I Speak, Therefore I am: Seventeen Thoughts About Language

If I told you you were swallows, you would immediately object that that is not true. But then if I asked you what it means to say that it isn’t true, the answer would not be so immediate. But everything is clear. Modern linguistics, armed with the weapons forged from logic and mathematics, tells us that a fundamental aspect of the meaning of a sentence like “All the readers of this book are swallows” is represented as a relation between two sets, the set of readers of this book and the set of swallows—a relation in which the set of readers of this book is contained in the set of swallows. To say that the sentence “All the readers of this book are swallows” is true or false then amounts, at a certain level of representation, to the possibility of saying whether this combination of the two sets associated with the sentence is correct. If the sentence had been “All the readers of this book are not swallows,” on the other hand, this would have been a relation of separation: no individual in the first set is also in the second.

Aristotle tied the notion of truth—and therefore falsity—to language so closely that not even today, 2,300 years later, can we get away from treating the problem in these terms, although today the approach is constrained and enhanced by new techniques (see Chierchia 1995, Chierchia & McConnell 1995 for references). But, as everyone knows, Aristotle was Plato’s pupil, and for Aristotle the bond between language and truth goes further, exploiting the knowledge of the harmonic fit between noun and verb recognized by his teacher. In fact, Aristotle said that there are no combinations or separations—and hence truth or falsity—without the combination of noun and verb, or, more precisely, subject and predicate. It seems that we are saying the same thing using different words: “subject” instead of “noun,” and “predicate” instead of “verb”—but it’s not so. The analogy between noun and verb on the one hand and subject and predicate on the other breaks down when we bring into consideration, in Greek and other languages that have it, the verb “to be.” Why? Let us proceed step by step. When he tried to find a way to define a sentence, Aristotle realized that it isn’t possible to give a general account: sentences can be used, for example, to order, to ask, to beg, to pray, to be ironic, to hypothesize, to implore, and to describe. So, for various reasons connected above all with his interest in deductive reasoning, Aristotle concentrated on this last fundamental capacity of language and stated that a sequence of words is a sentence only if it means something true or false. So, “that woman thinks” is a sentence, but “that woman who thinks” isn’t. In this way, he goes on to put logical flesh on the Platonic bones of noun-verb agreement. He distinguishes between that which is attributed and that to which something is attributed. That something to which a property is attributed is the thing that exists underneath all the properties given to it: it is “the thing that stands underneath,” “the sub-stance” or, in linguistic terms, the “subject.” These are all terms whose etymology reveals a “stereoscopic” view of reality. And for the property attributed to the substance Aristotle uses the term designating the accused in a trial or in speeches in the public square, the agorà: categoroúmenon or, in Boethius’s Latin translation, “predicate,” to speak in front of others. Subject and predicate, then, are the two pillars of linguistic, logical, ontological, and mathematical thought: without these, there would be no rational thought. And, as in the case of harmonic agreement between noun and verb, no one has managed to derive these properties from simpler facts (yet).

But now let’s go back to the verb “to be,” if only very briefly. Aristotle saw that in order to have a sentence a subject and a predicate weren’t enough; a verb is needed too. But why? Aren’t verbs predicates? No: Aristotle realized that in his language (and in our too) not all verbs are predicates, or rather that there is one verb that is not a predicate of anything, a verb that does not express any kind of property—the verb “to be.” This is a verb whose only function is to express tense, which is obligatorily applied as a third element alongside a subject and a predicate to make a sentence when tense cannot be expressed by means of the predicate. For example: just as I can say “A picture on the wall was the cause of the riot,” so I can say “A picture on the wall caused the riot.” In one case the verbal inflection (-d) expresses tense along with the property (cause-); in the other, since the predicate is a noun (cause), in order to express tense we need a verb “empty” of meaning, but nonetheless able to indicate tense (was). But the story of the verb “to be” is long, tortuous, and complicated, involving the logical duels of the Middle Ages, the battle over the mechanical nature of thought a few centuries later in the Baroque era, and leading up to the mathematics of the nineteenth and twentieth centuries. This story would take up at least a whole separate album (see Moro 2010 for a history of “be” and the appendix of Moro 1997 for a sketchy version).