Prototypes and negations: one interpretation of Nowicki’s intuition

All social structures are made of meanings. So it is no surprise that a problem in the philosophy of language would wreck havoc in practical legal cases. Recently, Elizabeth Nowicki (from Concurring Opinions) has advocated a certain interpretation of the legal phrase, “not in good faith”, inspired in part by the recent Disney ruling. Her interpretation can be summarized as follows: “The “not in good faith” universe is larger than the “bad faith” universe; the phrases “bad faith” and “not in good faith” are not interchangeable. A plaintiff/shareholder can show that a director acted “not in good faith” by showing something much more “Caremark-esque” than “Enron-esque.” A director does not need to affirmatively act in bad faith (fraud, lying, deceit) in order to violate his obligation to act “in good faith.” She just has to. . . fail to act in good faith.” This interpretation sits between a rock and a hard place: the rock is (seemingly) mainstream logic, and the hardplace is natural language.

I think she’s right, after a fashion, and want to show how her underlying intuitions are onto something, both logically and pragmatically. But we have to start with an analysis of language before we make sense of the law.

Take this utterance as true:

(1) That film wasn’t bad.

What does it mean, logically? What does it mean, pragmatically?

In mainstream logic, there is a rule called “double negation”. It says that two “not”s which follow each other are logically identical to there not being a not in the first place. Take the following utterance:

(2) That film wasn’t not good.

It is synonymous with (1), because the meaning of “bad” seems to be “not good”. Now we can apply the rule of double negation. This isn’t a very special rule: all we need to do is take the two “nots” in (2) and eliminate them, and it gives us the entailment

(2`) That film was good.

This rule is intuitively supported by the fact that in conversation people tend to get tired of constant double-negations, since double negations require a few mental jumps to decode. Presumably, people would be satisfied if the speaker were just to say what they meant in a straightforward way, possibly much like (2`).

But natural language tells us that this interpretation (2`) of (1) is incomplete. While (2`) is one possible interpretation of (1), it’s not the only interpretation. Indeed, we usually mean something quite different when we say “That film wasn’t bad”: we mean to say that it was, basically, mediocre.

We usually assume that people are trying to make the most out of their utterances. If people choose to speak in a certain way, we usually assume that there’s a relevant reason behind the choice of words. So when we encounter a double-negation, like in (1), we assume that there’s a special secret message that the sender is trying to communicate. In this case, the special secret message would be, “It’s not bad, but if I thought it were any good I would have said so, so the only option left is that it’s somewhere between good and bad”.

We might be forced to say that natural language has failed us (as the logical positivists may have done), or we might be tempted to say that logic has failed us (as the logical intuitionists may have done). Though both of these options are plausible, after Paul Grice, they seem terribly outdated. Or we might claim that the negation of “good” is not in fact “bad”; rather, that it is “bad or mediocre”. And that might be appropriate for some contexts, but discussions concerned with law demand a kind of Manicheanism that makes a genuine third option inadvisable. So I don’t want to do any of these things. Rather, I’d like for us to have a more nuanced understanding of language.

Take any lexical word: say, “birds”. That word summons up all kinds of birds to the mind: geese, robins, etc. But some birds are unusual, and they sit at the periphery of the meaning of the word: for example, penguins, or ostriches. The word, “birds”, has graded membership. At the center are the common birds, the prototypical birds; at the outskirts are the anomalous birds who don’t seem to fit with the rest.

With that in mind, take the utterance:

(3) That’s not a bird!

uttered by someone in reference to some mystery animal. If they were to say that about a robin, they’d be clearly wrong. And if they were to say it about a penguin, then they’d also be wrong — but it wouldn’t necessarily be obvious or clear. So graded membership doesn’t seem to matter in that case. But the word “birds” deals with a well-defined class of things. On the other hand, when we start using terms like “good” and “bad”, the class is not well-defined at all; we’re stuck with intuitions about what makes up a kind of thing, but there’s semantic uncertainty.

In the case of the film, while you’re still stuck with two categories — the Good and the Bad — it would seem that you might be able to say that there are items at the periphery of each. Imagine you have a list of movies in your head: some which are clearly good (i.e., The Usual Suspects), some which are barely good (Jeepers Creepers), some which are barely bad (The Transporter), and some which are awful (24 Hours in London). [For the sake of argument, take it for granted that these films fall into the categories that I've listed, even if you have different tastes.]

Let’s revisit (1). If you were to walk out of Jeepers Creepers and say, “That movie wasn’t bad”, you would both be satisfying the intuition that it was mediocre (and satisfy natural language), and at the same time be in line with mainstream logic. Because while you’d still be really saying that “That movie was good”, it would be at the periphery of “good”.

We’ve gotten the idea by now. So let’s clear the table of these mundane statements and get back to the original phrase, “not in good faith”. Literally, this entails “not in non-bad faith”, and logically entails, “bad faith”. But there is nuance and latitude to what is covered by the meanings of “good faith” and “bad faith”; both terms can extend to cover what we’d ordinarily just think of as mediocre behaviors which we can classify as being, for all intents and purposes, between clear ideas of what make up good and bad faith behavior. But, because we need to be totally clear, we need to make it fall under one category and not the other.

This diagram makes it appear as though there is an in-between category (represented by “y”). And on first blush, we can attribute there to be one there. But because we need to be clear, we have to decide to move “y” into one category or the other. So negligence, for instance, might fall into the outskirts of the class we call “bad faith” behavior (call it “putatively bad faith” behavior).

On the surface of it, this might seem like a claim that is hostile to Nowicki’s. However, it is not. First, I have put my analysis in terms of behavior and not action, while she used only actions; this makes our accounts somewhat incommensurable. Second, to put my conclusions in her terms, the “not in good faith” universe is indeed larger than the “obviously bad faith” universe in relevant ways; however, the phrases “bad faith” and “not in good faith” are interchangeable in all relevant respects.

This interpretation may or may not satisfy Elizabeth Nowicki (it may, perhaps, be a bit too theoretical for the courtroom); and it may or may not satisfy most philosophers of language and logic. But it is, I think, at least one possibility to consider. It has the attraction of being able to accomodate intuitions from all sides.

[This post is indebted to Dan Sperber & Deirdre Wilson (relevance theory), L.E.J. Brouwer (logical intuitionism), Eleanor Rosch & George Lakoff (prototype semantics), and Kit Fine (supervaluationism and precisification).]

License

This work is published under a Creative Commons Attribution-Noncommercial 2.5 License.