At 1:01 pm -0700 15/9/00, J Matthew Pearson wrote:
>Nik wrote:
>> Marcus Smith wrote:
>> > 6. All languages with dominant VSO order have SVO as an alternative or
>>as the
>> > only alternative basic order.
>> All?  There are none that are purely VSO?  Because Watakassí is purely
>> verb-initial.
>Greenberg may have been referring to the fact that verb-initial languages
>always allow topicalized and/or focussed noun phrases to precede the verb.

The fronting of topic or, less often, the focus occurs in other other
language types besides VSO.  We should not, maybe, contrast the fact that
there has been found no VSO language in which the V must always come first,
whilst there are SVO langs in which the V must always be last.  What,
maybe, we should look at is _fronting_ in both sorts of language.

It seems to me, on reflection, that the Welsh word order is more strictly:
[focus] Verb [S] [O] (where [] shows optionality, or the possibility of
zero forms).  Actually a better notation is really required since if, e.g.
S is the focus, then [S] is 'emptied'.

Not all sentences have 'focus' and those that begin with the verb have zero
or null focus, i.e. the 'unmarked' sentences.

This would make Welsh (and Gaelic) a "V2" language (an awkward term IMO and
one which, to people of my generation & older on this island, has
unfortunate sinister associations).

German is considered a V2 language (we must get a better terminology!) but
works rather differently than Welsh. It is usually stated that the 'normal'
word order in a main clause (I'm not considering German subordinate clauses
here) is SVO and that if something else is fronted then the SV is
"inverted".  Now, I suggest that the German main clause has, in fact, the
word order; [topic] V [S] [O].  When a topic does not change, the
grammatical subject more often than not corresponds to the topic, hence
main clauses tend to begin SV.

Where a so-called VSO language has _topic_ fronting, then the most common
variation will be SVO.  Indeed, whether a language is to be classified as
VSO with topic fronting, or as a "V2" language then becomes, I guess, the
degree to which the SVO order and VSO order dominate.

>In some
>languages, this fronting is restricted to subjects, hence the "only
>basic order" addendum.

This, I see, as a restricted form of topic fronting.

What I'm suggesting is that the distinction between the so-called V2
languages and VSO languages is unreal.  That there are no VSO langs, in the
strict sense, to contrast with SOV langs.  What we have is a distinction
between those languages which place the verb first & those that place it
last in the 'non-fronted' part of the sentence.

Oh dear, I need a term for the 'fronted element' and the rest. For the
moment, I'll use the terms Prolog uses in list processing: Head, Tail. (I
could use CAR and CDR, but only LISPers would understand that :)

Remembering that the 'head' can be null, we have four language types:

     HEAD   TAIL
    topic   V[S][O]
    topic   [S][O]V
    focus   V[S][O]
    focus   [S][O]V

Of course that does *not* account for all language types.  It doesn't
account, e.g. for English!  In English if we want to emphasize the topic,
we may front it - but we still keep the subject before the verb, e.g.
Yesterday I went to the market.

Oddly, however, not all topics can be fronted.  "To the market I went
yesterday" does not sound 'natural' to me and is AFAIK would not be used.

No doubt the professional linguists will take issue with some of these
ramblings.  I have not by any means developed these thoughts into a
coherent theory.  They are just thoughts that have come to me as I've been
thinking around Greenberg's 'Universal #6' and Matthew's reply and I share
them with you  :)


A mind which thinks at its own expense
will always interfere with language.
                   [J.G. Hamann 1760]