The ability to learn spatial properties of novel environments is an important aspect of our everyday lives—whether learning a new city you moved to or navigating through an unfamiliar airport. Spatial knowledge acquisition includes learning both metric and non-metric spatial properties of environments; metric properties include quantitative distances and directions and non-metric properties include sequence and connectivity. These spatial properties include the identities and locations of landmarks, the turns in a route, and the distances and directions between places (Goldin & Thorndyke, 1982; Thorndyke & Hayes-Roth, 1982). Environments can be learned directly by people sensing and moving through the environment or indirectly via symbolic sources of information, such as maps or language (Montello & Freundschuh, 1995). And of course, individuals differ in how well and how easily they learn spatial knowledge about environments. But an interesting and important question remains: Do people with good environmental spatial skills express that skill only when they intentionally pay attention to spatial properties or does their greater skill express itself without intentional effort? In this research, we examine how the individual-difference trait of “sense-of-direction” (SOD) expresses itself when people directly learn spatial properties of a new neighborhood. We specifically examine this relationship as a function of whether people receive intentional or incidental instructions to learn the spatial properties.
Directly acquiring spatial knowledge in the environment
Two accounts have been proposed to explain the process of acquiring spatial knowledge in new environments (microgenesis) from direct experience, without symbolic sources: the “dominant framework” proposed by Siegel and White (1975) and an “alternative framework” proposed by Montello (1998). Siegel and White (1975) contended that the learning of a new environment progressed through three hierarchical “stages.” First, individuals learn landmarks, which are salient point-like structures that lack distance and directional information (i.e., non-spatial identity). At this stage, landmark knowledge consists of familiarity with the appearance and perhaps the names of the landmarks, but no knowledge of how those landmarks are spatially related (assuming they are out of sight from each other). After acquiring a sufficient level of landmark knowledge, individuals learn the route(s) that connect the landmarks into a sequence of movements (non-metric only). For example, route knowledge might include knowing that landmark A comes before landmark B along a specific route and that at landmark A one needs to turn right to get to landmark B. Finally, after acquiring a sufficient number of routes and beginning to metrically scale them, individuals relate the landmarks and routes to one another as part of a metric spatial configuration, referred to as survey knowledge. The strict hierarchical nature of these stages meant that individuals could not have survey knowledge without first passing through landmark knowledge and then route knowledge. With only landmark and route knowledge (at least in its initial form), a person would not have accurate metric knowledge of environmental layout, such as knowledge of direct distances and directions between landmarks.
In contrast to what Montello (1998) termed the dominant framework, he proposed an alternative framework in which individuals continuously acquire all three forms of knowledge from the beginning of a single episode traveling in an environment, without the need to pass from one stage to another. For example, most people have probably experienced walking in an unfamiliar city for 30 min and knowing more than just the identities of landmarks they have passed—most would have some ability to retrace a route back to their starting location. Many people could probably estimate distances and directions towards at least some of the landmarks they passed along the route—albeit not very precisely or perfectly accurately after only one travel experience. Thus, it is likely that people acquire some level of route and survey knowledge after just minutes of exposure to a new environment and a single travel episode. The alternative framework contends that landmark, route, and survey knowledge are acquired more or less simultaneously, as soon as an individual starts experiencing a new environment. Of course, the accuracy and completeness of this knowledge increases with experience, potentially indefinitely.
In the current study, participants were taken on a walk through an unfamiliar housing development. If the dominant framework accurately describes spatial microgenesis, we should not expect any participants to acquire much metric knowledge about distances and directions; it is questionable if they would even acquire much information about sequences of places along routes. However, if the alternative framework more accurately describes spatial microgenesis, we expect that participants will acquire not only route knowledge but some metric survey knowledge. Only in the latter case should we expect a substantial relationship between the accuracy with which participants estimate metric spatial properties and their SOD.
Individual differences in acquiring spatial knowledge in the environment
Regardless of which framework describes spatial knowledge acquisition better, one would not expect all individuals to acquire knowledge at the same speed, with the same accuracy, and so on, even if they had similar levels of experience in the environment. While this is an active research topic (Wolbers & Hegarty, 2010), we presume that individuals differ in acquiring knowledge due to some combination of innate or learned spatial abilities and/or acquired strategies for learning and estimating spatial properties. In fact, Ishikawa and Montello (2006) found that individuals, who were driven along a novel route once a week for ten weeks, showed radically different patterns of spatial knowledge acquisition. These differences were particularly salient in their survey knowledge; even though some participants did not acquire accurate survey knowledge after ten trips, these participants did learn the identities of landmarks, the order of landmarks along two test routes, and distances between landmarks along the routes. In essence, even though that study was designed to compare the dominant and alternative frameworks, in the end it showed that individual differences were so substantial that no single framework was likely to describe the learning process well for everyone. Some people’s learning seemed best characterized by the dominant framework, others by the alternative. Some did not show much learning at all over the ten weeks. This finding suggests that we should see considerable variance among our participants in their acquisition of metric survey knowledge, which in turn implies ample variance to support sizeable correlations with self-reported sense-of-direction.
Cognitive effort and intention to learn
There is a long history of research on the role of intention and effort in learning different kinds of information (Craik & Lockhart, 1972; Hasher & Zacks, 1979; Postman, 1964), some of which has focused on learning spatial information (Mandler, Seegmiller, & Day, 1977; Naveh-Benjamin, 1987). Research on the impact of active versus passive exploration of environments while acquiring spatial knowledge has been mixed (for a review, see Chrastil & Warren, 2012). Lindberg and Gärling (1983) found no differences in survey knowledge after incidental or intentional learning across three exposures to the environment. However, all participants showed performance increases across the three exposures, suggesting that all participants were attending to the spatial properties of the environment. More recently, Van Asselen, Fritschy, and Postma (2006) investigated learning differences within a building and found that landmark identification and ordering did not differ between the incidental and intentional learners. Participants who learned intentionally were more accurate in retracing the route they learned and drawing the route on a map. In a similar study focused on landmark knowledge, landmark identification again did not differ between learning conditions but landmark placement on a map showed a benefit of intentional learning (Wenczel, Hepperle, & von Stülpnagel, 2016), but another study failed to replicate these findings (Von Stülpnagel & Steffens, 2013). Overall, these findings suggest that landmark knowledge and route knowledge might be relatively effortless to acquire, whereas survey knowledge might be more effortful. That is, the level of effortful processing required to learn the spatial properties of an environment might depend on the type of spatial knowledge being acquired. To investigate this possibility, the current research will assess participants’ landmark, route, and survey knowledge after incidental or intentional learning of a novel environment.
A classic method to determine whether a cognitive process requires automatic versus effortful processing is to manipulate the intention to learn (Hasher & Zacks, 1979). This is typically done by instructing some participants to try to learn a certain type of information and not instructing others. If performance is more accurate after intentional than incidental instructions, one can conclude that processing the information requires conscious attention and explicit processing. In contrast, if there are no performance differences whether learning was intentional or incidental, then one can conclude that the cognitive process must be automatic. In the current research, we apply this logic to investigate the interplay between spatial learning intentionality and individual differences in acquiring environmental spatial knowledge from direct experience. Importantly, we will contrast intentional with incidental learning of spatial aspects of the environment (e.g., landmark, route, and survey knowledge) but will not contrast intentional with incidental attention to the environment per se. This was done to mimic the restaurant scenario (see the Significance statement) in which an individual is looking around and attending to the environment but not attending to the spatial properties of the environment (e.g., incidental spatial learning). We will accomplish this by instructing all participants to attend to the environment (using a cover story about attitudes toward architectural and natural features) but instructing only half of the participants that they must learn the spatial layout of the environment and will be tested on it (i.e., intentional spatial learning). By manipulating intentionality in this way, we ensure that all participants are attending to the environment but the groups differ in their intentionality to learn spatial properties.
Ishikawa and Montello (2006) reported that the accuracy and speed with which survey knowledge was acquired by participants were strongly related to their self-reported SOD. SOD is the hypothesized ability to find your way within environmental-scale spaces. It has primarily been assessed by self-report measures, such as by answering the simple question “How good is your sense-of-direction?” (Kozlowski & Bryant, 1977) or by averaging responses to several questions, such as questions about getting lost, learning distances and directions, using maps, and following cardinal directions. Using the multi-item self-report survey known as the Santa Barbara Sense-of-Direction (SBSOD) scale (Hegarty, Richardson, Montello, Lovelace, & Subbiah, 2002), Ishikawa and Montello found that SBSOD scores related mostly to how well participants learned survey relations, such as straight-line directions between landmarks on their test routes. Those who reported having a good SOD learned survey knowledge substantially more accurately and quickly; those who reported having a poor SOD learned them less accurately and quickly, in some cases, virtually failed to learn them at all. In contrast, participants differed very little in their ability to acquire landmark and route knowledge as a function of their SBSOD score; all individuals—regardless of their reported SOD—were able to accurately order named landmarks after one exposure to the route. In fact, most participants were able to accurately estimate distances between landmarks along the route after only one trip, even if they reported a poor SOD. In the current study, we assessed several types of spatial knowledge and related participants’ performance to their self-reported SOD.
Previous research has rarely examined different types of spatial knowledge when examining individual differences in environmental spatial knowledge (e.g., Fields & Shelton, 2006; Hegarty, Montello, Richardson, Ishikawa, & Lovelace, 2006; Montello & Pick, 1993; Schinazi, Nardi, Newcombe, Shipley, & Epstein, 2013), but when multiple measures of spatial knowledge have been related to self-reported SOD, some measures of spatial knowledge relate to SOD and others do not. In the Ishikawa and Montello study, participants easily acquired accurate knowledge of landmark identities and routes, including metric distances along the routes, and these measures were not related to SOD. For their measure of landmark knowledge, Ishikawa and Montello used four landmarks per route (a total of eight for two routes) and they taught participants verbal labels for the landmarks. Naming landmarks could have introduced verbal processing into the processing of spatial information, which might have drawn upon cognitive skills that individuals with poor SOD are not particularly poor at. Support for this idea comes from dual-task paradigms in which verbal tasks interfere with aspects of landmark, route, and survey knowledge (Labate, Pazzaglia, & Hegarty, 2014; Saucier, Bowman, & Elias, 2003; Wen, Ishikawa, & Sato, 2011). In order to address this issue, the current experiment used eight landmarks along a route and the experimenter did not associate the landmark with verbal labels. Instead, the experimenter referred to the landmark scenes by using photographs of each landmark when testing participants’ spatial knowledge. This ensured that while participants may have associated the landmarks with verbal labels, those verbal labels were unique to each participant and not influenced by any verbal label given by the experimenter.
Interaction between cognitive effort and SOD
The main purpose of the current study is to investigate whether SOD relates to the acquisition of environmental spatial knowledge differently as a function of learning intentionality. This is important because it addresses the question of whether the skills associated with having a good SOD are better characterized as mental abilities (such as memory capacity or mental processing speed) or as strategies (such as paying attention to turns you take or watching the sun as you walk). Mental abilities would typically express themselves implicitly whether a person attempts to apply them or not—they do not require conscious effort to influence knowledge processing. Strategies, on the other hand, can be consciously retrieved by a spatial thinker and accurately described to another person (such as to a researcher during a protocol analysis). Even as a strategy becomes easier to apply with repeated use, people choose to use it when they are trying to solve a particular problem for which they think it is relevant. Note that the distinction here between implicit and explicit does not map perfectly onto the learned-innate distinction. Strategies are presumably learned, but mental abilities may be innate, learned, or (most likely) result from an interaction of innate and learning influences.
The question of how SOD skills relate to learning effort and automaticity is not only theoretically important but is also relevant to the prospect of training people to have a better SOD. If SOD skills are due to explicitly applied strategies for spatial problem-solving, then it will likely be easier and more straightforward to train individuals for better skill (e.g., Hegarty, Keehner, Cohen, Montello, & Lippa, 2007; Thorndyke & Stasz, 1980). It may still be possible to improve mental abilities expressed without conscious application, however, given appropriate training experiences (cf. Uttal et al., 2013). This may be true even for innate abilities; innate does not mean unchangeable, although it would typically mean less easily changeable. We expect that training mental abilities would be considerably less straightforward than simply telling people to use a specific strategy while solving a problem.
In sum, if SOD reflects learned strategies under conscious control, we should find at least a modest main effect of spatial learning intentionality on spatial knowledge acquisition, because people with good SOD would learn better under intentional instructions than incidental (it is unclear if people with poor SOD could learn the spatial layout better when intentionally attending to spatial properties of the environment or not). In particular, however, we should find an interaction between learning intentionality and SOD, because individual differences in spatial knowledge acquisition would be diminished when the spatial layout of an environment was experienced incidentally, without the intention to learn it. When people with varying SODs are not told to pay attention to spatial properties (incidental learning), they would not be as likely to apply their strategies and thus would not differ much in the spatial knowledge they acquire. When the spatial layout of an environment is experienced intentionally, however, individuals with good SOD would learn spatial aspects of environments better because they would better apply particular strategies to learning when told to do so. Therefore, we would expect that we might find a main effect of learning intentionality but that we would definitely find an interaction between SOD and learning intentionality.
In contrast, if SOD is a reflection of an individual’s implicit mental ability to acquire spatial knowledge, then it should not matter whether instructions to learn spatial properties are intentional or incidental (i.e., learning intentionality would have no effect). Those with good SOD would learn more and more accurately whatever the instructions they receive. It could even be that what underlies a good SOD is the unprovoked tendency to attend to spatial properties without being told to do so. That is, maybe people with a good SOD are always attending to the spatial properties of an environment, even when only given incidental instructions. Those with a poor SOD never or rarely attend to spatial properties, so even when told to attend to spatial properties, they do not know how to or they cannot make use of the spatial properties they notice. Whatever the nature of the implicit abilities underlying SOD, we would expect people with good SOD to learn better than those with poor SOD whether they receive incidental or intentional instructions. That is, we expect a main effect of SOD, no main effect of learning intentionality, and no interaction between SOD and learning intentionality.
In the current research, we investigate how environmental spatial knowledge that is acquired directly is related to: (1) SOD; (2) learning intentionality; and (3) the interaction between SOD and learning intentionality.