Spatial Cognition and the Development of an Awareness Space Model

Activity spaces are generally described as a geographic extent or surface that encapsulates

an individual’s interactions with specific locations. However, such a working definition fails to provide context for the cognitive processes at play which determine an individual’s

understanding of their environment and subsequently what they can or cannot do in said

environment. Heft, for example, argues that most psychological research in spatial cognition falls

into one of three perspectives: “1) an information-processing approach; (2) a tradition stemming from Piaget and Inhelder’s work on children’s development of spatial cognition; and (3) a nativist approach to spatial cognition with its roots in Cartesian and Kantian thought” (Heft, 2013, p. 18). These three perspectives, Heft argues, summarise the different ways in which

environment, of which activity spaces are thought to be a subset. Generally speaking the first two

perspectives argue that spatial cognition is developed iteratively through experience and

observation which in turn generates a mental construct that represents the surrounding

environment, while the third posits some form of inherent ability. Where the Tolman and

Piagetian perspectives differ is how those constructs are organized: the Tolman viewpoint being

one of a set of possible actions and their alternatives relative to a place (Tolman, 1948), and the

Piagetian perspective being one of cultivating a logical mental construct of the environmental

structure in strict relational terms (Piaget, Inhelder, & Piaget, 1997). The nativist perspective,

finally, proposes that spatial cognition occurs a priori; it is not a process that is learned but is

instead inherent. However, spatial cognition research very rarely adheres to any one perspective

strictly, instead typically being a combination of all three (Heft, 2013).

These perspectives all differ to various degrees regarding how spatial information is

processed and subsequently stored; the behaviourist perspective describes it as a set of actions

and their alternatives whereas the Piagetian perspective emphasises the malleable nature of

spatial cognition and its development; and finally the nativist perspective simply offers that

somehow spatial cognition is an inherent quality. Where these three perspectives seem to agree is

that the highest form of spatial cognition lies in the configurational understanding of the

environment – i.e. the individual understands how components of the environment are related.

This configurational understanding is typically described as “survey knowledge” or a cognitive map (Heft, 2013). Thus these perspectives disagree on the exact processes involved, but agree

that the “end product” of spatial cognition is some mental construct that stores the individual’s understanding of their environment and its relationships in memory for later recall.

The exact process by which the environment gets “parsed” or understood by an individual and subsequently stored is not yet agreed upon. Empirical research from animal studies as well

as child orientation tasks point towards competing processes. For example, in child orientation

tasks it has been found that geometric cues were more influential than “non-geometric” or “local” cues, such as coloured walls, for child participants (Hermer & Spelke, 1996). Children were placed into an enclosure and shown a toy, which was then hidden behind a sheet. The

participants were then gently spun around to disorientate them, and their search behaviour for the

toy was observed. Typically children either searched the correct corner of the enclosure, or the

geometric similar opposite corner - even when local cues were present. This implies that the

children were utilizing geometric cues – corner locations – to orientate themselves and not local

cues such as coloured walls.

Subsequent studies have illustrated that the size of the enclosure, among other things,

influences whether or not subjects utilize geometric or local cues to orientate themselves.

Learmonth, Nadel and Newcombe (2002) illustrated that once the experimental enclosure

reaches a certain size, children began to use local cues and would no longer confuse diagonally

adjacent corners. Thus orientation is not an either or scenario with regard to the importance of

geometric cues vs local landmarks, but rather that circumstance dictates which orientation

method is used. Finally, Heft (2013) draws attention to one other important consideration in the

development and manifestation of spatial cognition: personal history.

A number of animal studies investigated whether or not the rearing conditions of the

animal influenced whether geometric or local cues were more important in orientation. The

empirical results of which imply that the method of orientation used was in fact contingent on

Gould, 2009). This body of work serves to illustrate the plasticity associated with spatial

cognition. The relative importance of geometric cues versus local cues appear to vary according

to circumstance and an individual’s past experience.

Recent work within neurophysiology describes specific mechanistic theories for spatial

cognition and provides some context for interpreting the adaptive nature of spatial cognition and

its method of storage outlined above. Hartley and Burgess (2002) describe a cognitive model

based around neurobiology. They argue that different components of the brain handle different

spatial representations based on their unique demands. For example, they describe how:

“Hippocampal processes are concerned with large distances and long timescales, whereas parietal processes are concerned with short timescales and the space immediately surrounding

the body” (Hartley & Burgess, 2002, p. 2).

These observations lead to the development of a cognitive model based on frames of

reference. An egocentric frame of reference is centred on the self; whereas an allocentric frame

of reference is centred on the “world”. These two reference frames correspond to two different mental representations of the environment corresponding to short term and long term

respectively. The reason for this division is best illustrated by example: an egocentric

representation of the environment would be one that held the self as central to all relationships;

as the individual moves through space such representations would require constant updating and

obviously would be a very “expensive” means of representing the world. The allocentric representation sidesteps this issue by providing a means to store long-term spatial relationships

between entities irrespective of their location to the self. If we return to the orientation task

example, whether or not an individual employs local cues vs geometric cues may be thought of

depending on the contextual demands of the task. This short – long term mechanist model of

spatial cognition shares many similarities to the modal model of memory of Atkinson and

Shiffrin (Atkinson & Shiffrin, 1968).

Thus far, however, the examples and discussion have only dealt with spatial knowledge

resulting from direct experiences and interactions. Participants in the examples provided have

only needed to process one source of data: their immediate surroundings and by mediation of

their past experiences, utilize that information to complete a simple task. Obviously in more

complex behaviour evaluating more data streams becomes a necessity and the resulting

conceptualization becomes more complex.

This complexity was explored by Brown and Moore (1970), who were interested in

studying inter-urban migration. They identified three domains of information important in any

search activity: 1) the total information available to the searcher; (2) the initial information

possessed initially by the searcher; and (3) the way in which the searcher combines these two

sources of information. Thus for a given house-seeker that individual’s ability to find a new

house is restricted by (1) the available information regarding new houses, because they cannot

consider a house they do not know about (i.e. they are not aware of it as a possibility); (2)

knowledge about houses they have generated through experience (i.e. their specific taste in house

or perceived needs); and (3) the strategy they put into action given the information they have

aggregated together; as not all individuals search for houses in the same way. Brown and Moore

describe the syntheses of these three domains as an awareness space.

An awareness space is “the locations within the total urban space about which the intended

migrant household has knowledge (or knowledge above some threshold level) before search

space can be thought of as the collection of knowledge an individual holds about their

environment and its apparent opportunities. The concept of a mental representation of the

environment is not a unique one within the literature. Dijst (1999) describes this construct as the

“perceptual action space”; Horton and Reynolds named it the “action space” (1971), and Lynch (1984) called it a “mental map”. For consistency’s sake, “Awareness space” will hereafter be

used as the term to describe the construct that holds an individual’s understanding of their environment.

This formulation is consistent with Dijst who said: “Theoretically, the actual action space is situated within the potential action space. The perceptual action space covers the actual action

space. The potential action space can be covered entirely by the perceptual action space (Dijst,

1999, p. 196)”. While Dijist’s terminology can be confusing, it succinctly relates three important

concepts together which are consistently mentioned within the literature: (1) that there exists

some construct that encapsulates the full understanding an individual holds about their

environment [awareness space], (2) that there exists some subset of locations from said construct

where the individual could be at any given time [potential activity space], and (3) those locations

where the individual actually is or has been [activity space]. The potential activity space