LINGUIST List 6.1342

Mon Oct 2 1995

Disc: Languages With No Between-word Delimiters

Editor for this issue: Ljuba Veselinova <>


  1. Christopher Upward, Subject: 6.1264, Sum: Languages With No Between-word Delimiters

Message 1: Subject: 6.1264, Sum: Languages With No Between-word Delimiters

Date: Mon, 25 Sep 1995 16:19:50 Subject: 6.1264, Sum: Languages With No Between-word Delimiters
From: Christopher Upward <>
Subject: Subject: 6.1264, Sum: Languages With No Between-word Delimiters

Some rather vague historical comments - I don't have references to hand.

 This message uses Cut Spellng - for an outline of the system, see

I undrstand that th irish wer th orijnl inventrs of word-space delimitrs,
using them for latn, perhaps because it was a foren languaj for them, and
hence too dificlt to read in its norml unspaced script. Th irish then pasd
th trik on to th northumbrians, ho bilt it into ther english in th 7th
century. English was thus posbly th first vernaculr to insert spaces
between words. Howevr, th anglo-saxns wer ofn unsure wher one word endd
and th next began, so ther was a lot of variation (as ther stil is today -
do we rite 'on to' or 'onto'?).

Latn without word-spaces was of corse dificlt to read, and fluent silent
readng was most unusul. Ther is mor than one story about ho first tryd to
read silently: one story has it that Julius Caesar used to read militry
dispachs silently, so that militry secrets wudnt be overherd. Anothr story
has it that one of th mor importnt saints (Jerome? Augustine of Hippo? - I
forget wich) astonishd his felo eclesiastics by readng silently - an
unprecedentd feat.

Altho latn didnt insert spaces between words, it is not entirely tru to say
that words wer nevr delimitd: th betr class of roman inscriptions (eg,
Trajans Colum) used mid-line dots between words. But as with Old English,
th conventions as to wher th word boundris fel wer not yet fixd: wer
'prefixs' to be red as joind to or seprat from th foloing morfeme?

As far as I no, no languaj used word-spaces befor irish-latn. Certnly
ancient greek didnt, and so presumebly its semitic precursrs didnt eithr.

I gess ther is som controvrsy about how 'natrl' word-boundris ar. I hav
herd that ilitrat amazonian indians ar capabl of distinguishng words. But
many word boundris ar merely matrs of ritn convention (wy 'upon', not 'up
on'? wy 'up to', not 'upto'? wy 'fireman', 'fire-alarm', 'fire
insurance'?) and hence to som degree arbitry. Altho modrn jermn is famus
for formng long compound words, it too somtimes has problms deciding wen
two morfemes shud be compoundd or not: shud one rite 'Rad fahren' (= to
weel ride - ie, to ride a bicycl) or 'radfahren' (= to weelride - ie, to
rideabicycl). French also has uncertntis on this point, but both french
and jermn ar tryng to rationlize som of ther most notorius uncertntis in
recent spelng reform proposals.

Th botm line of word-spacing no dout has to do with 'th gramr of lejbility'.

I beleve ethiopian had a uniqe word-sepration markr, but I dont hav my
infrmation on it to hand.

Christopher UPWARD
Aston University 0121-359 3611
Fax 0121-359 6153
Home address and telephone
 	61 Valentine Road
 	Birmingham B14 7AJ
 	tel. 021-444 2837
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue