"simple juxtaposition"

[Martin Haspelmath]

I disagree with David and Östen:

David Gil wrote:
> Re the Västerbotten dialect: I would tend to agree with Östen Dahl 
> that, as compounds, they don't really belong in the same boat as true 
> syntactic juxtapositions.
Östen Dahl wrote:
> I also think that "simple juxtaposition" is not a wholly adequate 
> label for these constructions, which are rather to be seen as a kind 
> of incorporation involving among other things "compound" prosody.
Whether "simple juxtaposition" is an appropriate label for Västerbotten 
(Swedish) Pelle-äpple 'Pelle's apple' or not depends on the definition 
of "simple juxtaposition" as a comparative concept. Typologists are free 
to define their comparative concepts in whatever way they want, and they 
cannot assume that "juxtaposition" exists as a pre-established category 
(innate or otherwise given in advance, independently of the linguist).

Recall that this is how Andrew Spencer and Irina Nikolaeva define the 
concept "simple/pure juxtaposition":

"by means of pure juxtaposition, without any other morphosyntactic 
marking (agreement, adpositions, case marking etc.)"

This is not a very precise definition (it is unclear what exactly is 
meant by "morphosyntactic marking", and especially by "etc."). However, 
the interpretation that many readers would think of first is in terms of 
"overt marking" (usually by segmental marking, but possibly by stem 
change, as in the Welsh example). But there is no overt marking in 
Swedish, so this does fit Andrew's and Irina's definition. Östen points 
to "compound prosody", implying that "simple juxtaposition" should not 
have "compound prosody". But such a move does not work in typology, 
because "compound prosody" is not a universally applicable notion. 
Comparative concepts need to be defined in terms of universally 
applicable concepts.

Östen also suggests the label "incorporation", but how this is different 
from juxtaposition is unclear. Often it is thought of in terms of 
non-referentiality of the incorporee, but in Västerbotten the incorporee 
can evidently be referentil (Pelle-äpple).

Martin

P.S. I think the term "simple (or pure) juxtaposition" is somewhat 
confusing, because it suggests that "complex juxtaposition" also exists. 
In fact, however, juxtaposition is universally understood in the Spencer 
& Nikolaeva sense: as expression of a relationship between A and B by 
putting A next to B without any overt coding. Juxtaposition is thus 
"simple"/"pure" by definition.