Jason Ginsburg

535 Assignment 14

Pfau, pp. 101-171

Due: Tuesday, March 9, 2004

1. What evidence does Pfau say that Garrett gives to argue that, within the ‘formulation’ or ‘translation’ portion of the production process, there are discrete sub-levels?

According to Pfau, Garrett gives evidence from ‘exchange errors’. Word exchanges are usually exchanges of words that are of the same grammatical category, and they are usually words from different phrases. Because the grammatical category is usually the same, word exchanges appear to occur when syntax is being processed. Sound exchanges are exchanges of sounds from words of different grammatical categories, and these words usually occur within the same phrase. Because these exchanges occur between words of different grammatical categories, Garret argues that they occur separate from the processing of syntax. Garret also discusses exchange errors in which stem morphemes, but not affixes, are exchanged, providing evidence that morphemes are accessed separately from affixes. Also, because in example 4-3c (107) “the plural morpheme was produced correctly for the sentence as it was actually uttered,” it appears as though these exchanges occur before phonological specification of suffixes (Pfau: 106-107).

2. Garrett proposed two sub-levels. What does he call them? What happens at each of the two relevant sub-levels doing in the production process?

He calls them the ‘functional level’ and the ‘positional level’. The functional level is the level at which lemmas are selected. This level is associated with syntax, and it is the level at which word exchanges occur. The positional level is the level at which there is “retrieval of phonological form,” and it is the level at which sound exchanges occur (108).

3. Does a lemma, in Levelt’s model, contain the semantic features which distinguish concepts like CAT and DOG?

No, lemmas are only “syntactically specified.’’ However, the selection of lemmas “is triggered by semantic factors (115-116).” 

4. Pfau gives an example of a speech error that appears to support a Bobaljik-style analysis of do-support, according to which do-support is post-syntactic. What is the example?

The example is the bonsai didn’t die because I watered it, which the speaker intended to be the bonsai died because I didn’t water it. This example shows that the shift of Neg occurs before do-support, suggesting that do-support is post-syntactic. If Neg did not shift before do-support, then the bonsai not died because I did water it would be the expected form (127).

5. Because form-based substitutions happen after morphosyntactic agreement has been implemented, Pfau’s model predicts that we will never see accomodation with gender mismatches resulting from form-based substitutions. How many such examples does Pfau actually observe?

Pfau observes 1 example (1).