Yet, Bernstein's end goal was never coordination. A quick look at the titles of his book (On Dexterity and its Development), first essay (What is Dexterity?), and last essay (Dexterity and its Features) suggests that an account of dexterity was what he was after. So, do the humanoid robots demonstrate dexterity? What is dexterity anyway? And more importantly, what are the differences between dexterity and coordination? In this blog, I aim to provide a systematic approach to answering these questions, covering perspectives from the layman, Bernstein, and also some additional input from my end. Although our everyday definitions suggest that dexterity represents a subset of coordination, Bernstein regarded the two as distinct concepts. Coordination concerns the internal relations within a movement system; dexterity, by contrast, encompasses both these relations and the external motor problem the system seeks to solve. Unfortunately, I believe Bernstein's treatment of coordination is slightly simplistic. By building on his developed view on the two motor terms, I argue that coordination actually represents a subset of dexterity (which, given how we started, is somewhat of an unexpected reversal). For transparency, the reader should be informed that at the time of writing, I have read only five of Bernstein's seven essays. It is entirely possible that I have misrepresented Bernstein's views. Comments and discussion are always welcome!
Dictionary definitions of dexterity and coordination
Let's begin by outlining the nonscientific definitions of coordination and dexterity. According to the Oxford dictionary, coordination is 'the organisation of the different elements of a complex body or activity so as to enable them to work together effectively'. In a separate blog post, I argued that this showcases two features of coordination: 1) that coordination involves achieving global order among numerous more fundamental components, and 2) that coordination is functional (i.e., it serves a purpose). Meanwhile, dexterity is 'the ability to perform a difficult action quickly and skilfully with the hands'. In this light, one may be able to provide a synonymous definition of dexterity with coordination at its core, for example, 'dexterity is the ability to coordinate the elements in one's finger-hand-wrist system to perform a difficult action quickly and skilfully'. It is not difficult, then, to see why dexterity might be seen as an example of coordination. Put another way, dexterity is simply coordination applied to situations that require fine motor control, including activities involving the hands (e.g., piano playing, cardistry, surgery).
Bernstein's definitions of dexterity and coordination
Let's turn to Bernstein now. I've already addressed Bernstein's definition of coordination here, as well as the challenges to motor control that informed his definition. What about dexterity? Interestingly, Bernstein begins by outlining what dexterity isn't, arguing that it isn't simply good coordination or 'harmony in movements'. More concretely, Bernstein highlights that '[d]emand for dexterity is not in the movements themselves but in the surrounding conditions' and that '[t]here is no movement that would not place high demands on dexterity, given appropriate conditions'. Instead, dexterity is 'in finding a [quick and successful] motor solution [to a complex motor problem] for any situation and in any condition'. In contrast, coordination involves 'overcoming excessive degrees of freedom in our movement organs' or 'turning the movement organs into controllable systems'. The difference is subtle, but unmistakable. To Bernstein, coordination involves finding a way to restrain the countless directions of mobility of our skeletal-articular-muscular systems and the unpredictable forces that accompany every muscular impulse. This is an undeniably within-system approach that focuses on achieving large-scale order from the internal organisation of system components. On the other hand, dexterity extends this internal focus on movements outward, emphasising the ability to problem-solve in the face of unpredictable and novel situations.
Note the duality of the term 'problem-solving' -- it includes both the problem (i.e., the external motor task) and the solution (i.e., the coordinated movement). These dual aspects of problem-solving are inseparable when discussing dexterity -- there is a mutual relationship between the movement problem and the animal-movement system. For those well-versed in the tenets of ecological psychology, this should sound very familiar. Bernstein's problem-solution mutuality of dexterity strongly parallels the animal-environment mutuality advocated by James J. Gibson. For me, this is why these ideas are often mentioned in the same breath. Just as Gibson argued that one cannot talk about an animal isolated from their environmental context, it is difficult, if not impossible, to answer questions about movement and action without an inkling of the external motor problem to be solved.
Bringing the discussion back to the differences between dexterity and coordination, we can see clear ways in which semantics diverge between Bernstein's and the layman's perspective. While dexterity is limited to the hands and fine motor control in the latter view, Bernstein's dexterity applies similarly to gross motor control, so long as it involves motor flexibility and resourcefulness in the face of unpredictable motor problems. Yet, while Bernstein's view already provides much insight into how we should approach the topic of movement, I disagree with his (or more accurately, my interpretation of his) treatment of coordination. Here, I disagree that coordination only relates to the internal components of a movement system. Instead, I argue that, similar to dexterity, the notion of coordination inherently includes an external referent alongside the implemented movement!
Is coordination problem-solving?
There are a few clues that point to an unavoidable problem-solution mutuality in motor coordination. For one, recall the second insight gained from the dictionary definition of coordination -- that the ability to coordinate is functional and serves a goal. An action is only coordinated if it involves many moving parts AND effectively deals with a motor task. This leads us to the second clue -- Bernstein's own definition of coordination, in which he places great emphasis on degrees of freedom. Here, I think there is a tendency (and I am guilty of this too!) to mistakenly associate 'degrees of freedom' with 'more elementary components to control'. However, this does not do justice to the term's original use, at least in statistics. Recognised in the work of Carl Friedrich Gauss (who popularised the normal distribution, or bell curve), the number of degrees of freedom can be thought of as the number of values within a calculation that are free to vary, given that we already know the outcome of the calculation. For example, given an average score of 10 (i.e., the outcome of the calculation) and a sample of 5 scores, we calculate the degrees of freedom by taking the sample size minus 1, yielding 4 (free-varying values) in this example. More generally, the number of degrees of freedom is the number of parameters that can vary freely without violating specified constraints. In the case of motor control, the constraints refer to the motor problem itself. Hence, when we say that there are excessive degrees of freedom in the movement system, we are actually saying that there are infinitely many ways in which the skeletal-articular-muscular apparatus can coordinate such that a movement solution that addresses the constraints of the motor task emerges. Importantly, what this demonstrates is that the notion of degrees of freedom, at least when talking about movement systems, comes fully baked with both the problem and its many solutions (i.e., the problem-solution mutuality). It is then inaccurate to suggest that coordination focuses solely on the internal relations among parts of the movement system.
The final clue is more trivial, but interesting nonetheless. In various parts of his essays, Bernstein constantly describes coordination as harmonious, and not without good reason. Often, we perceive good coordination as harmonious, elegant, and beautiful, and it seems like such aesthetic judgements require no external motor problem to be passed. Indeed, if we are to see harmony as global order resulting from the consonant, organised internal relations between local components, then one might argue similarly that coordination requires nothing more than an appropriate internal organisation. But do these aesthetic concepts exist in isolation? Are beauty and elegance inherently possessed, or are they appreciated given a wider, ecological context? In music, consonant harmonies are often accentuated and shot into the limelight when played against the crunchy dissonance and tension of altered dominant 7th chords. In art and fashion, what is considered beautiful or stylish ebbs and flows with the currents of shorter-lived fads and longer-lasting trends. While not an external problem in and of itself, it can be argued that the aesthetic notions of harmony and beauty are relative rather than absolute, requiring an external referent to serve as a criterion for the internal organisation of coordination. In this respect, I am inclined to see Bernstein's tendency to describe coordination as harmonious as support for my claim that coordination, like dexterity, involves a problem-solution mutuality.
The dexterity-coordination difference
At this stage, I have established that both coordination and dexterity involve some form of problem-solving, where movement solutions are implemented in service of an external movement goal. If the difference between the motor terms lies not in the internal-focus versus external-focus distinction, where then does it lie? Let's go back to the humanoid robots introduced earlier. I maintain my position that each robot demonstrates high levels of coordination. Combining Bernstein's definition with a more explicit reference to problem-solving, the robots demonstrate the ability to overcome the excessive degrees of freedom in their movement apparatus and to control it adequately to solve certain movement problems. Yet a strong kick to their robotic guts would likely stop them from performing their routine, not least because they would struggle to get off the ground. Or suddenly request a change in style -- from the more staccato popping to the more flowing modern dance style -- and the robots will likely be unable to adapt on the fly. This, to me, is where the difference between coordination and dexterity lies. The robots show coordination but not dexterity because they are unable to quickly and efficiently solve problems in unpredictable, novel situations!
Of course, this insight is nothing groundbreaking. One only has to take a cursory glance through Bernstein's essays to realise that he indeed saw resourcefulness, adaptability, and flexibility in the face of novelty and unpredictability as the hallmark of dexterity. Where I hope to come in is by providing another angle in which one can situate the coordination and dexterity distinction. At the start of this blog, I claimed that coordination represented a subset of dexterity. What I meant by that is that while both coordination and dexterity involve problem-solving, the degree to which they do so varies drastically. For me, coordination emerges when a movement system demonstrates the capability to control its degrees of freedom to solve a single or, more broadly, a set of related movement problems. On the other hand, dexterity extends this and emerges when a movement system can continuously solve larger sets of unrelated, unexpected problems quickly and effectively. Put simply, coordination is in solving single, isolated problems; dexterity is revealed through solving a wider range of novel problems. In other words, the difference between coordination and dexterity is a matter of scale!
Concluding remarks
The prior representation of the difference between coordination and dexterity invites many questions. For example, when does a movement system begin to display dexterity rather than coordination? How quickly should the problem be solved, and how novel and unexpected should the problem be before we consider the system dexterous rather than coordinated? Is there a discrete line separating these domains, or does coordination and dexterity lie on a continuous spectrum? Is this even the right way to conceptualise their difference? Have I simply misrepresented Bernstein and mistakenly presented a strawman caricature of his work? Or might this have been what he was going after the entire time? If anything, I think my analysis of coordination as problem-solving strengthens the view that movements cannot be studied outside the context of the motor problem they are intended to solve. There is a problem-solution mutuality that movement scientists will have to account for when studying motor control. As always, it goes without saying that this blog entry is by no means conclusive and simply represents my thinking-out-loud process. I fully expect my views to shift as I engage in more literature from Bernstein and others in motor control.
References
Bernstein NA (1996) On dexterity and its development. In: Latash ML, Turvey MT (Eds.) Dexterity and Its Development, pp. 1–244, Erlbaum Publ.: Mahwah, NJ.
No comments:
Post a Comment