Origin of nominalising morphology.

[Larry Trask]

----------------------------Original message----------------------------
--On Sunday, June 15, 2003 2:07 pm +0000 Guy Deutscher <gd116 at HERMES.CAM.AC.UK> wrote:

> I have a question about the origin of nominalizing morphology, which I
> hope someone can help with. What prompts the question is the wish to
> understand how infinitives develop in language.

> But where
> does the morphology which takes a verb and turns it into an 'action
> nominal' come from in the first place? One would expect that it arises by
> grammaticalization, and ultimately from some lexical source. But what are
> these sources?

> Do people know of examples where one can actually
> observe the emergence of nominalizing morphology on verbs, or at least
> reconstruct it transparently?

I think we can reconstruct the origins of nominalizing morphology in Basque
with some confidence.

Basque has no infinitive, but it has a gerund which is of very frequent use.  The gerund can appear in any NP slot in a sentence, and it can take any required nominal case-marking.  For example, it can take the allative case-suffix <-ra> to express purpose:

  (1a) Etxe-ra    joan nahi du-t       aita   ikus-te-ra
       house-ALL  go   want AUX-1SgERG father see-GER-ALL
       'I want to go home to see my father.'

It is clear that the modern gerund is of recent origin.  For one thing, the
formation of the gerund exhibits considerable regional variation, though the gerund is always formed by adding a suffix to the stem of the verb. The western dialect Bizkaian uses only the suffix <-te>.  Most varieties use both <-te> and <-tze>, which are in complementary distribution, depending on the nature of the verb stem.  A west-central area uses <-keta>.  An eastern variety uses <-(e)ta>.  And old Bizkaian exhibits an apparent suffix
<-zaite>, long extinct.

Now, these gerund-forming suffixes are identical in form to ordinary noun-forming suffixes in the language, as follows.

The suffix <-te> commonly forms nouns of duration:

  <gose> 'hunger' --> <gosete> 'famine'
  <euri> 'rain' --> <eurite> 'rainy spell'

The suffix <-tze> commonly forms nouns of abundance:

  <jende> 'people' --> <jendetze> 'crowd, multitude'
  <diru> 'money' --> <dirutze> 'riches, treasure'

The suffix <-keta> has several functions, including 'abundance' and 'duration', but most commonly it forms nouns of activity:

  <zezen> 'bull' --> <zezenketa> 'bull-running' (as in Pamplona)
  <hitz> 'word' --> <hizketa> 'speech, conversation'

The eastern <-(e)ta> is quite possibly a variant of <-keta>, since <-keta> has a frequent variant <-eta>.  The archaic <-zaite> appears to consist of
<-te> attached to another suffix; I won't pursue it here, but in fact there
is little difficulty in seeing this as <-te> attached to the noun-forming suffix <-tza> (the extra /i/ would be expected here) (compare the universal
agent suffix, which is <-le> or <-tzaile>, according to the nature of the verb-stem).

So, there are good reasons for supposing that the modern gerund has been constructed by adding to the verb-stem a noun-forming suffix meaning variously 'duration', 'abundance' or 'activity'.  All these look semantically good.

Now, there is one possible technical objection.  The noun-forming suffixes just mentioned form ordinary nouns, with ordinary nominal properties.  But the gerund is wholly verbal in nature: it takes subjects and objects with ordinary case-marking; it takes adverbs; and in short it has only verbal properties.  It can't take nominal modifiers or specifiers like adjectives or determiners.  This looks awkward.  (What is nominalized is the entire gerund phrase, though any required case-marking appears on the gerund itself.)

But there's a wrinkle.  In northern (French) Basque, the gerund exhibits one nominal property: in certain circumstances, the direct object of a gerund stands, not in the absolutive case, like an ordinary direct object, but in the genitive case.  So, my example (1a) above appears as follows in northern Basque, where the allative case-suffix is <-rat> and the /k/ in <ikusi> 'see' is aspirated:

  (1b) etxera joan nahi dut aitaren ikhusterat

Here <aita> 'father' takes the genitive case-suffix <-en> (the /r/ is purely phonological), and the whole thing appears to be literally something
like "...for the seeing of my father".

This observation strongly suggests that the gerund was once an ordinary noun, with ordinary nominal properties, and that its acquisition of verbal properties is a secondary development.

I've written about this here:

R. L. Trask. 1995. 'On the history of the non-finite verb forms in Basque'.
In José Ignacio Hualde, Joseba A. Lakarra and R. L. Trask (eds), Towards a History of the Basque Language, Amsterdam: John Benjamins, pp. 207-234.

I hope this is the sort of thing you're looking for.


Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk