5.0246 Number Words -- The Teens (2/133)

Elaine Brennan & Allen Renear (EDITORS@BROWNVM.BITNET)
Wed, 24 Jul 1991 21:39:45 EDT

Humanist Discussion Group, Vol. 5, No. 0246. Wednesday, 24 Jul 1991.

(1) Date: Fri, 19 Jul 91 12:05:27 -1000 (110 lines)
From: David Stampe <stampe@uhccux.BITNET>
Subject: the teens &c.

(2) Date: Wed, 24 Jul 91 17:46:22 EDT (23 lines)
Subject: Summary of Responses on the Teens

(1) --------------------------------------------------------------------
Date: Fri, 19 Jul 91 12:05:27 -1000
From: David Stampe <stampe@uhccux.BITNET>
Subject: the teens &c.

In response to Eric Rabkin's query about teens, they are discussed in my
paper "Cardinal Number Systems", in the 1976 Chicago Linguistic Society
volume. My aim there was to explain univerals of number systems (their
math, prosody, syntax, and phonology) in terms of their uses, of which
the primary one is counting. Here's a brief outline, regrettably long:

A. Counting requires consecutive integers, i.e. addition. There are
(or were) languages with only addition, e.g. "two and one" = 3, "two
and two and one" = 5. The order is smallest additive last: "one and
two" would not be a number. Why? In language new information comes
last, and in counting what's new is always the lower "digits". (See
your digital watch.)

B. The units (simple words for consecutive integers) often referred
originally to handy counters such as fingers (4 or 5, or 8 or 10),
joints of the fingers (12), or fingers and toes (20). The highest
unit, which is used as a base for combining numbers (see B), refers
to these counters collectively. "Ten", "-teen", and "-ty" may have
derived from proto-Indo-European (and earlier) words for hand(s).

C. To count higher our limited memories require some "clumping", so we
use multiples of the base: "one HAND and one", "one HAND and two",
... "two HANDS and one", etc. For consecutiveness, these have the
structure "(UNIT x BASE) + UNIT", with the base a simple word, never
a complex as in *"UNIT x (BASE + UNIT)". Higher bases are nouns
for its multiples (twenty, thirty) or powers (hundred, thousand).

D. If additives have the order higher...lower (see A), then by ordering
multiplicatives lower...higher, any numeric expression can be parsed
and interpreted without explicit "times" or "plus" words, simply by
their orders and magnitudes. So we say "two hundred" for 2 x 100, to
reserve "hundred two" for 100 + 2, and therefore "one hundred two" is
unambigously interpretable as (1 100) + 2 = 102. The same rule of
"LARGE + SMALL, SMALL x LARGE" handles even "recursive" expressions
like "one hundred two thousand one hundred two", i.e.
(((1 100) + 2) 1000) + (1 100) + 2 = 102,102

E. BUT counting is talking, and talking takes time. So we abbreviate in
familiar contexts: "three ninety-five", "ten sixty-six", etc. And we
elide familiar numbers: "se'm" (7), "'le'm" (8). Systematically, in
number systems, "one" can often be left out where it's deducible,
as in "(one) hundred", and it's normally left out in "[one] ten",
and in the teens: "fourteen" in effect stands for "one-ty four".

Further, we shorten familiar phrases by eliminating rests between
words, thus making phrases into compounds: "twenty [rest] seven" ->
"twenty-seven" (contrast the unshortened "twenty [rest] thousand").
And we shorten familar compounds by making them simple words: German
"vi'erze`hn" (with two accents) -> Yiddish "fi'rtsn" (with one
accent), Latin "quattuor-decim" (with two accents) -> French
"quatorze" (with one). Shortening typically affects the most
familiar elements, and in counting those are the lower numbers. So
it is most likely to affect the teens, and then the higher decades.

F. There is another rule for new information: it is always accented. In
counting we say twenty-ONE, twenty-TWO, twenty-THREE. Note how the
accent shifts to the new information in counting (abnormally) by tens
from five: FIVE, FIFteen, TWENTY-five, THIRTY-five. And the teens,
which take the normal English "hind" word-accent in isolation, in
counting are accented on the new information: THIRteen, FOURteen.
New information changes the accent, but the grammar doesn't allow it
to change the order of the parts of phrases or compounds. But the
grammar can change...

G. English used to say "four-and-twenty", like German "vierundzwanzig";
and also like "fourteen", German "vierzehn", Latin "quattordecim".
In all these the rule Large + Small is violated. The reason is that
they are compounds, and compounds act more like words than phrases,
and words may have rules for accent different from phrases. In Old
English, German, and prehistoric Latin word accent was initial. And
so, in compound number words, the natural place for the small number,
which is accented as new information in normal counting, was initial.

H. In Classical Latin and early modern English, accent shifted to
(relatively) word-final. In the higher compounds in both languages,
the small number eventually shifted with it: twenty-four, viginti-
quattuor, and likewise in Romance: vingt-quatre. But the teens,
being more frequent, and thus more likely learned by rote, resisted
change -- just as frequent irregular forms, like forms of "be",
resist change. Nonetheless, in Romance, some higher (less frequent)
teens were rebuilt: after seize (a one-syllable shortening from Latin
sexdecim), new forms like dix-sept were substituted with the small
number under the final word accent.

In English, word accent shifted in the teens (fo'urte`en ->
fo`urte'en) as pronounced in isolation, but the word order didn't.
Maybe we should be saying "one-ty four". Besides conservatism, the
the reason why "fourteen", with its non-match of accent with new
information, is tolerated may be that in context our English "rhythm
rule" happens to make it right: "fo'urte`en do'llars".

I. To keep this over-long posting shorter, I've skipped suppletive
numbers, like eleven and twelve, where the "system" would lead us to
expect "oneteen" and "twoteen". Or subtractive (really anticipatory)
numbers like Latin "duodeviginti" (2 from 20 = 18), which didn't make
it into Romance. Usually we learn such numbers before we learn the
system, just as we learn "was" before we know enough of the English
verb tense system to produce "is-ed". For their origins, any good
dictionary will give etymologies, but who knows the deeper reasons
for their existence, persistence, and even productivity? There is
one system, that of Sora (a Munda language of India) which, though
derived historically from a purely decimal system, counts by twelves
and twenties, so that 56 is two-twenties twelve-and-four. Go figure.

David Stampe <stampe@uhunix.uhcc.hawaii.edu>, <stampe@uhunix.bitnet>
Dept. of Linguistics, Univ. of Hawaii/Manoa, Honolulu HI 96822
(2) --------------------------------------------------------------27----
Date: Wed, 24 Jul 91 17:46:22 EDT

In response to my "Teen Query" asking if anyone knew
why French starts teens with "dix" beginning with seventeen
while Spanish starts teens with "diez" beginning with
sixteen, I received four replies. The first three served
to show that the "teens" are even more complicated than
I had mentioned, pointing out, for example, that Latin
(at least in some forms) forms 18 and 19 by subtracting
from 20. Willard McCarty directed my attention to Karl
Menninger's excellent *Number Words and Number Symbols*
where, on pp. 84-86, the teens are examined. The book
as a whole, and that section, are enlightening, but I
must report that the variation between Spanish
and French is, according to Menninger, "unexplained."

Anyone out there looking for a dissertation topic?

Eric Rabkin esrabkin@umichum.bitnet
Department of English esrabkin@um.cc.umich.edu
University of Michigan office: 313-764-2553
Ann Arbor MI 48109-1045 dept : 313-764-6330