Week 2 - first steps into language

Question 1

What order to children produce their first sounds to their first words?

Answer 1

1Reflexive
2Cooing laughing
3Vocal play – start to produce sounds
4Babbling – phonemic sounds in their own language
5Jargon – some form of word used systematically but not a ‘real’ word in their native language
6First words

Question 2

Which consonant sounds are produced in early speech or late babbling?

What is the relevance of these to early speech?

Answer 2

• Stops [p, b, t, d, k, g]
• Nasals: [m, n]
• Glides: [w, j]

The early belief was that there was no relation between sounds of babble (Jakobson 1968’s tongue delirium)

However, we now know:

Consonants favoured in babbling are more likely to be produced correctly in first words (Vihman et al., 1985)

Frequencies of sounds used in late babbling and in first words are very similar (Boysoson-Bardies and Vihman, 1951)

Question 3

To what extent does babies’ babbling reflect ambient language?

What is the babbling drift?

Answer 3

To some extent early sounds reflect ambient language e.g., French babies babble contains more nasals than English babies (Polka et al., 2007)
Sounds used in babbling increasingly reflect phonemes used in the ambient language at first-word stages is approached (babbling drift)
Babbling starts to become more language-specific as babies get closer to producing their first words (babbling drift)

Question 4

Pre linguistic vocalisations: babbling >6 months

Answer 4

Reduplicated (canonical) babbling strings of identical syllables: [bababa]
Variegated (non-canonical) babbling syllable strings with varying consonants and vowels [badagubu], [bamido] – adult like prosody – more language specific than canonical babbling
The child’s intonation starts to follow patterns of native like prosody

Question 5

Babbling > 6 months

Answer 5

Distinguished from vocal play by presence of true syllables – sequences of CV syllables with adult like rhythm e.g., ‘dada’
CV (consonant vowel syllable) is the simplest type of syllable
Variegated babbling usually follows a period of reduplicated babbling e.g., ‘babeda’

Question 6

4-6 months

Answer 6

More experimental use of vocal apparatus testing or playing
Involving extremes of sounds (high, low, loud, quiet e.g., yells whispers squeals growls
Around 6 months there is an increase in phoneme repertoire – no longer just velars but also /m/ /n/ /d/ /p/ /b/

Question 7

Pre linguistic cooing 6-8 weeks

Answer 7

Sounds made up of velars /k/ /g/ and back vowels ‘oohs’ and ‘aahs’
Vowels predominate especially in early stages
Indicated happiness contentment and comfort
Progressive becoming more varied and longer

Question 8

What happens from birth to 2 months in terms of early language?

Answer 8

• Crying coughing burping sneezing
• Can include vibration of vocal folds (involuntary)
• Restricted in range by immature vocal anatomy of infant

Question 9

Which consonant sounds are produced in early speech or late babbling?

What are the relevance of these to early speech?

Answer 9

• Stops [p, b, t, d, k, g]
• Nasals: [m, n]
• Glides: [w, j]

The early belief was that there was no relation between sounds of babble (Jakobson 1968’s tongue delirium)

However, we now know:

Consonants favoured in babbling are more likely to be produced correctly in first words (Vihman et al., 1985)

Frequencies of sounds used in late babbling and in first words are very similar (Boysoson-Bardies and Vihman, 1951)

Question 10

What are the challenges for babies learning new words?

What methods do researchers use to investigate this?

Example: Jurczyk and Aslin (1995)

Answer 10

Finding the word
Matching the form to the meaning
Segmenting the speech steam – children need to isolate individual words

Preferential listening tasks. Example - Juscyzk and Aslin (1995)
Study 1
1. light is illuminated on one side of the child
2. when the child looks at the light, she hears a sound or sequence of sounds particular to that light
3. the child learns that her gaze controls what she hears
4. on one side familiarised words are presented
5. on the other non-familiarised words

Study 2
1. infants familiarised with lists of isolated words
2. infants heard sentences on each side
3. one one side sentences containing familiarised words
on the other side sentences containing non-familiarised words

Response times were measured in response to each steam in both tasks

DV – listening time e.g., how much they listened to each stream?

Results:

6-month-olds no difference
7.5-month-olds listen more intently to words they have heard before
Both studies showed that infants become able to segment words from fluent speech between 6 and 7.5 months
At this age, infants listened longer to the target words/sentences containing target words

Question 11

What are the two types of phonological processes?

Examples of both

Answer 11

1A structural process – whole word processes (affect the whole word because affect phonotactic structure)

Examples of structural phonological processes
• Weak syllable deletion e.g., nana for ‘banana’
• Final consonant deletion: nai for ‘night’
• Reduplication: wawa for ‘water’
• Consonant cluster reduction: pider for ‘spider’

2
Systemic processes – segment substitution processes (they do not affect the whole word by only a specific segment)

Examples of systemic phonological processes
• Velar fronting – ‘bet’ – get
• Stopping – tink – think
• Gliding - wun – run
• Assimilation tat – cat (consonant harmony, Vihman, 1988)

Question 12

Why are children’s initial word productions different from adults’?

Answer 12

without a full phoneme repertoire, children need other strategies to produce words

early on children produce the same word in different ways (instability of productions) and tend to use words reflecting their phonological competence to date

by 18th months – consistent processes which simplify syllable structure

Question 13

What are the phonotactics of early words?

Answer 13

Early words have a simple structure (Ingram, 1999)
Single syllable CV or reduplicated CVCV e.g., /dada/
Can get other more complex examples e.g., ‘Gladys’. ‘moon’ reported in Bloom (1973)
More complex forms such as CVCC and initial clusters later (after 24 months)
Phonologically really complex structures such as CCCVC not until 3 years
Phoneme repertoire remains restricted at this stage (those great variability
e.g., /m/ /n/ /b/ invariably present but other sounds are typically later acquired e.g., /th/ /r/, /l/

Question 14

What is the functional load hypothesis (Ingram, 1989)

Answer 14

• The degree to which a contrast is used in a language
• /v/ has a higher functional load in Estonian
• It is more used to mark differences, in that more binary pairs exist with /v/ in Estonian

Question 15

What is the order acquisition of sounds?

What are prelinguistic vocalizations?

IS the order of acquisition stable across languages?

Answer 15

• Nasals, glides, stops (bilabial and alveolar)
• Velar consonants
• Voiced fricatives and affricates

Pre linguistic vocalisations jargon stage > 10 months
•Strings of sounds and syllables uttered with a rich variety of stress/timing and intonation patterns. Sounds like adult intonation
•Melody of language without the words (Hoff, 2017)

Order of acquisition is broadly similar across languages
But exceptions e.g., /v/ relatively late in English but much earlier in Swedish, Estonian and Bulgarian

Question 16

Early changes to the lower vocal tract (lungs and intercostal muscles)?

Answer 16

Initially, ribs are perpendicular to the spine
From 3 months ribs starts to angle down and out
Better control of the muscles
Better control of air pressure
Facilitates nontidal breathing (controlled periods of sustained or controlled phonation necessary for speech)

Question 17

What are the early physiological changes to the vocal tract?

Answer 17

Larynx begins to lower from 4-6 months (continues until 3 years)
Vocal tract lengthens (average 8 cm in infants, 17 cm in adults)
The angle between the oral cavity and pharynx becomes more acute
All changes leave more space for tongue movement and alter the resonant qualities of the vocal tract
Permit a wider range of vowels

Question 18

How is the infant’s vocal tract different from the adult vocal tract?

Answer 18

High larynx
Different tongue shape
Relative positions of the epiglottis and vocal cords closer than in adults
The angle between the oral cavity and pharynx shallower
Shorter vocal tract (affects sounds newborns can make)

Question 19

How can the process of overextension be best described in young children’s production?

Child A produces the following string of words [fud] to only refer to the cereals she has every morning. Child B produces the same string of words to only refer to the cereals she and her father have every morning. What are these children using?

On what grounds do Saffran et al. argue that infants are good statisticians when segmenting speech? What do babies do?

Answer 19

The use of a single label for multiple concepts in the respective adult language

Both children have applied under extension to the word

They detected the presence of a new non-word in a sequence of non-words they had previously heard

Question 20

What is the noun bias? Evidence for this:
Is it universal?
Why do infants display a noun bias?

Answer 20

There is evidence of a noun bias

In English, an average child’s 50-word vocabulary contains: 40% nouns: 10% verbs, adjective, and function words (Bates et al., 1994)

When the child’s vocab exceeds 600 words: 40% nouns, 25% verbs and adjectives; about 15% functions words

Is the noun bias universal?

Also, a higher proportion of nouns are found in many other languages: Hebrew, French, Italian

But possibly not true for Chinese, Japanese, Korean where no noun bias or even a verb bias has been argued for (Tardif, 1996)

This is debated (Korean shows signs of a verb bias)

Why a noun bias?

Input: in adult speech to children, labels for different kinds of objects are more numerous than labels for actions, properties, or relations

Concepts for objects are easier to perceive than those for actions

Representing actions are more complex so the symbolic means for encoding these concepts (verb and adjectives) are recorded late

Question 21

What happens when children have acquired 50 words?

Answer 21

After 50 words children segment words more effectively and build more detailed phonological representations
vocabulary growth increases the need for more fine-grained representations
more phonemes and phonotactic structures
facilitates vocabulary growth

Question 22

The next challenge that children face in terms of learning new words is
• Matching the form to meaning

What are proto words?
How does babies’ word use differ from adults’?
What is over and under extension?

Answer 22

– sound sequences used as words by not obviously related to any real word. Proto words may appear before first ‘true’ words i.e., variegated babble or jargon phase or alongside first true words. May continue for some time.

Two ways in which they differ
• Over extension
• Under extension
The use of such processes reflects developing sematic representations and vocabulary and that these are not quite mature yet

Overextension
• From 1;6 until 2;6 children over-extend words in production i.e., apply words beyond the usual meaning
• Ball – first balls, then apples, grapes, or anything round
• Young children use the word more broadly than an adult would by extending it to refer to things with similar characteristics
• 33% of word uses are overextensions though only a few words account for these (Rescorla, 1980)
• Also find over-extension in comprehension, but this is rarer (although harder to test)

Under extension
• Much less common than over extensions (Anglin, 1983)
• Bottle to refer to only plastic baby bottles rather than any other bottles
• Number of over extensions decreases a vocabulary becomes larger

Question 23

Safran et al., 1996

Answer 23

Tried to teach 8-month-old children a mini artificial language
Example stimuli – four 3 syllable non-words used (tupiro, golabu, bidaku, padoti)

Words played sequentially without gaps to habituate children
After the habituation phase, the infants were exposed to two words from the original set plus two new words made from the same syllables but arranged in a different order

Indicates that they are able to segment the speech stream and that they are making assumptions which are guiding their expectations

Results:
• Infants listened more intensely to the new words
• Children were able to infer words from statistical features of input

Question 24

How do babies do this?

What can we conclude from this?

Answer 24

In stress-timed languages, adults use a Metrical Segmentation Strategy to segment speech into words e.g., stressed syllable often denotes the start of a new word.
Jusczyk et al., 1993: between 6- and 9-months infants develop a preference for dominant language stress pattern:

Strong/weak e.g., kingdom, butter

Weak/strong e.g., resent, guitar

9-month-olds preferred S/W to W/S relative to their native language – but 6-month-olds did not show a preference

Listened more to the familiarised words when the rhythm was S/W but not in the reverse

This suggests that children used the rhythmic patterns that are in their native language