Darwin and IE

Larry Trask larryt at cogs.susx.ac.uk
Tue Jan 26 16:16:49 UTC 1999

On Mon, 25 Jan 1999, H. Mark Hubey wrote:

> As for the specifics, a language that is isolating and has
> prepositions probably cannot become agglutinating. Look at English.
> Do you think that it will become agglutinating?

> No, it won't. Look at the way phrases are constructed; "a house", "a
> blue house", "a large blue house". How could it agglutinate? Do you
> think that the definite article will get glued as a prefix? Suppose
> it does and we have "ahouse". Then how will "a blue house" look
> like, "ahouse blue"? That would require a change in which the
> modifier would have to go after the noun. I don't see how that could
> happen unless all of a sudden all over the world, all English
> speakers caught on to this craze of poetic speaking like "four walls
> do not a prison make", "my house beautiful", "my fiancee's eyes
> blue", etc.  IT won't happen! Maybe if the language had
> postpositions, it could happen but differently.

As a matter of fact, *exactly* the scenario described here has happened
countless times in languages.  It has been observed in more languages
than you can shake a stick at, and it has happened in languages with
prepositions and in languages with postpositions -- or, more to the
point, in languages with preposed determiners and languages with
postposed determiners.  As it happens, one of the principal
investigators of this phenomenon is Joseph Greenberg, who published a
classic paper on it in 1978 and returned to the subject in 1991.  This
work, unlike Greenberg's mass comparisons, is generally admired by

Let's look at Basque.  In Basque, the order of elements in noun phrases
is Noun - Adjective - Determiner.  An example:

	etxe    zuri    hori
 	house	white	that
	`that white house'

In earlier Basque, there was no definite article, but there were three
demonstratives: <hau(r)> `this', <hori> `that (just there)', and *<har>
`that (over yonder'.  But then the distal *<har> came to be used
unstressed as a definite article; in the process, it was reduced in form
and independence to a suffix <-a>.  Here's the result in modern Basque:

	etxe `house'
	etxea `the house'
	etxe zuria `the white house'
	etxe zuri txikia `the little white house' (<txiki> = `little')
	etxe zuri txiki polita `the pretty little white house'
		(<polit> = `pretty')

See the result?  The article <-a> is simply suffixed to what is
otherwise the final item in the noun phrase, regardless of whether that
is a noun or an adjective.

And it really is an agglutinated suffix.  It is so tightly bound to the
preceding item, whatever that is, that, in western dialects, the usual
western rules applying to vowel sequences apply here as well.  In
western dialects, we have the changes <ea> --> <ia>, <ia> --> <iya>, and
<ua> --> <uwa> or <uba>.  So, in western dialects, we have:

	etxe `house'
	etxia `the house' (= <etxea> elsewhere)
	etxe zuriya `the white house' (= <etxe zuria> elsewhere)
	buru `head'
	buruwa OR buruba `the head' (= <burua> elsewhere)

So, *exactly* the change that Mr. Hubey has just declared impossible is
attested in Basque -- among many other languages.

Of course, this hasn't happened in English -- yet.  But it might.
Already I notice that many of my students -- and one or two of the
contributors to this list -- write `a lot of' as `alot of', suggesting
that they feel the article to be fused to the following item in this
case, at least.  There is nothing to stop English from doing the same
thing that Basque has done, but at the other end of the noun phrase.

More generally, English is showing signs of some very interesting
developments.  Not only do we now have nouns like `house-hunting' and
`baby-sitting', we even have verbs like `house-hunt' and `baby-sit'.
This isn't just agglutination: it's incorporation.  If this development
continues, English may wind up looking rather like an Eskimo language.
Every major change in a language starts from tiny beginnings like this

English has the same pathways open to it as any other language, and its
speakers may choose to follow any of those pathways, sooner or later.
Who can say which choices will be made by our descendants?  Declaring a
perfectly well-attested change to be impossible, merely because it
hasn't happened yet, is unhelpful.

Suppose we could talk to an Anglo-Saxon.  How would he respond if we
told him his language was about to lose its entire case-system for
nouns, its entire agreement system for adjectives, its subjunctive, and
virtually the entire set of verb-agreement endings?  Or if we told him
that his language was about to acquire a whole new set of modal
auxiliaries, a common verb-ending <-ing>, a complete set of overt
progressive verb-forms, a <going to> marker of futurity, and a host of
other grammatical innovations in this vein?  I'll bet he'd say "That's
impossible!" and walk away from the obvious lunatic.

Larry Trask
University of Sussex
Brighton BN1 9QH

larryt at cogs.susx.ac.uk

More information about the Indo-european mailing list