Meaning, Intention, and Mental States

Lawrence B. Solum, Artificial Meaning, 89 Wash. L. Rev. 69 (2014), available at SSRN.

This is a provocative and important essay that has implications Solum doesn’t spell out for some positions on meaning, communication, statutory interpretation, and the understanding—sometimes called the “construction”—of statutory texts. Solum is interested in communicative content, principally of directives. Most of his examples are of legal directives, or as he prefers to describe them, legal rules.

Solum begins with an important distinction between communicative content, the kind of meaning he is interested in discussing, and legal meaning, the legal contribution a text makes in its particular legal system. It is not uncommon for discussion of statutory interpretation to conflate the two or to focus entirely on the latter, but this is a mistake. In some legal systems, such as in the United States, the communicative content of a statute can cause it to fail to make any legal contribution (because, for example, the statute is unconstitutional). So getting clear on communicative content is a prolegomena to getting clear on much of statutory interpretation. Solum aims to make a significant contribution to this task by illuminating the lack of connection between communicative content, intention, and the mental states of individual legislators.

The core of the article is a thought experiment Solum calls “The Chinese Intersection,” in which an artificial intelligence system for a supremely complicated traffic intersection in Shanghai is described. The Shanghai Artificially Intelligent Traffic Authority, or SAITA, is created to govern a busy intersection involving several kinds of transportation systems. SAITA is programmed to monitor, on a moment-to-moment basis, vehicular, rail, and pedestrian traffic and, in order to insure the traffic’s smooth and safe flow, to alter signs, traffic lanes, and speed limits. SAITA can also make legal changes to the traffic regulations, post appropriate signs, and create legal texts. It can invent symbols for behavior newly outlawed or regulated; moreover, it has the capacity to create YouTube videos and engage in other campaigns of public education to inform about new regulations, new vocabulary, and so on. We are invited to envisage such a system operating successfully.

The beauty of this thought experiment is that if you find it coherent, you are forced to reject a number of philosophical claims, and to reevaluate others. For instance, you are forced to reject the conjunction of two claims: (1) marks don’t mean anything if not made with an intention to mean something (a view held by literary theorists Steven Knapp and Walter Benn Michaels, though Solum doesn’t note this), and (2) intentions are mental states. You are also forced to reject this conjunction of views: (2) intentions are mental states, (3) communication is either successful or unsuccessful relative to a communicator and a recipient (hearer, reader, or viewer), and (4) successful communication requires that the communicator’s “communication” intentions be recognized by the recipient. By hypothesis, SAITA has no mental states. Yet its marks, even when vocabulary has been invented for new offenses in signage and legal texts, do mean something to the pedestrians, motorists, train crews, etc., who navigate the Chinese Intersection. Moreover, SAITA’s success, especially where it is truly innovative, seems to depend heavily on its ability to communicate to the appropriate addressees, sometimes rapidly in real time, the changes it introduces.

Even if one rejects (2), the Chinese Intersection invites reconsideration of philosophical claims influential in discussions of statutory interpretation. For example, if there is nothing that can be an intention of SAITA (and is there?), one must reject claim (1), that there cannot be “intentionless meanings.” In addition, one must reject claims (3) or (4). The conjunction of both is a Gricean-inspired view too frequently uncritically adopted by those who think the communicative content of statutes depends entirely on legislative intention.

All Solum explicitly aims for in this article is the rejection of the contentions that communicative content is constitutively determined by the mental states of individual human minds, and that this is especially true about legislation. To bring the thought experiment closer to legislation as we know it, he alters the thought experiment so that SAITA is replaced by a group of individuals known as GAST, in which various teams take on various sub-tasks of SAITA, including the production of different sections of the traffic code. The story is that frequently, a text of new regulations is approved and implemented without any single individual having read the entire text. Solum treats it as a short step to the conclusion that meaning in law—that is, of legal texts—is just as “artificial” as the meaning produced by SAITA and GAST. There is the further suggestion that this meaning is semantic meaning.

I find this part of the article less successful for a number of reasons, but its success isn’t all that important. Sophisticated proponents of the view that the meaning of a statute is determined by some sort of legislative intention do not maintain—and, indeed, sometimes emphatically deny—that this legislative intention is reducible to the intentions of any single individual, whether understood as a mental state or in some other way.

There is, to be clear, no claim in this article to tackle the intentionalist view in its sophisticated form(s); Solum aims instead to refute what he calls a “folk theory” of meaning. As I’ve said, I think he achieves much more. Even those who do not share his intuitions about the presence of communication by SAITA are well advised to grapple seriously with his Chinese Intersection thought experiment.