Phonology: adjacent ordered rules (Kisseberth, 1970)

Question

I have a question when reading the paper Kisseberth(1970). It’s about when rules can be collapsed. My question is about "adjacent rules". I don't quite understand what "ordered before" and "ordered after" means.

The author says that for rules to be collapsed, not only the structure environment should be similar, the rules also need to be adjacently ordered.

‘That is, if two rules such as (4) and (5) are to be collapsed by standard notational conventions, there must not be some other rule which is ordered after (4) but before (5), or after (5) but before (4). According to the standard theory, the existence of such a rule blocks the collapsing of (4) and (5) into (6) - whatever the degree of structural similarity may be.’

(4)V -> ∅/__ V

(5)V -> ∅/VC[-long]CV

(6)V -> ∅/{ __V

       VC[-long]CV}

Thanks for help in advance!

Answer 1

One of the basic ideas in SPE-style phonology (which is the theory that's being used here) is that there's a long list of rules, which are applied in a "chain". The first rule in the chain takes the underlying form as input and produces an intermediate form as output, the second one takes that intermediate form as its input, and so on and so forth.

Rule X is "ordered before" rule Y if X comes before Y in the chain. Similarly, Y is "ordered after" X if Y comes after X in the chain. This is especially important for cases of "rule feeding". For example, imagine a rule that turns y into i, and another rule that turns i into e. In one order, ly turns into le; in the opposite order, ly turns into li. (Can you see why?)

Similarly, two rules are "adjacent" if they come next to each other in the chain. This is really only important for the purpose of collapsing rules: taking two similar rules and turning them into a single, more comprehensive rule.