Analogy like syllepsis
Thomas Aquinas argues that everything true we can say about God, we can say only analogously. I’ve noticed that people have trouble telling the difference between this claim and a much stronger one, namely, that everything we can say about God is, ultimately, false, even if useful for devotional purposes. Such an apophatic attitude tends to raise the hackles of more down-to-earth philosophers: if there’s something we can’t talk about, why talk about it? And these philosophers extend their approbation to Thomas as well. This seems to me unfair, for his position is really quite different from a simple quietism.
But I can also see why people are confused. They doubt Thomas’ answer, not to a question about God, but to a question about talk: of what philosophical use is analogical language?
As we use the word nowadays, an analogy is a ratio, a relation of one relation to another: HAND : PALM :: FOOT : SOLE. We express the mapping with words like “like”: “Hands are to palms as feet are to soles”; or, “Palms are like soles the way hands are like feet.” Such comparisons often feel informative. But what do they really tell us?
If analogies tell us that two things are alike, we might think that they let us draw conclusions about one based on what we know about the other. For example, when we realize that hands have fingers as well as palms, we can ask: HAND : PALM : FINGER :: FOOT : SOLE : ___? And we fill the blank in with TOE. But what about HAND : PALM : FIST :: FOOT : SOLE : ___? Nothing can fill in this blank; the analogy breaks down. So the analogy does not, in fact, let us draw any conclusions. It merely tells us that two things are in a certain respect alike, and prompts us to ask whether or not they are alike in some other way.
These kind of analogies are obviously useful for thinking; they help us come up with interesting hypotheses, for example, that the foot will have a feature corresponding to the hand’s finger or fist. But it seems as if they can’t do argumentative work. If someone denies, for example, that toes are like fingers, we might bring to his attention the analogy HAND : PALM :: FOOT : SOLE, and then ask if we cannot append … : FINGER :: … : TOE. Perhaps he will say yes. But he may say no, and pointing to the analogy is not enough to prove him wrong.
What would it even mean for him to be wrong? Suppose he does say no. If we ask why, he may reply: fingers are opposable, and toes are not. We of course agree that … : THUMB :: … : BIG TOE is not valid, or, at least, seems fishy. But we still want to say that fingers are like toes, and he still does not. What exactly are we disagreeing about? It seems that we’re disagreeing over whether or not to use use words like “like” to connect them; it’s not clear that we disagree about anything else.
This kind of analogy is just a way of putting similes and metaphors. “Achilles was a lion on the battlefield” is the same as ACHILLES : BATTLEFIELD :: LION : HUNT. Metaphorical language is essentially evocative, not meaningful; it prompts thoughts, but does not communicate thoughts in a way that allows disagreement. We cannot appeal to a metaphor in argument, we can only call it to our interlocutor’s attention; and we cannot disagree with a metaphor, we can only call it unhelpful.
This is the view of analogy at which we arrive when we see it as essentially a ratio, as A : B :: C : D. But in doing so, we lose sight of the analogical use of a word in which Thomas is so interested.
Analogical use stands between univocal and equivocal. It’s easy to see what is meant by the latter two. Take the sentences “I’m a fan of the Texas Rangers” and “You’re a fan of the Chicago Cubs.” These uses of “fan” are univocal because they mean it in the same way. We could combine them as follows: “We’re both fans: me of the Texas Rangers, you of the Chicago Cubs.” Take, on the other hand, “I took to the ballpark a fan of the Boston Red Sox” and “I took to the ballpark a fan to cool off with.” These uses of “fan” are equivocal because they mean it in unrelated ways. There is no valid way to combine them; the sentence “I took to the ballpark two fans, one of the Boston Red Sox and one to cool off with” is clear nonsense.
Now consider the sentences “His body was healthy,” “His food was healthy,” and “His saliva was healthy.” These uses of “healthy” are analogical because they mean it in different, but related ways. A body is healthy when there’s nothing wrong with it; food is healthy when it makes its consumer healthy; and saliva is healthy when it indicates that its salivator is healthy. The first use is primary; the second relates to the first causally; the third relates to the first symptomatically. These kinds of relations are termed an analogy of attribution.
Other kinds of different, but related uses exist in which there are not some uses whose meanings refer to other uses, but rather a commonality between all of the various meanings; these are called analogies of proportionality. I can “give salt,” “give an idea,” and “give counsel,” and these are different kinds of giving, but they have something in common. We might say: to give something to someone is to make it now his; sometimes giving results in my no longer having, as with salt; sometimes giving results in us now both having, as with an idea; and sometimes I can give that which I never have myself, as with counsel. So we can draw a proportion between the things we give: LOSE : SALT :: RETAIN : AN IDEA :: NEVER HAD : COUNSEL.
What seems to me significant about analogies, when described thsi way, is that sentences like “His body, food, and saliva were all healthy” and “Give neither counsel nor salt till you are asked for it” sound strange, but still make sense. Both are examples of syllepsis, a figure by which a single word is made to connect to two or more other words in the sentence, applying to them in different senses. A few more:
- She made no reply, up her mind, and a dash for the door.
- Eggs and oaths are soon broken.
- She went straight home, in a flood of tears and a sedan-chair.
- God and his creation both exist, are true, are good, ….
A thesis: Analogical language makes sylleptic language possible.
A corollary: To reject analogical language is to deny that sylleptic language is grammatically permissible.
Why should we permit syllepsis in our grammar?
Syllepsis is a special form of zeugma, and zeugma is, essentially, an application of the distributive property of language: A*B+A*C=A*(B+C). If we reject zeugma entirely, we can never again say “I went to Texas and California,” or even “I went to Texas and went to California”; the only way to express this thought would be “I went to Texas and I went to California.” This would be far too restrictive; we must allow some distribution. But we should reject equivocal zeugmas, which are, essentially, attempts to distribute across words that are not in fact the same–like concluding A*B+E*C=A(B+C) just because A and E sound alike.
To permit syllepsis, then, is to say that the two uses of, for example, “healthy,” in “healthy body” and “healthy food,” are the same word, not just similar words, even though the uses of the word are similar but not the same. This at first glance looks strange. Why would not an ideal language differentiate between these, for example, by using “healthy” for one and “healthful” for the other, such that we could recognize the similarity from their common stem, and yet recognize their dissimilarity from their differing suffixes?
To be sure, appending differing suffixes is often worthwhile, helpfully disambiguating the language. But we should not imagine that it would be possible in all cases. This would assume that we can easily differentiate between the various meanings of a word and assign a separate suffix to each; but to the contrary, nothing assures that these meanings are clearly separated, or finite in number, or organized along any particular axis. There is no reason to think that any system of generated-on-the-fly suffixes, or of modulated pitch, speed, volume, etc (or size, angle, color, etc), would suffice to communicate exhaustively the shades of meaning contained in our words. No notational system, however complex, can make it possible to mechanically extract the meaning of a sentence from its representation.
Syllepsis is not, I would say, an abuse of zeugma, any more than analogy is an abuse of the fact that we can use a word more than once. They yoke together disparate meanings of a word in a way that makes us uncomfortable, and gives us the feeling that, perhaps, this one word ought, in fact, to be two different ones; but if we split words whenever we have such a feeling, our language would grow more complex without becoming any less ambiguous.
There remains, to be sure, a further question: why, in the particular case of God, we would not be better off eliminating the analogical language? We could say, for example, that “Creation exists, while God exists*.” This would, in the case of God, seem especially helpful: it would allow us to say “God exists* but does not exist,” rather than affirming “God does and does not exist,” an explosive contradiction–or at least it would explode, if it were permissible to mechanically draw inferences from a sentence’s representation without considering what it actually means.
There are, I think, arguments to be made in favor of retaining the same word. But more importantly, to even ask this question is to stop debating the philosophical question whether God does and does not exist, and to begin debating the practical question whether this sentence is the best way to express this fact. The stakes are certainly much lower.