compounding and attributive modification

Martin Haspelmath haspelmath at EVA.MPG.DE
Mon Jan 30 13:46:00 UTC 2006

There's a paper by Laurie Bauer (in the 2-volume handbook "Language 
typology and language universals, ed. by Haspelmath et al., 695-707) on 
"Compounding" in a typological perspective, where he investigates this 
and related questions in a world-wide sample of 36 languages.

He finds that there seems to be a unidirectional implication from AdjN 
order to ModHead order in compounds: his sample has only a single 
language (from South America) with AdjN and HeadMod order (p. 697):

NAdj & HeadMod 10
NAdj & ModHead 11
AdjN & HeadMod 1
AdjN & ModHead 9

Thus, Ulwa's pattern (NAdj & ModHead) is actually the majority pattern 
in Bauer's sample. But note that such typological data cannot be used 
for or against a particular synchronic description anyway, because Bauer 
cannot claim to have looked only at "real" adjectives, to the exclusion 
of "apparent" adjectives that "in reality" are internally headed 
relative clauses. Maybe "in reality" all 11 languages in Bauer's sample 
are like Ulwa...


Andrew Koontz-Garboden wrote:

>Hello.  Does anyone know of any literature that addresses the
>relationship between headedness in compounds and headedness in
>noun/adjective attributive modification?  (Or, save that, does anyone
>simply have any idea/hunch whether there's any relationship?)  What I
>have in mind is the following.  In the compound in (1a), the head is
>man (ie, a frogman is a kind of man).
>a.  frogman
>b. the white house.
>Similarly for (1b), which is a kind of house.
>In the attributive construction in (2), the head is "dog" (i.e., an
>ugly dog is a kind of dog).
>(2) an ugly dog
>Someone suggested to me that there might be a relationship between
>headedness of compounds like those in (1) and attributive
>constructions in (2).  So, the question I pose is:  if compounds are
>right-headed, do attributive constructions tend also to be, and vice
>versa?  Similarly for left-headedness?
>Among the reasons I ask is that in Ulwa, a Misumalpan language I'm
>working on, there is a mismatch in headedness of constructions like
>those in (1) and (2), as shown in (3), a kind of woman, and (4), a
>kind of man.
>(3) was sirau
>    water maiden
>   `mermaid'
>(4) al yuuhka
>   man tall
>  `a tall man'
>I have independent reasons for actually believing that what look like
>attributive constructions (ie, (4)) may instead be internally headed
>relative clauses.  I'm wondering if the mismatch in headedness between
>compounding and attributive modification can be taken as a(n) (perhaps
>weak) argument.  So, if anyone knows of any literature that addresses
>this question (or has evidence one way or another), I'd be grateful.
>Andrew Koontz-Garboden
>Andrew Koontz-Garboden
>Department of Linguistics
>Margaret Jacks Hall, Bldg. 460
>Stanford University
>Stanford, CA 94305-2150
>andrewkg at

Martin Haspelmath (haspelmath at
Max-Planck-Institut fuer evolutionaere Anthropologie, Deutscher Platz 6	
D-04103 Leipzig      
Tel. (MPI) +49-341-3550 307, (priv.) +49-341-980 1616

More information about the Lingtyp mailing list