Moore’s Law: How Its Chinese Translation Overstates the Claim

After nearly forty years in IT, I have never liked the Chinese term 摩尔定律, usually used for “Moore’s Law”. Oddly enough, after moving to Australia and reading more technical material in English, I became somewhat more comfortable with the English name. The reason is that “law” in English does not quite carry the same force as 定律 in Chinese.

What we call Moore’s Law began as Gordon Moore’s observation of, and prediction about, the development of integrated circuits in 1965. He projected that the number of components on an integrated circuit would continue to double roughly every year. In 1975, he revised the interval to about two years. The idea later came to be commonly summarised as the number of transistors on a chip doubling approximately every two years.

This is plainly not a scientific law in the strict sense. It does not have the universality or necessity of a natural law. The trend depends on manufacturing processes, physical constraints, economic costs, market demand and investment across the industry. It may hold over a particular historical period, but it can also slow, change or come to an end as those conditions change.

The distinction I want to make is between 定律, a scientific law, and 规律, a regularity or pattern.

A regularity describes a relatively stable relationship in how things develop and change. It can depend on particular conditions and periods of history. A scientific law is generally a proposition with clearly defined conditions of application, systematic verification and a stronger claim to generality. The two are not wholly opposed, but neither should they be used interchangeably.

The English word “law” has a wider range of uses and can sometimes name an empirical regularity. “Moore’s Law” may still overstate the concept, but it has become the accepted name in the semiconductor industry. Using it does not necessarily mean treating the trend as a strict law of natural science.

The Chinese translation, to my ear, reinforces that misunderstanding. 摩尔定律 sounds like a technically established rule that has been scientifically proved and must hold. That makes it easier to forget that the original claim was an empirical prediction made under particular historical conditions.

There is another point worth noticing. One reason this trend continued to be borne out for so long is that the semiconductor industry adopted it as a development target. Research investment, industrial planning and technological competition all helped bring about the progress it predicted. It was therefore both an observation drawn from experience and an expectation shaped by what the industry chose to do.

For that reason, I prefer 摩尔规律, or “Moore’s regularity”, to 摩尔定律.

An English name that was already somewhat imprecise acquired an even stronger suggestion of scientific certainty in Chinese. This may look like a small choice of words, but it affects how we distinguish an empirical regularity, a scientific law and a trend in technological development. Translating a technical concept should do more than match words on the surface. It should, as far as possible, preserve the kind of claim the concept was meant to make.