Epoch and Epoche.

I’m still reading George Sand’s Spiridion (see this post), and a while back it became clear to me that it was that rarity among novels, one centered on intellectual striving and drama (there’s no love interest, not even a female character so far). The focus is the struggle between blind faith (of the Catholic variety) and a confrontation with pagan philosophy, Reformed Christianity, and even atheism; it takes place in an Italian monastery, mainly in the 18th century, and Sand makes the agony of the struggle convincing even to someone as far removed from it as most of us are in the 21st century. The title character, the late-17th-century monk who founded the monastery, was baptized by Bossuet himself, but found himself assailed by doubts, and the main narrator, a century later, says “Quelle situation terrible était donc la mienne! Au dix-huitième siècle j’avais été élevé dans le catholicisme du moyen âge; à vingt-cinq ans j’étais presque aussi ignorant de l’antiquité qu’un moine mendiant du onzième siècle. C’est du sein de ces ténèbres que j’avais voulu tout à coup embrasser d’un coup d’œil et l’avenir et le passé.” [What a terrible situation I was in! In the eighteenth century I had been brought up in the Catholicism of the Middle Ages; at twenty-five I was almost as ignorant of antiquity as a mendicant monk of the eleventh century. It is from the midst of those shadows that I had wanted to take in at a single glance both the future and the past.]

This was on my mind when I glanced over to the sideboard to my left (piled high with books that don’t fit on the shelves and that I am seriously intending to read in the foreseeable future) and noticed a book some generous soul had sent me for my birthday last July (thank you, whoever you are!), one I had still not gotten around to: The Crisis of the European Mind: 1680-1715, by Paul Hazard. It occurred to me that it bore directly on the themes of the novel, and when I opened it and saw on the first page of the preface “One day, the French people, almost to a man, were thinking like Bossuet. The day after, they were thinking like Voltaire,” I knew I had to read it, which I am now doing. (I’m always reading at least three books at any given time.)

Early on in the book, Hazard mentions “the word Pyrrhonism, which had created such a commotion in Pascal’s bosom,” so of course I had to look up Pyrrhonism, and that Wikipedia article told me that “Pyrrhonian skeptics … inferred the need for total suspension of judgment (epoché) on things.” I clicked on the latter link and found that “Epoché (ἐποχή, epokhē “suspension”) is an ancient Greek term which, in its philosophical usage, describes the theoretical moment where all judgments about the existence of the external world, and consequently all action in the world, is suspended.” A nice word, I thought, but I wasn’t clear on how it came to mean that; a glance at my pocket Greek dictionary reminded me that ἐπέχω (epekhō), the verb it’s based on, is one of those confusingly multivalent words (reminiscent of Irish cur): the pocket dictionary says “to have or hold upon; to place upon; to hold out, present, offer; to have opposite oneself; to keep shut; to hold fast, retain, hinder; to delay, retard; to reach to, extend; to have in one’s power.” It’s that “delay, retard” sense that’s involved here–”hold up” would be a helpful way to render it. So much for epochē, but what about epoch, which was obviously another derivative of the same Greek word? The OED explains:

Etymology: < late Latin epocha, < Greek ἐποχή stoppage, station, position (of a planet), fixed point of time, < ἐπέχειν to arrest, stop, take up a position, < ἐπί + ἔχειν to hold. Compare French époque, Italian epoca.

I. A fixed point in the reckoning of time.

1. Chronol. The initial point assumed in a system of chronology; e.g. the date of the birth of Christ, of the Hegira, of the foundation of Rome, etc.; an era n. Also, in wider sense, any date from which succeeding years are numbered. Now rare.

2. a. The beginning of a ‘new era’ or distinctive period in the history of mankind, a country, an individual, a science, etc. Phr., to make an epoch.

b. The date of origin of a state of things, an institution, fashion, etc.; occasionally, an event marking such a date. Obs.

3. In wider sense: A fixed point of time.

a. The date, or assigned position in chronological sequence, of a historical event.

b. [= French époque.] A precise date; the exact time at which an event takes place or is appointed to take place. Formerly gen.; now only with reference to natural phenomena (cf. 4).

c. A point of time defined by the occurrence of particular events or the existence of a particular state of things; a ‘moment’ in the history of anything.

4. Astron. The point of time at which any phenomenon takes place; an arbitrarily fixed date (often the first day of a century or half-century) for which the elements necessary for computing the place of a heavenly body are tabulated. Also, the heliocentric longitude of a planet at such a date (more fully, the longitude of the epoch).

II. A period of time. (Cf. similar use of era, term).

5. a. In early use, a chronological period dated from an ‘epoch’ in sense 1. In later use, a period of history defined by the prevalence of some particular state of things, by a connected series of events, or by the influence of some eminent person or group of persons.

b. A period in an individual’s life, or in the history of any continuous process.

c. Geol. A period or division of the history of the formation of the earth’s crust.

A complicated word!


  1. Sometimes words are mere labels that don’t mean what they say. (An oncologist to a friend of mine: “Just because the tumour is benign doesn’t mean it won’t kill you”.) Sometimes a word is genuinely descriptive, albeit some times as a metaphor e.g. malignant tumour.

    Given these oddities I sometimes wonder whether knowing the etymology of a word of the first class is any help at all.

  2. ‘Skepticism: An Anthology’ eds. Richard Popkin and José R. Maia Neto does an excellent job of explaining the origins of Phyrronism. Popkin also translated parts of Le Dictionnaire historique et critique, by Pierre Bayle, a book Sand read, and which Balzac, with typical fanaticism, considered practically sacred.

    Bayle was a lunatic. His book is genius. From Wiki.fr:

    Ce Dictionnaire se veut, en première intention, la correction des erreurs des auteurs des dictionnaires précédents (en particulier Louis Moréri). Mais Bayle précise son projet dans la préface:

    “Or voici de quelle manière j’ai changé mon plan, pour tâcher d’attraper mieux le goût du public. J’ai divisé ma composition en deux parties l’une est purement historique, un narré succinct des faits l’autre est un grand commentaire, un mélange de preuves et de discussions où je fais entrer la censure de plusieurs fautes, et quelquefois même une tirade de réflexions philosophiques en un mot assez de variété pour pouvoir croire que par un endroit ou par un autre chaque espèce de lecteur trouvera ce qui l’accommode1.”

  3. I’ve just gotten to the chapter on Bayle in the Hazard book, and now I’m looking forward to it even more!

  4. Trond Engen says:

    A woman’s order consists of organizing all objects into rectangular staples and hiding them in drawers where nobody can see them. A man’s order consists in having, from a given point, e.g. his best chair, all of his belongings in sight, and being immediately able to point any one of them out.

    Petter Wessel Zapffe, philosopher, in one of his lighter works (quoted from memory, translated ad hoc)

  5. John Emerson says:

    This is the definitive Popkin. It was the third of three editions; each edition has a different subtitle and greater scope:


    I also recommend his Bayle translation, which is full of all kinds of things, from abstruse information on unknown people to what amount to jokes to deadpan satire to serious discussions of major topics. About 400-500 pages; the whole dictionary was thousands of pages long.

  6. John Emerson says:

    Bayle was a major source for Voltaire and the others, but all the evidence is that he was some kind of Calvinist. His writing is deadpan and he never tips his hand. One kind of Pyrrhonian doubted all science and factual knowledge, but did not question the truths of faith, and that may have been Bayle’s position. But Popkin confessed bafflement.

  7. The German word Epoché (written just so) is a central notion in Husserl’s phenomenology, designating a technique also called Einklammerung (bracketing). The SEP article on Husserl says this about that:

    Husserl developed the method of epoché or “bracketing” around 1906. It may be regarded as a radicalization of the methodological constraint, already to be found in Logical Investigations, that any phenomenological description proper is to be performed from a first person point of view, so as to ensure that the respective item is described exactly as is experienced, or intended, by the subject. …

  8. The OED’s judgment Now rare on sense 1 is obsolete, as computer architects have revived it: see Wikipedia for a list of different computer epochs. The most widespread is probably “the Unix epoch”, January 1, 1970 at midnight Universal Time; Windows systems use January 1, 1980 at midnight local civil time. Perhaps the oddest is “the VMS epoch”, November 17, 1858, also at midnight Universal Time, which is halfway through Julian Day 2,400,000; the Julian day count in turn begins at noon on January 1, 4713 B.C.E. (by the Julian calendar). Since Ptolemy’s day, astronomers used dating systems beginning at noon, so as not to break up a night’s observation into two different calendar dates, but mostly abandoned that system in 1925.

  9. I should add that the plain ol’ German Epoche (stress on first syllable) means “epoch”.

  10. The OED’s first definition for “epoch” (“[t]he initial point assumed in a system of chronology”) is no longer “rare”. It’s widely used in software circles to refer to the timing system used by computers. For example the UNIX epoch occurred on January 1st, 1970.

  11. German Epoche (stress on second syllable) .

    dw: a google search for “unix epoch occurred” yielded 1200 hits, a search for “unix epoch began” 53200 hits. Without citing a source, the WiPe article on Epoch (reference date) begins thus:

    In the fields of chronology and periodization, an epoch is an instant in time chosen as the origin of a particular era. The “epoch” then serves as a reference point from which time is measured. Time measurement units are counted from the epoch so that the date and time of events can be specified unambiguously.

    I don’t think I have encountered “epoch” in IT in the sense of “instant” – but I’m not a Unix user, and in Windows / Java development here in Germany neither “epoch” nor Epoche turns up frequently in that sense. The new(ish) Java jodatime classes include many with “Instant” in their name, but not “Epoch”. The Javadoc of org.joda.time.DateTime does say this:

    Internally, the class holds two pieces of data. Firstly, it holds the datetime as milliseconds from the Java epoch of 1970-01-01T00:00:00Z

    So it may just be that mere users of IT time classes, such as myself, tend to think “period of time” when they read “epoch”, although this is apparently a misprision of a technical term in IT chronology. I wonder why whoever-it-were burdened us with this revived rare sense of “epoch”. Were they ignernt, or smart-alecks, or both ?

  12. .. neither “epoch” nor Epoche turns up much at all

  13. @Stu: That (to me) inexplicable (or at least unexplained (to me)) Ghit problem strikes again: if you click through, you see that “unix epoch occurred” actually yields only 15 hits and “unix epoch began” just 64 hits.
    Does anyone know why this happens?

  14. PaulB: At the bottom of the last page, after “clicking through”, the following text appears (in German, I can’t get it to show English): “To provide you with only the most relevant results, some results have been omitted that are very similar to the X (many) that you see. You can repeat the search to obtain all results.”

    The part of that last sentence beginning with “repeat” is a link that repeats the search. For instance, the “1200 hits” for “unix epoch occurred” initially showed only 2 pages to click through at the bottom. On repeating the search to get all results, I got 6 pages to click through.

    There’s another issue, too: the differences in number of hits reported by people is (I think) affected partly by browser settings such as “preferred languages to display” (or whatever the English is) in Firefox

  15. I’ll note that the whole subject of ‘time reference’ is a truly epic mess– this is the short version:


    For the long version, see Chapter 3 of


  16. I’m unclear on what distinction the OED is trying to draw between definitions 1 and 4. I work with astronomical data all the time, and the definition 1 matches perfectly with how I use the astronomical epoch data that I see. The primary use is as a dating standard; I seem plots all the time with body positions or measured radiation fluxes given as a function of days (or seconds), with some reference to the epoch, which just means the zero-point of that particular time measures. I have no particular interest in the planetary ephemeres mentioned in definition 4, however, so maybe somebody who worked directly with that kind of data would feel that having a whole separate sense for 4 is justified.

    Regarding the UNIX epoch, it’s no surprise that “UNIX epoch began” should be a common way or describing the origin date. I don’t think most programmers I know would think of the epoch as a single date, but rather as a meaning falling under the heading II. However, probably because of my experience with the astronomical data, I would have no trouble understanding “UNIX epoch occurred,” although I wouldn’t phrase it that way myself. That makes the UNIX epoch sound like an event when something happened. Thinking of it, as I do, as simply the zero point of a time coordinate, I would probably say the “UNIX epoch was” to reference 1/1/70 00:00:00 CUT.

  17. “Epoche” sounds a lot like the appraoch that informs the Madhyamika systems.

    “one of those confusingly multivalent words (reminiscent of Irish cur):”

    About like Englsih “get”.

    “the pocket dictionary says “to have or hold upon; to place upon; to hold out, present, offer; to have opposite oneself; to keep shut; to hold fast, retain, hinder; to delay, retard; to reach to, extend; to have in one’s power.”

    A lot like Englsih “seize”.

    Americanists have an efficient way of rendering these supposedly multivalent wards – they just reduce them to whatever semantic load they actually have, which can be pretty abstract, and then go with that. So you get verbs that mean things like “Cross a boundary between liquid and dry” because they show up in strings that denote “come ashore”, “wash up on shore” “push off from the beach”, “fall into the water”, “add to a pot of soup” etc.

    Again, rather like English “get” which can mean “come into the posession of..”, “Take posession of…”, “become X” and really just means “change of state”, kind of the aspectual equivalent of a modal.

  18. The modern, “external” world, so far removed from the epochal interiority of the Greek philosophical world view, renders this highly conceptual word nearly unintelligible in a 21st century born to Voltairean skepticism. Witness the IT types and their stumbling to ascertain its meaning through solely observable phenomenon. The female muses always knew better. It was the advent of the a-mused male which heralded the epoch of disbelief.

  19. @John Cowan: sorry to repeat the content of your comment — I didn’t see it when I posted.

  20. @Stu Clayton:

    I did a Google search for “unix epoch” (with quotes). The results are approximately evenly split between the OED’s senses 1 and 5. This presumably recapitulates the semantic development that resulted in sense 5 developing from sense 1 in the first place.

  21. To clarify my previous comment: I looked at the first page of Google hits only.

  22. I wonder why whoever-it-were burdened us with this revived rare sense of “epoch”. Were they ignernt, or smart-alecks, or both?

    Definitely smart-alecks. Specifically, they were the Zen Patriarchs of the Bell Labs dharma line.

    All operating systems older than Unix, to my knowledge, exposed only local time to programmers, and in broken down format (year, month, day, hour, minute, second, subsecond). Unix was the first to expose a count of seconds since a particular epoch, and to define that epoch using Universal Time. (The earliest versions had a different epoch and only a 16-bit time count, which therefore reset to zero about every 2.5 years.)

    The advantage of the Unix system was that timestamps on different computers could be compared by mere subtraction, and without having to take time zones into account. Someone at the Labs, a place where polymaths tended to congregate, undoubtedly knew the astronomical and historical uses of epoch and started to use it. It then spread via Sun Microsystems to Java, even Java running on Windows.

    The part of that last sentence beginning with “repeat” is a link that repeats the search.

    Useful tip from an Xoogler: You don’t actually need to go to the last page to do a complete search. Just display the first page and click on your browser’s address bar, which will look something like “https://www.google.com/search?q=…” Then insert the string “filter=0&” after the “?” and press Enter, and you’ll get an unfiltered search.

    Americanists have an efficient way of rendering these supposedly multivalent wards – they just reduce them to whatever semantic load they actually have, which can be pretty abstract, and then go with that.

    You can do that with English words too. For example, climb in some situations allows movement downwards, when human or animal limbs are involved; in other cases, like the movement of airplanes, it excludes downward movement. This can be captured by defining climb as ‘clamber or ascend’.

    I would probably say the “UNIX epoch was” to reference 1/1/70 00:00:00 CUT.

    Well, sort of. For one thing, UTC doesn’t extend back beyond 1972, when it was defined as TAI – 10. For another, Unix time ignores leap seconds and fractions thereof. So in effect the Unix epoch is 1/1/70 00:00:10 TAI, but the count of seconds since then ignores all pre-UTC and UTC adjustments to TAI. The Network Time Protocol literally does tick two consecutive seconds the same way over a leap second (though it provides an out-of-band flag).

  23. Both TOPS-10 and TENEX counted time from November 18, 1858 GMT (Smithsonian Astronomical Date Standard). From the UUO / JSYS, the LH was days and the RH was fraction of a day in the former case and seconds in the latter (IIRC).

  24. Ah, skewered by the MMcM polymathy!

    In my defense, I had little to do with Tens and Twenties, and much to do with Eights and Elevens.

Speak Your Mind