Like many English-speakers, I hesitate when faced with the necessity of discussing a single one of those dotted things that usually come in pairs; I say “die,” but I feel funny about it. Jonathon at Arrant Pedantry has a nice post on the topic, explaining how the plural -s went from voiceless to voiced but “remained voiceless in dice. Why?”
Well, apparently because people had stopped thinking of it as a plural and started thinking of it as a mass noun, much like corn and rice, so they stopped seeing the s sound on the end as the plural marker and started perceiving it as simply part of the word. Singular dice can be found back to the late 1300s, and when the sound change came along in the 1500s and voiced most plural –s endings, dice was left behind, with its spelling altered to show that it was unequivocally voiceless.
He discusses truce, bodice, pence, and other words, and in the comment thread John Cowan quotes a great Ambrose Bierce definition. Check it out. (Via Stan Carey’s Link love: language (41), which links to lots of other great stuff as well.)
I too found that page via Stan’s link love, and enjoyed the historical background. It is, or at least was, a non-issue here though, because in the NZE I grew up with “dice” was always the singular term too. To say “die” was a dead giveaway that the speaker was either American, or affecting an Americanism, since the only time the word was heard was on TV, in American shows. Whether die as the singular of dice has made inroads into current NZE I don’t know.
One die two dice is what I naturally say; one dice sounds really funny to me (central Mass., late 40s). When I was a kid I played a lot of strategy games and roleplaying games where the number of dice you rolled varied with the situation in the game, and I wonder if that has anything to do with it. Outside of that context — which excludes many popular games (e.g., Monopoly) where you always use the same number of dice — specifying the number of dice you need is not exactly a pressing concern for most people and might never come up in conversation.
Interesting article, and that looks like it might be a fun blog to explore.
Thanks for the link! I’m glad you enjoyed it.
Me too. I had never read your blog, but will do so again.
Dice is still dominant in NZ for the singular – die is only used by pedants.
I’d guess that most of us (except Ken) have very little experience with one die/dice, and so we have little reason to refer to just one. Most common games require two, Yahtzee requires five (?), and you would always hang two, never one, from your rear-view mirror. Think of the symbolism.
Question: what to tool-and-die people say. Presumably they say one die but is the plural /dais/ or /daiz/? It’s the latter, I suspect.
Singular die for me, and I’ve honestly never thought it was any stranger than mouse/mice. I didn’t realize it wasn’t common outside N. America.
Can’t speak for Ken but in the long-ago context in which I would have had pragmatic reasons to use a singular in conversation (Dungeons & Dragons, circa ’78-’81) there was a further ambiguity in that “dice” could be and were of various polyhedral shapes not limited to the standard cube. So “d6” (pron. dee-six) was in use to some extent as the unambiguous singular for the object under discussion here.
Oddly enough, I see from their website that the current proprietors of Cosmic Wimpout (sort of like Yahtzee-for-hippies? I played a bit in college) refer to the sets of five dice (with non-standard symbols on them) as “cubes” as if “dice” would be insufficiently specific.
Note, however, that if “dice” really were a standard mass noun there would presumably be a standard phrase usable for individuation, akin to “grain of sand” or “piece of furniture.” Plus, you can say in e.g. a Yahtzee context “five dice” where you can’t say “five scissors” or “five trousers” (or if you can, what is implied is five pairs, not two and a half pairs).
A Chancellor of the Exchequer was mocked a few years ago for mentioning, in his Budget speech, “one pence”.
What about foot/feet? When I was a etc etc, it was common to hear someone say “That’s four foot long” or “six inch” but never, that I remember, “four yard”. But you would hear “That’s a four mile walk”, a twenty acre field and so forth.
Well, it’s not “full fathoms five,” is it?
The Nine Inch Nails. – That was a pretext to insert this interesting but not exactly on-topic article about what it’s like being a freelance journalist at the London Daily Mail.
For me:
a four mile walk, not a four miles walk
but
I went four miles, not I went four mile
likewise
a six foot fence rather than a six feet fence
the tree grew three feet rather than the tree grew three foot
a forty acre field but not a fort acres field
he bought forty acres rather than he bought forty acre
and so on
Units of measure in the singular with a number is archaic, poetic, or obsolete. But when a noun modifies another noun, singularity is standard: we say doghouse, not dogs house, even if it houses several dogs.
Units of measure in the singular with a number is archaic, poetic, or obsolete.
That’s not right, John. For instance we construction folk do it (with -foot & -inch) all the time. And yes in the USA too. Not with metric, though; so perhaps it is on its way out.
Incidentally, it’s not originally singular—it’s the Old English genitive plural.
Is “yard” special? Soccer goals were “eight yards by eight foot”, in the lingo of my youth.
Maybe “foot” is special. I can’t see “six inch long” or “six yard long”, but I can see “six foot long”. Or “six foot wide” or “six foot tall”, for that matter.
“Six mile long” maybe, or maybe not.
I wonder if this way of using “foot” is at all influenced by the conveniently brief expression “six foot two”, “five foot three” etc., where the second number is inches.
The takeaway: Never say “die.”
Yes, maybe foot is special.
it’s the Old English genitive plural
Is that why you could say either “the nine-inch nails” or “the nails of nine inches in length”? Is that genitive? I’m getting confused.
Incidentally, it’s not originally singular—it’s the Old English genitive plural.
Hat, I’m having some pronoun trouble. What is not originally singular?
I say either for the singular, depending on the register and the audience. But I strongly prefer die. *”The dice is cast”? I think not.
The verb form indicates what is most natural in common or kitchen English: *”Die the turnips and add to the broth”? Nah. And “dice the turnips” can’t be from the plural form; we would not say: *”Cubes the turnips and add to the broth.” (SOED at “cube, v.”: “3 Cut into small cubes; dice. M20.”)
Part of what is going on seems to be that the niche /daj/ is already occupied by the verbs die and dye, and the noun dye, right?
I was playing D&D around the same time as J.W. Brewer and I recall the d6 usage as well, but more often I was playing the kind of strategy games that had a map, hundreds of little pieces (square-shaped pieces of cardboard that we called “counters”), and dozens of pages of rules. The dice in these games were almost always six-sided. I can’t verify it now but I’m sure the rules made the one die two dice distinction and we picked it up from there. With games like Monopoly or Yahtzee, there was an oral tradition: you learned how to play from other people and probably never read the rules. With strategy games, not only did you have to read the rules to learn how to play (because so few people played them) but you had to refer to them constantly (because they were often ridiculously complicated).
What is not originally singular?
The apparent singular in constructions like “five foot tall.”
Noetica: “dice the turnips” can’t be from the plural form; we would not say: *”Cubes the turnips and add to the broth.”
“Dice” and “cube” here are from verbs which are derived (by conversion, without other change) from the nouns. The verb “dice” ‘to cut into dice-like pieces’ is from a now plural noun (like “people”), rather than a plural form of a noun, since it has become unanalyzable for the majority of speakers, as detailed by several commenters above. “Cubes” could not be the base for a verb, since it is obviously a plural form, transparently analyzable as cube + -s. Only the basic, singular cube can be used in this role. Similarly, as someone wrote above, in compound formation the added (first) noun is the basic (singular) form, eg toothbrush not *teethbrush, or oatmeal not *oatsmeal. This is not a matter of logic (nobody brushes just one tooth) but of the morphological structure that English shares with its Germanic relatives (eg in German Zahnbrüste ‘toothbrush’, while “teeth” is Zähne).
Sure, Marie-Lucie. I find no discord between our views.
To people (LME, from French peupler) is interesting and perhaps harder to account for, since as a plural people is anomalous more in the manner of cattle, not in the manner of mice or of pease. (A cat does not mice, it mouses.)
SOED records a second verb die, from the early 18C: “Provide with a die; shape using a die.” Dieing is given for the participle.
Ah yes, those would be the Zähne that the Haifisch wears in his Gesicht.
As for people v., it is plainly derived from singular people n. ‘folk, populus‘.
it is plainly derived from singular people
OK. And that makes it fit in a less plain manner with the central case looked at here: dice. Note that to people does not mean “to supply with a people”, but “to supply with people”.
If we made a verb meaning “to supply with cattle”, would it be to cattle, to cow, to ox, or what? (To invaccinate ☺? Already taken, according to OED. To vacciplete.)
When you make an integrated circuit, you dice the silicon wafer with a dicer, and then each die is mounted on a lead frame and connected using a die bonder.
Hmm. Cowherd, cow-herder, etc.; but cowman is rare and cattleman is common. Cowboy, but *cattleboy?
The Diceman, not The Dieman. Also huntsman, not *huntman; batman and batsman have distinct meanings. But those are horse of different colours.
Noetica: I like your comment about to people.
In French, the word peuple, as in un peuple ‘a population, nation, ethnic group, etc’, le peuple ‘the population, especially the mass of the people as opposed to the upper classes’, are always singular. But in (at least British) English, collective nouns referring to some human groups, like family or committee can be used in a plural sense (as in The family were at church, an example I learned many years ago), and this is probably what happened with people, which went from a singular collective noun to a plural one, now on its way to replace persons, at least when no precise number is specified.
Two cows
Two cattle?
Two head of cattle?
When you playfully pull someone’s trousers down, you pants them, you don’t pant them.
Noetica: Cowboy, cattleman, etc: in such cases, once a word is formed and becomes current, there is no need to make up a rival one. But different countries or regions needing a word for the same thing might independently make up different words.
I don’t know what a batsman does. As for huntsman, I have no idea why the -s is there.
gunman
swordsman
The OED says that huntsman, batsman (the cricket equivalent of a batter), craftsman, tradesman, bondsman, frontiersman, helmsman, headsman (cf. headman), herdsman (formerly herdman, but changed when herd in the same sense was lost save in cowherd, shepherd, swineherd), kinsman, linesman (cf. lineman), marksman, oarsman, salesman, sportsman, statesman, swordsman, townsman, tribesman, clansman, woodsman, yachtsman are all derived from the genitive form of the first noun: thus huntsman < hunt’s man, and so on. Some of these alternate with non-genitive forms having the same meaning: bondman, oarman, townman. Since the ‘s genitive is now a clitic rather than an inflection (that umbrella is the young lady I go with’s, e.g.), such compounds can no longer be made.
Thanks for the research, JC. I thought of the genitive, since obviously huntsman could not include hunts as a plural. I guess the s-less words, such as fireman and policeman, are much more recent. Gunman is interesting, since (as far as I know) it means “man using a gun to commit a crimeof violence”, not “gun aficionado” or “gun dealer” which would be plausible meanings.
According to the OED, the original meaning of gunman was simply ‘person armed with or using a gun’ as opposed to a bow, sword, or pike: “not only a penman, but a gunman, a rodman, and a horseman” (1888). (Rodman ‘fisherman’, by the way, alternates with rodsman.) Not until the 20th century does the association with illegal violence become the dominant one, perhaps because every soldier was by then a gunman in the older sense. The OED records a unique use in 1881 of gunmen ‘gun makers’.
marksman: EtymOnlDic says this form with genitive -s is attested from 1600s while the form markman goes back a little further.
There must have been a period when both forms were in use, until one won out over the other. This often happens, especially over a period of one or two generations, since young and old coexist but do not speak in exactly the same way.
From people to populate:
For years I’ve been annoyed that in (now older versions) of Windows, the Control Panel ‘add/remove software’ function informs the user that “The list is being populated.” Why couldn’t they have used ‘compiled’?
Googling [ list populate ] returns 16 million hits. Languages evolve.
this form with genitive -s is attested from 1600s
Attested in the context of a dictionary simply means “occurring in a written document”. Some very common words are attested relatively late, because for a long time nobody bothered to used them in writing (or dared to do so: witness the “four-letter words”, some of them known to practically all speakers for centuries but until recently strongly discouraged in writing).
Nowadays there are so many written documents, and such efficient searching technology, that it is possible to find the exact day some newly formed word appeared in print, or even was recorded in speech (see how many instances of this are documented regularly on Language Log). The farther back we go in history, the fewer documents have survived, so the fact that a word’s first appearance in a surviving written document occurred in a certain year does not mean that this was the first time it was used, it just means that we can be certain that it was used in that particular year (or even day and month, for instance in a dated letter or will).
Populating a list.
Hah, the fact that dice as a fossilized plural has irregular voiceless /s/ never occurred to me before, even though it is obvious on two seconds of examination. (I have so far not really managed to acquire stable syllable-final voiced fricatives in my L2 idiolect of English.)
Hah, the fact that dice as a fossilized plural has irregular voiceless /s/ never occurred to me before, even though it is obvious on two seconds of examination.
That’s the kind of thing that makes linguistics so much fun!
Before I did, I simply mapped English ss/c to my /sː/. That completely prevented me from noticing the plural |z| in dice, too.
For me, the genitive singular is voiced die’s (daiz), and the genitive plural is voiceless dice’ (dais).
Other meanings of die (as well as the homophonous dye) have a standard voiced /z/ plural.
> people […] started thinking of it as a mass noun
This is really hard for me to grok. Why would people think of something that is so obviously discrete objects as a mass noun? I’ve been using the word with my 3-year-old recently and lots of opportunities to use it in unambiguously count-noun ways come up, like “you made a nice 4-sided die” or “how many dice do you have?”. Did anyone ever say “nice 4-sided piece of dice” or “how much dice”?
Cf. peas(e).
Pease is an interesting case, because there really are two physically different forms of pea matter. The individual seeds of the pea plant are clearly countable objects. However, the most traditional English was of eating peas is boiled and then squished into porridge (and then left for nine days, perhaps?); in this form, the use of a mass noun pease makes a lot of sense.
I like to imagine that the move from mass noun to count noun was caused by greater sophistication in how the British ate their legumes.
There’s also the opposite case: broccoli.
But food item nouns are really different because it’s common for them to be used in both mass and count ways. Would you like some more cucumbers? More pieces of cucumber? More cucumber?
Nothing looks more countable than an ear of maize, but in North America where we call it corn, we still maintain the mass nature of the noun from the inherited use of it for small grains such as wheat, rye, oats, and barley (no longer current here), and so we must say two ears of corn, not *two corns.
I would think that the unacceptable *two corns would be confusingly ambiguous. Why would it presumptively mean “two ears of corn” rather than “two kernels of corn”? I suppose it could sort itself out in time, the way we know that “two peas” doesn’t mean two pods-full. Another way of thinking about it is that the ear may seem the “natural” unit pre-harvest, but not subsequently, since on-the-cob has never been the primary way (for Anglophones at least) of consuming maize and most other uses (including grinding down into cornmeal etc) require the stripping of the kernels off the cobs as the next step after taking the ears off the stalk and husking them.
This is precisely the logic of classifier languages, where *two apples is too vague to be intelligible: you must say two round-things of apple, two slices of apple, two cups of apple, or whatever. Classifiers can become detached from their semantics: in Japanese we have two long-thin-things of book, just like two long-thin-things of beer, because books were originally scrolls and the use of the classifier persists even when essentially all books are now codices. (Tough talk in Japanese bars: “Gimme a scroll of beer.”) In Burmese, though, the classifiers are pretty much still fully semantic and reflect the local culture: two baskets of chicken, *two baskets of mosquito, *two baskets of cow because Burmese carry chickens in baskets but not mosquitoes, still less cows.
“Corns” is supposedly used in Hawaiian Creole English (Pidgin). That’s what I’ve read anyway. I don’t know if it’s still current.
@John Cowan: Nothing looks more countable than an ear of maize
Exactly. Because they’re ears. Ears are countable. But good luck counting a pound of corn kernels. It’s easier to weigh it.
@John Cowan: in Japanese we have two long-thin-things of book
Not sure what you’re thinking of here. In Japanese book is 本 with the classifier 冊, which is pretty much only used for books. 本 itself is a classifier for long things. In Mandarin, however, 本 is a classifier for books 書/书, but not a classifier (or at least not the common classifier) for long thin things, which would usually be 條/条.
Huh. Read that somewhere, too long ago to rediscover the source. I’m sure there are other examples of semantic shifting leaving the classifier high and dry, though.
The thing that I find most interesting about classifiers is the specificity of some of them. For example, you can have an ear of grain, but not really of anything else. Most commonly, the grain involved is maize. The pragmatic reason for this is that an ear of corn is a reasonably size unit in which to buy or eat the grain; ears of wheat or barley are too small to be useful individually. The ear is not the only specific unit of corn though; a volumetric measure specific to maize is the crib of corn. (Perhaps unsurprisingly,* Google Ngrams indicates a clear decline in this usage for 1950-2000, compared with 1910-1950.)
My favorite example of an extremely specific unit/classifier though is rasher. It means a slice of bacon, nothing more, nothing less. (The OED informs me that it is also the name of the game fish, but considers it an unrelated meaning, not a development.) In my idiolect, application of rasher to anything other than a slice of pork belly would have to be metaphorical, although the OED does cite some instances with other meats or even non-meat foodstuffs. (The 1861 “rashers of smoked whale” stands out among these, as do two references to rashers of watermelon.)
* I think “perhaps unsurprisingly” is one of the most useful phrases to use in academic writing. It allows an author to introduce something that is not necessarily obvious, yet which, upon reflection, might seem to be a natural consequence of other conclusions that have elsewhere been drawn.
If anyone’s wondering, here’s the German cognate – no relation between the body parts of plants and animals.
From Proto-Germanic *ahaz < *akos.
And aks/ax in North Germanic — one of the less obvious cognates with English.