Consider the following prefixed words (from Bauer, Lieber, and Plag 2013: 600, data from COCA). Each word has two prefixes. First use the online $O E D$ or another dictionary to identify which prefixes are native and which are non-native. Then discuss what problems, if any, these examples pose for the Stratal Ordering Hypothesis.
ex-betrothed
hyper-unemployment
mini-outbreak
polyunsaturated
post-midnight
1. Aronoff is aware of examples like these, and is forced to argue that there are two different -ize suffixes and two different -er suffixes that are homophonous, that is, that sound identical. It is generally accepted, though, that this is not a strong defense of the Unitary Base Hypothesis, and that the hypothesis is therefore almost certainly incorrect.
2. For the moment, we can set aside how the phonological information is stated in this sort of theory.
3. This is a bit of a simplification. Although WP is realizational, other kinds of morphological theory are realizational as well. See Chapter 11 for two examples.
4. In (16) ' $e$ ' indicates that the subject NP is empty, and ' $t$ ' stands for 'trace', which marks the place from which a constituent has been moved.
5. In COCA, examples with these suffixes are sometimes hyphenated and sometimes not. Here I have preserved the hyphenation found in COCA.
6. In the literature on morphology, the term Level Ordering Hypothesis has frequently been used, for reasons we do not need to go into here.
7. Note that the term base in this context is used in a broader sense than I have used it in this book. For Plag and Giegerich the base of a word is whatever simple or complex form an affix attaches to.
8. This is of course itself a complex word, but as its internal structure is not relevant to the issue we're discussing here, we will ignore that internal structure.