Pinker's Anti-Whorfian Hypothesis

Stephen Pinker’s new book, The Stuff of Thought: Language as a Window on Human Nature will be published this Tuesday, 9/11. The book’s virtues are what any reader of Pinker expects—wit, clarity, and more clarity. For many language mavens the big story is likely to be that Pinker, a one-time member in good standing of the Chomsky school, has broken with its number one doctrine; Pinker places semantics above syntax. I sympathize with that bias, yet the book also has a more regrettable feature that I have also come to expect from Pinker. It tells a story that is hard to credit in the face of how evolution works. And, oh, yes, if his theory of language is correct, this blog’s long focus on the roles of attention, joint attention, and listening have been a waste of time.

Pinker calls his subject conceptual semantics and his thesis holds “that word meanings are represented in mind as assemblies of bacic concepts in a language of thought.” [p. 91] To decider that sentence, read it backwards:

language of thought: underlying the language we speak/write/sign is a more primal system that Pinker calls, perhaps metaphorically, a language of thought.

basic concepts: this primal “language” is composed of a set of fundamental abstractions, such as “red” or “square” or “move.” These atoms of thought are concepts, not concrete discoveries of the senses. Pinker quotes Leibnitz approvingly, “There is nothing in the intellect that was not first in the senses… except the intellect itself.” [p. 93] These basic concepts are innate.

assemblies: the atoms of thought can combine into molecules, according to fixed rules, that define complex relationships. The assemblies are not innate; only the basic concepts are born into us. The rules for combining concepts, however, do seem to be innate.

word meanings are represented: Normally, we talk about language representing things in the world. The verb to swear in represents the act of taking or giving an oath of office. But Pinker’s representation is not of something in the world. When we get psychological about language, we often say the world stands for a mental image or perception of something, say Lyndon Johnson being sworn in on Air Force One or a witness being sworn in before testifying. But Pinker’s words neither point to the world not its sensual traces. They refer to the concepts that are built into or assembled by the brain.

That’s Pinker’s theory: the words we use are determined by the concepts that are built into the brain; the rules for assembling concepts determine the limits and possibilities of speech. The core of the argument is that the flow goes from thoughts to language, not the other way around.

Pinker complains that the Whorfian hypothesis—language controls thought—so dominates pop psychology that whenever he told people he was working on a book about thought and language they assumed he was writing about how language controls thought. Even the PR sheet sent out with the book’s review copies makes the same mistake, boasting that the book describes many “issues that show how language impacts how we think and feel.”

No, no. The book shows how our thoughts and feelings impact language. And I am 100% with Pinker on the direction of flow—thought to speech—although I don’t believe Pinker has studied his Proust enough to know how easily habit puts an end to thought. (Pinker does quote Proust, but it’s the part everybody quotes.)

As for its importance to this blog, Pinker’s flow implies that concepts evolved before words. We cannot have evolved any words first because, without their underlying concepts they would be meaningless (“word meanings are represented … in … concepts”). One idea this blog has toyed with is that the first words might have begun as emotion-building babble. Making a sweet sound like mama could strengthen bonds with a caregiver. Later, the sound might have become associated with an individual and been transformed into a word, but in Pinker’s view this scenario is only possible if the concept for that mama person already existed.

Well, I’m not married to the mama story and I like to think I’d give it up, even as a plaything, in the face of good evidence or sharp reasoning against it. But in Pinker’s case we have a problem, for his theory demands that the first words, however they arose, followed the existence of their concepts. Assuming the innate concepts are part of our general primate inheritance, an account of speech origins based on Pinker’s theory of meaning has to account for the evolution of the ability to transform these abstract concepts into concrete words.

Fortunately, Occam’s Razor can quickly shave away this ugly stubble.

Continued in tomorrow's post. (Note: in order to move quickly through Pinker's book, posts will be every Monday and Tuesday.)

Comments

[MyaR]

I just found your blog via Cognitive Daily, and I must say, I'm very grateful that you're reading Pinker's new book so I don't have to -- I invariably end up throwing them across the room about halfway through. I think I'll go find one of your books instead. Cheers!

[Malatesta]

Just off the cuff, doesn't some of this sound like Fodor's "all concepts are innate" hypothesis and the "conceptual semantics" rings of Jackendoff's work?

Either way, it sounds interesting though I usually tend to disagree with Pinker.
------------------------
BLOGGER: Pinker specifically separates his idea from both Fodor and Jackendoff. Fodor's all concepts are innate makes no evolutionary sense and Pinker calls it too extreme. He also discusses and rejects Jackendoff.