TLDR: William Deming (1950)

If you are referring to the perception that Japanese people are unbelievably devoted to their narrow job descriptions and take pride in that. You'd be surprised to know it's a very recent perception and invention. At heart, East Asian society has always trumpeted being an all-round genius, which gives the Chinese proverb (here in Japanese) 文武兼備 (double proficiency in words and war), and traditional samurai heroes like Musashi Miyamoto were known for being both unparalleled swordsmen and profound philosophers. You have Kaibara Ekken, the Edo period's Dutch medical/philosophy/botanical genius, who was also a wandering ronin in his early days. The pillars of Juntendo, the Dutch medical training institute in Nagasaki, were covered in lacerations because medical students simultaneously trained in sword arts while being doctors. Even in the 1930s, Japan was dominated not by specialised industrial titans, but gigantic vertical conglomerates called zaibatsu with integrated supply chains that did everything from financial services to mining to industrial manufacturing. Even the so-called 'new' zaibatsu like Nissan, which were perceived to be very specialised compared to the 'old' zaibatsu like Mitsubishi, still did everything from rubber plantations to cars. In that era, industrial communication was very much one-sided. Workers reported to managers, who reported to managers above them, but in no way were managers expected or incentivised to listen to what the workers had to say.

In the 1920s, Walter Shewart at Bell Technologies was starting to adopt applied statistical methods to mechanical processes. The idea is that a well controlled process should have remarkably little variation from some base level of growth. If you have periods of high and low quality, then some part of the process is not well-controlled, and this leads to inefficiency. In the case of Bell Technologies, the goal was to achieve 100% connectivity, with no disconnects or fluctuations in the telecommunications infrastructure.

Stewart's disciple, William Deming realised that this method could be generalised to a broader principle of quality control, rather than just a mechanical one. The policy recommendation was to actively track all significant deviations and immediately implement small improvements at every level. It was argued that ensuring quality would increase the reliability of products and boost demand, while ensuring efficiency would lower costs. In combination, this would lead to lower costs for everyone involved. Deming's ideas were not controversial, but American businesses perceived the hassle of implementing such a micro-managed system to be unnecessary. The dominant model of the time was the Ford assembly line, which revered a master class of engineer-scientists while the assembly line workers themselves were specifically meant to be replaceable and homogenous, concentrating costs where they were deemed to be most efficient. The standardised assembly line was considered the most efficient model available at the time, and in many ways it was, dramatically lowering implementation costs and reducing the haphazard effects of skill variance.

Post-war Japan though, was willing to give it a try. As part of the Marshall Plan, various management consultants were brought in to try and turn the ravaged economy around. Japanese engineers and scientists were introduced to his ideas by American engineer Henrh Sarasohn, who then lobbied hard to bring Deming in and influenced management across the country to attend his lecture series. Consider how crazy it would have been to convince all the business leaders of Japan to gather on the foothills of Mt Fuji in Hakone for a week to attend lectures by some random American guy immediately after the Americans bombed the country to hell and back.

They succeeded. Industrial companies started the adoption of these principles in 1950, and by 1955 had developed specific departments designed to further optimise them. This led to the birth of the Kaizen system, popularised by Toyota, which initially emphasised a four step Plan -> Do -> Check -> Act process implemented at every level from the janitor to the CEO. Every person was incentivised to improve his or her department's own process, typically through monetary incentive and promotions, but over time through social sanction and peer approval. This meant a radical overhaul of the manager's role, no longer was he there to dictate policy, but rather to facilitate dialogue, exchange of ideas, and implement new suggestions. The worker's role was not just to do his job, but also to optimise, improve, and philosophise it. This required a level of pride and specialisation that Japanese society had not yet seen, the kind of pride that would allow a sales clerk to tell a CEO that a company's management policy needed to be changed, and the CEO would listen. Famously, the kanban system of inventory management was developed by sales clerks at supermarkets.

Over time, this was supplemented with all kinds of new models like flow kaizen, geinichi genbutsu, and just in time processes (just Google them, not going to explain what they are). The first results came around 1960, when it emerged that with comparable capital stock per worker and broadly identical production technologies, the output per effective worker in Japan was already double that of the US. By the 1970s, Japanese car companies had basically completely displaced almost every domestic brand in the US and had become famous for quality and safety (but not really performance). In contrast, the Ford Pinto was prone to explosions, which the engineers knew about but management did not care about; and the Chevrolet Vega was prone to rust because materials engineers did not communicate with designers.

So we don't jerk them off too much, it should be noted that the Kaizen system has all but been adopted by every country in the world, and Japan no longer has much real advantage in efficiency, though the cultural perception remains. If anything, Japanese specialisation and productivity has gone backwards, with significantly reduced employer loyalty and increasingly inefficient processes owing to persistent miscommunication. Countries making great strides in productivity gains, like China, are typically doing so by rewarding technological innovation and implementation over mechanical and management optimisation. They are willing to sacrifice some level of individual efficiency if it means constantly being at the forefront of the technological frontier. If Japan is any indication, we could well see the birth of some kind of technology cult in Chinese culture over the next few years.