Feeding/counterfeeding/bleeding/counterbleeding.

I have a confession to make. I never really felt that I understood this terminology.

Perhaps it had something to do with the circumstances under which I took John McCarthy’s phonology 101 class. It was at the ungodly hour of 9:30AM! In those days I didn’t own an umbrella, so whenever it rained–and it rains pretty often in Amherst–I would lie in bed waiting for the rain to stop before heading off to class. So obviously my attendence was not what it should have been.

But whatever the reason I have always been in denial about this terminology. Which is why Eric’s recent discussion touched a nerve. I finally decided I needed to confront these demons and put them to bed once and for all.

The idea which is at the heart of Eric and Colin’s bet is that there is a verb feed which describes the action of one rule upon another, and to this we can prefix counter- to get another verb which also describes the action of some rule upon another one.

I don’t think things work this way. In fact, I don’t even think this makes sense.

The prefix counter- apparently comes from an Old French source, and probably made its way into English through words like counterfeit or countermand where the semantic decomposition is not really transparent, but the general meaning “against”–the OF affix is from Latin contra and became ModF contre–is nevertheless apparent. However the exact meaning of the prefix in English is not completely clear. Think of cases like balance/counterbalance or sign/countersign, where the “counter-” forms describe actions where there is some sense of opposition, but the exact relation is not really consistent.

One more important point about this prefix is that it isn’t limited to verbs. It is just as likely (much more likely?) to be used with nouns (countermeasure, counterpart, counterintelligence) or adjectives (counterproductive, counterintuitive). In fact counterfeeding and counterbleeding are, I would claim, instances of the latter. Feeding and bleeding are really adjectives used to describe certain types of rule interaction; counterfeeding and counterbleeding are adjectives that describe the ‘opposite’ orders. If the verbs counterfeed and counterbleed exist at all–something that I doubt–they would be backformations.

This still leaves the question to clarify the meaning of ‘opposite’ that is at work here. Feeding orders need to meet two conditions:

1. Rule A precedes Rule B
2. The applicability of Rule B (in some forms) depends on the application of Rule A

Note that this is not meant to be a precise definition. For that the vague terms ‘application/applicability’ would have to be better specified. Bleeding order has the same two conditions–except that the effect of the application of A is different.

But now consider the definition for counterfeed provided by Kentsowicz (1994), and quoted in Eric’s post:

counterfeed means “fails to feed”

This is a real clunker. A rule can “fail to feed” if it fails to meet either condition 1, or condition 2, or both. But the only relations we really would consider counterfeeding are those that meet 2, but fail to meet 1, i.e., those where there is interapplicability between the two rules, but the ordering is wrong. To ask in such a situation, which rule “fails to feed” the other is an entirely pointless exercise. It’s as if you eat desert before the main meal. The two courses are out of order, but there is no sense to asking which of the two is out of order. Or, in fact, which of the two is ‘doing’ the wrong order to the other.

Don’t get me wrong. I am not saying it isn’t possible to define that one of the two rules is the doer and the other the doee. But I don’t think this is useful. I would maintain that any definitions offered for the putative verbs counterfeed and counterbleed in any phonology textbook are arbitrary stipulations by the authors themselves and are not grounded in any actual facts of phonology, or usage by phonologists. Responsible phonologists should not use these terms.

So what should one use? Well if it can’t be helped, just say “Rules A and B are in a counterfeeding relation.” But I really don’t know why one needs these terms, and I wouldn’t bother to teach them to students. The class of counterfeeding rule relationships does not identify any known natural class of phonological objects.

Where does the terminology come from? I hate to blame anyone, but I generally associate the terms with the work of Paul Kiparsky. In fairness to Kiparsky, the insight he was really trying to capture was that opacity in phonology is unnatural. Considering the paradigm he was working in, this was a remarkable insight, and this terminology must have seemed like the best way within that theory to come to grips with this insight.

But that was then. We have since moved on. Counterfeeding, counterbleeding, R.I.P.