You don't have to be a mathematician to love symbols.
As I was saying, this year began with the publication of a series of essays joined under the title Attention and Meaning and they explore the relationship between attention and language. One of the book’s essays, titled “The Evolution of a Hierarchy of Attention,” was authored by yours truly. Many of the things I say in the paper will be familiar to regular readers of this blog, but I was at last forced to work out the different capacities of attention required to get us from a generic great ape to modern speakers. I had had the project in the back of my mind for years, but this blog mostly builds on the work of others. It was only when I was invited to contribute to the volume that I finally got around to the work.
Starting with the assumption that natural language is a tool for drawing attention to topics, we can move on to the different evolutionary steps needed to support that tool.
Had from the beginning: Homo primus inherited a couple of attention systems from its Australopithecus ancestor. First was the posterior attention circuit, or reflexive circuit. Something in the environment brings an animal to attention. We can see this operating in our pet dog when we call, “Hey, Fido,” and the dog turns to look at us. Thus, there is no need to look for some special evolutionary power that enables us to attend to a subject of a sentence. The first talker could have said, “Lion,” and the auditors would have got it.
A second inherited system is the anterior attention circuit that lets an animal stay focused on a task while shifting attention. Thus, Australopithecus could shift attention from berry to berry without losing sight of the essential task of picking berries. The first talker could have expressed a predicate like, “Over there,” and the listener would have understood the subject.
Evidence of these assertions comes from (a) the fact that these circuits are common to primates and mammals generally; (b) apes can express subjects or predicates when they use sign language; and (c) toddlers can say these things.
At this level, the key difference between a great ape and a human is not in powers of attention. It lies in the motivation to share information and the ability to trust the auditor not to use the information to the speaker’s disadvantage.
Integrated circuits: The two attention circuits are separated in all primates excepting humans. We have a third circuit that connects the two attention circuits, making it easy for us to maintain a conversational thread and to attend to both subject and predicate at the same time. Probably, simple conversations and sentences are possible without the integrative circuit, but shared speech is much more efficient with the circuit. We therefore must assume that the first speakers were not terribly fluent and became more efficient only after some further evolution.
Working memory: If we take a sentence like, “The dog I bought yesterday has fleas,” we can find two topics combined into one sentence: topic 1, The dog… has fleas; topic 2, I bought [the dog] yesterday. The link between the two topics is the dog. We don’t have to repeat the topic, however, because we remember it. Working memory enables us to interrupt topic 1 with subtopic 2 and still understand the sentence when the speaker returns attention to topic 1. You could say, The dog—I bought a dog yesterday—has fleas, and be understood.
The second topic must be a subtopic of the first, that is working memory is overtaxed if the first topic is interrupted by a new one. Suppose I said, The dog—I bought a cat yesterday—has fleas. This sentence is much more confusing than the redundant one even though both sentences are the same length. Working memory does not just enable one to hold more things in memory. It requires a continuity of topic. We cannot talk about two distinct topics at the same time.
It is possible to understand the dog/cat sentence simply by ignoring the cat part as a confused irrelevance. Perhaps early fluent, speakers were able to handle relevant subtopics by ignoring them, but modern speakers require the ability to handle subtopics and that required the evolution of a working memory system.
Three hypothetical circuits: My paper also proposes the possibility of three other circuits which, as far as I know, are not yet discovered. They may reflect purely cultural processes and require no hardware, although in each case I suspect there is biological support for these universal powers of attention.
- Emotional awareness circuit: Speakers can direct attention to remembered feelings. Is there a language that provides no way of saying I am hungry? It would seem as basic as “pass the salt,” but it requires an awareness beyond the five senses that Aristotle spoke of. Can a well fed chimpanzee recall what it is to be hungry? I don’t know. Koko the gorilla was able to tell his keepers that he had a tooth ache, which suggests pretty strongly that Koko not only felt the pain but knew he felt it. So Homo primus probably arrived with some awareness of subjective states. It also seems reasonable to suppose that as we have become much more communal, we have grown aware of many states and subtleties that gorillas do not notice. Much literary art is devoted to finding ways to speak about unnamed subjective states, so it is possible that all the emotional awareness we have developed since language’s start has been the result of cultural labors, but that work probably had some biological support as well.
- Metaphorical substitute circuit: When speaking about ideas as though they were things—e.g., justice, God’s will—speakers typically need to direct attention to a physical substitute for the ideal. For example, “Your infidelity shattered my trust,” treats two abstractions, Infidelity and trust, as though they were things capable of acting and being acted upon. Hence, the verb shattered is offered as a substitute for whatever change actually occurred. There is no reason to suppose that apes can think anything similar, and indeed there are people who cannot free themselves from literal thinking enough to understand metaphorical sentences. I am reasonably confident that circuitry supporting metaphorical thought evolved after language began.
- Cultural emotions circuit: Culture, in this context, refers to a people’s symbolic environment. I seem to recall somebody, was it Terrence Deacon?, saying that we are sautéed in symbols. They define our environment. Other species are not even lightly sprinkled with them. Different populations of chimpanzees use different tools and have their own ways of making them, but it takes quite a tone-deaf philistine to think those contingent distinctions are a precursor of human culture. Cultural symbols draw attention to states that are neither physical nor subjective. A wedding ring, for example, points to a cultural status; a cross can alert people to a site’s sanctified state; other symbols remind people of a political power’s presence. Along with their cultural history, these symbols may rest on a biological history that primes humans to respect and be moved by them. They use of symbols assumes an emotional awareness and metaphorical understanding that certainly was not present when language began. Thus, archaeologists may have gotten it backward. Instead of seeing the rise of symbols as a marker of the start of language, they are more likely to indicate the completion of a long evolution of attentive powers required to support modern language.