Language

1
Language

• Phonology
• categorical perception
• Morphology
• past tense learning
• Syntax
• Tacit vs. explicit knowledge
• Language acquisition

2
Areas of language study

• Phonology - the system of sounds in a language
• Syntax - grammatical arrangement of words in
  sentences
• Semantics - meanings of sentences and words
• Pragmatics - social rules of language use that
  contribute to sentence meaning

3
Figure 10.01
4
Phonology

• Speech is made up of small units called phonemes
• All words in English can be made from
  approximately 40 phonemes
• Some languages have many more phonemes
• The phonetic alphabet has symbols for phonemes
  for all languages

5
Speech production

• Speech is produced by interrupting airflow as we
  breathe out
• place of articulation tongue, lips, and teeth
  move to various positions to restrict the airflow
• manner of production air flow may be completely
  or partially stopped, will flow through different
  cavities (m sound flows through the nose)
• voicing is produced by vibrating the vocal folds
  (as with the letter z)

6
Figure 10.02
7
Speech production (cont.)

• All phonemes can be made up of combinations of
  these features
• However, there are no markers in the speech
  stream where one phoneme ends and the other
  begins
• speech segmentation - we need to slice the speech
  stream into recognized phonemes
• coarticulation - phonemes overlap in production

8
Figure 10.03
9
Categorical speech perception

• We tend to hear speech sounds only as members of
  a category (e.g., ba or pa, not something in
  between)
• The sounds for b and p differ on only one
  feature, voicing (b is voiced, p is not). This
  difference is reflected in the voice onset time
  (VOT) gap between when the air starts moving and
  voicing begins.

10
Speech categories

• VOT boundary for b vs. p is around 30
  milliseconds
• People generally do not hear within-category
  differences
• 0 msec and 20 msec VOT perceived as b
• 60 and 80 msec VOT perceived as p
• People are good at detecting across-category
  differences
• 20 and 40 msec sound different (b vs. p)

11
Figure 10.04
12
Categorical speech perception

• Infants (0 to 6 months) show categorical
  perception for all possible language contrasts
• the newborn is a universal learner
• By age 1 year, we lose the ability to perceive
  contrasts not used in our native language
• makes 2nd language learning difficult!

13
Morphology

• Morphemes are the smallest units of language that
  carry meaning
• When children learn verbs, they learn the past
  tense for each verb (go-went talk-talked)
• Later, they learn the rule that adding the
  morpheme -ed produces the past tense

14
Past tense learning

• Children go through a stage (ages 2-3) where they
  make over-regularization errors (go - goed, hold
  - holded)
• With feedback, they learn which verbs are special
  cases. Eventually, they apply the rule only where
  appropriate, as adults do.

15
Syntax

• People can easily distinguish between grammatical
  and ungrammatical sentences. This suggests that
  speakers know the rules that are used for putting
  words together to make sentences
• These rules are not explicitly taught
• Language learning is not merely imitation --
  children produce many predictable errors that
  they do not hear in adult speech

16
Figure 10.06
17
Tacit knowledge

• Tacit knowledge is knowledge that we possess and
  use, but cannot articulate
• Humans are said to have tacit knowledge of the
  rules of syntax -- we know how to put words
  together to form grammatical sentences

18
Tacit vs. explicit knowledge (1)

• Grammar
• Folk physics (how to calculate a moving objects
  trajectory)

• Prescriptive language rules (dont say aint)
• Equations for object motion

19
Tacit vs. explicit knowledge (2)

• We may sometimes convert tacit knowledge to
  explicit knowledge. An example would be studying
  linguistics and learning the rules that we apply
  in understanding sentences.
• Explicit knowledge may also be converted to tacit
  knowledge, e.g., as we become skilled drivers, we
  stop using explicit rules

20
What is a grammar?

• A person who has acquired a language has
  internalized the grammar of that language. This
  knowledge of language is an example of tacit
  knowledge.
• This knowledge is different from a persons
  ability to speak the language, as discussed by
  Chomsky
• competence vs. performance

21
Language acquisition research

• Studies of child language have examined how
  different aspects of language develop
• lexical development (word learning)
• speech development (changes in ability to
  discriminate sounds increase in length of
  utterances)
• syntactic development (understanding and use of
  forms such as the passive)
• social use of language

22
Correlational vs. experimental studies

• A correlational study measures naturally
  occurring relationships, for example
• between language skill and amount of reading in
  the childs environment
• between amount of motherese and childs language
  skill
• An experiment would manipulate treatments of
  different groups to test a specific hypothesis,
  for example
• exposure to complex sentences leads to
  acquisition of syntax

23
Correlational vs. experimental studies

• Advantages of experiments
• increased power
• can infer that the treatment caused the change in
  the outcome (dependent) variable
• Advantages of correlational studies
• studies naturally occurring relationships
• can study variables for which manipulation is not
  possible (we cant control the amount and type of
  language input that a child gets)

24
Environmental Factors in language acquisition

• Children must have language exposure to develop a
  full language. It is thought that exposure
  serves to trigger the acquisition of various
  aspects of language
• parameter setting
• Motherese (special way of talking to children
  that exaggerates intonation and uses clear, short
  sentences) may assist children in developing
  language

25
Innate mechanisms (1)

• Elements of grammar may be innate, and require
  only minimal linguistic input for a full language
  to be acquired.
• With language fragments as input, children of
  immigrant workers have developed creole languages
  that are much more sophisticated than the input.
  Bickerton attributes the development of creoles
  to an innate language bioprogram.

26
Innate mechanisms (2)

• Congenitally deaf children acquire a gestural
  language called homesign, which shows
  similarities to language acquisition of children
  with normal hearing. The two- and three-sign
  utterances express semantic relations that are
  similar to those used by children in two- and
  three-word stages.
• This evidence supports the language bioprogram
  hypothesis.

27
Innate mechanisms (3)

• Parameter setting only minimal input is needed
  to set these parameters.
• Null-subject parameter some languages allow the
  subject to be dropped English does not
• This view predicts that once a parameter has been
  set, the child will consistently use the correct
  forms
• distinction between competence and performance
  errors might be due to limits on articulation or
  working memory, not lack of knowledge

28
Evidence of innate mechanisms for language
acquisition

• Children acquire language rapidly and with little
  effort
• Children go through a progression that is
  remarkably similar across children and across
  languages
• Poverty of the stimulus argument (associated with
  Chomsky). Language input is
• finite
• degenerate (interruptions, speech errors, pauses)
• does not provide negative evidence

29
Modularity in language processing

• Although the human language system seems to
  handle sounds, structure, and meaning in an
  integrated way, this system has modules (separate
  parts)
• One line of evidence for modularity is that brain
  damage can cause deficits in specific aspects of
  language processing, e.g., Wernickes aphasia
  loss of specific semantic classes of words

30
Aphasias

• Brocas aphasia disrupted language production,
  usually nouns strung together, missing function
  words such as prepositions
• Wernickes aphasia production is fluent but
  often gibberish. Comprehension is severely
  disrupted.

31
Summary

• Language is generative
• morphemes can be added to create new words
• we can create an infinite number of sentences
  from a limited vocabulary
• The language system is modular - its different
  parts function somewhat independently

32
Sections not on exam

• Sense and reference
• Phrase structure
• Movement rules
• Trace theory