Let’s take three phrases: “yellow truck,” “yellow dress,” “yellow flower.” On the face of it we have one constant—“yellow”—and three variables—“truck,” “dress,” “flower.” I’d like to suggest that this is an illusion, and that there are in reality four variables here, including “yellow.”
To make things harder for us, I’m going to assume that in all cases we’re looking at what, technically, we could call the same yellow, say primrose yellow. Nevertheless, I am not convinced that “yellow” in “yellow truck” is exactly the same word as “yellow” in “yellow dress.”
It appears as if “yellow” the constant has a modifying effect on “truck” the variable. And that is certainly clear. A truck is just a truck until the word “yellow” gives it a nice shiny coat of paint. But what has happened to “yellow”? Suppose for a moment that our truck is the first yellow truck the world has ever seen. Now, for the first time, “yellow” is something that can apply to “truck.” This modifies the meaning of “yellow.”
There is a distinction here. While the modification of “truck” by “yellow” is unmistakable, even dramatic, the modification of “yellow” by “truck” is minuscule, barely perceptible. (In fact, from an everyday point of view, completely imperceptible.) Nevertheless, “yellow” has expanded its meaning. It has been modified by what it modifies. This is what I would call “semantic reciprocation.”
The idea of semantic reciprocation is that there are no constant meanings. All words live in a state of dynamic and unstable tension with each other. And all meanings impact all other meanings.
The meaning of each word is driven entirely by context—a dual context of the real world situation in which it is spoken, and the surrounding words. And because no context is identical to any other, the meaning of every word is altered by every new use. But with some historic exceptions—for example, “digital”—this alteration is unimaginably tiny: we cannot observe it, except cumulatively, over long periods of time, when retrospectively we notice that the meaning of a word has drifted. (So we have here a process of evolution, somewhat akin to Darwin’s.)
We think of ourselves as remembering words. We can also think of words as a kind of memory. Each word is the memory of all its uses to date. And just as we are changed by the new memories we accumulate, so are words.
This train of thought was prompted by reading a book about epistemology, Esther Meek’s Longing to Know. In her book, Meek runs together the meaning of “know a fact”, “know a skill” and “know a person.” For her these are all cases of “know”—a constant that she can then analyze with confidence. To my mind, the error here is somewhat crass. (One wonders, is there a French edition?). But there is a subtler and more informative error inside just one of these three sets, “know a person”. Meek claims that she “knows” God in the same way that she “knows” her auto-mechanic. To my mind, the meaning of “know” is modified by its object in these different cases, rather significantly. And that punctures the attempt to know what we mean by “know” in some general or constant way.
Although it is slow in its effects, semantic reciprocation is immediately visible to a small sector of the population, poets and their readers. If you write a poem, you will discover that changing one word alters the meaning of all the others. That’s because poetry, uniquely, instills a heightened awareness of the meaning of each word, and—here is the point—of the inherent instability of meaning.
* * *
FOOTNOTE: It might be argued that what is altered in the poem is not the “meanings” of the other words but something hazier, perhaps their tonal value. This evokes a familiar distinction between connotation and denotation. In a future post I’ll try to show that the distinction is spurious. There is only connotation, of which what we call “denotation” is a peculiarly condensed form.