The Art of Training Young People
Apprenticeship; or, how to resolve the precarious acquisition of craft.
“You're Overpriced"
Benjamin Graham to Warren Buffet when he offered to work for him for free. Graham would hire Buffet three years later starting him on his career as an investor.1
Young people are pretty useless. Humans have one of the longest periods of juvenile dependence of any animal. It takes 4 months after we're born before we can hold our heads up on our necks and it takes at least another 8 months before we have the coordination needed to feed ourselves. We limit people from voting in elections before they are 18 and increasingly there are arguments we would be wise to not consider someone a fully formed adult until they are 25 and their pre-fontal cortex has stopped developing2.
Alongside this long developmental arc there is an even more pernicious learning curve for competence. Our ability to learn things and bootstrap new skills is clearly phenomenal. If it weren’t you could never explain the heights of human art, science, and craft that we see exhibited every day. But to get there takes energy, good teachers, a conducive environment, and most importantly time. Kids who haven't yet had the chance to learn in this manner are, for lack of a better word, useless3.
So, when Warren Buffet went to Benjamin Graham - the legendary inventor of value investing - and offered to work for him for free Graham was not wrong to say that free was overpriced. Rather than Buffet’s free labour representing a boon, bringing on the young upstart would have been a drag for Graham as he exerted all kinds of energy to teach the youngster the tricks of the trade.
This seems like a pretty brutal chicken and egg problem. If young people are so useless how can we justify the cost of creating a learning environment for them so they can start to provide value.
One of the first ways we tried to solve this problem was with apprenticeships.
The History of Apprenticeship
The idea that a young person should be bound to a specific master so they can learn the fundamentals of a craft is a human universal. It appears in some of the oldest records we have from China4, the Middle East, and the Mediterranean. However, the form of apprenticeship that we commonly reference today was a product of early modern Europe, and in particular England.
The basic structure looked like this: A master craftsman - a person skilled in a trade like tanning, carpentry, cordwaining5, or tinsmithing - needed additional labour to expand their business. At the same time, a young person who needed to gain an economic foothold would look to learn from that master. To formalize the arrangement the two would sign articles of indenture tying the apprentice to the master’s house.
The period of apprenticeship began young, usually around the age of 13, so the contracts were often brokered by the apprentice’s parent or guardian. This protected the apprentice from signing something they couldn’t understand, and offered some guarantee to the craftsman that they would not be put out by a feckless youth breaking a contract.
Commonly the tenure for the apprenticeship was in the range of 7 years. During this time the apprentice lived with the master, receiving free room and board as well as some very modest form of compensation. In the final years the apprentice would 'graduate' to become a journeyman or a freeman6 - a recognised skilled craftsman in their own right - and receive the wages appropriate to that role while still in the master’s shop.
Benjamin Franklin describes his apprenticeship in precisely these terms:
"[I] at last was persuaded, and signed the indentures when I was yet but twelve years old. I was to serve as an apprentice till I was twenty-one years of age, only I was to be allowed journeyman's wages during the last year. In a little time I made great proficiency in the business, and became a useful hand to my brother."
The Autobiography of Benjamin Franklin
This pattern was well understood and almost everyone, regardless of social position, interacted with apprenticeships in some manner or another. However, the institution remained ill defined.
Apprenticeships had been entered into the law in 1563 starting with The Statute of Artificers enacted by the parliament of Queen Elizabeth I. In principle, the law meant that the length of contract could be enforced and, nominally, guilds had the ability to set wages. In practice, apprenticeships were essentially unmanaged and varied dramatically from one town to the other, let alone between professions. Ultimately each apprenticeship contract was a unique agreement between apprentice and master.
As Joel Mokyr puts it the apprenticeship market was similar in this respect to the marriage market. Everyone understood what it was in the abstract and knew it was essential to the operation of society but it had a flexibility to encompass nearly every kind of human interaction within a particular economic frame.
Although it was widespread apprenticeship was a brutal business. Some data shows that as many as half of all apprenticeships would not finish their term. Some of this was due to unforeseen circumstances such as serious economic hardships or the death of either the apprentice or master. More often though the contract was broken early by either an unhappy master reneging on the bargain or a fleeing apprentice taking their skills elsewhere. Again, as a typical case, this can be seen with Benjamin Franklin who ran away from his brother in Boston after just 5 years to head to Philadelphia as he felt his brother had "demeaned me too much"7.
Nonetheless, despite this rough hewn system, and the high attrition rate, apprenticeships flourished and played an increasingly important role in European economies.
In fact there have been arguments that apprenticeships laid the foundation for the industrial revolution in England. The system created a stock of “brilliant tinkerer[s] with little formal education but with excellent mechanical intuition, good hands, and a quick mind”8 that were essential in the early stages of the industrial revolution.
Over time, as part of an ostensible desire to decrease exploitation on both sides, there was pressure to standardize the model and bring it under institutional control. This happened, first, through self-administration by the guilds of master craftsmen, then increasingly by the courts, and finally through rounds of government legislation9.
In the US, as Benjamin Franklin's story shows, apprenticeships were alive and well even before the declaration of independence. However, they took much longer to standardize than in the old world, the first law governing apprenticeships in the states was only passed in 1911 by Wisconsin, and the law governing apprenticeships at the federal level - The National Apprenticeship Act - was only passed in 1937 when the Department of Labour was put in charge of regulating apprenticeship as part of the series of reforms that took place under FDR.
Despite the important start for apprenticeship the historical trend for the form in the last century has not been good. In the US in 1947 there were nationally 94,238 registered entrants to apprenticeship programs. In 2010 there were only 109,989 despite a doubling of the population. With limited exceptions this kind of trend has also been true across Europe.
There were three reasons why the apprenticeship model fell out of favour. The first was that the dynamics of the market changed.
Apprenticeships work so well because the skills of the craftsman are holistic. If you go to a master carpenter you expect to employ someone who not only knows how to make a chair of a certain set of dimensions but someone who can “work with wood” along with all that such working entails. This calls for a kind of skill that is complex, generative, and fully embodied.
As industrial society developed however, rather than the well rounded craftsman of the early modern times we needed more individuals that could operate in a factory setting doing one thing very well. In such a model the need for a long period of training fell away. As Joel Mokyr describes it:
[The Apprenticeship] labor market was affected by the growing division of labor: the finer the division of labor, the simpler the tasks and the easier it would be to get an untrained beginner to be productive (even though the master himself had to acquire supervisory and managerial skills). As markets expanded, the division of labor became finer and the demand for unskilled labor increased even if the workers were termed “apprentices.” The closer the relationship was to one of pure wage labor as opposed to training, the less reluctant the master was to take on more apprentices.10
The second shift was a cultural change that accompanied this economic change. We began to prioritize technical, formal and propositional knowledge over general, informal, and procedural knowledge. Instead of skills being learned on the job we imagined a world where the same skills could be taught just as well, if not better, in a classroom setting.
Lastly, as apprenticeships become more strictly codified and governed the possibility of creating apprenticeships in novel industries decreased. The economic arrangement of an apprenticeship, as such, is flexible to nearly any skill that humans could express. However, after the system was set down legally if you were unwilling or unable to go through the hoops to register your novel industry like marketing, supply chain management, or software development then it was a system that became much harder to participate in.
In this move away from the apprenticeship model it feels that we may have lost something. Our young people didn't stop being useless. Are the techniques that we use today to learn our crafts just as good as the apprenticeship models of old?
"The fundamental learning situation is one in which a person learns by helping someone who really knows what he is doing." —Christopher Alexander et al., A Pattern Language
Why does Apprenticeship work?
Ultimately the apprenticeship model was created to solve a labour market problem. The master, needs to be able to find cheap labour but the available pool of trained freeman is too small. The apprentice, needs to gain skills that will make them employable but still needs to support themselves - at least with food and shelter - while they remain an economic drag.
The solution to these two problems is to create a contract bridging the two positions through a combination of time and the right learning environment.
First, the 7 year period of indenture ensures there is enough time in the relationship for both parties to benefit.
At the beginning of the contract, the master loses money while they are training the apprentice. However, they recoup these loses later by continuing to pay apprenticeship wages, once the apprentice is operating at a level closer to a journeyman.
The apprentice on the other hand can lock-in the wages they need for sustenance while still being confident that they will receive the education they need. They will have the direct attention of a single master for the full period of indenture, and more importantly, the incentives are aligned. The better the master is able to train the apprentice at their craft during their tenure the more value they will recoup from having the apprentice work for them later on.
Second, both sides believe that the environment created by the apprenticeship will be optimal for the apprentice to learn.
Today we associate almost all of our learning with what goes on in schools and laboratories, or with ‘book learning’. In this environment we deal with a series of formal propositions: “Force = Mass x Acceleration”, “Concrete is good for fire-resistance and sound-proofing", “A dollar today is worth more than a dollar in the future”, “Python is a particularly adept language for working with big data and complex mathematics”. We memorize and manipulate these propositions to derive ‘rational’ solutions to the problems we face.
The truth however is that the vast majority of what we need to know to operate effectively in the world is not graspable in a propositional way like this.
Most knowledge is tacit or non-formal it either has not, or perhaps even cannot, be written down as a proposition11. The learning may involve a set of non-formalised heuristics and a distinct pattern language for the craft but nothing that can be used in an algorithmic manner.
Instead, when an apprentice learns from a master, rather than learning formulas, proofs, and taxonomies they take in this tacit knowledge through a kind of “monitored participation”12 . In other words they participate in the workshop of the master, and attempt to mimic them with feedback to let them know whether that participation and mimicry has been effective.
To start with they are given the most menial and low-risk tasks like cleaning and organising or making deliveries. Then as their competence is proved they ‘graduate’ to following more complex behaviours.
The other reason this has to be the case is that the skills that the apprentice learns are multi-faceted, dealing with every aspect of a particular craft, so they require a holistic learning environment. This is why the requirement to provide the apprentice room and board was seen originally not just as an arrangement to keep them fed and sheltered but an essential part of a fulsome education13.
The importance of this tacit and holistic knowledge along with the long timespan needed to take it in is why the cultural assumption we have that schooling can replace the apprenticeship style relationship is a bad one.
The situation today has not substantially changed from prior centuries. Although we live in an industrialized and services oriented society we still have huge problems figuring out what to do with our youth. This is, of course, backed up by the experience of every person that has left school to get a job and immediately realised they were woefully unprepared14. But more importantly, the gap between youth and adult employment15 remains stubbornly high, and appears to even be getting worse over time.
This situation is similar globally and in some places worse. In China in the last few months sharp increases in youth unemployment (when including those in school) jumped as high as 20% and even the fully adjusted rate has remained around 15%16.
Within this climate some people have again attempted to promote apprenticeships as a solution.
In the US the first major change was an update to the NAA made in 2008 as one of the final acts of George Bush's term. The change introduced a managerial rather than simply regulatory role with the Department of Labour and created a group responsible for the active promotion of apprenticeships. Despite this new role however apprenticeship numbers initially remained sluggish.
The definition of an apprenticeship in the new law was relatively broad: a length of at least 2000 hours (~1 year), minimum age of 16 etc. and there was new funding and power for the DoL but they were unable to face the core issue that over time the set of Registered Apprenticeships had become restrictive and mismatched to the economy. There was no meaningful way to increase the number of apprenticeships in a mature apprenticeship market like steel manufacturing if the industry as a whole was stagnant.
This changed in 2017 when a Trump executive order created Industry-Recognized Apprenticeship Programs. These essentially allowed any company to begin operating an apprenticeship program as long as it was vetted by an approved party like a trade group, education institution etc. In it's early years this appears to have allowed for an expansion of apprenticeships by adding new fields. Though it remains to be seen how much of this recent success is simply an artefact of the pandemic and associated changes in labour markets.
In recent years there has also been a revival of interest in the apprenticeship format - though not always using those words - among some of the larger tech companies. Tesla has introduced multiple programs to train high school graduates in advanced manufacturing and Shopify's Dev Degree program gives the same cohort a way to earn a CS degree while working at Shopify for 4 years and earning a salary17.
Anecdotes from the Dev Degree imply that it still is faced with the problems of traditional apprenticeships that attrition is high with many of the students not going on to continue working for the company. However, this doesn’t imply that the program will not be a success. Just as with grocery sales or credit card debt, there will be some rate of loss that is acceptable to the employer. And, without doubt, the ecosystem of software engineering as a whole will be healthier from all the people that picked up the holistic skills of a developer then chose to go use them somewhere else.
From Apprentice to Freeman
"You shall obey—someone and for a long time: else you will perish and lose the last respect for yourself" - Nietzche
Apprenticeships are a path to a thick skilfulness in a craft and a real solve for the problems of training and helping the next generation of young workers become productive members of the workforce, but they are also more than this.
The reason that in early modern Europe an apprentice was called a freeman or journeyman at the end of their tenure was that they were qualified to be a 'free' citizen or to ‘journey’ out into the world. They were prepared to live and work in a city without restriction. The apprenticeship had liberated them not just economically but socially.
Once at university, in a fit of religious zeal, I was forcefully expounding to a professor and friend about how I thought I needed to become a monk in order to find enlightenment. I reasoned that without the complete seperation from society that the monastic life promised there would be no way to fully dedicate myself to that journey. In response my professor wryly explained that, contrary to popular opinion monks were in fact a lot like the army.
In the army you gather a group of people who for one reason or another are usually at the fringes of society or looking for a significant change in their life: socially outcast, ejected from home, on the cusp of university, under-ememployed etc. and give them food, shelter, community, a habit of discipline, and - most importantly - an identity. The army on surface separates people from the broader society but in reality it provides them with a universally recognized place within it.
I thought that I needed to go become a monk because it was the particular skills of a monk that would lead me on the right religious path and that there would be no way to develop those skills in society. It turns out in retrospect that what I was really craving was to grow up and join the society as a recognized individual with an identity and the monastery seemed to be a path to do this.
As Jordan Peterson puts it:
"Discipline should therefore be regarded as a skill that may be developed through adherence to strict ritual, or by immersion within a strict belief system of hierarchy of values. Once such discipline has been attained, it may escape the bounds of its developmental precursor."
Like the army or the monastery apprenticeships are a sacrifice. The apprentice sacrifices many of the best years of their youth and the multiplicity of career options that they could see before them to undergo a specific course of training. Yet, just like the army or the monastery apprenticeship also represents a gift of freedom. At the end of their tenure the apprentice has been given an identity and a chance to navigate society as they see fit.
In our noble drives to democratize access to the liberal arts and our cultural valourization of the dropout entrepreneur we have inadvertently looked down on the creative capacity, economic windfall, and social liberation that can come from apprenticeships. As we seek as individuals and as a society to find new avenues for building capability we would be wise to rediscover this form.
It may not look exactly like the apprenticeships of old. Certainly the Tesla Manufacturing Development Program and the Shopify Dev Degree wouldn’t perfectly fit the mold. But, if we can find ways to allow young people to make long-term commitments to become part of a tradition of tacit knowledge and skilfulness we will all be richer for it.
And, for the young people, they should be encouraged that if they are willing to make the kind of sacrifice that an apprenticeship entails the economic and social rewards are more than worth it.
Some Takeaways
Apprenticeships are a time tested solution to not only the challenges of youth unemployment but more profound issues of skill development.
By enforcing a long tenure and being part of a holistic work and life environment the apprentice picks up the tacit knowledge that is needed to excel at a craft and in a workplace. Something that school and ‘book learning’ is structurally incapable of. As AI becomes capable of automating all work that can be captured as propositions it will be increasingly important to cultivate these kinds of environments.
For the masters, or companies, that run them apprenticeships are a risky business. However, they can solve a labour market problem and more importantly they create an on-ramp for the kind of skilled labour that could one day provide a huge surplus value.
There is momentum behind experimentation with apprenticeships today as the legal structure has been liberalized and there is a cultural shift with companies realizing that school can not replace training on the job. To accelerate this we will need to lean into giving employers the freedom to offer lower starting wages and longer tenure of contract as these are the core elements of what make apprenticeships a worthwhile trade.
Apprenticeships require serious sacrifice from both the master and the apprentice. The kind of sacrifices that we are not comfortable with today. However, the long-term economic and social rewards are significant. And, the path to mastery is one that ultimately leads to personal achievement and freedom.
Acknowledgements
Thanks to Brian Hagood, David Zhao and Chris Spoke for reading early version of this essay. I could not have written this without the paper The Economics of Apprenticeship by Joel Mokyr.
To be clear this is not a statement about whether or not we should believe in young people and ask them to try for outlandish goals - I agree with Tyler Cowen that for the most part young people are underestimated. It is simply an observation about the time span needed to really develop new skills.
中国成语 "一日为师,终身为父 "就是一个例子。
There are some subtle differences between these two to do with how people related to their guilds and their ability to get work but for now suffice to say that they qualified the apprentice as someone with full rights to practice their craft and a stamp of approval that they could practice it well. Though not quite yet themselves considered a "master craftsman"."
Note though that he also took advantage of a legal quirk that had led his brother to end his original articles of indenture. The possibility of some repercussions were real.
Morgan Kelly, Joel Mokyr, Cormac Ó Gráda, Precious Albion: A New Interpretation of the Industrial Revolution
At this point the guilds very rapidly lost their status and importance in setting the standards.
Joel Mokyr, The Economics of Apprenticeship
Michael Polanyi defined tacit knowledge as: ““the observance of a set of rules not known to the person following them.”
The phrase if from Mokyr. For more see Situated Learning by Jean Lave and Etienne Wenger
This would often go even further. As Mokyr writes of apprenticeship contracts: “In many cases he [The Apprentice] was socialized in other subjects, such as piety, literacy, and good manners.”
It is also why the Income Share Agreement (ISA) model for education practiced by some companies like Bloom Tech (formerly Lamda School) may be fundamentally flawed. As evidenced by the recent Consumer Financial Protection Bureau ruling.
US unemployment numbers are based entirely on survey results for those “seeking work” so although it’s not perfect this should account for those who are in full time education or otherwise not in need of a job.
其中一个原因可能是中国特色的 996 工作日和躺平的现象。
Disclosure: Long SHOP
Great essay on a wonderful topic
I'm surprised at two things I didn't see:
1) Family apprenticeships. Many trades ran in the family. This makes financial incentives strongly aligned. Training often begins at birth. This still exists in some rare niches; certain crafts, performers.
2) Schooling as the destruction of apprenticeship. Besides for the preference for propositional knowledge you mentioned, there's also the simple fact that schools make it hard for kids to work. A thirteen-year-old is unable to spend their best hours in an apprenticeship learning --- they're stuck in a classroom. I wonder how much a) child labor laws and b) compulsory schooling are correlated with the decline of apprenticeships.
I'm looking forward to reading much more from you! This is really excellent!