Concepts and categorization in language acquisition

From How Emotions Are Made
Jump to: navigation, search

Chapter 5 endnote 21, from How Emotions are Made: The Secret Life of the Brain by Lisa Feldman Barrett.
Some context is:

Syllables that co-occur relatively rarely are more likely to be part of different words. Babies learn these regularities extremely quickly, even within a few minutes of exposure. [...] Whether the patterning of certain sound concepts (phonemes) is learned from experience or triggered by experience (i.e., is innate) is a matter of great debate.

Language acquisition is an example of the great nativism/empiricism debate.[1][2] In language, the debate began with Chomsky[3] (language is innate) vs. Skinner[4] (language is constructed).

In some ways, how children learn language may be a great analogy for how they learn to make emotions, since the brain is faced with a similar set of problems in each. In language learning:

  1. The brain receives a continuous stream of information and has to break it up into phonemes, even though there are usually no clear, audible breaks between the phonemes (and a speech sound can blend phonemes).
  2. The information is often ambiguous in its meaning.
  3. The information is variable and based on context... a "D" sounds different in Dad and Death (acoustical signal is different).  And yet a human perceiver hears a similar "D" in each instance.
  4. The information is degraded (many of the sound waves a human receives cannot be understood out of context).
  5. Some parts of the acoustic signal is not meaningful (i.e., is noise).
  6. Vocal cords do not start and stop vibrating with different phonemes; there is gradation from one phoneme (category) to another.

Yet even with this graded acoustical function, statistical analyses will produce categories.  Just because classification analyses can categorize continuous signal into categories does not imply that separate and discrete categories exist in the signal.  Categories need not live in the physical aspects of the signal if they exist in the head of the perceiver.

The result is very fast categorizations that often involve disambiguating the information that is present or "filling in" missing information.


Notes on the Notes

  1. Also see https://en.wikipedia.org/wiki/Statistical_learning_in_language_acquisition 
  2. And, https://en.wikipedia.org/wiki/Language_acquisition_device
  3. Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
  4. Skinner, Burrhus Frederic. (1957) 2014. Verbal Behavior. BF Skinner Foundation. See also Wikipedia.