Ora Matushansky (CNRS/Université Paris-8) and Joost Zwarts (Utrecht University)

THE PARTIAL NOMINALITY OF AXIAL PARTS

Axial prepositional complexes are widespread cross-linguistically:

(1)     a.       El    libro  está   de-l-ante           de la    mesa.                                  Spanish, Fábregas 2007
the  book  is       from-the-front   of the  table
The book is in front of the table.

1. hu   haya    mi-taxat          la-bayit/ha-bayit.                       Hebrew, Botwinik-Rotem 2008
   he   was      from-bottom   dir+def-house/ def-house
   He was under the house.
2. S-pered-i                      ot      dom-a        roslo  derevo.
   down.from-front-loc   from  house-gen  grew  tree
   A tree grew in front of the house.                           Russian, Mitrofanova and Minor 2013

(2)     Maria     a-mami       î-gûrû ri-a    metha.                                       Kîîtharaka, Muriungi 2006
1.Maria  sm1-sleep   5-top   5-as  9.table
Maria is sleeping/lying on top of the table.

Svenonius 2006, 2010 and many others treat axial elements like ante, taxat, pered, and gûrû as purely functional. We argue that they are lexical (and generally nominal), albeit deficient, because this naturally accounts for alternations like on top of your head versus at the top of your head, idiosyncratic restrictions on the ground with some axial parts (e.g., on board this train/aircraft/spaceship/*theater), their nominal syntax (e.g., overt definite articles (1a), their gender agreement), and various similarities with weak definites (e.g., noun-dependent article drop, lack of modification and impossibility of pluralization (Ross 1996, Svenonius 2006)).

In order to explain how axial complexes are built around axial nouns, we propose a semantic decomposition of an axial complex:

(3)     [locative [def₁ [projective [def₂ [axial [possessive [ground]]]]]]]

We show how this decomposition accounts for the different ways in which axial complexes are realized cross-linguistically.