Interpretive Speech

(Note: This post is the sixth in a series of posts exploring the contents of Jean-Louis Dessalles' book, Why We Talk.  For a summary of what came before see Why We Talk: Summary.)

The obvious difference between protolanguage (presumably spoken by Homo erectus) and normal speech is the presence of syntax. But what does that give us? If erectus says ‘me Ishmael,’ what does sapiens gain by saying ‘Call me Ishmael’? Dessalles asserts, “Syntax is not an obvious and expected extension of protolanguage, but rather an unexpected and peculiar development of prehuman communication” (p. 194).

A more technical difference between protosemantics and the semantics of human speakers is predication. “Protosemantics,” writes Dessalles, “is solely referential, whereas semantics is predicative” (p. 212). We can see the difference in the Ishmael variations above:

me Ishmael: This remark points. In fact, I find it hard to imagine somebody saying these words and not pointing to himself.

Call me Ishmael: This time I imagine the speaker first tapping my shoulder, for the reader is the unvoiced subject of this sentence. In this case I’m receiving instructions although more frequently sentences are about something. Sentences typically point at least twice first as a reference (you) and then something about the reference (call me Ishmael).

Protolanguage points; full language points and then points out. The one word leopard points and the natural reply is where. A sentence—that leopard is stalking impala—points to the leopard and then points out something about it.

Protolanguage can handle that second sentence. Alley Oop could say leopard and then say impala. The listeners look and draw their own conclusions. The syntactical sentence imposes an interpretation on the scene. Naturally we wonder what was accomplished by the interpretation that would make natural selection replace a protolanguage-speaking species with a different kind of talker.

There are two separate ways that languages handle predication. One, typical of English, uses word order. In that case a sentence like I have not yet seen him has a rigid word order. In that example yet could be moved to the end, but everything else is fixed (with a few rhetorical fiddlings possible). The other way to handle predication is through inflections, changes in word form. The Swahili version of that sentence is sijamwona, in which many syllables have been added to one word to indicate the same thing expressed through word order in English. As an aside, Dessalles raises the question of whether these two syntactical systems (word order and morphology) evolved separately.

One can imagine variety p of Homo heidelbergensis evolving a word-order-based syntax and variety q of the same species evolving an inflection-based syntax. Then you would have had the interesting case of two varieties of Homo unable to learn each other’s language (although both could fall back on a pidgin). If so, they seem to have crossbred enough to give their descendants both powers.

This kind of thinking is purely speculative, but if you don’t enjoy playing with ideas of this sort you are going to have a dull time pondering speech origins. So much of it consists of playing with whispy threads.

More pressing is the search for something to explain the introduction of syntactical expression. Homo erectus seems to have been doing fine with protolanguage. What changed? To discover what might have changed, Dessalles analyzes the “atoms of meaning” in proper sentences. This task is always a bit daunting because its essential ambition is to outsmart language.

Speech looks at topics from the outside and points out things about the topic. Behind that hill stands a fine old tree. In a sentence like that, we stand apart from the hill and therefore can point out something about the hill. I’m using a spatial example and am also deliberately using spatial metaphors (“from the outside,” “stand apart,” etc.) because the ability I’m describing is very much a navigational one. There are two ways to navigate through space: procedurally and improvisationally. Procedural navigation follows a route. Go to A then to B then to C. You need no understanding of space itself to navigate this way. Improvisational navigation lets you move about freely and appropriately because you understand space itself. It is as though you were somehow outside of yourself, not just aware of yourself but aware of yourself and your setting, so that you can move about appropriately even though you have never followed that particular route before and never seen anybody else follow it either.

The challenge then for Dessalles is to stand outside language (so he can point out the atoms of meaning) while working inside language to communicate what he has to tell us. It may not be as difficult as solitaire judo, but it ain’t easy. On the whole, Dessalles seems to do a good job, probably because he is alert to the navigational quality of our thinking. He quotes the German linguist Walter Porzig:

Language translates all non-visual relationships into spatial relationships. All languages do this without exception, not just one or a group. This is one of the invariable characteristics of human language. (p. 239)

And he uses this statement to justify the two particles he identifies that combine into his atoms of meaning. One is a frame or reference, the still point in the relationship. The other is the part that either moves or is located, called the theme.

In the sentence I used above—Behind that hill stands a fine old tree—the theme is the ‘fine old tree’ (thing located) and the reference frame is ‘behind that hill’ (still point). This sentence is literally about space, so of course the atom works well. A more abstract sentence was proposed by Derek Bickerton to argue that thinking is not visual: My trust in you has been shattered forever by your unfaithfulness. Dessalles system works well here too. The theme is ‘my trust in you’ (thing changed) and the reference frame is ‘your unfaithfulness’ (still point).

There are sentences that do not fit this pattern. An especially famous one is cogito ergo sum. The relationship it postulates is logical rather than spatial. Mathematical statements as a whole do not fall under Dessalles’ theme/reference analysis, but this point should not be surprising because experiments have shown that chimpanzees can be taught to use logical symbols. Surely Homo erectus was capable of at least as much logical thought as an ape.

Dessalles has done an excellent job of clarifying just what it is that is functionally new in language. Syntax does not just combine elements; it organizes the elements in a way that lets us interpret them as a kind of spatial gestalt. Protolanguage gives us images procedurally, one after another. Hill. Tree. Real language, combines elements (even abstract ones) into a spatial gestalt, putting seemingly unrelated items “my trust,” “your unfaithfulness” together into a metaphorical space.

Syntax provides a way of transforming space’s three dimensions into one, and syntax’s most abstract properties like hierarchy and recursion are the means of squeezing that spatial richness into a string of words.

Comments

[Brian Mihalic]

The point about language creating a spatial gestalt makes me wonder if this is related to the danger of having a cell phone conversation while driving. Recent research shows the problem is as much the distraction of the conversation as fumbling with the device. Perhaps our brains have difficulty navigating both a real and a linguistic landscape at the same time.
---------------------------
BLOGGER: Wouldn't that be an interesting finding.