Lelis Rivera & Alberto Romero
TITULACIÓN DE COMUNIDADES NATIVAS DEL PUTUMAYO
Parallel to the manner in which the interpretation component of the grammar takes the output of the syntax as its input, the phonological component takes the same output of the syntax and maps it to a phonological representation. The postsyntac- tic derivation applies morphological operations which realize the syntactic terminal nodes and subsequently the structure is linearized and flattened and phonological processes operate on the output.
Consider the nouns leikur ‘game’ and kassi ‘box’ whose morphological segmen-
tation is given below.
(60) a. leik-ur √play-nInfl[masc,nom,sg] ‘game’
b. kass-i √box-nInfl[masc,nom,sg] ‘box’
Two morphemes have overt phonological exponents in each noun, the root and the nominal inflection nInfl. The exponent is decided in each case by Vocabulary Items of the following type.11
11It should be noted that the effect of the notion declension class is derived in this system by assuming that sets like {√play, √horse, ...} are reusable. Thus, even if my Vocabulary Items do not make reference to features along the lines of Class-IV, the system allows for systematic correlations between nouns which take a particular nominative nInfl exponent and a particular ac- cusative exponent. We could equivalently write something like [class:IV], usingIVas a diacritic. A class is a diacritic phenomenon where the diacritic is a shorthand for set membership.
(61) a. nInfl[masc,nom,sg] ↔ -ur / {√play, √horse, ...}__
b. nInfl[masc,nom,sg] ↔ -i
The Vocabulary Item in (61a) states that the phonological exponent for the mor- pheme nInfl[masc,nom,sg] is-ur when nInfl is concatenated with the roots√play, √
horse, and some others. The Vocabulary Item in (61b) states that the same
morpheme is generally realized as -i. The choice between VIs is constrained by the
Subset Principle.
(62) PF Subset Principle (Halle 1997:128)
The phonological exponent of a Vocabulary Item is inserted into a position if the item matches all or a subset of the features specified in that posi- tion. Insertion does not take place if the Vocabulary Item contains features not present in the morpheme. Where several Vocabulary Items meet the conditions for insertion, the item matching the greatest number of features specified in the terminal morpheme must be chosen.
It is a consequence of the Subset Principle that the most specific match is chosen in each case. Note that the context counts for specificity so that a VI is selected if
it matches the features [masc,nom,sg] and also some aspect of the context like a
concatenated root, rather than just the features.
Consider the PF derivation ofleikur ‘game’ from the root√playwhich is given
head n which makes it into a noun and the nominal inflection morpheme nInfl has been adjoined to n. (63) a. n n √ play n nInfl [masc,nom,sg]
b. √play_n, n_nInfl (Concatenation)
c. √play_[n,-
∅], [n,-∅]_nInfl (Vocabulary Insertion)
d. √play_nInfl (Pruning)
e. √play_[nInfl,-ur] (Vocabulary Insertion)
f. √play-ur (Chaining)
g. /leikur/ (PF output)
The tree in (63a) is converted into concatenation statements which encode linear
adjacency (63b). Vocabulary Insertion starts by inserting a -∅ exponent for the
nominalizer due to a Vocabulary Item as follows, leading to (63c).
(64) n ↔-∅
A linear configuration in which two morphemes are connected via a zero expo- nent undergoes an operation of Pruning which removes the zero exponent from the representation (Embick 2010). This operation leads to (63d).
(65) Pruning
√
root_[x,-∅], [x,-∅]_Y → √
root-Y
Vocabulary Insertion proceeds. The-ur exponent is selected according to the Subset
Principle because it is the most specific candidate for insertion (63e). Finally,
the noun is as in (63g). The root might enter the derivation with a phonological form or it might undergo Vocabulary Insertion like other morphemes (Harley 2014). The issue will not be explored in the dissertation but whichever answer is correct, I will in general not show the VI step for the root in PF derivations.
I assume with Embick and Marantz (2008:6) that every root must combine with
a category-defining functional head. For this reason, a noun always has an n mor-
pheme and a verb a v morpheme, even if there is no overt derivational morphology.
(66) Categorization assumption
Roots cannot appear (cannot be pronounced or interpreted) without being categorized; they are categorized by merging syntactically with category- defining functional heads.
A category-defining head may vary in its interpretive realization depending on its
feature content and its context. In many cases, I only write n for nominalizer and
leave implicit the potential need for different flavors of this head. For example,
an agentive nominalizer -er in sing-er might contain a feature [+agent] or it
might receive a special interpretation in the context of agent-associated functional structure, e.g., a Voice head or something similar. When a special emphasis is placed on a particular flavor of a category-defining head, I use a subscript to indicate this. For example, a causative nominalizer which is introduced in Chapter 3 is referred
to realize special flavors as the vanilla n at PF unless they make reference to the relevant feature distinctions.