<OT> New Posting: ROA-693

roa at ruccs.rutgers.edu roa at ruccs.rutgers.edu
Fri Nov 5 12:04:39 PST 2004


ROA 693-1104

Nonlocal Reduplication

Jason Riggle <jriggle at uchicago.edu>

Direct link: http://roa.rutgers.edu/view.php3?roa=693


Abstract:
Marantz (1982) observed that reduplicative affixes generally
copy the string of segments beginning with the edge to which
the affix is attached and proceeding into the word. Though
Marantz described ‘edge-in association’ as a tendency, many
subsequent researchers have (at least tacitly) taken it
to be an inviolable principle in reduplication (e.g. McCarthy
and Prince 1996, Kager 1999, Nelson 2003 and others). This
assumption is incompatible with the existence of nonlocal
patterns of reduplication where unreduplicated surface material
intervenes between the two surface copies.

To maintain the inviolability of edge-in association such
patterns must be analyzed as showing covert locality or
as non-instances of reduplication. On the other hand, if
edge-in association is merely a tendency, nonlocal reduplicative
patterns can be given a straightforward analysis with locality
constraints and a relaxed definition of the ‘base’ -the
string the reduplicant is obliged to copy.

(1) The base generalized:  
    Everything in the output that isn’t the reduplicant
is the base.

(2) LOCALITY:  
    No segment that isn’t itself in the correspondence relation
M1 R M2
    may intervene between two segments corresponding via R. 
    – One mark is assigned per segment y that lies between
x and x' in S
      where x R x', unless there's a y' in S and y R y'. 

LOCALITY says that only segments that are themselves in
B/R-correspondence may separate corresponding elements in
the base and reduplicant. Under the generalized definition
of basehood, reduplicant placement will be determined solely
by LOCALITY and the constraints generally responsible for
affix placement (e.g. ALIGN). Reduplicant content can then
be determined by B/R-MAX constraints indexed to salient
elements like stems, edges, and stressed syllables. These
constraints generate a typology of reduplication that includes
nonlocal patterns like the ones in Koryak and Creek but
is overwhelmingly composed of reduplicative patterns that
obey Marantz’s generalization.

Comments: This paper appers in the proceedings of NELS 34 at Stony Brook
Keywords: Reduplication, Locality, Creek
Areas: Phonology
Type: Manuscript

Direct link: http://roa.rutgers.edu/view.php3?roa=693



More information about the Optimal mailing list