Programm 79. StuTS

There are two different views on cyclic interaction: a) Morphology interprets syntax
(Bobaljik 2000, among others), where in a first cycle, syntactic structure-building takes place,
and in a second cycle, morphological realisation applies to the fully built syntactic structure,
and b) Cyclic Spellout (see see Uriagereka 1999, Chomsky 2001), where once a morphological
phase P is completed in syntactic structure-building, the structure built so far is transferred to
the interfaces and morphological realisation applies to it, such that syntactic structure-building
and morphological realisation are intertwined. I show that the latter approach can predict extended exponence, where a feature is realised by multiple affixes, by counter-bleeding without having to resort to secondary contextual features or enrichment (Müller 2004). In a Cyclic Spellout approach, as heads are merged in the syntax, features are incrementally added to the syntactic structure that provides the context for vocabulary insertion. Now
certain heads have the property that their maximal projection is a morphological phase and
must be spelled out before other heads can be merged, i.e., before all features are present in
the syntactic structure, leading to premature insertion of more generic exponents where more
specific ones would be expected.

See attached file.
Background:
In (1), the German plural suffix /er/ and the German plural dative suffix
/n/ occur together (1b), but /n/ cannot occur on its own (1c). Given the Specificity condition
of the Subset Principle (e.g. Halle and Marantz 1993), the less specific plural marker /er/ is
expected to be blocked by the more specific plural dative marker /n/, counter to fact.
(1) a. Kind-er
child-PL
b. Kind-er-n
child-PL-PL.DAT
c. *Kind-n
child-PL.DAT
The standard approach is to assume that the shared feature is a
primary feature on the less specific and a secondary, contextual feature on the more specific
exponent. This relies on a distinction between primary and secondary features and a concept of
feature discharge, which arguably makes a wrong prediction for affix order in cases of extended
exponence like (1): It predicts that the more specific marker is inserted before and thus closer to
the stem than the generic marker, counter to the cross-linguistic observation made in Grofulović
& Müller (2023) that in such cases, it is the more generic exponent that tends to be closer to
the stem, which means from a derivational point of view that it must be inserted first. nother
solution proposed by Müller (2004) and Grofulovi´c & Müller (2023) is enrichment, where
prior to vocabulary insertion, the feature shared by two exponents is copied such that it can be
discharged twice. This approach involves look-ahead if feature copying only happens before the
insertion of exponents whose lexical entries have a feature in common. If, on the other hand, all
features are copied independently of whether markers in an extended exponence relation will
be inserted or not, then the analysis overgenerates.

Info

Tag: 16.05.2026
Anfangszeit: 15:20
Dauer: 00:30
Raum: DOR 24 1.102
Track: Morphology
Sprache: en

Links:

• iCalendar

Dateien

• cyclic_spellout.pdf