The Repugnant Conclusion and why it Violates our Moral Intuitions

The “repugnant conclusion” is an uncomfortable bit of philosophical reasoning first presented by Derek Parfit.
For a detailed discussion of the repugnant conclusion, check out this article from Stanford philosophy.
Here’s a quick summary.
Consider a population where we can measure “well-being” which aggregates everything that makes life worth living into one, rolled-up measure.
(People often get tripped up here, so consider that this includes everything in one aggregate measure. This measure is not quite the same as “happiness.” Pain, loss, sorrow, profundity, fulfillment, pleasure, etc. are all included here – as are the trade-offs between doing something unpleasant to achieve a positive outcome, etc. Any sort of Malthusian population dynamics are also accounted for in this aggregate measure of well-being, as are any amounts of individual dissatisfaction based upon inequality. You may dispute that a single measure can capture all of this information, but, for the sake of this thought experiment, assume that it can.)
Assume we have a population A with a given well-being (the height of the rectangle is the total well-being, and the width of the rectangle is the population).
Then, assume we have a population A+, which is the same as population A, except we are adding an additional population with lower well-being than population A (although everyone still has a relatively high well-being).
Population A+ seems obviously “better” than population A (or at least not worse), as there are more people with positive lives – and the total well-being of the entire group is higher.
Then, assume that there is a population B with the same number of people as population A+.
However, their well-being is averaged across the population such that the total well-being is the same as A+. Rather than having two groups with different well-beings, we now have one group with the same well-being.
This seems to be better (or, again, at least not worse) than A+, which is better, or at least not worse, than A.
If we continually perform this operation, we end up with a population Z with very low well-being (but still net positive lives). 
So, we’ve show that a very large population with lives just barely worth living is preferable to – or at least not worse than – a smaller population with a much higher well-being for each individual.

Feels rather repugnant, doesn’t it?
There have been a lot of attempts to reason around the repugnant conclusion, but many of these attempts themselves result in bizarre conclusions themselves (termed the “sadistic conclusion” and the “absurd conclusion” – see the Stanford article for further discussion.)
I think that the best way around the repugnant conclusion is by questioning the additive properties of well-being. This is touched on here in the Stanford article.
If you asked how many ants would have to have perfect lives to make it acceptable to kill one human being, most people would say “none.” In effect, they are saying that no number of ant lives is worth one human life.
If we frame this in terms of well-being, we can say something like, “Some amount of ant well-being is probably worth sacrificing some amount of human well-being. However, there is a threshold of human well-being below which no amount of increased ant well-being can justify the trade-off.”
I think we can have a similar understanding of the “well-being” of human lives. A single life above a certain threshold of well-being is worth “more” than any number of lives below that threshold of well-being. Below certain thresholds, we cannot simply add lives to get a better outcome.
Think of this like Class A and Class B voting shares in a corporate structure. We could theoretically create an organization such that, while Class B shares are able to vote, they cannot change the outcome of a decision if the majority of Class A shares vote in a certain direction.
We can imagine “tiers” of well-being that are created based upon something like diagnostic criteria for complicated syndromes and disorders. Something like the Beighton score for joint hypermobility – if you have 5 of these 9 criteria, then your life and well-being cannot be offset be anyone with less than these numbers.
While I think this is sufficient to avoid the repugnant conclusion in a philosophical sense, I don’t think that this is why most people find the repugnant conclusion so repugnant.
Instead, I think our intuitions follow something closer to a desert-based approach to human well-being.
Simply by virtue of being human, each individual deserves a certain quality of life.
So, imagining all of these people with net positive lives that are just barely worth living violates our intuitions.
This approach, however, doesn’t quite avoid the repugnant conclusion. Instead, it just moves the line for what we would consider a “net positive life.”
So, now we just have a different threshold above which lives must exist in order to be “positive,” and we can still run through the operations to generate a repugnant conclusion.
However, instead of imagining a huge population of people existing on “potatoes and Muzak,” we are now imagining a huge population existing on – I don’t know – slightly burnt toast, watery coffee, slow WiFi and Trapt.
While maybe not quite as repugnant, I think we now just have a “distasteful conclusion.” Still, I do think this is why most people’s moral intuitions are immediately violated when they first understand the repugnant conclusion.

Coaching the Complicated vs Coaching the Complex

As coaches improve their skillset, they become more adept at pattern-matching and quickly being able to diagnose flaws in someone’s movement or approach to a training session

They also develop more “soft skills” in communication, approaching clients, dealing with emotions, and generally navigating the interpersonal issues surrounding coaching.
We often talk about “best practices” – meaning that there are in, in fact, “best” ways to assess our clients, prescribe a program, and engage in ongoing communication.
However, after reading a bit about Dave Snowden’s Cynefin framework, I think that the domain of “best practices” is not necessarily appropriate for thinking about coaching.
When a coach is getting start, they must go through the process of learning how to handle “simple” scenarios. 
If clients are rounding their backs on deadlifts, tell them to stop rounding or go lighter. If clients are going out too fast on long workouts, tell them to back off and slow down. If clients are struggling to put their arms over their head, have them foam roll their upper back and do serratus wall slides.
Each of these scenarios can be put into a basic flowchart of decision-making.
However, everyone who has been coaching for awhile recognizes that this kind of simple decision-making is often inadequate and, in some cases, can be counter-productive.
A lot of these scenarios are “complicated” – meaning that there is a right answer somewhere, but it may be difficult to get to and often requires deep knowledge and expertise. 
And, there’s often more than one way to get to a correct answer.
Why can’t someone get their arms over their head. 
Let’s dig into their breathing patterns. Let’s have them do a variety of different scapular motor control exercises. Let’s have them work on repositioning their ribcage. Let’s have them get soft tissue work to release facilitated and inhibited muscles. Let’s have them try both unilateral and bilateral movements.
The underlying issue could be any of these things, none of these things, or some combination of all of them.
An expert is able to conduct a thorough assessment and come up with a plan that will ideally end with the “correct” outcome for the individual.
A lot of movement correction falls into this bucket.
But, as we move up the layers of abstraction into coaching an individual rather than just fixing a specific movement issue, we are now playing with complexity
Rather than thinking of the human body as something like a Ferrari where you have a schematic and you can figure out the “broken” piece and repair it in order to get everything back on track, we are now have a dynamic system with emergence, bidirectional causality and many layers of competing systems.
There is no schematic, every change you make to the system causes multiple different feedback loops to kick in, and there are huge contributions of randomness and chaos to the actual outcome of coaching someone.
So, what is our role as a coach in such a complex system?
How do we help individuals stick to a workout routine and change their behavior surrounding nutrition?
How do we help athletes maximize their potential in sport – especially when the competitive landscape is constantly changing?
How do we navigate the self-sabotage, lack of consistency, deception, and self-deception typical of working with both every day people and high level athletes?
So, coaches need to “probe” through conversation and try out different prescriptions to see what is working and what is creating buy-in, then attempt to execute on that.
Coaches must also enlist clients to do their own probing to figure out what works for them and to more accurately characterize the obstacles that are holding them back – not just from an exercise perspective, but, more importantly, from a behavioral perspective.
And, when we make changes, we must also recognize that inputs into the system change the system itself, and that there is also always the lurking specter of black swan events like injuries, overtraining, and large life events outside of the gym.
So, some coaches live in flow charts:
Don’t round your back when you deadlift.
Eat one gram of protein per pound of bodyweight.
Push your knees out when you squat.
Some coaches live in nuanced assessments and detailed program design:
Here’s your two week testing period to figure out your relative strengths and weaknesses.
This is a 12 week squatting progression with accessory work meant to strengthen your upper back and improve your front squat relative to your back squat.
These are your macros. Make sure you weigh and measure all of your food, and we will check on that every week.
But, the highest order coaches combine the flow charts with nuanced assessments and detailed designs – and also enlist the clients themselves to probe at what is working and what’s not working so that the real challenges can be uncovered and solved through bottom-up behavior change rather than top-down design.

Scott Young (Author of Ultralearning)

Scott Young

I’ve long been a Scott Young fan from my early days of reading blogs (miss you, Google Reader), so I was thrilled to get the chance to interview him about his new book Ultralearning. 
As a relentless consumer of information and a sometimes autodidact, I’ve found Scott’s blog to be very insightful in terms of approaching new projects and learning skills like coaching and coding without going through a formal educational process. 
With the current ubiquity of information – including entire college curriculums, endless video interviews with world-class experts, and entire industries of online courses – we should be able to learn just about anything we want. 
However, as anyone who has either attempted to learn a new skill or, God forbid, teach someone else a skill has experienced, learning is really, really hard
How can we actually transfer what we learn from theoretical lectures and books to real-life application? How can we practice skills in a way that makes us better at the skill itself – not just at random drills?
Check out the full conversation with Scott to learn:
  • Why listening to lectures, watching videos and reading books doesn’t tend to translate into applicable learning in most areas – and what to do instead so that you can actually transfer the things you learn into real-life practice
  • How to find the right balance between practicing the skill that you’re learning and doing drills – since some drills don’t translate effectively but only practicing the full skill can prevent you from improving on weaknesses
  • What really separates those who are willing to self-educate from those who aren’t – and why the common barriers (“I’m not smart enough,” “I’m not talented enough,” etc.) aren’t accurate

Check out the episode at the links below. If you enjoyed the episode, the best way to support the show is to share with your friends, so send them a link.

Listen Here

Check out more from Scott here:


Show Notes:

  • [1:50] Scott seems to understand that learning is most successful in an environment of doing and not in one of reading, lecture attendance, and video watching. However, he’s published a book about learning – so what exactly is its purpose and why did he choose to write it?
  • [4:21] There’s a lot more to learning than simply practicing, as certain skills seem to involve endless amounts of practice while others have more apparent, speedy transfer. Learning becomes more difficult when the type of practice performed deviates from how the skill is used in a real-life setting. Scott gives some examples of when these transfer problems arise and how transfer problems can arise even in learning about theoretical ideas.
  • [8:46] Directness and actual application are significant in order to learn all skills, but the order in which they’re performed matters. A learning strategy is likely transferring effectively when exposure to a skill is direct prior to performing any sort of drill and, once drills are introduced, it becomes important to return back to those situations of direct exposure regularly.
  • [13:01] Skills can be built up individually while lacking functionality outside of largely abstract situations, meaning that drills must be specific and relative to real performance of the skill. Feedback on those drills (and, generally, on performance of the skill being learned) shows to be a nonessential piece of the learning process.
  • [20:12] We can get knowledge into our heads, but accessing a learned skill isn’t done by pulling out a ‘saved’ memory from the brain and feedback is self-generated through realization of what is not able to be recalled – that aspect of retrieval is vital to performance of any skill, making the sophistication of recall more effective than repeated exposure.
  • [24:38] Studies may not be representative of all populations since skills vary so greatly in context – amount of acquired knowledge and ease of retrieval positively correlate, and sample sizes tend to be small. Giving learners opportunities to apply what they’ve learned can be a step toward bridging the gap in education where people review and ‘understand’ concepts but cannot seem to make any real change behaviorally.
  • [25:54] Experience is one of the many reasons experts perform better than novices at almost any skill – an expert’s experience in a particular skill allows them to chunk things together and to see prior patterns, obvious mistakes, and recognition of solutions to problems more readily than a novice, who likely attempts to piece together a multitude of individual parts of a larger concept.
  • [33:13] Learning a skill in order to solve problems rather than to simply know the information and to have it ‘stored’ can improve one’s ability to transfer. Autonomy is a necessity though: being able to apply a skill that you don’t really want to use is unlikely, no matter what super effective strategies or level of established intelligence or personality traits are present. Anyone can learn almost anything if they want to.
  • [43:04] Many people have negative experiences with learning and associate learning struggles with failure. Once you know how to put together a puzzle, it isn’t a puzzle anymore, but confidence and persistent engagement are keys to keep trying at that puzzle.
  • [53:10] Knowledge decay isn’t as serious as many believe because large ideas are retained – making a habit of performing physics problems or speaking in a particular language can help in maintaining those learned skills, but even more abstractly reminding yourself of formulas that exist can be helpful.
  • [57:38] Attitudes surrounding learning are the difference between either merely knowing about many concepts and drowning in self-doubt or having the confidence to succeed in complicated areas of work such as ultra learning. Can we make it prestigious to be a motivated self-educating person?
  • [1:04.35] Being able to copy someone else’s behavior or learn how someone else performs well at a particular skill by being able to watch and communicate with them about the subject can enhance and expedite the learning process. However, it’s possible that this is true in skills with more clearly defined ‘rights’ and ‘wrongs.’
  • [1:14.10] How to get Scott’s book if you want it. And you probably want it. And you probably also want to check out some more Scott content. So here’s how to get all of that.

Links and Resources Mentioned

You May Also Like These Posts