Jason Stanley and John Krakauer, "Motor skill depends on knowledge of facts," Frontiers in Human Neuroscience 7 (Article 503), 29 August 2013. doi: 10.3389/fnhum.2013.00503
MY TAKE-AWAY POINTS
The authors affirm that skill requires both acuity and knowledge. Via philosophical analysis, and via analysis of HM case, the authors undermine folk distinction between practical and theoretical activity. Paper emphasizes that despite common belief, cognitive neuroscience findings do not support that distinction.
Cognitive neuroscience distinguishes declarative and procedural knowledge. While declarative knowledge is verbally articulable knowledge, "procedural knowledge" is a misnomer; it is not really knowledge, but is better termed "motor acuity" (= decrease in variability and increase in smoothness of movements).
Many problems happen when equating these cognitive neuroscience terms with philosophical terms. It is very misleading to equate "declarative knowledge" with the philosophical term "propositional knowledge." Propositional knowledge is knowledge of facts, but that knowledge need not be verbally articulable. Furthermore, one should not equate "procedural knowledge" with "motor skill" for that leaves out the contribution of propositional knowledge to motor skill.
Propositional knowledge provides a scaffold for development of motor skills in three areas: 1) acuity of selected action components; 2) new actions (e.g., new techniques of running, jumping, etc); 3) ability to select the right action from a repertoire.
GOING THROUGH THE PAPER, SECTION BY SECTION
INTRODUCTION
Initial dialectic. Platonism about skill: you must be able to explain principles guiding action; otherwise it is just a habit. Against this extreme position, many now think skill does not involve knowledge of rules.
Cognitive neuroscience seems to mirror these philosophical assumptions in their concentration on experiments that show that active reflection on principles impedes performance and that retention of knowledge of facts not needed for skill (HM case).
There are some questionable assumptions at work here involving hasty or indeed wrong identification of neuroscience categories (declarative and procedural knowledge) and mental kinds (propositional knowledge; theoretical vs practical ability). Cognitive neuroscience should not equate procedural "knowledge" (a misnomer; should be "motor acuity" or decrease in error and increase in smoothness of movement) with "motor skill" tout court. And it should not oppose procedural knowledge to declarative knowledge (= verbal articulation) The problem is that "procedural knowledge" has been applied to non-knowledge aspects of skill (motor acuity). But it doesn't follow that skills do not require knowledge of facts. Thus we have made an attribute of a component of skill (that motor acuity is non-knowledge based) into an attribute of the whole of skill.
Note that intellectual ability may involve implicit (non-intentional) abilities.
MOTOR SKILL, PERCEPTUAL DISCRIMINATION AND INTENTION
There are two necessary conditions for belonging to functional category of "skill": 1) Practice-related improvement in goal-directed action and 2) knowing what to do to initiate action manifesting the skill. To have a skill one must know what to do to initiate the action that displays the skill; this is propositional knowledge. S and K do NOT say skills ONLY require know-what to do to initiate actions. But knowledge like that of initiating an action can be injected into course of activity. So they are claiming that skills require SOME knowledge.
PROPOSITIONAL vs DECLARATIVE KNOWLEDGE
For S and K, propositional knowledge includes the possession of know-what to do to initiate action displaying skill. However, this form of propositional knowledge does not require the ability to verbally articulate that knowledge (even though possession of other forms of propositional knowledge can and often are shown by verbalizing their contents). On the other hand, the cognitive neuroscience category of declarative knowledge does require ability to verbally articulate that knowledge. So the problem is the false identification of declarative knowledge and propositional knowledge.
THE LESSONS OF HM: MOTOR ACUITY, ACTION SELECTION AND KNOWLEDGE
The HM case shows multiple memory systems in brain. This is very important, but has been harmful in that it allowed false mapping of true cog NS distinction of declarative vs procedural "knowledge" onto bad folk distinctions of knowledge vs skill and theory vs practice.
HM cases show dissociation btw improved motor performance and ability to recall task aspect. However, HM required explicit instruction each day, so the idea that HM had an entirely procedural motor skill is undermined. If you really have a skill, you don't need instruction to start over every day. What is being learned each day is not skill but only a part of skill, motor acuity.
Displaying both propositional knowledge and motor skill involves both propositional knowledge and perceptual / motor acuity, so there is no good distinction between theoretical and practical activity.
CONCLUSIONS
For Dreyfus, skill moves you from guidance by knowledge-based decisions to perception of right action, so that expert performance is a species of perceptual acuity. Further, for Dreyfus, expertise is automatic or habitual and knowledge drops away. But for S and K, the expert uses knowledge of activity to dictate to non-knowledge components, so it is combination of knowledge and acuity that leads to skilled performance. Thus even if initial scaffolding for basic components drops away that doesn't mean you don't need new knowledge to continue refinement of skill and to add new non-knowledge components.