Training From Scratch (III) – Stability (1)

Stability training has been FUBAR-ed by ‘functional fitness’ and this post will level the ground so we can build a stable house later. (Around 5.700 words, estimated reading time: 27-30min)

Imagine that you are hanging out with a bunch of people who know something about exercise, and following what you read on the last post of this series, you drop the words "training" and "old adults" in the conversation. Of course, you're expecting everyone to pick up with sarcopenia, grip strength, or knee extension strength. To your dismay, however, everybody is soon talking about "center of mass," "base of support," "risk of fall", "balance training", and "stability." 
Then the shit really hits the fan when the audience splits into two groups: those who extoll the benefits of 'unconventional' training tools and those who advocate sticking to traditional resistance training. The former are the more vocal, and mention gear with colorful names like Battle Ropes and BOSU® or ethnic artifacts you've never heard of, like Swiss balls and Bulgarian bags [insert a dirty joke here].
You become suspicious. Something sinister is afoot. The way they are throwing fancy words and PubMed abstracts at one another reminds you of how chimps throw feces. Unfortunately, you suspect that, unlike the chimps, it may not be an indication that they are smart.
Suddenly, it dawns on you that they are talking about stability, but NOBODY EVEN MENTIONED BIOMECHANICS! The realization that you are surrounded by ignoramuses that almost convinced you that balance training matters drenches you in a cold sweat. You leave the room in horror. That was a close call...

As you’ll have guessed, today’s topic is stability. As the short horror story above suggests, I might as well have titled this post “the science and bullshit of stability”. And indeed, the science and bullshit of stability deserve their own miniseries, much like the science and bullshit of lifting. But I I can take care of it in two episodes, so decided to make it part of the Training From Scratch series. And honestly, I could have crammed everything in a single post. And it was initially my intention, but the draft ended up unusually long and technical with over 7.000 words (and still missing most of the material that ended up in this here post). Also, there were more asides than main text.

But there was a somewhat natural articulation in the argumentation, and so I decided to split the post in two. There will still be some substantial conclusions in this first post but the conclusions will be most destructive. Then again, that’s the purpose of this series (empty the bowl and all that Zen shit) and one more reason to embed the miniseries in the series.

Also, promise, in the second part, I’ll rebuild on what I crashed down. Cross my heart, and all that.

But for now, let’s get started with some serious Analytic Fitness™ methodology — and a bad pun because I can never resist one.

How to put stability back on its feet

The scientific basis for the study of stability is so straightforward that it’s surprisingly strange that almost everybody misses the point.

When I say ‘almost everybody’ I am lumping together not only the fitness-and-rehab people (personal trainers, physical therapists, and even some well-meaning M.D.) and the sport-and-performance people (sports coaches and competitive athletes) but also a good chunk of the exercises science researchers.

Fitness-and-rehab people are responsible for a mistake that has hatched the most pervasive stability-related bullshit to date (I’ll come to that) but even sport-and-performance people and some researchers have managed to miss a point or two on their own.

Even worse, the culprit is not a subtle issue resulting from the ignorance of possibly relevant confounding factors (like with grip strength, for instance) or considerations pertaining to nonexistent stuff (like the ‘law of irradiation’) although it eventually leads to something equivalent. Rather, the cause of the widespread confusion about stability is homonymy.

A linguistic phenomenon.

Homonymy is tricky to define in general, but it’s not the trickiest linguistic phenomenon to deal with either (try anaphora for a real snag). In practice, though, it may require some ingenuity, and even playing hero.

Playing hero: The two concepts of probability stability

Rudolf Carnap comes second on my list of all-time philosophical heroes (if anyone asks, David Hume comes first).

I already mentioned him in the Science and Bullshit of Lifting series, for something that turned up as a failure (finding a cretiron to tell apart science and pseudo-science). But hey, he’s a hero, and heroes sometimes fail against mere mortals (Achilles, anyone?). That being said, Carnap also scored more than a few successes. One of them was a general theory of synonymy (in the late 1940s-early 1950s) that would become the foundation of formal semantics. That’s another story, although it’s not unrelated to the topic of the day, which is homonymy.

Carnap1 may not have been a success but Carnap2 sure was

So, back to homonymy. In the early 1940s, discussions about the nature of probability were all the rage. The mathematics of probabilities had been systematized in the early 1930s but some big decisions had been made to do so and not everybody agreed. Plus, not everybody understood, which led to some big-time talking-past-one-another between mathematicians.

That’s where Carnap weighed in.

In 1945, Carnap proposed a distinction between two concepts of probability which remains to this day a masterpiece of philosophical analysis and cleared the air for the generations to come. He also proposed a notational convention that never took hold, between probability1 and probability2 (didn’t I mention that heroes fail?).

Following on Carnap’s footsteps (let’s not be too modest) I’m about to establish a distinction between stability1 and stability2 and hope that the distinction will stick in the reader’s mind even if the notation doesn’t. Also following Carnap, I’m building on older material. In fact, short of the notation, the distinction is more-or-less already in McGill’s Low Back Disorders.

Which is also a reminder that analytic stuff, like parasitic stuff, wouldn’t exist without what it is an analysis of. Then again, analytic stuff is therapeutic stuff which makes it more of a symbiote than a parasite. (Save for contemporary analytic philosophy, which tends to be a parasite of the public funding system.)

Stability as a functional concept

Defining either version of ‘stability’ in a physical activity context would be straightforward if it wasn’t for a huge smokescreen. Analytically speaking, stability is a functional concept. Unfortunately, in health-and-fitness contexts, the understanding of the phrase ‘functional concept’ may vary with the relative degree of exposure to :

  • mainstream fitness industry promotional material in any of its common forms (advertisement, magazine articles, YouTube videos, Facebook/Instagram feeds, etc.); and
  • abstract algebra in any of its common pure or applied forms, including in the latter case through engineering, computer science, or analytic philosophy.

And to make things easier, I’ll fix the notion of ‘functional concept’ using a staple of analytic philosophy (which, after all, was my way into abstract algebra): making up shit and playing around with words — also called a ‘thought experiment’.

In spite of all the ills I think of thought experiments (because they make that kind of nonsense pass for legit research) they are actually useful in real science: Newton’s cannonball never had to be fired and unlike Sherrington, Schrödinger never had to kill any cat. A useful feature of thought experiments is that they allow exploring extreme cases that we would never actually meet in real-life. Like so.

Thought Experiment #1: Full-on mainstream fitness without functional algebra. You walk into a gym where you are met by Joe PT Functional (‘PT’ stands interchangeably for ‘personal trainer’ and ‘physical therapist’), and you ask him what a ‘functional concept’ is. Since Joe PT has had zero exposure to abstract algebra, he immediately assumes that you’re talking about a ‘functional training concept’.

Ex hypothesis, Joe PT has also had zero exposure to analytic philosophy.[1] We can thus assume (without loss of generality) that his notion of ‘concept’ is non-standard and nonverbal, allowing him to bypass linguistic explanation and proceed immediately with illustration, with something like that whatever-it-is on the picture.

Joe PT is a fictional character, but the jackass on the picture above is real. Furthermore, with the exception of the breathing-impairment mask, every component of whatever-that-is he’s doing is motivated by the concept of stability I’ll introduce later. So keep it in mind.

Thought Experiment #2: Full-on abstract algebra without mainstream fitness. By contrast with Joe PT, Jane DM is a very methodical person, who has a good grasp on important things (as indicated by DM, which stands for ‘discrete mathematics’). To make things interesting, let’s assume that you ask her straight on to illustrate a ‘functional concept’. Now, because she’s not Joe PT, she knows that ‘illustrating’ does not mean ‘bypassing verbal explanation’. And since she knows that you are trained in Analytic Fitness™, she thinks that it’s better to start with some generalities and let you figure out the specifics, so doodles something like that:


Jane DM’s doodle is actually more useful to understand stability than Joe PT’s whatever-it-is. If we’d tell her that stability is a functional concept, she’d immediately get that stability is a function of something else and as every function can be described by its inputs and outputs. And if we asked her to define stability, she’d ask us for a set of inputs and a set of outputs and she would get to work.

And so, let’s do just that.

Analyzing stability

The first notion of stability, or stability1, is the one that almost everybody gets.

Again, by “everybody”, I mean everyone involved in physical activity stuff from the fitness-and-rehab PT/PT (should I write PT_1 and PT_2?) and the sport-and-performance athletes and coaches, to the exercise-science researchers. And by “almost gets”, I mean that they get the first half of the functional definition: the “input” part.

Honestly, I’m not sure everybody gets the “output” part. But I’m not sure they don’t get it either, as a general rule. There are however clear-cut cases where they don’t get it and we’ll see why.

What everybody gets (and what they miss out)

Most of the details of the discussion of the output part can be left for geeks-only asides. But I sure hope that you’ll read them, and to give you some reasons to do so, here’s the full input-output definition:

  • Inputs: relative positions of a center of mass (CoM) to the base of support (BoS) as given by a Cartesian coordinate system.
  • Output: an angle between certain lines originating from or passing through the Cartesian coordinates of the CoM and BoS.

In practice, we can abstract from the need of an explicit Cartesian coordinate system, as in the diagram below that I borrowed from McGill’s Low Back Disorders 3E (p. 157).

In order to decode this, we need a few auxiliary notions omitted by McGill and not represented on the diagram. For now, let’s focus on the left-hand side diagram (a) and pay no mind to angle θ.

First, look the at Big Triangle’s CoM : the projection of the position of Big Triangle’s CoM on the abscissa (the x-axis in a Cartesian Coordinate system, not represented in the picture) is called the center of gravity (CoG) of Big Triangle. If you confuse CoM and CoG, it’s not dramatic, but try not to anyway.

Second, look at the points of contact between Big Triangle, Left Small Triangle and Right Small Triangle. In the context of this discussion, Big Triangle’s base of support is an imaginary line segment that spans between the point of contact between Big Triangle and Left Small Triangle and the point of contact between Big Triangle and Right Small Triangle (but see aside below).

Now, try to picture the projection of that line segment on the abscissa. In diagram (a), Big Triangle’s CoG is smack dab in the middle of the projection of Big Triangle’s BoS which incidentally is the farther out Big Triangle’s CoG can be of each point of that segment at the same time.

The reason why this matters is that Big Triangle is the simplified 2D representation of a human, and that the usual recommendation to maintain a human’s stability1 is to pay attention to “keep the center of mass over the base of support” which is short for “keep the center of gravity over the geometric center of the projection of base of support”. Below is a short video of Steve Cotter using that kind of cue for coaching the pistol squat. (Tip: watch it a 2x speed).

Here’s the thing: everything that Cotters says makes perfect sense. But the cue to “sink over that base” when on one leg (03:10) for preserving knee stability incurs a risk of mixing up stability and balance. (And I’m not saying that Cotter mixes things up, because he isn’t, but it’s a risk.)

To see what the mix-up is, let’s have a look at the right-hand side diagram (b) but still paying no mind to angle θ or P (which represents a force, I’ll come back to that). Big Triangle is a 2D model of a human standing on one leg. Thus, Big Triangle’s BoS is a single point, which is (trivially) also a line segment.

Also now, low and behold, the cue of “keeping the CoM over the BoS” goes to the shitter. If you’d do that, you’d actually decrease the stability1 of BT (we’ll see why later) which is a stand-in for the stability1 of a human being standing on one leg.

This is not surprising since single-limb ‘stability exercises’ are meant to teach ‘balance’, which is why they are also called ‘balance’ exercises. However, stability1 and balance are not synonymous even by a long shot. ‘Balance’ is really the ability to counteract instability1 with micro-movements, whose effect would be represented by as many forces P, P’, P”, etc.

And this is why I said earlier that everybody almost gets stability1. They get the input, which is the same as with the functional definition of balance. But they don’t get the output. Below are some ruminations on the theme of single-limb exercises and a generalization factoring θ and P.

The Geometry of stability1. The geometry of stability can get really hairy, really fast. For instance, if we’d consider 3D applications (static humans), the BoS would be a 2D polygon (a projection on the ‘ground plan’). And with 4D applications (moving humans) it would be a 3D solid (‘shrinking’ or ‘expanding’ at time points). That’s why I stuck to the 2D model. And actually, if you listen carefully to Cotter (and even though he conflates sometimes balance problems with stability problems) and other people who (almost) know what they are talking about, coaching cues assuming that we are 2D are good enough.

‘Exercising’ stability1. Let’s assume that single-limb ‘stability exercises’ are really stability1 exercises, and let’s take ‘exercise’ as in ‘mathematical exercise’ or (more generally) as a problem-solving situation. And let’s illustrates this with a new thought experiment, which also contains a riddle.

Thought Experiment #3: Jane DM learns stability. Joe PT asks Jane DM to stand on one leg, and then to “get more stable”. Jane DM immediately figures out that, if Joe PT really means stability1, the solution to this problem is to increase the size of the projection of her BoS on the X-Y plane of a 3D Cartesian coordinate system and keep her CoG within it, but as far as possible of each point of its perimeter at the same time. However, she’s smart enough to realize that there are two equally good solutions to this problem, but that they cannot be ranked without secondary criteria. In order to rank the two solutions, she asks Joe PT what the purpose of the exercise is. Joe PT answers that it’s a functional exercise aimed at making her functional in activities of daily life. Now, Jane DM can rank the two solutions and choose the best: she drops on two legs.

[Here’s the riddle: can you figure out what was the other solution, and why Jane DM did not pick it?]

Thought Experiment #3 illustrates that the obvious purpose of single-limb ‘stability’/’functional’ exercises is not stability1. If it were, stability1 training would just be the art of re-establish stable1 bipedal stance, which would be tantamount to stepping out of the unstable platform and reestablishing solid ground contact. (The word “duh” applies here.) Notice however that decreasing the height of the CoM (equivalently, the distance between the CoM and the CoG) when standing on one leg increases the amount of force that has to be exerted to topple you over (which is the rationale behind Cotter’s cue) but that’s really stability2 and will have to wait for the next episode.

Beyond single-limb: stability1 and balance as optimization problems. There is a lot of confusion between the general notion of stability1, its application to bipedal humans, and ‘balance’, which is why even a simplified definition for stability1 cannot abstract from θ or P. For instance, assuming that BT is an abstract model of a 2D human, diagram (a) represents the solution to an optimization problem: BT is (most) stable because for every pair of points p1 and p2 on the BoS, the angle θ (resp: θ’) between the line parallel to the ordinate axis (the Y-axis) going through p1 (resp: p2) and the line going through the CoG (equivalently, the CoM) also parallel to the ordinate axis, are as far as they can be both at the same time from 0°.

By contrast, also assuming that BT is an abstract model of a 2D human, diagram (b) may represent two different problems: one is a balance problem and the other a stability1 problem. The stability1 problem is to increase the base of support, which amounts to generate a force P’ opposite P sufficient to topple down BT and re-establish contact with RST (Jane DM’s solution). The balance problem can be solved by either generating a force P’ counteracting P to maintain θ as is, or ‘yielding’ to P until θ=0° and then generating a force P” to keep θ=0°. The latter is in turn, a solution to yet another optimization problem, namely, find a position where, whatever force P threatens balance, the force P’ necessary to counteract it will be no greater than necessary (equivalently: find a position such as maintaining it requires the least amount of energy).

About problems One issue here is the concept of ‘problem’. Analytically speaking, problems are identified by the set of potential solutions. Hence, describing a ‘stability’ exercise by recommending a solution that involves balance makes it a ‘balance’ exercise after all. Of course, there’s nothing wrong with that, but increasing the challenge to balance puts a cap on how much other physical qualities can be challenging, and that, in turn, is an issue.

Summing up, almost everybody gets the input of the stability1 function but some mix up the output and mistake ‘stability1 training’ with training the ability to counteract P-forces, while it should be the ability to optimize θ-angles.

Then again, that is Jane DM-level analysis and not everybody is that smart. This brings me to how this confusion has spawned other mistakes and eventually bullshit.

Adding insult to injury: the category mistake

Whatever-it-is that Joe PT did in Thought Experiment #1 is inspired by two things:

  • a confusion between stability1 and balance, and:
  • the notion that increasing the intensity of an activity increases the training effect from this activity.

We’ve seen how (1) is possible, and (2) is essentially the rationale behind progressive overload (that I discussed in this post in the aside about the Law of Adaptation) so there’s nothing wrong with it.

Together (1) and (2) entail the notion that if the base of support is at its smallest, stability1 will be trained at maximal intensity while still ‘functionally’ (more on that soon). Which is, in turn, a dramatic category mistake perhaps responsible for the worst trends of the fitness industry since the 1990s (I’ll speculate about that in conclusion).

The concept of category mistake was introduced by British philosopher Gilbert Ryle in the 1940s, but it was already used in the 1920s-30s by the Vienna Circle, chief among them, Carnap (again). Here’s an example adapted from Ryle: someone has been told that there’s a marching army coming his way, watches the soldier passing by, and then asks “where’s the army? I only saw the soldiers!”

That’s a category mistake: it asks about something (the army) as if it were a different kind of thing, or category, as what it is (a bunch of soldiers) just because the words are different.

“Where’s the legion? I only see legionnaires…”

Also, if it’s a kid asking, it’s cute. If it’s a grown-up asking, it’s stupid.

Now, assume that you want to train an army to march, what would you do? That’s easy: you would train the soldiers to march. Similarly, training stability1 in motor patterns is training those motor patterns. The stability1 demand of a motor pattern is ‘built-in’: there is no point isolating that demand to train it at a greater intensity, because if you change this, you change that (I told you that anaphora was a snag).

Resolving anaphora, if you change the demand on balance you change the motor pattern so much so that you compromise transfer (aka specificity) for a bunch of performance parameters. You’ve probably figured that out. But the fine prints with a thought experiment that you can use to convince your friends who bought into Joe PT-type bullshit to throw their balance board through the window.

You’re in the army now. Let’s consider the problem of improving the yomping time of an army, or to make things more manageable, of a platoon P of about 50 goons. To further simplify, let’s agree to: (1) leave the officers and non-coms out of the discussion, which identifies P for all practical purposes with the proper subset of the platoon comprising only enlisted personnel; and: (2) define P as a set ‘in extension,’ that is, by the list of its enlisted members, in any order, which yields something like: P = {Private Snafu, Private Tarfu, …, Private Fubar}.

Let’s also define in extension another platoon, P’= {Private Bohica, Private Bimble, …, Private Buff}. Now we can we get to Thought Experiment #4.

Thought Experiment #4: Yomping with Joe PT and Jane DM. For years, the army to which P and P’ belongs has implemented sub-par physical training for improving yomping performance (for instance: the USMC Physical Fitness Test or the USMC Combat Fitness Test). Realizing the lack of specificity of this training, the high command hires Joe PT to design a ‘functional’ program to improve the yomping performance of the platoon P and Jane DM to do the same with platoon P’. Joe PT has hung out a lot with Jane DM since Thought Experiment #3, but still thinks she’s wrong about balance boards. And so, he figures out a solution inspired by ‘increasing the stability1 demands of a movement pattern’. He begins by demanding the same yomping performance P used to have on a flat-terrain yomping test but on sandy terrain. When his men perform poorly, he figures that he could: (1) remove the least-performing elements of P and borrow high-performing soldiers from other platoons; and (2) re-test whether the platoon yomps faster or not, and is thus closer to his increased demand. Private Snafu being the slowest, he is the first to go. Joe PT asks around, and Jane DM (who has figured out the flaw in Joe PT’s method) lets him borrow her slowest yomper, Private Bohica. At the next yomping test, Joe PT’s platoon is slightly faster. Emboldened by this success, Joe PT substitutes P‘s second-slowest yomper, Private Tarfu, with Private Bimble, also from P’ (and formerly its second-slowest yomper). Joe PT reiterates (1) and (2) until he is not testing P but P’, which happens to have the same yomping time on sandy terrain as P used to have on flat terrain. Realizing that he cannot do better, Joe PT hands over his new platoon to the officers. In the meantime, Jane PT has trained P with her own ‘functional’ methods, and they now match the performance of P’ both on flat and sandy terrain. She also hands over her platoon to the officers, who laugh their ass out with Jane DM, immediately fire Joe PT for breach of contract (since he never trained P), and give Jane DM his share of the contract money, because she has improved the yomping performance of both.

Decoding the analogy Thought Experiment #4 illustrates the twisted logic behind let’s-increase-the-stability-demand-of-an-exercise-by-doing-it-on-increasingly-unstable-surfaces by the following analogy: an unstable surface changes a motor pattern in such a way that the blend of physical qualities that is necessary to perform the motor pattern on the unstable surface is different from the blend of physical qualities necessary to perform the motor pattern on the stable surface. Similarly, Joe PT changed the surface and then had to change the blend of soldiers. And yet, it’s actually Jane DM’s training that was responsible for the observed performance of both P and P’. The part where Jane DM has trained P to improve their time on both flat and sandy terrain without training on sandy terrain at all translates the finding that strength training of young adults on stable surfaces improves stability1 about as much as training on unstable surfaces. At the same time, it also improves other performance markers (like speed and strength) that do not improve at all if strength training is performed on unstable surfaces. By analogy, the yomping performance of P’ has not improved, while the yomping performance of P has.[2]

Thought Experiment #4 is a little roundabout but if your friends do not get the argument and seem intent to keep their balance board, ask them to explain to you what’s wrong with the argument, or throw the balance board. I’ll bet they throw the balance board.

If you didn’t read the fine prints, here’s where the Joe PTs of the world get things wrong: you cannot increase the intensity of stability1 without decreasing the intensity of other physical demands that matter more for performance, be it for athletic purposes or for activities of daily life.

Also, it’s the result of a category mistake, and because people who make it are grown-ups, it’s not cute.

The Jane DMs, on the other hand, are on the right track. Their approach to ‘functional training’ manipulates input-output relations. We’ve already seen some of that elsewhere: training an army to march is ‘functional’ in that sense, as it tries to improve the output (max yomping speed) by manipulating the input (load, aerobic fitness, and strength), and I assumed that she was doing just that in Thought Experiment #4. Speaking of functional, let’s conclude with…

A recipe for ‘functional training’

By now, you have presumably come to the natural conclusion that any training aimed at improving any performance whatsoever is ‘functional’. And if you have not, just pretend.

Functional training follows the following recipe:

  • Operationalize the performance to be improved. This step amounts to find a way to measure it, and it’s usually straightforward for sports.
  • Hypothesize a functional relation between the performance and other stuff. This step is usually grounded in some knowledge base suggesting the hypothesis that some ‘input’ drills would (measurably) improve the ‘output’ performance operationalized at (1).
  • Practice ‘input’ drills and measure the ‘output’ performance. This step is the trial-and-error phase: if performance defined in step (1) improves, the ‘input’ practice is validated; if not, start over with a new hypothesis at step (2).

Defined that way, ‘functionality’ is just another term for the principle of specificity, which is one of the basic principles of exercise science. As always, the devil is in the details: ‘functional training’ is a fitness-and-rehab buzz-phrase that is supposed to pertain to the ‘activities of daily life’ (ADLs). This complicates matters a bit, but I’ll leave that for the fine prints.

Functionality as specificity (I): practicalities. The concept of Specificity and the adjective ‘specific’ are well-defined in exercise science and can be substituted (respectively) to the concept of Functionality and the adjective ‘functional’ in exercise-science contexts without loss of meaning. Since the possibility of substitution without loss of meaning is the litmus test of synonymy they are therefore synonymous in those contexts (they are not in others). An advantage of this substitution is that it yields another litmus test, namely as to whether the occurrence of “functional” was the likely consequence of some bullshitting (for instance, when the claim blatantly contravenes to the principle of specificity, as with Tabata-related bullshit). Unfortunately, the substitution test alone works only relative to sports performance. But ‘functionality’ encompasses both sports performance and ADLs, and the latter introduces qualitative outcomes (like improved quality-of-life). Quantifying these outcomes cannot be done without decisions about what kind of performance is indicative or predictive of improved quality of life. These decisions being underdetermined by science and evidence, they are vulnerable to the biases of analysts. An example would be the performance criteria for ‘improved back function’, namely increased spinal range of motion (ROM), which is easily measurable. Increased spinal ROM is a desirable outcome in some athletic performance (for instance, gymnastics) but trying to improve it typically delays rehabilitation of low back disorders, and may even lead to incorrect diagnosis, with far-reaching implications for health policies.[3]

Functionality as specificity (II): theory. The extension of the concept of specificity to ADLs can be viewed as a Carnapian explication of the notion of ‘functional’ gains. In a nutshell, the concept of ‘explication’ introduced by my hero Rudolf Carnap allows the substitution of a scientifically exact concept to a pre-scientific notion even without complete synonymy provided that the conceptual gains of doing so exceed the conceptual losses. Carnap’s theory of explication provides criteria to appraise conceptual gains and losses. However, in and of itself, it does not provide criteria for choosing one explication over another. For instance, the debate on ‘improved back function’ can be understood as an opposition between rival explications (in Carnap’s sense) of ‘back function’ in daily life (ROM is one, and the other is more complicated and will have to wait for the definition of stability2). See the Wikipedia page for a quick summary of Carnap’s theory (nothing wrong with it), a note by Branden Fitelson for full quotes with an example (§1) and some philosophical nitpicking (§ 2 sq. TL;DR: Carnap comes out on top) and the section on explication in the Stanford Encyclopedia of Philosophy (in the ‘Analysis’ entry) for another quick summary within a not-so-quick summary of the broader context.

Conclusion (for today): The irony of ‘functional training’

The irony of the confusion between stability1 and balance and the category mistake about stability1 is that ‘functional fitness’ was initially the fitness-and-rehab response to bodybuilding.

During the 1990s, at the height of the bodybuilding craze (which was also the golden age of steroids), isolation exercises were the main paradigm for general fitness. But in the late 1990s-early 2000s, under the influence of physical therapists (like Gray Cook) and biomechanists (like McGill), the fitness-and-rehab world gradually realized the lack of efficiency of isolation exercises for rehabilitation and preventions of injuries both in everyday life and sports practice.

Stuart McGill mostly stuck to his science, while Gray Cook spearheaded the ‘functional exercise’ movement, eventually proposing a metric for ‘functionality’ and a certification program (Functional Movement Screen™ or FMS™ ) aimed at teaching how to measure and improve ‘functionality’ ( but it’s most likely just bullshit [4])

And so, ironically, ‘functional fitness’ ended up trying to isolate physical qualities at the expense of others (Pavel Tsatsouline makes a similar point in this video). In retrospect, CrossFit® may very well have been a reaction to ‘functional fitness’ in the same way as ‘functional fitness’ had been a reaction to bodybuilding.

Didn’t I tell you that all the confusion around stability1 was responsible for the worse trends in the fitness industry since the 1990s? Ok, the part about CrossFit® is a bit of speculation on my part, but it’s a tempting hypothesis.

In the follow-up, I’ll look into Stuart McGill’s effort to spreading awareness about stability2 and fight a confusion between stability1 and stability2 that plagues both practice and research in exercise science and sports medicine, and threatens any ‘evidence-based’ suggestion about what ‘stability training’ should be.

I will also argue that we can survive all that mess and train stability{1,2} in one fell swoop without giving a fuck about balance boards, BOSU® and ‘functional’ exercises, with just a few hypotheses about what matters for ADLs.



[1]^ Proof: The proof is by reductio ad absurdum. Assumes, for reductio purposes, that Joe PT has been exposed to analytic philosophy, noted AP(Joe PT). Whoever has been exposed to analytic philosophy has also been exposed to abstract algebra which we can abbreviate: “For all x, If AP(x), then AA(x)” where AA(x) is short for “x has been exposed to Abstract Algebra”. Hence, we have “If AP(Joe PT) then AA(Joe PT)”. By reductio assumption, we have AP(Joe PT). And by the latter two and Modus Ponens (P, If P then Q => Q), we have AA(Joe PT): Joe PT has been exposed to abstract algebra, contrary to hypothesis. Therefore, we must abandon our reductio assumption and conclude that not-AP(Joe PT): Joe PT has not been exposed to analytic philosophy. QED.

[2]^Behm et al. (2015) Effects of Strength Training Using Unstable Surfaces on Strength, Power and Balance Performance Across the Lifespan: A Systematic Review and Meta-analysis, Sports Medicine 45:1645–1669, DOI: 10.1007/s40279-015-0384-x. In order for the analogy to be correct, we need to assume that Joe PT’s tests do not have notable training effect since yomping on sandy terrain actually carries over to yomping on flat terrain. The test could be spaced out in such a manner that although they train Joe PT’s platoon, the platoon detrains between tests (perhaps because Joe PT buys into some myths about recovery, that I’ll cover in a future installment of that series).

[3]^See S. McGill (2017), Low Back DIsorders (3rd ed), chap. 1 (esp.14-15) for ROM as a criterion. Chap. 1-2 examine the impact of ‘normal function’ criteria on public health policies, including patients being denied compensations on the grounds that their back pain is ‘psychological’ when ROM is restored but the pain does not subside.

[4]^ There’s about 10 years of research on FMS™ with several meta-analyses. So far, they do not fully agree, but there’s a trend: those who support FMS™ do so with explicit methodological reserves (like this one) that do not undermine those who recommend abandoning FMS™ altogether (this one and that one). Another warning sign is that there is a joint FMS™-RKC® certification administered by Dragon Door (see here and here for Dragon Door’s history of bullshit).:

Leave a Reply