Language acquisition II:

1
Language acquisition II

• Putting words together

2
The two-word stage

• At about 18 to 20 months, children's first
  multi-word utterances appear 2 words first

• I shut.
• See baby.
• More hot.
• Boot off.
• Airplane allgone.
• Papa away.
• All wet.
• No bed.
• Hi Sam.

• All dry.
• I sit.
• No pee.
• More cereal.
• Other pocket.
• Mail come.
• Our car.
• All messy.
• Bye-bye car.

3
Early Semantics

• As with single words, there is much semantic
  consistency in the two-word stage
• Early sentences in cultures around the world tend
  to focus on a few themes the appearance,
  disappearance, and movement of objects comments
  about object properties requests for and
  rejections of objects or activities and
  who/what/where questions

4
Early Syntax Production

• Syntax is already almost perfect! (within the
  limits of two words)
• 95 of sentences produced in this two-word stage
  have the proper syntactical order
• Two-words phrases also have all the components of
  complex sentences, but not at the same time
• If a child is looking at his mummy fixing a toy
  on the table which has just been given to her by
  daddy you might get subject-verb (Mommy fix),
  subject-object (Mommy toy), subject-location
  (Mummy table), verb-location (put table), and so
  on.
• There seems to be no syntactic component that
  infants can't use- the problem seems to be just
  that they can't string more than one relation
  together.

5
Early Syntax Comprehension

• As with single word production, there is a
  dissociation between comprehension and production
• At the two-word stage the infant already can
  understand very complicated syntax.
• How do we know?
• Anecdotal evidence Just hang out with a kid this
  age!
• Experimental evidence Show two TV screens with
  different images, and play a sentence describing
  just one ('Big Bird is tickling Ernie' vs. 'Ernie
  is tickling Big Bird').
• The child attends to the screen being described.

6
From 2 to 3 words

• After a two word stage, you might expect a
  three-word stage- but you'd be wrong.
• When a child passes the two-word stag (usually
  between the age of 2 and 3.5 years) things get so
  complicated so fast that no one has yet found a
  clear sequential pattern for what is happening
• It is as if language is suddenly on-line and it
  just bursts forth
• Moreover, some children zoom through the two word
  stage in just a few months

7
On the lack of a 3-word stage

• There are well-studied cases, by no means
  extraordinary, of children apparently having
  mastered syntax totally by the age of 2
• For example, Roger Brown reported on a child who
  produced these sentences before her second
  birthday
• I got peanut butter on the paddle.
• I sit in my high chair yesterday.
• Fraser, the doll's not in your briefcase.
• Fix it with the scissor.
• Sue making more coffee for Fraser.
• Clearly, she is not perfect there are a few
  errors in the sentencesbut no one has ever found
  a single grammatical rule that children at this
  stage usually get wrong
• At the highest estimate, children make errors
  not more than 8 of the time when an error is
  makeable, and some estimates range as low as
  0.1.

8
A case study Auxiliary verbs

• To get a feel for how amazing this is, lets
  consider just one example the use of auxiliary
  words- verbs that go with other verbs,
• For example can, should, must, be, and
  have in sentences like He should have eaten,
  He can be brushing his own teeth now or He
  must have left his mittens in the car.

9
A case study Auxiliary verbs

• It has been estimated that there are 24 billion
  billion logically-possible orderings of auxiliary
  words that can legally appear in the same
  sentence
• Of those, only about 100 are actually grammatical
  in English we can't say He have should eaten
  or He be can brushing his teeth now or He have
  must left it
• So chance performance is 0

10
A case study Auxiliary verbs

• Moreover, the problem is made even more difficult
  because some misorderings seem like they might be
  quite plausible by analogy
• He seems happy ? Does he seem happy?
• He is happy ? Does he be happy?
• He did eat ? He didn't eat
• He did a few things ? He didnt a few
  things.
• I like going ? He likes going
• I can go ? He cans go
• I am going ? He bes ams going.

11
A case study Auxiliary verbs

• So, the base rate odds of doing it right are
  minuscule, and the base rate odds of making
  errors are huge, plus there are many temptations
  by analogy to do it wrong
• One woman studied 66,000 productions of auxiliary
  sentences by young infants, in order to examine
  the pattern of their errors
• It turned out to be very easy to do- there were
  virtually no errors in those sentences!

12
What do infants do wrong?

• The errors that infants do make are limited
  largely to over-generalizations of very common
  rules
• They sometimes regularize irregular plurals I
  have two mouses or My sister is missing two
  tooths
• They sometimes regularize irregular verb forms
  I runned back to Mummy or I finded Daddy.
• The only way to get these right is to memorize
  them, since they are irregular
• These errors suggest that children have access to
  the abstract rules- the child who says ''runned"
  or "mouses" cannot possibly be repeating it,
  since s/he will never have heard it it must be
  deduced from an implicit understanding of the
  rule.

13
Adults do it too

• Adults also regularize if the irregular word is
  infrequent enough words like trod (not
  treaded), strove (not strived), dwelt
  (not dwelled), and smote (not smited) are
  often regularized
• Historical linguists have shown that this
  certainly does happen, because Old and Middle
  English have about twice as many irregular verbs
  (360) as we do now
• The rest have become regularized as the error has
  become standard.
• As a general rule, errors are more likely on less
  frequent aspects of language, and they are likely
  to be regularization errors

14
The problem with passivity

• Another example of where low frequency leads
  English children astray is in comprehending
  passive sentences, in which the S is not at the
  beginning of the sentence as it is in most
  sentences
• E.g. The cat was chased by the dog is a legal
  OVS order, rather than the usual SVO
• Children have trouble with passive sentences, and
  often mistake them for SVO sentences, so they
  will match a cat chasing a the dog to the
  sentence The cat was chased by the dog.

15
Some other interesting errors

• We can get some idea about how children parse
  language at this stage from their other errors
• "I am heyv" (reponse to 'Behave!")
• Daddy, when you go tinkle you're an eight, and
  when I go tinkle I'm an eight, right? from
  urinate
• In both cases, the child makes a plausible
  miscalculation about where word boundaries are.

16
Creoles and pidgins

• We can get further insight into what is happening
  inside children's heads from a very interesting
  linguistic phenomenon creolization
• A pidgin arises when two peoples who don't speak
  the same language are suddenly and without
  advance warning forced to live and work together
  their ad hoc communication system is a pidgin
• Creolization is the process by which a pidgin
  language is regularized
• Creolization offers us unique insight into
  linguistic construction in real time.

17
What pidgin?

• Pidgins arise depressingly often in this
  sorrowful world
• The best-known cases sprung from the slave trade
  in America, and from indentured servitude in the
  South Pacific
• They have also arisen in other less horrific
  situations there are pidgins between people who
  occasionally get a chance to barter but have no
  formal trade, such as Russian and Scandinavian
  sailors before the fall of the Soviet Union

18
What pidgin?

• Pidgins tend to draw heavily from a third
  language, usually the language of the boss
• Pidgin utterances are short almost never more
  than 4 words at a time never more than 1 idea at
  a time (very much like chidren's two-word stage)
• Pidgin has no grammar at all word order is
  completely free
• There are no tense markers (or any other
  morphological affixes), no articles, no
  prepositions
• Many pidgin utterances don't even contain a verb
  i.e. Big expensive flour Russia this year

19
Pidgin Examples (Papua)

• (Incidentally, most examples of pidgins on the
  WWW are really creoles.)
• Sapos yu kaikai planti pinat, bai yu kamap
  strong olsem phantom.
• If you eat plenty of peanuts, you will come up
  strong like the phantom.
• Yu pren tru bilong mi. Inap yu ken helpim mi
  nau?
• You are a true friend of mine. Are you able to
  help me now?
• Em i go we?"
• Where did he go?'

20
Pidgin Examples (Rus-sonorsk)

• R What say? Me no understand.
• N Expensive, Russiangoodbye.
• R Nothing. Four half.
• N Give four, nothing good.
• R No brother. How me sell cheap? Big expensive
  flour on Russia this year.
• N You no say true.
• R Yes. Big true, me no lie, expensive flour.
• N If you buyplease four pud. If you no
  buythen goodbye.
• R No, nothing brother, please throw on deck.

21
So care yu pidgin?

• Our interest in pidgins in this context is due to
  the fact that there have been a few cases in
  which pre-linguistic children have been raised in
  situations in which pidgin was the only language
• This happens when they are raised primarily by a
  caretaker who speaks pidgin
• Something remarkable happens in that situation
  the language-learning infants regularize the
  pidgin, add some features, and turn it into creole

22
What creole?

• Unlike pidgins, Creoles include rule-based tense
  markers, prepositions and a hard-coded word order
  they are bona fide languages
• Creole rules may bear no strong resemblance to
  the other ancestral languages in the mix, nor to
  the dominant language, nor to the native language
• BUT they do bear a striking resemblance to other
  creoles which have arisen in other situations on
  other places on the planet!

23
Creole example (Bislama )

• Bislama is a mixture of English, French, and
  Melanesian words used on the pacific island of
  Vanuatu
• Tufala i stap yet long Betlehem, nao i kam kasem
  stret taem blong Meri i bonem pikinini. Nao hem i
  bonem fasbon pikinin blong hem we hem i boe. Hem
  i kavremapgud long kaliko, nao i putum hem i slip
  long wan bokis we oltaim ol man oli stap putum
  gras long hem, blong ol anamol oli kakae. Tufala
  i mekem olsem, from we long hotel, i no gat ples
  blong tufala i stap. - Luk 26-7.
• "The two of them were in Bethlehem, now it came
  the exact time for Mary she births child. Now him
  he born firstborn of her that him he boy. She she
  coverup (him) good in cloth, now she put him he
  lay in one box where always all men they are
  putting grass in him, for all animals they eat
  (it). The two of them they made same, because at
  hotel, it no got place for the two of them to
  stay." - Luke 26-7

24
Case study Hawaii

• Creolization happened quite recently (just at
  the end of the last century) in Hawaii, when
  plantations boomed and uneducated plantations
  workers were brought in from around the world
• Some of their children- who were raised with
  pidgin as their first language- are still alive,
  and have been studied by linguist Derek Bickerton
• Hawaiian Creole includes rule-based tense
  markers, prepositions and a hard-coded word order

25
(No Transcript)
26
The claim

• Bickerton's claim- not without critics, but with
  good evidence- is that the children exposed to
  the pidgin were the source of its regularization
• The suggestion is that there are default
  settings on language which children are innately
  drawn to use (UG?), even if they have to invent
  the language in doing so, in the absence of a
  compelling reason to use other settings
• The pre-linguistic infant is thus seen as a
  bottleneck which places strong constraints on
  what is possible in language, and which may be
  the source of the regularities we see across
  languages