U. of Rochester researchers found evidence that infants use complex statistical techniques to begin to discern words early in life -- evidence of importance of language acquisition to the learning process.
Researchers developed a bogus language of 3-syllable nonsense "words" and played a computer-generated voice reading 2-minute random combinations of these "words" to 24 eight-month-old infants. A test (described below) was applied, indicating that a baby recognizes something new by a longer attention-span. This test showed that the second time babies heard these sequences, they parsed-out the "words" (as Netscape or any Internet browser parses out commands in HTML).
The sound clusters the babies heard did not vary "in pitch, rhythm, or in any other way". So, seemingly, the babies' only learning strategy for parsing "word"-boundaries was a relatively sophisticated analysis of those sound-patterns they heard most frequently. "Infants have access to a powerful mechanism for the computation of statistical properties of the language output", researchers report in Dec. 13, 1996 issue of Science.
U. of California of San Diego researchers (in an accompanying article) hypothesized: NOT INBORN LANGUAGE ABILITY, BUT LEARNING ACTIVATES LANGUAGE ACQUISITION.
The above-mentioned test is described in The Language Instinct (1994), by Steven Pinker, a colleague of MIT linguist, Noam Chomsky, who has long hypothesized an "inborn language ability", labeled by Pinker "an instinct".
Pinker tells us that psychologists Peter Eimas and Peter Jusczyk devised a clever way to detect attentiveness of infants. (Eimas, P. D.; Siqueland, E. R.; Jusczyk, P; Vigorito, J; 1971, "Speech perception in infants", Science, 171, 303-306.) Eimas and Jusczyk put a switch inside a nipple, connecting the switch to a tape recorder emitting sounds: WHEN BABY SUCKS, TAPE PLAYS.
Now,
I'll CHARLOTTE the whole process. (The term "Charlotte" is my epinome, honoring my favorite teacher, Charlotte the Spider, who taught me about CONNECTIONS. So, "to charlotte a process" shall mean "to articulate the process in terms of its critical connections" -- especially those metathinkal, semiotic, strategic.)
- WHEN RECORDER REPEATS A GIVEN SOUND ABOVE AN OCCURRENCE- THRESHOLD, BABY SLOWS ITS SUCKING ON THE NIPPLE.
- WHEN RECORDER CHANGES FROM REPETITION TO A NEW SOUND, BABY SUCKS RAPIDLY.
Elsewhere I've defined the semiotic signs of INDICATOR and SIGNAL. Extending the definitions of Charles S. Peirce (founder of Semiotics: Theory of Signs, I define INDICATOR as A CORRELATED PAIR OF SIGNS: ONE HIGHLY OBSERVABLE, LOW IN INFORMATION CONTENT; OTHER, LOW IN OBSERVABILITY, HIGH IN INFORMATION CONTENT. (Example: lightning, distant thunderstorm. Again: pinked litmus paper, acidity in fluid.) Peirce defined a SIGNAL AS AN INDICATOR UNDER CONTROL. I articulate it as AN INDICATOR UNDER PHYSICAL AND LINGUISTIC CONTROL. (Example: the electricity of the lightning is put under the PHYSICAL CONTROL of an electric circuit, with a "telegraph key" as switch or circuit breaker; and this is under the LINGUISTIC CONTROL of The Morse Code. Hence, a telegraphic SIGNAL.) Dig?
Here, in the "sucking experiment", I articulate an INDICATOR consisting of observable sucking rate, correlated with what experimenters realize are repetitions or changes. (Can you articulate a better INDICATOR here? BMG: Be My Guest! I'm just asking for an INDICATOR in this charlotting.) This INDICATOR comes under the PHYSICAL CONTROL of the nipple-switch-recorder circuit; and under the LINGUISTIC CONTROL OF PARSING (by expert, not baby) OF THE INPUT INTO VARIOUS TYPES OF UNITS. I'll explain.
Suppose one of those "bogus words", each composed of "three nonsense syllables", is "novwazxij". This consists of three syllables or morphemes:"nov"; "waz"; "xij". Each syllable (morpheme) is composed of three phonemes (smallest units -- "gnomons" -- that can form a syllable/morpheme), resulting in 9 distinct phonemes for this particular "word": "n", "o", "v", "w", "a", "z", "x", "i", "j".
Apparently, babies noticed changes like that of "nov" becoming "nav" and vice-versa by sucking-response; a change like that of "xij" becoming "hij" and vice-versa; etc. -- hence, were distinguishing phonemes. And noted differences like that of "novwazxij" vs. "vonwazxij", etc., so were distinguishing syllables (morphemes. And the researchers tell us they recognized "word-boundaries", hence, words. Then, all of this -- phonemes, morphemes, words -- goes into the above LINGUISTIC CONTROL