Here's a desperately needed canon of construction

The Choose Life Canon: If a statute is ambiguous, and interpreting it one way will save many more people's lives than interpreting it the other way, interpret it so it saves more people's lives.

For an illustration, see FDA v. Brown and Williamson Tobacco Corp., 529 U.S. 120 (2000). The FDA had interpreted Congress's delegation to it of authority to regulate "drugs" and "devices" to include the authority to regulate tobacco products such as cigarettes. Consequently it established regulations aimed at reducing the likelihood that children would take up smoking. The tobacco companies sued, arguing that that there was no Congressional grant of regulatory authority to the FDA.

The arguments included many interpretive rules, and on the basis of several of these, the majority sided with the tobacco companies. On the basis of several other rules, four justices dissented. Underlying parts of Justice Breyer's dissent was an idea that looks like a more particular instance of the canon I'm proposing: "In my view, where linguistically permissible, we should interpret the FDCA in light of Congress’ overall desire to protect health."

Maybe if the majority had accepted the Choose Life Canon, hundreds of thousands of lives might have been saved.

I suppose the principle can be generalized, as follows:

If a statute is ambiguous, and interpreting it one way causes greater social benefit than interpreting it the other way, interpret it so it causes greater social benefit.

The philosophical basis for such a canon is pretty straightforward: the purpose of the law is the common good, so the courts, which uphold the law, should err on the side of the common good.

There is also a division of powers rationale: The purpose of the legislature is to promote the common good, so the court ought to suppose that the legislation is aimed at the common good.

A paradox in bankruptcy law

And now for something completely different: a possible legal paradox.

(The further thoughts I promised on the SCAT hypothesis are coming soon.)

Consider the following situation:

A debtor walks into a lawyer's office and declares that he's bankrupt. The lawyer advises him that a declaration of bankruptcy might be more legally effective if done in a formal legal proceeding. The debtor says okay, and asks the lawyer to represent him in such as proceeding. She agrees. The debtor and the lawyer sign a contract where the lawyer agrees to represent the debtor for a flat fee of $2000, payable in quarterly installments of $500, with the entire agreement being subject to the court's agreeing to appoint the lawyer as the debtor's representative – a necessary procedure under bankruptcy law.

The lawyer petitions the bankruptcy court to be appointed as the debtor's representative. The court reasons as follows:

If we appoint the lawyer, she becomes a creditor of the debtor she's representing.
If she's a creditor, she has a potential conflict of interest, since her duties to the client include trying to discharge as many of his debts as possible.
Since a lawyer cannot be in such a potential conflict of interests with her client, we cannot permit the appointment.

But by not appointing her, the court prevents the conflict of interest from arising. And in the absence of a potential conflict of interest, there is no reason not to appoint the lawyer, and the court ought to appoint her.

And repeat.

Of course, the court has the discretion to disqualify the lawyer for any other good reason, including that representation would give rise to a paradox. But this gives rise to the same problem: if the court doesn't appoint her, it does not give rise to a paradox, meaning the court should appoint her.

This may fall short of a genuine paradox, since the lawyer probably does not have an actual right to be appointed, subject only to a good reason not to appoint her. The court would certainly be right to disqualify her. I just don't know why.

The SCAT hypothesis of meaning and the textualist fallacy

The SCAT hypothesis of meaning holds that the components of meaning are Structure, Context, And Text, not necessarily in that order. I name this hypothesis and make it explicit because it is common to suppose, to the contrary, that meaning is determined only by text. This erroneous supposition I call the "textualist fallacy."

It is easy to show that meaning is more than the sum of the words in a text. The fact that "Oscar is taller than Nigel" does not mean the same thing as "Nigel is taller than Oscar," despite the fact that they contain exactly the same words, is proof enough. This, however, can be explained by invoking linearity rather than structure or context, so it does not prove that structure and context are components of meaning.

But there is abundant evidence that structure contributes to meaning. For example, the phrasal ambiguity as in "old men and women adore me" (the subject is either old men and all women, or old men and old women) is not explained by word order, but is explained by assuming an internal structure for the sentence. If "old" is grouped with "men" and the result grouped with "and women," then we get the first meaning. If, on the other hand, "men" is first grouped with "and women" and the result is grouped with "old," we get the second meaning. Since different structures give rise to different meanings, structure must play a role in determining the meaning.

There is also abundant evidence that context contributes to meaning. In "John's pen," for example, we don't know the nature of the relationship between John and the pen until we have some context. In a context-free environment like I just presented, it is natural to interpret the relation as ownership, such that John owns the pen. But put in other contexts, many other relations are possible. John could be the creator of the pen, the inventor of the kind of pen, the person who is currently holding the pen, or even just the person who pointed to the pen.

The examples of the contribution of structure and the contribution of context are single examples of phenomena that are pervasive. To see the pervasiveness of context, take the meaning of any extended text, comparing it to the meaning of each individual word, for example by looking it up in a dictionary. Many words will have multiple definitions (and even the best dictionaries significantly understate the number of senses of many words), but the extended text will permit far fewer meanings than the product of the definitions of the words. This is because context will eliminate many of the theoretically possible meanings.

To see the pervasiveness of structure, read an introduction to linguistic analysis, such as Steven Pinker's The Language Instinct.

The SCAT hypothesis represents the narrowest possible theory consistent with the basic realities of language. It may be that other factors, like linearity, will have to be added.

The textualist fallacy is not something I just invented for pedagogical reasons. People fall for it all the time. Even legal scholars fall for it. In separate posts I will show how a recent scholarly paper falls for the textualist fallacy, and I will suggest that even the Supreme Court fell for it in a landmark opinion.