“Kinesthetic spatial cognition” can be defined as referring to the perception, memory, and recall of spatial information via the kinesthetic perceptual-motor system.

A great deal of research has demonstrated that spatial cognitive processes and verbal cognitive processes use separate cognitive resources. This has formed the basis of multi-channel models of information processing according to which cognitive attention can be allocated simultaneously to separate verbal and spatial tasks (eg. Allport et al., 1972; McLeod, 1977). This has been developed into the model of “working memory” which includes the “visuo-spatial scratch pad” for spatial rehearsal and processing and the “articulatory loop” for verbal rehearsal and processing (eg. Baddeley, 1986). Demonstrations of separate verbal and spatial cognitive processes are briefly reviewed here (for details, see Appendix III).

Much of the research probing multi-channels of information processing comes from “dual-task interference” studies in which a subject undertakes two separate tasks simultaneously, or a second task is undertaken while information from the first task is held in memory for later recall. The typical result is that two concurrent verbal tasks will interfere with each other, and two concurrent spatial tasks will interfere with each other, however a spatial task and a verbal task can often be performed as well simultaneously as either task can be performed individually. This result is interpreted as indicating separate spatial versus verbal cognitive resources (Allport et al., 1972; Baddeley et al., 1975; Baddeley and Liberman, 1980; Brooks, 1967; 1968; 1970; Farmer et al., 1986; Logie, 1986; Phillips and Christie, 1977b; McLeod, 1977; Morris, 1987; Pritchard and Hendrickson, 1985; Salthouse, 1974; 1975).

Other evidence for multi-channel models comes from studies of patients with neurological disease or injury. Often the ability to solve one type of cognitive task has been damaged while ability for other types of tasks remains normal. This is also interpreted as indicating the use of separate cognitive resources devoted to the different types of tasks. Much of this work has also contributed to the general notion of specialisation of the right cerebral hemisphere for spatial tasks and the left cerebral hemisphere for verbal tasks (De Renzi and Nichelli, 1975; De Renzi et al., 1977; Fried et al., 1982; Hanley et al., 1991; Kosslyn, 1987; Paivio and te Linde, 1982; Paivio and Ernest, 1971).

The right-brain spatial, left-brain verbal specialisation is not a fixed relationship but appears to be based on more fundamental differences in processing styles of the two cerebral hemispheres such as sequential processes of the left hemisphere versus holistic processes of the right hemisphere (Bradshaw and Nettleton, 1981; Luria, 1970; Trevarthen, 1978). There are also considerable differences between subjects (Ojemann, 1979; Gur and Reivich, 1980). In many cases the hemispheric superiority is minimal (ie. both hemispheres perform verbal or spatial tasks equally well), or is even reversed. For example, Levy and Reid (1976) found a correlation in which right-handed normal writers and left-handed inverted writers exhibited the normal right-hemisphere spatial, left-hemisphere verbal superiority. In contrast, right-handed inverted writers and left-handed normal writers had reversed hemispheric superiority.

Separate modes of spatial versus verbal cognitive processing are also posited by the “dual-coding hypothesis” which proposes that the spatial imagery system and the verbal system comprise two distinct symbolic systems which are involved in cognition (Paivio, 1978; 1979, p. 233). When both a verbal and a spatial code are learned for the same item then it is said to be dual-coded and can usually be remembered better. Dual-coding can also facilitate the recall of body movements when a verbal label is learned together with a motor action (Ho and Shea, 1978; Shea, 1977; Winter and Thomas, 1981)

In addition, a kinesthetic-motor code (“enactment”) can be identified which also facilitates verbal memory when it is dual-coded together with a verbal phrase (Cohen, 1981; 1983; Cohen et al., 1987; Engelkamp, 1986; 1988a; 1988b; Engelkamp and Zimmer, 1984; 1990; Nilsson and Cohen, 1988; Saltz, 1988; Saltz and Donnenwerth-Nolan, 1981). It appears that the motor enactment learning strategy leads to such good verbal recall performance that other factors have no further beneficial effects. Thus, enactment learning is believed to provide an “optimal encoding” (Nilsson and Cohen, 1988, p. 427) or an “inherent richness” (Cohen et al., 1987, p. 110).

Similar types of dual-coding have a long history of use in various strategies for improving memory from the ancient Greeks through to modern times. Many of these mnemonic strategies utilise a large group of imagined spatial locations (eg. rooms in a building). An image of each item-to-be-remembered is visualised as being at each of the locations. During recall subjects imagine a walk through the building and recall the image present at each of the locations. A variety of these memory strategies have been reviewed by Bower (1970a), Yates (1966), and Paivio (1979, pp. 153–175).

Ho and Shea (1978) point out that the superior memory performance from dual-coding can be explained according to the “levels of processing” model of memory (Craik and Lockhart, 1972). The dual-code increases the “depth” of processing by encouraging both both elaboration and distinctiveness of the item-to-be-remembered which are both necessary to improve memory performance (Craik, 1983; Eysenck, 1979; Hunt and Einstein, 1981; Hunt and Seta, 1984).

However, in some cases when verbal labels are attached to stimuli the memory for those stimuli does not necessarily improve. Rather, information in the verbal labels appears to be relied upon and the actual details of the stimuli may be forgotten (Carmichael et al., 1932; Daniel, 1972; Goldstein and Chance, 1970; Hall, 1977; Hirtle and Jonides, 1985; Hirtle and Mascolo, 1986; Klatzky et al., 1982; Nagae, 1980; Pezdek and Evans, 1979; Price, 1968; Ranken, 1963; Schooler and Engster–Schooler, 1990).

Spatial information can be perceived through visual, audio, and kinesthetic perceptual-motor systems which may exhibit separate perceptual, retention, and retrieval characteristics (Connolly and Jones, 1970; Jones and Connolly, 1970; Diewert and Stelmach, 1977; Newell et al., 1979; Posner, 1967; Reeve et al., 1986; For details see Appendix IV). It is also well documented that visual spatial information tends to dominate auditory or kinesthetic spatial information (Adams et al., 1977; Klein and Posner, 1974; Laszlo and Baker, 1972; Pick et al., 1969; Posner et al., 1976; Reeve et al., 1986; Willott, 1973).

In other cases information from all the perceptual systems appears to be integrated into a single unified spatial representation. Accordingly, what was initially conceived of as “visual memory” (eg. Baddeley et al., 1975; Phillips and Christie, 1977a) was later refined into a unified concept of “spatial memory” which is devoted to spatial information from any perceptual system (Baddeley and Lieberman, 1980). In many cases spatial perception and performance is identical regardless of which perceptual system is used (Bairstow and Laszlo, 1978a; Solso and Raynis, 1979), and spatial “images” are not necessarily visual (Kerr, 1983; Millar, 1990).

Two fundamental types of spatial information can be distinguished. Location spatial information refers to loci or targets which are independent of the body movements which might be used to perceive or recall these. Whereas configuration spatial information refers to figures which are specific to particular body parts (Smyth et al., 1988; Smyth and Pendleton, 1989; 1990). Location versus configuration types of spatial information are analogous to “space” versus “shape” as identified in “Laban Movement Analysis” (Dell, 1970; Hackney, 1989; Maletic, 1987) or to “spatial progression” versus “body design” as developed in “choreological studies” (Preston–Dunlop, 1980, pp. 87-93; 1981, pp. 54-60; 1984, p. x).

IIC.31 Spatial Cognition.

“Cognition” is defined as “a generic term for any process whereby an organism becomes aware or obtains knowledge . . . It includes perceiving, recognizing, conceiving, judging, reasoning” (English and English, 1974), or as “the activity of knowing: the acquisition, organization and use of knowledge” (Neisser, 1976, p. 1). In common usage cognition might be considered to be different than perception but research has revealed that perception involves cognitive processes of interpreting sensory data relative to past experiences and current knowledge. That is, perception is interpretation (for a review see Eysenck and Keane, 1990, pp. 84-95).

The concept of “visual cognition” is used in psychology and includes the perception, memory, and retrieval of visual shapes and their locations (eg. Pinker, 1984). In a similar way the notion of “vestibular memory” is sometimes used to refer to task performance based on spatial information arising from vestibular sensations (Israel and Berthoz, 1992, p. 197). The concept of “spatial cognition” is also widely used (eg. Brésard, 1988; Thinus-Blanc, 1988; Sadalla et al., 1980) to refer to the cognitive processing of any type of spatial task.

The notion of “kinesthetic spatial cognition” developed here is analogous to “visuo-spatial cognition” (Phillips, 1983) and refers to perception, memory, and retrieval of spatial information which has risen from kinesthetic stimulations. This includes “kinesthetic memory” (eg. Keele, 1968) and “motor memory” (eg. Housner and Hoffman, 1979) as used in motor control and cognitive research.

IIC.32 Body Movement as Cognitive.

The conception of kinesthetic-motor cognition is supported by many researchers who consider motor actions to inherently involve high-order cognitive processes rather than being controlled solely by lower-order sensory-motor processes. Paillard (1987, pp. 60-63) summarises the difference between “two classes of information processing by the nervous system”. Whereas the sensorimotor mode is primarily automatic, unconscious, and driven directly by sensory data, the cognitive mode is evidenced by conscious attention (though this may also be automatic) and is driven by internal computations and memorial cues:

The sensorimotor processing mode directly relates, via external loops, sensory information (gathered by sense organs from the physical world) to motor activities directly driven by this information . . .
The cognitive processing mode concerns the internal dialogue between the cognitive apparatus and stored mental representations of the physical environment, under the supervision of a conscious evaluator and the monitoring of attentional and intentional processes. (Paillard, 1987, p. 63)

Paillard (1987) confronts the dispute over whether spatial information is processed by sensorimotor responses to the immediate environment (eg. automatic reflexes), or by cognitive operations. This is described as a debate of “trenchant opposition between behaviourist and cognitivist theories” (p. 43). Paillard reviews a variety of fundamental spatial tasks and skills such as maintaining the perceived stability of visual space despite eye and body movements, visually fixating a stable target despite eye and body movements, smooth pursuit visual tracking of a moving target, quick programmed eye movements to fixate on a newly appearing target, bodily pointing at a visual target, and processes of adaptation to the displacement of the visual field as a result of looking through prisms. In each of these spatial skills Paillard describes how they can be accomplished by sensorimotor functions (perceptual responding based on automatic reflexes) and cognitive operations (conscious responding based on mental representations). It is clear from the discussion that both sensorimotor and cognitive modes of responding are active for all types of spatial processes.

Likewise, in their studies of locomotor patterns, Baratto and Colleagues (1986) develop a “motor cognitive model” (p. 79) which considers “movement as a cognitive process” (p. 81) in which high-order motor cognitive processes control lower-order motor subsystems. They reevaluate several pathological movement behaviors according to this model. For example, patients with central nervous system lesions have difficulty in motor learning whereas patients with peripheral impairments (eg. insufficient muscular force, limited joint motion) still have a large capacity for motor learning, “expressing a high degree of ‘motor intelligence’” (p. 80).

Other researchers also refer to high-order “motor knowledge” (Camurri et al., 1986, p. 88) which is processed at the “motor cognitive level, ie. at a level of motor or visuo-motor reasoning” (Morasso et al., 1983, p. 84). These “higher, cognitive levels” of the motor system can be considered to include “motor programming, organization, planning, and anticipation” (Thomassen, 1992, p. 250). Accordingly, researchers refer to “the perception and processing of kinesthetic spatial information” (Bairstow and Laszlo, 1980, p. 1), simply “kinesthetic information” (Keele and Ells, 1972) or “efferent information” and “proprioceptive information” in which the movement plan is considered to be “knowledge” represented as an “image or template of the motor commands” (Kelso, 1977b, pp. 42, 44). This leads to common hypotheses of “motor engrams”, a “motor image of a movement”, a “motor image of space” (Bernstein, 1984, pp. 99-102, 109) or simply as a “motor image” (Housner and Hoffman, 1979; Marteniuk, 1973; Posner, 1967).

One type of evidence for cognitive processes in kinesthetic-motor learning and memory come from studies is which Subjects imagine that they are moving. In these cases the motor image will interfere with other cognitive tasks just as if the body movements had actually been performed (Johnson, 1982; Marteniuk, 1986). Thus, movements, and movement images, appear to play a role in higher-order cognitive processes:

Viewed like this, movement learning has information processes held in common with a large number of cognitive problem solving tasks. Thus, . . . at least early in movement acquisition, there are cognitive information processes that underlie movement learning. (Marteniuk, 1986, p. 74)

Gardner (1983; 1990) distinguishes similar notions in terms of six types of “intelligence”. “Linguistic intelligence” includes semantics (verbal meanings), syntax (word order), pragmatics (word functions), and phonology (word sounds). “Musical intelligence” includes knowledge of pitch, rhythm, melody, harmony, timbre, and orchestration. “Logical-Mathematical intelligence” includes knowledge of number, quantity, sets, and operations. “Spatial intelligence” includes knowledge of form, transformations, structure, and kinesthetic and visual manipulation of real-world objects. “Bodily intelligence” includes knowledge of bodily movement capabilities. “Personal intelligence” includes knowledge of emotional, psychological, and social capabilities.

Aylwin (1988) also distinguished between visual, verbal, and motor “cognitive styles” (also called “representational styles” or “modes of thought”). Subjects’ visual, verbal, or motor cognitive styles could be determined by analysing their free associations to various stimuli. These cognitive styles were then positively correlated with Subjects’ performance measures on several different personality and aptitude tests.

Laban (1952) made a similar distinction between “action memory” which includes “the mastery of movement” as opposed to “verbal memory” which includes a knowledge of intellectual facts. He stresses that even though bodily movement actions can never be separated from cognitive thinking, that knowledge about the “logic of action” has not been well developed compared to knowledge about “verbal logic”. This “logic of action” cannot be learned from verbal descriptions only but must be experienced through actual bodily movement.

In another perspective, Jordan and Rosenbaum (1989) discuss how the motor system serves to move the sensory receptors to new places, thus allowing more information to be gathered by the perceptual systems. This information is then cognitively assessed in order to determine where, when, and how to move next. Because of this fundamental interrelationship between sensory perception and motor action they assert that “cognitive science, insofar as it regards perception as one of its core problems, cannot afford to ignore action” (p. 727).

Whiting (1986) also reviews the importance of human body movement within psychology and probes the question of why a subfield of psychology concerned with human movement has not been differentiated. It is pointed out that since virtually all behavioral and cognitive processes involve body movement that there is a “dualistic thinking implicit in trying to separate out movement from cognition” (p. 116). The practical, doing, “knowing how”, procedural knowledge has been separated from the conceptual, thinking, “knowing that”, propositional knowledge. Whiting questions this distinction and also the devaluation which is given to the non-verbal processes. He argues that cognitive abstract thought is essential for the learning of body movements and for their successful application in appropriate contexts (pp. 121–125).

Aspin (1977) considers “Knowing how”, as traditionally referring to procedural knowledge such as a knack or skill in games requiring bodily training and having a benefit of therapy or relief but which also lacks cognitive content. Whereas “knowing that” is traditionally considered to refer to propositional knowledge such as serious intellectual education and the rigor of academic disciplines which brings cognitive understanding. The distinction between these types of knowledge can be traced back to the ancient Greeks and was made explicit in modern times by Ryle (1949). However, Aspin (1977) disputes the basis of this distinction as a fallacy of the philosophical perspective of “‘essentialism’” which assumes that concepts (in this case “knowledge”) can be reduced to one or more essences. It is argued that language cannot be divorced from context and so a multitude of different types of “knowledge” (not only knowing-how and knowing-that) can be identified according to the particular context in which they occur (pp. 28-29). Furthermore, the traditional conception of knowing-that versus knowing-how cannot be sharply distinguished since one will always be based on the other. Knowing-that (propositional knowledge) will always be based on knowing-how to justify and support the knowledge and knowing-how to use a symbol system to express it. Conversely, Know-how (procedural knowledge) will always be based on knowing-that certain criteria constitutes success and that certain aspects of the overall context must be considered for successful performance. Therefore, there is a “fundamental connectedness” and interdependence between these types of knowledge, not a sharp distinction (pp. 23-28).

Psychologists have been traditionally most interested in studying propositional verbal processes, and more recently spatial processes have received attention, but body movement processes have been neglected. This is true in spite of their most basic role in all forms of cognition (eg. subvocal articulation movements as the basis for verbal perception and rehearsal; Baddeley, 1990, pp. 71-81; Baddeley et al., 1981; Hintzman, 1965; 1967; Levy, 1971).

Whiting (1986) begins to integrate body movement into psychological study. A hierarchy of body movements is conceived, including “movement elements” (single action of a group of motor effectors), “modular units” (a series of movement elements), and “movement actions” (an “orchestrated” group of modular units) (p. 131). As learning proceeds the control of movement is thought to shift to higher levels in the hierarchy and so becomes more and more subject to cognitive input. Thus, a “cognitive motor system” (p. 133) is proposed:

. . . it is wished to postulate a cognitive-motor representation system as part of a more general cognitive representation system responsible for upper-level control of human physical actions. The general conception therefore of a cognitive motor system is in terms of a hierarchical organization, the various levels of which reflect a shift from representations of, at the lowest level, specific movement elements, to, at the highest level, motor plans or cognitions of more general operations. (Whiting, 1986, p. 133).

Also, in studies of motor-enactment (see IIC.10) the typical finding is that acting-out the content a verbal sentence greatly increases the likelihood that that sentence can later be verbally recalled. Saltz (1988) also found that covert motor enactment (ie. very small, not noticeable, movements used to act-out the sentence) resulted in as good verbal sentence recall as overt motor enactment. Since motor processing (ie. “M–processing”) results in automatic deep learning of verbal phrases, it can be seen as a separable memory subsystem (Engelkamp, 1988a; Engelkamp and Zimmer, 1984; Saltz and Donnenwerth-Nolan, 1981):

At a theoretical level, the existence of covert M-processing would support the writer’s assumption that motoric factors can be conceptualized as an aspect of cognitive representational systems. (Saltz, 1988, p. 412)

The exception to the predominance of psychological research into propositional knowledge is research in motor skill learning, but here the laboratory experiments use minor actions which are separated from cognitive requirements of a surrounding environmental context and so have little (if any) ecologic validity. When body movement is observed in its natural setting, and when movement actions occur in complex situations, then cognitive knowledge of the surrounding situation is critical for successful selection and implementation of bodily movements.

Whiting (1986) reasons that one factor leading to this neglect in studying body movements may be the methodological difficulties involved in trying to quantify their attributes. Even though body movement is familiar to everyone it is also elusive and its “vocabulary is difficult to codify” (p. 124). Likewise, In Morasso’s (1983b, p. 187) attempts to use verbal descriptions of three-dimensional arm/hand trajectories it is noted that “simple experiments of this kind reveal the dramatic inadequacy of natural language to express movements and spatial relations”.

A good example of this problem can be seen in the lexicon of “motor knowledge”, or “motor language” presented by Cammurri and Colleagues (1986, pp. 104, 116-124) which consists of an assemblage of dance and movement terms without any consistent underlying analysis of their interrelationships. This problem can be informed by the movement categories and terminology developed in choreutics. The first steps toward a more explicit taxonomy of motor knowledge is taken in this present research.

IIC.33 Kinesthetic Basis for Spatial Knowledge.

The varied nature of what can be considered to be “spatial knowledge” can be grasped by reviewing the wide variety of what is considered to be “spatial” stimuli and tasks. This review reveals that the kinesthetic perceptual-motor system is fundamental for the learning and recall of spatial knowledge. This notion that kinesthetic-motor activity is a principle method for gathering perceptual information about space can be traced back to the 1800s (Viviani and Stucchi, 1992, p. 230). A variety of “spatial” stimuli have been used in experimental research. Many of these overtly utilise the kinesthetic-motor activity, for example spatial positioning, abstract and skilled body movements, spatial localization, Corsi blocks, and many cases where figures are physically traced or drawn (for details see Appendix V). In some cases the sense of “touch” is used which is itself a type of kinesthesia since touching requires body movements and arises from the same receptors as does kinesthesia (see IIA.11). When learning about large-scale environments “Actual locomotion in space appears to be an almost essential condition for the construction of spatial representations” (Siegel and White, 1975, p. 26).

Covert kinesthetic-motor processes can also be identified within “visual” spatial tasks (eg. seeing various figures, arrays, or matrices). The only purely “visual” space would be the array of stimuli reaching the retina of an unmoving eye. However, in natural visual perception the eyes rarely remain still. Eye movements appear to be essential to visual spatial perception. Efferent data about motor commands sent to the six extrinsic muscles of each eye and kinesthetic stimulation arising from those muscles provides information about eye movements and eye positions. These play a critical role in spatial perception.

When the eyes move then the visual stimulation moves across the retina. A classic question in psychology concerns how a perceiver is able to distinguish whether the environment is moving or whether one’s own eyes are moving. This perceptual problem is generally believed to be resolved by an efferent copy of the motor commands sent to the extrinsic eye muscles. This data is provided to visual perceptual processes which determine whether the visual motion across the retina is attributable to environmental motion, body motion, or a combination of both (eg. Grüsser, 1986b). In Helmholtz’ early conception of efferent data it was conceived as knowledge of the “‘effort of will’” (Matin, 1972, p. 368). Grüsser (1986a) reviews the history of similar ideas of “interaction theories of visual perception” from pre-Socratic philosophers to to its development as the “reafference principle” by Von Holst and Mittelstaedt (eg. Jeannerod et al., 1979; Von Holst, 1954).

Kinesthesia and the “oculomotor system” (extrinsic eye muscles) are generally thought of as two separate systems (eg. Craske and Crawshaw, 1974a, p. 106). Eye movements must be correlated with retinal vision during the process of visual perception. Other, more typically “kinesthetic” receptions often must also function closely with vision. For example, if the visual array is moving across the retina as a result of hip/knee/ankle articulations (eg. walking), while the extrinsic eye muscles hold the eyes still, then kinesthetic stimulations (including efferent data) provide information which leads to a perception of a moving body within a stable environment (rather than vice versa). Kinesthesia (including efferent data) relative to any body movements (including eye movements) plays an essential role in “visual” spatial perception.

Kinesthesia and efferent data about eye and body movements within visual perception are sometimes referred to as the “oculomotor” aspects (Israel and Berthoz, 1992, p. 196), the “extraretinal signal” (Jeannerod, 1983, p. 4; Skavenski et al., 1972), or as “extraretinal information” (Paillard, 1987). Matin (1972) notes that “The terms extraretinal source, signal or influence are intended to refer to any channel in which information is not derived from stimulation of the retina by light” (p. 332), that is, information about movements of the eyes or other body-parts. Likewise, when head/neck movements are included together with eye moves it can be termed the “eye-head motor system” (Bizzi, 1974, p. 106).

Support for the role of efferent data within visual perception can be found in experimental effects. When the eye is moved with an exterior apparatus by the Experimenter then Subjects are unaware of their eye having moved and so exterior objects, rather than the eye, are perceived to have moved. A similar experience can be obtained by pulling on the skin around one’s own eye causing it to move and resulting in a visual impression of the surrounding environment in motion. Conversely, if a Subject executes motor commands for eye movements, but the eye is restrained by an exterior apparatus, then Subjects still perceive (erroneously) that their eyes have moved and so exterior objects (which are actually stationary) are also perceived to move (Brindley and Merton, 1960; Irvine and Ludvigh, 1936). In another type of task, when viewing a visual picture the movements of the eyes are critical for selectively focusing on various aspects of the picture during learning. Recognition memory performance for pictures has been shown to positively correlate with the number of visual fixations each picture received during learning, regardless of the overall amount of time spent looking at the picture (Loftus, 1972).

These effects indicate that the role of efferent data (central feedback, or “outflow”) appears to be more important than sensory information (peripheral feedback, or “inflow”) (Matin, 1972, p. 368; Skavenski et al., 1972). However the response of muscle spindle receptors is minimal when movements are imposed externally rather than produced voluntarily (see APX. II.43). Thus, in normal conditions it may be that efferent data and muscle spindle response work together. Matin (1972, pp. 371-373) discusses the possibility of this “hybrid mechanism” which includes both afferent and efferent information within visual-kinesthetic spatial perception.

IIC.34 Kinesthetic-motor Mechanism for Spatial Calibration.

The kinesthetic mechanism for spatial perception is also evident in the phenomenon of “adaptation to displaced vision” (for details see Appendix VI). When a discrepancy is introduced between perceptual systems (eg. by wearing prisim-glasses which displace or reverse the visual image) then the interpretation of kinesthetic stimulations will adapt so that the different perceptual systems “read” the same (Gibson, 1966, p. 122; Moulden, 1971; Rock and Harris, 1967). This ability to adapt is referred to as “perceptual-motor plasticity” and is necessary to develop a calibration among the spatial stimulations arising from different perceptual systems (Held, 1968; Held and Freedman, 1963) and also to calibrate spatial information arising from different body-parts (Kenny and Craske, 1981; Lackner, 1973; Putterman et al., 1969; Rock and Harris, 1967, p. 101). A kinesthetic-motor mechanism is evident in this perceptual spatial calibration since active voluntary movement (rather than passive) produces the greatest and fastest adaptation (Held and Freedman, 1963; Held and Gottlieb, 1958; Held and Hein, 1958). Voluntary movement is not always necessary for adaptation (Howard et al., 1965; Mather and Lackner, 1975) but it may be that the increased kinesthetic and efferent data arising from voluntary movements elicits the greatest adaptation (Lackner, 1977a). These adaptation effects indicate that kinesthetic information (including efferent data) is at the basis for calibrating the different perceptual systems so that a unitary spatial perception is derived.

IIC.35 Kinesthetic-motor Mechanism for Spatial Rehearsal and Memory.

Since kinesthetic-motor activity is associated with spatial perception it is a logical extension to suggest that it is also associated with the mental rehearsal (eg. imagination) of spatial information. It is well known that rapid eye movements accompany dreaming while asleep. These usually occur together with other body movements and they are believed to be involved in the visual imagery accompanying dreaming (Aserinsky and Kleitman, 1953; 1955). Likewise, Deckert (1964) found that the eye movements which occurred when Subjects visualised a swinging pendulum (with eyes closed) were virtually identical to eye movements occurring when actually watching a pendulum. This kind of evidence leads investigators to believe that eye movements and other bodily orientation movements serve as a mechanism for spatial imagination (eg. Berlyne, 1965).

Incidental observations of movements during spatial cognitive processes have also been noted. Byrne (1974) observed that when Subjects were imagining a spatial matrix that “some Subjects made noticeable finger or head and eye movements that traced out the matrix” (p. 57) and noted twice how subjects made statements about needing to use eye movements to scan the spatial images (pp. 56, 58). Thus, a kinesthetic-motor mechanism underlying the use spatial imagination is suggested (p. 59). Brooks (1967) also observed that Subjects frequently glanced away from the answer sheet when recalling an imagined matrix and that “Subjects explained the glancing away as ‘getting the pattern back’” (p. 294).

Idizikowski and Colleagues (1983) tested a hypothesis that extraneous eye movements might disrupt spatial imagination. Reflexive rotational nystagmus eye movements had no effect on the simultaneous performance of Brooks’ matrix (for description of Brooks’ matrix see Appendix V). However, they suggested that the spatial “rehearsal-controlling process may not be the eye movements themselves but rather the central system involved in their voluntary control” (p. 231). Voluntary visual tracking, with or without corresponding visual field motion across the retina, did result in a decrement to Brooks’ spatial matrix recall, but not to Brooks’ verbal material. Voluntary stationary eye fixation had no effect on performance. These results indicate that voluntary eye movement disrupts the spatial imaging process.

Quinn and Ralston (1986) built a large matrix (0.4 m²) wherein Subjects could move their hand through the squares while learning and imagining the Brooks’ matrix material. Moving the Subjects’ hands either actively (Subject voluntarily moves) or passively (Experimenter manipulates Subject’s relaxed arm) into the wrong squares resulted in a decrement to the matrix recall, while moving into the correct squares or tapping in place did not result in any change to recall performance. This indicates that incompatible movements disrupt the spatial imagining process.

IIC.36 Kinesthetic-motor Basis for Cognition in General.

Many theorists go even further and purport that kinesthetic and motor activity plays a fundamental role in all types of cognition. McGuigan (1978) offers an extensive review of research which has reported consistent patterns of muscular activity throughout the body during various cognitive tasks. McGuigan proposes that skeletal muscular activity related to cognition generates “nonlinguistic coding . . . a more primitive kind of symbolism than that needed for language” (p. 88). When certain stimuli arouse certain muscular activities the kinesthetic information from the muscular activities becomes part of the mental representation of the stimuli. When perceiving the same stimuli again, or simply thinking about, imagining, or dreaming about the stimuli, the muscular activity is repeated.

Weimer (1977, p. 302) distinguishes the “motor theory” from the “muscle theory”. According to the muscle theory, activity of the muscles are utilised within cognitive processes (eg. McGuigan, 1978; see above). However, the motor theory is about central nervous system activity rather than the peripheral muscles. Viviani and Stucchi (1992) describe how according to the motor theory “motor processes enter into the genesis of percepts” (p. 230). The motor actions used to explore the environment and gather sensory data will be included with the perceptions of that environment, but the motor theory is “not contingent upon actual execution of motor actions” (p. 231) but is based on the cognitive “activation of stored motor routines” (Jordan and Rosenbaum, 1989, p. 727):

A motor theory of memory need not necessarily involve actual movements. Visual or auditory input might be converted to motor commands that control muscle movements, and the motor commands may be remembered whether or not the movement is actually initiated. (Keele, 1968, p. 387)

Watson (1924, pp. 264-265) was an early advocate of body movements being at the basis of thought and verbal processes and points out that movements are intelligently organised in the infant long before words, and that motor actions are constantly forming without words throughout life. Words are thought to gradually develop in order to name the underlying motor processes and eventually the verbal organisation becomes dominate. Motor processes are thought to make up the Freudian “unconscious” and the emotional processes which are conceived to be more basic to thought than words. Laban also expresses a similar conception of motor actions as the basis for cognition:

The words of language, giving names to objects and thoughts, conceivably sprang into being in remote times from movement impulses which were made audible. Thinking is certainly a kineto-dynamic process, and its trace-forms [ie. movement pathways] (presumably complicated shadow-forms [ie. covert pathways], noticeable in free space lines) will one day be discovered. (Laban, 1966, p. 124)

Coren (1986) reviews the history of motor theories of perception (“efferent theories”) and identifies their fundamental conception that the primary function of the brain is to produce motor actions:

This postulate is that the brain, viewed objectively, is primarily a mechanism for governing motor activity. Its raison d’être is the transformation of sensory patterns into patterns of motor coordination. This viewpoint is, of course, quite out of keeping with the generally accepted notion that the major functions of the brain are the manufacture of ideas, feelings, the storage of memory, and the interpretation of sensations into a conscious representation of the external environment. Such subjective phenomena . . . may simply be epiphenomena - the byproduct of brain activity - rather than its targeted functional result. . . . When reduced to its essence, the fundamental interpretive task of the brain . . . is to transform the sensory inputs into motor programs that allow the organism to interact with the external environment. (Coren, 1986, p. 394)

The traditional view of motor actions as interpreting the mind can be contrasted with this integrated view of motor actions as constructing the mind:

From the time of Aristotle it has been taught that the motor system is the chattel of the sensory system. Nourished by the senses the motor system obediently expresses in automation and relatively uninteresting fashion the cleverly contrived ideas of the higher mental processes, themselves offshoots of the sensory mechanisms. In this view, action is interpretive of the sensory mind . . . [As an alternative to this] a constructive theory of mind [is advocated] in which it is argued that higher mental processes in addition to perception are skilled acts that reflect the operating principles of the motor system. In short, experience is constructed in a fashion intimately related to the construction of coordinated patterns of movement. (Turvey, 1977, pp. 211-212)

Similar to this is the “principle of somatotopy” (Trevarthen, 1978, pp. 110-113) according to which kinesthetic-motor activity plays a critical role in all areas of brain development. Therefore, the actual structure of the brain can be described as “body-shaped maps” and “mechanisms of the brain [are believed] to be laid down anatomically in close correspondence of motor function”.

Gyr and Colleagues (1979) reason that the kinesthetic-motor mechanism in adaptation to displaced vision (see IIC.34) supports the notion that there is “an active involvement of areas of the motor cortex and motor-sensory feedback systems in the perceptual process” and so “neurological processing of sensory input is dependent on the organization of motor activity” (p. 59). Since both sensory and motor aspects are involved, they refer to these as “sensorimotor theories” of perception.

The influence of motor commands on perception is well known in studies of visual perception. Motor commands which produce eye movement (ie. extraretinal information) appear to play a vital part in the visual perception of whether the exterior environment is perceived to be moving or whether the body is perceived to be moving (see IIC.33). Coren (1986) has also demonstrated that efferent motor commands operate to produce varieties of visual illusions in configurations of line segments (eg. Mueller-Lyer illusion).

In specific, the motor theory of speech perception posits that the perception of spoken language is not based on auditory recognition of sounds which are translated into meaningful words, but that language is perceived from a recognition of the physical movements which would have to be made in order to make those sounds (Viviani and Stucchi, 1992, p. 230). This may consist of a “decoder” which perceives audio phonemes according to neoromotor commands of the muscles which would be used to produce the sounds (Liberman et al., 1967). That is, speech perception requires “the hearer to utilise the same central neural machinery that would be involved in speaking” (Weimer, 1977, p. 283). The motor commands which would be necessary to make the speech sound are used to interpret the phonic stimuli, but these motor actions are known internally rather than having to be fully enacted (Ibid, p. 282). Likewise, Sperling (1967) developed a model in which the visual image of a letter is stored in memory and rehearsed as a “‘program of motor-instructions’” (p. 291). This is sometimes conceived in terms of subvocal articulatory movements which appear to form the basis of verbal rehearsal rather than the auditory sounds of the words (Baddeley, 1990, pp. 71-81; Baddeley et al, 1981; Hintzman, 1965; 1967; Levy, 1971).

The motor theory of speech perception can be extended to the perception of other types of stimuli. The perceptual systems are seen by Weimer (1977) as “motor skills production systems” since the act of perceiving is seen as an act of constructing motor actions which correspond with the stimulus (p. 283). This process of constructing motor actions during perception is equally applicable to the processes of imagination and hallucination (p. 289).

The motor theory of perception is most readily applicable to spatial perception because of the critical role of body movements during spatial learning. According to the motor theory, knowing the location of an object in space means knowing the motor commands and movements necessary to reach the object (Viviani and Stucchi, 1992, p. 232). This agrees with Piaget’s view that “in order to know objects, the subject must act upon them” and so “knowledge is constantly linked with actions or operations” (Piaget, 1970, p. 104). The development of abstract spatial representations through motor actions has been explored in detail in the classic work on childrens’ spatial learning by Piaget and Inhelder (1967). Accordingly, in sociological studies Hall (1966, p. 108) points out that “perception of space is dynamic because it is related to action - what can be done in a given space - rather than what is seen by passive viewing”.

Shapes of objects are originally learned by correlating visual appearance of seeing the shape with the kinesthetic experience of touching the shape. When later seeing the same shape the kinesthetic-motor actions of touching such a shape are recalled (without actually touching the shape or doing the movements again). Festinger and Colleagues (1967) found support for this theory by training Subjects to associate a visually straight line with a curving body movement by having Subjects touch curved objects while looking through prisms which made the objects appear straight. This visual/motor relationship led to perceptions of curved body movements as indicating a straight line in visual space. They conclude that visual perception is based on “preprogrammed efferent [motor] instructions that are activated by the visual input” (p. 34).

This interior knowledge of motor processes which serves as the basis of motor theories of perception is referred to by Viviani and Stucchi (1992) as “implicit motor competencies”. That is, an implicit understanding of biomechanical “motor regularities and constraints” which limit the variety of body movement characteristics (p. 235). The influence of implicit motor competencies on perception has been shown is a variety of studies.

When a sequence of similar stimuli are presented in slightly different spatial locations the Observer will frequently perceive this as a single stimuli which is moving through the space. This is known as “apparent motion” (Johansson, 1975; Kolers and Pomerantz, 1971). A common example is the perceived movement of lights on a theatre sign resulting from a sequence of lights turning on and off in sequence. The typical effect is that apparent motion will be perceived along the shortest possible path from the first stimuli to the next (eg. Shepard, 1984). This same type of apparent motion will be perceived when viewing a sequence of human body poses. However, in this case the pathway of the illusory movement will conform to anatomic constraints even if this is not the shortest distance between the position of a limb in one photo to the next photo (Shiffrar and Freyd, 1990). This indicates an implicit understanding of biomechanical constraints which exerts an automatic influence on perception.

When lights are attached to the major joints of a human body (eg. knees, hips, etc.) and these are observed in darkness (ie. the lights can be seen but the actual physical body cannot be seen) Subjects will instantly recognise movements of the lights as being a human body (Johansson, 1973; 1977). These findings indicate an implicit knowledge of human movement constraints. What is more significant for the motor theory of perception is that Subjects would have had the greatest amount of visual experience seeing their friends’ gaits, and the least amount of visual experience seeing their own gait, yet when viewing the gaits on a video recording (showing only a group of twelve lights, one at each of the major skeletal joints, and thus unable to see the actual physical body) Subjects recognised their own gait better than they recognised their friends’ gaits (Beardsworth and Bukner, 1981). This supports the idea that Subjects recognised their own gaits best because of their implicit knowledge of their own personal motor processes.

Viviani and Stucchi (1992) review a biomechanical constraint which specifies a “functional relation” between the velocity of the hand (in a multi-joint movement) and the degree of curvature of the hand’s pathway. This relation is precisely specified in a ratio known as the “two-thirds power law” (p. 236). In general terms the hand’s velocity will decrease in regions of high curvature and will increase where the curvature is more flat (p. 239). They characterise this constraint as a “biological signature” and review evidence that perception will implicitly conform to this signature.

Viviani and Stucchi (1991) showed Subjects a spot of light moving in an ellipse or random curved patterns. When the velocity of the moving light was actually constant Subjects perceived that its velocity was changing. Subjects attempted to adjust the velocity of the light until it appeared to be moving at a constant speed throughout the pathway (the velocity was automatically allowed to be variable from one part of the path to another, unbeknownst to Subjects). When the light appeared to Subjects to be moving at a constant speed, in actuality the velocity modulations and the degree of pathway curvature were conforming with the two-thirds power law. Thus it appears that the motion of the light was perceived according to implicit knowledge of motor constraints.

Subjects were also shown recordings of movements of lights attached to the major joints of a human body and asked to adjust the velocities of the lights until the body movement looked as “natural” as possible. In this case they also adjusted the velocities to conform with the two-thirds power law (Viviani and Stucchi, 1992, pp. 240–242).

“Kinesthetic spatial cognition” is defined as cognitive processes (eg. imagery, mental manipulations) which are performed on kinesthetic spatial information. Support for this concept is built-up from psychological theory. A great deal of research has distinguished spatial cognition from verbal cognition as using separate cognitive resources. Spatial information can arise from separate visual, audio, and kinesthetic perceptual-motor systems but is eventually represented in a unitary spatial memory system. Kinesthetic-motor knowledge is considered by many researchers to inherently require cognitive processing rather than consisting solely of sensory-motor responding. Kinesthetic-motor activity has long been identified as being at the basis of all spatial learning and is hypothesised to function as a spatial rehearsal mechanism (eg. eye movements). Body movements also appear to serve as a mechanism whereby spatial information arising from different receptors is compared and calibrated so that the various spatial sensations “read” the same. Many theorists also purport that kinesthetic-motor information is at the basis of all types of cognitive processes (including verbal). This concept of “kinesthetic spatial cognition” has not been heretofore explicitly developed in cognitive psychology and so constitutes new knowledge. This provides a cognitive and motor control context in which to reevaluate choreutics.