Letters Blogatory Opposes Abolition of the Leap Second

Readers, I am going to subject you today to an editorial on a topic that periodically gets me worked up even though it has nothing to do with private international law: the proposed abolition of the leap second. I’m sorry to say that the United States is the main proponent of the change. Here is the story.

What is a Leap Second?

Today, the world’s time standard is coordinated universal time, or UTC. Why UTC instead of CUT? Well, in French, it’s temps universel coordonné, which could be abbreviated TUC. UTC, which stands for neither the French name for the standard nor the English name, was apparently a compromise between the Francophone world and the Anglophone world.

The Royal Observatory, Greenwich
The Royal Observatory, Greenwich. Credit: Shizhao
The old way of keeping time, known as Universal Time, or UT (the English won that round!), was based on astronomical observations of the sun. We observed the length of the mean solar day, and then divided that day into hours, minutes, and seconds. So the second was in a sense defined to be 1/86,400 of the solar day and depended on the period of the earth’s rotation about its axis. To be sure, the solar day is not the only natural way to measure time: we could also measure time by the length of the year—by the period of the Earth’s revolution around the Sun rather than its rotation about its own axis. Astronomers and other scientists made use of this so-called ephemeris time because they knew that the length of the solar day was not perfectly uniform: the earth’s rotation is slowing down over time. And they had a system of converting UT to ephemeris time, or in other words, a system of tying the natural length of the day to the natural length of the year.

US Naval Observatory atomic clock
The master atomic clock ensemble at the US Naval Observatory, Washington, D.C. Credit: Coenen.
But with the development of highly regular atomic clocks in the 1950s, we began to think of the second not as a subdivision of a fundamental unit of time like a day or a year, but as the fundamental unit of time in its own right, and to define the second with reference to the number of “ticks” of the atomic clock. And since the period of the Earth’s rotation was not perfectly uniform, it became necessary to insert occasional “leap seconds” into UTC to keep the official time in line with the natural period of the day. (The more familiar leap year serves a similar function, keeping the official time in line with the natural period of the year). There have been 25 leap seconds since 1972, and thus UTC is now 25 seconds behind the time scale that relies only on the ticking of the atomic clock without reference to days and years, called International Atomic Time, or TAI (the French must have won that round).

Why Do Some People Want to Get Rid of the Leap Second?

GPS Satellite In Orbit
Artistic Impression of a GPS Satellite In Orbit. Credit: NASA
Now that computers rely on highly precise time synchronization throughout the Internet and where airplanes, ships, and cars rely on highly precise measurements of time to determine their position using GPS, there is a hue and cry each time a leap second is declared (the next leap second will be in June 2015). A group called the Civil GPS Interface Committee has declared the leap second a hazard to navigation. IT people seem to be unable to ensure that their computers can properly account for leap seconds, leading to computer glitches and crashes. And so the United States is proposing that the leap second be abolished. The effect of this proposal would be to remove the link in principle between days and years on the one hand and the inexorable ticking of the atomic clock on the other. To be sure, because leap seconds are so rare, in practice time as measured by atomic clocks will be more or less the same as time measured by the sun or the stars for many, many years to come.

Why Should You Care?

Days and years are meaningful for human beings. We live our lives in days and years. In my view we should take the meaningful units of measurement as the fundamental units of measurement, and then subdivide as appropriate. The SI definition of the second, useful in atomic clocks, “the duration of 9192631770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom,” is meaningless for human beings. It might just as well be 9,192,631,771 or 9,192,631,772, or 10,000,000,000 periods. And even granted that we can measure the absolute interval of time between two events with such precision, it’s meaningless, on a human scale, to do so. I see no reason why we should move from a standard of time that is meaningful to a standard of time that is meaningless, especially because over time the result will be that the natural, meaningful definition of time we all use everyday will become progressively more “wrong.”

Well, what about high-speed computers and other technology that require a very high degree of precision and synchronization? It’s a technical problem that the people who have a vested interest in ultra-precise time should bear the cost of fixing. And while I’m no expert, it seems to me there is a pretty easy answer: let the airlines and the Internet companies use TAI. In other words, let them simply stop adjusting for for leap secondsuse a time scale that does not include leap seconds, while the rest of the world continues to use UTC as it is now defined. Let the atomic clocks TAI become progressively more wrongout of sync with official civil time.

It seems to me we have our priorities backwards when we worry about how to keep our everyday natural time in sync with our atomic clocks instead of vice versa!

Further Reading

I had thought about quoting at length Hannah Arendt’s prologue to The Human Condition, a great book, which at the dawn of the space age expressed similar thoughts about cutting the necessary link between humankind and the earth (she didn’t have the measurement of time in mind). But instead of quoting it, I will just say: go read it! It’s a great book.

17 responses to “Letters Blogatory Opposes Abolition of the Leap Second”

  1. Interested readers can delve further into this subject through the presentations and preprints of two international workshops that were organized on this topic in 2011 and 2013: http://futureofutc.org

    1. Rob, thank you for the reference to the workshop materials!

      1. The detailed transcriptions of workshop discussions also provide insight into the many overlapping layers of technical, legal, historical and social issues pertinent to leap seconds and to timekeeping in general, for example:


        1. Thanks, Rob. The mention of zmanim was particularly interesting to me. I had considered including in my post a discussion of religious calendrical calculations, using the example of the observation of the new moon in the Jewish past as described in chapters 1 and 2 of the tractate Rosh Hashana. The text describes a whole system of observations of the new moon by laypeople, who would travel to Jerusalem and make their reports to a court, which would then declare the beginning of the new month. The reference seemed too parochial, so I left it out. But the emphasis on the importance of accurate measurement of time based on observations of nature (in that case, of the moon rather than the sun or the stars) seems to me to go to something deep in people of many cultures, whether or not religious.

          Another way, maybe, to put this. Everyone knows that in the very long run time has to be conceptualized linearly. After a certain time the Sun won’t exist. After a certain time, the position of the stars as we observe them will change. And I suppose that when very short times are the focus, time also has to be conceptualized linearly. But on human time scales, time, as measured in days or years, is cyclical, because the sun rises and sets every day and because it returns to its place in the sky every year. How, if at all, is this understanding reflected in the work of the expert groups considering the issue today?

  2. Gerry Ashton

    I think it will be quite difficult to keep TAI in a walled garden. Technical, legal, and personal matters will be intersecting more, not less, in the future. But I also feel it just isn’t right to divorce time from the Sun.

    1. Gerry, thank you for the comment. I wonder whether decisionmaking groups composed of technical experts, engineers, and representatives of national governments have the time of day (as it were) for the very natural thought that “it just isn’t right to divorce time from the Sun.” How do such thoughts even get evaluated in a technical discussion among experts focused on things like internet synchronization and hazards to navigation?

      1. I have seen that the folks making the proposal understand the consequences of abandoning leap seconds, but many of them are constrained not to speak anything contrary to the policies of the agencies and national governments under whose auspices they work. Furthermore, the culture of the ITU-R is one of diplomats and bureaucrats operating in camera, so ironically the folks who make things happen at the International Telecommunications Union are not good at communicating plainly and openly.

        1. Thanks, Steve, for the comment. I have attended meetings of technical delegations from national governments in my own field as an invited observer, and I agree with you that there’s little plain talk, let alone poetry, at such gatherings. But time is one of the subjects that calls for plain talk and maybe even some poetry.

  3. In response to a comment on the LEAPSECS mailing list, where real experts discuss these things and where there seems to be some consternation that a lawyer would jump into the conversation, I’ve tried to clarify what I think it would mean for people who need really precise measurements of time to use TAI. I don’t mean to suggest that TAI should be redefined or that UTC should be redefined. I just mean that those who need a timescale that doesn’t contain leap seconds could use TAI for their own purposes, but that UTC would continue as it is now and that UTC would be the official civil time.

    “But it’s crazy to say that the internet and navigation systems should use their own non-official time!” Maybe, but not crazier in my view than defining official time in a way that lets it drift away from solar time. And anyway, the people who write the software and hardware that has trouble with leap seconds have it in their power to avoid the result if they want by fixing the software or hardware problems that make leap seconds difficult to deal with.

    By the way, if you want to see the LEAPSECS list discussion, it’s here. My favorite comment: “Yeah, he’s entitled to his opinion, just like everybody else, but he doesn’t have any special standing for his opinion, which as others have pointed out, interfaces badly with reality.”

  4. Nero Imhard

    Perhaps even more importantly, there’s an element of betrayal in fundamentaly changing an existing definition. It breaks your ability to meaningfully choose the right time scale for your purpose.

    1. Thank you, Nero, for the comment! I am not an expert (as should be apparent) and am not sure I understand why what you say is so. Can you explain?

      1. For example, astronomers and astronautical engineers have made the natural assumption throughout their software and systems that Coordinated Universal Time is indeed a flavor of Universal Time, that is, an approximation to Greenwich Mean Time. If UTC is redefined to no longer reflect Earth orientation then a lot of our software will break. The software is not at fault, the naive standards process is.

        Also see: http://www.cacr.caltech.edu/futureofutc/aas223/presentations/2-1-ISOterminologyAAS.pdf

        1. Thanks Rob. Does this suggest that eliminating the leap second is just trading one set of technical problems (software that cannot handle discontinuities, for example) for another (software that assumes that UTC approximates UT)?

          1. Hi Ted. Yes, that would be my understanding. It might also be worth pointing out that redefining UTC would not improve access to unsegmented timescales (that is, without leaps) since projects and systems can already use TAI, GPS and related timescales. Rather, all it would accomplish is to eliminate straightforward access to mean solar time. As a result confusion would increase, not decrease.

  5. […] the fabulous Spanish language spin-off blog, Cartas Blogatorias, or frolics and detours like my editorial in favor of the leap second or my annual April Fool’s post, please take a minute and nominate […]

  6. […] way of thinking. Can I stretch a bit and compare my thought on this to what I’ve written previously about the leap second? As you probably know, a second, traditionally, was defined as 1/86,400 of a day (there are 24 […]

Leave a Reply

Your email address will not be published. Required fields are marked *

Thank you for commenting! By submitting a comment, you agree that we can retain your name, your email address, your IP address, and the text of your comment, in order to publish your name and comment on Letters Blogatory, to allow our antispam software to operate, and to ensure compliance with our rules against impersonating other commenters.