Week 6: Rules

English Morphology

Fernanda Barrientos

2024-11-26

Recap: Morphological relationships

Rules

• Sometimes, word-formation processes are very clear-cut: a morphological meaning can be easily mapped to a form.
  • This is usually the case of affixation
• Other times, patterns can be less clear, such as ablaut, templatic morphology, reduplication, etc.
• This difference is also known as concatenative vs. non-concatenative morphology

Two approaches to morphological rules

• Concatenative morphology is far more common as a word formation process, within and across languages
  • A morpheme-based approach to rules favors a view of language as a linear process
• But since non-concatenative processes are not possible to analyze as a sequence of morphemes, we need a different approach for them
  • A word-based approach is a wider, more encompassing framework

Morpheme-based approach 1: rewrite rules

• A word form can be represented by rewrite rules.
• For that, we need first to make generalizations of the following type:

Morpheme-based approach

• Considering the generalizations above, we can form rules of the type A \(\rightarrow\) B + C, where A is the word form, and B and C the morphemes that form it.
• … Which allows us to analyze morphologically complex words thus:

Morpheme-based approach 2: lexical entries by morpheme

• … Likewise, we can assume lexical entries for morphemes which show:
  • The phonological form
  • The combinatory potential of affixes, or the word class
  • The rough meaning

Morpheme-based approach 3: trees

• We can use a hierarchical structure in the form of a tree, just as syntactic theory does

Word-based approach to rules

• As stated above, this model allows us to express both concatenative and non-concatenative models
• The lexical entries for word forms give the phonological form, the word class and the meaning
• We can express generalizations through a word schema
  • But before knowing what word schemas are, we need to see what lexical entries look like within a word-based approach

Lexical entries

• Unlike the lexical entries assumed in the morpheme-based model, each entry consists of a full word form
• We have no information on combinatory potential because the entries are not separate morphemes

Word-based approach to rules: word schema

• The advantage of this approach is that it captures generalizations between forms
• A word schema for English plurals will look like this:

Word-based approach to rules: word schema

• … Which in turn allows us to make the following generalizations.
• The idea is that the relationship between two sets of words (singulars and plurals) is expressed in a relatively simple way:

• This is called a morphological correspondence: to each singular there is a plural, and vice versa (hence the double-pointed arrow)
• The first part is the word schema for any singular noun, whereas the last bit is the one for plurals.

Word-based approach to rules: word schema

• … And as promised: this works quite well with non-concatenative patterns too! 😎
  • (3.28) shows how a causative is formed in Urdu via vowel shortening
  • (3.29) shows plural formation in Somali through partial reduplication

Summary

• Morphological patterns can be captured in different ways, which implies different approaches
  • The morpheme-based approach (concatenative in nature)
  • The word-base approach (deals with both types of patterns, even though it may be an overshot).

Next week

• Read Haspelmath and Sims, Ch. 3
• Finish the exercises
• Attend the tutorial