Bernstein's Level of Movement Construction A Contemporary Perspective

Human Movement Science 57 (2018) 111–133
Contents lists available at ScienceDirect
Human Movement Science

journal homepage: www.elsevier.com/locate/humov
Bernstein’s levels of movement construction: A contemporary

T
perspective
⁎
Vitor L.S. Profetaa, , Michael T. Turveya,b
a
Center for the Ecological Study of Perception and Action, University of Connecticut, Storrs, CT, USA
b
Haskins Laboratories, New Haven, CT, USA
AR TI CLE I NF O AB S T R A CT
Keywords: Explanation of how goal-directed movements are made manifest is the ultimate aim of the field
Bernstein classically referred to as “motor control”. Essential to the sought-after explanation is compre-
Control hension of the supporting functional architecture. Seven decades ago, the Russian physiologist
Coordination and movement scientist Nikolai A. Bernstein proposed a hierarchical model to explain the con-
Synergy
struction of movements. In his model, the levels of the hierarchy share a common language (i.e.,
Movement construction
they are commensurate) and perform complementing functions to bring about dexterous
movements. The science of the control and coordination of movement in the phylum Craniata has
made considerable progress in the intervening seven decades. The contemporary body of
knowledge about each of Bernstein’s hypothesized functional levels is both more detailed and
more sophisticated. A natural consequence of this progress, however, is the relatively in-
dependent theoretical development of a given level from the other levels. In this essay, we revisit
each level of Bernstein’s hierarchy from the joint perspectives of (a) the ecological approach to
perception-action and (b) dynamical systems theory. We review a substantial and relevant body
of literature produced in different areas of study that are accommodated by this ecological-dy-
namical version of Bernstein’s levels. Implications for the control and coordination of movement
and the challenges to producing a unified theory are discussed.
1. Introduction
Nicolai Bernstein’s essay Levels of Construction of Movements was written some 70 years ago. An English translation from the
original Russian by Mark L. Latash was published in 1996. The publication was in a volume entitled Dexterity and its Development
(Latash & Turvey, 1996) that included contributions by a dozen contemporary movement scientists. At the core of Levels of Con-
struction of Movements is a hierarchical neural model. The model is framed in terms of the evolution of the nervous system in the
phylum Craniata (see Margulis & Schwartz, 1998). It proposes that each level of the hierarchy evolved to solve a particular class of
movement problems.
Anatomically, the hierarchy is built bottom-up with older levels at the base of the hierarchy. Functionally, the hierarchy operates
top-down with the upper and younger of any two levels taking advantage of the functional capabilities of the lower and older level so
as to decrease the upper level’s involvement. Any movement requires at least two levels, establishing a relationship of leading and
background levels. The “leading” refers to the upper level that controls the given movement and the “background” refers to the level
or levels that provide the necessary support for the given movement.
⁎
Corresponding author.
E-mail address: vitor.da_silva_profeta@uconn.edu (V.L.S. Profeta).
https://doi.org/10.1016/j.humov.2017.11.013
Received 9 August 2017; Received in revised form 24 November 2017; Accepted 27 November 2017
0167-9457/ © 2017 Elsevier B.V. All rights reserved.
V.L.S. Profeta, M.T. Turvey Human Movement Science 57 (2018) 111–133
The neural-based treatment that Bernstein gave to the construction of movement was necessary to bring different movement
demands into a common language. Despite Bernstein’s efforts to apprehend the wholeness of the system of movement organization,
the enterprise, both past and present, has been to study the levels independently. Considerable progress has been made in this
endeavor. In our opinion, it would be fruitful (a) to summarize this progress and (b) to identify, as Bernstein did, a means to reconcile
the levels of movement construction in terms of a single and coherent model.
In this essay, we present what might be termed an ecological-dynamical perspective on Bernstein’s model of movement con-
struction or assembly. An ecological perspective is grounded in the assumption of mutuality between an animal and its environment.
More specifically, understanding any given animal’s behavior requires an appreciation of what behaviors the animal’s environment
affords (Gibson, 1979): the animal-environment system constitutes the unit of analysis. In turn, a dynamical system approach to
movement relies on mathematical concepts of nonlinear dynamics to describe and interpret coordination (e.g., Kugler, Kelso, &
Turvey, 1980), which is understood as an emergent state produced by self-organizing processes (Kelso, 1995; Kugler & Turvey, 1987).
The ecological and the dynamical perspectives seek lawful regularities in behavior and advocate parsimony in the promotion of local
executive control.
Our goal is to establish the state of the art at each of Bernstein’s levels from the ecological-dynamical perspective. More than
offering definitive answers, we offer a viewpoint and suggest (to movement scientists in general and, in particular, those abiding an
ecological-dynamical perspective) a way to grasp the big picture of the coordination and control of movement.
The remaining of this essay is organized in four sections. In Section 2, we revisit Levels of Construction of Movements identifying its
primary themes. In Section 3, we present the bulk of the essay, a re-examination of each of Bernstein's levels in the light of con-
temporary conceptual advances and experimental results. In Section 4, we offer some thoughts towards reconciliation of the levels as
a modern version of Bernstein’s original attempt. In Section 5, we conclude presenting an agenda to push further the ecological-
dynamical perspective presented in this essay.
2. Bernstein’s hierarchy of movement construction
Fig. 1 shows schematic relations among the levels comprising Bernstein’s hierarchical account of how movements are constructed.
The first level is the Level of Tonus. This level establishes the communication between the neural and muscular systems. In Bernstein’s
words, this level speaks the “muscle language”. Residing in the spinal cord, the first level’s function is to prepare the movement
apparatus to respond adequately to commands (influences, instructions, constraints) coming from upper levels of movement con-
struction by changing the excitability of sensory and motor cells. Consequently, it is never a leading level in the performance of a
purposive movement. It is always strictly a background level. The form that a movement takes in solving a particular movement
challenge is never defined at the Level of Tonus. Without the Level of Tonus, however, the movement apparatus would be unprepared
to respond adequately to upper-level commands and the intended movement form would never be made manifest.
The Level of Synergies (or the Level of Muscular-Articular Links) is the second level of movement construction in Bernstein’s
hierarchy. Residing in the middle brain, it constrains the degrees of freedom of the motor apparatus and guarantees coherence in
movements by controlling large choirs of muscles. Synergies are shaped by the inflow of proprioceptive information about the whole
body. Besides its relevance and unique function, the Level of Synergies happens to be, like the Level of Tonus, a background level
when one performs any movement oriented to an external goal. Its function is to correct the details of movements. That is the case
because the Level of Synergies is not informed by (does not receive afference from) either the visual or auditory systems, thus it is not
Fig. 1. Schematic representation of functional interaction among Bernstein's levels of construction of movements. Arrows indicate directions of dominance between
levels, from leading to background levels. See text for details.
112
able to anticipate or compensate for any non-mechanical environmental perturbation.

The Level of Space is the third level in Bernstein’s hierarchical model. In contrast to Levels 1 and 2, Level 3 can lead to purposive
movements. It can exhibit dexterity1. The Level of Space refers to how one performs a goal-directed movement in one’s surroundings.
In accord with its duties, the Level of Space has an inflow of optical, acoustical, and haptic stimulation that generates the space field,
a perception of the external space based on the interweaving of different sensory sources and previous experiences. The space field
performs a dual function: objective perception of the relations between the body and external space (i.e., an agent’s surrounding
layout of surfaces, stationary or moving), and the ability to use external space. The space field supports (a) classes of movements that
entail displacing the body or segments of the body from one location to another and (b) classes of movements that entail object
manipulation. The latter two classes of movement problems are resolved by two different neural subsystems, the extrapyramidal and
the pyramidal, respectively. Furthermore, anticipatory behaviors (i.e., movements employed to synchronize with up-coming en-
vironmental conditions) are under the control of the Level of Space. Importantly, when acting upon the Level of Synergies, the Level
of Space does not impinge how a movement is produced; instead, it informs the Level of Synergies that a given goal-directed
movement is possible. A limitation of the Level of Space is its inability to control sequences of movements. The fourth and highest
level of the hierarchy, the Level of Action, performs this role.
The Level of Action2 resides in the frontal cortex. It is responsible for controlling sequences of movements (i.e., ‘actions’ in
Bernstein’s terms) organized to accommodate a behavioral problem that requires many steps in its solution. Following this definition,
the order of elements in a sequence is of singular import to reaching the goal of the action, since each individual movement provides
an intermediate response to a specific aspect of the movement problem. Besides that, different elements can be organized in different
ways to achieve the same goal, this assures flexibility of the Level of Action. In that sense, there are innumerable ways to solve a
particular movement problem, although each way has its own intrinsic organization. This last fact highlights an important char-
acteristic of actions: the meanings of the parts, the specific roles that they play, can only be defined and understood in the context of
the whole functional unit that they compose.
3. The hierarchy of movement construction from an ecological- dynamical perspective
3.1. Level of tonus
For Bernstein, the background muscular-contraction level (the Level of Tonus) constitutes basic priming in the preparation of the
construction of movement patterns. The word tonus follows etymologically from words meaning tuning or fine adjustment (Walsh,
1992), and muscular tonus inherits this sense as a property that guarantees that the movement apparatus is prepared to perform
qualitatively different movements. Despite its theoretically intuitive relevance, tonus is ill-defined and there is little agreement
among movement scientists as to its precise nature. As pointed out by Fenn and Garvey (1934), p. 383) in the decade prior to
Bernstein’s “Levels”, tonus is a “convenient term which includes many different properties such as elasticity, viscosity, and con-
tractility….”
3.1.1. The orthodox view of tonus

In accord with Bernstein’s ideas, tonus has been understood as a minimal level of muscle tension in the resting condition regulated
by neural activity that resists stretching and maintains a given body or limb posture. This traditional view of tonus is coordinate with
two complementary ideas, namely, that muscle fibers traverse from one myotendinous junction to another and that the only way
transmission of tension is produced within a sarcomere is serial, to adjacent sarcomeres. Underlying these two ideas is the assumption
of independence between myofibrils and the connective tissue adjacent to them. The latter implies (a) that the level of muscular
tension observed at rest is the result of a basal level of neural excitation and muscle mechanical properties, and (b) that changes in
tension are, primarily, due to changes in excitation. As anticipated above, the neural-centered view of tonus expressed by the
foregoing (a) and (b) ignores contributions of connective tissue in the maintenance of the overall basal tension of the body’s multiple
muscle-to-muscle and muscle-joint complexes. As will become evident, however, findings at different levels of analysis point to a
substantial contribution of connective tissue. We review some of these findings in what follows.
3.1.2. Connective tissue’s contribution to tonus

The main question raised in this section is how the tension produced within a muscle is transmitted to the myotendinous junction,
yielding tension on bones such that movements are produced and postures are maintained. In Section 3.1.1 we claimed that the
orthodox view of tonus is not supported by experimental results, which demonstrate a non-obvious contribution of connective tissue.
Connective tissue allows lateral force transmission and its contribution has been demonstrated at different levels of analyses. That is,
regarding tension distribution, connective tissue re-direct force to parallel paths increasing the complexity of interactions involved in
movement production at different scales. Huijing (2007) suggested three ways of lateral tension transmission, namely: in-
tramuscularly, intermuscularly, and extramuscularly (see Turvey & Fonseca, 2009, 2014 for overviews). We discuss each of them
separately.
1
Bernstein defined dexterity as a capacity to respond quickly and adaptively to environmental demands.
2
The term action in Bernstein’s model has a different meaning compared with what proponents of the ecological-dynamics perspective mean. For the latter, action
means goal-directed movements guided by perception. We will use action in interchangeable ways, making explicit distinctions when we see it is necessary.
113
Fig. 2. Schematic representation of the vinculin and dystrophin complexes. Each of these complexes bind to actin molecules contributing to the tension produced
within a sarcomere to be transmitted to laminin in the extra-cellular matrix (see text for details).
3.1.2.1. Transmission of intramuscular lateral tension. The transmission mode of intramuscular lateral tension refers to the fact that the
forces produced by sarcomeres, the fundamental functional units of striated muscle, are also transmitted to the extracellular matrix
(ECM) surrounding the myofibrils. This implies the force observed at a myotendinous junction is not only due to a sequential
shortening of sarcomeres.
Studies in molecular biology have shown that besides attaching to Z-disks3, actin also binds to both vinculin and dystrophin
complexes that are embedded in the sarcolemma. These two protein complexes4 connect with elements of the ECM, suggesting that
the tension generated within the sarcomere is transmitted to the ECM (Patel & Lieber, 1997). Dystrophin is associated with the actin
cytoskeleton and the dystroglycan complex (α and β units), which is associated, in turn, with laminin in the ECM (Monti, Roy,
Hodgson, & Edgerton, 1999). Similarly, vinculin also seems to mediate force transmission from actin to laminin (Monti et al., 1999;
Patel & Lieber, 1997). In this case, the actin cytoskeleton links to vinculin via talin, and vinculin links to an integrin complex (α and β
subunits) that also links to laminin. A schematic representation of these mechanical linkages qua associations is presented in Fig. 2.
The tension exerted on laminin is transferred to the endomysium through the basement membrane (Huijing, 1999; Trotter,
Richmond, & Purslow, 1995). This raises the possibility of the endomysium contributing additionally to the lateral force distribution.
Evidence that such can be the case is provided by Street’s (1983) investigations of lateral force transmission due to connective tissue
within a muscle. In one experiment, one “half” of a muscle was dissected until only a few remaining deep fibers remained exposed,
leaving the other “half” with all transected fibers. A transducer was attached to the tendon of the intact “half”. With this preparation,
lateral force transmission was tested in two situations. One was fixing both ends of the muscle. The other was leaving only the
dissected “half” free. The dissected part was then stimulated and the force magnitudes transmitted to the tendon in the experimental
conditions were compared. Street observed that the force transmitted to the transducer with the dissected “half” free was about 85%
of the force generated when it was fixed. Of importance, if force were transmitted only serially, then the expectation would be a
shortening of stimulated fibers without a considerable amount of force registered by the transducer at the tendon of the intact “half”.
Additional evidence comes from Street’s second experiment. The central portion of a muscle was dissected, leaving only few deep
fibers connecting its two ends that remained intact. In this case, both ends of the muscle were fixed. Markers were then put on the
surface of the cells of the exposed area commensurate with the sarcomeres’ lengths. After stretching the muscle, it was observed that
the lengths of the sarcomeres were not constant. Sarcomeres of the dissected areas closer to the intact ends were shorter than
sarcomeres at the center. Because the length of a sarcomere refers to the amount of tension exerted upon it, one would expect the
absence of lateral force transmission to result in a constant length of the sarcomeres throughout a muscle fiber. However, that proved
not to be the case.
3.1.2.2. Transmission of intermuscular lateral tension. The transmission mode of intermuscular lateral tension refers to force
3
A Z-disk is a dark disk bisecting the transversely isotropic band of a striated muscle fiber.
4
Cytoskeletal proteins are polymers that compose the cytoskeleton and flagella/cilia of cells. They include tubulin, actin and lamin as components of microtubules,
microfilaments and intermediate filaments, respectively.
114
transmission between two adjacent muscles via connective tissues linking their muscle bellies. At the macro-anatomical level, the
perspective that muscle force is transmitted only serially assumes that the joint torque due to muscle activation is dependent on the
geometrical configuration of the muscle’s torque arm. This assumption underlies surgical distal tendon transfer procedures in which
the tendon of a muscle is attached to the tendon of its antagonist. The goal of this procedure is to turn the muscle that had the tendon
transferred into an antagonist of its self, i.e. if a muscle is classified as a flexor, it is expected that it will act as an extensor. It has been
shown, however, that after the surgery a transferred muscle preserves its original function (Riewald & Delp, 1997). It is suggested that
epimysium and scar tissue are the probable cause of such a counterintuitive result.
The hypothesis of force transmission via epimysium following tendon transfer was tested in a controlled setting with an animal
model (Maas & Huijing, 2012). Roughly, they evaluated whether the stimulation of the flexor carpi ulnaris (FCU) with rat’s wrist
placed at different angles would result in flexion or extension torque. They identified the neutral position of the wrist and set the joint
passively at different angles. In addition, they excited the muscle belly surface of FCU with values just above the excitation threshold.
There were three different conditions. One was characterized by only a minimal disruption of the connective tissue surrounding FCU
(the control condition) and two were characterized by transferring the distal tendon of FCU of rats to the distal tendon of their
extensor carpi radialis (ECR). In one condition the surrounding connective tissue was preserved. In another it was completely re-
moved. Results indicated that when the distal tendon of FCU was transferred to the distal tendon of ECR keeping surrounding
connective tissue intact, stimulation of FCU with wrist extended generated flexion torque. In contrast, stimulation of FCU with the
wrist flexed generated extension torque, a result opposite to what should be the case if torque was defined by the muscle moment-
arm.
The explanation of the foregoing proposed by Maas and Huijing (2012) relies on the orientation of the connective tissue at the
wrist’s position given the muscle’s length. When the wrist is placed passively at flexion, the extensors (transferred FCU included) are
longer than the flexors, an outcome favoring tension transmitted by connective tissue to extend the joint after stimulation of FCU. In
contrast, when the wrist is extended the flexors are longer than the extensors and the force transmitted by connective tissue favors
flexion (Fig. 3). Additional support for the transmission of lateral force through the epimysium is provided by changes in muscular
passive tension due to synergist length change (see Maas, Meijer, & Huijing, 2005).
3.1.2.3. Transmission of extramuscular lateral tension. The transmission mode of extramuscular lateral tension refers to the
connections of the epimysium of a muscle and its non-muscular surroundings (Maas & Sandercock, 2010). Evidence in favor of
muscular tension transmission to non-adjacent muscles via fascia has also been reported. The first evidence came from studies with
cadavers. Vleeming, Pool-Goudzwaard, Stoeckart, van Wingerden, and Snijders (1995) observed that the thoracolumbar fascia
attaches to many structures, including the arms, the pelvis, the spine, and the legs suggesting the possibility of fascia-mediated
tension transmission among these structures. As a test, they applied traction to latissimus dorsi and measured its effect on the
contralateral gluteus maximus. Their results showed that when traction was applied to the lower part of latissimus dorsi, tension
Fig. 3. Hypothetical mechanism proposed by Maas and Huijing (2012) of wrist's torque generated by stimulation of surgically transferred of flexor carpi ulnaris (FCU)
of rats, the black diagonal ellipse, to the tendon of the extensor carpi radialis (ECR) when FCU adjacent connective tissue is maintained. Wrist at neutral position
without benefiting any torque (Top). When rat’s wrist is at extension (Middle), wrist flexors are longer than wrist extensors and stimulation of the transferred FCU is
biased to transfer force to flexors, leading wrist flexion. In contrast, when rat’s wrist is flexed, (Bottom), wrist extensors are longer than wrist flexors and stimulation of
transferred FCU is biased to transfer force to extensors, yielding wrist extension.
115
Fig. 4. Pictorial representation of the fiber alignment of the thoracolumbar fascia (dark gray area) with the fascia of gluteus maximus (a), the fascia of gluteus medius
(b), the fascia of external oblique (c), the fascia of latissimus dorsi (d), the posterior superior iliac spine (1), the sacral crest (2), part of the lateral raphe (lr). Letters and
numbers hold for both sides of the figure. Arrows indicate, from cranial to caudal, the direction of traction of the trapezius, cranial, and caudal parts of latissimus dorsi,
gluteus medius, and gluteus maximus muscles, respectively. Same arrow arrangement applies to right side of figure. Adapted with permission from Vleeming, A., Pool-
Goudzwaard, A. L., Stoeckart, R., van Wingerden, J. P., & Snijders, C. J. (1995). The posterior layer of the thoracolumbar fascia| its function in load transfer from spine
to legs. Spine, 20(7), 753–758.
transmission was observed in contralateral gluteus maximus (Fig. 4).

Carvalhais et al. (2013) evaluated whether latissimus dorsi might transmit tension to contralateral gluteus maximus in intact
individuals through the thoracolumbar fascia. They showed that actively increasing tension in latissimus dorsi, without a significant
electromyographic activity of contralateral gluteus maximus, shifted hip resting position towards lateral rotation. In addition, the
amount of passive torque per change in joint angle increased.
3.1.3. Tonus as a mechanical property

The foregoing review of the transmission of lateral muscular tension suggests that the resistance of a muscle to stretch cannot be
considered independent of the resistance to stretch of the associated connective tissue. In this regard, we follow Turvey and Fonseca
(2014) and redefine tonus as a mechanical property of the muscular-connective tissue-skeletal (MCS) system.
It is important to note that understanding tonus as primarily a mechanical property as opposed to a strictly neural property does
not change the function that tonus serves. It guarantees responsiveness of the muscle-joint complexes in the production of move-
ments. That said, the understanding of tonus as a mechanical property invites a rethinking of the level’s structural organization as
conceived by Bernstein (1996). More explicitly, it invites an alternative characterization of that structural organization.
3.1.3.1. Tonus and tensegrity. The discussion presented above outlines two requirements for an alternative model of the architectural
organization of the level of tonus. First, the model must recognize the tautness of the MCS system at rest. Second, given the nested
organization of connective tissue (ranging from microbiological to macro-anatomical levels), the model must account for tension
propagation across levels. Tensegrity systems can account for both requirements.
A tensegrity system is characterized by the capacity to sustain its mechanical integrity by continuous tension over discontinuous
compression (Levin, 2006). The basic elements of a tensegrity system are cables (tensile elements) and struts (compressed elements).
Struts are embedded in a net of cables that exerts continuous tension over them (Fig. 5a). The system is stable such that when a
perturbation (i.e., external pressure) is imposed upon it, its shape changes (Fig. 5b) so as to redistribute the tension over the entire
system. Importantly, this is the case because local tension never reaches zero (Fig. 5c) even in the absence of an imposed force. In
other words, a tensegrity system is prestressed (Levin, 2006).
It has been suggested that the muscular-connective tissue-skeletal system is both morphologically and functionally like a ten-
segrity system (Levin, 2006; Turvey & Fonseca, 2014). Morphologically, bones are the compressed elements, whereas muscles and
connective tissues form the net of tensed elements. Functionally, much like a conventional tensegrity system, the MCS system exhibits
prestress as demonstrated by Souza, Fonseca, Gonçalves, Ocarino, and Mancini (2009). In their study, they tested the hypothesis that
within the MCS system there is a passive co-tension and only the net momenta can be equal to zero. They measured tension in a chain
of synergic muscles without affecting directly the tension of the antagonist of one of the chain’s members. More specifically, the
116
Fig. 5. Schematic representation of a tensegrity system (an icosahedron) and its mechanical behavior. A) An icosahedron without any external force imposed upon it,
B) An external force (black arrow) is applied upon one of the rods of the icosahedron increasing tension in the cable. In response, the icosahedron changes its shape re-
distributing the tension in its cables, and C) A characteristic external force-overall tension in the icosahedron curve. Of importance, stress never reaches zero, even
when no external force is applied as it is the case pictured in A. Tension response to increasing level of external force (as represented in B) is nonlinear and has a “J”
shape. Based on Levin, M. F. (2006). Tensegrity: The new biomechanics. In M. Hutson & R. Ellis (Eds.), Textbook of musculoskeletal medicine (pp. 69–80). Oxford,
England: Oxford University Press
procedure consisted of performing passive dorsiflexion and plantar flexion of the ankle at four different knee angles. Souza et al.’s
model of two prestressed ideal nonlinear springs predicted that increasing the stiffness of one spring would increase the stiffness of
the other. They observed an increase in both plantar and dorsiflexion stiffness as the knee assumed more extended positions, sup-
porting the model. From the tensegrity perspective on the MCS system, tonus is equated to prestress and the fast tension redistribution
that it allows accounts for the preparedness characteristics of the level of tonus (Turvey & Fonseca, 2014).
Preparedness is not only a reactive aspect of the movement apparatus, it also relates to the notion of tuning, i.e. anticipatory
adjustments that predisposes a system to behave in a way (Easton, 1972) and gives active support to movements (Belen’kii, Gurfinkel,
& Paltev, 1967). Tuning demonstrates the influence of other levels of the hierarchy on the Level of Tonus through informational
constraints. Furthermore, the phenomenon of tuning outlines another important aspect of the Level of Tonus, the interaction of the
mechanical tensegrity system with the nervous system (for one of the few exceptions, refer to Silva, Fonseca, & Turvey, 2010). The
concentration of spindles along tension pathways (Kokkorogiannis, 2004) led Turvey and Fonseca (2014) to hypothesize that the
nervous system responds to tension distribution within the biotensegrity system. More studies, however, need to be conducted to fully
support this hypothesis.
3.1.3.2. Multiscale tensegrity systems and fractals. Tensegrity-like systems have been observed not only at the macro-anatomical level
of the movement system as a whole but also at its microbiological level. Chen and Ingber (1999) advanced that microtubules are the
compressed elements, while actin and intermediate filaments constitute the net of tensile elements. This suggests that the
architectural organization of biotensegrity systems is analogous to the nested anatomical organization of connective tissues. It is
not unreasonable to assume, therefore, that biotensegrity systems can serve as models of the functional tension distributions
exhibited by nature’s wide variety of movement systems (see Scarr, 2014; Turvey & Fonseca, 2014). A popular suggestion is that the
icosahedron (a well-studied tensegrity architecture) works as the building block for a nested tensegrity system (e.g., Levin, 2006). The
symmetry of the individual icosahedron is such that icosahedra can be combined to form a nested tensegrity system (Fig. 6) with
scale-invariant properties (Levin, 2006). That is, an assemblage of icosahedra behaves as a single icosahedron. Thus, as required,
biotensegrity properties, including prestress, are preserved across scales.
Scale invariance is the hallmark of the fractal formalism (Feder, 1988). This suggests the possibility of combining the conceptions
of biotensegrity and fractals into a single theoretical model. Because of the variability observed across levels of biological systems,
heterogeneity is the rule, not the exception, and the biotensegrity hypothesis requires the expansion of the fractal formalism, from the
usual monofractals (i.e., systems characterized by fluctuations that follow a single power-law scaling) to multifractals (i.e., a system
with fluctuations that follow a variety of power-laws) (Feder, 1988; Palatinus, Kelty-Stephen, Kinsella-Shaw, Carello, & Turvey, 2014;
Turvey & Fonseca, 2014). The multifractal formalism emphasizes interaction across levels of analysis (Kelty-Stephen, Palatinus,
Saltzman, & Dixon, 2013).
In sum, the foregoing suggests that tonus should be considered a mechanical property of the MCS system functionally equivalent
to prestress in tensegrity systems. Due to the nested organization and scale invariance of the biotensegrity architecture tonus can be
equated with the prestress at the level of the MCS system. Although tonus is characterized at a given scale, it is the manifest form of
influences at multiple spatiotemporal scales that constitute the whole nested biotensegrity organization of the movement systems that
characterize the phylum Craniata.
3.2. Level of synergies
In his hierarchy of the construction of movements, Bernstein (1996) ascribed to the level of synergies the expertise to compress
the dimensionality of the movement apparatus. He viewed synergies as large choirs of muscles, yielding coherence and harmony of
movements. Over the years the understanding of synergies has changed, leading to the advent of different research programs with
distinct concepts and methods.
117
Fig. 6. An assembly of icosahedra generating a nested tensegrity system that preserves the properties of a single icosahedron. Letters “C” and “T” refer to compressed
and tensile elements, respectively. Reproduced with permission from Palatinus, Z., Kelty-Stephen, D. G., Kinsella-Shaw, J., Carello, C., & Turvey, M. T. (2014). Haptic
perceptual intent in quiet standing affects multifractal scaling of postural fluctuations. Journal of Experimental Psychology: Human Perception and Performance, 40(5),
1808.
3.2.1. Modularization
One of the perspectives on synergies assumes that redundant degrees of freedom of the movement apparatus is a computational
problem and synergies are the biological solution to reduce computational demands. The core of this perspective is that the spinal
cord contains clusters of neurons constituting modules that constrain activation of muscles into fixed patterns or muscle synergies
(d’Avella & Bizzi, 1998; Saltiel, Wyler-Duda, D'Avella, Tresch, & Bizzi, 2001). Variations on the level of stimulation of a module affect
the level of activation of muscles belonging to the same synergy. The computational advantage provided by this organizational style
is that control relies on selecting the right synergies and adequate scaling factors. Following Easton (1972); ∗∗see also Turvey, 2007),
muscle synergies are building blocks (or bases in the mathematical sense) that when linearly combined give origin to complex
movements (Bizzi, Cheung, d'Avella, Saltiel, & Tresch, 2008).
A methodology for decomposing a complex movement into constituent synergies involves four basic steps (Tresch & Jarc, 2009).
First, electromyographic (EMG) signals of a large number of muscles contributing to a complex movement are recorded and reduced
(e.g., the integral of the signal of each muscle is computed). Second, a dimension-reduction computational analysis (usually Non-
negative Matrix Factorization, NMF5) is applied to extract basic patterns from the signal, grouping muscles into synergies, and
identifying their associated scaling factors. Third, the basic patterns are combined to reconstruct the original EMG signal following an
optimization criterion (e.g., minimal least-square residuals) with the expectation that relatively few muscle synergies suffice to
explain the results. Fourth, the identified synergies are related with task-relevant variables, accounting for sub-goals of a complex
movement. For instance, Routson, Kautz, and Neptune (2014) applied NMF to investigate differences in locomotion between normal
individuals and post-stroke patients. They found that the EMG activity of normal individuals constituted four sequentially ordered
synergies, each one associated with a specific phase of the gait cycle. Adaption of steady-state walking to different constraints
happened by modifying the temporal relation among these four synergies. In contrast, post-stroke patients demonstrated poor gait
patterns that were associated with a smaller number of synergies. More specifically, synergies merged leading to a less functional
relation between synergy activation and phases of the gait cycle.
Two classes of muscle synergies have been identified, so-called synchronous synergies and varying synergies. Synchronous sy-
nergies refer to a group of muscles functioning as a single unity with constant and simultaneous co-activation (Ajiboye & Weir, 2009;
Tresch, Saltiel, & Bizzi, 1999), whereas varying synergies refer to groups of muscles that preserve a temporal shifted relation of their
activation (d’Avella, Saltiel, & Bizzi, 2003). The foregoing investigations have also shown that whereas some muscular synergies are
task-specific, others are not (e.g., d’Avella & Bizzi, 2005). The latter, to the contrary, are deployed in many different movements
conducted to solve a variety of different tasks.
It is important to note that although muscle synergies are organized in a task-dependent way, each synergy is a fixed and task-
5
For comparisons among different dimension reduction analyses see Tresch, Cheung, and d'Avella (2006) and Ting and Chvatal (2011).
118
Fig. 7. Three main components of a synergy according to Latash (2008).
independent pattern of muscular organization. That is, muscle synergies themselves are task-independent, but adaptability to task-
specific demands is a result of how a set of synergies is assembled.
3.2.2. Synergies and the uncontrolled manifold hypothesis

A different perspective on synergy is provided by M. Latash and colleagues (Latash, 2008, 2010; Latash, Scholz, & Schöner, 2007),
who considered not only the assembling aspect of synergies but also their intrinsic adaptive property. Latash (2008) outlined three
aspects of a synergy: (1) a group of elements that work together to achieve a common goal, i.e. they share a common purpose, (2) the
same group of elements presents flexible interactions to account for ongoing environmental changes and stabilize performance, and (3)
the organization of these elements is task-dependent, i.e. the same elements can be assembled in different ways to account for different
tasks. A schematic representation of this definition is presented in Fig. 7.
Synergies are embedded in a hierarchical architecture of control, in which upper levels of the hierarchy set boundary conditions,
constraining the degrees of freedom of lower levels facilitating the formation of synergies (Latash, 2010). Of importance, lower levels
always present more degrees of freedom than upper levels and, consequently, any sort of solution to a motor problem can emerge
given the many different ways their elements can interact under the same imposed constraints. The potential to bring about new
solutions is, therefore, maximum when the potential of (horizontal) interactions is also maximum [consonant with the principle of
maximum interaction (Latash, 2008)]. This implies that extra degrees of freedom (those that are not directly involved in achieving a
synergy’s function) are not a source of computational problems, but a resource for adaptation as held by the principle of abundance
(see Gelfand & Latash, 1998, 2002).
The fact that upper levels do not prescribe solutions for lower levels, accords the lower levels a certain degree of autonomy. That
is, synergies can perform their functions absent direct influence from higher levels. This means that, to the degree that the possible
interactions among elements of a synergy are sufficient to stabilize a performance variable, interactions with upper levels will be kept
minimal in accord with the principle of minimal (vertical) interaction (Latash, 2008).
A consequence of the scheme presented above is that variability in the relations among the elements that constitute a synergy (i.e.,
the elemental variables) expresses not only the deleterious effects of uncorrelated white noise but also adaptive response variations of
the movement system. The challenge to movement scientists turns out to be how to tease apart these two types of variability. Most of
this research program has been carried out under the umbrella of the Uncontrolled Manifold (UCM) hypothesis (Scholz & Schöner,
1999). The UCM hypothesis states that the variance of the elemental variables can be partitioned into two orthogonal components.
One component is “good” for performance (aligned with the null space of the uncontrolled manifold), in the sense that it does not
disturb the value of a particular performance variable. The other component is “bad” for performance (orthogonal to the null space of
the Uncontrolled Manifold) in the sense that it captures those configurations of elemental variables that push the performance
variable away from the desired value.
A fundamental assumption in the research program carried out by the proponents of the UCM hypothesis is that as large the
“good” variability is relative to the “bad” variability, as stronger the synergistic control is. This assumption has been used to test
different hypotheses of control. For instance, Domkin, Lacsko, Djupsjöbaka, Jaric, and Latash (2005) investigated synergy control in a
bimanual pointing task. They tested whether the task is controlled by a synergy that involves both arms (i.e., bimanual synergy) or by
a superimposition of two individual arm synergies (i.e., unimanual synergy). They demonstrated that the ratio of “good” variability to
“bad” variability was larger for the bimanual synergy and concluded that control in this task relies on the bimanual synergy.
One of the challenges to deploying the UCM hypothesis as a means of coming to terms with synergies is the identification of the
right elemental variables—those that correspond to the description of the system of interest at the right level of analysis. Latash et al.
(2007) established three basic criteria to identify and select the elemental variables. First, they need to affect performance variables.
Second, one of the elemental variables can (in principle) be controlled without affecting the others. And third, their effects are task-
specific. In the case of the bimanual pointing task described above, joint angles were chosen as the elemental variables.
A further aspect is that some elemental variables are composed variables or modes (Latash, Danion, Scholz, Zatsiorsky, & Schöner,
2003). That is, they are a combination of smaller elements that behave in a very specific manner every time the same mode is
selected6. In some studies, matrix factorization analyses have been applied to identify modes before applying UCM analysis. Danna-
6
A mode is like a muscular synergy presented in Section 3.2.1: when a mode is activated, all its elements are brought to work together. In contrast, a mode is not
strictly neural (see the text for more details).
119
dos-Santos, Slomka, Zatsiorsky, and Latash (2007) applied principal component analysis onto EMG data of lower limb and torso
muscles to identify modes of control during a task involving self-regulated postural sway. The activation of each mode was related to
consequences at the center of pressure displacement (i.e., the performance variable). Then, they applied UCM analysis on the modes
obtained from PCA to characterize the synergy underlying postural sway control. Modes, it should be noted, are not synergies in M.
Latash’s sense. That is the case because to M. Latash (and different from the modularization approach presented in Section 3.2.1),
synergies involve not only a set of elemental variables sharing responsibilities but also a flexible interaction among those elemental
variables oriented to stabilize a performance variable. That is why Danna-dos-Santos et al. (2007) applied the UCM analysis after
identifying modes through PCA.
3.2.3. Dynamical synergies

So far, we presented two perspectives on synergies, namely: synergies as modules (i.e., hard-wired structures) and synergies
within the UCM framework (i.e., the focus is on the consequences of using synergies to control a movement). Both perspectives have
proved to be very fruitful inquiries but they do not fully resonate with the ecological-dynamical perspective advocated in this essay.
From an ecological-dynamical point of view, modules are hard-constraints that do play a role on the emergence of behavior but they
are not synergy per se. In turn, UCM approach has not explored the possibility of the same elemental variables presenting different
qualitative/topological organizational state, which could, in principle, lead to the same quantitative synergistic control.
In what follows, we introduce the ecological-dynamical approach advanced in this essay. Our concerns are how to express
synergies in terms of emergent structures through a formalism that relates synergy and its elements and guarantees parsimony in the
interaction between levels of the hierarchy of movements, assuring commensurability. Inherent structure-function relation, as well as
context-dependence of synergies, are also considered.
3.2.3.1. Relational quantities and self-organization. Another perspective on synergies is that they are relational quantities or variables
that express invariances among their constitutive elements, usually described by nonlinear dynamical equations (Kelso & Fuchs,
2016). In a series of experimental and theoretical studies, it was shown that shoulder, elbow, and wrist can be considered as identical
oscillators presenting different invariant relations corresponding to subgroups of a larger dihedral symmetry group7 (Buchanan,
Kelso, & de Guzman, 1997; de Guzman, Kelso, & Buchanan, 1997). Participants were asked to track continuously arcs displayed on a
computer screen with different degrees of curvature at a pace dictated by a metronome. Participant could not move his or her trunk
and only movements in the sagittal plane were allowed. Results showed that different invariant relations among shoulder, elbow, and
wrist spontaneously emerged as a function of the level of curvature by virtue of self-organizing processes. In other words, a new
qualitative relational state (i.e., synergy) spontaneously emerges according to the value of a relevant parameter (e.g., curvature) of
the task.
Results of the preceding kind suggest that a synergy is a temporary functional grouping of structural elements softly assembled for
a particular purpose involving whatever elements are available and befitting the task (Kugler & Turvey, 1987). Important to the
current discussion is that as a soft-assembly emerging from self-organizing processes, a synergy expresses a form of control without a
controller (Turvey, 2009), since, by definition, self-organizing processes do not have an organizing agent.
The dynamical perspective on synergies expands the concept of synergy beyond the level of synergies itself. Synergies are re-
lational quantities whose constitutive elements are also relational quantities (Turvey, 1990). That is, synergies are composed by other
synergies, which implies that there is no absolute definition of what constitutes a synergy and its elements (Kelso, 2009)8. For
example, an invariant relation among joints of an upper limb as described by the dihedral symmetry group (see above) can be
considered a synergy and the individual muscles that cross each joint can be considered its elements. On the other hand, when this
synergy is embedded in a bimanual coordination task, it becomes an element of a larger synergy expressing a lawful-invariant
relation between the limbs as described by Haken, Kelso, and Bunz (1985).
3.2.3.2. Cooperation among contextual roles. An aspect of the individual synergy, complementary to the notion of quantitative
relations, is the synergy’s disposition to preserve its outcome despite the fact that the states of its components change qualitatively.
For example, a hallmark of synergies is immediate adjustments of linked elements (whether proximal or distal to each other, whether
synchronous or not) so as to preserve functional integrity when a perturbation is introduced. Bernstein (1967, p. 69) illustrated this
hallmark of synergies as follows:
“Movements react to changes in one single detail with changes in a whole series of others which are sometimes very far removed
from the former both in space and in time, and leave untouched such elements as are closely adjacent to the first detail, almost
merged with it. In this way movements are not chains of details but structures that are differentiated into details; they are structurally
whole, simultaneously exhibiting a high degree of differentiation of elements and differing in the particular forms of the relationships
between these elements”.
Kelso, Tuller, Vatikiotis-Bateson, and Fowler (1984) carried out a series of experiments that demonstrated remote immediate
adjustment. The experiments consisted basically of introducing a mechanical perturbation to a participant’s jaw when uttering “bab”
or “baz”. They recorded EMG signals of muscles of the tongue and kinematic features of the upper lip, lower lip, and jaw. Results
7
Roughly, a dihedral symmetry group is a group of rotations and reflections in the plane that preserves a regular polygon that they are applied on (Golubitsky &
Stewart, 2000).
8
A different notion of nested synergies is presented in Latash (2010).
120
Fig. 8. A minimal example of synergy as impredicativity. At the top and bottom are represented synergy and informational coupling functions. Each one of these
functions is described by an ordered pair in braces: the first member is the input of the function, while the second member is its output. Arrows are drawn from each set
to its members. Since synergy and informational coupling point to one another, a closed loop is created, indicating that the system is autonomous. Reproduced with
permission from Turvey, M. T. (2009). Nature of motor control: Not strictly “motor”, not quite “control”. In: D. Sternad (Ed.) Progress in motor control: A multi-
disciplinary perspective (pp. 3–6). New York, NY: Springer Verlag.
demonstrated that depending on which utterance was being pronounced when the jaw was disturbed, adaptations could occur at
tongue, upper lip, or lower lip. More specifically, the lips responded to perturbations of “bab”, whereas the tongue responded to
perturbations of “baz”. It is seemingly the case that individual speech elements (continuing Bernstein’s choice of terms) adapt
according to the circumstances. Each applies its own interactive capacity within the embedding context to preserve the functional
integrity of the synergy (Kelso, 2009; Turvey, 2007). Apparently, the cooperation characterizing the elements of synergies is less
about the individual elements as such and more about their context-dependent functions (Turvey & Fonseca, 2009).
An implication of the foregoing is that the relation between a synergy and its elements is one of circular entailment. That is (a) the
role an element plays changes as the synergy constituted by the element changes, and (b) the function the synergy performs also
changes as the role of an element changes (Turvey, 2007). In short, one can only know the element(s) of a synergy and the function of
a synergy in relation to each other. Circular entailment implies that something contains itself in its own definition.
Henry Poincaré coined the term impredicative as a label for self-referenced mathematical objects (see Kline, 1980). For many years
impredicative definitions and logical forms were excluded from mathematical formalism. Such was the case because of Well-Founded
Set Theory, the principal foundation of modern mathematics, rejects circularity in the definition of its objects (Chemero & Turvey,
2007). An alternative to deal with impredicativities is Hyperset Theory or Non-Well-Founded Set Theory (Aczel, 1988; Barwise &
Moss, 1996). Hyperset Theory is a branch of Set Theory in mathematics that deals with sets containing themselves—that is, sets that
are self-referencing (Chemero & Turvey, 2006). On grounds of the foregoing, we can define a synergy provisionally as a temporary
assembly of a system of impredicative context-dependent relations that emerges by virtue of self-organizing processes to satisfy specific
functions or goals. A minimal example of synergy as impredicativity consonant with Bernstein’s original level of synergies is shown in
Fig. 8.
3.3. Level of space
In broad terms, the level of space refers to how members of the Phylum Craniata of the Kingdom Animalia perform movements with
respect to environmental challenges. It is the first level in Bernstein’s hierarchy that addresses dexterity, the ability to produce fast and
accurate responses to environmental demands. The level of space accounts for both perceiving the environment and acting in it.
Bernstein suggested that both processes would take place in the “space field”, a reliable representation of the environment that
integrates different sources of stimuli. The space field is both an objective perception of the relations between body and surroundings,
alias the “external space”, and an ability to use this external space. Thus, Bernstein assumed a pragmatic view of perception9; in
121
perceiving the environment one also perceives what to do with it.
3.3.1. Affordances: objectivity and usefulness of space

Bernstein’s pragmatic stance on the conception of space is that the meaning of a perceived object refers to what to do with it, i.e.
what kind of action it supports. In this regard, meaning is intrinsic to the act of perceiving and meaning is not separated from the
perceived object (Reed & Jones, 1982). This highlights the fundamental aspect of the relation between perception and action10; one
perceives to act and acts to perceive (Gibson, 1979). Consider the example provided by Dotov, de Wit, and Nie (2012) about walking
into a cluttered room. As one progresses, some objects become occluded as others come into sight. Each position that one takes in the
room provides a specific relational configuration between oneself and the objects that furnish the room; the states of affairs are
uniquely defined. In consequence, the indefinitely many actions that one can do in the cluttered room are also defined relationally. If
an object is not occluded and is at a reachable distance from the current position occupied by an individual, then all the individual
needs to do is to move one’s hand directly to the object. However, if the object is still reachable, but occluded by another object, then
one’s hand has to take a detour to grasp it. What the latter prosaic example shows is that the possibilities for action provided by a
locale in the environment are relative to the individual occupying that locale.
The mutual relationship between individual and environment implies that the unit of analysis of purposive movements is the
whole system—the individual-environment system. The challenge is to define this system such that its wholeness is preserved but the
distinctions between its two units are recognized. This can be expressed as “A situation or event X affords action Y for animal Z on
occasion O if certain relevant mutual compatibility relations between X and Z obtain” (Shaw, Turvey, & Mace, 1982, p. 196). This
definition highlights the fact that the action Y (grasping) materializes only if the situation X (current environmental conditions) and
individual Z are compatible on dimensions relevant to Y. Formally, this can be defined as: the object X has a property p that affords Y.
This possibility for action only exists because there is an individual Z that possesses a complementary property q (an effectivity, e.g.,
capability to grasp). When Xp and Zq are juxtaposed they form a system Wpq, which exhibits the behavior Y (Turvey, 1992).
The definition of the affordance-effectivity duality presented above accounts for both needs advanced by Bernstein, namely, the
objectivity of what is perceived and its support for action. Affordances are real and actions are performed in respect to them.
Furthermore, the concepts of affordance and effectivity allow for treating an additional issue raised by Bernstein: the control of an
action is oriented to a desired future state in the relationship between individual and environment (Feigenberg & Latash, 1996). Note
that perceiving a thing’s affordances or a situation’s affordances means perceiving what can be done with the thing, what can be done
in the situation. In other words, affordance points to possible states of affairs that can be achieved if one engages in a given activity
(Turvey, 1992). Affordances ground prospective control (see Reed, 1996).
3.3.2. Information in the sense of specification

To actualize an affordance through an effectivity one needs to perceive the affordance, which from the ecological-dynamical point
of view is done by virtue of detecting information. Information in ecological terms relies on (a) an invariant relation between one or
more patterned energy distributions (e.g., optical, acoustical, mechanical, etc.) and the circumstances of the environment and/or
organism that generated them, and (b) the ability of a perceiving-acting system to detect them. In mainstream ecological psychology,
perception is the ability to detect specifying information11. In that sense, information has a dual characteristic; it points to the
environment as well as to the individual. Information is about affordances and information is for the use of the perceiver (Michaels &
Carello, 1981).
The scientific investigation of the study of perception-action has relied on identifying variables that specify an action. For ex-
ample, Fink, Foo, and Warren (2009) demonstrated (by manipulating the trajectories of fly balls in a virtual environment) that the
outfielder problem of how a player gets to the right place at the right time to catch the ball is solved by relying on a strategy of optical
acceleration canceling (Fig. 9). The strategy of so moving as to cancel the ball’s optical acceleration exemplifies how it can be the case
that the player’s change is defined intrinsically by the player’s relationship to the environment (for a review see Michaels & Zaal,
2002). For the limiting case: “If the vertical optical acceleration increases, move forward; if it decreases, move backward.” In can-
celing out the vertical optical acceleration, the outfielder is in the proper relation to the environment for catching the ball. In short,
the strategy of canceling optical acceleration realizes prospective control by relating the outfielder’s behavior to the specifying
variable.
3.3.3. Affordance-based control

Strategies of control such as canceling optical acceleration can be generally expressed as12
α̇ = −bτ (τcurrent
̇ ̇ )
−τideal (1)
9
In its original formulation, the space field is a sort of (objective) representation oriented to the use of one’s surrounding environment. The notions of objectivity and
usability are aligned with the pragmatic view underlying an ecological-dynamical approach to perception and action. In contrast, the notion of a representation as the
structural support to the observed functions is not. Once again, we embrace Bernstein’s functional perspective of construction of movements, but we disagree with his
understanding of which kind of structures/processes support these functions.
10
As mentioned before, the word action has two distinct meanings in this essay. When discussing the level of space, we use it as synonymous with goal-directed
movement.
11
For alternative ecological perspectives on information see Chemero (2009) and Vaz (2015).
12
Eq. (1) was derived from the discussion presented by Harrison, Turvey, and Frank (2016), sections Fixed Points and Nullclines and Modeling Affordance-Based
Control.
122
Fig. 9. Pictorial representation of the Optic Acceleration Cancellation (OAC) strategy. Optical acceleration increases when ball is about to fall ahead of individual eyes
location (Left), maintains constant when the ball is about to fall exactly where eyes are (Middle), and decreases when ball is about to land behind one (Right).
where α̇ is the rate of change in an action variable, bτ is a strength parameter, and τ̇current is the rate of change in some perceptible
variable. When τcurrent
̇ ̇ , then α̇ = 0 and further rate change in α is not necessary. Eq. (1) is an expression of the fundamental
= τideal
notion that animals and other organisms guide a given behavior by nullifying the difference between the current value of a perceptual
variable and its ideal value. In dynamical systems theory, the latter value is the system’s attractor location13. Because this approach is
based on the referent value of the specifying information variable, Fajen (2007) suggests use of the term “information-based control”.
For Fajen (2007) (see also Fajen & Matthis, 2011 and Fajen, 2013), information-based control as a theoretical perspective answers
the question “What will happen if the current conditions persist?” For example, in the outfielder problem, if the current value of
optical acceleration is canceled the ball will be caught. However, as underscored by Fajen (2005), the preceding formulation is not
fully ecological: perception-action is grounded in affordances. The laws of control generalized in Eq. (1) do not inform about the
boundary conditions that determine whether a given goal-directed act is, or is not, afforded the outfielder by the current circum-
stances. Following Fajen (2007, p. 390) the affordance for flyball catching by the outfielder is of the form “It is (not) still possible to
reach the landing location on time within the limits of [my] running capabilities”. Due to the emphasis ascribed to affordance in his
model, Fajen (2007) termed this approach as “affordance-based control”.
Fig. 10 illustrates Fajen's perspective of the reliance of control on affordance in the context of driving simulations. From an
information-based control point of view, the ideal deceleration to avoid collision is specified by tau-margin. Fig. 10 contrasts a tau-
margin specified deceleration profile (dashed line) with a participant’s actual deceleration profile (continuous line). Clearly, the
predicted and the actual trajectories are quite distinct, indicating that a strategy other than tau-margin was used by the participant.
Fig. 10 also illustrates a situation that given actual deceleration one cannot stop before the collision. More importantly, the
impossibility of avoiding collision is realized about one second before collision (indicated by the arrow). Such will be the case because
given the car’s limited deceleration capacity the driver is not able to reduce velocity to zero before covering the distance up to a
collide-with-able object. Fajen's example shows that maximum deceleration (a capacity of the system) embedded in a given context
(current velocity and distance from the collide-with-able object) defines a boundary that separates “avoidable collision” from “un-
avoidable collision”. Oudejans, Michaels, Bakker, and Dolné (1996) obtained results compatible with Fajen’s perspective in the
context of the outfielder problem, although they did not cast their discussion in the terms of affordance-based control.
A critical aspect of affordance-based control is that, in order to perceive affordances, one has to be perceptually attuned to one’s
own limitations. Through manipulations of the gain of brake pedals, Fajen (2005) demonstrated that individuals are sensitive to
different gain values and adjust deceleration accordingly. Fajen attributed this rapid adaptation to a process of calibration interpreted
as the expression of information in units of the driver’s capabilities. In this regard, affordance-based control is commensurate with the
contention that the utility of affordances is with respect to the continuous control of action (Stoffregen, 2000).
3.3.4. Reconciling information-based control and affordance-based control

The limitations of Eq. (1) with respect to explaining the influences of boundary conditions on control disposed Fajen (2007) to
reject it as representative of affordance-based control. Taking a different tack, Harrison and colleagues (Harrison et al., 2016) have
proposed an extension of affordance-based control aimed at reconciling the information-based and affordance-based approaches.
13
See Warren (2006) for a more extensive discussion on this subject.
123
Fig. 10. Top: Data showing actual deceleration (continuous line) and ideal value of acceleration at each instant (dashed line) in Fajen’s breaking studies. The border
between white and gray areas indicates value of maximum deceleration. Arrow indicates the instant when collision avoidance is not possible given speed at which one
is moving, braking gain, and distance to the road sign. Bottom: Actual tau value (continuous line) and ideal tau value (dashed line). Reproduced with permission from
Fajen, B. R. (2007). Affordance-based control of visually guided action. Ecological Psychology, 19(4), 383–410.
They proposed a model aligned with information-based control but sensitive to boundary conditions. Specifically, they suggested that
the attractor strength is inversely proportional to the distance from a boundary condition as shown in Eq. (2)
τ̇ −τ ̇
α̇ = −bτ ⎛ current ideal ⎞
⎜ ⎟
τ ̇ − ̇
⎝ ideal τmax ⎠ (2)
where the denominator is the scaling factor of the attractor strength b . The numerator is the attractor location as in Eq. (1). The
denominator shows that the closer one is to a boundary, the stronger is the attractor, which constrains the number of ways the given
task can be accomplished. The fact that the attractor strength is smaller the further from the attractor one is located opens the
possibility for other factors (e.g., individual preferences) to play a role in defining the actual solution exhibited by the system. Those
additional factors are soft constraints and their influence is shown in Eq. (3),
τ̇ −τ ̇
α̇ = −bτ ⎛ current ideal ⎞ + bβ (βcurrent−βpreferred )
⎜ ⎟
̇ −τmax
⎝ τideal ̇ ⎠ (3)
where bβ is the strength of the soft constraint β (understood as any factor that does not define boundary conditions but does affect
behavior, e.g. speed of locomotion) and βpreferred is the individual preferred value of β , that is, the individual bias. Important to the
present discussion is that when the agent is away from a boundary condition, the hard constraint is weak and the soft constraint
dominates. With approach to the boundary, the hard constraint becomes more prevalent (for details see Harrison et al., 2016).
Different sets of solutions can emerge from the competition between soft and hard constraints.
Despite the fact that Harrison et al. demonstrated a mathematical connection between the information-based and affordance-
based approaches, they did not attend explicitly to the issue of how informational variables are used. The model is still in devel-
opment. Its current formulation is sufficient, however, to evaluate how information constrains movements that accord with affor-
dances—that is, accord with what Fajen (2007) claims should be the case. Of additional significance is the fact that the Harrison et al.
model is also flexible enough to account for learning along the lines suggested by Direct Learning Theory (Jacobs & Michaels, 2007),
which provides insights into how information constrains action degrees of freedom (Michaels, Gomes, & Benda, 2016).
It should be noted that information (consequently, affordances) does not prescribe the form a movement takes, this is the business
of the Level of Synergies (see Section 3.2). The Level of Space, in turn, sets constraints upon the Level of Synergies that reflect the
agent’s intention. In doing so, it guides the agent to keep track of the task’s goal. In other words, the Level of Space deals with what is
performed instead of how it is performed.
3.4. Level of action
The level of action refers to the organizing of a sequence of movements (an action in Bernstein’s terms) with respect to an
environmental goal. That is, it deals with action problems that need more than one movement to be solved. Each movement that
composes the sequence can be performed individually, but when embedded in a sequence, it is changed in some of its features.
124
Fig. 11. Schematic representation of the Rosenbaum task. Basic layout configuration (top) and relation thumb orientation at grasping and final orientation of the bar.
3.4.1. The fundamental problem at the level of action

Bernstein emphasized that the overall organization of an action is such that the final movement of the chain of events realizes the
goal of the action. Some of the constituent movements, if considered apart from the whole organization that they compose, would
seem dysfunctional. When analyzed in consideration of the whole action, however, they are transparently functional. Bernstein’s
point is that the meaning of component movements unfolds only in the context of the whole action.
Consider the experimental setup presented in Fig. 11, which provides the basic features of what can be called the Rosenbaum task.
It is a much-studied task for a very good reason. Fig. 11 shows the initial condition. A bar lies horizontally in a cradle. On each side
there is an empty platform. The task is to reach out to the bar, grasp it, and place it vertically on one of the empty platforms according
to the experimenter’s instructions—placing the bar vertically on one of the platforms with the assigned color, black or white, at the
bottom. Fig. 11 also shows the experimental results of Rosenbaum and Jorgensen (1992). When the bar sat vertically with the white
half at the bottom, participants grasped the bar with the thumb pointing towards the black half (overhand). However, when the bar
sat vertically with the black half at the bottom, participants grasped the bar with the thumb pointing towards the white half (un-
derhand).
Of importance, in achieving the goal of the task, the act of grasping is constrained by the acts of transporting and unloading. That
is, the second part of the action (what happens after grasping) constrains the first part (from movement onset to grasping). Similar
results have been obtained with grasping a circular knob and rotating it at a specific angle (Herbort & Butz, 2010). Notice that in this
case there is no strong constraint on how to grasp the knob as there is with the bar on the cradle. Even though such is the case,
grasping (the first act) is still affected by rotation (the second act).
3.4.2. A contemporary perspective on the level of action

As highlighted in the preceding subsection, the primary goal of scientists interested in understanding the level of action is to
explain how a successor movement affects quantitatively and qualitatively features of a predecessor movement. This has been
equated with the problem of planning14 and has been studied extensively under the umbrella of representational/computational
models. For instance, Rosenbaum and colleagues suggested that the solution to the Rosenbaum task is essentially a computational
process that involves planning a sequence of postures (Rosenbaum, Meulenbroek, Vaughan, & Jansen, 2001). Resolving this se-
quential movement problem is essentially a matter of pre-selecting the right postures stored in memory. In a two-stage process, a
promising posture is identified in memory and a possible better posture is generated from that. This process is repeated until a
deadline imposed by temporal task constraints is reached and the movement is performed. Criteria of optimization such as comfort
are applied to select among candidate postures. The model is able to reproduce most kinematic features of manual movements
including more complex scenarios such as avoidance of objects.
Concerns regarding the adoption of a computational perspective on motor control have been expressed elsewhere. They involve
the problem of origin (Bickhard, 2015), the problem of abduction or inferring causes from consequences (Fodor, 2000; Turvey, 2015;
14
In the computational approaches of motor control, planning, anticipation, and prediction, although not exactly synonymous, are terms used to address problems
involving prospectivity.
125
Turvey, Shaw, Reed, & Mace, 1981), and the physical infeasibility that often accompanies assumed processes (Feldman, 2015),
among others. The reader is encouraged to consult these references for a complete treatment of philosophical/theoretical issues with
representational/computational approaches in general.
It is important to underscore, however, that despite all the possible inconsistencies of the aforementioned approaches, an inquiry
into the Level of Action has been almost exclusively conducted within the representational/computational framework. Most of our
limited understanding of this level was developed in experimental designs motivated by representational/computational thinking.
Despite its limitations, what progress has been made in respect to the Level of Action is due primarily to this research program.
3.4.3. Dynamical approaches to sequence

One factor that contributes to the almost exclusive treatment of the Level of Action in terms of representational/computational
perspectives is the attributed dependence of action on memory. This attribution is anchored in the notion that perception does not
inform about the particulars of the next movement, only about the present ongoing movement. And further, there is the historical
assumption of guessing about the future by using past experiences stored in memory (Helmholtz, 1868/1977). This assumption is
often apparent in dynamical approaches, even though the dynamical perspective is frequently promoted as contrary to the re-
presentational/computational approach.
Sandamirskaya and Schoner (2010) report a robotic implementation of a model based on dynamical field theory that produces
sequences of movements. Their model consists of a stack of ordinal neuronal activation fields obtained after teaching the robot a
sequence of stimuli to be recognized. The layers of the stack are related such that the peak in an ordinal layer can persist without
continuous stimulation. A peak is a stable solution. When it emerges, it inhibits activation in the preceding layer and plays a role in
the excitation of the succeeding layer. Each ordinal field is coupled to an output field, which remains as a stable state of the system
until a condition-of-satisfaction is achieved. When that happens, a peak in the successor ordinal layer emerges inhibiting the current
peak and, consequently, the output layer that it is associated with. The ordinal layer that exhibits the new stable solution is associated
with a different output layer that generates a different behavior. Thus, from a dynamical field theory point of view, order is a result of
inhibition/excitation processes among layers. It should be noted, however, that the sequence of layers is pre-defined and the serial
activation of the individual layers maintains relative independence with the exception of short transitory periods. Patently, despite its
sophistication, this is not the kind of solution expected to solve the Rosenbaum task.15 The question becomes whether this is the case
because (a) the Rosenbaum task can only be tackled by representational/computational strategies or (b) the Rosenbaum task requires a
different class of interpretation. We assume the latter. Accordingly, we would like to provide a way of casting the problem within the
ecological-dynamical perspective. What follows is a broad sketch of how the latter might be achieved.
3.4.4. A philosophical/theoretical argument

A question of some significance is whether it is possible, in principle, to ground problems at Bernstein’s Level of Action in the
ecological-dynamical perspective. Although our main concern in this essay has been to rethink Bernstein's levels of movement
construction in terms of the ecological-dynamical perspective, the perspective’s limited inquiry into the Level of Action demands from
us an indication of how the rethinking might proceed. Although the endeavor is limited to a theoretical exercise, we would like to
stress that it is based on a literature of relevance that has cumulated over the years without being brought to bear directly on the
issue. Our argument is based on two fairly recent ideas that have been developed independently in the scope of the ecological-
dynamical perspective; both of them address problems involving prospectivity: strong anticipation and the four modes of in-
tentionality. We begin with strong anticipation.
Strong anticipation or anticipatory coupling (e.g., Stepp, 2009; Stepp & Turvey, 2010, 2015, 2017; Voss, 2000, 2001, 2016) refers
to the possibility of systems coordinating with the future states of other systems in a lawful way16. The paradigm of strong antici-
pation entails formalizing in mathematical terms unidirectional coupling between (at least) two systems, a driver system x and a
driven system y where the latter is affected by the former but not vice versa. In this canonical formulation, the current state of the
“driver” is due to its own dynamics and the current state of the “driven” is a result of the current state of the “driver” and its own past.
This is expressed by,
x ̇ = f (x ) (4)
y = g (y ) + k (x −yτ )
where f and g are the intrinsic dynamics of systems x and y, respectively, k is a coupling strength, and yτ is y(t–τ). It is important to
note that the driven system anticipates the driver system due to the delay term τ. Hence, the delay is not a factor to be nullified;
rather, it is a key factor that allows for anticipation (Stepp & Turvey, 2010, 2015).
Results of studies with simulations (Dubois, 2003), physical systems (Sivaprakasam, Shahverdiev, Spencer, & Shore, 2001), and
humans (Stepp, 2009; Stepp & Turvey, 2017), have shown that, within a certain range of delays between the driver system and the
15
Despite the robustness and applicability of Sandamirskaya and Schöner’s model (and other dynamical models that do not rely explicitly on memory to produce
sequences of behaviors, e.g., Frank, 2014; Saltzman & Caplan, 2015) it remains the case that they do not handle the problem at the level of action described here when
applied to an organism getting about in its environment.
16
The term strong anticipation was coined by Daniel M. Dubois (see Dubois, 2003) to represent systems that can build their present behavior from past, present, and
possible future states anticipated from these systems themselves. The term contrasts with weak anticipation, which refers to systems that build their present states
considering past, present, and predicted future states obtained from models of the systems themselves.
126
Fig. 12. Four modes of awareness suggested by Shaw (2001). Reproduced from Shaw, R. (2001). Processes, acts, and experiences: Three stances on the problem of
intentionality. Ecological Psychology, 13(4), 275–314.
driven system, the driven system anticipates the states of the driver, even when the dynamics of the driver is chaotic. Even signals
that are not generated by a deterministic dynamical system can be reasonably predicted in real time by anticipatory coupling of a
simple relaxation system (Voss, 2016). For so-called anticipatory relaxation dynamics, no knowledge about the model that generated
the signal to be predicted is required. The latter fact greatly expands the domain of strong anticipation. It holds promise for both a
general framework capable of explaining the biological phenomenon of anticipation across all levels of analysis (Stepp & Turvey,
2010, 2015, 2017) and a mathematical formalism with the physical coherence necessary to the scientific formulation of direct
perception. Stepp and Turvey (2015) proposed a generalization of strong anticipation formalism. Of importance, they conjectured
that relations between the driver and driven system can present many different forms, including the possibility of the driven system
be coupled to many upcoming states of the driver system. Despite the body of theoretical and empirical work17 done on strong
anticipation, at its current stage, strong anticipation does not provide an explanation about how information constrains/regulates
action. That is why we now turn to the notion of intentionality.
Intentionality has four fundamental modes of reference to which we now turn (Shaw, 2001). They comprise two pairs:
(1) exteroperception (“about or of surroundings”) and exproprioperception (“about or of surroundings relative to self”).
(2) proprioperception (“about or of self”) and proexteroperception (“about or of self relative to surroundings”).
Fig. 12 summarizes the four modes and their coordination. Exteroperception and exproprioperception constitute a coordinate pair
such that the act of attention that yields perception of the environmental layout also yields the perception of the environmental
layout relative to the perceiver. Likewise, proprioperception and proexteroperception constitute a coordinate pair such that the act of
attention that yields perception of the body’s layout (the current posture) also yields the perception of the body’s layout relative to the
environmental layout. Of importance, Shaw (2001) outlines the symmetry between proexteroperception and exproprioperception:
the same energy array carries information about self relative to surroundings and about surroundings relative to self. In sum, in-
tentionality (qua being aware of) happens in the act of perceiving and perceiving is always with respect to both the environment and
the self (Gibson, 1979; Mace, 2005).
It is important to notice that perception is defined as detection of specifying information, which, in turn, reveals affordances.
Affordances are inherently dealing with the future (see Turvey, 1992). For instance, before catching a ball one has to perceive that the
ball is catch-able. Consequently, attuning to conditions that allow actualization of affordances implies attuning to future states of
affairs (i.e., prospective control).
3.4.5. Rosenbaum problem redux

We can now return to the question of whether the conjunction of strong anticipation and intentionality’s four modes helps
reconcile the problem at hand (to reiterate, how in a simple two-component action, the second component affects the first). From
strong anticipation, we can assume that the environmental layout is the driver system (i.e., it possesses relative independent dy-
namics and affects another system), while the perceiver-actor is the driven system (i.e., the system that needs to anticipate future
states of another system). However, in this case, the driven system can exert an influence on the driver system as well. That is, in this
case, the current state of the driver system is a result of its instantaneous relation with the driven system. Washburn, Kallen, Coey,
Shockley, and Richardson (2015) provided evidence for this bidirectional coupling in a context of strong anticipation involving two
agents. Assuming Stepp and Turvey’s (2015) assertion that there is breadth in perception pointing to the future (perception’s pro-
spective nature), a perceiver-actor can know about the different possibilities of environmental layout configurations. By detecting
17
As it was pointed out by one of the reviewers, studies on strong anticipation in perception-action systems has been conducted primarily on continuous tasks,
while Rosenbaum task is a discrete and sequential task. Although conceptually, strong anticipation is not limited to continuous task, results of these studies have to be
considered carefully.
127
Fig. 13. Schematic representation of the nested-temporal structure of the ambient array proposed to Rosenbaum task redux. After instructions of a researcher, a
subject perceives the necessary relation between the bar laying on the cradle and the possible state of the bar sitting on the platform. Furthermore, one perceives the
transformation necessary to actualize the latter (intended) state. This is a matter of exteroperception of a future state of environment layout. Subject also perceives
that, in order to apply the transformation required to achieve researcher’s instructions, the bar on the cradle relates in a specific way with his or her right hand
(exproprioperception). Due to symmetry in the ambient array, it is also perceived that subject’s right hand also relates in a specific way with the bar laying on the
cradle (proexteroperception). In addition, parts of the body are organized to allow the right supporting postures (proprioperception).
information that specifies the relation between objects and self, one is made aware of which kind of relation can be established with
the objects, including those relations that affect the environmental configuration in a way as perceived/intended before (Figure 13).
In the context of Rosenbaum task, the argument would require the following. First, based on the instructions of the experimenter
the participant perceives the possibility of a given environmental layout (a matter of exteroperception) in a specific way (e.g., set the
bar on the platform with the white half pointing towards the ceiling). Second, given an intention to change the environmental layout,
the participant perceives what kind of relations between hand and bar (a matter of a symmetrical relation between exproprio- and
proextero-perception) can be taken to actualize the desired configuration (e.g., grasping underhand). Finally, a set of postures needs
to be sustained in order to enact the necessary relations between hand and bar. Fig. 13 presents a schematic representation of the
argument.
In this scheme, the continuity of the ambient optic array and the prospective nature of perception are essential assumptions that
would allow the performance of a sequence of movements to be accomplished with the task specified in the act of perceiving. Of
course, the challenge is how to design an experiment that shows that the ambient array containing nested information arrangement
about possible future states of affairs and that this information is detectable, and that once this information is detected it constrains
how states of affairs evolve. Recently this aspect of the ecology of organisms begun to be experimentally explored in terms of a
hierarchy of affordances (e.g., Wagman, Caputo, & Stoffregen, 2016).
An implication of the foregoing is that there is no clear distinction between Bernstein’s levels of space and action. Both are
grounded in information about body and environment in the sense of specificity to body and environment. The level of action outlines
Gibson’s (1979) observation that an organism’s environment is a nested organization, and since the information available in the
ambient array (be it optical, acoustical, mechanical, chemical or conjunctions thereof) is specific to its environmental source, it is also
a nested organization. Although some classes of actions do not fall into the description provided above, most of the tasks of interest to
students of biological movement systems do fall into that description. Thus, from the ecological-dynamical point of view, the levels of
space and action do not, in the most relevant cases, distinguish as sharply as Bernstein’s exposition of the two levels implies.
4. Re-reading Bernstein’s levels of construction of movements
Bernstein remains one of the most thoughtful and influential thinkers in the study of how coordinated movements of the human
body (and the bodies of the phylum Craniata in general) are achieved. In this hierarchical model of the construction of movements,
Bernstein hypothesized a division of labor in which each of four levels (tonus, synergy, space, and action) solves a specific class of
problems. In this essay, we have reviewed the main aspects of Bernstein's hypothesis and provided our own understanding of each
level from an ecological-dynamics point of view. In this, the final section, we provide a broader picture of our understanding of the
construction of movements. We will adopt a Socratic strategy; given the incompleteness of the answers, we have given to the
questions raised.
4.1. What are the levels of movement construction?
Bernstein (1996) was careful in his use of the word levels since it applies to the neuro-anatomical perspective that grounds his
theory. For Bernstein, a level is a neuro-anatomical structure uniquely suited to a specific class of movement problems. We have
emphasized the functional aspect of movements without referring to the anatomy that supports them. As mentioned above, we accept
Bernstein’s functional ideas, but we do not relate functions to any readily circumscribed anatomical structures. The levels of con-
struction of movement are manifest as resources available to produce a movement: tonus is a material basis for movement, space and
action set informational constraints, and synergies are abstract relations connecting them.
128
4.2. What are synergies?
For general purposes an apposite definition of synergy is that provided by Webster’s 3rd New International Dictionary: “Working
together; combined action or cooperation”. Synergy, however, is not necessarily the most apposite label for the functional entity in
question. Concinnity is possibly a better label. Webster defines it thusly “Harmony or fitness in the adaptation of parts to a whole or to
each other”. Concinnity’s definition is more closely allied with the language of self-organizing systems. Further candidate terms with
a similar alliance are heterarchy (McCulloch, 1945) and coalition (Turvey, Shaw, & Mace, 1978). A particularly intuitive candidate
term is coordinative structure (Easton, 1972; see Turvey, 1977): A group of muscles so constrained as to act as a single functional
unit. For simplicity, and in concert with Bernstein (1996), we have adhered to the term synergy in the present manuscript thus far and
will continue to do so in the remainder of the manuscript. We can, however, take a step beyond the dictionary definition and the
commonplace understanding. The step in question follows from recognizing that muscles are more than motors (Dickinson et al., 2000)
and a synergy is a group of devices so constrained as to act as a single functional unit. Muscles function as motors, rulers, springs, struts,
tuners, and brakes and synergies (at the behavioral level) are task-specific organizations of them.
Patently, synergies play a critical role in Bernstein’s understanding of the construction of movements. They are key to the solution
of the degrees of freedom problem (Bernstein, 1967; Kugler & Turvey, 1987; Latash, 2008; Turvey, 1990, 2007)—a functional
challenge manifest at each level of the nested movement apparatus. Any given synergy is composed of other synergies, making the
concepts of macro and micro relative (Kelso, 2000). A synergy at any one level of analysis comprises multiple elements and, in turn, is
itself an element, but at a higher level. This latter notion is important because it allows us to consider the degrees of freedom problem
at multiple scales of analysis and encourages the provision of an answer that has a common principle of operation everywhere,
assuring coherence for the system. In taking this stance, we are assuming that there is no level of synergies per se. Synergies are
abstractions at many levels of analysis, spontaneously emerging due to informational and material constraints. At the behavioral
level, such constraints are imposed by the levels of tonus, space, and action. In turn, synergies constrain these levels as well.
4.3. What is the relation between Synergies and Tonus?
It has been proposed that synergies are created taking advantage of the resources of the level of tonus (Bernstein, 1996). In the
foregoing, we have equated the architecture of the level of tonus with a multi-fractal biotensegrity system, characterized by prestress
at all levels (Turvey & Fonseca, 2014). Prestress guarantees the unity of the system and fast adaptation to mechanical perturbations
via rapid tension re-distribution. The latter mechanism is hypothesized to operate at any scale, from cells in the small (e.g., Chen &
Ingber, 1999) to the muscle-connective tissue-skeletal system in the large (e.g., Souza et al., 2009). In principle, this feature of the
level of tonus provides the necessary conditions for the nesting of self-organizing processes that yield synergies.
It is in the foregoing sense that a synergy is a softly assembled (perforce temporary) biotensegrity system, with an unknown and
likely complex geometry. The geometry of a softly assembled tensegrity system is task-specific, which suggests that paths of tension
re-distribution after perturbation are also task-specific even if the elements that compose the synergy are the same. We speculate that
the latter aspect is why perturbation of the jaw in two different utterances exhibits two different kinds of compensation (Kelso et al.,
1984).18
4.4. What is the relation between the Synergy and Space?
Information is included in the constraints that govern the emergence and enactment of synergies. The organizational form of a
goal-directed movement is affected by information although not prescribed by it. The way information affects movement depends on
the interplay between hard and soft-constraints (Harrison et al., 2016). A hard constraint is one essential to the success of a given
activity and hence inviolable. A soft constraint is simply a non-essential preference formative of the resulting activity. It is integral to
neither success nor failure.
The basic strategy that information-based control relies on is nullifying the difference between the current value of a detected
variable and its ideal value. For the outfielder with respect to a fly ball, the rule is to move so as to cancel the ball’s vertical optical
acceleration. In so doing the outfielder scales his or her speed of locomotion without affecting the overall movement pattern requisite
to catching the ball (Fink et al., 2009). In addition, environmental information can induce synergy re-organization. Bruggeman, Zosh,
and Warren (2007) asked their participants to walk in a virtual environment with virtual-heading direction displaced from the actual
walking. They found that participants produced a kind of “crab walking”, that is, during locomotion, their legs crossed the midline of
the body. This procedure, of which the participants were unaware, differs from regular walking. It led to stabilization of the optic
flow and consequent regulation of direction of displacement.
The level of space constrains synergies, but the reverse is also true: synergies constrain information pickup. Actually, this outlines
one of the fundamental assumptions of the ecological-dynamics perspective on perception-action capabilities: perception constrains
action and action constrains perception (Gibson, 1979). Evidence for this assumption is found in the structure of postural exploratory
movements at the mm scale (Palatinus et al., 2014). The experiment in question requires a little more detail than that given in the
preceding summaries.
It is well established that one can perceive the lengths of hand-held objects without vision, simply by wielding them. The same is
18
This hypothesis receives additional support from Scarr and Harrison (2017).
129
true of objects attached to the torso (for summaries of both handheld and torso-attached see Carello & Turvey, 2015). The individual
participant maintained a quiet standing posture on a force platform with a rod attached at its center point to the participant’s
shoulders (specifically, at the 1st thoracic vertebra). Rod length varied across trials. Depending on the trial the participant had to
perceive the rod’s whole length or the rod’s partial length, left or right of the mid-shoulder point of the rod’s attachment.19
On any given trial of the experiment by Palatinus and colleagues, a rod was attached to the back of the torso by means of a hollow
plastic tube affixed at the level of the junction between the first thoracic and seventh cervical vertebrae. (Rods were 72, 96 or 120 cm
in length.) Blinders limited the field of view to the straight-ahead. Perceived length, whole and part, differed as actual lengths, whole
and part. Of special interest to Palatinus and colleagues was whether the variability of the center of pressure (COP) at the mm scale
differed between the two conditions. Multifractal Detrended Fluctuation Analysis (Ihlen, 2012; Kelty-Stephen et al., 2013) revealed
that it did: the spectral range of COP fluctuations when participants intended to perceive partial length was larger than when
participants intended to perceive the whole length.
The results of Palatinus et al. (2014) are particularly important to the present discussion because they make evident a further
aspect that is underemphasized in Bernstein’s original formulation. The perception of the full or partial length of a rod attached to the
body is a matter of exteroperception (Shaw, 2001; Turvey & Fonseca, 2014), a mode of awareness attributed commonly to vision and
hearing. However, Palatinus et al. showed that exteroperception is available through the flesh. That is, the tension distribution in the
biotensegrity structure that supports the level of tonus is also informative about the level of space. This observation has been made
over the years within the domain of dynamic or effortful touch (Carello & Turvey, 2004, 2015), which includes a study of a person
suffering from peripheral neuropathy (specifically, lacking discriminative touch) (Carello, Kinsella-Shaw, Amazeen, & Turvey, 2006).
In that sense, the level of space resembles the notion of perceptual systems (see Gibson, 1966).
4.5. What is the relation between Synergies and Action?
In the discourse on speech production, it has been suggested that “an account of a speech that addresses issues of sequencing but
not issues of concinnity would be far from complete” (Turvey, 2007, p. 659). The latter points to the need for a common framework to
explain two of the most important problems in action-perception (alias motor control): the serial order problem (Lashley, 1951) and
the degrees of freedom problem (Bernstein, 1967). Expressed in terms of Bernstein’s hierarchy, the latter is the problem of how the
level of action and the level of synergy interact. Above, we addressed this problem taking advantage of a well-known task, referred to
here as the Rosenbaum task. This task, it will be recalled, refers to how a preceding movement is performed to make a succeeding
movement available. Said differently it is a task that refers to how the currently occurring is constrained by the subsequently
occurring. From an ecological-dynamical point of view, it means how information constrains action, what makes the problem, at least
in principle, similar to that of the level of space. More specifically, one should look for how the nesting of information sequentially
constrains synergies. Such an inquiry is pursuable by combining the conceptual tools of modes of awareness proposed by Shaw (2001)
and the experimental paradigm and technical tools developed within the strong anticipation framework (Stepp, 2009, 2010, 2015,
2017).
5. Conclusion
In the previous sections, we revisited Bernstein’s hierarchy of construction of movements and we presented theoretical and
experimental support to the ecological-dynamical perspective advanced in this essay, including how the different levels of the
hierarchy interact. In what follows, we outline three basic questions that characterize a basic agenda to those abiding the ecological-
dynamical perspective presented above. 1) How does lateral force transmission at the macroanatomical level of analysis contributes to
synergy formation? 2) How does specifying information constrain synergy formation? 3) How do possible future states of individual
surroundings affect sequential transitions between synergies? Strategically, these three basic questions express problems that cannot
be revealed addressing each level separately. Finally, we emphasize the how in these three questions because as it was stressed in
Section 1, an ecological-dynamical perspective strives for identifying lawful regularities, i.e. regular ways of how variables relate.
Acknowledgements
Vitor L. S. Profeta was supported by Coordination of Improvement of Higher Education Personnel (CAPES, Brazil) [Grant Number
991/12-0, 2012].
Michael T. Turvey was supported in part by National Science Foundation [Grant BCS-1344725 (INSPIRE Track 1)].
The authors thank Gabriela Baranowski Pinto, Paula Lanna Silva, and Steven Harrison for important discussions and comments on
early versions of this paper and Claudia Carello for her help with the figures. We also appreciated the comments and criticisms of
anonymous reviewers.
19
Prior research (Palatinus, Carello, & Turvey, 2011) had revealed the feasibility of conditions that allowed whole-body wielding of, and perception of, shoulder-
supported rods.
130
References
Aczel, P. (1988). Non-well-founded sets. Stanford: CSLI.

Ajiboye, A. B., & Weir, R. F. (2009). Muscle synergies as a predictive framework for the EMG patterns of new hand postures. Journal of Neural Engineering, 6(3), 036004.
Barwise, K. J., & Moss, L. (1996). Vicious circles. On the mathematics of non-wellfounded phenomena. Stanford: CSLI.
Belen’kii, V. Y., Gurfinkel, V. S., & Paltev, Y. I. (1967). Elements of control of voluntary movements. Biofizika, 12(1), 135–141.
Bernstein, N. A. (1967). The co-ordination and regulation of movements. London: Pergamon Press.
Bernstein, N. A. (1996). On the construction of movements. In M. L. Latash, & M. Turvey (Eds.). On dexterity and its development (pp. 3–244). Mahwah, NJ: Lawrence
Erlbaum.
Bickhard, M. H. (2015). Toward a model of functional brain processes I: Central nervous system functional micro-architecture. Axiomathes, 25(3), 217–238.
Bizzi, E., Cheung, V. C. K., d'Avella, A., Saltiel, P., & Tresch, M. (2008). Combining modules for movement. Brain Research Reviews, 57(1), 125–133.
Bruggeman, H., Zosh, W., & Warren, W. H. (2007). Optic flow drives human visuo-locomotor adaptation. Current Biology, 17(23), 2035–2040.
Buchanan, J. J., Kelso, J. S., & de Guzman, G. C. (1997). Self-organization of trajectory formation. Biological Cybernetics, 76(4), 257–273.
Carello, C., Kinsella-Shaw, J., Amazeen, E. L., & Turvey, M. T. (2006). Peripheral neuropathy and object length perception by effortful (dynamic) touch: A case study.
Neuroscience Letters, 405(3), 159–163.
Carello, C., & Turvey, M. T. (2004). Physics and psychology of the muscle sense. Current Directions in Psychological Science, 13(1), 25–28.
Carello, C., & Turvey, M. T. (2015). Dynamical (effortful) touch. Scholarpedia, 10(4), 8242.
Carvalhais, V. O. C., Ocarino, J. M., Araújo, V. L., Souza, T. R., Silva, P. L. P., & Fonseca, S. T. (2013). Myofascial force transmission between the latissimus dorsi and
gluteus maximus muscles: An in vivo experiment. Journal of Biomechanics, 46(5), 1003–1007.
Chemero, A. (2009). Radical embodied cognitive science. Cambridge, MA: The MIT Press.
Chemero, A., & Turvey, M. T. (2006). Complexity and “closure to efficient cause”. In ALIFE X: workshop on artificial autonomy. MIT Press, Cambridge, MA.
Chemero, A., & Turvey, M. T. (2007). Complexity, hypersets, and the ecological perspective on perception-action. Biological Theory, 2(1), 23–36.
Chen, C. S., & Ingber, D. E. (1999). Tensegrity and mechanoregulation: from skeleton to cytoskeleton. Osteoarthritis and Cartilage, 7(1), 81–94.
d’Avella, A., & Bizzi, E. (1998). Low dimensionality of supraspinally induced force fields. Proceedings of the National Academy of Sciences, 95(13), 7711–7714.
d’Avella, A., & Bizzi, E. (2005). Shared and specific muscle synergies in natural motor behaviors. Proceedings of the National Academy of Sciences of the United States of
America, 102(8), 3076–3081.
d’Avella, A., Saltiel, P., & Bizzi, E. (2003). Combinations of muscle synergies in the construction of a natural motor behavior. Nature Neuroscience, 6(3), 300–308.
Danna-dos-Santos, A., Slomka, K., Zatsiorsky, V. M., & Latash, M. L. (2007). Muscle modes and synergies during voluntary body sway. Experimental Brain Research,
179(4), 533–550.
de Guzman, G. C., Kelso, J. S., & Buchanan, J. J. (1997). Self-organization of trajectory formation. Biological Cybernetics, 76(4), 275–284.
Dickinson, M. H., Farley, C. T., Full, R. J., Koehl, M. A. R., Kram, R., & Lehman, S. (2000). How animals move: an integrative view. Science, 288(5463), 100–106.
Domkin, D., Lacsko, J., Djupsjöbaka, M., Jaric, S., & Latash, M. L. (2005). Joint angle variability in 3D manual pointing: uncontrolled manifold analysis. Experimental
Brain Research, 163, 44–57.
Dotov, D. G., de Wit, M., & Nie, L. (2012). Understanding affordances: history and contemporary development of Gibson’s central concept. AVANT. Pismo Awangardy
Filozoficzno-Naukowej, 2, 28–39.
Dubois, D. M. (2003). Mathematical foundations of discrete and functional systems with strong and weak anticipations. Lecture Notes in Computer Science, 2684,
110–132.
Easton, T. A. (1972). On the normal use of reflexes: The hypothesis that reflexes form the basic language of the motor program permits simple, flexible specifications of
voluntary movements and allows fruitful speculation. American Scientist, 60(5), 591–599.
Fajen, B. R. (2005). Calibration, information, and control strategies for braking to avoid a collision. Journal of Experimental Psychology: Human Perception and
Performance, 31(3), 480.
Fajen, B. R. (2007). Affordance-based control of visually guided action. Ecological Psychology, 19(4), 383–410.
Fajen, B. R. (2013). Guiding locomotion in complex, dynamic environments. Frontiers in Behavioral Neuroscience, 7.
Fajen, B. R., & Matthis, J. S. (2011). Direct perception of action-scaled affordances: The shrinking gap problem. Journal of Experimental Psychology: Human Perception
and Performance, 37(5), 1442.
Feder, J. (1988). Fractals. New York: Plenum Press.
Feigenberg, I. M., & Latash, L. P. (1996). NA Bernstein: The reformer of neuroscience. In M. L. Latash, & M. Turvey (Eds.). Dexterity and its development (pp. 247–275).
Mahwah, NJ: Lawrence Erlbaum.
Feldman, A. G. (2015). Referent control of action and perception. Challenging conventional theories in behavioral neuroscience. New York: Springer.
Fenn, W. O., & Garvey, P. H. (1934). The measurement of the elasticity and viscosity of skeletal muscle in normal and pathological cases; a study of so called “muscle
tonus”. Journal of Clinical Investigation, 13(3), 383.
Fink, P. W., Foo, P. S., & Warren, W. H. (2009). Catching fly balls in virtual reality: A critical test of the outfielder problem. Journal of Vision, 9(13) 14-14.
Fodor, J. (2000). The mind doesn’t work that way: The scope and limits of computational psychology. Cambridge, MA: MIT Press.
Frank, T. D. (2014). A nonlinear physics model based on extended synergetics for the flow of infant action during infant-mother face-to-face communication.
International Journal of Scientific World, 2(2), 62–74.
Gelfand, I., & Latash, M. (1998). On the problem of adequate language in movement science. Motor Control, 2, 303–312.
Gelfand, I., & Latash, M. (2002). On the problem of adequate language in biology. In M. Latash (Ed.). Progress in Motor Control II: Structure-function relations in voluntary
movements(pp. 209–228). Urbana, IL: Human Kinetics.
Gibson, J. (1966). The senses considered as perceptual systems. Boston: Houghton – Mifflin.
Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton – Mifflin.
Golubitsky, M., & Stewart, I. (2000). The symmetry perspective. From the equilibrium to chaos in phase space and physical space. Boston: Birkhäuser.
Haken, H., Kelso, J. A. S., & Bunz, H. (1985). A theoretical model of phase transition in human movements. Biological Cybernetics, 51, 347–356.
Harrison, H. S., Turvey, M. T., & Frank, T. D. (2016). Affordance-based perception-action dynamics: A model of visually guided braking. Psychological Review, 123(3),
305.
Helmholtz, von H. (1868/1977). On the origin and significance of the axioms of geometry. In R. S. Cohen and Y. Elkana (Eds.), Herman von Helmholtz:
Epistemological writings. Boston: Reidel.
Herbort, O., & Butz, M. V. (2010). Planning and control of hand orientation in grasping movements. Experimental Brain Research, 202(4), 867–878.
Huijing, P. A. (1999). Muscle as a collagen fiber reinforced composite: A review of force transmission in muscle and whole limb. Journal of Biomechanics, 32(4),
329–345.
Huijing, P. A. (2007). Epimuscular myofascial force transmission between antagonistic and synergistic muscles can explain movement limitation in spastic paresis.
Journal of Electromyography and Kinesiology, 17(6), 708–724.
Ihlen, E. A. (2012). Introduction to multifractal detrended analysis in Matlab. Frontiers in Physiology, 3, 1–18.
Jacobs, D. M., & Michaels, C. F. (2007). Direct learning. Ecological Psychology, 19(4), 321–349.
Kelso, J. S. (1995). Dynamic patterns. The self-organization of brain and behavior. Cambridge, MA: MIT press.
Kelso, J. S. (2000). Principles of dynamic pattern formation and change for a science of human behavioral. In R. B. C. Bergman, & L. N. Nilsson (Eds.). Developmental
science and the holistic approach (pp. 63–83). UK: Taylor & Francis.
Kelso, J. S., & Fuchs, A. (2016). The coordination dynamics of mobile conjugate reinforcement. Biological Cybernetics, 110(1), 41–53.
Kelso, J. S. (2009). Synergies: atoms of brain and behavior. In D. Sternad (Ed.). Progress in motor control: A multidisciplinary perspective (pp. 83–91). New York, NY:
Springer Verlag.
131
Kelso, J. S., Tuller, B., Vatikiotis-Bateson, E., & Fowler, C. A. (1984). Functionally specific articulatory cooperation following jaw perturbations during speech:
Evidence for coordinative structures. Journal of Experimental Psychology: Human Perception and Performance, 10(6), 812.
Kelty-Stephen, D. G., Palatinus, K., Saltzman, E., & Dixon, J. A. (2013). A tutorial on multifractality, cascades, and interactivity for empirical time series in ecological
science. Ecological Psychology, 25(1), 1–62.
Kline, M. (1980). Mathematics: The loss of certainty. New York, NY: Oxford University Press.
Kokkorogiannis, T. (2004). Somatic and intramuscular distribution of muscle spindles and their relation to muscular angiotypes. Journal of Theoretical Biology, 229,
263–280.
Kugler, P. N., Kelso, J. S., & Turvey, M. T. (1980). On the concept of coordinative structures as dissipative structures: I. Theoretical lines of convergence. In G. E.
Stelmach & J. Requin (Eds.), Tutorials in motor behavior. Amsterdam: North Holland.
Kugler, P. N., & Turvey, M. T. (1987). Information, natural law, and the self-assembly of rhythmic movement. Hillsdale, NJ: L. Erlbaum Associates.
Lashley, K. S. (1951). The problem of serial order in behavior. In L. A. Jeffress (Ed.). Cerebral mechanisms in behavior (pp. 112–136). New York: Wiley.
Latash, M. (2008). Synergy. New York, NY: Oxford University Press.
Latash, M. (2010). Motor synergies and the equilibrium point hypothesis. Motor Control, 14(3), 294–322.
Latash, M. L., Danion, F., Scholz, J. F., Zatsiorsky, V. M., & Schöner, G. (2003). Approaches to analysis of hand writing as a task of coordinating a redundant motor
system. Human Movement Science, 22(2), 153–171.
Latash, M. L., Scholz, J. P., & Schöner, G. (2007). Toward a new theory of motor synergies. Motor Control, 11(3), 276–308.
Latash, M. L., & Turvey, M. (Eds.). (1996). On dexterity and its development. Mahwah, NJ: Lawrence Erlbaum.
Levin, M. F. (2006). Tensegrity: The new biomechanics. In M. Hutson, & R. Ellis (Eds.). Textbook of musculoskeletal medicine (pp. 69–80). Oxford, England: Oxford
University Press.
Maas, H., & Sandercock, T. G. (2010). Force transmission between synergistic skeletal muscles through connective tissue linkages. Journal of Biomedicine and
Biotechnology, pp. 1–9, doi: 10.1155/2010/575672.
Maas, H., & Huijing, P. A. (2012). Mechanical effect of rat flexion carpi ulnaris after tendon transfer: Does it generate a wrist extension moment? Journal of Applied
Physiology, 112, 607–614.
Maas, H., Meijer, H. J., & Huijing, P. A. (2005). Intermuscular interaction between synergists in rat originates from both intermuscular and extramuscular myofascial
force transmission. Cells Tissues Organs, 181(1), 38–50.
Mace, W. M. (2005). James J. Gibson's ecological approach: Perceiving what exists. Ethics & The Environment, 10(2), 195–216.
Margulis, L., & Schwartz, K. V. (1998). Five kingdoms. An illustrated guide to the phyla of life on earth. New York, NY: W. H. Freeman. (Original work published 1982).
McCulloch, W. S. (1945). A heterarchy of values determined by the topology of nervous nets. The Bulletin of Mathematical Biophysics, 7(2), 89–93.
Michaels, C. F., & Carello, C. (1981). Direct perception. Englewood Cliffs, NJ: Prentice-Hall1–208.
Michaels, C. F., Gomes, T. V., & Benda, R. N. (2016). A direct-learning approach to acquiring a bimanual tapping skill. Journal of Motor Behavior, 48(1–18), 2016.
Michaels, C. F., & Zaal, F. T. J. M. (2002). Catching fly balls. In K. Davids, G. J. P. Savelsbergh, S. J. Bennett, & J. van der Kamp (Eds.). Interceptive actions in sport:
Information and movement (pp. 172–183). London: Routledge.
Monti, R. J., Roy, R. R., Hodgson, J. A., & Edgerton, V. R. (1999). Transmission of forces within mammalian skeletal muscles. Journal of Biomechanics, 32(4), 371–380.
Oudejans, R. R., Michaels, C. F., Bakker, F. C., & Dolné, M. A. (1996). The relevance of action in perceiving affordances: Perception of catchableness of fly balls. Journal
of Experimental Psychology: Human Perception and Performance, 22(4), 879.
Palatinus, Z., Carello, C., & Turvey, M. T. (2011). Principles of part–whole selective perception by dynamic touch extend to the torso. Journal of Motor Behavior, 43(2),
87–93.
Palatinus, Z., Kelty-Stephen, D. G., Kinsella-Shaw, J., Carello, C., & Turvey, M. T. (2014). Haptic perceptual intent in quiet standing affects multifractal scaling of
postural fluctuations. Journal of Experimental Psychology: Human Perception and Performance, 40(5), 1808.
Patel, T. J., & Lieber, R. L. (1997). Force transmission in skeletal muscle: From actomyosin to external tendons. Exercise and Sport Sciences Reviews, 25(1), 321–364.
Reed, E. (1996). Encountering the world. New York: Oxford University Press.
Reed, E., & Jones, R. (1982). Selected essays of James J. Gibson. Hillsdale, NJ: LEA Publishers.
Riewald, S. A., & Delp, S. L. (1997). The action of the rectus femoris muscle following distal tendon transfer: Does it generate knee flexion moment? Developmental
Medicine & Child Neurology, 39(2), 99–105.
Rosenbaum, D. A., & Jorgensen, M. J. (1992). Planning macroscopic aspects of manual control. Human Movement Science, 11(1), 61–69.
Rosenbaum, D. A., Meulenbroek, R. J., Vaughan, J., & Jansen, C. (2001). Posture-based motion planning: Applications to grasping. Psychological Review, 108(4), 709.
Routson, R. L., Kautz, S. A., & Neptune, R. R. (2014). Modular organization across changing task demands in healthy and poststroke gait. Physiological Reports, 2(6),
e12055.
Saltiel, P., Wyler-Duda, K., D'Avella, A., Tresch, M. C., & Bizzi, E. (2001). Muscle synergies encoded within the spinal cord: evidence from focal intraspinal NMDA
iontophoresis in the frog. Journal of Neurophysiology, 85(2), 605–619.
Saltzman, E., & Caplan, D. (2015). A graph-dynamic perspective on coordinative structures, the role of affordance-effectivity relations in action selection, and the self-
organization of complex activities. Ecological Psychology, 27(4), 300–309.
Sandamirskaya, Y., & Schoner, G. (2010). Dynamic field theory of sequential action: A model and its implementation on an embodied agent. In Development and
Learning, 2010. ICDL 2010. 9th IEEE International Conference on (pp. 133–138).
Scarr, G. (2014). Biotensegrity: The structural basis of life. Handspring Publishing Limited.
Scarr, G., & Harrison, H. (2017). Examining the temporo-mandibular joint from a biotensegrity perspective: A change in thinking. Journal of Applied Biomedicine, 15,
55–62.
Scholz, J. P., & Schöner, G. (1999). The uncontrolled manifold concept: Identifying control variables for a functional task. Experimental Brain Research, 126(3),
289–306.
Shaw, R. (2001). Processes, acts, and experiences: Three stances on the problem of intentionality. Ecological Psychology, 13(4), 275–314.
Shaw, R., Turvey, M. T., & Mace, W. (1982). Ecological psychology: The consequence of a commitment to realism. Cognition and the Symbolic Processes, 2, 159–226.
Silva, P. L., Fonseca, S. T., & Turvey, M. T. (2010). Is tensegrity the functional architecture of the equilibrium point hypothesis? Motor Control, 14, e35–e40.
Sivaprakasam, S., Shahverdiev, E. M., Spencer, P. S., & Shore, K. A. (2001). Experimental demonstration of anticipating synchronization in chaotic semiconductor
lasers with optical feedback. Physical Review Letters, 87(15), 154101.
Souza, T. R., Fonseca, S. T., Gonçalves, G. G., Ocarino, J. M., & Mancini, M. C. (2009). Prestress revealed by passive co-tension at the ankle joint. Journal of
Biomechanics, 42(14), 2374–2380.
Stepp, N. (2009). Anticipation in feedback-delayed manual tracking of a chaotic oscillator. Experimental Brain Research, 198(4), 521–525.
Stepp, N., & Turvey, M. T. (2010). On strong anticipation. Cognitive Systems Research, 11(2), 148–164.
Stepp, N., & Turvey, M. T. (2015). The muddle of anticipation. Ecological Psychology, 27(2), 103–126.
Stepp, N., & Turvey, M. T. (2017). Anticipation in manual tracking with multiple delays. Journal of Experimental Psychology: Human Perception and Performance, 44(5),
914–925.
Stoffregen, T. A. (2000). Affordances and events. Ecological Psychology, 12(1), 1–28.
Street, S. F. (1983). Lateral transmission of tension in frog myofibers: A myofibrillar network and transverse cytoskeletal connections are possible transmitters. Journal
of Cellular Physiology, 114(3), 346–364.
Ting, L. H., & Chvatal, S. A. (2011). Decomposing muscle activities in motor tasks: methods and interpretations. In F. Danion, & M. L. Latash (Eds.). Motor control:
Theories, experiments, and applications (pp. 102–138). Oxford: Oxford University Press.
Tresch, M. C., Cheung, V. C., & d'Avella, A. (2006). Matrix factorization algorithms for the identification of muscle synergies: Evaluation on simulated and experi-
mental data sets. Journal of Neurophysiology, 95(4), 2199–2212.
Tresch, M. C., & Jarc, A. (2009). The case for and against muscle synergies. Current Opinion in Neurobiology, 19(6), 601–607.
132
Tresch, M. C., Saltiel, P., & Bizzi, E. (1999). The construction of movement by spinal cord. Nature Neuroscience, 2, 162–167.
Trotter, J. A., Richmond, F. J., & Purslow, P. P. (1995). Functional morphology and motor control of series-fibered muscles. Exercise and Sport Sciences Reviews, 23(1),
167–214.
Turvey, M. T. (1977). Preliminaries to a theory of action with reference to vision. Perceiving, Acting and Knowing, 211–265.
Turvey, M. T. (1990). Coordination. American Psychologist, 45(8), 938.
Turvey, M. T. (1992). Affordances and prospective control: An outline of the ontology. Ecological Psychology, 4(3), 173–187.
Turvey, M. T. (2007). Action and perception at the level of synergies. Human Movement Science, 26(4), 657–697.
Turvey, M. T. (2015). Quantum-like issues at nature's ecological scale (the Scale of Organisms and Their Environments). Mind and Matter, 13(1), 7–44.
Turvey, M. T., & Fonseca, S. T. (2014). The medium of haptic perception: A tensegrity hypothesis. Journal of Motor Behavior, 46(3), 143–187.
Turvey, M. T., & Fonseca, S. (2009). Nature of motor control: perspectives and issues. In D. Sternad (Ed.). Progress in motor control: A multidisciplinary perspective (pp.
93–123). New York, NY: Springer Verlag.
Turvey, M. T., Shaw, R. E., & Mace, W. (1978). Issues in the theory of action: Degrees of freedom, coordinative structures and coalitions. In J. Requin (Ed.). Attention
and performance VII (pp. 557–595). Hillsdale, NJ: Erlbaum.
Turvey, M. T., Shaw, R. E., Reed, E. S., & Mace, W. M. (1981). Ecological laws of perceiving and acting: In reply to Fodor and Pylyshyn (1981). Cognition, 9(3),
237–304.
Turvey, M. T. (2009). Nature of motor control: Not strictly “motor”, not quite “control”. In D. Sternad (Ed.). Progress in motor control: A multidisciplinary perspective (pp.
3–6). New York, NY: Springer Verlag.
Vaz, D. V. (2015). Direct perception requires an animal-dependent concept of specificity and of information. Ecological Psychology, 27(2), 144–174.
Vleeming, A., Pool-Goudzwaard, A. L., Stoeckart, R., van Wingerden, J. P., & Snijders, C. J. (1995). The posterior layer of the thoracolumbar fascia| its function in load
transfer from spine to legs. Spine, 20(7), 753–758.
Voss, H. U. (2000). Anticipating chaotic synchronization. Physical Review E, 61(5), 5115.
Voss (2001). Dynamical long-term anticipation of chaotic states. Physical Review Letter, 87, 0141021–0141024.
Voss, H. U. (2016). Adaptation and self-organizing systems. Physical Review E, 93, 030201.
Wagman, J. B., Caputo, S. E., & Stoffregen, T. A. (2016). Sensitivity to hierarchical relations among affordances in the assembly of asymmetric tools. Experimental Brain
Research, 234, 2922–2923.
Walsh, E. G. (1992). Muscles, masses and motion: The physiology of normality, hypotonicity, spasticity and rigidity. Cambridge University Press 20.
Warren, W. H. (2006). The dynamics of perception and action. Psychological Review, 113(2), 358.
Washburn, A., Kallen, R. W., Coey, C. A., Shockley, K., & Richardson, M. J. (2015). Harmony from chaos? Perceptual-motor delays enhance behavioral anticipation in
social interaction. Journal of Experimental Psychology: Human Perception and Performance, 41(4), 1166.
20
This reference is misidentified in Turvey and Fonseca (2014) as Walsh(1992). The measurement of muscle tone. Paraplegia, 30, 507-508.
133

Bernstein's Level of Movement Construction A Contemporary Perspective

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bernstein's Level of Movement Construction A Contemporary Perspective

Uploaded by

Copyright:

Available Formats

Human Movement Science 57 (2018) 111–133

Contents lists available at ScienceDirect

Human Movement Science

Bernstein’s levels of movement construction: A contemporary

2. Bernstein’s hierarchy of movement construction

able to anticipate or compensate for any non-mechanical environmental perturbation.

3. The hierarchy of movement construction from an ecological- dynamical perspective

3.1. Level of tonus

3.1.1. The orthodox view of tonus

3.1.2. Connective tissue’s contribution to tonus

transmission was observed in contralateral gluteus maximus (Fig. 4).

3.1.3. Tonus as a mechanical property

3.2. Level of synergies

Fig. 7. Three main components of a synergy according to Latash (2008).

3.2.2. Synergies and the uncontrolled manifold hypothesis

3.2.3. Dynamical synergies

3.3. Level of space

perceiving the environment one also perceives what to do with it.

3.3.1. Aﬀordances: objectivity and usefulness of space

3.3.2. Information in the sense of speciﬁcation

3.3.3. Aﬀordance-based control

3.3.4. Reconciling information-based control and aﬀordance-based control

3.4. Level of action

3.4.1. The fundamental problem at the level of action

3.4.2. A contemporary perspective on the level of action

3.4.3. Dynamical approaches to sequence

3.4.4. A philosophical/theoretical argument

3.4.5. Rosenbaum problem redux

4. Re-reading Bernstein’s levels of construction of movements

4.1. What are the levels of movement construction?

4.2. What are synergies?

4.3. What is the relation between Synergies and Tonus?

4.4. What is the relation between the Synergy and Space?

4.5. What is the relation between Synergies and Action?

Aczel, P. (1988). Non-well-founded sets. Stanford: CSLI.

You might also like