cognitive psychologists in a different direction, the search for the expression of a general set of representational and computational rules for navigational behaviors across situations and species.
In this respect, the concept of representational/computational rules seems reminiscent of equipotentiality, a core principle of behaviorism that maintained that all reinforcers were equally effective for all behaviors of all species. Garcia and colleagues (Garcia, Ervin, & Koelling, 1966; Garcia & Koelling, 1966) discredited that notion in a series of studies on taste avoidance in rats that were seminal to the concept of domain specificity. They demonstrated that organisms were genetically programmed in an evolutionarily adaptive manner, such that particular behaviors conditioned only to particular reinforcers in particular species. Rats, scavengers by nature, could only be conditioned to avoid a particular flavor by the negative reinforcer of induced nausea.
A model that seeks to apply a single set of rules to navigational behavior across species will also be misled by unique, species-specific mechanisms. Bees and some other insects seem to navigate by using the sun as a compass, but are actually following planes of polarized light (von Frisch, 1967). Migrating locusts would appear to defy representational/computational analysis in that individuals are seen flying in different directions within the swarm, but this is the locusts' way of keeping the group on course, despite wind shifts (Rainey, 1962).
Finally, it is reasonable to assume that the evolutionary development of animal and human navigational abilities began with simpler mechanisms than are described in a representational/computational analyses. This does not exclude from consideration the higher order abstractions unique to humans that have enhanced our navigational capacities to the point where we can travel to space and back, but these were derived in relatively recent evolutionary time, as a function of our emergent general analytical abilities. It does not follow that these analyses are somehow embedded in mechanisms that have worked effectively from prehistory.
Navigation is the process by which an animal uses available cues to travel to predetermined locations. The nature and extent of travel, however, varies greatly among species. Some cover relatively short distances in their lifetimes, whereas others migrate halfway around the world. Attempts to explain these differences in terms of fitness requirements have generally resided in the domain general principle of optimization; that is, the presumption that evolved behavioral characteristics reflect optimal trade-offs between costs and benefits to the animal's fitness.
Alcock (1984, pp. 199–203), however, has pointed out an essential problem with optimization theory; it is usually impossible to measure evolutionary costs and benefits in the same units. For example, the benefits of migrations include greater availability of food sources and facilitation of breeding, whereas the costs include expenditure of energy and danger from predators along the way. Thus, the only means by which we can conclude that the positive value of the benefits exceeds the negative value of the costs is by the fact that migratory behavior has selected in for the animal. This, of course, is a textbook example of circular reasoning.
In fact, Alcock expresses skepticism about the cost–benefit explanations of long-range migrations (pp. 241–244), and regards the phenomenon as a continuing theoretical challenge. In general, Alcock prefers a “qualitative” approach to evolutionary analyses, whereby correspondence is established for a particular species between a particular behavioral trait and particular aspect of its adaptation. Within Cosmides & Tooby's (1992) domain-specificity model, this would be phrased in terms of the correspondence of the design features of a specific cognitive mechanism and the task requirements of a specific adaptive problem.
Alcock does concede that quantitative analyses based on optimization may have utility for more circumscribed areas of behavior, where costs and benefits may be measured in the same currency—for example, calories lost and gained in foraging strategies. Janson (2000), however, has described some of the constraints of quantitatively based laboratory research on optimization of foraging behavior that limits its ability to generalize to behavior in vivo. He points out that laboratory experiments usually expose the animal to a single cluster of food sites that can be visited during a limited time interval; thus the only available adaptive responses for the animal are to eat at all sites and minimize travel distances between them. Janson's (1998) own studies of brown capuchin monkeys showed that the foraging behavior of these animals in their own habitat were quite different than in the laboratory. Capuchins in the wild use a variety of foraging strategies, dictated by the longer-term goal of consuming a fixed daily food requirement with the shortest required overall travel distance. Furthermore, they alter their strategies as distributions of food resources change, and they do not consume what cannot be digested during the course of the day. According to Janson, capuchin monkeys and many other species can find and ingest food faster than they can digest it; hence, there are minimal fitness returns in increasing food intake beyond that which can be digested within a limited time period.
Janson (1998) concluded that in order to devise a model to predict how an animal will forage, one must first determine what the animal knows about its environment. Most researchers implicitly assume that it is very little, but some food storing birds seem to remember hundreds of seed caches (Balda & Kamil, 1988; Hilton & Krebs, 1990) and many primate species, as well, show precise recall of specific resources (Janson, 1998; Menzel, 1991). Janson also concludes that evolutionary hypotheses should be conceptualized in the context of the animal's unique attributes and requirements, an approach compatible to both Alcock's and the domain specificity perspectives.
As frequently noted, in order to navigate, you must have a map and a compass. The map, physical or mental, indicates where you are in relation to your goal. The compass refers, figuratively, to the cues you will use to get there. Generally speaking, we know more about compasses than maps; that is, there are much data about the cues animals use to reach a goal, but less about how they decide where the goal is.
Navigational cues fall into two distinct strategies, which we will call orientation versus landmark but which have also been referred to as Euclidean or geometric versus topographic, dead reckoning versus episodic, and allocentric versus egocentric. The orientation strategy, most effective for journeys over long distances, requires the animal to maintain a sense of its own position in relation to various global markers. These include the sun, the stars, wind direction, the earth's magnetic field and barometric changes, and usually involve the individual's proprioceptive bodily cues and biological clock. When navigating indoors by an orientation strategy, the subject uses the configuration of the structure. In contrast, the landmark strategy involves the learning and recall of visual markers and their relationships to each other along the route, including objects, turning points, and details of the terrain.
Many species possess the capacity for both strategies, but the one that is best developed and most frequently used depends on the animal's ecological requirements (Alcock, 1984; Drickamer & Vessey, 1986). Thus, migratory and homing animals primarily use an orientation strategy, whereas animals that stay closer to home generally use a landmark strategy. Strategy is also a function of the greater availability of landmark or global cues, as illustrated in the differential navigational processes of tropical forest versus desert-dwelling ants. The former use a landmark strategy for foraging, whereby the ants learn and follow the markings on the forest canopy above them. The latter utilize their sustained access to the sun by deriving compass information from its position. We know that these species are using these strategies by studies showing that if a forest ant is experimentally displaced to a point that it has not previous traveled, it will not be able to readily find its way back, whereas a desert ant will proceed directly on the correct path (Holldobler, 1980).
Animals that primarily use an orienting strategy often have more than one method at their disposal and use these interchangeably, in an adaptive manner. Thus, honeybees and homing pigeons orient by the sun when it is visible, but during overcast days or nights will revert to methods that utilize the magnetic fields of the earth. Migrating birds generally use a sun-based orientation strategy for most of their journey, but they revert to a landmark strategy when approaching home. The orienting strategy enables them to navigate across long distances where landmarks are not available, such as over the sea, or where frequencies of landmarks along the way are too copious to recall. The landmark strategy enables them to hone in on their precise destination when approaching the completion of their journey
Humans are, historically and prehistorically, foragers, hunters, and colonizers. Thus, we engage in short, intermediate, and long forays, thus using both orientation and landmark strategies.
To demonstrate this to yourself, point to some other state or country, far from your own. Now point to a place close to home, where you go on a regular basis, such as a grocery or a friend's residence. You will probably use an orientation strategy for the first task, by constructing a mental image of a map that includes your own and the target's location and taking an estimate of your current compass bearing in relation to the target. On the other hand, you will most likely use a landmark strategy for the second task, based on the landmarks on the route you customarily take to your target (adapted from Thorndike and Hayes-Roth, 1980).
Across studies and measures, the heritability of spatial abilities, including those directly involved in navigation, ranges at about 0.50 (Bouchard, Segal, & Lykken, 1990; Defries et al., 1976; Plomin, Pederson, Lichtenstein, & McClearn, 1994; Tambs, Sundet, & Magnus, 1984; Vandenberg, 1969).
Regarding neural mechanisms, studies with humans (e.g., Maguire et al., 1998; Maguire, Frackowiak, & Frith, 1996; Maguire, Woollett, & Spiers, 2006), monkeys (Ono & Nishijo, 1999; Rolls, Robertson, & Georges-Francois, 1997), and rats (e.g., Eichenbaum, Stewart, & Morrisa, 1990; Thinus Blanc, Save, Pucet, & Buhot, 1991), have shown that navigational processes in general are associated with the functions of the hippocampal formation, which includes the hippocampus and the adjacent cortex in the most medial area of the temporal lobe. Other studies have shown different neurological processes within the hippocampus for tasks involving orientation and landmark strategies, which supports the notion that these evolved in a domain-specific manner. O'Keefe and Nadel (1978) first demonstrated that navigation by rats in an environment that is defined only by the shape of the enclosure, thereby requiring an orientation strategy, activates different types of neurons than those involved in landmark-based tasks. The former are called place cells, and include neurons that encode the animal's location and specific bodily movements in relation to the geometric properties of the environment (Muller, Bostock, Taube, & Kubie, 1994; Taube 1995, 1998; Taube, Muller, & Ranck, 1990).
Pizzamiglio, Guariglia, and Cosentino (1998) presented clinical data in humans that also demonstrated the dual neurological bases of orientation and landmark strategies. Two right-hemisphere-damaged subjects were unable to orient themselves to an enclosure when its shape was the only information available, but improved considerably when a visual object was added as a cue. Two other subjects, with a different lesion site in the right hemisphere than the first two, could orient themselves using only the shape of the enclosure. Unlike right-brained-intact controls, however, these subjects did not increase their performance with the addition of the cue.
Domain specificity can also be inferred from differences in the developmental stages when the two navigational strategies emerge. Children from about 2 years of age use landmark strategies, whereas rudimentary orientation strategies do not appear until about the age of 8 (Anooshian & Young, 1981; Blades & Medllicott, 1992; Scholnick, Fein, & Campbell, 1990). Landau & Gleitman (1985), however, performed a study with a congenitally blind girl of 31 months, in which she was led to various landmarks in a room and back again, each time, to a starting point, and then asked to navigate from landmark to landmark. Gallistel (1990, pp. 99–100) interpreted her success, despite her inability to see the landmarks in relation to each other, as demonstration of orientating behavior in a very young child. As previously defined, however, the landmark strategy does not necessarily require visual contact with markers. It does require prior knowledge of their positions in relation to each other, which can come from a map, or virtual travel, or, apparently, from a guided, tactual tour.
Though evolutionary theorists are primarily interested in universals, group differences often provide the first clues about these. Thus, the theory of evolution by natural selection began with Darwin's observations of subgroup differences within bird and amphibian populations in the Galapagos Islands.
Generally, the focus is on interspecies differences, but any ubiquitous group difference amenable to explanation in terms of natural selection may be relevant. Thus, the pervasive bias favoring males in spatially related tasks, both in humans and infrahuman species, ultimately led to evolutionary based theory and data on the nature of human navigational processes.
Studies of human spatial sex differences have shown a male advantage across a variety of measures, including field dependence, mental rotations, embedded figures, map reading, maze learning, and estimating the speed of a moving object. The magnitude of the sex difference varies among measures, with three-dimensional tasks showing greater differences than two-dimensional tasks and three-dimensional mental rotations tests yielding the largest and most reliable differences (Halpern, 1992; Linn and Peterson, 1985; McGee, 1979; Phillips and Silverman, 1997; Rahman and Koerting, 2008; Saucier et al., 2002; Saucier, Lisoway, Green, & Elias, 2007). Meta-analysis (Voyer, Voyer, & Bryden, 1995) has shown that the average difference between sexes for three-dimensional mental rotations, across dozens of studies, is a robust 0.94 by Cohen's d, indicating that the mean performance of males is nearly one standard deviation above that of females.
The male advantage in spatial tasks is highly consistent across human geographic populations. Though most studies have been conducted in North America, the sex diff-erence across various tests has been replicated in Japan (Mann, Sasanuma, Sakuma, & Masaki, 1990; Silverman, Phillips, & Silverman, 1996), England (Lynn, 1992); Scotland (Berry, 1966; Jahoda, 1980); Ghana (Jahoda, 1980); Sierra Leone (Berry, 1966); India, South Africa, and Australia (Porteus, 1965). A more recent set of studies, demonstrating the universality of sex differences across dozens of diverse cultures, will be described in a later section.
Although the sex difference has been reported in children as young as preschoolers (McGuinness & Morley, 1991), the consensus is that it does not appear reliably across tasks until early adolescence, which is generally attributed to accelerated hormonal differentiation (Burstein, Bank, & Jarvik, 1980; Johnson & Meade, 1987). One study (Willis & Schaie, 1988) has shown that from this age, the magnitude of the difference tends to be constant throughout the lifespan.
The sex difference extends also across species. Studies with wild and laboratory rodents have shown that males consistently outperform females in maze learning tasks (Barrett & Ray, 1970; Binnie-Dawson & Cheung, 1982; Gaulin & FitzGerald, 1986; Joseph, Hess, & Birecree, 1978; Williams & Meck, 1991).
Sex hormones, in terms of both organizational and activational effects, have been implicated in spatial sex differences in humans and animals (Choi & Silverman, 2002; Gouchie & Kimura, 1991; Hampson & Kimura, 1992; Janowsky, Oviatt, & Orwoll, 1994; Kimura, 1999; Kimura & Hampson, 1993; Nyborg, 1983, 1984; Phillips & Silverman, 1997; Puts et al., 2010; Reinisch, Ziemba-Davis, & Saunders, 1991; Silverman & Phillips, 1993; Silverman, Kastuk, Choi, & Phillips, 1999; Williams & Meck, 1991). Studies have shown a decrease in spatial abilities with increased estrogen levels, consistent with the direction of the sex difference. Corresponding increases in spatial performance with increased testosterone levels, however, occur reliably for females, but not males. Males have shown direct, inverse, and nil effects across studies. Nyborg (1983) attempted to explain this paradox in terms of the fact that plasma testosterone is, under some circumstances, converted to brain estrogen. Silverman et al. (1999) explained the differences in results in terms of the difficulty levels of the tasks used. Puts et al. (2010) concluded that hormonal influences in males are limited to organizational effects.
The first systematic, evolutionary based theory of spatial sex differences was by Gaulin and FitzGerald (1986). The core of the theory was that spatial abilities were more strongly selected for in males than females in polygynous species, for the reason that polygynous males require navigational skills to maintain large home ranges (the area within which an animal freely travels on a regular basis), in which to seek potential mates and resources to attract mates.
The investigators tested their theory with two species of voles; one, meadow voles, which are polygynous, and the other, pine voles, which feature an open promiscuous style. Findings were consistent with predictions; sex differences both in the direction of larger home ranges and superior maze learning ability for males occurred solely for meadow voles. Jacobs, Gaulin, Sherry, & Hoffman (1990) compared sex differences in size of hippocampus between these species and found, again as expected, proportionally larger male hippocampi in meadow voles but no sex difference in pine voles.
Does Gaulin and FitzGerald's mating strategy theory pertain to humans? Moderate polygyny is characteristic of our species (Symons, 1979) and a review of the cross-cultural literature on sex differences in home range size showed a near universal male bias beginning at the toddler stage (Gaulin & Hoffman, 1988). Additional support comes from Ecuyer-Dab & Robert's (2004a) finding that men tended to possess larger home ranges than women, as measured by retrospective and direct accounts of their comings and goings over extended time periods. Ecuyer-Dab and Robert posited also that if there was a “functional relation” between spatial abilities and home range size in males but not females, as inferred from Gaulin and FitzGerald's model, correlations between the two variables should be found only in the male. This was demonstrated in their study and also in earlier studies of African children (Munroe & Munroe, 1971; Nerlove, Munroe & Munroe, 1971).
Silverman and Eals (1992) questioned, however, whether the relationship between home range size and reproductive success applied to the human case. The only data that pertain to the question has shown that females exhibit greater natal dispersal (the distance travelled by an individual from natal site to first place of breeding) than their male counterparts (Koenig, 1989), which would contradict Gaulin and FitzGerald's theory.
Silverman and Eals (1992; Eals & Silverman, 1994) posed an alternative theory, in which the critical factor in selection for human spatial sex differences was division of labor during the Pleistocene. During that era, considered to be the most significant in human evolution, males primarily hunted, whereas females functioned as plant food gatherers, keepers of the habitat, and caretakers of the young (Tooby & Devore, 1987).
Silverman and Eals noted that the various spatial tests showing the strongest male bias (e.g., field independence, mental rotations, maze learning) corresponded to attributes that would enable navigation by orientation. This would be essential for successful hunting, which requires the pursuit of prey animals across unfamiliar territory and the capacity to return by a fairly direct route. They contended further that, if spatial attributes associated with hunting evolved in males, it is feasible that spatial specializations that would have facilitated their own roles in the division of labor would have evolved in females.
For food gathering, success would have required finding edible plants within diverse configurations of vegetation and locating them again in ensuing growing seasons; that is, the capacity to rapidly learn and remember the contents of object arrays and the relationships of objects to one another within these arrays. Success in gathering would also be increased by peripheral perception and incidental memory for objects and their locations, inasmuch as this would allow one to assimilate such information nonpurposively, while attending to other matters. Incidental object-location memory would also be useful in tending to the domicile and offspring.
There is supporting physical evidence for this analysis. Women have larger visual fields than do men; that is, they can see farther out on the periphery while fixating on a central point (Burg, 1968). They are also better than men at scanning, excelling in various tests of perceptual speed (Kimura, 1999, pp. 87–88).
Silverman and Eals (1992; Eals & Silverman, 1994) developed several methods to compare sexes on their ability to learn spatial configurations of object arrays, all of which generally supported the hypothesized female advantage. These findings have been partially or fully replicated in multiple laboratories and with diverse research designs (e.g., Choi & Silverman 1996; Dabbs, Chang, Strong, & Milun, 1998; Eals & Silverman, 1994; Gaulin, Silverman, Phillips, & Reiber, 1997; Hassan & Rahman, 2007; James & Kimura, 1997; McBurney, Gaulin, Devineni, & Adams, 1997; McGivern et al., 1997; Neave, Hamilton, Hutton, Tildesley, & Pickering, 2005; Spiers, Sakamoto, Elliot, & Bauman, 2008) with the most consistent differences occurring for incidental location recall.
Comparing the two theories of spatial sex differences, Gaulin and FitzGerald's (1986) has an advantage in that it applies to both humans and nonhuman species. Silverman and Eals' (1992), however, provide testable hypotheses about female spatial specializations. Silverman and Eals suggested that the theories may be reconciled if it is presumed that in a given species or subspecies, any difference in selection pressures between sexes related to spatial behavior may result in an evolved dimorphism. Ecuyer-Dab and Robert (2004b), however, presented a more precise means of integrating the theories, described in a later section.
Numerous studies have shown that males tend to use an orientation strategy in navigational tasks, whereas women use a landmark strategy (e.g., Bever, 1992; Choi & Silverman, 1996, 2003; Dabbs et al., 1998; Galea & Kimura, 1993; Holding & Holding, 1989; Joshi, MacLean & Carter, 1999; Lawton, 1994, 1996, 2001; Lawton & Kallai, 2002; McGuinness & Sparks, 1983; Miller & Santoni, 1986; Moffat, Hampson, & Hatzipantelis, 1998; Schmitz, 1997; Ward, Newcombe, & Overton, 1986). Specifically, males use distances and cardinal directions; that is, north, south, east, and west, whereas females rely more on landmarks and relative directions, such as right, left, in front of, and behind. These differences have been demonstrated with a variety of methods, including learning routes from maps or photographs, walking through mazes, retracing computer simulated routes on virtual mazes, drawing maps, giving directions, and finding one's way back after being led along an unfamiliar indoor or outdoor route.
Evolved mechanisms tend to remain quite broadly defined at this early stage of theoretical development, and more exact definitions will entail a long-term, continuing process of theoretical refinement and data gathering. Some recent research by the present authors and their colleagues, however, may provide a first approximation of an evolved mechanism for navigation by orientation.
Silverman et al. (2000) conducted a study in which subjects were led, individually, on a circuitous route through a heavily wooded area. During the walk, they were stopped periodically and required to set an arrow pointing to the place from which they began. Eventually, they were asked to lead the experimenters back to the starting point by the most direct route. Men's performances surpassed women's on all of these measures, and overall performance scores were significantly related across sexes to three-dimensional mental rotations scores, but not to nonrotational spatial abilities nor to general intelligence. Moreover, mental rotations scores emerged as the sole significant predictor in a multiple regression analysis that included sex as an antecedent variable, suggesting that the variance in orientation ability associated with sex appears wholly attributable to mental rotations abilities.
What do mental rotations tests measure that may function as an evolved mechanism for navigation by orientation? According to the investigators, both mental rotations and navigation by orientation require that the individual maintain the integrity of a space while exposed to it from various viewpoints. A number of studies (Cochran & Wheatley, 1989; Freedman & Rovagno, 1981; Schulz, 1991) have shown that the sole method for solving mental rotations problems with any degree of efficacy is by visualizing the rotation in three-dimensional space of one object while comparing it to another. In this manner, the subject mentally peruses the periphery of the object from various perspectives while maintaining a mental representation of its whole. Silverman et al. (2000) suggested that this is comparable, in terms of the processes involved, to searching for or following a prey while maintaining a mental representation of the boundaries of one's route.
Based on this explanation, Silverman et al. (2000) considered that the evolved mechanism at the core of the relationship between mental rotations and navigation by orientation appeared to be space constancy, whereby, “the properties of objects tend to remain constant in consciousness although our perception of the viewing conditions may change” (Coren & Ward, 1989, p. 406). (See also Bisiach, Pattini, Rusconi, Ricci, & Bernardini, 1997; Niemann & Hoffmann, 1997; Probst, Brandt, & Degner, 1986).
Further refinements of the role of space constancy in navigation by orientation may be informed by perceptual and neuropsychological studies. A complete explanation of the evolved mechanism, however, will also require consideration of environmental interactions, inasmuch as innately based behaviors are always expressed in an environmental context. In this vein, there is evidence that exposure to lines and angles during an early critical period is salient to the development of shape constancy (Allport & Pettigrew, 1957).
The question arises about whether the diverse navigational strategies of men and women represent one mechanism or two; that is, does the females' use of a landmark strategy represent an attempt to compensate for less-developed orientation abilities, or is it part of a separate evolved mechanism related to greater proficiency in recalling object locations. The latter view would apply both to the female's greater use of landmarks than distances and to their greater use of relative rather than cardinal directions. Relative directions are more efficacious for recalling and describing the locations of objects in relation to one another within a relatively small space, whereas cardinal directions are more suitable for processing and describing the vectors denoting longer distances.
Many investigators accept the compensation interpretation (e.g., Galea & Kimura, 1993; Lawton, 1994; Miller & Santoni, 1986; Moffat et al., 1998). Silverman and Eals' hunter-gatherer theory, however, would suggest dual mechanisms. So, also, do the prior cited neurophysiological studies demonstrating different neural processes underlying orientation and landmark strategies. As well, Gur et al. (2000) has shown differential brain site activation between men and women engaged in a spatial task.
Additionally, Choi and Silverman (1996) found that in a route-learning task in which the sexes performed equally well, success was predicted by preferences for landmarks and relative directions for females only, and preferences for distances and cardinal directions for males only. Similarly, Saucier et al. (2002) administered laboratory and field navigational tasks in which participants were required, at the direction of the experimenter, to use either an orientation or a landmark strategy. The sexes did not differ in performance, but males did better when using an orientation strategy, whereas females had higher scores with a landmark strategy. In both these studies, the observation that sexes performed equally well when using their own strategies of choice suggests that the use of landmarks is not a default strategy, but an expression of a well-developed mechanism in itself.
What would be the nature of a separate evolved mechanism mediating the unique spatial attributes of women? Silverman and Phillips (1998, p. 603) suggest that it entails “a more inclusive attentional style,” whereas Kimura (1999, p. 15) uses the term “efficient perceptual discriminations.” There is an alternative view, however, that involves imagery rather than attention or perception. Eals and Silverman (1994) found that the markedly greater abilities of females to recall object locations within arrays pertained also to unfamiliar objects, for which they did not have verbal referents. This may suggest that females have the ability to encode and recall entire scenes, in detail, by a process akin to eidetic imagery. Data purporting to show a greater “power of visualizing” for females were first reported more than a century ago by Galton (1883), and similar reports have appeared through the years (e.g., Anastasi, 1958; Sheehan, 1967).
Thus, whereby the navigational mechanism for males enables them to create mental maps of extended spaces, to which they had never been directly exposed, the corresponding mechanism for females gives them the ability to mentally construct and recreate detailed maps of smaller, previously observed spaces. This would appear to represent a highly adaptive dimorphism for the evolution of hunters and gatherers, favoring Silverman and Eals' theory, but there is a complicating factor. Rats, who are not hunters and gatherers, also feature the sex difference in navigational strategy. When navigating in radial-arm mazes, males are capable of using distal cues such as the shape of the room, whereas females require landmarks (Williams, Barnett, & Meck, 1990; Williams & Meck, 1991).
Ecuyer-Dab and Robert (2004b) presented a revised theory, emanating from both Gaulin and FitzGerald's and Silverman and Eals's, which may account for cross-species parallels. They proposed a twofold selection process underlying spatial sexual dimorphisms. For males, the critical selection factor is male-male reproductive competition, which tends to be more intense in polygynous societies and which would have favored the evolution of spatial abilities essential for both orientation-type navigational strategies and the effective use of projectiles. These skills would have enabled resource provision of mates and offspring by means of hunting and also aided in agonistic encounters between males competing for reproductive opportunities.
For females, on the other hand, the paramount selection factor for the evolution of a landmark strategy is the need for physical security for themselves and their offspring. The greater capacity to learn and recall details of the proximate environment, which is the basis of a landmark strategy, would have facilitated navigation within a relatively narrow home range and keener attention to cues regarding the presence of predators and other dangers. Such attributes would have also aided in finding and recalling possible hiding places or escape routes, which would have been particularly important when pregnant or tending small children. In this model, the greater capacity for food gathering of the female is a by-product, rather than the essential selection factor, in the evolution of her spatial specializations.
Ecuyer-Dab and Robert provide examples of sex-specific spatial strategies that appear to operate for these purposes in various nonhuman species, and contend that these observations probably generalize to most mammals. Thus, they have provided a theory that can encompass the findings from both Gaulin and FitzGerald's and Silverman and Eals's, and would appear to offer the most productive venue for further research.
A prime indicator of the evolutionary origins of a human trait is its generality across countries and cultures. In 2005, the British Broadcasting Company (BBC), gathering information for a documentary, administered a series of psychological tests via the Internet to more than a quarter of a million participants in 226 countries throughout the world. Included in the test battery was a brief version of a standard measure of three-dimensional mental rotations, which yields the strongest and most reliable male advantage, and a test developed for group administration of object location memory developed by Silverman and Eals (1992), which consistently favors females. For the latter, the individual observed a drawing of an array of objects for a specified time; then was presented with a drawing containing the same objects, but with half in different locations, and asked to circle the objects that have been moved.
Participants in the BBC study identified their genders and their ethnic groups from seven ethnic categories listed. Thus, the opportunity was available to assess the generality of the hunter-gatherer theory of spatial sex differences across countries and ethnic groups throughout the planet.
Silverman, Choi, and Peters (2007) performed this study using the data of 40 countries; those with a minimum of 100 each of males and females partaking in both tests. For mental rotations, significant differences (at p < .05) favoring males were found for all of the seven ethnic groups and 40 countries. For object location memory, significant differences favoring females were found for all of the seven ethnic groups and 35 of the 40 countries. The five remaining countries showed trends in the predicted direction, with one approaching significance at p = .07.
Evolutionary psychologists do not eschew the role of socialization in behavioral development, acknowledging that genetic effects are always manifested in environmental contexts. Thus, Lippa, Collear, and Peters (2010), using the BBC data from 53 countries for mental rotations and line angle judgment (another visuospatial test showing a male advantage), explored the question of whether spatial sex differences could be attributed, at least in part, to stereotypic sex roles. The latter was assessed from the United Nations gender-related development index, which rates countries on gender equity across three dimensions: health and longevity, standard of living, and knowledge and education.
Male scores were significantly higher for both tests in all countries. Regarding the effect of gender equity, however, results were in the opposite direction to the hypothesis; sex differences were significantly greater, rather than smaller, in countries with higher equity ratings.
The authors offered the possible explanation that women in highly developed, gender-egalitarian societies may be exposed more to evidence of spatial sex differences, leading to more pervasive stereotypes that could be debilitating in their effects on performance. An alternative explanation, however, is that inadequate educational systems in less-developed countries may not afford the environmental supports necessary in either gender for the sufficient development of adaptive, sex specific, spatial specializations.
Having considered the mechanisms of navigation, we turn now to a related issue; the choice of where to go. Questions about landscape preference have traditionally come from the study of aesthetics, but evolutionary based theories and data suggest that this was also a critical aspect of survival for both human and animal life.
Deer mice, a common North American rodent, can be divided into two types according to whether they inhabit grasslands or forests. Wecker (1963) built an outdoor enclosure, half consisting of a grassland and half a forest environment. He released two samples of grassland deer mice into the center of the enclosure, one a group of wild-caught mice and the other their laboratory-reared offspring, to see which environment they would prefer. Both groups showed strong affinities for the grasslands.
These findings and similar others (Klopfer, 1963; Thorpe, 1945) may suggest that landscape preferences are manifestations of a species' habitat selection, which would depend on ecological conditions such as availability of food and water, shelter, weather, and protection from predators.
This analysis can be extended to our own species. Human reactions to landscapes, positive or otherwise, tend to be immediate, unequivocal, and emotional, a response pattern that is presumed to have evolved from the needs of our forbearers to make rapid decisions regarding the benefits versus dangers of potential new habitats (Orians & Heerwagen, 1992). The capacity of favored landscapes to evoke positive emotions is well recognized in the conventional wisdom and has been well documented in research. For example, heart rates have been shown to decrease during the viewing of video clips of natural, but not urban scenes (Laumann, Garling, & Stormark, 2003). Postoperative patients in recovery rooms with pleasant, natural views have speedier and more positive recoveries than patients without such views (Ulrich, 1984), and numerous other studies demonstrate the psychological and physiological restorative influences of exposure to natural landscapes (e.g., Fuller, Irvine, Devine-Wright, Warren, & Gaston, 2007; Grahn & Sigsdotter, 2003; Kaplan, 1995; Parsons, Tassinary, Ulrich, Hebl, & Grossman-Alexander, 1998; Rappe & Kivelä, 2005; Ulrich, 1983;)
Although the task of identifying the habitats and consequent landscape preferences of deer mice and most other animal species seems relatively straightforward, humans have been unique in their ability to colonize a diverse range of environments. One approach to this problem, taken by Orians (1980) is to posit that human landscape preferences evolved in the habitat where the species presumably originated, the African savanna. The savanna biome features clumps of acacia trees scattered across wide grassy plains. This would have provided the human inhabitant with a readily identifiable and accessible place for the gatherer to acquire quality food; the fruits of the trees. Trees could also be used to keep watch for both prey and predators, and escape from the latter. They also serve as protection from the sun. Furthermore, the plains are suitable for grazing animals, which provide opportunity for the hunter (Orians & Heerwagon, 1992).
Orians and Heerwagon (1992) had subjects in the United States, Argentina, and Australia rate the attractiveness of acacia trees, which varied in terms of trunk height, branching pattern, and canopy density and shape. Acacia trees rated most attractive by all samples were those from areas of the savanna considered to be high quality in their general adaptive value for humans. These were characterized by moderately dense canopies and trunks which bifurcated near the ground, which would contribute both to ease of climbing and concealment.
Orians and Heerwagon also pointed to the recurrent nature of tropical savanna themes in landscape art. They quoted Humphrey Repton (1907, p. 105), a 19th-century pioneer of landscape architecture, who stated: “Those pleasing combinations of trees which we admire in forest scenery will often be found to consist of forked trees, or at least trees placed so near each other that the branches intermix.…”
Balling and Falk (1982) provided further supporting data for the savanna hypothesis, albeit equivocal. They used a series of 20 slides, encompassing five biomes; savanna, desert, and deciduous, tropical rain, and coniferous forest. Subject groups ranging in age from 8 to 70 rated each for desirability, both as a place to live and to visit. Overall, the slides of the savanna were rated significantly higher on both criteria, but beginning at age 15, savanna, deciduous, and coniferous forest landscapes were virtually tied for highest preference scores. Balling and Falk interpreted their data as supporting an innately based preference for the savanna, but one that may be altered by experience over the life span.
Appleton (1975) proposed an alternative theory concerning the adaptive basis of landscape preferences. Based on his analyses of landscape paintings, Appleton concluded that those with high attraction value contained a balance between prospect and refuge features. Prospect features, such as elevated landforms, provide an overall view of the landscape and facilitate the search for food, water, and prey. Refuge features, such as groupings of trees, permit the individual to see without being seen and function mainly in the interests of security. In this theory, the most important aspect of the habitat is the spatial arrangement of environmental attributes; that is, landforms, trees, open spaces, and water, inasmuch as their arrangement determines whether prospect and refuge opportunities can be effectively utilized.
Attempts to test Appleton's theory by means of comparative landscape judgments have yielded equivocal results (e.g., Clamp & Powell, 1982; Heyligers, 1981), which is possibly a function of variation in spatial arrangements of prospect and refuge features (Appleton, 1988). Supporting data for the theory were obtained by Mealey and Theis (1995), however, based on their contention that the relative attraction value of prospect and refuge should vary within individuals in accordance with their moods. Positive moods, they maintained, would induce a need to explore and take risks for the sake of future benefits, and would thus be associated with prospect. Negative moods, conversely, would give rise to a need for security and rest, and would thereby be associated with refuge. As predicted, subjects reporting positive moods preferred landscapes with vast expanses and overviews, whereas subjects reporting mood dysphoria preferred landscapes with enclosed, protected spaces.
Kaplan and Kaplan's (1982) notions stand in contrast to the emphasis on both security and simplicity in the theories described earlier. In addition to coherence, referring to organization of the scene, and legibility, meaning how easily one can navigate within the landscape depicted, the authors included mystery and complexity as key, evolutionary-based elements of attraction in human landscape preference. Their reasoning was based on the adaptive function of curiosity for the species, particularly the predisposition to seek new information about the environment that can facilitate its mastery, and the greater likelihood of finding new information in complex designs.
Some studies based on Kaplan and Kaplan's notions have revealed preferences for “mystery,” as represented by winding forest paths or obscure coves (Herzog, 1988; Kaplan, 1992). As for complexity, moderate levels appear to evoke the strongest preference levels (see Ulrich, 1983).
As in the case of spatial navigation, the ultimate issue for an evolutionary theory of landscape preferences is how best to conceptualize the evolved cognitive mechanisms that mediate these. All theories seem to imply a mechanism similar to the ethological construct of innate schemata; that is, a mental image of an ideal landscape that serves as standard for judgment. The theories differ, however, in the substance of this image. Orians maintains that it is a copy of the specific landscape in which humans evolved, whereas both Appleton's and Kaplan and Kaplan's imply that it can be any landscape type that contains features that signal fitness-related opportunities. Modern measures of brain site activation may provide a more precise methodology for the comparison of emotional responses to landscapes, and thereby help resolve this issue.
There is also a general methodological problem that needs to be addressed. Wilson, Robertson, Daly, and Wilson (1995) point to the confounds that may readily attend any attempt to compare the preference values of specific features between scenes, which has been the customary method of testing hypotheses derived from both Appleton's and Kaplan and Kaplan's theories. For example, scenes considered be high in mystery by the experimenter may simply have lower and more pleasing brightness levels.
Wilson et al. (1995) suggest a methodology whereby the same scene is manipulated so that just one feature is modified at a time, thereby allowing greater control of potential confounds. By this means, they were able to establish that small and subtle cues as to the water quality of seascapes have a marked effect on attraction value.
Finally, it is worth noting that landscape preference studies that have included urban landscapes (Kaplan, Kaplan, & Wendt, 1972; Laumann et al., 2003; Parsons et al., 1998; Purcell, Lamb, Peron, & Falchero, 1994; Ulrich, 1981, 1983) have universally found strong preferences favoring rural scenes of any type, by both urban and rural dwellers. Ulrich (1983) concluded that the distributions of preference ratings between rural and urban scenes barely overlap, even when ordinary rural scenes are compared to urban scenes that are particularly picturesque.
This, in itself, provides broad support for an evolutionary perspective on landscape aesthetics in that it suggests the profound influence of prehistoric origins, even when pitted directly against life experiences. It points, also, to the potential adverse effects of living in the “unnatural” environment of high-density urban centers, particularly in light of the previously described data on the effects of landscape exposure on psychological and physical well-being.
The application of the evolutionary model to human navigation and landscape preference represents a relatively recent movement in the behavioral sciences, though the burgeoning theory and data reviewed in this chapter is testimony to its relevance. This review has also highlighted two major aspects of the movement. One is the salience of an ethological approach, which has provided compelling insights about analogous processes mediating human and animal adaptations in these areas. The second is the conceptual utility of evolutionary psychology's model of mind as comprised of evolved, domain-specific mechanisms.