Daniel wrote:
> I was just considering word boundaries and thought that a change of
> script should indicate the start of a new word (assume a space somehow
> vanished). But are there any exceptions to the rule? Can people here
> think of any words that must be spelled with (or likely to spelled
> with) letters of more than one system of writing?

In general, I don't think that a script switch is a valid clue of a word
boundary. However, you should perhaps define what "scripts" and "words" are,
for your purpose.

For instance, do you consider Japanese writing as a single script or as a
combination of two/three/four scripts? If you consider hiragana, katakana
and kanji as three separate scripts, then words written in miltiple scripts
are the norm in Japanese (e.g., verb stems are normally written in kanji,
while tense endings are in hiragana.

Also the concept of "word" can be defined in several different ways, not
necessarily related with grammar. E.g., would you consider a Russian phrase
like "вUnicode-шpифтах" as one, two or three words? (BTW: the example comes
from an actual web page: http://www.sensi.org/~alec/unicode/ru-off97.html).

> Do any nations
> allow for the legal registration of a company name in mixed scripts?

I think that the Latin alphabet is often used in company names throughout
the world, although the legal acronyms ("Ltd", "Inc.", etc.) normally remain
in the local script and language.

I notice that the names of Greek companies often seem to be in a mixture of
Latin and Greek alphabets: Greek letters easily recognized by foreigners
(e.g.: α, ß, δ, ε, ι, κ, ς, τ, υ) are retained, while other letters are
substituted by their Latin equivalents (e.g.: g, z, t, h, l, m, n, p, r, f,
h). Of course, this could be just a graphical stylism used in logos: perhaps
the registered name is all in one script.

_ Marco