some dozen, some thousand

Question

Some [number] [plural noun] means "approximately [number] [plural noun]". For example:

• Some seven speakers preceded me, so no one listened to my speech.

= About seven speakers preceded me, so no one listened to my speech.

• He had some thirty jellybeans and threw up.

= He had about thirty jellybeans and threw up.

---

Some [plural noun] means (roughly) "more than one [noun]". For example:

• Some speakers preceded me, so no one listened to my speech.

= A number of speakers preceded me, so no one listened to my speech.

• He had some jellybeans and threw up.

= He had jellybeans and threw up.

---

But some words that mean numbers aren't syntactically numbers (numerals) but instead nouns, like dozen and score and hundred and quintillion and googol: a dozen eggs, two score years, several hundred people, a few quintillion seconds, but no *a two eggs, *two three years, *several four people, or *a few five seconds.

---

What does some [that kind of word] [plural noun] mean? For example:

• Some dozen speakers preceded me, so no one listened to my speech.

• He had some hundred jellybeans and threw up.

Does some have the meaning in these sentences that it normally does before a number, "approximately", or that it normally has before a plural noun, "more than one"? In other words, do those example sentences mean:

About a dozen speakers preceded me, so no one listened to my speech.

He had about a hundred jellybeans and threw up.

— or do they mean:

Dozens of speakers preceded me, so no one listened to my speech.

He had hundreds of jellybeans and threw up.

—? (But of course those are just two example sentences: I'm asking about words like dozen and hundred in general.)

Answer 1

In some cases, at least, some dozen means "some twelve", not "some number of dozens":

• "some dozen or two of varieties" [link]
• "some dozen or sixteen lines" [link]
• "some dozen open-air localities" [link]
  • My confidence about the meaning of this last one is because this book very frequently uses phrases like "some half dozen", "some three dozen", "some dozen and a half", etc., to the point that it seems to be an idiosyncrasy of the author. So bare "some dozen" is almost certainly part of that.

And while searching for the above, I also found this paragraph making the same claim:

But before substantives determined or to be conceived quantitatively some appears in the singular or plural, not so much to mark the fraction of a quantum, as the quantum itself, as indefinite, inexact, or approximate: My father had the full view of your flourishing style some hour before I saw it (Ben Jons, Ev. Man in his Hum 3, 1.). About some half hour hence (Shaksp., Cymb. 1, 2.). If I may counsel you, some day, or two, Your highness shall repose you at the Tower (Rich. III. 3, 1.). I would detain you here some month or two (Merch. of V. 3, 2.). Some dozen Romans of us (Cymb. 1, 7.). Bastards, some dozen, or more (Ben Jons., Fox 1, 1.). Some five and twenty years (Shaksp., Rom. a. Jul. 1, 5.). Some six years ago or more (Carl., Fred. the Gr, 5, 5.). Some four miles distant from one of our northern manufacturing towns . . was a wide and desolate common (Bulw., Maltrav. 1, 1.). Surrounded by some fifty or sixty fathoms of iron cable (Dickens, Christm. Car. 1.). All the ladies save some twenty score (Byron, D. Juan 8, 130.). If here an, a should appear instead of some in most cases, no essential difference, will arise. Some and an are distinguished syntactically by the formers answering, except with collective numbers, to an original plural form. [From An English Grammar: Methodical, Analytical, and Historical, by Professor Maetzner, translated to English by Clair James Grece.]

---

Here are some examples with hundred, thousand, etc., that also seem to have this same sense:

• "Love be planted in my heart even the more, blossom and multiply some thirty some sixty some hundred fold." [link]
• "At one instant they [='the broken clouds'] were suddenly cleared away, and discovered to the view a lake that lay probably some hundred feet below, and looking like a small basin of water, or a round hole through the world, so that the sky was visible beyond." [link]
  • Seems to be giving an estimate; the "probably" would be strange if the estimate were not at least somewhat precise.
• "The drive, some thousand miles, would take him two days." [link]
  • The drive in question seems to be to Houston from somewhere in Arizona, so one thousand miles is a reasonable approximation. (And two days is a reasonable estimate for an unaccompanied driver.)
• "In contrast, there are many incentives and attractions in receiving European and North American countries that have increased the number of skilled Arab migrants to some million." [link]
  • The same page uses "some 2 million workers", "some $232 billion", and so on, which is why I'm reasonably confident that it's this sense of some.

---

Also perhaps relevantly, the decline of some dozen, some hundred, and some thousand over the past 150 years correlates pretty well with the decline of the unambiguous some half dozen (see Google Books Ngram).

---

None of the above, however, demonstrates that some dozen/hundred/etc. never means "some number of dozens/hundreds/etc.". So the phrase could still be ambiguous. But I've found only one case where it does seem to mean that:

twenty-eight thousand and some hundred guilders. [link]

and the more I think about this case, the more I think it does actually mean something like "28,100 or so", with the same sense as the above example.

Answer 2

"some" has quite the history.

1. Importantly, its root - *sem - contributes the initial m in mile:

   from Latin mīlle passus, pl. mīlia (“thousands”), from reconstructed Proto-Indo-European *smih₂ǵʰéslih₂ (“one thousand”), from *sm̥- (“one”) and *ǵʰes- (“hand”; “tobtake, give in exchange) (en.wiktionary: mille), from the numeral stem *sem- (“one, together”).

2. The reconstructed root of one is not present in the Anatolian and Tocharian branches of Indo-European. It will be helpful to illustrate this in more detail:

   • Tocharian B (Adams 2013) ṣe: "one", "same", "some", "together";

   ... directly equatable with Greek heîs (< PIE *sm-s or possibly *sem-s). More distantly we have Armenian mi (reflecting *sm-ihxos) ‘one,’ Gothic sin-teins ‘daily,’ Old English sinnihte ‘eternal night,’ Latin sem-per ‘always,’ etc. (P:902-904; MA:399). This connection goes back to Sieg/Siegling, 1908:927, and Meillet, 1911-12:284-5.

   This Old English sinnihte is similar to the construction of fortnight "2 weeks; 14 nights" and sennight, OE seofeniht "1 week". The meaning "eternal" is also present in Thocharian B ṣek "always". Adams argues that, "The semantic development is ‘once for all’ to ‘always’ and one should compare, mutatis mutandis, Latin sem-per (Meillet and Lévi, 1911:451; cf. MA:410). Otherwise VW:453-454 (from *sgh-) […]

   Adams admits many more derivations, especially ṣemeske "only, sole; private" may be of note. This seems to match the first part of the question, some dozen, approximately. "up to" is collocated in the phrase […] śak täṅte.

   Problematically, however, the initial of Toch. B śak "ten" < *déḱm̥- "is difficult, since we might ordinarily expect either ts- or c-" (Adams)—cp. Toch. śak-śe "eleven", śak-wi ‘twelve’, śkänte-wäte "twelfth" etc. and English dozen (en.WT): Old French douzaine, doze + -aine, from Lat. douodecim; -aine "(noun) indicates approximation, suffixed to a number". Wartburg (FEW: duodecim) refers to [A Gl. 9, 60 n. 4] (sigel resolved in Beiheft, not available to me rn) for the suffix, though separating Old Provencal dotzen from dozen – right to tax, of wine etc. (translation mine) makes it similar to tithe "tenth" and to Tocharian B śkanno "tithe", as **śkanto is derivable as feminine from śkante "tenth" (Adams)—a gloss of Sanskrit daśamá "tenth", from *déḱm̥-.

   Discussion of possible relations between **déḱm̥- "10", **ḱm̥t- "100" and *sem- vel sim. in (Eskes 2020) does not account for *dḱ > *s or *só / *to &c. as far as I can tell. Their notion of conditioned **d > **h1 (Kortlandt Effect) should at least be relatable to the root of one < **h1oy-nos, which leaves a lot to be explained.

   • Hittite (Kloekhorst, 2008): ši- (numeral) "one"; šani (adj.) "the same, one and the same"

   The Anaotlian numerals are difficult because they are almost always written logographically. Kloekhorst reconstructs *sih2- "one" without further etymology. He notes formal problems for a derivation of šani from *sem (pace Eichner). However, he provisionally eccepted *som- in another lemma:

   • =(š)šan

   sentence particle indicating superposition (‘over’ , ‘upon’ , ‘on’ etc.); indicating contiguity or close proximity; accompanying ‘for (the benefit of)’ or ‘about, concerning’ ; accompanying ideas of measuring or counting; indicating ‘of, from’ (only OH).

   The list of ultimately uncertain, superficially similar lemmas is of course too long to brood force our way through.

   In addition, both Tocharian and Hittite admit borrowing from Indo-Iranian. E.g. Toch B śkä-maiyya – of ten powers – in loan translation to Prakrit daśabala, an epithet of Buddha.

   Eventually, nothing much on the collective sense of assemble turned up in my search, as German adv. insgesamt "in sum, total, all told" would suggest, though it should fit the opening question's second example, "[sum] seven speakers preceded me, so no one listened to my speech." Toch. B ṣäms "count" is close enough, cp. (Adams) Sanskrit samayati "puts in order", Old Norse semja "put together, put in order, unite", Toch. B ṣämla- "count" also in collocation: "countless".

3. someone comes possibly from rebracketing so many - *some any. Point in case, German so manche is formally equivalent to so many, but it patently means some, a couple, any; also manch einer "some people" is formally equivalent to the "many a" construction, e.g. manchenorts "many a place" (cf. collinsdictionary german-english: manch). It could be the other way around or complete coincidence. The fact is that it works.

   This pertains to the second part of the question, exactly:

   Some speakers preceded me, so no one listened to my speech.

   [So many] speakers preceded me, so no one listened to my speech.

   He had some jellybeans and threw up.

   He had [so many] jellybeans [he] threw up.

   The latter is an example of stranding, I believe. The former is remarkable for the repeated so.

   Further relations to the Old English indefinite pronoun man "one", German man, jemand "somebody", jederman "anybody" are obscure.