Post #300. The counter from the folks at Typepad.com tells me this is my 300th post. In words it adds up to a full novel by Charles Dickens. My persistence on this beat has only been possible because serious papers relevant to speech origins keep appearing. I have not yet hit the wall and hope I can keep on running.
One of the persistent mysteries of speech origins is the growth of complex sentences. Chomsky and his disciples argue that language began with full syntax and complex sentences in place. Their point is that baby speech—one, two, or three words combined without syntax—does not provide the kind of running start that makes syntactical speech any easier than it would be if you just jumped straight to the syntax. Thus, there is no reason to assume that the human lineage ever went through a stage of speaking baby talk or pidgin (also known as protolanguage). The logic is strong, but the penalty for this argument is that it ignores the fact that speech in children does begin with simple sentences and it forces theorists to abandon the idea that language evolved as a form of communication between individuals. Instead it began as a tool for superior thinking. (See: Chomsky’s Theory of Language Origins)
This blog presses the view that words pilot attention and that speech is a tool for creating joint attention. That idea does nicely with basic utterances, especially childish sentences like biggie doggie, but it does not really explain how or why we got around Chomsky’s objection and began spouting sentences with a subject and a verb. So I was happy to see that the January issue of Experimental Brain Research has a paper by New Zealander Michael C. Corballis that argues “grammatical language evolved primarily to communicate episodes” [p. 557]. If right, the Chomskyan objection will finally have a strong retort.
A few days ago Ken Arneson posted a comment on this blog that asserted, “Language is a mechanism for deliberately transferring declarative memories from one person to another.” His point is related to the Corballis proposition that language communicates episodes and I want to acknowledge him here, but Corballis takes the trouble of developing his case. (His article, “Mental time travel and the shaping of language,” can be found here.)
The focus of Corballis’s paper is “episodic memory” which he defines as “memory for events,”  but then he immediately goes on to make the points that have always made me dislike the term. The most critical objection is that there is no distinction between the works of the proposed episodic memory and imagining other kinds of episodes.
recent brain-imaging studies show that remembering past events and imagining future ones activate the same core network in the brain, and people with amnesia following hippocampal damage have as much trouble imagining new experiences as remembering past ones. 
Another point is that episodic memories are extremely unreliable:
people probably remember only a tiny fraction of actual past episodes, and events are often remembered inaccurately, even to the point that people will claim with some certainty to remember events that did not in fact happen. 
No wonder then that Corballis replaces the term “episodic memory” with another name. Unfortunately the term he uses, “mental time travel,” is much too metaphorical to carry much authority and is limiting as well because instead of thinking about the future we can think about alternate pasts or events in fairy-tale kingdoms or epics long long ago in galaxies far far away. So instead of adopting Corballis’s term, I am going to claim my blogger’s prerogative and substitute “episodic thinking.” (If he had titled his paper ‘Episodic thinking and the shaping of language,’ the work’s gravitas would have been far less surprising.)
When we describe episodes from our past, we are not pulling data from stored picture albums and sound recordings. Episodic thinking
… is an act of construction. This is clearly true of imagining the future, but we also construct our pasts, often embellishing the true facts. As [Ulrich] Neisser put it, “Remembering is not like playing a tape or looking at a picture; it is more like telling a story.” 
Since I once began a book with the sentence, “Remembering is an act of imagination,” you can be sure that I will like what Corballis and Neisser are saying. But Corballis has pointed us directly into the target. The really distinctive form of human thought is not symbolic thought but episodic thinking.
Corballis devotes over a page to the question of whether episodic memory is uniquely human and shows plenty of evidence that other birds and mammals may remember enough details of an action to make it an episode. (An episode is a web of associations, a set of facts, joined in time. It is distinguished from a generic memory in which a single item is associated with an emotion, perception, or action.) I would be quite surprised if someone was able to establish that chimpanzees and gorillas never recall episodes from their own experience and use them as guides to behavior. I’m talking, and Corballis is talking, about something more. Episodic thinking provides
the material from which to construct detailed scenarios … [it] provides an extra degree of detail that enables us to plan with greater precision and specificity than is possible through more general adaptive mechanisms. 
Again, Corballis is so focused on the role of the future that he ignores other possible benefits of episodic thinking, but we get the idea. Episodic thinking provides a level of detail, a sense of the individual here and now, that symbols, which are generalized by nature, cannot express. It enables us to give specific directions, provide realistic entertainment, produce myths, etc. Of course these benefits can only really become important when episodic thinking is joined with language. Corballis devotes the last part of his paper to discussing the co-evolution of speech and episodic thinking. Language, he says,
provides a means by which people can create the equivalent of episodic memories in others, and therefore contribute to their episodic thinking. [557; note that when pressed even Corballis uses the phrase ‘episodic thinking.’]
This idea points toward a way of resolving the question of what baby talk contributes to the development of complex sentences. Single words pilot attention to the here and now. Complex sentences enable us to imagine an absent or alternate reality in detail. You cannot pilot attention in complex ways until you can reliably pilot it in simple ways.
Chomsky is not wrong, just too abstract. He’s like the man who gives directions to the gas station by saying, “Follow along course X until you come to object Y and then shift to course Z.” It’s true, but beside the point. A more episodic account is required.