internally headed relative clauses

nigel vincent nbvint at NESSIE.MCC.AC.UK
Tue Mar 13 12:12:31 UTC 2001

Dear Colleagues,
I can't add anything to the discussion of the interesting pattern of case
marking that Wolfgang Schutze reports for Udi, but I was surprised by a
remark in the middle of his message, viz:

"I know that some people probably relate this technique to what is
sometimes called 'relative clause internal head' constructions (in
formal approaches) [personally, I don't think that such a formal label
does explain anything]."

My surprise arises from the fact that I had always thought that the term
'internal headed' had arisen in typologically oriented work on relatives as
a label for the pattern found in languages like Quechua and Navaho in which
the 'head' (in the sense of the noun which is modified by the relative
clause) appears according to standard distributional tests to be inside the
relative clause rather than outside. Certainly Keenan in his survey of
relatives for the Shopen CUP collection (1985) treats them as a separate
category, and notes a correlation with SOV order. Where does the idea arise
then that the term 'internal headed' is due to formal approaches?
In fact I think formal approaches are rather embarrassed by this kind of
typological diversity in relative clause constructions (there is much less
of a formal problem about resumptive strategies which are simply the overt
'spelling out' of an empty category). Hence the attempt by Cole ('Natural
Language and Linguistic Theory' 5 (1987): 277-302) to show that these
constructions do indeed have an external head. This idea is picked up by
Kayne in his book 'The Antisymmetry of Syntax' (1994), since Kayne for
independent reasons is led to the conclusion that Cole's analysis of
Quechua and Lakhota applies to English. In other words, those formalists
who have looked at this problem have tended to argue for a single unified
analysis of relative clauses, and not to put any weight on a terminological
distinction between external and internal headedness.

I make these comments because I believe very firmly that typologists and
formalists should be engaged in a dialogue with understanding on both
sides, and that dismissive remarks of the kind in Wolfgang's message are
not very helpful, especially when, as far as I can see, they are not even
factually accurate. I'm sure some people will immediately react with
instances of dimissive and ill-informed remarks on the part of formalists
vis-a-vis typologists, but that is not really the point. As the proverb has
it, 'two wrongs don't make a right'. The issue seems to me rather to be
whether formal approaches can in principle contribute to the advancement of
knowledge and understanding. My view on this one has always been
unequivocally 'yes', and to that extent I've never seen any intrinsic
conflict between formalism on the one hand and typological work on the
other. Am I alone (at least amongst subscribers to this list and to ALT) in
holding this view?

Best wishes
Nigel Vincent

Nigel Vincent	               Tel: +44-(0)161-275 3194
Department of Linguistics      Fax: +44-(0)161-275 3187
University of Manchester       e-mail: nigel.vincent at
Manchester M13 9PL
UK		    	         Visit our web-page:

More information about the Lingtyp mailing list